Antioxidants For Male Sub Fertility (Review)

Antioxidants for male subfertility (Review)
Showell MG, Brown J, Yazdani A, Stankiewicz MT, Hart RJ
This is a reprint of a Cochrane review, prepared and maintained by The Cochrane Collaboration and published in The Cochrane Library 2011, Issue 6 http://www.thecochranelibrary.com
Antioxidants for male subfertility (Review) Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
TABLE OF CONTENTS HEADER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ABSTRACT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . PLAIN LANGUAGE SUMMARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . SUMMARY OF FINDINGS FOR THE MAIN COMPARISON . . . . . . . . . . . . . . . . . . . BACKGROUND . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . OBJECTIVES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . METHODS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . RESULTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 4. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 5. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 6. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 8. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 9. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 10. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 11. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 12. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 13. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 14. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 15. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 16. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 17. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 18. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 19. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 20. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 21. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 22. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 23. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DISCUSSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 24. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Figure 25. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . AUTHORS CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ACKNOWLEDGEMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . REFERENCES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CHARACTERISTICS OF STUDIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DATA AND ANALYSES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 1.1. Comparison 1 Antioxidant(s) versus control, Outcome 1 Live Birth per couple randomised. . . . . Analysis 1.2. Comparison 1 Antioxidant(s) versus control, Outcome 2 Pregnancy rate per couple randomised. . . . Analysis 1.3. Comparison 1 Antioxidant(s) versus control, Outcome 3 Adverse event: Miscarriage rate per couple randomised. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Analysis 1.4. Comparison 1 Antioxidant(s) versus control, Outcome 4 DNA fragmentation. . . . . . . . . . Analysis 1.5. Comparison 1 Antioxidant(s) versus control, Outcome 5 Sperm total motility at 3 months or less. . . Analysis 1.6. Comparison 1 Antioxidant(s) versus control, Outcome 6 Sperm total motility at 6 months. . . . . Analysis 1.7. Comparison 1 Antioxidant(s) versus control, Outcome 7 Sperm total motility at 9 months or more. . Analysis 1.8. Comparison 1 Antioxidant(s) versus control, Outcome 8 Sperm concentration at 3 months or less. . . Analysis 1.9. Comparison 1 Antioxidant(s) versus control, Outcome 9 Sperm concentration at 6 months. . . . . Analysis 1.10. Comparison 1 Antioxidant(s) versus control, Outcome 10 Sperm concentration at 9 months or more. Analysis 1.11. Comparison 1 Antioxidant(s) versus control, Outcome 11 Total sperm motility over time. . . . .
1 1 2 2 5 6 6 9 13 14 15 16 17 19 19 21 23 24 25 27 28 29 30 31 32 33 34 35 36 37 38 38 39 40 41 41 42 46 91 98 99 101 102 102 105 106 107 109 110 111

i
Analysis 1.12. Comparison 1 Antioxidant(s) versus control, Outcome 12 Sperm concentration over time. . . Analysis 1.13. Comparison 1 Antioxidant(s) versus control, Outcome 13 Side effects. . . . . . . . . . Analysis 1.18. Comparison 1 Antioxidant(s) versus control, Outcome 18 Pregnancy. . . . . . . . . . . Analysis 2.1. Comparison 2 Live birth versus pregnancy, Outcome 1 Live birth. . . . . . . . . . . . Analysis 3.1. Comparison 3 Head to head antioxidant(s), Outcome 1 Pregnancy per couple randomised. . . Analysis 3.2. Comparison 3 Head to head antioxidant(s), Outcome 2 Sperm total motility at 3 months or less. Analysis 3.3. Comparison 3 Head to head antioxidant(s), Outcome 3 Sperm total motility at 6 months. . . . Analysis 3.4. Comparison 3 Head to head antioxidant(s), Outcome 4 Sperm total motility at 9 months or more. Analysis 3.5. Comparison 3 Head to head antioxidant(s), Outcome 5 Sperm concentration at 3 months or less. Analysis 3.6. Comparison 3 Head to head antioxidant(s), Outcome 6 Sperm concentration at 6 months. . . . Analysis 3.7. Comparison 3 Head to head antioxidant(s), Outcome 7 Sperm concentration at 9 months or more. Analysis 3.8. Comparison 3 Head to head antioxidant(s), Outcome 8 Side effects. . . . . . . . . . . APPENDICES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . WHATS NEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . HISTORY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . CONTRIBUTIONS OF AUTHORS . . . . . . . . . . . . . . . . . . . . . . . . . . . DECLARATIONS OF INTEREST . . . . . . . . . . . . . . . . . . . . . . . . . . . . SOURCES OF SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . DIFFERENCES BETWEEN PROTOCOL AND REVIEW . . . . . . . . . . . . . . . . . . . INDEX TERMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . .
113 114 117 118 119 120 121 122 123 124 125 125 126 135 135 135 135 136 136 136
ii
[Intervention Review]
Antioxidants for male subfertility

Marian G Showell1 , Julie Brown1 , Anusch Yazdani2 , Marcin T Stankiewicz3 , Roger J Hart4 and Gynaecology, University of Auckland, Auckland, New Zealand. 2 Clinical Research and Development, Queensland Fertility Group, Woolloongabba, Australia. 3 Reproductive Medicine, Flinders Reproductive Medicine, Bedford Park, Australia. 4 School of Womens and Infants Health, The University of Western Australia, King Edward Memorial Hospital and Fertility Specialists of Western Australia, Subiaco, Australia Contact address: Marian G Showell, Obstetrics and Gynaecology, University of Auckland, Park Road Grafton, Auckland, New Zealand. m.showell@auckland.ac.nz. Editorial group: Cochrane Menstrual Disorders and Subfertility Group. Publication status and date: Edited (no change to conclusions), published in Issue 6, 2011. Review content assessed as up-to-date: 21 August 2010. Citation: Showell MG, Brown J, Yazdani A, Stankiewicz MT, Hart RJ. Antioxidants for male subfertility. Cochrane Database of Systematic Reviews 2011, Issue 1. Art. No.: CD007411. DOI: 10.1002/14651858.CD007411.pub2. Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd.
1 Obstetrics
ABSTRACT Background Between 30% to 80% of male subfertility cases are considered to be due to the damaging effects of oxidative stress on sperm. Oral supplementation with antioxidants may improve sperm quality by reducing oxidative stress. Objectives This Cochrane review aimed to evaluate the effect of oral supplementation with antioxidants for male partners of couples undergoing assisted reproduction techniques (ART). Search strategy We searched the Cochrane Menstrual Disorders and Subfertility Group Register, CENTRAL (The Cochrane Library), MEDLINE, EMBASE, CINAHL, PsycINFO and AMED databases (from their inception until Febuary 2010), trial registers, sources of unpublished literature, reference lists and we asked experts in the eld. Selection criteria We included randomised controlled trials comparing any type or dose of antioxidant supplement (single or combined) taken by the male partner of a couple seeking fertility assistance with placebo, no treatment or another antioxidant. The outcomes were live birth, pregnancy, miscarriage, stillbirth, sperm DNA damage, sperm motility, sperm concentration and adverse effects. Data collection and analysis Two review authors independently assessed studies for inclusion and trial quality, and extracted data. Main results We included 34 trials with 2876 couples in total. Live birth: three trials reported live birth. Men taking oral antioxidants had an associated statistically signicant increase in live birth rate (pooled odds ratio (OR) 4.85, 95% CI 1.92 to 12.24; P = 0.0008, I2 = 0%) when compared with the men taking the control. This result was based on 20 live births from a total of 214 couples in only three studies.
Antioxidants for male subfertility (Review) Copyright 2011 The Cochrane Collaboration. Published by John Wiley & Sons, Ltd. 1
Pregnancy rate: there were 96 pregnancies in 15 trials including 964 couples. Antioxidant use was associated with a statistically signicant increased pregnancy rate compared to control (pooled OR 4.18, 95% CI 2.65 to 6.59; P < 0.00001, I2 = 0%). Side effects: no studies reported evidence of harmful side effects of the antioxidant therapy used. Authors conclusions The evidence suggests that antioxidant supplementation in subfertile males may improve the outcomes of live birth and pregnancy rate for subfertile couples undergoing ART cycles. Further head to head comparisons are necessary to identify the superiority of one antioxidant over another.
PLAIN LANGUAGE SUMMARY Antioxidants for male subfertility Oxidative stress may cause sperm cell damage. This damage can be reduced by the bodys own natural antioxidant defences. Antioxidants can be part of our diet and taken as a supplement. It is believed that in many cases of unexplained subfertility, and also in instances where there may be a sperm-related problem, taking an oral antioxidant supplement may increase a couples chance of conceiving when undergoing fertility treatment. This review identied 34 randomised controlled trials involving 2876 couples. Pooled ndings support increases in live births and pregnancy rates with the use of antioxidants by the male partner. Further work is recommended to conrm these ndings.
S U M M A R Y O F F I N D I N G S F O R T H E M A I N C O M P A R I S O N [Explanation]
Antioxidant/s versus control for male subfertility Patient or population: patients with male subfertility Settings: Intervention: Antioxidant/s versus control Outcomes Illustrative comparative risks* (95% CI) Relative effect (95% CI) No of Participants (studies) Quality of the evidence (GRADE) Comments
Assumed risk Control
Corresponding risk Antioxidant/s control 90 per 1000 (38 to 200) versus
Live Birth per couple 20 per 1000 randomised Follow-up: 6 - 24 months Pregnancy rate per cou- 31 per 1000 ple randomised Follow-up: 3-24 months Adverse event: Miscar- 19 per 1000 riage rate per couple randomised Follow-up: 6-18 months Total sperm motility over time - Total sperm motility at 3 months or less Follow-up: 4-24 weeks
OR 4.85 (1.92 to 12.24)
214 (3 studies)
low1,2
118 per 1000 (78 to 174)
OR 4.18 (2.65 to 6.59)
964 (15 studies)
moderate3
29 per 1000 (6 to 124)
OR 1.54 (0.32 to 7.3)
242 (3 studies)
moderate4
The mean Total sperm motility over time - Total sperm motility at 3 months or less in the intervention groups was 13.47 higher (3.45 to 23.49 higher)
454 (10 studies)
very low5,6
Sperm concentration over time - Sperm concentration at 3 months or less Follow-up: 4-36 weeks
The mean Sperm concentration over time Sperm concentration at 3 months or less in the intervention groups was 6.79 higher (5.09 lower to 18.66 higher)
302 (7 studies)
very low7,8
*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; OR: Odds ratio; GRADE Working Group grades of evidence High quality: Further research is very unlikely to change our confidence in the estimate of effect. Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate. Low quality: Further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate. Very low quality: We are very uncertain about the estimate.
1 2 3 4 5 6 7 8
The trials lacked adequate explanation for methods of randomisation, allocation concealment and blinding wide confidence intervals Only 6 trials gave adequate explanation for randomisation and allocation concealment Total number of events = 7 Very high heterogeneity and remained high after sensitivity analysis Wide confidence intervals Very high heterogeneity Wide confidence intervals
BACKGROUND
there are advocates who state that this should be part of a standard assessment of the male partner when a couple presents with subfertility (Boe-Hansen 2006); although it is recognised that the technique has its limitations and hence strict laboratory control and standardisation is required (Boe-Hansen 2005). Sperm DNA fragmentation does not appear to inuence fertilisation in in vitro fertilisation (IVF), although a negative correlation of sperm DNA damage with embryo or blastocyst development has been described (Evenson 2006; Li 2006; Tarozzi 2007). Women undergoing intrauterine insemination with a sperm DNA fragmentation index < 30%, as measured by the SCSA, are seven times more likely to achieve a pregnancy than those couples where the male partner had a higher degree of sperm DNA damage (Evenson 2006). The evidence for the effect of a high degree of sperm DNA damage upon pregnancy outcome is less clear. A meta-analysis of sperm DNA fragmentation assessed by the SCSA determined that if the sperm DNA fragmentation was < 30% the couple were twice as likely to conceive in an IVF cycle than if it was greater than 30%, though the evidence for a benet in women undergoing intracytoplasmic sperm injection (ICSI) was unclear (Evenson 2006). However, a meta-analysis of SCSA papers published in the same year demonstrated conicting results (Li 2006). This meta-analysis found that there was no effect of sperm DNA damage assessed by the SCSA assay upon the outcome of IVF or ICSI (Li 2006). The meta-analysis also reviewed the effect of sperm DNA damage as assessed by the TUNEL assay. This demonstrated a reduced pregnancy rate in women undergoing IVF when the male partner had a high degree of sperm DNA damage, but no difference if they were undergoing ICSI (Li 2006). It also appears that miscarriage is more likely in women undergoing assisted reproduction when the sperm DNA damage is high (Borini 2006).
Description of the condition

Eighty million people worldwide are affected by the inability to have children (Tournaye 2006). Delayed conception affects 15% of couples trying to conceive (Attia 2007) and male factor subfertility accounts for up to 50% of these cases. A review by Tremellen (Tremellen 2008) notes that one man in 20 will be affected by subfertility. Some 30% to 80% of male factor subfertility cases are said to be due to the damaging effects of oxidative stress (Tremellen 2008). Oxidative stress occurs when reactive oxygen species (ROS) overcome the semens natural antioxidant defences and cause cellular damage (Tremellen 2008). Some studies have suggested that sperm production and quality has decreased over the past few decades (Stankiewicz 2003). The increased levels of ROS are thought to be due to environmental factors such as high temperatures, electromagnetic radiation, pesticides and pollution; and lifestyle factors of advanced age, alcohol consumption, smoking, stress, obesity and poor diet. Other factors include infections, autoimmunity and chronic disease (Aitken 2007; Alvarez 2003; Tremellen 2008). Spermatozoal membranes are rich in poly-unsaturated fatty acids and are susceptible to oxygen damage from lipid peroxidation (Sheweita 2005). Abnormal spermatozoa and contaminating leukocytes generate ROS (Sikka 1995). Antioxidants that are naturally found in semen include vitamins E and C, superoxide dismutase, glutathione and thioredoxin (Omu 2008). These antioxidants act as free radical scavengers that help to overcome ROS. Subfertile men have been identied as having lower levels of antioxidants in their semen compared to fertile men (Tremellen 2007). A study by Bykova (Bykova 2007) showed that ROS levels were signicantly higher in infertile sperm samples when compared with healthy controls and that the infertile men who provided these samples may benet from an antioxidant supplement. ROS are thought to cause fertility problems in two ways, rstly by damaging the sperm membrane thus affecting the sperm motility and the ability of the spermatozoa to break down the oocyte membrane; and secondly by altering the sperm DNA. Spermatozoal DNA integrity is one of the major determinants of normal fertilisation and embryo growth in natural and assisted conception (Agarwal 2003; Aitken 2004; Tarozzi 2007). Indeed, many men with normal seminal parameters may have a high degree of sperm DNA damage and this correlates with a poor chance of natural conception (Boe-Hansen 2006). Sperm DNA damage or integrity can be assessed in a number of ways. These include sperm chromatin structural assay by ow cytometry (SCSA); enzymatic labelling of broken DNA strands, the terminal deoxynucleotide transferase-mediated nick end-labelling assay (TUNEL); and microscopic observations of DNA fragments, the Comet assay. The greatest experience with and standardisation of exists for the SCSA (Aitken 2007; Evenson 2007). Indeed
Description of the intervention

Antioxidants are both biological (enzymes) and chemical substances that reduce oxidative damage. The chemical antioxidants are both natural and synthetic. These can be derived from nutritional sources and from supplementation (Sikka 1995). The predominant supplementary antioxidants that are studied in male subfertility clinical trials are vitamin E, vitamin C, carotenoids, ubiquinol and the micronutrients folate and zinc (Eskenazi 2005). A paper by Tremellen (Tremellen 2008) discussed the use of a combination of vitamins and minerals to improve pregnancy rates for subfertile men. Polyunsaturated fatty acids (PUFAs) are sources of antioxidants and are commonly taken in the community as nutritional supplements. PUFAs have varying effects in male fertility. They provide antioxidants and also increase the plasma uidity of the sperm membrane which acts to assist with conception, however this uidity makes the sperm susceptible to reactive oxygen species and lipid peroxidation that can damage the sperm (Wathes 2007). Wathes states that It appears that PUFAs are a two edged sword - some
5
are essential, but too many are potentially harmful (Wathes 2007, page198). An open study by Comhaire attempted to overcome the double edged sword of essential fatty acid supplements by also treating the subfertile men in their study with antioxidant supplements of acetyl cysteine or beta-carotene and alpha-tocopherol (Comhaire 2000). PUFAs are classied into omega-3, omega-6 and omega-9. Omega-9 is synthesised by animals but omegas-3 and 6 need to be supplemented in the diet. The main sources of omega-6 are vegetable oils. Sources of omega-3 are vegetable and sh oils (Wathes 2007). Antioxidants are widely available and inexpensive when compared to other fertility treatments however this review did not include any cost benet analysis.
The main questions asked in this review were whether: 1. supplemented oral antioxidants improve outcomes for couples with male factor infertility or unexplained infertility who are undergoing ART; 2. different types of antioxidants given to men have varying effects in this setting.
METHODS
Criteria for considering studies for this review
How the intervention might work

Antioxidants are known to scavenge and dispose of ROS, suppress their formation and also act to oppose the actions of ROS (Sikka 1995). The dietary intake of antioxidants has been shown to be critically important for semen quality. Some clinical trials of antioxidant supplementation have shown an associated increase in fertilisation rates, possibly by reducing oxidative stress, lipid peroxidation potential and ROS levels (Eskenazi 2005).
Types of studies
Inclusion criteria Only randomised controlled trials were eligible for inclusion. The participants were randomised to antioxidant versus placebo, an alternative antioxidant, or no treatment. Only pre-crossover data were used from randomised crossover trials, as achieving outcomes such as pregnancy and live birth precluded couples entering the next trial phase (Dias 2006).
Why it is important to do this review

Currently there is limited evidence that antioxidant supplementation improves outcomes for subfertile couples. Although some clinical trials of supplemental antioxidants have suggested benets in treating male subfertility there are other trials that fail to demonstrate the same benet (Agarwal 2004). Tournaye (Tournaye 2006) described the consensus on the treatment of unexplained male subfertility with antioxidants as potentially benecial but stated a need for further evaluation. The purpose of this Cochrane review was to assess the effects of antioxidants on men with documented sperm DNA damage and men with impaired semen parameters from appropriate clinical trials that use the clinically relevant parameters of live birth and pregnancy. The review also assessed the effectiveness of different antioxidants and dosages on these outcomes.
Exclusion criteria Any quasi-randomised trials Types of participants
Inclusion criteria Trials that included men who were part of a couple with male factor subfertility or unexplained subfertility and who were undergoing ART using their own gametes. In situations where individuals were randomised again following failed cycles the data would not be pooled in a meta-analysis unless individual data could be excluded.
OBJECTIVES
The objective of this review was to determine whether supplemented oral antioxidants compared to placebo, no treatment or another antioxidant improve outcomes for couples with a subfertile male partner. Couples were undergoing assisted reproduction technologies (ART) such as in vitro fertilisation (IVF), intrauterine insemination (IUI) or intracytoplasmic sperm injection (ICSI). Exclusion criteria Trials that included men taking any other fertility enhancing drugs Trials that included men who had had chemotherapy treatment
6
Types of interventions Trials were included if they investigated any type of chemical supplementary antioxidant Trials were included if they investigated individual or combined antioxidants (including head to head comparisons) Any dose of antioxidant was included Trials using only the oral route of administration were included
Types of outcome measures
Primary outcomes
This register also contains unpublished trial abstracts. These were found by handsearching 20 relevant journals and conference proceedings. (2) The following databases were searched using the Ovid platform: Cochrane Central Register of Controlled Trials (CENTRAL) (The Cochrane Library 2010, Issue 3) (Appendix 3); MEDLINE (inception to 22.08.10) (Appendix 1); EMBASE (2008 to 22.08.10) (Appendix 2); CINAHL (Cumulative Index to Nursing and Allied Health Literature) (1980 to 22 August 2010) (Appendix 4); PsycINFO (inception to 22.08.10) (Appendix 5); AMED (Allied and Complementary Medicine) (inception to 22 August 2010) (Appendix 6). Both indexed and free text terms were used. The MEDLINE search was combined with the Cochrane highly sensitive search strategy for identifying randomised trials, which appears in the Cochrane Handbook of Systematic Reviews of Interventions (Version 5.0.1 chapter 6, 6.4.11). The EMBASE and CINAHL searches were combined with trial lters developed by the Scottish Intercollegiate Guidelines Network (SIGN) (www.sign.ac.uk/mehodology/lters.html#random). Searches were not limited to any one language.
Live birth rate per couple randomised
Secondary outcomes
Pregnancy rate per couple Miscarriage rate per couple, or spontaneous abortion Stillbirth rate per couple Level of sperm DNA damage after treatment Sperm concentration Sperm motility Adverse effects or side effects associated with antioxidant supplementation, or as was reported by the trial. Withdrawal numbers due to individual adverse effects were reported
Searching other resources Appropriate journals were handsearched for trial conference abstracts. These journals included Human Reproduction which contains abstract supplements for the European Society of Human Reproduction and Embryology (ESHRE) and Fertility and Sterility which contains abstract supplements for the American Society for Reproductive Medicine (ASRM). Lists of journals handsearched by the Menstrual Disorders and Subfertility Group (MDSG) are found in the MDSG Module. Research registers were searched for ongoing and recently completed trials: ClinicalTrials.gov, a service of the US national Institutes of Health (http://clinicaltrials.gov/ct2/home) (last searched 6 August 2010); The World Health Organization International Trials Registry Platform search portal (www.who.int/trialsearch/ Default.aspx) (last searched August 2010). Conference abstracts were found on the ISI Web of Knowledge: http://isiwebofknowledge.com/ (last searched 6 August 2010). OpenSIGLE database was searched for European grey literature: http://opensigle.inist.fr/ (last searched 6 August 2010). Citation lists from review articles and other relevant publications were searched. Personal communication with manufacturers, experts and specialists in the eld.
7
Search methods for identication of studies

See the Cochrane Menstrual Disorders and Subfertility Group (MDSG) for methods used in reviews, as stated in their Module (http://www.mrw.interscience.wiley.com/cochrane/ clabout/articles/MENSTR/frame.html). All reports that described randomised controlled trials of oral antioxidant supplementation for subfertile men were to be found using the following strategy.
Electronic searches (1) The Menstrual Disorders and Subfertility Group Specialised Register of controlled trials was searched by the Groups Trial Search Coordinator using the terms male subfertility or In vitro fertilisation or IVF or Intracytoplasmic sperm injection or ICSI AND antioxidants or vitamin e or vitamin or ascorbic acid or zinc or folate or selenium or glutathione or ubiquinol or carnitine or astaxanthin or coenzyme Q10 or lycopene or Menevit or carnitine or carnitine or ascorbic acid or zinc or fatty acids or oil or sh oils or plant extracts in the titles, abstracts and keywords (from inception to 22.08.10) (Appendix 7).
Data collection and analysis
Selection of studies Two people were responsible for independently selecting the trials. Titles and abstracts from the searches were scanned. We obtained the full text of those articles that appeared to be eligible for inclusion. On assessment they were then placed in included (Characteristics of included studies), excluded (Characteristics of excluded studies), ongoing (Characteristics of ongoing studies) studies or studies awaiting assessment (Characteristics of studies awaiting classication). One of these individuals is a content expert. Any disagreements were resolved through consensus or by a third party. Studies were appraised in an unblinded fashion, as recommended by the Cochrane Menstrual and Fertility Disorders Group. Further information, where required, was sought from the authors
The authors aimed to minimise reporting bias, publication bias, within trial reporting bias and their potential impact by ensuring that a comprehensive search for eligible studies was carried out. Authors were also alert to duplication of data. As there were 10 or more studies in the analysis, a funnel plot was used to explore the possibility of small study effects (a tendency for estimates of the intervention effect to be more benecial in smaller studies). Care was taken to search for within study reporting bias, such as trials failing to report obvious outcomes or reporting them in insufcient detail to allow inclusion. Measures of treatment effect The dichotomous data for live birth, pregnancy rate, miscarriage, stillbirth, and adverse events were expressed as odds ratios (OR) with 95% condence intervals (95% CI) and combined in a metaanalysis with Rev Man software using the Peto method and a xedeffect model (Higgins 2006). The OR has mathematically sound properties that are consistent with benet or harm and work well in small samples with rare events. This effect measure is appropriate when considering subfertility. For continuous data (for example sperm quality measurements) mean differences between treatment groups were calculated with associated standard deviations and 95% condence intervals. These were displayed on forest plots, if possible. Attempts were made to contact all authors of the included trials that reported data in a form that was not suitable for meta-analysis, for example data reported in medians or ranges. Where additional data were not forthcoming the data were reported under other data and not included in the meta-analysis. Unit of analysis issues
Data extraction and management The studies that appeared to meet the inclusion criteria were independently assessed by the two review authors (MS and JB) using data extraction forms. Any discrepancies were resolved with discussion. The data extraction forms included methodological quality and allocation information. This information was included in the review and presented in the characteristics of included and excluded studies tables (see Characteristics of included studies; Characteristics of excluded studies) following the guidance of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2006).
Assessment of risk of bias in included studies The included studies were assessed for risk of bias using the Cochrane risk of bias assessment tool to assess: sequence generation; allocation concealment; blinding of participants, providers and outcome assessors; completeness of outcome data; selective outcome reporting; and other potential sources of bias. MS and JB assessed these six criteria; any disagreements were resolved by consensus or by discussion with a third author. The conclusions were presented in the Risk of bias table (Characteristics of included studies) and incorporated into the interpretation of review ndings by means of sensitivity analyses. Where identied studies failed to report the primary outcome of live birth, but did report interim outcomes such as pregnancy, informal assessment was undertaken on whether those reporting the primary outcomes had typical values for the interim outcomes. If other potential sources of bias were identied in advance, these were specied and the number of criteria increased accordingly. Sensitivity analyses were specied later, under the appropriate heading.
The outcomes of live birth and pregnancy were analysed as per couple randomised. The sperm outcomes analyses were per man randomised. We included both parallel and crossover trials, using only the rst phase of the crossover trial. Dealing with missing data We attempted to contact the authors of the trials with any missing data by email or post. If there was no reply, and if possible, we reported the data in terms of intention to treat. If this was not possible the trials were placed in studies awaiting assessment. Assessment of heterogeneity The authors considered whether the clinical and methodological characteristics of the included studies were sufciently similar to provide a meaningful summary in a meta-analysis. Heterogeneity between the treatment effects of different studies were studied by looking at the points on the forest plot, the overlap of condence intervals (a poor overlap indicates heterogeneity), and the Chi2 statistical test for heterogeneity. A low P value (or a
8
large Chi2 statistic relative to its degree of freedom) shows evidence of heterogeneity of treatment effects, or that the differences were not likely to be by chance (Higgins 2006). If there was only a small number of trials or they had small sample sizes then a P value of 0.10 rather than 0.05 was used to show heterogeneity. To more formally quantify the variations between the studies the I2 statistic was used (Higgins 2006). This statistic describes the variation in effect estimates due to heterogeneity rather than by chance, as a percentage. If a value over 50% was found we assumed that there was large heterogeneity.
The aim here was to dene analyses that were comprehensive and mutually exclusive so that all eligible study results could be be slotted into one stratum only. Comparisons were specied so that any trials falling within each stratum could sensibly be pooled for meta-analysis. Stratication allowed for consideration of effects within each stratum as well as, or instead of, an overall estimate for the comparison. Subgroup analysis and investigation of heterogeneity Subgroup analysis was performed on the following. Trials that reported antioxidants (individual and combined) given to men with documented elevated levels of sperm DNA damage, as dened by the trials, who were undergoing fertility treatment (ART). A minimum of 80% of participants in the trials had to satisfy this criterion for inclusion. If individual data were available these were used in preference. Trials that reported antioxidants (individual and combined) given to men with impaired semen parameters who were undergoing fertility treatment (ART). A minimum of 80% of participants in the trials had to satisfy this criterion for inclusion. If individual data were available these were used in preference. Treatment was divided into the subgroups of individual and combined antioxidants for live birth and pregnancy outcomes. Subgroup analysis was also performed on those studies that reported live birth and pregnancy in order to assess any overestimation of effect and reporting bias. Subgroup analysis was also used in the sperm outcomes of motility and concentration over time, at: three, six and nine months. Sensitivity analysis We performed sensitivity analysis (using the random-effects model in RevMan software) for the primary outcomes to assess whether the ndings from the analysis were robust. Sensitivity analysis considered whether conclusions were any different if eligibility was restricted to studies without high risk of bias. We also stated that sensitivity analysis would be performed on unpublished studies. These studies may not have been peer reviewed and thus could be of lower quality. However there were no unpublished trials in this review. Initially trials were subgrouped according to placebo or no treatment, however after statistical advice all the trials comparing antioxidants to placebo and no treatment were grouped as one then a sensitivity analysis was performed to assess any evidence of difference in effect between the placebo and the no-treatment group. These results are described in the Risk of bias in included studies section of the review.
Assessment of reporting biases If there was a question of publication bias, a funnel plot was drawn. A gap on either side of the graph would have given a visual indication that some trials have not been identied (Higgins 2006). This is often due to the difculties in locating unpublished trials. If there were outcomes with no data the review included empty data tables. This indicates that further clinical trials need to be conducted in this area. It also allows for updating, when new data may be found.
Data synthesis Statistical analysis of the data was carried out using Review Manager 5 (RevMan 5). Pregnancy outcomes are considered positive and higher numbers of pregnancy rates are considered a benet. The outcomes of miscarriage and adverse events are negative effects and higher numbers would be considered harmful. These aspects have to be considered when assessing the summary graphs (Attia 2007). The data from primary studies were combined using a xed-effect model in the following comparisons. 1. Antioxidants versus control (placebo or no treatment), stratied by type of antioxidant. Further analyses of sperm motility and concentration were carried out by sub grouping trials over time: at three, six and nine months. Subgrouping was also performed on the outcomes of live birth and pregnancy. Trials that reported both of these outcomes were subgrouped into live birth and pregnancy in order to assess any overestimation of effect in the pregnancy outcome. Subgrouping was also performed on those trials that reported both live birth and pregnancy and trials that only reported pregnancy in order to assess any reporting bias in the pregnancy outcome. 2. Antioxidants versus antioxidants or head to head stratication by type of antioxidant. An increase in the odds of a particular outcome, which may be benecial (for example live birth) or detrimental (for example adverse effects), were displayed graphically in the meta-analyses to the right of the centre line and a decrease in the odds of an outcome to the left of the centre line.
RESULTS
9
Description of studies
See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of studies awaiting classication; Characteristics of ongoing studies. Results of the search Two review authors (MS and RH) scanned the abstracts and titles of the results of the search strings. The MEDLINE search produced 406 abstracts; there were six abstracts from CENTRAL, three from CINAHL, 62 from EMBASE, 107 from the MSDG database and three from PsycINFO. Two conference abstracts were found from handsearching the conference proceedings of the European Society for Human Reproduction and Embryology (ESHRE) and the American Society for Reproductive Medicine (ASRM). One title was found from reference lists in reviews. Five non-English trials were assessed for inclusion: two Chinese, one Bulgarian, one Japanese and one Iranian. The two Chinese trials (Li 2005a; Li 2005 ), the Japanese trial (Akiyama 1999) and the Iranian trial (Peivandi 2010) were included in the analysis. The Bulgarian study (Nikolova 2007) was excluded as it did not have random allocation (see Characteristics of excluded studies). After removal of inappropriate and duplicate studies 53 trials remained: 34 trials were included, 15 were excluded and four trials were ongoing. Included studies There were 34 included trials in this review. Full details of these trials can be seen in the table of included studies (Characteristics of included studies). The trials came from 19 different countries and randomised 2876 couples. Eight trials were based in Italy (Balercia 2005; Balercia 2009; Biagiotti 2003; Cavallini 2004; Galatioto 2008; Lenzi 2003; Lenzi 2004; Lombardo 2002). Three trials were from Iran (Safarinejad 2009; Safarinejad 2009a; Peivandi 2010). Two trials each were from USA (Dawson 1990; Sigman 2006), Tunisia (Keskes-Ammar 2003; Nozha 2001), United Kingdom (Kessopoulou 1995; Scott 1998), Hungary (Micic 1988; Zavaczki 2003), Kuwait (Omu 1998; Omu 2008) and China (Li 2005; Li 2005a). The remaining trials were from single countries: Japan (Akiyama 1999), Turkey (Ciftci 2009), Canada (Conquer 2000), France (Greco 2005), Mexico (Merino 1997), Germany (Rolf 1999), Saudi Arabia (Suleiman 1996), Australia (Tremellen 2007), Hong Kong (Wang 1983), Netherlands (Wong 2002) and Belgium (Zalata 1998). Data were able to be extracted from 28 of these included trials. The six remaining trials either did not report any data or the continuous data were reported in medians or ranges (Biagiotti 2003; Lombardo 2002; Merino 1997; Nozha 2001; Wong 2002; Zalata 1998), see Characteristics of included studies and the analyses (Analysis 1.14; Analysis 1.15; Analysis 1.16; Analysis 1.17; Analysis 3.9; Analysis 3.10). Attempts were made to contact all authors of the included trials for further details and clarication.
The duration of treatment period ranged from three weeks with three weeks follow up (Dawson 1990) to 26 weeks treatment and a 24-month follow up (Wong 2002). The longest follow-up period was in the trial by Kessopoulou (Kessopoulou 1995), with a threeweek treatment period and a two-year follow-up period. Three trials (Kessopoulou 1995; Sigman 2006; Tremellen 2007) enrolled men who, as a couple, were undergoing IVF or ICSI. There were no trials that enrolled men who were part of a couple undergoing intrauterine insemination (IUI).
Participants
Most trials included couples who had attended a fertility clinic, with a fertile partner and had been trying to conceive with regular intercourse for over one year. The mean age of the participants ranged from 20 to 52 years. Most men in the included trials had a decient level of spermatozoa in the seminal uid (oligospermia) or a low motility of sperm in the seminal uid (asthenospermia). Trials excluded men with any inammatory disease, antibody problems or chromosomal problems and most trials stated that they did not enrol men who smoked, took any additional medication or drank alcohol. Details of inclusion and exclusion criteria are found in Characteristics of included studies.
Interventions
A wide variety of antioxidants were used in the included trials. The three comparisons were: antioxidants versus placebo, antioxidants versus no treatment and antioxidants versus antioxidants (or head to head). The comparison antioxidants versus placebo included the antioxidants: vitamin E, combined antioxidants plus minerals, L-acetylcarnitine, L-carnitine, L-acetyl carnitine plus L-carnitine, pentoxifylline and magnesium. The comparison antioxidants versus no treatment included: zinc, combined antioxidants plus minerals, pentoxifylline, zinc plus vitamin E, zinc plus vitamin E plus vitamin C. In the analysis placebo and no treatment groups were grouped as one and the comparison labelled antioxidants versus control. The third comparison, head to head, was implemented in an attempt to assess whether one antioxidant may be more effective than another, as stated in the protocol. However in some cases they may have countered each other as the treatment was also included in the control (Balercia 2005; Omu 2008; Safarinejad 2009; Scott 1998). The antioxidants used were: L-acetyl carnitine, L-carnitine, L-acetyl carnitine plus L-carnitine, ethyl cysteine, vitamin E, docosahexaenoic acid, vitamin C, selenium, vitamin B, combined antioxidants plus minerals and N-acetyl cysteine.
Outcomes
The primary outcome for this review was as follows.

10
Live birth per couple. Three trials reported live birth (Kessopoulou 1995; Omu 1998; Suleiman 1996) in the antioxidant versus control groups. Secondary outcomes for this review were as follows. Pregnancy rate per couple, as reported by 16 trials (Balercia 2005; Balercia 2009; Cavallini 2004; Galatioto 2008; Lenzi 2003; Lenzi 2004; Li 2005; Omu 1998; Peivandi 2010; Rolf 1999; Safarinejad 2009a; Sigman 2006; Suleiman 1996; Tremellen 2007; Wang 1983; Zavaczki 2003). Li 2005 (Li 2005) used head to head comparisons. Balercia 2005 (Balercia 2005) was multi-armed and used both head to head and a control while the remaining trials investigated antioxidant versus control. Miscarriage rate, as reported by three trials (Omu 1998; Suleiman 1996; Tremellen 2007) in the antioxidant versus control comparison. DNA fragmentation, as reported by one trial (Greco 2005). This trial compared antioxidant versus control. The sperm motility and concentration outcomes were divided into three groups: measurement at three, six and nine months or more; the included trials measured motility and concentration at these different points in time. This analysis was only useful in directly comparing the same trials reporting at the three time points and not in comparing results of meta analyses that included different subsets of trials. Sperm motility at three months or less was reported by 13 trials (Akiyama 1999; Ciftci 2009; Conquer 2000; Dawson 1990; Greco 2005; Keskes-Ammar 2003; Li 2005; Micic 1988; Omu 2008; Peivandi 2010; Scott 1998; Sigman 2006; Zavaczki 2003). Sperm motility at six months was measured by four trials (Balercia 2005; Balercia 2009; Safarinejad 2009; Suleiman 1996). Sperm motility at nine months or more was reported by three trials (Balercia 2005; Balercia 2009; Safarinejad 2009a). Sperm concentration at three months or less was reported by nine trials in total (Akiyama 1999; Conquer 2000; Ciftci 2009; Greco 2005; Peivandi 2010; Rolf 1999; Wang 1983; Zavaczki 2003). Sperm concentration at six months was reported as an outcome by a total of ve trials (Balercia 2005; Balercia 2009; Lenzi 2004; Li 2005a; Safarinejad 2009; Wang 1983). Sperm concentration at nine months or more was reported on by three trials (Balercia 2005; Balercia 2009; Safarinejad 2009a). Side effects were subgrouped according to the type of side effect, as described by the trials. There were only seven trials that reported side effects. Failure by the remaining included studies to
report adverse effects could be construed as selective reporting. The side effects reported were mainly gastrointestinal upsets (Cavallini 2004; Kessopoulou 1995; Safarinejad 2009a; Sigman 2006; Tremellen 2007; Zavaczki 2003). These were reported as mild epigastria, nausea, gastric acidity, gastro-oesophageal reux, constipation and diarrhoea. One of the six trials (Cavallini 2004) also reported on the side effect of euphoria. The trial by Li (Li 2005) reported that no side effects were found in either the treatment or control groups. It should also be noted that the trial by Cavallini (Cavallini 2004) had unexplained differences in randomisation and analysis numbers for this outcome, and therefore introduced some reporting bias. Excluded studies In total 15 trials were excluded. Six trials (Cavallini 2004a; Ebisch 2006; Pryor 2003; Rolf 1998; Tremellen 2006; Zalata 1998a) were excluded as they reported the initial analysis or were a conference proceeding of a later, included trial. Five trials (Comhaire 2005; Micic 2001; Nikolova 2007; Pawlowicz 2001; Stanislavov 2009) were excluded because they did not use a randomised sequence to allocate the participants to the treatment and control groups. Four trials (Ebisch 2003; Elgindy 2008; Vicari 2001; Vicari 2001a) did not have appropriate inclusion criteria for population, outcomes or controls. Refer to the table Characteristics of excluded studies.
Studies awaiting classication

There is one trial (Wang 2010) awaiting classication. It has not yet been assessed.
Ongoing studies
Four trials were ongoing (Revel 2006; Sadeghi 2008; Tsafrir 2010; Gonzalez 2009).
Risk of bias in included studies

Adequate sequence generation All included trials were randomised. Four trials (Akiyama 1999; Kessopoulou 1995; Lenzi 2003; Peivandi 2010) had a randomised crossover design. In the meta-analysis only the rst phase data were used as all trials reported rst and second phase data separately. Including crossover data may introduce reporting bias due to the fact that it is unusual for a trial to report usable data at each phase. The remaining 30 trials had a randomised parallel group design. Only 10 trials described their methods of randomisation (Biagiotti 2003; Galatioto 2008; Keskes-Ammar 2003; Rolf 1999; Safarinejad 2009; Safarinejad 2009a; Scott 1998; Sigman 2006; Tremellen 2007; Wong 2002). The predominant method was by computer generated blocks. Some degree of selection bias may
11
have been present in the trials due to the high numbers of included trials that did not describe their method of randomisation (24 out of 34 trials). One study (Li 2005) had a large imbalance between the intervention and control groups at the randomisation stage; 150 men were randomised: 90 into the treatment group and 60 into the control group. This imbalance was not explained in the report. Attempts were made to contact the author but as yet there has been no reply. Allocation The methods of allocation concealment were generally quite poorly described in the included studies. Therefore there was some risk of selection bias in the trials under review. Nine trials (Cavallini 2004; Ciftci 2009; Galatioto 2008; Rolf 1999; Safarinejad 2009; Sigman 2006; Tremellen 2007; Wong 2002; Peivandi 2010) described their method of allocation. These methods included anonymous coloured boxes, sealed opaque envelopes and numbered bottles. Blinding Seventeen trials (Balercia 2005; Balercia 2009; Cavallini 2004; Kessopoulou 1995; Lenzi 2003; Lenzi 2004; Li 2005; Lombardo 2002; Peivandi 2010; Rolf 1999; Safarinejad 2009; Safarinejad 2009a; Scott 1998; Sigman 2006; Suleiman 1996; Tremellen 2007; Wong 2002) were described as randomised, double blind controlled trials. The researchers and patients were blinded. Two trials (Ciftci 2009; Dawson 1990) stated that a placebo was used as the control but only the patients were blinded. Other trials (Conquer 2000; Greco 2005; Merino 1997; Wang 1983; Zavaczki 2003) used a placebo as the control but did not discuss blinding. One trial (Galatioto 2008) was open label but stated that the pharmacy was blinded during the randomisation process. One trial (Keskes-Ammar 2003) was described as open label however the numeric code was withheld from the investigators and patients. Suleiman 1996 reported that if a couple became pregnant then the treatment was stopped; however they did not appear to stop the placebo. This suggests that the investigators had knowledge of whether the patients were in the placebo or antioxidant. Incomplete outcome data Only two trials (Balercia 2009; Galatioto 2008) actually stated that they used intention to treat in their analysis, however most of the included trials accounted for the participants that withdrew from their trials and then analysed the groups in an intention-totreat fashion. Four trials had over 20% withdrawal from their trials. One trial ( Keskes-Ammar 2003) had over 50% drop out rate; these men were accounted for and the main reason given was non-compliance. Cavallini 2004 had 30% dropout rate and reasons were provided for 53 out of the 55 drop outs; these reasons included refusal
due to the chance of taking a placebo and preference for assisted reproduction techniques. There also remained some confusion in this trial on the total numbers randomised and analysed. Li 2005a and Suleiman 1996 had slightly over 20% withdrawal from their trials. One trial (Suleiman 1996) had a large imbalance in numbers. There were found to be 52 in the treatment group and 35 in the placebo once the code had been broken at the end of the trial. There was no indication of how the randomisation was performed therefore the trial did not perform intention to treat. The dropouts were however accounted for: many couples had left the region and some simply failed to continue.
Selective reporting There was great potential for reporting bias throughout this review as no outcome had been consistently reported by all studies and the key outcomes were reported by a minority of studies. The trial by Cavallini (Cavallini 2004) had unexplained differences in randomisation and analysis numbers and therefore may have introduced some reporting bias. Four trials (Balercia 2005; Balercia 2009; Scott 1998; Sigman 2006) gave pregnancy data although pregnancy as an outcome was not stated a priori in the protocols. Li 2005a stated that pregnancy was an outcome but did not provide the data. One trial (Wang 1983) gave follow-up data on the treatment group but no data were supplied on the control group. Three trials reported on the adverse event of miscarriage ( Tremellen 2007; Suleiman 1996; Omu 1998). Six trials reported on side effects, which were mainly gastrointestinal (Cavallini 2004; Kessopoulou 1995; Safarinejad 2009a; Sigman 2006; Tremellen 2007; Zavaczki 2003). Cavallini (Cavallini 2004) also reported euphoria as a side effect. The majority of included trials did not report side effects and this could be construed as a reporting bias.
Other potential sources of bias Funding sources were stated by nine trials (Conquer 2000; Kessopoulou 1995; Lenzi 2003; Lombardo 2002; Omu 1998; Peivandi 2010; Rolf 1999; Wang 1983; Zavaczki 2003). Two trials (Conquer 2000; Wang 1983) stated that funding was from a commercial source and the remaining seven obtained funding through non-commercial avenues or university grants. However three of these non-commercially funded trials (Kessopoulou 1995; Lenzi 2003; Zavaczki 2003) received the medication from commercial interests. Publication bias was assessed using a funnel plot (Figure 1). Pregnancy outcome was used for the funnel plot as there were only two studies in the primary outcome of live birth. The plot appeared to be symmetrical and therefore we assumed that there was little publication bias.
12
Figure 1. Funnel plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.2 Pregnancy rate per couple randomised.
Trials comparing an antioxidant to a placebo or to no treatment were grouped together and compared in the subgroups. Sensitivity analysis was performed on the outcomes whereby the placebo data were removed in order to assess whether the no treatment trials were biasing the results. This was only possible for the outcome of sperm motility at three months and the result remained statistically signicant. The other outcomes either contained only placebo trials or contained only one no-treatment trial, therefore sensitivity analysis could not be performed.
Live birth per couple randomised
Effects of interventions
See: Summary of ndings for the main comparison Antioxidant/s versus control for male subfertility Antioxidant(s) versus control
Overall there was a statistically signicant difference in favour of the antioxidant treated group compared with the control group for live birth (pooled OR 4.85, 95% CI 1.92 to 12.24; P = 0.0008, I2 = 0%) (Analysis 1.1). Three trials reported on this outcome (Kessopoulou 1995; Omu 1998; Suleiman 1996). In total there were 20 live births from a total of 214 couples randomised, 18 (15%) in the antioxidant group (n = 116) and two (2%) in the control group (n = 98) (Figure 2). The I2 statistic in the analysis was 0% however sensitivity analysis was performed on the trial by Suleiman 1996 as this trial had more than 20% of participants drop out of the trial. There remained a statistically signicant difference in favour of the antioxidant-treated group compared with the control group for live birth (pooled OR 3.94, 95% CI 1.14 to 13.55; P = 0.03, I2 = 0%).
13
Figure 2. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.1 Live birth per couple randomised.
Two trials (Kessopoulou 1995; Suleiman 1996) reported on this outcome comparing the intervention of vitamin E versus placebo. Kessopoulou 1995 included couples undergoing IVF or ICSI. There was a statistically signicant difference (pooled OR 6.44, 95% CI 1.72 to 24.04; P = 0.006, I2 = 0%) in the number of live births between the antioxidant and placebo groups that favoured the antioxidant over the placebo. One trial (Omu 1998) reported on live births per couple randomised and compared zinc versus no treatment. In total there were 10 live births from 97 couples randomised, eight in the antioxidant group (n = 49) and two in the no treatment group (n = 48). There was a statistically signicant difference (OR 3.67, 95% CI 1.00 to 13.51; P = 0.05) in the number of live births between the antioxidant and no treatment, in favour of the antioxidant.
Pregnancy rate per couple randomised
Fifteen trials reported the outcome of pregnancy rate per couple randomised (Balercia 2005; Balercia 2009; Cavallini 2004; Galatioto 2008; Lenzi 2003; Lenzi 2004; Omu 1998; Peivandi 2010; Rolf 1999; Sigman 2006; Safarinejad 2009a; Suleiman 1996; Tremellen 2007; Wang 1983; Zavaczki 2003). There were 964 couples in total, 515 in the antioxidant group and 449 in the control group. In total there were 96 pregnancies, 82 (16%) in the antioxidant group and 14 (3%) in the control group. There was a statistically signicant difference (pooled OR 4.18, 95% CI 2.65 to 6.59; P < 0.00001, I2 = 0%) (Analysis 1.2) in the number of pregnancies between the antioxidant and control groups in favour of the antioxidant group over the control group (Figure 3).
14
Figure 3. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.2 Pregnancy rate per couple randomised.
15
The trial by Cavallini 2004 included men with variocoele associated oligoasthenospermia whereas many of the other included trials excluded men with variocoele (see Characteristics of included studies). Heterogeneity was however low in this analysis and therefore sensitivity analysis was not performed. Sensitivity analysis was performed on those trials (Cavallini 2004; Suleiman 1996; Wang 1983) that reported over 20% dropouts. There was a statistically signicant difference (pooled OR 3.52, 95% CI 2.08 to 5.95; P < 0.00001, I2 = 0%) in favour of the antioxidant group over the control group. Two of the 15 trials included couples undergoing IVF or ICSI. When these two trials (Sigman 2006; Tremellen 2007) were removed the effect of antioxidants over control remained as a statistically signicant difference for the remaining trials (pooled OR 4.52, 95% CI 2.71 to 7.53; P < 0.00001, I2 = 0%), in favour of the antioxidant over control. When these two trials (Tremellen 2007; Sigman 2006) were subgrouped together there was also a statistically signicant difference in favour of the antioxidant over the control (pooled OR 3.06, 95% CI 1.10 to 8.47; P = 0.03, I2 = 7%). After removing those trials that used a placebo as a control, as opposed to no treatment, there remained a statistically signicant difference (pooled OR 4.97, 95% CI 1.63 to 15.14; P = 0.005, I 2 = 0%) although the condence intervals widened, in favour of the antioxidant group over the no-treatment group. Two trials (Rolf 1999; Wang 1983) were not estimable due to no pregnancies in either the treatment or control group. Two trials (Galatioto 2008; Omu 1998) reported on pregnancy
rate per couple randomised, comparing antioxidants versus no treatment. In total there were 14 pregnancies from 130 couples randomised, 12 in the treatment group (n = 69) and two in the no-treatment group (n = 70). There was a statistically signicant difference (pooled OR 4.97, 95% CI 1.63 to 15.14; P = 0.005) in the number of pregnancies between the antioxidant group and the no-treatment group, in favour of the antioxidant group. No heterogeneity was seen in this analysis (Chi2 = 0.07, I2 = 0%). Two further comparisons were created to assess any overestimate of benet or reporting bias that may exist in the pregnancy analysis. Live birth versus pregnancy: two (Omu 1998; Suleiman 1996) out of three trials that reported on live birth and also reported on pregnancy rate were analysed (see Figure 4). The live birth analysis showed a statistically signicant difference in favour of the antioxidants (pooled OR 4.73, 95% CI 1.82 to 12.27; P = 0.001, I2 = 0%). This analysis had slightly lower evidence of effect than when these same trials were analysed in terms of the pregnancy outcome (pooled OR 5.53, 95% CI 2.34 to 13.08; P < 0.0001, I2 = 0%) (see Analysis 2.1). The results for the two analyses appear similar therefore there is no clear evidence of overestimation of benet in the the full (all studies) pregnancy analysis. If the results for pregnancy suggested a larger effect than the corresponding result for live birth then we would have assumed that the full pregnancy analysis overestimated the benet of antioxidants. In the trials that reported on both live birth and pregnancy 22% of the treatment group became pregnant and 17% had a live birth, while 2% of the control became pregnant and 2% had a live birth.
Figure 4. Forest plot of comparison: 2 Live birth versus pregnancy, outcome: 2.1 Live birth.
16
Pregnancy: the trials (Omu 1998; Suleiman 1996) that reported on both live birth and pregnancy (pooled OR 5.53, 95% CI 2.34 to 13.08; P < 0.0001, I2 = 0%) (Analysis 1.18; Figure 5) were compared to all other trials reporting pregnancy (Balercia 2005; Galatioto 2008; Lenzi 2003; Lenzi 2004; Peivandi 2010; Rolf 1999; Safarinejad 2009a; Sigman 2006; Tremellen 2007; Wang 1983; Zavaczki 2003) (pooled OR 4.40, 95% CI 2.78 to 6.97; P < 0.00001, I2 = 0%). The results for the two analyses appear similar thus giving us some condence that the reporting bias in this outcome was low. In the trials that reported both live birth and pregnancy 22% became pregnant in the antioxidant group and 2% in the control, while in the pregnancy-only reported trials 15% of the antioxidant group became pregnant while 3% became pregnant in the control group Figure 5. Forest plot of comparison: 1 Antioxidant(s) versus control, outcome: 1.16 Live birth and pregnancy.
17
Two trials (Galatioto 2008; Tremellen 2007) compared combined antioxidants. There were 25 pregnancies in total, 21 in the antioxidant group (n = 58) and 4 in the control (n = 42). There was a statistically signicant difference in pregnancy rates (pooled OR 3.97, 95% CI 1.39 to 11.33; P = 0.01, I2 = 0%) in favour of the combined antioxidants over the control. One trial (Balercia 2005) compared L-acetyl carnitine to placebo. There were three pregnancies in total, two in the antioxidant group (n = 15) and one in the placebo group (n = 5). The placebo group was divided by three as this trial had multiple arms and the placebo data appeared three times in the analysis. The analysis showed no statistically signicant difference (OR 0.61, 95% CI 0.04 to 9.64; P = 0.72) in pregnancy rates between the antioxidant and placebo groups. Three trials (Balercia 2005; Lenzi 2003; Peivandi 2010) compared L-carnitine to control. There were 12 pregnancies in total, 11 in the antioxidant group (n = 73) and 1 in the control group (n = 63). There was a statistically signicant difference in pregnancy rates (pooled OR 5.08, 95% CI 1.51 to 17.09; P = 0.009, I2 = 29%) in favour of the antioxidant group over the control group. Four trials (Balercia 2005; Cavallini 2004; Lenzi 2004; Sigman 2006) assessed L-acetyl carnitine + L-carnitine versus placebo. There were 21 pregnancies in total, 19 (n = 96) in the antioxidant group and 3 in the placebo (n = 87). There was a statistically signicant difference signicance in pregnancy rates (pooled OR 4.48, 95% CI 1.77 to 11.36; P = 0.002, I2 = 0%) in favour of the antioxidant group over the placebo group. Two trials (Balercia 2009; Safarinejad 2009a) reported on coenzyme versus placebo. There were nine pregnancies in total (n = 272). All pregnancies occurred in the Balercia 2009 trial and none in the Safarinejad 2009a trial. Balercia 2009 was the only trial to be considered for this outcome as there were no pregnancies in the second trial. Balercia 2009 showed no statistically signicant difference (pooled OR 2.16, 95% CI 0.53 to 8.82; P = 0.28) in pregnancy rates between the antioxidant and placebo groups.
Only one trial (Wang 1983) reported on pentoxifylline versus placebo and there were no pregnancies in either group, therefore the results were not estimable. One trial (Zavaczki 2003) reported on magnesium versus placebo. There was one pregnancy in the antioxidant group (n = 10) and none in the placebo (n = 10). The results showed no statistically signicant difference in pregnancy rates between the antioxidant group and the placebo group (OR 7.39, 95% CI 0.15 to 372.38; P = 0.32). One trial (Suleiman 1996) reported on vitamin E versus placebo. There were 11 pregnancies in the antioxidant group (n = 52) and none in the placebo group (n = 35). There was a statistically signicant difference (OR 6.64, 95% CI 1.84 to 23.93; P = 0.004) in the pregnancy rate in favour of the antioxidant group over the placebo group. However this trial had a large degree of attrition due to drop outs leading to an imbalance between the groups; the condence intervals were quite large. One trial (Rolf 1999) reported on vitamin C + vitamin E versus placebo. There were no pregnancies in either group therefore the results were not estimable. One trial (Omu 1998) reported on pregnancy rate with the comparison of zinc versus no treatment. A statistically signicant difference was seen (OR 4.75, 95% CI 1.49 to 15.20; P = 0.009) in favour of the antioxidant group over the no-treatment group.
Adverse event: miscarriage rate per couple randomised
Three trials (Omu 1998; Suleiman 1996; Tremellen 2007) reported on miscarriage per couple randomised. In total there were ve miscarriages in the antioxidant group (n = 139) and two in the control group (n = 103). There was no statistically signicant difference (pooled OR 1.54, 95% CI 0.32 to 7.30; P = 0.59, I2 = 20%) (Analysis 1.3) in miscarriage rate per couple randomised between the antioxidant and placebo groups. The condence intervals of the three trials overlapped the line of no effect (Figure 6).
18
Figure 6. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.3 Adverse event: miscarriage rate per couple randomised.
Stillbirth rate per couple randomised
There were no trials of antioxidants versus control that reported on stillbirths per couple randomised.
Level of sperm DNA fragmentation
There was only one trial (Greco 2005) that reported on the outcome of sperm fragmentation levels. There was a statistically signicant difference (MD -13.80, 95% CI -17.50 to -10.10; P 0.00001) (Analysis 1.4) in favour of the antioxidant group (vitamin C + vitamin E) over the placebo group (Figure 7).
Figure 7. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.10 DNA fragmentation.
19
Total sperm motility at three months or less
Ten trials (Ciftci 2009; Conquer 2000; Dawson 1990; Greco 2005; Micic 1988; Omu 2008; Peivandi 2010; Scott 1998; Sigman 2006; Zavaczki 2003) reported on total sperm motility at three months or less. In total there were 514 participants, 302 in the antioxidant group and 212 in the control group. There was a statistically signicant difference (MD 31.26, 95% CI 31.17 to 31.35; P < 0.00001, I2 = 96%) (Analysis 1.5) in favour of the antioxidant group over the control group (Figure 8). The heterogeneity with
the I2 statistic at 96% was very high. Sensitivity analysis was performed on the trial of Peivandi 2010 as the data in this trial were outlying. The standard deviations had unusually low values when compared to other trials. The trial by Peivandi also carried 99.8% of the weight in the analysis. A random-effects model was used and after sensitivity analysis the statistically signicant difference remained, in favour of the antioxidant group over the control (MD 11.72, 95%CI 6.94 to 16.49; P < 0.00001) and the I2 dropped to 56%. As heterogeneity remained high we were unable to pool these data and only subtotals were used in the analysis.
20
Figure 8. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.5 Sperm total motility at 3 months or less.
21
Total sperm motility at six months
Seven trials (Balercia 2005; Balercia 2009; Lenzi 2004; Safarinejad 2009; Safarinejad 2009a; Sigman 2006; Suleiman 1996) reported on total sperm motility at six months. In total there were 963 participants, 625 in the antioxidant group and 338 in the control group. This difference in numbers between the treatment and control groups was due to the splitting of the control arm of the three or four-arm trials. There was a statistically signicant differ-
ence (MD 4.19, 95% CI 3.81 to 4.56; P < 0.00001) (Analysis 1.6) in favour of the antioxidant group over the control group (Figure 9). Substantial heterogeneity was found in this analysis (Chi2 = 90.87, I2 = 89%). As heterogeneity was large, a sensitivity analysis was performed by removing the study with over 20% drop out (Suleiman 1996) but there was little change to the I2 statistic. Therefore, due to unexplained heterogeneity, the totals were removed in this analysis and only subtotals used.
22
Figure 9. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.6 Sperm total motility at 6 months.
23
Only one comparison (Balercia 2005; Lenzi 2004; Sigman 2006), L-carnitine + L-acetyl carnitine versus placebo at six months, had condence intervals that crossed the line of no effect. One comparison, (Balercia 2009; Safarinejad 2009a) coenzyme Q10 versus placebo at six months, carried 42.9% of the weight of this analysis and was in favour of the antioxidant group over the placebo group (MD 4.50, 95% CI 3.93 to 5.07; P < 0.00001, I2 = 0%). Three trials (Balercia 2005; Lenzi 2004; Sigman 2006) reported on total sperm motility with the comparison of L-carnitine + Lacetyl carnitine versus placebo at six months. In total there were 97 participants, 57 in the antioxidant group and 40 in the placebo group. There was a statistically signicant difference in the total sperm motility in favour of the antioxidant over the placebo group (MD 5.48, 95% CI 0.44 to 10.52; P = 0.03). Substantial heterogeneity was found in this analysis (Chi2 = 8.61, I2 = 77%).
Total sperm motility at nine months or more
Three trials (Balercia 2005; Balercia 2009; Safarinejad 2009a) reported on total sperm motility at nine months or more. In total there were 332 participants, 181 in the antioxidant and 151 in the placebo group. There was a statistically signicant difference in treatment effect in favour of the antioxidant over the control group (MD 1.38, 95% CI 0.81 to 1.95; P < 0.00001) (Analysis 1.7; Figure 10). Heterogeneity was found in this analysis (Chi2 = 10.98, I2 = 64%). Three comparisons had condence intervals that crossed the line of no effect. One comparison, coenzyme Q10 (Balercia 2005; Safarinejad 2009a) versus control at 56 weeks, carried 98.9% of the weight of this analysis and was in favour of the antioxidant group over the control group (MD 1.31, 95% CI 0.74 to 1.88; P < 0.00001). Heterogeneity was moderate in the analysis, probably due to the variation of antioxidants, therefore a random-effects model was used. Only subtotals were able to be used in the analysis.
Figure 10. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.7 Sperm total motility at 9 months or more.
24
Sperm concentration at three months or less
Seven trials ( Ciftci 2009; Conquer 2000; Greco 2005; Peivandi 2010; Rolf 1999; Wang 1983; Zavaczki 2003) reported on sperm concentration at three months or less. In total there were 320 participants, 162 in the antioxidant and 158 in the placebo group. There was no statistically signicant difference (MD 6.04, 95% CI -5.42 to 17.50; P = 0.30) (Analysis 1.8) between the antioxidant
and the placebo groups (Figure 11). Heterogeneity was high (I2 = 93%) in this analysis. A random-effects model was used and a sensitivity analysis performed for the trial Peivandi 2010. This trial had outlying data as the standard deviations were much lower than for the other studies. When this trial was removed from analysis the I2 statistic became 0% and there was no statistically signicant difference between the groups (MD 2.64, 95% CI -0.52 to 5.81; P = 0.10, I2 = 0%).
Figure 11. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.8 Sperm concentration at 3 months or less.
25
Two trials (Greco 2005; Rolf 1999) reported on sperm concentration with the comparison of vitamin C + vitamin E versus placebo at two months. In total there were 95 participants, 47 in the antioxidant group and 48 in the placebo group. A statistically significant difference between the antioxidant and placebo group could not be found (MD 1.36, 95% CI -10.01 to 12.72; P = 0.82). Moderate heterogeneity was found in this analysis (Chi2 = 2.07, I 2 = 52%).
Sperm concentration at six months
Six trials (Balercia 2005; Balercia 2009; Lenzi 2004; Safarinejad 2009; Safarinejad 2009a; Wang 1983) reported on sperm con-
centration at six months in eight different comparisons. In total there were 825 participants, 536 in the antioxidant and 289 in the placebo group. There was a statistically signicant difference in treatment effect in favour of the antioxidant over the placebo group (MD 5.25, 95% CI 4.43 to 6.08; P < 0.00001) (Analysis 1.9; Figure 12). Heterogeneity was moderate in this analysis (Chi 2 = 19.05, I2 = 53%). The raised heterogeneity was probably due to the variation in antioxidants. Five comparisons had condence intervals that crossed the line of no effect. Due to the heterogeneity a random-effects model was applied and subtotals were used in the analysis.
26
Figure 12. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.9 Sperm concentration at 6 months.
27
Two trials (Balercia 2005; Lenzi 2004) reported on sperm concentration with the comparison L-carnitine + L-acetyl carnitine versus placebo at six months. In total there were 76 participants, 45 in the antioxidant group and 31 in the placebo group. A statistically signicant difference between the antioxidant and placebo group could not be found (MD 0.63, 95% CI -5.93 to 7.18; P = 0.85). There was no heterogeneity in this analysis (Chi2 = 0.19, I2 = 0%). Two trials (Balercia 2009, Safarinejad 2009a) reported on sperm concentration with the comparison of coenzyme Q10 versus placebo at six months. In total there were 272 participants, 136 in the antioxidant group and 136 in the placebo group. There was statistically signicant difference in favour of the antioxidant over the placebo (MD 5.50, 95% CI 4.34 to 6.67; P < 0.00001). Heterogeneity was moderate (Chi2 = 1.92, I2 = 48%) in this analysis.
Sperm concentration at nine months or more
Three trials (Balercia 2005; Balercia 2009; Safarinejad 2009a) reported on sperm concentration at nine months or more in ve different comparisons. In total there were 332 participants, 181 in the antioxidant group and 151 in the placebo group. There was a statistically signicant difference in favour of the antioxidant over the placebo (MD 1.61, 95% CI 0.61 to 2.61; P = 0.002) (Analysis 1.10), see Figure 13. Heterogeneity could not be found in this analysis (Chi2 = 3.79, I2 = 0%). One trial (Safarinejad 2009a) carried 96.1% of the weight of this pooled analysis as it had a larger sample size and tighter condence intervals.
Figure 13. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.10 Sperm concentration at 9 months or more.
Two trials (Balercia 2009; Safarinejad 2009a) reported on sperm concentration with the comparison coenzyme Q10 versus placebo at nine months or more. A statistically signicant difference was seen in favour of the the antioxidant over the placebo group (MD
1.53, 95% CI 0.51 to 2.55; P = 0.003). One trial (Safarinejad 2009a) carried 96.1% of the weight of this analysis.
28
Total sperm motility and sperm concentration over time
Trials were also subgrouped according to sperm motility over time. A random-effects model was used due to high heterogeneity in the following analyses. At three months or less (MD 31.26, 95% CI 31.17 to 31.36; P < 0.00001) (Analysis 1.11; Figure 14) a statistically signicant difference in favour of the antioxidant, however the I2 statistic was 97%. When sensitivity analysis was performed for Peivandi 2009 (Peivandi 2010) due to outlying data the heterogeneity remained high at 71%. Figure 14. Forest plot of comparison: 1 Antioxidant(s) versus control, outcome: 1.17 Total sperm motility over time.
29
At six months there was a statistically signicant difference in treatment effect in favour of the antioxidant over the control group (MD 5.34, 95% CI 4.91 to 5.77; P < 0.00001) and the I2 statistic was 84%. Sensitivity analysis was performed for Sulieman (Suleiman 1996) due to >20% drop outs however there was only a slight decrease in heterogeneity. At nine months there was also a statistically signicant difference in treatment effect in favour of the antioxidant over the placebo (MD 1.35, 95% CI 0.78 to 1.92; P<0.00001, I2 = 68%).
Sperm concentration over time: at three months or less (MD 12.74, 95% CI 10.23 to 15.26; P < 0.00001) (Analysis 1.12; Figure 15) a statistically signicant difference in favour of the antioxidant over the control however the I2 statistic was 94%. When sensitivity analysis was performed for Peivandi 2009 (Peivandi 2010) due to outlying data the I2 statistic became 0% and the result changed to become not statistically signicant (MD 2.70, 95% CI -0.47 to 5.88; P = 0.10);
Figure 15. Forest plot of comparison: 1 Antioxidant(s) versus control, outcome: 1.18 Sperm concentration over time.
at six months (MD 6.32, 95% CI 5.40 to 7.24; P < 0.00001) a statistically signicant difference of the antioxidant
over the control however the I2 statistic = 69%; at nine months or more (MD 1.56, 95% CI 0.55 to 2.57; P
30
= 0. 003, I2 = 0%) a statistically signicant difference in favour of the antioxidant over the control.
Side effects
Gastrointestinal side effects
Six trials (Cavallini 2004; Kessopoulou 1995; Safarinejad 2009a; Sigman 2006; Tremellen 2007; Zavaczki 2003) reported on gastrointestinal side effects with the comparison of antioxidants versus placebo. In total there were 426 participants, 219 in the antioxidant group and 207 in the placebo group. A statistically signicant difference between the antioxidant and placebo groups could not be found (pooled OR 1.63, 95% CI 0.48 to 5.58; P = 0.44) (Analysis 1.13; Figure 16). There was no heterogeneity (Chi 2 = 2.49, I2 = 0%).
Figure 16. Forest plot of comparison: 1 Antioxidant(s) versus placebo, outcome: 1.11 Side effects.
Euphoria One trial (Cavallini 2004) reported on euphoria as a side effect with the comparison of antioxidants versus placebo. A statistically signicant difference between the antioxidant and placebo group could not be found (OR 1.21, 95% CI 0.16 to 9.01; P = 0.85) ( Analysis 1.13; Figure 16). Antioxidant(s) versus antioxidant(s) - head to head The total summary statistic was disabled for this analysis as there was no reason to pool direct comparisons.
Pregnancy rate per couple randomised
Two trials (Balercia 2005; Li 2005 ) reported on pregnancy rates with three different comparisons (see Analysis 3.1; Figure 17). All three comparisons crossed the line of no effect. There were 10 pregnancies form 90 couples (11%) in the L-acetyl carnitine + Lcarnitine group and 2 pregnancies from 60 couples in the control group (3%). The controls were L-acetyl carnitine, carnitine and vitamin E + vitamin C.
31
Figure 17. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.1 Pregnancy per couple randomised.
One trial (Balercia 2005) reported on the outcome of pregnancy rates with the comparison of L-acetyl carnitine + L-carnitine (treatment) versus L-acetyl carnitine (control). There were two pregnancies in the L-acetyl carnitine + L carnitine group and two in the L-acetyl carnitine. . Balercia (Balercia 2005) also reported on the outcome of pregnancy rate with the comparison of L-acetyl carnitine + L-carnitine versus L-carnitine. There were three pregnancies in the treatment group and two in the control group.
Sperm total motility at three months or less
Seven trials (Akiyama 1999; Conquer 2000; Dawson 1990; Keskes-Ammar 2003; Li 2005; Omu 2008; Scott 1998) reported on sperm motility at three months or less in nine different comparisons (Analysis 3.2) and (Figure 18).
32
Figure 18. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.2 Sperm total motility at 3 months or less.
33
Only one trial (Dawson 1990) showed a statistically signicant difference in measure of effect. This study compared ascorbic acid (vitamin C) 200 mg/day versus vitamin C 1000 mg/day. There was a statistically signicant difference (MD -43.00, 95% CI 67.10 to -18.90; P = 0.0005) in favour of the 1000 mg/day in sperm motility.
Sperm total motility at six months
Two trials (Balercia 2005, Safarinejad 2009) reported on ve different comparisons of antioxidants versus antioxidants (Analysis 3.3; Figure 19). Figure 19. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.3 Sperm total motility at 6 months.
The comparisons of L-acetyl carnitine + L-carnitine versus L-carnitine (Balercia 2005) (MD -3.46, 95% CI -11.41 to 4.49; P = 0.39) and L-acetyl carnitine + L-carnitine versus L-acetyl carnitine (Balercia 2005) (MD 0.64, 95% CI -7.58 to 8.86; P = 0.88) showed no statistically signicant difference in effect between these
two different antioxidants. Safarinejad (Safarinejad 2009) also reported on three other comparisons. Selenium versus N-acetyl cysteine, with a statistically signicant
34
difference in favour of selenium (MD 1.30, 95% CI 0.39 to 2.21; P = 0.005). Selenium versus selenium + N-acetyl cysteine with a statistically signicant difference in favour of selenium + N-acetyl cysteine over selenium alone (MD -3.10, 95% CI -4.16 to -2.04; P < 0.00001). N-acetyl cysteine versus selenium + N-acetyl cysteine with a statistically signicant difference in favour of selenium plus N-acetyl cysteine (MD -4.40, 95% CI -5.31 to -3.49; P < 0.00001).
Sperm total motility at nine months
Only one trial (Balercia 2005) reported on sperm motility at nine months using the two comparisons of L-acetyl carnitine + L-carnitine versus L-carnitine (MD -5.27, 95% CI -12.61 to 2.07; P = 0.16) and L-acetyl carnitine + L-carnitine versus L-acetyl carnitine (MD -1.57, 95% CI -7.70 to 4.56; P = 0.62) with no evidence of effect in either of these comparisons (Analysis 3.4; Figure 20).
Figure 20. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.4 Sperm total motility at 9 months or more.
Sperm concentration at three months
Three trials (Akiyama 1999; Conquer 2000; Li 2005a) reported on sperm concentration at three months (Analysis 3.5; Figure 21). The trial by Akiyama showed no statistically signicant difference in effect between the antioxidant ethyl cysteine versus vitamin E (MD 2.20, 95% CI -16.65 to 21.05; P = 0.82). The trial by Conquer compared docosahexaenoic acid (DHA) 400g/day and DHA 800 g/day also with no statistically signicant difference in effect (MD -6.80, 95% CI -41.87 to 28.27; P = 0.70). The trial by Li 2005a reported on sperm concentration at three months comparing L-carnitine versus vitamin E + vitamin C with a statistically signicant difference (MD 15.50, 95% CI 12.49 to 18.51; P < 0.00001) in favour of vitamin E + vitamin C over L-carnitine.
35
Figure 21. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.5 Sperm concentration at 3 months or less.
Sperm concentration at six months
Two trials (Balercia 2005; Safarinejad 2009) reported on sperm concentration at six months using ve comparisons (Analysis 3.6; Figure 22).
36
Figure 22. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.6 Sperm concentration at 6 months.
The trial by Balercia reported on two comparisons for this outcome: L-acetyl carnitine + L-carnitine versus L-carnitine (MD 8.13, 95% CI -24.42 to 8.16; P = 0.33) and L-acetyl carnitine + L-carnitine versus L-acetyl carnitine (MD -2.17, 95% CI -17.40 to 13.06; P = 0.78). Neither of these comparisons showed any statistically signicant difference in effect between the antioxidants. Safarinejad 2009 reported on three comparisons. Selenium versus N-acetyl cysteine showed no statistically signicant difference in effect. Selenium versus selenium + N-acetyl cysteine showed a statistically signicant difference favouring selenium + N-acetyl cysteine (MD -4.50, 95%CI -6.90 to -2.10; P < 0.0002).
Safarinejad 2009 also reported on N-acetyl cysteine versus selenium + N-acetyl cysteine. There was a statistically signicant difference (MD -5.30, 95% CI -7.29 to -3.31; P < 0.00001) in favour of selenium + N-acetyl cysteine.
Sperm concentration at nine months
Balercia (Balercia 2005) reported on the outcome of sperm concentration at nine months with two comparisons (Analysis 3.7; Figure 23).
37
Figure 23. Forest plot of comparison: 3 Antioxidant(s) versus antioxidant(s), outcome: 3.7 Sperm concentration at 9 months or more.
L-acetyl carnitine + L-carnitine versus L-carnitine showed no statistically signicant difference in effect between the two antioxidants (MD -6.13, 95% CI -18.44 to 6.18; P = 0.33). L-acetyl carnitine + L-carnitine versus L-acetyl carnitine also showed no statistically signicant difference in effect between the two antioxidants (MD 2.06, 95% CI -8.33 to 12.45; P = 0.70).
DISCUSSION Summary of main results

This systematic review and meta analysis aimed to investigate whether supplemental antioxidants given to subfertile men made a difference to the outcomes of live birth, pregnancy and the adverse events of miscarriage or stillbirth. Seminal parameters were also analysed. The comparisons used were antioxidants versus control and head to head antioxidants. In the antioxidant versus control comparison there was a statistically signicant difference in effect in favour of antioxidants in the live birth and pregnancy outcomes. These antioxidants did not appear to be associated with any increase in miscarriage. There were however only three trials (Omu 1998; Suleiman 1996; Tremellen 2007) in the outcome of live birth. One of these trials Suleiman 1996 had drop outs. The reasons were explained but we were unaware of how many of the drop outs were from the treatment or control group. Therefore we could not use intention-to-treat analysis for this trial. Fifteen trials reported on pregnancy rate (Balercia 2005; Balercia 2009; Cavallini 2004; Galatioto 2008; Lenzi 2003; Lenzi 2004; Omu 1998; Peivandi 2010; Rolf 1999; Safarinejad 2009a; Sigman 2006; Suleiman 1996; Tremellen 2007; Wang 1983; Zavaczki 2003) in the antioxidant versus control comparison and two (
Balercia 2005; Li 2005) in the head to head comparisons. Balercia 2005 was multi-armed and also reported pregnancy rate per couple using the antioxidant versus control comparison. There were no trials reporting on the outcome of stillbirth. The data for the outcomes of total sperm motility at three, six and nine months were unable to be pooled due to heterogeneity. A random-effects model was used and sensitivity analysis was performed. Heterogeneity remained moderate therefore we were not able to pool this data and subtotals were used. Sperm concentration at three months or less had high heterogeneity and did not show a statistically signicant difference in effect after sensitivity analysis was performed. The I2 statistic became 0% in the sensitivity analysis and therefore the data were able to be pooled. The analysis of the outcome of sperm concentration at six months also showed high heterogeneity. A random-effects model was used and subtotals were used as the data were unable to be pooled. The analysis of sperm concentration at nine months showed a statistically signicant difference in effect in favour of the antioxidants over the control group. No heterogeneity was evident in this analysis. The high heterogeneity in the analysis of the sperm parameters may in some way be due to the variations of the types of antioxidants. The main side effects reported were gastrointestinal and there was no evidence that the antioxidant was more harmful than the placebo. It is difcult to draw conclusions from the head to head comparisons until there are more randomised trials that compare the same antioxidants against controls. One antioxidant did not appear to have any effect on pregnancy rate per couple or sperm parameters over those of another antioxidant.
Overall completeness and applicability of evidence
38
The live birth analysis in the antioxidant versus control group was based on only three trials (Kessopoulou 1995; Omu 1998; Suleiman 1996). The Suleiman trial had some imbalance in participant numbers due to loss to follow up. The analysis remained statistically signicant when sensitivity analysis was performed on the Suleiman trial. Pregnancy rate evidence between antioxidants and control is incomplete. Only 15 of 34 trials reported this outcome. Several of these trials had high and selective losses to follow up. However heterogeneity between the trials is low. The outcomes of sperm motility and concentration also appear to be incomplete and not applicable. Only 15 of 34 trials reported on sperm motility and 16 reported on sperm concentration Two trials (Galatioto 2008; Tremellen 2007) used combined antioxidants (three or more antioxidants) versus control. Only one trial (Greco 2005) reported DNA fragmentation in the comparison of antioxidants versus control therefore there were no pooled data available for this outcome. The adverse events of stillbirth, miscarriage and side effects appear to be poorly reported. The head to head comparison does not provide constructive information as there was no reason to pool direct comparisons. Subgrouping of antioxidants or different doses of antioxidants was un-
able to be performed in the treatment versus treatment groups as there were only single trials analysing these differences. Therefore this review was unable to show any difference in effect between different antioxidants or different doses of the same antioxidant.
Quality of the evidence

All three studies reporting live birth (Kessopoulou 1995; Omu 1998; Suleiman 1996) had unclear methods of sequence generation and allocation concealment; one trial (Suleiman 1996) also had some imbalance in numbers at analysis due to drop outs and in another (Omu 1998) the comparison was no treatment and not placebo. Figure 23 (Figure 24) shows the review authors judgements about the methodological quality of the trials included in this review. All included trials were described as randomised however only 30% gave information on how the randomisation was achieved. Allocation concealment was described in only 24% of the trials. Blinding was better described with over 50% of the trials being double and occasionally single blinded; 10% of trials stated that there was no blinding. Incomplete outcome data were addressed by most of the studies but selective reporting of outcomes was unclear in the majority of trials.
Figure 24. Methodological quality graph: review authors judgements about each methodological quality item presented as percentages across all included studies.
The seven trials that described their methods more thoroughly (Galatioto 2008; Greco 2005; Peivandi 2010; Rolf 1999; Safarinejad 2009; Sigman 2006; Tremellen 2007) can be seen graphically in gure 24 (Figure 25).
39
Figure 25. Methodological quality summary: review authors judgements about each methodological quality item for each included study.
40
Potential biases in the review process

Some bias in the review process may have arisen due to the inclusion of trials that have had drop outs of participants of > 20% and imbalances between the treatment and control groups.
conceiving. More research is required to further substantiate these conclusions.
Implications for research

In this review there were only three trials (two comparing antioxidant versus placebo and one versus no treatment) reporting on live birth, the most important outcome from the perspective of the couple experiencing difculty with conception. Only one trial reported on DNA fragmentation. A low degree of DNA fragmentation is thought to increase a the likelihood of achieving a pregnancy. More well designed randomised controlled trials that report on the outcomes of live birth, pregnancy, DNA fragmentation and adverse events are needed. Only two trials (Galatioto 2008; Tremellen 2007) used combined antioxidants (three or more antioxidants) versus control. The results were in favour of the antioxidant over the control however there is a need for more randomised controlled trials on this intervention in order to make any conclusions on whether a combination of antioxidants would have statistical signicant difference over a single antioxidant versus placebo. Further randomised controlled trials are needed to assess whether one antioxidant is more effective than another by head to head comparisons. There is also a gap in the evidence for the assessment of different doses of an antioxidant having differing effects. This review was only able to include single trials measuring different doses and therefore meta analysis of this comparison was unable to be performed. Evidence to date suggests that the side effect prole is low however, as there were few studies, more data are required to determine any adverse events and the side effect prole of these supplements.
Agreements and disagreements with other studies or reviews

Two reviews described the effects of L-carnitine and L-acetylcarnitine on subfertile men. The systematic review and meta-analysis by Zhou (Zhou 2007) compared L-carnitine and L-acetylcarnitine therapy to placebo treatment and found signicant improvements in pregnancy rate and total sperm motility. No signicant difference was found in the outcome of sperm concentration. The descriptive review by Patel (Patel 2008) discusses the improvement in pregnancy rates with oral intake of antioxidants, however Patel states that randomised controlled trials (RCTs) have not shown an effect on sperm motility and that there is a need for more RCTs in men with oxidative stress. These results were similar for the trials reporting on L-carnitine and acetyl carnitine interventions in this review. Agarwal (Agarwal 2004) in an overview of the literature discusses a range of antioxidants, and combinations of these, and their effect on male subfertility. Agarwal notes that vitamin E and a combination of vitamin E with other antioxidants such as N-acetylcysteine, vitamin A and fatty acids appears to improve pregnancy rates in asthenozoospermic men. Carntines also appear to have an effect on pregnancy rates. This review agrees with the carnitine discussion but there were only single trials in the vitamin E interventions versus placebo so a summary statistic for the outcome was unable to be obtained. Another review that was published recently (Ross 2010) also showed improvement in pregnancy rate after antioxidant therapy. A Cochrane review of antioxidants for female subfertility has been published concurrently (Clarke 2009). The ndings of both reviews are consistent in that combined antioxidants improved pregnancy rate.
ACKNOWLEDGEMENTS
Cochrane Menstrual Disorders and Subfertility Group (MDSG). I would like to make special mention of the MDSG editors who were very thorough and helpful in editing this review.
AUTHORS CONCLUSIONS Implications for practice

When trying to conceive as part of an assisted reproductive program, it may be advisable to encourage the male partner to take an oral antioxidant supplement to improve his partners chance of
Many thanks to the translators of the non-English trials - Ichiro Omori (Japanese paper), Shaofu Li (two Chinese papers), Ivan Sola (Bulgarian paper) and Pawel Kanturski (Polish paper) and Dr Peviandi (Iranian paper). Thanks also to Stephan Bontekoe who kindly helped with some of the text.
41
REFERENCES
References to studies included in this review

Akiyama 1999 {published data only} Akiyama M. In vivo scavenging effect of ethylcysteine on reactive oxygen species in human semen. Japanese Journal of Urology 1999; Vol. 90, issue 3:4218. Balercia 2005 {published data only} Balercia G, Regoli F, Armeni T, Koverech A, Mantero F, Boscaro M. Placebo-controlled double-blind randomized trial on the use of L-carnitine, L-acetylcarnitine, or combined L-carnitine and L-acetylcarnitine in men with idiopathic asthenozoospermia. Fertility and Sterility 2005; Vol. 84, issue 3:66271. Balercia 2009 {published data only} Balercia G, Buldreghini E, et al.Coenzyme Q10 treatment in infertile men with idiopathic asthenozoospermia: a placebo-controlled, double-blind randomized trial. Fertility and Sterility 2009;91(5):178592. Biagiotti 2003 {published data only} Biagiotti G, Cavallini G, Modenini F, Vitali G, Magli C, Ferraretti A. Prostaglanadins pulsed down-regulation enhances carnitine therapy performance in severe idiopathic oligoasthenospermia. Human Reproduction 2003; Vol. 18 Suppl 1:202. Cavallini 2004 {published data only} Cavallini G, Ferraretti AP, Gianaroli L, Biagiotti G, Vitali G. Cinnoxicam and L-carnitine/acetyl-L-carnitine treatment for idiopathic and varicocele-associated oligoasthenospermia [see comment]. Journal of Andrology 2004; Vol. 25, issue 5:76170. Ciftci 2009 {published data only} Ciftci H, Verit A, Savas M, Yeni E, Erel O. Effects of N-acetylcysteine on semen parameters and oxidative/ antioxidant status. Urology 2009;74(1):736. Conquer 2000 {published data only} Conquer JA, Martin JB, Tummon I, Watson L, Tekpetey F. Effect of DHA supplementation on DHA status and sperm motility in asthenozoospermic males. Lipids 2000;35(2): 14954. Dawson 1990 {published data only} Dawson EB, Harris WA, Powell LC. Relationship between ascorbic acid and male fertility. World Review of Nutrition and Dietetics 1990; Vol. 62:126. Galatioto 2008 {published data only} Galatioto GP, Gravina GL, Angelozzi G, Sacchetti A, Innominato PF, Pace G, et al.May antioxidant therapy improve sperm parameters of men with persistent oligospermia after retrograde embolization for varicocele?. World Journal of Urology 2008; Vol. 26, issue 1:97102. Greco 2005 {published data only} Greco E, Iacobelli M, Rienzi L, Ubaldi F, Ferrero S, Tesarik J. Reduction of the incidence of sperm DNA fragmentation by oral antioxidant treatment. Journal of Andrology 2005; Vol. 26, issue 3:34953.
Keskes-Ammar 2003 {published data only} Keskes-Ammar L, Feki-Chakroun N, Rebai T, Sahnoun Z, Ghozzi H, Hammami S, et al.Sperm oxidative stress and the effect of an oral vitamin E and selenium supplement on semen quality in infertile men. Archives of Andrology 2003; Vol. 49, issue 2:8394. Kessopoulou 1995 {published data only} Kessopoulou E, Powers HJ, Sharma KK, Pearson MJ, Russell JM, Cooke ID, et al.A double-blind randomized placebo cross-over controlled trial using the antioxidant vitamin E to treat reactive oxygen species associated male infertility. Fertility and Sterility 1995; Vol. 64, issue 4: 82531. Lenzi 2003 {published data only} Lenzi A, Lombardo F, Sgro P, Salacone P, Caponecchia L, Dondero F, et al.Use of carnitine therapy in selected cases of male factor infertility: a double-blind crossover trial [see comment]. Fertility and Sterility 2003; Vol. 79, issue 2: 292300. Lenzi 2004 {published data only} Lenzi A, Sgro P, Salacone P, Paoli D, Gilio B, Lombardo F, et al.A placebo-controlled double-blind randomized trial of the use of combined l-carnitine and l-acetyl-carnitine treatment in men with asthenozoospermia [see comment]. Fertility and Sterility 2004; Vol. 81, issue 6:157884. Li 2005 {published data only} Li Z, Chen GW, Shang XJ, Bai WJ, Han YF, Chen B, et al.A controlled randomized trial of the use of combined L-carnitine and acetyl-L-carnitine treatment in men with oligoasthenozoospermia. Zhong Hua Nan Ke Xue 2005; Vol. 11, issue 10:7614. Li 2005a {published data only} Li Z, Gu R, Liu Y, Xiang Z, Cao X, Han Y, et al.Curative effect of L-carnitine supplementation in the treatment of male infertility. Academic Journal of Shanghai Second Medical University 2005; Vol. 25, issue 3:2924. Lombardo 2002 {published data only} Lombardo F, Gandini L, Agarwal A, Sgro P, Dondero F, Lenzi A. A prospective double blind placebo controlled cross over trial of carnitine therapy in selected cases of male infertility. Fertility and Sterility 2002; Vol. 78 Suppl 1: 689. Merino 1997 {published data only} Merino G, Martinez Chequer JC, Barahona E, Bermudez JA, Moran C, Carranza-Lira S. Effects of pentoxifylline on sperm motility in normogonadotropic asthenozoospermic men. Archives of Andrology 1997; Vol. 39, issue 1:659. Micic 1988 {published data only} Micic S, Hadzi-Djokic J, Dotlic R, Tulic C. Pentoxifyllin treatment of oligoasthenospermic men. Acta Europaea Fertilitatis 1988; Vol. 19, issue 3:1357. Nozha 2001 {published data only} Nozha CF, Leila AK, Zouhir S, Hanen G, Khled Z, Tarek R. Oxidative stress and male infertility: comparative study
42
of combined vitamin E/selenium treatment versus vitamin B. Human Reproduction 2001; Vol. 17, issue Abstract book 1:1112. Omu 1998 {published data only} Omu AE, Dashti H, Al-Othman S. Treatment of asthenozoospermia with zinc sulphate: andrological, immunological and obstetric outcome. European Journal of Obstetrics, Gynecology, and Reproductive Biology 1998; Vol. 79, issue 2:17984. Omu 2008 {published data only} Omu AE, Al-Azemi MK, Kehinde EO, Anim JT, Oriowo MA, Mathew TC. Indications of the mechanisms involved in improved sperm parameters by zinc therapy. Medical Principles and Practice 2008; Vol. 17, issue 2:10816. Peivandi 2010 {published data only} Peivandi S, Abasali K, Narges M. Effects of L-carnitine on infertile mens spermogram; a randomised clinical trial. Journal of Reproduction and Infertility 2010;10(4):331. Rolf 1999 {published data only} Rolf C, Cooper TG, Yeung CH, Nieschlag E. Antioxidant treatment of patients with asthenozoospermia or moderate oligoasthenozoospermia with high-dose vitamin C and vitamin E: a randomized, placebo-controlled, double-blind study [see comment]. Human Reproduction 1999; Vol. 14, issue 4:102833. Safarinejad 2009 {published data only} Safarinejad MR, Safarinejad S. Efcacy of selenium and/ or N-acetyl-cysteine for improving semen parameters in infertile men: a double-blind, placebo controlled, randomized study. The Journal of Urology 2009;181(2): 74151. Safarinejad 2009a {published data only} Safarinejad ME. Efciacy of coenzyme Q10 on semen parameters, sperm function and reproductive hormones in infertile men. The Journal of Urology 2009;182:23748. Scott 1998 {published data only} Scott R, MacPherson A, Yates RW, Hussain B, Dixon J. The effect of oral selenium supplementation on human sperm motility. British Journal of Urology 1998; Vol. 82, issue 1: 7680. Sigman 2006 {published data only} Sigman M, Glass S, Campagnone J, Pryor JL. Carnitine for the treatment of idiopathic asthenospermia: a randomized, double-blind, placebo-controlled trial. Fertility and Sterility 2006; Vol. 85, issue 5:140914. Suleiman 1996 {published data only} Suleiman SA, Ali ME, Zaki ZM, el-Malik EM, Nasr MA. Lipid peroxidation and human sperm motility: protective role of vitamin E. Journal of Andrology 1996; Vol. 17, issue 5:5307. Tremellen 2007 {published data only} Tremellen K, Miari G, Froiland D, Thompson J. A randomised control trial examining the effect of an antioxidant (Menevit) on pregnancy outcome during IVFICSI treatment. The Australian & New Zealand Journal of Obstetrics & Gynaecology 2007;47:21621.
Wang 1983 {published data only} Wang C, Chan CW, Wong KK, Yeung KK. Comparison of the effectiveness of placebo, clomiphene citrate, mesterolone, pentoxifylline, and testosterone rebound therapy for the treatment of idiopathic oligospermia. Fertility and Sterility 1983;40(3):35865. Wong 2002 {published data only} Wong WY, Merkus HM, Thomas CM, Menkveld R, Zielhuis GA, Steegers-Theunissen RP. Effects of folic acid and zinc sulfate on male factor subfertility: a double-blind, randomized, placebo-controlled trial. Fertility and Sterility 2002; Vol. 77, issue 3:4918. Zalata 1998 {published data only} Zalata A, Christophe A, Horrobin D, Dhooge W, Comhaire F. Effect of essential fatty acids and antioxidants dietary supplementation on the oxidative DNA damage of the human spermatozoa. ESHRE 14th annual meeting. Goteborg, 1998:2701. Zavaczki 2003 {published data only} Zavaczki Z, Szollosi J, Kiss S, Koloszar S, Fejes I, Kovacs L, Pal A. Magnesium-orotate supplementation for idiopathic infertile male patients: a randomized, placebo-controlled clinical pilot study. Magnesium Research 2003;16(2):1316.
References to studies excluded from this review

Cavallini 2004a {published data only} Cavallini G, Ferraretti AP, Gianaroli L, Magli MC, Biagiotti G, Vitali G. Cinnoxicam + carnitines for idiopathic dyspermia. The 20th Annual Meeting of the European Society of Human Reproduction and Embryology 2004: i234. Comhaire 2005 {published data only} Comhaire FH, El Garem Y, Mahmoud A, Eertmans F, Schoonjans F. Combined conventional/antioxidant Astaxanthin treatment for male infertility: a double blind, randomized trial. Asian Journal of Andrology 2005; Vol. 7, issue 3:25762. Ebisch 2003 {published data only} Ebisch IM, van Heerde WL, Thomas CM, van der Put N, Wong WY, Steegers-Theunissen RP. C677T methylenetetrahydrofolate reductase polymorphism interferes with the effects of folic acid and zinc sulfate on sperm concentration. Fertility and Sterility 2003; Vol. 80, issue 5:11904. Ebisch 2006 {published data only} Ebisch IM, Pierik FH, DE Jong FH, Thomas CM, Steegers-Theunissen RP. Does folic acid and zinc sulphate intervention affect endocrine parameters and sperm characteristics in men?. International Journal of Andrology 2006; Vol. 29, issue 2:33945. Elgindy 2008 {published data only} Elgindy EA, El-Huseiny AM, Mostafa MI, Gaballah AM, Amed TA. N-Acetylcysteine: Could it be an effective adjuvant therapy in ICSI cycles. Fertility and Sterility 2008;90 Suppl 1(American Society for Reproductive
43
Medicine 64th Annual Meeting. November 8-12, 2008, San Francisco CA):356 Abstract no: P738. Micic 2001 {published data only} Micic S, Lalic N, Bojanic N, Nale DJ. Oligospermic men treated by carnitine. 17th World Congress on Fertility and Sterility. Melbourne, 2001:38. Nikolova 2007 {published data only} Nikolova V, Stanislavov R, Vatev I, Nalbanski B, Punevska M. Sperm parameters in male idiopathic infertility after treatment with prelox. Akusherstvo i Ginekologiia 2007; Vol. 46, issue 5:712. Pawlowicz 2001 {published data only} Pawlowicz P, Stachowiak G, Bielak A, Wilczynski J. Administration of natural anthocyanins derived from chokeberry (aronia melanocarpa) extract in the treatment of oligospermia in males with enhanced autoantibodies to oxidized low density lipoproteins (oLAB). The impact on fructose levels. Ginekologia Polska 2001;72(12):9838. Pryor 2003 {published data only} Pryor JL, Stacy L, Glass S, Campagnone J, Sigman M. Randomized double blind placebo controlled trial of carnitine for the treatment of idiopathic asthenospermia. Fertility and Sterility 2003; Vol. 80 Suppl 3:48. Rolf 1998 {published data only} Rolf C. Antioxidant treatment with high dose vitamin C and vitamin E of patients with asthenozoospermia a randomized placebo controlled double blind study. ESHRE 14th annual meeting. Goteborg, 1998:72. Stanislavov 2009 {published data only} Stanislavov R, Nikolova V, Rohdewald P. Improvement of seminal parameters with Prelox : a randomized, doubleblind, placebo controlled, cross over trial. Phytotherapy Research 2009;23:297302. [DOI: 10.1002] Tremellen 2006 {published data only} Tremellen K, Froiland D, Miari G, Thompson J. A randomised control trial examining the effect of an antioxidant on sperm function pregnancy outcome during IVF treatment. The Fertility Society of Australia 25th Annual scientic meeting 2006 2006; Vol. 46 Suppl 2:A2. Vicari 2001 {published data only} Vicari E, Calogero AE. Effects of treatment with carnitines in infertile patients with prostatovesiculoepididymitis and elevated sperm oxidative stress. ESHRE 17th Annual Meeting. Lausanne: Human Reproduction, 2001; Vol. 16 Suppl 1:1067. Vicari 2001a {published data only} Vicari E, Rubino C, De Palma A, Longo G, Lauretta M, Consoli S, Arancio A. Antioxidant therapeutic efciency after the use of carnitine in infertile patients with bacterial or non bacterial prostato-vesiculo-epididymitis. Archivio Italiano di Urologia, Andrologia 2001;73(1):1525. Zalata 1998a {published data only} Zalata A, Christophe A, Horrbin D, Dhooge W, Comhaire F. Protection of sperm function and DNA by essential fatty
acids and antioxidants dietary supplementation. Fertility and Sterility 1998;70(3 Abstracts of 1998 Meetings):287.
References to studies awaiting assessment

Dimitriadis 2010 {published data only} Dimitriadis F, Tsambalas S, Tsounapi P, Kawamura H, Vlachopoulou E, Haliasos N, Gratsias S, Watanabe T, Saito M, Miyagawa I, Sokitis N. Effects of phosphodiesterase5 inhibitors on Leydig cell secretory function in oligoasthenospermic infertile men: A randomized trial. BJU International 2010;106(8):11811185. Wang 2010 {published data only} Wang YW, Yang SW, Qu CB, Huo HX, Li W, Li JD, Chang XL, Cai GZ. L- carnitine: safe and effective for asthenozoospermia. National Journal of Andrology 2010;16 (5):4202.
References to ongoing studies

Gonzalez 2009 {unpublished data only} Gonzalez Sanchez R. Assessment of the efcacy of dietary supplement spermotrend in the treatment of male infertility. ClinicalTrials.gov. Kumar 2010 {unpublished data only} Kumar R. Evaluation of oxidative stress as a cause for idiopathic male infertility and the role of addyzoa. ClinicalTrials.gov. Revel 2006 {published data only} Revel A. A preliminary study on effect of omega-3 on human sperm. ClinicalTrials.gov. [CTG: NCT00479960] Sadeghi 2008 {published data only} Sadeghi M. Effects of coenzyme Q10 (CoQ10) supplementation on semen quality and seminal oxidative stress of idiopathic oligoasthenoteratozoospermic (iOAT) infertile men. World Health Organization International Clinical Trials Registry Platform Search Portal 2008. [ISRCTN: ISRCTN29954277] Tsafrir 2010 {published data only} Tsafrir A. Is a carnitine based food supplement (PorimoreTM) for infertile men superior to folate and zinc with regard to pregnancy rates in intrauterine insemination cycles?. Information obtained from ClinicalTrials.gov on March 01, 2010.
Additional references
Agarwal 2003 Agarwal A, Said T. Role of sperm chromatin abnormalities and DNA damage in male infertility. Human Reproduction Update 2003;9(4):33145. Agarwal 2004 Agarwal A, Nallella K, Allamaneni S, Said T. Role of antioxidants in treatment of male infertility: an overview of the literature. Reproductive Biomedicine Online 2004;8(6): 61627.
44
Aitken 2004 Aitken RJ, Koopman P, Lewis SE. Seeds of concern. Nature 2004;432(7013):4852. Aitken 2007 Aitken RJ, De Iuliis GN. Origins and consequences of DNA damage in male germ cells. Reproductive Biomedicine Online 2007;14(6):72733. Alvarez 2003 Alvarez 2003. Nuture vs nature: How can we optimise sperm quality?. Journal of Andrology 2003;24(5):6408. Attia 2007 Attia AM, Al-Inany HG, Farquar C, Proctor M. Gonatrophins for idiopathic male factor male subfertility. Cochrane Database of Systematic Reviews 2004, Issue 4. [DOI: 10.1002/14651858.CD005071.pub3; : ] Boe-Hansen 2005 Boe-Hansen GB, Ersboll AK, Christensen P. Variability and laboratory factors affecting the sperm chromatin structure assay in human semen. Journal of Andrology 2005;26(3): 3608. Boe-Hansen 2006 Boe-Hansen GB, Fedder J, Ersboll AK, Christensen P. The sperm chromatin structure assay as a diagnostic tool in the human fertility clinic. Human Reproduction 2006;21(6): 157682. Borini 2006 Borini A, Tarozzi N, Bizzaro D, Bonu MA, Fava L, Flamigni C, Coticchio G. Sperm DNA fragmentation: paternal effect on early post-implantation embryo development in ART. Human Reproduction 2006;21(11):287681. Bykova 2007 Bykova M, Athayde K, Sharma R, Jha R, Sabanegh E, Agarwal A. Dening the reference value of seminal reactive oxygen species in a population of infertile men and normal healthy volunteers. Fertility and Sterility 2007;88 Suppl 1 (P-597):305. Clarke 2009 Clarke J, Showell MG, Hart RJ, Agarwal A, Gupta S. Antioxidants for female subfertility. Cochrane Database of Systematic Reviews 2009, Issue 2. [DOI: 10.1002/ 146518858] Comhaire 2000 Comhaire FH, Christophe AB, Zalata AA, Dhooge WS, Mahmoud AM, Depuydt CE. The effects of combined conventional treatment, oral antioxidants and essential fatty acids on sperm biology in subfertile men. Prostaglandins, Leukotrienes, and Essential Fatty Acids 2000;63(3):15965. Dias 2006 Dias S, McNamee R, Vail A. Evidence of improving quality of reporting of randomised controlled trials in subfertility. Human Reproduction 2006;21(10):261727. El-Taieb 2009 El-Taieb MAA, Herwig R, Nada EA, Greilberger J, Marberger M. Oxidative stress and epididymal sperm
transport, motility and morphological defects. European Journal of Obstetrics, Gynecology, and Reproductive Biology 2009;144 Suppl 1:199203. Eskenazi 2005 Eskanazi B, Kidd SA, Marks AR, Sloter E, Block G, Wyrobeck AJ. Antioxidant intake is associated with semen quality in healthy men. Human Reproduction 2005;20(4): 100612. Evenson 2006 Evenson D, Wixon R. Meta-analysis of sperm DNA fragmentation using the sperm chromatin structure assay. Reproductive Biomedicine Online 2006;12(4):46672. Evenson 2007 Evenson DP, Kasperson K, Wixon RL. Analysis of sperm DNA fragmentation using ow cytometry and other techniques. Society for Reproduction and Fertility 2007;65 Suppl:93113. Higgins 2006 Higgins JPT, Green S, editors. Cochrane Handbook for Systematic Reviews of Interventions 4.2.6 (updated Sept 2006). The Cochrane Library Issue 4 Sept 2006. Chichester UK: John Wiley and Sons Ltd, 2006. Li 2006 Li Z, Wang L, Cai J, Huang H. Correlation of sperm DNA damage with IVF and ICSI outcomes: a systematic review and meta-analysis. Journal of Assisted Reproduction and Genetics 2006;23(9-10):36776. Omu 2008 Omu AE, Al-Azemi MK, Kehinde EO, Anim JT, Oriowo MA, Mathew TC. Indications of the mechanisms involved in improved sperm parameters by zinc therapy. Medical Principles and Practice 2008;17:10816. Patel 2008 Patel SR, Sigman M. Antioxidant therapy in male infertility. Urologic Clinics of North America. 2008;35(2):31930. [MEDLINE: 18423251] Ross 2010 Ross C, Morriss A, Khairy M, Khalaf Y, Braude P, Coomarasamy A, El-Toukhy T. A systematic review of the effect of oral antioxidants on male infertility. Reproductive Biomedicine Online 2010;20(6):71123. [PUBMED: 20378409] Sheweita 2005 Sheweita S, Tilmisany A, Al-Sawaf H. Mechanisms of male infertility: Role of antioxidants. Current Drug Metabolism 2005;6(5):495501. Sikka 1995 Sikka S, Rajasekaran M, Hellstrom W. Role of oxidative stress and antioxidants in male infertility. Journal of Andrology 1995;16(6):4648. Stankiewicz 2003 Stankiewicz M, Norman R. Zinc for male infertility. Cochrane Database of Systematic Reviews 2003, Issue 4. [DOI: 10.1002/14651858.CD004633]
45
Tarozzi 2007 Tarozzi N, Bizzaro D, Flamigni C, Borini A. Clinical relevance of sperm DNA damage in assisted reproduction. Reproductive Biomedicine Online 2007;14(6):74657. Tournaye 2006 Tournaye H. Evidence-based management of male subfertility. Current Opinion in Obstetrics & Gynecology 2006;18:2539. Tremellen 2008 Tremellen K. Oxidative stress and male infertility - a clinical perspective. Human Reproduction Update 2008;14(3):
24358. Wathes 2007 Wathes DC, Abayasekara DR, Aitken RJ. Polyunsaturated fatty acids in male and female reproduction. Biology of Reproduction 2007;77(2):190201. Zhou 2007 Zhou X, Liu F, Zhai S. Effect of L-carnitine and/or Lacetyl-carnitine in nutrition treatment for male infertility: a systematic review. Asia Pacic Journal of Clinical Nutrition 2007;16 Suppl 1:38390. [MEDLINE: 17392136] Indicates the major publication for the study
46
CHARACTERISTICS OF STUDIES Characteristics of included studies [ordered by study ID]

Akiyama 1999 Methods Participants Randomised crossover trial. Single centre Country: Japan Male infertility Mean age: 36 years (treatment group age range 24-49, control 30-37) n=10 recruited Inclusion criteria: Male infertility (ROS >5*10000 counts/10000000 viable spermatozoa) Exclusion criteria: Azoospermia, pyospermia Duration of study: 8 months Ethylcysteine 600mg/day for 3 months (n=5) versus Vitamin E 600mg/day (n=5) With a one month wash out, then crossover for another 3 months. Only data from the rst phase was used in data analysis Semen parameters In Japanese. Data extraction translated by Ichiro, a colleague of Samantha Roberts, 29/ 01/09 Author contacted no further information is available.
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement Patients were divided randomly
Random sequence generation (selection Unclear risk bias) Allocation concealment (selection bias) Unclear risk
No details No details
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias Low risk
No incomplete outcome data
Low risk Unclear risk
No statement regarding funding Can not denitively assess other bias
47
Balercia 2005 Methods Participants Randomised trial, double blind. No details of randomisation or concealment Country: Italy Infertile men recruited from an andrology clinic mean age 30 (range 24-38 years) n=60 recruited Inclusion criteria: Primary infertility > 2 years after regular intercourse with a fertile woman; 20-40 years of age; normal rheologic characteristics; sperm count >20 x 10 6 /ml; sperm motility < 50%; normal sperm morphological features > 30%; seminal WBC <1x106 /ml; negative sperm culture and chlamydia and mycoplasma urealyticum; normal serum gonadotropins; T, E2 and PRL; absence of infectious or genital disease; no anatomic abnormalities of the genital tract; absence of systemic diseases or treatment with other drugs within the 3 months before enrolment in the study. Absence of smoking, alcohol or recreational drug use or of occupational chemical exposure. Total duration of study 9 months L carnitine 3g/day orally (n=15) versus L-acetyl carnitine 3g/day orally (n=15) versus L carnitine 2g/day + L-acetyl carnitine 1g/day (n=15) versus placebo (n=15) Duration of treatment: 6 months Semen parameters
Interventions
Outcomes Notes Risk of bias Bias
Authors judgement
Support for judgement Placebo controlled double blind randomised trial No methods of randomisation mentioned No details double blind study. Patients blinded but no details as to who else, assume it is the clinician 1 withdrawal from the L carnitine 2g/day + L-acetyl carnitine 1g/day group Pregnancy was not in the original list of outcomes but was reported in the results
Random sequence generation (selection Unclear risk bias)
Allocation concealment (selection bias)
Unclear risk
Blinding (performance bias and detection Low risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Low risk
High risk
48
Balercia 2005
(Continued)
Other bias
Unclear risk
Can not denitively assess other bias
Balercia 2009 Methods Participants Double blind, placebo controlled randomised trial Country: Italy (University of Marche, Ancona) Infertile men recruited from andrology clinic Mean age: 32 (range 27-32 years) N= 60 men recruited Inclusion criteria: Age 20-40 years, infertility >2 years, regular sexual intercourse with a potentially fertile female; Normal rheologic characteristics (appearance, consistency and liquefaction) of semen and volume and pH in normal range, sperm count >20 x 10 6 /mL, sperm motility <50% (WHO 1999), normal morphology >30%, seminal white blood cells <1 x 106 /mL and a negative sperm culture and chlamydia and M.urealyticum detection. Normal levels of gonadotropins absence of genital disease and anatomical abnormalities of the genital tract including variocoele and antibodies. Absence of systemic disease or treatment with other drugs within 3 months of being enrolled in the study. Absence of smoking, alcohol and drug addiction and exposure to occupational chemicals. Exclusion criteria: Transient decrease in semen quality during run in and those who had sudden improvement in semen parameters during run in. Coenzyme Q10 100mg 2x/day (n=30) versus placebo (n=30) Duration of study: 9 months Primary: Semen parameters Secondary: Pregnancy rate
Interventions
Outcomes
Notes Risk of bias Bias Authors judgement Support for judgement randomised - no details. At end of trial the paper mentions - after opening randomisation list page 1789 No details Double blind - no details
Unclear risk
Blinding (performance bias and detection Low risk bias) All outcomes
49
Balercia 2009
(Continued)
Incomplete outcome data (attrition bias) All outcomes
Low risk
5 patients dropped out of the study 2 from the treatment group and 3 from the placebo group this was discovered after opening the randomisation list at the end of the study. Intention to treat was carried out. Pregnancy was not in the original list of outcomes but was reported in the results, although pregnancy was not an end point for this controlled study, because of many possible interferences, it is interesting to note that six pregnancies occurred with patients given CoQ10 Can not denitively assess other bias
Selective reporting (reporting bias)
High risk
Other bias
Unclear risk
Biagiotti 2003 Methods Participants Randomised study - conference proceeding Country: Italy, Andrology clinic in Bologna Population: severe idiopathic oligoasthenospermia (sperm concentration >5000/l) Mean age: Group a and b 35 (range 30-40 years), Group c 31 (range 24-34) N= 42 Inclusion criteria: severe idiopathic oligoasthenospermia (sperm concentration >5000/ l) Exclusion criteria: Genomic, hormonal or inammatory diseases Duration of study: ? a-Acetyl-carnitine 1 gr/day + L-carnitine 2gr/day + cinnoxicam (n=14) versus b-ALC + LC (n=14) versus c-no therapy (n=14) Duration of treatment: ? Semen parameters Conference abstract - no data given. Contacted authors but no reply re questions as yet.
Interventions
Authors judgement
Support for judgement randomised (1patient = 1 block) analysis of variance Was this at the time of se50
Random sequence generation (selection Low risk bias)

Biagiotti 2003
(Continued)
quence generation or at data analysis? Allocation concealment (selection bias) Unclear risk Not mentioned Not mentioned
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias Unclear risk
Unclear
Unclear risk Unclear risk
Unclear Can not denitively assess other bias
Cavallini 2004 Methods Randomised controlled trial. Allocation concealment: anonymous colour coded boxes Country: Italy Population: Idiopathic plus variocoele associated oligoasthenospermia (OAT) Mean age: 34 years (range 27-40) N= 325 Inclusion criteria: OAT men with deciencies in all sperm patterns whose chief complaint was primary couple infertility >12 months with regular intercourse. Normal sperm appearance, consistency, liquefaction, volume, pH. Female partner without fertility problems. Exclusion criteria: Azoospermia, seminal white blood cell concentration more than 1000,000/mL, positive urethral chlamydia swab test, oligospermia < 5000,000mL, hormonal alterations, age >40 yrs, presence of anti sperm antibodies, drug, tobacco or alcohol abuse, ongoing medical treatments, presence of hydrocoele, diabetes,hypertension, x ray exposure in previous 8 months, peptic ulcer, unexplained gastric pain, previous hypersensitivity to NSAIDS or carnitines, carnitine metabolism deciency, bilateral variocoele, prostate abnormalities, previous or current testicular pathology, testicle echographic abnormalities. Duration of study: 9 months Group 1: Placebo, starch tablets 2x/day + glycerine suppository (1 every 4 days) (n=47) Group 2: L-carnitine 1x 2g/day plus acetyl-L-carnitine 500 x 2mg/day plus glycerine suppository (n=39) Group 3: L-carnitine 1x 2g/day plus acetyl-L-carnitine 500 x 2mg/day plus glycerine suppository plus cinnoxicam suppository 1 x 30mg (every 4 days) n=(44). Cinnoxicam is a non steroidal antiinammatory therefore this arm (Group 3) was not included in meta analysis as per protocol. Duration of treatment: 6 months
Participants
Interventions
51
Cavallini 2004
(Continued)
Outcomes
Primary: Sperm parameters Secondary: Pregnancy, side effects Continuous data taken from Cavallini 2004a excluded conference abstract no data for placebo group unit of analysis variocoeles therefore cannot extract data that were presented as median (interquartile range) Author contacted regarding uneven numbers and missing placebo and continuous data Author replied that raw data were not available due to computer crash
Notes
Risk of bias Bias Authors judgement Support for judgement the patients were randomised into 3 groups no details of how randomisation was carried out drug placebos identical in appearance, anonymized carnitine and cinnoxicam and glycerine suppository containers; and lled and sealed anonymous color coded boxes, the color code was disclosed to physicians by pharmacists and by IRB at the end of the research All study personnel and participants were blinded to treatment assignment for the duration of the study 325 randomised but only 185 accounted for. 55 drop outs from 185 i.e. 42%, 53 reasons given for the drop outs Unclear Can not denitively assess other bias
Low risk
Blinding (performance bias and detection Low risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes High risk
Selective reporting (reporting bias) Other bias
Ciftci 2009 Methods Randomised controlled trial Allocation concealment: Sealed envelopes Country: Turkey Population: Men attending fertility clinic with idiopathic infertility. Normal sperm parameters Mean age: Treatment group 33.14.5, Control 32.83.7 years N=120 recruited
52
Participants
Ciftci 2009
(Continued)
Inclusion criteria: Men attending fertility clinic with idiopathic infertility. Normal sperm parameters Exclusion criteria: Cryptorchidism, vasectomy, abnormal liver functioning, smoking, alcohol consumption Duration of study: 3 months Interventions N-acetylcysteine 600mg/day (n=60) versus placebo (n=60) Duration of treatment: 3 months Primary outcomes: Total antioxidant capacity, peroxide levels, oxidative stress Secondary outcomes: Other semen parameters Attempt made to contact author regarding data reported in SDs or SEs email sent 24.09.10
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement patients were randomly allocated to the study group (60 men) or control group (60 men) No mention of how the randomisation was carried out. Using sealed envelopes, these patients were randomly allocated... The patients in the study and control groups were unaware of whether they were receiving the drug or placebo. ? researchers blinded All men recruited were analysed. No withdrawals No statement of funding Report includes all expected outcomes Baseline data were similar for both groups Can not denitively assess other bias
Low risk
Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias)
Low risk
Unclear risk
Other bias
Unclear risk
53
Conquer 2000 Methods Participants Randomised placebo controlled trial. 3 arms Country: Canada Population: healthy asthenozoospermic individuals who were patients of an infertility clinic Mean age: placebo=35.2, DHA 400mg=38.3, DHA 800mg=34.4 N=28 Inclusion criteria: Asthenozoospermic, sperm motility <50% of total sperm Exclusion criteria: Not stated Duration of study: Not stated Fatty acid omega 3 Docosahexaenoic acid (DHA) 400mg/day (n=9) versus DHA 800mg/day (n=10) versus placebo (n=9) Duration of treatment: 3 months Sperm parameters Data with SEMs converted to SDs Placebo arms split
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement The 28 subjects were randomly assigned to ... Not stated Placebo capsules made with corn oil/soy oil however blinding not stated.
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Low risk
All men randomised were in the analysis, no drop outs. All specied outcomes were reported Outcomes reported Funding from Martek Bioscience corporation (also provided the placebo capsules)
54
Dawson 1990 Methods Participants Randomised controlled trial Country: USA Population: males with sperm agglutination Mean age: age range 25-45 years N=30 Inclusion criteria: sperm agglutination over 25%, negative sperm antibodies, physically normal, no inammatory disease Exclusion criteria: unclear Duration of study: 4 weeks Ascorbic acid (AA)1000mg (n=10) versus AA 200mg (n=10) versus placebo (n=10) Duration of treatment: 3 weeks Seminal parameters Placebo numbers split by 2 Data were given in standard error converted to standard deviation
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement By random selection, three groups of 10 subjects each..
Not stated Each subject was told he was receiving AA and expected improvement in sperm quality
No drop outs
All specied outcomes were reported supplement and placebo were provided by HoffmannLa Roche, Inc, Nutley, NJ.
55
Galatioto 2008 Methods Randomised controlled, intention to treat, single centre study. Central allocation - pharmacy, blinded Power calculation performed Country: Italy Population: Men with persistent oligospermia (5-20m/ml) Mean age:Treatment group=32 years (27.5-35.5), Control=33(23-36) N=42 Inclusion criteria: having performed a retrograde embolization with concomitant oligospermia, persistent oligospermia and infertility > 12 months Exclusion criteria: smoking, alcohol consumption, taking any fertility drugs within 3 months prior to the study, serious medical or psychiatric condition, abnormal hormonal prole, sperm infection N-acetylcysteine (NAC) 600mg and vitamins-minerals (Vit C, Vit E, Vit A, thiamine, riboavin, piridoxin, nicotinamide, pantothenate, biotin, cyanocobalamin, ergocalciferol, calcium, magnesium, phosphate, iron, manganese, copper, zinc. (n= 20) versus no treatment (n=22) Duration: 12 months after end of study which ran for 90 days Primary: seminal parameters Secondary: pregnancy and adverse effects Attempted to contact author regarding median data. No response as yet.
Participants
Interventions
Outcomes
Notes Risk of bias Bias
Authors judgement
Support for judgement Subjects were randomly assigned to either antioxidant therapy or no medical therapy. Randomisation number was assigned by random allocation software using a block randomisation design All steps of randomisation process were performed blindly in the pharmacy of our hospital. Control is no treatment
Low risk
Blinding (performance bias and detection High risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Low risk
intention to treat
Low risk
Does not appear to be any selective reporting

56
Galatioto 2008
(Continued)
Other bias
Unclear risk
Can not denitively assess other bias
Greco 2005 Methods Double blind randomised controlled trial. Trialists and patients blinded. Methods were unclear Country: France Population: infertile males Mean age: ? N=64 Inclusion criteria: Tunel assay showed a presence of fragmented DNA15%of ejaculated spermatozoa Exclusion criteria: Variocele, genitourinary inammation, infection, smoking Duration of study: ? Vitamin C 500mg 2x/day + vitamin E 500mg 2x/day (n=32) versus placebo (n=32) Duration of treatment: 2 months Sperm parameters
Participants
Interventions
Authors judgement
Support for judgement The study participants were randomised into 2 groups No mention of allocation concealment The study was double-blinded with both the authors and the patients unaware of which of the patients was in the treatment or control arm of the study All specied outcomes are assessed. No drop outs Does not appear to be any selective reporting Can not denitively assess other bias
Random sequence generation (selection Low risk bias) Allocation concealment (selection bias) Unclear risk
Low risk
Low risk
Other bias
Unclear risk
57
Keskes-Ammar 2003 Methods Participants Randomised controlled trial - open labelled Country: Tunisia Population: Infertile men Mean Age: 35.56.8 (SD) Recruited: N=78 Randomised: N=54 Inclusion criteria: Infertile men who had been married one year Exclusion criteria: ? Duration of study: 10 months Vitamin E (400mg/day) + selenium (225mg/day) for 3 months (n=12) versus vitamin B (4.5g/day) (n=8) Duration of treatment: 3 months Semen parameters Attempted to contact authors regarding high attrition rate >50% ?78 men randomised or 78 recruited then only 54 (28 in intervention and 26 in control) Then only 20 analysed due to non-compliance.
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement Randomisation was performed with random numbers and the numeric code was withheld from researchers and patients open label The trial was randomised and open However the numeric code was withheld from researchers and patients High attrition rate was due to non-compliance, no intention to treat. MDA post treatment has no SD no explanation Can not denitively assess other bias
Unclear risk
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Low risk
Unclear risk
Other bias
Unclear risk
58
Kessopoulou 1995 Methods Double blinded randomised placebo cross over trial Power calculation performed Country: UK, Shefeld Population: Men with high levels of reactive oxygen species (ROS) attending Obstetrics and Gynecology department for infertility Couples were undergoing IVF Mean Age: ? Median age: 32 years Recruited: N= 30 Inclusion criteria: Attending fertility clinic, high levels of ROS in semen. Female partner has tubal patency and is ovulating Exclusion criteria: Men with antisperm antibodies, >20% spermatozoa with Ig (immuno globin A) and/or IgG antibodies and sperm concentration <5x 06 mL Duration of study: 2 years Vitamin E 300mg 2x/day (n=15) versus identical placebo (n=15) Duration of treatment: 3 months Primary outcomes: Semen parameters Secondary outcomes: Adverse effects. Live birth Attempted to contact author regarding median data, no response as yet Only rst phase data used in analysis
Participants
Interventions
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement The study was a randomised double blind placebo controlled trial. The randomisation was performed by the manufacturer The randomisation was performed by the manufacturer the code was blind for the researcher and patients. The code was broken at the end of the trial All outcomes are reported
Unclear risk
Low risk
Does not appear to be any selective reporting Funding grant by EEC. Tablets provided by pharmaceutical company
59
Other bias
Unclear risk
Lenzi 2003 Methods Randomised placebo controlled, double blind crossover trial. Power calculation performed Country: Italy Population: Male factor infertility - oligoasthenoteratozoospermia (OAT). Mean age: ? Range: 20-40 years N=100 Inclusion criteria: Men are aged between 20-40 years with infertility lasting longer than 2 years. Regular sexual intercourse with a gynaecologically normal female partner with no female infertility. Absence of endocrine disease, genital infections, obstructive cryptorchism, antisperm antibodies, normal sperm parameters with no signicant differences after 3 tests. Mild oligospermia. Sperm concentration 10-20 x 106 /mL and motility 1030% Exclusion criteria: Not mentioned Duration of study:10 months L-carnitine 2g/day (n=43) Versus placebo (n=43) Duration of treatment: 6 months Semen parameters and pregnancy rate First phase data only used in analysis Attempted to contact author regarding standard deviations, how many were in each group for the 1st phase and how many of the 4 who went to assisted reproduction did so in the 1st phase and what do they mean by 172 cycles. No response yet.
Participants
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement we report on a randomised placebo controlled cross over trial. No mention of method of randomisation No mention Double blinded. seemingly identical placebo
Unclear risk
14 withdrew - 4 went onto assisted reproduction, 6 did not return for 2nd period and 4 due to pregnancy in rst phase. Therefore should only be ?4 at the most lost from 1st phase. No intention to treat All withdrawals accounted for for whole trial however how many were lost in the
60
Lenzi 2003
(Continued)
rst phase in rst phase Selective reporting (reporting bias) Unclear risk Although all outcomes are reported but no standard deviations given Funding stated - University grant. Tablets from Sigma Tau SRL
Other bias
Unclear risk
Lenzi 2004 Methods Placebo controlled, double blind randomised trial. No mention of method of randomisation or allocation concealment When codes were broken at the end of the study Power calculation performed Country: Italy Population: Infertile males with oligoasthenoteratozoospermia Mean age: Not stated. Age range 20-40 years N=60 Duration of study: 8 months Inclusion criteria: oligoasthenoteratospermia, age between 20-40 years, infertility >2years with regular intercourse. No endocrine disease, cryptorchidism, genital infections/obstructions, variocoele or testicular hypertrophy, antisperm antibodies Exclusion criteria: None Duration of study: 8 months L-carnitine 2g/day + L-acetyl-carnitine 500mg 2x/day (n=30) versus placebo (n=26) Duration of treatment: 6 months Semen parameters and pregnancy rate Attempted to contact author regarding 8 month follow up data. No reply as yet.
Participants
Interventions
Authors judgement
Support for judgement placebo controlled double blind randomised trial Mentions coding - When codes were broken at the end of the study Double
61
Lenzi 2004
(Continued)
Unclear risk
4 men withdrew from the placebo group. 60 randomised 56 analysed. No intention to treat Unclear Can not denitively assess other bias
Li 2005 Methods Participants Double blinded randomised parallel trial Country: Eastern China Population: infertile men with oligoasthenospermia Mean Age: Treatment 305.5 (23-45 years), Control 323.5 (24-46 years) N=150 Inclusion criteria: No smoking or alcohol. Any fertility medication needed to be stopped 2 weeks before Exclusion criteria: Nil Duration: 3 months L-carnitine 2g/day + acetyl-L-carnitine 1g/day (n=85) (90 with intention to treat) versus vitamin E (100 mg tid) + vitamin C (100mg tid) (n=53) (60 with intention to treat) Duration of treatment: 3 months Seminal parameters and pregnancy rate per couple Withdrawal: 5 from treatment group and 7 from control Query whether randomised as there is imbalance between the groups. Contact author re methods of randomisation (large imbalance of numbers), concealment and whether SD or SEs used and query that this is the same trial as Li 2005a Translated by Shaofu Li 10/11/08
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement Randomised controlled trial. Methods not described. Unexplained differences in the number of men in the treatment and control groups Allocation concealment not described Double
Unclear risk
62
Li 2005
(Continued)
High risk
Unexplained imbalance between the number of men randomised into the treatment and control groups Unclear whether data given in standard deviations or standard errors Unexplained imbalance between the number of men randomised into the treatment (n90) and control groups (n60)
Unclear risk
Other bias
Unclear risk
Li 2005a Methods Participants Randomised trial. No information on concealment Country: Eastern China Population: Infertile men with oligoasthenospermia Mean age: 293.5 (23-40 years) N=80 Inclusion criteria: No smoking or alcohol. Any fertility medication needed to be stopped 2 weeks before Exclusion criteria: Nil Duration: Not stated L-carnitine 2g/day (n=40) versus vitamin E 100mg + vitamin C 100mg tid (n=40) Duration of treatment: 3 months Seminal parameters and pregnancy Attempted to contact author re methods of randomisation, concealment and whether SD or SEs used and whether this is the same trial as Li 2005. Also asked whether there were any data on pregnancy rate. Translator replied 22.09.09 no pregnancy data were available in the text of the trial. Translated by Shaofu Li 10/11/08
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement No mention of methods of randomisation
No mention of allocation concealment
63
Li 2005a
(Continued)
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Unclear risk
No mention of blinding
Withdrawal: 8 from treatment (n=32) and 9 from control (n=31). 21% loss to follow up. No intention to treat No data on pregnancy outcome Can not denitively assess other bias
High risk Unclear risk
Lombardo 2002 Methods Participants Randomised controlled crossover trial Country: Italy Population: male infertility, oligoasthenospermia Mean age: Not stated N=100 Inclusion criteria: Age 20-40 years. Infertility > 2 years. 3 baseline semen analysis demonstrating concentration 10-20 106 /mL , %motility 10-30%, forward progression <15%, abnormal morphological forms <70%, curvilinear velocity 10-30/second + linearity <4 Exclusion criteria: not mentioned Duration: 10 months L-carnitine 2g/day (n=?) versus placebo (n=?) Duration of treatment: 2 months Semen parameters Abstract only Attempted to contact author re 1st phase data, outcomes, randomisation, concealment and whether there was a full publication of the trial.
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement randomised -all patients had an initial 2 months run in period and then randomised to 2 months of carnitine or placebo. No mention of allocation concealment
Unclear risk
64
Lombardo 2002
(Continued)
Blinding (performance bias and detection Low risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Unclear risk
double blind crossover trial
86 patients completed the trial out of 100. Need to see full trial for the reasons for withdrawals and intention to treat Unclear Funding stated - support: None
Merino 1997 Methods Participants Randomised controlled trial Country: Mexico Population: Male idiopathic asthenozoospermia attending a fertility clinic Mean age: 30.86 (20-40 years) N=47 Inclusion criteria: Free of urogenital symptoms, idiopathic asthenozoospermia, not received drugs in prior 6 months and healthy Exclusion criteria: Variocoele, inammatory diseases, endocrine disorders Duration of study: 6 months Pentoxifylline 1200mg - 400mg/3xday (n=25) versus placebo (n=22) Duration of intervention: 6 months Seminal parameters and hormonal assays Attempted to contact author to ask if had data in means + SD and not medians + range as published. Letter sent 15.09.09. Letter returned to sender.
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement The 47 men were divided at random
No mention of allocation concealment No mention of blinding however the trial is placebo controlled
Blinding (performance bias and detection Unclear risk bias) All outcomes
65
Merino 1997
(Continued)
Unclear risk
No withdrawals
Unclear risk
?all outcomes reported however seminal outcomes reported in medians Can not denitively assess other bias
Other bias
Unclear risk
Micic 1988 Methods Randomised controlled trial Methods not stated. No mention on methods of concealment Country: Belgrade, Hungary Population: idiopathic oligoasthenospermia Mean age:292 in treatment group, 263 in control group N=90 Inclusion criteria: idiopathic oligoasthenospermia and no pregnancy for 2 years Exclusion criteria: Variocoele, inammatory diseases, endocrine disorders Duration of study: 3 months Pentoxifylline 1200mg/day (n=51) versus no treatment (n=39) Duration of treatment: 3 months Seminal parameters Attempted to contact author regarding extractable data for pregnancy rate plus methods of of randomisation and concealment also dont know if adverse events are single or multiple events. No reply as yet.
Participants
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement A randomised group of 90 men
No mention of allocation concealment No mention of blinding
Blinding (performance bias and detection High risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Unclear risk
Dont know if adverse events are single or multiple. Pregnancy data not extractable
66
Micic 1988
(Continued)
All randomised were analysed Can not denitively assess other bias
Nozha 2001 Methods Randomised comparative study No mention of allocation concealment Abstract only Country: Tunisia Population: Infertile males with oligoasthenoteratozoospermia Mean age: not stated N=? Inclusion criteria: males with oligoasthenoteratozoospermia. Exclusion criteria: none mentioned Duration of study: not stated Vitamin E (400mg) + selenium (200mg)/day (n=12) versus vitamin B (B2 , B6 and B12 ) (n=8) Duration of treatment: 3 months Seminal parameters Attempted to contact authors regarding methods of randomisation and data - no extractable data from the abstract. No reply as yet.
Participants
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement In a prospective randomised comparative study No mention of allocation concealment No mention of blinding
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias Unclear risk
Unclear
Unclear Can not denitively assess other bias

67
Omu 1998 Methods Randomised controlled trial Open trial - control is no treatment Country: Kuwait Population: Men with asthenozoospermia attending infertility and andrology clinic Mean age: 37.87.9 in treatment group, 38.18.2 in control N= 100 Inclusion criteria: Men with asthenozoospermia. Spermatozoal motility impaired with >40% non mobile sperm. Have been trying to conceive for at least one year. Plus no obvious female factor Exclusion criteria: None mentioned Duration of study: 12 months Zinc 250mg 2xday (n=49) versus no treatment (n=48) Duration of treatment: 3 months Seminal parameters Attempted to contact authors regarding methods randomisation and concealment questioned. No reply as yet.
Participants
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement Randomised. 100 men with Asthenozoospermia were randomised into two groups No mention of allocation concealment No mention of blinding
Unclear risk
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Other bias Unclear risk
100 men randomised, 97 analysed, drop outs are accounted for Funding - Kuwait University
Unclear risk
68
Omu 2008 Methods Participants Randomised controlled 4 armed trial, open as one arm of the trial is no therapy Country: Kuwait Population: Men with Asthenozoospermia attending infertility clinic in Kuwait Mean age: 351 N= 45 Inclusion criteria: Asthenozoospermia with normal sperm concentration (20-250mill/ ml) but with 40% or more immotile sperm Exclusion criteria: Asthenozoospermia but sperm concentration of <20 mill/ml Duration of study: No stated Zinc 200mg 2x/day (n=11) versus zinc 200mg + vitamin E 10mg 2x/day (n=12) versus zinc 200mg + vitamin E 10mg + vitamin C 5mg 2x/day (n=14) versus no therapy (n=8) Duration of intervention: 3 months Seminal parameters Attempted to contact author re methods of randomisation - states that 8 men served as non- therapy control. No reply as yet.
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement the 45 asthenozoospermic men were randomised into four groups No mention of allocation concealment No mention of blinding
All outcomes are reported. No drop outs
Appears to be free of selective reporting Can not denitively assess other bias
69
Peivandi 2010 Methods Participants Double blind randomised crossover trial Country: Iran Population: Infertile men with at least two abnormal spermiograms N=30 Exclusion criteria: variocoele, testicular atrophy, ejaculatory disorders, use of medications, azoospermia, endocrinological disorders, ICSI candidacy or other causes of infertility Duration of 1st phase: 8 weeks L-carnitine (2g/day) versus placebo Sperm parameters Abstract in English, full text in Arabic. Contacted the author and he is lling out the data extraction sheets. Author responded but data queries remain contacted again re SDs and pregnancies in 1st phase of cross over. Author responded saying that the data was given in SDs and there were 3 pregnancies in the 1st phase.
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement patients were randomly allocated to two groups of A and B sealed opaque envelopes Double blind outcome assessor was blinded
Random sequence generation (selection Unclear risk bias) Allocation concealment (selection bias) Low risk
Blinding (performance bias and detection Low risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Unclear risk
loss to follow up was not accounted for
Unclear risk
Pregnancy was not our major goal but when we found a positive pregnancy test after treatment with carnitine, we reported the results Standard deviations appear to be unusual Can not denitively assess other bias
Other bias
Unclear risk
70
Rolf 1999 Methods Randomised placebo controlled double blind study No mention of allocation concealment Power calculation performed Country: Munster Germany Population: men with infertility for over one year. Mean age: treatment 36.15.0, placebo 35.24.8 N= 33 Inclusion criteria: Asthenozoospermia (<50% motile) diagnosed after 2 examinations, normal or reduced sperm concentration (>20 x 106 per ejaculate) and without infection of access glands Exclusion criteria: None mentioned Duration of study: 8 weeks Vitamin C 1000mg + vitamin E 800mg/day (n=15) versus placebo (n=16) Duration of treatment: 8 weeks Primary: Semen parameters Secondary: Pregnancy rate and adverse effects Contacted author about the allocation concealment and pregnancy and adverse effects were outcomes in their protocol. Rolf replied saying that pregnancy and adverse effects were stated in the protocol.
Participants
Interventions
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement Randomisation was performed with random numbers without further stratication by the pharmacist and the code was withheld from researchers and patients Pharmacist performing randomisation and code withheld from patients and researchers. However no mention of type of containers or envelopes Double - patients and researchers
Unclear risk
Blinding (performance bias and detection Low risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Low risk
All data reported. 2 patients withdrew form the trial - results from two patients were rejected from analysis. 1 from the treatment group due to poor compliance and 1 from the placebo group due to genital tract
71
Rolf 1999
(Continued)
infection. Intention to treat Selective reporting (reporting bias) Low risk All semen outcomes reported and author states (email 22.09.09) that pregnancy and adverse effects were set a priori in the protocol Funding stated - Health Ministry
Other bias
Unclear risk
Safarinejad 2009 Methods Double blind placebo controlled randomised study. Randomised using permuted blocks Allocation concealment: Sealed envelopes Blinding: Double, identical coating of tablets Power calculation performed Country: Iran Population: Idiopathic oligoasthenoteratospermia, asthenospermia or teratospermia of 2 years duration Mean age: 31 years (25-48 years) Recruited: N= 548 Randomised: N= 468 Inclusion criteria: Sperm count >5 x 106 /ml, over 2 years of failed conception, no female fertility problems, no history of possible cause for male infertility Exclusion criteria: Abnormal testes, history of cancer or chemotherapy, testosterone or antiandrogen use, use of selenium or N-acetyl-cystine supplements, abnormal hormone levels, genital disease, genital inammation or variocoele, history of genital surgery, major surgery, central nervous system injury, a known sperm defect or retrograde ejaculation. Y chromosome abnormalities, sexually transmitted disease, genitourinary infection, leukocytospermia, smoking, any environmental exposures to reproductive toxins. Medical, neurological or psychological problems. A history of drug or alcohol abuse, hepatobiliary disease or signicant renal insufciency. Any endocrine abnormality, a body mass index (BMI) of 30kg/m2 or over, participation in another investigational study and a likelihood of being unavailable for follow up. Duration of study: 56 weeks Selenium 200 g/day (n=116) versus N-acetylcysteine 600mg/day (n=118) versus selenium 200 g/day + N-acetylcysteine 600mg/day (n=116) versus placebo (n=118) Duration of treatment: 26 weeks or 6.5 weeks Analysed: n= 105 in selenium group (loss 11), n= 106 in placebo group (loss 12), n=105 in N-acetylcysteine group (loss 13) and n= 104 in selenium + N-acetylcysteine group (loss 12)
Participants
Interventions
72
Safarinejad 2009
(Continued)
Outcomes
Primary outcome: Semen parameters Secondary outcomes: adverse events Attempted to contact authors regarding side effect data that had not yet been added to the review due to the query of multiple comparisons. Also to ask whether data is in SD or SE as requested by statistician 24.09.10.
Notes
Risk of bias Bias Authors judgement Support for judgement randomisation table generated by the method of random permuted blocks. Patient randomisation numbers were allocated to each site in ascending sequence in blocks. Assignment to treatment groups was performed using a sealed envelope technique. Eligible patients were randomly assigned to double blind.. Placebo pills were coated with titanium oxide to ensure an identical appearance and smell. All withdrawals were accounted for in each treatment group. Withdrawal was mainly due to withdrawal of consent followed by lost to follow up and lastly for reasons of missing data. No intention to treat The published report includes all expected outcomes Can not denitively assess other bias
Low risk
Low risk
Low risk
Other bias
Unclear risk
Safarinejad 2009a Methods Randomised controlled trial, double blind Allocation concealment: Not mentioned Power calculation performed Country: Tehran, Iran Population: Infertile males between 21 and 42 years with idiopathic oligoasthenoteratospermia. Mean age: Treatment group 289, Placebo 2810. (Range 21-42 years) Recruited: N= 268
73
Participants
Safarinejad 2009a
(Continued)
Randomised: N=212 Inclusion criteria: Minimum 2 years unprotected intercourse with 2 years unwilling childlessness. male infertility diagnosed if 1 or more standard semen parameters were below cutoff levels accepted by World Health Organisation (WHO). A fertile female partner. No known medical condition that could account for infertility, testicular volume 12 ml or greater. No medical therapy for at least 12 weeks before the study begins. Only patients seeking medical attention for infertility were included. Exclusion criteria: Azoospermia or severe oligospermia (sperm count less than 5 million/ ml. An history of epypidymo-orchitis, prostatitis, genital trauma, testicular torsion, inguinal or genital surgery. Any genital or central nervous system disease, endocrinopathy, cytotoxic drugs, immunosuppressants, anticonvulsives, androgens, antiandrogens, a recent history of Sexually transmitted disease. Psychological or physiological abnormalities that would impair sexual functioning or ability to produce sperm samples. Drug, alcohol or substance abuse. Liver disease, renal insufciency or chromosome abnormalities. occupational and environmental exposures to reproductive toxins. A body mass index (BMI) of 30kg/m2 or over, participation in another investigational study and a likelihood of being unavailable for follow up. Duratoin of study: 20 months, from February 2005 until October 2006 Interventions Coenzyme Q10 (CoQ10) 300mg/day (n=106) versus placebo (n=106) Duration of treatment: 26 weeks or 6.5 weeks Analysed n= 98, 8 withdrawals in treatment group, analysed n=96, 10 withdrawals in the placebo group Primary outcomes: semen parameters and testicular volume Secondary outcomes: adverse effects and hormone levels
Outcomes
Notes Risk of bias Bias Authors judgement Support for judgement Each eligible patient received a randomisation number, which was determined by a computer generated schedule. Therafter a randomisation table was generated by the method of random permuted blocks. Individuals who were geographically and operationally independent of the study investigator performed the study randomisation Allocation concealment: not mentioned The clinician prescriber and the patients were blinded to the treatment condition. To maintain and guarantee blinding
74
Unclear risk
Safarinejad 2009a
(Continued)
CoQ10 and placebo were identical in appearance. Participant data collected during this trial were kept condential and locked in a secure ofce area. Randomisation codes were opened only after all patients had completed the whole study protocol. Incomplete outcome data (attrition bias) All outcomes Low risk All patients who dropped out of the trial were accounted for - 8 from treatment group and 10 from placebo group for reasons such as withdrawal of consent, missing data and loss to follow up All semen, hormone and adverse effects outcomes reported. Pregnancy outcome not stated as an outcome initially but it was reported on in the discussion Funding source not stated. Can not denitively assess other bias
Unclear risk
Other bias
Unclear risk
Scott 1998 Methods Double blind randomised trial Allocation concealment: no mention Country: Glasgow, UK Population: Men attending subfertility clinic with low sperm motility Mean age: 33.3 years0.64 Recruited: N=69 Analysed: N=64 Inclusion criteria: Low sperm motility Exclusion criteria: Not mentioned Duration of study: 3 months and two weeks Selenium 100g/day (n=16) versus selenium 100g + vitamin A 1mg + vitamin C 10mg + vitamin E 15mg/day (n=30) versus placebo (n=18) Duration of treatment: 3 months Analysed = 64, 5 withdrew 1 from selenium group, 4 from selenium + vit A, C and E group. All withdrawals were due to non-compliance Primary outcome: semen parameters Secondary outcome: Pregnancy rates given although not stated as an initial outcome
Participants
Interventions
Outcomes
75
Scott 1998
(Continued)
Notes
Uneven numbers, multivitamin numbers are double the other groups Need to ask author if they have separate numbers for pregnancy data. Currently have 5 pregnancies in the 2 treatment groups and none in placebo Who was blinded, was the placebo identical when group 2 contained so many different vitamins Was there any allocation concealment? Author has retired and is not contactable
Risk of bias Bias Authors judgement Support for judgement As the patients entered the trial they were randomly allocated to one of three treatments, which had in turn been randomised within each block of four numbers and blinded using a numeric code. Unclear as to why the uneven nature of the numbers in the groups i.e. 30 in multivitamin group and 16 in Selenium, 18 in placebo Not mentioned Double blind
Unclear risk
Numbers of withdrawals and reasons (non compliance) were reported Sperm counts and motility were reported as per trial methods. Pregnancy rates were given in the results of the trial, these had not been specied in the methods. The pregnancy results given as a combination of the 2 vitamin groups. There were no pregnancies in the placebo group Can not denitively assess other bias
Unclear risk
Other bias
Unclear risk
76
Sigman 2006 Methods Randomised double blind trial Allocation concealment: adequate Country: Minneapolis, USA Population: Males aged 18-65 years IVF Mean age: 36.25.8 (SD), 35.37.5 (SD) Recruited: N=26 Analysed: N=21, 5 withdrawals Inclusion criteria: Males 18-65 years with infertility of at least six months duration, sperm concentration of at least ve million sperm/mL, motility of 10%-50%, absent pyospermia and normal FSH and testosterone levels. Exclusion criteria: History of post pubertal mumps, cryptorchism, vasal or epididymal surgery, history of medication or chemotherapy. recent alcohol, chronic marijuana. Use of testosterone or steroids. Exposure to environmental toxins. Recent history of fever or diabetes, liver failure, renal failure, endocrine disorder, untreated variocoele, urogenital infection, or prior vasectomy reversal. Duration of study: 24 weeks Carnitine (L-carnitine 2000mg +L-acetylcarnitine 1000mg/day) (n=12) versus placebo (n=9) Primary outcome: semen parameters Secondary outcomes: Pregnancy - 2 pregnancies, 1 in the treatment arm after IVF and one in placebo through intercourse Author replied 21.09.09 saying The published 2006 trial is the published version of the 2003 abstract (Pryor 2003) and giving details of randomisation and concealment. Author says he will try and nd out about the 5 patients that dropped out. Why did - 5 additional patients entered the study but dropped out before completion - when did these patients enter and were they randomised? One of these 5 dropped out because of pregnancy three months after starting carnitine Pryor paper excluded as it is the same study as Sigman, author also gave details of randomisation and allocation concealment, author will try to nd info on 5 patients who dropped out.
Participants
Interventions
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement Patients were randomised to receive carnitine or placebo The randomisation was done by a 3rd party a company that oversaw the trial. We sent the patient number of new recruited patients in to them, they assigned them a study number that was associated with a collection of medication/placebo. The author replied to randomisation query
77
Sigman 2006
(Continued)
23.09.09 saying that the protocol stated that - treatments will be assigned randomly to a subject number. The numbers will range from 1-84 for study centre 1 and 85-168 for study centre 2. Randomisation of treatments for each centre will be done independently. One half of subject numbers will be placebo, the other half, active ingredient. Allocation concealment (selection bias) Low risk The investigators and study sites had the study medication/placebo packets identied by number only. They were blinded to what was in the medication/placebo packets. We were sent the code at the conclusion of the trial. The author replied to a query on allocation concealment on 23.09.09 saying that the protocol stated that - Integrated Data Solutions, Inc. will keep the randomisation code in a separate sealed envelope for each site until the end of the study. The randomisation lists will be provided to the packaging company for packaging of the packets into patient medication boxes. Both the investigators and the patient were blinded to the treatment arm assignment.
Blinding (performance bias and detection Low risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Low risk
5 additional patients entered the study but dropped out before completion. One of these dropped out because of pregnancy three months after starting carnitine. Author replied to query re drop outs - I have data on one drop out at my site - the drop out occurred after randomisation to carnitine. The drop out occurred before the rst follow-up study visit. The other four drop outs were from the other study site - I am trying to get that data for you (23.09.09) Baseline data were similar Pregnancy data were not an outcome but is reported in the paper All outcomes of interest were reported Can not denitively assess other bias
Unclear risk
Other bias
Unclear risk
78
Suleiman 1996 Methods Double blind randomised controlled trial No mention of allocation concealment Country: Saudi Arabia Population: Men attending a fertility centre Mean age: 34.8 (27-52 years), placebo 33.2 (22-45 years) N=110 Inclusion criteria: Asthenospermic (>/=20 x 106 /ml). sperm motility </= 40%, normal sperm count, leucocyte concentration <5%, normal fructose concentration. Normal female Exclusion criteria: None mentioned Duration of study: 6 months Vitamin E 100mg 3x/day (n=52) versus placebo 3x/day (n=35) Duration of treatment: 6 months Primary outcome: motility and MDA concentration Secondary outcome: live birth, pregnancy, miscarriage Method of randomisation not stated. Large imbalance between the treatment (n=52) and placebo group (n=35) at analysis. Drop outs were accounted for in text.
Participants
Interventions
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement Either 100mg vitamin E or a placebo was prescribed in a random double blind fashion. Method of randomisation not stated. Large imbalance between the treatment (n= 52) and placebo group (n=35) at analysis Not mentioned Double blinded. Does not report on who was blinded
Unclear risk
All men who dropped out the trial were accounted for. Although the exact gures for each group is unclear - A total of 110 patients were enrolled in the study, but some of the patients dropped out and some left the region and failed to continue. When the experiment was terminated, 52 patients were found to have taken vitamin E and 35 patients to have taken the placebo.
79
Suleiman 1996
(Continued)
Unclear risk
All outcomes stated in the methods were reported in results Unequal reporting bias If the semen sample improved and the patients spouse became pregnant, the treatment was stopped; otherwise it was continued for 6 months. The placebo was given for 6 months Unequal reporting ?? incomplete outcome data addressed
Other bias
Unclear risk
Tremellen 2007 Methods Randomised double blind controlled trial. randomisation by computer generated blocks. Using a 2:1 ratio (treatment 2: placebo 1) Allocation concealment: numbered bottles delivered to the site with all members of the trial blinded to sequence, identical placebo Power calculation performed Country: Australia Population:male factor infertility undergoing IVF Mean age: treatment group - 37.15.1, placebo group - 35.54.3 Recruited: N= 82 Randomised: N= 60 Inclusion criteria: men with sperm samples showing oxidative stress and a signicant level of DNA fragmentation (>25% Tunel positive) Exclusion criteria: Female partner with diminished ovarian reserve or if the female partner is aged over 39 years Duration of study: 1.5 years Menevit (folate 0.5mg, garlic 1000mg, lycopene 6mg, vitamin E 400 IU, vitamin C 100mg, zinc 25mg, selenium 26 gm, palm oil). One capsule per day versus placebo (identical in appearance and taste - containing palm oil) Duraton of treatment: 3 months prior to IVF cycle Primary outcome: embryo quality Secondary outcomes: pregnancy, fertilization rate, side effects The study reports only on 37 men in the treatment group and not 38. There is no explanation of the one man missing from this analysis.
Participants
Interventions
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement
80
Tremellen 2007
(Continued)
The randomisation schedule was computer generated in blocks of six by Bayer Consumer Care Australia. Using a 2:1 ratio There were no signicant differences between the active and the placebo group in terms of important baseline prognostic characteristics... the appropriately numbered bottles of capsules delivered to the clinical site without any participant knowing the treatment sequence. Patients were allocated the next numerical treatment package (one to sixty as they became eligible for enrolment double blind placebo controlled trial
Low risk
All withdrawals were accounted for, 2 from the intervention group, 4 from placebo all due to the couples not going through to embryo transfer The study reports only on 37 men in the treatment group and not 38 (page 219 ? typo). There is no explanation of the one man missing from this analysis Intention to treat stated All specied outcomes are reported Pregnancy only followed to 13 weeks Can not denitively assess other bias
Low risk
Other bias
Unclear risk
Wang 1983 Methods Participants Randomised placebo controlled trial Country: Hong Kong Population: New patients attending infertility clinic Mean age: 33.74.4 years (28-45 years) N=46 Inclusion criteria: Men with Idiopathic oligospermia attending fertility clinic with normal hormones. They had not received treatment with pharmacologic agents for 1 year prior to entry into the trial. None of the female partners had irreversible causes of female infertility Exclusion criteria: Not mentioned
81
Wang 1983
(Continued)
Duration of study: 6-9 months Interventions Pentoxifylline 400mg 3 x/day (n=11) versus placebo (n=7) Duration of treatment: 6 months Other interventions not included in meta analysis were clomiphene citrate, mesterolone and testosterone enanthate Primary outcome: Semen parameters and pregnancy Pentoxifylline supplied by Hoechst Aktiengesellshaft Attempted to contact the author re motility data as it was stated in the paper that this was an outcome. Also asked about control data for post treatment period and methods of randomisation + allocation concealment. Author replied saying she no longer has the data.
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement the subjects were randomly assigned to one of the therapy groups.. No mention of methods of randomisation No mention of allocation concealment No blinding - one of the interventions was an intra muscular injection while the others were oral Sperm motility data not given when motility was discussed as an outcome of interest No post-treatment control data given however there is post-treatment given for treatment group No intention to treat reported Side effects reported in the results but this was not stated as an initial outcome Funding - Hoechst Aktiengesellshaft
Unclear risk
Blinding (performance bias and detection High risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) High risk
High risk
Other bias
Unclear risk
82
Wong 2002 Methods Double blind randomised placebo controlled interventional study. Computer generated randomisation Allocation concealment: central pharmacy coding Country: Netherlands Population: Fertile and subfertile men Mean age: 34.33.9 years Recruited: N=258 subfertile Randomised: N=103 Inclusion criteria for subfertile group: failure to conceive after 1 year regular unprotected intercourse and sperm concentration of 5 - 20 million/ml Exclusion criteria for subfertile group: chromosomal disorders, cryptorchidism, vasectomy, use of folic acid or zinc supplements in the previous 3 months, vitamin B deciency Duration of study: 1 year Folic acid 5mg/day (n=22) versus placebo (n=25) Zinc sulphate 66 mg/day (n=23) versus placebo (n=25) Zinc sulphate 66 mg/day + folic acid 5 mg/day (n=24) versus placebo (n=25) Duration of treatment: 26 weeks Semen parameters Data in median and range. Attempted to contact authors regarding means and standard deviations. Letter returned to sender.
Participants
Interventions
Outcomes Notes
Risk of bias Bias Authors judgement Support for judgement eligible fertile and subfertile men were randomly assigned according to a simple computer-generated randomisation schedule in four blocks to receive folic acid and placebo, zinc sulphate and placebo, zinc sulphate and folic acid, or placebo and placebo, which resulted in eight subgroups. At the end of the trial, the research fellow received the randomisation list that matched the codes from the hospital pharmacy.
83
Wong 2002
(Continued)
Unclear risk
capsules were coded by the hospital pharmacy according to the randomisation list. Double blind Neither the research fellow and the participants knew whether the participants received folic acid, zinc sulphate or placebo capsules Folic acid and placebo capsules were yellow and identical in appearance. Zinc sulphate and placebo capsules were white and identical in appearance 9 men withdrew from the subfertile arm of the trial, 1 due to side effects (gastrointestinal) and 8 due to lack of motivation. It is unclear which treatment groups these men were randomised to All outcomes except for semen volume is given as medians The protocol is given in methods section main outcomes are semen parameters and these are presented in the report - Our main goal was to investigate the effects of the interventions on semen variables in fertile and subfertile men. Can not denitively assess other bias
Unclear risk
Low risk
Other bias
Unclear risk
Zalata 1998 Methods A randomised pilot study Allocation concealment not mentioned Country: Belgium Population: Men attending andrology clinic Mean age: Not given N=22 Inclusion criteria: None given Exclusion criteria: None given Duration of study: ?4-6 months Acetylcysteine 600mg/day (n=5) versus EFA (DHA 1g, y-linolenic acid, arachidonic acid 100mg) 100mg/day + antioxidant oil mixture and tocopherol + B-carotene (n=12) versus
84
Participants
Interventions
Zalata 1998
(Continued)
Acetylcysteine + EFA + antioxidants (n=5) Duration of treatment: 4-6 months Outcomes Notes Sperm parameters Abstract only. No extractable data. Attempted to contact authors re availability of data as means, if published?, methods of randomisation and allocation concealment.
Risk of bias Bias Authors judgement Support for judgement A prospective randomised pilot study No details of randomisation given No details of allocation concealment No details
Blinding (performance bias and detection Unclear risk bias) All outcomes Incomplete outcome data (attrition bias) All outcomes Selective reporting (reporting bias) Other bias Zavaczki 2003 Methods Unclear risk
No details
Unclear, abstract only Unclear
Randomised placebo controlled clinical study Allocation concealment not mentioned Country: Hungary Population: Subfertile men attending andrology Department of Obstetrics and Gynaecology, University of Szeged Mean age: Treatment group 29.6, placebo group 28.3 years Recruited: N =26 Randomised: N=20 Inclusion criteria: Unsuccessful attempt at pregnancy for over one year. A healthy female partner examined by a gynaecologist. Sperm volume < 2ml and/or sperm concentration <20mill/ml and/or morphology ratio <30% and/or motility <50%. No genital tract infection, no bacteria or fungi in urine or semen. Hormones are within physiological range. Intact renal function. No excessive magnesium intake Exclusion criteria: None mentioned Duration of study: 3 months
Participants
85
Zavaczki 2003
(Continued)
Interventions
Magnesium 3000mg/day (n=10) versus placebo (n=10) Number analysed: N=14 Duration of treatment: 90 days Primary: Semen parameters Secondary: pregnancy and side effects Attempted to contact authors regarding methods of randomisation and allocation concealment
Outcomes
Notes
Risk of bias Bias Authors judgement Support for judgement Patients were randomised and divided into Magnesium or Placebo groups Methods of randomisation are not mentioned Not mentioned The members of Group P received the same number of placebo tablets which closely resembled the Magnerot tablets. 20 were randomised and 14 were analysed. To date 26 patients have participated in the study and 20 men (10 in both groups) have completed the program of treatment. Six patients (2 in group M and 4 in group P were excluded from the program, including ve cases for poor compliance, since they did not attend the control meeting at the end of treatment. One patient from Group M experienced severe diarrhoea and so his treatment was halted. All data for outcomes in the trial were given, unsure if these were prespecied Funding - OKTA program. Tablets supplied by pharmaceutical company
Unclear risk
Unclear risk
Other bias
Unclear risk
86
Characteristics of excluded studies [ordered by study ID]
Study Cavallini 2004a Comhaire 2005 Ebisch 2003 Ebisch 2006 Elgindy 2008 Micic 2001 Nikolova 2007
Reason for exclusion Superceded by full paper of the trial - Cavallini 2004. Continuous data were used in Cavallini 2004 analysis. Used non randomised controls recruited from another unrelated trial. Inappropriate outcomes. Polymorphisms. This study is part of included study Wong 2002. Antioxidant given to the women and not to the men. Not randomised. 105 men in the treatment group and 35 in control. Abstract only. Trial is not randomised, allocation method is by alternation. Translated from Bulgarian by Ivan Sola. 50 of them were randomly invited to participate depending on their order of attendance to the clinic. Not a randomised controlled trial. Conference proceeding superceded by Sigman 2006. Conference proceeding superceded by Rolfe 1999. Trial design is not random. The trial uses alternate allocation, odd and even numbers. Appears to be a report of the trial Nikolova 2007. Conference proceeding superceded by Tremellen 2007. Inapproriate control (anti-inammatory). Treatment is not compared to placebo or another antioxidant. Inappropriate comparison. The same antioxidant is compared at different times - L-carnitine +acetyl-carnitine versus L-carnitine +acetyl-carnitine. Same study as included trial Zalata 1998.
Pawlowicz 2001 Pryor 2003 Rolf 1998 Stanislavov 2009
Tremellen 2006 Vicari 2001 Vicari 2001a
Zalata 1998a
Characteristics of studies awaiting assessment [ordered by study ID]

Dimitriadis 2010 Methods Participants Interventions Randomised controlled trial 75 infertile men with oligioasthenospermia Vardenal (10mg/day)- Group A, sildenal (50mg/day)- Group B, L-carnitine (1000mg/day)- Group C, no treatment- Group D
87
Dimitriadis 2010
(Continued)
Outcomes Notes Wang 2010 Methods Participants Interventions Outcomes Notes
sperm motility, sperm concentration
Random controlled trial 135 men with asthenozoospermia L-carnitine + vitamin E vs vitamin E Pregnancy, adverse events, motile sperm. In Chinese - paper will need translation
Characteristics of ongoing studies [ordered by study ID]

Gonzalez 2009 Trial name or title Methods Assessment of the efcacy of dietary supplement Spermotrend in the treatment of male infertility Randomised, double blind (subject, caregiver, investigator), placebo control, parallel assignment, efcacy study Subfertile men Spermotrend (vitamins plus other antioxidants) versus placebo Parameters of seminal analysis at weeks 24 Fertilisation achievement Presence of mild or severe adverse effects September 2009 Rogelio Gonzalez Sanchez, MD 53 7 838 2626 ext 277 Gynecologic and Obstetric Hospital Havana, Cuba, 10400 Havana, Cuba, 10400
Participants Interventions Outcomes
Starting date Contact information
Notes
88
Kumar 2010 Trial name or title Kumar R. Evaluation of oxidative stress as a cause for idiopathic male infertility and the role of Addyzoa in its treatment Randomised, parallel group, placebo controlled trial. Method of generating randomisation sequence: Random number table method of allocation concealment: Pre-numbered or coded identical containers. Blinding and masking: participant, investigator,outcome assessor and data-entry operator/statistician blinded. Inclusion criteria: Infertile male, normal semen volume with either oligospermia (sperm density between 5 and 20 million/mL) or asthenospermia or teratospermia by WHO criteria Exclusion criteria: Identiable cause for semen abnormality (varicocele, cryptorchidism, orchitis, radiation, chemotherapy, recent febrile illness, hormonal abnormality etc) Intervention1: Addyzoa capsules: 2 BD for 3 months Control Intervention1: placebo: 2 BD for 3 months Improvement in semen parameters timepoint: 3 months and 6 months Pregnancy timepoint: Upto 12 months 15-03-2009 Name: Dr Rajeev Kumar Address: Department of Urology,AIIMS, Ansari Nagar 110029, New Delhi, India Telephone: 01126594884 Email: rajeev02@gmail.com Main ID:CTRI/2009/091/000551. In WHO International Clinical trials Registry Platform.
Methods
Participants
Interventions
Outcomes
Starting date Contact information
Notes
Revel 2006 Trial name or title Revel A. A preliminary study on effect of omega-3 on human sperm. ClinicalTrials.gov. [ClinicalTrials.gov: NCT00479960] Double blind randomised crossover trial Men with oligospermia, asthenoospermia or teratozoospermia Omega 3 fatty acids Sperm count, sperm lipid transition determination Enrollment due to be June 2007 Ariel Revel: arielrevel2@gmail.com Abraham Benchetrit: benchet@012.net.il Email sent to Drs Revel and Benchetrit 17.09.09 asking whether they have any data ready for their trial
Methods Participants Interventions Outcomes Starting date Contact information
Notes
89
Sadeghi 2008 Trial name or title Sadeghi M. Effects of coenzymeQ10 (CoQ10) supplementation on semen quality and seminal oxidative stress of idiopathic oligoasthenoteratozoospermic (iOAT) infertile men. World Health Organization International Clinical Trials Registry Platform Search Portal 2008. [ISRCTN: ISRCTN29954277] Randomised double blind placebo controlled trial Men with idiopathic oligoasthenoteratozoospermia with at least one year infertility Coenzyme10 (Ubiquinone) 200mg 2x/day for 3 months Sperm morphology, motility and sperm DNA fragmentation 1.10.08 sadeghi@avicenna.ac.ir letter sent to author regarding data 17.09.09
Methods Participants Interventions Outcomes Starting date Contact information Notes
Tsafrir 2010 Trial name or title Is a carnitine based food supplement (PorimoreTM) for infertile men superior to folate and zinc with regard to pregnancy rates in intrauterine insemination cycles? Randomised, open label, active control, parallel assignment, efcacy study Infertile men aged 18-50 years Carnitine supplement versus zinc and folate supplement Pregnancy rate and sperm quality parameters February 2010 - expected completion December 2012 Avi Tsafrir, MD tel: 972-6555111 avits@szmc.org.il Shaare-Zedek IVF and Infertility Unit Jerusalem 91031
Methods Participants Interventions Outcomes Starting date Contact information
Notes
90
DATA AND ANALYSES
Comparison 1. Antioxidant(s) versus control
Outcome or subgroup title 1 Live Birth per couple randomised 1.1 Vitamin E versus placebo 1.2 Zinc versus no treatment 2 Pregnancy rate per couple randomised 2.1 Combined antioxidants versus placebo/no treatment 2.2 L acetyl carnitine versus placebo 2.3 L carnitine versus placebo 2.4 L acetyl carnitine + L carnitine versus placebo 2.5 Coenzyme Q10 versus placebo 2.6 Pentoxifylline versus placebo 2.7 Magnesium versus placebo 2.8 Vitamin E versus placebo 2.9 Vitamin C plus Vitamin E versus placebo 2.10 Zinc versus no treatment 3 Adverse event: Miscarriage rate per couple randomised 3.1 Combined antioxidants versus placebo 3.2 Vitamin E versus placebo 3.3 Zinc versus no treatment 4 DNA fragmentation 4.1 Vitamin C + vitamin E versus placebo at 2 months 5 Sperm total motility at 3 months or less 5.1 Docosahexaenoic acid (DHA) 400g/day versus placebo 5.2 Docosahexaenoic acid (DHA) 800mg/day vs placebo 5.3 Ascorbic acid (Vitamin C) 200mg/day versus placebo 5.4 Ascorbic acid (Vitamin C)1000mg/day versus placebo
No. of studies 3 2 1 15 2 1 3 4 2 1 1 1 1 1 3 1 1 1 1 1 10 1
No. of participants 214 117 97 964 100 20 136 183 272 18 20 87 31 97 242 58 87 97 64
Statistical method Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Random, 95% CI)
Effect size 4.85 [1.92, 12.24] 6.44 [1.72, 24.04] 3.67 [1.00, 13.51] 4.18 [2.65, 6.59] 3.97 [1.39, 11.33] 0.61 [0.04, 9.64] 5.08 [1.51, 17.09] 4.48 [1.77, 11.36] 2.16 [0.53, 8.82] 0.0 [0.0, 0.0] 7.39 [0.15, 372.38] 6.64 [1.84, 23.93] 0.0 [0.0, 0.0] 4.75 [1.49, 15.20] 1.54 [0.32, 7.30] 0.48 [0.06, 4.03] 5.43 [0.32, 93.28] 7.24 [0.14, 364.94] Subtotals only -13.80 [-17.50, 10.10] Subtotals only -7.80 [-30.56, 14.96] -15.20 [-34.31, 3.91] 2.0 [-24.07, 28.07] 45.0 [15.25, 74.75]
91
14
Mean Difference (IV, Random, 95% CI)
1 1 1
15 15 15
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI)
5.5 Ascorbic acid (vitamin C) + Vitamin E versus placebo at 2 months 5.6 L-carnitine +l-acetyl carnitine versus placebo at 3 months 5.7 Selenium versus placebo at 3 months 5.8 Combined antioxidants versus placebo at 3 months 5.9 N-acetylcysteine versus placebo at 3 months 5.10 Magnesium versus placebo at 90 days 5.11 L-carnitine versus placebo 5.12 Pentoxifyllin versus no treatment at 3 months 5.13 Zinc versus no treatment at 3 months 5.14 Zinc + Vitamin E versus no treatment at 3 months 5.15 Zinc + Vitamin E + Vitamin C versus no treatment at 3 months 6 Sperm total motility at 6 months 6.1 L-carnitine versus placebo at 6 months 6.2 L-acetyl carnitine versus placebo at 6 months 6.3 L-carnitine + L-acetyl carnitine versus placebo at 6 months 6.4 Selenium versus placebo at 26 weeks (6 months) 6.5 N-acetyl-cysteine versus placebo at 26 weeks (6months) 6.6 Selenium plus N-acetyl-cysteine versus placebo at 26 weeks (6months) 6.7 Coenzyme Q10 versus placebo at 6 months. Safarinejad trial =26 weeks 6.8 Vitamin E versus placebo at 6 months 7 Sperm total motility at 9 months or more 7.1 L-carnitine versus placebo at 9 months 7.2 L-acetyl carnitine versus placebo at 9 months
64
2.90 [-7.76, 13.56]
21
-9.0 [-39.49, 21.49]
1 1 1 1 1 1 1 1 1
25 39 120 20 30 90 14 15 17
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI)
14.90 [-1.04, 30.84] 11.7 [-1.79, 25.19] 10.96 [7.91, 14.01] 14.5 [-6.01, 35.01] 31.30 [31.21, 31.39] 12.77 [9.23, 16.31] 25.0 [9.68, 40.32] 26.0 [9.03, 42.97] 26.0 [8.85, 43.15]
7 1 1 3
20 20 97
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI)
Subtotals only 21.13 [11.50, 30.76] 17.03 [6.90, 27.16] 6.12 [-7.37, 19.60]
1 1 1
155 157 155
3.20 [2.33, 4.07] 1.90 [1.03, 2.77] 6.3 [5.43, 7.17]
272
4.50 [3.93, 5.07]
1 3 1 1
87
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI)
13.0 [7.02, 18.98] Subtotals only 11.54 [1.66, 21.42] 7.84 [-1.41, 17.09]
20 20
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI)
92
7.3 L-carnitine + L-acetyl carnitine versus placebo at 9 months 7.4 Coenzyme Q10 versus placebo at 9 months 8 Sperm concentration at 3 months or less 8.1 Docosahexaenoic acid (DHA) 400g/day versus placebo 8.2 Docosahexaenoic acid (DHA) 800g/day versus placebo 8.3 Magnesium versus placebo at 90 days 8.4 Vitamin C + Vitamin E versus placebo at 2 months 8.5 N-acetylcysteine versus placebo at 3 months 8.6 Pentoxifylline versus placebo at 3 months 8.7 L-carnitine versus placebo 9 Sperm concentration at 6 months 9.1 L-carnitine versus placebo at 6 months 9.2 L-acetyl carnitine versus placebo at 6 months 9.3 L-carnitine + L-acetyl carnitine versus placebo at 6 months 9.4 Selenium versus placebo at 26 weeks (6 months) 9.5 N-acetyl-cysteine versus placebo at 26 weeks (6months) 9.6 Selenium plus N-acetyl-cysteine versus placebo at 26 weeks (6months) 9.7 Coenzyme Q10 versus placebo at 6 months 9.8 Pentoxifylline versus placebo up to 6 months 10 Sperm concentration at 9 months or more 10.1 L-carnitine versus placebo at 9 months 10.2 L-acetyl carnitine versus placebo at 9 months 10.3 L-carnitine + L-acetyl carnitine versus placebo at 9 months
20
6.27 [-3.36, 15.90]
2 7 1
272 320 18
-0.04 [-3.62, 3.55] 6.04 [-5.42, 17.50] -5.30 [-41.09, 30.49] 1.50 [-35.23, 38.23]
19
1 2 1 1 1 6 1 1 2
20 95 120 18 30
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI)
5.20 [-2.61, 13.01] 1.36 [-10.01, 12.72] -0.47 [-6.70, 5.76] 4.30 [-0.69, 9.29] 29.5 [25.39, 33.61] Subtotals only 11.80 [-4.81, 28.41] 5.84 [-10.31, 21.99] 0.63 [-5.93, 7.18]
20 20 76
1 1 1
155 157 155
4.10 [1.94, 6.26] 3.30 [1.24, 5.36] 8.60 [6.40, 10.80]
2 1 3 1 1 1
272 18 332 20 20 20
Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
3.86 [-2.09, 9.81] 2.80 [-2.58, 8.18] 1.61 [0.61, 2.61] 9.27 [-1.51, 20.05] 1.08 [-8.16, 10.32] 3.14 [-7.54, 13.82]
93
10.4 Coenzyme Q10 versus placebo at 9 months or more 11 Total sperm motility over time 11.1 Total sperm motility at 3 months or less 11.2 Total sperm motility at 6 months 11.3 Total sperm motility at 9 months or more 12 Sperm concentration over time 12.1 Sperm concentration at 3 months or less 12.2 Sperm concentration 6 months 12.3 Sperm concentration at 9 months or more 13 Side effects 13.1 gastrointestinal 13.2 euphoria 14 Sperm motility at 3 months or less 14.1 L-carnitine + Acetyl-carnitine versus placebo (median and interquartile range) 14.2 Combined antioxidants versus no treatment 14.3 Vitamin E versus placebo 14.4 L-carnitine versus placebo 14.5 Pentoxifylline versus Placebo 15 Sperm motility at 6 months 15.1 L-carnitine + Acetyl-carnitine versus placebo (median and interquartile range) 15.2 Pentoxifylline versus placebo 15.3 Folic acid versus placebo 15.4 Zinc versus placebo 15.5 Zinc + folic acid versus placebo 16 Sperm concentration at 3 months or less 16.1 L-carnitine + Acetyl-carnitine versus placebo (median and interquartile range) 16.2 Vitamin E versus placebo
2 16 10 7 3 12 7 6 3 6 6 1
272
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Mean Difference (IV, Random, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Other data Other data
1.53 [0.51, 2.55] Subtotals only 13.47 [3.45, 23.49] 6.28 [4.36, 8.21] 1.07 [-2.29, 4.43] Subtotals only 6.79 [-5.09, 18.66] 5.12 [2.53, 7.72] 1.56 [0.55, 2.57] Subtotals only 1.63 [0.48, 5.58] 1.21 [0.16, 9.01] No numeric data No numeric data
454 700 302
302 610 302
426 86
Other data Other data Other data Other data Other data Other data
No numeric data No numeric data No numeric data No numeric data No numeric data No numeric data
Other data Other data Other data Other data Other data Other data
No numeric data No numeric data No numeric data No numeric data No numeric data No numeric data
Other data
No numeric data
94
16.3 L-carnitine versus placebo 17 Sperm concentration at 6 months 17.1 L-carnitine + acetyl-carnitine versus placebo 17.2 Folic acid versus Placebo 17.3 Zinc versus Placebo 18 Pregnancy 18.1 Pregnancy rate of trials that also report live birth 18.2 Pregnancy rate only reported
Other data Other data Other data Other data Other data Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI)
No numeric data No numeric data No numeric data No numeric data No numeric data 4.40 [2.78, 6.97] 5.53 [2.34, 13.08] 4.02 [2.34, 6.92]
15 2 13
963 184 779
Comparison 2. Live birth versus pregnancy
Outcome or subgroup title 1 Live birth 1.1 live birth 1.2 pregnancy
No. of studies 2 2 2
No. of participants
Statistical method Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI)
Effect size Subtotals only 4.73 [1.82, 12.27] 5.53 [2.34, 13.08]
184 184
Comparison 3. Head to head antioxidant(s)
Outcome or subgroup title 1 Pregnancy per couple randomised 1.1 L acetyl carnitine + L carnitine versus L acetyl carnitine 1.2 L acetyl carnitine + L carnitine versus L- carnitine 1.3 L acetyl carnitine + L carnitine versus Vitamin E + Vitamin C 2 Sperm total motility at 3 months or less 2.1 Ethylcysteine 600mg/day vs Vitamin E 2.2 Docosahexaenoic acid (DHA) 400g/day vs Docosahexaenoic acid 800mg/day
No. of studies 2 1
No. of participants
Statistical method Peto Odds Ratio (Peto, Fixed, 95% CI)
Effect size Subtotals only 2.66 [0.27, 25.80]
22
Peto Odds Ratio (Peto, Fixed, 95% CI)
1 1
23 150
Peto Odds Ratio (Peto, Fixed, 95% CI) Peto Odds Ratio (Peto, Fixed, 95% CI)
3.89 [0.51, 29.76] 2.86 [0.86, 9.49]
7 1 1 10 19
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
Subtotals only -1.90 [-41.97, 38.17] 7.40 [-11.35, 26.15]
95
2.3 Ascorbic Acid 200mg/day versus Ascorbic Acid 1000mg/day 2.4 Vitamin E + Selenium versus Vitamin B at 3 months 2.5 Zinc versus Zinc + Vitamin E at 3 months 2.6 Zinc versus Zinc + Vitamin E + Vitamin C at 3 months 2.7 Zinc + Vitamin E versus Zinc + Vitamin E + Vitamin C at 3 months 2.8 Selenium versus combined antioxidants 2.9 L acetyl carnitine + L carnitine versus Vitamin E + Vitamin C 3 Sperm total motility at 6 months 3.1 L-aceytl carnitine +L-carnitine versus L-carnitine at 6 months 3.2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 6 months 3.3 Selenium versus N-acetyl-cysteine at 26 weeks (6 months) 3.4 Selenium versus selenium plus N-acetyl-cysteine at 26 weeks (6 months) 3.5 N-acetyl-cysteine vs selenium plus N-acetyl-cysteine at 26 weeks 4 Sperm total motility at 9 months or more 4.1 L-aceytl carnitine +L-carnitine versus L-carnitine at 9 months 4.2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 9 months 5 Sperm concentration at 3 months or less 5.1 Ethylcysteine 600mg/day vs Vitamin E 5.2 Docosahexaenoic acid (DHA) 400g/day versus Docosahexaenoic acid (DHA) 800g/day
20
Mean Difference (IV, Fixed, 95% CI)
-43.0 [-67.10, 18.90] 0.0 [-10.71, 10.71] -1.0 [-13.00, 13.00] -1.0 [-19.66, 17.66]
1 1 1
54 18 12
18
0.0 [-18.97, 18.97]
1 1
46 150
Mean Difference (IV, Fixed, 95% CI) Mean Difference (IV, Fixed, 95% CI)
3.20 [-10.13, 16.53] 23.05 [20.23, 25.87]
2 1
22
Subtotals only -3.46 [-11.41, 4.49]
23
0.64 [-7.58, 8.86]
176
1.30 [0.39, 2.21]
116
-3.10 [-4.16, -2.04]
176
-4.40 [-5.31, -3.49]
1 1 22
Subtotals only -5.27 [-12.61, 2.07]
23
-1.57 [-7.70, 4.56]
3 1 1 10 19
Subtotals only 2.20 [-16.65, 21.05] -6.80 [-41.87, 28.27]
96
5.3 L-carnitine versus Vitamin E + Vitamin C 6 Sperm concentration at 6 months 6.1 L-aceytl carnitine +L-carnitine versus L-carnitine at 6 months 6.2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 6 months 6.3 Selenium versus N-acetyl-cysteine at 26 weeks (6 months) 6.4 Selenium versus selenium plus N-acetyl-cysteine at 26 weeks (6 months) 6.5 N-acetyl-cysteine vs selenium plus N-acetyl-cysteine at 26 weeks 7 Sperm concentration at 9 months or more 7.1 L-aceytl carnitine +L-carnitine versus L-carnitine at 9 months 7.2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 9 months 8 Side effects 8.1 L acetyl carnitine + L carnitine versus Vitamin E + Vitamin C 9 Sperm concentration at 6 months 9.1 Zinc versus Folic acid 9.2 Zinc versus Zinc + Folic acid 10 Sperm motility at 6 months 10.1 Zinc versus Folic acid
1 2 1
63
15.5 [12.49, 18.51] Subtotals only -8.13 [-24.42, 8.16]
22
23
-2.17 [-17.40, 13.06] 0.80 [-1.10, 2.70]
176
116
-4.5 [-6.90, -2.10]
176
-5.3 [-7.29, -3.31]
1 1 22
Subtotals only -6.13 [-18.44, 6.18]
23
2.06 [-8.33, 12.45]
1 1
150 150
Odds Ratio (M-H, Fixed, 95% CI) Odds Ratio (M-H, Fixed, 95% CI)
0.0 [0.0, 0.0] 0.0 [0.0, 0.0]
Other data Other data Other data Other data Other data
No numeric data No numeric data No numeric data No numeric data No numeric data
97
Analysis 1.1. Comparison 1 Antioxidant(s) versus control, Outcome 1 Live Birth per couple randomised.
Review: Antioxidants for male subfertility
Comparison: 1 Antioxidant(s) versus control Outcome: 1 Live Birth per couple randomised
Study or subgroup
Experimental n/N
Control n/N
Peto Odds Ratio Peto,Fixed,95% CI
Weight
1 Vitamin E versus placebo Kessopoulou 1995 Suleiman 1996 1/15 9/52 0/15 0/35 5.6 % 43.8 % 7.39 [ 0.15, 372.38 ] 6.33 [ 1.56, 25.63 ]
Subtotal (95% CI)

Total events: 10 (Experimental), 0 (Control)
67
50
49.4 %
6.44 [ 1.72, 24.04 ]
Heterogeneity: Chi2 = 0.01, df = 1 (P = 0.94); I2 =0.0% Test for overall effect: Z = 2.77 (P = 0.0056) 2 Zinc versus no treatment Omu 1998 8/49 2/48 50.6 % 3.67 [ 1.00, 13.51 ]
Subtotal (95% CI)

Total events: 8 (Experimental), 2 (Control) Heterogeneity: not applicable Test for overall effect: Z = 1.96 (P = 0.050)
49
48
50.6 %
3.67 [ 1.00, 13.51 ]
Total (95% CI)
116
98
100.0 %
4.85 [ 1.92, 12.24 ]
Total events: 18 (Experimental), 2 (Control) Heterogeneity: Chi2 = 0.36, df = 2 (P = 0.84); I2 =0.0% Test for overall effect: Z = 3.34 (P = 0.00084) Test for subgroup differences: Chi2 = 0.35, df = 1 (P = 0.55), I2 =0.0%
0.001 0.01 0.1 Favours control
10 100 1000 Favours antioxidant
98
Analysis 1.2. Comparison 1 Antioxidant(s) versus control, Outcome 2 Pregnancy rate per couple randomised.
Comparison: 1 Antioxidant(s) versus control Outcome: 2 Pregnancy rate per couple randomised
Study or subgroup
Experimental n/N
Control n/N
1 Combined antioxidants versus placebo/no treatment Galatioto 2008 Tremellen 2007 1/20 20/38 0/22 4/20 8.17 [ 0.16, 413.39 ] 3.75 [ 1.26, 11.16 ]
Subtotal (95% CI)

58
42
3.97 [ 1.39, 11.33 ]
Heterogeneity: Chi2 = 0.14, df = 1 (P = 0.71); I2 =0.0% Test for overall effect: Z = 2.57 (P = 0.010) 2 L acetyl carnitine versus placebo Balercia 2005 2/15 1/5 0.61 [ 0.04, 9.64 ]
Subtotal (95% CI)

Total events: 2 (Experimental), 1 (Control) Heterogeneity: not applicable Test for overall effect: Z = 0.35 (P = 0.72) 3 L carnitine versus placebo Balercia 2005 Lenzi 2003 Peivandi 2010
15
0.61 [ 0.04, 9.64 ]
2/15 6/43 3/15
1/5 0/43 0/15
0.61 [ 0.04, 9.64 ] 8.37 [ 1.61, 43.58 ] 8.57 [ 0.82, 89.45 ]
Subtotal (95% CI)

73
63
5.08 [ 1.51, 17.09 ]
Heterogeneity: Chi2 = 2.81, df = 2 (P = 0.25); I2 =29% Test for overall effect: Z = 2.63 (P = 0.0086) 4 L acetyl carnitine + L carnitine versus placebo Balercia 2005 Cavallini 2004 Lenzi 2004 Sigman 2006 5/15 9/39 4/30 1/12 1/5 1/47 0/26 1/9 1.83 [ 0.21, 15.73 ] 7.50 [ 2.01, 27.98 ] 7.20 [ 0.95, 54.34 ] 0.74 [ 0.04, 13.02 ]
Subtotal (95% CI)

96
87
4.48 [ 1.77, 11.36 ]
Heterogeneity: Chi2 = 2.99, df = 3 (P = 0.39); I2 =0.0% Test for overall effect: Z = 3.17 (P = 0.0015) 5 Coenzyme Q10 versus placebo
0.002
0.1
10
500
Favours control
Favours antioxidant
(Continued . . . )
99
Study or subgroup
Experimental n/N
Control n/N 3/30 0/106
(. . . Continued) Peto Odds Ratio

Peto,Fixed,95% CI 2.16 [ 0.53, 8.82 ] 0.0 [ 0.0, 0.0 ]
Balercia 2009 Safarinejad 2009a
6/30 0/106
Subtotal (95% CI)

136
136
2.16 [ 0.53, 8.82 ]
Heterogeneity: Chi2 = 0.0, df = 0 (P = 1.00); I2 =0.0% Test for overall effect: Z = 1.08 (P = 0.28) 6 Pentoxifylline versus placebo Wang 1983 0/11 0/7 0.0 [ 0.0, 0.0 ]
Subtotal (95% CI)

Total events: 0 (Experimental), 0 (Control) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) 7 Magnesium versus placebo Zavaczki 2003
11
0.0 [ 0.0, 0.0 ]
1/10
0/10
7.39 [ 0.15, 372.38 ]
Subtotal (95% CI)

Total events: 1 (Experimental), 0 (Control) Heterogeneity: not applicable Test for overall effect: Z = 1.00 (P = 0.32) 8 Vitamin E versus placebo Suleiman 1996
10
10
7.39 [ 0.15, 372.38 ]
11/52
0/35
6.64 [ 1.84, 23.93 ]
Subtotal (95% CI)

Total events: 11 (Experimental), 0 (Control) Heterogeneity: not applicable Test for overall effect: Z = 2.89 (P = 0.0038) 9 Vitamin C plus Vitamin E versus placebo Rolf 1999
52
35
6.64 [ 1.84, 23.93 ]
0/15
0/16
0.0 [ 0.0, 0.0 ]
Subtotal (95% CI)

Total events: 0 (Experimental), 0 (Control) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001) 10 Zinc versus no treatment Omu 1998
15
16
0.0 [ 0.0, 0.0 ]
11/49
2/48
4.75 [ 1.49, 15.20 ]
Subtotal (95% CI)

49
48
4.75 [ 1.49, 15.20 ]
Total (95% CI)

515
449
4.18 [ 2.65, 6.59 ]
Heterogeneity: Chi2 = 9.41, df = 13 (P = 0.74); I2 =0.0% Test for overall effect: Z = 6.14 (P < 0.00001) Test for subgroup differences: Chi2 = 3.47, df = 7 (P = 0.84), I2 =0.0%
0.002
0.1
10
500
Favours control
Favours antioxidant
100
Analysis 1.3. Comparison 1 Antioxidant(s) versus control, Outcome 3 Adverse event: Miscarriage rate per couple randomised.
Comparison: 1 Antioxidant(s) versus control Outcome: 3 Adverse event: Miscarriage rate per couple randomised
Study or subgroup
Experimental n/N
Control n/N
Weight
1 Combined antioxidants versus placebo Tremellen 2007 2/38 2/20 54.1 % 0.48 [ 0.06, 4.03 ]
Subtotal (95% CI)

Total events: 2 (Experimental), 2 (Control) Heterogeneity: not applicable Test for overall effect: Z = 0.67 (P = 0.50) 2 Vitamin E versus placebo Suleiman 1996
38
20
54.1 %
0.48 [ 0.06, 4.03 ]
2/52
0/35
30.1 %
5.43 [ 0.32, 93.28 ]
Subtotal (95% CI)

Total events: 2 (Experimental), 0 (Control) Heterogeneity: not applicable Test for overall effect: Z = 1.17 (P = 0.24) 3 Zinc versus no treatment Omu 1998
52
35
30.1 %
5.43 [ 0.32, 93.28 ]
1/49
0/48
15.8 %
7.24 [ 0.14, 364.94 ]
Subtotal (95% CI)

49
48
15.8 %
7.24 [ 0.14, 364.94 ]
Total (95% CI)

139
103
100.0 %
1.54 [ 0.32, 7.30 ]
Heterogeneity: Chi2 = 2.50, df = 2 (P = 0.29); I2 =20% Test for overall effect: Z = 0.54 (P = 0.59) Test for subgroup differences: Chi2 = 2.50, df = 2 (P = 0.29), I2 =20%
0.01
0.1
10
100
Favours antioxidant
Favours control
101
Analysis 1.4. Comparison 1 Antioxidant(s) versus control, Outcome 4 DNA fragmentation.

Comparison: 1 Antioxidant(s) versus control Outcome: 4 DNA fragmentation
Study or subgroup
Experimental N Mean(SD)
Control N Mean(SD)
Mean Difference IV,Fixed,95% CI
Weight
1 Vitamin C + vitamin E versus placebo at 2 months Greco 2005 32 9.1 (7.2) 32 22.9 (7.9) 100.0 % -13.80 [ -17.50, -10.10 ]
-50
-25
25
50
Favours antioxidant
Favours control
Analysis 1.5. Comparison 1 Antioxidant(s) versus control, Outcome 5 Sperm total motility at 3 months or less.
Comparison: 1 Antioxidant(s) versus control Outcome: 5 Sperm total motility at 3 months or less
Study or subgroup
Control N Mean(SD)
Mean Difference IV,Random,95% CI
Weight
1 Docosahexaenoic acid (DHA) 400g/day versus placebo Conquer 2000 9 39.4 (24.3) 5 47.2 (18.6) 100.0 % -7.80 [ -30.56, 14.96 ]
Subtotal (95% CI)

Heterogeneity: not applicable
100.0 % -7.80 [ -30.56, 14.96 ]
Test for overall effect: Z = 0.67 (P = 0.50) 2 Docosahexaenoic acid (DHA) 800mg/day vs placebo Conquer 2000 10 32 (16.1) 5 47.2 (18.6) 100.0 % -15.20 [ -34.31, 3.91 ]
Subtotal (95% CI)

10
100.0 % -15.20 [ -34.31, 3.91 ]
Test for overall effect: Z = 1.56 (P = 0.12) 3 Ascorbic acid (Vitamin C) 200mg/day versus placebo Dawson 1990 10 51 (22.1) 5 49 (25.3) 100.0 % 2.00 [ -24.07, 28.07 ]
Subtotal (95% CI)

10
100.0 %
2.00 [ -24.07, 28.07 ]
-50
-25
25
50
Favours control
Favours antioxidant
(Continued . . . )
102
Study or subgroup
Control N Mean(SD)
Weight
(. . . Continued) Mean Difference

IV,Random,95% CI
Test for overall effect: Z = 0.15 (P = 0.88) 4 Ascorbic acid (Vitamin C)1000mg/day versus placebo Dawson 1990 10 94 (32) 5 49 (25.3) 100.0 % 45.00 [ 15.25, 74.75 ]
Subtotal (95% CI)

10
100.0 %
45.00 [ 15.25, 74.75 ]
Test for overall effect: Z = 2.96 (P = 0.0030) 5 Ascorbic acid (vitamin C) + Vitamin E versus placebo at 2 months Greco 2005 32 41.6 (22) 32 38.7 (21.5) 100.0 % 2.90 [ -7.76, 13.56 ]
Subtotal (95% CI)

32
32
100.0 %
2.90 [ -7.76, 13.56 ]
Test for overall effect: Z = 0.53 (P = 0.59) 6 L-carnitine +l-acetyl carnitine versus placebo at 3 months Sigman 2006 12 28.6 (38.1) 9 37.6 (33) 100.0 % -9.00 [ -39.49, 21.49 ]
Subtotal (95% CI)

12
100.0 % -9.00 [ -39.49, 21.49 ]
Test for overall effect: Z = 0.58 (P = 0.56) 7 Selenium versus placebo at 3 months Scott 1998 16 30.2 (22.8) 9 15.3 (17.4) 100.0 % 14.90 [ -1.04, 30.84 ]
Subtotal (95% CI)

16
100.0 %
14.90 [ -1.04, 30.84 ]
Test for overall effect: Z = 1.83 (P = 0.067) 8 Combined antioxidants versus placebo at 3 months Scott 1998 30 27 (20.3) 9 15.3 (17.4) 100.0 % 11.70 [ -1.79, 25.19 ]
Subtotal (95% CI)

30
100.0 %
11.70 [ -1.79, 25.19 ]
Test for overall effect: Z = 1.70 (P = 0.089) 9 N-acetylcysteine versus placebo at 3 months Ciftci 2009 60 31.29 (8.62) 60 20.33 (8.43) 100.0 % 10.96 [ 7.91, 14.01 ]
Subtotal (95% CI)

60
60
100.0 %
10.96 [ 7.91, 14.01 ]
Test for overall effect: Z = 7.04 (P < 0.00001) 10 Magnesium versus placebo at 90 days Zavaczki 2003 10 33.5 (29.8) 10 19 (14.4) 100.0 % 14.50 [ -6.01, 35.01 ]
Subtotal (95% CI)

10
10
100.0 %
14.50 [ -6.01, 35.01 ]
Test for overall effect: Z = 1.39 (P = 0.17) 11 L-carnitine versus placebo Peivandi 2010 15 48.3 (0.16) 15 17 (0.09) 100.0 % 31.30 [ 31.21, 31.39 ]
Subtotal (95% CI)

15
15
100.0 %
31.30 [ 31.21, 31.39 ]
-50
-25
25
50
Favours control
Favours antioxidant
(Continued . . . )
Study or subgroup
Control N Mean(SD)
Weight

IV,Random,95% CI
Test for overall effect: Z = 660.35 (P < 0.00001) 12 Pentoxifyllin versus no treatment at 3 months Micic 1988 51 31.21 (10.66) 39 18.44 (6.35) 100.0 % 12.77 [ 9.23, 16.31 ]
Subtotal (95% CI)

51
39
100.0 %
12.77 [ 9.23, 16.31 ]
Test for overall effect: Z = 7.07 (P < 0.00001) 13 Zinc versus no treatment at 3 months Omu 2008 11 49 (12) 3 24 (12) 100.0 % 25.00 [ 9.68, 40.32 ]
Subtotal (95% CI)

11
100.0 %
25.00 [ 9.68, 40.32 ]
Test for overall effect: Z = 3.20 (P = 0.0014) 14 Zinc + Vitamin E versus no treatment at 3 months Omu 2008 12 50 (18) 3 24 (12) 100.0 % 26.00 [ 9.03, 42.97 ]
Subtotal (95% CI)

12
100.0 %
26.00 [ 9.03, 42.97 ]
Test for overall effect: Z = 3.00 (P = 0.0027) 15 Zinc + Vitamin E + Vitamin C versus no treatment at 3 months Omu 2008 14 50 (20) 3 24 (12) 100.0 % 26.00 [ 8.85, 43.15 ]
Subtotal (95% CI)

14
100.0 %
26.00 [ 8.85, 43.15 ]
Test for overall effect: Z = 2.97 (P = 0.0030)
-50
-25
25
50
Favours control
Favours antioxidant
104
Analysis 1.6. Comparison 1 Antioxidant(s) versus control, Outcome 6 Sperm total motility at 6 months.
Comparison: 1 Antioxidant(s) versus control Outcome: 6 Sperm total motility at 6 months
Study or subgroup
Control N Mean(SD)
Weight
1 L-carnitine versus placebo at 6 months Balercia 2005 15 64.53 (8.41) 5 43.4 (9.85) 100.0 % 21.13 [ 11.50, 30.76 ]
Subtotal (95% CI)

15
100.0 % 21.13 [ 11.50, 30.76 ]
Test for overall effect: Z = 4.30 (P = 0.000017) 2 L-acetyl carnitine versus placebo at 6 months Balercia 2005 15 60.43 (10.46) 5 43.4 (9.85) 100.0 % 17.03 [ 6.90, 27.16 ]
Subtotal (95% CI)

15
100.0 %
17.03 [ 6.90, 27.16 ]
Test for overall effect: Z = 3.30 (P = 0.00098) 3 L-carnitine + L-acetyl carnitine versus placebo at 6 months Balercia 2005 Lenzi 2004 Sigman 2006 15 30 12 61.07 (9.07) 31.11 (13.46) 32.3 (24.2) 5 26 9 43.4 (9.85) 29.55 (9.5) 40 (33) 38.4 % 43.9 % 17.7 % 17.67 [ 7.89, 27.45 ] 1.56 [ -4.48, 7.60 ] -7.70 [ -33.24, 17.84 ]
Subtotal (95% CI)
57
40
100.0 %
6.12 [ -7.37, 19.60 ]
Heterogeneity: Tau2 = 98.36; Chi2 = 8.61, df = 2 (P = 0.01); I2 =77% Test for overall effect: Z = 0.89 (P = 0.37) 4 Selenium versus placebo at 26 weeks (6 months) Safarinejad 2009 116 26.1 (2.9) 39 22.9 (2.2) 100.0 % 3.20 [ 2.33, 4.07 ]
Subtotal (95% CI)

116
39
100.0 %
3.20 [ 2.33, 4.07 ]
Test for overall effect: Z = 7.22 (P < 0.00001) 5 N-acetyl-cysteine versus placebo at 26 weeks (6months) Safarinejad 2009 118 24.8 (2.9) 39 22.9 (2.2) 100.0 % 1.90 [ 1.03, 2.77 ]
Subtotal (95% CI)

118
39
100.0 %
1.90 [ 1.03, 2.77 ]
Test for overall effect: Z = 4.30 (P = 0.000017) 6 Selenium plus N-acetyl-cysteine versus placebo at 26 weeks (6months) Safarinejad 2009 116 29.2 (2.9) 39 22.9 (2.2) 100.0 % 6.30 [ 5.43, 7.17 ]
Subtotal (95% CI)

116
39
100.0 %
6.30 [ 5.43, 7.17 ]
Test for overall effect: Z = 14.21 (P < 0.00001) 7 Coenzyme Q10 versus placebo at 6 months. Safarinejad trial =26 weeks
-20
-10
10
20
Favours control
Favours antioxidant
(Continued . . . )
105
Study or subgroup
Experimental N Mean(SD) 39.41 (6.8) 27.6 (2.2)
Control N 30 106 Mean(SD) 34.93 (8.04) 23.1 (2.1)
Weight

IV,Random,95% CI 4.48 [ 0.71, 8.25 ] 4.50 [ 3.92, 5.08 ]
Balercia 2009 Safarinejad 2009a
30 106
2.3 % 97.7 %
Subtotal (95% CI)
136
136
100.0 %
4.50 [ 3.93, 5.07 ]
Heterogeneity: Tau2 = 0.0; Chi2 = 0.00, df = 1 (P = 0.99); I2 =0.0% Test for overall effect: Z = 15.41 (P < 0.00001) 8 Vitamin E versus placebo at 6 months Suleiman 1996 52 48.9 (15.5) 35 35.9 (12.8) 100.0 % 13.00 [ 7.02, 18.98 ]
Subtotal (95% CI)

52
35
100.0 %
13.00 [ 7.02, 18.98 ]
-20
-10
10
20
Favours control
Favours antioxidant
Analysis 1.7. Comparison 1 Antioxidant(s) versus control, Outcome 7 Sperm total motility at 9 months or more.
Comparison: 1 Antioxidant(s) versus control Outcome: 7 Sperm total motility at 9 months or more
Study or subgroup
Control N Mean(SD)
Weight
1 L-carnitine versus placebo at 9 months Balercia 2005 15 54.27 (8.96) 5 42.73 (10.02) 100.0 % 11.54 [ 1.66, 21.42 ]
Subtotal (95% CI)

15
100.0 %
11.54 [ 1.66, 21.42 ]
Test for overall effect: Z = 2.29 (P = 0.022) 2 L-acetyl carnitine versus placebo at 9 months Balercia 2005 15 50.57 (5.71) 5 42.73 (10.02) 100.0 % 7.84 [ -1.41, 17.09 ]
Subtotal (95% CI)

15
100.0 %
7.84 [ -1.41, 17.09 ]
Test for overall effect: Z = 1.66 (P = 0.097) 3 L-carnitine + L-acetyl carnitine versus placebo at 9 months Balercia 2005 15 49 (7.8) 5 42.73 (10.02)
-20 -10 0 10 20
100.0 %
6.27 [ -3.36, 15.90 ]
Favours control
Favours antioxidant
(Continued . . . )
106
Study or subgroup
Control N Mean(SD)
Weight

IV,Random,95% CI
Subtotal (95% CI)

15
100.0 %
6.27 [ -3.36, 15.90 ]
Test for overall effect: Z = 1.28 (P = 0.20) 4 Coenzyme Q10 versus placebo at 9 months Balercia 2009 Safarinejad 2009a 30 106 32.93 (6.33) 24.2 (2.1) 30 106 35.3 (8) 22.8 (2.2) 38.1 % 61.9 % -2.37 [ -6.02, 1.28 ] 1.40 [ 0.82, 1.98 ]
Subtotal (95% CI)
136
136
100.0 %
-0.04 [ -3.62, 3.55 ]
Heterogeneity: Tau2 = 5.33; Chi2 = 4.00, df = 1 (P = 0.05); I2 =75% Test for overall effect: Z = 0.02 (P = 0.98)
-20
-10
10
20
Favours control
Favours antioxidant
Analysis 1.8. Comparison 1 Antioxidant(s) versus control, Outcome 8 Sperm concentration at 3 months or less.
Comparison: 1 Antioxidant(s) versus control Outcome: 8 Sperm concentration at 3 months or less
Study or subgroup
Control N Mean(SD)
Weight
1 Docosahexaenoic acid (DHA) 400g/day versus placebo Conquer 2000 9 37.8 (36.9) 9 43.1 (40.5) 6.2 % -5.30 [ -41.09, 30.49 ]
Subtotal (95% CI)

6.2 % -5.30 [ -41.09, 30.49 ]
Test for overall effect: Z = 0.29 (P = 0.77) 2 Docosahexaenoic acid (DHA) 800g/day versus placebo Conquer 2000 10 44.6 (41.1) 9 43.1 (40.5) 6.0 % 1.50 [ -35.23, 38.23 ]
Subtotal (95% CI)

10
6.0 %
1.50 [ -35.23, 38.23 ]
Test for overall effect: Z = 0.08 (P = 0.94) 3 Magnesium versus placebo at 90 days Zavaczki 2003 10 16.1 (10.2) 10 10.9 (7.4) 14.7 % 5.20 [ -2.61, 13.01 ]
Subtotal (95% CI)
10
10
-50 -25 0 25 50
14.7 %
5.20 [ -2.61, 13.01 ]
Favours control
Favours antioxidant
(Continued . . . )
107
Study or subgroup
Control N Mean(SD)
Weight

IV,Random,95% CI
Heterogeneity: not applicable Test for overall effect: Z = 1.30 (P = 0.19) 4 Vitamin C + Vitamin E versus placebo at 2 months Greco 2005 Rolf 1999 32 15 27.5 (24.6) 20.6 (13.5) 32 16 20.3 (21.2) 25 (17.8) 13.7 % 13.7 % 7.20 [ -4.05, 18.45 ] -4.40 [ -15.48, 6.68 ]
Subtotal (95% CI)
47
48
27.4 %
1.36 [ -10.01, 12.72 ]
Heterogeneity: Tau2 = 34.83; Chi2 = 2.07, df = 1 (P = 0.15); I2 =52% Test for overall effect: Z = 0.23 (P = 0.82) 5 N-acetylcysteine versus placebo at 3 months Ciftci 2009 60 21.85 (11.08) 60 22.32 (22) 15.0 % -0.47 [ -6.70, 5.76 ]
Subtotal (95% CI)

60
60
15.0 %
-0.47 [ -6.70, 5.76 ]
Test for overall effect: Z = 0.15 (P = 0.88) 6 Pentoxifylline versus placebo at 3 months Wang 1983 11 12.8 (4.7) 7 8.5 (5.6) 15.3 % 4.30 [ -0.69, 9.29 ]
Subtotal (95% CI)

11
15.3 %
4.30 [ -0.69, 9.29 ]
Test for overall effect: Z = 1.69 (P = 0.091) 7 L-carnitine versus placebo Peivandi 2010 15 46 (3.62) 15 16.5 (7.26) 15.4 % 29.50 [ 25.39, 33.61 ]
Subtotal (95% CI)

15
15
15.4 %
29.50 [ 25.39, 33.61 ]
Test for overall effect: Z = 14.08 (P < 0.00001)
Total (95% CI)
162
158
100.0 %
6.04 [ -5.42, 17.50 ]
Heterogeneity: Tau2 = 217.11; Chi2 = 107.29, df = 7 (P<0.00001); I2 =93% Test for overall effect: Z = 1.03 (P = 0.30) Test for subgroup differences: Chi2 = 100.86, df = 6 (P = 0.0), I2 =94%
-50
-25
25
50
Favours control
Favours antioxidant
108
Analysis 1.9. Comparison 1 Antioxidant(s) versus control, Outcome 9 Sperm concentration at 6 months.
Comparison: 1 Antioxidant(s) versus control Outcome: 9 Sperm concentration at 6 months
Study or subgroup
Control N Mean(SD)
Weight
1 L-carnitine versus placebo at 6 months Balercia 2005 15 45.53 (21.42) 5 33.73 (14.36) 100.0 % 11.80 [ -4.81, 28.41 ]
Subtotal (95% CI)

15
100.0 % 11.80 [ -4.81, 28.41 ]
Subtotal (95% CI)

15
100.0 % 5.84 [ -10.31, 21.99 ]
Test for overall effect: Z = 0.71 (P = 0.48) 3 L-carnitine + L-acetyl carnitine versus placebo at 6 months Balercia 2005 Lenzi 2004 15 30 37.4 (16.42) 22.09 (9.05) 5 26 33.73 (14.36) 22.17 (16.95) 18.9 % 81.1 % 3.67 [ -11.41, 18.75 ] -0.08 [ -7.36, 7.20 ]
Subtotal (95% CI)
45
31
100.0 %
0.63 [ -5.93, 7.18 ]
Heterogeneity: Tau2 = 0.0; Chi2 = 0.19, df = 1 (P = 0.66); I2 =0.0% Test for overall effect: Z = 0.19 (P = 0.85) 4 Selenium versus placebo at 26 weeks (6 months) Safarinejad 2009 116 27.6 (6.4) 39 23.5 (5.8) 100.0 % 4.10 [ 1.94, 6.26 ]
Subtotal (95% CI)

116
39
100.0 %
4.10 [ 1.94, 6.26 ]
Test for overall effect: Z = 3.72 (P = 0.00020) 5 N-acetyl-cysteine versus placebo at 26 weeks (6months) Safarinejad 2009 118 26.8 (5.3) 39 23.5 (5.8) 100.0 % 3.30 [ 1.24, 5.36 ]
Subtotal (95% CI)

118
39
100.0 %
3.30 [ 1.24, 5.36 ]
Test for overall effect: Z = 3.15 (P = 0.0017) 6 Selenium plus N-acetyl-cysteine versus placebo at 26 weeks (6months) Safarinejad 2009 116 32.1 (6.8) 39 23.5 (5.8) 100.0 % 8.60 [ 6.40, 10.80 ]
Subtotal (95% CI)

116
39
100.0 %
8.60 [ 6.40, 10.80 ]
Test for overall effect: Z = 7.66 (P < 0.00001) 7 Coenzyme Q10 versus placebo at 6 months Balercia 2009 30 44.93 (19.3) 30 46.37 (19.77)
-20 -10 0 10 20
24.7 %
-1.44 [ -11.33, 8.45 ]
Favours control
Favours antioxidant
(Continued . . . )
109
Study or subgroup
Experimental N Mean(SD) 26.4 (4.4)
Control N 106 Mean(SD) 20.8 (4.3)
Weight

IV,Random,95% CI 5.60 [ 4.43, 6.77 ]
Safarinejad 2009a
106
75.3 %
Subtotal (95% CI)
136
136
100.0 %
3.86 [ -2.09, 9.81 ]
Heterogeneity: Tau2 = 11.88; Chi2 = 1.92, df = 1 (P = 0.17); I2 =48% Test for overall effect: Z = 1.27 (P = 0.20) 8 Pentoxifylline versus placebo up to 6 months Wang 1983 11 11.1 (5.8) 7 8.3 (5.6) 100.0 % 2.80 [ -2.58, 8.18 ]
Subtotal (95% CI)

11
100.0 %
2.80 [ -2.58, 8.18 ]
-20
-10
10
20
Favours control
Favours antioxidant
Analysis 1.10. Comparison 1 Antioxidant(s) versus control, Outcome 10 Sperm concentration at 9 months or more.
Comparison: 1 Antioxidant(s) versus control Outcome: 10 Sperm concentration at 9 months or more
Study or subgroup
Control N Mean(SD)
Weight
1 L-carnitine versus placebo at 9 months Balercia 2005 15 39.4 (13.93) 5 30.13 (9.3) 0.9 % 9.27 [ -1.51, 20.05 ]
Subtotal (95% CI)

15
0.9 %
9.27 [ -1.51, 20.05 ]
Subtotal (95% CI)

15
1.2 %
1.08 [ -8.16, 10.32 ]
Test for overall effect: Z = 0.23 (P = 0.82) 3 L-carnitine + L-acetyl carnitine versus placebo at 9 months Balercia 2005 15 33.27 (13.62) 5 30.13 (9.3) 0.9 % 3.14 [ -7.54, 13.82 ]
Subtotal (95% CI)
15
5
-20 -10 0 10 20
0.9 %
3.14 [ -7.54, 13.82 ]
Favours control
Favours antioxidant
(Continued . . . )
110
Study or subgroup
Control N Mean(SD)
Weight

IV,Fixed,95% CI
Heterogeneity: not applicable Test for overall effect: Z = 0.58 (P = 0.56) 4 Coenzyme Q10 versus placebo at 9 months or more Balercia 2009 Safarinejad 2009a 30 106 44.18 (20.43) 22.8 (3.8) 30 106 49.56 (20.47) 21.2 (3.8) 0.9 % 96.1 % -5.38 [ -15.73, 4.97 ] 1.60 [ 0.58, 2.62 ]
Subtotal (95% CI)
136
136
97.1 %
1.53 [ 0.51, 2.55 ]
Heterogeneity: Chi2 = 1.73, df = 1 (P = 0.19); I2 =42% Test for overall effect: Z = 2.95 (P = 0.0032)
Total (95% CI)
181
151
100.0 %
1.61 [ 0.61, 2.61 ]
Heterogeneity: Chi2 = 3.79, df = 4 (P = 0.44); I2 =0.0% Test for overall effect: Z = 3.14 (P = 0.0017) Test for subgroup differences: Chi2 = 2.05, df = 3 (P = 0.56), I2 =0.0%
-20
-10
10
20
Favours control
Favours antioxidant
Analysis 1.11. Comparison 1 Antioxidant(s) versus control, Outcome 11 Total sperm motility over time.
Comparison: 1 Antioxidant(s) versus control Outcome: 11 Total sperm motility over time
Study or subgroup
Control N Mean(SD)
Weight
1 Total sperm motility at 3 months or less Ciftci 2009 Conquer 2000 Dawson 1990 Greco 2005 Micic 1988 Omu 2008 Peivandi 2010 Scott 1998 60 10 10 32 51 14 15 30 31.29 (8.62) 32 (16.1) 94 (32) 41.6 (22) 31.21 (10.66) 50 (20) 48.3 (0.16) 27 (20.3) 60 9 10 32 39 8 15 18 20.33 (8.43) 47.2 (18.6) 49 (25.3) 38.7 (21.5) 18.44 (6.35) 24 (12) 17 (0.09) 15.3 (17.4)
-50 -25 0 25 50
12.4 % 9.6 % 7.0 % 11.0 % 12.3 % 10.2 % 12.5 % 10.9 %
10.96 [ 7.91, 14.01 ] -15.20 [ -30.92, 0.52 ] 45.00 [ 19.72, 70.28 ] 2.90 [ -7.76, 13.56 ] 12.77 [ 9.23, 16.31 ] 26.00 [ 12.62, 39.38 ] 31.30 [ 31.21, 31.39 ] 11.70 [ 0.87, 22.53 ]
Favours control
Favours antioxidant
(Continued . . . )
111
Study or subgroup
Experimental N Mean(SD) 28.6 (38.1) 33.5 (29.8)
Control N 9 10 Mean(SD) 37.6 (33) 19 (14.4)
Weight

IV,Random,95% CI -9.00 [ -39.49, 21.49 ] 14.50 [ -6.01, 35.01 ]
Sigman 2006 Zavaczki 2003
12 10
5.8 % 8.2 %
Subtotal (95% CI)
244
210
100.0 % 13.47 [ 3.45, 23.49 ]
Heterogeneity: Tau2 = 208.58; Chi2 = 359.85, df = 9 (P<0.00001); I2 =97% Test for overall effect: Z = 2.63 (P = 0.0084) 2 Total sperm motility at 6 months Balercia 2005 Balercia 2009 Lenzi 2004 Safarinejad 2009 Safarinejad 2009a Sigman 2006 Suleiman 1996 15 30 30 116 106 12 52 61.07 (9.07) 39.41 (6.8) 31.11 (13.46) 29.2 (2.9) 27.6 (2.2) 32.3 (24.2) 48.9 (15.5) 15 30 26 118 106 9 35 43.4 (9.85) 34.93 (8.04) 29.55 (9.5) 22.9 (2.2) 23.1 (2.1) 40 (33) 35.9 (12.8) 6.5 % 14.5 % 7.7 % 31.4 % 31.6 % 0.6 % 7.8 % 17.67 [ 10.89, 24.45 ] 4.48 [ 0.71, 8.25 ] 1.56 [ -4.48, 7.60 ] 6.30 [ 5.64, 6.96 ] 4.50 [ 3.92, 5.08 ] -7.70 [ -33.24, 17.84 ] 13.00 [ 7.02, 18.98 ]
Subtotal (95% CI)
361
339
100.0 %
6.28 [ 4.36, 8.21 ]
Heterogeneity: Tau2 = 2.95; Chi2 = 37.94, df = 6 (P<0.00001); I2 =84% Test for overall effect: Z = 6.41 (P < 0.00001) 3 Total sperm motility at 9 months or more Balercia 2005 Balercia 2009 Safarinejad 2009a 15 30 106 49 (7.8) 32.93 (6.33) 24.2 (2.1) 15 30 106 42.73 (10.02) 35.3 (8) 22.8 (2.2) 17.8 % 31.9 % 50.3 % 6.27 [ -0.16, 12.70 ] -2.37 [ -6.02, 1.28 ] 1.40 [ 0.82, 1.98 ]
Subtotal (95% CI)
151
151
100.0 %
1.07 [ -2.29, 4.43 ]
Heterogeneity: Tau2 = 5.75; Chi2 = 6.27, df = 2 (P = 0.04); I2 =68% Test for overall effect: Z = 0.62 (P = 0.53)
-50
-25
25
50
Favours control
Favours antioxidant
112
Analysis 1.12. Comparison 1 Antioxidant(s) versus control, Outcome 12 Sperm concentration over time.
Comparison: 1 Antioxidant(s) versus control Outcome: 12 Sperm concentration over time
Study or subgroup
Control N Mean(SD)
Weight
1 Sperm concentration at 3 months or less Ciftci 2009 Conquer 2000 Greco 2005 Peivandi 2010 Rolf 1999 Wang 1983 Zavaczki 2003 60 10 32 15 15 11 10 21.85 (11.08) 44.6 (41.1) 27.5 (24.6) 46 (3.62) 20.6 (13.5) 12.8 (4.7) 16.1 (10.2) 60 9 32 15 16 7 10 22.32 (22) 43.1 (40.5) 20.3 (21.2) 16.5 (7.26) 25 (17.8) 8.5 (5.6) 10.9 (7.4) 16.0 % 6.4 % 14.6 % 16.4 % 14.6 % 16.3 % 15.6 % -0.47 [ -6.70, 5.76 ] 1.50 [ -35.23, 38.23 ] 7.20 [ -4.05, 18.45 ] 29.50 [ 25.39, 33.61 ] -4.40 [ -15.48, 6.68 ] 4.30 [ -0.69, 9.29 ] 5.20 [ -2.61, 13.01 ]
Subtotal (95% CI)
153
149
100.0 %
6.79 [ -5.09, 18.66 ]
Heterogeneity: Tau2 = 218.93; Chi2 = 106.32, df = 6 (P<0.00001); I2 =94% Test for overall effect: Z = 1.12 (P = 0.26) 2 Sperm concentration 6 months Balercia 2005 Balercia 2009 Lenzi 2004 Safarinejad 2009 Safarinejad 2009a Wang 1983 15 30 30 116 106 11 37.4 (16.42) 44.93 (19.3) 22.09 (9.05) 32.1 (6.8) 26.4 (4.4) 11.1 (5.8) 15 30 26 118 106 7 33.73 (14.36) 46.37 (19.77) 22.17 (16.95) 23.5 (5.8) 20.8 (4.3) 8.3 (5.6) 4.8 % 5.8 % 9.4 % 31.9 % 33.9 % 14.2 % 3.67 [ -7.37, 14.71 ] -1.44 [ -11.33, 8.45 ] -0.08 [ -7.36, 7.20 ] 8.60 [ 6.98, 10.22 ] 5.60 [ 4.43, 6.77 ] 2.80 [ -2.58, 8.18 ]
Subtotal (95% CI)
308
302
100.0 %
5.12 [ 2.53, 7.72 ]
Heterogeneity: Tau2 = 4.82; Chi2 = 16.26, df = 5 (P = 0.01); I2 =69% Test for overall effect: Z = 3.87 (P = 0.00011) 3 Sperm concentration at 9 months or more Balercia 2005 Balercia 2009 Safarinejad 2009a 15 30 106 33.27 (13.62) 44.18 (20.43) 22.8 (3.8) 15 30 106 30.13 (9.3) 49.56 (20.47) 21.2 (3.8) 1.5 % 1.0 % 97.6 % 3.14 [ -5.21, 11.49 ] -5.38 [ -15.73, 4.97 ] 1.60 [ 0.58, 2.62 ]
Subtotal (95% CI)
151
151
100.0 %
1.56 [ 0.55, 2.57 ]
Heterogeneity: Tau2 = 0.0; Chi2 = 1.87, df = 2 (P = 0.39); I2 =0.0% Test for overall effect: Z = 3.02 (P = 0.0025)
-20
-10
10
20
Favours control
Favours antioxidant
113
Analysis 1.13. Comparison 1 Antioxidant(s) versus control, Outcome 13 Side effects.

Comparison: 1 Antioxidant(s) versus control Outcome: 13 Side effects
Study or subgroup
Experimental n/N
Control n/N
1 gastrointestinal Cavallini 2004 Kessopoulou 1995 Safarinejad 2009a Sigman 2006 Tremellen 2007 Zavaczki 2003 2/39 0/15 0/106 0/12 3/37 2/10 2/47 1/15 0/106 0/9 0/20 1/10 1.21 [ 0.16, 9.01 ] 0.14 [ 0.00, 6.82 ] 0.0 [ 0.0, 0.0 ] 0.0 [ 0.0, 0.0 ] 4.94 [ 0.44, 55.27 ] 2.11 [ 0.19, 23.05 ]
Subtotal (95% CI)

219
207
1.63 [ 0.48, 5.58 ]
Heterogeneity: Chi2 = 2.49, df = 3 (P = 0.48); I2 =0.0% Test for overall effect: Z = 0.78 (P = 0.44) 2 euphoria Cavallini 2004 2/39 2/47 1.21 [ 0.16, 9.01 ]
Subtotal (95% CI)

39
47
1.21 [ 0.16, 9.01 ]
Test for subgroup differences: Chi2 = 0.06, df = 1 (P = 0.81), I2 =0.0%
0.002
0.1
10
500
Favours antioxidant
Favours control
Analysis 1.14. Comparison 1 Antioxidant(s) versus control, Outcome 14 Sperm motility at 3 months or less. Sperm motility at 3 months or less
Study
Heading one
Heading two
L-carnitine + Acetyl-carnitine versus placebo (median and interquartile range) Cavallini 2004 L-carnitine + Acetyl-carnitine Median=22.3 (n=39) Interquartile range =28.4-15.2 Placebo Median= 14.0 (n=47) Interquartile range=17.4-5.1
Combined antioxidants versus no treatment Galatioto 2008 Combined antioxidants % of motile sperm =58% (n=20) No treatment % of motile sperm =51% (n=22)
114
Sperm motility at 3 months or less
(Continued)
Vitamin E versus placebo Kessopoulou 1995 Vitamin E Median=7 (n=15) min/max= -27-34 Placebo Median=7 (n=15) min/max= -33-36
L-carnitine versus placebo Lenzi 2003 L-carnitine Mean=11(n=43) no sd given Placebo Mean=8.8 (n=43) no sd given
Pentoxifylline versus Placebo Merino 1997 Pentoxifylline Median=35.5 (n=25) Range=31.5-39.5 Placebo Median=33.5 (n=22) Range=28.5-37.5
Analysis 1.15. Comparison 1 Antioxidant(s) versus control, Outcome 15 Sperm motility at 6 months. Sperm motility at 6 months
Study
Heading one
Heading two
L-carnitine + Acetyl-carnitine versus placebo (median and interquartile range) Cavallini 2004 L-carnitine + acetyl-carnitine Median=23.6 (n=39) Interquartile range=28.9-16.0 Placebo Median=13.2 (n=47) Interquartile range=18.6-9.0
Pentoxifylline versus placebo Merino 1997 Pentoxifylline Median=42 (n=25) Range=38-46 Placebo Median=31.5 (22) Range=28-35
Folic acid versus placebo Wong 2002 Folic acid Median=35 (n=22) Range=5-65 Placebo Median=30 (n=25) Range=5-80
Zinc versus placebo Wong 2002 Zinc Median=35 (n=23) Range=10-65 Placebo Median=30 (n=25)
115
Sperm motility at 6 months
(Continued)
Range=5-80 Zinc + folic acid versus placebo Wong 2002 Zinc + folic acid Median=35 (n=24) Range 5-70 Placebo Median=30 (n=25) Range=5-80
Analysis 1.16. Comparison 1 Antioxidant(s) versus control, Outcome 16 Sperm concentration at 3 months or less. Sperm concentration at 3 months or less
Study
Heading one
Heading two
L-carnitine + Acetyl-carnitine versus placebo (median and interquartile range) Cavallini 2004 L-carnitine + acetyl-carnitine Median=20.9 (n=39) Interquartile range=25.6-14.8 Placebo Median=12.3 (n=47) Interquartile range=16.0-9.1
Vitamin E versus placebo Kessopoulou 1995 Vitamin E Median= -15 (n=15) min/max= -58-59 Placebo Median=0 (n=15) min/max= -37-160
L-carnitine versus placebo Lenzi 2003 L-carnitine Mean= 9 (1st phase data) (n=43) no sd given Placebo Mean=5.3 (n=43) no sd given
Analysis 1.17. Comparison 1 Antioxidant(s) versus control, Outcome 17 Sperm concentration at 6 months. Sperm concentration at 6 months
Study
Heading one
Heading two
L-carnitine + acetyl-carnitine versus placebo Cavallini 2004 L-carnitine + acetyl-carniitne Median=20.6 (n=39) Interquartile range=24.9-15.1 Placebo Median=10.9 (n=47) Interquartile range=15.1-9.0
Folic acid versus Placebo

Sperm concentration at 6 months
(Continued)
Wong 2002
Folic acid Median=14 (n=22) Range=0.9-130
Placebo Median=9 (n=25) Range=0.8-80
Zinc versus Placebo Wong 2002 Zinc Median=16 (n= 23) Range=0.6-80 Placebo Median=9 (n= 25) Range=0.8-80
Analysis 1.18. Comparison 1 Antioxidant(s) versus control, Outcome 18 Pregnancy.

Comparison: 1 Antioxidant(s) versus control Outcome: 18 Pregnancy
Study or subgroup
Experimental n/N
Control n/N
1 Pregnancy rate of trials that also report live birth Omu 1998 Suleiman 1996 11/49 11/52 2/48 0/35 4.75 [ 1.49, 15.20 ] 6.64 [ 1.84, 23.93 ]
Subtotal (95% CI)

101
83
5.53 [ 2.34, 13.08 ]
Heterogeneity: Chi2 = 0.14, df = 1 (P = 0.71); I2 =0.0% Test for overall effect: Z = 3.89 (P = 0.00010) 2 Pregnancy rate only reported Balercia 2005 Balercia 2009 Cavallini 2004 Galatioto 2008 Lenzi 2003 Lenzi 2004 Peivandi 2010 Rolf 1999 Safarinejad 2009a 9/45 6/30 9/39 1/20 6/43 4/30 3/15 0/15 0/106 3/15 3/30 1/47 0/22 0/43 0/26 0/15 0/16 0/106
0.005 0.1 1 10 200
1.00 [ 0.23, 4.26 ] 2.16 [ 0.53, 8.82 ] 7.50 [ 2.01, 27.98 ] 8.17 [ 0.16, 413.39 ] 8.37 [ 1.61, 43.58 ] 7.20 [ 0.95, 54.34 ] 8.57 [ 0.82, 89.45 ] 0.0 [ 0.0, 0.0 ] 0.0 [ 0.0, 0.0 ]
Favours control
Favours antioxidant
(Continued . . . )
117
Study or subgroup
Experimental n/N
Control n/N 0/10 4/20 0/7 0/10

Peto,Fixed,95% CI 7.39 [ 0.15, 372.38 ] 3.75 [ 1.26, 11.16 ] 0.0 [ 0.0, 0.0 ] 7.39 [ 0.15, 372.38 ]
Sigman 2006 Tremellen 2007 Wang 1983 Zavaczki 2003
1/10 20/38 0/11 1/10
Subtotal (95% CI)

412
367
4.02 [ 2.34, 6.92 ]
Heterogeneity: Chi2 = 6.96, df = 9 (P = 0.64); I2 =0.0% Test for overall effect: Z = 5.02 (P < 0.00001)
Total (95% CI)

513
450
4.40 [ 2.78, 6.97 ]
Heterogeneity: Chi2 = 7.48, df = 11 (P = 0.76); I2 =0.0% Test for overall effect: Z = 6.32 (P < 0.00001) Test for subgroup differences: Chi2 = 0.38, df = 1 (P = 0.54), I2 =0.0%
0.005
0.1
10
200
Favours control
Favours antioxidant
Analysis 2.1. Comparison 2 Live birth versus pregnancy, Outcome 1 Live birth.
Comparison: 2 Live birth versus pregnancy Outcome: 1 Live birth
Study or subgroup
Experimental n/N
Control n/N
Weight
1 live birth Omu 1998 Suleiman 1996 8/49 9/52 2/48 0/35 53.6 % 46.4 % 3.67 [ 1.00, 13.51 ] 6.33 [ 1.56, 25.63 ]
Subtotal (95% CI)
101
83
100.0 %
4.73 [ 1.82, 12.27 ]
Total events: 17 (Experimental), 2 (Control) Heterogeneity: Chi2 = 0.31, df = 1 (P = 0.58); I2 =0.0% Test for overall effect: Z = 3.19 (P = 0.0014) 2 pregnancy Omu 1998 Suleiman 1996 11/49 11/52 2/48 0/35 54.9 % 45.1 % 4.75 [ 1.49, 15.20 ] 6.64 [ 1.84, 23.93 ]
Subtotal (95% CI)
101
83
0.005 0.1 1 10 200
100.0 %
5.53 [ 2.34, 13.08 ]
Favours control
Favours antioxidant
(Continued . . . )
118
Study or subgroup
Experimental n/N
Control n/N
Weight

Peto,Fixed,95% CI
Total events: 22 (Experimental), 2 (Control) Heterogeneity: Chi2 = 0.14, df = 1 (P = 0.71); I2 =0.0% Test for overall effect: Z = 3.89 (P = 0.00010)
0.005
0.1
10
200
Favours control
Favours antioxidant
Analysis 3.1. Comparison 3 Head to head antioxidant(s), Outcome 1 Pregnancy per couple randomised.
Comparison: 3 Head to head antioxidant(s) Outcome: 1 Pregnancy per couple randomised
Study or subgroup
Antioxidant a n/N
Antioxidant b n/N
Weight
1 L acetyl carnitine + L carnitine versus L acetyl carnitine Balercia 2005 2/7 2/15 100.0 % 2.66 [ 0.27, 25.80 ]
Subtotal (95% CI)

Heterogeneity: not applicable Test for overall effect: Z = 0.84 (P = 0.40)
15
100.0 %
2.66 [ 0.27, 25.80 ]
Total events: 2 (Antioxidant a), 2 (Antioxidant b)
2 L acetyl carnitine + L carnitine versus L- carnitine Balercia 2005 3/8 2/15 100.0 % 3.89 [ 0.51, 29.76 ]
Subtotal (95% CI)

15
100.0 %
3.89 [ 0.51, 29.76 ]
3 L acetyl carnitine + L carnitine versus Vitamin E + Vitamin C Li 2005 10/90 2/60 100.0 % 2.86 [ 0.86, 9.49 ]
Subtotal (95% CI)

90
60
100.0 %
2.86 [ 0.86, 9.49 ]
Test for subgroup differences: Chi2 = 0.08, df = 2 (P = 0.96), I2 =0.0%
0.01
0.1
10
100
Favours antioxidant b
Favours antioxidant a
119
Analysis 3.2. Comparison 3 Head to head antioxidant(s), Outcome 2 Sperm total motility at 3 months or less.
Review: Antioxidants for male subfertility Comparison: 3 Head to head antioxidant(s) Outcome: 2 Sperm total motility at 3 months or less
Study or subgroup
Antioxidant b N Mean(SD)
Weight
1 Ethylcysteine 600mg/day vs Vitamin E Akiyama 1999 5 40.9 (30.1) 5 42.8 (34.4) 100.0 % -1.90 [ -41.97, 38.17 ]
Subtotal (95% CI)

100.0 %
-1.90 [ -41.97, 38.17 ]
Test for overall effect: Z = 0.09 (P = 0.93) 2 Docosahexaenoic acid (DHA) 400g/day vs Docosahexaenoic acid 800mg/day Conquer 2000 9 39.4 (24.3) 10 32 (16.1) 100.0 % 7.40 [ -11.35, 26.15 ]
Subtotal (95% CI)

10
100.0 %
7.40 [ -11.35, 26.15 ]
Test for overall effect: Z = 0.77 (P = 0.44) 3 Ascorbic Acid 200mg/day versus Ascorbic Acid 1000mg/day Dawson 1990 10 51 (22.1) 10 94 (32) 100.0 % -43.00 [ -67.10, -18.90 ]
Subtotal (95% CI)

10
10
100.0 % -43.00 [ -67.10, -18.90 ]
Test for overall effect: Z = 3.50 (P = 0.00047) 4 Vitamin E + Selenium versus Vitamin B at 3 months Keskes-Ammar 2003 28 39 (19) 26 39 (21) 100.0 % 0.0 [ -10.71, 10.71 ]
Subtotal (95% CI)

28
26
100.0 %
0.0 [ -10.71, 10.71 ]
Test for overall effect: Z = 0.0 (P = 1.0) 5 Zinc versus Zinc + Vitamin E at 3 months Omu 2008 6 49 (12) 12 50 (18) 100.0 % -1.00 [ -15.00, 13.00 ]
Subtotal (95% CI)

12
100.0 %
-1.00 [ -15.00, 13.00 ]
Test for overall effect: Z = 0.14 (P = 0.89) 6 Zinc versus Zinc + Vitamin E + Vitamin C at 3 months Omu 2008 6 49 (12) 6 50 (20) 100.0 % -1.00 [ -19.66, 17.66 ]
Subtotal (95% CI)

100.0 %
-1.00 [ -19.66, 17.66 ]
Test for overall effect: Z = 0.11 (P = 0.92) 7 Zinc + Vitamin E versus Zinc + Vitamin E + Vitamin C at 3 months Omu 2008 12 50 (18) 6 50 (20) 100.0 % 0.0 [ -18.97, 18.97 ]
Subtotal (95% CI)
12
6
-50 -25 0 25 50
100.0 %
0.0 [ -18.97, 18.97 ]
Antioxidant b
Antioxidant a
(Continued . . . )
120
Study or subgroup
Weight

IV,Fixed,95% CI
Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P = 1.0) 8 Selenium versus combined antioxidants Scott 1998 16 30.2 (22.8) 30 27 (20.3) 100.0 % 3.20 [ -10.13, 16.53 ]
Subtotal (95% CI)

16
30
100.0 %
3.20 [ -10.13, 16.53 ]
Test for overall effect: Z = 0.47 (P = 0.64) 9 L acetyl carnitine + L carnitine versus Vitamin E + Vitamin C Li 2005 90 38.29 (9.67) 60 15.24 (7.89) 100.0 % 23.05 [ 20.23, 25.87 ]
Subtotal (95% CI)

90
60
100.0 %
23.05 [ 20.23, 25.87 ]
Test for overall effect: Z = 16.00 (P < 0.00001)
-50
-25
25
50
Antioxidant b
Antioxidant a
Analysis 3.3. Comparison 3 Head to head antioxidant(s), Outcome 3 Sperm total motility at 6 months.
Comparison: 3 Head to head antioxidant(s) Outcome: 3 Sperm total motility at 6 months
Study or subgroup
Antioxidant a N Mean(SD)
Weight
1 L-aceytl carnitine +L-carnitine versus L-carnitine at 6 months Balercia 2005 7 61.07 (9.07) 15 64.53 (8.41) 100.0 % -3.46 [ -11.41, 4.49 ]
Subtotal (95% CI)

15
100.0 % -3.46 [ -11.41, 4.49 ]
Test for overall effect: Z = 0.85 (P = 0.39) 2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 6 months Balercia 2005 8 61.07 (9.07) 15 60.43 (10.46) 100.0 % 0.64 [ -7.58, 8.86 ]
Subtotal (95% CI)

15
100.0 %
0.64 [ -7.58, 8.86 ]
Test for overall effect: Z = 0.15 (P = 0.88) 3 Selenium versus N-acetyl-cysteine at 26 weeks (6 months) Safarinejad 2009 58 26.1 (2.9) 118 24.8 (2.9) 100.0 % 1.30 [ 0.39, 2.21 ]
Subtotal (95% CI)
58
118
-10 -5 0 5 10
100.0 %
1.30 [ 0.39, 2.21 ]
(Continued . . . )
121
(. . .
Study or subgroup Antioxidant a N Heterogeneity: not applicable Test for overall effect: Z = 2.80 (P = 0.0052) 4 Selenium versus selenium plus N-acetyl-cysteine at 26 weeks (6 months) Safarinejad 2009 58 26.1 (2.9) 58 29.2 (2.9) 100.0 % Mean(SD) Antioxidant b N Mean(SD) Mean Difference IV,Fixed,95% CI Weight
Continued)
-3.10 [ -4.16, -2.04 ]
Subtotal (95% CI)

58
58
100.0 % -3.10 [ -4.16, -2.04 ]
Test for overall effect: Z = 5.76 (P < 0.00001) 5 N-acetyl-cysteine vs selenium plus N-acetyl-cysteine at 26 weeks Safarinejad 2009 118 24.8 (2.9) 58 29.2 (2.9) 100.0 % -4.40 [ -5.31, -3.49 ]
Subtotal (95% CI)

118
58
100.0 % -4.40 [ -5.31, -3.49 ]
Test for overall effect: Z = 9.46 (P < 0.00001) Test for subgroup differences: Chi2 = 81.67, df = 4 (P = 0.00), I2 =95%
-10
-5
10
Analysis 3.4. Comparison 3 Head to head antioxidant(s), Outcome 4 Sperm total motility at 9 months or more.
Review: Antioxidants for male subfertility Comparison: 3 Head to head antioxidant(s) Outcome: 4 Sperm total motility at 9 months or more
Study or subgroup
Weight
1 L-aceytl carnitine +L-carnitine versus L-carnitine at 9 months Balercia 2005 7 49 (7.8) 15 54.27 (8.96) 100.0 % -5.27 [ -12.61, 2.07 ]
Subtotal (95% CI)

15
100.0 % -5.27 [ -12.61, 2.07 ]
Test for overall effect: Z = 1.41 (P = 0.16) 2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 9 months Balercia 2005 8 49 (7.8) 15 50.57 (5.71) 100.0 % -1.57 [ -7.70, 4.56 ]
Subtotal (95% CI)

15
100.0 %
-1.57 [ -7.70, 4.56 ]
Test for overall effect: Z = 0.50 (P = 0.62) Test for subgroup differences: Chi2 = 0.57, df = 1 (P = 0.45), I2 =0.0%
-20
-10
10
20
122
Analysis 3.5. Comparison 3 Head to head antioxidant(s), Outcome 5 Sperm concentration at 3 months or less.
Review: Antioxidants for male subfertility Comparison: 3 Head to head antioxidant(s) Outcome: 5 Sperm concentration at 3 months or less
Study or subgroup
Weight
1 Ethylcysteine 600mg/day vs Vitamin E Akiyama 1999 5 20.1 (14.8) 5 17.9 (15.6) 100.0 % 2.20 [ -16.65, 21.05 ]
Subtotal (95% CI)

100.0 % 2.20 [ -16.65, 21.05 ]
Test for overall effect: Z = 0.23 (P = 0.82) 2 Docosahexaenoic acid (DHA) 400g/day versus Docosahexaenoic acid (DHA) 800g/day Conquer 2000 9 37.8 (36.9) 10 44.6 (41.1) 100.0 % -6.80 [ -41.87, 28.27 ]
Subtotal (95% CI)

10
100.0 % -6.80 [ -41.87, 28.27 ]
Test for overall effect: Z = 0.38 (P = 0.70) 3 L-carnitine versus Vitamin E + Vitamin C Li 2005a 32 34.6 (7.4) 31 19.1 (4.5) 100.0 % 15.50 [ 12.49, 18.51 ]
Subtotal (95% CI)

32
31
100.0 % 15.50 [ 12.49, 18.51 ]
-100
-50
50
100
123
Analysis 3.6. Comparison 3 Head to head antioxidant(s), Outcome 6 Sperm concentration at 6 months.
Comparison: 3 Head to head antioxidant(s) Outcome: 6 Sperm concentration at 6 months
Study or subgroup
Weight
Subtotal (95% CI)

15
100.0 %
-8.13 [ -24.42, 8.16 ]
Test for overall effect: Z = 0.98 (P = 0.33) 2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 6 months Balercia 2005 8 37.4 (16.42) 15 39.57 (19.99) 100.0 % -2.17 [ -17.40, 13.06 ]
Subtotal (95% CI)

15
100.0 % -2.17 [ -17.40, 13.06 ]
Test for overall effect: Z = 0.28 (P = 0.78) 3 Selenium versus N-acetyl-cysteine at 26 weeks (6 months) Safarinejad 2009 58 27.6 (6.4) 118 26.8 (5.3) 100.0 % 0.80 [ -1.10, 2.70 ]
Subtotal (95% CI)

58
118
100.0 %
0.80 [ -1.10, 2.70 ]
Test for overall effect: Z = 0.82 (P = 0.41) 4 Selenium versus selenium plus N-acetyl-cysteine at 26 weeks (6 months) Safarinejad 2009 58 27.6 (6.4) 58 32.1 (6.8) 100.0 % -4.50 [ -6.90, -2.10 ]
Subtotal (95% CI)

58
58
100.0 %
-4.50 [ -6.90, -2.10 ]
Test for overall effect: Z = 3.67 (P = 0.00024) 5 N-acetyl-cysteine vs selenium plus N-acetyl-cysteine at 26 weeks Safarinejad 2009 118 26.8 (5.3) 58 32.1 (6.8) 100.0 % -5.30 [ -7.29, -3.31 ]
Subtotal (95% CI)

118
58
100.0 %
-5.30 [ -7.29, -3.31 ]
-20
-10
10
20
124
Analysis 3.7. Comparison 3 Head to head antioxidant(s), Outcome 7 Sperm concentration at 9 months or more.
Review: Antioxidants for male subfertility Comparison: 3 Head to head antioxidant(s) Outcome: 7 Sperm concentration at 9 months or more
Study or subgroup
Weight
Subtotal (95% CI)

15
100.0 % -6.13 [ -18.44, 6.18 ]
Test for overall effect: Z = 0.98 (P = 0.33) 2 L-acetyl carnitine + L-carnitine versus L-acetyl carnitine at 9 months Balercia 2005 8 33.27 (13.62) 15 31.21 (8.6) 100.0 % 2.06 [ -8.33, 12.45 ]
Subtotal (95% CI)

15
100.0 % 2.06 [ -8.33, 12.45 ]
Test for overall effect: Z = 0.39 (P = 0.70) Test for subgroup differences: Chi2 = 0.99, df = 1 (P = 0.32), I2 =0.0%
-20
-10
10
20
Analysis 3.8. Comparison 3 Head to head antioxidant(s), Outcome 8 Side effects.

Comparison: 3 Head to head antioxidant(s) Outcome: 8 Side effects
Study or subgroup
Experimental n/N
Control n/N
Odds Ratio M-H,Fixed,95% CI
Odds Ratio M-H,Fixed,95% CI
1 L acetyl carnitine + L carnitine versus Vitamin E + Vitamin C Li 2005 0/90 0/60 0.0 [ 0.0, 0.0 ]
Total (95% CI)

Total events: 0 (Experimental), 0 (Control) Heterogeneity: not applicable Test for overall effect: Z = 0.0 (P < 0.00001)
90
60
0.0 [ 0.0, 0.0 ]
0.01
0.1
10
100
125
Analysis 3.9. Comparison 3 Head to head antioxidant(s), Outcome 9 Sperm concentration at 6 months. Sperm concentration at 6 months
Study
Heading one
Heading two
Zinc versus Folic acid Wong 2002 Zinc Median=16 Range=0.6-80 Folic acid Median=14 Range=0.9-130
Zinc versus Zinc + Folic acid Wong 2002 Zinc Median= 16 Range=0.6-80 Zinc + Folic acid Median=12 Range=0.5-180
Analysis 3.10. Comparison 3 Head to head antioxidant(s), Outcome 10 Sperm motility at 6 months. Sperm motility at 6 months
Study
Heading one
Heading two
Zinc versus Folic acid Wong 2002 Zinc Median=35 Range=10-65 Folic acid Median=35 Range=5-65
APPENDICES Appendix 1. MEDLINE search strategy

1 exp male infertility/ (19113) 2 (asthenozoospermia or oligospermia or azoospermia).tw. (4464) 3 exp Spermatozoa/ (48125) 4 Sperm$.tw. (85274) 5 (male$ adj2 subfertil$).tw. (484) 6 (male$ adj2 infertil$).tw. (5719) 7 (subfertil$ adj2 men).tw. (356) 8 (infertil$ adj2 men).tw. (2400) 9 (male$ adj2 fertility).tw. (2706) 10 semen.tw. (16927) 11 oligoasthenoteratozoospermi$.tw. (185) 12 or/1-11 (108373)
13 exp antioxidants/ or free radical scavengers/ (280006) 14 (antioxidant$ or radical scavengers).tw. (69599) 15 exp vitamins/ or exp ascorbic acid/ or exp dehydroascorbic acid/ or exp vitamin a/ or exp vitamin e/ or exp vitamin u/ or exp alphatocopherol/ or exp beta carotene/ or exp beta-tocopherol/ or exp gamma-tocopherol/ (234713) 16 vitamin$.tw. (111181) 17 exp Zinc/ (41085) 18 exp Selenium/ (13920) 19 (Glutathione$ or folate).tw. (81376) 20 exp Ubiquinone/ (5352) 21 ubiquin$.tw. (5228) 22 coenzyme q10.tw. (1462) 23 exp Carnitine/ (6812) 24 (carnitine$ or carotenoid$).tw. (17142) 25 (astaxanthin$ or lycopene$).tw. (2711) 26 menevit.tw. (1) 27 multivitamin$.tw. (1967) 28 (betacarotene$ or beta carotene$).tw. (7738) 29 ascorbic acid.tw. (18104) 30 n-acetylcysteine.tw. (5514) 31 alpha-tocopherol$.tw. (10653) 32 exp Pentoxifylline/ (3302) 33 Pentoxifylline$.tw. (3079) 34 (sh adj2 oil$).tw. (5536) 35 omega$.tw. (23211) 36 exp fatty acids/ or exp sh oils/ or exp cod liver oil/ or exp fatty acids, omega-3/ or exp plant oils/ (321518) 37 fatty acid$.tw. (109608) 38 (plant adj4 oil$).tw. (899) 39 l-arginine$.tw. (25627) 40 avonoid$.tw. (11562) 41 carotenoid$.tw. (8425) 42 riboavin$.tw. (5897) 43 pycnogenol$.tw. (177) 44 lutein$.tw. (27811) 45 lipoic acid$.tw. (1986) 46 or/13-45 (968481) 47 46 and 12 (8395) 48 randomized controlled trial.pt. (280587) 49 controlled clinical trial.pt. (80011) 50 randomized.ab. (190341) 51 placebo.tw. (118552) 52 clinical trials as topic.sh. (146404) 53 randomly.ab. (138335) 54 trial.ti. (82364) 55 (crossover or cross-over or cross over).tw. (43907) 56 or/48-55 (666167) 57 (animals not (humans and animals)).sh. (3345663) 58 56 not 57 (615399) 59 47 and 58 (247) 60 from 59 keep 1-247 (247)
127
Appendix 2. EMBASE search strategy

1 exp male infertility/ (17017) 2 (asthenozoospermia or oligospermia or azoospermia).tw. (3803) 3 exp Spermatozoa/ (10504) 4 Sperm$.tw. (62512) 5 (male$ adj2 subfertil$).tw. (451) 6 (male$ adj2 infertil$).tw. (5173) 7 (subfertil$ adj2 men).tw. (302) 8 (infertil$ adj2 men).tw. (2207) 9 (male$ adj2 fertility).tw. (2138) 10 semen.tw. (11264) 11 oligoasthenoteratozoospermi$.tw. (192) 12 or/1-11 (74872) 13 vitamin$.tw. (83637) 14 exp Zinc/ (44487) 15 exp Selenium/ (15822) 16 (Glutathione$ or folate).tw. (72891) 17 exp Ubiquinone/ (3388) 18 ubiquin$.tw. (4120) 19 coenzyme q10.tw. (1640) 20 exp Carnitine/ (6361) 21 (carnitine$ or carotenoid$).tw. (14110) 22 (astaxanthin$ or lycopene$).tw. (2452) 23 menevit.tw. (6) 24 multivitamin$.tw. (1730) 25 (betacarotene$ or beta carotene$).tw. (6736) 26 ascorbic acid.tw. (13434) 27 n-acetylcysteine.tw. (5536) 28 alpha-tocopherol$.tw. (9008) 29 exp Pentoxifylline/ (8389) 30 Pentoxifylline$.tw. (3181) 31 (sh adj2 oil$).tw. (5366) 32 omega$.tw. (19649) 33 fatty acid$.tw. (85842) 34 (plant adj4 oil$).tw. (937) 35 l-arginine$.tw. (24461) 36 avonoid$.tw. (11342) 37 carotenoid$.tw. (6265) 38 riboavin$.tw. (3601) 39 pycnogenol$.tw. (202) 40 lutein$.tw. (21048) 41 lipoic acid$.tw. (1791) 42 exp antioxidant/ (46224) 43 exp vitamin/ or exp ascorbic acid/ or exp carotenoid/ or exp multivitamin/ or vitamin b group/ (246879) 44 exp fatty acid/ (244186) 45 exp vegetable oil/ (25806) 46 exp sh oil/ (7175) 47 exp cod liver oil/ (573) 48 exp omega 3 fatty acid/ (8974) 49 or/13-48 (735430) 50 12 and 49 (5848) 51 Clinical Trial/ (579096)
52 Randomized Controlled Trial/ (181492) 53 exp randomization/ (27531) 54 Single Blind Procedure/ (9051) 55 Double Blind Procedure/ (76202) 56 Crossover Procedure/ (22447) 57 Placebo/ (138119) 58 Randomi?ed controlled trial$.tw. (37822) 59 Rct.tw. (3250) 60 random allocation.tw. (669) 61 randomly allocated.tw. (10813) 62 allocated randomly.tw. (1393) 63 (allocated adj2 random).tw. (573) 64 Single blind$.tw. (7937) 65 Double blind$.tw. (88810) 66 ((treble or triple) adj blind$).tw. (149) 67 placebo$.tw. (116382) 68 prospective study/ (90830) 69 or/51-68 (760080) 70 case study/ (6847) 71 case report.tw. (127165) 72 abstract report/ or letter/ (526689) 73 or/70-72 (658112) 74 69 not 73 (733696) 75 50 and 74 (544) 76 (2009$ or 2010$).em. (774638) 77 75 and 76 (59) 78 from 77 keep 1-59 (59)
Appendix 3. CENTRAL search strategy

1 exp male infertility/ (433) 2 (asthenozoospermia or oligospermia or azoospermia).tw. (180) 3 exp Spermatozoa/ (315) 4 Sperm$.tw. (1726) 5 (male$ adj2 subfertil$).tw. (52) 6 (male$ adj2 infertil$).tw. (266) 7 (subfertil$ adj2 men).tw. (19) 8 (infertil$ adj2 men).tw. (91) 9 (male$ adj2 fertility).tw. (42) 10 semen.tw. (510) 11 oligoasthenoteratozoospermi$.tw. (6) 12 or/1-11 (2087) 13 exp antioxidants/ or free radical scavengers/ (8109) 14 (antioxidant$ or radical scavengers).tw. (2829) 15 exp vitamins/ or exp ascorbic acid/ or exp dehydroascorbic acid/ or exp vitamin a/ or exp vitamin e/ or exp vitamin u/ or exp alphatocopherol/ or exp beta carotene/ or exp beta-tocopherol/ or exp gamma-tocopherol/ (9080) 16 vitamin$.tw. (7367) 17 exp Zinc/ (930) 18 exp Selenium/ (345) 19 (Glutathione$ or folate).tw. (1838) 20 exp Ubiquinone/ (200) 21 ubiquin$.tw. (70) 22 coenzyme q10.tw. (189)
23 exp Carnitine/ (358) 24 (carnitine$ or carotenoid$).tw. (955) 25 (astaxanthin$ or lycopene$).tw. (231) 26 menevit.tw. (1) 27 multivitamin$.tw. (392) 28 (betacarotene$ or beta carotene$).tw. (959) 29 ascorbic acid.tw. (747) 30 n-acetylcysteine.tw. (444) 31 alpha-tocopherol$.tw. (850) 32 exp Pentoxifylline/ (356) 33 Pentoxifylline$.tw. (589) 34 (sh adj2 oil$).tw. (998) 35 omega$.tw. (885) 36 exp fatty acids/ or exp sh oils/ or exp cod liver oil/ or exp fatty acids, omega-3/ or exp plant oils/ (13281) 37 fatty acid$.tw. (4996) 38 (plant adj4 oil$).tw. (48) 39 l-arginine$.tw. (775) 40 avonoid$.tw. (221) 41 carotenoid$.tw. (380) 42 riboavin$.tw. (242) 43 pycnogenol$.tw. (50) 44 lutein$.tw. (1402) 45 lipoic acid$.tw. (108) 46 or/13-45 (36854) 47 46 and 12 (253) 48 from 47 keep 1-253 (253)
Appendix 4. CINAHL search strategy

CINAHL search from inception until 13.08.08 1 exp male infertility/ (0) 2 (asthenozoospermia or oligospermia or azoospermia).tw. (35) 3 (male$ or men or man).tw. (60126) 4 exp Spermatozoa/ (269) 5 Sperm$.tw. (585) 6 or/1-5 (60531) 7 exp antioxidants/ or free radical scavengers/ (3528) 8 (antioxidant$ or radical scavengers).tw. (2606) 9 exp vitamins/ or exp ascorbic acid/ or exp dehydroascorbic acid/ or exp vitamin a/ or exp vitamin e/ or exp vitamin u/ or exp alphatocopherol/ or exp beta carotene/ or exp beta-tocopherol/ or exp gamma-tocopherol/ (11605) 10 vitamin$.tw. (6740) 11 exp Zinc/ (1056) 12 (zinc or selenium).tw. (1529) 13 exp Selenium/ (524) 14 exp Glutathione Peroxidase/ or exp folic acid/ (2362) 15 (Glutathione$ or folate).tw. (1745) 16 exp Ubiquinone/ (328) 17 (ubiquin$ or folic acid).tw. (988) 18 coenzyme q10.tw. (149) 19 exp Carnitine/ (271) 20 (carnitine$ or carotenoid$).tw. (609) 21 (astaxanthin$ or lycopene$).tw. (223) 22 menevit.tw. (2)
23 or/7-22 (18483) 24 6 and 23 (1359) 25 exp fertilization in vitro/ or exp sperm injections, intracytoplasmic/ or exp Insemination, Articial, Homologous/ (947) 26 (in vitro fertilisation or intracytoplasmic sperm injection$).tw. (148) 27 (intrauterine adj3 insemination$).tw. (45) 28 (ivf or icsi or iui).tw. (332) 29 in-vitro fertilisation.tw. (90) 30 in vitro fertilization.tw. (287) 31 ART.tw. (5558) 32 Articial reproduc$ technique$.tw. (1) 33 or/25-32 (6666) 34 24 and 33 (8) 35 exp clinical trials/ (61758) 36 Clinical trial.pt. (32434) 37 (clinic$ adj trial$1).tw. (14019) 38 ((singl$ or doubl$ or trebl$ or tripl$) adj (blind$3 or mask$3)).tw. (8276) 39 Randomi?ed control$ trial$.tw. (12024) 40 Random assignment/ (18434) 41 Random$ allocat$.tw. (1272) 42 Placebo$.tw. (11470) 43 Placebos/ (4474) 44 Quantitative studies/ (4044) 45 Allocat$ random$.tw. (74) 46 or/35-45 (85043) 47 34 and 46 (3) 48 from 47 keep 1-3 (3) CINAHL search from 13.08.08 until 22.02.10
# S5
Query S1 and S4
Limiters/Expanders Search modes - Boolean/Phrase
Last Run Via Interface - EBSCOhost Search Screen - Advanced Search Database - CINAHL Plus Interface - EBSCOhost Search Screen - Advanced Search Database - CINAHL Plus Interface - EBSCOhost Search Screen - Advanced Search Database - CINAHL Plus
Results 31
S4
S2 or S3
Search modes - Boolean/Phrase
1167
S3
(sperm) or (MH Sperm Search modes Count) or (MH Spermato- - Boolean/Phrase zoa)
1074
131
(Continued)
S2
male infertility
Search modes - Boolean/Phrase
Interface - EBSCOhost Search Screen - Advanced Search Database - CINAHL Plus Interface - EBSCOhost Search Screen - Advanced Search Database - CINAHL Plus
168
S1
(MH Antioxidants+) or Search modes (MH Berries+) or - Boolean/Phrase (MH Chlorophyll) or (MH Flavonoids+) or (MH Lycopene) or (MH Polyphenols+)
8581
Appendix 5. PsycINFO search strategy

1 exp male infertility/ (0) 2 (asthenozoospermia or oligospermia or azoospermia).tw. (26) 3 exp Spermatozoa/ (0) 4 Sperm$.tw. (1871) 5 (male$ adj2 subfertil$).tw. (6) 6 (male$ adj2 infertil$).tw. (114) 7 (subfertil$ adj2 men).tw. (0) 8 (infertil$ adj2 men).tw. (46) 9 (male$ adj2 fertility).tw. (89) 10 semen.tw. (268) 11 oligoasthenoteratozoospermi$.tw. (1) 12 or/1-11 (2254) 13 exp antioxidants/ or free radical scavengers/ (736) 14 (antioxidant$ or radical scavengers).tw. (1433) 15 exp vitamins/ or exp ascorbic acid/ or exp dehydroascorbic acid/ or exp vitamin a/ or exp vitamin e/ or exp vitamin u/ or exp alphatocopherol/ or exp beta carotene/ or exp beta-tocopherol/ or exp gamma-tocopherol/ (2281) 16 vitamin$.tw. (3287) 17 exp Zinc/ (317) 18 exp Selenium/ (0) 19 (Glutathione$ or folate).tw. (1280) 20 exp Ubiquinone/ (0) 21 ubiquin$.tw. (52) 22 coenzyme q10.tw. (49) 23 exp Carnitine/ (0) 24 (carnitine$ or carotenoid$).tw. (310) 25 (astaxanthin$ or lycopene$).tw. (16) 26 menevit.tw. (0) 27 multivitamin$.tw. (115) 28 (betacarotene$ or beta carotene$).tw. (64) 29 ascorbic acid.tw. (264) 30 n-acetylcysteine.tw. (65) 31 alpha-tocopherol$.tw. (127)
32 exp Pentoxifylline/ (0) 33 Pentoxifylline$.tw. (40) 34 (sh adj2 oil$).tw. (87) 35 omega$.tw. (908) 36 exp fatty acids/ or exp sh oils/ or exp cod liver oil/ or exp fatty acids, omega-3/ or exp plant oils/ (1906) 37 fatty acid$.tw. (1589) 38 (plant adj4 oil$).tw. (18) 39 l-arginine$.tw. (555) 40 avonoid$.tw. (87) 41 carotenoid$.tw. (153) 42 riboavin$.tw. (92) 43 pycnogenol$.tw. (9) 44 lutein$.tw. (1034) 45 lipoic acid$.tw. (52) 46 or/13-45 (11595) 47 46 and 12 (29) 48 from 47 keep 1-29 (29)
Appendix 6. AMED search strategy

1 exp male infertility/ (0) 2 (asthenozoospermia or oligospermia or azoospermia).tw. (12) 3 exp Spermatozoa/ (32) 4 Sperm$.tw. (151) 5 (male$ adj2 subfertil$).tw. (4) 6 (male$ adj2 infertil$).tw. (107) 7 (subfertil$ adj2 men).tw. (2) 8 (infertil$ adj2 men).tw. (7) 9 (male$ adj2 fertility).tw. (24) 10 semen.tw. (82) 11 oligoasthenoteratozoospermi$.tw. (0) 12 or/1-11 (285) 13 exp antioxidants/ or free radical scavengers/ (1009) 14 (antioxidant$ or radical scavengers).tw. (1775) 15 exp vitamins/ or exp ascorbic acid/ or exp dehydroascorbic acid/ or exp vitamin a/ or exp vitamin e/ or exp vitamin u/ or exp alphatocopherol/ or exp beta carotene/ or exp beta-tocopherol/ or exp gamma-tocopherol/ (2001) 16 vitamin$.tw. (1947) 17 exp Zinc/ (92) 18 exp Selenium/ (83) 19 (Glutathione$ or folate).tw. (574) 20 exp Ubiquinone/ (0) 21 ubiquin$.tw. (28) 22 coenzyme q10.tw. (62) 23 exp Carnitine/ (14) 24 (carnitine$ or carotenoid$).tw. (157) 25 (astaxanthin$ or lycopene$).tw. (36) 26 menevit.tw. (0) 27 multivitamin$.tw. (52) 28 (betacarotene$ or beta carotene$).tw. (91) 29 ascorbic acid.tw. (374) 30 n-acetylcysteine.tw. (25) 31 alpha-tocopherol$.tw. (73) 32 exp Pentoxifylline/ (0)
33 Pentoxifylline$.tw. (7) 34 (sh adj2 oil$).tw. (142) 35 omega$.tw. (173) 36 exp fatty acids/ or exp sh oils/ or exp cod liver oil/ or exp fatty acids, omega-3/ or exp plant oils/ (1076) 37 fatty acid$.tw. (588) 38 (plant adj4 oil$).tw. (689) 39 l-arginine$.tw. (101) 40 avonoid$.tw. (903) 41 carotenoid$.tw. (91) 42 riboavin$.tw. (16) 43 pycnogenol$.tw. (15) 44 lutein$.tw. (47) 45 lipoic acid$.tw. (35) 46 or/13-45 (6342) 47 46 and 12 (24) 48 from 47 keep 1-24 (24)
Appendix 7. MDSG Specialised Register search strategy

Keywords CONTAINS antioxidants or antioxidant levels or vitamin or vitamin A or vitamin B or Vitamin-B-12 or Vitamin-B-12-Therapeutic-Use or vitamin B6 or vitamin C or Vitamin D or vitamin E or vitamins or selenium or folic acid or glutathione or Menevit anti-oxidant or carnitene or carnitine or ascorbic acid or zinc or fatty acids or oil or sh oils or plant extracts or Title CONTAINS fatty acids or oil or sh oils or plant extracts or antioxidants or antioxidant levels or vitamin or vitamin A or vitamin B or Vitamin-B-12 or Vitamin-B-12-Therapeutic-Use or vitamin B6 or vitamin C or Vitamin D or vitamin E or vitamins or selenium or folic acid or glutathione or Menevit antioxidant or carnitene or carnitine or ascorbic acid or zinc AND Keywords CONTAINS IVF or ICSI or idiopathic asthenospermia or idiopathic oligozoospermia or in-vitro fertilisation or in-vitro fertilisation procedure or in vitro fertilization or intracytoplasmic sperm injection or intracytoplasmic morphologically selected sperm injection or superovulation or superovulation induction or IUI or insemination, intrauterine or Intrauterine Insemination or ART or articial insemination or Sperm or sperm DNA integrity or sperm damage or sperm quality or sperm parameters or oligo-asthenozoospermia or Oligoasthenospermia or oligoasthenoteratozoospermia or oligospermia or oligozoospermia or asthenospermia or asthenozoospermia or assisted reproduction techniques or azoospermia
Appendix 8. The World Health Organization International Trials Registry Platform search portal
1) Antioxidants AND men 2) Vitamins AND men 3) Antioxidant* AND male 4) Vitamin* AND male 5) Infertiltiy AND men
Appendix 9. ClinicalTrials.gov trials register

1) Antixidants AND men 2) Vitamins AND men
134
Appendix 10. OpenSigle

1) Antioxidant* 2) Vitamin* 3) Infertility AND Men
WHATS NEW
Last assessed as up-to-date: 21 August 2010.
Date 3 May 2011
Event Amended
Description 2.1 Analysis edited to xed effect Peto. The conclusions remain the same.
HISTORY
Protocol rst published: Issue 4, 2008 Review rst published: Issue 1, 2011
Date 8 March 2011 21 December 2010 4 May 2007
Event Amended Amended New citation required and major changes
Description Changed summary of ndings table to reect quality of studies Minor edits made - no changes to conclusions Substantive amendment
CONTRIBUTIONS OF AUTHORS
MGS: initiated, conceptualised and wrote the protocol. Performed the searches, selected trials for inclusion, assessed quality, performed data extraction, entered data and wrote the nal review. RH: advised and supervised both the protocol and review, helped select trials for inclusion and wrote the implications for practice and research and assisted with the abstract. JB: co-drafted the protocol. Selected trials for inclusion, assessed quality and performed data extraction. Julie also provided advice on the data analysis and helped with incorporating the editorial comments into the review. AY: co-drafted the protocol and provided Word document concerning sperm DNA for background and provided technical advice. MS: co-drafted the protocol and provided technical advice.
135
DECLARATIONS OF INTEREST
None known
SOURCES OF SUPPORT Internal sources

Cochrane Menstrual Disorders and Subfertility Group, Not specied.
External sources
No sources of support supplied
DIFFERENCES BETWEEN PROTOCOL AND REVIEW

Sperm outcomes of concentration and motility were added as these two sperm outcomes are thought to reect the oxidative process. A study by El-Taieb (El-Taieb 2009) states that increased ROS generation and reduced antioxidant capacity is negatively correlated with sperm concentration and motility in infertile men. The comparisons antioxidant versus placebo and antioxidants versus no treatment were combined as one comparison antioxidants versus control and then it was stated under sensitivity analysis whether exclusion of those that failed to use placebo would have altered the conclusions - as per statistical advice in editorial comments. Subgrouping and sensitivity analysis were performed on the outcomes of live birth and pregnancy in order to assess the potential of overestimation of benet and reporting bias. Sensitivity analysis was performed on trials that enrolled couples undergoing IVF or ICSI or IUI.
INDEX TERMS Medical Subject Headings (MeSH)

Antioxidants [ therapeutic use]; DNA Damage; Infertility, Male [ drug therapy; etiology]; Live Birth; Oxidative Stress [ drug effects]; Pregnancy Rate; Sperm Count; Sperm Motility [drug effects]; Spermatozoa [drug effects]
MeSH check words

Female; Humans; Male; Pregnancy
136

Antioxidants For Male Sub Fertility (Review)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Antioxidants For Male Sub Fertility (Review)

Uploaded by

Copyright:

Available Formats

Antioxidants for male subfertility (Review)

Showell MG, Brown J, Yazdani A, Stankiewicz MT, Hart RJ

1 1 2 2 5 6 6 9 13 14 15 16 17 19 19 21 23 24 25 27 28 29 30 31 32 33 34 35 36 37 38 38 39 40 41 41 42 46 91 98 99 101 102 102 105 106 107 109 110 111

Antioxidants for male subfertility

Assumed risk Control

Corresponding risk Antioxidant/s control 90 per 1000 (38 to 200) versus

OR 4.85 (1.92 to 12.24)

118 per 1000 (78 to 174)

OR 4.18 (2.65 to 6.59)

964 (15 studies)

29 per 1000 (6 to 124)

OR 1.54 (0.32 to 7.3)

454 (10 studies)

Description of the condition

Description of the intervention

Criteria for considering studies for this review

How the intervention might work

Why it is important to do this review

Exclusion criteria Any quasi-randomised trials Types of participants

Types of outcome measures

Live birth rate per couple randomised

Search methods for identication of studies

Data collection and analysis

The primary outcome for this review was as follows.

Studies awaiting classication

Risk of bias in included studies

Live birth per couple randomised

Pregnancy rate per couple randomised

Stillbirth rate per couple randomised

Level of sperm DNA fragmentation

Total sperm motility at three months or less

Total sperm motility at six months

Sperm concentration at three months or less

Sperm concentration at nine months or more

Total sperm motility and sperm concentration over time

Gastrointestinal side effects

Pregnancy rate per couple randomised

Sperm total motility at nine months

Sperm concentration at three months

Sperm concentration at six months

DISCUSSION Summary of main results

Overall completeness and applicability of evidence

Quality of the evidence

Potential biases in the review process

conceiving. More research is required to further substantiate these conclusions.

Implications for research

Agreements and disagreements with other studies or reviews

AUTHORS CONCLUSIONS Implications for practice

References to studies included in this review

References to studies excluded from this review

References to studies awaiting assessment

References to ongoing studies

CHARACTERISTICS OF STUDIES Characteristics of included studies [ordered by study ID]

No incomplete outcome data

Low risk Unclear risk

No statement regarding funding Can not denitively assess other bias

Outcomes Notes Risk of bias Bias

Random sequence generation (selection Unclear risk bias)

Allocation concealment (selection bias)

Can not denitively assess other bias

Random sequence generation (selection Unclear risk bias)

Allocation concealment (selection bias)

Incomplete outcome data (attrition bias) All outcomes