OtolaryngologyHead and Neck Surgery (2006) 134, 357-364

ORIGINAL RESEARCH

Antibiotics to Improve Recovery Following Tonsillectomy: A Systematic Review

Muthuswamy Dhiwakar, MS, MRCS, Chee Y. Eng, MRCS, DOHNS, Sivasubramaniam Selvaraj, MSc, and William S. McKerrow, FRCS, Inverness, United Kingdom
OBJECTIVE: To determine if antibiotics improve recovery following tonsillectomy. STUDY DESIGN: Data sources: Electronic databases Medline, Embase, and Cochrane Controlled Trials Register were searched using relevant search terms. Additional trials, if any, were retrieved by searching the references from all identied trials, reviews, correspondences, editorials, and conference proceedings. No language restriction was applied. Study selection: Systematic review of trials in which antibiotic was administered as a study medication intraoperatively and/or postoperatively, in children or adults undergoing tonsillectomy or adenotonsillectomy. Only randomized, placebo-controlled, double-blind trials attaining preset quality scores were included. Outcomes analyzed: 1) pain, need for analgesia, fever, halitosis, and return to normal diet and activities; 2) secondary hemorrhage using 2 parameterssignicant hemorrhage (ie, warranting readmission, blood transfusion, or return to theatre for hemostasis) and total hemorrhage; and 3) adverse events. RESULTS: Five trials met the eligibility criteria. Antibiotics signicantly reduced the number of subjects manifesting fever (relative risk [RR]: 0.62, 95% condence interval [CI]: 0.45, 0.85) and duration of halitosis ( 1.94 [ 3.57, 0.30] days), and marginally reduced the time taken to resume normal activity ( 0.63 [ 1.12, 0.14] days), but had no signicant effect in reducing pain scores ( 0.01 [ 0.60, 0.57]) or need for analgesia. Similarly, there was no signicant difference in the time taken to resume normal diet or incidence of signicant and total hemorrhage, although data was underpowered to detect differences for these outcomes. In the antibiotic group 4 patients developed an adverse reaction (3 cases of rash and 1 case of oropharyngeal candidiasis), while in the control group 1 patient had an adverse reaction (rash). The RR of antibiotic-related adverse events was 2.45 (0.45, 13.31).
From the Departments of OtolaryngologyHead and Neck Surgery (Drs Dhiwakar, Eng, and McKerrow) and Biostatistics (Dr Selvaraj), Raigmore Hospital.

CONCLUSION: Antibiotics appear to be effective in reducing some, but not all, morbid outcomes following tonsillectomy, and may increase the risk of adverse events. Further trials are needed to better dene the role of antibiotics in facilitating post-tonsillectomy recovery. EBM rating: A-1a 2006 American Academy of OtolaryngologyHead and Neck Surgery Foundation, Inc. All rights reserved.

onsillectomy continues to be one of the most common surgical procedures performed in children and adults. Despite improvement in surgical and anesthetic techniques, postoperative morbidity remains a signicant clinical problem. The most common problems encountered in the postoperative period include pain and the consequent need for analgesics, as well as the inability to resume normal diet and activity for a few hours to several days. Postoperative hemorrhage is reported to occur in 1.9% to 3.3% of cases,1,2 and may incur additional morbidity in the form of readmission, blood transfusion, and return to theatre for hemostasis. After surgery, the tonsillar fossa is left open to heal by secondary intention and is contaminated by oropharyngeal bacteria.3 Several authors argue that this predisposes to an inammatory reaction and infection and contributes to postoperative morbidity. They therefore recommend prophylactic antibiotics to reduce the morbidity,3-5 and its routine administration intraoperatively and/or in the postoperative period has become frequent in otolaryngology practice.6,7 A recent survey of American otolaryngologists revealed 79% of the responders to routinely prescribe antibiotics in the perioperative period, with decreased pain, decreased inammation, and faster healing being cited as the most common
Reprint requests: M. Dhiwakar, 49 Cedarwood Drive, Inverness, UK IV2 6GU. E-mail address: dhiwamahi@yahoo.com.

0194-5998/$32.00 2006 American Academy of OtolaryngologyHead and Neck Surgery Foundation, Inc. All rights reserved. doi:10.1016/j.otohns.2005.12.016

358

OtolaryngologyHead and Neck Surgery, Vol 134, No 3, March 2006

reasons.7 Other studies, however, report no benet due to antibiotics,8,9 and the subject remains contentious. Given the high volume and priority of tonsillectomy in otolaryngology practice and the current emphasis on evidence-based medicine, the effectiveness of antibiotic in facilitating recovery following tonsillectomy needs to be evaluated by a systematic review.

Selection Criteria
Trials included were those in which antibiotic was administered as a study medication intraoperatively and/or postoperatively, in children or adults undergoing tonsillectomy or adenotonsillectomy. Only randomized, placebo-controlled, double-blind trials attaining a quality score of 3 or more were included in the review. Excluded from the review were trials in which antibiotic was administered topically or, where explicitly stated, in the immediate preoperative period (ie, within 7 days before surgery). Also where explicitly stated, patients undergoing the following procedures were excluded: unilateral tonsillectomy, tonsillar biopsy, tonsillectomy for known carcinoma, tonsillectomy in conjunction with palatal surgery, and hot tonsillectomy for peritonsillar abscess.

METHODS
Search Strategy
Two reviewers (MD and CYE) independently extracted trials for possible inclusion. No language restriction was applied for the literature search. Electronic databases searched were Medline (1966-2004; search terms tonsil and tonsillectomy, each searched in combination with antibiotic), Embase (1980-2004; search terms tonsil, tonsillectomy, antibiotic, and antibiotic therapy [each term exploded] used in all possible combinations), and Cochrane Controlled Trials Register (CENTRAL; search terms tonsil and tonsillectomy, each searched in combination with antibiotic). Abstracts from the following conference proceedings were searched: Otolaryngological Research Society (1996 April 2004), Netherlands Society for Otorhinolaryngology and Cervico-Facial Surgery (1995April 2003), British Association of OtolaryngologistsHead and Neck Surgeons (19962002), American Academy of OtolaryngologyHead and Neck Surgery Foundation (2003-2004), British Association for Paediatric Otorhinolaryngology (1994-2003), British Academic Conference in Otolaryngology (2003), and The American Laryngological, Rhinological & Otological Society (1999-2004). Additional articles, if any, were retrieved by searching the reference lists from all identied studies, reviews, correspondences, and editorials.

Outcomes
Outcomes evaluated were 1) pain, analgesia consumed, fever, halitosis, and time taken to resume normal diet and activities; 2) secondary hemorrhage using 2 parameters signicant hemorrhage (ie, warranting readmission, blood transfusion, or return to theatre for hemostasis) and total (any documented) hemorrhage; and 3) adverse events. Hemorrhage occurring within 24 hours of surgery was classied as primary, and any hemorrhage occurring beyond this time period was classied as secondary. If an eligible trial did not evaluate or report any of the outcomes detailed above, that trial was excluded from the analysis of that particular outcome. Intention-to-treat analysis was attempted, wherein all participants randomized into a trial, irrespective of which (or how much) treatment they actually received and regardless of other protocol irregularities such as ineligibility, were included for analysis.

Statistics
If not explicitly stated, standard deviation (SD) values for continuous outcomes were estimated using raw data, mean, t values, and/or P values. Summary measures were calculated where possible for combinable data. Given the expected variability in participants, interventions, outcomes studied, and trial design and quality, DerSimonian and Laird random-effects models were used to generate summary measures. RevMan version 4.2 was used for the analysis.

Quality Assessment of the Studies

The quality of each trial was graded according to the validated and well-established criteria proposed by Jadad et al,10 and the CONSORT statement for improving the quality of randomized trials.11 One point was given for each of the following 5 criteria, for a maximum score of 5: 1) trial described as randomized; 2) all 3 steps of the randomization process, ie, randomization sequence generation, allocation concealment, and implementation, described and appropriate; 3) trial reported as double-blind; 4) double-blinding method described and appropriate; and 5) completeness of follow-up and specication of the numbers of and reasons for withdrawals and dropouts. Where necessary, the rst author of the relevant trial was contacted for additional information regarding methodology and/or results. Two reviewers (MD and CYE), who were blinded (by editing copies of the article) to the authors, institution, journal, and results, independently assessed trial quality. Any disagreement was resolved by the senior reviewer (WM).

RESULTS
Study Selection
Twenty-two trials evaluating the role of intraoperative and/or postoperative antibiotics in reducing morbidity following tonsillectomy were identied by the systematic search. The following were excluded from the review: 4 uncontrolled12-15 and 3 nonrandomized controlled16-18 trials, 3 trials examining topical antibiotics,19-21 and 2 trials with inappropriate control groups (1 compared antibiotic vs another antibiotic,22 1 compared systemic vs topical antibi-

Dhiwakar et al

Antibiotics to Improve Recovery Following . . .

359

Table 1 Quality assessment of included trials Randomization Study 1. OReilly et al9 2. Telian et al3 3. Grandis et al5 4. Mann et al28 5. Cannon8 Yes/No Yes Yes Yes Yes Yes Method By central pharmacy ; Geigy table of random numbers & sealed envelopes By central pharmacy; method unclear By central pharmacy; method unclear Computer-generated random numbers, rest unclear By central pharmacy; method unclear Double-blinding Yes/No Yes Yes Yes Yes Yes Method Dropout details Total score* 4 4 3 3 3

Identical placebo Not given Identical placebo Identical placebo Near-identical placebo Placebo Given** Not given Not given Not given

*maximum score 5 (See text for details). after communication with author. **article specied number of and reasons for dropout. author conrmed dispensation of placebo as a capsule, but could not conrm if it was exactly identical to amoxicillin capsule. no reply from author yet. double blinding satisfactory, as both study medication and placebo were single intravenous doses.

otic23). Also excluded were 4 randomized controlled, nonblinded trials.4,24-26 One randomized controlled trial (translated from Portuguese) wherein only the outcome assessor was blinded was also excluded.27 Thus a total of 5 randomized, placebo-controlled, double-blind trials, evaluating the impact of systemic antibiotics on post-tonsillectomy morbidity,3,5,8,9,28 were identied. All 5 studies attained the minimum quality score of 3 and were included for analysis (Table 1).

Table 2 illustrates important data from the included studies. All studies compared a short course of systemic antibiotics vs placebo. The study by Mann et al included 2 additional arms, in which patients received 2 different topical antibiotics to the tonsillar fossa.28 These additional arms were unsuitable for analysis and therefore excluded. Intention-to-treat analysis was possible in two trials,3,9 while only per-protocol analysis (analysis of the results of only those participants who completed the trial and who

Table 2 Summary of data from the 5 included trials Distribution N Study 1. OReilly9 Age range (yr) 16-53 NR 200 Dropout n (%) 105 (52.5)* Antibiotic 46 Control 49 Antibiotic drug, dose, route, & duration Amoxicillin, IV 250 mg at induction, followed by 250 mg tds oral for 10 days Amoxicillin, IV at surgery and qds for 24 hrs, followed by oral tds for 7 days Ticarcillin clavulanic acid, IV 3.1 gm at surgery, 6 & 12 hrs after surgery; followed by amoxicillin clavulanic acid 250 mg tds oral for 7 days Amoxicillin, tds oral for 7 days Cefonicid, IV 1 gm at surgery Follow-up (days) 10

2. Telian et al3

Children (NR)

100

0 (0)

45

40

7-14

3. Grandis et al5

12-48

198

97 (49)*

51

50

4. Mann et al28 5. Cannon8

17 13-40

NR 50

NR* 4 (8)

8 24

10 22

7 7

NR, not reported; IV, intravenous; NR, total number of patients after randomization; tds, three times a day; qds, four times a day. weight-adjusted dosage: IV 500 mg if 20 kg, IV 1 gm if 20 kg; oral 125 mg if 20 kg, oral 250 mg if 20 kg. *comparable numbers between study and control groups lost to follow-up.

360
Table 3 Meta-analysis of pain scores* Antibiotic Study Mann28 Grandis5 OReilly9 Total N 8 51 46 105

OtolaryngologyHead and Neck Surgery, Vol 134, No 3, March 2006

Placebo N 10 50 49 109 Mean (SD) 6.88 (2.02) 5.94 (2.02) 3.47 (2.02) WMD (random) 95% CI 0.24 0.18 0.21 0.01 ( ( ( ( 2.28, 1.03, 0.67, 0.60, 1.80) 0.67) 1.09) 0.57)

Mean (SD) 6.64 (2.32) 5.76 (2.32) 3.68 (2.32)

WMD, weighted mean difference. *converted to a scale of 1-10 over 5 days.

complied with their allocated treatment) was possible in the rest. Anesthetic technique was not controlled in any study. Surgical technique was adequately controlled in 2 studies,3,8 both of which reported the use of dissection and snare of the tonsils, with electrocautery for hemostasis. Surgical technique was quasi-controlled in 2 studies.9,28 One reported using mostly electrodissection for tonsillectomy,9 while the other did not mention the tonsillectomy technique, and instead only mentioned the hemostatic method (electrocautery).28 In one trial,5 surgical technique was not controlled. All of the following 3 exclusion criteria1) antibiotic administered within 1 week preoperatively; 2) medical condition requiring perioperative antibiotic; and 3) allergy to antibiotic studiedwere explicitly applied in every trial, except one.9

were measured and converted to a scale of 1 to 10. Mean pain scores along with mean of the measured SE values were calculated, which allowed estimation of SD for each group. Similarly, line graphs of pain scores in the studies by OReilly et al9 and Grandis et al5 were used to derive mean pain scores on a scale of 1 to 10 over 5 days. As no other data were published in these 2 trials, the SD values from Mann et als trial28 were imputed. Telian et al3 reported the percentage of patients manifesting fever, which was converted to the nearest number for meta-analysis.

Outcome Analyses
Mann et als trial had a small sample size (Table 2), limiting its power to detect any differences in outcomes.28 Hence, this trial was excluded from the analyses, except where meta-analysis was possible.

Data Extraction
No study explicitly stated SD values for any outcome measure. Raw data was available from bar graphs for the outcomes reported by Cannon.8 In the trial by Telian et al,3 for time taken to resume normal diet and activity, and for resolution of halitosis, pooled SD were estimable from published mean differences and t and P values. In the trial by Grandis et al,5 for time taken for resolution of halitosis, pooled SD was calculable from the published mean difference and P value. Mann et al28 reported mean pain scores for each of 5 days as a bar graph on a scale of 0 to 100 with standard error (SE) bars. These bars, including the SE bars,

Pain
Three studies assessed pain using linear pain scores, and none found a signicant reduction with antibiotics. OReilly et al9 used a scale of 1 to 5 over 10 days, Mann et al28 used a scale of 0 to 100 over 5 days, and Grandis et al5 used a scale of 1 to 10 over 7 days. Meta-analysis of pain scores converted to a scale of 1 to 10 over 5 days similarly showed no reduction in pain with antibiotics (P 0.97; Table 3 & Fig 1). However, Telian et al found the mean number of days with continuous subjective pain to be improved with antibiotics (3.3 vs 4.4; P 0.05).

Table 4 Meta-analysis of number of subjects with fever* Study Telian3 Grandis5 Total Figure 1 Forest plot of differences in mean pain scores (on a converted scale of 1-10 in the rst 5 days). Test for heterogeneity P 0.80. WMD, weighted mean difference. Antibiotic (n/N) 23/45 8/51 96 Placebo (n/N) 32/40 14/50 90 RR (random) 95% CI 0.64 (0.46, 0.88) 0.56 (0.26, 1.22) 0.63 (0.46, 0.85)

RR, relative risk. *temperature 99.9F in the rst 7 days.

Dhiwakar et al

Antibiotics to Improve Recovery Following . . .

361

Figure 2 Forest plot of differences in number of subjects with fever (temperature 99.9F in the rst 7 days). Test for heterogeneity P 0.74. RR, relative risk.

Figure 3 Forest plot of differences in number of days taken to resume normal activity. Test for heterogeneity P 0.27. WMD, weighted mean difference.

Need for Analgesia

Two studies assessed the need for analgesics. Neither found a signicant reduction with antibiotic use. OReilly et al9 reported a mean of 43% and 46% subjects in the antibiotic and placebo groups, respectively, to consume additional analgesics. A respective mean of 63% and 51% contacted primary care physicians for analgesia. Grandis et al5 found no difference between the antibiotic and placebo groups in the mean number of days wherein more than 5 doses of pain medicine was taken (1.8 vs 2.4 respectively), or in the mean number of doses of pain medication (19.45 vs 19.23).

(weighted mean difference: 0.56 [ 1.62, 0.50] days; P 0.30), although data was of insufcient power to detect any difference. Grandis et al5 reported antibiotics to facilitate earlier return to normal diet, which just reached statistical signicance (P 0.05).

Time Taken to Resume Normal Activity

Antibiotics marginally reduced the number of days taken to resume normal activity (P 0.02; Table 5 & Fig 3). Grandis et al5 similarly reported earlier return to work or school with antibiotics (P 0.045).

Fever
Antibiotics reduced the number of patients manifesting fever (temperature 99.9F) in the rst 7 days (P 0.002; Table 4 & Fig 2). Telian et al3 also reported a signicant reduction in the mean number of oral temperature recordings more than 100F (1.5 vs 2.9; P 0.05) and more than 101.5F (0.02 vs 0.23; P 0.05), with use of antibiotics. Similarly, Cannon8 reported a signicant reduction in the number of patients with fever ( 99F; 6 vs 16; P 0.003). Grandis et al5 reported reduction in the mean number of days with fever ( 99.9F), although it did not reach statistical signicance (0.35 vs 0.51; P 0.05).

Halitosis
Antibiotics reduced the number of days with halitosis (P 0.02; Table 6 & Fig 4). Telian et al3 also noted a reduction in the number of days with severe odor (0.4 vs 2.1; P 0.001). Similarly, Grandis et al5 reported a reduction in the cumulative mouth odor scores with antibiotics (P 0.05).

Secondary Hemorrhage
Four studies evaluated for postoperative hemorrhage and reported incidence rates.3,5,9,28 Correspondence with the rst author of these trials claried the distinction between signicant and insignicant but documented hemorrhage. Cannon reported that the rate of postoperative hemorrhage was the same in both groups, but did not give the exact incidence rates.8 This trial was hence excluded from the meta-analyses for hemorrhage. No incidence of primary hemorrhage was reported in any trial. No fatality occurred

Time Taken to Resume Diet

Meta-analysis of the data from the trials by Telian et al3 and Cannon8 revealed antibiotics not to signicantly reduce the number of days required to resume soft or normal diet

Table 5 Meta-analysis of number of days taken to resume normal activity Antibiotic Study Telian3 Cannon8 Total N 45 24 69 Mean (SD) 4.20 (1.92) 6.29 (1.20) N 40 22 62 Placebo Mean (SD) 5.20 (1.92) 6.71 (0.90) WMD (random) 95% CI 1.00 ( 1.82, 0.18) 0.42 ( 1.03, 0.19) 0.64 ( 1.20, 0.09)

WMD, weighted mean difference.

362

OtolaryngologyHead and Neck Surgery, Vol 134, No 3, March 2006

Table 6 Meta-analysis of number of days with halitosis Antibiotic Study Telian3 Grandis5 Cannon8 Total N 45 51 24 120 Mean (SD) 2.60 (2.85) 2.50 (2.90) 2.86 (1.96) N 40 50 22 112 Placebo Mean (SD) 6.10 (2.85) 4.20 (2.90) 3.39 (2.81) WMD (random) 95% CI 3.50 1.70 0.53 1.94 ( ( ( ( 4.71, 2.29) 2.83, 0.57) 1.94, 0.88) 3.57, 0.30)

WMD, weighted mean difference.

and no bleeding event reoccurred in the same patient. Antibiotics did not signicantly reduce secondary hemorrhage rates, although all trials were underpowered to distinguish differences. A total of 9 signicant hemorrhages (RR with antibiotics: 0.72 [0.05, 10.38; P 0.81]) and 27 total hemorrhages (RR: 0.90 [0.38, 2.12; P 0.81]) occurred in 314 participants.

DISCUSSION
The present systematic review suggests that antibiotics reduce post-tonsillectomy fever and halitosis, and marginally reduce the time taken to resume normal activity. The reduction in fever with antibiotics is likely due to the amelioration of bacteremia, which is recognized to occur during and immediately after tonsillectomy.29 Similarly, the improvement in halitosis is attributable to a reduction in the count of malodorous bacteria that tend to contaminate the postoperative tonsillar fossa.4,5 Nevertheless, the severity of halitosis as a morbid outcome and the clinical signicance of the reduction in halitosis and time taken to resume normal activity by 0.64 day are unclear. Antibiotics appear to have no signicant effect in reducing pain or need for analgesia and have an equivocal effect in reducing the time taken to resume normal diet and secondary hemorrhage. This challenges the widely held rationale for routine prescription of antibiotics. A causal relationship between bacterial inammation of the tonsillar fossa and postoperative morbidity such as pain, need for analgesia, inability to resume normal swallowing, and secondary hemorrhage is yet to be proven. The difculty of establishing such a correlation is complicated by the variety of commensals harbored in the oropharyngeal mucosa. This renders bacterial culture results of the postoperative tonsillar fossa as reported in a few studies3-5 difcult to interpret. It is also not known if the dosage and antibacterial spectrum of antibiotics used in the trials have been adequate or if there exists a dose-response effect. The potential for adverse events also needs to be considered while prescribing antibiotics. Although no major adverse reaction has been documented in the included trials, cases of anaphylaxis and death continue to be reported as a result of administering antibiotics.30,31 An incidence of 0.18% of adverse events following outpatient prescription of penicillin has been reported.32 Given the frequency and volume of tonsillectomy as a surgical procedure, this small risk of minor and major adverse events may translate into signicant harm when balanced against the limited effectiveness of antibiotics.

Adverse Events
There was no major adverse event reported in either group. With regard to minor adverse events, in the antibiotic group, there were 3 cases manifesting a rash3,5,28 and 1 developed oropharyngeal candidiasis.3 In comparison, 1 patient in the control group developed a rash.5 The RR of antibioticrelated adverse events was 2.45 (0.45, 13.31; P 0.30).

Miscellaneous Outcomes
With antibiotic administration, no signicant reduction in otalgia,5,28 trismus,3,8 eschar formation in the tonsillar fossa,3,8 or weight loss3 was achieved. Cannon8 reported weight loss of 5.3 pounds and 8.1 pounds in the antibiotic and placebo groups respectively, but did not comment whether the difference was statistically signicant. Grandis et al were the only investigators reporting diarrhea as a side effect, which occurred in 3 and 2 subjects in the antibiotic and placebo groups respectively.5

Figure 4 Forest plot of differences in number of days with halitosis. Test for heterogeneity P 0.006. WMD, weighted mean difference.

Dhiwakar et al

Antibiotics to Improve Recovery Following . . .

363

We acknowledge several limitations in this review. The most obvious is the lack of good-quality studies that limits the evidence and underpowers the data to detect differences in certain outcomes. Secondly, despite our active search, it is possible that unpublished trials have been missed, as in any systematic review. Unpublished trials are well recognized to be associated with null or insignicant results.33,34 This potentially biases the results in favor of antibiotics. However, in trials examining the role of antibiotics, a null result is as signicant and clinically important as a positive one, and should not have adversely impacted on the potential for publication. Thirdly, only one trial9 was conrmed to fulll all essential steps for a robust randomization process. Studies with inadequate allocation concealment may overestimate treatment effect by 37%,34 which implies that the results of this review are potentially skewed in favor of antibiotics. Lastly, the overall dropout rate is high with a large potential for attrition bias. However, within each study the number of patients lost to follow-up is comparable between the study and control groups (Table 2), possibly limiting the bias.

REFERENCES
1. Wei JL, Beatty CW, Gustafson RO. Evaluation of posttonsillectomy hemorrhage and risk factors. Otolaryngol Head Neck Surg 2000; 123(3):229 35. (Grade B). 2. Lowe D, van der Meulen J; National Prospective Tonsillectomy Audit. Tonsillectomy technique as a risk factor for postoperative haemorrhage. Lancet 2004;364(9435):697702. (Grade B). 3. Telian SA, Handler SD, Fleisher GR, et al. The effect of antibiotic therapy on recovery after tonsillectomy in children. A controlled study. Arch Otolaryngol Head Neck Surg 1986;112(6):610 5. (Grade A). 4. Colreavy MP, Nanan D, Benamer M, et al. Antibiotic prophylaxis post-tonsillectomy: is it of benet? Int J Pediatr Otorhinolaryngol 1999;50(1):1522. (Grade A). 5. Grandis JR, Johnson JT, Vickers RM, et al. The efcacy of perioperative antibiotic therapy on recovery following tonsillectomy in adults: randomized double-blind placebo-controlled trial. Otolaryngol Head Neck Surg 1992;106(2):137 42. (Grade A). 6. Kay DJ, Mehta V, Goldsmith AJ. Perioperative adenotonsillectomy management in children: current practices. Laryngoscope 2003;113(4):5927. (Grade D). 7. Krishna P, LaPage MJ, Hughes LF, et al. Current practice patterns in tonsillectomy and perioperative care. Int J Pediatr Otorhinolaryngol 2004;68(6):779 84. (Grade D). 8. Cannon CR. The efcacy of a single dose antibiotic regimen in adults undergoing tonsillectomy. J Miss State Med Assoc 1996;37(11):817 21. (Grade A). 9. OReilly BJ, Black S, Fernandes J, et al. Is the routine use of antibiotics justied in adult tonsillectomy? J Laryngol Otol 2003;117(5): 3825. (Grade A). 10. Jadad AR, Moore RA, Carroll D, et al. Assessing the quality of reports of randomized clinical trials: Is blinding necessary? Control Clin Trials 1996;17:112. (Grade D). 11. Moher D, Schulz KF, Altman D; CONSORT Group. The CONSORT statement: revised recommendations for improving the quality of reports of parallel-group randomized trials. JAMA 2001;285(15):1987 91. (Grade D). 12. Al-Kindy SA. Do antibiotics decrease post-tonsillectomy morbidity? Saudi Med J 2002;23(6):7057. (Grade C). 13. Juarbe C. Efcacy of ceftibuten for pediatric patients undergoing adenotonsillectomy. Bol Asoc Med P R 2000;92(9 12):110 4. (Grade C). 14. Handler SD, Miller L, Richmond KH, et al. Post-tonsillectomy hemorrhage: incidence, prevention and management. Laryngoscope 1986; 96(11):12437. (Grade C). 15. Zakirullah. Post-tonsillectomy and adenotonsillectomy morbidity & complications at District Headquarter Hospital Daggar. J Ayub Med Coll Abbottabad 2001;13(3):4 6. (Grade C). 16. Lee WC, Duignan MC, Walsh RM, et al. An audit of prophylactic antibiotic treatment following tonsillectomy in children. J Laryngol Otol 1996;110(4):3579. (Grade B). 17. Al-Tamimi SF. Should antibiotics be mandatory in perioperative tonsillectomy and/or adenotonsillectomy? J Coll Phys Surg Pak 2000; 10(6):207 8. (Grade B). 18. Szmeja Z, Golusinski W, Kruk-Zagajewska A, et al. The evaluation of ceftibuten (cedax) effectiveness in the prevention of perioperative infections. Otolaryngologia Polska 1997;51(4):390 5. (Grade B). 19. King JT. Alleviation of pain and prevention of infection after tonsillectomy. Trans Indiana Acad Ophthalmol Otolaryngol 1967;50:25 8. (Grade B). 20. Lukan N. Fusafungine after tonsillectomy. Laryngorhinootologie 2002; 81(6):426 9. (Grade A). 21. Jurkiewicz D, Wojdas A, Rapiejko P, et al. Application of the fusafungine in perioperative period in patients after tonsillectomy. Polski Merkuriusz Lekarski 2004;16(94):358 61. (Grade B). 22. Jones J, Handler SD, Guttenplan M, et al. The efcacy of cefaclor vs amoxicillin on recovery after tonsillectomy in children. Arch Otolaryngol Head Neck Surg 1990;116(5):590 3. (Grade A).

CONCLUSION
The present systematic review suggests that antibiotics reduce post-tonsillectomy fever and halitosis, and marginally reduce the time taken to resume normal activity, but have no signicant effect in reducing pain or need for analgesia. Antibiotics have an equivocal role in reducing the time taken to resume normal diet and secondary hemorrhage, and may increase the risk of adverse events. The ndings of this review will be important to health care professionals for reevaluating current practice, taking informed decisions regarding adjuvant therapy for tonsillectomy, and calculating cost-benet ratios. Any putative benet of antibiotic needs to be carefully weighed against the possible emergence of resistant bacteria, fungal colonization/ infection, risk of other adverse events, and side effects. It is incumbent upon the clinician to fully inform the patient of the potential benets, risks, and cost, as well as the lacunae in knowledge, before embarking on any proposed treatment. Further adequately powered well-designed trials are recommended to conrm and clarify our ndings and to better dene the role of antibiotics in facilitating post-tonsillectomy recovery. In future studies continuous outcome values need to be explicitly stated with mean and SD, which would facilitate meta-analyses.
The authors gratefully acknowledge the valuable help of the Cochrane Collaboration in constructing the title and locating and procuring journal copies for this review. The principal author (MD) had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

364

OtolaryngologyHead and Neck Surgery, Vol 134, No 3, March 2006

29. Soldado L, Esteban F, Delgado-Rodriguez M, et al. Bacteraemia during tonsillectomy: a study of the factors involved and clinical implications. Clin Otolaryngol Allied Sci 1998;23(1):63 6. (Grade A). 30. Kuehm SL, Doyle MJ. Medication errors: 1977 to 1988. Experience in medical malpractice claims. N J Med 1990;87(1):2734. (Grade C). 31. Chao TC. Adverse drug reactionstales of a forensic pathologist. Ann Acad Med Singapore 1993;22(1):86 9. (Grade C). 32. Apter AJ, Kinman JL, Bilker WB, et al. Represcription of penicillin after allergic-like events. J Allergy Clin Immunol 2004;113(4):764 70. (Grade B). 33. Dickersin K, Min YI. NIH clinical trials and publication bias. Online J Curr Clin Trials [serial online] 1993;Doc No 50. (Grade A). 34. Moher D, Pham B, Jones A, et al. Does quality of reports of randomised trials affect estimates of intervention efcacy reported in metaanalyses? Lancet 1998 Aug 22;352(9128):609 13. (Grade A).

23. Akbas Y, Pata YS, Unal M, et al. The effect of fusafungine on post-operative pain and wound healing after pediatric tonsillectomy. Int J Pediatr Otorhinolaryngol 2004;68(8):1023 6. (Grade A). 24. Aslam M, Mustafa G. Prophylactic antibiotics for tonsillectomy: experience at Rawalpindi General Hospital. J Coll Physicians Surg Pak 1998;8(1):36 8. (Grade A). 25. Linden BE, Gross CW, Long TE, et al. Morbidity in pediatric tonsillectomy. Laryngoscope 1990;100(2 Pt 1):120 4. (Grade A). 26. Khan MHZ, McCombie AW, Swift AC. Prophylactic antibiotic for tonsillectomy? Pakistan Journal of Otolaryngology 1994;10:122 6. (Grade A). 27. Ramos CC, Goncalves MER, Rapoport PB, et al. Prophylactic antibiotic therapy after tonsillectomy. Study with amoxicillin-clavulanic acid. Rev Bras Otorinolaringol 2000;66(6):62730. (Grade A). 28. Mann EA, Blair EA, Levy AJ, et al. Effect of topical antibiotic therapy on recovery after tonsillectomy in adults. Otolaryngol Head Neck Surg 1999;121(3):277 82. (Grade A).