World J Surg (2011) 35:25962602
DOI 10.1007/s00268-011-1298-x
Local Anesthesia: A Strategy for Reducing Surgical Site
Infections?
Jay S. Lee Awori J. Hayanga James J. Kubus
Henry Makepeace Max Hutton Darrell A. Campbell Jr.
Michael J. Englesbe
Published online: 8 October 2011
Societe Internationale de Chirurgie 2011
Abstract
Background Surgical site infections (SSIs) contribute to
increased morbidity, mortality, and hospitalization costs. A
previously unidentified factor that may reduce SSIs is the
use of local anesthesia. The objective of this study was to
determine if the use of local anesthesia is independently
associated with a lower incidence of SSIs compared to
nonlocal anesthesia.
Methods Using the American College of Surgeons
National Surgical Quality Improvement Program database
(20052007), we identified all patients undergoing surgical
procedures that could be performed using local or general
anesthesia, depending on the preference of the surgeon.
Logistic regression was used to identify factors indepen-
dently associated with the use of local anesthesia. Pro-
pensity matching was then used to match local and
nonlocal anesthesia cases while controlling for patient and
operative characteristics. SSI rates were compared using a
v2 test.
Results Of 111,683 patients, 1928 underwent local
anesthesia; and in 109,755 cases the patients were given
general anesthesia where a local anesthetic potentially
could have used. In the unmatched analysis, patients with
local anesthesia had a significantly lower incidence of SSIs
than patients with nonlocal anesthesia (0.7 vs. 1.4%,
J. S. Lee
A. J. Hayanga
J. J. Kubus
H. Makepeace

D. A. Campbell Jr.
M. J. Englesbe (&)
Department of Surgery, University of Michigan Medical School,
2926A Taubman Center, 1500 East Medical Center Drive,
Ann Arbor, MI 48109-5331, USA
e-mail: englesbe@umich.edu
M. Hutton
Department of Vascular Health, Allegiance Health, Ann Arbor,
MI, USA
123
P = 0.013). Similarly, after propensity matching, the
incidence of SSIs in patients given local anesthesia was
significantly lower than for that of patients given nonlocal
anesthesia (0.8 vs. 1.4%, P = 0.043).
Conclusions Use of local anesthesia is independently
associated with a lower incidence of SSIs. It may provide a
safe, simple approach to reducing the number of SSIs.
Introduction
Surgical site infection (SSI) is the most common infection
among surgical patients and the second most commonly
reported nosocomial infection overall [15]. SSIs contrib-
ute to an increased morbidity, mortality, hospital length of
stay (LOS), and costs [69]. Several innovative efforts
have been initiated to decrease the incidence of SSIs. These
efforts emphasize the proper timing of antibiotic adminis-
tration [1012], the importance of glycemic control [13,
14], avoidance of hypothermia [15, 16], and maintenance
of tissue oxygenation [17, 18]. However, no single effort
has succeeded in completely eliminating SSIs, even among
clean cases [19]. Currently, the incidence of SSIs for
clean cases in the American College of Surgeons National
Surgical Quality Improvement Program (ACS NSQIP)
database is 1.55%. Although a comparatively low rate,
when examined in the context of the total number of
operative cases performed annually in the United States the
problem of SSIs assumes great importance.
Despite ongoing clinical efforts to curtail SSIs, the
inability to reduce their incidence to zero suggests that
there may be contributing factors that have thus far gone
unappreciated. To identify these factors, we conducted site
visits to several hospitals in the Michigan Surgical Quality
Collaborative (MSQC) group of hospitals. MSQC is a
World J Surg (2011) 35:25962602
34-hospital group that works collaboratively to improve the
quality of surgical care. All MSQC participating hospitals
use the ACS NSQIP to assess surgical quality, which
allows comparative evaluation of results. Using this sys-
tem, we assessed surgical quality on a spectrum ranging
from best performers, defined as those with a very low mance of local anesthesia rather than general anesthesia.
incidence of SSIs, to worst performers, defined as those The independent variables for the logistic regression
with the highest incidence of SSIs. We then conducted site
visits to both the best and worst performing hospitals in an
attempt to establish best practices for the prevention of variable was the type of anesthesia used for the operative
SSIs.
During one such visit to a best performer hospital were used to verify the adequacy of the model. Linear
with a low incidence of SSIs, we recognized that more
local anesthesia was used during simple operative cases
than at any other site. This finding generated the hypothesis
that the use of local anesthesia could be a factor indepen-
dently associated with a low incidence of SSIs. To test this
hypothesis we used the ACS NSQIP public use file (PUF)
to examine the incidence of SSIs in surgical cases per-
formed under local anesthesia versus those in which gen-
eral anesthesia was used. To manage potential variations in
the case mix and selection bias, we used propensity-mat-
ched case analysis.
Materials and methods
Patients
All patients who had undergone outpatient general or
vascular surgery (n = 111,683) in the ACS NSQIP data-
base between 2005 and 2007 were included in the study. To
participate in the ACS NSQIP program, hospitals must pass
and continually undergo audits that guarantee the quality of
the extracted data. Parameters collected for each patient
included demographic information, description of medical
co-morbidities, and operative and preoperative laboratory
data. These categories have been previously described [20].
An occurrence of SSI was defined as a superficial inci-
sional, deep incisional, or organ-space infection in any
patient within 30 days of an operative procedure. Specific
surgical procedures were identified for study that could
potentially be performed under local anesthesia with sup-
plementation or general inhalation anesthesia, depending
on the preference of the surgeon. A nonlocal anesthesia
group was defined as patients who were given a general
anesthetic, intravenous (IV) anesthesia (with intubation or
laryngeal mask airway), and/or epidural, spinal, regional,
or monitored anesthesia care without the use of local
anesthesia at the operative site. The local anesthesia group
was defined by their having been administered a local
anesthetic at the operative site without the use of system-
ically administered anesthesia.
2597
Propensity matching
After defining the cases to be studied and the type of
anesthesia used, we used a logistic regression model to
identify factors independently associated with the perfor-
included patient characteristics, co-morbidities, and oper-
ative and preoperative laboratory data. The dependent
procedure. The c-statistic and HosmerLemeshow test
regression was used to impute missing preoperative labo-
ratory values from the existing laboratory values after
adjusting the model for time to prevent cases from drop-
ping out of the analysis. The SAS Greedy 5 ? 1 Digital
Match algorithm was used to match local and nonlocal
anesthesia cases for the propensity analysis [21]. To gen-
erate propensity scores, we used the full logistic regression
model. The c-statistic was 0.90. The HosmerLemeshow
P value was 0.082, which suggested that the model was
accurate in predicting whether a patient had local or non-
local anesthesia. After matching local and nonlocal anes-
thesia cases, a v2 or t-test was used to determine whether
there were any significant differences in the co-morbidities
or process of care measures between the local and nonlocal
anesthesia groups. Nonsignificant P values for differences
in patient co-morbidities and process of care measures for
these matched cases were interpreted as success in con-
trolling for patient and operative characteristics. Finally, a
v2 test was used to compare the local and nonlocal anes-
thesia matched groups to estimate whether patients treated
with local anesthesia had different SSI outcomes than
patients treated with nonlocal anesthesia.
Results
Using the ACS NSQIP PUF, we identified two groups of
patients who had undergone procedures that arguably could
have been done under local or general anesthesia. Using
these CPT codes we identified 1928 patients who under-
went the specified procedure under local anesthesia, and
109,755 patients who underwent this type of procedure
under general anesthesia. The most common procedure
types done under local anesthesia are detailed in Table 1.
These two groups were then compared with regard to
co-morbidities and operative characteristics (Table 2).
Nonlocal cases involved younger patients (51.2 vs.
55.1 years), had a significantly higher percentage of men
(39.3 vs. 35.6%), smokers (18.9 vs. 14.9%), dyspnea (6.3
vs. 4.1%), and congestive heart failure (CHF) (2.0 vs.
0.4%) compared to local cases. Nonlocal anesthesia cases
123
2598 World J Surg (2011) 35:25962602

Table 1 Most common NSQIP procedures performed under local perioperative factors between the two groups of patients
anesthesia (Table 3). The independent variables significantly associ-
Surgical description No. ated with receiving a local or nonlocal anesthetic were as
follows: work RVUs, age, weight, presence of CHF, a
Excision of cyst, fibroadenoma, or other benign or malignant 381
history of revascularization/amputation, dialysis, operating
tumor, aberrant breast tissue, duct lesion, nipple or areolar
lesion (open; male or female; 1 or more lesions) time, clean wounds, American Society of Anesthesiologists
Repair initial inguinal hernia (age 5 years or older), reducible 370 (ASA) class, and certain laboratory values including
Excision of breast lesion identified by preoperative placement 348 sodium, albumin, alkaline phosphatase, hematocrit (Hct),
of radiological marker (open; single lesion) and the international normalized ratio (INR). Patients who
Endovenous ablation therapy of incompetent vein in extremity 156 underwent procedures under local anesthesia were overall
(inclusive of all imaging guidance and monitoring); first vein less complex; they had cleaner wounds, were older, and
treated had lower body weights, lower ASA class, lower Hct, and
Mastectomy, partial (e.g., lumpectomy, tylectomy, 153 shorter operative duration. They were also more likely to
quadrantectomy, segmentectomy)
suffer from CHF and to have undergone dialysis and less
Repair umbilical hernia (age 5 years or older), reducible 73
likely to have a history of revascularization/amputation.
Excision of infected graft in extremity 52
They were more likely to have higher serum sodium,
Ligation or banding of angioaccess arteriovenous fistula 51
albumin, alkaline phosphatase, and INR levels.
Repair recurrent inguinal hernia, any age; reducible 36
Using results from the logistic regression analysis,
NSQIP National Surgical Quality Improvement Program propensity matching was performed to define closely
matched groups of patients who had had either local
Table 2 Co-morbidities and operative characteristics of patients (n = 1343) or nonlocal (n = 1343) anesthesia. After pro-
given local and nonlocal anesthesia pensity matching the groups for analysis (local versus
Variable Local anesthesia Nonlocal anesthesia P
nonlocal anesthesia), there were no longer significant dif-
(n = 1928) (n = 10,9755) ferences in potentially clinically relevant covariates
(Table 4). Importantly, the case RVU and the operative
Age 55.1 51.2 <0.001
duration were the same in the two groups. In addition,
Male sex 686 (35.6%) 43115 (39.3%) 0.004 when comparing the five most common procedures in each
Smoker 288 (14.9%) 20706 (18.9%) <0.001 matched group (Table 5) the same five procedures are
Alcohol use 38 (2.0%) 2190 (2.0%) 0.502 found for both groups.
Dyspnea 79 (4.1%) 6909 (6.3%) <0.001 Figure 1 shows the SSI rates for the unmatched and
Functional status: 1913 (99.2%) 108,642 (99.0%) 0.462 propensity-matched comparison of local and nonlocal
independent
anesthesia types. In the unmatched analysis, patients given
Ventilator 0 (0%) 35 (0%) 0.544
local anesthesia had a significantly lower incidence of SSIs
COPD 32 (1.7%) 2114 (1.9%) 0.223
than patients given nonlocal anesthesia (0.7 vs. 1.4%,
CHF 8 (0.4%) 180 (0.2%) 0.017
P = 0.013). Similarly, after propensity matching, the
Previous cardiac 78 (4%) 3344 (3%) 0.007 incidence of SSIs for patients with local anesthesia was
surgery
significantly lower than for that of patients with nonlocal
Diabetes 188 (9.8%) 9362 (8.5%) 0.031
anesthesia (0.8 vs. 1.4%, P = 0.043).
Pneumonia 1 (0.1%) 34 (0.1%) 0.456
Cancer 20 (1.0%) 1344 (1.2%) 0.262
Operating time 41.7 64.6 <0.001
(min)
Discussion
a
RVU 7.0 9.5 <0.001
Prevention of SSIs is a national priority. The Surgical Care
*Boldface type indicates significance Improvement Program, sponsored by the Center for
a
Mean work RVU (relative value units) is the metric for technical Medicare and Medicaid Services, mandates that certain
effort, stress, mental challenge, and complexity of an operation
independent of the preoperative risk factors used by the NSQIP
evidence-based SSI prevention strategies be used in sur-
gical patients throughout the country. Whether this
approach will yield measurable benefit remains to be seen.
also had significantly longer operating times and more This report is an account motivated by the collaborative
complex cases as measured by the relative value unit initiative of 34 hospitals in Michigan, the MSQC, which
(RVU) compared to the local cases. has also targeted reducing the incidence of SSI as a prior-
Logistic regression was used to adjust for the observed ity. We performed site visits to hospitals with the lowest
differences in co-morbidities, patient characteristics, and SSI to identify which practices, recommended or not, offer

123
World J Surg (2011) 35:25962602 2599

Table 3 Significant factors from the logistic regression model used to calculate propensity score (c-statistic = 0.90)
Covariate Coefficient P OR 95% CI for OR
Lower Upper

Work RVU -0.105 \0.000 0.901 0.885 0.917

Age 0.012 \0.000 1.012 1.007 1.016
Weight -0.003 \0.000 0.997 0.995 0.998
History of CHF 0.920 0.034 2.509 1.072 5.875
History of revascularization or amputation -0.823 0.027 0.439 0.211 0.912
Dialysis 1.390 \0.000 4.014 2.521 6.393
Operating time -0.016 \0.000 0.984 0.982 0.986
Clean wound 1.430 \0.000 4.177 3.396 5.138
ASA class (systemic disease)
Mild -6.397 \0.000 0.002 0.001 0.002
Severe -6.299 \0.000 0.002 0.001 0.002
Life-threatening -6.241 \0.000 0.002 0.001 0.003
Sodium 0.048 \0.000 1.049 1.022 1.078
Albumin 0.257 0.010 1.293 1.063 1.573
Alk Phos 0.002 0.011 1.002 1.000 1.004
Hematocrit -0.039 \0.000 0.961 0.945 0.978
INR 0.264 0.017 1.302 1.048 1.619
OR odds ratio, CI confidence interval, CHF congestive heart failure, ASA American Association of Anesthesiologists, Alk Phos alkaline
phosphatase, INR international normalized ratio (prothrombin time)
Odds ratio reflects the odds of getting a local anesthetic rather than a nonlocal anesthetic. Odds ratio C 1 indicates that the co-morbidity,
laboratory, or operative factor was associated with the local anesthetic; odds ratio \ 1 indicates that the co-morbidity, laboratory, or operative
factor was associated with a nonlocal anesthetic
Nonsignificant covariates included steroid use, weight loss, bleeding disorder, preoperative or intraoperative transfusion, chemotherapy,
radiotherapy, cancer, white blood cell (WBC) count, platelet count, sex, diabetes, smoking status, alcohol use, dyspnea, preoperative ventilator
use, chronic obstructive pulmonary disease (COPD), pneumonia, emergency operation, and creatinine level

an opportunity to lower the incidence of SSIs at the com-
munity hospital level. An evaluation of performance during
visits to one site showed that the use of local anesthesia, in
contrast to general/spinal anesthesia, was associated with a
lower incidence of SSIs. Based on this observation, we
developed a hypothesis that the use of local anesthesia was
associated with a decrease in the incidence of SSIs.
A robust outcome metric tool, the ACS NSQIP system,
was used to test the hypothesis. The ACS NSQIP is one of
the most comprehensive outcome measurement systems
available. This system provides 30-day follow-up, a precise
definition of SSI, and regular audit of results. The large
national ACS NSQIP database allowed for detailed eval-
uation of more than 100,000 cases that could have been
performed under either local or general/spinal anesthesia,
depending on the judgment of the surgeon. The initial
unadjusted comparison of the two groups showed sub-
stantial differences in patient characteristics. As would be
expected, there is a substantial bias involved in deciding
which patient receives which type of anesthesia. To control
for this bias, propensity matching was used. This statistical
methodology provides a mechanism to construct closely
matched groups of patients, differing only in the dependent
variable of interest. In this study, the dependent variable of
interest was the use of local anesthesia. Our results,
regardless of whether propensity-based, showed that
patients undergoing simple procedures under local anes-
thesia experienced fewer SSIs than patients undergoing the
same procedure under general anesthesia.
Our observation that surgery done under local anesthesia
was associated with a lower incidence of SSIs is important
because it suggests a potential new strategy for reducing
the incidence of SSIs: reduce the level of inhalational
anesthesia needed by infiltrating the wound with a long-
acting local anesthetic agent. Inhalational anesthetic agents
are known to have immunosuppressive effects [22, 23],
which include inhibition of proinflammatory cytokine
release [24, 25]. It follows that reducing the level of
inhalational agent needed would result in less immuno-
suppression and potentially fewer SSIs. In addition, it has
been observed that the use of local anesthesia is associated
with increased oxygen tension and tissue perfusion at the
wound site, each of which would be expected to diminish
the incidence of SSIs [26]. An equally if not more

123
2600 World J Surg (2011) 35:25962602

Table 4 Clinical characteristics of propensity-matched groups

Variable Local anesthesia Nonlocal P
(n = 1343) anesthesia
(n = 1343)

Male sex 518 (38.6%) 533 (39.7%) 0.304

Smoker 206 (15.3%) 191 (14.2%) 0.233
Alcohol 28 (2.1%) 36 (2.7%) 0.188
Dyspnea 61 (4.5%) 54 (4.0%) 0.759
Nonindependent 12 (0.9%) 17 (1.3%) 0.288
functional status
Ventilator 0 0
COPD 22 (1.6%) 22 (1.6%) 0.56 Fig. 1 Surgical site infection (SSI) rates in unadjusted and
CHF 6 (0.4%) 4 (0.3%) 0.377 propensity-matched local anesthesia and nonlocal anesthesia outpa-
Previous cardiac 61 (4.5%) 52 (3.9%) 0.221 tient cases. In the unmatched analysis, patients given local anesthesia
surgery had a significantly lower incidence of SSIs than patients given
nonlocal anesthesia (0.7 vs. 1.4%, P = 0.013). Similarly, after
Diabetes 127 (9.5%) 112 (8.3%) 0.171
propensity matching, the incidence of SSIs for patients given local
Pneumonia 0 (0%) 1 (0.1%) 0.5 anesthesia was significantly lower than for that of patients given
Cancer 16 (1.2%) 11 (8.8%) 0.22 nonlocal anesthesia (0.8 vs. 1.4%, P = 0.043)
Operating time (min) 45.7 45.1 0.565
Work RVU 7.2 7.2 0.866 et al. [23], Grimmond and Brownridge [24], and Sakuragi
et al. [25], all of whom demonstrated bactericidal activity
These data represent the characteristics of patients undergoing general
and vascular surgery operations within the NSQIP. Note that no of bupivacaine against methicillin-resistant S. aureus
significant differences existed between the two groups following (MRSA).
propensity matching Our study has several limitations. It was not a pro-
spective randomized trial, and we acknowledge that even
propensity matching might not identify factors accounting
important aspect of our findings involves the possible for a selection bias in the type of anesthesia used. In this
effect of the local anesthetic agent itself on the prolifera- analysis, we also grouped procedures using epidural or
tion of bacteria in the operative site, nicely summarized spinal anesthesia with those cases using general inhala-
elsewhere [27]. Lidocaine exhibits a dose-dependent inhi- tional anesthesia to isolate the possible beneficial effect of
bition of growth of both gram-positive and gram-negative local anesthesia administered on its own. There is growing
organisms such as Staphylococcus aureus and Pseudomo- evidence that epidural anesthesia also has fewer immuno-
nas aeruginosa, respectively, both of which account for suppressive effects than general inhalational anesthesia
[30% of the SSI and nosocomial infections in general [1]. [2731]. Thus, rather than minimizing differences between
Noda et al., through quantitative analyses, demonstrated the local and nonlocal groups, exclusion of these cases
the bactericidal activity of both lidocaine and bupivacaine from the nonlocal cohort would likely have accentuated
against both S. aureus and P. aeruginosa. They noted that differences between nonlocal and local groups in terms of
bupivacaine had greater antibacterial activity than lido- the SSI rate. Also, we were not able to identify patients in
caine [22]. These results were suppplemented by Johnson the local anesthesia cohort who were given a local

Table 5 Comparison of the

Surgical procedures Local anesthesia Nonlocal anesthesia
five most common procedures
(no.) (no.)
in propensity-matched groups
Repair of initial inguinal hernia (age [ 5 years), reducible 303 224
Excision of cyst, fibroadenoma, or other benign or malignant 265 163
tumor aberrant breast tissue, duct lesion, or nipple lesion
(male or female; one or more lesions)
Excision of breast lesion identified by preoperative 235 152
placement of radiologic marker (open, single lesion)
Repair of umbilical hernia (age [ 5 years), reducible 52 95
Partial mastectomy (e.g., lumpectomy, tylectomy, 64 67
quadrantectomy, segmentectomy)

123
World J Surg (2011) 35:25962602
anesthetic agent plus epinephrine. Adding epinephrine to
the local agent might offset the beneficial effect of local
anesthesia on SSIs by causing vasoconstriction. Therefore,
if we had been able to study only local cases in which
epinephrine was not used we would likely have seen more,
rather than less, pronounced differences in the SSI rate
between nonlocal and local cases. In addition, although we
adjusted for work RVU and operating time during pro-
pensity matching, it was not possible to account for all
operative variables in our analysis, such as hernia size.
Prospective studies are thus needed to fully account for
operative complexity. Finally, we noted that there was a
relatively low incidence of SSIs in each of the matched
groups. Given the small number of events in each group, it
is possible for type I errors to occur. Future studies should
evaluate groups with a larger number of events to confirm
the findings of this work.
We recognize that surgeons long accustomed to using
general anesthesia for the types of cases we identified
might not readily begin to use local anesthesia for these
cases on the basis of our findings. We suggest, however,
that routine wound infiltration of a long-acting local
anesthetic as an adjunct to general anesthesia might reduce
the level of inhalational agent needed and at the same time
provide an anti-infection agent in the wound itself, the
combination of which would lower the incidence of all
types of SSI, possibly even for larger cases than the types
we describe. The routine use of a long-acting local anes-
thetic might also reduce the need for postoperative opiates
delivered through a patient-controlled anesthesia device,
the use of which has been associated with a higher inci-
dence of SSIs [32].
Conclusions
Our findings support the development of a prospective
randomized controlled trial in which general surgical cases
would be divided into two groups: (1) patients given a
general inhalational agent plus wound infiltration with
placebo and (2) patients given general anesthesia plus
wound infiltration with a long-acting local anesthetic
without epinephrine. The endpoint would be the incidence
of SSIs. If the results of such a study favored routine use of
local anesthesia, we will have identified a safe and easy
approach to reducing the incidence of SSIs.
References
1. Horan TC, Gaynes RP, Martone WJ et al (1992) CDC definitions
of nosocomial surgical site infections, 1992: a modification of
2601
CDC definitions of surgical wound infections. Am J Infect
Control 20:271274
2. Emori TG, Gaynes RP (1993) An overview of nosocomial
infections, including the role of the microbiology laboratory. Clin
Microbiol Rev 6:428442
3. Khuri SF, Henderson WG, DePalma RG et al (2005) Determi-
nants of long-term survival after major surgery and the adverse
effect of postoperative complications. Ann Surg 242:326341
Discussion 341343
4. Dimick JB, Chen SL, Taheri PA et al (2004) Hospital costs
associated with surgical complications: a report from the private-
sector National Surgical Quality Improvement Program. J Am
Coll Surg 199:531537
5. Mahmoud NN, Turpin RS, Yang G et al (2009) Impact of surgical
site infections on length of stay and costs in selected colorectal
procedures. Surg Infect (Larchmt) 10:539544
6. Campbell DA Jr, Henderson WG, Englesbe MJ et al (2008)
Surgical site infection prevention: the importance of operative
duration and blood transfusionresults of the first American
College of Surgeons-National Surgical Quality Improvement
Program best practices initiative. J Am Coll Surg 207:810
820
7. Steinberg JP, Braun BI, Hellinger WC et al (2009) Timing of
antimicrobial prophylaxis and the risk of surgical site infections:
results from the trial to reduce antimicrobial prophylaxis errors.
Ann Surg 250:1016
8. Weber WP, Marti WR, Zwahlen M et al (2008) The timing of
surgical antimicrobial prophylaxis. Ann Surg 247:918926
9. Mangram AJ, Horan TC, Pearson ML et al (1999) Guideline for
prevention of surgical site infection, 1999: hospital infection
control practices advisory committee. Infect Control Hosp Epi-
demiol 20:250278 Quiz 279280
10. Stone HH, Hooper CA, Kolb LD et al (1976) Antibiotic pro-
phylaxis in gastric, biliary and colonic surgery. Ann Surg
184:443452
11. Bratzler DW, Houck PM, Richards C et al (2005) Use of anti-
microbial prophylaxis for major surgery: baseline results from the
National Surgical Infection Prevention Project. Arch Surg
140:174182
12. Classen DC, Evans RS, Pestotnik SL et al (1992) The timing of
prophylactic administration of antibiotics and the risk of surgical-
wound infection. N Engl J Med 326:281286
13. Van den Berghe G, Wouters P, Weekers F et al (2001) Intensive
insulin therapy in the critically ill patients. N Engl J Med
345:13591367
14. Van den Berghe G, Wouters PJ, Bouillon R et al (2003) Outcome
benefit of intensive insulin therapy in the critically ill: insulin
dose versus glycemic control. Crit Care Med 31:359366
15. Kurz A, Sessler DI, Lenhardt R (1996) Perioperative normo-
thermia to reduce the incidence of surgical-wound infection and
shorten hospitalization: study of wound infection and temperature
group. N Engl J Med 334:12091215
16. Melling AC, Ali B, Scott EM et al (2001) Effects of preoperative
warming on the incidence of wound infection after clean surgery:
a randomised controlled trial. Lancet 358:876880
17. Maragakis LL, Cosgrove SE, Martinez EA et al (2009) Intraop-
erative fraction of inspired oxygen is a modifiable risk factor for
surgical site infection after spinal surgery. Anesthesiology
110:556562
18. Greif R, Akca O, Horn EP et al (2000) Supplemental perioper-
ative oxygen to reduce the incidence of surgical-wound infection:
Outcomes Research Group. N Engl J Med 342:161167
19. Mazuski JE, Sawyer RG, Nathens AB et al (2002) The surgical
infection society guidelines on antimicrobial therapy for intra-
abdominal infections: an executive summary. Surg Infect
(Larchmt) 3:161173
123
2602
20. Khuri SF, Daley J, Henderson W et al (1998) The Department of
Veterans Affairs NSQIP: the first national, validated, outcome-
based, risk-adjusted, and peer-controlled program for the mea-
surement and enhancement of the quality of surgical care
National VA Surgical Quality Improvement Program. Ann Surg
228:491507
21. Parsons LS (2001) Reducing bias in a propensity score matched-
pair sample using greedy matching techniques. In: SUGI 26
SUGI Global Forum SAS Proceedings. Long Beach, CA,
pp 214226
22. Noda H, Saionji K, Miyazaki T (1990) Antibacterial activity of
local anesthetics. Masui 39:9941001
23. Johnson SM, Saint John BE, Dine AP (2008) Local anesthetics as
antimicrobial agents: a review. Surg Infect (Larchmt) 9:205213
24. Grimmond TR, Brownridge P (1986) Antimicrobial activity of
bupivacaine and pethidine. Anaesth Intensiv Care 14:418420
25. Sakuragi T, Ishino H, Dan K (1996) Bactericidal activity of
clinically used local anesthetics on Staphylococcus aureus. Reg
Anesth 21:239242
26. Akca O, Melischek M, Scheck T et al (1999) Postoperative pain
and subcutaneous oxygen tension. Lancet 354:4142
123
World J Surg (2011) 35:25962602
27. Moraca RJ, Sheldon DG, Thirlby RC (2003) The role of epidural
anesthesia and analgesia in surgical practice. Ann Surg
238:663673
28. Hole A, Unsgaard G, Breivik H (1982) Monocyte functions are
depressed during and after surgery under general anaesthesia but
not under epidural anaesthesia. Acta Anaesthesiol Scand
26:301307
29. Porter JS, Bonello E, Reynolds F (1996) The effect of epidural
opioids on maternal oxygenation during labour and delivery.
Anaesthesia 51:899903
30. Yeager MP, Glass DD, Neff RK et al (1987) Epidural anesthesia
and analgesia in high-risk surgical patients. Anesthesiology
66:729736
31. Tuman KJ, McCarthy RJ, March RJ et al (1991) Effects of epi-
dural anesthesia and analgesia on coagulation and outcome after
major vascular surgery. Anesth Analg 73:696704
32. Horn SD, Wright HL, Couperus JJ et al (2002) Association
between patient-controlled analgesia pump use and postoperative
surgical site infection in intestinal surgery patients. Surg Infect
(Larchmt) 3:109118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.