You are on page 1of 6

CE: A.G.

; ANNSURG-D-16-01402; Total nos of Pages: 6;



Randomized Controlled Trial of Two Alcohol-based Preparations

for Surgical Site Antisepsis in Colorectal Surgery
Robyn B. Broach, PhD,  Emily C. Paulson, MD, MSCE,y Charles Scott, PhD,z and Najjia N. Mahmoud, MD 

Although skin flora represents the majority of pathogens in

Objective: To compare 2 alcohol-based, dual-action skin preparations for
clean cases, in colorectal surgery, both bowel flora and skin con-
surgical site infection (SSI) prevention in elective colorectal surgery.
taminants play a role. The risk for SSI originating from these bacteria
Background: Colorectal surgery is associated with the highest SSI rate
can be reduced by skin preparation in the operating room using an
among elective surgical procedures. Although evidence indicates that alco-
antiseptic agent containing alcohol, chlorhexidine, or solubilized
hol-based skin preparations are superior in SSI prevention, it is not clear if
iodine used alone or in combination forms. A randomized, controlled
different alcohol-based preparations are equivalent in clean-contaminated
trail by Darouiche et al6 clearly identified the superiority of chlor-
colorectal procedures.
ehexidine-alcohol compared with povidone-iodine alone for the
Methods: We performed a blinded, randomized, noninferiority trial compar-
prevention of SSIs in clean-contaminated cases.
ing iodine povacrylex-alcohol (IPA) and chlorhexidine-alcohol for elective,
Less clear, however, is which, if any, alcohol-based skin
clean-contaminated colorectal surgery. The primary outcome was the pres-
preparation is most effective. In the study by Swenson et al,7 both
ence or absence of SSI, defined as superficial or deep SSI, within 30 days
IPA (iodine povacrylex and isopropyl alcohol) and povidone iodine
postdischarge. A 6.6% noninferiority margin was chosen.
with isopropyl alcohol sequentially applied were significantly more
Results: Between January 2011 and January 2015, 802 patients were random-
effective than chlorhexidine-alcohol in preventing SSI in a hetero-
ized with 788 patients included in the intent to treat analysis (396 IPA and 392
geneous cohort of general surgery patients. However, a study recently
chlorhexidine-alcohol). The difference in overall SSI rate between IPA
published by Tuuli et al8 demonstrated that chlorhexidine-alcohol
(18.7%) and chlorhexidine-alcohol (15.9%) was 2.8% (P ¼ 0.30). The upper
resulted in significantly lower risk of SSI than iodine-alcohol (with-
bound of the 2.5% confidence interval of this difference was 8.9%, which is
out povacrylex) after Caesarian sections. These, and other conflicting
greater than the prespecified noninferiority margin of 6.6%. Other endpoints,
reports in differing patient populations, make it difficult to conclude
including individual SSI types, time to SSI diagnosis, and length of stay were
which alcohol-based skin preparations may be most effective in
not different between the 2 arms.
colorectal surgery patients, who have the highest risk of SSI.
Conclusions: In patients undergoing elective, clean contaminated colorectal
The opportunity for improvement in SSI rates in colorectal
surgery, the use of IPA failed to meet criterion for noninferiority for overall
surgery is greater than all other surgical procedures because of both
SSI prevention compared with chlorhexidine-alcohol. Photodocumentation of
high surgical volume nationally and high infection rates. Concom-
wounds and rigorous tracking of outcomes up to 30 days postdischarge
itant reduction of cost, length of stay, and morbidity represents a
contributed to high fidelity to current standard SSI descriptions and wound
significant step forward. Demonstration of comparative effectiveness
between agents is essential. We designed this randomized, blinded,
Keywords: colorectal surgery, Skin Antisepsis, surgical site infection prospective trial to test the hypothesis that preoperative skin prep-
aration in colorectal surgery with IPA is noninferior to skin prep-
(Ann Surg 2017;xx:xxx–xxx) aration with chlorhexidine-alcohol. The use of photodocumentation,
exclusion of noncolorectal procedures, blinded SSI analysis, and 30
S urgical site infection (SSI) is the second most common of the
hospital acquired infections (HAI), representing 22% of all HAIs
and accounting for 374,000 infections among hospitalized patients
day outpatient outcomes are unique and overcome some of the
obstacles encountered by prior trials and retrospective reviews.
yearly in the United States.1,2 Colorectal surgery is associated with
the highest SSI rate among elective surgical procedures, with con- METHODS
servative estimates ranging from 5% to 25%.3,4 It is recognized that
SSI is associated with prolonged length of stay and increased Study Design
morbidity, mortality, and cost.1,5 A randomized, blinded, prospective trial designed to evaluate
whether skin preparation with IPA is noninferior to that with
chlorhexidine-alcohol in colorectal surgery was performed in Uni-
versity of Pennsylvania health system between January 2011 and
From the Department of Surgery, Hospital of the University of Pennsylvania, 3400 January 2015 ( NCT01233050; full protocol avail-
Spruce Street, 4 Silverstein, Philadelphia, PA; yDepartment of General able upon request). The study was approved by the Institutional
Surgery, VA Medical Center, Philadelphia, PA; and zCBS Squared, Inc., Review Board of the University of Pennsylvania and informed
CBS Squared, Inc., Flourtown, PA.
Disclosure: This study was funded by 3 M, Inc., ( consent was obtained from all patients before enrollment (Fig. 1).
NCT01233050). The authors declare no conflict of interests. The study was stopped in January 2015 after prespecified accrual
Supplemental digital content is available for this article. Direct URL citations numbers were met.
appear in the printed text and are provided in the HTML and PDF versions of
this article on the journal’s Web site (
Reprints: Najjia N. Mahmoud, MD, Department of Surgery, Hospital of the Patients
University of Pennsylvania, 3400 Spruce St., 4th Floor Silverstein, Philadel- Eligible patients included those 18 years of age or above
phia, PA 19104. E-mail: undergoing an elective clean-contaminated colorectal procedure.
Copyright ß 2017 Wolters Kluwer Health, Inc. All rights reserved.
ISSN: 0003-4932/16/XXXX-0001 Exclusion criteria were antibiotic use within 5 days before surgery,
DOI: 10.1097/SLA.0000000000002189 infected or dirty wound classification, preoperative plan to leave the

Annals of Surgery  Volume XX, Number XX, Month 2017 | 1

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: A.G.; ANNSURG-D-16-01402; Total nos of Pages: 6;

Broach et al. Annals of Surgery  Volume XX, Number XX, Month 2017

FIGURE 1. Screening, randomization,

and follow up of study participants.

incision open, ongoing radiation or chemotherapy, history of lapa- representation in the 2 treatment arms. The study statistician created
rotomy within 60 days, current abdominal wall infection, and known a computer generated list of random assignments by hospital block.
allergy to chlorhexidine gluconate or iodine. Patients were excluded The clinical trial coordinator delivered sequentially numbered,
if they were participating in any concomitant preoperative antibiotic randomized opaque envelopes to the sites in a block. The study
or skin antisepsis trial. Women who were pregnant or breast feeding coordinator opened the sequentially numbered, opaque envelope
were excluded. Patients were recruited into the study by attending containing the randomization assignment postinduction of anesthesia
surgeons or advanced practitioners in the preoperative clinic setting when in the operating room.
simultaneously with discussions about, and consent for, surgery. Training in use of each skin preparation was conducted by the
study coordinator and the operating room staff compliance and
Interventions quality officers. Every surgeon, resident, fellow, and nurse involved
The patients were randomized in a 1:1 ratio to undergo skin in a study case was required to attend or view an online training
preparation with either iodine povacrylex-alcohol, a 26 mL single- module detailing the appropriate use of each skin preparation. The
use applicator containing iodine povacrylex [0.7% available iodine]/ study coordinator or attending surgeon was present for every prep-
74% isopropyl alcohol (w/w) or chlorhexidine-alcohol, a 26 mL aration application to ensure quality control. A single applicator was
single-use applicator containing 2% chlorhexidine gluconate (w/v) used for most patients. Those who were morbidly obese required a
and 70% isopropyl alcohol (v/v). Both skin preps contain isopropyl second applicator. The IPA arm involved single pass application
alcohol that has strong immediate antimicrobial effects but no whereas the chlorhexidine-alcohol arm involved several passes of the
appreciable residual activity on skin. Each contains a second com- applicator in a circular motion. Both were allowed to dry for 3
ponent that offers residual activity up to 48 hours. minutes before draping. All preparation sticks were used according
Block randomization, based on hospital site, was used to to manufacturer’s instructions by attending surgeons, residents or
insure that each site would have statistically proportionate fellows who underwent live and video training.

2 | ß 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: A.G.; ANNSURG-D-16-01402; Total nos of Pages: 6;

Annals of Surgery  Volume XX, Number XX, Month 2017 Skin Antisepsis in Colorectal Surgery

Outcomes 30  5 days after hospital discharge. The questionnaires were

The primary outcome was the occurrence of any surgical site acquired during postoperative visits or through phone interviews
infection within 30  5 days postdischarge. This included the occur- at 15  5 days and at 30  5 days postdischarge. Follow up was
rence of superficial skin infection or deep skin infection. Centers for discontinued if a wound infection was confirmed by culture or by
Disease Control and Prevention (CDC) definitions were used to CDC defined clinical diagnosis. In addition, if patients were seen in
classify superficial and deep skin infections.9–11 Although the CDC the office at POD 22–25 and did not have an SSI, and had a well
does not recognize cellulitis as a postoperative wound complication, healed wound, they were discharged from the study and asked to
we believe that it is a clinically meaningful outcome. Therefore, follow up if problems developed. All photographs and other data
cellulitis was included as a secondary outcome measurement. Other were reviewed, in blinded fashion, twice by the PI, once within 2
secondary endpoints included organ space infection, infectious months of the assessment, and a second time at the end of the study to
pathogen, length of stay, and time to SSI. make a final determination regarding SSI classification.

Clinical Assessment Statistical Analysis

Preoperative evaluation included a medical history, physical The 30-day postoperative SSI rate (superficial SSI þ deep
exam, and routine laboratory testing. Perioperative information SSI) using chlorhexidine-alcohol was assumed to be 20% for the
including prophylactic antibiotics, vital signs, and other relevant power calculation.12 Based on the Darouiche study, we assigned a
information was obtained from anesthesia and nursing records. The maximal benefit to chlorhexidine alcohol, compared with povidone
surgical site was assessed using a standard assessment form and iodine, of 41% (relative risk 0.59); that is, chlorhexidine alcohol
photographed by a blinded assessor daily starting from postoperative reduces the rate of SSI compare with povidone iodine by 41%. FDA
day 3 to discharge, including weekend days (see assessment form guidance on noninferiority studies recommends a maximal non-
document, Supplemental Digital Content 1, inferiority margin up to 50% of the maximal benefit of one of the
SLA/B194). Vital signs, laboratory values, relevant postoperative treatment arms (chlorhexidine-alcohol); in this case, an SSI rate of
events, and wound culture data, if available, were also recorded by 27% (ie, a maximum noninferiority margin of 7%). We postulated
the blinded assessor. During postoperative clinic visits, participants that iodine-alcohol would be noninferior to chlorhexidine-alcohol in
completed a questionnaire to help identify possible wound compli- preventing SSI, specifying a noninferiority margin of 6.6%, within
cations diagnosed and treated at another facility (see questionnaire the 50% of maximal benefit margin. A sample size of 400 subjects
document, Supplemental Digital Content 2, per treatment group would provide 90% statistical power to detect a
SLA/B194). Photodocumentation of the incision was also obtained noninferiority margin of 6.6% using a one-sided a of 0.025. The
by a blinded assessor (Fig. 2). Patients were monitored for up to patients were included for analysis in a modified intent to treat

FIGURE 2. Photographic examples from

the study illustrating the classifications.
A, Healthy incision; (B) cellulitis; (c)
superficial SSI; and (D) deep SSI.

ß 2017 Wolters Kluwer Health, Inc. All rights reserved. | 3

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: A.G.; ANNSURG-D-16-01402; Total nos of Pages: 6;

Broach et al. Annals of Surgery  Volume XX, Number XX, Month 2017

TABLE 1. Baseline Characteristics of the Patients and Procedures (mITT)

Iodine Povacrylex Alcohol ChlorhexidineAlcohol
Characteristics (n ¼ 396) (n ¼ 392) P
Sex (male) 48.7% (193) 48.5% (190) 0.94c
Age (years) 56.8  15.8 57.0  16.7 0.61F
BMI 28.1  5.8 27.9  5.5 0.61F
Diabetes (yes) 12.2% (48) 14.8% (57) 0.27c
Neoplasia 44.5% (175) 47.6% (185) 0.40c
Inflammatory bowel disease 27.3% (108) 25.0% (98) 0.46c
Diverticulitis 12.4% (49) 12.5% (49) 0.96c
Benign 16.9% (67) 14.8% (58) 0.41c
Medical therapy
Steroid use 9.8% (38) 10.6% (41) 0.70c
Nonsteroidal immunosuppression 17.5% (69) 16.4% (64) 0.67c
Prior abdominal/pelvic radiation 11.8% (46) 15.8% (61) 0.11c
Prior abdominal surgery 48.2% (190) 51.5% (201) 0.35c
Duration of procedure (minutes) 179.69  77.71 181.04  76.71 0.81F
Surgery technique (laparoscopic) 52.8% (209) 46.7% (183) 0.09c
Body temperature 96.7  1.5 96.9  1.4 0.19F
Length of incision (cm) 13.79  6.88 14.14  6.66 0.50F
Blood transfusion 9.8% (39) 7.6% (30) 0.28c
Glucose level 100.0  38.4 102.8  42.0 0.33F
Creation or reversal of ostomy 29.8% (118) 31.9% (125) 0.53c
Hospital location 0.94c
A 62.4% (247) 61.2% (240)
B 26.5% (105) 27.6% (108)
C 11.1% (44) 11.2% (44)
Prophylactic antibiotics
SCIP compliant 90.9% (360) 92.1% (361) 0.55c

c indicates Chi-square test; F, F test; mITT, modified intent to treat; SCIP, SCIP, Surgical Care Improvement Project.

(mITT) manner. This sample was defined as those subjects who were The baseline patient characteristics by treatment group are
randomized and underwent a clean-contaminated surgery and had no displayed in Table 1. There were no significant differences among the
reoperation before SSI within 30 days postdischarge. The differences patients in each group with regard to demographics, diagnoses,
between patient factors between the 2 treatment arms were evaluated surgical techniques, preoperative medical therapies, or perioperative
using the F test for continuous variables and x2 test for categorical antibiotics. The types of operation were also comparable (data not
variables. shown). Of note, all patients received antibiotics within an hour of
To examine the primary outcome, the overall rate of wound incision and almost all received antibiotics consistent with the
infection was compared by treatment arm using the Fisher exact test. Surgical Care Improvement Project guidelines as specified in
Secondary outcomes, including the rate of each individual type of the protocol.
wound infection and organ space infection were evaluated separately The overall rate of SSI (superficial and deep SSI) was 17.3%
between subjects in each treatment arm. Length of stay was com- and did not differ significantly between the treatment arms (Table 2).
pared using an F test. Time from date of surgery to date of wound The rate of overall SSI was 18.7% in the IPA arm and 15.9% in the
infection was graphed using the product-limit method with subjects chlorhexidine-alcohol group (P ¼ 0.30). The difference in SSI rate
censored at the date of follow up if no SSI was observed. Differences between chlorhexidine-alcohol and IPA was 2.8%. The upper bound
between groups were evaluated using the log-rank test. All P values of the 2.5% confidence interval was 8.9%, which is greater than the
were 2-sided. The organisms identified in wound cultures were prespecified noninferiority margin of 6.6%. When examined indi-
compared descriptively. As not all subjects with infection were vidually, there were no differences in the rate of superficial SSI or
cultured and some subjects had more than one microorganism deep SSI or cellulitis between the 2 treatment groups. Cellulitis was
identified, no statistical test was performed. diagnosed in 4.8% of the IPA patients and in 3.6% of the chlorhex-
idine-alcohol patients.
Secondary endpoints including organ space infection, length
RESULTS of stay, and time to diagnosis of wound infection were analyzed.
A total of 802 subjects were randomized (402 to IPA and 400 Organ space infections were identified in 4.0% of IPA patients and
to chlorhexidine-alcohol). Ten subjects were excluded because of 5.1% of chlorhexidine-alcohol patients (P ¼ 0.50). Length of stay
reoperation within 30 days (4 IPA and 6 chlorhexidine-alcohol); 4 was identical between the 2 treatment arms. The average length of
were excluded because the case was aborted and no resection was stay for patients in the IPA arm was 6.8 days (s.d. 3.7), compared with
done. In total, 788 patients (396 patients in the IPA arm and 392 in the 7.0 (s.d. 3.9) days for patients in the chlorhexidine-alcohol arm (P ¼
chlorhexidine-alcohol arm) were included in the mITT analysis 0.45).
(Fig. 1). Eighteen patients in the IPA arm and 12 patients in the Time to infection is displayed graphically in Figure 3. The
chlorhexidine-alcohol arm did not complete the 15 and/or 30 day average time from surgery to diagnosis was 8.7 days (s.e 0.6 days) in
follow up. They were included, however, in the mITT cohort for the IPA arm and 8.0 days (s.e. 0.7 days) in the chlorhexidine-alcohol
analysis based on the information available at time of latest contact. arm (P ¼ 0.21). The result is notable for the wide range of times in

4 | ß 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: A.G.; ANNSURG-D-16-01402; Total nos of Pages: 6;

Annals of Surgery  Volume XX, Number XX, Month 2017 Skin Antisepsis in Colorectal Surgery

TABLE 2. SSI Rates by Treatment Arm (mITT)

Iodine povacrylexalcohol Chlorhexidinealcohol Difference
(n ¼ 396) (n ¼ 392) P (95% CI)
Overall SSI (superficial þ deep) 18.7% (74) 15.9% (62) 0.301 2.8 (–3.2,8.9)
Cellulitis þ overall SSI 23.5% (93) 19.4% (76) 0.193 4.1 (–2.5,10.6)
Superficial SSI 11.6% (46) 10.2% (40) 0.569 1.4 (–3.6,6.4)
Deep SSI 7.1% (28) 5.6% (22) 0.466 1.5 (2.5,5.3)
Cellulitis 4.8% (19) 3.6% (14) 0.478 1.2 (–2.0,4.4)
Organ space 4.0% (16) 5.1% (20) 0.500 –1.1 (–2.3,4.4)
CI indicates confidence interval; mITT, modified intent to treat; SSI, surgical site infection.

both arms; there were infections diagnosed as soon as 3 days post- prospectively, with rigorous documentation procedures in place,
operatively, and as long as 37 days after surgery. the use of IPA compared with chlorhexidine for patients undergoing
Finally, we examined the bacteria identified in the cultured clean-contaminated elective colorectal fails to meet the criterion for
wounds. Overall, only 69 wounds were cultured. In 59 wounds, 1 noninferiority for our primary outcome, overall SSI.
organism was identified; in 9 wounds, 3 organisms were isolated; Two previous prospective trials have examined the relation-
and, in 1 wound, 3 organisms were found. There was no difference in ship between skin preparation and SSI in surgical patients and have
the types of bacteria between the treatment arms (see table, Supple- informed this study to a large degree. Findings from these trials have
mental Digital Content 3, In both been seminal in driving the national change to alcohol-based skin
groups, Escherichia coli was the most frequent organism isolated, preps for general surgery patients. A study by Swenson et al7
accounting for almost 33% of the organisms identified. compared 3 different dual action skin preparations, all employing
There were no adverse events or serious adverse events isopropyl alcohol, in general surgery (including colorectal) patients
attributed to either skin preparation during the trial. Complications, with SSI as primary outcome. The results demonstrated that both IPA
other than SSI, not related directly to the skin preparation were not and povidone iodine with isopropyl alcohol applied sequentially
different between the 2 groups (data available upon request). were significantly more effective than chlorhexidine-alcohol. This
was followed by a trial from Darouiche et al6 in 2010 that compared
DISCUSSION chlorhexidine-alcohol with single component 10% povidone iodine
Optimal preoperative skin preparation is an important aspect in general surgery patients. This trial demonstrated a 41% reduction
of SSI prevention. Organizations focused on outcome improvement in SSI in the chlorhexidine-alcohol arm. Although elective colorectal
such as the National Quality Forum now recommend use of alcohol procedures comprised a percentage of patients in these trials, there
based skin preparations for surgical procedures but do not specify were a large number of low risks (cholecystectomy) and clean
which to use.13,14 There are numerous studies examining the efficacy procedures (hernia repair) included. The heterogeneity of the groups
of preoperative skin preps in both clean and clean-contaminated made subset analysis of the highest risk group difficult.
procedures, but none prospectively examining only colorectal pro- To better evaluate the comparative efficacy of 2 alcohol based
cedures—those with the highest risk. With relatively a few excep- preps, a study dedicated to those cases at highest risk was required.
tions, most recent studies examining SSI in abdominal surgery are The inclusion of cellulitis as an outcome in the analysis was
retrospective in nature or rely on administrative data from national important. Although it does not meet CDC criteria for SSI, it is a
and state databases. Our group found that when examined clinically meaningful complication that frequently results in use of

FIGURE 3. Time to wound infection, by

treatment arm.

ß 2017 Wolters Kluwer Health, Inc. All rights reserved. | 5

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.
CE: A.G.; ANNSURG-D-16-01402; Total nos of Pages: 6;

Broach et al. Annals of Surgery  Volume XX, Number XX, Month 2017

antibiotics. We defined cellulitis as ‘‘erythema, warmth, and pain more objective means of diagnosis compared with review of clinical
requiring antibiotics’’ and only scored erythema as cellulitis if documentation only.
antibiotics were prescribed and the patient had all 3 symptoms. In conclusion, this study statistically fails to support the
Photodocumentation was used in addition to clinical criteria to noninferiority of iodine povacrylex-alcohol to chlorhexidine-alcohol
document and diagnose SSI. Distinctions between cellulitis, super- skin preparation in regard to superficial or deep SSI between patients
ficial, and deep SSI were ultimately aided by visual cues provided by undergoing elective colorectal surgery. Rigorous photodocumenta-
high resolution photographs in addition to knowledge of culture and tion aided standardized assessments, and adherence to the protocol
treatment data (Fig. 2). Blinded review of all study photos allowed was high. It is clear that alcohol based skin preparations are an
for standardization of diagnosis across subjects and sites and important part of SSI prevention in elective colorectal surgery. It is
represents a unique study feature that ensured fidelity to CDC also clear that the absolute differences in SSI prevention between
definitions. Photodocumentation made the most vague and problem- these preparations are not large. However, in light of these findings,
atic of the CDC definitions—‘‘diagnosis by attending surgeon’’ and until further studies are available, chlorhexidine-alcohol is
unnecessary and unused.13 favored to prevent superficial and deep SSI in elective clean con-
The rate of SSI in this study was comparable with other taminated colorectal cases.
prospective datasets but higher when compared with that from
retrospective studies. Although not powered to examine differences
in colorectal procedures, the Darouiche group found rates of SSI ACKNOWLEDGMENTS
ranging from 15.1 to 22.0%.6 Their protocol included 30 day out- The authors would like to thank clinical investigators Robert
comes assessed by weekly telephone encounters to determine wound D Fry, Skandan S Shanmugan, Joshua IS Bleier, Brian R Kann, and
problems. Swenson’s group did not specifically examine rates of Cary B Aarons for their contributions to this study.
wound infection in colorectal surgery but did a subset analysis of
clean contaminated procedures and found that the rate of SSI ranged REFERENCES
from 5.9% to 10.7% among the 3 test groups.7 A recently published 1. Dimick JB, Chen SL, Taheri PA, et al. Hospital costs associated with surgical
complications: a report from the private-sector National Surgical Quality
retrospective review of skin preparation data from the Michigan Improvement Program. J Am Coll Surg. 2004;199:531–537.
Surgical Quality Collaborative examining only colorectal surgical 2. Lissovoy G, Fraeman K, Teerlink JR, et al. Hospital costs for treatment of
cases and encompassing a variety of skin preparations showed a low acute heart failure: economic analysis of the REVIVE II study. Eur J Health
rate of overall SSI, ranging from 3.9% to 7.1%.13 This database relies Eco. 2009;11:185–193.
on inpatient data. It is exceedingly difficult to capture outpatient 3. Hedrick TL, Sawyer RG, Hennessy SA, et al. Can we define surgical site
clinical data using even specific administrative datasets because of infection accurately in colorectal surgery? Surg Infect. 2014;15:372–376.
nonuniformity in recording, varying medical records, and variability 4. Ju MH, Ko CY, Hall BL, et al. A comparison of 2 surgical site infection
monitoring systems. JAMA surgery. 2015;150:51–57.
among abstractors. By contrast, in this trial, the rate of superficial and 5. Anderson DJ, Kaye KS, Classen D, et al. Strategies to prevent surgical
deep SSI ranged from 15.9% to 18.7%. Data, however, were col- site infections in acute care hospitals. Infect Control Hosp Epidemiol.
lected in the office with patients returning within 30 days of 2008;29(Suppl 1):605–627.
discharge for the questionnaire, photographs, and direct examination. 6. Darouiche RO, Wall MJ, Itani KMF, et al. Chlorhexidine-alcohol versus
The data highlight the fact that most SSIs are detected after discharge povidone-iodine for surgical-site antisepsis. New Engl J Med. 2010;362:
and are therefore routinely missed by administrative databases. 18–26.
Length of stay averaged 6.8 days with average time to diagnosis 7. Swenson BR, Hedrick TL, Metzger R, et al. Effects of preoperative skin
preparation on postoperative wound infection rates: a prospective study of 3
of wound infection of 8.7 days. Most wound infections were diag- skin preparation protocols. Infect Control Hosp Epidemiol. 2009;30:964–971.
nosed in the outpatient population. The higher rate of SSI in this 8. Tuuli MG, Liu J, Stout MJ, et al. A randomized trial comparing skin antiseptic
study likely represents a higher rate of detection rather than a higher agents at cesarean delivery. New Engl J Med. 2016;374:647–655.
rate of infection. 9. Mangram AJ, Horan TC, Pearson ML, et al. Guideline for Prevention of
In contrast to other studies, a majority of the infectious agents Surgical Site Infection, 1999. Am J Infect Control. 1999;27:97–134.
identified by wound culture were gram negative bowel flora; with E. 10. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of
coli representing 33% of the isolates identified.15 This reflects the health care-associated infection and criteria for specific types of infections in
the acute care setting. Am J Infect Control. 2008;36:309–332.
fact that all of the procedures in this trial were large intestine 11. CDC/NHSN. CDC/NHSN Surveillance Definitions for Specific Types of
operations. Clearly, wound infections in these cases are more com- Infections. Available at: http://wwwcdcgov/nhsn/PDFs/pscManual/9pscSSI-
plex, can be polymicrobial, and are less influenced by skin flora than currentpdf 2013-2015.
clean procedures such as hernia repair. 12. Smith RL, Bohl JK, McElearney ST, et al. Wound infection after elective
This trial has several limitations. The generalizability of colorectal resection. Ann Surg. 2004;239:599–607.
findings may be limited by the fact that the trial was conducted 13. Kaoutzanis C, Kavanagh CM, Leichtle SW, et al. Chlorhexidine with iso-
within one health system in 3 hospitals by colorectal specialists. propyl alcohol versus iodine povacrylex with isopropyl alcohol and alcohol-
versus nonalcohol-based skin preparations: the incidence of and readmissions
Although the trial arms were racially diverse and balanced by gender, for surgical site infections after colorectal operations. Dis Colon Rectum.
most colorectal surgery in the United States is done by general 2015;58:588–596.
surgeons in a range of hospital settings and sizes. A major limitation 14. Talsma A, Chenoweth CE. Review of evidence for alcohol-based skin
of all studies of SSI is misdiagnosis based on the often subjective preparation agents. Infect Control Hosp Epidemiol. 2012;1059–1060.
nature of the data. The risk of misdiagnosis should have been equal 15. Hidron AI, Edwards JR, Patel J, et al. Antimicrobial-resistant pathogens
between the 2 cohorts given that diagnosis was made by a blinded associated with healthcare-associated infections: Annual Summary of Data
Reported to the National Healthcare Safety Network at the Centers for Disease
reviewer familiar with CDC definitions of SSI. Moreover, the use of Control and Prevention, 2006-2007. Infect Control Hosp Epidemiol.
photodocumentation coupled with clinical information provided a 2008;29:996–1011.

6 | ß 2017 Wolters Kluwer Health, Inc. All rights reserved.

Copyright © 2017 Wolters Kluwer Health, Inc. Unauthorized reproduction of this article is prohibited.