Multicenter Assessment of Antibiotic Prophylaxis Spectrum on Surgical Infections in Head and Neck Cancer Microvascular Reconstruction

Transcription

1 Original Research Head and Neck Surgery Multicenter Assessment of Antibiotic Prophylaxis Spectrum on Surgical Infections in Head and Neck Cancer Microvascular Reconstruction Otolaryngology Head and Neck Surgery 2018, Vol. 159(1) Ó American Academy of Otolaryngology Head and Neck Surgery Foundation 2018 Reprints and permission: sagepub.com/journalspermissions.nav DOI: / Michael P. Veve, PharmD 1,2, Joshua B. Greene, MD 2, Amy M. Williams, PhD 2, Susan L. Davis, PharmD 1,2, Nina Lu, MD 3, Yelizaveta Shnayder, MD 3, David X. Li 4, Salem I. Noureldine, MD 4, Jeremy D. Richmon, MD 4, Lawrence O. Lin 5, Matthew M. Hanasono, MD 5, Patrik Pipkorn, MD 6, Ryan S. Jackson, MD 6, Joshua D. Hornig, MD 7, Tyler Light 8, Mark K. Wax, MD 8, Yin Yiu, MD 9, James Bekeny, MD 9, Matthew Old, MD 9, David Hernandez, MD 10, Urjeet A. Patel, MD 10, and Tamer A. Ghanem, MD, PhD 1,2 Sponsorships or competing interests that may be relevant to content are disclosed at the end of this article. Abstract Objective. To characterize and identify risk factors for 30-day surgical site infections (SSIs) in patients with head and neck cancer who underwent microvascular reconstruction. Study Design. Cross-sectional study with nested case-control design. Setting. Nine American tertiary care centers. Subjects and Methods. Hospitalized patients were included if they underwent head and neck cancer microvascular reconstruction from January 2003 to March Cases were defined as patients who developed 30-day SSI; controls were patients without SSI at 30 days. Postoperative antibiotic prophylaxis (POABP) regimens were categorized by Gram-negative (GN) spectrum: no GN coverage, enteric GN coverage, and enteric with antipseudomonal GN coverage. All POABP regimens retained activity against anaerobes and Gram-positive bacteria. Thirty-day prevalence of and risk factors for SSI were evaluated. Results. A total of 1307 patients were included. Thirty-day SSI occurred in 189 (15%) patients; median time to SSI was 11.5 days (interquartile range, 7-17). Organisms were isolated in 59% of SSI; methicillin-resistant Staphylococcus aureus (6%) and Pseudomonas aeruginosa (9%) were uncommon. A total of 1003 (77%) patients had POABP data: no GN (17%), enteric GN (52%), and antipseudomonal GN (31%). Variables independently associated with 30-day SSI were as follows: female sex (adjusted odds ratio [aor], 1.6; 95% CI, ), no GN POABP (aor, 2.2; 95% CI, ), and surgical duration 11.8 hours (aor, 1.9; 95% CI, ). Longer POABP durations (6 days) or antipseudomonal POABP had no association with SSI. Conclusions. POABP without GN coverage was significantly associated with SSI and should be avoided. Antipseudomonal POABP or longer prophylaxis durations (6 days) were not protective against SSI. Antimicrobial stewardship interventions should be made to limit unnecessary antibiotic exposures, prevent the emergence of resistant organisms, and improve patient outcomes. 1 Wayne State University, Detroit, Michigan, USA 2 Henry Ford Health System, Detroit, Michigan, USA 3 Department of Otolaryngology, School of Medicine, University of Kansas, Kansas City, Kansas, USA 4 Johns Hopkins University, Baltimore, Maryland, USA 5 The University of Texas MD Anderson Cancer Center, Houston, Texas, USA 6 Washington University School of Medicine, St Louis, Missouri, USA 7 Medical University of South Carolina, Charleston, South Carolina, USA 8 Oregon Health Sciences University, Portland, Oregon, USA 9 The Ohio State University Wexner Medical Center, Columbus, Ohio, USA 10 Northwestern University, Chicago, Illinois, USA This article was presented at the 2016 AAO-HNSF Annual Meeting & OTO EXPO; September 18-21, 2016; San Diego, California. This article was also presented as a poster at the Infectious Diseases Society of America 2016 IDWeek; October 2016; New Orleans, Louisiana. Corresponding Author: Tamer A. Ghanem, MD, PhD, Department of Otolaryngology, Henry Ford Hospital, 2799 West Grand Boulevard, Detroit, MI 48202, USA. TGHANEM1@hfhs.org

2 60 Otolaryngology Head and Neck Surgery 159(1) Keywords surgical site infections, head and neck cancer microvascular reconstruction, antibiotic postoperative prophylaxis, antimicrobial stewardship Received December 30, 2016; revised November 9, 2017; accepted January 10, Surgical site infections (SSIs) in head and neck cancer microvascular reconstruction remain a significant postoperative complication, despite the increased use of postoperative antibiotic prophylaxis (POABP). 1-3 SSIs in this setting are associated with poor outcomes, including prolonged hospitalizations, development of additional health care associated infections, and death. 2-4 While advancements in head and neck reconstruction have continued to develop, there remain few data identifying modifiable risk factors for SSI prevention Antibiotic prophylaxis has reduced the incidence of SSI in patients who have undergone extensive surgical revision of the head and neck, but current surgical infection prophylaxis guideline recommendations may fail to account for the complexity of microvascular reconstruction relative to other head and neck procedures regarding spectrum and duration. 15 Previous head and neck SSI data identify polymicrobial infections involving anaerobic mouth flora and Gram-positive bacteria, but more recent findings suggest enteric Gramnegative (GN) bacteria as concerning pathogens Employing POABP without appropriate coverage, including GN bacteria, likely leads to increased incidence of SSI. Additional concerns exist for methicillin-resistant Staphylococcus aureus (MRSA) and Pseudomonas aeruginosa, 2 pathogens with severe public health implications. 15,19-21 Despite limited data surrounding the causative organisms of head and neck SSI in the oncology population, broad- or extended-spectrum antibiotic prophylaxis is routinely prescribed. 4,6,22 Strategies to prevent antimicrobial overuse, a key component of antimicrobial stewardship, include using the narrowest spectrum and shortest duration of antimicrobials needed to prevent infection. 23,24 Additionally, optimal POABP durations in this setting are unknown and, when combined with overuse of extended-spectrum antimicrobials, can lead to poor outcomes (eg, SSI), selection for resistant organisms, or severe adverse reactions (eg, Clostridium difficile associated diarrhea) Our hypothesis was that receiving POABP without GN coverage would lead to increased 30-day SSI. The objectives of this study were to characterize 30-day SSI and identify modifiable risk factors for SSI in relation to POABP spectrum and duration. Secondary objectives were to describe the time course to SSI and infection microbiology. Methods Study Population This multicenter cross-sectional study with a nested casecontrol design included hospitalized patients who received head and neck cancer ablation and microvascular reconstruction from January 2003 to March The study took place at 9 tertiary care hospitals throughout the United States, and Institutional Review Board approval was obtained from each site: Henry Ford Hospital, University of Kansas Medical Center, Johns Hopkins Hospital, The University of Texas MD Anderson Cancer Center, Washington University School of Medicine, Medical University of South Carolina Medical Center, Oregon Health Sciences University Hospital, The Ohio State Wexner Medical Center, and Northwestern Memorial Hospital. Additional inclusion criteria were as follows: age 18 years, documented oncologic disease, and available medical record data. Patients were excluded if they had an existing infection prior to surgery, required an infection-related operating room visit within 48 hours after initial surgery, or received surgery on an existing tumor within 90 days prior to microvascular reconstruction. Data Source The following data were extracted from each patient s electronic medical record: select characteristics (eg, age, sex), pertinent comorbid conditions (eg, diabetes mellitus, smoking and alcohol status, previous radiation therapy), surgery details (eg, flap and resection type, duration, POABP selection and duration), patient outcomes (eg, 30-day SSI, time to SSI, unexpected return to operating room), and SSI microbiology. The American Society of Anesthesiologists (ASA) classification system was used to identify presurgical health status. 27 All patients received pre-, peri-, and postoperative wound preparation and antimicrobial prophylaxis with redosing per individual institutional guidelines as the standard of care. All data were collected with a standardized electronic case report form created through REDCap (Research Electronic Data Capture, Vanderbilt University) and hosted on secure internal servers. Key Definitions All SSIs were retrospectively identified from real-time assessment of the treating surgeon within 30 days postprocedure and in compliance with the reporting standards of the ACS NSQIP (American College of Surgeons National Surgical Quality Improvement Program). 27 SSI involved the skin and subcutaneous tissue or deep soft tissues of incision and coincided with at least 1 of the criteria per the Centers for Disease Control and Prevention (CDC) National Health Safety Network for superficial or deep incisional SSI. 2 Infection sites were categorized to flap harvest (donor), recipient, nondonor or recipient sites (eg, stoma or tracheostomy, drain, flap, and/or incision sites), or multiple sites. Superficial incisional SSI criteria included any of following: Purulent drainage from the superficial incision If an organism was identified from an aseptically obtained specimen from the superficial incision or subcutaneous tissue by a culture- or nonculture-

3 Veve et al 61 Table 1. Spectrum of Activity for Commonly Used Antibiotics in Head and Neck Cancer Surgical Prophylaxis. POABP Classification: Antibiotic Name No Gram-negative coverage Clindamycin Metronidazole Linezolid, vancomycin Enteric Gram-negative coverage Cefazolin, cephalexin, ceftriaxone Amoxicillin/clavulanate, ampicillin/sulbactam, cefoxitin, cefotetan, ertapenem, moxifloxacin Doxycycline, trimethoprim/sulfamethoxazole Enteric Gram-negative with antipseudomonal coverage Aztreonam, gentamicin Cefepime, ciprofloxacin, levofloxacin Imipenem/cilastatin, meropenem, piperacillin/tazobactam Spectrum of Activity Gram-positive (including MRSA), anaerobes Anaerobes only Gram-positive (including MRSA) only Gram-positive, enteric Gram-negative Gram-positive, enteric Gram-negative, anaerobes Gram-positive (including MRSA), enteric Gram-negative Enteric Gram-negative with antipseudomonal Gram-positive, enteric Gram-negative with antipseudomonal Gram-positive, enteric Gram-negative with antipseudomonal, anaerobes Abbreviations: MRSA, methicillin-resistant Staphylococcus aureus; POABP, postoperative antibiotic prophylaxis. based microbiologic testing method performed for purposes of clinical diagnosis or treatment A superficial incision that was deliberately opened by a surgeon, attending physician, or other designee and culture- or nonculture-based testing was not performed and the patient had at least 1 of the following signs or symptoms: pain or tenderness, localized swelling, erythema, or heat Diagnosis of a superficial incisional SSI by the surgeon, attending physician, or other designee Deep incisional SSI criteria included any of the following: Purulent drainage from the deep incision A deep incision that spontaneously dehisced or was deliberately opened or aspirated by a surgeon, attending physician, or other designee and either (1) an organism that was identified by culture- or nonculture-based microbiological testing performed for the purposes of clinical diagnosis or treatment, or (2) at least 1 of the following signs or symptoms: fever (.38 C) or localized pain/tenderness An abscess or other evidence of infection involving the deep incision, as detected on gross anatomic or histopathologic examination or imaging test For the nested case-control sampling, cases were defined as patients who developed a 30-day SSI; controls were defined as patients without infection at 30-day postprocedure. Postoperative antibiotic prophylaxis was categorized a priori per the GN bacterial spectrum of activity, as determined by 2 infectious diseases pharmacists (M.P.V. and S.L.D.) blinded to the study: no GN coverage, enteric GN coverage, and enteric GN with antipseudomonal coverage. Table 1 lists all POABP regimens, including activity against anaerobic and Gram-positive bacteria, as well as examples of commonly used POABP antibiotics with spectrum of activity. The longest duration of any single or combination of antimicrobials was considered the maximum POABP duration. Statistical Analyses This study was designed to detect the prevalence of 30-day SSI in patients with head and neck cancer via descriptive measures. A nested case-control study was performed to identify risk factors for 30-day SSI; the primary variable of interest was POABP spectrum of activity based on previously published data. 4 Given the exploratory nature of this research, no formal sample size calculations were performed. Baseline patient characteristics were compared with bivariate analyses: categorical variables were compared by the Pearson s chi-square or Fisher s exact test; continuous variables were compared by the Student s t test or the Mann-Whitney U test. Comparisons were considered statistically significant if calculated P values were \.05. Secondary analyses were performed on variables associated with development of 30-day SSI from bivariate analyses. Classification and regression tree (CART) analyses were performed to identify breakpoints in continuous variables associated with a significant difference in 30-day SSI; for skewed continuous data, breakpoints were determined with quartiles. Any dichotomized variable that had an association with 30-day SSI (P \.2) or was deemed clinically relevant a priori was considered for inclusion into a multivariable logistic regression model. Variables were manually entered into the multivariable model with a backward-stepwise approach to determine independent associations with 30-day SSI while controlling

4 62 Otolaryngology Head and Neck Surgery 159(1) Table 2. Baseline Patient Characteristics. a Demographics Total (n = 1307) 30-d SSI (n = 189) No SSI (n = 1118) P Value Age, y b 62 (53-70) 63 (53-72) 61 (53-70).22 Male 894 (68) 116 (61) 778 (70).025 Diabetes mellitus 205 (16) 28 (15) 177 (16).72 Previous radiation therapy 452 (35) 65 (34) 387 (35) d chemotherapy 139 (11) 12 (6) 127 (11).04 ASA class (84) 162 (86) 929 (83).40 Smoking status, current 375 (28) 65 (34) 310 (28).06 Alcohol status, current 553 (42) 78 (41) 475 (43).75 Abbreviations: ASA, American Society of Anesthesiologists; SSI, surgical site infection. a Values are presented as n (%) unless noted otherwise. b Median (interquartile range). for potential confounders. The multivariable model was examined for goodness of fit with the Hosmer-Lemeshow test. All statistical analyses were performed with SPSS 23.0 for Macintosh (IBM Corp, Armonk, New York). Results A total of 1307 patients were included. Baseline patient characteristics are listed in Table 2. The most frequent cancer resection sites were as follows: 452 (35%), oral or oropharynx cavity soft tissue; 402 (31%), mandible or maxillary bony or anterior base of skull reconstruction; 169 (13%), external skin defect reconstruction; 135 (10%), laryngeal or hypopharyngeal reconstruction; 32 (2%), temporal bone or lateral skull base; and 117 (9%), other or not listed. The most frequent free-flap sites used included the following: 418 (32%), radial forearm fasciocutaneous; 341 (26%), anterolateral thigh free; 203 (16%), free fibula; 64 (5%), osteocutaneous radial forearm; 34 (3%), latissimus dorsi; 32 (3%), rectus abdominis; 18 (2%), osteocutaneous parascapular; 29 (2%), multiple sites; and 168 (12%), other or not listed. Pre-, peri-, and postoperative surgical preparation (eg, wound care and antimicrobial prophylaxis) was performed for all patients per individual institutional guidelines. A total of 189 (15%) patients developed 214 thirty-day SSIs; the median surgical duration was 9.6 hours (interquartile range [IQR], ); and the median time to SSI development was 11.5 days (IQR, 7-17). The most common surgical infection sites were as follows: 150 (70%), recipient site; 43 (20%), donor site; and 21 (10%), nonrecipient or donor site. Twenty-five (13%) patients had multiple infection sites. The most common recipient infection sites included oral or oropharynx cavity soft tissue (33%), mandible or maxillary bony or anterior base of skull reconstruction (28%), laryngeal or hypopharyngeal reconstruction (7%), external skin defect reconstruction (10%), temporal bone or lateral skull base (2%), and other or not listed (20%). For donor site infections, the most common had free flaps harvested from radial forearm fasciocutaneous (32%), free fibula (21%), osteocutaneous radial forearm (7%), rectus abdominis (5%), jejunum (2%), and other or not listed (22%). The majority of recipient site infections were deep (57%), but superficial infections were more common (82%) in donor sites. All nonrecipient or donor SSIs were superficial. Of 189 patients with 30-day SSI, 111 (59%) had positive surgical wound cultures, and a total of 176 organisms were obtained. Of all patients with 30-day SSI, 38 (20%) cultures were polymicrobial. Cultured organisms were as follows: 75 (40%), Gram-positives; 47 (25%), GNs; 20 (11%), Candida spp; 2 (1%), anaerobes; and 32 (18%), other organisms (Table 3). MRSA and P aeruginosa were isolated at 11 (5%) and 17 (15%), respectively. Seventy-eight (41%) patients were culture negative or had no cultures taken but still qualified for 30-day SSIbasedonpreviouslydiscussedcriteria. Postoperative antibiotic prophylaxis data were available for 1003 (77%) patients: 169 (17%), no GN; 524 (52%), enteric GN; 309 (31%), antipseudomonal GN. A total of 934 (93%) patients received an antibiotic with anaerobic bacterial coverage; 435 (43%), any anti-mrsa coverage; and 195 (20%), a nonlincosamide anti-mrsa antibiotic (eg, vancomycin or linezolid). The median duration of all POABP was 6 days (IQR, 3-7). Patients who received no- GN POABP were more likely to develop 30-day SSI (P \.001), have shorter median POABP durations (3 days [IQR, 1-7] vs 7 days [IQR, 4-8], P \.001), smoke at time of surgery (P =.09), and have an ASA classification 3 (P =.02). Infection with MRSA or P aeruginosa SSI was no different between patients who received an anti-mrsa (P =.54) or antipseudomonal (P =.42) antibiotic and those who did not, and no variables had an association for 30-day SSI with either organism. The results of bivariate analyses and clinical rationale dictated which variables were selected for inclusion into the final parsimonious multivariable model (Table 4): no-gn POABP, POABP duration 6 days, surgical duration 11.8 hours, ASA classification 3, and female sex. Additional bivariate analyses displayed that enteric GN POABP, but not antipseudomonal POABP, had a protective association for 30-day SSI (crude odds ratio [OR] 0.58; 95% CI, ); no-gn POABP was included in the model after

5 Veve et al 63 Table 3. Microbiological Data from 30-Day Surgical Site Infections (n = 189). Organisms Patients, n (%) No cultures or negative cultures 78 (41) Polymicrobial infection 38 (20) Monomicrobial infection 73 (39) Gram-positive 75 (40) Enterococcus faecium 2 (1) Enterococcus faecalis 4 (2) Streptococcus pneumoniae 3 (2) Methicillin-resistant Staphylococcus aureus 11 (6) Methicillin-susceptible S aureus 12 (6) Coagulase-negative Staphylococcus spp 18 (10) Streptococcus spp 25 (13) Gram-negative 47 (25) Acinetobacter baumannii 2 (1) Proteus spp 2 (1) Citrobacter spp 3 (2) Escherichia coli 4 (2) Enterobacter spp 6 (3) Klebsiella pneumoniae 5 (3) Pseudomonas aeruginosa 17 (9) Other Gram-negatives 8 (4) Other species 54 (29) Anaerobes 2 (1) Candida spp 20 (11) Other organisms 32 (18) considerations for goodness-of-fit testing. POABP with clindamycin monotherapy was associated with 30-day SSI (crude OR, 2.7; 95% CI, ) but was also excluded from the final model to prevent collinearity with other variables (eg, narrow POABP). Additional variables were excluded from the final model due to unmet clinical or statistical criteria. Variables found to independently predict 30- day SSI were as follows: receipt of no-gn POABP (adjusted OR [aor], 2.2; 95% CI, ), surgical duration 11.8 hours (aor, 1.9; 95% CI, ), and female sex (aor, 1.6; 95% CI, ). POABP duration 6 days did not retain a protective association for 30-day SSI (P =.08). A number of subanalyses were performed to delineate 30-day SSI risk when accounting for 2 potentially confounding subpopulations: (1) those with donor site infections and (2) cases that may be considered clean (eg, external skin resections). When the data were stratified to exclude 30-day donor SSI, all the variables obtained from the original bivariate analyses (Table 4) retained similar magnitude and associations for 30-day SSI, except female sex (P =.39). Extended POABP was associated with 30-day SSI (OR, 2.0; 95% CI, ), and external skin defect reconstruction was protective against 30-day recipient SSI (OR, 0.46; 95% CI, ); all other resection types had no associations. When analyses were stratified to exclude patients with external skin resections or cases perceived as clean, the results of the multivariable model displayed findings very similar to the original cross section, except that prolonged surgical duration was not associated with 30-day SSI (Table 5). Discussion The prevalence of 30-day SSI after head and neck cancer ablation and microvascular reconstruction in this study was 15%, which is similar to previously reported data of 18% to 42%. 4,6,7,10,18-20 Additionally, this study demonstrates that antibiotic prophylaxis with Gram-positive, enteric GN, and anaerobic coverage, but not antipseudomonal GN coverage, is associated with decreased 30-day SSI. Longer durations of POABP (6 days) and other traditional risk factors for infection (eg, current smoking status at time of surgery, diabetes mellitus, and previous radiation or chemotherapy) were not associated with 30-day SSI. Last, specific risk factors for MRSA and P aeruginosa, 2 uncommon organisms in this study, were not identified. These findings provide new insight on postoperative management in this patient population. The appropriate POABP spectrum in head and neck cancer microvascular reconstruction has been debated, but the existing evidence is limited. Weber et al performed a blinded randomized controlled trial that compared intravenous ampicillin/sulbactam (Gram-positive, GN, and anaerobic coverage) and clindamycin (Gram-positive and anaerobic coverage only) for patients who underwent head and neck cancer microvascular reconstruction. 28 The investigators found a lower incidence of SSI among patients who received ampicillin/sulbactam versus clindamycin (13% vs 27%, P =.02). While these findings were demonstrated in a number of additional studies, 4,19,29 some investigators found no difference in SSI when using GN POABP and not Many of these data focus on clindamycin POABP as a significant risk factor for SSI, suggesting that GN coverage leads to better patient outcomes. Identification of an ideal POABP duration has not been established and is often arbitrarily dichotomized to short and long durations. Johnson et al performed a randomized trial comparing 1 and 5 days of cefoperazone for patients with cancer who received free-flap reconstructive surgery. 37 The authors found no difference in SSI development (18.9% vs 25%, P..05) and concluded that POABP.1 day is unnecessary. Mitchell et al performed a retrospective analysis of free-flap cancer among patients receiving 24 hours of POABP versus prolonged courses; long POABP durations did not have a protective effect against SSI (P =.16). 30 Additional studies have not identified benefits from longer POABP durations. 25, The evidence behind employing extended-spectrum agents with activity against MRSA and P aeruginosa is also controversial. These organisms were previously identified with SSI, patients with prolonged hospitalizations, and those who are severely immunosuppressed, such as those with

6 64 Otolaryngology Head and Neck Surgery 159(1) Table 4. Variables Associated with 30-Day Surgical Site Infection in Bivariate and Multivariable Analyses. a Odds Ratio (95% CI) Characteristic Unadjusted Adjusted P Value No-GN POABP 2.0 ( ) 2.2 ( ) \.001 POABP duration 6 d 0.70 ( ) 0.71 ( ).08 Surgical duration 11.8 h 1.9 ( ) 1.9 ( ).001 Female sex 1.4 ( ) 1.6 ( ).01 ASA class ( ) 1.3 ( ).29 Smoking status, current 1.4 ( ) Alcohol status, current 0.95 ( ) Previous radiation therapy 0.99 ( ) Diabetes mellitus 0.93 ( ) 30-d history of chemotherapy 0.53 ( ) Abbreviations: ASA, American Society of Anesthesiologists; GN, Gram-negative; POABP, postoperative antibiotic prophylaxis. a Dashes ( ) indicate not tested. Table 5. Variables Associated with 30-Day Surgical Site Infection in Bivariate and Multivariable Analyses, Excluding Perceived Clean Cases. a Odds Ratio (95% CI) Characteristic Unadjusted Adjusted P Value No-GN POABP 2.3 ( ) 2.5 ( ) \.001 POABP duration 6 d 0.64 ( ) 0.74 ( ).15 Surgical duration 11.8 h 1.3 ( ) 1.4 ( ).11 Female sex 1.5 ( ) 1.6 ( ).03 ASA class ( ) 1.1 ( ).82 Smoking status, current 1.1 ( ) Alcohol status, current 0.86 ( ) Previous radiation therapy 1.0 ( ) Diabetes mellitus 1.1 ( ) 30-d history of chemotherapy 0.56 ( ) Abbreviations: ASA, American Society of Anesthesiologists; GN, Gram-negative; POABP, postoperative antibiotic prophylaxis. a Dashes ( ) indicate not tested. head and neck cancer. Both organisms also have the capability of colonizing skin or the prosthetic materials utilized in oncology during intravenous chemotherapy (eg, indwelling central venous catheters). The true incidence of these organisms in this patient population remains questionable or underreported, but the perceived risk likely contributes to extended POABP use. Durand et al performed a retrospective analysis on patients with head and neck cancer and found the prevalence of MRSA and P aeruginosa SSI to be 20% and 13%, respectively. 19 A key finding was that patients who were previously colonized with MRSA were more likely to develop MRSA SSI than patients who were not. Additional studies identified MRSA and P aeruginosa as SSI pathogens but usually in variable rates. 4,7,21 The CDC and World Health Organization global guidelines for preventing SSIs do not suggest administering antibiotic prophylaxis beyond the surgical duration. 43,44 While the evidence behind peri- versus postoperative antibiotic prophylaxis is debated in this setting, recent literature suggests the commonality and discordance of this practice. 4,30,42,45,46 Notably, Bartella et al performed a prospective comparison of antibiotic prophylaxis protocols in major head and neck cancer surgery, and they found significantly decreased SSI in patients who received peri- and postoperative antibiotics when compared with perioperative antibiotics alone (4% vs 36%, P =.011) or perioperative antibiotics with increased local antiseptic care (4% vs 36%, P =.011). 47 Langerman et al found similar results when extending POABP with ampicillin/sulbactam beyond the day of surgery (OR, 0.28; 95% CI, ). 36 These data suggest that wound classification and, thus, antibiotic prophylaxis recommendations may fail to account for surgical procedure duration, nature of radiated versus nonradiated field, and contamination of hardware and/or blood vessels to saliva during procedures. Therefore, before applying general recommendations for cases, it is important to

7 Veve et al 65 distinguish head and neck cancer microvascular procedures as having unique risk factors for SSI, necessitating at least some duration of POABP. While our study provides relevant information on POABP spectrum and durations in head and neck cancer microvascular reconstruction, a number of limitations should be addressed. The retrospective design of cross-sectional and case-control studies is noted but appropriate given our research question. A standardized electronic case report form was used to prevent information bias. Blinded adjudication was not used to confirm a diagnosis of 30-day SSI, but all SSIs coincided with ACS NSQIP and CDC National Health Safety Network reporting and definitions at time of diagnosis and would have been treated clinically. 2,27 The potential for selection bias, including POABP selection and dosing strategies, may have been present given the geographic differences among the 9 institutions. Surgical wounds in this study may have been misclassified as clean-contaminated (eg, external skin reconstruction or patients with fungating skin cancers). Subanalyses excluding these cases showed very similar findings to the total population; this suggests that the likelihood of wound misclassification is low. The multicenter design and large number of patients included in our study provide for significant heterogeneity that is well representative of the general population. In our multivariable analyses, antibiotic spectrum was analyzed on an aggregate level rather than individually, and specific optimal antibiotic regimens may not be clearly defined. Given the retrospective study design, we utilized this pragmatic method of categorization to allow for greater external validity for other clinicians, based on antibiotic availability or institution formulary restrictions. Additionally, specific comorbid conditions that may be associated with poor outcomes, such as chronic steroid use, obesity, or level of diabetes mellitus control, were not available for inclusion to our multivariable model due to a lack of data granularity. ASA classification 3 was selected as a surrogate variable for some of these comorbid conditions. All of the cases included in analyses were considered cleancontaminated based on the type of surgical procedure and surgeon consensus, but there is a lack of guidance on case classification for this type of surgery, which may have an impact on outcome analyses. 15 The POABP duration breakpoint of 6 days was selected according to data set distribution and not through CART: POABP duration data followed a bimodal distribution, and CART-identified breakpoints were unable to be identified. Breakpoints identified from continuous data are representative of only the data set utilized and should be interpreted as such. Conclusions The results of our study highlight that POABP is an important modifiable risk factor for preventing 30-day SSI in head and neck cancer microvascular reconstruction. Based on our data and existing evidence, short durations of broadspectrum antibiotics that cover Gram-positive, GN, and anaerobic organisms (eg, ceftriaxone plus metronidazole or ampicillin/sulbactam) are recommended. 48 These findings also have implications for patients with severe beta-lactam allergies, who should still receive appropriate prophylaxis with GN coverage, such as the combination of clindamycin and gentamicin, and are otherwise at increased risk for SSI. 49 Areas that warrant further emphasis are risk stratification tools for multidrug-resistant organisms, development of consistent wound classification definitions for free tissue transfer of the head and neck, and obtainment of quality surgical cultures when infection is suspected. While our data set clearly shows that POABP 6 days is not protective against developing postoperative SSI, POABP for even shorter durations should be used on the basis of current guidelines for case classification until randomized trials or additional data show a clear duration benefit. Acknowledgments We acknowledge Jamie L. Wagner, PharmD, School of Pharmacy, University of Mississippi, Jackson, Mississippi, USA, for her assistance with this study. Author Contributions Michael P. Veve, substantial contributions to the conception/ design of the work; analysis and interpretation of data, drafting the work and revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Joshua B. Greene, substantial contributions to the conception/design of the work; or the acquisition and interpretation of data for the work, drafting the work or revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Amy M. Williams, substantial contributions to the conception/design of the work; analysis and interpretation of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Susan L. Davis, substantial contributions to the conception or design of the work; analysis and interpretation of data, drafting the work or revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Nina Lu, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Yelizaveta Shnayder, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; David X. Li, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Salem I. Noureldine, substantial contributions to acquisition of data, drafting the work/ revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Jeremy D. Richmon, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Lawrence O. Lin, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be

8 66 Otolaryngology Head and Neck Surgery 159(1) published, agreement to be accountable for all aspects of the work; Matthew M. Hanasono, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Patrik Pipkorn, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Ryan S. Jackson, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Joshua D. Hornig, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Tyler Light, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Mark K. Wax, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Yin Yiu, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; James Bekeny, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Matthew Old, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; David Hernandez, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Urjeet A. Patel, substantial contributions to acquisition of data, drafting the work/revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work; Tamer A. Ghanem, substantial contributions to the conception/design of the work; the acquisition, analysis, and interpretation of data, drafting the work and revising it critically for important intellectual content, final approval of the version to be published, agreement to be accountable for all aspects of the work. Disclosures Competing interests: Yelizaveta Shnayder, HylaPharm share owner, advisor. Sponsorships: None. Funding source: None. References 1. Urken ML, Weinberg H, Buchbinder D, et al. Microvascular free flaps in head and neck reconstruction: report of 200 cases and review of complications. Arch Otolaryngol Head Neck Surg. 1994;120: Centers for Disease Control and Prevention. Surgical site infection (SSI) event definitions. pdfs/pscmanual/9pscssicurrent.pdf. Published January Accessed October Cunha TF, Soares Melancia TA, Zagalo Fernandes Ribeiro CM, et al. Risk factors for surgical site infection in cervicofacial oncological surgery. J Craniomaxillofac Surg. 2012;40: Wagner JL, Kenny RM, Vazquez JA, et al. Surgical prophylaxis with gram-negative activity for reduction of surgical site infections after microvascular reconstruction for head and neck cancer. Head Neck. 2016;38: Hanasono MM. Reconstructive surgery for head and neck cancer patients. Adv Med. 2014;2014: Yarlagadda BB, Deschler DG, Rich DL, et al. Head and neck free flap surgical site infections in the era of the Surgical Care Improvement Project. Head Neck. 2016;38(suppl 1):E392-E Karakida K, Aoki T, Ota Y, et al. Analysis of risk factors for surgical-site infections in 276 oral cancer surgeries with microvascular free-flap reconstructions at a single university hospital. J Infect Chemother. 2010;16: Lotfi CJ, de Cassia Calvalcanti R, Costa e Silva AM, et al. Risk factors for surgical-site infections in head and neck cancer surgery. Otolaryngol Head Neck Surg. 2008;138: Barry B, Lucet JC, Kosmann MJ, et al. Risk factors for surgical wound infections in patients undergoing head and neck oncologic surgery. Acta Otolaryngol. 1999;53: Penel N, Lefebvre D, Fournier C, et al. Risk factors for wound infection in head and neck cancer surgery: a prospective study. Head Neck. 2001;23: Dor P, Klastersky J. Prophylactic antibiotics in oral, pharyngeal and laryngeal surgery for cancer: a double-blind study. Laryngoscope. 1973;83: Ketcham AS, Bloch JH, Crawford DT, et al. The role of prophylactic antibiotic therapy in control of staphylococcal infections following cancer surgery. Surg Gynecol Obstet. 1962; 114: Becker GD, Parell GJ. Cefazolin prophylaxis in head and neck cancer surgery. Ann Otol Rhinol Laryngol. 1979;88: Mandell-Brown M, Johnson JT, Wagner RL. Cost-effectiveness of prophylactic antibiotics in head and neck surgery. Otolaryngol Head Neck Surg. 1984;92: Bratzler DW, Dellinger EP, Olsen KM, et al. Clinical practice guidelines for antimicrobial prophylaxis in surgery. Am J Health-Syst Pharm. 2013;70: Johnson JT, Yu VL, Myers EN, et al. Efficacy of two thirdgeneration cephalosporins in prophylaxis for head and neck surgery. Arch Otolaryngol. 1984;110: Brook I, Hirokawa R. Microbiology of wound infection after head and neck cancer surgery. Ann Otol Rhinol Laryngol. 1989;98: Man LX, Beswick DM, Johnson JT. Antibiotic prophylaxis in uncontaminated neck dissection. Laryngoscope. 2011;121: Durand ML, Yarlagadda BB, Rich DL, et al. The time course and microbiology of surgical site infections after head and neck free flap surgery. Laryngoscope. 2015;125: Yang CH, Chew KY, Solomkin JS, et al. Surgical site infections among high-risk patients in clean-contaminated head and neck reconstructive surgery: concordance with preoperative oral flora. Ann Plast Surg. 2013;71(suppl 1):S55-S60.

9 Veve et al Skitarelić N, Morović M, Manestar D. Antibiotic prophylaxis in clean-contaminated head and neck oncological surgery. J Craniomaxillofac Surg. 2007;35: Simons JP, Johnson JT, Yu VL, et al. The role of topical antibiotic prophylaxis in patients undergoing contaminated head and neck surgery with flap reconstruction. Laryngoscope. 2001;111: Chambers RM, Davis SL. Expanding the reach of antimicrobial stewardship. Pharmacotherapy. 2013;33: Barlam TF, Cosgrove SE, Abbo LM, et al. Implementing an antibiotic stewardship program: guidelines by the Infectious Diseases Society of America and the Society for Healthcare Epidemiology of America. Clin Infect Dis. 2016;62:e51-e Russell MD, Goldberg AN. What is the evidence for use of prophylactic antibiotics in clean-contaminated head and neck surgery? Laryngoscope. 2012;122: Brown KA, Fisman DN, Moineddin R, et al. The magnitude and duration of Clostridium difficile infection risk associated with antibiotic therapy: a hospital cohort study. PLoS One. 2014;9(8):e American College of Surgeons National Surgical Quality Improvement Program. Data variables and definitions. puf_userguide_2014.ashx. Published October Accessed October Weber RS, Raad I, Frankenthaler R, et al. Ampicillin-sulbactam vs clindamycin in head and neck oncologic surgery: the need for gram-negative coverage. Arch Otolaryngol Head Neck Surg. 1992;118: Callender DL. Antibiotic prophylaxis in head and neck oncologic surgery: the role of gram-negative coverage. Int J Antimicrob Agents. 1999;12(suppl 1):s21-s Mitchell RM, Mendez E, Schmitt NC, et al. Antibiotic prophylaxis in patients undergoing head and neck free flap reconstruction. JAMA Otolaryngol Head Neck Surg. 2015;141: Johnson JT, Yu VL. Antibiotic use during major head and neck surgery. Ann Surg. 1988;207: Johnson JT, Yu VL, Myers EN, et al. An assessment of the need for gram-negative bacterial coverage in antibiotic prophylaxis for oncological head and neck surgery. J Infect Dis. 1987;155: Johnson JT, Yu VL. Role of aerobic gram-negative rods, anaerobes, and fungi in wound infection after head and neck surgery: implications for antibiotic prophylaxis. Head Neck. 1989;11: Pool C, Kass J, Spivack J, et al. Increased surgical site infection rates following clindamycin use in head and neck free tissue transfer. Otolaryngol Head Neck Surg. 2016;154: Murphy J, Isaiah A, Dyalram D, Lubek JE. Surgical site infections in patients receiving osteomyocutaneous free flaps to the head and neck: does choice of antibiotic prophylaxis matter? J Oral Maxillofac Surg. 2017;75(10): Langerman A, Thisted R, Hohmann S, et al. Antibiotic and duration of perioperative prophylaxis predicts surgical site infection in head and neck surgery. Otolaryngol Head Neck Surg. 2016;154: Johnson JT, Schuller DE, Silver F, et al. Antibiotic prophylaxis in high-risk head and neck surgery: one-day vs five-day therapy. Otolaryngol Head Neck Surg. 1986;95: Piccart M, Dor P, Klastersky J. Antimicrobial prophylaxis of infections in head and neck cancer surgery. Scand J Infect Dis Suppl. 1983;39: Bhathena HM, Kavarana NM. Prophylactic antibiotics administration head and neck cancer surgery with major flap reconstruction: 1-day cefoperazone versus 5-day cefotaxime. Acta Chir Plast. 1998;40: Carroll WR, Rosenstiel D, Fix JR, et al. Three-dose vs extended-course clindamycin prophylaxis for free-flap reconstruction of the head and neck. Arch Otolaryngol Head Neck Surg. 2003;129: Khariwala SS, Le B, Pierce BH, et al. Antibiotic use after free tissue reconstruction of head and neck defects: short course vs long course. Surg Infect (Larchmt). 2016;17: Sephr A, Gutierrez Santos B, Chou C, et al. Antibiotics in head and neck surgery in the setting of malnutrition, tracheotomy, and diabetes. Laryngoscope. 2009;119: World Health Organization. Global guidelines for the prevention of surgical site infection. Published November Accessed November Berríos-Torres SI, Umscheid CA, Bratzler DW, et al. Centers for Disease Control and Prevention guideline for the prevention of surgical site infection, JAMA Surg. 2017;152: Cohen LE, Finnerty BM, Reiffel Golas A, et al. Perioperative antibiotics in the setting of oropharyngeal reconstruction: less is more. Ann Plast Surg. 2016;76: Reiffel AJ, Kamdar MR, Kadouch DJ, et al. Perioperative antibiotics in the setting of microvascular free tissue transfer: current practices. J Reconstr Microsurg. 2010;26: Bartella AK, Kamal M, Teichmann J, et al. Prospective comparison of perioperative antibiotic management protocols in oncological head and neck surgery. J Craniomaxillofac Surg. 2017;45: Veve MP, Davis SL, Williams AM, McKinnon JE, Ghanem TA. Considerations for antibiotic prophylaxis in head and neck cancer surgery. Oral Oncol. 2017;74: Blumenthal KG, Ryan EE, Li Y, Lee H, Kuhlen JL, Shenoy ES. The impact of a reported penicillin allergy on surgical site infection risk. Clin Infect Dis. 2017:cix794. doi: /cid/cix794