- Email: [email protected]
Previous Methicillin-Resistant Staphylococcus aureus Infection Independent of Body Site Increases Odds of Surgical Site Infection after Ventral Hernia Repair
Previous Methicillin-Resistant Staphylococcus aureus Infection Independent of Body Site Increases Odds of Surgical Site Infection after Ventral Hernia Repair Jenny Ousley, BS, Rebeccah B Baucom, MD, Melissa K Stewart, MD, Sharon E Phillips, MSPH, Michael D Holzman, MD, MPH, FACS, Jesse M Ehrenfeld, MD, MPH, Kenneth W Sharp, MD, FACS, William H Nealon, MD, FACS, Benjamin K Poulose, MD, MPH, FACS Methicillin-resistant Staphylococcus aureus infections can be difficult to manage in ventral hernia repair (VHR). We aimed to determine whether a history of preoperative MRSA infection, regardless of site, confers increased odds of 30-day surgical site infection (SSI) after VHR. STUDY DESIGN: A retrospective cohort study of patients undergoing VHR with class I to III wounds between 2005 and 2012 was performed using Vanderbilt University Medical Center’s Perioperative Data Warehouse. Preoperative MRSA status, site of infection, and 30-day SSI were determined. Univariate and multivariate analyses adjusting for confounding factors were performed to determine whether a history of MRSA infection was independently associated with SSIs. RESULTS: A total of 768 VHR patients met inclusion criteria, of which 46% were women. There were 54 (7%) preoperative MRSA infections (MRSA positive); 15 (28%) soft tissue, 9 (17%) bloodstream, 4 (7%) pulmonary, 3 (6%) urinary, and 5 (9%) other. Overall SSI rate was 10% (n ¼ 80), SSI rate in the MRSA-positive group was 33% (n ¼ 18), compared with 9% (n ¼ 62) in controls (p < 0.001). Multivariate analysis demonstrated that a history of MRSA infection significantly increased odds of 30-day SSI after VHR by 2.3 times (95% CI, 1.1-4.8; p ¼ 0.035). Other factors associated with postoperative SSI were performance of myofascial release, increasing BMI, length of operation, open repair, and clean-contaminated wound classification. CONCLUSIONS: A history of site-independent MRSA infection confers significantly increased odds of 30-day SSI after VHR. Additional investigation is needed to determine perioperative treatment regimens that might decrease odds of SSI in VHR, and optimal prosthetic types and techniques for this population. (J Am Coll Surg 2015;221:470e477. 2015 by the American College of Surgeons)
BACKGROUND:
Ventral hernia repair (VHR) is one of the most common procedures performed by general surgeons. In 2006, these hernia repairs were estimated to cost the United States $3.2 billion in hospital expenses alone.1 Due to a
higher-than-expected rate of surgical site infections (SSI) compared with other clean procedures and concern about mesh infections, a history of MRSA soft tissue infection presents a specific dilemma when considering VHR.2
Disclosure Information: Nothing to disclose. Disclosures outside the score of this work: Dr Poulose received a grant from Bard-Davol and is paid as a consultant to Ariste Medical. Dr Holzman has been paid by private attorneys for expert testimony in malpractice cases. All other authors have nothing to disclose. Support: Ms Ousley received research support from the Medical Scholars Program at the Vanderbilt University School of Medicine, Nashville, TN. Dr Nealon received travel support from Vanderbilt University Medical Center. Presented at the American College of Surgeons 100th Annual Clinical Congress, Surgical Forum, San Francisco, CA, October 2014.
Received February 21, 2015; Revised April 6, 2015; Accepted April 24, 2015. From the Division of General Surgery, Department of Surgery (Ousley, Baucom, Stewart, Holzman, Sharp, Nealon, Poulose), Department of Biostatistics (Phillips), and Department of Anesthesiology (Ehrenfeld), Vanderbilt University Medical Center, Nashville, TN. Correspondence address: Jenny Ousley, BS, Division of General Surgery, Department of Surgery, Vanderbilt University Medical Center, 1161 Medical Center Dr, D-5203 Medical Center North, Nashville, TN 37232. email: [email protected]
ª 2015 by the American College of Surgeons Published by Elsevier Inc.
470
http://dx.doi.org/10.1016/j.jamcollsurg.2015.04.023 ISSN 1072-7515/15
Vol. 221, No. 2, August 2015
Ousley et al
Abbreviations and Acronyms
EMR LOS OR PDW SSI VHR VUMC
¼ ¼ ¼ ¼ ¼ ¼ ¼
electronic medical record length of stay odds ratio perioperative data warehouse surgical site infection ventral hernia repair Vanderbilt University Medical Center
Much controversy exists about whether these patients should be managed differently, and a true correlation between a previous MRSA infection and postoperative SSIs has not been clearly defined. Some groups have suggested screening and eradication of MRSA in carriers to reduce postoperative infections, yet more than half of the patients with MRSA SSI in one of these studies were negative on nasal swab screening.3 In addition, the prevalence of MRSA continues to increase nationwide (66.4 per 1,000 inpatients in 2010 compared with 46.3 inpatients in 2006), with MRSA being the primary organism causing skin and soft tissue infections in the United States.4,5 Staphylococcus aureus infections are the most common contributor to SSIs, and constitute 30% to 50% of SSIs that develop after clean procedures.6,7 A 2005 study by Anderson and colleagues8 found that MRSA was the most common organism isolated in SSIs from 26 hospitals, representing 17% of all SSIs and 53% of SSIs attributable to S aureus. Faraday and colleagues9 found that a history of preoperative skin infections, no matter how remote from the date of operation, was associated with a 3-fold higher risk of deep and organ space SSI and infectious death after high-risk clean procedures. Although this study and others evaluated the links between preoperative skin and wound infections and SSI, there remains a paucity of knowledge about whether preoperative MRSA infections independent of body site contribute to the odds of SSI.10,11 The objective of this study was to determine whether a preoperative MRSA infection at any body site increases the odds of 30-day SSI after VHR. If a history of site-independent MRSA infection is found to increase the odds of postoperative SSI, efforts could then focus on appropriate interventions to reduce these odds.
METHODS Study design and data sources A retrospective cohort study was performed that included adult patients at Vanderbilt University Medical Center (VUMC) who underwent VHR between July 2005 and May 2012. Patients were identified by CPT codes
MRSA History and Surgical Site Infection
471
(49560, 49561, 49566, 49568, 49570, 49572, 49585, 49587, 49590, 49652, 49653, 49654, 49655, 49657) using Vanderbilt’s Perioperative Data Warehouse (PDW). The PDW is an institutional database containing clinical and administrative data on patients undergoing procedures. Vanderbilt University Medical Center’s electronic medical record (EMR) was queried and abstracted to supplement PDW data when needed. This study included elective repairs of both primary and incisional hernias. Small umbilical hernias and laparoscopic repairs were included, along with open VHRs, because mesh is often used in all of these cases, and concern exists about the use of prosthetic material in patients with a history of MRSA. Patients meeting any of the following criteria were excluded from analysis in this study: death within 30 days of operation without an SSI developing; active SSI at the time of operation; active mesh infection at the time of operation; and patients with class IV wounds. The Social Security Death Index was used to determine death status. Study data were collected and managed using REDCap electronic data capture tools, which is supported by the UL1 TR000445 grant from National Center for Advancing Translational Sciences/NIH and hosted at VUMC.12 The Human Research Protection Program at VUMC approved this study protocol. Determination of preoperative methicillin-resistant Staphylococcus aureus status Preoperative MRSA exposure was defined as a documented history of MRSA infection at any body site, up to and including the date of operation. Nasal swab screening is not routinely performed at VUMC, and our definition of MRSA positive focused on patients with a history of an active infection rather than the asymptomatic carrier state. Methicillin-resistant S aureus status was determined initially by a free-text query of the VUMC EMR for the term MRSA, followed by indepth review of preoperative clinic notes, anesthesia evaluations, and laboratory cultures for any mention of MRSA, S aureus, methicillin, or methicillin resistance. Culture results, including date and site of infection, as well as antibiotic sensitivities, were obtained from the VUMC laboratory, when available, for patients with a history of MRSA infection (MRSA positive). Referring hospital culture results were reviewed for patients without culture results at VUMC. Outcomes measures and confounding variables The primary outcome measure was 30-day SSI, as defined by the CDC and NSQIP criteria.13 The EMR (including progress notes, discharge summaries, emergency department visits, and outpatient clinic notes) was evaluated
472
Ousley et al
MRSA History and Surgical Site Infection
J Am Coll Surg
Table 1. Characteristics of Patients Undergoing Ventral Hernia Repair between 2005 and 2012 at Vanderbilt University Medical Center Characteristic
Age, y, mean (SD) Sex, n (%) Male Female Race/ethnicity, n (%) White African American Hispanic Other Comorbidities Smoker <30 d, n (%) Chronic steroids, n (%) Diabetes mellitus, n (%) BMI, kg/m2, mean (SD) Immunocompromised, n (%) ASA classification, n (%) I II III IV
MRSA positive, (n ¼ 54)
53 (14.65)
MRSA negative, (n ¼ 714)
p Value
47 (87.04) 6 (11.11) 1 (1.85) 0
54 (13.11) 0.553 0.786 383 (53.64) 331 (46.36) 0.077 613 (85.85) 54 (7.56) 4 (0.56) 43(5.60)
9 5 14 32
109 39 115 32
30 (55.56) 24 (44.44)
(16.67) (9.26) (25.93) (9.60)
9 (16.67) 14 (25.93) 40 (74.07) 0 0
(15.27) (5.46) (16.11) (7.93)
0.783 0.247 0.063 0.722
82 (11.48) 0.256 0.010 339 (47.61) 369 (51.83) 3 (0.42) 1 (0.14)
ASA, American Society of Anesthesiologists.
for documentation of an SSI. For analysis, a dichotomous variable was created for SSI, which included superficial incisional, deep incisional, and organ space infections. Hospital length of stay, in days, was analyzed as a secondary outcome measure. The EMR and PDW were also used to collect information about potential confounding variables, including perioperative antibiotic compliance; history of steroid use; diagnosis of diabetes mellitus; operative details, such as type of mesh used, operative technique, and concomitant procedures; and wound classification. For the purpose of this study, laparoscopic procedures that were converted to open were analyzed as open procedures. Operative notes were specifically evaluated for mention of myofascial release and, for the purpose of this study, our definition included any type of separation of abdominal wall components, such that fascia was separated from muscle. Other clinical information collected from the PDW included BMI, age, American Society of Anesthesiologists class, race, sex, length of operation, LOS, and intraoperative complications. Intraoperative complications were defined as hemorrhage requiring transfusion, peritoneal access injury, and injury to bowel (including partial and full thickness injuries),
Figure 1. Preoperative MRSA infections by site (n ¼ 54). Cultures for 18 patients were not available; those sites of infection are unknown and were not included in this figure.
bladder, liver, stomach, or major vessel requiring operative intervention. Smoking status within 30 days of operation was collected from preoperative anesthesia evaluations. Statistical analysis Descriptive statistics including means, medians, SDs, and ranges were determined. Comparisons of proportions and means were made using the Pearson chi-square, Wilcoxon rank sum, and Fisher’s exact tests, as appropriate, with a 2-tailed a of 0.05 considered significant. Multivariable logistic regression was used to determine the association between preoperative MRSA and SSI, adjusting for smoking, BMI, wound classification, laparoscopic vs open procedure, length of operation, and concurrent myofascial release performed during operation. Finally, ordinary least squares regression was used to determine the influence of preoperative MRSA infection on LOS. Analyses were performed using STATA software, version 13.1 (Stata Corp).
RESULTS Demographics There were 768 patients who met inclusion criteria; 46% (n ¼ 355) were female and mean SD age was 53.5 13.1 years. Comorbidities and demographics for the study population are illustrated in Table 1. There were 54 patients (7%) who had a history of MRSA infection before VHR, 3 of which were MRSA mesh infections. Body mass index, race, sex, smoking status, and chronic steroid use were similar between the exposure and control groups.
Ousley et al
Vol. 221, No. 2, August 2015
Table 2.
MRSA History and Surgical Site Infection
473
Operative Details for Overall Population
Operative characteristic
Length of operation, min, mean (SD) Operative approach, n (%) Open Laparoscopic Wound classification, n (%) Clean Clean contaminated Contaminated Mesh placed, n (%) Type of mesh, n (%) Permanent synthetic Absorbable synthetic Biologic Myofascial release, n (%) Fascia closed, n (%) Incarcerated hernia, n (%) Intraoperative complication, n (%) Subcutaneous flaps, n (%) SCIP-compliant antibiotics, n (%)
MRSA positive (n ¼ 54)
MRSA negative (n ¼ 714)
p Value
208 (102.46)
135 (83.16)
<0.001 0.001
53 (98.15) 1 (1.85)
567 (79.63) 145 (20.37)
40 8 6 43
(74.07) (14.81) (11.11) (79.63)
590 105 19 554
(82.63) (14.71) (2.66) (77.59)
16 1 26 16 47 50 12 24 52
(37.21) (2.33) (60.47) (29.63) (87.04) (92.59) (22.22) (44.44) (96.30)
437 9 105 63 486 484 36 107 703
(78.74) (1.62) (18.92) (8.85) (68.35) (67.98) (5.06) (15.01) (98.46)
0.013
0.728 <0.001
<0.001 0.004 <0.001 <0.001 <0.001 0.231
SCIP, Surgical Care Improvement Project.
The sites of preoperative MRSA infection are depicted in Figure 1; soft tissue and bloodstream infections were the most common. Other sites of infection included cerebrospinal fluid, maxillary sinus, Jackson-Pratt drain, and those infections for which culture results were unavailable. Operative details Operative details for the study population are depicted in Table 2. Mean SD length of operation was 140 87 minutes. Those with a history of a preoperative MRSA infection had a mean length of operation that was 73 minutes longer than the control group (208 vs 135 minutes; p < 0.001). Overall rate of open procedures was 80%; there were 10 VHRs that were attempted laparoscopically but converted to open procedures, and these were considered open VHRs for the purpose of our analysis. The frequency of open VHR was higher in the MRSA-positive group than controls (98% vs 80%; p ¼ 0.001), as was the rate of myofascial release (30% vs 9%; p < 0.001). The cohort included 175 patients who underwent umbilical hernia repair, 51% of whom had mesh placed during the operation. The MRSA-positive group had a higher rate of intraoperative complications (22% vs 5%; p < 0.001). The most common intraoperative complication was bowel injury, which occurred in 5% (n ¼ 42) of the study population and included both full and partial thickness (serosal) injuries.
There was a significant difference between the perioperative antibiotics administered to MRSA-positive patients when compared with the MRSA-negative group (p < 0.001). The MRSA-negative patients most often received cefazolin (67%) or cefoxitin (11%). The most common antibiotics in the MRSA-positive patients were cefazolin (43%), vancomycin (13%), clindamycin (9%), and cefoxitin (9%). Outcomes The overall rate of SSI in the study population was 10% (n ¼ 80). The rate of SSI in the exposure (MRSA-positive) group was 33% (n ¼ 18) compared with 9% (n ¼ 62) in the control group (p < 0.001). In 19% (n ¼ 15) of patients with SSIs, the organism cultured from the SSI was MRSA. The rate of MRSA SSI was 33% (n ¼ 6) in the MRSA-positive group compared with 15% (n ¼ 9) in the control group (p ¼ 0.07). There was no difference in the distribution of superficial incisional, deep incisional, and organ space infections between MRSA-positive and MRSA-negative populations (p ¼ 0.991). The breakdown of SSIs was as follows: 47% superficial incisional, 24% deep incisional, and 28% organ space infections. After 2 years of follow-up, there were 12 instances of mesh infection, representing a 2% mesh infection rate in both the MRSA-positive group (1 of 54) and MRSA-negative group (11 of 714) (p ¼ 0.586).
474
Ousley et al
MRSA History and Surgical Site Infection
Figure 2. Forest plot depicting factors associated with 30-day surgical site infection (SSI). MRSA, BMI, open ventral hernia repair (VHR), clean-contaminated wound class (Wound class: CC), and myofascial release were significant risk factors for SSI. Odds ratios (OR) are plotted with the vertical dashed line representing no effect (OR ¼ 1), and the horizontal lines represent 95% CIs. (Note that the 95% CI for open repair (2.9-168.9) has been eliminated for clarity). Wound Class: Cont, contaminated wound class.
Figure 2 is a graphical representation of the multivariable logistic regression model, which depicts the factors associated with postoperative SSI, including odds ratios (ORs) and 95% CIs. Covariates include MRSA status, smoking status, BMI, wound classification, length of operation, operative approach, and myofascial release. Preoperative MRSA infection increased the odds of postoperative SSI by 2.3 times (OR ¼ 2.26; 95% CI, 1.064.83; p ¼ 0.035). Body mass index (OR ¼ 1.05; 95% CI, 1.02-1.09; p ¼ 0.001), open VHR (OR ¼ 22.17; 95% CI, 2.90-168.94; p ¼ 0.003), increasing length of operation (OR ¼ 1.00; 95% CI, 1.00-1.01; p < 0.001), myofascial release (OR ¼ 3.43; 95% CI, 1.776.63; p < 0.001), and clean-contaminated wound class (OR ¼ 2.46; 95% CI, 1.24-4.87; p ¼ 0.010) were also independently associated with SSI. We also evaluated whether a previous MRSA soft tissue infection was associated with higher odds of postoperative SSI when compared with a previous MRSA infection at another site and found no difference (p ¼ 1). Median LOS for the MRSA-positive patients was 4.3 days (interquartile range 2.3 to 6.3 days) compared with 1.2 days for the MRSA-negative patients (interquartile range 0.3 to 4.0 days). Figure 3 is a graphical representation of a multivariable ordinary least squares regression model evaluating the effect of various factors on LOS.
J Am Coll Surg
Figure 3. Forest plot depicting number of days length of stay (LOS) increased for each variable in the regression model. Effect sizes (days) are shown; the vertical dashed line represents the line of no effect (0 days), and the horizontal lines represent 95% CI. VHR, ventral hernia repair; Wound Class: CC, clean contaminated wound class; Wound Class: Cont, contaminated wound class.
Preoperative MRSA infection significantly increased LOS by 1.5 days (95% CI, 0.46-2.54; p ¼ 0.005) after adjusting for operative approach, length of operation, wound classification, smoking, BMI, and myofascial release. Open VHR (effect 1.75 days; 95% CI, 1.072.43; p < 0.001), increasing length of operation (effect 0.02 days; 95% CI, 0.017-0.024; p < 0.001), myofascial release (effect 2.39 days; 95% CI, 1.46-3.33; p < 0.001), and increasing wound classification (clean contaminated, effect 2.43 days; 95% CI, 1.65-3.20; p < 0.001; contaminated, effect 1.74 days; 95% CI, 0.27-3.20; p ¼ 0.020) were significantly associated with increased LOS.
DISCUSSION Our findings indicate that a history of preoperative MRSA infection at any body site significantly increased the odds of 30-day SSI after VHR by 2.3 times. MRSA infection was also associated with longer operative times, a higher rate of intraoperative complications, and increased LOS by 1.5 days. Although a number of studies have been conducted on the topic of MRSA colonization, few studies have evaluated the impact of site-independent preoperative MRSA infections on surgical outcomes. This is the first study to evaluate a history of MRSA infection at any body site and its impact on postoperative SSI after VHR. Although SSIs are a feared complication in patients with a history of preoperative wound infection, few
Vol. 221, No. 2, August 2015
Ousley et al
studies have been conducted to assess the impact that preoperative infections in general have on surgical outcomes. A 2012 study by Blatnik and colleagues14 found that patients with previous wound infections (including MRSA and non-MRSA infections) did not exhibit higher odds of postoperative SSIs developing. The findings of this study, along with the gaps in the current literature, highlight the need for additional investigation into the relationship between preoperative infections and postoperative development of SSI in hernia patients. Given the known and perceived morbidity of MRSA infections, this particular subset of patients warrants additional evaluation. One interesting finding of our study was that although MRSA infection at any body site increased the odds of SSI after VHR, MRSA infection did not confer increased odds of mesh infection during the 2-year follow-up period. This is an important finding because much of the concern surrounds the use of mesh and future mesh infection in MRSA-positive patients. Due to the retrospective nature of the study, we were unable to assess the rate of recurrence in the population. However, multiple earlier studies have demonstrated an association between hernia formation and SSI, so we would expect the rate might be higher in the MRSA-positive group.15,16 The overall SSI rate in our study (10%), which included predominantly open VHRs, is comparable with other reported rates for open VHR.14,17,18 In a 2011 study, Pofahl and colleagues3 confirmed that preoperative eradication of the MRSA carrier state decreased the incidence of a postoperative MRSA SSI developing, yet more than half of patients with MRSA SSI in the study had negative screening results preoperatively. These results suggest that other risk factors exist for development of MRSA SSIs aside from the MRSA carrier state, and the results did not evaluate non-MRSA SSIs. A subsequent 2012 study by Ramirez and colleagues7 found that preoperative colonization was not a significant risk factor for SSI in patients undergoing gastrointestinal operations. Based on these studies, the impact of MRSA carrier state remains unclear. The findings of our study suggest that a history of active MRSA infection, rather than the carrier state, predisposes to subsequent SSIs. The mechanism for this increased risk is unclear, and it might represent a combination of both host and bacterial factors. Although the risk of an SSI developing differs between open and laparoscopic VHRs, mesh is used in both operations. Therefore, we believed it was vital to include both groups of patients for risk adjustment. For the same reason, umbilical hernia repairs were also included. Even when umbilical hernias were excluded from the multivariable model, the significance of MRSA infection was unchanged.
MRSA History and Surgical Site Infection
475
The use of myofascial release was an important independent predictor of SSI in this patient population, increasing the odds of an SSI developing by 3.4 times. This is perhaps secondary to alteration of blood flow to the surgical wound and development of multiple surgical planes. It might also be an indicator of a wound more likely to be under tension or a larger hernia repair. Additionally, this might be a marker of a more complex repair. The MRSA-positive patients more frequently underwent open repair with myofascial release, and this might be partially due to surgeon preference at our institution. Several surgeons favor placing prosthetic material in the well-vascularized retromuscular space rather than intraperitoneal position in patients who are MRSA positive. Additional evaluation of the association of myofascial release with SSI could have a significant impact on preoperative planning, particularly in patients who are at high risk for SSI. Our findings also correlate with previous studies that demonstrate an increased risk of SSI with increasingly contaminated wound classification and higher BMI. Berger and colleagues,19 in a study designed to identify risk factors for SSI and surgical site occurrence after VHR, found that BMI 40 and class 4 wounds were risk factors for SSI. Van Walraven and colleagues20 also confirmed that SSI risks increased with increasing BMI and in class 3 and 4 wounds. The findings of this study are clinically relevant for several reasons. First, MRSA infections are increasingly common, both in the community and inpatient settings. Existing literature reveals that MRSA has become the most common contributor to skin and soft tissue infections seen in emergency departments in many US cities.21 Second, these results demonstrate the potential to decrease morbidity and hospital costs in patients undergoing VHR by identifying patients with a preoperative history of MRSA and developing interventions that might reduce the odds of SSI. The differences in perioperative antibiotics given to MRSA-positive compared with MRSA-negative patients in our study was not surprising, but the study was not designed to draw generalizable conclusions about the impact this has on SSI. Additional study is needed to determine whether a longer duration of postoperative antibiotics or a more aggressive perioperative antibiotic protocol would reduce the odds of SSI in MRSA-positive patients. The information from this study can be used to power more definitive, prospective studies. In addition to optimizing the perioperative antibiotic regimen in MRSA-positive patients, our results suggest that selection of operative technique used during VHR can also be tailored for this population to reduce SSIs. Laparoscopic VHR is known to carry a much lower risk of postoperative SSI than open VHR. Although it is not
476
Ousley et al
MRSA History and Surgical Site Infection
clear whether myofascial release was an indicator of complexity of VHR in this study, or whether the technique itself actually alters factors related to wound healing and blood flow, the results suggest that alteration of surgical technique can reduce the odds of SSI in patients at high risk. This study could not accurately determine the type of myofascial release performed in every case, and prospective efforts, including registries, are better positioned to answer these detailed technical issues. Controversy exists about the appropriate selection of prosthetic and bioprosthetic materials for use in patients with an increased risk of infection. During the early 2000s, the use of bioprosthetic materials expanded greatly in VHR, particularly in operative fields with the potential for contamination; this has since become the prosthesis favored by many surgeons for situations with a high risk of infection.22 The appropriate selection of these materials for use in patients with a previous history of MRSA infection warrants additional study to determine whether certain types of prosthetics are associated with lower odds of SSI in this particular patient population. This study is limited by its retrospective nature, introducing ascertainment bias and unbalanced comparison groups. The 2 exposure groups, MRSA-positive and MRSA-negative, had different characteristics in this observational study. In general, the MRSA-positive group had more comorbidities compared with the MRSAnegative patients, with a significantly increased proportion of MRSA-negative patients undergoing laparoscopic repair (imparting less chance of SSI in this group). The MRSA-positive patients tended to undergo more complex repairs, raising the possibility that a history of MRSA infection is simply a marker for patients with more complicated hernias. We attempted to account for this in multivariable logistic regression, but some residual bias very likely existed when evaluating the SSI outcomes between the MRSA-positive and MRSA-negative groups. This significant limitation highlights the need for prospective data from a large dataset to answer this completely. We were unable to verify laboratory cultures for 18 of the preoperative MRSA infections. Many of these patients no longer had valid contact information in the medical record, were deceased, had unknown dates of infection, or records were no longer available. However, we made the assumption that when a health care provider recorded in the medical record that the patient had a history of MRSA infection they had reasonable proof of such history. It is also likely that some patients with a history of MRSA infection were undetected by our inclusion criteria preoperatively. Therefore, our sample of patients with preoperative MRSA infections might actually be under-
J Am Coll Surg
representative. One strength of our study was the ability to perform a free-text search within the medical record rather than relying solely on culture data. To gain a better understanding of the impact of MRSA, all patients undergoing VHR should be queried about previous MRSA infection, and this information recorded in the medical record. Finally, although MRSA infections are becoming increasingly common, the generalizability of our findings might be limited, as this study was conducted at a single tertiary care referral center with a predominantly Caucasian population.
CONCLUSIONS Patients with a site-independent preoperative MRSA infection are significantly more likely to have an SSI develop after VHR, and they have longer lengths of hospital stay when compared with those without preoperative MRSA infection. Based on these results, the assessment of past MRSA infections, no matter how remote from the date of current operation, is warranted before VHR. Patients and surgeons should be aware that the odds of SSI for these patients are 2.3 times that of patients without a history of MRSA. In addition, more research is needed to determine the optimal perioperative treatment of these patients to improve their outcomes. Author Contributions Study conception and design: Ousley, Baucom, Phillips, Holzman, Sharp, Nealon, Poulose Acquisition of data: Ousley, Baucom, Stewart, Ehrenfeld Analysis and interpretation of data: Ousley, Baucom, Phillips, Poulose Drafting of manuscript: Ousley, Baucom, Poulose Critical revision: Ousley, Baucom, Stewart, Phillips, Holzman, Ehrenfeld, Sharp, Nealon, Poulose Acknowledgment: Ms Ousley received funding from the Medical Scholars Program at the Vanderbilt University School of Medicine to complete this project. REFERENCES 1. Poulose BK, Shelton J, Phillips S, et al. Epidemiology and cost of ventral hernia repair: making the case for hernia research. Hernia 2012;16:179e183. 2. Breuing K, Butler CE, Ferzoco S, et al. Incisional ventral hernias: review of the literature and recommendations regarding the grading and technique of repair. Surgery 2010;148: 544e558. 3. Pofahl WE, Ramsey KM, Nobles DL, et al. Importance of methicillin-resistant Staphylococcus aureus eradication in carriers to prevent postoperative methicillin-resistant Staphylococcus aureus surgical site infection. Am Surg 2011;77: 27e31.
Vol. 221, No. 2, August 2015
Ousley et al
4. Jarvis WR, Jarvis AA, Chinn RY. National prevalence of methicillin-resistant Staphylococcus aureus in inpatients at United States health care facilities, 2010. Am J Infect Control 2012;40:194e200. 5. Dantes R, Mu Y, Belflower R, et al. National burden of invasive methicillin-resistant Staphylococcus aureus infections, United States, 2011. JAMA Intern Med 2013;173: 1970e1978. 6. Humphreys H. Staphylococcus aureus: the enduring pathogen in surgery. Surgeon 2012;10:357e360. 7. Ramirez MC, Marchessault M, Govednik-Horny C, et al. The impact of MRSA colonization on surgical site infection following major gastrointestinal surgery. J Gastrointest Surg 2013;17:144e152. 8. Anderson DJ, Sexton DJ, Kanafani ZA, et al. Severe surgical site infection in community hospitals: epidemiology, key procedures, and the changing prevalence of methicillin-resistant Staphylococcus aureus. Infect Control Hosp Epidemiol 2007;28:1047e1053. 9. Faraday N, Rock P, Lin EE, et al. Past history of skin infection and risk of surgical site infection after elective surgery. Ann Surg 2013;257:150e154. 10. Engemann JJ, Carmeli Y, Cosgrove SE, et al. Adverse clinical and economic outcomes attributable to methicillin resistance among patients with Staphylococcus aureus surgical site infection. Clin Infect Dis 2003;36:592e598. 11. Hicks CW, Blatnik JA, Krpata DM, et al. History of methicillin-resistant Staphylococcus aureus (MRSA) surgical site infection may not be a contraindication to ventral hernia repair with synthetic mesh: a preliminary report. Hernia 2014;18:65e70. 12. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)da metadata-driven methodology and
13. 14.
15. 16. 17. 18.
19.
20. 21. 22.
MRSA History and Surgical Site Infection
477
workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377e381. American College of Surgeons NSQIP. Chapter 4: ACS NSQIPdClassic and Essential Variables and Definitions. Chicago, IL: American College of Surgeons; 2012. Blatnik JA, Krpata DM, Novitsky YW, Rosen MJ. Does a history of wound infection predict postoperative surgical site infection after ventral hernia repair? Am J Surg 2012;203: 370e374; discussion 374. Sanders DL, Kingsnorth AN. The modern management of incisional hernias. BMJ 2012;344:e2843. Bucknall TE, Cox PJ, Ellis H. Burst abdomen and incisional hernia: a prospective study of 1129 major laparotomies. Br Med J (Clin Res Ed) 1982;284:931e933. White TJ, Santos MC, Thompson JS. Factors affecting wound complications in repair of ventral hernias. Am Surg 1998;64: 276e280. Medina M, Sillero M, Martı´nez-Gallego G, DelgadoRodrı´guez M. Risk factors of surgical wound infection in patients undergoing herniorrhaphy. Eur J Surg 1997;163: 191e198. Berger RL, Li LT, Hicks SC, et al. Development and validation of a risk-stratification score for surgical site occurrence and surgical site infection after open ventral hernia repair. J Am Coll Surg 2013;217:974e982. van Walraven C, Musselman R. The Surgical Site Infection Risk Score (SSIRS): a model to predict the risk of surgical site infections. PLoS One 2013;8:e67167. Moran GJ, Krishnadasan A, Gorwitz RJ, et al. Methicillinresistant S. aureus infections among patients in the emergency department. N Engl J Med 2006;355:666e674. Carbonell AM, Criss CN, Cobb WS, et al. Outcomes of synthetic mesh in contaminated ventral hernia repairs. J Am Coll Surg 2013;217:991e998.
BACKGROUND:
Ventral hernia repair (VHR) is one of the most common procedures performed by general surgeons. In 2006, these hernia repairs were estimated to cost the United States $3.2 billion in hospital expenses alone.1 Due to a
higher-than-expected rate of surgical site infections (SSI) compared with other clean procedures and concern about mesh infections, a history of MRSA soft tissue infection presents a specific dilemma when considering VHR.2
Disclosure Information: Nothing to disclose. Disclosures outside the score of this work: Dr Poulose received a grant from Bard-Davol and is paid as a consultant to Ariste Medical. Dr Holzman has been paid by private attorneys for expert testimony in malpractice cases. All other authors have nothing to disclose. Support: Ms Ousley received research support from the Medical Scholars Program at the Vanderbilt University School of Medicine, Nashville, TN. Dr Nealon received travel support from Vanderbilt University Medical Center. Presented at the American College of Surgeons 100th Annual Clinical Congress, Surgical Forum, San Francisco, CA, October 2014.
Received February 21, 2015; Revised April 6, 2015; Accepted April 24, 2015. From the Division of General Surgery, Department of Surgery (Ousley, Baucom, Stewart, Holzman, Sharp, Nealon, Poulose), Department of Biostatistics (Phillips), and Department of Anesthesiology (Ehrenfeld), Vanderbilt University Medical Center, Nashville, TN. Correspondence address: Jenny Ousley, BS, Division of General Surgery, Department of Surgery, Vanderbilt University Medical Center, 1161 Medical Center Dr, D-5203 Medical Center North, Nashville, TN 37232. email: [email protected]
ª 2015 by the American College of Surgeons Published by Elsevier Inc.
470
http://dx.doi.org/10.1016/j.jamcollsurg.2015.04.023 ISSN 1072-7515/15
Vol. 221, No. 2, August 2015
Ousley et al
Abbreviations and Acronyms
EMR LOS OR PDW SSI VHR VUMC
¼ ¼ ¼ ¼ ¼ ¼ ¼
electronic medical record length of stay odds ratio perioperative data warehouse surgical site infection ventral hernia repair Vanderbilt University Medical Center
Much controversy exists about whether these patients should be managed differently, and a true correlation between a previous MRSA infection and postoperative SSIs has not been clearly defined. Some groups have suggested screening and eradication of MRSA in carriers to reduce postoperative infections, yet more than half of the patients with MRSA SSI in one of these studies were negative on nasal swab screening.3 In addition, the prevalence of MRSA continues to increase nationwide (66.4 per 1,000 inpatients in 2010 compared with 46.3 inpatients in 2006), with MRSA being the primary organism causing skin and soft tissue infections in the United States.4,5 Staphylococcus aureus infections are the most common contributor to SSIs, and constitute 30% to 50% of SSIs that develop after clean procedures.6,7 A 2005 study by Anderson and colleagues8 found that MRSA was the most common organism isolated in SSIs from 26 hospitals, representing 17% of all SSIs and 53% of SSIs attributable to S aureus. Faraday and colleagues9 found that a history of preoperative skin infections, no matter how remote from the date of operation, was associated with a 3-fold higher risk of deep and organ space SSI and infectious death after high-risk clean procedures. Although this study and others evaluated the links between preoperative skin and wound infections and SSI, there remains a paucity of knowledge about whether preoperative MRSA infections independent of body site contribute to the odds of SSI.10,11 The objective of this study was to determine whether a preoperative MRSA infection at any body site increases the odds of 30-day SSI after VHR. If a history of site-independent MRSA infection is found to increase the odds of postoperative SSI, efforts could then focus on appropriate interventions to reduce these odds.
METHODS Study design and data sources A retrospective cohort study was performed that included adult patients at Vanderbilt University Medical Center (VUMC) who underwent VHR between July 2005 and May 2012. Patients were identified by CPT codes
MRSA History and Surgical Site Infection
471
(49560, 49561, 49566, 49568, 49570, 49572, 49585, 49587, 49590, 49652, 49653, 49654, 49655, 49657) using Vanderbilt’s Perioperative Data Warehouse (PDW). The PDW is an institutional database containing clinical and administrative data on patients undergoing procedures. Vanderbilt University Medical Center’s electronic medical record (EMR) was queried and abstracted to supplement PDW data when needed. This study included elective repairs of both primary and incisional hernias. Small umbilical hernias and laparoscopic repairs were included, along with open VHRs, because mesh is often used in all of these cases, and concern exists about the use of prosthetic material in patients with a history of MRSA. Patients meeting any of the following criteria were excluded from analysis in this study: death within 30 days of operation without an SSI developing; active SSI at the time of operation; active mesh infection at the time of operation; and patients with class IV wounds. The Social Security Death Index was used to determine death status. Study data were collected and managed using REDCap electronic data capture tools, which is supported by the UL1 TR000445 grant from National Center for Advancing Translational Sciences/NIH and hosted at VUMC.12 The Human Research Protection Program at VUMC approved this study protocol. Determination of preoperative methicillin-resistant Staphylococcus aureus status Preoperative MRSA exposure was defined as a documented history of MRSA infection at any body site, up to and including the date of operation. Nasal swab screening is not routinely performed at VUMC, and our definition of MRSA positive focused on patients with a history of an active infection rather than the asymptomatic carrier state. Methicillin-resistant S aureus status was determined initially by a free-text query of the VUMC EMR for the term MRSA, followed by indepth review of preoperative clinic notes, anesthesia evaluations, and laboratory cultures for any mention of MRSA, S aureus, methicillin, or methicillin resistance. Culture results, including date and site of infection, as well as antibiotic sensitivities, were obtained from the VUMC laboratory, when available, for patients with a history of MRSA infection (MRSA positive). Referring hospital culture results were reviewed for patients without culture results at VUMC. Outcomes measures and confounding variables The primary outcome measure was 30-day SSI, as defined by the CDC and NSQIP criteria.13 The EMR (including progress notes, discharge summaries, emergency department visits, and outpatient clinic notes) was evaluated
472
Ousley et al
MRSA History and Surgical Site Infection
J Am Coll Surg
Table 1. Characteristics of Patients Undergoing Ventral Hernia Repair between 2005 and 2012 at Vanderbilt University Medical Center Characteristic
Age, y, mean (SD) Sex, n (%) Male Female Race/ethnicity, n (%) White African American Hispanic Other Comorbidities Smoker <30 d, n (%) Chronic steroids, n (%) Diabetes mellitus, n (%) BMI, kg/m2, mean (SD) Immunocompromised, n (%) ASA classification, n (%) I II III IV
MRSA positive, (n ¼ 54)
53 (14.65)
MRSA negative, (n ¼ 714)
p Value
47 (87.04) 6 (11.11) 1 (1.85) 0
54 (13.11) 0.553 0.786 383 (53.64) 331 (46.36) 0.077 613 (85.85) 54 (7.56) 4 (0.56) 43(5.60)
9 5 14 32
109 39 115 32
30 (55.56) 24 (44.44)
(16.67) (9.26) (25.93) (9.60)
9 (16.67) 14 (25.93) 40 (74.07) 0 0
(15.27) (5.46) (16.11) (7.93)
0.783 0.247 0.063 0.722
82 (11.48) 0.256 0.010 339 (47.61) 369 (51.83) 3 (0.42) 1 (0.14)
ASA, American Society of Anesthesiologists.
for documentation of an SSI. For analysis, a dichotomous variable was created for SSI, which included superficial incisional, deep incisional, and organ space infections. Hospital length of stay, in days, was analyzed as a secondary outcome measure. The EMR and PDW were also used to collect information about potential confounding variables, including perioperative antibiotic compliance; history of steroid use; diagnosis of diabetes mellitus; operative details, such as type of mesh used, operative technique, and concomitant procedures; and wound classification. For the purpose of this study, laparoscopic procedures that were converted to open were analyzed as open procedures. Operative notes were specifically evaluated for mention of myofascial release and, for the purpose of this study, our definition included any type of separation of abdominal wall components, such that fascia was separated from muscle. Other clinical information collected from the PDW included BMI, age, American Society of Anesthesiologists class, race, sex, length of operation, LOS, and intraoperative complications. Intraoperative complications were defined as hemorrhage requiring transfusion, peritoneal access injury, and injury to bowel (including partial and full thickness injuries),
Figure 1. Preoperative MRSA infections by site (n ¼ 54). Cultures for 18 patients were not available; those sites of infection are unknown and were not included in this figure.
bladder, liver, stomach, or major vessel requiring operative intervention. Smoking status within 30 days of operation was collected from preoperative anesthesia evaluations. Statistical analysis Descriptive statistics including means, medians, SDs, and ranges were determined. Comparisons of proportions and means were made using the Pearson chi-square, Wilcoxon rank sum, and Fisher’s exact tests, as appropriate, with a 2-tailed a of 0.05 considered significant. Multivariable logistic regression was used to determine the association between preoperative MRSA and SSI, adjusting for smoking, BMI, wound classification, laparoscopic vs open procedure, length of operation, and concurrent myofascial release performed during operation. Finally, ordinary least squares regression was used to determine the influence of preoperative MRSA infection on LOS. Analyses were performed using STATA software, version 13.1 (Stata Corp).
RESULTS Demographics There were 768 patients who met inclusion criteria; 46% (n ¼ 355) were female and mean SD age was 53.5 13.1 years. Comorbidities and demographics for the study population are illustrated in Table 1. There were 54 patients (7%) who had a history of MRSA infection before VHR, 3 of which were MRSA mesh infections. Body mass index, race, sex, smoking status, and chronic steroid use were similar between the exposure and control groups.
Ousley et al
Vol. 221, No. 2, August 2015
Table 2.
MRSA History and Surgical Site Infection
473
Operative Details for Overall Population
Operative characteristic
Length of operation, min, mean (SD) Operative approach, n (%) Open Laparoscopic Wound classification, n (%) Clean Clean contaminated Contaminated Mesh placed, n (%) Type of mesh, n (%) Permanent synthetic Absorbable synthetic Biologic Myofascial release, n (%) Fascia closed, n (%) Incarcerated hernia, n (%) Intraoperative complication, n (%) Subcutaneous flaps, n (%) SCIP-compliant antibiotics, n (%)
MRSA positive (n ¼ 54)
MRSA negative (n ¼ 714)
p Value
208 (102.46)
135 (83.16)
<0.001 0.001
53 (98.15) 1 (1.85)
567 (79.63) 145 (20.37)
40 8 6 43
(74.07) (14.81) (11.11) (79.63)
590 105 19 554
(82.63) (14.71) (2.66) (77.59)
16 1 26 16 47 50 12 24 52
(37.21) (2.33) (60.47) (29.63) (87.04) (92.59) (22.22) (44.44) (96.30)
437 9 105 63 486 484 36 107 703
(78.74) (1.62) (18.92) (8.85) (68.35) (67.98) (5.06) (15.01) (98.46)
0.013
0.728 <0.001
<0.001 0.004 <0.001 <0.001 <0.001 0.231
SCIP, Surgical Care Improvement Project.
The sites of preoperative MRSA infection are depicted in Figure 1; soft tissue and bloodstream infections were the most common. Other sites of infection included cerebrospinal fluid, maxillary sinus, Jackson-Pratt drain, and those infections for which culture results were unavailable. Operative details Operative details for the study population are depicted in Table 2. Mean SD length of operation was 140 87 minutes. Those with a history of a preoperative MRSA infection had a mean length of operation that was 73 minutes longer than the control group (208 vs 135 minutes; p < 0.001). Overall rate of open procedures was 80%; there were 10 VHRs that were attempted laparoscopically but converted to open procedures, and these were considered open VHRs for the purpose of our analysis. The frequency of open VHR was higher in the MRSA-positive group than controls (98% vs 80%; p ¼ 0.001), as was the rate of myofascial release (30% vs 9%; p < 0.001). The cohort included 175 patients who underwent umbilical hernia repair, 51% of whom had mesh placed during the operation. The MRSA-positive group had a higher rate of intraoperative complications (22% vs 5%; p < 0.001). The most common intraoperative complication was bowel injury, which occurred in 5% (n ¼ 42) of the study population and included both full and partial thickness (serosal) injuries.
There was a significant difference between the perioperative antibiotics administered to MRSA-positive patients when compared with the MRSA-negative group (p < 0.001). The MRSA-negative patients most often received cefazolin (67%) or cefoxitin (11%). The most common antibiotics in the MRSA-positive patients were cefazolin (43%), vancomycin (13%), clindamycin (9%), and cefoxitin (9%). Outcomes The overall rate of SSI in the study population was 10% (n ¼ 80). The rate of SSI in the exposure (MRSA-positive) group was 33% (n ¼ 18) compared with 9% (n ¼ 62) in the control group (p < 0.001). In 19% (n ¼ 15) of patients with SSIs, the organism cultured from the SSI was MRSA. The rate of MRSA SSI was 33% (n ¼ 6) in the MRSA-positive group compared with 15% (n ¼ 9) in the control group (p ¼ 0.07). There was no difference in the distribution of superficial incisional, deep incisional, and organ space infections between MRSA-positive and MRSA-negative populations (p ¼ 0.991). The breakdown of SSIs was as follows: 47% superficial incisional, 24% deep incisional, and 28% organ space infections. After 2 years of follow-up, there were 12 instances of mesh infection, representing a 2% mesh infection rate in both the MRSA-positive group (1 of 54) and MRSA-negative group (11 of 714) (p ¼ 0.586).
474
Ousley et al
MRSA History and Surgical Site Infection
Figure 2. Forest plot depicting factors associated with 30-day surgical site infection (SSI). MRSA, BMI, open ventral hernia repair (VHR), clean-contaminated wound class (Wound class: CC), and myofascial release were significant risk factors for SSI. Odds ratios (OR) are plotted with the vertical dashed line representing no effect (OR ¼ 1), and the horizontal lines represent 95% CIs. (Note that the 95% CI for open repair (2.9-168.9) has been eliminated for clarity). Wound Class: Cont, contaminated wound class.
Figure 2 is a graphical representation of the multivariable logistic regression model, which depicts the factors associated with postoperative SSI, including odds ratios (ORs) and 95% CIs. Covariates include MRSA status, smoking status, BMI, wound classification, length of operation, operative approach, and myofascial release. Preoperative MRSA infection increased the odds of postoperative SSI by 2.3 times (OR ¼ 2.26; 95% CI, 1.064.83; p ¼ 0.035). Body mass index (OR ¼ 1.05; 95% CI, 1.02-1.09; p ¼ 0.001), open VHR (OR ¼ 22.17; 95% CI, 2.90-168.94; p ¼ 0.003), increasing length of operation (OR ¼ 1.00; 95% CI, 1.00-1.01; p < 0.001), myofascial release (OR ¼ 3.43; 95% CI, 1.776.63; p < 0.001), and clean-contaminated wound class (OR ¼ 2.46; 95% CI, 1.24-4.87; p ¼ 0.010) were also independently associated with SSI. We also evaluated whether a previous MRSA soft tissue infection was associated with higher odds of postoperative SSI when compared with a previous MRSA infection at another site and found no difference (p ¼ 1). Median LOS for the MRSA-positive patients was 4.3 days (interquartile range 2.3 to 6.3 days) compared with 1.2 days for the MRSA-negative patients (interquartile range 0.3 to 4.0 days). Figure 3 is a graphical representation of a multivariable ordinary least squares regression model evaluating the effect of various factors on LOS.
J Am Coll Surg
Figure 3. Forest plot depicting number of days length of stay (LOS) increased for each variable in the regression model. Effect sizes (days) are shown; the vertical dashed line represents the line of no effect (0 days), and the horizontal lines represent 95% CI. VHR, ventral hernia repair; Wound Class: CC, clean contaminated wound class; Wound Class: Cont, contaminated wound class.
Preoperative MRSA infection significantly increased LOS by 1.5 days (95% CI, 0.46-2.54; p ¼ 0.005) after adjusting for operative approach, length of operation, wound classification, smoking, BMI, and myofascial release. Open VHR (effect 1.75 days; 95% CI, 1.072.43; p < 0.001), increasing length of operation (effect 0.02 days; 95% CI, 0.017-0.024; p < 0.001), myofascial release (effect 2.39 days; 95% CI, 1.46-3.33; p < 0.001), and increasing wound classification (clean contaminated, effect 2.43 days; 95% CI, 1.65-3.20; p < 0.001; contaminated, effect 1.74 days; 95% CI, 0.27-3.20; p ¼ 0.020) were significantly associated with increased LOS.
DISCUSSION Our findings indicate that a history of preoperative MRSA infection at any body site significantly increased the odds of 30-day SSI after VHR by 2.3 times. MRSA infection was also associated with longer operative times, a higher rate of intraoperative complications, and increased LOS by 1.5 days. Although a number of studies have been conducted on the topic of MRSA colonization, few studies have evaluated the impact of site-independent preoperative MRSA infections on surgical outcomes. This is the first study to evaluate a history of MRSA infection at any body site and its impact on postoperative SSI after VHR. Although SSIs are a feared complication in patients with a history of preoperative wound infection, few
Vol. 221, No. 2, August 2015
Ousley et al
studies have been conducted to assess the impact that preoperative infections in general have on surgical outcomes. A 2012 study by Blatnik and colleagues14 found that patients with previous wound infections (including MRSA and non-MRSA infections) did not exhibit higher odds of postoperative SSIs developing. The findings of this study, along with the gaps in the current literature, highlight the need for additional investigation into the relationship between preoperative infections and postoperative development of SSI in hernia patients. Given the known and perceived morbidity of MRSA infections, this particular subset of patients warrants additional evaluation. One interesting finding of our study was that although MRSA infection at any body site increased the odds of SSI after VHR, MRSA infection did not confer increased odds of mesh infection during the 2-year follow-up period. This is an important finding because much of the concern surrounds the use of mesh and future mesh infection in MRSA-positive patients. Due to the retrospective nature of the study, we were unable to assess the rate of recurrence in the population. However, multiple earlier studies have demonstrated an association between hernia formation and SSI, so we would expect the rate might be higher in the MRSA-positive group.15,16 The overall SSI rate in our study (10%), which included predominantly open VHRs, is comparable with other reported rates for open VHR.14,17,18 In a 2011 study, Pofahl and colleagues3 confirmed that preoperative eradication of the MRSA carrier state decreased the incidence of a postoperative MRSA SSI developing, yet more than half of patients with MRSA SSI in the study had negative screening results preoperatively. These results suggest that other risk factors exist for development of MRSA SSIs aside from the MRSA carrier state, and the results did not evaluate non-MRSA SSIs. A subsequent 2012 study by Ramirez and colleagues7 found that preoperative colonization was not a significant risk factor for SSI in patients undergoing gastrointestinal operations. Based on these studies, the impact of MRSA carrier state remains unclear. The findings of our study suggest that a history of active MRSA infection, rather than the carrier state, predisposes to subsequent SSIs. The mechanism for this increased risk is unclear, and it might represent a combination of both host and bacterial factors. Although the risk of an SSI developing differs between open and laparoscopic VHRs, mesh is used in both operations. Therefore, we believed it was vital to include both groups of patients for risk adjustment. For the same reason, umbilical hernia repairs were also included. Even when umbilical hernias were excluded from the multivariable model, the significance of MRSA infection was unchanged.
MRSA History and Surgical Site Infection
475
The use of myofascial release was an important independent predictor of SSI in this patient population, increasing the odds of an SSI developing by 3.4 times. This is perhaps secondary to alteration of blood flow to the surgical wound and development of multiple surgical planes. It might also be an indicator of a wound more likely to be under tension or a larger hernia repair. Additionally, this might be a marker of a more complex repair. The MRSA-positive patients more frequently underwent open repair with myofascial release, and this might be partially due to surgeon preference at our institution. Several surgeons favor placing prosthetic material in the well-vascularized retromuscular space rather than intraperitoneal position in patients who are MRSA positive. Additional evaluation of the association of myofascial release with SSI could have a significant impact on preoperative planning, particularly in patients who are at high risk for SSI. Our findings also correlate with previous studies that demonstrate an increased risk of SSI with increasingly contaminated wound classification and higher BMI. Berger and colleagues,19 in a study designed to identify risk factors for SSI and surgical site occurrence after VHR, found that BMI 40 and class 4 wounds were risk factors for SSI. Van Walraven and colleagues20 also confirmed that SSI risks increased with increasing BMI and in class 3 and 4 wounds. The findings of this study are clinically relevant for several reasons. First, MRSA infections are increasingly common, both in the community and inpatient settings. Existing literature reveals that MRSA has become the most common contributor to skin and soft tissue infections seen in emergency departments in many US cities.21 Second, these results demonstrate the potential to decrease morbidity and hospital costs in patients undergoing VHR by identifying patients with a preoperative history of MRSA and developing interventions that might reduce the odds of SSI. The differences in perioperative antibiotics given to MRSA-positive compared with MRSA-negative patients in our study was not surprising, but the study was not designed to draw generalizable conclusions about the impact this has on SSI. Additional study is needed to determine whether a longer duration of postoperative antibiotics or a more aggressive perioperative antibiotic protocol would reduce the odds of SSI in MRSA-positive patients. The information from this study can be used to power more definitive, prospective studies. In addition to optimizing the perioperative antibiotic regimen in MRSA-positive patients, our results suggest that selection of operative technique used during VHR can also be tailored for this population to reduce SSIs. Laparoscopic VHR is known to carry a much lower risk of postoperative SSI than open VHR. Although it is not
476
Ousley et al
MRSA History and Surgical Site Infection
clear whether myofascial release was an indicator of complexity of VHR in this study, or whether the technique itself actually alters factors related to wound healing and blood flow, the results suggest that alteration of surgical technique can reduce the odds of SSI in patients at high risk. This study could not accurately determine the type of myofascial release performed in every case, and prospective efforts, including registries, are better positioned to answer these detailed technical issues. Controversy exists about the appropriate selection of prosthetic and bioprosthetic materials for use in patients with an increased risk of infection. During the early 2000s, the use of bioprosthetic materials expanded greatly in VHR, particularly in operative fields with the potential for contamination; this has since become the prosthesis favored by many surgeons for situations with a high risk of infection.22 The appropriate selection of these materials for use in patients with a previous history of MRSA infection warrants additional study to determine whether certain types of prosthetics are associated with lower odds of SSI in this particular patient population. This study is limited by its retrospective nature, introducing ascertainment bias and unbalanced comparison groups. The 2 exposure groups, MRSA-positive and MRSA-negative, had different characteristics in this observational study. In general, the MRSA-positive group had more comorbidities compared with the MRSAnegative patients, with a significantly increased proportion of MRSA-negative patients undergoing laparoscopic repair (imparting less chance of SSI in this group). The MRSA-positive patients tended to undergo more complex repairs, raising the possibility that a history of MRSA infection is simply a marker for patients with more complicated hernias. We attempted to account for this in multivariable logistic regression, but some residual bias very likely existed when evaluating the SSI outcomes between the MRSA-positive and MRSA-negative groups. This significant limitation highlights the need for prospective data from a large dataset to answer this completely. We were unable to verify laboratory cultures for 18 of the preoperative MRSA infections. Many of these patients no longer had valid contact information in the medical record, were deceased, had unknown dates of infection, or records were no longer available. However, we made the assumption that when a health care provider recorded in the medical record that the patient had a history of MRSA infection they had reasonable proof of such history. It is also likely that some patients with a history of MRSA infection were undetected by our inclusion criteria preoperatively. Therefore, our sample of patients with preoperative MRSA infections might actually be under-
J Am Coll Surg
representative. One strength of our study was the ability to perform a free-text search within the medical record rather than relying solely on culture data. To gain a better understanding of the impact of MRSA, all patients undergoing VHR should be queried about previous MRSA infection, and this information recorded in the medical record. Finally, although MRSA infections are becoming increasingly common, the generalizability of our findings might be limited, as this study was conducted at a single tertiary care referral center with a predominantly Caucasian population.
CONCLUSIONS Patients with a site-independent preoperative MRSA infection are significantly more likely to have an SSI develop after VHR, and they have longer lengths of hospital stay when compared with those without preoperative MRSA infection. Based on these results, the assessment of past MRSA infections, no matter how remote from the date of current operation, is warranted before VHR. Patients and surgeons should be aware that the odds of SSI for these patients are 2.3 times that of patients without a history of MRSA. In addition, more research is needed to determine the optimal perioperative treatment of these patients to improve their outcomes. Author Contributions Study conception and design: Ousley, Baucom, Phillips, Holzman, Sharp, Nealon, Poulose Acquisition of data: Ousley, Baucom, Stewart, Ehrenfeld Analysis and interpretation of data: Ousley, Baucom, Phillips, Poulose Drafting of manuscript: Ousley, Baucom, Poulose Critical revision: Ousley, Baucom, Stewart, Phillips, Holzman, Ehrenfeld, Sharp, Nealon, Poulose Acknowledgment: Ms Ousley received funding from the Medical Scholars Program at the Vanderbilt University School of Medicine to complete this project. REFERENCES 1. Poulose BK, Shelton J, Phillips S, et al. Epidemiology and cost of ventral hernia repair: making the case for hernia research. Hernia 2012;16:179e183. 2. Breuing K, Butler CE, Ferzoco S, et al. Incisional ventral hernias: review of the literature and recommendations regarding the grading and technique of repair. Surgery 2010;148: 544e558. 3. Pofahl WE, Ramsey KM, Nobles DL, et al. Importance of methicillin-resistant Staphylococcus aureus eradication in carriers to prevent postoperative methicillin-resistant Staphylococcus aureus surgical site infection. Am Surg 2011;77: 27e31.
Vol. 221, No. 2, August 2015
Ousley et al
4. Jarvis WR, Jarvis AA, Chinn RY. National prevalence of methicillin-resistant Staphylococcus aureus in inpatients at United States health care facilities, 2010. Am J Infect Control 2012;40:194e200. 5. Dantes R, Mu Y, Belflower R, et al. National burden of invasive methicillin-resistant Staphylococcus aureus infections, United States, 2011. JAMA Intern Med 2013;173: 1970e1978. 6. Humphreys H. Staphylococcus aureus: the enduring pathogen in surgery. Surgeon 2012;10:357e360. 7. Ramirez MC, Marchessault M, Govednik-Horny C, et al. The impact of MRSA colonization on surgical site infection following major gastrointestinal surgery. J Gastrointest Surg 2013;17:144e152. 8. Anderson DJ, Sexton DJ, Kanafani ZA, et al. Severe surgical site infection in community hospitals: epidemiology, key procedures, and the changing prevalence of methicillin-resistant Staphylococcus aureus. Infect Control Hosp Epidemiol 2007;28:1047e1053. 9. Faraday N, Rock P, Lin EE, et al. Past history of skin infection and risk of surgical site infection after elective surgery. Ann Surg 2013;257:150e154. 10. Engemann JJ, Carmeli Y, Cosgrove SE, et al. Adverse clinical and economic outcomes attributable to methicillin resistance among patients with Staphylococcus aureus surgical site infection. Clin Infect Dis 2003;36:592e598. 11. Hicks CW, Blatnik JA, Krpata DM, et al. History of methicillin-resistant Staphylococcus aureus (MRSA) surgical site infection may not be a contraindication to ventral hernia repair with synthetic mesh: a preliminary report. Hernia 2014;18:65e70. 12. Harris PA, Taylor R, Thielke R, et al. Research electronic data capture (REDCap)da metadata-driven methodology and
13. 14.
15. 16. 17. 18.
19.
20. 21. 22.
MRSA History and Surgical Site Infection
477
workflow process for providing translational research informatics support. J Biomed Inform 2009;42:377e381. American College of Surgeons NSQIP. Chapter 4: ACS NSQIPdClassic and Essential Variables and Definitions. Chicago, IL: American College of Surgeons; 2012. Blatnik JA, Krpata DM, Novitsky YW, Rosen MJ. Does a history of wound infection predict postoperative surgical site infection after ventral hernia repair? Am J Surg 2012;203: 370e374; discussion 374. Sanders DL, Kingsnorth AN. The modern management of incisional hernias. BMJ 2012;344:e2843. Bucknall TE, Cox PJ, Ellis H. Burst abdomen and incisional hernia: a prospective study of 1129 major laparotomies. Br Med J (Clin Res Ed) 1982;284:931e933. White TJ, Santos MC, Thompson JS. Factors affecting wound complications in repair of ventral hernias. Am Surg 1998;64: 276e280. Medina M, Sillero M, Martı´nez-Gallego G, DelgadoRodrı´guez M. Risk factors of surgical wound infection in patients undergoing herniorrhaphy. Eur J Surg 1997;163: 191e198. Berger RL, Li LT, Hicks SC, et al. Development and validation of a risk-stratification score for surgical site occurrence and surgical site infection after open ventral hernia repair. J Am Coll Surg 2013;217:974e982. van Walraven C, Musselman R. The Surgical Site Infection Risk Score (SSIRS): a model to predict the risk of surgical site infections. PLoS One 2013;8:e67167. Moran GJ, Krishnadasan A, Gorwitz RJ, et al. Methicillinresistant S. aureus infections among patients in the emergency department. N Engl J Med 2006;355:666e674. Carbonell AM, Criss CN, Cobb WS, et al. Outcomes of synthetic mesh in contaminated ventral hernia repairs. J Am Coll Surg 2013;217:991e998.