Surgical Site Infections Associated with Methicillin-Resistant Staphylococcus aureus: Do Postoperative Factors Play a Role?
- Dr. Farrin A. Manian, Div. of Infectious Diseases, St. John's Mercy Medical Center, 621 S. New Ballas, Tower B, Ste. 3002, Creve Coeur, MO 63141 (manianfa{at}aol.com).
-
Presented in part: 12th Annual Scientific Meeting of the Society for Healthcare Epidemiology of America, Salt Lake City, Utah, April 2002 (abstract 248).
Abstract
Patients with surgical site infections (SSIs) who underwent surgery during the period of September 1997 through December 1999 and January through July 2001 were retrospectively studied to compare patients infected with methicillin-resistant Staphylococcus aureus (MRSA) with those infected with organisms other than MRSA. Of patients with SSI who had known culture results, 77 (28.5%) of 270 had cultures that yielded MRSA. On univariate analysis, age of ⩾70 years, duration of surgery of ⩾4 h, duration of postoperative antibiotic treatment of >1 day, and discharge to a long-term care facility (LTCF) were significantly associated with MRSA SSI (P <.05 for all). On multivariate analysis, only discharge to an LTCF (odds ratio [OR], 2.3; P =.04) and duration of postoperative antibiotic treatment of >1 day (OR, 2.0; P =.03) were significantly associated with MRSA SSI; there was also a trend toward MRSA SSI being associated with use of a surgical drain for >1 day (P =.078). Postoperative factors may play a more important role in the causation of MRSA SSI than has previously been appreciated.
Methicillin-resistant Staphylococcus aureus (MRSA) is an increasingly important nosocomial pathogen that causes >50% of hospital-acquired S. aureus infections in the United States [1]. MRSA has also become an important cause of postoperative infection [2]. Isolation of MRSA from surgical sites has been associated with a lower frequency of primary healing and delayed healing [3]. At St. John's Mercy Medical Center (SJMMC; St. Louis, MO), the percentage of surgical site infections (SSIs) caused by MRSA increased from 9% in 1995 to 30% in 2000 (authors' unpublished observation). Despite the increasing importance of MRSA in nosocomial infections, published studies specifically addressing risk factors associated with MRSA SSIs in settings where they are endemic have been lacking. The following study was performed to help identify factors that may be uniquely associated with MRSA SSIs compared with other SSIs in hopes of designing potential interventions aimed at reducing the incidence of endemic MRSA SSI at our medical center.
Patients And Methods
Setting. SJMMC is an 859-bed tertiary care hospital located in Creve Coeur, a suburb of St. Louis, Missouri. More than 20,000 surgical procedures are performed annually, and approximately two-thirds of the procedures are performed for outpatients without anticipated hospitalization during the postoperative period. Patients with MRSA SSI are routinely placed in private rooms or cohorted with other patients who have MRSA infection or colonization. In the intensive care units (ICUs), in addition to universal precautions and private rooms, gowns and gloves are routinely used on entry into the room of patients who are colonized or infected with MRSA. The majority (90%) of patients transferred to long-term care facilities (LTCFs) are admitted to the skilled nursing or rehabilitation facilities affiliated with SJMMC.
Study design. We conducted a retrospective cohort study of all patients with SSI who underwent surgery at our medical center during the periods of September 1997 through December 1999 and January 2001 through July 2001 (total of 35 months). SSI was defined in accordance with the Centers for Disease Control and Prevention (CDC) criteria [4]. Surveillance for SSI was performed prospectively as part of the quality-improvement activities of the medical center by 3 infection-control practitioners; traditional methods of SSI detection (i.e., in-house surveillance, hospital admissions, and review of microbiological laboratory reports) and postdischarge surveillance via monthly mailed questionnaires to all surgeons (detailed elsewhere [5, 6]) were used.
The following preoperative characteristics were recorded and analyzed for each patient who had an SSI: age, sex, American Society of Anesthesiology risk category, diabetes status, whether the patient had undergone dialysis, MRSA colonization or infection during the previous year, and the type and timing of antibiotic surgical prophylaxis (relative to the incision time). Intraoperative variables studied included surgical site class (according to the CDC categories [4]), estimated blood loss, duration of surgery, surgeon, and surgical group. The latter comprised 4 categories: inpatients (i.e., those who had been hospitalized for ⩾24 h before surgery), patients with same-day admissions (i.e., those who were admitted to the hospital on the same day of their surgery), patients with emergency department admissions, and outpatients. Postoperative variables included duration of antibiotic therapy after surgery, presence of surgical drain and duration of use, immediate postoperative ICU stay, and discharge disposition (i.e., discharge to home, LTCF, or another hospital or death).
Definitions. The classification of the depth of SSI (e.g., superficial, deep, or organ space) was based on published CDC criteria [4]. MRSA was defined as an S. aureus isolate with an MIC of oxacillin of ⩾4 μg/mL. An MRSA SSI was defined as a case associated with isolation of either pure or mixed growth of MRSA from a culture of a surgical site specimen.
Statistics. Statistical analyses were performed using the Statistica (StatSoft) and InStat (GraphPad) software programs. The χ2 test with Yates' correction and Fisher's exact test were used for comparison of categorical data. The Mann-Whitney U test was used for comparison of continuous data. Two-tailed tests of statistical significance were performed, and P <.05 was considered to be statistically significant. Stepwise logistic regression analysis of selected variables from univariate analysis was performed to delineate factors that may be independently associated with MRSA SSI.
Results
During the study periods, 73,154 surgical procedures were performed, of which 293 (0.4%) were associated with SSI (1.2% were associated with SSI for patients who required immediate postoperative hospitalization, and 0.03% were associated with SSI for outpatients). Of 23,671 procedures performed among hospitalized patients, 75 (0.3%) were associated with MRSA SSI; this rate was significantly higher than the rate for outpatients (2 [0.004%] of 49,483 procedures; relative risk, 78; 95% CI, 19–317; P <.0001).
Of the 293 SSIs, 161 (55%) occurred in boys and men. The median patient age was 58 years (range, newborn to 95 years). Surgical group composition was as follows: inpatient group, 99 patients (34%); same-day admission group, 157 (54%); emergency department group, 22 (8%); and outpatient group, 15 (5.1%). The median interval between surgery and diagnosis of SSI was 14 days (range, 1–252 days).
The distribution of SSIs according to type of surgery is shown in table 1. Orthopedic surgery associated with the placement of a prosthetic device was the most common type of procedure, followed by gastrointestinal-colorectal, vascular, and spinal procedures. Prophylactic antibiotics were used by 287 patients (98%). Cefazolin was administered to 199 patients with SSI (68%), vancomycin was administered to 63 (22%), cefotetan was administered to 10 (3%), and miscellaneous antibiotics were administered to 15 (5%). Of 257 patients with SSIs associated with antibiotic prophylaxis and for which the timing of antibiotic administration was documented, 243 (95%) received prophylaxis ⩽2 h before the incision was performed. Of the 289 patients with SSIs who had documented dates of diagnosis of SSI and discharge from the hospital, 189 (65%) received the diagnosis after discharge.
Table 1
Distribution of surgical site infections (SSIs), by surgical procedure.
The bacteriology of the surgical site cultures is shown in table 2. MRSA (pure or mixed growth) accounted for 28.5% of all SSIs, followed by methicillin-susceptible S. aureus (MSSA; pure or mixed growth; 19%) and miscellaneous organisms that consisted primarily of aerobic, gram-negative bacilli (21%). No samples for culture were obtained from 23 patients with known SSI; these patients were omitted from further analysis comparing MRSA and non-MRSA SSI groups. After undergoing outpatient surgery or after hospital admission involving surgery, 203 (69%) of 293 patients were discharged home, 55 (19%) were transferred to an LTCF (85% were transferred to a skilled nursing facility, and 15% were transferred to a rehabilitation facility), 11 (4%) died, and 24 (8%) were still in the hospital at the time that study data were collected.
Table 2
Distribution of organisms isolated from infected surgical sites.
Of 262 SSIs for which the type of SSI was known, 106 (40%) were superficial space infections, 108 (41%) involved deep space, and 48 (18%) were organ-space infections. Of the 74 patients with MRSA SSI for whom the type of SSI was known, 31 (42%) were superficial space infections, 32 (43%) involved deep space, and 11 (15%) were organ-space infections; these rates were not significantly different than those for the non-MRSA SSIs (40%, 40%, and 20%, respectively).
The frequency of selected characteristics of patients in the MRSA and non-MRSA groups is shown in table 3. Age of ⩾70 years, duration of surgery of ⩾4 h, duration of postoperative antibiotic therapy of >24 h, postoperative use of surgical drains >24 h (compared with patients with no drains or drains in place for ⩽24 h), postoperative hospital stay of >3 days, and discharge to an LTCF were significantly associated with MRSA SSI; of 77 patients with MRSA SSI, 19 (25%) received diagnoses of SSI after transfer to an LTCF, compared with 19 (9.9%) of 193 patients with non-MRSA SSI (OR, 3.0; 95% CI, 1.5–6.1; P =.003). Vancomycin prophylaxis (administered ⩽2 h after incision) was inversely associated with MRSA SSI, reaching borderline statistical significance (P =.058). The following variables were not significantly associated with MRSA SSI: diabetes, dialysis, history of MRSA colonization or infection, preoperative hospital stay, preoperative hospital surgical group (including the inpatient group), duration of hospitalization before surgery of >3 days, surgical site class (clean and clean-contaminated vs. others), estimated blood loss, and immediate postoperative ICU stay. There was also no significant association between specific surgeons and MRSA SSI (data not shown).
Table 3
Comparison of the frequency of various characteristics of patients with methicillin-resistant Staphylococcus aureus (MRSA) surgical site infection (SSI) and those with SSI due to organisms other than MRSA (non-MRSA).
The proportion of SSIs diagnosed before discharge from the hospital was identical in MRSA and non-MRSA groups (26 [34%] of 77 vs. 65 [34%] of 193; P = 1.00). Among hospitalized patients, the median number of days of preoperative hospital stay among patients with MRSA SSI and among those with non-MRSA SSI was not significantly different (2.5 days for the MRSA group and 3.0 days for the non-MRSA group; P =.95). There was also no significant difference between the 2 groups with regard to the median number of days between surgery and diagnosis of SSI (14 days for both groups) and the median duration of ICU stay during the immediate postoperative period (2 days for both groups).
We then compared patients who had MRSA SSI with patients who had methicillin-susceptible S. aureus (MSSA) SSI (including pure MSSA cases as well SSIs associated with MSSA plus growth of other organisms on culture) with respect to the previously discussed variables. One patient had SSI that was simultaneously associated with MRSA and MSSA growth on culture of surgical site specimens; this patient was included in the MRSA group. Categorical attributes significantly associated with MRSA SSI on univariate analysis are shown in table 4. Patients with MRSA were more likely to have been discharged to an LTCF, to have received antibiotics for >24 h after the operation, to have had drains in place for >24 h after the operation, and to have had >3 days of hospitalization immediately after surgery. The median interval between surgery and diagnosis of MSSA SSI was significantly longer than the median interval between surgery and diagnosis of MRSA SSI (26 days vs. 14 days, respectively; P =.006). There were no other significant differences between MRSA and non-MRSA groups with respect to the remaining variables (data not shown).
Table 4
Comparison of selected characteristics of patients with methicillin-resistant Staphylococcus aureus (MRSA) surgical site infection (SSI) and those with methicillin-susceptible S. aureus (MSSA) SSI.
The results of multivariate analysis of risk factors associated with MRSA SSI compared with all other SSIs using logistic regression are shown in table 5. Seven variables that reached (or that nearly reached) statistical significance on univariate analysis (P <.06) shown in table 3 were entered into the model. Only discharge to an LTCF and duration of postoperative antibiotic therapy of >24 h were independently associated with MRSA SSI, with a trend toward duration of drain use of >24 h also being associated with such infections (P =.078).
Table 5
Findings of multivariate logistic regression analysis of selected variables associated with methicillin-resistant Staphylococcus aureus surgical site infection.
Discussion
MRSA is an increasingly important nosocomial pathogen. In a recent study from New York City hospitals, MRSA accounted for 29% of all nosocomial infections caused by S. aureus, with an attributable mortality rate of 21% (compared with 8% for MSSA) and an attributable cost of $3700 more than that for MSSA [7]. Risk factors for MRSA infection have included diabetes, receipt of dialysis, peripheral vascular disease, prolonged antibiotic exposure, and extended length of stay in the ICU [8–10]. In a recent study of S. aureus bacteremia at the time of hospital admission, nursing home residence, hospitalization during the 6 months before admission, receipt of antimicrobial agents in the 3 months before hospital admission, and presence of indwelling urinary catheters were significantly associated with MRSA infection [11]. Surprisingly, however, very little is known about the factors specifically associated with MRSA SSI. We are aware of only a single previously published study that compared risk factors related to MRSA SSI compared with those of non-MRSA SSI [12]. This study involved 35 patients with MRSA SSI in an outbreak in an Australian hospital in 1979 and 1980 that had a relatively high total SSI rate of >10%. Given the potentially disastrous effect of SSI on the health of patients with its attendant increased cost of care, as well as the likelihood that MRSA infection may add to the burden of rather than replace MSSA nosocomial infection [13], a more recent study of factors that may be unique to endemic MRSA SSI is essential. To our knowledge, ours is the most comprehensive study to date to have compared MRSA SSI with SSIs caused by other organisms.
We found that MRSA SSIs are significantly associated with discharge to an LTCF (even after exclusion of patients whose SSI was diagnosed before discharge) and duration of postoperative antibiotic therapy of >24 h. Use of surgical drains for >24 h was also independently associated with MRSA SSI; this nearly reached statistical significance on logistic regression analysis. None of the other preoperative patient attributes, such as age, American Society of Anesthesiology risk category, presence of diabetes, and lack of vancomycin use as surgical prophylaxis, were independently associated with MRSA SSI. Similarly, intraoperative factors, such as surgical site class, inpatient status before surgery, duration of surgery, surgeon, and estimated blood loss, did not appear to be independent predictors of MRSA SSI among infected patients. We also failed to determine that immediate postoperative ICU stay was an important predictor of MRSA SSI. Whether this finding is related to the presence of only private rooms in our ICU, an appropriate nurse-patient ratio, or other factors is not clear. Previously, nursing staff workload was reported to be a determinant in the spread of MRSA infection in an adult ICU [14].
We also found significant differences between MRSA and MSSA SSIs with regard to several postoperative variables, including discharge to an LTCF, use of surgical drains for >24 h, duration of antibiotic treatment of >24 h, and hospital stay of >3 days after surgery. Again, major differences between SSIs caused by these 2 organisms seem to revolve around variables associated with the postoperative period. We cannot readily explain the longer interval between surgery and diagnosis of SSI associated with MSSA compared with MRSA SSIs. One possible explanation may be that, because of its association with postoperative stay in an LTCF, patients with MRSA might have been receiving closer monitoring, which, therefore, led to earlier diagnosis of SSI.
The results of our study may be useful in at least 2 ways. First, the presence of risk factors for MRSA SSI may help in deciding which patient may need appropriate isolation or empirical therapy with vancomycin before surgical site culture results are available. For example, in our study, patients who presented with SSI after a postoperative stay at an LTCF had a 45% risk (95% CI, 32%–59%) of having MRSA isolated from their surgical site, whereas those who had received postoperative antibiotics for >24 h had a 40% risk (95% CI, 30%–50%).
Our study also raises some questions about the pathophysiological characteristics of many MRSA SSIs. It is interesting that no preoperative or intraoperative factors were associated with isolation of MRSA from surgical sites among patients with SSI. It is generally held that the great majority of SSIs occur as a result of intraoperative contamination of the surgical site, and that “the fate of the wound relative to subsequent infection is determined at the time of wound closure” [15, p. 147]. Given our finding of a strong association between MRSA SSI and postoperative factors when compared with other SSIs, it is distinctly possible that the surgical site of many patients with MRSA SSI actually became contaminated after surgical site closure. It is likely that prolonged antibiotic pressure (e.g., for >24 h after the operation), LTCF stay, and having drains left in place for more than a day all facilitate postoperative exposure of surgical sites to MRSA when this organism is endemic in certain institutions. Ross [12] reported a significant association between MRSA SSI in an epidemic setting and postoperative antibiotic administration. Rezende et al. [11] recently reported a strong association between MRSA bacteremia at the time of hospital admission and nursing home residence. Surgical site drains, which have long been considered a potential conduit for organisms that contaminate the surgical site after its closure [16, 17], have also been previously demonstrated to be independently associated with MRSA SSI [12].
Several limitations of our study are worth discussing. First, we did not study previous antibiotic therapy (e.g., during the previous several months) as a risk factor for MRSA colonization and subsequent infection. However, given the retrospective nature of our study and the fact that the majority of patients underwent surgical procedures in nonhospital settings, this information would have been difficult to obtain reliably from chart review. Furthermore, recent reports of the ineffectiveness of nare decolonization with mupirocin for prevention of S. aureus SSIs [18, 19] raise some doubts regarding the relative importance of preoperative staphylococcal colonization in the causation of S. aureus SSIs. Second, given the relatively large number of variables studied (20), it is possible that significant association of some of the variables with MRSA SSI occurred by chance alone. Finally, our data set was based on observations in one medical center only, and the majority of patients were transferred to 1 of 2 LTCFs (skilled nursing and rehabilitation facilities) associated with SJMMC. However, given the relatively large number of SSIs during nearly 3 years at a community-based facility, our results are probably generalizable to many health care centers that have a case mix similar to ours.
Several postoperative variables are associated with MRSA SSIs compared with SSIs caused by organisms other than MRSA. These include postoperative transfer to an LTCF, duration of antibiotic treatment of >24 h after surgery, and possibly also placement of surgical drains for >24 h. Knowledge of these variables may help predict which patients with SSI are likely to have MRSA isolated from their surgical sites. In addition, our findings raise questions regarding the pathophysiological characteristics of MRSA SSI and suggest that postoperative factors may be more important than appreciated in its causation. Further studies involving multiple health care centers are needed to help corroborate our conclusions.
Acknowledgments
We thank Marlene Northway and Tammy Manian for their invaluable assistance with data entry.
- Received October 9, 2002.
- Accepted December 11, 2002.
- © 2003 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
