Clostridium difficile infection (CDI) following surgical procedures is associated with morbidity, mortality, and cost. Length of stay is increased by 1 week if surgery is complicated by a CDI, and mortality has been reported as high as 35% in trauma surgery patients.Reference Maltenfort, Rasouli, Morrison and Parvizi 1 , Reference Sharma, Bomireddy and Phillips 2 Cost of care more than doubles if the postoperative patient contracts CDI.Reference Maltenfort, Rasouli, Morrison and Parvizi 1 Therefore, it is essential to reduce the risk of CDI in surgical patients.
Previous studies have reported CDI rates for individual specialties or after specific procedures with small sample sizes.Reference Maltenfort, Rasouli, Morrison and Parvizi 1 , Reference Damle, Cherng and Flahive 3 – Reference Krapohl, Morris and Cai 7 To our knowledge, there have been no reports of CDI across multiple surgical specialties. Differences across surgical specialties could be due to overall health of the patient population, complexity of surgery, perioperative antibiotic selection, pre- and postoperative exposure to additional antibiotics, or many other factors. The aims of our study were to determine the incidence of postoperative CDI across four surgical specialties (orthopedic surgery, neurosurgery, trauma surgery, and general surgery) and to examine risk factors for CDI.
METHODS
This study was conducted at a 592-bed tertiary-care academic center, and we were granted an exemption by the institutional review board. Using the hospital’s quality improvement database, all admissions to 4 hospital units (ie, orthopedic surgery, neurosurgery, trauma surgery, and general surgery) from January 2014 to July 2016 were reviewed. Patients who underwent an inpatient surgical procedure were included. Exclusion criteria included a documented CDI in the 2 months prior to admission or within 72 hours of admission. A case patient was defined as a patient who underwent an inpatient surgical procedure and subsequently developed healthcare-associated CDI (HA-CDI), defined as a positive PCR for C. difficile toxin gene recorded more than 72 hours after admission and within 12 weeks of discharge. Controls were patients selected from the same group of patients who underwent an inpatient surgical procedure but did not develop HA-CDI. The electronic medical record was used for data extraction. Variables of interest were extracted by a single author.
Statistical Methods
Control patients were matched in triplicate based on age, sex, and admitting department. If more than 3 controls fit the matching criteria for a case, then 3 were chosen at random.
Data analysis included summarizing the cases with means (±standard deviation [SD]) for normally distributed continuous variables, medians (with interquartile range [IQR]) for nonnormally distributed continuous variables, and frequency (%) for categorical variables. Associations between collected data and subject group were first assessed with univariable mixed-effect logistic regression (MELR) models with matched pair number as a random effect followed by multivariable MELR analyses with a priori defined covariates of body mass index (BMI), American Society of Anesthesiologists (ASA) classification, and antibiotic use in the previous 6 months. A P value≤.05 was considered significant. R version 3.3 software (R Foundation for Statistical Computing, Vienna, Austria) was used for statistical analyses.
RESULTS
Incidence
Over the 30-month study period, there were 11,310 admissions to the 4 included hospital units during which the patient had an inpatient operation. In 52 cases, the patient had a positive C. difficile polymerase chain reaction (PCR) test recorded more than 72 hours after admission and within 12 weeks of discharge. The incidence rate of hospital-acquired CDI was 0.80 cases per 1,000 patient days. Comparing specialties, trauma surgery had the highest rate of CDI at 9.5 cases per 1,000 admissions (Table 1).
TABLE 1 Incidence of Clostridium difficile infection (CDI) by Hospital Unit
NOTE. OR, odds ratio; CI, confidence interval.
a Cases per 1,000 admissions.
Risk Factor Analysis
From univariable analysis, the following variables were associated with statistically significant increased odds of CDI: number of admissions in past year (OR, 2.25; P<.001), ASA classification of 4–5 versus 1–2 (OR, 10.45; P<.001), antibiotic use in the previous 6 months (OR, 2.80; P=.002), proton pump inhibitor or H2 receptor blocker use in the previous 6 months (OR, 2.14; P=.026), antibiotics continued more than 24 hours postoperatively (OR, 5.44; P<.001), and nonperioperative antibiotics (OR, 3.59; P<.001) (Table 2).
TABLE 2 Risk Assessment of Subject and Procedure Characteristics and OutcomesFootnote a
NOTE. SD, standard deviation; IQR, interquartile ratio; OR, odds ratio; CI, confidence interval; BMI, body mass index, CCI, Charlson comorbidity index, ASA, American Society of Anesthesiologists; CKD, chronic kidney disease; PPI, proton pump inhibitor; SNF, skilled nursing facility.
a Data reported as mean (SD), median (IQR) or OR (95% CI).
b Multivariable analysis controls for BMI, ASA classification, and previous antibiotic use in the past 6 months.
c ASA classification: 1, a normal healthy patient; 2, a patient with mild systemic disease; 3, a patient with severe systemic disease; 4, a patient with severe systemic disease that is a constant threat to life; 5, a moribund patient who is not expected to survive without the operation.
Controlling for BMI, ASA classification, and antibiotic use in the previous 6 months, the following variables were associated with statistically significant increased odds of CDI by multivariable analysis: number of admissions in past year (OR, 2.33; P=.001), ASA class of 4–5 versus 1–2 (OR, 15.39; P<.001), antibiotic use in the previous 6 months (OR, 3.74; P<.001), antibiotics continued more than 24 hours postoperatively (OR, 3.38; P=.026), and nonperioperative antibiotics (OR, 2.20; P=.038).
Outcomes
Length of stay during the postoperative admission was significantly increased on univariable analysis (OR, 1.16; P<.001) as well as multivariable analysis (OR, 1.14; P<.001) (Table 2). Length of stay was evaluated as a continuous variable, and for every additional admission in the previous year, odds of CDI increased by 14%. Discharge to skilled nursing facility versus home was also more likely for patients with CDI on univariable analysis (OR, 2.61; P=.020).
DISCUSSION
Our study examined incidence of and risk factors for CDI across multiple surgical specialties in one tertiary-care center. We found rates 4–5 times higher in the trauma and general surgery units than in the orthopedic and neurosurgery units. This variability may be attributable to differences in patient factors such as overall health, antibiotic use, or healthcare setting exposure as well as surgical factors (eg, intra-abdominal trauma surgeries versus orthopedic joint surgeries).
We found that having an ASA class of 4 or 5 increases the odds of CDI 15-fold compared to an ASA class of 1 or 2. This suggests that the complexity of a patient’s illness may increase the risk of CDI. Interestingly, the complexity of the surgery itself, as represented by operative time and blood loss, does not correlate with increased odds of CDI.
If the perioperative antibiotic is continued more than 24 hours after surgery, outside of the perioperative window, the odds of CDI increase 3.34-fold. This finding has been reported previously and highlights the importance of minimizing antibiotic use to only the perioperative period if possible.Reference Shah, Pass, Cox, Lanham and Arnold 8 Use of antibiotics while in the hospital other than the perioperative antibiotic was also associated with 2.2 times greater odds of CDI. Antibiotic exposure as long as 6 months prior to surgery increases the odds of CDI more than 3-fold. Although the surgeon cannot change the fact that their patient received antibiotics, it should alert him or her to the fact that the patient may be at higher risk for postoperative CDI.
Previous studies have shown a correlation between CDI and hospital admission in the previous 3 months.Reference Zacharioudakis, Zervou, Pliakos, Ziakas and Mylonakis 9 Our study reports that this association extends to 12 months. We found that the number of hospital admissions in the past year increases the odds of CDI by 133% for each admission. Given the extensive environmental contamination with C. difficile in healthcare settings, this is a biologically plausible finding.
This study has several limitations. First, the hospital database only tracks patients within the hospital system. Therefore, CDI diagnoses, admissions, and medications from other facilities were not tracked. This study has a relatively small sample size of 52 cases, which limits the power of our statistical analysis. Further, this case-control study is subject to the limitations of all case-control studies.Reference Lewallen and Courtright 10
In conclusion, we found a variable incidence of CDI among surgical specialties at a single tertiary-care hospital. Our analysis suggests that higher ASA class, pre- or postoperative antibiotic use (outside of the perioperative period), and more admissions in the past year are associated with increased odds of CDI. Future studies should examine antibiotic stewardship efforts in the surgical population for reducing CDI.
ACKNOWLEDGMENTS
Financial support. None reported.
Conflicts of interest. All authors report no conflicts of interest relevant to this article.
