Evaluation of Timing of Antimicrobial Surgical Prophylaxis on Rates of Surgical Site Infections

Introduction

Surgical site infections (SSIs) are among the most common healthcare-associated infections, with an incidence rate of ∼0.5–3.0% in the United States over the past decade.^1,2 Patients who incur a SSI are more likely to have a longer hospital length of stay, contributing to an estimated US Health care cost of 3.5–10 billion dollars annually.^3–5 In addition, patients that develop a SSI have nearly a twofold higher risk of mortality compared with patients without postoperative SSIs.^6,7 As such, SSIs remain a major cause of morbidity, mortality, and increased health care costs.

Current guidelines from the American College of Surgeons, Surgical Infection Society, and American Society of Health Systems Pharmacists (ASHP) recommend that antimicrobial prophylaxis (AP) be administered within one hour prior to incision, or within two hours for vancomycin or fluoroquinolones because of longer infusion times. ⁸ Successful prophylaxis requires the delivery of the antimicrobial to the operative site before contamination occurs. Thus, the antimicrobial agent should be administered at such a time to provide serum and tissue concentrations exceeding the minimum inhibitory concentration for the probable organisms associated with that procedure. Although national guidelines clearly support administration of AP within 60 min prior to incision to reduce the risk of SSI, some studies suggest that despite appropriately timed administration of antibiotics, infection rates still vary among intervals within the 60-min window.^9–11 To date, there are limited and conflicting data as to whether there is a “most optimal” time for administration within the 60-min window. This study assessed the impact of timing of antimicrobial administration on rates of SSIs within a large health care system.

Patients and Methods

This was a multicenter, retrospective study of admitted patients that underwent an abdominal hysterectomy (HYST), colorectal surgery (COLO), or craniotomy (CRANI) between January 1, 2021 and February 28, 2021. Definitions of operative procedures used for this study are in accordance with those outlined by the National Healthcare Safety Network. COLO was defined as incision, resection, or anastomosis of the large intestine, including large-to-small and small-to-large bowel anastomosis, and did not include rectal operations. HYST was defined as removal of the uterus through an abdominal incision. CRANI was defined as incision through the skull to excise, repair, or explore the brain and did not include taps or punctures.

Patients were included in the analysis if they were at least 18 years of age and received AP within 60 min of incision. Patients were excluded if they had a documented preexisting infection prior to the first incision, required emergency surgery such as in the event of trauma, or developed a SSI secondary to endemic fungi such as Blastomyces, Histoplasma, Coccidiodes, Paracoccidioides, Cryptococcus, and/or Pneumocystis.

Utilizing the electronic medical record, patients were identified via a standard SSI surveillance report generated for each procedure type. Reports included patient demographics, admission and procedure date, ICD-10 procedural code, and National Healthcare Safety Network (NHSN) wound classification (i.e., clean, clean-contaminated, dirty). ¹² Additional electronic chart review was conducted to collect information on past medical history, antibiotic allergies, antibiotic therapy, timing of administration of preoperative AP, duration of perioperative antibiotics, duration of surgical procedure, length of inpatient stay, and incidence of in-hospital mortality.

The primary outcome, incidence of SSI, was compared between patients who received AP 0–30 min versus 31–60 min prior to incision. AP timing was on the basis of the end of the infusion of antibiotics, with the default administration of prophylactic cefazolin as an IV push over 3 min. The recommended NHSN SSI surveillance periods of 30 days post COLO or HYST and 90 days post CRANI were used when evaluating for incidence. SSIs were further classified per NHSN as superficial incisional, deep incisional, or organ space. ¹² Secondary outcomes included appropriateness of antimicrobial selection, dosing, and duration, length of stay (LOS), in-hospital mortality, and 30-day readmissions. A predefined subgroup analysis evaluated the same primary and secondary outcomes for each 15-min interval (0–15 min, 16–30 min, 31–45 min, and 46–60 min).

A power and sample size analysis estimated that a total study population of 300, with individual group sample sizes of 250 (AP 0–30 min) versus 50 (AP 31–60) patients would afford 80% power to detect a 12.5% difference in SSI rates. Continuous data for the primary analysis were compared using a Student’s t-test if normally distributed or a Mann-Whitney U test if non-normally distributed. Continuous data for the subgroup analysis were compared using an analysis of variance if normally distributed, or a Kruskal-Wallis test if non-normally distributed. Categorical data were evaluated with a Pearson chi-square test or Fisher exact test where appropriate. All analyses were conducted using SPSS v. 26 (IBM, Armonk, NY 2019), using an alpha level of 0.05 such that results yielding p < 0.05 were considered statistically significant.

Results

A total of 517 patient charts were reviewed for eligibility, of which 277 were included in the primary and subgroup analyses. Figure 1 shows reasons for patient exclusion, as well as the number of patients included in each AP timing group for the primary analysis. Baseline demographics were similar between the two groups, with the majority of patients being Caucasian females with a mean age of 57 years and no statistically significant differences in comorbidities. However, with respect to procedural characteristics, there were significantly more patients who received cefazolin (91.4% vs. 77.3%, p = 0.036) and underwent an abdominal hysterectomy (42.1% vs. 22.7%, p = 0.048) in the 0–30-min group, whereas patients in the 31–60-min group were more likely to have undergone colorectal surgery (45.5% vs. 36.5%, p = 0.048) or craniotomy (21.5% vs. 31.8%, p = 0.048). Surgical wound classifications (26.0% vs. 31.8% with “clean” wound; 71.9% versus 65.9% with “clean/contaminated” wound, p = 0.720) and operative duration (3.67 vs. 3.75 h, p = 0.772) did not differ between the two groups (Table 1).

OPEN IN VIEWER

FIG. 1.

CONSORT diagram for study. *Duplicate patients: multiple surgeries within the study time period. AP: antimicrobial prophylaxis.

Overall, SSIs occurred in 6.0% (14/233) versus 4.5% (2/44) patients in the 0–30-min and 31–60-min groups, respectively (p = 0.703). In-hospital mortality (1.7% vs. 2.3%, p = 0.582), median LOS (3 vs. 4 d, p = 0.089), and 30-day readmission rates (11.8% vs. 13.6%, p = 0.731) were similar between the groups (Table 3).

A subgroup analysis was conducted, evaluating the same primary and secondary outcomes between groups of 15-min intervals (0–15, 16–30, 31–45, and 46–60 min prior to incision). The duration of operative procedure was significantly prolonged in the 46–60-min group (3.42 vs. 4.03 vs. 3.50 vs. 4.55 h, respectively; p = 0.006), and cefazolin was used more commonly in the 0–15- and 16–30-min groups (91.2% vs. 91.6% vs. 76.5% vs. 81.8%, p = 0.022) (Table 2). There were no differences in incidence of SSIs in patients who received AP 0–15, 16–30, 31–45, or 46–60 min prior to incision (4.4% vs. 8.4% vs. 2.9% vs. 9.1%, p = 0.487). Upon analysis of secondary outcomes, there were also no significant differences in median LOS (3 vs. 3 vs. 3 vs. 6 d, p = 0.093), in-hospital mortality (2.9% vs. 0.0% vs. 2.9% vs. 0.0%, p = 0.368), or readmission within 30 days (13.9% vs. 8.4% vs. 11.8% vs. 18.2%, p = 0.558) between the four subgroups (Table 3).

Sixteen patients in this study incurred a SSI. Eight had cultures taken, of which five were positive for bacterial growth. Organisms were isolated from surgical wounds (3/5), blood and cerebrospinal fluid (CSF) (1/5), and bone (1/5). Three cultures (60%) identified bacteria that were found to be resistant to the antibiotic used for preoperative AP for CRANIs, which was cefazolin for all three cases (Table 4).

Discussion

In this retrospective analysis, no significant difference in incidence of SSIs between patients receiving AP 0–30 versus 31–60 min prior to incision was observed. It is possible that no difference was found because of unbalanced sample sizes, the study not meeting its predefined power parameters, and/or differences in baseline characteristics (e.g., procedure type and antibiotic used). Existing literature has investigated similar outcomes with various surgical populations with findings both conflicting to and supporting ours. Steinberg et al. evaluated patients undergoing cardiac surgery, arthroplastic hip and knee procedures, or hysterectomy, ultimately finding a trend that patients had a lower incidence of SSIs if they received AP with short-acting cephalosporins or antibiotics with short infusion times 0–30 min before incision compared with those who received AP 31–60 min prior to incision, although the findings were not statistically significant. ⁹ Conversely, Weber et al. evaluated timing of AP on incidence of SSI in patients undergoing a visceral, trauma, or vascular procedure and found a statistically lower incidence of SSI in patients who were administered AP 30–59 min prior to incision compared with those who received AP 0–29 min prior to incision when using cefuroxime as AP. ¹⁰ Lastly, a cohort study conducted by de Jonge et al. evaluated incidence of SSIs in patients receiving cephalosporins with or without anaerobic coverage where appropriate as AP for general, orthopedic, or gynecological surgery. Utilizing a multivariable logistic regression, the authors found no conclusive evidence on incidence of SSIs between patients who received AP with the aforementioned regimen 0–30 and 31–60 min prior to incision, similar to our findings. ¹³

When comparing study designs of previous literature with this study, the inclusion of specific procedure types significantly varies. The three procedure types included in this study were chosen because they were the most common procedure types in our health system and therefore results would yield pertinent internal data. It is crucial to recognize that each procedure type carries a different risk of postoperative SSIs. Although previous literature has evaluated various types of procedures and AP, the incidence in SSIs ranged from 1.6% to 5.4%. In our study, SSIs ranged from 4.5% to 6.0%, higher than what has previously been reported. This may be because of our study evaluating a significantly smaller population compared with populations included in previous studies.^9,10,13,14 The prevalence of pathogenic microbes may vary depending on the operating room, procedure, and patient-specific factors. The antibiotics used for AP is mostly consistent, with both previous studies and this study having a short-acting cephalosporin used the vast majority of the time, and therefore focusing efforts on mainly the incidence of SSIs on the basis of timing of the short-acting cephalosporin.

The majority of patients (84.1%) received AP within 30 min prior to incision, which may have been strategic and necessary because of time constraints and scheduling issues when trying to complete several other crucial tasks in the operating room. The most common antibiotic used was cefazolin, likely because it is commonly formulated as a syringe, and therefore has the benefit of quicker administration as a 3-min IV push. In addition, having the patient receive AP minutes before incision in the operating room more likely ensures that AP is delivered within the recommended 60-min window.

To our knowledge, this is the first study that includes a direct evaluation of incidence of SSIs on the basis of AP in 15-min intervals within the 60-min pre-incision window. In the predefined subgroup, analysis, the subgroup that received AP 46–60 min prior to incision coincidentally had the highest overall percentage of SSIs and the longest operative duration, although not statistically significant. This finding may be because of longer operative durations being associated with higher risk of SSI. Notably, fewer patients received AP 46–60 min prior to incision relative to other timing groups, which may have also impacted the outcomes of the statistical analysis. Compared with previous literature, Sommerstein et al. found that the administration of AP 10–25 min prior to incision had lower incidence of SSI compared with administration of AP 30–55 min prior to incision when utilizing cefuroxime as AP in patients undergoing hernia repair, knee or hip arthroplasty, cardiac surgery, laminectomy, spondylodesis, colon surgery, cholecystectomy, cesarean delivery, gastric bypass, or hysterectomy. ¹⁴ In our secondary analysis, significant differences were seen in the primary antibiotic used, which limits the comparison as antibiotics with longer administration times and half-lives (e.g., levofloxacin) were the second most commonly used antibiotic.

Of the 16 patients in this study that incurred a SSI, 8 had cultures collected, of which 5 were positive for bacterial growth. According to NHSN, the most common causative organisms for SSIs include S. aureus, coagulase-negative Staphylococcus spp., Enterococcus spp., E. coli, P. aeruginosa, Enterobacter spp., and Klebsiella pneumoniae. ¹⁵ Similar findings were seen in this study with S. epidermidis, E. coli, and E. faecalis isolated in three out of five of the positive cultures, suggesting that the majority of culture-positive SSIs were because of bacteria commonly seen in SSIs. Notably, of the culture-positive SSIs, two out of the five cases also included bacteria that are associated with common hospital-acquired infections (HAIs), specifically S. marcescens and K. aerogenes. ¹⁶ With respect to the documented culture-positive SSIs, the infecting bacteria was most often resistant to preoperative AP. In all three cases of resistant causative bacteria, patients received cefazolin as AP for CRANIs, in accordance with institutional protocol and ASHP national guidelines. Although evaluating a different procedure, a similar finding was found in a study by McCullough et al., which evaluated the efficacy of perioperative AP in patients receiving breast reduction surgery. The study found that when first-generation cephalosporins were used as prophylaxis, SSI organisms showed resistance rates to preoperative cefazolin of 20.5%. ¹⁷ Determining adequate agents for AP will continue to be a challenge when balancing appropriate coverage and preventing the selection of development of pathogenic organisms with broader resistance.

The study has several limitations. The study’s sample size was smaller than the power estimate calculated, and incidence of SSI was lower than anticipated. Thus, the study was underpowered to have detected a statistically significant difference in the primary outcome. In addition, this was a retrospective review, for which there are inherent biases. Certain variables had missing or unknown data (e.g., race and BMI), and in some cases, there was a lack of validation of an SSI by a qualified diagnostician. Lastly, the differences in pharmacokinetics and pharmacodynamics between beta-lactams (e.g., cefazolin) and fluoroquinolones (e.g., levofloxacin) could have potentially influenced outcomes, as antibiotic selection significantly varied among the groups.

Conclusions

Ultimately, the results of the primary endpoint support current national guidelines and clinical practice within our health system. Future investigation of a different interval (i.e., AP 15–45 min prior to incision) may be beneficial on the basis of pharmacokinetics of routinely prescribed antimicrobials for surgical prophylaxis. Subgroup analyses on the basis of antibiotic used for AP and type of procedure may support the use of specific antibiotics and timing of AP for the procedure being performed.