Moving Closer to Zero: Effect of “Just in Time” Dosing of Prophylactic Antibiotics on Surgical Site Infections

Abstract

Background

A local quality initiative to improve compliance with surgical antibiotic prophylaxis measures resulted in a high percentage of patients receiving antibiotics within minutes of surgical incision. Studies examining the association between timing of prophylaxis and the risk for surgical site infection (SSI) have produced heterogeneous results.

Objective

To examine risk factors for SSI, including “just in time” dosing of antibiotic prophylaxis (dose administered within 5 minutes of incision).

Methods

This was a retrospective matched case-control study. Case patients developed SSI in the 30 days following a clean or clean-contaminated surgical procedure. Control patients did not develop SSI following similar procedures and were matched to ensure comparable baseline risk. We assessed the rate of guideline-compliant antibiotic prophylaxis and calculated odds ratios (ORs) to determine the association of patient covariates with the risk for SSI.

Results

Forty case patients and 104 controls were included in the study. The rate of appropriate prophylaxis was high in both groups (98% and 94% for case and control groups, respectively). Approximately 15% of case and control patients received antibiotic prophylaxis within 5 minutes of incision, thus, “just in time” dosing did not appear to increase the risk for SSI (OR, 0.814; 95% CI, 0.274–2.415). There was a nonsignificant association between receipt of vancomycin and SSI (OR, 2.844; 95% CI, 0.926–8.737).

Conclusion

“Just in time” dosing of prophylactic antibiotics was not associated with increased risk for SSI. Further study is needed to clarify the impact of antibiotic choice on the risk for subsequent SSI.

Surgical site infection (SSI) is the most common and economically burdensome health care–associated infection in the United States, costing the health care system up to $3.3 billion annually.^1–3 By National Healthcare Safety Network (NHSN) definitions, SSI can range from superficial to deep tissue or organ infection arising between 30 and 90 days following select surgical procedures. ⁴ Recognizing the significant morbidity and costs associated with SSI, clinical guidelines and quality improvement measures have been developed to assist institutions in reducing SSI rates.^1,4–6 The NHSN outlines SSI surveillance approaches and facilitates risk-adjusted comparisons of SSI rates through the NHSN risk index. The Surgical Care Improvement Project (SCIP) provides a set of core measures with an emphasis on the use of perioperative antibiotics to prevent SSI. ⁶ SCIP antibiotic measures address the appropriate selection, dosing, timing, and duration of antibiotic prophylaxis.

SSI-reduction initiatives should be closely evaluated for intended and unintended consequences. We implemented a variety of measures to reduce the incidence of SSI at our institution over several years, including practice changes to improve compliance with timing of antibiotic prophylaxis measures. Incorporated into the process change was a shift in prophylactic antibiotic administration from the preoperative holding area to the operative suite, with administration coordinated by the anesthesiology service. This resulted in near 100% compliance with timely administration, but some patients received prophylactic antibiotics, including vancomycin, just minutes before surgical incision. To ensure our SSI-prevention program did not have an unintended adverse effect, we evaluated the relationship between SSI risk and the timing of antibiotic prophylaxis, including administration of antibiotics within 5 minutes of surgical incision (“just in time” dosing).

Methods

This was a retrospective case-control study conducted at University of Florida Health Shands Hospital, a 939-bed academic medical center in Gainesville, Florida. The study was approved by the local institutional review board.

Study Population

Case patients were adults who developed NHSN-defined SSI following NHSN wound class I or II procedures (clean or clean-contaminated) between May 2011 and July 2012. ⁵ The pool of eligible controls was comprised of patients who did not develop SSI after a class I or II procedure during the study period. Class I procedures involve uninfected operative wounds that do not pass through the respiratory, alimentary, genital, or urinary tracts. In class II procedures, these tracts are entered under controlled conditions without unusual contamination. Case patients were identified through our infection control database and control patients were identified through a query of our institution's decision support services database. Our infection control surveillance software was updated in mid 2011. This dictated the beginning of our study period and the number of cases we were able to identify. Patients were excluded if they were younger than 18 years of age, if antibiotics other than preoperative prophylactic antibiotics were administered in the 24 hours prior to surgery, or if additional surgeries were performed in the 30 days prior to the SSI.

Data Collection and Matching

The electronic medical record (EMR) was used to retrospectively collect demographic, laboratory, and antibiotic therapy data. The times of surgical incision and antibiotic administration were documented by operating room personnel and were available in the EMR. The Charlson Comorbidity Index (CCI) score was calculated by weighting and summing International Classification of Diseases, 9th ed (ICD-9) comorbidities noted at the time of the procedure. Each case patient was matched with up to 3 control patients based upon the following criteria: age within 5 years, identical CCI score, and the presence/absence of each component of the NHSN risk index (operation lasting longer than the 75th percentile of the duration for the operative procedure, wound class III or IV, American Society of Anesthesiologists' Classification of Physical Status score ≥3).

Outcomes and Statistical Analysis

Our main outcome was the association between SSI and initiation of prophylactic antibiotics within 5 minutes prior to surgical incision (“just in time” dosing). We also evaluated guideline compliance with antibiotic choice and timing measures. Timing was considered appropriate if the infusion began within 60 minutes prior to incision or within 120 minutes for fluoroquinolones and vancomycin. We calculated odds ratios (ORs) with 95% confidence intervals to assess the association of patient covariates with the development of SSI. Continuous variables were converted into categorical variables to allow for the calculation of ORs (ie, albumin < 3.5g/dL). Data are described using median and interquartile range or percentages, where appropriate. We analyzed all data with IBM SPSS version 20.0 (IBM, Armonk, NY).

Results

A total of 40 cases and 104 well-matched controls were included in the study. Patient characteristics and risk factors for SSI are shown in Table 1. Orthopedic surgery was the most common procedure in both groups. Analysis of antibiotic therapy revealed high rates of appropriate preoperative antibiotic selection and administration timing in the case and control groups (98% and 94%, respectively). Each case of inappropriate prophylaxis was due to no prophylactic antibiotic administration being documented.

Table 1.

Demographics and clinical characteristics of case and control patients ^a

Variable	Case (n=40)	Control (n=104)	Odds ratio [95% CI]
Demographics
Male sex	26 (55%)	53 (51%)
Age, years	60 (45–67)	59 (45–66)
Weight, kg	87(75–109)	83 (68–102)
BMI	29.2 (25.6–34.5)	28.2 (24.5–32.9)
Comorbidities
ASA class	3 (2–3)	3 (2–3)
Charlson Comorbidity Index	0 (0–3)	0 (0–2)
Weight >100 kg	15 (36%)	27 (26%)	1.711 [0.788–3.7171
BMI >25	34 (85%)	75 (72%)	2.191 [0.832–5.768]
Diabetes	11 (28%)	18 (17%)	1.812 [0.767–4.283]
Malignancy	4 (10%)	14 (13%)	0.714 [0.220–2.316]
Laboratory values
Max glucose, mg/dL	126 (106–181)	140 (107–186)
Glucose >180 mg/dL	10 (25%)	27 (26%)	0.922 [0.379–2.242]
Albumin, g/dL	3.7 (3.1–4.3)	4.2 (3.6–4.5)
Albumin <3.5 g/dL	9 (43%)	13 (22%)	2.712 [0.939–7.829]
Type of surgical procedure
Wound class	1 (1–1)	1 (1–1)
Orthopedic	16 (40%)	50 (48%)	0.673 [0.319–1.421]
Vascular	4 (10%)	17 (16%)	0.569 [0.179–1.807]
Gastrointestinal	6 (15%)	16 (15%)	0.971 [0.351–2.687]
Genitourinary	3 (8%)	8 (8%)	0.973 [0.245–3.868]
CNS	4 (10%)	6 (6%)	1.815 [0.484–6.804]
Breast	4 (10%)	5 (5%)	2.200 [0.560–8.649]
Other	3 (8%)	2 (2%)	4.135 [0.664–25.73]
Procedural duration
Median duration, minutes	182 (140–227)	155 (106–232)
Procedure >2 hours	33 (83%)	73 (70%)	2.002 [0.800–5.011]
Procedure >3 hours	20 (50%)	42 (40%)	1.476 [0.709–3.073]
Antibiotic therapy
Documented prophylaxis a	39 (98%)	98 (94%)	2.388 [0.278–20.48]
Time from incision, minutes	14 (10–25)	15 (8–23)
0–5	5 (13%)	15 (15%)	0.814 [0.274–2.415]
0–10	9 (23%)	28 (29%)	0.750 [0.316–1.780]
0–15	23 (59%)	47 (48%)	1.560 [0.736–3.306]
0–30	33 (83%)	86 (88%)	0.767 [0.321–1.971]
Re-dosing, when indicated	7/9 (78%)	17/23 (74%)	1.235 [0.199–7.673]
β-lactam only	31 (79%)	84 (86%)	0.646 [0.247–1.689]
Vancomycin only	3 (7.7%)	5 (4.8%)	1.550 [0.352–6.824]
β-lactam + Vancomycin	4 (10%)	2 (2.0%)	5.486 [0.962–31.28]
Vancomycin in any combination	7 (18%)	7 (7%)	2.844 [0.926–8.737]

Note: Values provided as median (interquartile range) or n(%). ASA = American Society of Anesthesiologists; BMI = body mass index; CNS = central nervous system.

^aAll administrations within the recommended time frame (within 1 hour prior to incision, or 2 hours for vancomycin or fluoroquinolones).

“Just in time” dosing was not significantly associated with increased odds for developing SSI (OR, 0.814; 95% CI, 0.274–2.415). Further analysis did not identify associations between other dosing time periods and the risk for developing SSI (ie, patients receiving antibiotics within 0–10 minutes, 0–15 minutes, etc). Of interest, over half of all patients received therapy within 15 minutes of incision and approximately 85% of patients in both groups received therapy within 30 minutes of incision, including patients receiving vancomycin (7/11; 64%). There was a nonsignificant association between receipt of vancomycin and SSI (OR, 2.844; 95% CI, 0.926–8.737). Most patients received vancomycin at a dose of 1,000 mg, which equates to a median dose of 11.3 mg/kg (interquartile range [IQR], 9.4–13.6mg/kg). Only 1/11 (9%) vancomycin patients received a dose of greater than or equal to 15 mg/kg. Other covariates unrelated to antibiotic timing were not significantly associated with increased odds for developing SSI.

Discussion

Studies examining the relationship between timing of antibiotic prophylaxis and SSI risk have produced heterogeneous results.^7–11 Antibiotic administration close to the time of incision has been linked to lower risk for SSI ⁹ and higher risk for SSI ¹⁰ ; it was also found to have no significant association with SSI risk. ¹¹ This matched case-control study was designed to determine whether antibiotic dosing just before surgical incision (ie, within 5 minutes) was associated with the risk for SSI. We did not identify timing of prophylactic antibiotics as a significant risk factor in our diverse surgical population.

In a recent report, Hawn et al also failed to identify a meaningful relationship between timing of antibiotic prophylaxis and SSI in a large Veterans Affairs cohort. ¹¹ However, there was a higher occurrence of SSI in orthopedic surgery patients receiving prophylaxis with vancomycin. This finding was further explored in another publication. ¹² In a cohort of 18,830 arthroplasty patients, vancomycin was associated with a higher incidence of SSI when compared to cefazolin (2.3% vs 1.3%) and clindamycin in penicillin-allergic patients (2.0% vs 1.0%).

In our study, patients who developed SSI were 2.8 times more likely to have received prophylactic vancomycin than controls. This interaction did not reach statistical significance due to our small sample size, but it warrants discussion given similar findings by other researchers. Ponce et al hypothesized that vancomycin dosing could be inadequate if patients do not receive does of greater than or equal to 15 mg/kg. ¹² Dosing information was not available in their study. In our population, patients rarely received doses greater than or equal to 15 mg/kg. Thus, our results support the notion that underdosing of vancomycin could contribute to the risk for SSI. A prospective study of cardiac surgery patients receiving vancomycin found the highest rate of SSI when vancomycin was administered 0 to 15 minutes prior to incision. ¹⁰ Roughly half of our patient population received prophylactic therapy within this time frame, a high rate in comparison to previous literature. ¹⁰ Vancomycin prophylaxis may be less efficacious at institutions like ours that administer prophylactic antibiotics in the operating suite.

Our study has several limitations. This was a retrospective study, so we relied on the accuracy of prior coding and documentation. Our sample size was small and we could not completely match our control and case groups in a 3:1 fashion as planned (some case patients only had 2 suitable matches). The matching procedure minimized differences in baseline characteristics, which was desirable since our main focus was the impact of antibiotic timing. However, matching hindered our ability to assess risk factors for SSI other than antibiotic timing. An ongoing randomized control trial (NCT01790529) aims to further clarify the optimal timing of prophylaxis (<30 minutes vs 30–75 minutes prior to incision). ¹³

Quality initiatives to decrease the incidence of SSI have led to high rates of guideline-concordant preoperative antibiotic prophylaxis. There is an intuitive concern that “just in time” antibiotic administration, although administratively acceptable, may not achieve the same prophylactic efficacy. For patients given antibiotics within the recommended time frame, we did not identify timing of antibiotic prophylaxis as a significant risk factor for SSI. Given the high compliance with quality measures, other patient- or drug-related factors are likely to drive SSI rates within our institution and abroad. Larger, prospective studies would be ideal to further explore the interplay between SSI and these variables, including the impact of antibiotic dose and choice (vancomycin-based vs other prophylactic regimens).