Introduction
Cefazolin is largely the antibiotic of choice in surgical infection prophylaxis due to its tolerable side effective profile and potent activity against Gram-positive bacteria and some enteric Gram-negative bacilli that are common in most surgical procedures. Reference Bratzler, Dellinger and Olsen1,Reference Calderwood, Anderson and Bratzler2 Patients with a large body habitus or obesity represent a high-risk population for surgical site infection (SSI). Reference Meijs, Koek, Vos, Geerlings, Vogely and de Greeff3 Due to concerns of adequate cefazolin tissue penetration in obesity, collective surgical infection prophylaxis guidelines endorse cefazolin 3 g as surgical prophylaxis in patients weighing ≥120 kg even though there are limited outcomes data to support this practice. Reference Bratzler, Dellinger and Olsen1,Reference Ban, Minei and Laronga4 This recommendation is based on scant pharmacokinetic obesity data as well as the low cost and favorable safety profile of cefazolin, but few studies have evaluated the impact of cefazolin dose on SSI in this population.
The pharmacokinetic rationale for cefazolin 3 g doses in obese patients primarily stems from a single study that evaluated cefazolin exposures in gastric bypass patients with BMI ≥40 who received a 2 g preoperative dose with redosing at 3 h. Reference Edmiston, Krepel and Kelly5 The authors concluded that cefazolin 2 g exposures were insufficient based upon a selected pharmacokinetic target of ≥32 mcg/mL; however, mean intraoperative serum concentrations prior to their 3-h redosing interval ranged from 17.1 to 24.3 mcg/mL, exceeding the current Clinical Laboratory Standards Institute (CLSI) susceptibility breakpoint of 2 mcg/mL for methicillin-susceptible Staphylococcus aureus and Gram-negative Enterobacterales, including Escherichia coli. Reference Edmiston, Krepel and Kelly5,6 These data suggest that, in fact, cefazolin 2 g dosing results in therapeutic exposures throughout the duration of surgical procedure in obese patients and that 3 g dosing may not offer additional benefit.
National guidelines acknowledge the uncertain evidence supporting cefazolin 3 g doses, stating “Although data are conflicting regarding the role of 3 g of cefazolin dosing in reducing SSI in obese patients, multiple studies have shown a benefit compared to 2 g dosing in this patient population, with few adverse events from a single dose of 3 g versus 2 g of cefazolin.” Reference Calderwood, Anderson and Bratzler2 This statement does not address the logistical impact of implementing this practice, which increases the complexity of care by requiring administration of separate 1 g and 2 g premixed bags for injection or pharmacy compounding to create a 3 g dose. The study objective was to evaluate the impact of cefazolin dose on postoperative SSI among patients ≥120 kg who underwent elective colorectal surgery.
Methods
Study design
A multicenter, retrospective cohort study was performed utilizing data from a previously described large, validated database of elective colorectal surgeries in Michigan, the Michigan Surgical Quality Collaborative (MSQC). Reference Kuriakose, Vu and Karmakar7–10 The cohorts were defined as (i) patients who received perioperative cefazolin 3 g with metronidazole 0.5 g and (ii) patients who received perioperative cefazolin 2 g with metronidazole 0.5 g.
Setting
The MSQC represents a partnership among 73 Michigan hospitals with a robust data infrastructure and flexible platform for the promulgation of best practices in surgical quality improvement. 10 The MSQC represents 90% of eligible hospitals and is represented by community hospitals and academic centers in urban and rural locations. Data input into the MSQC database are prospectively abstracted by site-specific trained clinical data abstractors following strict guidelines and with responsibility for the accuracy of submitted data. The MSQC monitors site data collection through multiple means to assure data reliability. These include inter-rater reliability audits, abstractor certification, and continuing education. Strengthening the Reporting of Observational Studies in Epidemiology reporting guidelines were followed, and the Trinity Health Ann Arbor institutional review board approved this study.
Participants
Data queries from the MSQC database were obtained and reviewed to meet study parameters. Adult patients who received elective colorectal surgery between July 2012 and June 2021 were included (Supplement Table 1). Patients with body weight <120 kg, who received emergent or urgent surgeries, had missing follow-up and antibiotic information, had class I (clean) and class IV (dirty) wound classifications, received a small bowel resection, or did not receive either cefazolin 3 g or 2 g with metronidazole 0.5 g for surgical prophylaxis were excluded. Variables analyzed included age, gender, race, body mass index (BMI), weight, height, diabetes mellitus, smoking status, corticosteroid use, hypertension, dialysis status, wound classification, American Society of Anesthesiologist (ASA) classification, procedural groupings and approach, receipt of mechanical and oral bowel preparation (MOABP), and surgery duration (Supplement Table 2). Analyses of surgery duration excluded procedures <1 h and >8 h in duration.
Outcomes and definitions
The primary outcome was SSI, which is defined as an infection diagnosed within 30 days following the principal operative procedure. SSI was categorized into deep, organ space, or superficial SSIs in accordance with international guidelines (Supplement Table 3) and compared between cefazolin 3 g and 2 g dosing cohorts. Secondary outcomes included the proportion of patients who developed postoperative Clostridioides difficile infection (CDI) 30 days post-procedure.
Cumulative cefazolin doses between 3.5 h before incision and up to 1 h after were analyzed to determine 2 g versus 3 g cohorts. Recommended timing was defined as prophylaxis of all doses of cefazolin and metronidazole initiated within 1 h prior to incision. Post-incision doses were noted. Secondary analyses included subgroups of only patients with available MOABP data, patients receiving MOABP, patients receiving both MOABP and recommended timing, and by procedure group.
Statistical methods
Descriptive statistics (ie, proportion [%], median [IQR]) were used to describe the patient cohorts. Bivariate analyses were performed to compare categorical variables using Pearson’s χ2 or Fisher’s exact tests. Continuous variables were compared using the unpaired 2-tailed t tests and the Mann–Whitney test analyzed continuous outcomes.
To determine variables independently associated with 30-day SSI, variables associated with the outcome from bivariate analysis (P < .2) were entered into a multivariable logistic regression model. The primary exposure variable of interest was a receipt of cefazolin 3 g as surgical prophylaxis. Variables included in the model were restricted to an event-to-variable ratio of 10:1; the model fit was assessed using the Hosmer–Lemeshow goodness of fit test. The categorical variables were evaluated for collinearity using Pearson’s χ2 test prior to model inclusion; variable selection was prioritized based on the magnitude of association. Data were analyzed using IBM SPSS Statistics v. 26.0 (IBMCorp, Armonk, NY, USA) and statistical significance was defined as a P ≤ .05.
Results
A total of 35,138 patients met inclusion criteria and were assessed for eligibility. Following exclusions, 581 patients (1.7%) were included (Supplement Figure 1). Of these, 367 (63.1%) patients received cefazolin 3 g plus metronidazole prophylaxis, and 214 (36.8%) received cefazolin 2 g plus metronidazole. Demographic and clinical characteristics of the cohorts are listed in Table 1. Colectomy was the most common colorectal procedure performed (446 [76.7%]), and an open surgical approach was the most frequent (284 [48.8%]). MOABP was given in 253 (83.5%) and 129 (76.3%) cases in the cefazolin 3 g and 2 g cohorts with available MOABP data (P = .057). The only differences between cohorts were significantly fewer patients in the 3 g cohort received recommended timing (324 [88.3%] vs 200 [93.5%], P = .043) and a higher receipt of at least 1 of the prophylaxis antibiotics after incision (22 [6%] vs 5 [2.3%]; P = .043).
Demographics and clinical characteristics of patients who received cefazolin and metronidazole
Note. ASA, American Society of Anesthesiologists; SD standard deviation; MOABP, mechanical and oral antibiotic bowel preparation.
a Data are patients, no. (%) unless otherwise specified.
b Data are summarized as means (SD).
c Analysis performed on 472 procedures (303 3 g prophylaxis, 169 2 g prophylaxis) with available data for receipt of MOABP.
d Analysis performed on 556 (356 3 g prophylaxis, 200 2 g prophylaxis) procedures with available data for surgery duration.
e Data are summarized as median (interquartile range).
A total of 39 patients (6.7%) experienced an SSI (Table 2). There were no differences in overall SSIs (23 [6.3%] vs 16 [7.5%], P = .574) or CDI (2 [0.5%] 3 [1.4%]; P = .363) between the cefazolin 3 g and 2 g cohorts, respectively. There was a nonsignificant higher proportion of patients in the cefazolin 3 g cohort that developed deep SSIs compared to the 2 g cohort (5 [1.4%] vs 0 [0%]; P = .164).
Outcomes by cohort of patients who received cefazolin and metronidazole
a Data are no. (%).
Results of the bivariate analyses and clinical rationale dictated the variables selected for inclusion into the multivariable logistic regression model. When accounting for age, smoking status, and surgical duration, cefazolin 3 g prophylaxis (adjOR, .64; 95% CI, .32–1.29) was not associated with a reduction in SSIs (Table 3). Smoking within the past year and increased procedure duration were associated with increased SSIs.
Variables associated with 30-day surgical site infection
Note. CI, confidence interval; MOABP, mechanical and oral bowel preparation; ASA, American Society of Anesthesiologists.
a Adjusted analysis performed on 556 procedures with 37 surgical site infections.
b The baseline value for wound classification is clean/contaminated procedures.
c The baseline value for ASA classification is an ASA score of 1 and 2.
d The baseline value for the procedural group is colectomy procedures.
e The baseline value for surgical approach is open procedures in the unadjusted analysis.
f Surgical time divided by 15 min.
Results were similar when limited to a subgroup of only the 472 procedures (303 3 g cohort, 169 2 g cohort) with available MOABP data. There were no differences in overall SSIs (21 [6.9%] vs 12 [7.1%], P = .945) between the cefazolin 3 g and 2 g cohorts, respectively. Results were also similar in the subgroup of patients who received MOABP (382, 65.7%). There were 253 (66.2%) patients who received cefazolin 3 g and 129 (33.8%) who received cefazolin 2 g within this subgroup (Supplement Table 4). There were no differences in demographic data or outcomes in this subset including SSIs (Supplement Tables 4, 5). Cefazolin 3 g prophylaxis (adjOR, .94; 95% CI, .38–2.31) was not associated with a reduction in SSIs in this subgroup in the multivariable logistic regression model (Supplement Table 6).
A separate subgroup was performed of patients who received both MOABP and recommended prophylaxis timing, in which there were also no differences in unadjusted SSIs (Supplement Table 7). When analyzed by procedure group, there were no SSI differences between the 3 g and 2 g cohorts in colectomy, other colon procedures, or proctectomy procedures (Supplement Table 8). This included SSIs with the receipt of MOABP and with the receipt of MOABP and receipt of recommended prophylaxis timing in all procedures.
Discussion
This study found that cefazolin 3 g dosing had no impact on reduced SSIs when compared to 2 g dosing in obese patients who received elective colorectal surgery (in addition to metronidazole). Additionally, patients who received cefazolin 3 g were less likely to receive recommended antibiotic administration timing prior to the procedure, which likely represents operational and workflow challenges when utilizing a cefazolin 3 g dose or care disparities in obese patients. Antibiotic administration timeliness prior to incision was not found to be a predictor of SSI among this cohort. Overall, these findings suggest that cefazolin 3 g surgical prophylaxis does not have an added benefit over 2 g dosing within obese populations who receive elective colorectal surgery and that operational challenges may lead to delays in antibiotic administration.
Pharmacokinetic and pharmacodynamic data that have explored cefazolin 3 g dosing in obesity for surgical infection prevention are based on small studies with contradictory results. Additionally, these data do not align with lowered, contemporary breakpoints and typical minimum inhibitory concentrations (MIC) for Gram-positives associated with SSI. Four primary pharmacokinetic studies suggested that cefazolin 2 g doses result in inadequate blood and/or tissue concentrations in obese patients who underwent bariatric or cesarean section procedures. Reference Edmiston, Krepel and Kelly5,Reference Swank, Wing, Nicolau and McNulty11–Reference Pevzner, Swank, Krepel, Wing, Chan and Edmiston13 When considering the time of unbound plasma or interstitial space fluid tissue cefazolin concentrations above the MIC (f T ≥ MIC) between opening and closure of the wound as the target pharmacokinetic/pharmacodynamic index, these studies demonstrated decreased cefazolin 2 g exposures in obese adults, suggesting the need for increased cefazolin dose adjustments. Reference Hussain, Curtain, Mirkazemi, Gadd, Peterson and Zaidi14 A major interpretive limitation to these data is that the exposure to MIC target values explored is inconsistent or irrelevant when considering common organism susceptibility profiles and cefazolin redosing practices with surgical duration. These studies suggest that cefazolin 2 g dosing provides adequate target attainment against most pathogens with actual MIC values seen in clinical practice, with the results of one study Monte Carlo simulation suggesting that 2 g dosing would provide adequate drug exposures in obese patients with organism MIC values from ≤1 to 4 mcg/mL for up to 180 min. Reference Brill, Houwink and Schmidt12 Although a reduction in cefazolin exposure is observed at greater BMI, other pharmacokinetic data suggest cefazolin concentrations at the end of surgery may be sufficient to achieve 4 times the MIC90 of 1 mcg/mL and contemporary CLSI breakpoint of 4 mcg/mL for Staphylococcus aureus. 6,Reference Housman, Sutherland and Nicolau15–Reference Stitely, Sweet and Slain19 In the context of patients with more resistant flora and higher MIC values, one could argue that they should instead receive an alternative antibiotic altogether rather than larger cefazolin doses.
In addition to the limited pharmacokinetic/pharmacodynamic data discussed, the observational studies that have evaluated cefazolin prophylaxis dosing on SSI in obese patients are variable and require specific interpretation. Morris and colleagues performed a retrospective cohort study of patients who received cefazolin prophylaxis for hip or knee arthroplasty to assess the impact of cefazolin dose on SSI. Reference Morris, Roberts, Grae and Frampton20 Patient data were extracted from a validated New Zealand surgical quality improvement database and investigators defined optimal cefazolin dose by patient weight strata: cefazolin 1 g in <80 kg, cefazolin 2 g from 80 to 120 kg, and cefazolin 3 g or more grams in ≥120 kg. Of the 2,011 patients with a weight ≥120 kg, just 27.2% received cefazolin 3 g dosing; the authors concluded that patients who were underdosed were more likely to develop SSI (adjOR 2.19; 95% CI, 1.61–2.99) and that 3 g dosing should be used in obesity. Reference Morris, Roberts, Grae and Frampton20,Reference Forse, Karam, MacLean and Christou21 However, the authors also found no differences in the proportions of SSI types in patients who were underdosed versus those who received recommended doses (P = .64, data not shown per authors). The analysis also included patients <120 kg who received suboptimal dosing, and these results ultimately reflect a small proportion of obese patients who received optimal dosing. Rondon and colleagues performed a retrospective cohort study of obese patients undergoing total knee arthroplasty and found that 95.9% of patients ≥120 kg received cefazolin 2 g or less for surgical prophylaxis. Reference Rondon, Kheir, Tan, Shohat, Greenky and Parvizi22 The investigators concluded that underdosed patients more commonly experience periprosthetic joint infections compared with adequately dosed patients at 1 year (1.51% vs .86%; P = .002); however, approximately two-thirds of patients classified as underdosed were <120 kg. Reference Rondon, Kheir, Tan, Shohat, Greenky and Parvizi22 Additionally, since microbiology data are not reported, it is unclear if optimal cefazolin dosing would have prevented SSI or if these patients needed alternative therapies with activity against methicillin-resistant organisms. Two additional observational studies have determined that cefazolin 3 g dosing had no impact on SSI when compared to 2 g dosing in obese patients receiving elective surgeries, although they may have been underpowered. Reference Hussain, Curtain, Mirkazemi, Gadd, Peterson and Zaidi14,Reference Peppard, Eberle, Kugler, Mabrey and Weigelt23
Although this study provides useful information related to cefazolin prophylaxis dose on SSI in obese populations, there are several limitations that should be addressed when interpreting the results. This study was a retrospective, observational analysis that has inherent limitations to its design but is representative of a large, multicenter, prospectively validated database with extensive data from one surgical specialty and from one state in the United States. The primary limitation is that this study was underpowered to detect SSI and CDI differences given the incidence of SSI was relatively low; however, the magnitude of SSI difference between cefazolin 3 g and 2 g groups was approximately 1% and with unclear clinical significance. In addition, patient data were extracted from the medical record and evaluated using manual chart review with clear, objective definitions and represents a study strength when compared to other large database studies prone to misclassification. Despite this, there are likely several variables not included that may have impacted SSI results, including glycemic control and normothermia. Intraoperative and postoperative redosing were also not included; however, the population median surgical duration was significantly less than the 4-h recommended intraoperative dosing time for cefazolin and suggests intraoperative dosing would be unnecessary for most cases, whereas postoperative dosing is not recommended in elective colorectal cases. Reference Bratzler, Dellinger and Olsen1 These findings suggest that cefazolin 3 g dosing may not provide added benefit in elective colorectal surgery, but cefazolin prophylaxis dose has not been evaluated to the same capacity in other surgical procedures. However, colorectal surgery represents a higher SSI risk when compared to others. The authors feel these data can be safely generalized to other surgical procedure types.
In conclusion, cefazolin 3 g in combination with metronidazole surgical prophylaxis was not associated with reduced SSI in patients weighing ≥120 kg in elective colorectal surgery procedures; however, this study was underpowered to detect a difference in SSI. Elective colorectal surgery patients may be best served by receiving evidence-based care, including MOABP combined with cefazolin 2 g dosing with metronidazole that is appropriately timed pre-operatively. Future evaluations could leverage large datasets to compare cefazolin 3 g and 2 g dosing strategies, including other surgical procedures and urgent/emergent cases, and may also include assessments of intraoperative and postoperative redosing.
Supplementary material
The supplementary material for this article can be found at https://doi.org/10.1017/ice.2024.215.
Acknowledgments
None.
Financial support
None reported.
Competing interests
Dr Cleary reported receiving honoraria from Intuitive Surgical, Inc. outside the submitted work. No other authors report potential conflicts of interest.
Disclosure
This work was presented in part as a poster at the 2023 IDWeek meeting in Boston, MA, USA.
