Surgical site infection rate is higher following hip and knee arthroplasty when cefazolin is underdosed

Abstract

While many guidelines recommend higher doses of cefazolin for patients with higher body weights, there are scant outcome data showing the benefit of higher doses. Surgical site infection (SSI) rates by dose of cefazolin used for surgical prophylaxis after hip or knee arthroplasty were analyzed. Analysis of patient data entered into New Zealand's national, prospective, surveillance and quality improvement SSI Improvement Programme database for the period July 2013 through December 2017 was conducted. The US Centers for Disease Control and Prevention's National Healthcare Safety Network SSI definitions were used, and patients were followed for 90 days after surgery. Underdosing was defined as use of 1 g of cefazolin in patients weighing 80 kg or more or a cefazolin dose of <3 g in those weighing 120 kg or more. There were 38,288 procedures where cefazolin was used for prophylaxis; patient body weight was known for all these procedures. Of the 1,840 patients who received 1 g of cefazolin, 676 (37%) weighed 80 kg or more. Of the 2,011 patients weighing 120 kg or more, 1,464 (73%) were underdosed. After multivariable analysis, male gender, higher total surgical risk scores, performance of revision and hip arthroplasties, and cefazolin underdosing were associated with higher SSI rates. For the 2,106 underdosed patients, the odds ratio for SSI was 2.19 (95% confidence interval, 1.61-2.99; P < 0.0001). The number of higher-weight patients needed to treat to prevent 1 SSI was 83, with an estimated cost of <NZ$500 to prevent 1 infection costing an estimated NZ$40,000. Patients undergoing hip or knee arthroplasty and with weights of ≥80 kg and those with weights of ≥120 kg should receive cefazolin doses of 2 g and ≥3 g, respectively, for SSI prophylaxis. The question of whether a dose of ≥4 g is needed in patients weighing 120 kg or more or who are above a given body mass index threshold (eg, >35 kg/m2 or >40 kg/m2) remains unanswered.