A procalcitonin-based algorithm to guide antibiotic use in patients with acute pancreatitis (PROCAP): a single-centre, patient-blinded, randomised controlled trial

Abstract

Background: Differentiating inflammation from bacterial infection in patients with acute pancreatitis can be difficult. Procalcitonin can distinguish infection from inflammation, and algorithms based on procalcitonin measurement can differentiate bacterial sepsis from a systemic inflammatory response. We aimed to test the hypothesis that a procalcitonin-based algorithm to guide initiation, continuation, and discontinuation of antibiotics could lead to reduced antibiotic use without an adverse effect on outcome in acute pancreatitis.

Methods: PROCAP was a single-centre, patient-blinded, randomised controlled trial done at the Manchester Royal Infirmary (Manchester, UK). Eligible participants were aged 18 years or older and had a clinical diagnosis of acute pancreatitis. Participants were randomly assigned (1:1) to procalcitonin-guided care or usual care using web-based randomisation software. The randomisation sequence was stratified by disease severity and admission pathway, using variable block sizes of 4, 6, or 8. Patients, but not clinicians, were masked to group assignment. In the procalcitonin-guided care group, procalcitonin testing was conducted on days 0, 4, 7, and weekly thereafter. Guidance was to stop or not start antibiotics following a test value of less than 1·0 ng/mL and to start or continue antibiotics following a test value of 1·0 ng/mL or more. In the intervention group, any empirical clinical decision to use antibiotics was preceded by measurement of procalcitonin. Otherwise, both groups received standard care. The primary outcome was use of antibiotics during the index admission to hospital. All analyses were done in the intention-to-treat population. This study was registered with the International Standard Randomised Controlled Trial registry, ISRCTN 50584992.

Findings: Between July 29, 2018, and Nov 13, 2020, 369 patients were screened, of whom 260 were enrolled and randomly assigned to a treatment group (132 to procalcitonin-guided care and 128 to usual care). 59 (45%) of patients in the procalcitonin-guided care group were prescribed antibiotics compared with 79 (63%) in the usual care group (adjusted risk difference -15·6% [95% CI -27·0 to -4·2]; p=0·0071). The odds ratio for the treatment effect was 0·49 (95% CI 0·29 to 0·83; p=0·0077). There was no significant difference between groups in terms of the number of clinical infections or hospital-acquired infections per patient. Four (3%) patients in the procalcitonin-guided care group and three (2%) patients in the usual care group died; all deaths were related to underlying severe pancreatitis. There was no difference in adverse events between the groups.

Interpretation: Our findings suggest that procalcitonin-guided care can reduce antibiotic use without increasing infection or harm in patients with acute pancreatitis. Procalcitonin-based algorithms to guide antibiotic use should be considered in the care of this group of patients and be incorporated into future guidelines on the management of acute pancreatitis.

Funding: None.