Standard Versus Prolonged Antibiotic Prophylaxis After Pancreatoduodenectomy (SPARROW)

Rationale: The additional value of pre-emptive antibiotic treatment after pancreatoduodenectomy is undetermined as previous research reported conflicting results regarding infectious complications. Prolonged antibiotic prophylaxis (formally pre-emptive antibiotic treatment) after pancreatoduodenectomy might reduce the rate of surgical site infections in patients with a high risk for contaminated bile (predominantly patients with preoperative biliary drainage or an ampullary malignancy). Current national and international guidelines lack clear recommendations regarding pre-emptive antibiotic treatment leading to substantially varying antibiotic prophylactic regimes between institutes.

Objective: This trial evaluates the additional value of pre-emptive antibiotic treatment on clinically relevant organ/space surgical site infections (OSIs) in patients undergoing pancreatoduodenectomy with a high risk for contaminated bile.

Study design: This multicenter, randomized controlled, superiority trial compares perioperative versus pre-emptive antibiotic treatment during five postoperative days after pancreatoduodenectomy in patients with a high risk for contaminated bile.

Study population: Adult patients undergoing pancreatoduodenectomy with a high risk for contaminated bile (patients with preoperative biliary drainage or an ampullary malignancy). Patients with a contraindication for the study antibiotics or a preoperative indication for antibiotics (e.g. cholangitis of preoperative abscesses) are excluded.

Intervention: Participants will be randomized to either perioperative prophylaxis (cefazolin, metronidazole and a single dose of 5-7mg/kg gentamicin, control arm) or additional cefuroxime and metronidazole for five postoperative days (experimental arm).

Main study endpoints: The primary endpoint are organ/space infections (OSIs) within 90 days after surgery requiring a therapeutic intervention. Secondary endpoints are OSIs, isolated OSIs, wound infections, postoperative pancreatic fistula, bile or enteric anastomotic leakage, post pancreatectomy hemorrhage, delayed gastric emptying, bacteremia, Clostridium difficile infection, major morbidity (Clavien-Dindo ≥III), reintervention, ICU admission, length of hospital stay, readmission, and in-hospital and 90-day mortality. Besides, switch of postoperative antibiotics, antibiotic sensitivity patterns and concordance between perioperative bile and postoperative surgical site cultures are analyzed.

Sample size: The sample size is calculated for superiority to achieve an OSI difference of 15% (40% vs 25%). With a 80% power (1-β) and a two-sided significance level (α) of 5.0%, a sample of 304 evaluable patients is required for superiority. Assuming a 3% non-resection rate due to metastatic disease and a 3% loss-of-follow-up rate, an expected number of 322 included patients are needed to reach the sample size of 304 evaluable patients to demonstrate superiority for the intervention.