Surgical drains do not alter the spectrum of complications in distal pancreatectomy. An analysis of the NSQIP pancreatectomy database

Abstract

Background: Distal pancreatectomy remains the procedure of choice for masses within the body and tail of the pancreas however post-operative pancreatic fistulas (POPF) continue to be the major source of morbidity. While we reasoned that drains may not reduce POPF, some evidence suggests they may reduce the frequency of related severe complications. We sought to evaluate the role of surgical drains in mitigating the frequency of severe complications related to POPF in distal pancreatectomy. Methods: We performed a retrospective cohort study of the procedure targeted, Pancreatectomy NSQIP database from 2014–2016. We included all patients who underwent non-emergent, distal pancreatectomy for any reason. We excluded patients with sepsis, or clinical instability. The primary endpoint was a modified clinically relevant POPF (mCR-POPF). This included any POPF that resulted in the need for percutaneous drainage, or was associated with sepsis, organ failure, reoperation, or death. We also looked at multiple secondary endpoints including percutaneous drainage, sepsis, septic shock, acute renal failure, re-operation, cardiac arrest, organ space infection and mortality. Multivariable logistic regression models were built for the primary and each secondary endpoint. Results: Based on our criteria, we identified 5,055 distal pancreatectomies eligible for analysis. Of these, 4679 had no mCR-POPF, while 376 did. In addition, drains were left in 4334 patients (85.7%). On multivariable logistic regression, surgical drain placement was not associated with mCR-POPF (OR 0.86, 95% CI 0.637–1.161, P = 0.325). Additionally, for each of the secondary endpoints evaluated, there was no evidence that surgical drain placement was associated with either an increase or a decrease in any of the outcomes (Percutaneous drainage OR 1.151, 95% CI 0.875–1.514, p 0.3143, Septic Shock OR 1.240, 95% CI 0.888–1.941, p = 0172, mortality OR 1.503, 95% CI 0.456 – 4.955, p = 0.503). Conclusion: Based on this analysis of the NSQIP procedure targeted pancreatectomy database, we did not find any evidence that surgical drains alter the frequency of mCR-POPF, or with related moderate to severe complications. As a result, the practice of standard drainage in the setting of distal pancreatectomy should be critically evaluated.Table 2Final adjusted odds ratios for the effect of surgical drain placement on each outcome based on multivariable logistic regression models.OutcomesOR95% CIp-valueRisk factors controlled formCR-POPF0.860.631.160.33Age, CHF hx, OP time, BMI, reconstruction, ASAPercutaneous drain1.150.881.510.31COPD hx, CHF hx, Chemo, OP time, BMI, ASA, reconstructionSepsis (including septic shock)1.310.891.940.17Sex, COPD hx, OP time, ASASeptic shock1.240.562.750.60Age, dialysis, OP time, ASARenal failure (very few events, 0.4%)3.240.4324.260.25ASATransfusion1.580.902.7580.11Dialysis, steroid, OP time, BMI, ASAReturn to OR0.980.641.5140.94COPD hx, OP time, ASA30-day mortality1.500.464.9550.50Age, steroid, OP timeCardiac arrest1.840.437.8200.41Age, OP timeOrgan space infection1.020.771.3420.89Sex, smoking, dialysis, steroid, OP time, BMI, ASA, reconstruct>=30 day admission0.640.281.4740.29COPD hx, OP time, ASA30-day mortality1.570.475.1980.46OP time, logit of predicted mortality Open table in a new tab Odds ratios for the effect of surgical drains on each outcome when adjusted for the indicated confounders. mCR-POPF (modified clinically relevant POPF), CHF (Congestive heart failure), OP time (operative time), BMI (Body Mass Index), COPD (Chronic Obstructive Pulmonary Disease), ASA (American Society of Anesthesiologist Class).