To Err Is Robot—An analysis of complications following robotic donor hepatectomy

Abstract

Robotic donor hepatectomy (RDH) has been reported with lower morbidity than its open counterpart. Donor safety remains the primary concern, precluding its wide adoption. We aimed to evaluate donor complications following RDH and identify their predictive factors. Out of 348 live donor liver transplants performed between 2018 and 2021, the prospective data of 202 RDH were analyzed for complications by the modified Clavien-Dindo grading system. Multivariate analysis of donor and operative parameters was done to identify factors predicting complications, and CUSUM (cumulative sum) analysis was done to evaluate the effect of a learning curve. Out of 202 RDH (mean age: 37.5 [±10.4]; f [female]:m [male]-133:69; mean body mass index: 25.2±3.84), 196 (97%) were modified right lobe grafts. Conversion to open occurred in 7 (3.4%) (5-bleeding, 1-hepatic duct injury, and 1-portal vein kink). Postoperative complications occurred in 33 (16.3%), the most common being bile leak (5.9%) and bleeding (3.9%). Grades IIIa, IIIb, and IVa complications were seen in 3.4%, 3.4%, and 0.9% of patients, respectively. Reoperation was required in 3 cases for PVT, narrowing of IVC, and biliary peritonitis, respectively. At follow ups of 4 years, these patients are doing well. Although in univariate analysis, higher blood loss and body mass index appeared to be significant, multivariate analysis did not reveal any donor factor that could predict complications (biliary anatomy, portal anatomy, blood loss, body mass index, duration of surgery, or future liver remnant volume). The number of overall complications (21.7% vs. 9.9%; p =0.020; OR, 2.53) came down significantly in the second half. On RA-CUSUM (risk-adjusted CUSUM) analysis, we identified that it took around 130 cases for our unit to collectively overcome the learning curve. Although RDH appears to be safe, critical complications can occur in a minority of cases. Safety lies in flattening the learning curve.