Lessons Learned From Two Prematurely Terminated, Randomized Trials on Biliary Drainage in Perihilar Cholangiocarcinoma

Abstract

The recently published INTERCPT-study was after the DRAINAGE-study, the second randomized controlled trial that compared endoscopic biliary drainage versus percutaneous biliary drainage in patients with perihilar cholangiocarcinoma (PHC).1Elmunzer B.J. et al.Clin Gastroenterol Hepatol. 2020 May 23; (S1542-3565(20)30696-0)Google Scholar,2Coelen R.J.S. et al.Lancet Gastroenterol Hepatol. 2018; 3: 681-690Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar Unfortunately, both studies have been prematurely stopped, raising several concerns. The INTERCPT-study was started in October 2017 and stopped in March 2019 by the Data Safety Monitoring Board because of slow accrual. In 18 months, only 13 patients were included out of 51 screened individuals among 16 centers in the United States, while the calculated sample size was 184. The authors described several factors that may have caused slow enrollment; many patients had been referred after initial drainage had already been undertaken, precluding enrollment in the trial. This was also encountered in the Dutch multicenter DRAINAGE-trial, which also randomized patients with PHC to endoscopic biliary drainage or percutaneous biliary drainage.2Coelen R.J.S. et al.Lancet Gastroenterol Hepatol. 2018; 3: 681-690Abstract Full Text Full Text PDF PubMed Scopus (88) Google Scholar This trial was stopped because of higher overall mortality in the percutaneous biliary drainage group. Within a 2.5-year period, 54 out of 261 screened patients could be included. A national, multidisciplinary clinical pathway, succeeded in more patients being referred to our center without previous biliary drainage. We believe that awareness among gastroenterologists and surgeons in referring centers of assessment of imaging before inserting biliary drains, is key to staging and to devising a treatment strategy. Although both randomized trials had the same objective (ie, to identify the optimal, initial biliary drainage technique in PHC), primary endpoints differed. In the INTERCPT-study, the primary endpoint was successful biliary drainage, defined as 50% reduction in bilirubin level within 3 weeks without additional drainage procedures. In the DRAINAGE-trial, the primary outcome was the number of severe complications of biliary drainage occurring between randomization and surgery. Therapeutic success was a secondary endpoint and was defined as normal-caliber bile ducts on ultrasound and a 20% decrease in bilirubin level within 1 week. Although difficult to compare because of different definitions, therapeutic success of biliary drainage was achieved in 6 out of 13 patients (46%) in the INTERCPT-study versus 38 out of 54 patients (70%) in the DRAINAGE-trial. Remarkably, the high mortality rate in the DRAINAGE-trial led to early study termination, whereas high mortality and complication rates were also observed in the INTERCPT-study: 10 of 13 patients (77%) experienced adverse events requiring hospital admission (vs 65% in the DRAINAGE-trial) and 8 of 13 patients (62%) died within 3 months follow-up (vs 26% in the DRAINAGE-trial). The unexpectedly high mortality rates in both trials require special attention. In the DRAINAGE-trial, with an all-cause mortality design, mortality seemed unrelated to drainage in 5 of 14 patients (2 myocardial infarction and 3 disease progression). Six patients died because of postoperative complications. The causes of death in the INTERCPT-trial have not been specified. An important difference between both trials was that the DRAINAGE-trial included patients with tumors deemed resectable, whereas the INTERCPT-study also included patients requiring biliary drainage in the palliative setting. This may have resulted in an even more heterogenous and complex trial population, hampering conduct and progress of this study. The authors of the INTERCPT-study encountered logistic issues in enrolling patients on-site and that institutional/clinician bias in favor of 1 of the 2 drainage techniques prevailed. The DRAINAGE-trial was conducted in 4 Dutch tertiary centers, whereas the INTERCPT-trial included 16 centers scattered among the United States. Effective collaboration among surgery, gastroenterology, and interventional radiology departments is more likely in a limited number of dedicated, participating centers. Remarkably, 2 randomized studies including patients with PHC both showed outcomes that contradict only retrospective data available in literature. Especially in this complex disease, retrospective cohort studies are biased by patient selection and inadequate registration of complications emphasizing the need for more prospective cohort studies. Whether a new randomized trial with the same objective will be attempted in the near future is questionable, in the wake of the 2 recently failed trials. Nevertheless, in any new trial examining the mode of biliary drainage in PHC, we strongly advise to focus on drainage-specific complications as primary outcome.