A systematic review of methodologies, endpoints, and outcome measures in randomized trials of radiation therapy-induced nausea and vomiting

Clinical trials in radiation therapy-induced nausea and vomiting (RINV) appear to have varied methodologies, endpoints, and outcome measures. This complicates trial comparisons, weakens practice guideline recommendations, and contributes to variability in supportive care patterns of practice. We systematically reviewed RINV trials to describe and compare their pertinent design features.

Ovid versions of the Cochrane Central Register of Controlled Trials, the Cochrane Database of Systematic Reviews, EMBASE, and MEDLINE to January/February 2017 were searched for adult phase III trials of RINV management strategies. Key abstracted data included trial interventions and eligibility criteria, standard radiation therapy (RT) metrics, symptom assessment procedures, symptom definitions and grading systems, pre-specified and reported endpoints, and other outcome measures.

From 1166 references identified in the initial database search, we selected 34 trials for analysis that collectively randomized 4529 patients (median 61, range 11–1492). Twenty-eight trials (82%) were published prior to the year 2000. Twenty-seven trials (79%) involved multiple fraction RT and 7 (21%) single fraction RT. Twenty-four trials (71%) evaluated prophylactic interventions, 9 (26%) rescue interventions, and 1 trial did not specify. Thirty-three trials (97%) evaluated pharmacologic interventions. Twenty trials (59%) had patient report symptoms, 5 (15%) healthcare professionals or researchers, and 10 (29%) did not specify. Nausea was not defined in any trial but was reported as a stand-alone symptom in 26 trials (76%) and was graded in 20 (59%), with categorical qualitative scales being the most common method. Vomiting was defined in 3 trials (9%), was reported as a stand-alone symptom in 17 (47%), and was graded in 7 (21%), with continuous numerical scales being the most common method. Retching was defined in 3 trials, was not reported as a stand-alone symptom in any trial, and was graded in 1 (3%). Twenty-one trials (62%) created compound symptom measures that combined individual symptoms. Fifteen trials (44%) reported “emetic episode/event” measures but only 9 defined them. Seventeen trials (50%) reported complicated endpoints (e.g., “response,” “control,” “success”) that combined multiple symptom or compound symptom measures, but 7 did not define them comprehensively. Ten trials (29%) defined a primary endpoint a priori.

Methodologies, endpoints, and outcome measures varied considerably among 34 randomized trials in RINV.