Quality of life assessment and reporting in colorectal cancer: a systematic review of phase 3 trials published between 2012 and 2018

Introduction: We previously reported that quality of life (QoL) is not included among endpoints, and QoL results are significantly underreported in a high proportion of recently published phase 3 trials in oncology. In this study, our aim was to describe QoL prevalence and heterogeneity in QoL reporting in colorectal cancer phase 3 trials. Methods: We included all primary publications of phase 3 trials evaluating anticancer drugs in colorectal cancer published between 2012 and 2018 by 11 major journals. We analyzed QoL inclusion among endpoints, presence of QoL results, and methodology of QoL analysis. Results: In total, 67 publications were identified (20 in adjuvant/neoadjuvant setting, 26 in first-line or maintenance and 21 in second and further lines). In 41 (61.2%) publications, QoL was not listed among endpoints: 17/20 (85.0%) in adjuvant/neoadjuvant setting, 13/26 (50.0%) in first line, and 11/21 (52.4%) in second and further lines. The proportion of trials not including QoL among endpoints was similar over time: QoL was not listed in 26/43 (60.5%) publications between 2012 and 2015 vs. 15/24 (62.5%) between 2016 and 2018. Out of 26 primary publications of trials including QoL among endpoints, QoL results were not reported in 10 (38.5%). Overall, no QoL data were available in 51/67 (76.1%) primary publications: 19/20 (95.0%) in adjuvant/neoadjuvant setting, 18/26 (69.2%) in first line, and 14/21 (66.7%) in second and further lines. The proportion of trials without available QoL results in primary publication was similar over time: QoL was not reported in 32/43 (74.4%) publications between 2012 and 2015 vs. 19/24 (79.2%) between 2016 and 2018. Overall, 7 secondary QoL publications were found. For trials with QoL among endpoints but no QoL results in the primary publication, the probability of secondary publication was 0%, 33.3%, and 50.0% after 1, 3 and 5 years, respectively. In trials conducted in metastatic patients, QoL data were not available in 12/18 (66.7%) trials with primary endpoints of overall survival and in 20/29 (69.0%) trials with other primary endpoints. QoL data were not available in 20/32 (62.5%) trials with a positive result: 5/6 (83.3%) in adjuvant/neoadjuvant setting, 7/13 (53.8%) in first line, and 8/13 (61.5%) in second and further lines. QoL was not reported in primary publication in 19/26 (73.1%) of for-profit trials vs. 32/41 (78.0%) for no-profit trials. Of 7 secondary QoL publications, 1 was of for-profit trials vs 6 of no-profit trials. In 21 trials with available QoL results, most common QoL tools were EORTC QLQ-C30 (14, 66.7%); EORTC colorectal cancer module (3, 14.3%); EQ5D (8, 38.1%); FACT-C (3, 14.3%) and other FACT tools (4, 19.0%). Common methods of analysis were mean scores over time (10, 47.6%), mean changes (6, 28.6%), time to deterioration (5, 23.8%) and proportion of responders or worsening patients (4, 19.0%). Conclusion: QoL was absent in a high proportion of recently published phase 3 trials in colorectal cancer, even in no-profit trials and in trials of second or further lines, where attention to QoL should be particularly high. Furthermore, the methodology of QoL analysis was found to be heterogeneous in terms of types of instruments, analysis, and presentation of results.