Toxicity May Be Underreported in Phase 3 Oncology Trials

Fifty-one percent of trials reported SAEs, and 55% reported total AEs.

Toxicity may often be underreported in two-arm, superiority-design, phase 3 oncology trials, and may frequently be minimized in its interpretation by researchers, according to findings from a cross-sectional analysis published in the Journal of Clinical Oncology.

Of the 407 trials from 2002 to 2020 included in this study, 51% reported serious adverse events (SAEs), 88% (n = 358) of trials reported total deaths, and 84% (n = 340) reported discontinuation due to toxicity.

Additionally, 55% reported total AEs, although 32% (n = 131; 95% CI, 28%-37%) of the trials had complete toxicity reporting (CTR). CTR was more commonly observed in trials sponsored by industry (37%) compared with cooperative group sponsorship (4%). Of note, all of the trials that reported CTR were industry sponsored, compared with 3% of (4/31) of trials sponsored by a cooperative group.

Ninety-eight percent of trials found to be in accordance with previously established toxicity reporting guidelines were industry sponsored, and 6% were cooperative-group sponsored.

“These findings, specifically the lack of CTR in the majority of trials, are consistent with previously published research that suggests that AEs are not just incompletely reported but reported on the basis of widely variable criteria,” wrote the authors of the study.

Toxicity minimizing language (TML) was used in 46% (n = 186; 95% CI, 41%-51%) of trials. No factors related to the trial, including the source of sponsorship, were linked with the likelihood of TML use.

In the 407 trials that researchers assessed, a trial was found to completely report toxicity if it reported total adverse events (AEs), total SAEs, total deaths, and discontinuation of study therapy due to toxicity. Researchers found 44% of trials (95% CI, 40%-49%) to be in concordance with guidelines from a prior study on toxicity reporting, defined as meeting the aforementioned criteria with the exception of total AEs.

Likewise, terms that could soften the perception of toxicity in a trial including “acceptable,” “safe,” “well tolerated,” “tolerable,” “feasible,” “favorable toxicity profile,” and “manageable” constituted toxicity minimizing language.

Of note, 43% of trials that assessed therapies that were granted FDA approval completely reported toxicity. Forty percent of therapies that were approved by the FDA based on the trials did not concord with prior toxicity reporting guidelines, and 35% of approved drugs based on these trials used TML.

Medical writers assisted with reports for 227 trials (56%) altogether, 102 of which had CTR, meaning those reports accounted for 78% of all trials where toxicity was completely reported. Guideline concordance and TML were identified in 72% and 58%, respectively, of trials where medical writers were involved.

TML terms were most often used in the discussion (44%) and abstract (19%) of trial reports, and the most frequent TML terms found in the trials were “tolerable” (24%), “manageable” (16%), and “acceptable” (12%).

“Therefore, in addition to what is already recommended by current guidelines, we encourage CTR, which includes [total AEs] along with total SAEs, total deaths, and study therapy discontinuations because of toxicity,” researchers noted.

Even after adjusting for confounders, industry-funded trials published more recently had greater odds of CTR (adjusted OR = 1.25; 95% CI, 1.14-1.37; P < .01), as did use of medical writers (adjusted OR = 2.25; 95% CI, 1.35-3.75; P < .01). Similarly, publication year (adjusted OR = 1.18; 95% CI, 1.09-1.28; P < .01) and use of medical writers (adjusted OR = 1.92; 95% CI, 1.20-3.08; P = .01) were also associated with greater odds of guideline concordance in industry-funded trials.

Studies involving thoracic cancers were most common in the CTR group compared to other disease types, comprising 30% of CTR trials (OR = 3.94; P < .01), followed by hematologic cancers (19%; OR = 2.90; P < .01).

Reference

Miller AM, Passy AH, Sherry AD, et al. Incomplete Toxicity Reporting and Use of Toxicity-Minimizing Language in Phase III Oncology Trials. JCO Oncol Pract. Published online January 8, 2025. doi:10.1200/OP-24-00735