JCO Article Insights: Low and Moderate Grade Adverse Events and the Patient Experience in Clinical Trials

In this JCO Article Insights episode, Subodh Selukar summarized findings from the original article published in the January 2024 JCO issue: “The Importance of Low and Moderate Grade Adverse Events on Patients’ Treatment Experience and Treatment Discontinuation” and accompanying editorial “Patient Experience, Adverse Event Reporting, and Clinical Trial Design”. The summary provides information regarding low and moderate grade adverse events and the patient experience in clinical trials. TRANSCRIPT Welcome to the JCO Article Insights episode for the January 2024 issue of Journal of Clinical Oncology. This is Subodh Selukar, your host, and today I will be providing a summary on 2 articles focused on low and moderate grade adverse events. The first article, titled “The Importance of Low and Moderate Grade Adverse Events on Patients’ Treatment Experience and Treatment Discontinuation” by Dr. O’Connell and colleagues, investigated low and moderate grade adverse events and the patient experience in clinical trials. Their article is accompanied by an editorial entitled “Patient Experience, Adverse Event Reporting, and Clinical Trial Design” by Dr. Neuman.  In clinical trials, a standardized system for reporting adverse events is the Common Terminology Criteria for Adverse Events (or “CTCAE”) established by the NCI, the United States National Cancer Institute. The CTCAE categorizes adverse events at 5 severity grades across 26 system organ classes. However, some clinical trials may only report adverse events at grade 3 or higher, with one possible rationale being that low and moderate grades are unlikely to affect patient safety or key trial endpoints. In Dr. O’Connell’s article, the team investigated how the numbers of grade 1 and 2 adverse events related to patient self-reported side-effect burden and treatment discontinuation. To do this, they analyzed data from the Phase 3 trial E1912 conducted by ECOG-ACRIN comparing two treatments for chronic lymphocytic leukemia. They chose this trial as an example because the study data included all adverse event grades throughout the duration of treatment for each patient. The authors studied side-effect burden based on GP5, which is the fifth item in the FACT-G subscale in the Functional Assessment of Cancer Therapy. GP5 rates the patient’s agreement with the statement “I am bothered by side effects of treatment” in the past 7 days, and it has previously been connected with adverse event grade and treatment discontinuation. For treatment discontinuation, the authors focused on those discontinuations that were recorded as being due to “adverse events, side effects or complications.” They found that, for each adverse event grade, there were, on average, more adverse events in cycles that ended with a patient discontinuing treatment compared to other cycles. Next, they used Bayesian models to assess how the numbers of grade 1 and grade 2 adverse events in a treatment cycle were associated with the odds of higher side-effect bother and odds of treatment discontinuation, after adjusting for cycle number, treatment and occurrence of grade 3 or higher adverse events within the cycle. Baseline GP5 was also included in the models, and these models also accounted for the inclusion of multiple cycles for each patient. When adjusting for baseline GP5, treatment, cycle and presence of grade 3 or 4 adverse events, both the number of grade 1 and the number of grade 2 adverse events were each strongly associated with increasing side-effect bother. The adjusted odds of treatment discontinuation were also higher with more grade 2 adverse events. However, with the same adjustment variables, the odds of treatment discontinuation were actually lower with larger numbers of grade 1 adverse events. In their primary analysis, they focused on adverse events that were attributed to treatment, so they excluded non-treatment-related adverse events from the counts. Sensitivity analyses including these adverse events have similar conclusions but with a weaker magnitude of effect. They attributed this to issues like existing adverse events not causing new bother. Next, the authors analyzed whether symptomatic versus asymptomatic adverse events affected these results by re-fitting the models and separating the predictors into numbers of asymptomatic and symptomatic grade 1 or 2 adverse events. In these results, they found no evidence for associations between numbers of asymptomatic adverse events at any grade and side-effect bother. On the other hand, they found strong evidence for associations with symptomatic adverse events of grade 2 and 3 or higher both for side-effect bother as an outcome and with treatment discontinuation. Asymptomatic grade 2 adverse events were associated with treatment discontinuation but not side-effect bother, and symptomatic grade 1 adverse events were associated with side-effect bother but not treatment discontinuation.  ·       The authors conclude that adverse events of all grades, especially symptomatic adverse events, should be recorded regularly in cancer clinical trials. Formal patient reported outcomes are not typically collected as frequently as adverse events are recorded, so identifying patients with a high number of lower grade adverse events could be used to facilitate early supportive care to improve patient quality of life and reduce the likelihood for treatment discontinuation. ·       They also highlight their result identifying lower odds of treatment discontinuation with larger numbers of grade 1 adverse events. They provide one explanation that patients may perceive grade 1 adverse events being associated with treatment efficacy, but this perception changes with higher grades.  In their call to collect more lower grade adverse events, the authors acknowledge that recording more adverse events may be time-consuming and burdensome for sites and recommend cost-benefit analyses to develop future guidelines. ·       This balance between the benefits and costs of increased adverse event data collection is the focus of Dr. Neuman’s editorial. Dr. Neuman acknowledges that Dr. O’Connell’s article provides a convincing argument for how low grade adverse event information is valuable, but notes the clinical trial context that current efforts at the NCI are to more efficiently conduct cancer research, which could be supported by streamlining data collection. ·       Requiring the collection of low grade adverse events could have important impacts to trial logistics. Due to the high volume of low grade adverse events, reporting all low grade events could delay reporting higher grade and more serious adverse events; and it would require an increase in the effort of clinical trial research staff, which would be difficult if not accompanied by an increase in reimbursement to sites. ·       Dr. Neuman suggests 3 approaches to balance the costs and benefits of collecting low and moderate grade adverse events. First, investigators could consider limiting low-grade adverse event reporting to the experimental arm. The standard of care regimens may not always have low-grade adverse event data available, but this may still be justified when there is extensive clinical experience with the standard of care. However, this approach is only practical when the experimental arm is not blinded. ·       A second approach for moderating the effort in collecting low-grade adverse events is to limit collection to symptomatic adverse events, connecting with Dr. O’Connell’s example E1912 dataset. This approach could be addressed by prespecifying types of symptomatic adverse events that would be most impactful during the trial design phase. ·       Dr. Neuman’s third suggestion is to plan for a follow-up study after the phase 3 trial to collect low-grade adverse event data and their impact on patients’ experiences and treatment discontinuation. This would be beneficial by only requiring low-grade adverse events in an experimental regimen that has successfully passed phase 3. However, a new study would require funding and site enthusiasm, which could prove challenging. ·       Overall, Dr. Neuman emphasizes that investigators should develop trial-specific considerations and engage with the relevant stakeholders during study design. Because of the complexity of adverse events in these patient populations, the best uses of grade 1-2 adverse events will likely continue to develop in the future. In their article, Dr. O’Connell’s team studied grade 1 and 2 adverse events as separate predictors, but I would be curious to know how the accumulation and trajectory of these adverse events affect the patient experience. For example, even if the severity does not rise to grade 3, an increasing trend in a patient’s adverse event severities could signal the treating physician to modify study dose or to discontinue the treatment. I’m not sure if that type of information was available in their trial E1912, but perhaps that could be a factor to consider for the future. And, of course, it will be important to assess how these grade 1-2 adverse events relate to the patient experience in different studies, especially across different cancer patient populations, acknowledging that this is inherently challenging to study because the data to inform this research is not universally available. As Dr. Neuman indicates, trial-specific goals and expertise will remain critical when considering the data collection for a given trial.  That concludes this episode of JCO Article Insights regarding a summary of the article “The Importance of Low and Moderate Grade Adverse Events on Patients’ Treatment Experience and Treatment Discontinuation” by Dr. O’Connell and colleagues and the editorial entitled “Patient Experience, Adverse Event Reporting, and Clinical Trial Design” by Dr. Neuman. This is Subodh Selukar. Thank you for your attention and stay tuned for the next episode of JCO Article Insights.