Understanding discordance in systematic reviews for treating chemotherapy-induced peripheral neuropathy: A focused umbrella review.

Abstract

e24111 Background: Despite the abundance of publications on chemotherapy-induced peripheral neuropathy (CIPN), we are still lacking effective management options. CIPN-focused systematic reviews (SRs) are increasingly encountered but have not aided much in clarifying treatment guidance. Additionally, there is a rising concern that SRs are now over-saturating the clinical literature, leading to redundant and misleading publications. Umbrella reviews (URs) are an attempt to appraise SRs but do not yet have firm methodologic frameworks. The aims of our UR were to appraise the discordance in recently published CIPN SRs and use content analysis as a method for identifying risks of discordance. Methods: This is a focused UR as part of a larger study to characterize the number of efficacy trials for established CIPN. Following PRISMA guidelines, we conducted a systematic literature search of databases MEDLINE, Embase, and Cochrane Library for publications in English from 1/1/2023 to 11/21/2023. Inclusion criteria were published systematic reviews attempting a comprehensive effort to synthesize any treatment outcomes for established CIPN. The AMSTAR 2 framework was used to appraise results for bias. Measures of randomized controlled trial (RCT) overlap are presented using a novel measure of time-adjusted CA, which accounts for RCT publication dates and SR search dates. Grounded theory coding was used for content analysis on efficacy. Results: We identified 18 SRs containing CIPN treatment RCTs published in less than 1 year, resulting in 32 unique CIPN RCTs. Seven SRs assessed exercise interventions, covering 6 exercise RCTs (41% time-adjusted CA). Five SRs assessed acupuncture, with 9 RCTs (50% time-adjusted CA). One SR searched for all CIPN treatments for pain but missed 4 RCTs. One SR was concerning for spurious data. Bias was variable and imperfect across SRs. Contextual analysis identified the themes of SRs combining preventative and treatment RCTs, confusing comparative vs. controlled trials, and frequent use of p-value only reporting to conclude treatment efficacy as major risks of SR discordance. Conclusions: UR was useful in identifying 32 CIPN RCTs that were not comprehensively identified by SRs. We presented a time-adjusted CA measure to indicate concordance percentage. A higher percentage indicates more re-citation of the same RCTs across SRs, with likely better chances of more concordant conclusions. This measure showed that SRs for CIPN treatment are discordant across the topics of exercise and acupuncture interventions. Content analysis revealed multiple potential causes of discordance. However, SR redundancy was helpful for identifying more unique RCTs. CIPN SR content standards should be developed to minimize discordancy and improve clinical applicability. Clinicians should be cautious when applying CIPN SR recommendations given the high discordance.