Oncological and Survival Endpoints in Cancer Cachexia Clinical Trials: Systematic Review 6 of the Cachexia Endpoints Series.

Background: In patients receiving anti-cancer treatment, cachexia results in poorer oncological outcomes. However, there is limited understanding and no systematic review of oncological endpoints in cancer cachexia (CC) trials. This review examines oncological endpoints in CC clinical trials.

Methods: An electronic literature search of MEDLINE, Embase and Cochrane databases (1990-2023) was performed. Eligibility criteria comprised participants ≥ 18 years old; controlled design; ≥ 40 participants; and a cachexia intervention for > 14 days. Trials reporting at least one oncological endpoint were selected for analysis. Data extraction was performed using Covidence and followed PRISMA guidelines and the review was registered (PROSPERO CRD42022276710).

Results: Fifty-seven trials were eligible, totalling 9743 patients (median: 107, IQR: 173). Twenty-six (46%) trials focussed on a single tumour site: eight in lung, six in pancreatic, six in head and neck and six in GI cancers. Forty-two (74%) studies included patients with Stage III/IV disease, and 41 (70%) included patients receiving palliative anti-cancer treatment. Ten studies (18%) involved patients on curative treatment. Twenty-eight (49%) studies used pharmacological interventions, 29 (50%) used oral nutrition, and two (4%) used enteral or parenteral nutrition. Reported oncological endpoints included overall survival (OS, n = 46 trials), progression-free survival (PFS, n = 7), duration of response (DR, n = 1), response rate (RR, n = 9), completion of treatment (TC, n = 11) and toxicity/adverse events (AE, n = 42). Median OS differed widely from 60 to 3468 days. Of the 46 studies, only three reported a significant positive effect on survival. Seven trials showed a difference in AE, four in TC, one in PFS and one in RR. Reported significances were unreliable due to missing adjustments for extensive multiple testing. Only three of the six trials using OS as the primary endpoint reported pre-trial sample size calculations, but only one recruited the planned number of patients.

Conclusion: In CC trials, oncological endpoints were mostly secondary and only few significant findings have been reported. Due to heterogeneity in oncological settings, nutritional and metabolic status and interventions, firm conclusions about CC treatment are not possible. OS and AE are relevant endpoints, but future trials targeting clinically meaningful hazard ratios will required more homogeneous patient cohorts, adequate pre-trial power analyses and adherence to statistical testing standards.