Impact of comprehensive geriatric assessment on the risk of adverse events in the older patients receiving anti-cancer therapy: a systematic review and meta-analysis

BACKGROUND: to assess the efficacy of comprehensive geriatric assessment (CGA) for preventing treatment-related toxicity in older people undergoing non-surgical cancer therapies. METHODS: MEDLINE, EMBASE and Cochrane library databases were searched from inception till January 2022 to identify randomised controlled trials (RCTs) on the incidence of toxicity measured by the Common Terminology Criteria for Adverse Events (primary outcome) and that of therapeutic modifications, early treatment discontinuation, progression-free survival, overall survival and hospitalisation (secondary outcomes). RESULTS: analysis of six RCTs published from 2016 to 2021 recruiting 2,126 participants (median age: 71-77) who received chemotherapy as the major therapeutic approach revealed 51.7% and 64.7% of Grade 3+ toxicity in the CGA and control (i.e. standard care) groups, respectively (RR = 0.81, 95% CI: 0.7-0.94, P = 0.005, I2 = 65%, certainty of evidence COE: moderate). There were no significant differences in the incidence of early treatment discontinuation (RR = 0.88, P = 0.47; I2 = 63%,1,408 participants, COE: low), initial reduction in treatment intensity (RR = 0.99, P = 0.94; I2 = 83%, 2055 participants, COE: low), treatment delay (RR = 1.06, P = 0.77, I2 = 0%, 309 participants, COE: moderate), hospitalisation (RR = 0.86, P = 0.39, I2 = 41%, 914 participants, COE: moderate), progression-free and overall survival with or without CGA. However, there was an association between CGA and a lower incidence of dose reduction during treatment (RR = 0.73, P < 0.00001, 956 participants, COE: moderate). CONCLUSIONS: our results demonstrated that comprehensive geriatric assessment may be associated with a lower incidence of treatment-related toxicity and dose reduction compared to standard care in older people receiving non-surgical cancer treatments. Further large-scale studies are warranted to support our findings.