Are aging clocks based on routine clinical indicators trustworthy and applicable? A systematic review and critical appraisal

Aging clocks based on routine clinical indicators have emerged as a cost-effective tool for assessing biological age. This systematic review aims to summarize the characteristics and critically appraise these available aging clocks. Studies that developed aging clocks for adults (≥18 years) based on routine clinical indicators were retrieved from six databases (PubMed, EMBASE, Web of Science, CNKI, Wanfang Data, and Sinomed) up to June 18, 2024. The PROBAST + AI tool was used to assess the methodological quality, risk of bias, and applicability of included aging clocks. All the results were narratively summarized. Fifty-nine studies involving 81 aging clocks were included, of which 71 (87.7%) were developed using single-country datasets predominantly from China, the United States, Korea, and the United Kingdom. Notably, 31 aging clocks (38.3%) were developed with neither internal nor external validation. The majority of aging clocks were rated as having high concern regarding quality and high risk of bias, even including those published in high-impact journals. Only three aging clocks (3.7%) from two studies were rated as having low concern regarding quality and applicability during development, and two of these (4.0%) from one study further demonstrated low risk of bias and low concern for applicability during model evaluation. Future research should prioritize validating the promising aging clocks in target populations rather than developing new ones, adhere to the PROBAST + AI and TRIPOD + AI guidelines for methodological rigor and transparent reporting, and provide reproducible and user-friendly model codes and tools.