Prognostic factors in mid-life for predicting subsequent frailty: systematic literature review.

Objective

This systematic review aimed to identify, describe, and critically appraise the evidence for prognostic factors and models in mid-life (defined as a mean population age of 40-64 years) associated with frailty onset in later life. STUDY

Design

AND

Setting

MEDLINE, Embase, PsycINFO, CINAHL, Web of Science, LILACS and The HTA databases were searched for eligible studies involving mid-life populations without frailty that assessed prognostic factors for incident frailty with ≥5 years of follow-up (up to January 2025). Studies reporting the development or validation of a prediction model to be used in mid-life for predicting frailty were also included. Risk of bias was assessed using the Quality In Prognosis Studies (QUIPS) and Prediction model Risk of Bias Assessment Tool (PROBAST) tools. A qualitative synthesis of the evidence was conducted, using an adapted version of the GRADE framework. Members from the public contributed to this review through a Patient and Public Involvement (PPI) Advisory Group.

Results

Thirty-five prognostic factor studies (including >2.3 million individuals) and one prediction model (n=287) were included. Fifty-two prognostic factors were identified across six categories: demographic, lifestyle, general health, social networks or support, psychological, work-related and environmental. Twenty-eight studies were rated at high risk of bias in at least one QUIPS domain. No prognostic factors for frailty onset with high-quality evidence were identified. Moderate-quality evidence was found for older age, being single, smoking, depression, perceived unmet healthcare needs, area deprivation, lower occupational level, and lower wealth. The sole prediction model showed that musculoskeletal factors accounted for a significant proportion of the predicted risk of frailty, but it was rated at high risk of bias. Eighteen additional factors were suggested by PPI contributors for future investigation.

Conclusion

Robust studies are needed to improve our understanding of those at higher risk of developing frailty later in life to inform potential targets for preventive interventions. We identified eight prognostic factors with moderate-quality evidence. Smoking, depression, and unmet healthcare needs are modifiable and should be considered as candidate predictors for inclusion in future prediction models. Members of the PPI advisory group emphasised the importance of investigating additional prognostic factors not captured in this review as priorities for future research.