Frailty is very common in old age and often associated with adverse events. Transitioning between frailty states is possible in both directions (improvement and worsening) offering targets for interventions. Frailty is more prevalent in women, but little is known about the impact of gender on frailty transition. The aim of this study is to identify gender differences for frailty transition in older adults and to develop gender-stratified prognostic prediction models for frailty transition. We performed a longitudinal analyses of the Berlin Initiative (cohort) Study with a frailty follow-up of 2.1 years. Description of frailty transition using the frailty phenotype and development of prognostic prediction models using multivariable logistic regressions for transition (improvement or worsening) stratified by gender following the TRIPOD statement were performed. In total, the study population consisted of 1158 community-dwelling adults with a mean age of 84.4 years and of whom 55% were women. Out of 1158 participants 225 (19%) were robust, 532 (46%) prefrail and 401 (35%) frail. After 2.1 (IQR 2.0-2.3) years, half of the participants had transitioned between frailty states. Men worsened more often and those who were already frail died more often than women. Gender-stratified prediction models for frailty transition demonstrated that some predictors (age, self-rated health, cognitive impairment, baseline frailty status) were included in all models. While stroke, diabetes mellitus, smoking and glomerular filtration rate were unique predictors in the models for females, osteoarthritis, hospitalization and education were predictors in the models for males. There are gender differences in frailty transition rates, patterns and prediction. This supports the importance of considering gender when addressing frailty and targeting interventions in old age.
