Joint latent class mixed model trajectories of estimated glomerular filtration rate related to mortality and physical activity in older adults

Abstract Deterioration of renal function occurs with age, leading in many individuals to chronic kidney disease, considerable disability and premature mortality. The trajectories of kidney function decline and potential protective factors remain poorly understood, especially in older individuals. We therefore investigated eGFR-decline trajectories related to mortality as well as the influence of physical activity patterns on trajectory membership in a population-based cohort of older individuals. 1,506 community-dwelling adults aged $$\:\ge\:$$ 65 years were recruited in 2009 and followed up for up to three times. Trajectories of eGFR-decline and their relation with mortality were determined using joint latent class models. The association of physical activity (PA) and sedentary behavior (SB), determined by accelerometer measurements at baseline, with class-membership was studied with logistic regression. There were 2,812 observations on eGFR. We identified two trajectories, one starting with a lower mean baseline eGFR value and showing a stronger decline over time and higher mortality. In adjusted models, we found robust inverse associations of time spent in PA and total steps with being in this trajectory (Odds ratios with 95%-confidence intervals of 0.72 [0.60, 0.87] for 10 min and 0.73 [0.62, 0.86] for 1000 steps); SB was positively associated with this trajectory (2.51 [1.65, 3.81] for 100 min). PA was negatively and SB was positively associated with the worse trajectory. This may indicate that interventions for increasing PA may potentially be viable strategies to prevent impaired renal function and slow age-related functional decline in older individuals and should be investigated in future studies.