Population attributable fraction of dementia risk in individuals with reduced kidney function

Abstract

AbstractBackgroundCommunity‐based studies regarding the association between chronic kidney disease and dementia risk are conflicting. In a large Swedish population‐based cohort, we explored the incidence and population attributable fractions (PAFs) of dementia risk in patients with chronic kidney disease (CKD).MethodIn the Stockholm CREAtinine Measurements (SCREAM) project, we studied the associations of baseline estimated glomerular filtration rate (eGFR) with the incidence of dementia among 329,822 residents of Stockholm, aged ≥65 years and with no prior history of dementia. The primary outcome was dementia incidence based on ICD‐10 codes or use of anti‐dementia drugs. Secondary outcomes were incidence of the dementia subtypes Alzheimer’s and vascular dementia. The association between eGFR and dementia risk was examined with Cox proportional hazards models. The weighted contributions of kidney disease (eGFR<60ml/min/1.73m2) and other plausible comorbidities to the risk of dementia were quantified using the population attributable fraction (PAF).Result18,983 cases of dementia (5.8% of participants) were detected throughout 1,185,304 person‐years. Dementia incidence rates (IR) were progressively higher with lower eGFR: from 6.56/1000 person‐years in persons with eGFR 90‐104ml/min to 30.28/1000 person‐years in those with eGFR<30ml/min). The HR of dementia risk was highest for the presence of CKD [eGFR<60ml/min; HR 1.54(1.28‐1.85)], followed by having a history of stroke [1.51(1.44‐1.58)], diabetes [1.23(1.18‐1.28)], atrial fibrillation [1.09(1.05‐1.14)], and myocardial infarction [1.07(1.01‐1.14)]. The presence of diagnosed hypertension and history of cancer [0.96(0.93‐0.99); and 0.87(0.84‐0.90) respectively] were associated with a lower dementia risk. Via PAF analyses we estimated that 11% [(95% CI (7‐14%)] of the dementia cases were attributed to eGFR<60 ml/min, followed by stroke [4% of cases attributed, 95%CI (4‐5%)] and diabetes [3% of cases, 95% CI (2‐3%)]. PAFs were around 1% for comorbidities such as atrial fibrillation and myocardial infarction. Risk magnitudes were stronger for vascular dementia than for Alzheimer. Sensitivity analyses across predefined strata or after excluding early events yielded similar results.ConclusionReduced kidney function is a strong risk factor for development of dementia, particularly vascular dementia. The proportion of dementia cases attributed to CKD is higher than that attributed to other dementia risk factors such as cardiovascular disease and diabetes.