No abstract

Background: Longitudinal studies predictably experience non-random attrition over time. Among older adults, risk factors for attrition may be similar to risk factors for outcomes such as cognitive decline and dementia, potentially biasing study results. Objective: To characterize participants lost to follow-up which can be useful in the study design and interpretation of results. Methods: In a longitudinal aging population study with 10 years of annual follow-up, we characterized the attrited participants (77%) compared to those who remained in the study. We used multivariable logistic regression models to identify attrition predictors. We then implemented four machine learning approaches to predict attrition status from one wave to the next and compared the results of all five approaches. Results: Multivariable logistic regression identified those more likely to drop out as older, male, not living with another study participant, having lower cognitive test scores and higher clinical dementia ratings, lower functional ability, fewer subjective memory complaints, no physical activity, reported hobbies, or engagement in social activities, worse self-rated health, and leaving the house less often. The four machine learning approaches using areas under the receiver operating characteristic curves produced similar discrimination results to the multivariable logistic regression model. Conclusions: Attrition was most likely to occur in participants who were older, male, inactive, socially isolated, and cognitively impaired. Ignoring attrition would bias study results especially when the missing data might be related to the outcome (e.g. cognitive impairment or dementia). We discuss possible solutions including oversampling and other statistical modeling approaches.

Background and objectivesIn the general population, sleep disorders are associated with mortality. However, such evidence in patients with CKD and ESKD is limited and shows conflicting results. Our aim was to examine the association of sleep apnea with mortality among patients with CKD and ESKD.Design, setting, participants, & measurementsIn this prospective cohort study, 180 patients (88 with CKD stage 4 or 5, 92 with ESKD) underwent in-home polysomnography, and sleep apnea measures such as apnea hypopnea index (AHI) and nocturnal hypoxemia were obtained. Mortality data were obtained from the National Death Index. Cox proportional hazard models were used for survival analysis.ResultsAmong the 180 patients (mean age 54 years, 37% women, 39% with diabetes, 49% CKD with mean eGFR 18±7 ml/min per 1.73 m2), 71% had sleep apnea (AHI>5) and 23% had severe sleep apnea (AHI>30). Median AHI was 13 (range, 4–29) and was not significantly different in patients with advanced CKD or ESKD. Over a median follow-up of 9 years, there were 84 (47%) deaths. AHI was not significantly associated with mortality after adjusting for age, sex, race, diabetes, body mass index, CKD/ESKD status, and kidney transplant status (AHI>30: hazard ratio [HR], 1.5; 95% confidence interval [95% CI], 0.6 to 4.0; AHI >15 to 30: HR, 2.3; 95% CI, 0.9 to 5.9; AHI >5 to 15: HR, 2.1; 95% CI, 0.8 to 5.4, compared with AHI≤5). Higher proportion of sleep time with oxygen saturation <90% and lower mean oxygen saturation were significantly associated with higher mortality in adjusted analysis (HR, 1.4; 95% CI, 1.1 to 1.7; P=0.007 for every 15% higher proportion, and HR, 1.6; 95% CI, 1.2 to 2.1; P=0.003 for every 2% lower saturation, respectively). Sleep duration, sleep efficiency, or periodic limb movement index were not associated with mortality.ConclusionsHypoxemia-based measures of sleep apnea are significantly associated with increased risk of death among advanced CKD and ESKD.

Background/Objective Poor air quality is implicated as a risk factor for cognitive impairment and dementia. Few studies have examined these associations longitudinally in well‐characterized population‐based cohorts with standardized annual assessment of both mild cognitive impairment (MCI) and dementia. We investigated the association between estimated ambient fine particulate matter (PM2.5) and risk of incident MCI and dementia in a post‐industrial region known for historically poor air quality. Setting/Participants Adults aged 65+ years in a population‐based cohort (n = 1572). Measurements Census tract level PM2.5 from Environmental Protection Agency (EPA) air quality monitors; Clinical Dementia Rating (CDR)®. Design We estimated ambient PM2.5 exposure (μg/m3, single‐year and 5‐year averages) by geocoding participants' residential addresses to census tracts with daily EPA PM2.5 measurements from 2002 to 2014. Using Bayesian spatial regression modeling adjusted for age, sex, education, smoking history, and household income, we examined the association between estimated PM2.5 exposure and risk of incident MCI (CDR = 0.5) and incident dementia (CDR ≥ 1.0). Results Modeling estimated single‐year exposure, each 1 μg/m3 higher ambient PM2.5 was associated with 67% higher adjusted risk of incident dementia (hazard ratio [HR] = 1.669, 95% credible interval [CI]: 1.298, 2.136) and 75% higher adjusted risk of incident MCI (HR = 1.746, 95% CI: 1.518, 2.032). Estimates were higher when modeling 5‐year ambient PM2.5 exposure for incident dementia (HR = 2.082, 95% CI: 1.528, 3.015) and incident MCI (HR = 3.419, 95% CI: 2.806, 4.164). Conclusions Higher estimated ambient PM2.5 was associated with higher risk of incident MCI and dementia, particularly when considering longer‐term exposure, and independent of demographic characteristics and smoking history. Targeting poor air quality may be a reasonable population‐wide intervention to reduce the risk of cognitive impairment in older adults, particularly in regions exceeding current recommendations for safe exposure to PM2.5.