MO902: Age-Specific Impacts of Risk Factors on All-Cause Mortality in Dialysis Patients

Abstract

Abstract BACKGROUND AND AIMS Chronic haemodialysis patients expect longer lifespan than previously; they have various backgrounds of mortality risk factors. However, age specific impacts of risk factors on mortality in dialysis patients have rarely been evaluated. Therefore, we aimed to analyze age-specific association between risk factors and all-cause mortality and evaluated differences in the association among age groups. METHOD We conducted a post-hoc analysis of the 461 hypertensive chronic haemodialysis (HD) patients participated in the Olmesartan Clinical Trial in Okinawan Patients Under OKIDS (OCTOPUS) between June 2006 and June 2011. The participants were additionally followed up until 31 July 2018. We stratified the participants into 3 groups (younger patients, n = 156; middle-aged patients, n = 160; elderly patients, n = 145) based on tertile of their ages at their enrollment to the trial. We used age-stratified Cox regression models to assess the association of baseline predictors with all-cause mortality for each age group. In addition, we calculated hazard-ratio based population attributable fraction (PAF) for baseline predictors and all-cause mortality across age groups. RESULTS A median follow-up period of participants was 10.2 years. Means and standard deviations of age for three age groups; younger, middle-aged and elderly patients, were 45.9 ± 6.5,  61.2 ± 3.9 and 72.5 ± 3.3,  respectively. The number (%) of deceased patients for the three group was 40 (25.6%), 65 (40.6%) and 106 (73.1%); hazard ratios estimated with Cox regression adjusted for gender and diabetes for the three groups were 1(Reference), 1.43 (0.96–2.14; P = 0.075) and 3.49 (2.41–5.05; P < 0.001), respectively. Age-stratified Cox-regression analyses showed the following associations; diabetes was stably associated with all-cause mortality in all age groups; history of stroke, history of coronary intervention, systolic blood pressure ≥ 170 mmHg and serum phosphate ≥ 6.0 mg/dL had greater impacts on all-cause mortality in younger patients compared with elderly patients; and hypoalbuminemia and hypokalemia had higher impact on all-cause mortality in elderly patients (Figure 1A). We calculated hazard-ratio based PAF for the risk factors to evaluate the difference in impacts of risk factors on all-cause mortality among the age groups; systolic blood pressure ≥ 170mmHg and serum phosphate ≥ 6.0 mg/dL were strongly associated with deaths especially in younger patients (Figure 1B). The total of PAF in elderly patients was almost half of the total of PAF in younger patients. CONCLUSION In chronic haemodialysis patients, associations between risk factors and all-cause mortality are not uniform among different age groups. Vascular disease associated risk factors and hyperphosphatemia were associated with greater risk of morality in younger patients. Hypoalbuminemia and hypokalemia were associated with higher risk of mortality in elderly patients. Proportion of deaths due to modifiable risk factors in elderly HD patients is almost half of that in younger HD patients.