About one-third of the total population of patients starting HD in the study period comprised very elderly patients (≥75 years). This group had a higher proportion of women, hypertension as the primary cause of ESRD, vascular access via a catheter, capable of partial self-care or bedridden status, and lower BMI and serum albumin level compared with the two groups of younger HD patients. The all-cause mortality rate was higher in very elderly patients than in those aged < 75 years. Older age, presence of cancer, catheter use, and low Kt/V, BMI, and serum albumin level were associated with an increased mortality risk in very elderly patients. Using the nCI to investigate the associations between comorbid conditions and all-cause mortality, we found that patients with an nCI score > 3 had a 58.2% higher mortality risk compared with those with an nCI score of zero.
Despite recent technical advances, the mortality rate remains high in elderly HD patients9. This may partially reflect the shorter life expectancy of the older general population. A comparison of life expectancy between the HD population and the general population in the Japanese Society for Dialysis Therapy Renal Data Registry (JRDR) database found a life expectancy ratio of almost 50% for the HD population relative to the general population irrespective of age10. This finding suggests that the shorter life-span of older HD patients relates to the shorter life expectancy of the older general population.
Given that old age is no longer considered to be a contraindication for starting HD, investigating the clinical and functional risk factors for adverse prognosis is important in the elderly population in the context of treatment decisions. Studies have evaluated the outcomes of and risk factors for maintenance HD in elderly patients11. Age, low BMI, cardiovascular disease, diabetes, central venous catheter, early start of HD, frailty, functional impairment, cognitive impairment, and falls have been suggested as risk factors for mortality12–17. However, there is no consensus about the factors affecting mortality in this population.
The distribution of clinical parameters sometimes differs between age groups, and clinical parameters in elderly people can differ from those in their younger counterparts. These differences can affect the outcomes of treatment and the associations between clinical parameters and outcome. A study using the JRDR database found different distributions of clinical parameters by age group; for example, low serum levels of hemoglobin, albumin, creatinine, phosphate, and total cholesterol, and low body weight were prevalent among the older age groups18. In our study, the distributions of sex, primary cause of ESRD, BMI, Kt/V, and serum levels of hemoglobin, albumin, calcium, and phosphorus differed between age groups. A lower BMI, Kt/V, and albumin level were associated with increased mortality in the very elderly patients in our study, and we also found that they more frequently had a central venous catheter for vascular access.
The type of vascular access used for HD has been shown to influence HD patient mortality in many observational studies19–21. However, few studies have examined the impact of vascular access type on the survival of very elderly HD patients. We found that central catheter use was significantly related to the survival rate of very elderly patients undergoing HD. Other clinical parameters and comorbid conditions are associated with mortality in ESRD patients. In the current study, we found that the nCI, which includes several comorbid conditions without an age component, is a useful predictor of mortality in elderly HD patients. Very elderly patients with a low nCI score, indicating fewer comorbidities, had better survival. This finding suggests that adequate dialysis rather than conservative treatment should be considered preferentially, especially in elderly patients with fewer comorbid conditions.
Recent studies have suggested that HD provides a survival benefit over conservative management in the elderly population22–24. However, survival time in elderly patients who receive conservative treatment may be comparable with HD, as shown by one study that reported similar numbers of hospital-free days in conservative and HD groups25. Older patients may experience poorer QOL than younger patients, especially in terms of functional impairment26. The benefits of HD and its effects on QOL in elderly patients continue to be debated6,27. In addition, QOL issues have not been reported in many studies because of their controversial nature in measurement and randomized controlled trials are not feasible.
Observational studies remain the best way to examine and compare the prognosis of elderly HD patients and treatment methods. A systematic review of relevance of impairments in elderly patients starting HD noted that impairments in multiple physical and mental functions at HD initiation are related to poor outcomes28. These results suggest that the distribution of vulnerability and clinical characteristics can differ between age groups and may be related to the outcomes. We found a high prevalence of capable status of self-care in the very elderly, who had a high cumulative survival rate. However, bedridden status was not a significant risk factor for mortality in the multivariate Cox analysis. This might be explained by the small number of bedridden patients or the low risk of falls or infection in bedridden patients because of their low activity levels. Further studies are needed using standardized assessment tools focused specifically on the ESRD population.
To our knowledge, this is the first study to analyze mortality rates and prognostic factors in a large cohort of very elderly patients (≥75 years) starting HD in Korea. However, our study has some limitations. First, given its retrospective nature, data from a registry lack detailed information. Second, this study included the Korean ESRD population, and the results of this study cannot be generalized. Third, information about QOL and fragility was lacking in the registry. Fourth, we used gastrointestinal disease as a general category that included gastric and duodenal ulcers instead of gastrointestinal bleeding when calculating the nCI score. Thus, the primary question about which very elderly patients will derive survival and other benefits remains unanswered.
In conclusion, a large proportion (about one-third) of the patients starting HD included in our study were very elderly (≥75 years). Among these, patients with a lower nCI had a significant survival benefit compared with those with a higher nCI. And the survival rate was lower in patients with vascular access via a catheter than in those with an arteriovenous fistula or graft. These findings suggest that age should not be a barrier to starting HD in very elderly patients with few comorbid diseases and use of an arteriovenous fistula or graft as a vascular access should be considered carefully in these patients.