2115

Understanding predictors of mortality in End Stage Renal Disease patients on dialysis: a retrospective observational study

Emma Murphy1,2,5, Michael J Germain4, Hugh Cairns3, Irene J Higginson2, Fliss EM Murtagh2,

1NIHR GSTFT/KCL Biomedical Research Centre, Guy’s and St Thomas’ NHS Foundation Trust, London, UK

2King’s College London, Department of Palliative Care, Policy and Rehabilitation, Cicely Saunders Institute London, UK

3Kings College Hospital NHS Foundation Trust London, UK

4Baystate Medical Center and Tufts University School of Medicine, Springfield, USA

5University of Southampton, UK

Background: In patients with end stage renal disease (ESRD), the rate of deaths preceded by dialysis withdrawal is high. However, rates vary across studies and national renal registries. Different classifications of causes of death make it difficult to compare mortality rates and factors predictive of death preceded by dialysis withdrawal. The study of dialysis withdrawal epidemiology is restricted by the lack of consensus over the basic definition and classification of dialysis withdrawal.

Aim: To understand variations associated with mortality rates preceded by dialysis withdrawal, propose a unified classification of dialysis withdrawal based on trajectories and causal criteria, examine mortality rates and factors associated with death preceded by dialysis withdrawal.

Methods: Retrospective cohort study. Data from one UK renal unit were used to examine mortality rates preceded by dialysis withdrawal and identify variables associated with death preceded by dialysis withdrawal and death on dialysis not preceded by dialysis withdrawal among a cohort of adult ESRD patients on dialysis. Descriptive mortality rates, unadjusted and adjusted, and rate ratios are reported. Factors associated with deaths preceded by dialysis withdrawal are examined using poisson regression.

Findings: The cohort included 1175 participants from the year 2005 to 2012 of whom 497 died. Patient age ranged from 18 to 88 years. 59% had 1-2 co-morbid conditions, diabetes was present in nearly 40% of the total study population and 20% were diagnosed with ischaemic heart disease. The most common cause of renal disease classified in the total population was diabetic nephropathy and aetiology uncertain (30.1% and 20.2% respectively). Median survival was 6.3 years (95% CI 5.6-7.0). Crude death rate – all cause, all ages was 154 per 1000 person years, the age-standardized death rate was 67.7 per 1000 person years. Crude death rate preceded by dialysis withdrawal in 80+ age groups was 65.4 per 1000 person years, in comparison the crude death rate on dialysis not preceded by withdrawal was 320.3 per 1000 person years. The mortality rate ratio for death preceded by dialysis withdrawal relative to diabetes mellitus, as estimated by Poisson regression, was 1.7(95% CI: 1.1-2.6). The rate ratio for death preceded by dialysis withdrawal relative to age («65 v's »65) was 6.1 (95%CI 3.7-10.8).

Conclusion: Mortality rates preceded by dialysis withdrawal should be interpreted with caution because of differences in classification. In addition, without age adjustment, observation about causes of death preceded by dialysis withdrawal are likely to be misleading. All cause specific mortality rates in populations should be adjusted in order to reliably describe mortality patterns. The different age profiles of people 1) whose death is preceded by dialysis withdrawal and 2) deaths on dialysis not preceded by dialysis withdrawal are important as this determines the needs of the individuals affected, and also determine where people are likely to receive end of life care.

1