Burden of dialysis patients on the healthcare system: from quantification of cardiovascular events to cost-effectiveness

Abstract

Very few studies investigated the interaction of incident diabetes and end-stage renal disease (ESRD) on the risks of cardiovascular (CV) events life expectancy and expected years of life lost (EYLL) In addition although treatment for the dialysis population is resource intensive a cost-effectiveness analysis between hemodialysis (HD) and peritoneal dialysis(PD) by matched pairs is still lacking To investigate the impact of ESRD and diabetes on the risks of CV events and LE and EYLL after CV events we first determined the age- and sex-specific incidences twenty-year risks of incident CV events including acute myocardial infarction (AMI) stroke and congestive heart failure (CHF) stratified by the presence of diabetes de novo diabetes after ESRD or ESRD by using two representative national cohorts Individuals were excluded if aged below 18 years or the presence of previous CV events or malignancy before enrollment Cox proportional hazard models were also constructed with adjustments for competing risk of mortality A total 648 851 non-ESRD individuals and 71 397 ESRD patients including 53 342 and 34 754 diabetic patients respectively were followed up during 1998-2009 A monotonic risk pattern of CV-related incidences was noted with the presence of diabetes ESRD or both respectively after stratified by age and sex De novo diabetes showed similar increased risks for CV incidences especially AMI and stroke There is a multiplicatively synergistic effect of diabetes and ESRD for CV related risks especially for AMI and stroke of which the adjusted hazard ratios (aHRs) [95% confidence intervals] were 5 24 [4 83-5 68] and 2 43[2 32-2 55] respectively in comparison with people without diabetes or ESRD; de novo diabetes after ESRD had similar effects with aHRs of 4 12[3 49-4 87] and 1 75[1 57-1 95] respectively Second we further followed up these individuals with incident CV events till mortality or the end of 2009 We estimated age- and sex-specific survival rates by the Kaplan-Meier method which were further extrapolated to lifetime to estimate the LE and EYLL based on an assumption of constant excess hazard Cox proportional hazard models were also constructed to validate the results Of 35 793 patients with incident CV events the LE and EYLL of the non-ESRD/non-diabetes group ranged from 2 08-8 07 and 3 07-10 25 years in patients aged ? 65 years respectively and 4 09-17 28 and 6 60-22 45 years in those aged under 65 Diabetes and ESRD would result in additional losses of 0 57-6 59 and 2 32-8 74 EYLL while de novo diabetes after ESRD led to the highest EYLL After multivariate adjustment we found a monotonic increase in the relative hazard with the presence of diabetes ESRD and both The effect of de novo diabetes after ESRD on mortality was still the highest especially among people followed for more than 4 5 years Finally we performed the cost-effectiveness analysis between different dialysis modalities 4285 pairs of incident HD and PD patients were selected from a national cohort after matching for clinical characteristics and propensity scores Lifetime survival and healthcare expenditure were estimated by data of 14-year follow-up and subsequently extrapolated to lifetime under the assumption of constant excess hazard A cross-sectional EQ-5D survey were performed in 12 dialysis units which resulted in 179 matched pairs The product of survival probability and the mean utility value at each time point estimated by kernel-smoothing method were summed up throughout lifetime to obtain the quality-adjusted life expectancy (QALE) The results revealed the estimated life expectancy between HD and PD were nearly equal (19 11 versus 19 08 years) The mean utility values QALE were also similar whereas average lifetime healthcare costs were higher in HD than PD (237 795 versus 204 442 USD) and the cost-effectiveness ratios for PD and HD were 13 681 and 16 643 USD per quality-adjusted life year respectively In conclusion our results indicated diabetes and ESRD synergistically increase risks of CV events Diabetes especially de novo diabetes after ESRD and ESRD itself are both associated with increased risk of mortality and extra loss of LE after various CV events Intensive proactive and/or reactive preventions are warranted to reduce the CV events and potential loss of life in these cases When considering the cost-effectiveness between dialysis modality PD is a more cost-effective therapy than HD of which the major determinant was the total medical costs paid for the dialysis modality and its associated complications

Date of Award	2016 Jul 19
Original language	English
Supervisor	Junne-Ming Sung (Supervisor)