Chapter 6: Mortality
- Mortality rates continue to decrease for dialysis and transplant patients, having fallen by 28% and 40%, respectively, since 1996.
- Adjusted mortality rates in 2013 per 1,000 patient-years were 138, 169, and 35 for ESRD, dialysis, and transplant patients, respectively. By dialysis modality, mortality rates were 172 for hemodialysis patients and 152 for peritoneal dialysis patients, per 1,000 patient-years.
- Patterns of mortality during the first year of dialysis differ substantially by modality. For hemodialysis patients, reported mortality is very high in month 2, but declines thereafter. In contrast, mortality rises slightly over the course of the year for peritoneal dialysis patients.
- The relationship between race and mortality differs considerably by age among dialysis patients. Among dialysis patients younger than 45 years old, Whites have mortality rates comparable to Blacks; however, in older age groups, Whites have higher mortality rates than their Black counterparts.
- Dialysis patients continue to have substantially higher mortality, and fewer expected remaining life years, compared to the general population and Medicare populations with cancer, diabetes, or cardiovascular disease. However, the relative and absolute decline in mortality for dialysis patients in the past 15 years has been greater than for these other diagnostic groups.
Mortality analyses in this chapter are based on both end-stage renal disease (ESRD) data and general population data. ESRD data are from the USRDS ESRD Database. General population data are based on the Medicare 5 percent standard analytical files and U.S. Census mortality data. Note that universal reporting of ESRD patient deaths to the Centers for Medicare & Medicaid (CMS) is required via CMS form 2746 as a condition of coverage for dialysis units and transplant centers. In addition, mortality ascertainment is augmented by Social Security Death Master File data to the extent allowed by regulation.
For analyses in this chapter, the term “incident” refers to patients new to ESRD, while “prevalent” refers to patients receiving ESRD treatment on a specific date, and “period prevalent” includes patients treated for ESRD over a period of time. Modality is assigned as of the earliest date within the range used in the analysis, without use of the 60-day stable modality rule (i.e., the requirement of 60 days on a modality for change in modality assignment) or the 90-day rule for outcomes (attribution of outcomes for up to 90 days after a change in modality).
The decline in mortality shown in this chapter has important implications for both patients and resource allocation, as increasing ESRD patient lifespan is likely contributing to the ongoing increase in the size of the prevalent ESRD population.
See the ESRD Analytical Methods chapter for an explanation of analytical methods used to generate the figures and tables in this chapter.
Mortality Among ESRD Patients, Overall and by Modality
Overall mortality rates among ESRD (dialysis and transplant) patients continue to decline, with steeper declines in more recent years. Over the last two decades, the adjusted death rate fell by 7% from 1996 to 2003, and by 23% from 2004 to 2013 (Figure 6.1.a). The trend was similar for dialysis (hemodialysis and peritoneal dialysis) patients, with the mortality rate falling by 5% from 1996 to 2003 and by 23% from 2004 to 2013. Among transplant patients, mortality fell by 12% from 1996 to 2003 and by 28% from 2004 to 2013. Since 1996, the net reduction in mortality was 30% for all ESRD patients, including 28% for dialysis patients and 40% for transplant patients.
By dialysis modality, among hemodialysis patients the adjusted mortality rate fell by 2% from 1996 to 2003 and by 22% from 2004 to 2013. Among peritoneal dialysis patients, the mortality rate fell by 21% from 1996 to 2003 and by 34% from 2004 to 2013 (Figure 6.1.b). The net reductions in mortality from 1996 to 2013 were 25% for hemodialysis patients and 49% for peritoneal patients.
Adjusted mortality rates in 2013 were 138, 169, and 35 per 1,000 patient-years for ESRD, dialysis, and transplant patients, respectively. By dialysis modality, mortality rates were 172 for hemodialysis patients and 152 for peritoneal dialysis patients, per 1,000 patient-years.
Figure 6.1 Adjusted all-cause mortality (deaths per 1,000 patient-years) by treatment modality (a) overall, dialysis, and transplant, and (b) hemodialysis and peritoneal dialysis, for period-prevalent patients, 1996-2013
Mortality by Duration of Dialysis, Including Trends Over Time
Among hemodialysis patients, from 1996-2011 the average yearly death rate was highest during the first year, then dropped to its lowest point during the second year, and then tended to rise for more than 5 years afterward (Figure 6.2). Among peritoneal dialysis patients, mortality rates tended to increase over the first five years after starting dialysis. For both hemodialysis and peritoneal dialysis patients, mortality rates tended to be higher after 5 years than between 2-5 years on dialysis. The patterns of death rates according to time since dialysis initiation have been fairly similar over calendar time (comparing cohorts based on calendar year of initiation of treatment), within modality.
Figure 6.2 Adjusted all-cause mortality (deaths per 1,000 patient-years) by treatment modality, cohort (year of ESRD onset), and number of years after start of dialysis among incident (a) hemodialysis patients and (b) peritoneal dialysis patients, 1996, 2001, 2006, and 2011
Mortality During the First Year of ESRD
Among patients starting hemodialysis in 2012, reported all-cause mortality peaked at 400 deaths per 1,000 patient-years in month 2, and decreased thereafter to 200 per 1,000 patient-years in month 12 (Figure 6.3). Note that the steep rise in hemodialysis mortality rates between months 1 and 2 may reflect ‘data reporting issues’; e.g., some patients who die soon after starting dialysis related to ESRD might not be registered as being ESRD and included in the CMS database (Foley et al., 2014). The extent to which this occurs is currently unknown.
Among patients with peritoneal dialysis as initial renal replacement modality, mortality does not peak early but instead tends to increase gradually during the first year on dialysis. Mortality at month 12 among these patients was 119 per 1,000 patient-years. Peritoneal dialysis patients may not experience an early peak in mortality, in part, because patients beginning ESRD via peritoneal dialysis are a highly selected group, in many cases being younger, healthier, and having undergone substantial pre-ESRD planning.
Post-transplant mortality among the <2% of patients who initiate ESRD treatment with a kidney transplant peaks in month 1, followed by a generally decreasing trend for the remainder of the first year (not shown).
Figure 6.3 Adjusted mortality (deaths per 1000 patient-years) by treatment modality and number of months after treatment initiation among ESRD patients, 2012
Mortality by Age and Race
Mortality rates among ESRD patients increase with rising age, as expected. Mortality rates differ by race, but this difference is not constant within age groups or by modality. For example, White patients on dialysis had comparable mortality rates to Black/African American patients among those aged 0-44 years old, but higher mortality than Blacks at older ages.
Table 6.1 Adjusted all-cause mortality (deaths per 1,000 patient-years) by patient age and race among ESRD patients, 2012
Cause-Specific Mortality Rates
The largest category of known cause-specific mortality for dialysis patients is deaths due to cardiovascular disease (CVD), which comprises 41% of the deaths and 53% of the deaths with known causes. The cause of death information (based on CMS 2746) is missing or unknown for 23% of dialysis patients and 68% of transplant patients.
Table 6.2 Unadjusted percentages of deaths due to cardiovascular disease (CVD), infection, other specified causes, and with missing data, by modality among ESRD patients, 2012
Survival Probabilities for ESRD Patients
Despite improvements in survival on dialysis over the years, adjusted survival for hemodialysis patients who were incident in 2008 is only 55% at three years after ESRD onset (Table 6.3.a.). For peritoneal dialysis patients, adjusted survival is 66% at three years. These results illustrate the extreme vulnerability of these patients relative to the general population.
Survival has improved between the 2000 and 2008 incident ESRD cohorts for all modalities. For example, five-year survival rose from 35% to 40% among hemodialysis patients, from 37% to 50% among peritoneal dialysis patients, from 66% to 75% among deceased donor transplant patients, and from 75% to 87% among living donor transplant patients. Adjusted survival was consistently higher in the transplant population than in dialysis patients, and among living donor transplant recipients than deceased donor recipients.
In the 2008 incident ESRD cohort, adjusted survival was consistently higher (Table 6.3.b) among younger patients, among Asians and Blacks compared to other races, and patients among with primary cause of ESRD designated as glomerulonephritis compared to patients with diabetes or hypertension.
Table 6.3 Adjusted survival (%) by (a) treatment modality and incident cohort year (year of ESRD onset), and (b) age, sex, race, and primary cause of ESRD, for ESRD patients in the 2008 incident cohort (initiating ESRD treatment in 2008)
Expected Remaining Lifetime: Comparison of ESRD Patients to the General U.S. Population
The differences in expected remaining lifetime between the ESRD and general populations are striking (Table 6.4). Dialysis patients younger than 80 years old are expected to live less than one-third as long as their counterparts without ESRD, and dialysis patients aged 80 years and older are expected to live less than one-half as long as their counterparts without ESRD. Transplant patients fare considerably better, with expected remaining lifetimes for people under the age of 75 estimated at 67% to 84% of expected lifetimes in the general population.
Table 6.4 Expected remaining lifetime (years) by age, sex, and treatment modality of prevalent dialysis patients, prevalent transplant patients, and the general U.S. population (2012), based on USRDS data and the National Vital Statistics Report (2013)
Mortality Rates: Comparisons of ESRD Patients to the Broader Medicare Population
Comparison to the General Medicare Population
Dialysis patients over the age of 75 years experienced mortality rates 3.9 times higher for males and 3.8 times higher for females than for males and females in the general Medicare population (Table 6.5). Among kidney transplant patients, mortality rates were 2.5-3.3 times higher than for the general Medicare population aged 65-74, and 1.4 times higher at age 75 and older.
Table 6.5 Adjusted mortality (deaths per 1,000 patient-years) by age, sex, treatment modality, and Medicare comorbidity among ESRD patients and people covered by Medicare in 2013, based on USRDS and CMS data, 2013
Comparison to Comorbidity-Specific Medicare Patients
From 1996 to 2013, adjusted mortality among ESRD patients aged 65 years and older declined by 48%, from 338 to 174 per 1,000 patient-years (Table 6.6). Among dialysis patients, adjusted mortality fell 37%, from 349 to 219. Among transplant patients, adjusted mortality fell 9%, from 79 to 72. The decline in mortality for dialysis patients was greater than for other major diagnostic groups, including cancer, diabetes, CHF, CVS/TIA, and AMI. Adjusted mortality fell 34% for cancer and 32% for diabetes, but somewhat less for cardiovascular conditions, at 23% for heart failure, 27% for cerebrovascular accident/transient ischemic attack (CVA/TIA), and 20% for acute myocardial infarction (AMI).
In 2013, mortality rates among dialysis patients aged 65years and older ranged from 1.7 times higher than for heart failure patients to 4.0 times higher than for patients with diabetes. For transplant patients aged 65 and older, the mortality rate was within the range of mortality rates for Medicare patients with the other listed conditions.
Table 6.6 Adjusted mortality (deaths per 1,000 patient-years) by calendar year, treatment modality, and Medicare comorbidity among ESRD patients and comorbidity-specific Medicare populations aged 65 & older, 1996-2013
Foley RN, Chen SC, Solid CA, Gilbertson DT, Collins AJ. Early mortality in patients starting dialysis appears to go unregistered. Kidney Int 2014;86(2):392-398.
National Vital Statistics Report. Table 7. Life expectancy at selected ages, by race, Hispanic origin, race for non-Hispanic population, and sex: United States, 2012. 2015;63(9):29. http://www.cdc.gov/nchs/data/nvsr/nvsr63/nvsr63_09.pdf. Accessed October 2, 2015.