Long-term data on the ESRD population continue to show a flattening of incident rates, with a growth of only 2.0 percent in the number of new patients between 2004 and 2005. Data suggest that increased use of ACE inhibitors and ARBs, better glycemic control in diabetic patients, and better control of blood pressure may be contributing to this stabilization. Whether the trend continues depends on the impact of the baby boomers as they reach retirement age.

Five percent of incident dialysis patients in 2005 were returning after a failed kidney transplant or the recovery of kidney function. Peritoneal dialysis accounted for 6.6 and 5.3 percent of the incident and prevalent dialysis populations, proportions that have continued to decline over the past decade from peaks of 13 and 11 percent. The number of kidney transplants reached a high of 17,424 in 2004, while the prevalent transplant population, at 143,693, rose 5.6 percent from 2004 to 2005, despite continued growth in the number of dialysis patients on the transplant wait list.

Data on clinical care show that the percent of prevalent dialysis patients with a functioning fistula grew to 41.2 percent in 2004, exceeding the target of 40 set by K/DOQI in 1997, and placement rates have continued to climb. The mean hemoglobin level was 12.0 g/dl through December, 2005, with only 18 percent of patients having a single month’s hemoglobin less than 11 g/dl—the lowest number since epoetin was introduced in June, 1989. Transfusion rates, not reported in many years, have reached their lowest level since the introduction of ESAs in 1989, at just 0.51 percent of the prevalent dialysis population in 2005 compared to 19 percent in 1980. The use of preventive care, however, still needs improvement. Twenty-one percent of patients with diabetes receive no HbA1c testing, and 37 percent no lipid testing, within a year.

Incident-based mortality rates continue to fall for those on treatment two or more years, with even first-year rates for hemodialysis patients showing a decline when adjusted for age, gender, race, and cause of ESRD. With more complete adjustments based on data from the Medical Evidence form, including BMI, hemoglobin, and eGFR, the first-year death rate has not changed over the past decade. Use of dialysis catheters is still an important issue, since over 80 percent of patients have a catheter at their first outpatient dialysis.

Faster growth in general Medicare versus ESRD costs has led to a decline in the percent of the Medicare budget spent on ESRD. The program, however, still grew 9.1 percent, in absolute dollars, from 2004 to 2005.

Costs for CKD patients with diabetes and CHF are rising in both the general Medicare and dually-enrolled populations, and racial differences in the prevalence of and costs associated with CKD, diabetes, and CHF support the observation that ESRD disproportionately affects African Americans.

The impact of CKD and ESRD is thus far greater than previously seen, and obviously of central importance in the budgets of state departments of health. Detection and prevention as well as treatment may improve through greater awareness of these diseases and through use of Part D medication coverage.

Trends in ESRD patient counts and spending  

During 2005, 106,912 new dialysis and transplant patients began ESRD therapy, for an adjusted rate of 347 per million population (Table p.a). More than 485,000 patients were receiving treatment on December 31, 2005, for an adjusted rate of 1,569 per million population—341,319 were on dialysis, and 143,693 had a functioning transplant. A total of 17,429 kidney transplants were performed during the year, and 85,790 patients died.

Between the 1997–2001 and 2001–2005 periods, the average annual percent change in rates per million for the hemodialysis population fell from 3.4 to 0.5 among incident patients, and from 4.4 to 1.7 among prevalent patients. In the transplant population, incident growth slowed from 7.3 to 2.9 percent, and prevalent growth from 4.7 to 3.8. Both incident and prevalent rates in the peritoneal dialysis population continue to fall, though not as quickly as in the late 1990s.

Medicare spending per patient year rose 3.2 percent between 2004 and 2005; after adjustments for inflation, however, spending actually fell 0.2–0.9 percent. Total Medicare costs reached $21.3 billion, while non-Medicare costs rose to $10.7 billion. Medicare spending per ESRD patient year in 2005 reached $59,417—$69,758 for hemodialysis, $50,847 for peritoneal dialysis, and $25,015 for transplant.

While the number of new dialysis patients has grown steadily since the late 1980s, the annual rate of growth has slowed, staying under 4 percent since 2000 (Figure p.2; because these data are from CMS’s annual End-Stage Renal Disease Facility Survey, they differ slightly from the numbers in Table p.a). The number of patients returning to dialysis from transplant reached 5,357 in 2005, and in this decade has been growing at rates between 2 and 13 percent. Fewer patients restart dialysis each year, but these numbers have been increasing as well, to 2,479 in 2005. The total number of patients starting or restarting dialysis in 2005, then, reached nearly 111,000—2.3 percent more than in the previous year.

The incident ESRD population in 2005 was 13.1 percent larger than in 2000, with a one-year growth after 2004 of 2.0 percent (Figure p.3). The prevalent populations have grown at higher rates—20.5 and 32.3 percent after 2000 for dialysis and transplant, respectively, and 3.3 and 5.6 percent after 2004.

In 2005, total ESRD costs reached $32 billion—1.6 percent of the nearly $2 trillion spend by the United States on healthcare that year (Figure p.5). As total healthcare dollars have increased, this proportion has grown as well, from 10.3 percent in 1991

Trends in Modalities  

The 2005 incident ESRD population numbered 106,442 (excluding those with unknown modality; Figure p.6). Growth since 2000 has been nearly 13 percent, but ranges from a 40 percent increase in the number of new patients beginning with a transplant to an 8 percent fall in the number starting therapy with peritoneal dialysis.

While the number of patients has grown, however, the overall incident rate has flattened out during the past several years, rising only 2.8 percent since 2000 to reach 347 per million population. By modality, the rate for patients beginning therapy on hemodialysis has risen 4.1 percent. The rate for transplant patients, in contrast, has grown 17.5 percent, while for those on peritoneal dialysis it is nearly 15 percent lower than in 2000.

In terms of the prevalent population, there were nearly 484,000 patients receiving ESRD therapy at the end of 2005—24 percent more than in 2000 (Figure p.7). Nearly 144,000 of these patients had a functioning transplant. But while this is 32 percent higher than in 2000, the prevalent transplant population still accounts for nearly the same proportion of the entire prevalent population as it did five years ago—29 percent, compared to 28 percent in 2000.

In 2005, the prevalent rate of ESRD was 1,569 per million population, 12 percent higher than in 2000. The rate for hemodialysis patients has grown at nearly the same pace, rising 10 percent during the five years to reach 1,016 in 2005. As in the incident population, the most dramatic changes have been in the peritoneal dialysis and transplant populations, with the rate falling 6 percent in the former (to 84), and increasing nearly 21 percent in the latter (to 466).

Geographic patterns show that incident and prevalent hemodialysis rates in 2005 were highest in the Gulf Coast states and states located along the eastern seaboard, reaching an average of 465 and 1,663 per million population, respectively, in the upper quintile (Figure p.8). Peritoneal dialysis rates for incident patients are highest in North Dakota, Montana, Idaho, the southern states of Arkansas, Mississippi, Alabama, and South Carolina, and in West Virginia and Connecticut, and average 38.9 in the upper quintile. Prevalent rates are highest in Idaho and the aforementioned southern and eastern states as well, in addition to New Mexico, and average 131.7 in the upper quintile.

The highest transplant rates, both incident and prevalent, are evident in the same geographic areas—namely states located in the Upper Midwest and Great Lakes area, reaching 13.2 and 628 per million population, respectively, in the upper quintile. Prevalent transplant rates are also highest for patients in Pennsylvania and Delaware.

At the end of 2005, there were 64,943 patients listed for a kidney-only transplant, an increase of 7.6 percent from the previous year (Figure p.9). The numbers of white and African American patients on the wait list grew 7.2 and 6.2 percent, while for patients of other races the list grew 14.2 percent.

The median wait time for a transplant is 669 days overall; this ranges, however, from 557 days among white patients to 852 for African Americans and 804 for patients of other races—53 and 44 percent longer, respectively, than for whites.

Counts of patients on the wait list for renal transplantation are highest in California, Texas, Alabama, Florida, North Carolina, Illinois, Michigan, Pennsylvania, and New Jersey, and in these areas average over 4,200 in the upper quintile. Several of these states have high waiting times as well. In California, Alabama, Illinois, and New York, for example, patients in the upper quintile wait an average of nearly three years for a transplant

Arteriovenous (AV) fistulas are linked to lower rates of infection and increased dialysis efficiency, and the NKF’s Kidney Disease Outcomes Quality Initiative (K/DOQI) calls for more fistula use. Between 1998 and 2004, the proportion of patients using fistulas increased from 27.8 to 41.2 percent (Figure p.11). Rates of fistula placements per 1,000 patient years, moreover, have more than doubled since 1991, from 54.4 to 122.1 , and catheter use, which peaked in the late 1990s, has returned to early 1990s levels; graft use has fallen 52.0 percent since 1991 (Figure p.12).

Infectious complications per patient year are highest  in those with catheters as their first access, and sepsis complications follow suit (Figure p.13). Peritonitis rates in PD patients were 1.6 per patient year in 2004, 26 and 17 percent higher than infection and sepsis rates in those with catheters.

Dialysis adequacy, as represented by the urea reduction ratio (URR) and Kt/V, has improved dramatically (Figures p.14–15). In 1993, only 43 percent of hemodialysis patients had a URR of 65 or greater (corresponding to a delivered Kt/V of 1.2). Since then, the proportion of patients with URRs meeting or exceeding 65 has more than doubled, and in 2004 reached 89 percent. Adequacy in PD patients has improved as well. In 1996, 45 percent had a mean weekly Kt/V of 2 or more; by 2005 this had grown to 72 percent.

K/DOQI guidelines formerly set a target hemoglobin of 11–12 g/dl; in 1991, nearly 85 percent of prevalent dialysis patients fell below this level (Figures p.16–17). Mean levels have since risen steadily, and by the end of 2005 only 19 percent of patients were below the target. Recent clinical trials (CHOIR and CREATE) have raised concerns about high hemoglobin levels, citing evidence of potential harm. The USRDS has shown that, as of 2005, the mean hemoglobin in prevalent dialysis patients is 12 g/dl, with more than one in five patients exceeding 13 g/dl in any single month. New K/DOQI guidelines in 2006 indicated no benefit related to morbidity and mortality with a hemoglobin above 13 g/dl; and 2007 recommendations returned to the target of 11 g/dl. We further address hemoglobin “overshooting” in Chapter Five.

The rise in mean hemoglobin can be directly related to higher EPO doses. As of 2005, the mean weekly dose exceeded 19,000 units (Figure p.17). These costs present a serious fiscal challenge to the ESRD program (See Chapter Eleven, Figures 11.21–22).

Figure p.18 presents data showing changes in the use of blood transfusions since the early days of the dialysis program. In the late 1970s and early 1980, more than 15 percent of outpatient dialysis patients received at least one transfusion during a three-month period. After EPO’s introduction, however, this number fell, and has remained at less than 1 percent since 2002.

Trends in hospitalization and mortality  

Adjusted event probabilities show that, at the end of one year, the probability of hospitalization among new dialysis patients is three times greater than the probability of death—0.66 versus 0.24 (Figure p.20).

Geographic variations in hospital admissions per patient year show no dramatic change from 1993 (Figure p.21). Rates have fallen slightly in the western half of the country, but in regions represented by the lower quintile, the mean number of admissions is almost unchanged, at 1.42 in 1993 and 1.40 in 2005. In areas represented by the upper quintile, mean admissions have decreased from 2.57 to 2.46.

All-cause hospitalization rates for prevalent ESRD patients have remained stable, changing only 1.8 percent since 1993 (Figure p.22). Hospitalizations for cardiovascular disease have increased by 7.8 percent, while infectious hospitalizations continue to rise, and in 2005 were 26.8 percent higher than in 1993. This latter growth appears in part to be related to utilization rates for catheters which, while decreasing, remained more than three times higher than the rate for fistulas in 2005. In hemodialysis patients, all-cause hospitalization rates have remained steady, while rates for cardiovascular disease and infection have increased by 10.7 and 38.6 percent, respectively. The continuing rise in infectious hospitalizations is disturbing, and warrants further investigation by the USRDS. Hospitalizations for events related to vascular access have decreased by 34 percent, perhaps reflecting a shift in outpatient prescriptions.

In the peritoneal dialysis population, all-cause hospitalization rates have decreased by 8.9 percent since 1993, and rates for cardiovascular disease have fallen 7.5 percent. Hospitalizations for vascular access have fallen 22.5 percent since 1999.

Rates have also decreased in the transplant population—6.1 percent for all-cause hospitalizations, and 21.9 percent for those related to cardiovascular disease. Rates for infectious hospitalizations, however, which had been relatively steady, rose sharply in 2005. It is too early to determine if this spike is a significant phenomena, one that will require increased vigilance by the USRDS.

Overall, prevalent mortality rates for dialysis patients have been slowly declining in the past five years, and patient vintage (time on ESRD therapy) has a dramatic effect on these rates (Figure p.23). In the mid-1980s, for example, patients with less time on dialysis had higher mortality rates compared to those with longer vintage. In 1985, the rate in those with less than two years on hemodialysis was 282.9 per 1,000 patient years, compared to 238.8 in patients on the therapy for five years or more. In 1992, mortality rates for the youngest and oldest vintages overlapped, and thereafter the lowest rates have occurred in patients of younger vintage. Mortality in this population fell 27 percent between 1985 and 2005, in contrast to a 7 percent increase in patients of longer vintage. This variation over time is more pronounced in the peritoneal dialysis population. In patients with less than two years of therapy, for instance, mortality rates decreased 47 percent—from 374.2 to 198.3—between 1985 and 2005. Conversely, for patients with a vintage of five or more years, the rate grew 14.2 percent, from 349.4 to 399.0.

Since 1980, first-year ESRD mortality rates have fallen 26 percent, from 305.3 to 225.3 per 1,000 patient years (Figure p.24). The rate of decline has recently slowed, however—to 4.8 percent between 1996 and 2004, and to 4.1 since 2000. Dramatic changes have occurred in the peritoneal dialysis and transplant populations, in which first-year mortality rates have fallen by 61 and 82 percent since 1980, in 2004 standing at 165.6 and 69.4 per 1,000 patient years at risk; in patients treated with peritoneal dialysis, changes in this group may in part be due to selection bias. In this year’s chapter on Emerging Issues we provide a detailed examination of first-year mortality rates and some of the factors which may influence them.

Between the 1991–1995 and 1996–2000 periods, five-year survival probabilities increased slightly across all modalities (Figure p.25). Patients who initiate ESRD therapy with a transplant remain more than twice as likely as their dialysis counterparts to survive five years, with a probability in 1996–2000 of 0.75, compared to 0.35 in hemodialysis patients and 0.34 in those on peritoneal dialysis. In the early years of treatment, patients on peritoneal dialysis have a slight survival advantage over those treated with hemodialysis; at approximately three years, however, this advantage disappears, and their probabilities of survival became nearly equal.

Trends in ESRD expenditures  

Per person per year (PPPY) costs during the transition to ESRD spike sharply in the month following initiation (Figure p.26). In non-diabetics, for example, costs in this month are 2.5 times higher than in the month prior, at $14,155 compared to $5,640. Similar growth occurs for patients with diabetes, with costs rising from $5,216 to $13,745.

Medicare spending for ESRD reached $21.3 billion in 2005, a 9.1 percent increase from 2004 (Figure p.27). Since 2002, ESRD spending has accounted for 6.4 percent of the Medicare program, which reached a total of $331.4 billion in 2005.

In 2005, diabetic patients age 65 and older were the most expensive to care for, with PPPY costs nearing $75,000; costs in patients this age but with no diabetes were nearly 17 percent less, at $64,094 (Figure p.28). In patients younger than 65, diabetic costs were more than a third higher than those for non-diabetic patients, at $62,094 compared to $46,705.

Total ESRD expenditures in 2005, which exceeded $19 billion, were more than three-fold higher than those in 1991, and inpatient, outpatient, and physician/supplier costs have grown in a similar fashion (Figure p.29). The largest growth has occurred in the outpatient setting, with a nearly four-fold increase from $1.9 billion in 1991 to $7.2 billion in 2005. Inpatient costs rose from $2.2 billion to nearly $7 billion, while costs for physican/supplier services rose from $1.2 billion to $4.1 billion.

PPPY costs for ESRD grew 67 percent between 1991 and 2005, from $35,821 to $59,681 (Figure p.30). Costs for other services have followed suit, rising 43.2, 83.0, and 57.8 percent for inpatient, outpatient, and physician/supplier services, respectively.

By modality, PPPY cost increases have been highest for services related to transplant (Figure p.31). In 2005, for example, physician/supplier costs for patients with a functioning graft were more than three times higher than those in 1991—at $6,815 versus $2,108. And outpatient costs for patients experiencing a graft failure more than doubled between 1991 and 2005.

Costs for injectables appear to be moderating, evidenced by only a 2 percent rise in ESA expenditures between 2004 and 2005 (Figure p.32). Costs for IV vitamin D hormone and IV iron fell 12.7 and 18.3 percent, respectively. Overall PPPM costs for injectables are greatest in the southeastern states, parts of the Upper Midwest, and the Northwest (Figure p.33).

Figure p.34 shows how PPPM costs between 2004 and 2005 were affected by reimbursement changes—an increase in dialysis payments, and a decrease in those for non-ESA injectables.

Total Medicare expenditures for patients on dialysis topped $17 billion in 2005, growth of 6.6 percent over the previous year (Table p.b). Costs for patients with a functioning graft reached $1.14 billion, and those for patients receiving a transplant rose to $984 million—one-year growth of 10.8 and 6.5 percent, respectively. For those with a functioning graft this growth was slightly higher than in the 2003–2004 period; for patients on dialysis or with a transplant during the year, in contrast, growth slowed from 9.8 and 11.6 percent, respectively. Total per patient year Medicare costs for transplant events within the year reached nearly $103,000 in 2005, a high cost balanced out by yearly expenditures of only $17,273 for patients with a functioning graft. These costs were 2.3 and 3.8 percent higher, respectively, than in 2004; per patient year dialysis costs rose 3.4 percent, to $68,585, while costs for patients with a graft failure during the year rose 4 percent to $79,704.