61

INTRODUCTION 62

COMPLEXITY OF THE PATIENT POPULATION 64cardiovascular comorbidity · selected comorbidities

PRE-ESRD EPO USE & ANEMIAAT INITIATION 66pre-esrd epo use & hemoglobin at initiation ·geographic variations in pre-esrd epo use &hemoglobin at initiation

BIOCHEMICAL & PHYSICAL CHARACTERISTICS 68blood urea nitrogen · serum creatinine · estimatedglomerular filtration rate · body mass index

ESTIMATED GFRS & OUTCOMES 70gfr by charlson score · event curves for hospitaliza-tion & survival · percent survival

INSURANCE COVERAGEOF INCIDENT PATIENTS 72geographic variations in insurance coverage ·patient distribution by insurance coverage

SUMMARY 74

Chapter threePatient characteristics

...I have a duty to speak the truth as I see it and to

share not just my triumphs, not just the things that

felt good, but the pain, the intense, often unmitigating

pain. It is important to share how I know survival is

survival and not just a walk through the rain.

Audre Lorde, Black Women Writers at Work

3

chap

ter

62

20

03

a

nn

ua

l

da

ta

r

ep

or

t

P A T I E N T C H A R A C T E R I S T I C S

The complexity of the ESRD popu-lation has continued to evolve over the last two decades. In addition to dra-

matic changes in patient distribution by age, gender, and race, comorbid conditions have had an

increasingly profound impact on patient survival after the start of therapy. In 1980 diabetes was the cause of

ESRD in only 13.1 percent of new patients (see Reference Table A.1). Now, however, it is the primary diag-

nosis in 64 percent of Hispanic-Mexican patients, 55 percent of Hispanics of other origins, and 43 percent

of non-Hispanics. � Patient selection for ESRD treatment also began to alter dramatically when in 1982

the payment system for dialysis changed to a prospective system under the “Composite Rate.” In the next

decade the proportion of ESRD patients initiating dialysis with a primary diagnosis of diabetes increased

more than 2.5 times, to 33.7 percent. And in the 1990s this proportion increased by another one-third, to

44.3 percent. � The USRDS has been assessing the broader prevalence of diabetes not only as a primary

contributor to ESRD, but as a complicating condition. In the 2002 ADR, the chapter by the Cardiovascular

SSC evaluated the graded association between cardiovascular mortality and diabetes as a primary cause of

ESRD, diabetes as a secondary complicating condition, and diabetes documented from medical services one

year after the start of ESRD therapy. Understandably, patients with diabetes as a primary diagnosis have the

most widespread organ system disease and the lowest survival. Those with diabetes as a secondary condi-

tion have the next lowest survival, and survival is best in relatively new diabetes patients. This year we have

determined the total burden of diabetes—both primary and secondary—in patients who survive the first

year of ESRD therapy, finding it to be 80 percent in Native Americans, 73 percent in Hispanics, 61 percent

in Asians, 59 percent in blacks, and 58 percent in whites. Overall, 59 percent of new patients enter the pro-

gram with some form of diabetes, making the burden of the disease far greater than previously considered.

� Diabetic and non-diabetic patients carry other comorbid conditions as well, the occurrence of which has

evolved over time. In patients surviving the first year of therapy, for example, the proportion with cardio-

vascular comorbidity increased dramatically in the late 1980s; growth has since slowed, but rates are still

rising. There is also an increased number of patients entering therapy with cancer as a complicating diagno-

sis, most evident in patients with multiple myeloma (see Figure 2.31). These patients clearly add to the

CHAPTER HIGHLIGHTS � Figure 3.3 In 2000, 45 percent of diabetics and 37 percent of non-diabetics who survived one year on dialysis had cardio-

vascular comorbidity, compared to only 25 and 18 percent in 1984. � Figure 3.8 The mean hemoglobin at initiation rose from 9.2 g/dl in May 1995

to 10.0 g/dl in May 2002, a 9 percent increase, while the percent of patients using EPO in the pre-ESRD period grew from 22 to 32. � Figure 3.11 The

percent of patients initiating dialysis with a hemoglobin less than 11 g/dl decreased from 84 in May 1995 to 74 in May of 2002, a 12 percent reduc-

tion. Even more dramatically, the percent initiating with a hemoglobin level less than 10 g/dl fell from 68 to 53, a 22 percent reduction over the

seven-year period.

63

1995 1996 1997 1998 1999 2000 2001 2002

Perc

ent o

f pat

ient

s

40

50

60

70

80

90

All

Native American

White Black

Asian Hispanic

Percent of patients0 20 40 60 80 100

Hispanic-Mexican

Hispanic-other

Non-Hispanic

Diabetes

Hypertension

Glomerulonephritis

Cystic kidney

Other/unknown

3.1 · Primary diagnosis of new ESRD patients at initiation, by Hispanic ethnicityincident ESRD patients, 1998–2002 combined; data from Medical Evidence form.

complexity of the population under ESRD treatment. Information

on the Medical Evidence form allows us to assess some of the care pro-

vided to chronic kidney disease patients prior to ESRD. Only thirty-two

percent of new ESRD patients, for example, receive EPO treatment be-

fore beginning therapy. And while mean hemoglobins at initiation rose

overall from 9.2 g/dl in May 1995 to 10 g/dl in June 2002, these levels

are still well below the minimum target set by the National Kidney

Foundation’s K/DOQI guidelines of 11 g/dl. The lowest hemoglobin

levels continue to occur in children and in blacks. Serum creatinine

levels at initiation are declining, while estimated glomerular filtration

rates (eGFRs) are increasing—indicating, perhaps, that patients are

beginning ESRD therapy earlier in the course of their chronic kidney

disease. We have assessed the association between eGFR and incident

patient comorbidity, and found a direct relationship which suggests

that this earlier initiation of therapy may be due to increased comorbidity, with conditions such as fluid

overload and congestive heart failure amplifying the need for renal replacement therapy. Based on these

findings, eGFR at initiation appears to reflect severity of disease. Because risks of mortality and hospitaliza-

tion are greater for those with high eGFR levels at the start of therapy, these clinical associations complicate

any interpretation of earlier dialysis initiation. Additionally, there are emerging concerns over the accu-

racy of the estimating equation from the Modification of Diet in Renal Disease Trial (MDRD). Severely

malnourished patients with marked loss of muscle mass may appear to have lower serum creatinines and

higher eGFRs, which would in turn be associated with a lower survival. Additional research is needed to

validate this formula in patients with severe renal failure to determine the best method of assessing residual

renal function. Insurance coverage of the incident population may have important implications for the

delivery of care in both the pre-ESRD period and the early months of dialysis. The greatest proportion of

patients with no coverage at the start of therapy occurs among blacks, followed by Hispanics and Asians. At

initiation, almost 60 percent of ESRD patients are covered by Medicare, Medicaid, or a combination of

both. These dual-eligible patients, many of whom live in the southern states, may be particularly vulnerable

to cutbacks in Medicaid payments, particularly for prescription drugs. They may also be disproportionately

affected by shortfalls in the 2003 state budgets. The composition of the incident ESRD population con-

tinues to change, placing increasing demands on the provider care system. This complexity affects not only

the individual patients, but also the insurance coverage needed to pay for their care. In future USRDS stud-

ies we will focus on the degree of insurance coverage and how it influences care and associated outcomes.

3.2 · Incident pts with primary & secondary diabetes, by race/ethnicityincident ESRD patients with a first ESRD service date betweenMay 1995 & June 2002, surviving one year plus 90 days; diabetesdetermined from inpatient claims for diabetes within one year +90 days of first service date, or indicated on the Medical Evidenceform as a primary diagnosis or comorbid condition.

3

chap

ter

64 P A T I E N T C H A R A C T E R I S T I C S

1984 1986 1988 1990 1992 1994 1996 1998 2000

Pe

rce

nt

of

pa

tie

nts

ho

spit

aliz

ed

Pe

rce

nt

of

pa

tie

nts

wit

h c

om

orb

idit

y

0

10

20

30

40

50

1984 1986 1988 1990 1992 1994 1996 1998 2000

Diabetics Non-diabetics

CHF

ISHD

PVD

ISHD

PVD

All CV disease All CV disease

CHF

0

10

20

30

40

50

60

White

Black

Native American

Asian

Hispanic

All

All cardiovascular disease Congestive heart failure

1995 1996 1997 1998 1999 2000 2001 2002 1995 1996 1997 1998 1999 2000 2001 20020

10

20

30

40Ischemic heart disease Peripheral vascular disease

Complexity of thepatient populationBecause comorbid conditions were not included on the Medical

Evidence form until its revision in 1995, it is difficult to assess

long-term trends in the comorbidity of patients beginning ESRD

therapy. We are able, however, to use inpatient hospitalization

records to track cardiovascular complications since 1984. In dia-

betic patients surviving at least one year after the start of ESRD,

the percent of patients hospitalized with cardiovascular disease

increased 25 percent between 1990 and 2000; for congestive

heart failure, ischemic heart disease, and peripheral vascular

disease, growth was approximately 50 percent (Figure 3.3). The

3.3 · Cardiovascular comorbidity in incident patients who survive one year following ESRD initiation; from REBUS inpatient diagnosis codes

3.4 · Cardiovascular comorbidity at initiation, by race/ethnicity; from the Medical Evidence form

20

03

a

nn

ua

l

da

ta

r

ep

or

t

percent of non-diabetics hospitalized with cardiovascular disease

increased 17 percent overall. Between 1996 and 2000, however,

rates for both patient cohorts leveled off.

Since 1995, the proportion of new ESRD patients with a cardio-

vascular comorbidity reported on the Medical Evidence form

increased from 48.6 to 50.6 percent of the population (Figure

3.4). The amount of congestive heart failure was generally steady

across racial and ethnic categories, while that of peripheral vascu-

lar disease decreased in most populations—1.4 percent overall,

and 14.5 percent among Asians. Rates of ischemic heart disease

increased 15.9 percent overall, and nearly 22 percent in blacks.

The greatest changes, however, occurred among Native Ameri-

cans. Between 1995 and 2002 the proportion of these patients

65

0

1

2

3

4

5

6

7

8White

Black

Native American

Asian

Hispanic

All

Cancer Alcohol dependence

Smoking Chronic obstructive pulmonary disease

Pe

rce

nt

of

pa

tie

nts

wit

h c

om

orb

idit

y

0

2

4

6

8

10

1995 1996 1997 1998 1999 2000 2001 20020

1

2

3

4

5

6

1995 1996 1997 1998 1999 2000 2001 20020.0

0.5

1.0

1.5

2.0

2.5Inability to ambulate Inability to transfer

3.5 · Selected comorbidities at initiation, by race/ethnicity; from the Medical Evidence form

with any kind of cardiovascular comorbidity increased 33 per-

cent, and 64 percent for those with ischemic heart disease. The

differences between the number of patients beginning ESRD

therapy with a cardiovascular comorbidity, and those with car-

diovascular hospitalizations in the first year, points to the poor

survival of these patients in the first year of treatment (see pages

109 and 111).

Most patients beginning ESRD therapy with a diagnosis of cancer

are white (Figure 3.5). Between 1995 and 2002, rates of cancer

among new patients decreased for Native Americans, but in-

creased for other patients—as much as 43 percent for patients of

Hispanic ethnicity. Rates of alcohol dependence and smoking

decreased among almost all populations, though smoking in-

creased 4.3 percent among Asians and 23 percent among Native

Americans. Rates of COPD decreased slightly for Asians, but

among Native Americans increased 61 percent. In 2002, patients

were more likely to be able to ambulate on their own and to ar-

rive at therapy without assistance; the proportion of patients un-

able to ambulate or transfer decreased by this time for all patients

but those of Asian descent, in whom they increased 11.4 and 6.7

percent, respectively.

Because comorbidity reporting on the Medical Evidence form is

voluntary, reported comorbidities can be underestimated by

30–50 percent. These trends therefore need to be viewed with

caution, and supporting data are needed to assess them more

completely.

� Figure 3.3 incident ESRD patients surviving one year plus 90 days after ini-tiation; hospitalization data from REBUS inpatient hospitalization diagnosiscodes. Because the USRDS is now using a different method of identifying primarypayors, data in this figure differ slightly from those presented in Figure 2.4 of the2002 ADR. � Figures 3.4–5 incident ESRD patients with a first service datebetween May 1995 & June 2002; data from Medical Evidence form.

3

chap

ter

66

20

03

a

nn

ua

l

da

ta

r

ep

or

t

P A T I E N T C H A R A C T E R I S T I C S

All White Black N Am Asian Hispanic

Pe

rce

nt

of

pa

tie

nts

re

ceiv

ing

EP

O

0

10

20

30

40

50

All White Black N Am Asian Hispanic

Male Female

1996 1997 1998 1999 2000 2001 2002

He

mo

glo

bin

(g

/dl)

9.0

9.2

9.4

9.6

9.8

10.0

10.2

10.4

Pe

rce

nt

rece

ivin

g E

PO

20

22

24

26

28

30

32

34

Hgb:with EPO

Hgb: without EPO Hgb: overall

% receiving EPO

All 0-19 20-44 45-64 65-74 75+

Pe

rce

nt

of

pa

tie

nts

re

ceiv

ing

EP

O

0

10

20

30

40

50

All 0-19 20-44 45-64 65-74 75+

Male Female

36.8 (43.5)31.5 to <36.827.6 to <31.523.9 to <27.6below 23.9 (20.5)

36.8 (43.5)31.5 to <36.827.6 to <31.523.9 to <27.6below 23.9 (21.2)

36.8 (42.3)31.5 to <36.827.6 to <31.523.9 to <27.6below 23.9 (19.8)

1996

2002

1996

2002

1996

2002

1996

2002

3.7 · EPO use prior to initiation, by race/ethnicity & gender

3.8 · Mean hemoglobin at initiation, by EPO treatment

Pre-ESRD EPO use & hemoglobin at initiation

3.6 · EPO use prior to initiation, by age & gender

3.9 · Geographic variations in the percent of patients receiving EPO prior to initiation, by race

Pre-ESRD EPOuse & anemia atinitiationWhile the proportion of patients who

receive EPO prior to ESRD therapy has

increased since 1996, rates of EPO use

remain low (Figures 3.6–7), particularly

in relation to the low hemoglobin levels

seen in patients beginning therapy. By

age, male pediatric patients are the most

likely to receive pre-ESRD EPO, and male

patients age 20–44 the least. Among fe-

males, rates are more consistent across

ages. By race, the lowest rates of EPO

therapy occur among black and Native

American males, while the highest are

seen in females of Asian ancestry.

Since early 1995 the overall mean hemo-

globin at initiation has increased from 9.2

to 10 g/dl (Figure 3.8). Patients treated

with EPO prior to beginning ESRD

therapy have hemoglobin levels 0.4–0.5

g/dl higher than those receiving no EPO.

The percentage of patients receiving EPO

prior to initiation increased from 21.6 to

32.4 between 1995 and 2002.

Between May 1995 and June 2002, the

percent of patients with initial hemoglo-

bins above 11 g/dl—the K/DOQI target—

rose from 15.9 to 26.1, a 64 percent

increase (Figure 3.11). The proportion of

patients with levels of 12 g/dl and above

rose 80 percent.

Patients age 75 and above have the highest

hemoglobin levels at initiation, while lev-

els are lowest in pediatric patients (Figure

All White Non-white

67

10-<11

12+

1996 1997 1998 1999 2000 2001 2002P

erc

en

t o

f p

ati

en

ts

0

10

20

30

40

50

60

70

80

90

100

9-<10

<9

11-<12

All White Black N Am Asian Hispanic All White Black N Am Asian Hispanic8.4

8.8

9.2

9.6

10.0

10.4 Male Female

1996 1997 1998 1999 2000 2001 20029

10

11

12

Hemodialysis

Peritoneal dialysis

Transplant

All

All 0-19 20-44 45-64 65-74 75+ All 0-19 20-44 45-64 65-74 75+

He

mo

glo

bin

(g

/dl)

He

mo

glo

bin

(g

/dl)

He

mo

glo

bin

(g

/dl)

8.4

8.8

9.2

9.6

10.0

10.4Male Female

10.25+ (10.39) 10.11 to <10.25 9.96 to <10.11 9.78 to <9.96 below 9.78 (9.60)

10.25+ (10.40) 10.11 to <10.25 9.96 to <10.11 9.78 to <9.96 below 9.78 (9.66)

10.25+ (10.43) 10.11 to <10.25 9.96 to <10.11 9.78 to <9.96 below 9.78 (9.47)

1996

2002

1996

2002

1996

2002

1996

2002

Mean hemoglobin at initiation

3.11 · Patient distribution, by mean monthly hemoglobin (g/dl) at initiation

3.12 · Mean hemoglobin at initiation, by age & gender

3.13 · Mean hemoglobin at initiation, by race/ethnicity & gender

3.14 · Mean hemoglobin at initiation, by first modality

3.10 · Mean hemoglobin (g/dl) at initiation

All

White

Non-white

3.12). By race, hemoglobin levels are high-

est among white patients (Figure 3.13).

Among males they are lowest in blacks

and Hispanics, and among females the

lowest levels are found in blacks and Na-

tive Americans.

Patients who begin ESRD therapy on he-

modialysis consistently have the lowest

initial hemoglobin levels, while those

starting therapy with a transplant have the

highest (Figure 3.14).

� Figures 3.6–3.8, 3.11–3.14 incident ESRDpatients with a first service date between May 1995& June 2002; data from Medical Evidence form.� Figures 3.9–10 incident ESRD patients, 2000–2001 combined, by HSA, unadjusted; data fromMedical Evidence form.

3

chap

ter

68

20

03

a

nn

ua

l

da

ta

r

ep

or

t

P A T I E N T C H A R A C T E R I S T I C S

95 96 97 98 99 00 01 02

BU

N (

mg

/dl)

80

85

90

95

100 White

Black

N Am

Asian

Hispanic

All

95 96 97 98 99 00 01 0280

85

90

95

100 0-19

20-44

45-64

65-74

75+

All

BU

N (

mg

/dl)

95 96 97 98 99 00 01 02

Se

rum

cre

ati

nin

e (

mg

/dl)

6

8

10

12 White

Black

N Am

Asian

Hispanic

All

95 96 97 98 99 00 01 02

Se

rum

cre

ati

nin

e (

mg

/dl)

5

7

9

11

13

0-19

20-44

45-64

65-74

75+

All

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

BU

N (

mg

/dl)

60

70

80

90

100

110 White Native AmericanBlack Asian Hispanic

Se

rum

cre

ati

nin

e (

mg

/dl)

5

7

9

11

13White Native AmericanBlack Asian Hispanic

1996

2001

1996

2001

Blood urea nitrogen (BUN) at initiation

3.15 · by race/ethnicity 3.16 · by age

3.17 · by race/ethnicity & age

Serum creatinine at initiation

3.18 · by race/ethnicity 3.19 · by age

3.20 · by race/ethnicity & age

Biochemical& physicalcharacteristicsBetween 1995 and 2002, mean blood urea

nitrogen (BUN) levels among new ESRD

patients decreased 7.7 mg/dl (Figure 3.15).

In black and Asian patients, this decrease

was 8.5 and 8.0 mg/dl, respectively. By age

group, the largest change was seen in pa-

tients age 75 and older, whose mean BUN

levels fell 9.3 percent, from 95.1 to 86.3 mg/

dl (Figure 3.16). Similar patterns are appar-

ent in data on initial serum creatinine lev-

els, which decreased 1.7 mg/dl for blacks

during this period, and 1.4 mg/dl for pa-

tients age 75 and older (Figures 3.18–19).

Patients with low serum creatinine levels

are often assumed to have high levels of

residual renal function. This is not always

the case, however, as many of these pa-

tients may have poor nutrition and low

muscle mass. Caution should therefore be

used in interpreting the data as a reflec-

tion of earlier initiation.

The hypothesis that patients are beginning

therapy at an earlier stage of chronic kid-

ney disease may, however, be supported by

the steady increase in estimated glomeru-

lar filtration rates (eGFRs) at initiation

(Figures 3.21–23). Since 1995 the mean

eGFR at the beginning of ESRD therapy

has increased 2.1 ml/min/1.73 m2 overall,

to 9.6. By race, the lowest increase—1.6

ml/min/1.73 m2—has occurred in Asians,

while the mean eGFR of whites has in-

creased 2.2 ml/min/1.73 m2. Pediatric pa-

tients tend to have the highest eGFRs, and

patients age 20–44 the lowest. The greatest

increase in initial levels since 1995 has

occurred in patients age 75 and older,

from 8.2 to 10.7 ml/min/1.73 m2. Con-

cerns have recently been raised, however,

regarding the methods used to estimate

GFR in patients with advanced renal fail-

ure (particularly diabetic patients, for

whom the formulas have not been vali-

dated), and these should be acknowledged

when considering the data.

Body mass index (BMI) among new

ESRD patients increased an average of 1.8

69

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

All0-19

20-4445-64

65-7475+

95 96 97 98 99 00 01 02

eG

FR

(m

l/m

in/1

.73

m2)

eG

FR

(m

l/m

in/1

.73

m2)

eG

FR

(m

l/m

in/1

.73

m2)

6

7

8

9

10

11 White

Black

Native American

Asian

Hispanic

All

95 96 97 98 99 00 01 026

7

8

9

10

11

65-74

75+

All

0-19

20-44

45-64

95 96 97 98 99 00 01 02

BM

I (kg

/m2)

22

24

26

28

30White

Black

Native American

Asian

Hispanic

All

95 96 97 98 99 00 01 02

BM

I (kg

/m2)

20

22

24

26

28

30 0-19

20-44

45-64

65-74

75+

All

4

6

8

10

12 White Native AmericanBlack Asian Hispanic

BM

I (kg

/m2)

18

20

22

24

26

28

30White Native AmericanBlack Asian Hispanic

1996

2001

1996

2001

Estimated glomerular filtration rate (eGFR) at initiation

3.21 · by race/ethnicity 3.22 · by age

3.23 · by race/ethnicity & age

Body mass index (BMI) at initiation

3.24 · by race/ethnicity 3.25 · by age

3.26 · by race/ethnicity & age

kg/m2 between 1995 and 2002 (Figure

3.24), a 7 percent increase. Asian and pe-

diatric patients consistently have the low-

est BMIs, and their levels changed the least

over the period. Native American patients

and those age 45–64, in contrast, present

with the highest BMIs, and their indices

have increased 2.5 and 1.9 kg/m2, respec-

tively.

These increasing BMIs in ESRD patients

are a source of concern. Chronic ESRD

patients constitute a biased sample of pa-

tients, as they are healthy enough to have

survived to ESRD. It is not clear, therefore,

how trends in BMIs should be interpreted,

particularly when high indices in the gen-

eral population are associated with an

elevated risk of death, while the reverse is

true in the population with ESRD.

� Figures 3.15–26 incident ESRD patients with afirst service date between May 1995 & June 2002;data from Medical Evidence form.

� Figures 3.22–23 eGFR calculation for ages 0–18from Schwartz et al., & for ages 19 & above fromLevey et al.

3

chap

ter

70

20

03

a

nn

ua

l

da

ta

r

ep

or

t

P A T I E N T C H A R A C T E R I S T I C S

eG

FR

(m

l/m

in/1

.73

m2)

7

8

9

10

11

12

13

65-<70

70-<75

75-<80

80-<85

85+

Age Gender

All 0 1 2 3 4 5 6 7+ All 0 1 2 3 4 5 6 7+7

8

9

10

11

12

13

White

Black

Native American

Asian

Charlson score

Hispanic

Non-Hispanic

Race Ethnicity

All

Male

Female

Estimated glomerularfiltration rates & outcomesOn the previous spread we illustrate recent changes in biochemi-

cal markers at the initiation of ESRD therapy. It is unclear, how-

ever, whether lower BUN and creatinine levels and higher eGFRs

show that patients are beginning dialysis earlier because they

have greater comorbidity, or because they are receiving better

care in the pre-ESRD period. A number of investigators have

attempted to interpret eGFR at initiation as an indication of early

referral for dialysis. This idea may, however, be complicated by

the advanced comorbidity linked in observational studies to

higher eGFR levels. To investigate this question we look here at

eGFR levels in relation to comorbidity, and also at outcomes in

the twelve months following the start of ESRD therapy.

We calculated Charlson scores (a measure of comorbidity) for

patients age 67 or older who had two years of medical services

prior to starting ESRD therapy in 2000. In both males and fe-

3.27 · Differences in eGFR at initiation, by Charlson score, age, gender, race, & ethnicity

males, across age groups, and across most races and ethnicities,

there is an almost linear relationship between higher Charlson

scores and higher eGFR levels (Figure 3.27). This provides strong

evidence that lower BUN and creatinine levels at the initiation of

therapy, along with higher eGFR levels, are related not to im-

proved pre-ESRD care, but rather to increased comorbidity.

This hypothesis that higher eGFRs at the initiation of ESRD

therapy are a surrogate for increased comorbidity and, therefore,

predict increased mortality competes with the hypothesis that

earlier initiation of ESRD therapy leads to improved survival. To

test these theories, we used probability models to determine the

risk of hospitalization and mortality. We obtained Medical Evi-

dence form data on age, gender, race, primary diagnosis, and

ethnicity, and calculated body mass index and eGFR from values

supplied on the form. We then looked at outcomes in a one-year

followup period.

Probabilities of hospitalization and mortality follow similar pat-

terns, increasing with higher eGFR levels (Figures 3.28–30). Pa-

71

Months

1 2 3 4 5 6 7 8 9 10 11 12

Pro

ba

bil

ity

of

rem

ain

ing

wit

ho

ut

a h

osp

ita

liza

tio

n

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

eGFR <5

eGFR 5-<7

eGFR 7-<10

eGFR 10+

1 2 3 4 5 6 7 8 9 10 11 12

Unadjusted (Kaplan-Meier) hospitalization Adjusted hospitalization

Months

1 2 3 4 5 6 7 8 9 10 11 12

Pro

ba

bil

ity

of

surv

iva

l

0.7

0.8

0.9

1.0

eGFR <5

eGFR 5-<7

eGFR 7-<10

eGFR 10+

1 2 3 4 5 6 7 8 9 10 11 12

Unadjusted (Kaplan-Meier) survival Adjusted survival

Months

1 2 3 4 5 6 7 8 9 10 11 12

Pro

ba

bil

ity

of

surv

iva

l

0.6

0.7

0.8

0.9

1.0

eGFR <5

eGFR 5-<7

eGFR 7-<10

eGFR 10+

All

1 2 3 4 5 6 7 8 9 10 11 12

BlackWhite

3.28 · Event-free probability for first-year first hospitalization, by eGFR at initiation

3.29 · Survival probability, by eGFR at initiation

3.30 · Survival probability, by eGFR at initiation & race

tients with the highest eGFRs at the start

of dialysis therapy are the most likely to be

hospitalized in the first year of ESRD, and

the least likely to survive that year. This

pattern is consistent in both white and

black patients, though blacks have higher

survival rates. These data, then, strongly

support the hypothesis that eGFR at ini-

tiation is a surrogate for disease severity,

and a predictor of poor patient outcomes.

Placed within a clinical context, it seems

reasonable to assume that patients and

physicians are more likely to initiate dialy-

sis at later stages if patients are in other

ways stable. It is equally likely that physi-

cians initiate patients earlier due to indi-

cations of fluid overload, acidosis,

hyperkalemia, declining nutritional status,

or pericarditis, the usual indications for

starting renal replacement therapy. These

types of selection bias, intrinsic in obser-

vational data, complicate any interpreta-

tion of outcomes associated with eGFR.

� Figure 3.27 incident ESRD patients age 67 &older, 2000. Estimated GFR levels in older & frailpatients may be inaccurate secondary to low musclemass. Charlson score calculations use claims fromtwo years prior to first ESRD date. Estimated GFRcalculated using the four-variable Levey formula,with age, gender, race, & creatinine taken from theMedical Evidence form. The Charlson score is ameasure of overall comorbidity; categories of comor-bidity are determined using inpatient hospitalclaims. � Figures 3.28–29 incident dialysis pa-tients, 1999–2000 combined; adjusted data areadjusted for age, gender, race, ethnicity, primarydiagnosis, & BMI. � Figure 3.30 incident dialysispatients, 1999–2000 combined; unadjusted.

3

chap

ter

72

20

03

a

nn

ua

l

da

ta

r

ep

or

t

P A T I E N T C H A R A C T E R I S T I C S

10.44+ (12.51) 8.31 to <10.44 6.62 to <8.31 5.23 to <6.62 below 5.23 (4.42)

17.0+ (20.8) 14.1 to <17.0 11.7 to <14.1 9.5 to <11.7 below 9.5 (8.0)

34.8+ (40.5) 27.6 to <34.8 22.5 to <27.6 17.2 to <22.5 below 17.2 (13.8)

15.3+ (20.4) 11.8 to <15.3 10.2 to <11.8 8.8 to <10.2 below 8.8 (7.6)

9.58+ (14.61) 6.90 to <9.58 5.36 to <6.90 4.08 to <5.36 below 4.08 (3.33)

13.04+ (16.81) 9.86 to <13.04 7.84 to <9.86 6.10 to <7.84 below 6.10 (5.08)

3.31 · Geographic variations in insurance coverage at initiation (percent of patients)

Insurance coverageof incident patientsThe highest proportion of ESRD patients covered solely by Medic-

aid live in California, Arizona, portions of Washington and Oregon,

and areas of the Gulf Coast and Atlantic Seaboard; coverage is al-

most three times higher in these areas than in the lowest quintile

(Figure 3.31). Medicare-only coverage is most common in the

South and along the East Coast, and least common in Upper Mid-

DVA or other EGHP with Medicare as secondary payor

Medicaid only Medicare and Medicaid

Medicare with other secondary payor Medicare only

west and western states; patients here are more likely to have

combined coverage from Medicare and another payor.

Patients whose Medicare coverage is accompanied by EGHP or

other insurance coverage are most likely to be white, while mi-

nority populations are disproportionately represented among

patients with Medicare or Medicare/Medicaid coverage, and

among those with no insurance coverage (Figure 3.32). The dis-

tribution of Hispanic patients by insurance coverage is similar to

that of other minority populations.

73

0

20

40

60

80

100

Insurance type (see table for codes)

None DVA/oth EGHP EGHP + M M/caid M & M M w/oth M/care

Pe

rce

nt

of

pa

tie

nts

0

5

10

15

20

Race

Hispanic ethnicity

95 96 97 98 99 00 01 020

20

40

60

80

100

Pe

rce

nt

of

pa

tie

nts

0

20

40

60

80

100All White Black

Native American Asian Hispanic

95 96 97 98 99 00 01 02 95 96 97 98 99 00 01 02

Unknown

Asian

Native American

Black

White

None

DVA/oth

EGHP

EGHP + M

M/caid

M & M

M w/oth

M/care

3.32 · Patient distribution, by insurance coverage at initiation, race, & ethnicity

3.33 · Trends in insurance coverage at initiation, by race/ethnicity

Insurance groupsNone · No insuranceDVA/oth · DVA or

otherEGHP · Employer

Group HealthPlan only

EGHP + M · EGHPwith Medicaresecondary

M/caid · Medicaidonly

M & M · Medicare &Medicaid

M w/oth · Medicarewith othersecondary

M/care · Medicareonly

Between 1995 and 2002 the proportion of

patients covered solely by Medicaid at the

start of therapy fell 12.3 percent, to 10.8

percent (Figure 3.33). Almost one-fifth of

Native Americans, however, were covered

by Medicaid in 2002, an increase of more

than 50 percent. Similar dramatic increases

occurred in other types of coverage for

Native Americans, as the number with no

insurance dropped from more than a third

to only 7 percent. The number of Asian

patients with no insurance dropped almost

43 percent during the same period.

� Figure 3.31 incident ESRD patients, 2001, byHSA, unadjusted; data from Medical Evidence form.� Figure 3.32 incident ESRD patients, 2001; datafrom Medical Evidence form. � Figure 3.33 incidentESRD patients with a first service date between May1995 & June 2002; data from Medical Evidence form.

3

chap

ter

74

summarychapter

20

03

a

nn

ua

l

da

ta

r

ep

or

t

P A T I E N T C H A R A C T E R I S T I C S

Maps: National means & patient populationsFigure number 3.9 3.9 3.9 3.10 3.10 3.10 3.31 3.31 3.31 3.31 3.31 3.31

All White Non-wh. All White Non-wh. DVA M/EGHP M/caid M & M M w/oth M/care

Overall value for all pts 32.1 33.6 29.5 9.9 10.1 9.7 12.4 7.2 10.8 11.9 21.1 13.8

Total patients 41,881 26,599 15,282 39,183 24,806 14,377 11,728 6,818 10,218 11,213 19,925 13,053

Overall value for pts mapped 33.0 34.2 30.5 10.0 10.1 9.7 11.8 7.7 9.5 12.6 22.3 14.4

Missing HSA/state: pts dropped 8,369 4,356 4,013 7,892 4,076 3,816 1,401 108 1,918 204 383 481

INTRODUCTION Figure 3.1 Diabetes as a primary diagnosis accounts for

64 percent of ESRD cases in Mexican Hispanics, 55 percent in Hispanics of

other origins, and 43 percent in non-Hispanics. Figure 3.2 The total burden

of diabetes, both primary and secondary, in ESRD patients surviving one year

is 80 percent for Native Americans, 73 percent for Hispanics, 61 percent for

Asians, 59 percent for blacks, and 58 percent for whites. COMPLEXITY OF THE

PATIENT POPULATION Figure 3.3 In 2000, 45 percent of diabetics and 37

percent of non-diabetics who survived one year on dialysis had cardiovascu-

lar comorbidity, compared to only 25 and 18 percent in 1984. EPO USE &

ANEMIA TREATMENT Figures 3.6–7 EPO use prior to dialysis is lowest in

those age 20–44, and in blacks and Native Americans. Figure 3.8 The mean hemoglobin at initiation rose from 9.2 g/dl in

May 1995 to 10.0 g/dl in May 2002, a 9 percent increase, while the percent of patients using EPO in the pre-ESRD period

grew from 22 to 32. Figure 3.11 The percent of patients initiating dialysis with a hemoglobin less than 11 g/dl decreased

from 84 in May 1995 to 74 in May of 2002, a 12 percent reduction. Even more dramatically, the percent initiating with a

hemoglobin level less than 10 g/dl fell from 68 to 53, a 22 percent reduction over the seven-year period. Figure 3.12

Female pediatric patients and those age 20–44 continue to have the lowest hemoglobins at the initiation of ESRD treat-

ment—9.1 and 9.4 g/dl, respectively. Figure 3.13 Native American women and black patients of both genders have the

lowest hemoglobin levels at the start of therapy. Figure 3.14 At the beginning of therapy, hemoglobin levels in perito-

neal dialysis patients are more than one-half g/dl higher than those in hemodialysis patients. BIOCHEMICAL & PHYSICAL

CHARACTERISTICS Figures 3.15–20 Since 1995, blood urea nitrogen and serum creatinine levels have fallen by 8.2 and

17.4 percent, respectively. Figures 3.21–23 Estimated glomerular filtration rates have increased 27 percent over the last

eight years, so that individuals are starting therapy on average with 9.6 ml/min/1.73 m2 of residual renal function. Fig-

ures 3.24–26 Overall body mass index has increased 7 percent over the last seven years, from 25.7 to 27.5 kg/m2. ESTI-

MATED GFR & OUTCOMES Figure 3.27 Estimated GFRs and Charlson scores appear to be linearly related within age,

gender, race, and ethnicity. Figures 3.28–30 Survival probability appears to be inversely related to eGFR. INSURANCE

COVERAGE OF INCIDENT PATIENTS Figures 3.31–33 Minority populations are disproportionately covered by Medic-

aid only, Medicare and Medicaid, and Medicare without secondary insurance. Hispanic populations also contain a large

proportion of patients who have no insurance coverage at the initiation of dialysis.