1 · chronic kidney disease - usrds home … 3 48 chronic kidney disease 1 introduction chronic...

18
1 chronic kidney disease The wrong of unshapely things is a wrong too great to be told; I hunger to build them anew and sit on a green knoll apart, With the earth and the sky and the water, remade, like a casket of gold For my dreams of your image that blossoms a rose in the deeps of my heart. william butler yeats “the lover tells of the rose in his heart”

Upload: lyhanh

Post on 25-May-2019

232 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

1chronic kidney disease

The wrong of unshapely things is a wrong too great to be told;

I hunger to build them anew and sit on a green knoll apart,

With the earth and the sky and the water, remade, like a casket of gold

For my dreams of your image that blossoms a rose in the deeps of my heart.

william butler yeats“the lover tells of the rose in his heart”

Page 2: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

CCHRONIC KIDNEY DISEASE3

48

1 Introduction

Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the 2002 publication of the National Kidney Foundation’s Chronic Kidney Disease Guidelines. This docu-ment created a CKD classification system, using five stages defined by increasing evidence of kid-ney damage, as shown by microalbuminuria and estimated glomerular filtration rates (eGFRs).

Applying this new classification system to NHANES III data, Coresh and Levey et al. esti-mated that eight million people in the U.S. had an eGFR less than 60 ml/min/.73 m2, and another 2 million had evidence of microalbuminuria.

The USRDS uses diagnosis codes from hos-pital and outpatient encounters to determine the size of the CKD population—a method which utilizes only those codes identified by providers. In the Précis we define the CKD population over a two-year period, looking at those who survive until the end of the year, and considering them point prevalent at the start of the next year. To approximate a period prevalent CKD population, we also include new patients identified with CKD in the next year; expenditures are determined in the second year.

In this chapter, however, we use a slightly dif-ferent definition for CKD patients, requiring sur-vival for the entire second year in order to give each patient equal opportunity for access to care. This definition may exclude the sickest CKD pa-tients, since they may have early mortality.

Figure . shows that in almost two-thirds of Medicare patients with CKD—identified by di-

agnosis codes over the two-year period—CKD is accompanied by diabetes, CHF, or both diseases combined. In employed CKD patients under 65, a similar proportion has diabetes, but far fewer patients have CHF alone or with diabetes. Over the past decade the CKD population has become more complex, with rates of diabetes alone and in combination with CHF growing 38 and 26 per-cent, respectively, in the Medicare population, while in the employed population rates of CKD accompanied by diabetes, CHF, or both diseases have grown 9–3 percent.

Figure .4 illustrates these changes in further detail. The percent of CKD patients with diabetes increased in the Medicare population from 6.7 in 992–993 to 23.7 in 2002–2003, and in the EGHP population from 23.8 in 999–2000 to 26.6 in 2002–2003. While the percent whose CKD is ac-companied by heart failure has remained steady or fallen, the burden of CKD with both diabetes and heart failure has grown.

Few patients who are at risk for CKD receive optimal assessment. Testing for microalbumin-uria or proteinuria is done in only 20 percent of diabetic patients, and testing rates are even lower in diabetic patients who also have CHF.

Preventive care in CKD patients is similarly infrequent. Measurement of at least two glyco-sylated hemoglobins each year, as recommended by the American Diabetes Association, occurs in only 56 percent of Medicare patients with both diabetes and CKD, and 35 percent of the EGHP population with the same diagnoses.

50 · prevalence of CKD

52 · assessment of populations at risk for CKD

54 · preventive

healthcare monitoring in CKD patients

56 · prescription drug

therapy in CKD patients

geographic variations in the prevalence of CKD · interactions of diabetes, CHF, & CKD

creatinine · proteinuria · renal ultrasound · calcium & phosphorus · parathyroid hormone glycosylated hemoglobin & lipid testing · use of diabetic testing supplies · vaccinations ACE-Is/ARBs · beta blockers · calcium channel blockers · diuretics · lipid-lowering agents · diabetes drugs

58 · racial disparities in the assessment

of CKD

60 · hospitalization rates in CKD & non-

CKD patients

62 · acute kidney failure:

hospitalization & outcomes

64 · summary

assessments for renal function · drug therapy · geographic variations in assessment rates all-cause & cause specific hospitalization rates · geographic variations in hospitalization rates

probability of hospitalization for AKF · ESRD & mortality rates following hospitalization for AKF

Page 3: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

Figure 1.6 The overall probability of microalbuminuria or proteinuria testing reaches only

0.04 in the Medicare population, and 0.01 for EGHP patients. Figure 1.14 In the Medicare

CKD population, the probability of receiving at least two HbA1c tests within a year is 0.56,

compared to 0.36 for EGHP patients. Figure 1.19 The use of ACE-Is and ARBs for renal pro-

tective treatment has almost doubled in the CKD population. Figure 1.32 In the Medicare

population, the overall hospitalization rate for CKD patients reached 1,068 per 1,000 pa-

tient years in 2003—nearly three times greater than that in patients without CKD. Figure

1.39 For patients with a prior diagnosis of CKD, the rate of death in the first three months

after hospitalization is 2.7–4 times greater than the rate of an ESRD diagnosis.

Medicare

CKD only: 423,200 36%

CKD+CHF: 200,900

17%

CKD+DM: 299,600

25%

EGHP (age <65)

CKD only: 9,990 62%

CKD+CHF: 8295%

CKD+DM+CHF: 1,159

7%

CKD+DM: 4,16626%

CKD total: 16,144(0.37% of total EGHP)

CKD total: 1,177,560(3.89% of total Medicare)

CKD+DM+CHF:

253,860 22%

99 01 03

CKD+DMCKD+CHF

CKD+DM+CHFCKD other

93 95 97 99 01 03

Perc

ent

of p

atie

nts

0

10

20

30

40

50

60

70 Medicare EGHP

49

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

Chapter highlights

Limited data show that the use of ACE inhib-itors and ARBs for renal protective treatment has almost doubled. CKD patients with diabetes are almost twice as likely to receive this treat-ment, and those with CHF almost three times as likely, as those with other diagnoses.

All-cause hospitalization rates are almost 50 percent greater in CKD patients than in those without the disease. Of note is the fact that hos-pitalizations for pneumonia and bacteremia/septicemia have changed little over the last ten years, in contrast to the progress made in car-diovascular hospitalizations.

Growing comorbidity increases the likeli-hood of hospitalization for acute kidney failure (AKF). Patients with CKD, diabetes, and CHF combined, for example, have a 38 percent chance of hospitalization for AKF within the next three years, compared to only 4 percent in non-CKD patients. After an AKF hospitalization, patients with a history of CKD are five times more likely to die than to progress to ESRD; for those with-out prior CKD, the likelihood is 20 times greater.

Assessment of CKD patients will change based on public policy initiatives by the CDC and CMS. The CDC, for example, will be adding CKD sur-veillance programs. ICD-9-CM and ICD-0-CM diagnosis codes for individual stages of CKD have been approved, and will be introduced in October, 2005. And performance standards are being considered by both the National Center for Quality Assurance and the Joint Commission for Accreditation of Health Care Organizations to address standardized methods for measuring serum creatinines, calculating estimated GFRs, conducting urine albumin testing, and defining care through NKF’s clinical practice guidelines.

.2 · Diabetes & CHF interac-tions in patients with CKDMedicare: general Medicare CKD patients continuously enrolled in Medicare Parts A & B for an entire calendar year; patients en-rolled in an HMO or diagnosed with ESRD during the year are excluded. EGHP: CKD patients younger than 65 & continuously enrolled in a fee-for-service-plan for an entire calendar year.

. · Counts of diabetes & CHF in the CKD populationMedicare: general Medicare CKD patients continuously en-rolled in Medicare Parts A & B in 2003 (numbers estimated from 5 percent sample); patients enrolled in an HMO or diagnosed with ESRD during the year are excluded. EGHP: CKD patients younger than 65 & continuously enrolled in a fee-for-service-plan in 2003.

Page 4: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

25.2 + (29.4)22.3 to <25.219.8 to <22.317.2 to <19.8below 17.2 (15.2)

36.7 + (41.6)32.3 to <36.728.6 to <32.325.1 to <28.6below 25.1 (21.9)

61.4 + (68.7)53.9 to <61.447.3 to <53.941.2 to <47.3below 41.2 (36.5)

65.5 + (70.1)60.9 to <65.556.0 to <60.949.5 to <56.0below 49.5 (39.8)

CHRONIC KIDNEY DISEASE3

50

1

1.3 general Medicare patients continuously enrolled in Medicare Parts A & B during period

Geographic variations in the rate of CKD in Medicare patients; per 1,000 patients, by HSA

Prevalence of chronic kidney disease

ince 993 the rate of recognized chronic kidney disease (CKD) has more than doubled across the country (Fig-ure .3). Areas showing the highest CKD rates are located

in the Gulf Coast states and the eastern third of the country. These patterns are generally similar to those found in incident ESRD rates noted in Figure 2.8. In the mid-Atlantic states, for instance, rates of ESRD and CKD appear to be similar. There are, however, some nota-ble geographic differences. One, for example, occurs in the south-western states, where high rates of ESRD exist but rates of CKD are inconsistent. A variation such as this may reflect regional differ-ences in how CKD is identified, i.e., use of diagnosis codes versus the actual estimated glomerular filtration rate and stage of CKD.

Figure .4 illustrates changes over time in the general Medi-care and EGHP populations with CKD, both alone and in com-bination with diabetes and CHF. It shows, first of all, the steady growth of the overall recognized CKD population. The Medicare CKD population, for example, grew 46 percent between 992–993 and 2002–2003. In the EGHP population the same amount

of growth occurred in a much shorter time period, with the num-ber of CKD patients increasing 48 percent between 999–2000 and 2002–2003. In the Medicare population the greatest growth, of almost 250 percent, has been seen among patients with both CKD and diabetes; in the EGHP population growth has been highest among those with CHF as well.

The figure also shows the evolution of the disease burden in the recognized CKD population. In the general Medicare population, for example, the percentage of patients whose CKD is not accom-panied by either diabetes or CHF has dropped from 4 to 33 since the early 990s, while the percentage who have CKD together with congestive heart failure has fallen from 24 to 9. Diabetes, in con-trast, now has a larger impact on the CKD population, occurring in 24 percent of patients compared to 7 percent in 992–993. And the most complicated disease burden—CKD in combination with both diabetes and CHF—is increasing as well. Nineteen percent of CKD patients in 992–993 had both additional diagnoses; by 2002–2003 this had grown to nearly one in four patients.

1992–1993 1997–1998

2002–2003 Percent change 1992–93 to 2002– 03

S

Page 5: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

Medicare, 1992–1993

CKD only: 257,98040.7%

CKD+CHF: 149,900

23.6%

CKD+DM+CHF:

119,960 18.9%

CKD+DM: 106,000

16.7%

CKD total: 633,840(2.34% of total Medicare)

Medicare, 1997–1998

CKD only: 307,14035.4%

CKD+CHF: 192,480

22.2%

CKD+DM+CHF:

193,22022.2%

CKD+DM: 175,840

20.2%

CKD total: 868,680(3.38% of total Medicare)

EGHP, 1999–2000

CKD only: 4,23363.6%

CKD+CHF: 3525.3%

CKD+DM+CHF:

4807.2%

CKD+DM: 1,58623.8%

CKD total: 6,651(0.47% of total EGHP)

EGHP, 2002–2003

CKD only: 9,78959.4%

CKD+CHF: 9115.5%

CKD+DM+CHF: 1,4008.5%

CKD+DM: 4,38226.6%

CKD total: 16,482(0.66% of total EGHP)

Medicare, 2002–2003

CKD only: 520,52033.4%

CKD+CHF: 289,380

18.6%

CKD+DM+CHF:

380,30024.4%

CKD+DM: 368,780

23.7%

CKD total: 1,558,980(5.62% of total Medicare)

eg

nah c t

ne creP

0

50

100

150

200

250

300

All CKDCKD+DM

CKD+CHFCKD+DM+CHF

CKD other

Percent change in population size

Medicare: 1992-93 to 2002-03

EGHP: 1999-2000 to 2002-03

51

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

1.4 Medicare CKD patients, & EGHP CKD patients younger than 65

Trends in the interactions of diabetes, congestive heart failure, & CKD

{Figure .3} per ,000 patients, by HSA, unadjusted. General Medicare patients continuously enrolled in Medicare Parts A & B during two consecutive years; patients enrolled in an HMO or diagnosed with ESRD during the period are excluded. {Figure .4} Medicare: general Medicare CKD patients continuously enrolled in Medicare Parts A & B for two consecutive years (numbers estimated from 5 percent sample); patients enrolled in an HMO or diagnosed with ESRD during the period are excluded. EGHP: CKD patients younger than 65 & continu-ously enrolled in a fee-for-service-plan for two consecutive years.

Page 6: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

All NCKD Diabetes CHF DM+CHF Other

All NCKD Diabetes CHF DM+CHF Other

All NCKD Diabetes CHF DM+CHF Other

ytiliba

borP

0.0

0.1

0.2

0.3

Medicare

EGHP

Medicare

EGHP

Medicare

EGHP

ytiliba

borP

0.0

0.1

0.2

0.3

ytiliba

bor P

0.00

0.02

0.04

0.06

0.08

0.10

0.286 + (0.404)0.187 to <0.2860.144 to <0.187

0.108 to <0.144below 0.108 (0.075)

0.246 + (0.307)0.203 to <0.2460.171 to <0.203

0.134 to <0.171below 0.134 (0.106)

0.057 + (0.067)0.051 to <0.0570.045 to <0.051

0.038 to <0.045below 0.038 (0.030)

CHRONIC KIDNEY DISEASE3

52

11.5 prevalent Medicare

& EGHP patientsProbability of & geographic variations (by HSA) in the assessment

of one or more serum creatinine specific tests within a year

1.6 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in the assessment of one or more microalbuminuria or proteinuria tests within a year

Assessment of populations at risk for CKD

n this spread we examine the med-ical assessments received in both Medicare and Employer Group

Health Plan (EGHP) populations without an existing diagnosis of chronic kidney dis-ease (CKD), and focus on patients whose diabetes and/or congestive heart failure places them at high risk for CKD.

For all of the diagnostic tests presented here, Medicare patients are more likely than patients with EGHP coverage to receive test-ing. The probability of receiving a targeted serum creatinine to assess kidney function, for example, is 0.3 in Medicare non-CKD patients, compared to 0.03 in their EGHP counterparts (Figure .5). The probability rises to 0.24 for Medicare patients with both diabetes and CHF, but reaches only 0.3 in the equivalent EGHP population. Patients residing in the Upper Midwest, the Ohio Valley, areas in New England and areas along the Atlantic Coast have the highest probability of receiving a targeted serum creatinine test within a year.

The overall probability of microalbu-minuria or proteinuria testing reaches only 0.04 in the Medicare population, and 0.0 for EGHP patients (Figure .6). The great-est likelihood—0.22 and 0.9, respectively—occurs in patients with diabetes but with-out accompanying CHF. Patients residing in the western third of the country, the Upper Midwest, and northern New England are most likely to be tested.

In the non-CKD population the proba-bility of a renal ultrasound is 0.04 and 0.0 for Medicare and ESRD patients, respec-tively (Figure .7). It rises to 0.0 and 0.07 in patients with combined diabetes and CHF, and is slightly higher for those with CHF alone than for those with diabe-tes alone. Renal ultrasounds appear to be

1.7 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in the use of renal ultrasound within a year

Medicare: all NCKD diabetes patients

Medicare: all NCKD diabetes patients

Medicare: all NCKD diabetes patients

O

Page 7: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

yti liba

borP

0.00

0.02

0.04

0.06

0.08

0.10

0.12Medicare

EGHP

Medicare

EGHP

Medicare

EGHP

ytiliba

bor P

0.00

0.01

0.02

ytiliba

bo rP

0.0

0.1

0.2

0.3

0.4

All NCKD Diabetes CHF DM+CHF Other

All NCKD Diabetes CHF DM+CHF Other

All NCKD Diabetes CHF DM+CHF Other0.094 + (0.121)0.073 to <0.0940.062 to <0.073

0.050 to <0.062below 0.050 (0.039)

0.0137 + (0.0209)0.0101 to <0.01370.0064 to <0.0101

0.0031 to <0.0064below 0.0031 (0.0017)

0.452 + (0.554)0.373 to <0.4520.319 to <0.373

0.267 to <0.319below 0.267 (0.212)

53

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

1.9

1.10 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in the performance of a battery of assessment tests within a year

prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in the assessment of parathyroid hormone within a year

more highly utilized in the eastern half of the nation as well as in areas in the South-west and Gulf Coast states.

The probability of calcium and phos-phorus assessment is 0.06 in Medicare non-CKD patients overall, and 0.02 in the EGHP population (Figure .8). The greatest proba-bility of assessment occurs in patients with both diabetes and CHF, at 0.2 for Medi-care patients and 0.08 for those with EGHP coverage. Geographically, the likelihood of a calcium and phosphorous assessment appears to be highest in regions along the Atlantic Seaboard and in the Southwest, the Midwest, and the Ohio Valley.

Not unexpected, parathyroid hormone testing is rare in the non-CKD popula-tion (Figure .9). The probability of testing exceeds 0.0 only in Medicare patients with diabetes and in Medicare and EGHP patients with both diabetes and CHF. Patients in the western third of the country are the most likely to be tested.

In the non-CKD Medicare population, the probability of receiving a battery of assessment tests reaches 0.38, 0.24, and 0.37 in patients with diabetes, CHF, and both diagnoses combined, respectively (Figure .0). Compared to that in the EGHP pop-ulation, the likelihood of testing is nearly four times greater in Medicare patients overall and in those with other diagnoses, .4 times greater for those with diabetes, .6 times for those with diabetes and CHF combined, and twice as high for those with CHF alone. Nationwide, the probability of patients receiving a battery of tests within a year is highest in the Upper Midwest, the Ohio Valley and scattered areas throughout New England.

These data suggest that assessment strat-egies in the at risk population are markedly underutilized, particularly in the EGHP populations.

{Figures .5–0} general Medicare: patients enter-ing Medicare before January , 2002, alive & remain-ing in the program through December 3, & with-out CKD diagnosed during 2002. Patients enrolled in an HMO, with Medicare as secondary payor, or with ESRD diagnosed during the year are excluded. EGHP: patients enrolled for the entire year 2002 in a fee-for-service plan, younger than 65, & without a diagnosis of CKD during 2002. Patients diagnosed with ESRD before or during the year are excluded.

1.8 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in the assessment of calcium & phosphorous within a year

For both populations, diabetes, CHF, & other comor-bidities are defined in 2002. Patients censored at end of plan & end of 2003; prevalent Medicare patients also censored at death. All testing tracked in 2003. In Figure .5, data on serum creatinine testing obtained from individual tests & panels. Maps by HSA, unad-justed. {Figure .5} CPT codes used for assess-ment of serum creatinine include 80069 & 82565. {Figure .0} testing includes serum creatinine, microalbuminuria or proteinuria, & calcium & phos-phorous testing.

Medicare: all NCKD diabetes patients

Medicare: all NCKD diabetes patients

Medicare: all NCKD diabetes patients

Page 8: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

All CKD CKD+CHF CKD+other

ytiliba

borP

0.0

0.1

0.2

0.3

Medicare

EGHP

All CKD CKD+CHF CKD+other

ytiliba

borP

0.0

0.1

0.2

0.3

0.4

0.5

0.6

All CKD CKD+CHF CKD+other

ytiliba

borP

0.0

0.1

0.2

0.3

0.4

0.5

All CKDCKD+DM

CKD+CHFCKD+DM+CHF

CKD other

All CKDCKD+DM

CKD+CHFCKD+DM+CHF

CKD other

ytiliba

borP

0.0

0.1

0.2

0.3

0.4Medicare

EGHP

ytiliba

bo rP

0.00

0.02

0.04

0.06

0.08

0.10

0.285 + (0.318)0.257 to <0.2850.236 to <0.257

0.209 to <0.236below 0.209 (0.180)

0.084 + (0.094)0.076 to <0.0840.068 to <0.076

0.061 to <0.068below 0.061 (0.055)

0.23 + (0.31)0.20 to <0.230.14 to <0.20

0.09 to <0.14below 0.09 (0.05)

0.63 + (0.70)0.59 to <0.630.55 to <0.59

0.47 to <0.55below 0.47 (0.38)

0.56 + (0.65)0.50 to <0.560.47 to <0.50

0.40 to <0.47below 0.40 (0.31)

Medicare

EGHP

Medicare

EGHP

Medicare

EGHP

CHRONIC KIDNEY DISEASE3

54

1prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in microalbuminuria or proteinuria testing within a year in diabetic CKD patients

Preventive healthcare monitoring in CKD patients

1.14 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in glycosylated hemoglobin monitoring (at least 2 tests) within a year in diabetic CKD patients

1.15 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in diabetic CKD patients receiving diabetic testing strips within a year

1.13

diabetic preventive care

ata here illustrate the likelihood of, and geographic variations in, preven-tive healthcare monitoring given to

patients with chronic kidney disease. For all types of preventive care examined here, the probability of receiving care is considerably higher for Medicare patients than for those with EGHP coverage.

The probability of a CKD patient receiv-ing calcium and phosphorus testing is 0.27 for those with Medicare coverage, and 0.7 for those under EGHP; the probability is slightly higher for patients with diabetes as well (Figure .). Evaluation of bone and

mineral metabolism through parathryoid hormone testing is even less frequent, with the probability reaching a high of only 0. in Medicare patients with CKD, diabetes, and CHF combined (Figure .2). The probabil-ity of receiving a PTH assessment is only 0.08 nationwide.

For patients with diabetes, the prob-ability of microalbuminuria or protein-uria testing reaches only 0.9 in the Medi-care CKD population, and 0.6 in EGHP patients with CKD (Figure .3). The prob-ability is slightly lower in patients who also have CHF, and higher in those with other

bone & mineral metabolism

prev. Medicare & EGHP patients

Prob. of & geo. variations in Ca & Ph testing w/i a year1.11

prev. Medicare & EGHP patients

Prob. of & geo. variations in PTH testing w/i a year1.12

D

Medicare: all CKD

Medicare: all CKD

Medicare: all CKD

Medicare: all CKD

Medicare: all CKD

Page 9: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

All CKDCKD+DM

CKD+CHFCKD+DM+CHF

CKD other

All CKDCKD+DM

CKD+CHFCKD+DM+CHF

CKD other

All CKDCKD+DM

CKD+CHFCKD+DM+CHF

CKD other

ytiliba

borP

0.0

0.1

0.2

0.3

0.4

0.5

0.6

ytiliba

bo r

P

0.00

0.05

0.10

0.15

ytiliba

borP

0.0

0.2

0.4

0.6

0.8

0.57 + (0.61)0.55 to <0.570.52 to <0.55

0.49 to <0.52below 0.49 (0.45)

0.57 + (0.61)0.55 to <0.570.52 to <0.55

0.48 to <0.52below 0.48 (0.43)

0.16 + (0.21)0.14 to <0.160.13 to <0.14

0.11 to <0.13below 0.11 (0.08)

Medicare

EGHP

Medicare

EGHP

Medicare

EGHP

55

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

Probability of & geographic variations (by HSA) in lipid monitoring within a year in CKD patients

prevalent Medicare & EGHP patients1.16

1.17

1.18 prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in CKD patients receiving pneumococcal pneumonia vaccine within two years

prevalent Medicare & EGHP patients

Probability of & geographic variations (by HSA) in CKD patients receiving an influenza vaccination within a year

diagnoses. Regionally, more patients resid-ing in the western third of the country and in upper New England receive microalbu-minuria or proteinuria testing.

The American Diabetes Association rec-ommends that people with diabetes receive 2–4 glycosylated hemoglobin tests each year. In the Medicare CKD population, the probability of receiving at least two HbAc tests within a year is 0.56, compared to 0.36 for EGHP patients (Figure .4). Testing for glycosylated hemoglobin is more common in the northern half of the country.

In the CKD population as a whole, the probability of lipid monitoring is 0.59 in Medicare patients, and 0.37 in EGHP patients (Figure .6). This rises to 0.7 and 0.50, respectively, in patients with diabetes alone, but falls to 0.6 and 0.36 in those with combined diagnoses of diabetes and CHF. Nationwide, the probability of receiving a test for lipids is highest on the West and East Coasts and in the Southwest.

Under Healthy People 200 guidelines, 90 percent of people should receive an influ-enza vaccination each year. In Medicare CKD patients, however, the probability of a vaccination is only 0.52–0.54 (Figure .7). In EGHP patients with CKD, the probability is even lower, ranging from 0.2 in patients with CHF or other diagnoses to 0.20 in those with diabetes. Some of these patients may, however, receive vaccinations through their employers. Patients in the Upper Mid-west and Ohio valley are the most likely to receive the flu vaccine.

Vaccinations for pneumococcal pneu-monia are even less frequent (Figure .8). The probability of receiving this vaccina-tion within a two-year period is only 0.3 in Medicare CKD patients (slightly higher for those with diabetes and CHF), and 0.03–0.07 in the EGHP population. Geograph-ically, the probability of receiving a pneu-mococcal vaccination is 0.2 for patients residing in areas represented by the top quintile.

{All figures} general Medicare: patients enrolled in an HMO, with Medicare as secondary payor, or with ESRD diagnosed during the year are excluded. EGHP: patients diagnosed with ESRD before or dur-ing the year are excluded. Both populations: patients censored at end of plan & end of 2003; Medicare patients also censored at death. Maps by HSA, unad-justed. {Figures .–2 & .6} general Medi-care: patients entering Medicare before January , 2002, alive & remaining in the program through December 3, & with CKD diagnosed during 2002. EGHP: patients enrolled for the entire year 2002 in a fee-for-service plan, younger than 65, & with CKD diagnosed during 2002. For both populations, diabe-tes, CHF, & other comorbidities are defined in 2002. First calcium & phosphorous testing, PTH testing, & lipid monitoring tracked in 2003. {Figures .3–5} general Medicare: patients entering Medicare before

vaccinations

January. , 2002, alive & remaining in the program through December 3, & with CKD & diabetes diag-nosed during 2002. EGHP: patients enrolled for the entire year 2002 in a fee-for-service plan, younger than 65, & with CKD & diabetes diagnosed during 2002. For both populations, CHF & other comorbid-ities are defined in 2002. All first testing tracked in 2003. {Figure .7} general Medicare: patients enter-ing Medicare before January , 2002, alive & remain-ing in the program through August 3, 2003, & with CKD diagnosed during 2002. EGHP: patients enrolled for the entire year 2002 & through August 3, 2003, in a fee-for-service plan, younger than 65, & with CKD

diagnosed during 2002. For both populations, diabe-tes, CHF, & other comorbidities are defined in 2002. First influenza vaccine tracked between Septem-ber and December 3, 2003. {Figure .8} general Medicare: patients entering Medicare before January , 200, alive & remaining in the program through December 3, & with CKD diagnosed during 200. EGHP: patients enrolled for the entire year 200 in a fee-for-service plan, younger than 65, & with CKD diagnosed during 200. For both populations, diabe-tes, CHF, & other comorbidities are defined in 200. First pneumococcal pneumonia vaccine tracked in 2002 & 2003.

Medicare: all CKD

Medicare: all CKD

Medicare: all CKD

Page 10: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

All CKD CKD+DM CKD+HTN CKD+CHF CKD other

stneita

p fo t

n ecreP

0

20

40

60

80NHANES III

NHANES 99-02

NHANES III

NHANES 99-02

*

All CKD CKD+DM CKD+HTN CKD+CHF CKD other

stneita

p fo t

necreP

0

20

40

All CKD CKD+DM CKD+HTN CKD+other

stneita

p fo t

n ecreP

0

5

10

15

20

25NHANES III

NHANES 99-02

Dihydropyridine calcium channel blocker Non-dihydrorpyridine calcium channel blocker

All CKD CKD+DM CKD+HTN CKD+other

**

stneita

p fo t

nec reP

0

10

20

30

40

50

60 Loop diuretic Thiazide-like diuretic

*

All CKDCKD+DM

CKD+HTN CKD otherCKD+CHF

All CKDCKD+DM

CKD+HTN CKD otherCKD+CHF

NHANES III

NHANES 99-02

CHRONIC KIDNEY DISEASE3

56

1 Prescription drug therapy in CKD patients

NHANES III & NHANES 1999–2002 participants, age 60 & older

CKD patients receiving ACE-Is/ARBs 1.19

NHANES III & NHANES 1999–2002 participants, age 60 & older

CKD patients receiving beta blockers1.20

NHANES III & NHANES 1999–2002 participants, age 60 & older

CKD patients receiving calcium channel blockers1.21

NHANES III & NHANES 1999–2002 participants, age 60 & older

CKD patients receiving diuretics1.22

iven the high percentage of CKD patients with diabetes and/or congestive heart failure,

there are several preventive and mainte-nance therapies that one would expect to see routinely prescribed. We use the NHANES III (988–994) and NHANES 999–2002 datasets to evaluate prescription drug use in CKD patients. These data rep-resent a cross-sectional evaluation of med-ication use in the past month, in contrast to the evaluation we performed on the Med-stat database for the 2004 ADR, in which cumulative use of medications in CKD Stages 3–5 over a one-year period of time was evaluated. It is important to note that CKD Stage 5 patients are excluded from these analyses because it is not possible to evaluate dialysis status. These data thus represent CKD Stages –4, with Stages and 2 determined by a calculated GFR plus a positive urine microalbumin. In addi-tion, because inclusion of all age groups leads to erratic results and there are too few patients younger than 60 to evaluate, this spread represents patients age 60 and older with Stage –4 CKD.

Overall, older CKD patients in 999–2002 were treated more aggressively with medications than in previous years. The use of angiotensin-converting enzyme inhibitors (ACE-Is) and angiotensin recep-tor blockers (ARBs) has more than dou-bled in CKD patients with diabetes, from 22 to 46 percent, and has almost doubled in CKD patients with hypertension (Figure .9). Further growth in use of these agents should be fueled by National Kidney Foun-dation (NKF) CKD practice guidelines, released in 2002, which support ACE-I and ARB use in CKD patients with diabetes, by 2004 guidelines advocating ACE-I and ARB use in CKD patients with hypertension, and by congestive heart failure guidelines pub-lished in 200 advocating ACE-Is and ARBs in CHF patients.

CHF and hypertension practice guide-lines advocate the use of beta blockers in the treatment of these diseases. Use of these agents has increased from 9 to 26 percent and 5 to 28 percent in older CKD patients with hypertension and congestive heart failure, respectively (Figure .20). Interest-ingly, beta blocker use has also increased tremendously in CKD patients with diabe-tes. Increased use overall, in hypertensive

G

Page 11: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

All CKD CKD+DM CKD+CVD CKD otherAll CKD CKD+DM CKD+CVD CKD other

stneita

p fo t

ne creP

0

10

20

30

40

NHANES III

NHANES 99-02

Statins Non-statins

* **

* ** *

*

Insulin TZDs Secretagogues MetforminInsulin TZDs Secretagogues Metformin

stneita

p fo t

necr eP

0

10

20

30

40

50NHANES III

NHANES 99-02

CKD Non-CKD

*

57

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

NHANES III & NHANES 1999–2002 participants, age 60 & older

CKD patients receiving lipid-lowering agents

patients, and in patients with diabetes may reflect lower blood pressure goals and the fact that beta blockers and diuretics were advocated by Joint National Committee on Prevention, Detection, Evaluation and Treatment 6 (JNC 6) guidelines in 997 as first-line therapy to reach new blood pres-sure goals.

Use of dihydropyridine calcium-chan-nel blockers (CCBs—amlodipine, felo-dipine, isradipine, nicardipine, nifedipine, nimodipine, nisoldipine) has risen substan-tially in older CKD patients with either dia-betes or hypertension (Figure .2). This is in stark contrast to nondihydropyri-dine CCB (diltiazem and verapamil) use, which has declined in the same timeframe. These results are not surprising, given the superior antihypertensive effect of the dihy-dropyridine CCB group.

This year we evaluate the use of both loop and thiazide (including metolazone) diuretics in CKD patients. Use of loop diuretics (bumetanide, furosemide or torse-

1.23

1.24 NHANES III & NHANES 1999–2002 participants, age 60 & older

Patients with diabetes who receive glucose-lowering agents

mide) has increased in all groups of older CKD patients (Figure .22). This may indi-cate that more CKD patients are being iden-tified and treated. In contrast, except for use in CKD patients with congestive heart fail-ure, use of thiazide diuretics has declined. The use of diuretics—particularly thiazide diuretics—may increase, since JNC 7 guide-lines published in 2003 strongly advocate the use of thiazide diuretics as a first-line agent, and new NKF clinical practice guide-lines on hypertension in CKD patients advo-cate diuretic use following institution of an ACE-I or ARB.

The use of lipid-lowering agents has dramatically increased in all older CKD patients, from virtually no use in the NHANES III study to over 28 percent in the NHANES 999–2002 population (Fig-ure .23). This most likely represents the influence of the National Cholesterol Edu-cation Program: Second Report of the Expert Panel on Detection, Evaluation, and Treatment of High Blood Cholesterol in

Adults (Adult Treatment Panel II—NCEP II), published in 994 and advocating more aggressive treatment of high cholesterol levels in the general population. In partic-ular, Figure .23 shows that statin use has increased up to ten-fold, while use of non-statins (bile acid sequestrants, fibrate acid derivatives, and niacin) actually fell slightly over the same period. In CKD patients with cardiovascular disease, statin use increased from 4 to 39 percent. Presumably, the use of statins will be even higher in the next NHANES dataset due to even more aggres-sive lipid goals as advocated by NCEP III, new NKF CKD guidelines on dyslipidemia, and increased safety data on statins.

We evaluated the use of glucose-low-ering medications in older CKD and non-CKD patients in the two NHANES pop-ulations (Figure .24). The use of insulin was substantially higher (30 percent and 28 percent versus 2 percent and 5 percent) in CKD Stage –4 patients than in non-CKD patients, most likely reflecting more severe diabetes in CKD patients. Accord-ingly, secretagogue use (sulfonylureas) was lower in CKD patients. Use of thiazolidine-dione (pioglitazone, rosiglitazone, and tro-glitazone, which was taken off the mar-ket in 2000) is approximately the same in CKD and non-CKD patients in the 999–2002 NHANES population. As expected, metformin use is substantially less in CKD patients than non-CKD patients (2 ver-sus 34 percent), as it is contraindicated in patients with kidney disease. The manufac-turer has recently tightened its recommen-dation for use of metformin in patients with kidney disease, and currently states that patients whose serum creatinine concen-trations exceed the upper limit of normal for their age should not receive metformin due to increased risk of lactic acidosis.

{Figures .9–24} NHANES III 988–994 & NHANES 999–2002 patients age 60 & older; patients with eGFRs of less than 5 ml/min/.73 m2

are excluded. *Sample size less than 30, or coefficient of variation is not less than 30 percent.

Page 12: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

All NCKD DM CHF DM+CHF Other

ytiliba

borP

0.0

0.1

0.2

0.3

White

Black

All NCKD DM CHF DM+CHF Other

ytiliba

bor

P

0.0

0.1

0.2

0.3

White

Black

All CHF Other

ytiliba

bor

P

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7CKD

Non-CKD

All CHF Other All CHF Other

White

Black

All patients CKD Non-CKD

Insulin TZDs Secretagogues MetforminInsulin TZDs Secretagogues Metformin

stneita

p fo t

ne cr eP

0

10

20

30

40

50

NHANES III

NHANES 99-02

White Other race

*

White

Black

0.274 + (0.394)0.192 to <0.2740.138 to <0.192

0.096 to <0.138below 0.096 (0.069)

0.271 + (0.399)0.188 to <0.2710.145 to <0.188

0.108 to <0.145below 0.108 (0.075)

0.248 + (0.314)0.203 to <0.2480.173 to <0.203

0.134 to <0.173below 0.134 (0.103)

0.260 + (0.372)0.200 to <0.2600.159 to <0.200

0.117 to <0.159below 0.117 (0.089)

CHRONIC KIDNEY DISEASE3

58

1

1.26

1.25 prevalent Medicare patients

Probability of & geographic variations in the assessment of serum creatinine within a year, by race & HSA

prevalent Medicare patients

Probability of & geographic variations in the assessment of microalbuminuria or proteinuria within a year, by race & HSA

1.28 NHANES III & NHANES 1999–2002 participants, age 60 & older

Diabetic patients receiving glucose lowering agents, by race

1.27 prevalent Medicare patients

Probability of diabetic patients receiving at least two glycosylated hemoglobin assessments within a year, by race

Racial disparities in the assessment of CKD

assessment of populations at risk for ckd

assessment in diabetic patientsn the non-CKD Medicare popu-lation as a whole, the probability of a serum creatinine assessment

is 0.4 for whites, and 0.2 for blacks. (Fig-ure .25). Compared to that in blacks, the probability of testing in whites is only .–.2 percent higher in those with diabetes, CHF, the two combined, or in those with other diagnoses. Geographically, the high-est probability of a serum creatinine test in white patients occurs in the Upper Mid-west, the Ohio Valley, and portions of New England. Patterns are generally similar for black patients.

The probability of microalbuminuria or proteinuria testing is only 0.04 and 0.06 for whites and blacks, respectively, without CKD, and reaches 0.22 and 0.9 in those with diabetes (Figure .26). In the non-CKD population overall and in those with CHF or other diagnoses, the probability of test-ing is greater for blacks, while among those with diabetes alone or in combination with CHF whites are more likely to receive test-ing. Nationwide, patients residing in the mid-sections and southern portions of the country are less likely to be tested.

I

White: all NCKD diabetes patients Black: all NCKD diabetes patients

White: all NCKD diabetes patients Black: all NCKD diabetes patients

Page 13: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

White

Black

OtherAllDM

CHFDM+CHF

OtherAllDM

CHFDM+CHF

OtherAllDM

CHFDM+CHF

ytiliba

borP

0.0

0.2

0.4

0.6

0.8CKD

Non-CKD

White

Black

All patients CKD Non-CKD

CKD+DM CKD+HTN CKD+CHF CKD+CVD

stneita

p fo t

necr e P

0

10

20

30

40

50

CKD+DM CKD+HTN CKD+CHF CKD+CVD

White Other race

NHANES III

NHANES 99-02

* **

* *

CKD+DM CKD+HTN CKD+CHF CKD+CVD

stneita

p fo t

n ecre P

0

20

40

60

80

NHANES III

NHANES 99-02

CKD+DM CKD+HTN CKD+CHF CKD+CVD

White Other race

59

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

prevalentMedicare patients

Probability of lipid assessments within a year, by race1.29

1.30

1.31 NHANES III & NHANES 1999–2002 participants age 60 & older

Use of ACE-Is/ARBs in patients with CKD, by race

NHANES III & NHANES 1999–2002 participants

Use of lipid lowering agents in patients with CKD, by race

In both the CKD and non-CKD dia-betic populations, and regardless of accom-panying diagnoses, white patients are more likely than black patients to receive at least two glycosylated hemoglobin tests within a year—the minimum recommended by the American Diabetes Association (Fig-ure .27). Among CKD patients, for exam-ple, the probability of testing in whites is 0.5 percent greater overall, and 5 and 8 percent higher, respectively, for those with CHF or with other diagnoses.

Non-white patients are slightly more likely than whites to receive TZDs, secreta-gogues, or metformin (Figure .28).

The probability of lipid assessment is 5 percent higher in CKD patients overall than in those without the disease (Figure .29). Among those with diabetes, CHF, or both diagnoses, however, the probability of testing is nearly equal. By race, and regard-less of diagnoses, the probability of testing is greater in whites than in blacks—0 per-cent higher in the CKD population overall, and 4 percent higher for those with diabe-tes or CHF.

In the NHANES 999–2002 popula-tion, use of lipid lowering agents is much more common in whites compared to peo-ple of other races. In contrast, 62.8 percent of non-white patients with congestive heart failure receive ACE inhibitors compared to 42.5 percent in the white population (Fig-ures .30–3).

{Figures .25–26} general Medicare: patients enter-ing Medicare before January , 2002, alive & remain-ing in the program through December 3, & with-out CKD diagnosed during 2002. Patients enrolled in an HMO, with Medicare as secondary payor, or with ESRD diagnosed during the year are excluded. Diabetes, CHF, & other comorbidities are defined in 2002. Patients censored at end of plan, death, & end of 2003. All testing tracked in 2003. In Figure .25, data on serum creatinine testing obtained from individual tests. CPT codes used for assessment of serum creat-inine include 80069, & 82565. Maps by HSA, unad-justed. {Figure .27} general Medicare patients enter-ing Medicare before January , 2002, alive & remaining in the program through December 3, & with diabetes diagnosed during 2002. Patients enrolled in an HMO, with Medicare as secondary payor, or with ESRD diagnosed during the year are excluded. CKD, CHF, & other comorbidities are defined in 2002. Patients censored at end of plan, death, & end of 2003. First glycosylated hemoglobin testing tracked in 2003. {Fig-ure .28} NHANES III 988–994 & NHANES 999–

lipid monitoring

2002 patients age 60 & older; patients with eGFRs of less than 5 ml/min/.73 m2 are excluded. *Sam-ple size less than 30 or coefficient of variation is not less than 30 percent. {Figure .29} general Medicare patients entering Medicare before January , 2002, alive & remaining in the program through December 3. Patients enrolled in an HMO, with Medicare as secondary payor, or with ESRD diagnosed during the year are excluded. CKD, CHF, & other comorbidities

are defined in 2002. Patients censored at end of plan, death, & end of 2003. First lipid monitoring tracked in 2003. {Figures .30–3} NHANES III 988–994 & NHANES 999–2002 patients age 60 & older; patients with eGFRs of less than 5 ml/min/.73 m2 are excluded. *Sample size less than 30 or coefficient of variation is not less than 30 percent.

Page 14: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

0

600

1,200

1,800

2,400

AllDiabetesCHFDiabetes+CHF

CKD: Medicare EGHP

93 94 95 96 97 98 99 00 01 02 03 0

600

1,200

1,800

2,400

00 01 02 03

Non-CKD

ksir ta sr ae y tneit a

p 0 00 ,1 rep et a

R

130 + (147)114 to <130104 to <114

92 to <104below 92 (83)

40.2 + (45.1)37.1 to <40.234.2 to <37.1

30.8 to <34.2below 30.8 (28.1)

CHRONIC KIDNEY DISEASE3

60

1

1.33 prevalent patients

Geographic variations in all-cause hospitalizations in patients with or without CKD, 2003, by HSA

n this spread we compare all-cause and cause-specific hospital-ization rates in Medicare patients

with and without CKD, and in their coun-terparts with EGHP coverage. For all types of hospitalization examined here, the highest rates occur in CKD patients with both diabetes and congestive heart failure (CHF) in the entry period. CKD patients with CHF alone have the next highest rates, followed by those with diabetes.

In the Medicare population, the over-all hospitalization rate for CKD patients reached ,068 per ,000 patient years in 2003—nearly three times greater than that

Adjusted hospitalization rates in CKD & non-CKD patients

1.32 prevalent patients

All-cause hospitalizations in patients with & without CKD, by prior comorbidity

in patients without CKD (Figure .32). The rate for patients with diabetes is 2.3 times higher in CKD patients, reaching ,276, while rates for patients with CHF or with both CHF and diabetes combined are .6–.7 times greater. In the EGHP population, the differences are even more dramatic. The overall 2003 rate of 494 in CKD patients is 6.5 times higher than in the non-CKD pop-ulation, while rates for patients with diabe-tes, CHF, or the two diseases combined are 3.8, 2.2, and 2.2 times greater, respectively, in patients with CKD.

Since 993, hospitalization rates have declined slightly among Medicare patients

with CKD—from 4 percent in those with CHF to 3 percent in those with diabetes. In the non-CKD cohort, rates have risen 6 percent overall and 2 percent in CHF patients, while falling 4.4–4.9 percent in patients with diabetes alone or in con-junction with CHF. Because of the smaller cohort size, changes in the EGHP popula-tion have been more variable.

Hospitalization rates for patients with or without CKD tend to be highest in the east-ern portions of the country (Figure .33). Rates for patients with CKD, however, are nearly three-fold higher than those found in patients without CKD.

Since 993, admissions for CHF have fallen 26 percent in the Medicare CKD pop-ulation overall, and 35 percent in patients with diabetes (Figure .34). Among non-CKD patients, in contrast, the overall rate has dropped only 7 percent, while the rate for diabetic patients has fallen 9 percent. Compared to their non-CKD counter-parts, Medicare CKD patients have hospi-talizations that are 6.5 times greater overall, four times higher for patients with diabe-tes during the entry period, and twice as high for those with CHF—whether or not it is accompanied by diabetes. In the EGHP population, the CHF hospitalization rate in CKD versus non-CKD patients is 8.5, 2.3, and 2.8 times higher for patients with dia-betes, CHF, and diabetes combined with CHF, respectively.

Rates of hospitalization for ASHD have also declined since 993—and to the great-est extent, as with CHF admissions, in the Medicare CKD population (Figure .35). Compared to those of non-CKD patients, overall ASHD admission rates are 2.6 times greater in Medicare patients, and 4.7 times higher in those with EGHP coverage.

In 2003, pneumonia admission rates reached 58 per ,000 patient years in Medi-care CKD patients overall, and 09 in patients with both diabetes and CHF—2.8 and .4 times higher, respectively, than in the non-CKD population (Figure .36). Among EGHP patients, rates of 8 overall and 93 for those with the combined diagno-ses were and four times greater, respec-tively, in CKD than in non-CKD patients. In the Medicare population, rates of hos-pitalization for pneumonia have remained stable since the mid-990s.

CKD Non-CKD

O

Page 15: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

0

200

400

600

AllDiabetesCHFDiabetes+CHF

CKD: Medicare EGHP

93 95 97 99 01 030

200

400

600

00 01 02 03

Non-CKD

ksir ta sra ey tn eit a

p 000,1 rep et aR

0

50

100

150

200

AllDiabetesCHFDiabetes+CHF

CKD: Medicare EGHP

93 95 97 99 01 030

50

100

150

200

00 01 02 03

Non-CKD

ksir ta sraey tneita

p 000,1 rep etaR

0

25

50

75

100

125

AllDiabetesCHFDiabetes+CHF

CKD: Medicare EGHP

93 95 97 99 01 030

25

50

75

100

125

00 01 02 03

Non-CKD

ks ir ta s raey tne ita

p 00 0, 1 rep e ta R

00 01 02 03

0

20

40

60

AllDiabetesCHFDiabetes+CHF

CKD: Medicare EGHP

93 95 97 99 01 030

20

40

60 Non-CKD

ksir ta sr aey tne ita

p 00 0, 1 rep e taR

61

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

1.34 1.35 prevalent patients

for ASHD

1.37 prevalent patients

for bacteremia/septicemia

prevalent patients

for congestive heart failure

1.36 prevalent patients

for pneumonia

cause-specific hospitalizations in patients with & without CKD, by prior comorbidity

Differences between the CKD and non-CKD populations are even greater for bacteremia/septicemia hospitalizations (Figure .37). For all patients and for those with diabetes, for example, rates are 4.6 and 3.0 times higher, respectively, in the Medicare popula-tion, and 27.5 and 7.5 times greater in those with EGHP coverage.

For both all-cause hospitalizations and the cause-specific rates examined here, the greatest differences between CKD and non-CKD patients consistently occur in the EGHP population—in

overall rates and in rates for patients diagnosed with diabetes dur-ing the entry period.

{Figures .32–37} Medicare: prevalent patients continuously enrolled in Medicare Parts A & B, with no HMO coverage, & alive during the one-year entry period; adjusted for age, gender, & race. EGHP: prevalent patients age 20–65 with fee-for-service coverage during the entire calendar year, & alive on the last day of the entry period; adjusted for age & gender. Patients diagnosed with ESRD before or during the entry period are omitted; 2003 patients used as reference cohort. Because of different age distributions in the two cohorts, rates are comparable only within the Medicare & EGHP cohorts, not between them. In Figures .32–33, rates exclude hospitaliza-tions related to pregnancy & childbirth. In Figure .33, maps by HSA, unadjusted.

Page 16: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40 All

Acute kidney failure as the principal diagnosis

Acute kidney failure as the secondary diagnosis

ytil iba

bor

p e vi talu

mu

C

0.00

0.02

0.04

0.06

0.08

0.10

0.12

0.14

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

All Diabetes CHF DM+CHF

1-year

2-year

3-year

1-year

2-year

3-year

1-year

2-year

3-year

With prior CKDWithout prior CKD

CHRONIC KIDNEY DISEASE3

62

1

1.38 Medicare Parts A & B patients

Probability of hospitalization for acute kidney failure

n this spread we examine hospitalizations for acute kid-ney failure, along with their subsequent outcomes. In the Medicare population without CKD, the probability of a

hospitalization for acute kidney failure is 0.034 at three years (Fig-ure .38). As we look at complicating diagnoses of diabetes, con-gestive heart failure, and diabetes combined with CHF, this three-year probability rises steadily—to 0.07, 0., and 0.6, respectively. Among patients with a diagnosis of CKD, the probability of an acute kidney failure hospitalization is, not surprisingly, far higher, from 0.22 overall to 0.39 for those with diabetes and CHF com-bined. In this CKD population, the probability of hospitalization in patients with both diabetes and CHF is 49 percent greater at one year, and 43 percent greater at three years, than that of patients with diabetes alone.

Similar patterns occur when we categorize hospitalizations by whether acute kidney failure is the primary or secondary diagno-sis at admission. For acute kidney failure as a primary diagnosis, the one-year probability of hospitalization in CKD patients ranges from 0.03 overall to 0.05 in those with diabetes and CHF, while the three-year probability ranges from 0.07 to 0.3. Probabilities are greater for hospitalizations in which acute kidney failure is the secondary admitting diagnosis, at three years reaching 0.8 overall and 0.32 for those with combined diabetes and CHF.

In Figures .39–42 we present data on ESRD and death follow-ing a hospitalization for acute kidney failure. Across outcomes and diagnoses, the highest event rates occur within the first three months after an acute kidney failure hospitalization, with rates then leveling out or decreasing slightly in the following 33 months.

For patients who have a prior diagnosis of CKD, the rate of death in the first three months after hospitalization is 2.7–4 times greater than the rate of an ESRD diagnosis. Patients without a CKD diagnosis prior to that hospitalization, however, face a death rate that is 3–20 percent greater than the risk of developing ESRD.

In the prevalent Medicare population as a whole, 2.6 percent of patients without prior CKD are diagnosed with ESRD in the first three months, compared to . percent of patients with an earlier CKD diagnosis. The three-month rates do not vary widely by diag-

Acute kidney failure: hospitalization & outcomes

O

Page 17: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

Months after AKF hospitalization

3 6 9 12 15 18 21 24 27 30 33 360

2

4

6

8

10

12

14

With prior CKDWithout prior CKD

3 6 9 12 15 18 21 24 27 30 33 360

10

20

30

40

50All: ESRD

stneita

p 001 rep etaR

Death

Months after AKF hospitalization

3 6 9 12 15 18 21 24 27 30 33 360

2

4

6

8

10

12

14

With prior CKDWithout prior CKD

3 6 9 12 15 18 21 24 27 30 33 360

10

20

30

40

50Diabetes: ESRD

stneita

p 001 rep etaR

Death

Months after AKF hospitalization

3 6 9 12 15 18 21 24 27 30 33 360

2

4

6

8

10

12

14

With prior CKDWithout prior CKD

3 6 9 12 15 18 21 24 27 30 33 360

10

20

30

40

50CHF: ESRD

stnei ta

p 001 r ep et aR

Death

Months after AKF hospitalization3 6 9 12 15 18 21 24 27 30 33 36

0

2

4

6

8

10

12

14

With prior CKDWithout prior CKD

3 6 9 12 15 18 21 24 27 30 33 360

10

20

30

40

50Diabetes+CHF: ESRD

stneita

p 001 rep etaR

Death

63

2005 Annual Data Report 31CHRONIC KIDNEY DISEASE

1.39 prevalent Medicare patients

All patients 1.40 prevalent

Medicare patientsPatients with

diabetes

1.41 prevalent Medicare patients

Patients with congestive heart failure 1.42 prevalent

Medicare patientsPatients with diabetes & congestive heart failure

outcomes following first hospitalization for acute kidney failure

nosis: 3.2 and 2.5, respectively, for patients with diabetes, 2.4 and 0.6 for those with CHF, and 3. and 2.0 for those with diabetes and CHF combined. Patients with CKD are, then, 3.9–4.4 times more likely than patients without prior CKD to be diagnosed with ESRD during the three months following an acute kidney failure hospitalization.

Results differ significantly for death after an acute kidney fail-ure hospitalization. Across diagnoses, three-month death rates are higher in patients without prior CKD than in those who have the

diagnosis—2.4 percent higher overall, and 9.4, 3.5, and 5.9 per-cent higher in those with diabetes, CHF, and the two combined, respectively. After this three-month period, however, rates are con-sistently greater in the CKD population.

{Figure .38} patients continuously enrolled in Medicare Parts A & B, with no HMO coverage during 2000, & alive on December 3, 2000. Patients diagnosed with ESRD before December 3, 2000 are excluded. {Figures .39–42} prevalent Medicare patients with hospitalization for acute kidney failure during 999–2000; adjusted for age, race, & gender. A one-year entry period prior to the first hospital-ization is used to define cardiovascular disease & diabetes.

Page 18: 1 · CHRONIC KIDNEY DISEASE - USRDS Home … 3 48 CHRONIC KIDNEY DISEASE 1 Introduction Chronic kidney disease (CKD), the precursor to ESRD, received little attention prior to the

Figure number 1.3 1.3 1.3 1.5 1.6 1.7 1.8 1.9 1.10 1.11 MC 92-3 MC 97-8 MC 02-3 Overall value for all pts 23.4 33.8 56.25 0 .20 0.20 0.06 0.09 0.01 0.38 0.27Total patients 1,354,207 1,283,428 1,385,876 235,798 235,798 235,798 235,798 235,798 235,798 52,121 Overall value for pts mapped 23.4 33.8 56.2 0.20 0.20 0.06 0.09 0.01 0.38 0.27Missing HSA/state: pts dropped 17,007 19,063 20,706 5,782 5,782 5,782 5,782 5,782 5,782 832

Figure number 1.12 1.13 1.14 1.15 1.16 1.17 1.18 1.25 1.25 1.26 Wh Bl WhOverall value for all pts 0.08 19.3 56.2 41.1 58.5 53.1 12.9 0.20 .018 0.21Total patients 52,121 24,531 24,531 24,531 52,121 45,414 44,942 191,934 30,744 191,934 Overall value for pts mapped 0.08 19.3 56.2 41.1 58.5 53.1 12.9 0.20 0.18 .21Missing HSA/state: pts dropped 832 547 547 547 832 756 750 3,483 363 3,483

Figure number 1.26 1.33 1.33 Bl w/CKD w/o CKDOverall value for all pts 0.18 107.7 36.3 Total patients 30,744 52.121 1,421,218Overall value for pts mapped 0.19 107.9 36.4Missing HSA/state: pts dropped 363 832 19,673

maps: national means & patient popUlations

CHRONIC KIDNEY DISEASE3

64

1 Chapter summary

Figure . In almost two-thirds of Medi-care patients age 65 and older with CKD, CKD is accompanied by diabetes, CHF, or both diseases combined.

Figure .4 The percent of CKD patients with just diabetes increased in the Medi-care population from 6.7 in 992–993 to 23.7 in 2002–2003, and in the EGHP population from 23.8 in 999–2000 to 26.6 in 2002–2003.

Figure .6 The overall probability of microalbuminuria or proteinuria testing reaches only 0.04 in the Medicare popu-lation, and 0.0 for EGHP patients. Fig-ure .8 The probability of calcium and phosphorus assessment is 0.06 in Medi-care non-CKD patients overall, and 0.02 in the EGHP population.

Figure . The probability of a CKD patient receiving calcium phospho-rus testing is 0.27 for those with Medi-care coverage, and 0.7 for those under EGHP. Figure .4 In the Medicare CKD population, the probability of diabetic patients receiving at least two HbAc tests within a year is 0.56, compared to 0.36 for EGHP patients.

Figure .9 The use of ACE-Is and ARBs for renal protective treatment has almost doubled in the CKD population. Figures .22–23 Use of diuretic therapy in CKD

introduction

prevalence of CKD

assessment of CKD in risk populations

preventive healthcare monitoring in CKD patients

prescription drug therapy in CKD patients

patients with CHF has grown, as has the use of lipid-lowering agents in both dia-betics and those with CVD.

Figure .27 In both the CKD and non-CKD diabetic populations, and regard-less of accompanying diagnoses, white patients are more likely than black patients to receive at least two glycosyl-ated hemoglobin tests within a year.

Figure .32 The overall hospitalization rate for Medicare CKD patients reached ,068 per ,000 patient years in 2003—nearly three times greater than that in non-CKD patients. The rate for diabetic patients is 2.3 times higher in those with CKD, reaching ,276.

Figure .38 In non-CKD Medicare patients, the probability of a hospital-ization for acute kidney failure is 0.034 at three years. Among patients with a diagnosis of CKD, the probability is far higher, from 0.22 overall to 0.39 for those with diabetes and CHF combined. Fig-ure .39 For patients with a prior diagno-sis of CKD, the rate of death in the first three months after hospitalization for AKF is 2.7–4 times greater than the rate of an ESRD diagnosis. Patients without a CKD diagnosis prior to that hospital-ization, however, face a death rate that is 3–20 percent greater than the risk of developing ESRD.

racial disparities in the assessment of CKD

hospitalization rates in CKD & non-CKD patients

acute kidney failure: hospitalization & outcomes