ORIGINAL CONTRIBUTION

Prevalence of Chronic Kidney Disease

in the United States
Josef Coresh, MD, PhD Context The prevalence and incidence of kidney failure treated by dialysis and trans-
Elizabeth Selvin, PhD, MPH plantation in the United States have increased from 1988 to 2004. Whether there have
been changes in the prevalence of earlier stages of chronic kidney disease (CKD) during
Lesley A. Stevens, MD, MS
this period is uncertain.
Jane Manzi, PhD
Objective To update the estimated prevalence of CKD in the United States.
John W. Kusek, PhD
Design, Setting, and Participants Cross-sectional analysis of the most recent Na-
Paul Eggers, PhD tional Health and Nutrition Examination Surveys (NHANES 1988-1994 and NHANES
Frederick Van Lente, PhD 1999-2004), a nationally representative sample of noninstitutionalized adults aged 20
years or older in 1988-1994 (n=15 488) and 1999-2004 (n=13 233).
Andrew S. Levey, MD
Main Outcome Measures Chronic kidney disease prevalence was determined based
on persistent albuminuria and decreased estimated glomerular filtration rate (GFR).

C
HRONIC KIDNEY DISEASE Persistence of microalbuminuria (⬎30 mg/g) was estimated from repeat visit data in
(CKD) is now recognized as NHANES 1988-1994. The GFR was estimated using the abbreviated Modification of
a common condition that el- Diet in Renal Disease Study equation reexpressed to standard serum creatinine.
evates the risk of cardiovas- Results The prevalence of both albuminuria and decreased GFR increased from 1988-
cular disease as well as kidney failure 1994 to 1999-2004. The prevalence of CKD stages 1 to 4 increased from 10.0% (95%
and other complications.1-3 The num- confidence interval [CI], 9.2%-10.9%) in 1988-1994 to 13.1% (95% CI, 12.0%-
ber of patients with kidney failure 14.1%) in 1999-2004 with a prevalence ratio of 1.3 (95% CI, 1.2-1.4). The prevalence
treated by dialysis and transplantation estimates of CKD stages in 1988-1994 and 1999-2004, respectively, were 1.7% (95%
(the end stage of CKD) has increased CI, 1.3%-2.2%) and 1.8% (95% CI, 1.4%-2.3%) for stage 1; 2.7% (95% CI, 2.2%-
dramatically in the United States from 3.2%) and 3.2% (95% CI, 2.6%-3.9%) for stage 2; 5.4% (95% CI, 4.9%-6.0%) and
209 000 in 1991 to 472 000 in 2004.4 7.7% (95% CI, 7.0%-8.4%) for stage 3; and 0.21% (95% CI, 0.15%-0.27%) and 0.35%
(0.25%-0.45%) for stage 4. A higher prevalence of diagnosed diabetes and hyperten-
The age-, sex-, and race-adjusted inci- sion and higher body mass index explained the entire increase in prevalence of albumin-
dence of end-stage renal disease in- uria but only part of the increase in the prevalence of decreased GFR. Estimation of GFR
creased by 43% during the decade fol- from serum creatinine has limited precision and a change in mean serum creatinine ac-
lowing 1991. 4 Estimation of the counted for some of the increased prevalence of CKD.
prevalence of earlier stages of CKD in Conclusions The prevalence of CKD in the United States in 1999-2004 is higher
the US population and ascertainment than it was in 1988-1994. This increase is partly explained by the increasing preva-
of trends over time is central to dis- lence of diabetes and hypertension and raises concerns about future increased inci-
ease management and prevention plan- dence of kidney failure and other complications of CKD.
ning, particularly given the increase in JAMA. 2007;298(17):2038-2047 www.jama.com
the prevalence of obesity, diabetes,5,6
and hypertension,7,8 the leading risk fac- nation Surveys (NHANES) have pro- Author Affiliations: Department of Epidemiology,
Bloomberg School of Public Health, and Welch Cen-
tors for CKD.5 vided a rigorous basis for estimating ter for Prevention, Epidemiology, and Clinical Re-
Earlier stages of CKD are defined CKD prevalence. These large nation- search, Johns Hopkins University, Baltimore, Mary-
based on the combination of kidney ally representative surveys conducted land (Drs Coresh, Selvin, and Manzi); Division of
Nephrology, Tufts-New England Medical Center, Bos-
damage (most often quantified using al- by the National Center of Health Sta- ton, Massachusetts (Drs Stevens and Levey); Na-
buminuria) and decreased kidney func- tistics (NCHS) include a laboratory as- tional Institute of Diabetes and Digestive and Kidney
Diseases, Bethesda, Maryland (Drs Kusek and Egg-
tion (quantified as glomerular filtra- sessment of albuminuria and serum cre- ers); and Department of Clinical Pathology, Cleve-
tion rate [GFR] estimated from the atinine allowing for identification and land Clinic Foundation, Cleveland, Ohio (Dr Van Lente).
Corresponding Author: Josef Coresh, MD, PhD, Johns
serum creatinine concentration).2 The staging of CKD regardless of the par- Hopkins University, 2024 E Monument, Ste 2-600, Bal-
National Health and Nutrition Exami- ticipant or their physician’s awareness timore, MD 21287 ([email protected]).

2038 JAMA, November 7, 2007—Vol 298, No. 17 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

of the condition. Initial prevalence es- view board and informed consent was atinine, age, sex, and race were related
timates for CKD stages in NHANES obtained from all participants.19,20 to GFR measured using urinary clear-
1988-1994 in adults have provided a In all NHANES, certain subgroups of ance of 125 I-iothalamate. 21 The for-
benchmark for kidney disease studies, the population were oversampled in- mula initially used serum creatinine
prevention efforts, and health care plan- cluding Mexican Americans, non- measured by a kinetic rate Jaffe method
ning.1,9 Later studies have compared Hispanic blacks, and elderly persons to but was later reexpressed to use stan-
NHANES 1988-1994 estimates with ensure adequate sample sizes of these dard creatinine22 as follows:
NHANES 1999-2000 data and found an groups. Individuals participated in an in- GFR=175⫻ (standardized serum cre-
increased prevalence of albuminuria but terview conducted at home and also in atinine)−1.154 ⫻(age)−0.203 ⫻0.742 (if the
no significant increase in the overall an extensive physical examination per- individual is female) or ⫻1.212 (if the
prevalence of CKD. The precision of formed at a mobile examination center, individual is black). Estimated GFR is
these trend estimates was constrained which included blood and urine collec- reported in mL/min/1.73 m2 and race
by the relatively small sample size of the tion. We limited the study population to is either black or not black. Values that
1999-2000 survey10 and limited data to persons examined in a mobile examina- exceeded 200 mL/min/1.73 m2 were
establish consistent calibration of the tion center who were aged 20 years or truncated at that level and individuals
creatinine assays over time. A recent older and who were not missing serum were classified using standard crite-
study calibrating serum creatinine in all creatinine measurements: 15 488 in ria.1 Individuals with estimated GFR be-
NHANES surveys from 1988 to 2004 NHANES 1988-1994 (7471 in 1988- low 15 mL/min/1.73 m2 (CKD stage 5)
permits a more rigorous examination 1991 and 8017 in 1991-1994) and were excluded because estimates of this
of the trends in the prevalence of CKD 13 233 in NHANES 1999-2004 (4101 in stage are likely to be unreliable due to
using standardized creatinine.11 Esti- 1999-2000, 4684 in 2001-2002, and the small number of individuals and the
mation of GFR from serum creatinine 4448 in 2003-2004). likelihood that many of these individu-
is the recommended approach for CKD als are ill or receiving dialysis and would
staging at this time and increasing evi- Measures of Kidney Function have a low response rate.
dence shows a strong association with and Kidney Damage A random spot urine sample was ob-
risk even when applied to the general Serum creatinine was measured using tained from participants examined at a
population.3,12-15 a kinetic rate Jaffe method. To appro- mobile examination center using a
We compare the prevalence of CKD priately estimate GFR, all serum cre- clean-catch technique and sterile con-
in NHANES 1988-1994 with NHANES atinine measurements were recali- tainers. Frozen nonhematuric speci-
1999-2004 and describe the distribu- brated to standardized creatinine mens were analyzed. Urine albumin and
tion of CKD stages and severity. The measurements obtained at the Cleve- creatinine concentrations were mea-
effect of the increasing prevalence of dia- land Clinic Research Laboratory sured in the same laboratory during all
betes and changes in hypertension and (Cleveland, Ohio) as detailed re- surveys. Albumin was measured by
obesity are examined as explanatory vari- cently.11 Briefly, frozen serum samples solid phase fluorescence immunoas-
ables for changes in CKD prevalence in (approximately 200 specimens from say and urine creatinine was mea-
the general US adult population. NHANES 1988-1994 and each later sured by the modified kinetic Jaffe
2-year survey [1999-2000, 2001- method using a Synchron AS/Astra Ana-
METHODS 2002, and 2003-2004]) were thawed lyzer (Beckman Coulter, Fullerton,
Study Population and reassayed for serum creatinine in California). Urinary albumin-to-
The NHANES are cross-sectional, mul- 2006. Regression models were devel- creatinine ratio was computed and is re-
tistage, stratified, clustered probabil- oped to correct serum creatinine ported in milligrams per gram. Albu-
ity samples of the US civilian noninsti- values in NHANES 1988-1994 and minuria is defined as an albumin-to-
tutionalized population conducted by 1999-2000 (standard creatinine = creatinine ratio of 30 mg/g or higher,
the NCHS, a branch of the Centers for –0.184 ⫹ 0.960 ⫻ NHANES 1988- with microalbuminuria defined as an
Disease Control and Prevention.16 The 1994 uncalibrated serum creatinine albumin-to-creatinine ratio of 30 mg/g
NHANES analyzed were conducted and standard creatinine = 0.147 to 299 mg/g, and macroalbuminuria de-
from 1988-1994 in 2 phases (1988- ⫹ 1.013 ⫻ NHANES 1999-2000 fined as an albumin-to-creatinine ra-
1991 and 1991-1994) and from 1999- uncalibrated serum creatinine). No cor- tio of 300 mg/g or higher.
2004 in 3 phases (1999-2000, 2001- rection was needed for the 2001-2002 Persistent albuminuria was defined
2002, and 2003-2004), and the data and 2003-2004 surveys. as kidney damage (CKD stages 1-2).
from the 2 phases and 3 phases, respec- The GFR was estimated using the ab- Repeated measurements obtained ap-
tively, were combined herein follow- breviated Modification of Diet in Re- proximately 2 weeks after the original
ing NCHS recommendations.17,18 The nal Disease (MDRD) Study formula. examination in a subset of 1241 par-
protocols for conduct of NHANES were This formula was developed based on ticipants in NHANES 1988-1994 were
approved by the NCHS institutional re- 1628 participants in which serum cre- used to estimate the persistence of mi-
©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 7, 2007—Vol 298, No. 17 2039

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

croalbuminuria as described previ- determined using answers to the ques- changes in the prevalence of CKD.
ously.1,9,10 Briefly, the proportion of tions: “Have you smoked at least 100 Among 8728 participants aged 20 to 39
individuals with persistent albumin- cigarettes in your life?” and “Do you now years without diagnosed hypertension
uria was calculated as the proportion smoke cigarettes?” Hypertension and or diabetes, the weighted mean stan-
with albuminuria at a single visit ⫻ diabetes were defined by a physician di- dardized serum creatinine levels were
50.9% for those with microalbumin- agnosis that was self-reported by the par- 0.04 mg/dL (to convert to µmol/L, mul-
uria and an estimated GFR of more ticipant because CKD is most strongly tiply by 88.4) lower in 1988-1994 than
than 90 mL/min/1.73 m2; 75.0% for related to these conditions. Fasting glu- in 1999-2004 (P ⬍ .001). In the con-
those with microalbuminuria and an es- cose (available for only half of the par- servative trends analysis, this value was
timated GFR of 60 to 89 mL/min/1.73 ticipants) and measured blood pres- added to the 1988-1994 data to make
m2; and 100% for those with macroal- sure were not used in the primary the mean value in the young healthy
buminuria regardless of estimated GFR. analysis but yielded similar results in group identical across surveys. The later
These same persistence estimates were sensitivity analyses (data not shown). surveys were chosen as the reference
applied to both surveys and all age, sex, group because their creatinine assay
and race groups thereby influencing Statistical Analyses yielded unbiased results compared with
only the prevalence of CKD stages 1 Analyses were performed incorporat- reference methods.11
and 2 but not trends over time or as- ing the sampling weights to obtain un-
sociations with other factors. Stan- biased estimates from the complex RESULTS
dard errors (SEs) incorporate sam- NHANES sampling design using Stata NHANES 1988-1994 included 15 488
pling variation in the persistence version 8.2 (StataCorp, College Sta- participants and NHANES 1999-2004
estimates above. All albuminuria tion, Texas). The SEs for all estimates included 13 233 participants aged 20
analyses excluded women who were were obtained using the Taylor series years or older examined at the mobile
pregnant or menstruating. (linearization) method following examination center with an estimated
The CKD stages were categorized NHANES recommended procedures GFR of at least 15 mL/min/1.73 m2.
based on the classification system es- and weights.16-18 Estimates and SEs were During the period between the sur-
tablished by the National Kidney Foun- obtained for each survey period. Com- veys, the US population became older
dation Kidney Disease Outcomes Qual- parisons across surveys and 95% con- and included a smaller proportion of
ity Initiative. 1 The CKD stages are fidence interval (CI) estimates for CKD non-Hispanic whites (TABLE 1). The
defined as follows: stage 1, persistent stages incorporating persistence data on shift in age distribution was less pro-
albuminuria with an estimated GFR microalbuminuria were made using nounced in individuals older than 60
higher than 90 mL/min/1.73 m2; stage bootstrap methods implemented in years in which CKD is more common.
2, persistent albuminuria with an esti- Stata. Adjustment was conducted using At the same time, the prevalence of self-
mated GFR of 60 to 89 mL/min/1.73 m2; the direct method for age and logistic reported diabetes and hypertension in-
stage 3, a GFR of 30 to 59 mL/min/ regression for multiple variables. Esti- creased as did the mean BMI and pro-
1.73 m2; and stage 4, a GFR of 15 to 29 mates from this study are nationally rep- portion of the population that is
mL/min/1.73 m 2 . NHANES 1999- resentative of the noninstitutionalized overweight and obese, which are all risk
2004 participants were asked if they US population of adults aged 20 years factors for CKD.
were aware they had “weak or failing or older. Prevalence estimates were ap- Mean albuminuria increased across
kidneys” and the responses were tabu- plied to the 2000 US census to obtain the surveys but mean albumin-to-
lated by the presence of markers of estimates of the number of individu- creatinine ratio was not different among
CKD. als with CKD in the United States in the young healthy individuals (12.2 mg/g
year 2000. In hypothesis testing, a P in 1988-1994 and 12.3 mg/g in 1999-
Assessment of Demographics value of less than .05 was the level of 2004). The mean serum creatinine con-
and Risk Factors for Kidney Disease significance used in this study. centration was higher in 1999-2004
NHANES included measurement of A sensitivity analysis (conservative compared with 1988-1994, correspond-
height and weight, which was used for trends analysis) was conducted by ad- ing to a lower mean estimated GFR in
the calculation of body mass index (BMI; justing serum creatinine so its mean 1999-2004. The conservative trends
calculated as weight in kilograms di- level in a young healthy subgroup was analysis, which added 0.04 mg/dL to the
vided by height in meters squared). In- identical between the 1988-1994 and serum creatinine concentration in
formation on age, sex, race/ethnicity 1999-2004 surveys. The goal of this NHANES 1988-1994, resulted in nearly
(categorized as non-Hispanic white, analysis was to determine if differ- identical mean serum creatinine con-
non-Hispanic black, Mexican Ameri- ences in mean serum creatinine across centrations and mean estimated GFRs
can, or all other) and smoking was based surveys (potentially indicating a re- across surveys.
on self-report during the interview por- sidual laboratory calibration differ- FIGURE 1 shows the distribution of
tion of the survey. Smoking status was ence) might explain some or all of the albuminuria (panel A) and estimated
2040 JAMA, November 7, 2007—Vol 298, No. 17 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

Table 1. Population Characteristics of US Adults Aged 20 Years or Older Based on NHANES 1988-1994 and NHANES 1999-2004 a
NHANES 1988-1994 (n = 15 488) NHANES 1999-2004 (n = 13 233)

No. of Mean (SE); No. of Mean (SE);

Participants % (SE) Median (IQR) Participants % (SE) Median (IQR)
Age, y 15 488 44.8 (0.5) 13 233 46.2 (0.3)
Age group, y
20-39 6367 45.7 (1.0) 4714 39.4 (0.8)
40-59 4194 31.7 (0.6) 3921 38.3 (0.7)
60-69 2174 11.4 (0.5) 2015 10.5 (0.4)
ⱖ70 2753 11.2 (0.7) 2583 11.9 (0.4)
Sex
Female 8214 52.2 (0.5) 6925 51.8 (0.4)
Male 7274 47.9 (0.5) 6308 48.2 (0.4)
Race/ethnicity
Non-Hispanic white 6450 76.9 (1.3) 6764 72.6 (1.7)
Non-Hispanic black 4168 10.3 (0.6) 2477 10.5 (1.0)
Mexican American 4250 5.1 (0.4) 3009 7.3 (0.9)
Other 620 7.7 (0.8) 983 9.6 (1.2)
Diabetes, self-report 1266 5.4 (0.3) 1278 6.8 (0.3)
Hypertension, diagnosed 4211 23.8 (0.7) 4120 27.1 (0.8)
Body mass index b 15 453 26.6 (0.1) 12 857 28.1 (0.1)
⬍25 6073 44.5 (0.9) 4083 34.4 (0.6)
25-29.99 5435 33.1 (0.6) 4640 34.8 (0.7)
ⱖ30 3945 22.3 (0.7) 4134 30.8 (0.7)
Urine albumin, mg/dL c 14 319 26.5 (1.7); 12 216 32.5 (2.1);
6.2 (2.7-12.8) 7.1 (3.7-14.1)
Urine creatinine, mg/dL c 14 319 129.0 (1.4); 12 216 129.7 (1.6);
117.6 (63.3-78.7) 118.0 (65.0-178.0)
Albuminuria (ACR), mg/g c 14 319 25.4 (1.7); 12 216 28.6 (2.3);
5.6 (3.5-10.1) 5.9 (3.9-11.0)
Standard serum creatinine, mg/dL 15 488 0.843 (0.003) 13 233 0.888 (0.003)
Estimated GFR, mL/min/1.73 m2 15 488 92.7 (0.5) 13 233 87.4 (0.4)
Conservative trends analysis
Serum creatinine, mg/dL 15 488 0.884 (0.003) NA NA
Estimated GFR, mL/min/1.73 m2 15 488 87.5 (0.5) NA NA
Abbreviations: ACR, albumin to creatinine ratio; GFR, glomerular filtration rate; IQR, interquartile range; NA, data not available; NHANES, National Health and Nutrition Examination Surveys.
SI conversion factor: To convert creatinine to µmol/L, multiply values by 88.4.
a Examined individuals excluding persons with an estimated GFR of less than 15.
b Calculated as weight in kilograms divided by height in meters squared.
c Only a total of 14 319 participants were available for these analyses because these data required a urine analysis and excluded pregnant or menstruating women.

Figure 1. Albumin to Creatinine Ratio and Estimated Glomerular Filtration Rate (GFR) in NHANES 1988-1994 and 1999-2004

A Albumin to Creatinine Ratio 1988-1994 B Estimated GFR 1988-1994

1999-2004 1988-1994 Conservative
45 12 trends analysis (+0.04 mg/dL)
1.5
40 1999-2004
10
35 1.0
30 8
Percentage

Percentage

25 0.5
6
20
0
15 4
30 40 50 60 70 80 90 100 100- ≥300
10 299
2
5
0 0
<5 10 20 30 40 50 60 70 80 90 100 100- ≥300 <5 10 20 40 60 80 100 120 140 160 180 200
299
Albumin to Creatinine Ratio, mg/g Estimated GFR, mL /min/1.73 m2

The inset in panel A shows the higher range of albuminuria in greater detail. Panel B includes data on GFR estimated from a conservative trends analysis in which 0.04
mg/dL was added to serum creatinine to eliminate the difference in mean GFR between surveys. The vertical line in panel B demarcates an estimated GFR of 60
mL/min/1.73 m2, which defines decreased GFR. For both panels, the statistical testing is done for the overall mean and for meaningful categories (ie, estimated GFR
⬍60 and albumin to creatinine ratio ⱖ30). NHANES indicates National Health and Nutrition Examination Surveys.

©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 7, 2007—Vol 298, No. 17 2041

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

GFR (panel B) in the 2 surveys. The dis- 1988-1994 estimate while prevalence mated GFR while all other subgroups
tribution of albumin to creatinine ra- estimates from each of the three 2-year showed no significant increase or de-
tio is skewed in both surveys. There is surveys had relatively wide 95% CIs. crease in albuminuria.
a somewhat higher prevalence of albu- Moderately reduced GFR increased in The prevalence estimate for each
minuria during 1999-2004 compared prevalence from 5.4% to 7.7% stage of CKD was higher in 1999-
with 1988-1994 in the ranges of 35 to (P ⬍ .001) and the prevalence of se- 2004 than in 1988-1994 with the dif-
49 mg/g and 300 mg/g or higher (panel verely reduced GFR increased from ference being statistically significant for
A). The distribution of estimated GFR 0.21% to 0.35% (P=.02; Table 2). Simi- CKD stages 2 through 4 and overall
is more symmetric with a higher preva- larly, the proportion of the overall (TABLE 3). Stratified analyses by sex and
lence of values below 60 mL/min/1.73 population with microalbuminuria on race showed similar trends. The over-
m2 in 1999-2004 than in 1988-1994. a single occasion increased from 7.1% all prevalence of CKD among men was
The conservative trends analysis aligns to 8.2% (P = .01). The prevalence of 8.2% in 1988-1994 and 11.1% in 1999-
the mean estimated GFR across sur- macroalbuminuria increased some- 2004. Among women, the prevalences
veys with implications for subsequent what from 1.1% to 1.3% but this dif- were 12.1% and 15.0%, respectively. By
sensitivity analyses. ference was well within the limits of ethnicity, the change was from 10.5%
The proportion of the US popula- random variation (P = .37). Subdivid- to 13.8% among non-Hispanic whites,
tion with mild, moderate, or severely ing the prevalence of albuminuria by 10.2% to 11.7% among non-Hispanic
reduced estimated GFR increased from different levels of estimated GFR blacks, and from 6.3% to 8.0% among
1988-1994 to 1999-2004 (TABLE 2). showed that the prevalence of micro- Mexican Americans. The age-adjusted
The combined prevalence estimate for albuminuria increased significantly prevalence odds ratios (ORs) for an es-
1999-2004 had similar precision to the among individuals with normal esti- timated GFR of less than 60 mL/min/

Table 2. Prevalence of Kidney Function and Albuminuria Categories in US Adults Aged 20 Years or Older Based on NHANES 1988-1994 and
NHANES 1999-2004 a
NHANES NHANES
1988-1994 1999-2004

No. of No. of P
Participants % (SE) Participants % (SE) Value
Kidney function (GFR), mL/min/1.73 m2
Normal (ⱖ90) 8600 51.9 (1.1) 5891 40.7 (1.0) ⬍.001
Mildly reduced (60-89) 5751 42.4 (1.0) 5946 51.2 (0.8) ⬍.001
Moderately reduced (30-59) 1088 5.4 (0.3) 1316 7.7 (0.3) ⬍.001
Severely reduced (15-29) 49 0.21 (0.03) 80 0.35 (0.05) .02
Albuminuria (ACR), mg/g b
Overall
Normal 12 655 91.8 (0.4) 10 636 90.5 (0.3) .01
Microalbuminuria 1353 7.1 (0.4) 1315 8.2 (0.3) .01
Macroalbuminuria 311 1.1 (0.1) 265 1.3 (0.1) .37
GFR ⱖ90 mL/min/1.73 m2
Normal 7182 94.1 (0.5) 4594 92.1 (0.5) .003
Microalbuminuria 532 5.3 (0.5) 435 7.2 (0.5) .004
Macroalbuminuria 76 0.6 (0.1) 49 0.7 (0.1) .53
GFR 60-89 mL/min/1.73 m2
Normal 4778 91.8 (0.5) 5143 91.8 (0.4) .96
Microalbuminuria 568 7.2 (0.4) 588 7.4 (0.4) .69
Macroalbuminuria 117 1.0 (0.1) 83 0.8 (0.1) .29
GFR 30-59 mL/min/1.73 m2
Normal 682 72.2 (1.9) 877 75.6 (1.6) .18
Microalbuminuria 243 22.0 (1.8) 270 18.3 (1.3) .10
Macroalbuminuria 96 5.8 (0.8) 104 6.1 (0.7) .82
GFR 15-29 mL/min/1.73 m2
Normal 13 37.5 (12.5) 22 34.0 (7.9) .81
Microalbuminuria 10 19.9 (8.6) 22 23.7 (6.1) .72
Macroalbuminuria 22 42.6 (8.8) 29 42.4 (8.2) .98
Abbreviations: ACR, albumin to creatinine ratio; GFR, glomerular filtration rate; NHANES, National Health and Nutrition Examination Surveys.
a Age-adjusted prevalence estimates for microalbuminuria and macroalbuminuria in 1988-1994 adjusted to the 1999-2004 age distribution in Table 1 are 7.2% and 1.2%, respectively.
b Women who were pregnant or in menses were excluded. Normal was defined as an ACR of less than 30; microalbuminuria, an ACR of 30 to 299; and macroalbuminuria, an ACR of 300
or greater.

2042 JAMA, November 7, 2007—Vol 298, No. 17 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

1.73 m2 were all between 1.4 and 1.5 bined were 0.2% in 1988-1994 and higher prevalence was partly explained
and statistically significant in men, 0.7% in 1999-2004 at aged 20 to 39 by the older age and high proportion of
women, non-Hispanic whites and years compared with 27.8% in 1988- minority groups (OR declined from 1.18
blacks, and Mexican Americans with a 1994 and 37.8% in 1999-2004 after the to 1.12 after adjustment). Further ad-
somewhat weaker association in the age of 70 years. The dramatic increase justment for the higher prevalence of di-
smaller number of individuals of other in prevalence is consistent with the agnosed diabetes and hypertension and
ethnicity. known decline of measured GFR with higher BMI explained practically all of
Trends over time were also similar age.23 At each age, the prevalence of de- the difference (OR declined to 1.03). In
within age categories, indicating the creased GFR was higher in the 1999- the fully adjusted models, the preva-
trends were not due to age differences in 2004 survey than in the 1988-1994 sur- lence of albuminuria was strongly asso-
the population (FIGURE 2). In both vey (P⬍.05 for ages 20-39, 40-59, and ciated with diagnosed diabetes (OR, 3.58;
NHANES surveys, the prevalence of ⱖ70 years). 95% CI, 3.12-4.12) and hypertension
CKD increased with age with stages 1 Differences in prevalence of de- (OR, 1.70; 95% CI, 1.10-1.92) as well as
and 2 increasing from 2% to 3% at age creased GFR and albuminuria between older age and all race/ethnicity groups
20 to 39 years to 9% to 10% after the 1988-1994 and 1999-2004 remain sub- other than non-Hispanic whites
age of 70 years. The prevalence of stage stantial after adjustment for changes in (P⬍.001) but not higher BMI (P=.12).
1 alone did not increase with age be- the age, sex, and race/ethnic composi- The prevalence OR of estimated GFR of
cause an increasing prevalence of al- tion of the US population during this pe- less than 60 mL/min/1.73 m2 in 1999-
buminuria was offset by a decreasing riod (TABLE 4). The higher prevalence 2004 compared with 1988-1994 was
proportion of individuals with a GFR of diagnosed diabetes, hypertension, and 1.47. Age adjustment had little impact,
of 90 mL/min/1.73 m2 or higher. The higher BMI explained some of the higher likely because the increase in the num-
prevalences for stage 3 and 4 com- prevalence. For albuminuria trends, the ber of older individuals was offset by a

Table 3. Prevalence of Chronic Kidney Disease (CKD) Stages in US Adults Aged 20 Years or Older Based on NHANES 1988-1994 and
NHANES 1999-2004
Prevalence, % (95% CI) Prevalence Ratio for Estimated No. of
CKD NHANES 1999-2004 US Adults in 2000,
Stage a NHANES 1988-1994 NHANES 1999-2004 to 1988-1994 (95% CI) No. in Millions (95% CI)
1 1.71 (1.28-2.18) 1.78 (1.35-2.25) 1.05 (0.85-1.30) 3.6 (2.7-4.5)
2 2.70 (2.17-3.24) 3.24 (2.61-3.88) 1.21 (1.03-1.41) 6.5 (5.2-7.8)
3 5.42 (4.89-5.95) 7.69 (7.02-8.36) 1.42 (1.25-1.62) 15.5 (14.1-16.8)
4 0.21 (0.15-0.27) 0.35 (0.25-0.45) 1.70 (1.11-2.51) 0.7 (0.5-0.9)
5 NA NA NA NA
Total 10.03 (9.16-10.91) 13.07 (12.04-14.10) 1.30 (1.19-1.43) 26.3 (24.2-28.3)
Abbreviations: CI, confidence interval; NA, data not included because patients with CKD stage 5 were excluded; NHANES, National Health and Nutrition Examination Surveys.
a Defined based on standard criteria1: stage 1, persistent albuminuria with glomerular filtration rate (GFR) higher than 90 mL/min/1.73 m2; stage 2, persistent albuminuria with GFR
of 60 to 89 mL/min/1.73 m2; stage 3, GFR of 30 to 59 mL/min/1.73 m2; stage 4, GFR of 15 to 29 mL/min/1.73 m2. The age-adjusted prevalence rates for CKD stages 1, 2, 3, and
4 in 1988-1994 adjusted to the 1999-2004 age distribution in Table 1 are 1.7%, 2.8%, 5.6%, and 0.2%, respectively, for a total of 10.3%.

Figure 2. Prevalence of Chronic Kidney Disease (CKD) Stages by Age Group in NHANES 1988-1994 and 1999-2004

Overall CKD Stages 1 and 2 CKD Stage 3 CKD Stage 4

50
Survey years
45
1988-1994
40 1999-2004
35
Prevalence, %

30
25
20
15
10
5
a a
0
20-39 40-59 60-69 ≥70 20-39 40-59 60-69 ≥70 20-39 40-59 60-69 ≥70 20-39 40-59 60-69 ≥70
Age Group, y Age Group, y Age Group, y Age Group, y

NHANES indicates National Health and Nutrition Examination Surveys.

a There were no cases in 1988-1994.

©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 7, 2007—Vol 298, No. 17 2043

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

similar increase in the number of younger CI, 1.02-1.34). After full adjustment in flected an increase in microalbumin-
individuals, allowing the percentage of the conservative trends analysis, the uria as well as an increase in the preva-
individuals aged 60 years or older to re- prevalence OR of decreased GFR be- lence of moderately and severely reduced
main relatively unchanged (Table 1). The tween surveys was 1.08 (95% CI, 0.94- estimated GFR. Overall, the preva-
prevalence OR increased further to 1.53 1.24), indicating that the differences in lence rate of CKD increased from 10.0%
after adjustment for age, sex, and race mean serum creatinine concentration, (95% CI, 9.1-10.9) to 13.1% (95% CI,
due to the lower prevalence of de- demographics, diagnosed diabetes, hy- 12.0-14.1). This increase was only
creased GFR among minority groups. pertension, and BMI between surveys slightly explained by the aging of the US
The OR decreased to 1.43 with addi- explain nearly all of the difference in population because the age-adjusted
tional adjustment for diagnosed diabe- prevalence of low GFR between 1988- prevalence estimate of CKD stages 1
tes, hypertension, and BMI. In the fully 1994 and 1999-2004 (P=.27). through 4 only increased to 10.3% when
adjusted model, the prevalence of low The proportion of individuals who the 1988-1994 prevalence was ad-
GFR was strongly associated with diag- reported being aware they had weak or justed to the 1999-2004 age structure.
nosed diabetes (OR, 1.54; 95% CI, 1.28- failing kidneys in 1999-2004 (not asked The increased prevalence of diagnosed
1.80) and hypertension (OR, 1.98; 95% in 1988-1994) was low. In CKD stage 3, diabetes and hypertension and obesity
CI, 1.73-2.67) as well as higher BMI (OR, 11.6% (SE, 2.0%) of men and only 5.5% explained some of the increase in preva-
1.08; 95% CI, 1.02-1.15 per 5-unit in- (SE, 0.8%) of women reported being lence. In contrast, these factors ex-
crement of BMI) and older age but was aware of having weak or failing kid- plained the entire increase in the preva-
lower among males, non-Hispanic neys. Even among men with CKD stage lence of albuminuria. Awareness of CKD
blacks, and Mexican Americans com- 3 and elevated albuminuria, awareness remains very low, even among individu-
pared with non-Hispanic whites of weak or failing kidneys was only 22.8% als with both reduced kidney function
(P⬍.001). (SE, 3.9%). In stage 4, there was no longer and albuminuria. The estimated num-
The conservative trends analysis a sex difference but the percentage was ber of individuals with CKD increased
showed that the difference in mean se- still only 42% (SE, 8%). Awareness rates even more dramatically because of the
rum creatinine concentration be- increased among women between 1999- growth in the US population aged 20
tween surveys accounts for much but 2000 and 2003-2004 but rates were low years or older from 178 million in 1990
possibly not all of the higher preva- in all survey periods. to 197 million in 2000.
lence of lower GFR in 1999-2004. In The increase in prevalence of CKD
this analysis, the prevalence of CKD in COMMENT is partly explained by the increase in a
1988-1994 was higher (1.5 for stage 1, Analysis of survey data from a represen- number of CKD risk factors, includ-
2.8 for stage 2, 6.7 for stage 3, and 0.23 tative sample of the US population shows ing an aging US population and an in-
for stage 4 for a total of 11.3). The that the prevalence rate of CKD is high. crease in the proportion of individuals
prevalence OR of estimated GFR less Estimates from 1999-2004 are higher with obesity, diagnosed diabetes, and
than 60 mL/min/1.73 m2 was 1.17 (95% than those in 1988-1994, which re- hypertension.5,24,25 The proportion of

Table 4. Logistic Regression of Albuminuria and Decreased Estimated Glomerular Filtration Rate (GFR) Comparing NHANES 1999-2004 With
NHANES 1988-1994 Before and After Adjustment
Conservative
Trends Trends Analysis a

OR (95% CI) P Value OR (95% CI) P Value

Albuminuria in NHANES 1999-2004 vs 1988-1994
Unadjusted 1.18 (1.03-1.34) .01
Adjusted for age 1.15 (1.00-1.32) .05
Sex and race b 1.12 (0.99-1.28) .08
Diagnosed diabetes and hypertension b 1.06 (0.93-1.21) .39
Body mass index b 1.03 (0.90-1.18) .63
Estimated GFR ⬍60 mL/min/1.73 m2 in NHANES 1999-2004 vs 1988-1994
Unadjusted 1.47 (1.27-1.69) ⬍.001 1.17 (1.02-1.34) .03
Adjusted for age 1.50 (1.31-1.73) ⬍.001 1.13 (0.99-1.30) .07
Sex and race b 1.53 (1.33-1.76) ⬍.001 1.15 (1.00-1.32) .05
Diagnosed diabetes and hypertension b 1.45 (1.27-1.67) ⬍.001 1.10 (0.96-1.26) .17
Body mass index b 1.43 (1.24-1.63) ⬍.001 1.08 (0.94-1.24) .29
Abbreviations: CI, confidence interval; NHANES, National Health and Nutrition Examination Surveys; OR, odds ratio.
a Serum creatinine among young healthy participants (aged 20-39 years without diabetes and hypertension) was adjusted to be identical across surveys by adding 0.04 mg/dL to
serum creatinine in NHANES 1988-1994.
b Indicates addition of variables to the model in the previous row.

2044 JAMA, November 7, 2007—Vol 298, No. 17 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

minority populations has also in- tion to the Roche Enzymatic method.28 serum creatinine concentration in the
creased although the prevalence of CKD The new calibration resulted in higher 1999-2004 survey compared with the
in these populations in NHANES is not values in the 1988-1994 and 1999- 1988-1994 survey, even among young
higher than that in non-Hispanic 2000 surveys (0.05 and 0.10 mg/dL, re- participants without diagnosed diabe-
whites. This lower CKD prevalence spectively, at 1.0 mg/dL).11 The previ- tes or hypertension. The persistence of
among ethnic minorities compared with ous calibration of the 1999-2000 survey the increase in CKD prevalence from
whites is in contrast to the higher di- was conducted during a single month 1988-1994 to 1999-2004 even after
alysis rates in these groups and re- of the survey rather than a random equalizing the mean serum creatinine
mains unexplained. One explanation sample of specimens from throughout concentration in a conservative trends
could be that the MDRD Study equa- the survey period. Thus, only recently analysis supports the conclusion that
tion performs differently in various eth- has there been data from a panel of the increase in CKD prevalence is real.
nic and racial groups. Mexican Ameri- masked frozen specimens from each However, CKD prevalence estimates are
cans and Native Americans in particular survey analyzed at a single laboratory sensitive to small differences in serum
were not represented in the MDRD to provide a direct comparison of se- creatinine concentration and the mag-
Study in which the equation was de- rum creatinine concentration across all nitude of the increase in prevalence
veloped. Nonetheless, the MDRD Study of the surveys. The high correlation be- could be smaller than that suggested by
equation was validated in 5504 indi- tween the original creatinine mea- a simple comparison across surveys.
viduals and showed little bias among sures and the assays performed for cali- The new data on trends in CKD pro-
blacks.26 An underestimate in whites bration (r=0.98 for 1988-1994, r=0.98 vide a larger context for trends in kid-
and women may explain some of the for 1999-2000, r =0.98 for 2001-2002, ney failure treated by dialysis and trans-
higher prevalence in these groups. It has and r = 0.99 for 2003-2004) lends sup- plantation. The age- and sex-adjusted
been hypothesized that the lower preva- port to our ability to make reliable com- incidence of end-stage renal disease in
lence of CKD among blacks could be parisons. The prevalence estimates pre- the United States increased 42% be-
due to hyperfiltration and faster pro- sented herein differ from those in a tween 1991 and 2001. During this time,
gression through CKD stage 3.10,27 How- recent brief report,29 which also ana- there has been a consistent slowing in
ever, differential errors in GFR estima- lyzed NHANES data, for 3 main rea- the rate of growth of treated kidney fail-
tion cannot be ruled out. Studies sons: we estimated CKD stage 1 or 2 ure rates from greater than 10% per year
measuring GFR in population-based based on persistent albuminuria rather in the 1980s to a decrease of 1.1% be-
samples of individuals of different eth- than counting all individuals with mi- tween 2002 and 2004. However, the
nicities and ages are needed. Adjust- croalbuminuria; we used the MDRD 2004 rate of 339 per million is still far
ment for all risk factors explained nearly Study equation expressed in standard higher than the Healthy People 2010
the entire increase in the prevalence of creatinine because it corresponds to the goal of 217 per million population.4
albuminuria but only a fraction of the calibration of NHANES serum creati- Furthermore, models of the treated kid-
increase in the prevalence of low esti- nine concentration to standard creati- ney failure epidemic suggest that even
mated GFR (Table 4). This is partly ex- nine; and we also estimated the effect with adjusted rates leveling off, the
plained by the stronger association of of a conservative trends analysis, which growing prevalence of diabetes and the
diabetes with albuminuria compared indicates much of the change in the aging of the population will result in a
with low GFR. In addition, albumin- prevalence of CKD is related to a subtle progressive increase in the number of
uria and low GFR show opposite asso- but influential difference in the esti- patients treated for kidney failure. It is
ciations with non-Hispanic black and mate mean serum creatinine concen- estimated that by 2015 there will be
Mexican American race/ethnicity. tration of the population. 136 000 patients with incident end-
The CKD prevalence estimates re- Despite an updated laboratory cali- stage renal disease per year and 712 000
ported herein are higher than those re- bration traceable to criterion standard patients with prevalent disease.30 Our
ported previously. For 1988-1994, the reference methods using a random analysis suggests that the increasing
prevalence based on the recent sample of specimens from each prevalence of diabetes is already lead-
calibration to standard serum creati- NHANES survey, small residual differ- ing to a measurable increase in the ear-
nine concentration is 10.0% com- ences in laboratory measurements lier stages of CKD (stages 1-4).
pared with a prevalence of 8.8% re- across surveys cannot be ruled out. In- Interpretation of the high preva-
ported previously using the same creased muscle mass or increased pro- lence of CKD in the US population
criteria. This difference is the result of tein intake could also have increased the should take into account its wide spec-
a difference between the initial calibra- mean serum creatinine concentration. trum of disease severity, etiology, and
tion of serum creatinine concentra- Laboratory or non–kidney-related ef- comorbid conditions. Within CKD
tion to the MDRD Study Beckman Syn- fects on serum creatinine concentra- stages 1 and 2, persistent microalbu-
chron CX3 method for assaying tion were of particular concern be- minuria outnumbers macroalbumin-
creatinine and the most recent calibra- cause there was a higher mean standard uria approximately 9 to 1. Although al-
©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 7, 2007—Vol 298, No. 17 2045

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

buminuria is the strongest risk factor suming the same persistence in the dif- Statistical analysis: Coresh, Selvin.
Obtained funding: Coresh, Kusek, Eggers, Levey.
for CKD progression, it is also associ- ferent surveys, the trends observed can- Administrative, technical, or material support: Stevens,
ated with an increased risk of cardio- not be the result of changes in estimated Eggers, Van Lente, Levey.
Study supervision: Coresh, Levey.
vascular morbidity and mortality.13 persistence rate of microalbuminuria. Financial Disclosures: None reported.
Chronic kidney disease stages 3 and 4 Differential nonresponse, particularly Funding/Support: The research for this study was sup-
ported by grants UO1 DK 053869, UO1 DK 067651,
are marked by complications of re- lower response among sicker individu- and UO1 DK 35073 from the National Institute of Dia-
duced GFR. Chronic kidney disease als, could have biased the present preva- betes and Digestive and Kidney Diseases. Dr Selvin was
stage 3 outnumbers CKD stage 4 ap- lence estimates. This bias may be most supported by grant K01 DK076595 from the Na-
tional Institute of Diabetes and Digestive and Kidney
proximately 20 to 1; within stage 3, the plausible for the more advanced stages Diseases.
upper half of estimated GFR outnum- of CKD and thus, the prevalence of Role of the Sponsor: Two of the coauthors, Drs Kusek
and Eggers, are employees of the National Institute
bers the lower half by 3 to 1. In one CKD stage 4 may be underestimated of Diabetes and Digestive and Kidney Diseases. The
study,13 CKD stage 4 was associated among individuals who volunteered to National Institute of Diabetes and Digestive and Kid-
ney Diseases had no input into the data collection, con-
with a 17.6% risk of progression to kid- participate in NHANES. In addition, duct, or management of the study. The National In-
ney failure and a 45.7% mortality risk both NHANES surveys only sampled stitute of Diabetes and Digestive and Kidney Diseases
during a mean follow-up of 3 years. By noninstitutionalized adults. Chronic staff commented on the analysis, interpretation of the
data, preparation, and review; and approved the manu-
contrast, in CKD stage 3, the risk for kidney disease prevalence in nursing script. The National Health and Nutrition Examina-
progression to kidney failure was 1.1% homes and its trends over time is un- tion Surveys were conducted by the US Centers for
Disease Control and Prevention.
but the mortality risk was 24.3% com- known but likely to be high. In 2000, Previous Presentation: This study was presented in part
pared with 0.07% and 10.2%, respec- 1.6 million (4.5%) adults older than age at the American Society of Nephrology; November 2,
2007; San Francisco, California.
tively, in patients without CKD. Bet- 65 years were in nursing homes.33 Additional Contributions: We thank the National Cen-
ter data are needed to allow physicians In conclusion, survey data suggest ter of Health Statistics for the data collection, Geral-
to combine information on estimated that the prevalence of CKD in the dine McQuillan, PhD, and David A. Lacher, MD, MEd,
for the laboratory calibration studies, and Lester R. Cur-
GFR, albuminuria, and other clinical United States is high and has in- tin, PhD, for advice on combining weights across sur-
characteristics to predict kidney dis- creased between 1988-1994 and 1999- veys. None of the individuals listed herein received any
compensation. All 3 work for the National Center for
ease progression, complications of re- 2004, from 10% to 13%, while aware- Health Statistics, Hyattsville, Maryland.
duced GFR and cardiovascular dis- ness of kidney disease among the
ease, as well as to establish the balance general public remains very low. The REFERENCES
of risks and benefits for the treatment increasing prevalence of diagnosed dia-
1. National Kidney Foundation. K/DOQI Clinical prac-
of different patient subgroups with betes and hypertension has contrib- tice guidelines for chronic kidney disease: evaluation,
CKD. uted to this increase, which may propa- classification and stratification. Am J Kidney Dis. 2002;
39(2)(suppl 1):S1-S266.
This study has a number of limita- gate to higher rates of complications and 2. Levey AS, Eckardt KU, Tsukamoto Y, et al. Defi-
tions including reliance on estimation kidney failure requiring dialysis or nition and classification of chronic kidney disease: a
of GFR, rather than direct measure- transplantation. Earlier stages ac- position statement from Kidney Disease: Improving
Global Outcomes (KDIGO). Kidney Int. 2005;67
ment using injection of an exogenous counted for most of the individuals with (6):2089-2100.
marker. Equations for estimating GFR CKD. Because individuals with early 3. Sarnak MJ, Levey AS, Schoolwerth AC, et al. Kid-
ney disease as a risk factor for development of car-
have limited precision compared with stages of CKD have a higher risk of car- diovascular disease: a statement from the American
measured GFR. Imprecision and bias diovascular disease morbidity and mor- Heart Association Councils on Kidney in Cardiovas-
cular Disease, High Blood Pressure Research, Clinical
are greater at higher GFR, limiting the tality than their risk of progression to Cardiology, and Epidemiology and Prevention.
accuracy of classification in the mildly kidney failure, cardiovascular risk fac- Hypertension. 2003;42(5):1050-1065.
4. US Renal Data Systems. USRDS 2006 Annual Data
decreased GFR group.26,31 In addition, tor management in this group is criti- Report: Atlas of End-Stage Renal Disease in the United
systematic differences between the mea- cal. The high prevalence of CKD over- States. Bethesda, MD: National Institutes of Health,
sured and estimated GFR may be in- all, and particularly among older National Institute of Diabetes and Digestive and Kid-
ney Diseases; 2007.
fluenced by the population in which the individuals and persons with hyper- 5. Gregg EW, Cheng YJ, Cadwell BL, et al. Secular
equation is applied.31,32 Such a bias tension and diabetes, suggests that CKD trends in cardiovascular disease risk factors accord-
ing to body mass index in US adults. JAMA. 2005;
might contribute to racial/ethnic dif- needs to be a central part of future pub- 293(15):1868-1874.
ferences in prevalence estimates, but lic health planning. 6. Mokdad AH, Ford ES, Bowman BA, et al. Preva-
lence of obesity, diabetes, and obesity-related health
any systematic difference would apply Author Contributions: Dr Coresh had full access to all risk factors, 2001. JAMA. 2003;289(1):76-79.
to all of the NHANES surveys and of the data in the study and takes responsibility for the 7. Fields LE, Burt VL, Cutler JA, Hughes J, Roccella
would be unlikely to affect the trends integrity of the data and the accuracy of the data analysis. EJ, Sorlie P. The burden of adult hypertension in the
Study concept and design: Coresh, Selvin. United States 1999 to 2000: a rising tide. Hypertension.
reported herein. The persistence of al- Acquisition of data: Coresh, Selvin, Manzi, Stevens, 2004;44(4):398-404.
buminuria was estimated based on lim- Eggers, Van Lente. 8. Hajjar I, Kotchen TA. Trends in prevalence, aware-
Analysis and interpretation of data: Coresh, Selvin, ness, treatment, and control of hypertension in the
ited data and assumed to be the same Stevens, Manzi, Kusek, Eggers, Levey. United States, 1988-2000. JAMA. 2003;290(2):199-
across surveys, age, and other sub- Drafting of the manuscript: Coresh, Selvin. 206.
Critical revision of the manuscript for important in- 9. Coresh J, Astor BC, Greene T, Eknoyan G, Levey
groups. The data are not sufficient to tellectual content: Coresh, Selvin, Stevens, Manzi, AS. Prevalence of chronic kidney disease and de-
test this assumption. However, by as- Kusek, Eggers, Van Lente, Levey. creased kidney function in the adult US population:

2046 JAMA, November 7, 2007—Vol 298, No. 17 (Reprinted) ©2007 American Medical Association. All rights reserved.

Downloaded From: on 08/26/2018

 CHRONIC KIDNEY DISEASE PREVALENCE

Third National Health and Nutrition Examination 18. National Center for Health Statistics; Centers. for 25. Kannel WB, Vokonas PS. Primary risk factors for
Survey. Am J Kidney Dis. 2003;41(1):1-12. Disease Control and Prevention. Analytic and coronary heart disease in the elderly: the Framing-
10. Coresh J, Byrd-Holt D, Astor BC, et al. Chronic reporting guidelines: the Third National Health ham Study. In: Wenger NK, Furberg CD, Pitt E, eds.
kidney disease awareness, prevalence, and trends and Nutrition Examination Survey, NHANES III Coronary Heart Disease in the Elderly. New York, NY:
among US adults, 1999 to 2000. J Am Soc Nephrol. (1988-94). http://www.cdc.gov/nchs/data/nhanes Elsevier Science Publishing Co Inc; 1986:60-92.
2005;16(1):180-188. /nhanes3/nh3gui.pdf. Accessibility verified Septem- 26. Stevens LA, Coresh J, Feldman HI, et al. Evalua-
11. Selvin E, Manzi J, Stevens LA, et al. Calibration ber 25, 2007. tion of the modification of diet in renal disease study
of serum creatinine in the National Health 19. US Department of Health and Human Services. equation in a large diverse population. J Am Soc
and Nutrition Examination Surveys (NHANES) Health statistics: plan and operation of the Third Na- Nephrol. 2007;18(10):2749-2757.
1988-1994, 1999, 2004. Am J Kidney Dis. tional Health and Nutrition Examination Survey, 27. Hsu CY, Lin F, Vittinghoff E, Shlipak MG. Racial
doi:10.1053/j.ajkd.2007.08.020. 1988-94. http://www.cdc.gov/nchs/data/series/sr_01 differences in the progression from chronic renal
12. Go AS, Lo JC. Epidemiology of non–dialysis- /sr01_032.pdf. Accessibility verified September 25, insufficiency to end-stage renal disease in the United
requiring chronic kidney disease and cardiovascular 2007. States. J Am Soc Nephrol. 2003;14(11):2902-
disease. Curr Opin Nephrol Hypertens. 2006;15 20. National Center for Health Statistics; Centers for 2907.
(3):296-302. Disease Control. Survey operations manuals, bro- 28. Levey AS, Coresh J, Greene T, et al. Expressing
13. Keith DS, Nichols GA, Gullion CM, Brown JB, Smith chures, and consent documents: 1999-current the Modification of Diet in Renal Disease Study equa-
DH. Longitudinal follow-up and outcomes among a NHANES. http://www.cdc.gov/nchs/about/major tion for estimating glomerular filtration rate with stan-
population with chronic kidney disease in a large man- /nhanes/currentnhanes.htm. Accessibility verified Sep- dardized serum creatinine values. Clin Chem. 2007;
aged care organization. Arch Intern Med. 2004;164 tember 25, 2007. 53(4):766-772.
(6):659-663. 21. Levey AS, Bosch JP, Lewis JB, Greene T, Rogers 29. Prevalence of chronic kidney disease and associ-
14. Manjunath G, Tighiouart H, Coresh J, et al. Level N, Roth D; Modification of Diet in Renal Disease Study ated risk factors–United States, 1999-2004. MMWR
of kidney function as a risk factor for cardiovascular Group. A more accurate method to estimate glomer- Morb Mortal Wkly Rep. 2007;56(8):161-165.
outcomes in the elderly. Kidney Int. 2003;63(3): ular filtration rate from serum creatinine: a new pre- 30. Gilbertson DT, Liu J, Xue JL, et al. Projecting the
1121-1129. diction equation. Ann Intern Med. 1999;130(6): number of patients with end-stage renal disease in the
15. Manjunath G, Tighiouart H, Ibrahim H, et al. Level 461-470. United States to the year 2015. J Am Soc Nephrol.
of kidney function as a risk factor for atherosclerotic 22. Levey AS, Coresh J, Greene T, et al. Using stan- 2005;16(12):3736-3741.
cardiovascular outcomes in the community. J Am Coll dardized serum creatinine values in the modification 31. Stevens LA, Coresh J, Greene T, Levey AS. As-
Cardiol. 2003;41(1):47-55. of diet in renal disease study equation for estimating sessing kidney function–measured and estimated glo-
16. US Department of Health and Human Services; glomerular filtration rate. Ann Intern Med. 2006; merular filtration rate. N Engl J Med. 2006;354(23):
Centers for Disease Control and Prevention. Na- 145(4):247-254. 2473-2483.
tional Health and Nutrition Examination Survey. http: 23. Davies DF, Shock NW. Age changes in glomer- 32. Rule AD, Larson TS, Bergstralh EJ, Slezak JM, Ja-
//www.cdc.gov/nchs/nhanes.htm. Accessed Janu- ular filtration rate, effective renal plasma flow, and tu- cobsen SJ, Cosio FG. Using serum creatinine to esti-
ary 11, 2007. bular excretory capacity in adult males. J Clin Invest. mate glomerular filtration rate: accuracy in good health
17. National Center for Health Statistics. National 1950;29(5):496-507. and in chronic kidney disease. Ann Intern Med. 2004;
Health and Nutrition Examination Survey (NHANES) 24. Ong KL, Cheung BM, Man YB, Lau CP, Lam KS. 141(12):929-937.
Analytic Guidelines. http://www cdc gov/nchs/about Prevalence, awareness, treatment, and control of hy- 33. US Census Bureau. PCT 16 group quarters popu-
/major/nhanes/nhanes2003-2004/analytical_guide- pertension among United States adults 1999-2004. lation by group quarters type. http://factfinder.census
lines htm. Accessed May 9, 2007. Hypertension. 2007;49(1):69-75. .gov. Accessibility verified September 26, 2007.

Medicine is the only profession that labors inces-

santly to destroy the reason for its existence.
—James Bryce (1838-1922)

©2007 American Medical Association. All rights reserved. (Reprinted) JAMA, November 7, 2007—Vol 298, No. 17 2047

Downloaded From: on 08/26/2018