Articles

Prevalence of chronic kidney disease in China:

a cross-sectional survey
Luxia Zhang*, Fang Wang*, Li Wang†, Wenke Wang†, Bicheng Liu†, Jian Liu†, Menghua Chen†, Qiang He†, Yunhua Liao†, Xueqing Yu†,
Nan Chen†, Jian-e Zhang, Zhao Hu, Fuyou Liu, Daqing Hong, Lijie Ma, Hong Liu, Xiaoling Zhou, Jianghua Chen, Ling Pan, Wei Chen,
Weiming Wang, Xiaomei Li, Haiyan Wang

Summary
Background The prevalence of chronic kidney disease is high in developing countries. However, no national survey of Lancet 2012; 379: 815–822
chronic kidney disease has been done incorporating both estimated glomerular filtration rate (eGFR) and albuminuria This online publication has
in a developing country with the economic diversity of China. We aimed to measure the prevalence of chronic kidney been corrected. The corrected
version first appeared at
disease in China with such a survey.
thelancet.com on August 17,
2012
Methods We did a cross-sectional survey of a nationally representative sample of Chinese adults. Chronic kidney See Editorial page 777
disease was defined as eGFR less than 60 mL/min per 1·73 m² or the presence of albuminuria. Participants completed See Comment page 783
a lifestyle and medical history questionnaire and had their blood pressure measured, and blood and urine samples *These authors contributed
taken. Serum creatinine was measured and used to estimate glomerular filtration rate. Urinary albumin and creatinine equally
were tested to assess albuminuria. The crude and adjusted prevalence of indicators of kidney damage were calculated †These authors contributed
and factors associated with the presence of chronic kidney disease analysed by logistic regression. equally
Peking University Institute of
Findings 50 550 people were invited to participate, of whom 47 204 agreed. The adjusted prevalence of eGFR less than Nephrology, Division of
Nephrology, Peking University
60 mL/min per 1·73 m² was 1·7% (95% CI 1·5–1·9) and of albuminuria was 9·4% (8·9–10·0). The overall prevalence
First Hospital, Beijing, China
of chronic kidney disease was 10·8% (10·2–11·3); therefore the number of patients with chronic kidney disease in (L Zhang MD, F Wang MD,
China is estimated to be about 119·5 million (112·9–125·0 million). In rural areas, economic development was Prof X Li MD, Prof H Wang MD);
independently associated with the presence of albuminuria. The prevalence of chronic kidney disease was high in north Division of Nephrology,
Sichuan Academy of Medical
(16·9% [15·1–18·7]) and southwest (18·3% [16·4–20·4]) regions compared with other regions. Other factors
Sciences and Sichuan
independently associated with kidney damage were age, sex, hypertension, diabetes, history of cardiovascular disease, Provincial People’s Hospital,
hyperuricaemia, area of residence, and economic status. Chengdu, China
(Prof L Wang MD, D Hong MD);
Division of Nephrology,
Interpretation Chronic kidney disease has become an important public health problem in China. Special attention
Chifeng Second Hospital,
should be paid to residents in economically improving rural areas and specific geographical regions in China. Chifeng, China
(Prof Wen Wang MD, L Ma MD);
Funding The Ministry of Science and Technology (China); the Science and Technology Commission of Shanghai; the Institute of Nephrology and
Division of Nephrology,
National Natural Science Foundation of China; the Department of Health, Jiangsu Province; the Sichuan Science and Zhongda Hospital, Southwest
Technology Department; the Ministry of Education (China); the International Society of Nephrology Research University, Nanjing, China
Committee; and the China Health and Medical Development Foundation. (Prof B Liu MD, H Liu MD);
Division of Nephrology, the
First Hospital of Xinjiang
Introduction will result in an even greater burden of chronic kidney University of Medicine,
In the past 10 years, chronic kidney disease has received disease in the future, and is likely to have substantial Uramuqi, China (Prof J Lui MD);
increased attention as a leading public health problem. socioeconomic and public health consequences in Division of Nephrology, the
The burden of chronic kidney disease is not restricted resource-poor countries.4 China is the world’s largest Affiliated Hospital of Ningxia
University of Medicine,
to its effect on demands for renal replacement therapy; developing country. Previous studies have indicated that Yinchuan, China
the disease has other major effects on the overall the prevalence of chronic kidney disease differs sub- (Prof M Chen MD, X Zhou MD);
population. For example, in the general population, stantially between geographical regions in China.6–9 This Division of Nephrology, the
and in high-risk diabetic or hypertensive populations, heterogeneity might be related to variability in lifestyles First Affiliated Hospital of
Zhejiang University,
outcomes, especially mortality and cardiovascular and economic development, or differences in study Hangzhou, China
events, are strongly affected by kidney involvement.1–3 methodology.6–9 National surveys of chronic kidney (Prof Q He MD, J Chen MD);
Therefore, through its effect on cardiovascular risk disease prevalence in a developing country that incor- Division of Nephrology, the
and outcomes as well as end-stage renal disease, porate estimated glomerular filtration rate (eGFR) and First Affiliated Hospital,
Guangxi University of
chronic kidney disease directly affects the global albuminuria, as well as the ethnic and economic Medicine, Nanning, China
burden of death caused by cardiovascular disease, the heterogeneity seen in China, are rare. (Prof Y Liao MD, L Pan MD);
most common cause of premature morbidity and The China National Survey of Chronic Kidney Division of Nephrology, the
First Affiliated Hospital, Sun
mortality worldwide.1 Disease, done from January, 2007, to October, 2010, was
Yat-sen University,
Chronic kidney disease is highly prevalent in devel- a cross-sectional study designed to provide reliable data Guangzhou, China
oping countries.4,5 Rapid increase in the prevalence of for the prevalence of chronic kidney disease and (Prof X Yu MD, W Chen MD);
risk factors such as diabetes, hypertension, and obesity associated factors in adults in China. Division of Nephrology, Ruijin

www.thelancet.com Vol 379 March 3, 2012 815

 Articles
Hospital, Shanghai Jiao Tong
University, Shanghai, China
(Prof N Chen MD,
Wei Wang MD); Division of
Nephrology, Taihe Hospital,
Hubei University of Medicine,
Shiyan, China
(Prof J-E Zhang MD); Division
of Nephrology, Qilu Hospital,
Shandong University, Jinan,
China (Prof Z Hu MD); and
Division of Nephrology, the
Second Xiangya Hospital,
Central South University,
Changsha, China
(Prof F Liu MD)
Correspondence to:
Prof Haiyan Wang,
Renal Division, Department of
Medicine, Peking University First
Hospital, Xicheng, Beijing, China
why@bjmu.edu.cn
See Online for appendix
For more on the definitions of
urban and rural see http://www.
stats.gov.cn/tjbz/cxfldm/2010/
index.html
816
Methods
Participants
We used a multistage, stratified sampling method to
obtain a representative sample of people aged 18 years
or older in the general population. We selected
13 provinces (Beijing, Sichuan, Inner Mongolia
Autonomous Region, Jiangsu, Xinjiang Uyghur
Autonomous Region, Ningxia Hui Autonomous Region,
Zhejiang, Guangxi Zhuang Autonomous Region,
Guangdong, Shanghai, Hubei, Hunan, and Shandong)
by probability proportional to size sampling (initially
including 15, to allow for non-responders) from different
geographical regions (south or north) in China
(appendix). We then selected one urban and one rural
district (defined according to the National Bureau of
Statistics in China) in each province. The local Centre
for Disease Control and Prevention for each province
identified three typical urban and three typical rural
districts. We then selected one of each for each province,
using computer-generated random numbers. We
selected three subdistricts (referred to as streets and
containing about 20 000–100 000 households in urban
districts, and referred to as townships and containing
about 5000–30 000 households in rural districts)
randomly (by simple randomisation using SAS software,
version 9.1) from each district. We then selected five
communities randomly (by simple randomisation using
SAS software) from each subdistrict. Finally, individuals
were randomly chosen (by simple randomisation using
SAS software) from each community. The ethics
committee of Peking University First Hospital approved
the study. All participants gave written informed consent
before data collection.
Screening protocol and assessment criteria
All on-site screenings were done between September,
2009, and September, 2010. Data were collected in
examination centres at local health stations or community
clinics in the participants’ residential area. All participants
completed a questionnaire documenting their socio-
demographic status (eg, age, sex, income, and education),
personal and family health history (eg, hypertension,
diabetes, and kidney disease), and lifestyle (eg, smoking
and alcohol consumption) with the assistance of medical
students, trained general practitioners, and nurses. History
of hepatitis, cardiovascular disease (myocardial infarction
or stroke), and nephrotoxic medications (non-steroidal
anti-inflammatory drugs and natural remedies and
preparations herbs containing aristolochic acid) were
noted. Anthropometric measurements (eg, weight and
height) were obtained. All study investigators and staff
members completed a training programme that taught the
methods and process of the study. A manual of procedures
was distributed, and detailed instructions for administration
of the questionnaires, taking of blood pressure and
anthropometric measurements, and biological specimen
collection and processing were provided.
Indicators of kidney damage and possible risk factors
were examined. All blood and urine samples were
analysed at the central laboratory in each province. All the
study laboratories successfully completed a standard-
isation and certification programme.
Urinary albumin and creatinine were measured
from a fresh morning spot urine sample or morning
urine sample stored at 4°C for less than 1 week.
Albuminuria was measured with immunoturbidimetric
tests. Urinary creatinine was measured with Jaffe’s kinetic
method. The urinary albumin to creatinine ratio (ACR;
mg/g creatinine) was calculated. Patients with an ACR
greater than 30 mg/g were defined as having albuminuria.
Albuminuria was defined as urinary albumin con-
centration of more than 20 mg/L.
To measure eGFR, blood was collected by venepuncture
after an overnight fast of at least 10 h. Serum creatinine
was measured by the same methods as was urinary
creatinine. eGFR was calculated with an equation
developed by adaptation of the Modification of Diet in
Renal Disease (MDRD) equation on the basis of data from
Chinese chronic kidney disease patients.10 Before the
study, the central laboratory in each province calibrated
creatinine measurements with samples at the laboratory
of Peking University First Hospital (Beijing, China),
where the modified equation was developed.10 To ensure
quality control creatinine was measured in 40 samples
with creatinine ranging from 48 μmol/L to 868 μmol/L,
both at the central laboratory in each province and at
the laboratory of Peking University First Hospital. From
13 linear regression models, representing 13 study centres
the range of the slopes, were 0·95 to 1·12 and the range of
the intercepts were –7·7 μmol/L to 3·9 μmol/L. Thus,
measurements from local laboratories were used directly
for calculation of eGFR. Reduced renal function was
defined as an eGFR of less than 60 mL/min per 1·73 m²:
eGFR = 175 × Scr−¹·²³⁴ × age–⁰·¹⁷⁹ [if female, × 0·79]
where Scr is serum creatinine concentration (in mg/dL)
and age in years.
Blood pressure was measured by sphygmomanometer,
three times at 5 min intervals. The mean of the three
readings was calculated, unless the difference between
the readings was greater than 10 mm Hg, in which case
the mean of the two closest measurements was used.
Hypertension was defined as a systolic blood pressure of
140 mm Hg or more, diastolic blood pressure of
90 mm Hg, any use of antihypertensive medication in
the past 2 weeks, or any self-reported history of
hypertension. Fasting blood glucose was measured
enzymatically with a glucose oxidase method. Diabetes
was defined as fasting plasma glucose of 7·0 mmol/L or
more, by hypoglycaemic agents despite fasting plasma
glucose, or any self-reported history of diabetes.
Serum total cholesterol, LDL cholesterol, HDL choles-
terol, triglycerides, and uric acid were measured with
www.thelancet.com Vol 379 March 3, 2012
 Articles

commercially available reagents. The laboratories used a
timed-endpoint colorimetric method to measure LDL
cholesterol and HDL cholesterol.
Statistical analysis
We estimated the prevalence of chronic kidney disease
according to age, sex, urban or rural residency, and
economic development. Assuming a prevalence of
chronic kidney disease of 13·0%,9 an α of 0·05, and a
relative error of sampling of 10%, we calculated a required
sample size of 50550 to allow for stratification by sex and
age, and a non-response rate of 10%.11 In addition to
crude prevalence, the prevalence estimates and com-
parisons were weighted to represent the total adult
population in China. Synthesised weights were calculated
from a sampling weight, a non-response weight, and a
population weight. These weights were used to adjust for
different selection probabilities, different response pro-
portions, and deviations in the sample compared with
the standard population, particularly for sex and age.
Data from the China Population Sampling Census in
2009, were used as the standard population.
Continuous data are presented as means with SDs
except for ACR, which is presented as median (IQR)
because of high skew. Categorical variables are presented
as proportions. Relevant characteristics are described and
stratified according to the presence of kidney damage (no
indicators of kidney damage [including no haematuria by
microscopic analysis of urine sediment], eGFR less than For more on measuring LDL
60 mL/min per 1·73 m², and albuminuria). Overlap was cholesterol and HDL cholesterol
see http://www.sekisuimedical.
less than 10% between indicators. jp/english/business/diagnostics/
The crude and adjusted prevalences of low eGFR, biochemistry/index.html
albuminuria, and total chronic kidney disease (defined as
low eGFR, albuminuria, or both) are reported in total, and
by sex and age (18–39 years, 40–59 years, 60–69 years,
≥70 years). Staging of chronic kidney disease was done on
the basis of the Kidney Disease Outcome Quality
Initiative,12 and stage 3 was further classified as stage 3A
and 3B with an eGFR 45 mL/min per 1·73 m² as the cutoff.
The adjusted prevalence of indicators of kidney damage
was then stratified by economic development (by tertiles
of gross domestic product per head) in rural and urban
areas. Furthermore, the adjusted prevalence of chronic
kidney disease was reported in different geographical
regions of China (east, south, middle, north, northwest,
and southwest). The prevalence and control rate of
hypertension and the prevalence and awareness of
diabetes are reported. Because results of postprandial
glucose are not available, we report the awareness of
diabetes instead of control of diabetes.
We analysed the association between indicators of
kidney damage and relevant covariates with logistic
regression models. We report the crude and multivariable
adjusted odds ratios (ORs) with 95% CIs. Covariates

Participants with no Participants with eGFR Participants with Total (n=47 204)
indicators of kidney damage <60 mL/min per 1·73 m² albuminuria (n=3517)
(n=41 165) (n=1185)
Age (years) 48·8 (15·0) 63·6 (14·7) 55·8 (15·3) 49·6 (15·2)
Men 18 059 (43·9%) 451 (38·1%) 1361 (38·7%) 20 148 (42·7%)
Rural residents 18 743 (45·5%) 615 (51·9%) 1597 (45·4%) 21 859 (46·3%)
Educated to high school or above 19 022 (46·2%) 340 (28·7%) 984 (28·0%) 20 950 (44·5%)
Have health insurance 34 354 (83·5%) 1037 (87·5%) 3056 (86·9%) 39 643 (84·0%)
Current smoker 9857 (24·0%) 230 (19·4%) 770 (21·9%) 11 094 (23·5%)
Self-reported HBV infection 1267 (3·1%) 32 (2·7%) 77 (2·2%) 1399 (3·0%)
Habitual drinker 3100 (7·5%) 83 (7·0%) 281 (8·0%) 3553 (7·5%)
Non-habitual drinker 7136 (17·3%) 128 (10·8%) 417 (12·0%) 7876 (16·7%)
Nephrotoxic medication used 1272 (3·1%) 79 (6·7%) 178 (5·1%) 1536 (3·3%)
History of CVD 948 (2·3%) 113 (9·5%) 179 (5·1%) 1220 (2·6%)
Hypertension 13 533 (32·9%) 717 (60·5%) 2152 (61·2%) 16 691 (35·4%)
Diabetes 2659 (6·5%) 226 (19·1%) 607 (17·3%) 3488 (7·4%)
Body-mass index (kg/m²) 23·8 (3·7) 24·4 (3·8) 24·7 (4·0) 23·9 (3·7)
Uric acid (μmol/L) 298·3 (91·0) 387·0 (129·5) 289·6 (96·2) 298·6 (93·1)
Triglyceride (mmol/L) 1·4 (1·5) 1·8 (1·4) 1·7 (1·8) 1·5 (1·5)
LDL cholesterol (mmol/L) 2·9 (0·9) 3·2(1·2) 3·0 (1·0) 2·9 (1·0)
HDL cholesterol (mmol/L) 1·4 (0·4) 1·4 (0·4) 1·4 (0·4) 1·4 (0·4)
Creatinine (μmol/L) 73·4 (15·5) 129·6 (66·3) 78·6 (32·4) 74·8 (20·6)
eGFR (mL/min per 1·73m²) 102·8 (26·5) 50·4 (10·9) 95·8 (29·4) 101·2 (27·4)
ACR (mg/g creatinine; median [IQR]) 6·0 (2·8–11·0) 10·5 (3·2–36·0) 57·2 (38·1–120·3) 6·6 (3·1–13·6)

Data are n (%) or mean (SD), unless stated otherwise. ACR presented as median because of high skew. Overlap between indicators of kidney damage was less than 10%.
eGFR=estimated glomerular filtration rate. HBV=hepatitis B virus. CVD=cardiovascular disease. ACR=albumin:creatinine ratio.

Table 1: Characteristics of participants according to indicators of kidney damage

www.thelancet.com Vol 379 March 3, 2012 817

 Articles

included in the multivariable logistic regression economic development and kidney damage varied by
models were age (change by 10 years), sex, history of urban or rural residency. Appropriate interaction terms
cardiovascular disease (yes vs no), hypertension (yes vs (tertiles of gross domestic product per head and urban or
no), diabetes (yes vs no), rural versus urban residents, rural residency)were generated to test whether inter-
education (≥high school vs <high school), current smoker actions were statistically significant. We did stratified
(yes vs no), alcohol intake (habitual drinker [drink once multivariable analyses if the interaction was significant.
per day or more] vs non-habitual drinker [six times per To examine a possible non-linear relation between
week to once per month] vs non-drinker [almost never]), albuminuria and covariates including systolic blood
self-reported hepatitis B virus infection (yes vs no), pressure, body-mass index, triglyceride concentration,
nephrotoxic medications (somedon, APC [both analgesic and LDL cholesterol concentration, we used restricted
mixtures containing phenacetin], ibuprofen, and herbal cubic spline regression.13 The four knots used in the
pills containing aristolochic acid [Long Dan Xie Gan, analyses were 100 mm Hg, 120 mm Hg, 140 mm Hg, and
Guan Xin Su He, and Pai Shi Ke Li]) used (yes vs no), 160 mm Hg for systolic blood pressure; 18·5 kg/m²,
body-mass index (<18·5 kg/m² vs 18·5–23·9 kg/m² 24·0 kg/m², 28·0 kg/m², and 30·0 kg/m² for body-mass
[reference] vs 24·0–27·9 kg/m², vs ≥28·0 kg/m²), index; 0·01 mmol/L, 0·02 mmol/L, 0·03 mmol/L, and
hyperuricaemia (>422 μmol/L for men, >363μmol/L for 0·04 mmol/L for triglycerides; and 0·07 mmol/L,
women), plasma triglyceride (a continuous variable), 0·09 mmol/L, 0·11 mmol/L, and 0·12 mmol/L for LDL
plasma LDL cholesterol (a continuous variable), and cholesterol. We used the likelihood ratio test to test for
plasma HDL cholesterol (a continuous variable). We non-linearity.14
also investigated whether the association between the We used Epidata software (version 3.1) for data
entry and management. All p values are two-sided, and a
p value of less than 0·05 was considered significant.
Kidney function Albuminuria CKD prevalence Analyses were done with SUDAAN (version 10) and SAS
(95% CI)
(version 9.1).
eGFR (mL/min n Prevalence n Prevalence
per 1·73 m²) (95% CI) (95% CI)
Role of the funding source
1 >90 29 244 65·2 (64·4–66·1) 1877 8·7 (8·0–9·3) 5·7 (5·2–6·1) The sponsor of the study had no role in study design,
2 60–89 16 775 33·0 (32·2–33·9) 1385 10·3 (9·3–11·2) 3·4 (3·1–3·7) data collection, data analysis, data interpretation, or
3 30–59 1106 1·6 (1·4–1·8) 221 21·1 (16·1–26·1) 1·6 (1·4–1·8) writing of the report. The corresponding author had full
3a 45–59 940 1·4 (1·2–1·5) 165 19·5 (14·2–24·98) 1·4 (1·2–1·5) access to all the data in the study and had final
3b 30–44 166 0·2 (0·1–0·3) 56 31·3 (16·2–46·4) 0·2 (0·1–0·3) responsibility for the decision to submit for publication.
4 15–29 59 0·1 (0·06–0·2) 25 34·3 (9·6–58·9) 0·1 (0·06–0·2)
5 <15 20 0·03 (0·01–0·05) 9 56·6 (22·6–90·5) 0·03 (0·01–0·05) Results
Total ·· 47 204 100 3517 9·4 (8·9–10·0) 10·8 (10·2–11·3) 50 550 people were invited to participate and
47 204 completed the survey and examination. The
Albuminuria was defined as a urinary albumin:creatinine ratio >30 mg/g creatinine. CKD was defined as eGFR
<60 mL/min per 1·73m² or albuminuria. All prevalences are adjusted for synthesised weights. eGFR=estimated response rate was 93% (93% for both men and women,
glomerular filtration rate. CKD=chronic kidney disease. 99% in the rural areas and 89% in the urban areas).
Urinary creatinine was not measured in 1501 (3·2%)
Table 2: Prevalence of indicators of kidney function, by disease stage
participants because of insufficient sample collection.

eGFR <60 mL/min Albuminuria CKD Hypertension Control of Diabetes Awareness of

per 1·73 m² hypertension diabetes
Rural
GDP per head, tertile 1 (0·2–2·3) 1·8% (1·3–2·2) 5·7% (4·9–6·4) 7·2% (6·4–8·0) 29·9% (28·3–31·6) 15·5% (13·1–17·9) 3·7% (3·1–4·4) 22·8% (15·8–29·7)
GDP per head, tertile 2 (2·4–3·6) 1·6% (1·3–2·0) 9·5% (8·4–10·6) 10·6% (9·5–11·7) 31·5% (29·9–33·1) 5·2% (3·9–6·5) 3·8% (3·2–4·4) 35·9% (28·3–43·5)
GDP per head, tertile 3 (2·4–7·5) 1·4% (1·0–1·7) 14·8% (13·4–16·3) 15·8% (14·3–17·3) 27·7% (25·9–29·4) 8·9% (7·0–10·8) 5·5% (4·6–6·3) 28·3% (21·1–35·5)
Total 1·6% (1·4–1·8) 10·1% (9·4–10·8) 11·3% (10·6–12·0) 29·9% (28·9–30·8) 9·1% (8·1–10·2) 4·3% (3·9–4·7) 29·7% (25·3–34·0)
Urban
GDP per head, tertile 1 (1·1–4·0) 2·7% (2·1–3·3) 7·6% (6·7–8·5) 10·0% (9·0–11·0) 27·3% (25·8–28·8) 18·0% (15·7–20·4) 6·7% (5·9–7·6) 68·7% (62·7–74·7)
GDP per head, tertile 2 (4·6–7·9) 1·9% (1·5–2·3) 7·0% (6·2–7·9) 8·2% (7·3–9·1) 28·4% (26·9–29·9) 18·1% (15·7–20·5) 7·4% (6·6–8·3) 64·5% (58·8–70·2)
GDP per head, tertile 3 (8·8–22·4) 1·3% (1·0–1·5) 5·3% (4·6–5·9) 6·3% (5·6–7·0) 43·7% (42·1–45·3) 29·3% (27·3–31·2) 7·5% (6·8–8·3) 68·4% (63·8–73·0)
Total 2·3% (1·9–2·6) 7·0% (6·4–7·5) 8·9% (8·2–9·6) 31·7% (30·6–32·8) 22·1% (20·6–23·6) 7·0% (6·5–7·6) 68·2% (64·2–72·2)

Data are adjusted prevalence (%; 95% CI). Albuminuria was defined as a urinary albumin to creatinine ratio >30 mg/g creatinine. CKD was defined as eGFR <60 mL/min per 1·73 m² or albuminuria. GDP per head
was based on data from the subdistrict, which is why the tertiles do not cover all values. eGFR=estimated glomerular filtration rate. CKD=chronic kidney disease. GDP=gross domestic product (per CN¥10 000).

Table 3: Adjusted prevalence of indicators of kidney damage by economic development

818 www.thelancet.com Vol 379 March 3, 2012

 Articles

Participants with low eGFR or albuminuria were older,
less educated, more likely to be women, and had higher
prevalences of cardiovascular disease, hypertension, and
diabetes, than did participants without indicators of
kidney damage (table 1).
The adjusted prevalence of eGFR less than 60 mL/min
per 1·73 m² was 1·7% (95% CI 1·5–1·9) and that of
albuminuria was 9·4% (8·9–10·0). The overall adjusted
prevalence of chronic kidney disease was 10·8%
(10·2–11·3). Table 2 shows the prevalences of indicators
of kidney damage at different disease stages. The
appendix shows these measures stratified by age and sex.
The prevalence of indicators of kidney damage increased
kidney disease was 12·5% (95% CI 10·6–14·4), and did
not vary significantly between rural and urban areas
(12·9% [95% CI 10·7–15·2] vs 10·5% [8·2–11·9]).
Age, being a woman, history of cardiovascular disease,
diabetes, and hyperuricaemia were all independently
associated with having an eGFR of more than 60 mL/min
per 1·73 m² (table 5). Factors independently associated
with albuminuria included increased age, being a
woman, hypertension, and diabetes. Use of nephrotoxic
eGFR <60 mL/min Albuminuria CKD
per 1·73 m²

with age. Women had a higher prevalence of chronic East 1·1% (0·9–1·3) 7·5% (7·0–8·1) 8·4% (7·8–9·0)
kidney disease (12·9%, 12·0–13·7) than did men (8·7%, South 1·3% (1·0–1·6) 6·0% (5·2–6·7) 6·7% (6·0–7·5)
95% CI 8·0–9·5). Middle 1·4% (1·0–1·8) 13·1% (11·6–14·5) 14·2% (12·8–15·7)
Fewer participants in rural areas had eGFR less than North 2·5% (1·9–3·2) 15·4% (13·6–17·1) 16·9% (15·1–18·7)
60 mL/min per 1·73 m², and more had albuminuria than Northwest 1·5% (1·0–2·1) 5·6% (4·5–6·7) 6·7% (5·6–8·9)
did those in urban areas (table 3). In rural areas, the Southwest 3·8% (2·9–4·6) 15·1% (13·4–16·9) 18·3% (16·4–20·1)
prevalence of albuminuria was higher in high tertiles of
Data are adjusted prevalence (%, 95% CI). Albuminuria was defined as a urinary
GDP per head than in low tertiles, and the prevalence of albumin to creatinine ratio >30 mg/g creatinine. CKD was defined as eGFR
For more on GDP per head see
eGFR less than 60 mL/min per 1·73 m² did not vary by http://www.stats.gov.cn/tjsj/
<60 mL/min per 1·73 m² or albuminuria. eGFR=estimated glomerular filtration
economic development. In urban areas, the prevalences rate. CKD=chronic kidney disease.
of eGFR less than 60 mL/min per 1·73 m² and albuminuria Table 4: Prevalence of chronic kidney disease stratified by
were low in sites with high economic development. High- geographical region
school or higher education was low in participants in
rural areas compared with those in urban areas (24·9%,
95% CI 24·4–25·5%, vs 61·3%, 60·7–61·9%; p<0·0001), eGFR <60 mL/min Albuminuria
and this trend also existed in the top tertile of GDP per per 1·73 m²
head (27·6%, 26·6–28·6%, vs 63·6%, 62·6–64·6%; Age change by 10 years 1·74 (1·59–1·91) 1·08 (1·02–1·15)
p<0·0001). In rural areas, control of hypertension and Sex (women vs men) 1·66 (1·17–2·37) 1·42 (1·17–1·71)
awareness of diabetes were not positively associated with Nephrotoxic drug use 1·64 (0·94–2·85) 1·31 (0·98–1·75)
economic development, and were significantly lower than History of cardiovascular disease 2·51 (1·24–5·10) 1·40 (0·91–2·14)
those in urban areas (table 3). Hypertension 1·08 (0·80–1·46) 2·71 (2·31–3·18)
The prevalence of chronic kidney disease varied greatly Diabetes 2·00 (1·35–2·97) 1·99 (1·60–2·48)
between geographical regions (table 4). Compared with Hyperuricaemia 9·30 (6·80–12·72) 0·88 (0·68–1·13)
other geographical regions, the prevalence of reduced renal Urban residents (vs rural residents) 1·12 (0·86–1·48) 0·97 (0·81–1·16)
function was high in north (2·5%, 95% CI 1·9–3·2) and Gross domestic product per head
southwest (3·8%, 2·9–4·6) China. Likewise, the preva- Rural residents
lence of albuminuria was higher in the north (15·4%, Tertile 1 (reference) 1·00 1·00
13·6–17·1) and southwest (15·1%, 13·4–16·9) than in other Tertile 2 1·93 (1·10–3·38) 1·11 (0·92–1·35)
regions. In the north, the prevalence of hypertension was
Tertile 3 0·79 (0·46–1·35) 1·74 (1·45–2·10)
similar to that of the total population (27·9%, 25·9–30·0%,
Urban residents
vs 30·2%, 29·5–31·0%), but control of hypertension
Tertile 1 (reference) 1·00 1·00
was substantially lower (6·2%, 4·3–8·1%, vs 12·0%,
Tertile 2 0·36 (0·23–0·56) 1·19 (0·95–1·48)
11·1–12·9%). For the southwest, the prevalence of hyper-
Tertile 3 0·16 (0·09–0·27) 1·05 (0·77–1·44)
tension (11·1%, 9·8–12·5%) was low and control of hyper-
tension (23·1%, 17·8–28·5%) was higher than that of the Data are multivariable-adjusted odds ratio (95% CI). Hypertension was defined as
total population. The prevalence of diabetes was lower in a systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg, the
use of antihypertensive drugs in the past 2 weeks irrespective of blood pressure,
the north (3·6%, 2·8–4·4%) and the southwest (3·8%, or by any self-reported history of hypertension. Hyperuricaemia was defined by
3·0–4·6%) than in the total population (4·9%, 4·5–5·2%). plasma uric acid concentration >422 μmol/L for men and >363 μmol/L for
The adjusted odds ratio (OR) of eGFR less than women. The range for tertiles 1–3 for GDP per head in rural areas was 0·2–1·2,
1·6–2·3, and 2·4–7·5 (per CN¥10 000). The range for tertiles 1–3 for GDP per head
60 mL/min per 1·73 m² was significant for the north
in urban areas were 1·1–4·0, 4·6–7·9, and 8·8–22·4 (per ¥10 000). GDP per head
(4·97, 95% CI 3·28–7·54) and for the southwest was based on data from the subdistrict, which is why the tertiles do not cover
(3·44, 2·23–5·31). The OR of albuminuria was signif- all values. eGFR=estimated glomerular filtration rate.
icantly different for the north (2·83, 2·36–3·38) and
Table 5: Factors associated with indicators of kidney damage
southwest (3·41, 2·77–4·20). Awareness of chronic

www.thelancet.com Vol 379 March 3, 2012 819

 Articles
medication was not significantly associated with an
increased risk of albuminuria (p=0·0602). Interactions
between urban or rural area, and economic development
were significant (p<0·0001) for both eGFR less than
60 ml/min/1·73 m² and albuminuria (data not shown),
therefore stratified analyses were done. After adjusting
for potential confounders, economic development was
positively associated with presence of albuminuria for
rural participants for tertile 3 of GDP per head.
The spline regression models showed a deviation from
linearity for the relation between systolic blood pressure
and albuminuria (for non-linearity p<0·0001; appendix).
The 95% CIs for ORs of body-mass index, triglyceride
concentration, and LDL cholesterol concentration usually
spanned 1 (appendix).
Discussion
In this representative sample of Chinese adults, the
prevalence of chronic kidney disease was 10·8%.
119·5 million adults aged 18 years or older have chronic
kidney disease as defined in this study, and only 12·5%
of them are aware of the condition. A major strength of
our study is that it is a nationwide survey in a developing
country with large economic heterogeneity and therefore
provides a unique dataset with which to explore the effect
of multiple demographic variables on the burden and
trends of chronic kidney disease. Furthermore, the
training programmes and the quality-control procedures
used ensure the credibility of the results.
Several studies have documented the high prevalence
of chronic kidney disease in developing countries
Panel: Research in context
Systematic review
We searched PubMed for cross-sectional surveys of chronic
kidney disease in China with the terms “renal insufficiency,
chronic“ and “China/epidemiology“. We included reports
published in English or Chinese between Jan 1, 2003, and
Nov 30, 2011. Only one national survey reported the
prevalence of reduced renal function in people aged
35–74 years, but proteinuria was not included in that study.
We identified four regional surveys of chronic kidney disease
(incorporating both reduced renal function and
albuminuria) in China. Three were from large cities and the
other from Tibet.
Interpretation
Our cross-sectional study assessed the prevalence of chronic
kidney disease in a representative sample of adults in China.
Chronic kidney disease has become an important public
health issue in China, possibly as a consequence of increased
diabetes and hypertension. Economically improving rural
areas should be closely monitored. Our findings add to the
current understanding of integrated prevention and
management of non-communicable diseases worldwide,
especially in developing countries.
820
including China (panel).4 However, most of these studies
were regional,15,16 or were focused on specific occu-
pations.17,18 To our knowledge, no national survey of
chronic kidney disease including both reduced eGFR and
albuminuria has been done in a developing country. We
used a standard design for a population survey and a
strict quality control procedure to ensure the represent-
ativeness of our study. The prevalence of chronic kidney
disease estimated in our study was similar to that in
developed countries such as the USA (13·0%) and
Norway (10·2%).19,20 We used the same cutoff to define
low eGFR as used in the US and Norwegan surveys, but
our definition of albuminuria was slightly different,
because it did not use sex-specific criteria. However, the
prevalences of stage 3 and stage 4 chronic kidney disease
in our study was low compared with those in developed
countries. For example, the prevalence of stage 3 chronic
kidney disease was 1·6% in our study, compared with
7·7% in the USA and 4·2% in Norway. One explanation
might be that hypertension and diabetes have increased
rapidly in the past 15–20 years in China,21 but for these
diseases to affect chronic kidney disease at a population
level might take another 10 years.21 This lag provides a
unique opportunity to slow the increase of chronic kidney
disease prevalence caused by the changing epidemiology
of hypertension and diabetes in China.
Unlike in developed countries, socioeconomic status
and the accompanying change in lifestyle has varied
substantially between rural and urban areas in China,
especially during the past decade. Several possible
explanations for the discrepancy between urban and
rural areas exist. First, the same pattern has been
reported for other non-communicable diseases including
diabetes, which is a risk factor for chronic kidney
disease. For example, in a national survey,22 the
prevalence of diabetes increased substantially with
economic development in rural areas (from 5·8% in
underdeveloped rural areas to 12·0% in developed rural
areas), whereas the prevalence of diabetes was similar
across economic development categories in urban areas
(10·4% in underdeveloped urban areas, 12·0% in
developed urban areas). Second, the difference between
rural and urban areas might be caused by the disparity
in health care between areas in China. A study23 using
data from the China National Nutrition and Health
Survey 2002 indicated that the treatment and the control
of hypertension were substantially lower in rural areas
than in urban areas, which is consistent with our results.
Finally, in our analysis, the increased OR of albuminuria
in economically developed rural areas still existed after
adjustment for status of diabetes and hypertension. We
suggest that factors not captured by our study might
contribute to the increased risk—eg, unhealthy lifestyle
caused by both rapidly increased economic development
and reduced health literacy (partly manifested by
education level). Our study shows that residents in
economically improving rural areas are a unique
www.thelancet.com Vol 379 March 3, 2012

population in terms of the burden of chronic kidney
disease in China.
In the analysis of different geographical regions, the
north and southwest were associated with high prevalences
of low eGFR and albuminuria. These associations were
still statistically significant after adjustment for multiple
confounders; two possible explanations exist. First,
unmeasured confounders might contribute to the positive
association between specific geographical regions and the
presence of chronic kidney disease. For example, the
results from the Chinese National Health and Nutrition
Examination Survey24 show that residents in the north and
southwest regions consumed a high percentage of red
meat, which is characteristic of a western diet. Furthermore,
diets in the north of China are typically high in sodium
and potassium.25 A typical western diet and high sodium
intake are significantly associated with increased micro-
albuminuria and rapid decrease in kidney function.26
Besides dietary factors, low birthweight, which is thought
to be associated with subsequent risk of chronic kidney
disease,27 might also contribute to the variation in chronic
kidney disease between regions. In a national survey in
China,28 the prevalence of low birthweight (<2·5 kg) in
the southwest was 5·0%, which was higher than that in
the east (3·4%) and south (4·3%). Genetic differences
might also contribute to the different risk of chronic
kidney disease between regions. Results of a genetic
analysis29 show a clear distinction between southern and
northern Chinese populations; and the genetic charac-
teristics of Han population in the southwest are reported
to be intermediate between northern and southern
Chinese populations.30 These environmental and genetic
factors were not fully captured in our study and deserve
further investigation.
Second, residual confounding might exist from the
covariates included in our multivariable logistic models.
For example, isolated impaired glucose is tolerated more
than is isolated impaired fasting glucose in Chinese
populations.22 However, in our study, diabetes was
defined by fasting blood glucose and self-reported history,
instead of by oral glucose-tolerance test.
Our multivariable analysis indicated that factors
associated with kidney damage include hypertension,
diabetes, and cardiovascular disease, which are all major
non-communicable diseases. These results are consistent
with previous studies from China.6,8,9 The spline regression
analyses suggests that above a systolic blood pressure of
120 mm Hg, the ORs for albuminuria were statistically
significant and increased gradually. However, the cross-
sectional design limited our ability to make a causal
inference. In view of the shared risk factors and causes of
chronic kidney disease and other non-communicable
diseases, incorporation of detection and management of
chronic kidney disease in the sustained global strategy is
crucial to combat the major non-communicable diseases.31
Our study had several limitations. First, the laboratory
tests were done at central laboratories of 13 study
www.thelancet.com Vol 379 March 3, 2012
Articles
provinces. Therefore, the possibility of variation between
laboratories exists. However, all participating laboratories
completed a standardisation and certification programme
before the study, and stringent quality-control procedures
were used during the study. Second, all the indicators of
chronic kidney disease were obtained on the basis of
single measurements; therefore the reported prevalence
of chronic kidney disease might be overestimated. Third,
the definition of hypertension and diabetes was partly
based on self-reported history and oral glucose-tolerance
test was not done to confirm the diagnosis of diabetes.
Finally, the cross-sectional design of the study makes
inference of a causal relationship between indicators of
kidney damage and associated factors impossible.
Our results show that chronic kidney disease has
become an important public health problem in China,
and special attention should be paid to residents in
economically improving rural areas. The rapid increase in
diabetes and hypertension, both of which are predicted to
drive epidemics of chronic kidney disease, will have
profound socioeconomic and public health consequences
in developing countries such as China. In view of the large
effect chronic kidney disease has on outcomes in diabetes,
hypertension, and cardiovascular disease, chronic kidney
disease should be an important consideration in any
strategy to address non-communicable diseases in
developing countries. Furthermore, the reasons for, and
potential interventions to address, the high-risk of chronic
kidney disease in the north and southwest of China
deserve further investigation.
Contributors
LZ and FW searched the literature, conceived the study, designed the
study, analysed the data, interpreted the results, and drafted the report. LW,
WenW, BL, JL, MC, QH, YL, XY, NC, J-eZ, ZH, FL, DH, LM, HL, XZ, JC,
LP, WC, and WeiW collected the data and revised the report. XL obtained
funding. HW conceived the study, organised and supervised the study,
interpreted the results, and revised the report.
Conflicts of interest
We declare that we have no conflicts of interest.
Acknowledgments
This study was supported by the National Key Technology R&D Program
from the Ministry of Science and Technology (China; 2007BAI04B10);
the Science and Technology Commission of Shanghai (08dz1900502 and
07JC14037); the Natural Science Funds of Ning Xia (NZ08102); the
Science and Technology Department, the National Natural Science
Funds (30660069); the National Natural Science Foundation of China;
the Department of Health, Jiangsu Province (H200936); the Key
Scientific and Technology Project from the Sichuan Science and
Technology Department (05SG1635); the Program for New Century
Excellent Talents in University from the Ministry of Education (China;
BMU2009131); the International Society of Nephrology Research
Committee; and the China Health and Medical Development
Foundation. We thank Yong Jiang of the Chinese Center for Disease
Control and Prevention for his contribution to the design of the study
and statistical support, and William G Couser for constructive
comments for the report.
References
1 Bello AK, Nwankwo E, El Nahas AM. Prevention of chronic kidney
disease: a global challenge. Kidney Int Suppl 2005; 98: S11–17.
2 Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic
kidney disease and the risks of death, cardiovascular events, and
hospitalization. N Engl J Med 2004; 351: 1296–305.
821
 Articles
3 Matsushita K, van der Velde M, Astor BC, et al. Association of
estimated glomerular filtration rate and albuminuria with all-cause
and cardiovascular mortality in general population cohorts:
a collaborative meta-analysis. Lancet 2010; 375: 2073–81.
4 Nugent RA, Fathima SF, Feigl AB, Chyung D. The burden of
chronic kidney disease on developing nations: a 21st century
challenge in global health. Nephron Clin Pract 2011; 118: c269–77.
5 Eknoyan G, Lameire N, Barsoum R, et al. The burden of kidney
disease: improving global outcomes. Kidney Int 2004; 66: 1310–14.
6 Chen N, Wang W, Huang Y, et al. Community-based study on CKD
subjects and the associated risk factors. Nephrol Dial Transplant
2009; 24: 2117–23.
7 Chen W, Liu Q, Wang H, et al. Prevalence and risk factors of
chronic kidney disease: a population study in the Tibetan
population. Nephrol Dial Transplant 2011; 26: 1592–99.
8 Chen W, Wang H, Dong X, et al. Prevalence and risk factors
associated with chronic kidney disease in an adult population from
southern China. Nephrol Dial Transplant 2009; 24: 1205–12.
9 Zhang L, Zhang P, Wang F, et al. Prevalence and factors associated
with CKD: a population study from Beijing. Am J Kidney Dis 2008;
51: 373–84.
10 Ma YC, Zuo L, Chen JH, et al. Modified glomerular filtration rate
estimating equation for Chinese patients with chronic kidney
disease. J Am Soc Nephrol 2006; 17: 2937–44.
11 Plan and operation of the Third National Health and Nutrition
Examination Survey, 1988–94. Series 1: programs and collection
procedures. Vital Health Stat 1 1994; 32: 1–407.
12 National Kidney Foundation. K/DOQI clinical practice guidelines
for chronic kidney disease: evaluation, classification, and
stratification. Am J Kidney Dis 2002; 39 (suppl 1): S1–266.
13 Greenland S. Dose-response and trend analysis in epidemiology:
alternatives to categorical analysis. Epidemiology 1995; 6: 356–65.
14 Greenland S, Rothman K. Fundamentals of epidemiologic data
analysis. In: Rothman KJ, Greenland S, Lash TL, eds.
Modern Epidemiology, 2nd edn. Philadelphia: Lippincott Williams
& Wilkins, 1998: 201–29.
15 Agarwal SK, Dash SC, Irshad M, Raju S, Singh R, Pandey RM.
Prevalence of chronic renal failure in adults in Delhi, India.
Nephrol Dial Transplant 2005; 20: 1638–42.
16 O’Donnell JK, Tobey M, Weiner DE, et al. Prevalence of and risk
factors for chronic kidney disease in rural Nicaragua.
Nephrol Dial Transplant 2010; 26: 2798–805.
822
17 Varma PP, Raman DK, Ramakrishnan TS, Singh P, Varma A.
Prevalence of early stages of chronic kidney disease in apparently
healthy central government employees in India.
Nephrol Dial Transplant 2010; 25: 3011–17.
18 Ito J, Dung DT, Vuong MT, et al. Impact and perspective on chronic
kidney disease in an Asian developing country: a large-scale survey
in north Vietnam. Nephron Clin Pract 2008; 109: c25–32.
19 Hallan SI, Coresh J, Astor BC, et al. International comparison of
the relationship of chronic kidney disease prevalence and ESRD
risk. J Am Soc Nephrol 2006; 17: 2275–84.
20 Coresh J, Selvin E, Stevens LA, et al. Prevalence of chronic kidney
disease in the United States. JAMA 2007; 298: 2038–47.
21 Zhang L, Wang H. Chronic kidney disease epidemic: cost and
health care implications in China. Semin Nephrol 2009; 29: 483–86.
22 Yang W, Lu J, Weng J, et al. Prevalence of diabetes among men and
women in China. N Engl J Med 2010; 362: 1090–101.
23 Wu Y, Huxley R, Li L, et al. Prevalence, awareness, treatment, and
control of hypertension in China: data from the China National
Nutrition and Health Survey 2002. Circulation 2008; 118: 2679–86.
24 Jin S. The Tenth Report of the Chinese National Health and
Nutrition Examination Survey—Nutrition and Health Status, 1st ed.
Beijing: People’s Medical Publishing House, 2008.
25 Liu LS. Epidemiology of hypertension and cardiovascular
disease—China experience. Clin Exp Hypertens A 1990; 12: 831–44.
26 Lin J, Fung TT, Hu FB, Curhan GC. Association of dietary patterns
with albuminuria and kidney function decline in older white
women: a subgroup analysis from the Nurses’ Health Study.
Am J Kidney Dis 2011; 57: 245–54.
27 White SL, Perkovic V, Cass A, et al. Is low birth weight an
antecedent of CKD in later life? A systematic review of
observational studies. Am J Kidney Dis 2009; 54: 248–61.
28 Lin L, Liu Y, Zhang X, Mi J, Cao L. Sampling survey on low-birth
weight in China in 1998. Zhonghua Yu Fang Yi Xue Za Zhi 2002;
36: 5 [in Chinese].
29 Chu JY, Huang W, Kuang SQ, et al. Genetic relationship of
populations in China. Proc Natl Acad Sci USA 1998; 95: 11763–68.
30 Shi L, Xu SB, Ohashi J, et al. HLA-A, HLA-B, and HLA-DRB1 alleles
and haplotypes in Naxi and Han populations in southwestern China
(Yunnan province). Tissue Antigens 2006; 67: 38–44.
31 Beaglehole R, Bonita R, Horton R, et al. Priority actions for the
non-communicable disease crisis. Lancet 2011; 377: 1438–47.
www.thelancet.com Vol 379 March 3, 2012