Articles

Acute kidney injury in China: a cross-sectional survey

Li Yang, Guolan Xing, Li Wang, Yonggui Wu, Suhua Li, Gang Xu, Qiang He, Jianghua Chen, Menghua Chen, Xiaohua Liu, Zaizhi Zhu, Lin Yang*,
Xiyan Lian, Feng Ding, Yun Li, Huamin Wang, Jianqin Wang, Rong Wang, Changlin Mei, Jixian Xu, Rongshan Li, Juan Cao, Liang Zhang, Yan Wang,
Jinhua Xu, Beiyan Bao, Bicheng Liu, Hongyu Chen, Shaomei Li, Yan Zha, Qiong Luo, Dongcheng Chen, Yulan Shen, Yunhua Liao,
Zhengrong Zhang, Xianqiu Wang, Kun Zhang, Luojin Liu, Peiju Mao, Chunxiang Guo, Jiangang Li, Zhenfu Wang, Shoujun Bai, Shuangjie Shi,
Yafang Wang, Jinwei Wang, Zhangsuo Liu, Fang Wang, Dandan Huang, Shun Wang, Shuwang Ge, Quanquan Shen, Ping Zhang, Lihua Wu,
Miao Pan, Xiting Zou, Ping Zhu, Jintao Zhao, Minjie Zhou, Lin Yang, Wenping Hu, Jing Wang, Bing Liu, Tong Zhang, Jianxin Han, Tao Wen,
Minghui Zhao, Haiyan Wang, ISN AKF 0by25 China Consortiums

Summary
Background Acute kidney injury (AKI) has become a worldwide public health problem, but little information is
available about the disease burden in China. We aimed to evaluate the burden of AKI and assess the availability of
diagnosis and treatment in China.
Methods We launched a nationwide, cross-sectional survey of adult patients who were admitted to hospital in 2013
in academic or local hospitals from 22 provinces in mainland China. Patients with suspected AKI were screened
out on the basis of changes in serum creatinine by the Laboratory Information System, and we reviewed medical
records for 2 months (January and July) to conrm diagnoses. We assessed rates of AKI according to
two identication criteria: the 2012 Kidney Disease: Improving Global Outcomes (KDIGO) AKI denition and an
increase or decrease in serum creatinine by 50% during hospital stay (expanded criteria). We estimated national
rates with data from the 2013 report by the Chinese National Health and Family Planning Commission and
National Bureau of Statistics.
Findings Of 2 223 230 patients admitted to the 44 hospitals screened in 2013, 154 950 (70%) were suspected of
having AKI by electronic screening, of whom 26 086 patients (from 374 286 total admissions) were reviewed with
medical records to conrm the diagnosis of AKI. The detection rate of AKI was 099% (3687 of 374 286) by KDIGO
criteria and 203% (7604 of 374 286) by expanded criteria, from which we estimate that 1429 million people with
AKI were admitted to hospital in China in 2013. The non-recognition rate of AKI was 742% (5608 of 7555 with
available data). Renal referral was done in 214% (1625 of 7604) of the AKI cases, and renal replacement therapy was
done in 593% (531 of 896) of those who had the indications. Delayed AKI recognition was an independent risk
factor for in-hospital mortality, and renal referral was an independent protective factor for AKI under-recognition
and mortality
Interpretation AKI has become a huge medical burden in China, with substantial underdiagnosis and undertreatment.
Nephrologists should take the responsibility for leading the battle against AKI.
Funding National 985 Project of China, National Natural Science Foundation of China, Beijing Training Program for
Talents, International Society of Nephrology Research Committee, and Bethune Fund Management Committee.

Introduction
Acute kidney injury (AKI) is a common disorder with a
high risk of mortality and development of chronic kidney
disease, although information from developing countries
is scarce.13 In 2013, the International Society of
Nephrology launched a global target of 0by25no
patient deaths due to untreated acute kidney failure by
2025to improve the diagnosis and treatment of AKI
globally.4 An important step to meet this target is to
measure and identify the burden and present situation of
AKI worldwide, for which data from China, the worlds
largest developing country, will provide valuable
information. Up to now, data about the prevalence of AKI
in China have been sparse. Single-centre studies have
reported that AKI complicates 241319% of all hospital
admissions,5,6 which is much lower than what is reported
by developed countries (718%).711

We designed the China national survey of AKI in adults

treated in hospital in 2013 to provide reliable data for
estimation of the burden of AKI, depict its characteristics,
and show how to recognise and treat AKI in the real
world of clinical practice in China.

Methods
Study design and participants
We did a cross-sectional survey in mainland China. We
included 22 of the 31 provinces, municipalities, and
autonomous regions in China in our survey, covering
82% of the countrys population and the four
geographical regions of China (north, northwest,
southeast, and southwest; appendix). In each region we
enrolled an academic hospital in the regions capital
city and a local hospital from a smaller city or rural
county. The appendix contains details on hospital

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

Lancet 2015; 386: 146571

See Editorial page 1419
See Comment page 1425
Peking University First
Hospital, Beijing, China
(Prof Li Yang MD,
Yafang Wang MD,
Jinwei Wang PhD,
Prof M Zhao MD,
Prof Haiyan Wang MD); The First
Aliated Hospital of
Zhengzhou University,
Zhengzhou, China
(Prof G Xing MD, Prof Z Liu MD);
Sichuan Provincial Peoples
Hospital, Chengdu, China
(Prof L Wang MD, F Wang MD);
The First Aliated Hospital of
Anhui Medical University,
Anhui, Hefei, China
(Prof Y Wu MD, D Huang MD);
The First Aliated Hospital of
Xinjiang Medical University,
Urumqi, China
(Prof Suhua Li MD, S Wang MD);
Tongji Hospital, Tongji Medical
College, Huazhong University
of Science and Technology,
Wuhan, China (Prof G Xu MD,
S Ge MD); Zhejiang Provincial
Peoples Hospital, Hangzhou,
China (Prof Q He MD,
Q Shen MD); The First Aliated
Hospital of Zhejiang University,
Hangzhou, China
(Prof J Chen MD, P Zhang MD);
General Hospital of Ningxia
Medical University, Yinchuan,
China (Prof M Chen MD,
L Wu MD); Ningde Municipal
Hospital, Fujian Medical
University, Ningde, China
(X Liu MD, M Pan MD); Meishan
City Peoples Hospital, Meishan,
China (Prof Z Zhu MD,
X Zou MD); The First College of
Clinical Medical Science, China
Three Gorges University,
Yichang, China
(Prof Lin Yang MD*, P Zhu MD);
The Second Aliated Hospital
of Kunming Medical University,
Kunming, China (Prof X Lian MD,
J Zhao MD); Shanghai Ninth
Peoples Hospital, School of
Medicine, Shanghai JiaoTong

1465

Articles

University, Shanghai, China

(Prof F Ding MD, M Zhou MD);
Jiangxi Provincial Peoples
Hospital, Nanchang, China
(Prof Y Li MD,
Prof Lin Yang MD); The Fourth
Aliated Hospital of Harbin
Medical University, Harbin,
China (Prof Huamin Wang MD,
W Hu MD); Lanzhou University
Second Hospital, Lanzhou,
China (Prof Jianqin Wang MD,
Jing Wang MD); Shandong
Provincial Hospital Aliated to
Shandong University, Jinan,
China (Prof R Wang MD,
Bing Liu MD); Shanghai
Changzheng Hospital, The
Second Military Medical

selection criteria. The study was approved by the ethics

committees of Peking University First Hospital and the
enrolled study hospitals.

Survey design
The survey for AKI was designed to have three steps.
First, we screened adult patients who were admitted to
hospital with suspected AKI on the basis of changes in
their serum creatinine reported by the Laboratory
Information System (a software-based laboratory and
information management system that can track and
analyse laboratory data in hospitals). We used the 2012
Kidney Disease: Improving Global Outcomes (KDIGO)
denition of AKI as the major screening criteria.12 For
patients who had repeated serum creatinine assay with
intervals longer than 7 days and those who had recovering

22 provinces in China in four geographical regions

44 study hospitals

Screen possible AKI

2 223 230 adult admissions in 2013

374 286 in January, 2013, and July, 2013

562 191 with two times serum creatinine assay (253%)

94 770 in January, 2013, and July, 2013 (253%)

Screening possible AKI by software on the basis

of change in serum creatinine by Laboratory
Information System

Identify AKI

154 950 cases of possible AKI in 2013

26 086 cases from January and July, 2013

Review medical records of cases from January and July

Investigate AKI

KDIGO: 3687 cases

Expanded criteria: change

in serum creatinine 50%
3917 cases additional
to KDIGO

AKI detection rate, according to:

KDIGO: 099% (3687/374 286)
KDIGO + expanded criteria: 203%
(7604/374 286)

7604 detected AKI cases

Complete investigating record

Cause
Comorbidity
Recognition
Treatment
Short-term outcome

Procedures

Figure: Study prole

AKI=acute kidney injury. LIS=Laboratory Information System.

1466

Exclusion criteria:
Chronic kidney disease stage 5
Nephrectomy
Kidney transplantation
Peak serum creatinine <53 mol/L
Serum creatinine change could not
be attributed to AKI

AKI, we expanded the screening criteria to an increase or

decrease in serum creatinine by 265 mol/L or more
during hospital stay.
Second, hospital medical records of patients with
suspected AKI were reviewed on a case-by-case basis to
conrm the diagnosis. Because of limitations on the
amount of research work that our network nephrologists
could aord, we only included patients with suspected
AKI treated in hospital during January, 2013, or
July, 2013. The identication criteria included the 2012
KDIGO denition of AKI (KDIGO criteria).10 For those
who had no repeated serum creatinine assay within
7 days or with recovering AKI, we expanded the
identication criteria to an increase or decrease in serum
creatinine by 50% during hospital stay (using serum
creatinine concentration at admission as a baseline;
expanded criteria). Patients who had chronic kidney
disease stage 5, nephrectomy, kidney transplantation, or
peak serum creatinine of less than 53 mol/L were
excluded.13 Patients who met the identication criteria
but whose serum creatinine change could not be
attributed to AKI (eg, a serum creatinine decrease after
amputation) were also excluded.
Third, for patients who were conrmed as having AKI
by our survey (detected AKI), investigating records were
completed to document the sociodemographic status;
comorbidities; clinical departments; diseases or
conditions that could cause renal hypoperfusion
(appendix) or urinary obstruction; nephrotoxic drugs
and environmental nephrotoxins (appendix); invasive
procedures and surgeries; critical illness; AKI classication; renal replacement therapy (RRT);14,15 renal referral;
length of hospital stay; hospital cost; and all-cause
in-hospital death. AKI staging (1 to 3)12 was done at
three timepoints: when AKI could be diagnosed, when
AKI was at peak (ie, the highest AKI stage that a patient
reached during their whole in-hospital stay), and when
AKI was recognised by the physicians in charge.
We dened renal recovery at discharge as full recovery
with serum creatinine decreased to below threshold or
to the baseline. We dened partial recovery as serum
creatinine decreased by 25% or more from peak
concentration but remaining higher than the threshold
or baseline. We dened failure to recover as patient still
dependent on dialysis or serum creatinine decreased by
less than 25% from peak concentration.
All study investigators were nephrologists and renal
fellows and had completed a training programme. The
appendix contains details of the training and working
process of the study.

We dened AKI as having been recognised by the

physicians in charge if any documentation were
available of increased serum creatinine concentration,
concerns about renal dysfunction, or treatment
adjustments; otherwise we dened the case as

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

Articles

levels (academic vs local), economic development (by

tertiles of gross domestic product per head),
geographical regions, and source of acquired disease
(hospital-acquired or community-acquired). The
association between all-cause in-hospital mortality (yes
vs no) and relevant covariates was also analysed with a
logistic regression model. The additional covariates,
compared with the association with under-recognition
of AKI, included cardiovascular disease (yes vs no),
diabetes (yes vs no), severe comorbidity (yes vs no), peak
serum creatinine (continuous), RRT indication (yes vs
no), and recognition of AKI (non-recognition, delayed
recognition, and timely recognition).
We used Epidata software (version 3.1, Epidata
Association, Odense, Denmark) for data entry and
management. All p values are two-sided, and a p value of
less than 005 was deemed signicant. Analyses were
done with SAS software (version 9.1, SAS Institute, Cary,
NC, USA).

non-recognition. Recognition rate refers to the

percentage of patients with AKI who had their AKI
recognised by the physicians in charge. We dened
recognition as timely if AKI was recognised by the
physicians in charge within 3 days of the point from
which AKI could be diagnosed and before the disorder
progressed to higher stages, otherwise we dened
recognition as delayed. We additionally collected
International Classication of Diseases, Tenth Edition,
Clinical Modication (ICD-10-CM) codes for each
patient treated in hospital to identify reported AKI.

Statistical analysis
We calculated detection rates of AKI (number of detected
AKI cases per number of admission) at the hospital level
(academic or local) and by geographical regions in two
categories according to the identication criteria for AKI.
We used data from the 2013 report by the Chinese
National Health and Family Planning Commission and
National Bureau of Statistics as the standard population.
We present continuous data as mean (SD) or median
(IQR) as appropriate and categorical variables as n (%).
We described characteristics of patients and the statues
of recognition and treatment of AKI, stratied by
hospital level, geographical region, and level of
economic development. We compared groups using
one-way ANOVA or Kruskal-Wallis test for continuous
variables and test for categorical variables.
We analysed relevant covariates that might associate
with under-recognition of AKI (non-recognition and
delayed recognition; yes vs no) with multivariable
logistic regression and report odds ratios with 95% CIs
and p values of Wald test. Covariates included in the
analysis were age (change by 10 years), sex (male vs
female), chronic kidney disease (yes vs no), renal referral
(yes vs no), AKI stages at detection and at peak, hospital
Total

The funder of the study had no role in the study design,

data collection, data analysis, data interpretation, or
writing of the Article. The corresponding author had full
access to all the data in the study and had nal
responsibility for the decision to submit for publication.

Results
We assessed 2 223 230 adult patients (aged 18 years)
admitted to 44 study hospitals in the survey: 1 541 151 from
academic hospitals and 682 079 from local hospitals.
In three regions, we were not able to enrol local hospitals
because of research recourse limitations and enrolled
additional local hospitals in the adjacent regions instead.
Of the 2 223 230 patients treated in hospital in 2013 who
were screened, 562 191 (253%) had serum creatinine

North

Northwest

Southeast

Southwest

135513

47167

9842

58926

14007

4676

1176

6024

2131

2 223 230 (159%)

704 271 (151%)

270 575 (230%)

974 848 (162%)

273 536 (128%)

Population (million)16
Hospital admissions (million)17
Screened cases*

Role of the funding source

University, Shanghai, China

(Prof C Mei MD, T Zhang MD);
Renshou County Peoples
Hospital, Renshou, China
(Prof Jixian Xu MD, J Han MD);
The Aliated Provincial
Peoples Hospital of Shanxi
Medical University, Taiyuan,
China (Prof R Li MD, T Wen MD);
Taixing Peoples Hospital,
Taixing, China (Prof J Cao MD);
Ordos Central Hospital, Ordos,
Inner Mongolia, China
(L Zhang MD); Xinganmeng
Peoples Hospital, Wulanhaote,
Inner Mongolia, China
(Yan Wang MD); Fuyang City
Peoples Hospital, Fuyang,
Zhejiang, China (Jinhua Xu MD);
Ningbo Yinzhou Second
Hospital, Ningbo, China
(Prof B Bao MD); Zhongda
Hospital, Southeast University,
Nanjing, China
(Prof Bicheng Liu MD);
Hangzhou Hospital of
Traditional Chinese Medicine,
Hangzhou, China
(Prof H Chen MD); The Second
Hospital of Hebei Medical
University, Shijiazhuang, China
(Prof Shaomei Li MD); Guizhou
Provincial Peoples Hospital,
Guizhou Medical University,
Guiyang, China (Prof Y Zha MD);
Peking University Shenzhen
Hospital, Shenzhen, China
(Prof Q Luo MD); Hengxian
Peoples Hospital, Hengxian,
China (D Chen MD); Miyun
County Hospital, Beijing, China
(Prof Y Shen MD); The First
Aliated Hospital of Guangxi
Medical University, Nanning,
China (Prof Y Liao MD); Puer

19578

Number of screened cases in January and July

Total

374 286

119 284

46 099

162 973

Academic hospital

257 498

86 721

42 372

101 313

45 930
27 092

Local hospital

116 788

32 563

3727

61 660

18 838

AKI detection rate (KDIGO criteria), n (%; 95% CI)

Total
Academic hospital

3687 (099%; 095102)

974 (082%; 077087)

404 (088%; 079096)

1759 (108%; 103113)

550 (120%; 110130)

2897 (113%; 108117)

802 (092%; 086099)

397 (094%; 085103)

1268 (125%; 118132)

430 (159%; 144174)

172 (053%; 045061)

7 (019%; 005033)

491 (080%; 073087)

120 (064%; 052075)

1250 (272%; 257287)

Local hospital

790 (068%; 063072)

AKI detection rate (expanded criteria), n (%; 95% CI)

Total

7604 (203%; 199208)

2097 (176%; 168183)

851 (185%; 172197)

3406 (209%; 202216)

Academic hospital

5662 (220%; 214226)

1692 (195%; 186204)

827 (195%; 182208)

2272 (224%; 215233)

871 (321%; 300343)

Local hospital

1942 (166%; 159174)

405 (124%; 112136)

24 (064%; 039090)

1134 (184%; 173195)

379 (201%; 181221)

AKI=acute kidney injury. KDIGO=Kidney Disease: Improving Global Outcomes. *Percentage is screened cases per total hospital admissions.

Table 1: Detection of AKI in adult patients admitted to hospital in China by region

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

1467

Articles

City Peoples Hospital, Puer,

China (Z Zhang MD); Zoucheng
City Peoples Hospital,
Zoucheng, China (X Wang MD);
Taihe Hospital of Traditional
Chinese Medicine, Taihe, China
(K Zhang MD); Shenzhen

measured two or more times, and 154 950 (70%) were

suspected to have AKI on the basis of electronic
screening. During the 2 months of January, 2013, and
July, 2013, 374 286 people were treated in hospital, of
whom 26 086 (70%) had suspected AKI and their
medical records were reviewed (gure). The overall
Total (N=7604)

Age (years)

Academic hospitals
(N=5662)

Local hospitals
(N=1942)

616 (174)

607 (174)

640 (172)

Aged 1839 years

876 (115%)

686 (121%)

Aged 4059 years

2341 (308%)

1831 (323%)

510 (263%)

Aged 6079 years

3120 (410%)

2277 (402%)

843 (434%)

Age group

Aged 80 years

p value
<00001
<00001

190 (98%)

1267 (167%)

868 (153%)

399 (205%)

Men

4955 (652%)

3685 (651%)

1270 (654%)

080

CA-AKI

4136 (544%)

2859 (505%)

1277 (658%)

<00001

601 (79%)

409 (72%)

192 (99%)

Ward of hospital admission

Renal

<00001

Surgical

1639 (216%)

1231 (217%)

408 (210%)

Medicinal

3195 (420%)

2266 (400%)

929 (478%)

Intensive-care unit

2169 (285%)

1756 (310%)

413 (213%)

Pre-renal

3936 (518%)

2866 (506%)

1070 (551%)

00007

Intrinsic-renal

2100 (276%)

1632 (288%)

468 (241%)

<00001

AKI classication

Post-renal

670 (88%)

472 (83%)

198 (102%)

00126

Unclassied

898 (118%)

692 (122%)

206 (106%)

006

Renal hypoperfusion

5914 (778%)

4400 (777%)

1514 (780%)

082

Nephrotoxic drugs

5444 (716%)

4087 (722%)

1357 (699%)

005

Injury factors

Environmental toxins

191 (25%)

129 (23%)

62 (32%)

Sepsis

483 (64%)

328 (58%)

155 (80%)

00006

Other critical illness

3224 (424%)

2484 (439%)

740 (381%)

<00001

Surgery

1629 (214%)

1406 (248%)

223 (115%)

<00001

3483 (458%)

2613 (461%)

870 (448%)

1950 (256%)

1422 (251%)

528 (272%)

2171 (286%)

1627 (287%)

544 (280%)

RRT indication

896 (118%)

732 (129%)

164 (84%)

Mortality*

927 (124%)

727 (131%)

200 (105%)

00027

18 (1029)

18 (1131)

16 (927)

<00001

AKI stage

Hospital stay* (days)

Hospital cost* (US$)

00263

020

5071 (232311 824)

6497 (2984
14 294)

2795 (15225609)

<00001

<00001

Comorbidity
Pre-existing CKD

1847 (243%)

1431 (253%)

416 (214%)

00006

Hypertension

3190 (420%)

2408 (425%)

782 (403%)

008

Cardiovascular disease*

2666 (351%)

1923 (340%)

743 (383%)

00006

Diabetes

1404 (185%)

1065 (188%)

339 (175%)

018

Malignancy

1418 (186%)

1144 (202%)

274 (141%)

<00001

Data are mean (SD), n (%), or median (IQR), unless stated otherwise. AKI=acute kidney injury.
CA-AKI=community-acquired AKI. RRT=renal replacement therapy. CKD=chronic kidney disease. *Data missing for
mortality in 129 cases (100 in academic hospitals vs 29 in local hospitals), for days of hospital stay in ve cases (5 vs 0), for
hospital cost in 1411 cases (1132 vs 279), and for cardiovascular disease comorbidity in one case (1 vs 0). Hospital cost is
equal to the entire cost of the hospital admission of a patient with AKI.

Table 2: Characteristics of patients with AKI according to hospital levels

1468

detection rate of AKI of the 374 286 people admitted to

hospital was 3687 (10%) by the KDIGO criteria (variation
of 0812% between geographical regions) and
7604 (20%) by the expanded criteria (1827%; table 1).
More than half of the people detected as having AKI
were older than 60 years (table 2). Most cases were treated
in medicinal and surgical departments, with some
treated in dedicated renal divisions or intensive care
units (table 2). Pre-renal AKI accounted for half the
cases, and the most common injury factors that might be
implicated in the development of AKI included renal
hypoperfusion and nephrotoxic drugs or environmental
nephrotoxins. 3707 (488%) of 7604 patients had severe
critical illness, and 2171 (286%) of 7604 of AKI episodes
reached stage 3. A comparison between the features of
AKI in academic and local hospitals, and between
dierent geographical regions is shown in table 2 and in
the appendix.
We recorded a very high non-recognition rate of AKI by
the physicians in charge in both academic and local
hospitals (table 3), with 5608 (742%) of 7555 patients with
identiable AKI not being recognised by physicians in
charge during hospital stay. Of the 1947 patients with AKI
that were recognised, 343 (176%) were given a delayed
diagnosis. Using multivariate analysis, we noted that how
developed the economy of the region a patient comes from
was an independent risk factor for under-recognition of
AKI, with under-recognition being more likely in patients
from less developed regions than in those from more
developed regions. Independent protective factors against
under-recognition of patients with AKI included a patient
having previously diagnosed chronic kidney disease,
having higher peak AKI stage, and being referred to a
renal specialist (appendix).
The rate of reported AKI, dened through ICD-10
codes, was 1273 (034%) of all 374 286 hospital admission,
1273 (167%) of 7604 detected AKI, and 789 (363%) of the
2171 patients with AKI stage 3. Referral to a renal specialist
was recorded in 1625 (214%) of the 7604 patients with
detected AKI. For those who had indications for RRT
(n=896), only 531 (593%) received RRT treatment, with
higher rates in patients who lived in more economically
developed regions (p<00001; appendix).
Of 7475 patients with AKI (129 of the 7604 patients with
AKI detected in total were missing information on
mortality), the all-cause in-hospital mortality was
927 (124%), whereas another 1203 (161%) were
discharged with severe AKI without further treatment.
We contacted the relatives of 556 of these patients for
telephone interview and noted that 363 (653%) of these
patients had died within 3 months after discharge. Of the
927 recorded in-hospital deaths, the most common causes
were multiple organ dysfunction, causing 313 (338%)
deaths, and sepsis, causing 124 (134%) deaths. Using
multivariate analysis, we noted that increased age,
cardiovascular disease, delayed AKI recognition, critical
illness, high AKI peak stage, high serum creatinine peak

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

Articles

concentrations, and RRT indication were independent

risk factors for mortality; whereas renal referral was the
independent protective factor for in-hospital death
(table 4). Factors that might be associated with in-hospital
mortality in the subgroups of various geographical
regions are shown in the appendix. Full renal recovery
was achieved at discharge in 2047 (325%) of the
6302 patients with AKI in the study (927 patients died
during hospital stay and 375 patients were missing
information for serum creatinine at discharge of the
7604 patients with AKI detected in total; appendix).
The additional patients with AKI identied by the
expanded criteria had more community-acquired AKI,
less severity of kidney injury, and lower in-hospital
mortality than patients with AKI who met KDIGO criteria
(appendix). However, 1861 (475%) of the 3917 additional
patients with AKI meeting the expanded criteria had AKI
stage 2 or 3 and in-hospital mortality was 223 (58%) of
3861 (56 of the additional patients were missing
information on mortality)a higher percentage than the
2033 (055%) of 366 682 patients who were not dened as
having AKI by our study (p<00001).

Discussion

Non-recognised Delayed
AKI
recognition of
AKI

Timely
recognition of
AKI

p value*

5608 (742%)

343 (46%)

1604 (212%)

NA

Academic hospitals (n=5616)

4199 (748%)

237 (42%)

1180 (210%)

Local hospitals (n=1939)

1409 (727%)

106 (55%)

424 (219%)

Tertile 1 (n=2245)

1712 (763%)

125 (56%)

408 (182%)

Tertile 2 (n=2729)

2061 (755%)

90 (33%)

578 (212%)

Tertile 3 (n=2581)

1835 (711%)

128 (50%)

618 (239%)

594 (680%)

29 (33%)

251 (287%)

Total (N=7555)
Hospital level

00431

GDP per head

<00001

Age (years)
1839 (n=874)

<00001

4059 (n=2321)

1715 (739%)

93 (40%)

513 (221%)

6079 (n=3101)

2354 (759%)

132 (43%)

615 (198%)

945 (751%)

89 (71%)

225 (179%)

CKD basis (n=5718)

1150 (626%)

121 (66%)

566 (308%)

Non-CKD (n=1837)

4458 (780%)

222 (39%)

1038 (182%)

80 (n=1259)
Disease factors

<00001

Data are %, unless stated otherwise. AKI=acute kidney injury. GDP=gross domestic product. CKD=chronic kidney
disease. *The p value represents the general statistical dierence between the three groups of AKI recognition.
49 cases had no information on their AKI recognition status.

Table 3: Recognition of AKI by physicians in charge during hospital stay

In view of the International Society of Nephrology AKF

0by25 initiatives,4 the burden of AKI in China needed
evaluation. Our survey suggests that the detection rate of
AKI in Chinese adults treated in hospital in 2013 was
about 1% according to KDIGO criteria and 2% by the
expanded criteria.
To overcome the potential obstacles of inadequate
serum creatinine assay and overlook of AKI by doctors,
we designed a three-step survey abiding by the principles
of broad screening and enrolment but with strict
identication criteria. We expanded the criteria of serum
creatinine change to the whole hospital stay and included
serum creatinine decrease; meanwhile data from the
patients with suspected AKI were reviewed by
nephrologists to ascertain the rationality of serum
creatinine change and dene its relationship with AKI.
This method enabled us to record the prevalence of AKI
as accurately as possible.
On the basis of data reported by the National Health
and Family Planning Commission of the Peoples
Republic of China,17 we estimate that about 14 million
patients with AKI according to the KDIGO criteria
(29 million according to the expanded criteria) are
estimated to have been treated in hospital in China
during 2013, consuming about US$13 billion for their
whole in-hospital cost10% of Chinas total medical
expense.18 The in-hospital mortality rate of patients with
AKI was 124%. If the 65% death rate of patients
discharged with severe AKI without further treatment
is included, an estimated 700 000 patients died with
AKI in China during 2013. However, only 167% of the
AKI cases had been reported to the health-care system
(by ICD-10 codes), which has obviously undermined

Expanded criteria*
OR (95% CI)

KDIGO AKI criteria

p value

OR (95% CI)

p value

Age (per 10 years increase)

135 (127142)

<00001

133 (125142)

<00001

Sex (male vs female)

121 (102143)

00307

121 (099149)

006

History of cardiovascular disease

(yes vs no)

120 (102142)

00302

125 (103153)

00263

Diabetes (yes vs no)

113 (093137)

023

111 (088139)

039

Chronic kidney disease (yes vs no)

085 (070105)

013

081 (063103)

009

Delayed vs timely recognition of

AKI

145 (104203)

00298

129 (089189)

018

Severe comorbidity (yes vs no)

525 (436631)

<00001

484 (386606)

<00001

AKI stage at peak

1

189 (153232)

<00001

189 (146244)

205 (154274)

<00001

195 (138275)

00001

165 (132207)

<00001

187 (142247)

<00001

RRT indication (yes vs no)

169 (134213)

<00001

146 (113190)

00042

Renal referral (yes vs no)

066 (053082)

00003

061 (047080)

00002

Academic vs local hospital

117 (097141)

009

114 (091143)

026

Peak serum creatinine (natural

logarithm transformed)

<00001

All variables listed in the table were included in the logistic regression model and adjusted for hospital-acquired or
community-acquired patients. AKI=acute kidney injury. KDIGO=Kidney Disease: Improving Global Outcomes.
OR=odds ratio. RRT=renal replacement therapy. *7431 cases were included in the analysis after excluding 129 cases,
which were missing the information for all-cause in-hospital mortality, one for history of cardiovascular disease, and
49 for delayed or timely recognition of AKI. 3591 cases were included in the analysis after excluding 73 cases
missing the information for all-cause in-hospital mortality and 26 for delayed or timely recognition of AKI.

Reference value.
Table 4: Multivariate logistic regression analysis for factors associated with all-cause in-hospital
mortality in AKI

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

1469

Articles

Panel: Research in context

Systematic review
We searched PubMed for articles published in English or Chinese between Jan 1, 2000,
and Dec 31, 2014, with the terms acute kidney injury (AKI), acute renal failure,
incidence, prevalence, and China. We identied four single-centre, retrospective,
cross-sectional studies, published between 2005 and 2014, describing the prevalence of
AKI in patients admitted to general hospitals.5,6,28,29 The earliest study used International
Classication of Diseases codes to identify patients with AKI and the detection rate was
012%.28 The other three studies used Acute Kidney Injury Network criteria to identify AKI.
The prevalence of AKI was reported as 241% and 319% in two studies,5,6 and a rate of
014% for hospital-acquired AKI was reported in the other study.29
Interpretation
To our knowledge, this studywhich screened 2 223 230 adult patients admitted to
44 hospitals of dierent levels, located in various geographical regions across China, and
identied 7604 cases of AKIis the rst nationally representative investigation of AKI. We
studied the detection rate, cause, short-term outcome, and the present status of
diagnosis and treatment of AKI. We noted that serum creatinine measurement has been
seriously inadequate and that AKI has been greatly underdiagnosed and undertreated,
irrespective of the classication or the location of hospitals, but with more severity in less
economical developed regions. Policy makers and health professionals in China have a
great opportunity to improve medical care and outcomes for patients with AKI.
Longhua New District Central
Hospital, Shenzhen, China
(L Liu MD); Tongren Hospital
Shanghai Jiao Tong University
School of Medicine, Shanghai,
China (Prof P Mao MD);
Zhongwei City Peoples
Hospital, Zhongwei, China
(C Guo MD); Huaxian Peoples
Hospital, Huaxian, China
(J Li MD); Suihua City First
Hospital, Suihua, China
(Prof Z Wang MD); Qingpu
Branch of Zhongshan Hospital,
Fudan University, Shanghai,
China (S Bai MD); and
Qingxuxian City Peoples
Hospital, Qingxu, China
(S Shi MD)
Prof Haiyan Wang died in
December, 2014
Correspondence to:
Prof Li Yang, Renal Division,
Peking University First Hospital,
Xicheng, Beijing, 100034, China
li.yang@bjmu.edu.cn
For 0by25 see http://
www.0by25.org/
See Online for appendix

1470

the clinical and nancial signicance of AKI to the

government and the public.
Even though we have shown a substantial burden of
medical care and mortality from AKI in China, our
gures probably vastly underestimate the problem. Only
253% of the patients admitted to hospital had repeated
serum creatinine assay, which is much lower than the
gure reported by developed countries (632676%).9,13,19
A meta-analysis2 published in 2013 on world AKI
prevalence has reported an inverse correlation between
the pooled rate of AKI and the percentage of country
gross domestic product spent on total health expenditure.
We assume that one possible explanation could be the
underdiagnosis of AKI in the less-developed countries.
Because the present method of detecting AKI is mainly
based on changes in serum creatinine by use of an assay
that for most inpatients is not repeated during their
hospital stay, the prevalence of AKI in China could
arguably have been substantially underestimated
throughout the country. Similar situations might also
exist in other developing countries where unexplained
low prevalence of AKI had been reported.2022 This
situation proves to be a big challenge for the precise
evaluation of the global burden of AKI.
One of the distinct features in this Chinese AKI survey
was the high proportion of nephrotoxic drug exposure
(716%) before or at the time AKI developed, compared
with the 2050% reported in developed countries.3,23,24 This
high proportion of nephrotoxic drug exposure is consistent
with the increasing incidence of drug-induced disease in
China25in part, caused by easy access to a wide range of
non-prescription drugs. Since whether a nephrotoxic drug
is the exact cause of AKI is dicult to establish because of

the existence of complicated comorbidities, we did not

dene drug-induced AKI in the present survey and
therefore report a high proportion of potentially drug
related AKI in all AKI cases compared with that reported
by the AKI registry in Shanghai in 2009.26
Because China has a broad territory, traversing various
geographical regions, it provides a unique opportunity to
investigate the eects of environmental, traditional, and
socioeconomic factors on the characteristics of AKI and
formulate a respective strategy. This study revealed that
southwest Chinalocated in the subtropical, monsoonal
climate zonewas associated with higher AKI detection
rate, higher nephrotoxic exposure, and relatively higher
in-hospital mortality than the rest of the country. Therefore
crucial steps to prevent AKI and improve its treatment will
be to identify the common nephrotoxins and dene the
critical risk factors that aect mortality. Patients with AKI
in northwest China, which is a region with a typical
continental climate and relatively low gross domestic
product per head, had a lower rate of nephrotoxic drugs
being implicated with AKI, a higher rate of communityacquired AKI, and less cardiovascular disease, but relatively
higher in-hospital mortality than those from north China,
which has a similar continental climate but a higher gross
domestic product per head. These dierences might be
due to patients from northwest China having less access to
as wide a range of drugs and invasive procedures and
being exposed to fewer environmental nephrotoxins than
those from north China. The high in-hospital mortality
and rate of community-acquired AKI suggest the need to
further improve the management of AKI in the region.
The huge burden of AKI and its heterogeneity make it
a particular challenge in China. However, the medical
service for AKI facing this challenge is highly
disadvantaged. Extremely high rates of underdiagnosis
and undertreatment of AKI have been reported in
Chinese hospitals, irrespective of their classication or
location within the country, but with the highest rates in
least economically developed regions. With the fast
development of the economy in China, the country has
had a booming increase in various new and eective
drugs and invasive procedures, although advances in
kidney protection have lagged far behind. Nephrologists
in China need to take responsibility for organising the
battle against AKI, including education, training,27 and
development of an AKI alarm system that can improve
the detection, diagnosis, and management of AKI by the
entire medical professional community.
This multicentre survey of AKI in China, with to our
knowledge the largest sample size of its kind so far and
covering dierent levels of hospitals and geographical
regions, has uncovered the present AKI situation in
China (panel). The expanded diagnostic criteria helped
to identify community-acquired AKI, which would be
missed by KDIGO criteria for delayed serum creatinine
repeated assay, and sustained progressive kidney
injury, which had an increase in serum creatinine that

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

Articles

was less rapid than what the KDIGO criteria required.

However, since serum creatinine measurement has
been seriously inadequate, this survey probably missed
many patients with AKI in reality, thus the true
prevalence of AKI cannot be identied and the data we
have recorded is probably insucient. Additionally, the
present study dened AKI according to criteria based
on changes in serum creatinine independently
of urinary output, which would have further
underestimated the prevalence of AKI. AKI was
identied and its cause determined on the basis of
medical records, which potentially could cause
misclassication of the true situation. These limitations
can only be resolved by a prospective study designed
on the basis of health-care professionals having
sucient knowledge about AKI and the setting up of
an eective AKI alarm system in the hospitals.

Contributors
Li Yang (Peking University First Hospital) conceived, designed, and
organised the study, interpreted the results, and drafted the manuscript.
Jinwei Wang analysed the data. All authors contributed to collecting the
data on site. Minghui Zhao obtained funding, helped organise the study,
and revised the manuscript. Haiyan Wang conceived, organised, and
supervised the study, interpreted the results, and revised the manuscript.

17

Declaration of interests
We declare no competing interests.

10

11
12

13

14
15
16

18

19

Acknowledgments
This study was supported by National Project 985 to Peking University
for Clinical Study on Cooperation; Beijing Training Program for Talents
(20110009001000002); National Natural Science Foundation of China
(81270777); International Society of Nephrology Research Committee;
Bethune Fund Management Committee; and Fresenius Medical Care.
We thank Haixia Li for the technical support in setting up the Laboratory
Information System laboratory screening system. We would like to
dedicate this Article to our respectable Prof Haiyan Wang.
References
1
Lewington AJ, Cerda J, Mehta RL. Raising awareness of acute
kidney injury: a global perspective of a silent killer. Kidney Int 2013;
84: 45767.
2
Susantitaphong P, Cruz DN, Cerda J, et al. Acute Kidney Injury
Advisory Group of the American Society of Nephrology. World
incidence of AKI: a meta-analysis. Clin J Am Soc Nephrol 2013;
8: 148293.
3
Li PK, Burdmann EA, Mehta RL; World Kidney day Steering
Committee 2013. Acute kidney injury: global health alert.
Transplantation 2013; 95: 65357.
4
Remuzzi G, Horton R. Acute renal failure: an unacceptable death
sentence globally. Lancet 2013; 382: 204142.
5
Fang Y, Ding X, Zhong Y, et al. Acute kidney injury in a Chinese
hospitalized population. Blood Purif 2010; 30: 12026.
6
Lu RH, Fang Y, Gao JY, et al. The incidence and risk factors
associated with prognosis of acute kidney injury in hospitalized
patients. Chin J Nephrol 2012; 28: 194200.
7
Nash K, Hafeez A, Hou S. Hospitalacquired renal insuciency.
Am J Kidney Dis 2002; 39: 93036.
8
Chertow GM, Burdick E, Honour M, Bonventre JV, Bates DW.
Acute kidney injury, mortality, length of stay, and costs in
hospitalized patients. J Am Soc Nephrol 2005; 16: 3365370.

20
21

22

23

24

25

26
27
28

29

Uchino S, Bellomo R, Goldsmith D, Bates S, Ronco C. An assessment

of the RIFLE criteria for acute renal failure in hospitalized patients.
Crit Care Med 2006; 34: 191317.
Thomas M, Sitch A, Dowswell G. The initial development and
assessment of an automatic alert warning of acute kidney injury.
Nephrol Dial Transplant 2011; 26: 216168.
Srisawat N, Kellum JA. Acute kidney injury: denition,
epidemiology, and outcome. Curr Opin Crit Care 2011; 17: 54855.
Kidney Disease: Improving Global Outcomes (KDIGO) Acute
Kidney Injury Work Group. KDIGO Clinical practice guideline for
acute kidney injury. Kidney Int Suppl 2012; 2: 1138.
Zeng X, McMahon GM, Brunelli SM, Bates DW, Waikar SS.
Incidence, outcomes, and comparisons across denitions of AKI in
hospitalized individuals. Clin J Am Soc Nephrol 2014; 9: 1220.
Palevsky PM. Renal replacement therapy in acute kidney injury.
Adv Chronic Kidney Dis 2013; 20: 7684.
Ricci Z, Ronco C. Timing, dose and mode of dialysis in acute
kidney injury. Curr Opin Crit Care 2011; 17: 55661.
National Bureau of Statistics of the Peoples Republic of China.
Report on the national population by the year of 2013. http://data.
stats.gov.cn/easyquery.htm?cn=C01 (accessed Sept 13, 2015;
in Chinese).
National Health and Family Planning Commission of the Peoples
Republic of China. Report on the total hospital admission during
the year of 2013. http://www.nhfpc.gov.cn/mohwsbwstjxxzx/s7967/2
01401/9ae55e07c5c54e3788e080dfede59f5d.shtml (accessed
Nov 25, 2014; in Chinese).
National Bureau of Statistics of the Peoples Republic of China.
Report on the national total medical expense during the year of 2013.
http://data.stats.gov.cn/easyquery.htm?cn=E0103 (accessed
Sept 13, 2015; in Chinese).
Shema L, Ore L, Geron R, Kristal B. Hospitalacquired acute kidney
injury in Israel. Isr Med Assoc J 2009; 11: 26974.
Naicker S, Aboud O, Gharbi MB. Epidemiology of acute kidney
injury in Africa. Semin Nephrol 2008; 28: 348353.
Singh TB, Rathore SS, Choudhury TA, Shukla VK, Singh DK,
Prakash J. Hospitalacquired acute kidney injury in medical,
surgical, and intensive care unit: a comparative study.
Indian J Nephrol 2013; 23: 2429.
Lombardi R, Yu L, Younes IM, Schor N, Burdmann EA.
Epidemiology of acute kidney injury in Latin America.
Semin Nephrol 2008; 28: 32029.
Schissler MM, Zaidi S, Kumar H, Deo D, Brier ME, McLeish KR.
Characteristics and outcomes in communityacquired versus
hospitalacquired acute kidney injury. Nephrology (Calton) 2013;
18: 18387.
Finlay S, Bray B, Lewington AJ, et al. Identication of risk factors
associated with acute kidney injury in patients admitted to acute
medical units. Clin Med 2013; 13: 23338.
INRUD China Center Group. Expert consensus on management of
medication errors in China. Adverse Drug Reactions J (Chinese) 2014;
16: 32126.
Miaolin C, Yucheng Y, Yun Z, et al. Analysis of drug-induced acute
renal failure in Shanghai. Natl Med J China 2009; 89: 74449.
International Society of Nephrology. 0 by 25. http://www.theisn.org/
initiatives/0x25-initiative?showall=&start=3(accessed Nov 25, 2014).
Wang Y, Cui Z, Fan M. Retrospective analysis on Chinese patients
diagnosed with acute renal failure hospitalized during the last
decade. Am J Nephrol 2005; 25: 51419.
Gao J, Chen MJ, Wang XY, Wang HX, Zhuo L. Risk factors and
prognosis of acute kidney injury in adult hospitalized patients:
a two year outcome. Minerva Urol Nefrol 2015; 67: 17985.

www.thelancet.com Vol 386 October 10, 2015

Downloaded from ClinicalKey.com at Universitas Kristen Duta Wacana May 17, 2016.
For personal use only. No other uses without permission. Copyright 2016. Elsevier Inc. All rights reserved.

1471