prevalence and control of diabetes in chinese adults. jama 宁光.pdf
PrevalenceandControlofDiabetesinChineseAdults YuXu,PhD;LiminWang,PhD;JiangHe,MD,PhD;YufangBi,MD,PhD;MianLi,PhD;TiangeWang,PhD; LinhongWang,PhD;YongJiang,MS;MengDai,BS;JieliLu,MD,PhD;MinXu,PhD;YichongLi,MS;NanHu,MS; JianhongLi,MS;ShengquanMi,PhD;Chung-ShiuanChen,MS;GuangweiLi,MD,PhD;YimingMu,MD,PhD; JiajunZhao,MD,PhD;LingzhiKong,MD;JialunChen,MD;ShenghanLai,MD,MPH;WeiqingWang,MD,PhD; WenhuaZhao,PhD;GuangNing,MD,PhD;forthe2010ChinaNoncommunicableDiseaseSurveillanceGroup IMPORTANCENoncommunicablechronicdiseaseshavebecometheleadingcausesof mortalityanddiseaseburdenworldwide. OBJECTIVEToinvestigatetheprevalenceofdiabetesandglycemiccontrolintheChinese adultpopulation. DESIGN, SETTING, AND PARTICIPANTSUsingacomplex,multistage,probabilitysampling design,weconductedacross-sectionalsurveyinanationallyrepresentativesampleof 98 658Chineseadultsin2010. MAIN OUTCOMES AND MEASURESPlasmaglucoseandhemoglobinA1clevelsweremeasured afteratleasta10-hourovernightfastamongallstudyparticipants,anda2-houroralglucose tolerancetestwasconductedamongparticipantswithoutaself-reportedhistoryof diagnoseddiabetes.Diabetesandprediabetesweredefinedaccordingtothe2010American DiabetesAssociationcriteria;whereas,ahemoglobinA1clevelof7.0%wasconsidered adequateglycemiccontrol. RESULTSTheoverallprevalenceofdiabeteswasestimatedtobe11.6%(95%CI,11.3%-11.8%) intheChineseadultpopulation.Theprevalenceamongmenwas12.1%(95%CI,11.7%-12.5%) andamongwomenwas11.0%(95%CI,10.7%-11.4%).Theprevalenceofpreviously diagnoseddiabeteswasestimatedtobe3.5%(95%CI,3.4%-3.6%)intheChinese population:3.6%(95%CI,3.4%-3.8%)inmenand3.4%(95%CI,3.2%-3.5%)inwomen.The prevalenceofundiagnoseddiabeteswas8.1%(95%CI,7.9%-8.3%)intheChinese population:8.5%(95%CI,8.2%-8.8%)inmenand7.7%(95%CI,7.4%-8.0%)inwomen.In addition,theprevalenceofprediabeteswasestimatedtobe50.1%(95%CI,49.7%-50.6%)in Chineseadults:52.1%(95%CI,51.5%-52.7%)inmenand48.1%(95%CI,47.6%-48.7%)in women.Theprevalenceofdiabeteswashigherinolderagegroups,inurbanresidents,andin personslivingineconomicallydevelopedregions.Amongpatientswithdiabetes,only25.8% (95%CI,24.9%-26.8%)receivedtreatmentfordiabetes,andonly39.7%(95%CI, 37.6%-41.8%)ofthosetreatedhadadequateglycemiccontrol. CONCLUSIONS AND RELEVANCETheestimatedprevalenceofdiabetesamonga representativesampleofChineseadultswas11.6%andtheprevalenceofprediabeteswas 50.1%.Projectionsbasedonsampleweightingsuggestthismayrepresentupto113.9million Chineseadultswithdiabetesand493.4millionwithprediabetes.Thesefindingsindicatethe importanceofdiabetesasapublichealthprobleminChina. JAMA.2013;310(9):948-958.doi:10.1001/jama.2013.168118 Editorialpage916 Supplementalcontentat jama.com AuthorAffiliations:Author affiliationsarelistedattheendofthis article. GroupInformation:2010China NoncommunicableDisease SurveillanceGroup,Investigatorsare listedattheendofthisarticle. TheCorrespondingAuthors:Guang Ning,MD,PhD,KeyLaboratoryfor EndocrineandMetabolicDiseasesof MinistryofHealth,Departmentof EndocrineandMetabolicDiseases, Rui-JinHospital,ShanghaiJiao-Tong UniversitySchoolofMedicine,197 Rui-Jin2ndRd,Shanghai,200025, China(gningsibs.ac.cn);Wenhua Zhao,PhD(whzhaoilsichina.org); andWeiqingWang,MD,PhD(wqingw hotmail.com). Research OriginalInvestigation 948jama.com Downloaded From: http:/jama.jamanetwork.com/ by a SHANGHAI JIATONG UNIVERSITY User on 09/06/2013 N oncommunicable chronic diseases have become the leadingcausesofmortalityanddiseaseburdenworld- wide. It was estimated that 34.5 million deaths glob- allywereduetononcommunicablediseasesin2010,whichre- flected a significant increase from 1990.1,2Mortality from diabetes doubled during this period and increased to 1.3 mil- lion deaths worldwide in 2010.1In addition, diabetes is a ma- jorriskfactorforischemicheartdiseaseandstroke,whichcol- lectively killed an estimated 12.9 million people globally in 2010.1,2As the most populous country, the rapid increase in morbidity and mortality from noncommunicable diseases in China contributed to this pandemic.3,4According to national data,noncommunicablediseasesaccountedforanestimated 80% of deaths and 70% of total disease burden in China in 2005.4 The prevalence of diabetes has increased significantly in recent decades and is now reaching epidemic proportions in China.5-8The prevalence of diabetes was less than 1% in the Chinese population in 1980.6In subsequent national surveys conductedin1994and2000-2001,theprevalenceofdiabetes was 2.5% and 5.5%, respectively.7,8The most recent national survey in 2007 reported that the prevalence of diabetes was 9.7%, representing an estimated 92.4 million adults in China withdiabetes.5Althoughdifferentsamplingmethods,screen- ing procedures, and diagnostic criteria were used, these data documentarapidincreaseindiabetesintheChinesepopula- tion. Recently, the American Diabetes Association (ADA) inte- gratedglycatedhemoglobinA1c(HbA1c)intothediagnosticcri- teriafordiabetesinitsupdated2010guidelines.9Justasthere is less than 100% concordance between fasting plasma glu- coseand2-hourplasmaglucosetests,thereisnotfullconcor- dance between HbA1cand either glucose-based test. There- fore, the prevalence of diabetes could be underestimated in thepreviousnationalsurveysbasedontheADA2010criteria. Furthermore, the previous national surveys could not assess diabetescontrolintheChinesepopulationbecauseHbA1cwas not measured. To estimate the prevalence and control of dia- betesinthegeneralChinesepopulation,wemeasuredHbA1c, fasting plasma glucose, and 2-hour plasma glucose in a large andnationallyrepresentativesampleof98 658adultswhowere 18 years or older in 2010. Methods China Noncommunicable Disease Surveillance 2010 in- cludedall162studysitesfromtheChineseCenterforDisease Control and Prevention s (CDC s) National Disease Surveil- lancePointSystem,whichwasdesignedtoselectanationally representativesampleofthegeneralpopulation,coveringma- jor geographic areas of all 31 provinces, autonomous regions, andmunicipalitiesinmainlandChina.10Thefirstlevelofsam- plingwasstratifiedby7geographicregions(Northeast,North, East,South,Southwest,NorthwestandCentralareas)and3mu- nicipalities(Beijing,Tianjin,andShanghai).Thesecondlevel of sampling was stratified by urban and rural locations. The thirdlevelofsamplingwasstratifiedby4socioeconomicstrata in rural areas and 3 population size strata in urban areas. The Surveillance Point System includes approximately 1% of the total Chinese population.10 At each site, a complex, multistage, probability sampling design was used to select participants who were representa- tiveofcivilian,noninstitutionalizedChineseadults.Onlyper- sonswhohadbeenlivingintheircurrentresidenceforatleast 6monthswereeligibletoparticipate.Inthefirststage,4sub- districts in urban areas or townships in rural areas were se- lected from each site with probability proportional to size. In the second stage, 3 neighborhood communities or adminis- trativevillageswereselectedwithprobabilityproportionalto size. In the third stage, households within each neighbor- hoodcommunityoradministrativevillagewerelisted,and50 households were randomly selected. In the final stage, 1 per- sonwhowasatleast18yearsoldwasselectedrandomlyfrom each household using a Kish selection table.11When the se- lected individual refused or was unavailable, a replacement household was selected from all households of similar com- position in the same neighborhood or village after excluding thealreadyselectedhouseholdsusingthesimplerandomsam- pling method. The replacements were used to ensure an ad- equate sample size within each selected neighborhood com- munityoradministrativevillageandtomaximizethenational representativenessofthesurveyedsampleswithregardtogeo- graphic distribution, economic development, and urbaniza- tion.Thehouseholdsinourstudywerecategorizedintosingle- person households, families of couples who were married or cohabitingadultswithorwithoutchildren,single-parentfami- lies,orhouseholdswith3ormorecohabitinggenerations.The household composition information was obtained from the government household registration system, which includes personal identifiers such as name, parents, spouse, and date ofbirthforeachmemberwithinahouseholdwhoisalocalper- manent resident. If the second household did not partici- pate, a third household was selected. All replacements were successfullyrecruitedbythethirdsampling.Ifnoavailablere- placementwasfoundinthesameneighborhoodorvillage,the nearest neighborhood or village was used. A total of 109 023 peoplewereselectedand98 658participatedinthesurvey.The overall response rate was 90.5% (replacement rate, 9.25%, eTable 1 in the Supplement). The study protocol was approved by the ethical review committee of the China CDC and other participating insti- tutes. Written informed consent was obtained from all study participants. Data collection was conducted in examination centers at localhealthstationsorcommunityclinicsintheparticipants residential area by trained staff according to a standard pro- tocol. A questionnaire including information on demo- graphic characteristics, medical history, and lifestyle factors was administered by trained interviewers. Current smoking wasdefinedashavingsmoked100cigarettesinone slifetime and currently smoking cigarettes. Current drinking was de- fined as alcohol intake more than once per month during the past12months.TheGlobalPhysicalActivityQuestionnairewas usedtoassessphysicalactivity.12Bodyweightandheightwere measuredaccordingtoastandardprotocolandbodymassin- DiabetesControlinChineseAdultsOriginalInvestigation Research jama.comJAMASeptember4,2013Volume310,Number9949 Downloaded From: http:/jama.jamanetwork.com/ by a SHANGHAI JIATONG UNIVERSITY User on 09/06/2013 dex(BMI),whichiscalculatedasweightinkilogramsdivided by height in meters squared. Waist circumference was mea- suredonstandingparticipantsmidwaybetweentheloweredge of the costal arch and the upper edge of the iliac crest. Over- weight was defined as a BMI of 25.0 to 29.9, and obesity was defined as a BMI of 30.0 or higher.13Central obesity was de- finedaswaistcircumference90cmormoreinmenand80cm ormoreinwomen.14Bloodpressurewasmeasuredatthenon- dominant arm 3 times consecutively with a 1-minute interval betweenthemeasurementswiththeparticipantinaseatedpo- sition after 5 minutes of rest using an automated device (OMRON Model HEM-7071, Omron Co). Blood samples were collected in all participants after an overnightfastofatleast10hours.Participantswithoutaself- reported history of diabetes were given a standard 75-g glu- cose solution, and plasma glucose was measured at 0 and 2 hours after administration during the oral glucose tolerance test.Bloodspecimensfortheglucosetestwerecollectedusing vacuum bloodcollection tubes containing anticoagulant so- dium fluoride and were centrifuged on site within 2 hours of collection. Plasma glucose was measured locally using glu- coseoxidaseorhexokinasemethodswithin24hours.Allstudy laboratoriessuccessfullycompletedastandardizationandcer- tification program. TheHemoglobinCapillaryCollectionSystem(Bio-RadLabo- ratories)wasusedtocollectcapillarybloodsamplesstrictlyac- cording to the manufacturer s instructions. Blood specimens preparedusingthisprocedurewerestableforupto4weeksat 2°C to 8°C. The capillary blood specimens were shipped and stored at 2°C to 8°C until HbA1cwas measured within 4 weeks after collection by high-performance liquid chromatography using the VARIANT II Hemoglobin Testing System (Bio-Rad Laboratories) at the central laboratory in the Shanghai Insti- tute of Endocrine and Metabolic Diseases, which was certifi- cated by the National Glycohemoglobin Standardization Pro- gram. Capillary HbA1cwas converted to venous values using a validated formula. In addition, we performed an internal vali- dationstudywithpairedsamplesfrom6648adultsthatshowed high agreement in HbA1cvalues from capillary whole blood samplespreparedwiththeHemoglobinCapillaryCollectionSys- tem vs the venous whole blood samples collected using EDTA tubes (capillary HbA1c= 0.0143 + 0.9983 × venous HbA1c). Serumsampleswerealiquotedandfrozenat80°Cwithin 2 hours of collection and shipped by air in dry ice to the cen- trallaboratory,whichwascertificatedbytheCollegeofAmeri- can Pathologists. Serum total cholesterol, low-density lipo- protein (LDL) cholesterol, high-density lipoprotein (HDL) cholesterol, and triglycerides were measured using an auto- analyser (Abbott Laboratories). A stringent quality assurance and quality control pro- gram was implemented to ensure the validity and reliability of study data. All investigators and research staff underwent aweeklongtrainingsessionontheuseofstandardizedproto- cols and instruments for data collection. Only certified staff wereallowedtocollectdata.Alllaboratoryequipmentwascali- bratedandblindedduplicatesampleswereused.Alldatawere double entered in a database and then compared and cor- rected for errors. According to the ADA 2010 criteria, diabetes was defined as(1)aself-reportedpreviousdiagnosisbyhealthcareprofes- sionals,(2)fastingplasmaglucoselevelof126mg/dLorhigher (toconverttomillimolesperliter,multiplyby0.0555),(3)2-hour plasmaglucoselevelof200mg/dLorhigher,or(4)HbA1ccon- centration of 6.5% or more. Prediabetes or categories of in- creasedriskofdiabetesweredefinedas(1)fastingplasmaglu- cose levels between 100 mg/dL and 125 mg/dL, (2) 2-hour plasma glucose levels between 140 mg/dL and 199 mg/dL, or (3) HbA1cconcentrations between 5.7% and 6.4% in partici- pants without a prior diabetes diagnosis. Awareness was de- fined as the proportion of individuals who reported a history of physician-diagnosed diabetes among all patients with dia- betes. Treatment was defined as the proportion of individu- als taking diabetes medications among all patients with dia- betes.Controlwasdefinedastheproportionofindividualswith anHbA1cconcentrationoflessthan7.0%amongpatientswith diabetes who were treated. Demographic and metabolic characteristics of study par- ticipants were described in means (95% CIs) for continuous variables and percentages (95% CIs) for categorical variables intheoverallpopulationandinsubgroupsofsex,location(ur- ban/rural),age,stagesofeconomicdevelopment,andcatego- ries of BMI and waist circumference. Prevalence and 95% CIs ofdiabetes,prediabetes,andsubtypesbyvariouscriteriawere estimatedbysubgroupsandoverall.Eachofthe162studysites was categorized into underdeveloped, intermediately devel- oped,ordevelopedaccordingtotheirregion sgrossdomestic product per capita in 2009. Age-standardized prevalences of Chinese adults with diabetes and prediabetes were also esti- mated in the overall population and among subgroups based on China 2010 census data. AllcalculationswereweightedtorepresenttheoverallChi- nese adult population aged 18 years or older. Weight coeffi- cients were derived from 2010 China population census data and the sampling scheme of the current survey to obtain na- tionalestimates.StandarderrorswerecalculatedusingtheTay- lor-lineariza