張景義,郭 靜,董 釗,武士芳,張飛飛,鄭寶霞
?
·論著·
糖化血紅蛋白和空腹血糖早期診斷糖尿病的截點(diǎn)研究
張景義1*,郭 靜1,董 釗1,武士芳1,張飛飛1,鄭寶霞2
目的 探討糖化血紅蛋白(HbA1c)和空腹血糖(FPG)早期診斷糖尿病的截點(diǎn)。方法 選取2012年1月—2016年4月在華北理工大學(xué)附屬開灤總醫(yī)院門診及住院進(jìn)行HbA1c和FPG測定及葡萄糖耐量試驗(yàn)(OGTT)的患者766例。使用日本HA-8160型全自動(dòng)糖化血紅蛋白分析儀,采用高壓液相色譜法測定HbA1c;使用HITACHI 7600型全自動(dòng)生化分析儀,采用己糖激酶法測定FPG。根據(jù)1999年WHO修訂的糖尿病診斷標(biāo)準(zhǔn),即以餐后2 h血糖(2 hPG)≥11.1 mmol/L為糖尿病組(n=628)和2 hPG<11.1 mmol/L為非糖尿病組(n=138)。繪制受試者工作特征(ROC)曲線,分析HbA1c和FPG診斷糖尿病的適宜截點(diǎn),并計(jì)算其靈敏度、特異度、陽性預(yù)測值、陰性預(yù)測值、陽性似然比、陰性似然比、約登指數(shù)及Kappa值。結(jié)果 HbA1c診斷糖尿病的ROC曲線下最大面積為0.918〔95%CI(0.895,0.942),P<0.001〕,得出的HbA1c截點(diǎn)為6.55%;FPG診斷糖尿病的ROC曲線下最大面積為0.901〔95%CI(0.877,0.926),P<0.001〕,得出的FPG截點(diǎn)為6.23 mmol/L。HbA1c=6.55%診斷糖尿病的靈敏度為90.1%,特異度為75.4%,陽性預(yù)測值為90.1%,陰性預(yù)測值為75.4%,陽性似然比為3.7,陰性似然比為0.1,約登指數(shù)為0.66,Kappa值為0.63。FPG=6.23 mmol/L診斷糖尿病的靈敏度為85.0%,特異度為81.9%,陽性預(yù)測值為84.9%,陰性預(yù)測值為81.9%,陽性似然比為4.7,陰性似然比為0.2,約登指數(shù)為0.67,Kappa值為0.65。HbA1c=6.55%和FPG=6.23 mmol/L聯(lián)合診斷糖尿病的靈敏度為76.6%,特異度為95.5%,陽性預(yù)測值為65.5%,陰性預(yù)測值為98.3%,陽性似然比為17.0,陰性似然比為0.3,約登指數(shù)為0.72,Kappa值為0.70。結(jié)論 HbA1c=6.55%和FPG=6.23 mmol/L聯(lián)合診斷糖尿病具有較好的特異度,且約登指數(shù)及Kappa值均較單獨(dú)HbA1c=6.55%、FPG=6.23 mmol/L高,可以作為早期診斷糖尿病的方法之一。
糖尿病;血紅蛋白A,糖基化;空腹血糖;診斷;截點(diǎn)
張景義,郭靜,董釗,等.糖化血紅蛋白和空腹血糖早期診斷糖尿病的截點(diǎn)研究[J].中國全科醫(yī)學(xué),2017,20(7):808-811,822.[www.chinagp.net]
ZHANG J Y,GUO J,DONG Z,et al.Cut-off points of glycosylated hemoglobin and fasting plasma glucose for early diagnosis of diabetes mellitus[J].Chinese General Practice,2017,20(7):808-811,822.
隨著生活水平的提高,糖尿病的發(fā)病率也在逐年升高,已經(jīng)成為繼腫瘤、心血管病之后第三大威脅人類健康的疾病[1],如果不能及時(shí)診斷糖尿病,那么會(huì)造成糖尿病并發(fā)癥的提早出現(xiàn)。因此,確定一個(gè)最佳的早期診斷糖尿病的截點(diǎn)變得尤為重要。1979年,美國國家糖尿病數(shù)據(jù)組(National Diabetes Data Group,NDDG)制定并發(fā)表了NDDG糖尿病診斷標(biāo)準(zhǔn)[2]。1980年,WHO對(duì)該標(biāo)準(zhǔn)進(jìn)行調(diào)整并正式公布[3]。1985年,WHO又對(duì)其標(biāo)準(zhǔn)進(jìn)行了補(bǔ)充修訂[4]。1997年美國糖尿病學(xué)會(huì)(American Diabetes Association,ADA)進(jìn)行了大量循證醫(yī)學(xué)糖尿病相關(guān)文獻(xiàn)的研究分析后,建議將糖尿病診斷的空腹血糖(fasting plasma glucose,FPG)截點(diǎn)從7.8 mmol/L降低至7.0 mmol/L[4]。1999年,WHO與國際糖尿病同盟會(huì)共同采納了ADA提出的將糖尿病FPG截點(diǎn)下調(diào)的建議[5],糖化血紅蛋白(hemoglobin A1c,HbA1c)與糖尿病的并發(fā)癥具有密切的聯(lián)系,與目前血糖檢測的方式比較,更能突顯其相對(duì)標(biāo)準(zhǔn)化的優(yōu)點(diǎn),許多學(xué)者和專家建議采用HbA1c進(jìn)行檢查和診斷糖尿病[6]。本研究旨在探討HbA1c和FPG早期診斷糖尿病的截點(diǎn),以期盡早發(fā)現(xiàn)糖尿病人群,減少漏診率,為社會(huì)減輕負(fù)擔(dān)。
1.1 研究對(duì)象 選取2012年1月—2016年4月在華北理工大學(xué)附屬開灤總醫(yī)院門診及住院進(jìn)行HbA1c和FPG測定及葡萄糖耐量試驗(yàn)(oral glucose tolerance test,OGTT)的患者766例。納入標(biāo)準(zhǔn):(1)年齡18~75周歲;(2)進(jìn)行OGTT前3 d,每天進(jìn)食量包含至少200 g的碳水化合物;(3)進(jìn)行OGTT前沒有發(fā)生酮癥酸中毒、高滲狀態(tài)等急性并發(fā)癥。排除標(biāo)準(zhǔn):(1)心肌梗死、腦血管意外、急性感染、燒傷以及創(chuàng)傷;(2)有嚴(yán)重貧血或血液系統(tǒng)疾??;(3)藥物源性高血糖。本研究雖未經(jīng)過醫(yī)學(xué)倫理委員會(huì)的批準(zhǔn),但患者均知情同意。
1.2 研究方法
1.2.1 HbA1c和FPG測定 所有患者禁食水10 h,于次日8:00空腹采集肘前靜脈血3 ml,置于真空抗凝采血管內(nèi),封口,使用日本HA-8160型全自動(dòng)糖化血紅蛋白分析儀,采用高壓液相色譜法測定HbA1c;使用HITACHI 7600型全自動(dòng)生化分析儀,采用己糖激酶法測定FPG。
1.2.2 OGTT 所有患者口服50%葡萄糖溶液150 ml,并在5~10 min內(nèi)服完,在30、60、120、180 min時(shí)分別依次抽取肘前靜脈血3 ml,并盡快送檢(30 min以內(nèi)送檢)[7]。根據(jù)1999年WHO修訂的糖尿病診斷標(biāo)準(zhǔn),即以餐后2 h血糖(2 hPG)≥11.1 mmol/L為糖尿病組(n=628)和2 hPG<11.1 mmol/L為非糖尿病組(n=138)[5]。
2.1 兩組一般資料比較 兩組性別、年齡間差異無統(tǒng)計(jì)學(xué)意義(P>0.05),而HbA1c、FPG水平間差異有統(tǒng)計(jì)學(xué)意義(P<0.05,見表1)。
2.2HbA1c、FPG及2hPG水平的相關(guān)性分析 糖尿病組2hPG水平為(18.7±5.0)mmol/L,非糖尿病組為(8.4±1.8)mmol/L。糖尿病組HbA1c、FPG水平均與2hPG水平呈直線正相關(guān)(r值分別為0.608、0.786,P<0.001),且HbA1c水平與FPG水平亦呈直線正相關(guān)(r=0.543,P<0.001);非糖尿病組HbA1c、FPG水平均與2hPG水平呈直線正相關(guān)(r值分別為0.476、0.478,P<0.001),且HbA1c水平與FPG水平亦呈直線正相關(guān)(r=0.443,P<0.001)。
2.3HbA1c、FPG影響因素的一元線性回歸分析 以HbA1c為因變量,2hPG、FPG分別為自變量,進(jìn)行一元線性回歸分析,得到回歸方程:HbA1c=3.390+0.285×2hPG、HbA1c=3.810+0.441×FPG(見表2)。以FPG為因變量,2hPG為自變量,進(jìn)行一元線性回歸分析,得到回歸方程:FPG=1.162+0.457×2hPG(見表2)。
表1 兩組一般資料比較
注:a為χ2值;HbA1c=糖化血紅蛋白,F(xiàn)PG=空腹血糖
表2 HbA1c、FPG影響因素的一元線性回歸分析
Table 2 Univariate linear regression analysis of the influencing factors of HbA1cand FPG
變量b95%CISEt值P值HbA1c 常量3.390(2.091,5.645)0.14822.956<0.001 2hPG0.285(0.012,1.288)0.01028.665<0.001HbA1c 常量3.810(2.301,4.957)0.13927.428<0.001 FPG0.441(0.052,2.304)0.01923.546<0.001FPG 常量1.162(1.058,3.594)0.2125.471<0.001 2hPG0.457(0.016,1.634)0.01238.722<0.001
注:2 hPG=餐后2 h血糖
2.4 HbA1c、FPG診斷糖尿病的ROC曲線分析 HbA1c、FPG診斷糖尿病的ROC曲線見圖1。結(jié)果顯示HbA1c診斷糖尿病的ROC曲線下最大面積為0.918〔95%CI(0.895,0.942),P<0.001〕,得出的HbA1c截點(diǎn)為6.55%;FPG診斷糖尿病的ROC曲線下最大面積為0.901〔95%CI(0.877,0.926),P<0.001〕,得出的FPG截點(diǎn)為6.23 mmol/L。HbA1c=6.55%診斷糖尿病的靈敏度為90.1%,特異度為75.4%,陽性預(yù)測值為90.1%,陰性預(yù)測值為75.4%,陽性似然比為3.7,陰性似然比為0.1,約登指數(shù)為0.66,Kappa值為0.63;FPG=6.23 mmol/L診斷糖尿病的靈敏度為85.0%,特異度為81.9%,陽性預(yù)測值為84.9%,陰性預(yù)測值為81.9%,陽性似然比為4.7,陰性似然比為0.2,約登指數(shù)為0.67,Kappa值為0.65;HbA1c=6.55%和FPG=6.23 mmol/L聯(lián)合診斷糖尿病的靈敏度為76.6%,特異度為95.5%,陽性預(yù)測值為65.5%,陰性預(yù)測值為98.3%,陽性似然比為17.0,陰性似然比為0.3,約登指數(shù)為0.72,Kappa值為0.70(見表3)。
糖尿病最大的危害是其并發(fā)癥,除了引起心臟、血管、腎臟、視網(wǎng)膜及神經(jīng)等病變外,還可引起急性會(huì)厭炎[8]、溶血性尿毒綜合征[9]以及糖尿病性心肌梗死[10]等多種病變。HbA1c是在高血糖作用下血紅蛋白發(fā)生的一種非酶促糖化反應(yīng)所生成的產(chǎn)物,是糖尿病診斷及篩查的重要指標(biāo)[11],而且不受近期飲食和抽血時(shí)間的影響,可以排除近期飲食、情緒和應(yīng)激等對(duì)血糖水平的影響[12-13],歐美等國家已將HbA1c的檢測用于2型糖尿病的篩查和診斷[14-15]。FPG檢測是糖尿病診斷的常用指標(biāo),反映短期內(nèi)血糖的變化,測量FPG影響因素較多。如果需要進(jìn)行OGTT,其過程較為繁瑣、費(fèi)時(shí),而且成本高。2010年ADA明確提出HbA1c≥6.5%或FPG≥7.0 mmol/L可診斷糖尿病,而多數(shù)學(xué)者認(rèn)為單獨(dú)應(yīng)用HbA1c或FPG均有超過15%的患者漏診[16]。
注:HbA1c=糖化血紅蛋白,F(xiàn)PG=空腹血糖,ROC曲線=受試者工作特征曲線
圖1 HbA1c、FPG診斷糖尿病的ROC曲線
Figure 1 ROC curves of HbA1cand FPG in all subjects for the diagnosis of diabetes mellitus
表3 不同血糖指標(biāo)截點(diǎn)對(duì)糖尿病的診斷價(jià)值
本研究結(jié)果顯示,HbA1c=6.55%時(shí),診斷糖尿病的陽性似然比為3.7,陰性似然比為0.1,約登指數(shù)為0.66,Kappa值為0.63;FPG=6.23 mmol/L時(shí),診斷糖尿病的陽性似然比為4.7,陰性似然比為0.2,約登指數(shù)為0.67,Kappa值為0.65;HbA1c=6.55%和FPG=6.23 mmol/L時(shí),診斷糖尿病的陽性似然比為17.0,陰性似然比為0.3,約登指數(shù)為0.72,Kappa值為0.70。陽性似然比>10或陰性似然比<0.1時(shí),診斷或排除某種疾病的可能性就明顯增加[17]。Kappa值越大,說明兩種診斷方法的一致性越好,其中Kappa≥0.75時(shí),兩種診斷方法的一致性較好;0.40≤Kappa<0.75時(shí),兩種診斷方法的一致性一般;Kappa<0.40時(shí),兩種診斷方法的一致性差[17]。說明HbA1c=6.55%和FPG=6.23 mmol/L聯(lián)合診斷糖尿病的結(jié)果較為可靠。進(jìn)一步分析HbA1c為6.55%時(shí),診斷糖尿病的靈敏度為90.1%,雖然略稍低于HbA1c為6.5%的90.6%,但是其特異度(75.4%)高于HbA1c為6.5%的74.1%,而且約登指數(shù)較HbA1c為6.5%高,Kappa值與HbA1c為6.5%一致,說明HbA1c為6.55%可能是早期診斷糖尿病的更可靠的參考值。FPG為6.23 mmol/L的靈敏度比FPG為7.0 mmol/L提高了12.2%,并且約登指數(shù)及Kappa值都較FPG為7.0 mmol/L高,說明FPG為6.23 mmol/L是早期診斷糖尿病的更可靠的參考值。HbA1c=6.55%和FPG=6.23 mmol/L聯(lián)合診斷糖尿病的約登指數(shù)及Kappa值均較單獨(dú)HbA1c=6.55%、FPG=6.23 mmol/L高,說明采用聯(lián)合指標(biāo)檢測能夠更準(zhǔn)確地早期診斷糖尿病。
總之,HbA1c=6.55%和FPG=6.23 mmol/L作為早期診斷糖尿病的截點(diǎn)能夠很好地提高特異度。特別是在無條件進(jìn)行OGTT的醫(yī)院,采用HbA1c和FPG聯(lián)合檢測進(jìn)行診斷優(yōu)于單一指標(biāo)。
作者貢獻(xiàn):郭靜進(jìn)行文章的構(gòu)思與設(shè)計(jì)、結(jié)果的分析與解釋、撰寫論文及英文的修訂;董釗進(jìn)行研究的實(shí)施與可行性分析;武士芳進(jìn)行數(shù)據(jù)收集;張飛飛進(jìn)行數(shù)據(jù)整理;鄭寶霞進(jìn)行統(tǒng)計(jì)學(xué)處理;張景義進(jìn)行論文的修訂,并負(fù)責(zé)文章的質(zhì)量控制及審校,對(duì)文章整體負(fù)責(zé),監(jiān)督管理。
本文無利益沖突。
[1]BEAGLEHOLE R,BONITA R,HORTON R,et al.Priority actions for the non-communicable disease crisis[J].Lancet,2011,377(9775):1438-1447.DOI:10.1016/S0140-6736(11)60393-0.
[2]National Diabetes Data Group.Classification and diagnosis of diabetes mellitus and other categories of glucose intolerance[J].Diabetes,1979,28(12):1039-1057.
[3]WHO Expert Committee on Diabetes Mellitus:second report[J].World Health Organ Tech Rep Ser,1980,646:1-80.
[4]Report of the Expert Committee on the Diagnosis and Classification of Diabetes Mellitus[J].Diabetes Care,1997,20(7):1183-1197.
[5]World Health Organisation.Definition,diagnosis,and classification of diabetes mellitus and its complications[R].Geneva:World Health Organisation,1999.
[6]MA X J,PAN J M,BAO Y Q.Combined assessment of glycated albumin and fasting plasma glucose improves the detection of diabetes in Chinese subjects[J].Clin Exp Pharmacol Physiol,2010,37(10):974-979.DOI:10.1111/j.1440-1681.2010.05417.
[7]American Diabetes Association.Standards of medical care in diabetes——2013[J].Diabetes Care,2013,36(Suppl 1):S11-66.DOI:10.2337/dc13-S011.
[8]GE R,MAO Y,ZHANG X L,et al.Cervical necrotizing fasciitis and a descending mediastinal abscess caused by acute epiglottitis with diabetes mellitus:a life-threatening complication[J].Diabetes Res Clin Pract,2012,95(2):e31-33.DOI:10.1016/j.diabres.2011.10.010.
[9]ROGOWSKA-KALISZ A,TKACZYK M,SZALAPSKA-ZAWODNIAK M.Diabetes mellitus as a rare complication of hemolytic uremic syndrome——case report[J].Pol Merkur Lekarski,2010,28(163):46-49.
[10]LYER S N,DRAKE A J,WEST R L,et al.Diabetes muscle infarction:a rare complication of long-standing and poorly controlled diabetes mellitus[J].Case Rep Med,2011,2011:407921.DOI:10.1155/2011/407921.
[11]KESAVADEV J,SHANKAR A,PILLAI P B,et al.Cost-effective use of telemedicine and self-monitoring of blood glucose via Diabetes Tele Management System(DTMS) to achieve target glycosylated hemoglobin values without serious symptomatic hypoglycemia in 1 000 subjects with type 2 diabetes mellitus——a retrospective study[J].Diabetes Technol Ther,2012,14(9):772-776.DOI:10.1089/dia.2012.0088.
[12]CHAMNAN P,SIMMONS R K,FOROUHI N G,et al.Incidence of type 2 diabetes using proposed HbA1cdiagnostic criteria in the European prospective investigation of cancer-norfolk cohort:implications for preventive strategies[J].Diabetes Care,2011,34(4):950-956.DOI:10.2337/dc09-2326.
[13]BONORA E,KIECHL S,MAYR A,et al.High-normal HbA1cis a strong predictor of type 2 diabetes in the general population[J].Diabetes Care,2011,34(4):1038-1040.DOI:10.2337/dc10-1180.
[14]PAJUNEN P,PELTONEN M.ERIKSSON J G,et al.HbA1cin diagnosing and predicting type 2 diabetes in impaired glucose tolerance:the Finnish Diabetes Prevention Study[J].Diabet Med,2011,28(1):36-42.DOI:10.1111/j.1464-5491.2010.03183.
[15]DEMDEN M C,SHAW J E,COLMAN P G,et al.The role of HbA1cin the diagnosis of diabetes mellitus in Australia[J].Med J Aust,2012,197(4):220-221.
[16]BAO Y,MA X,LI H,et al.Glycated haemoglobin A1cfor diagnosing diabetes in Chinese population:cross sectional epidemiological survey[J].BMJ,2010,340:c2249.DOI:10.1136/bmj.c2249.
[17]孫振球.醫(yī)學(xué)統(tǒng)計(jì)學(xué)[M].北京:人民衛(wèi)生出版社,2010:112. SUN Z Q.Medical statistics[M].Beijing:People ′s Health Publishing House,2010:112.
(本文編輯:崔沙沙)
Cut-off Points of Glycosylated Hemoglobin and Fasting Plasma Glucose for Early Diagnosis of Diabetes Mellitus
ZHANGJing-yi1*,GUOJing1,DONGZhao1,WUShi-fang1,ZHANGFei-fei1,ZHENGBao-xia2
1.DepartmentofEndocrinology,KailuanGeneralHospitalAffiliatedtoNorthChinaUniversityofScienceandTechnology,Tangshan063000,China2.CTRoom,KailuanGeneralHospitalAffiliatedtoNorthChinaUniversityofScienceandTechnology,Tangshan063000,China
*Correspondingauthor:ZHANGJing-yi,Professor,Mastersupervisor;E-mail:zhangjingyi63@sina.com
Objective To investigate the cut-off points of glycosylated hemoglobin(HbA1c) and fasting plasma glucose(FPG) for early diagnosis of diabetes mellitus.Methods The subjects enrolled in this study were 766 consecutive outpatients and inpatients of Kailuan General Hospital Affiliated to North China University of Science and Technology who underwent measurement of HbA1c,FPG and oral glucose tolerance test(OGTT) between January 2012 and April 2016.We used HA-8160 Automatic Glycosylated Hemoglobin Analyzer made in Japan with high pressure liquid chromatography to determine HbA1c,and HITACHI 7600 Automatic Biochemical Analyzer with hexokinase method to determine FPG.We divided the patients into diabetic group〔2-hour postprandial blood glucose(2 hPG) ≥11.1 mmol/L,n=628〕 and non-diabetic group(2 hPG<11.1 mmol/L,n=138) based on the Definition,Diagnosis and Classification of Diabetes Mellitus and Its Complications revised by WHO in 1999.We drew the receiver operating characteristic(ROC) curves of HbA1cand FPG,and investigated the optimal cut-off points of them for the diagnosis of diabetes mellitus with analyzing the sensitivity,specificity,positive predictive value,negative predictive value,positive and negative likelihood ratios,Youden′s index andKappavalue.Results The largest area under curve(AUC) of HbA1c′s ROC curve in the diagnosis of diabetes mellitus was 0.918〔95%CI(0.895,0.942),P<0.001〕,the cut-off point was 6.55%,with a sensitivity of 90.1%,specificity of 75.4%,positive predictive value of 90.1%,negative predictive value of 75.4%,positive likelihood ratio of 3.7,negative likelihood ratio of 0.1,Youden′s index was 0.66 andKappavalue of 0.63.The largest AUC of FPG′s ROC curve in the diagnosis of diabetes mellitus was 0.901〔95%CI(0.877,0.926),P<0.001〕,the cut-off point was 6.23 mmol/L,with a sensitivity of 85.0%,specificity of 81.9%,positive predictive value of 84.9%,negative predictive value of 81.9%,positive likelihood ratio of 4.7,negative likelihood ratio of 0.2,Youden′s index was 0.67 andKappavalue of 0.65.The cut off points of HbA1c(6.55%) and FPG(6.23 mmol/L) for the combined diagnosis of diabetes mellitus had a sensitivity of 76.6%,specificity of 95.5%,positive predictive value of 65.5%,negative predictive value of 98.3%,positive likelihood ratio of 17.0,negative likelihood ratio of 0.3,Youden′s index was 0.72 andKappavalue of 0.70.Conclusion HbA1c(6.55%) combined with FPG(6.23 mmol/L) can be one of the methods for early diagnosis of diabetes mellitus,for it has good specificity.Moreover,it has higher Youden′s index andKappavalue than HbA1c(6.55%) or FPG(6.23 mmol/L) used alone as the early diagnosis method of diabetes mellitus.
Diabetes mellitus;Hemoglobin A,glycosylated;Fasting plasma glucose;Diagnosis;Cut-off point
R 587.1
A
10.3969/j.issn.1007-9572.2017.07.010
2016-10-25;
2017-01-23)
1.063000 河北省唐山市,華北理工大學(xué)附屬開灤總醫(yī)院內(nèi)分泌科
2.063000 河北省唐山市,華北理工大學(xué)附屬開灤總醫(yī)院CT室
*通信作者:張景義,教授,碩士生導(dǎo)師;E-mail:zhangjingyi63@sina.com