Comparison Of HbA1c And OGTT Criteria To Diagnose Diabetes Among Chinese

BackgroundHistorically, the glucose measurement has been used to diagnose diabetes. Measurement of haemoglobin Ale (HbAlc) level is an integrated measure of circulating glucose levels and tracks well in subjects over time. Epidemiological studies have shown that HbAlc levels in nondiabetic adults can predict incident diabetes, cardiovascular disease morbidity and mortality, and total mortality. A reliable measure of chronic glycemic levels such as measurement of HbAlc levels, which captures the degree of glucose exposure over time and is related more intimately to the risk of complications than measurement of single or episodic measures of glucose levels, may be a better biochemical marker of diabetes and should be considered a diagnostic tool. Therefore, HbAlc≥6.5% was defined as one of the criteria for the diagnosis of diabetes by the American Diabetes Association (ADA).ObjectiveTo examine the sensitivity and specificity of HbAlc testing for the diagnosis of type 2 diabetes in high-risk adults in China and to compare the cardiovascular risk factors between 2 groups of patients with glycemic status classified by the 2 different tests.Research design and methodSubjectsWe estimated the prevalence of diabetes among Chinese adults in a national study from June 2007 through May 2008 and conducted the current study in Jinan, Shandong. We included subjects above 45 years but without diagnosed diabetes. The Ethics Committee of the Qilu Hospital approved the protocol. Informed consent was obtained from all volunteers.Study designAfter an overnight fast, participants underwent the oral glucose tolerance test (OGTT), and fasting and 2-h glucose levels were measured; a questionnaire was completed to document the presence of hypertension, cardiovascular disease, and whether there was a first-degree relative with diabetes. Height and weight were measured, body mass index (BMI) was calculated, and blood pressure was measured as described. Blood samples were taken from the antecubital vein for measurement of fasting plasma glucose (FPG) and levels of HbAlc, total cholesterol (TC), low-density lipoprotein cholesterol (LDL-c), triglycerides (TG) and high-density lipoprotein cholesterol (HDL-c). Plasma glucose was measured by the glucose dehydrogenase method and levels of HbAlc by high-performance liquid chromatography, TC by the cholesterol oxidase method, TG by the enzymatic method and HDL-c by the direct method. LDL-c level was calculated by the Friedewald equation for subjects with triacylglycerol concentrations< 4.53 mmol/L:LDL-c=(total cholesterol-HDL-c-triacylglycerols)/5.Following 2009 ADA criteria, subjects were classified into 4 groups as follows: OGTT-diabetes (FPG≥7.0mmol/l or 2-h PG≥11.1mmol/l; HbAlc-diabetes (HbAlc leve1≥6.5%); true-diabetes (both criteria positive); non-diabetes (neither of the criteria positive).HbAlc level for diagnosis was evaluted by sensitivity (percentage of subjects with OGTT-diabetes and a positive HbAlc screening test), specificity (percentage of subjects without OGTT-diabetes and a negative screening test), positive predictive value (PPV) (percentage of subjects with a positive HbAlc screening test with OGTT-diabetes), negative predictive value (NPV) (percentage of subjects with a negative screening test without OGTT-diabetes). To understand which test could best predict cardiovascular risk factors, we compared blood pressure, BMI and levels of TC, TG, LDL-c, and HDL-c among the 4 groups (true-diabetes, OGTT-diabetes, HbA1c-diabetes, nondiabetes).Statistical methodsAnalyses involved use of SPSS v11.5 for Windows (SPSS Inc., Chicago, IL). General characteristics of patients are summarized with means±SD for continuous data, median (interquartile range) for TG and HDL-C levels, and number (%) for categorical data. To compare the TG and HDL-C levels among the 4 groups, the Wilcoxon rank sum and Kruskal-Wallis tests were used. To compare blood pressure, BMI, TC, and LDL-C levels among the 4 groups, we used analysis of covariance, with age as a covariate to adjust for potential confounding, and obtained least-square means±SE. Two-sided p values of＜0.05 were considered significant. The receiver operating characteristic was used to describe the sensitivities and specifities of the HbA1c test in determining the presence of type 2 diabetes as defined by the OGTT.ResultsWe included 701 subjects with mean age 52±11 years (range 21-83 years) (392 [56%] women). The mean BMI was 26±3 kg/m2, and HbA1c level was 5.34±0.61%.Among 701 high-risk subjects who underwent OGTT,94 (13%) had diabetes, and 236 (34%) had pre-diabetes. Among the same 701 high-risk subjects who underwent HbA1c testing,65 (9%) had diabetes, and 123 (18%) pre-diabetes, so OGTT could detect more diabetic patients than HbA1c.The area under the receiver operating characteristics curve for detecting undiagnosed diabetes was 0.843 (95% confidence interval 0.798 to 0.888) for HbA1c (figurel). Compared with the OGTT, with HbAlc≥5.1%, the sensitivity was 98.7% and specificity 33%, PPV 18% and NPV 99%(Table 1). With HbAlc≥7.0%, the sensitivity was 24% and specificity 99.7%, PPV 92%, and NPV 89%. With HbAlc≥6.5%, the sensitivity was 48.2% and specificity 97.8%, PPV 69% and NPV 92%.There was not full concordance between HbAlc and OGTT (Table 2).30.7% of participants with HbAlc≥6.5% were classified as non-type 2 diabetes by OGTT criteria, and 52% subjects with type 2 diabetes according to the OGTT criteria had HbAlc<6.5%.To compare HbAlc and OGTT criteria with a type 2 diabetes complication, we considered cardiovascular risk factors.Subjects in the 3 diabetes groups were older than those in the non-diabetes group (Table 3). When adjusting for age, the values of every risk factor measured varied among the 4 groups with the exception of diastolic pressure and HDL-c concentrations. All risk factors, except for diastolic pressure, were higher for the true-diabetes group than for the nondiabetes group. All lipid levels, except HDL-c, were higher for the HbAlc-and OGTT-diabetes groups than for the nondiabetes group, with no difference between the HbAlc-and OGTT-diabetes groups.ConclusionsWe found high specificity and low sensitivity with use of HbAlc level≥6.5% to diagnose diabetes mellitus. The HbAlc test and OGTT did not reach full concordance. The 2 diagnostic criteria reflect similar levels of cardiovascular risk factors.Our results showed diabetes prevalence lower with the HbAlc-based diagnostic method than the OGTT criteria, we found HbAlc and OGTT criteria without full concordance. In the present study,30.7% of participants with A1C≥6.5% were not classified as diabetic by OGTT criteria and 52% of the participants with diabetes by OGTT criteria would be classified as normoglycemic by AlC; the corresponding probability of AlC≥6.5% among diabetic case subjects based on an OGTT was 68.3% in our study, which was higher than in Denmark, the U.K., Australia, Greenland, Kenya study(17-42%).To compare HbAlc and OGTT criteria with a type 2 diabetes macro vascular complication, we considered cardiovascular risk factors., In our data, levels of cardiovascular risk factors were similar for the HbAlc- and OGTT-diabetes groups and higher in the 2 groups than in the non-diabetes group. To our knowledge, this is the first study of cardiovascular risk factors between HbAlc- and OGTT-diagnosed subjects with diabetes.According to our data, only performing HbAlc instead of OGTT would miss 52% of those who are diabetes and have higher levels of cardiovascular risk factors.On the other hand, only performing OGTT would miss 30.7% of those with high HbAlc levels who are possibly diabetic retinopathy, so there is no single assay for hyperglycemia that can be considered the gold standard.In conclusion, the limited sensitivity of the AlC test may miss about half of patients who are diabetes and have higher levels of cardiovascular risk factors among Chinese. BackgroundAbout 60.7% of diabetic subjects remain undiagnosed because most of them have not typical symptoms in early stage of the disease.The diabetic microvascular or macrovascular complications have already existed in early stage, even in pre-diabetes stage, and some clinical trials have shown the diabetes and its complications could be delayed or even prevented by early interferences, so the diagnoses should be done as earlier as possible.The guideline screening for diabetic subjects in high-risk individuals has been presented by 2009 ADA, either fasting plasma glucose (FPG)≥7.0mmol/L or 2-h postload glucose (2hPG)≥11.lmmol/L define diabetes independently. The FPG level is easy to obtain and can be used for diabetes screening criteria, however, subjects with FPG<7.0mmol/Land diagnosed by the 2hPG criteria were ignored and a part of diabetes could not be detected.. OGTT is time wasting and inconvenient,so it is difficult to get physicians and patients to use the oral glucose tolerance test (OGTT). HbAlc has been suggested as the diagnostic or screening criteria for diabetes, however, several studies have shown diabetes prevalence was lower with the HbA1c-based diagnostic criteria. Using FPG and HbAlc screening diabetes were reported by several articles, but the IFG criteria were 6.1-6.9mmol/L and old diagnostic criteria only FPG>7.0mmol/L or 2-h postload glucose (2hPG)≥11.lmmol/L were used.Using 2010 ADA criteria, individuals with HbAlc≥6.5%, FPG≥7.0mmol/L or 2-h postload glucose (2hPG)≥11.1mmol/L can be diagnosed as diabetes independently, we evaluated whether the combination of FPG and HbA1c measurements enhanced the detection of diabetes in high-risk individuals. 1.ObjectiveThe aim of this study was to evaluate the use of HbA1c and FPG as predictors of type 2 diabetes and, accordingly, to develop a rational approach to screening for type 2 diabetes.2. Methods2.1 SubjectsThe national study from June 2007 through May 2008 to estimate the prevalence of diabetes among Chinese adults was conducted in China, and we conducted the study in Jinan and Zibo, Shandong. We collected individuals with risk factors for diabetes but without diagnosed diabetes from the total people. According to 2009 ADA,700 adults who were above 45years were included(5). The Ethics Committee of the Qilu Hospital approved the protocol. Informed consent was obtained from all individuals.2.2 Study designAfter an overnight fast, participants underwent an oral glucose-tolerance test, and fasting and 2-hour glucose levels were measured, detailed method has been reported (15). we measured HbAlc in the high-risk individuals. HbA1c was measured by high-pressure liquid chromatography, plasma glucose was measured by glucose dehydrogenase method. Using 2010 ADA criteria (14),Participants were classified as having diabetes if they had either HbAlc≥6.5% or FPG≥7.0mmol/L or 2hPG≥11.1mmol/L, FPG<5.6 mmol/L= normal fasting glucose(NFG); FPG 5.6-6.9 mmol/1 = IFG (impaired fasting glucose); 2-h postload glucose<7.8 mmol/L = normal glucose tolerance (NGT);2-h postload glucose 7.8-11.1 mmol/L = IGT (impaired glucose tolerance).2.3 Statistical methodsData are presented as mean±SD. Receiver operating characteristic curves were constructed to calculate sensitivity and specificity of HbAlc cut points for type 2 diabetes diagnosis.Cut points were defined as that point on the curve where the sum of sensitivity and specificity was highest. Comparisons between groups were performed using the chi-square test for categorical data.3. ResultsA total of 700 subjects aged 21-83 years (52±11) were studied. There were 392 (56%) women and 308 (44%) men.The body mass index was 26±3 kg/m2.3.1 The data of using FPG to screening diabetes were presented in table 1. There were 93 diabetes were detected, among them 46 subjects with FPG≥7.0mmol/L,47 diabetes with FPG< 7.0mmol/L.According to ROC analysis(figure 1), FPG≥5.6mmol/L predicted diabetes with a sensitivity of 78.0% and a specificity of 73.3%. FPG levels<4.8mmol/L and≥6.7mmol/L have 98.7% and 99% accuracy for predicting the absence and presence of type 2 diabetes, respectively. Based on the above results and reference (14), FPG was divided into the following groups:≥7.0mmol/L, 6.9-5.6mmol/L, 5.5-4.8mmol/L, and <4.8mmol/L (table 1). At FPG 6.9-5.6mmol/L, in order to find 35 patients with diabetes,193 people need to do OGTT; at FPG5.5-4.8mmol/L, in order to find 9 patients with diabetes, 368 people need to do OGTT. While 92 individuals with FPG<4.8mmol/L need not to do OGTT and can be rule out diabetes directly.3.2 The data of using HbA1c to screening diabetes were presented in table2. There were 98 diabetes were detected, among them 54 subjects with HbA1c≥6.5%,44 diabetes with HbA1c<6.5%. According to ROC analysis(figure 1), HbA1c≥5.6% predicted type 2 diabetes with a sensitivity of 86.0% and a specificity of 77.0%. HbA1c levels<5.0 and >6.5% have 98.7% and 99% accuracy for predicting the absence and presence of type 2 diabetes, respectively. Based on the above results and references(7,8), HbA1c was divided into the following groups:≥7.0%,6.9-6.5%,6.4-5.6%(impaired HbA1c),5.5-5.0%,<5.0%. (table 2). At HbA1c 5.6-6.4%, in order to find 30 patients with diabetes,157 people need to do OGTT; and at HbA1c 5.0-5.5%, in order to find 7 patients with diabetes,203 people need to do OGTT. While 284 individuals with HbAlc<5.0% do not need to do OGTT and can be rule out diabetes directly.33 The data of using FPG +HbA1c to screening diabetes were presented in table 3. There were 98 diabetes were detected, among them 64 subjects with HbA1c≥6.5% or FPG>7.0mmol/L, these people can be diagnosed as diabetes directly. Based on figure 1 results, we used HbA1c screening for diabetes in 179 IFG group. In individuals with HbAlc5.6-6.4%, in order to find 18 patients with diabetes, 68 people need to do OGTT; in individuals with HbAlc 5.5-5.0%, in order to find 6 patients with diabetes, 64 people need to do OGTT; while 47 individuals with HbAlc<5.0% do not need to do OGTT and can be rule out diabetes directly. We also used HbAlc to screen diabetes in individuals with FBG<5.6mmol/L. In individuals with HbAlc5.6-6.4%, in order to find 8 patients with diabetes, 80 people need to do OGTT; 372 individuals with HbAlc<5.6% do not need to do OGTT and can be rule out diabetes directly.3.4 The comparison of three methods to screen diabetes were presented in table 4. According to FPG, 49.5% patients with diabetes were directly diagnosed, and 14% individuals with risk factors for diabetes can be directly rule out diabetes, in order to find 82 patients with diabetes, 193 people need to do OGTT; while according to HbAlc or FPG+HbAlc, 52.7%, and 65.3% patients with diabetes were directly diagnosed respectively, and 40% and 60.5% individuals with risk factors for diabetes can be directly rule out diabetes, according to HbAlc, in order to find 88 patients with diabetes, 158 people need to do OGTT, according to FPG+HbAlc in order to find 82 patients with diabetes, only 68 people need to do OGTT. More patients with diabetes were directly diagnosed by FPG+HbAlc than by FPG(P<0.05), and more individuals with risk factors for diabetes were directly rule out diabetes by FPG+HbAlc than by FPG or HbAlc(P<0.01). FPG+HbAlc significantly reduced the number of people to do OGTT.4 ConclusionFPG combined with HbA1c may be a useful strategy to identify diabetes in individuals with risk factors for diabetes.According to our result, we can advice the procedure for diabetes screening as the following:1) give measurements of FPG and HbA1c in the individuals with high-risk factors defined by 2009 ADA,2) If FP≥7.0mmol/L or HbAlc≥6.5%,the individuals have diabetes; 3) if FBG< 4.8mmol/L or HbAlc<5.0%, or FPG<5.6mmol/L and HbAlc 5.5-5.0%, diabetes can be excluded.4) if FPG 5.6-6.9mmol/L and HbAlc 5.6-6.4%, OGTT is needed to find diabetes;5) if only IFG or impaired HbAlc, refering the high-risk factors, the more risk factors for diabetes the greater the likelihood should be. BackgroudObesity is a major risk factor for cardiovascular diseases and diabetes. Prospective epidemiological studies have shown increased body mass index (BMI) and waist circumference (WC) to be independent risk factors for type 2 diabetes mellitus, coronary artery disease (CAD), and hypertension. However, the definition of obesity lacks consensus, as does the specific aspects of obesity that contribute to cardiovascular disease. The major disagreement centers on whether it is the total amount or the distribution of adipose tissue that confers a greater risk of cardiovascular disease. Epidemiologic studies have shown that body mass index (BMI), a general measure of obesity, is a powerful predictor of cardiovascular disease. However, a growing body of evidence indicates that waist circumference (WC)—measure of central obesity—also provides information on the risk of cardiovascular disease, and other visceral adiposity measures such as ratio of waist to height (WHtR) appear to be better predictors of cardio-metabolic risk factors than BMI. Many studies of the association of total amount or distribution of adipose tissue and cardiovascular risk factors are performed in the United States and Europe; few data are from mainland China, so knowledge of risk factors are lacking in China.ObjectiveWe aimed to evaluate the predictive value of the body mass index (BMI), waist circumference (WC), and ratio of waist to height (WHtR) for the presence of several cardiovascular risk conditions -- hypertension, dyslipidemia, metabolic syndrome, and type 2 diabetes--in a Chinese population in Jinan, China.Methods Study ParticipantsThe China National Diabetes and Metabolic Disorders Study, conducted from June 2007 through May 2008, was a cross-sectional study designed to provide current, reliable data on the prevalence of diabetes and associated metabolic risk factors in the adult population in China13; we completed the survey in Jinan, Shandong. We used a stratified sampling method to select a representative sample of the general population in 6 districts of Jinan, Shandong. A total of 3400 individuals aged 20-74 years took part in the survey. At elephone appointment was made a week before the survey, and all participants received an agenda the day before the survey. After excluding data for 18 people lacking data on 2-hr plasma glucose levels, we included data for 3,006 people in the final analysis. The participants were mainly of Han ethnicity. The Ethics Committee of Qilu Hospital Shandong University approved the protocol. Informed consent was obtained from all subjects.After fasting overnight, subjects were required to arrive at the community clinic in every district before 7:00 and were asked to complete a questionnaire to document the presence of hypertension, cardiovascular disease, diabetes and the treatment of these diseases.Blood pressure and anthropometric measurementsBlood pressure, body weight, height, and WC were measured by standard methods 14. Body mass index was calculated as weight (Kg) divided by height squared (m2). Trained technicians performed the interview in community clinics in the subjects'residential areas.Laboratory testsA blood sample was drawn from the antecubital vein for measuring fasting plasma glucose (FPG), total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), triglycerides (TGs) and high density lipoprotein cholesterol (HDL-C). Thereafter, subjects received 75 g glucose orally, and 5 ml blood was collected at 2 hr for measurement of plasma glucose. Plasma glucose was measured by the glucose dehydrogenase method, TC by the cholesterol oxidase method, TGs by the enzymatic method and HDL-C by the direct method; LDL-C concentrations were calculated by the Friedewald equation. Statistical analysisData for subjects were analyzed separately by sex. Additionally, we analyzed high-risk age groups as defined by the American Diabetes Association16 with the ages≥45 yr in men and women. Continuous variables are expressed as mean±SD, and discrete variables are expressed as numbers and percentages. Comparisons between groups involved Student t test for continuous variables and chi-square test for categorical data. Sensitivity and specificity were examined by ROC analysis, and the areas under the ROC curve (AUC) cut-off values were calculated for each anthropometrical parameter and risk condition. An AUC of 1 indicates perfect separation between affected and nonaffected subjects, and an AUC of 0.5 indicates no discriminative value of the test. Individual cutoffs were defined as the point on the curve where the sum of sensitivity and specificity was highest. All data analyses involved use of SPSS v11.5 (SPSS Inc., Chicago, IL). Differences between 2 AUCs were tested with Z values: with Z> 1.96, P is< 0.05, and with Z> 2.58, P is< 0.01. For the comparison of corresponding AUCs for males and females, Z=(AA-AB)/√(SEA2+SEB2), where SE is standard error; for comparing AUCs for anthropometric indicators in predicting the same binary condition17, Z=(AA-AB)/√(SEA2+SEB2-2γSEA SEB), whereγis the correlation coefficient of AA and AB.ResultsBasic characteristics of the study subjectsThe characteristics of the study sample are in Table 1. Mean BMI, WC, WHtR, systolic and diastolic blood pressures, FPG, HDL-C, LDL-C, TG, and TC were higher among men than women. The age and 2-hr plasma glucose of OGTT were similar between women and men. The prevalence of metabolic syndrome and hypertension was higher in men than women, that of diabetes and dyslipidemia was similar between men and women (Table 1), and the prevalence increased in the high-risk age groups.Anthropometric variables and cardiovascular risk conditions The AUC cutoff values are in Table 2 for men and women. For men, the optimal cutoffs for BMI associated with hypertension, diabetes, dyslipidemia, and metabolic syndrome ranged from 24.5 to 25 kg/m2, for WC from 87.5 to 89.5 cm, and for WHtR from 0.52 to 0.53. For women, the optimal cutoffs for BMI varied from 24.5 to 25 kg/m2, for WC from 82.5 to 83.5 cm, and for WHtR from 0.52 to 0.53.ROC curves for the discriminating hypertension, diabetes, dyslipidemia, and metabolism syndrome by BMI, WC, and WHtR for males and females are in Figures 1-8) and their associations is in Table 3.For women, regarding diabetes and dyslipidemia, the AUC values for WHtR were significantly higher than for WC and BMI; regarding hypertension and metabolism syndrome, the AUC values for WHtR and WC were were similar, and both were higher than for BMI.For men, regarding diabetes and metabolic syndrome, the AUC values for WHtR were significantly higher than for WC and BMI; regarding dyslipidemia, the AUC values for WHtR and WC were similar, and both were higher than for BMI. However, regarding hypertension, the AUC value for BMI was significantly higher than for WHtR. In the high-risk age groups, there were significant differences for dyslipidemia for females, and for diabetes and metabolic syndrome for males.The AUC values for BMI, WC, and WHtR were all higher for women than for men for all risk factors except dyslipidemia; however, the AUC values for WC and WHtR were significently higher for women than men only for hypertension and metabolic syndrome.Conclusions A BMI of 24.5 kg/m2 for both men and women, a WC of 88.5 cm for men and 83.5 cm for women, and a WHtR of 0.52 for both men and women were found optimal cutoffs for defining overweight and central adiposity in this population. As compared with BMI, measures of central obesity, particularly WHtR, show a better association with obesity-related cardiovascular risk conditions for both sexes, except for hypertension in males, in Shandong, China.