One-hour glucose value as a long-term predictor of cardiovascular morbidity and mortality: the Malmö Preventive Project

in European Journal of Endocrinology
View More View Less
  • 1 Department of Endocrinology, Cardiovascular and Metabolic Preventive Clinic, Centre for Individualized Medicine in Arterial Diseases (CIMA), Odense University Hospital, Odense, Denmark
  • | 2 Cardiology Section, Department of Internal Medicine, Holbaek Hospital, Holbaek, Denmark
  • | 3 Department of Cardiology, Skåne University Hospital, Malmö, Sweden
  • | 4 Department of Clinical Sciences, Vascular Diseases, Lund University, Malmö, Sweden
  • | 5 Department of Clinical Sciences, Lund University, Skåne University Hospital, Malmö, Sweden

Free access

Objective

To examine the predictive capability of a 1-h vs 2-h postload glucose value for cardiovascular morbidity and mortality.

Design

Prospective, population-based cohort study (Malmö Preventive Project) with subject inclusion 1974–1992.

Methods

4934 men without known diabetes and cardiovascular disease, who had blood glucose (BG) measured at 0, 20, 40, 60, 90 and 120 min during an OGTT (30 g glucose per m2 body surface area), were followed for 27 years. Data on cardiovascular events and death were obtained through national and local registries. Predictive capabilities of fasting BG (FBG) and glucose values obtained during OGTT alone and added to a clinical prediction model comprising traditional cardiovascular risk factors were assessed using Harrell’s concordance index (C-index) and integrated discrimination improvement (IDI).

Results

Median age was 48 (25th–75th percentile: 48–49) years and mean FBG 4.6 ± 0.6 mmol/L. FBG and 2-h postload BG did not independently predict cardiovascular events or death. Conversely, 1-h postload BG predicted cardiovascular morbidity and mortality and remained an independent predictor of cardiovascular death (HR: 1.09, 95% CI: 1.01–1.17, P = 0.02) and all-cause mortality (HR: 1.10, 95% CI: 1.05–1.16, P < 0.0001) after adjusting for various traditional risk factors. Clinical risk factors with added 1-h postload BG performed better than clinical risk factors alone, in predicting cardiovascular death (likelihood-ratio test, P = 0.02) and all-cause mortality (likelihood-ratio test, P = 0.0001; significant IDI, P = 0.0003).

Conclusion

Among men without known diabetes, addition of 1-h BG, but not FBG or 2-h BG, to clinical risk factors provided incremental prognostic yield for prediction of cardiovascular death and all-cause mortality.

Abstract

Objective

To examine the predictive capability of a 1-h vs 2-h postload glucose value for cardiovascular morbidity and mortality.

Design

Prospective, population-based cohort study (Malmö Preventive Project) with subject inclusion 1974–1992.

Methods

4934 men without known diabetes and cardiovascular disease, who had blood glucose (BG) measured at 0, 20, 40, 60, 90 and 120 min during an OGTT (30 g glucose per m2 body surface area), were followed for 27 years. Data on cardiovascular events and death were obtained through national and local registries. Predictive capabilities of fasting BG (FBG) and glucose values obtained during OGTT alone and added to a clinical prediction model comprising traditional cardiovascular risk factors were assessed using Harrell’s concordance index (C-index) and integrated discrimination improvement (IDI).

Results

Median age was 48 (25th–75th percentile: 48–49) years and mean FBG 4.6 ± 0.6 mmol/L. FBG and 2-h postload BG did not independently predict cardiovascular events or death. Conversely, 1-h postload BG predicted cardiovascular morbidity and mortality and remained an independent predictor of cardiovascular death (HR: 1.09, 95% CI: 1.01–1.17, P = 0.02) and all-cause mortality (HR: 1.10, 95% CI: 1.05–1.16, P < 0.0001) after adjusting for various traditional risk factors. Clinical risk factors with added 1-h postload BG performed better than clinical risk factors alone, in predicting cardiovascular death (likelihood-ratio test, P = 0.02) and all-cause mortality (likelihood-ratio test, P = 0.0001; significant IDI, P = 0.0003).

Conclusion

Among men without known diabetes, addition of 1-h BG, but not FBG or 2-h BG, to clinical risk factors provided incremental prognostic yield for prediction of cardiovascular death and all-cause mortality.

Introduction

The strong influence of type 2 diabetes on the risk of cardiovascular morbidity and mortality is undisputed (1, 2), and effective measures to predict and prevent both type 2 diabetes and related cardiovascular complications are desirable.

Fasting blood glucose (FBG) alone is not sufficient in predicting hyperglycemia-related mortality, while the standard 2-h postload glucose value enables the detection of subjects with impaired glucose tolerance (IGT), who have a significantly increased risk of death (3). Furthermore, in overweight subjects with IGT, intensive lifestyle intervention and pharmacological treatment can prevent progression into overt type 2 diabetes and possibly ameliorate the risk of related cardiovascular complications (4, 5, 6). However, the 2-h OGTT has low sensitivity, has non-optimal specificity for incident diabetes and is time-consuming, and thus it is considered to carry limited cost-effectiveness and practicality (7). The past decade has seen several studies suggesting the superiority of a 1-h postload glucose value vs a 2-h postload glucose value for the prediction of type 2 diabetes (7, 8, 9, 10, 11, 12, 13, 14). Previous data from the Malmö Preventive Project have also shown that individuals at high risk for future type 2 diabetes can be identified using multivariable prediction models that include both FBG and blood glucose (BG) obtained during OGTT, with a significantly better predictive capability achieved for shorter OGTT regimens compared with both FBG and the classical 2-h OGTT (15). This might extend to cardiovascular morbidity and hyperglycemia-related mortality. Therefore, the primary purposes of the present study were (1) to evaluate the prognostic value of FBG and BG values obtained at different time points during OGTT, in predicting cardiovascular morbidity and mortality, both isolated and in addition to a multivariable prediction model; (2) to assess whether BG measured at 60 min during OGTT (BG60) provided better prognostic yield than FBG and BG measured at 120 min (BG120) and (3) to assess whether the prognostic value of BG60 was influenced by glucose tolerance.

Subjects and methods

Study population

The Malmö Preventive Project (1974–1992, n = 33 346) was a population-based cohort study conducted among inhabitants in Malmö, Sweden, belonging to pre-specified birth cohorts between 1921 and 1949 (16). Progressively older subjects were recruited later during the inclusion process. In all, 22 444 males and 10 902 females attended the screening program, with an overall attendance rate of 71% (75% for men). All subjects answered a self-administered questionnaire on lifestyle, family history of diabetes (first degree relatives) and cardiovascular disease, medical history and current medications. Current smoking was defined as smoking at least one cigarette daily, and sedentary lifestyle was defined as leisure time mostly spent on sedentary activities. Prevalent diabetes was defined as self-reported diabetes or according to the 1985 World Health Organization (WHO) criteria (17, 18). Prevalent cardiovascular disease was defined as self-reported cardiovascular disease (history of myocardial infarction (MI), stroke or transient ischemic attack) or according to national and local registries using the same diagnostic codes as for outcome (19). Height and weight in light indoor clothing were measured, from which body mass index (BMI) was calculated. Blood pressure was measured twice in the right arm after 10 min of supine rest, with the mean value recorded for analysis. Blood samples were obtained after ≥10-h overnight fasting with measurement of BG, serum lipids and serum creatinine. BG levels were measured in capillary whole blood from fingertip samples and analyzed using the glucose-oxidase method (1974–1977) or the hexokinase method (1977–1992) (20). Serum samples were analyzed using the laboratory’s standard methods. The Malmö Preventive Project was conducted in accordance with the Declaration of Helsinki.

Outcome

Follow-up time for each subject was defined as time elapsed from baseline screening to date of myocardial infarction, stroke, death, emigration or the last follow-up date up to 27 years. Cardiovascular events were recorded by national and local registries and comprised International Classification of Diseases (ICD-9) codes 410–414, 434, 430–438. Diagnosis of type 2 diabetes was recorded by 14 national and local registries and comprised International Classification of Diseases (ICD-10) codes E11.0–E11.9. The method to ascertain diagnosis of type 2 diabetes and cardiovascular events in the Swedish Inpatient Register has been previously validated (21). Positive predictive values were moderate for stroke and high for all other endpoints. Mortality follow-up was based on the national registry on causes of mortality at the Central Bureau of Statistics, Sweden.

Final study population

In participants without known diabetes, a 120-min OGTT was performed by ingestion of 30 g glucose per m2 body surface area (DuBois equation) in a 10% aqueous solution within 5 min. BG levels were determined at 0, 20, 40, 60, 90 and 120 min. Subjects without an OGTT with extra BG measurements (n = 27 728) were excluded from the present study, as were the 132 females of whom most did not have BG measured at the relevant intermediate time points. Remaining subjects with missing baseline variables (n = 70) or incomplete follow-up due to emigration (n = 70) were likewise excluded. Of the 5346 male subjects left, 111 had prevalent diabetes and 301 had previous cardiovascular disease or received cardiovascular medication (antihypertensives, antithrombotics, antiarrhythmics), leaving a final study population comprising 4934 males (Fig. 1). Of note, 315 subjects with IGT (defined locally as BG120 ≥7.0 mmol/L) who underwent lifestyle intervention (dietary advice and increased physical activity, most often including frequent visits at an outpatient clinic) for up to 12 years upon confirmation of their IGT status were not excluded (22).

Figure 1
Figure 1

Flowchart showing the study population selection.

Citation: European Journal of Endocrinology 178, 3; 10.1530/EJE-17-0824

Statistical analysis

Continuous variables were summarized by means and standard deviations (approximately normally distributed variables) or medians and 25th and 75th percentiles (non-normally distributed variables). Categorical variables were presented by frequencies and corresponding percentages. Group-wise comparisons were performed using one-way analysis of variance (ANOVA), the Kruskal–Wallis test or Pearson’s χ 2-test, as appropriate. In order to define potential explanatory variables for the endpoints, i.e., incident myocardial infarction (nonfatal and fatal), stroke (nonfatal and fatal), death from cardiovascular causes and all-cause mortality, univariable Cox proportional-hazards regressions were applied on the following demographic and clinical variables: age, smoking status, BMI, systolic blood pressure, diastolic blood pressure, total cholesterol, triglycerides, creatinine, family history of diabetes and sedentary lifestyle. Since the distributions of serum triglycerides and creatinine were moderately positively skewed, they were both natural log-transformed for the regression analyses. Statistically significant and clinically relevant variables were included in the final multivariable Cox models, and stepwise backward elimination using the likelihood-ratio test was applied for adjustment of these models. The final regression model (clinical prediction model) included variables that were statistically significant in the multivariable models for at least one of the four endpoints. For BG, standardized values were calculated as the absolute value of FBG, BG60 or BG120 for each subject divided by each of their standard deviations. Furthermore, we used interaction analyses to test whether glucose (in)tolerance influenced the prognostic value of BG60. This was done by including the interaction term for the continuous variable BG60 and the categorical variable, absence/presence of IGT, in the relevant regression models. The predictive abilities both alone and in addition to the clinical prediction model were tested with Harrell’s concordance index (C-index) (23). The ability of each BG measurement during OGTT to enhance prognostication was also evaluated using integrated discrimination improvement (IDI) (24). Follow-up time was 27 years for all subjects without events, and supplementary analyses were done at shorter follow-up times (5, 10 and 20 years) to ensure that the results were independent of follow-up time. Analyses were focused on BG measured at 0, 60 and 120 min during OGTT because of the overall better predictive capacity of BG60 compared with BG measured at 20 and 40 min. BG measured at 90 min was comparable with BG60, but since this measurement required 30 min extra, it seemed reasonable to exclude it (15). The significance level was 0.05 (two-sided), and given the exploratory nature of this study, no adjustments for multiple comparisons were made. All analyses were carried out using IBM SPSS Statistics 23 (IBM) and Stata/IC 14 (StataCorp LP).

Results

Characteristics of the study population

At baseline, median age was 48 (25th–75th percentile: 48–49) years, mean systolic blood pressure 130 ± 15 mmHg and mean BMI 24.8 ± 3.2 kg/m2. Total cholesterol was 5.8 ± 1.0 mmol/L. Subjects with IGT at baseline (n = 301) generally had a worse risk profile than their counterparts with NGT (Table 1). Similarly, a more adverse risk profile was present across BG60 tertiles (Table 1). During follow-up, 1381 composite cardiovascular events (myocardial infarction, stroke or cardiovascular death, whichever came first) were detected, corresponding to 12.3 cases per 1000 person-years. A total of 1517 individuals had died from any cause. IGT and the uppermost BG60 tertile appeared to be associated with poorer outcomes compared with NGT and the lower BG60 tertiles, respectively (Table 2).

Table 1

Baseline characteristics according to glucose tolerance status and BG60 tertile group at baseline, respectively. Categorical variables are presented as n (%), whereas continuous, approximately normally distributed variables are presented as mean ± s.d., and continuous, non-normally distributed variables are presented as median (IQR).

VariableAll subjects (n = 4934)NGT (n = 4633)IGT (n = 301)P valuesBG60 tertile 1 (n = 1641)BG60 tertile 2 (n = 1648)BG60 tertile 3 (n = 1645)P values
Age (years)48 (48–49)48 (48–49)48 (48–49)0.448 (48–49)48 (48–49)48 (48–49)<0.0001||
Active smoking2612 (53%)2487 (54%)125 (42%)<0.0001*803 (49%)863 (52%)946 (58%)<0.0001*
BMI (kg/m2)24.8 ± 3.224.7 ± 3.126.1 ± 3.5<0.000124.2 ± 2.924.8 ± 3.225.2 ± 3.4<0.0001§
Systolic BP (mmHg)130 ± 15129 ± 15140 ± 19<0.0001126 ± 13129 ± 15135 ± 17<0.0001§
TC (mmol/L)5.8 ± 1.05.8 ± 1.06.0 ± 1.10.0015.7 ± 1.05.8 ± 1.05.8 ± 1.0<0.0001§
Triglycerides (mmol/L)1.4 (1.0–1.9)1.4 (1.0–1.8)1.7 (1.2–2.3)<0.00011.3 (1.0–1.6)1.4 (1.1–1.8)1.5 (1.1–2.1)<0.0001||
Creatinine (pmol/L)94 (86–102)94 (86–102)91 (85–101)0.295 (87–103)94 (86–102)93 (85–101)0.001||
FBG (mmol/L)4.6 ± 0.64.6 ± 0.55.0 ± 0.6<0.00014.4 ± 0.54.6 ± 0.54.9 ± 0.6<0.0001§
60 min PLBG (mmol/L)8.1 ± 2.17.9 ± 2.010.6 ± 1.9<0.00016.0 ± 0.87.9 ± 0.510.4 ± 1.5<0.0001§
120 min PLBG (mmol/L)5.4 ± 1.55.1 ± 1.28.7 ± 0.8<0.00014.8 ± 1.15.2 ± 1.26.1 ± 1.7<0.0001§
Sedentary lifestyle2797 (57%)2605 (56%)192 (64%)0.01*881 (54%)924 (56%)992 (60%)0.001*

*Pearson’s 2-test; independent samples t-test; Mann–Whitney U test; §one-way ANOVA; ||Kruskal–Wallis test.

BP, blood pressure; FBG, fasting blood glucose; PLBG, post load blood glucose; BG60, blood glucose measured at 60 min during OGTT; IGT, impaired glucose tolerance; NGT, normal glucose tolerance.

Table 2

Impact of glucose tolerance status and BG60 tertile group on number of events and event rates per 1000 person-years.

Total (n = 4934)Normal glucose tolerance (NGT) (n = 4633)Impaired glucose tolerance (IGT) (n = 301)P values for difference between NGT and IGTBG60 tertile 1 (n = 1641)BG60 tertile 2 (n = 1648)BG60 tertile 3 (n = 1645)P values for difference between BG60 tertiles
All-cause mortality1517 (12.7)1407 (12.6)110 (15.1)0.02*418 (10.5)494 (12.4)605 (15.2)<0.0001*
Cardiovascular death638 (5.4)584 (5.2)54 (7.4)0.008*156 (3.9)221 (5.6)261 (6.6)<0.0001*
Myocardial infarction (nonfatal and fatal)763 (6.7)712 (6.7)51 (7.3)0.5*213 (5.6)253 (6.7)297 (7.8)<0.0001*
Stroke (nonfatal and fatal)455 (3.9)432 (3.9)23 (3.2)0.3*130 (3.3)151 (3.9)174 (4.5)0.03*
Nonfatal myocardial infarction592 (5.2)552 (5.2)40 (5.8)0.5*177 (4.7)188 (4.9)227 (6.0)0.02*
Fatal myocardial infarction171 (1.4)160 (1.4)11 (1.5)0.9*36 (0.9)65 (1.6)70 (1.8)0.002*
Nonfatal stroke419 (3.6)398 (3.6)21 (2.9)0.3*121 (3.1)138 (3.5)160 (4.1)0.052*
Fatal stroke36 (0.3)34 (0.3)2 (0.3)0.9*9 (0.2)13 (0.3)14 (0.4)0.6*

*Pearson’s 2-test.

BG60, blood glucose measured at 60 min during OGTT; IGT, impaired glucose tolerance; NGT, normal glucose tolerance.

The clinical prediction model

The following variables were statistically significant on univariable analysis for prediction of incident nonfatal plus fatal myocardial infarction and nonfatal plus fatal stroke: age, active smoking, BMI, systolic blood pressure, total cholesterol and triglycerides. Sedentary lifestyle and creatinine were significant on univariable analysis for prediction of cardiovascular death and all-cause mortality. The final Cox regression model (clinical prediction model) included age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. This model performed better than anyone isolated BG measurement for the prediction of all endpoints, with C-index for the former being highest for cardiovascular death (C-index = 0.682, 95% confidence interval (CI): 0.66–0.70), followed by myocardial infarction (C-index = 0.666, 95% CI: 0.65–0.69), all-cause mortality (C-index = 0.638, 95% CI: 0.62–0.65) and stroke (C-index = 0.612, 95% CI: 0.59–0.63).

Fasting blood glucose and postload glucose values

At 27 years, BG60 alone was a significant predictor of fatal myocardial infarction (hazard ratio (HR): 1.22, 95% CI: 1.09–1.36, P = 0.0006), nonfatal myocardial infarction (HR: 1.10, 95% CI: 1.02–1.18, P = 0.01), nonfatal plus fatal myocardial infarction (HR: 1.13, 95% CI: 1.06–1.20, P = 0.0001), nonfatal plus fatal stroke (HR: 1.14, 95% CI: 1.06–1.24, P = 0.0008), nonfatal stroke (HR: 1.14, 95% CI: 1.05–1.24, P = 0.002), death from cardiovascular causes (HR: 1.21, 95% CI: 1.14–1.28, P < 0.0001) and all-cause mortality (HR: 1.18, 95% CI: 1.14–1.23, P < 0.0001). BG120 and FBG alone did not predict any of the endpoints (Table 3). At 20 years of follow-up, BG120 alone predicted cardiovascular death (HR: 1.13, 95% CI: 1.02–1.25, P = 0.02), whereas FBG still did not predict any of the endpoints (P ≥ 0.21 for all), and BG60 remained superior to both (P < 0.0001).

Table 3

Hazard ratios for different standardized glucose values at 27-years of follow-up, unadjusted, age-adjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine.

All-cause mortality (n = 1517)Cardiovascular death (n = 638)Nonfatal and fatal MI (n = 763)Nonfatal and fatal stroke (n = 455)
HR (95% CI)P valueHR (95% CI)P valueHR (95% CI)P valueHR (95% CI)P value
Unadjusted
 FBG1.01 (0.96–1.06)NS1.01 (0.94–1.10)NS1.04 (0.97–1.12)NS1.01 (0.92–1.11)NS
 BG601.18 (1.14–1.23)<0.00011.21 (1.14–1.28)<0.00011.13 (1.06–1.20)0.00011.14 (1.06–1.24)0.0008
 BG1200.97 (0.92–1.02)NS1.04 (0.96–1.13)NS1.03 (0.96–1.10)NS0.96 (0.87–1.05)NS
Age adjusted
 FBG1.05 (1.00–1.11)NS1.08(1.00–1.16)NS1.07 (1.00–1.15)NS1.05 (0.96–1.15)NS
 BG601.18 (1.14–1.23)<0.00011.21 (1.14–1.28)<0.00011.13 (1.06–1.20)0.00021.14 (1.06–1.24)0.0008
 BG1200.97 (0.92–1.02)NS1.04 (0.96–1.13)NS1.03 (0.96–1.10)NS0.96 (0.87–1.05)NS
Fully adjusted
 FBG1.00 (0.94–1.05)NS0.98 (0.90–1.06)NS0.97 (0.90–1.05)NS0.98 (0.89–1.08)NS
 BG601.10 (1.05–1.16)<0.00011.09 (1.01–1.17)0.021.0 (0.93–1.07)NS1.06 (0.97–1.16)NS
 BG1200.97 (0.92–1.03)NS0.98 (0.91–1.07)NS0.97 (0.90–1.05)NS0.89 (0.80–0.98)0.02
Fully adjusted and further adjusted for FBG
 BG601.11 (1.06–1.17)<0.00011.10 (1.03–1.19)0.0091.01 (0.93–1.09)NS1.07 (0.98–1.18)NS
 BG1200.97 (0.92–1.03)NS0.99 (0.91–1.08)NS0.97 (0.90–1.05)NS0.89 (0.80–0.98)0.02

FBG, BG60 and BG120 are standardized using their standard deviations: 0.6, 2.1 and 1.5 mmol/L, respectively.

BG60, blood glucose measured at 60 min during OGTT; BG120, blood glucose measured at 120 min during OGTT; FBG, fasting blood glucose; MI, myocardial infarction.

Fatal myocardial infarction was detected in 171 subjects, and fatal stroke was detected in 36 subjects. Accordingly, fatal myocardial infarction contributed the most to cardiovascular death. C-index for BG60 alone was consistently greater than that for FBG or BG120 (Table 4); however, significant comparisons were only observed for all-cause mortality, cardiovascular death, nonfatal plus fatal myocardial infarction and nonfatal plus fatal stroke (P ≤ 0.02; P > 0.05 for the comparison with FBG (myocardial infarction) and for the comparison with BG120 (stroke)). The presence of IGT at baseline significantly interacted with the association between BG60 and all-cause mortality (P = 0.001), cardiovascular death (P = 0.0004) and nonfatal plus fatal stroke (P = 0.03), with higher predictive values of BG60 found among subjects with IGT at baseline (Fig. 2A, B, C and D). When BG60 and BG120 were both analyzed as continuous variables, BG120 modified the association between BG60 and myocardial infarction (P = 0.03).

Figure 2
Figure 2

(A) Prediction of all-cause mortality by blood glucose obtained at 60-min OGTT in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance. (B) Prediction of cardiovascular mortality by blood glucose obtained at 60 min during OGTT in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance. (C). Prediction of myocardial infarction (nonfatal + fatal) by blood glucose obtained at 60 min during OGTT test in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance. (D) Prediction of stroke (nonfatal + fatal) by blood glucose obtained at 60 min during OGTT in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance.

Citation: European Journal of Endocrinology 178, 3; 10.1530/EJE-17-0824

Table 4

Harrell’s C-indices for different glucose values at 27-years of follow-up alone and in addition to the clinical prediction model, which included age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine.

All-cause mortalityCardiovascular deathNonfatal and fatal MINonfatal and fatal stroke
Alone
 FBG0.5030.5030.5150.497
 BG600.5570.5670.5430.546
 BG1200.5150.5030.5060.516
  Pa< 0.001<0.001a0.060.02
  Pb0.002<0.0010.020.3
Together with the clinical prediction model
 FBG0.6380.6820.6670.612
 BG600.6410.6850.6660.613
 BG1200.6380.6820.6660.616
 FBG + BG600.6420.6850.6670.613
 FBG + BG1200.6380.6820.6660.616
 FBG + BG60 + BG1200.6430.6850.6660.618

P values are for comparisons of Harrell’s C-indices for BG60 vs Harrell’s C-indices for aFBG and bBG120, respectively.

BG60, blood glucose measured at 60 min during OGTT; BG120, blood glucose measured at 120 min during OGTT; FBG, fasting blood glucose; MI, myocardial infarction.

The intervention, carried out in a minority of the study subjects, did not alter the estimates and did not interact with the association between any BG measurement and any endpoint (P > 0.05 for all). Furthermore, incident type 2 diabetes during follow-up did not interact with any association between BG60 and the endpoints (P ≥ 0.50 for all).

Fasting blood glucose and oral glucose tolerance test in combination with the clinical prediction model

After adjusting for the variables included in the clinical prediction model, neither FBG nor BG120 predicted any of the endpoints. However, BG60 remained independently predictive of all-cause mortality (adjusted HR: 1.10, 95% CI: 1.05–1.16, P < 0.0001) and cardiovascular death (adjusted HR: 1.09, 95% CI: 1.01–1.17, P = 0.02) (Table 3). The predictive value of the fully adjusted BG60 model was also highest in subjects with IGT at baseline (Fig. 2A, B, C and D). Adjusting for FBG did not reduce the predictive value of BG60 (Table 3).

Adding BG60 to the clinical prediction model significantly improved the prediction of all-cause mortality (likelihood-ratio test; P = 0.0001) and cardiovascular death (likelihood-ratio test; P = 0.02), although with no significant C-index increment. Addition of FBG or BG120 did not provide model improvement (likelihood-ratio test; P ≥ 0.21 for all). BG60 was further associated with significant IDI in predicting all-cause mortality (P = 0.0003). Conversely, BG120 was associated with a significant IDI in predicting stroke (P = 0.02), but the HR for BG120 was <1 in the fully adjusted prediction model.

The overall results did not change using shorter follow-up (Table 5).

Table 5

Hazard ratios and Harrell’s C-indices for different standardized glucose values at 5, 10 and 20 years of follow-up, unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides, creatinine and fasting blood glucose.

All-cause mortality (n = 810 at 20 years)Cardiovascular death (n = 356 at 20 years)Nonfatal and fatal MI (n = 516 at 20 years)Nonfatal and fatal stroke (n = 198 at 20 years)
HR (95% CI)Harrell’s C-indexP-ValueHR (95% CI)Harrell’s C-indexP-ValueHR (95% CI)Harrell’s C-indexP-ValueHR (95% CI)Harrell’s C-indexP-Value
5 years of follow-up
 Unadjusted
  FBG1.04 (0.85–1.26)0.507NS1.08 (0.77–1.51)0.514NS1.03 (0.83–1.29)0.502NS0.93 (0.48–1.80)0.505NS
  BG601.26 (1.10–1.44)0.5800.0011.26 (1.00–1.59)0.574NS1.29 (1.12–1.48)0.609<0.0011.26 (0.80–1.98)0.585NS
  BG1200.95 (0.77–1.16)0.524NS1.10 (0.80–1.53)0.517NS1.19 (0.97–1.47)0.544NS1.17 (0.63–2.18)0.567NS
 Fully adjusted
  BG601.17 (0.99–1.39)0.693NS1.11 (0.79–1.56)0.777NS1.18 (0.98–1.42)0.756NS1.21 (0.70–2.07)0.713NS
  BG1200.92 (0.74–1.15)0.689NS0.99 (0.69–1.41)0.772NS1.15 (0.92–1.45)0.753NS1.15 (0.57–2.31)0.709NS
10 years of follow-up
 Unadjusted
  FBG1.11 (0.98–1.25)0.521NS1.22 (1.01–1.47)0.5470.041.08 (0.94–1.23)0.522NS0.73 (0.51–1.05)0.589NS
  BG601.24 (1.13–1.35)0.570<0.0011.29 (1.14–1.45)0.591<0.0011.17 (1.05–1.31)0.5570.0051.30 (1.05–1.61)0.5930.02
  BG1201.05 (0.93–1.18)0.499NS1.16 (0.96–1.39)0.520NS1.05 (0.92–1.20)0.515NS0.99 (0.69–1.41)0.502NS
 Fully adjusted
  BG601.13 (1.01–1.26)0.6670.031.13 (0.95–1.36)0.715NS1.01 (0.87–1.16)0.725NS1.28 (1.02–1.61)0.7950.03
  BG1201.02 (0.89–1.16)0.664NS1.03 (0.85–1.26)0.711NS0.99 (0.85–1.14)0.726NS1.10 (0.76–1.59)0.780NS
20 years of follow-up
 Unadjusted
  FBG1.04 (0.98–1.12)0.510NS1.07 (0.96–1.19)0.515NS1.05 (0.96–1.14)0.514NS1.02 (0.88–1.17)0.497NS
  BG601.21 (1.14–1.27)0.566<0.0011.24 (1.15–1.33)0.580<0.0011.12 (1.03–1.21)0.5390.0061.15 (1.02–1.29)0.5470.03
  BG1200.98 (0.91–1.05)0.515NS1.13 (1.02–1.25)0.5200.021.03 (0.95–1.13)0.505NS0.92 (0.79–1.06)0.530NS
Fully adjusted
  BG601.12 (1.05–1.19)0.656<0.0011.12 (1.01–1.24)0.6990.030.97 (0.88–1.07)0.690NS1.04 (0.90–1.21)0.662NS
  BG1200.96 (0.89–1.03)0.652NS1.03 (0.93–1.16)0.696NS0.97 (0.88–1.06)0.690NS0.85 (0.72–0.99)0.6620.04

FBG, BG60 and BG120 are standardized using their standard deviations: 0.6, 2.1 and 1.5 mmol/L, respectively.

BG60, blood glucose measured at 60 min during OGTT; BG120, blood glucose measured at 120 min during OGTT; FBG, fasting blood glucose; MI, myocardial infarction.

Discussion

Our results show that among apparently healthy, middle-aged men without known diabetes, BG measured at 60 min during OGTT provides better prognostic yield than both FBG and BG measured at 120 min, in predicting long-term cardiovascular morbidity and mortality, both isolated and in addition to a clinical prediction model.

The conventional 2-h postload glucose value is known to be a significant, independent and better predictor of cardiovascular morbidity and mortality than fasting glucose (3, 25, 26, 27, 28). We found a significant prediction of cardiovascular death with BG120, although only unadjusted, and no significant predictive value of FBG. However, the association between BG120 and cardiovascular death was only evident on supplementary analyses with a shorter follow-up time. The predictive capacity might have been diluted over time, given the old age, co-morbidity status and age-related events at the time of study completion. An additional explanation may be our relatively healthy (from a glucometabolic perspective) cohort, in which both fasting and 2-h glucose values generally were low and rarely reached the criteria for prediabetes. In contrast, other studies have largely evaluated persons with higher glucose values closer to or even fulfilling contemporary prediabetes criteria. The interventional program of lifestyle changes in most subjects with baseline IGT decreased mortality at 12 years (22), but there was no effect on long-term outcome (29). Furthermore, the local threshold for baseline IGT was rather low (7.0 rather than 7.8 mmol/L). This is likely to have resulted in a rather large number of subjects at moderate-to-high risk being captured. Still, excluding all subjects with IGT would have lowered the variability in observed BG levels substantially and excluded a large number of individuals at high risk for clinical events.

Despite the above, BG60 significantly predicted myocardial infarction, stroke, cardiovascular death and all-cause mortality, and after adjustment for other cardiovascular risk factors, BG60 still predicted cardiovascular death and all-cause mortality. Lending support to our results, Orencia and coworkers (30) demonstrated that 1-h postload glucose was an independent risk factor for all-cause mortality at long-term (22 years) follow-up in middle-aged men without diabetes. Higher postload glucose levels were also independent risk factors for the major cardiovascular diseases. However, there was no direct comparison with fasting or 2-h glucose measurements. This was also the case for a 44-year prospective study, in which a strong and graded relationship was found between 1-h glucose quartiles and both total and cardiovascular mortality, independently of traditional cardiovascular risk factors, in men without diabetes (31). In a third study by Meisinger and coworkers (32), it was likewise established that a higher 1-h postload glucose value was an independent long-term (up to 30 years) predictor of all-cause mortality. However, the investigators used a very high partition value (200 mg/dL = 11.1 mmol/L) and did not provide a comparison with fasting or 2-h glucose values. Data from the Finnish Diabetes Prevention Study showed that among persons with IGT, both 1-h and 2-h plasma glucose values and their temporal patterns were significantly associated with cardiovascular events (myocardial infarction, stroke, unstable angina, coronary artery bypass graft or percutaneous coronary intervention), although only the 2-h glucose remained significant in pairwise comparisons (33). Concerning reclassification abilities, Bergman and coworkers (7) found that a 1-h glucose value >155 mg/dL (>8.6 mmol/L) predicted mortality among individuals in whom the 2-h level was <140 mg/dL (<7.8 mmol/L). Finally, an elaborate study by Hulman and coworkers identified four distinct glucose patterns during an OGTT with glucose measurements at 0, 30 and 120 min. Among these healthy participants without diabetes at baseline, the risks of incident diabetes and of all-cause mortality were significantly increased with an elevated glucose level at 30 min, risks that were independent of glucose levels at 0 and 120 min (34). The same group of authors was also involved in another study of glucose response curves during OGTT, in which they observed a variety of responses despite similar fasting and 2-h glucose values. These responses differed with respect to their associations with cardiovascular and metabolic risk factors. This in and of itself suggests the usefulness of an intermediate, i.e., between 0 and 120 min, measurement of postload glucose (35).

Influence of impaired glucose tolerance at baseline

The interactions with IGT and the association between BG60 and all-cause mortality, cardiovascular death and stroke, may suggest that elevated BG60 does not merely precede elevated BG120, as a smaller effect of increased BG60 in subjects with increased BG120 would then be expected. Possibly, having both abnormal BG60 and BG120 increases the possibility of correct categorization. Still, the interactions for the continuous BG variables suggest other explanations. One very interesting hypothesis could be that high BG60 is particularly harmful, from a cardiovascular perspective, if allowed to stay elevated for at least 60 additional minutes. Accordingly, these interactions may reflect a dissociation between prediction of type 2 diabetes and cardiovascular events, as both BG60 and BG120 appear to be markers of prediabetes, while the mechanisms by which BG60 and BG120 predict cardiovascular events seemingly differ (33). Finally, it must be emphasized that the intervention program of lifestyle changes may have influenced the results, although given the interaction between BG60 and BG120, this does not seem to be the case at long-term follow-up.

Postload glucose values and clinical risk factors

Addition of BG60 to traditional cardiovascular risk factors significantly improved the predictive capability, and there was also significant IDI in the prediction of all-cause mortality, when adding BG60 to these risk factors. Conversely, Stern and coworkers (36) did not find a significant model improvement after addition of the standard 2-h postload glucose measurement to a clinical prediction model. The investigators thus suggested that prediction models should only include readily available clinical variables. Obviously, this is attractive because of the ease of administration, convenience, acceptability to screened subjects and lower cost. However, since the burden of type 2 diabetes and complications is rapidly rising, we believe that subjects at high risk of developing IGT and type 2 diabetes could be identified earlier using BG60. It is very likely that in subjects with an abnormal BG60, lifestyle intervention may prevent progression to IGT, type 2 diabetes, cardiovascular events and ultimately death.

Influence of development of type 2 diabetes during follow-up

Another question that appears is whether such elevated cardiovascular risk is restricted to individuals who develop diabetes during follow-up, but in the present study, incident type 2 diabetes did not significantly interact with the association between BG60 and incident cardiovascular events or death. Only few studies have explored this because of the absence of follow-up data on diabetes diagnosis; however, in accordance with our findings, Qiao and coworkers (37) demonstrated that baseline IGT, although defined by the standard 2-h OGTT, was an independent predictor for 10-year cardiovascular morbidity and mortality as well as total mortality, not confounded by the subsequent development of diabetes.

Limitations

A comparison between HbA1c and glucose values obtained during OGTT would have been desirable, but such measurements were not done. OGTT was performed only at baseline and therefore subjected to regression dilution bias (38). Furthermore, at the time of inclusion, a standard OGTT comprised a 30 g/m2 glucose load which, however, seems to yield results similar to the standard 75 g load (39). Our exclusion of a significant proportion of the original study population, although based on financial reasons (lack of extra BG measurements), constitutes a major limitation and prevents us from extending our results beyond middle-aged white men. Additionally, the lifestyle intervention performed in subjects with IGT at baseline may potentially have influenced the predictive capability of the 2-h glucose measurement, even though the follow-up data and analyses did not indicate this in neither the present nor in a previous study (29). In fact, excluding these subjects at the highest likelihood of developing type 2 diabetes and as such, the outcomes examined in this study, might have underestimated the true risk of complications. Despite significant differences in several comparisons between the prediction models, the discriminative values and their numerical differences were of limited magnitude. This may suggest limited clinical relevance. However, follow-up time was very long, which attenuated the discriminative ability of all models as previously shown (15). Furthermore, it should be remembered that prediction models generally tend to overestimate the actual event risk when applied to the population from which they were derived (40). Lastly, different between-study results might further be due to variations in sex ratio, ethnicity, sample sizes, inclusion criteria, co-variables included, glucose load used for OGTT, follow-up duration and definition of outcome variables.

Clinical perspectives

Shorter examination time is associated with lower cost and greater convenience for both patients and health care professionals. Although the introduction of HbA1c as a diagnostic tool for diabetes has led to performance of fewer OGTTs, it is well known that these methods display limited concordance (41). Therefore, a role for OGTT likely exists in subjects with a high normal or an only slightly elevated HbA1c, in order to reach a definitive diagnostic decision. With an increasing number of studies demonstrating a strong superior capacity compared with FBG and the 2-h glucose value in identifying subjects with a high risk of developing type 2 diabetes and related complications, it seems reasonable to consider the 1-h glucose value in order to benefit from early prevention in subjects with high risk of future type 2 diabetes and cardiovascular complications. Still, our findings need replication and external validation. In addition, interventional studies are necessary to prove the benefit of secondary prevention in high-risk individuals found through non-traditional methods.

In conclusion, in this prospective population-based cohort study, we found an OGTT-derived 60-min BG provided better prognostic value than FBG or 120-min postload BG, in predicting long-term cardiovascular morbidity and mortality, in apparently healthy, middle-aged, men without known diabetes.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of this study.

Funding

This study was funded by a research grant from The Heart and Lung Foundation of Sweden and The Danish Diabetes Academy supported by the Novo Nordisk Foundation.

Author contribution statement

Mette Nielsen contributed substantially to the conception of the hypothesis, the design of the work, analysis of the work, interpretation of data for the work, drafting the work, critical revision for important intellectual content, final approval of the version to be published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Manan Pareek contributed substantially to the conception of the hypothesis, the design of the work, analysis of the work, critical revision for important intellectual content, final approval of the version to be published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Margrét Leósdóttir, Karl-Fredrik Eriksson and Peter M Nilsson contributed substantially to the acquisition of data for the work, critical revision for important intellectual content, the final approval of the version to be published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. Michael Olsen contributed substantially to the conception of the hypothesis, the design of the work, analysis of the work, interpretation of data for the work, drafting the work, critical revision for important intellectual content, final approval of the version to be published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

References

  • 1

    Kannel WB & McGee DL. Diabetes and cardiovascular disease. The Framingham study. JAMA 1979 241 20352038. (https://doi.org/10.1001/jama.1979.03290450033020)

  • 2

    Stamler J, Vaccaro O, Neaton JD & Wentworth D. Diabetes, other risk factors, and 12-year cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial. Diabetes Care 1993 16 434444. (https://doi.org/10.2337/diacare.16.2.434)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3

    Glucose tolerance and mortality: comparison of WHO and American Diabetes Association diagnostic criteria. The DECODE study group. European Diabetes Epidemiology Group. Diabetes epidemiology: collaborative analysis of diagnostic criteria in Europe. Lancet 1999 354 617621.

    • Search Google Scholar
    • Export Citation
  • 4

    Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM & Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. New England Journal of Medicine 2002 346 393403. (https://doi.org/10.1056/NEJMoa012512)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P, Keinänen-Kiukaanniemi S, Laakso M, Louheranta A, Rastas M, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. New England Journal of Medicine 2001 344 13431350. (https://doi.org/10.1056/NEJM200105033441801)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6

    Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M & STOP-NIDDM Trail Research Group. Acarbose for prevention of type 2 diabetes mellitus: the STOP-NIDDM randomised trial. Lancet 2002 359 20722077. (https://doi.org/10.1016/S0140-6736(02)08905-5)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Bergman M, Chetrit A, Roth J & Dankner R. One-hour post-load plasma glucose level during the OGTT predicts mortality: observations from the Israel Study of Glucose Intolerance, Obesity and Hypertension. Diabetic Medicine 2016 33 10601066.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    Abdul-Ghani MA, Williams K, DeFronzo RA & Stern M. What is the best predictor of future type 2 diabetes? Diabetes Care 2007 30 15441548. (https://doi.org/10.2337/dbib6-1331)

  • 9

    Dankner R, Abdul-Ghani MA, Gerber Y, Chetrit A, Wainstein J & Raz I. Predicting the 20-year diabetes incidence rate. Diabetes/Metabolism Research and Reviews 2007 23 551558. (https://doi.org/10.1002/dmrr.728)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    Abdul-Ghani MA, Abdul-Ghani T, Ali N & Defronzo RA. One-hour plasma glucose concentration and the metabolic syndrome identify subjects at high risk for future type 2 diabetes. Diabetes Care 2008 31 16501655. (https://doi.org/10.2337/dbib8-0225)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Abdul-Ghani MA, Lyssenko V, Tuomi T, DeFronzo RA & Groop L. Fasting versus postload plasma glucose concentration and the risk for future type 2 diabetes: results from the Botnia Study. Diabetes Care 2009 32 281286. (https://doi.org/10.2337/dbib8-1264)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Priya M, Anjana RM, Chiwanga FS, Gokulakrishnan K, Deepa M & Mohan V. 1-hour venous plasma glucose and incident prediabetes and diabetes in Asian Indians. Diabetes Technology and Therapeutics 2013 15 497502. (https://doi.org/10.1089/dia.2013.0025)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    Alyass A, Almgren P, Akerlund M, Dushoff J, Isomaa B, Nilsson P, Tuomi T, Lyssenko V, Groop L, Meyre D. Modelling of OGTT curve identifies 1 h plasma glucose level as a strong predictor of incident type 2 diabetes: results from two prospective cohorts. Diabetologia 2015 58 8797. (https://doi.org/10.1007/s00125-014-3390-x)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Fiorentino TV, Marini MA, Andreozzi F, Arturi F, Succurro E, Perticone M, Sciacqua A, Hribal ML, Perticone F, Sesti G. One-hour postload hyperglycemia is a stronger predictor of type 2 diabetes than impaired fasting glucose. Journal of Clinical Endocrinology and Metabolism 2015 100 37443751. (https://doi.org/10.1210/jc.2015-2573)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Nielsen ML, Pareek M, Leosdottir M, Hojlund K, Eriksson KF, Nilsson P, Olsen MH. Follow-up duration influences the relative importance of OGTT and optimal timing of glucose measurements for predicting future type 2 diabetes. European Journal of Endocrinology 2016 174 591600. (https://doi.org/10.1530/EJE-15-1221)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    Berglund G, Eriksson KF, Israelsson B, Kjellstrom T, Lindgarde F, Mattiasson I, Nilsson JA, Stavenow L. Cardiovascular risk groups and mortality in an urban Swedish male population: the Malmo Preventive Project. Journal of Internal Medicine 1996 239 489497. (https://doi.org/10.1046/j.1365-2796.1996.483819000.x)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Diabetes mellitus. Report of a WHO Study Group. World Health Organization Technical Report Series 1985 727 1113.

  • 18

    Eriksson KF & Lindgarde F. Impaired glucose tolerance in a middle-aged male urban population: a new approach for identifying high-risk cases. Diabetologia 1990 33 526531. (https://doi.org/10.1007/BF00404139)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    Nilsson P, Nilsson JA, Hedblad B, Eriksson KF & Berglund G. Hyperinsulinaemia as long-term predictor of death and ischaemic heart disease in nondiabetic men: the Malmo Preventive Project. Journal of Internal Medicine 2003 253 136145. (https://doi.org/10.1046/j.1365-2796.2003.01064.x)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20

    Carroll JJ, Smith N & Babson AL. A colorimetric serum glucose determination using hexokinase and glucose-6-phosphate dehydrogenase. Biochemical Medicine 1970 4 171180. (https://doi.org/10.1016/0006-2944(70)90093-1)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, Heurgren M, Olausson PO. External review and validation of the Swedish national inpatient register. BMC Public Health 2011 11 450. (https://doi.org/10.1186/1471-2458-11-450)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 22

    Eriksson KF & Lindgarde F. No excess 12-year mortality in men with impaired glucose tolerance who participated in the Malmo Preventive Trial with diet and exercise. Diabetologia 1998 41 10101016. (https://doi.org/10.1007/s001250051024)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23

    Harrell FE Jr, Califf RM, Pryor DB, Lee KL & Rosati RA. Evaluating the yield of medical tests. JAMA 1982 247 25432546. (https://doi.org/10.1001/jama.1982.03320430047030)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24

    Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Statistics in Medicine 2008 27 157172; discussion 207212. (https://doi.org/10.1002/sim.2929)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Ford ES, Zhao G & Li C. Pre-diabetes and the risk for cardiovascular disease: a systematic review of the evidence. Journal of the American College of Cardiology 2010 55 13101317. (https://doi.org/10.1016/j.jacc.2009.10.060)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26

    Decode Study Group EDEG. Is the current definition for diabetes relevant to mortality risk from all causes and cardiovascular and noncardiovascular diseases? Diabetes Care 2003 26 688696.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 27

    de Vegt F, Dekker JM, Ruhe HG, Stehouwer CD, Nijpels G, Bouter LM, Heine RJ. Hyperglycaemia is associated with all-cause and cardiovascular mortality in the Hoorn population: the Hoorn Study. Diabetologia 1999 42 926931. (https://doi.org/10.1007/s001250051249)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28

    Meigs JB, Nathan DM, D’Agostino RB Sr, Wilson PW & Framingham Offspring S. Fasting and postchallenge glycemia and cardiovascular disease risk: the Framingham Offspring Study. Diabetes Care 2002 25 18451850. (https://doi.org/10.2337/diacare.25.10.1845)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 29

    Berglund G, Nilsson P, Eriksson KF, Nilsson JA, Hedblad B, Kristenson H, Lindgärde F. Long-term outcome of the Malmo preventive project: mortality and cardiovascular morbidity. Journal of Internal Medicine 2000 247 1929. (https://doi.org/10.1046/j.1365-2796.2000.00568.x)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 30

    Orencia AJ, Daviglus ML, Dyer AR, Walsh M, Greenland P & Stamler J. One-hour postload plasma glucose and risks of fatal coronary heart disease and stroke among nondiabetic men and women: the Chicago Heart Association Detection Project in Industry (CHA) Study. Journal of Clinical Epidemiology 1997 50 13691376. (https://doi.org/10.1016/S0895-4356(97)00201-1)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 31

    Strandberg TE, Pienimaki T, Strandberg AY, Salomaa VV, Pitkala KH, Tilvis RS, Miettinen TA. One-hour glucose, mortality, and risk of diabetes: a 44-year prospective study in men. Archives of Internal Medicine 2011 171 941943. (https://doi.org/10.1001/archinternmed.2011.183)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32

    Meisinger C, Wolke G, Brasche S, Strube G & Heinrich J. Postload plasma glucose and 30-year mortality among nondiabetic middle-aged men from the general population: the ERFORT Study. Annals of Epidemiology 2006 16 534539. (https://doi.org/10.1016/j.annepidem.2005.10.008)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Lind M, Tuomilehto J, Uusitupa M, Nerman O, Eriksson J, Ilanne-Parikka P, Keinänen-Kiukaanniemi S, Peltonen M, Pivodic A, Lindström J. The association between HbA1c, fasting glucose, 1-hour glucose and 2-hour glucose during an oral glucose tolerance test and cardiovascular disease in individuals with elevated risk for diabetes. PLoS ONE 2014 9 e109506. (https://doi.org/10.1371/journal.pone.0109506)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 34

    Hulman A, Vistisen D, Glumer C, Bergman M, Witte DR & Faerch K. Glucose patterns during an oral glucose tolerance test and associations with future diabetes, cardiovascular disease and all-cause mortality rate. Diabetologia 2018 61 101107. (https://doi.org/10.1007/s00125-017-4468-z)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 35

    Hulman A, Simmons RK, Vistisen D, Tabak AG, Dekker JM, Alssema M, Rutters F, Koopman AD, Solomon TP, Kirwan JP, et al. Heterogeneity in glucose response curves during an oral glucose tolerance test and associated cardiometabolic risk. Endocrine 2017 55 427434. (https://doi.org/10.1007/s12020-016-1126-z)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 36

    Stern MP, Fatehi P, Williams K & Haffner SM. Predicting future cardiovascular disease: do we need the oral glucose tolerance test? Diabetes Care 2002 25 18511856. (https://doi.org/10.2337/diacare.25.10.1851)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 37

    Qiao Q, Jousilahti P, Eriksson J & Tuomilehto J. Predictive properties of impaired glucose tolerance for cardiovascular risk are not explained by the development of overt diabetes during follow-up. Diabetes Care 2003 26 29102914. (https://doi.org/10.2337/diacare.26.10.2910)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 38

    Clarke R, Shipley M, Lewington S, Youngman L, Collins R, Marmot M, Peto R. Underestimation of risk associations due to regression dilution in long-term follow-up of prospective studies. American Journal of Epidemiology 1999 150 341353. (https://doi.org/10.1093/oxfordjournals.aje.a010013)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 39

    Eriksson KF & Lindgarde F. Prevention of type 2 (non-insulin-dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmo feasibility study. Diabetologia 1991 34 891898. (https://doi.org/10.1007/BF00400196)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 40

    Abbasi A, Peelen LM, Corpeleijn E, van der Schouw YT, Stolk RP, Spijkerman AM, van der A DL, Moons KG, Navis G, Bakker SJ, et al. Prediction models for risk of developing type 2 diabetes: systematic literature search and independent external validation study. BMJ 2012 345 e5900. (https://doi.org/10.1136/bmj.e5900)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 41

    Collaboration NCDRF. Effects of diabetes definition on global surveillance of diabetes prevalence and diagnosis: a pooled analysis of 96 population-based studies with 331 288 participants. Lancet Diabetes and Endocrinology 2015 3 624637. (https://doi.org/10.1016/S2213-8587(15)00129-1)

    • Search Google Scholar
    • Export Citation

 

     European Society of Endocrinology

Sept 2018 onwards Past Year Past 30 Days
Abstract Views 56 0 0
Full Text Views 618 144 12
PDF Downloads 430 190 20
  • View in gallery

    Flowchart showing the study population selection.

  • View in gallery

    (A) Prediction of all-cause mortality by blood glucose obtained at 60-min OGTT in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance. (B) Prediction of cardiovascular mortality by blood glucose obtained at 60 min during OGTT in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance. (C). Prediction of myocardial infarction (nonfatal + fatal) by blood glucose obtained at 60 min during OGTT test in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance. (D) Prediction of stroke (nonfatal + fatal) by blood glucose obtained at 60 min during OGTT in subjects with and without impaired glucose tolerance at baseline both unadjusted and fully adjusted for age, active smoking, BMI, systolic blood pressure, total cholesterol, triglycerides and creatinine. Hazard ratios are calculated for absolute, unstandardized glucose values. NGT, normal glucose tolerance; IGT, impaired glucose tolerance.

  • 1

    Kannel WB & McGee DL. Diabetes and cardiovascular disease. The Framingham study. JAMA 1979 241 20352038. (https://doi.org/10.1001/jama.1979.03290450033020)

  • 2

    Stamler J, Vaccaro O, Neaton JD & Wentworth D. Diabetes, other risk factors, and 12-year cardiovascular mortality for men screened in the Multiple Risk Factor Intervention Trial. Diabetes Care 1993 16 434444. (https://doi.org/10.2337/diacare.16.2.434)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 3

    Glucose tolerance and mortality: comparison of WHO and American Diabetes Association diagnostic criteria. The DECODE study group. European Diabetes Epidemiology Group. Diabetes epidemiology: collaborative analysis of diagnostic criteria in Europe. Lancet 1999 354 617621.

    • Search Google Scholar
    • Export Citation
  • 4

    Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM & Diabetes Prevention Program Research Group. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. New England Journal of Medicine 2002 346 393403. (https://doi.org/10.1056/NEJMoa012512)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Tuomilehto J, Lindstrom J, Eriksson JG, Valle TT, Hamalainen H, Ilanne-Parikka P, Keinänen-Kiukaanniemi S, Laakso M, Louheranta A, Rastas M, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. New England Journal of Medicine 2001 344 13431350. (https://doi.org/10.1056/NEJM200105033441801)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 6

    Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M & STOP-NIDDM Trail Research Group. Acarbose for prevention of type 2 diabetes mellitus: the STOP-NIDDM randomised trial. Lancet 2002 359 20722077. (https://doi.org/10.1016/S0140-6736(02)08905-5)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 7

    Bergman M, Chetrit A, Roth J & Dankner R. One-hour post-load plasma glucose level during the OGTT predicts mortality: observations from the Israel Study of Glucose Intolerance, Obesity and Hypertension. Diabetic Medicine 2016 33 10601066.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    Abdul-Ghani MA, Williams K, DeFronzo RA & Stern M. What is the best predictor of future type 2 diabetes? Diabetes Care 2007 30 15441548. (https://doi.org/10.2337/dbib6-1331)

  • 9

    Dankner R, Abdul-Ghani MA, Gerber Y, Chetrit A, Wainstein J & Raz I. Predicting the 20-year diabetes incidence rate. Diabetes/Metabolism Research and Reviews 2007 23 551558. (https://doi.org/10.1002/dmrr.728)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 10

    Abdul-Ghani MA, Abdul-Ghani T, Ali N & Defronzo RA. One-hour plasma glucose concentration and the metabolic syndrome identify subjects at high risk for future type 2 diabetes. Diabetes Care 2008 31 16501655. (https://doi.org/10.2337/dbib8-0225)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 11

    Abdul-Ghani MA, Lyssenko V, Tuomi T, DeFronzo RA & Groop L. Fasting versus postload plasma glucose concentration and the risk for future type 2 diabetes: results from the Botnia Study. Diabetes Care 2009 32 281286. (https://doi.org/10.2337/dbib8-1264)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 12

    Priya M, Anjana RM, Chiwanga FS, Gokulakrishnan K, Deepa M & Mohan V. 1-hour venous plasma glucose and incident prediabetes and diabetes in Asian Indians. Diabetes Technology and Therapeutics 2013 15 497502. (https://doi.org/10.1089/dia.2013.0025)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    Alyass A, Almgren P, Akerlund M, Dushoff J, Isomaa B, Nilsson P, Tuomi T, Lyssenko V, Groop L, Meyre D. Modelling of OGTT curve identifies 1 h plasma glucose level as a strong predictor of incident type 2 diabetes: results from two prospective cohorts. Diabetologia 2015 58 8797. (https://doi.org/10.1007/s00125-014-3390-x)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 14

    Fiorentino TV, Marini MA, Andreozzi F, Arturi F, Succurro E, Perticone M, Sciacqua A, Hribal ML, Perticone F, Sesti G. One-hour postload hyperglycemia is a stronger predictor of type 2 diabetes than impaired fasting glucose. Journal of Clinical Endocrinology and Metabolism 2015 100 37443751. (https://doi.org/10.1210/jc.2015-2573)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 15

    Nielsen ML, Pareek M, Leosdottir M, Hojlund K, Eriksson KF, Nilsson P, Olsen MH. Follow-up duration influences the relative importance of OGTT and optimal timing of glucose measurements for predicting future type 2 diabetes. European Journal of Endocrinology 2016 174 591600. (https://doi.org/10.1530/EJE-15-1221)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 16

    Berglund G, Eriksson KF, Israelsson B, Kjellstrom T, Lindgarde F, Mattiasson I, Nilsson JA, Stavenow L. Cardiovascular risk groups and mortality in an urban Swedish male population: the Malmo Preventive Project. Journal of Internal Medicine 1996 239 489497. (https://doi.org/10.1046/j.1365-2796.1996.483819000.x)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 17

    Diabetes mellitus. Report of a WHO Study Group. World Health Organization Technical Report Series 1985 727 1113.

  • 18

    Eriksson KF & Lindgarde F. Impaired glucose tolerance in a middle-aged male urban population: a new approach for identifying high-risk cases. Diabetologia 1990 33 526531. (https://doi.org/10.1007/BF00404139)

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 19

    Nilsson P, Nilsson JA, Hedblad B, Eriksson KF & Berglund G. Hyperinsulinaemia as long-term predictor of death and ischaemic heart disease in nondiabetic men: the Malmo Preventive Project. Journal of Internal Medicine 2003 253 136145. (https://doi.org/10.1046/j.1365-2796.2003.01064.x)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 20

    Carroll JJ, Smith N & Babson AL. A colorimetric serum glucose determination using hexokinase and glucose-6-phosphate dehydrogenase. Biochemical Medicine 1970 4 171180. (https://doi.org/10.1016/0006-2944(70)90093-1)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 21

    Ludvigsson JF, Andersson E, Ekbom A, Feychting M, Kim JL, Reuterwall C, Heurgren M, Olausson PO. External review and validation of the Swedish national inpatient register. BMC Public Health 2011 11 450. (https://doi.org/10.1186/1471-2458-11-450)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 22

    Eriksson KF & Lindgarde F. No excess 12-year mortality in men with impaired glucose tolerance who participated in the Malmo Preventive Trial with diet and exercise. Diabetologia 1998 41 10101016. (https://doi.org/10.1007/s001250051024)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 23

    Harrell FE Jr, Califf RM, Pryor DB, Lee KL & Rosati RA. Evaluating the yield of medical tests. JAMA 1982 247 25432546. (https://doi.org/10.1001/jama.1982.03320430047030)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 24

    Pencina MJ, D’Agostino RB Sr, D’Agostino RB Jr, Vasan RS. Evaluating the added predictive ability of a new marker: from area under the ROC curve to reclassification and beyond. Statistics in Medicine 2008 27 157172; discussion 207212. (https://doi.org/10.1002/sim.2929)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 25

    Ford ES, Zhao G & Li C. Pre-diabetes and the risk for cardiovascular disease: a systematic review of the evidence. Journal of the American College of Cardiology 2010 55 13101317. (https://doi.org/10.1016/j.jacc.2009.10.060)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 26

    Decode Study Group EDEG. Is the current definition for diabetes relevant to mortality risk from all causes and cardiovascular and noncardiovascular diseases? Diabetes Care 2003 26 688696.

    • PubMed
    • Search Google Scholar
    • Export Citation
  • 27

    de Vegt F, Dekker JM, Ruhe HG, Stehouwer CD, Nijpels G, Bouter LM, Heine RJ. Hyperglycaemia is associated with all-cause and cardiovascular mortality in the Hoorn population: the Hoorn Study. Diabetologia 1999 42 926931. (https://doi.org/10.1007/s001250051249)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 28

    Meigs JB, Nathan DM, D’Agostino RB Sr, Wilson PW & Framingham Offspring S. Fasting and postchallenge glycemia and cardiovascular disease risk: the Framingham Offspring Study. Diabetes Care 2002 25 18451850. (https://doi.org/10.2337/diacare.25.10.1845)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 29

    Berglund G, Nilsson P, Eriksson KF, Nilsson JA, Hedblad B, Kristenson H, Lindgärde F. Long-term outcome of the Malmo preventive project: mortality and cardiovascular morbidity. Journal of Internal Medicine 2000 247 1929. (https://doi.org/10.1046/j.1365-2796.2000.00568.x)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 30

    Orencia AJ, Daviglus ML, Dyer AR, Walsh M, Greenland P & Stamler J. One-hour postload plasma glucose and risks of fatal coronary heart disease and stroke among nondiabetic men and women: the Chicago Heart Association Detection Project in Industry (CHA) Study. Journal of Clinical Epidemiology 1997 50 13691376. (https://doi.org/10.1016/S0895-4356(97)00201-1)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 31

    Strandberg TE, Pienimaki T, Strandberg AY, Salomaa VV, Pitkala KH, Tilvis RS, Miettinen TA. One-hour glucose, mortality, and risk of diabetes: a 44-year prospective study in men. Archives of Internal Medicine 2011 171 941943. (https://doi.org/10.1001/archinternmed.2011.183)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 32

    Meisinger C, Wolke G, Brasche S, Strube G & Heinrich J. Postload plasma glucose and 30-year mortality among nondiabetic middle-aged men from the general population: the ERFORT Study. Annals of Epidemiology 2006 16 534539. (https://doi.org/10.1016/j.annepidem.2005.10.008)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 33

    Lind M, Tuomilehto J, Uusitupa M, Nerman O, Eriksson J, Ilanne-Parikka P, Keinänen-Kiukaanniemi S, Peltonen M, Pivodic A, Lindström J. The association between HbA1c, fasting glucose, 1-hour glucose and 2-hour glucose during an oral glucose tolerance test and cardiovascular disease in individuals with elevated risk for diabetes. PLoS ONE 2014 9 e109506. (https://doi.org/10.1371/journal.pone.0109506)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 34

    Hulman A, Vistisen D, Glumer C, Bergman M, Witte DR & Faerch K. Glucose patterns during an oral glucose tolerance test and associations with future diabetes, cardiovascular disease and all-cause mortality rate. Diabetologia 2018 61 101107. (https://doi.org/10.1007/s00125-017-4468-z)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 35

    Hulman A, Simmons RK, Vistisen D, Tabak AG, Dekker JM, Alssema M, Rutters F, Koopman AD, Solomon TP, Kirwan JP, et al. Heterogeneity in glucose response curves during an oral glucose tolerance test and associated cardiometabolic risk. Endocrine 2017 55 427434. (https://doi.org/10.1007/s12020-016-1126-z)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 36

    Stern MP, Fatehi P, Williams K & Haffner SM. Predicting future cardiovascular disease: do we need the oral glucose tolerance test? Diabetes Care 2002 25 18511856. (https://doi.org/10.2337/diacare.25.10.1851)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 37

    Qiao Q, Jousilahti P, Eriksson J & Tuomilehto J. Predictive properties of impaired glucose tolerance for cardiovascular risk are not explained by the development of overt diabetes during follow-up. Diabetes Care 2003 26 29102914. (https://doi.org/10.2337/diacare.26.10.2910)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 38

    Clarke R, Shipley M, Lewington S, Youngman L, Collins R, Marmot M, Peto R. Underestimation of risk associations due to regression dilution in long-term follow-up of prospective studies. American Journal of Epidemiology 1999 150 341353. (https://doi.org/10.1093/oxfordjournals.aje.a010013)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 39

    Eriksson KF & Lindgarde F. Prevention of type 2 (non-insulin-dependent) diabetes mellitus by diet and physical exercise. The 6-year Malmo feasibility study. Diabetologia 1991 34 891898. (https://doi.org/10.1007/BF00400196)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 40

    Abbasi A, Peelen LM, Corpeleijn E, van der Schouw YT, Stolk RP, Spijkerman AM, van der A DL, Moons KG, Navis G, Bakker SJ, et al. Prediction models for risk of developing type 2 diabetes: systematic literature search and independent external validation study. BMJ 2012 345 e5900. (https://doi.org/10.1136/bmj.e5900)

    • Crossref
    • PubMed
    • Search Google Scholar
    • Export Citation
  • 41

    Collaboration NCDRF. Effects of diabetes definition on global surveillance of diabetes prevalence and diagnosis: a pooled analysis of 96 population-based studies with 331 288 participants. Lancet Diabetes and Endocrinology 2015 3 624637. (https://doi.org/10.1016/S2213-8587(15)00129-1)

    • Search Google Scholar
    • Export Citation