• Order
  • Offers
  • Support
    • We are closed on Saturday 2nd July for customer service training. We're still working hard in the background to ensure confirmed Saturday deliveries are honoured.

      June 28, 2022

  • Sign In

Disclaimer: This is an example of a student written essay.
Click here for sample essays written by our professional writers.

Any scientific information contained within this essay should not be treated as fact, this content is to be used for educational purposes only and may contain factual inaccuracies or be out of date.

Clinical Utility of 30-Min Plasma Glucose and Oral Glucose Tolerance for Predicting Type 2 Diabetes

Paper Type: Free Essay Subject: Biology
Wordcount: 5677 words Published: 18th May 2020

Reference this



To examine the clinical utility of the 30-min plasma glucose (30-min PG) measurement during an oral glucose tolerance (OGTT) in predicting type 2 diabetes (T2DM).

Research Design and Methods

Data from a 3-year, randomized, controlled, community-based primary prevention trial among 578 Asian Indians with prediabetes were analyzed. Participants underwent OGTT with plasma glucose measurements at fasting, 30-min, and 2-h at baseline and annually until the end of the study. Cox regression models were constructed to calculate the risk of developing diabetes based on 30-min-PG levels. Improvement in prediction performance gained by adding an elevated level of 30-min PG to IFG and IFG + IGT models were calculated using the net reclassification (NRI), and the integrated discrimination improvement (IDI).


At the end of follow-up, 167 (30.4%) individuals had been diagnosed with T2DM by ADA criteria. 30-min-PG was associated with incident diabetes (adjusted hazard ratio (aHR): 1.96 [95%CI:1.38-2.78]) per SD. Based on the maximally selected log-rank statistics, the optimal 30-min PG cut point for predicting incident T2DM was 182 mg/dl. Subgroup analyses revealed that a trend toward a clinical association of elevated 30-min PG levels with T2DM incidence among the obese, individuals with a positive family history of diabetes, and in individuals allocated to the standard care group. The addition of elevated 30-min-PG to the IFG and IGT model significantly improved the prediction of incident diabetes (NRI: 0.51 [0.33- 0.69]).


30-min-PG is independently associated with T2DM incidence. The addition of 30-min PG significantly improved the prediction in models including IFG or IGT. Therefore, 30-min PG should be considered as part of the routine testing in addition to FPG and 2-h PG for better risk stratification.


Accurate identification of individuals at heightened risk of developing type 2 diabetes (T2DM) is pivotal for the prevention of both T2DM and its associated complications. Impaired fasting glucose (IFG; 100–125 mg/dl (5.6-7.0 mmol/l) and impaired glucose tolerance (IGT; 2 h-PG140–199 mg/dl (7.8-11.0 mmol/l)])(1), both of which can appear in isolation (i-IFG or i-IGT) or in combination (combined glucose intolerance (CGI)) during the oral glucose tolerance test (OGTT), are well-recognized markers of elevated risk of future T2DM. However, prospective epidemiological studies demonstrate the limitations of IFG and IGT in predicting risk, as ~30% of patients with prediabetes eventually convert to diabetes over 5-7 years (2). Therefore, in addition to prediabetes, other markers may be required to accurately identify those at highest risk for developing T2DM.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!

Essay Writing Service

Prospective epidemiological studies have consistently demonstrated that intermediary glucose measures, or 1-h time points during a standard oral glucose tolerance test (OGTT)(3; 4), and OGTT derived features such as glucose curve shapes (5), glucose peak time and size, or the number of peaks (6), are associated with heightened risk of incident diabetes, cardiovascular and overall mortality. However, only a few studies (7; 8) have examined the utility of 30-min plasma glucose (30-min-PG) on incident diabetes. The deranged 30-min PG response reflects inadequate first-phase insulin response and is the earliest detectable defect of pancreatic β-cell function in individuals destined to develop T2DM (9; 10). Therefore, assessing the intermediary glucose measurements at 30-minutes may have added benefit for identifying high-risk individuals. It is especially true for South Asians, as they exhibit long-term pancreatic β-cell compensation for chronic insulin resistance from childhood resulting in an inability to produce further β-cell compensation in response to compounding insulin resistance in their later life (11; 12). However, the practical implications of using 30-min PG for prediction of T2DM are less clear. Therefore, the aims of the present study were to 1)  assess the predictive role of 30-min-PG for progression of T2DM in individuals with prediabetes, and 2) examine the added predictive benefit of 30-min-PG on top of IFG or IGT.


Study Participants

Eligibility criteria, study methods, and primary results of the Diabetes Community Lifestyle Improvement Program (D-CLIP) have been reported in detail elsewhere (13; 14). Briefly, D-CLIP was a prospective, parallel-group, randomized controlled trial in participants with prediabetes in south-east India recruited between September 2009 and February 2012. The primary cohort consisted of 578 overweight/obese Asian Indians with prediabetes. They were randomized into two study groups: group-1 received standard care advice at baseline (n=295); group-2 received a step-up diabetes prevention program (n=283) which included six months of group-based, culturally-tailored lifestyle education classes plus metformin for participants who remained at highest risk of converting to diabetes at four months or later. The intervention classes followed a structured educational curriculum with 16 weekly active period classes followed by eight weeks of maintenance classes. The curriculum was adapted from the U.S. Diabetes Prevention Program (DPP) (15) and designed to reduce diabetes incidence through weight-loss (≥7% weight-loss), 150 minutes or more of moderate activity weekly, and reducing intakes of fat (to <30% of total energy) and total calories. Intervention group participants were prescribed a low dose of metformin (dose: 500 mg- twice daily) by the primary care physicians if individuals remained with or progressed to CGI or i-IFG + HbA1c > 5.7% at four months or later.

 For this study, we included 549 participants with complete OGTT data. The individuals were followed-up at 6-monthly intervals for the duration of the study (total of 36 months). At the annual follow-up, a standard OGTT was carried out with blood sampling at three intervals (fasting, 30 min, and 120 min). During the interim visits (4, 6, 18, 30 months), a venous fasting plasma glucose (FPG) was taken to minimize the study participants’ discomfort. The primary outcome was incident diabetes. Ascertainment of diabetes was based on the ADA guidelines: plasma glucose of a value of >126 mg/dl (>7.0 mmol/l) in the fasting state or >200 mg/dl (>11.1 mmol/l) 2-h  after a 75-g oral glucose load (1). The study showed that expert recommendations of adding low-dose of metformin in a stepwise manner in addition to a structured lifestyle education was an effective method for delaying incident diabetes among overweight Asian Indians with prediabetes (intervention (25.7%) vs. standard-care advice (34.9%): relative risk reduction (RR): 32% [95%CI: 7-50]).

Physical and Analytical determinations

At each visit, anthropometric, hemodynamic, and lifestyle measures were performed by trained personnel under standardized conditions, following WHO recommendations (16). The participants wore light clothes for all these measurements. Body weight was measured to the nearest 0.5 kg and height to the nearest 0.1 cm. Body mass index (BMI; kg/m2) was computed. Waist circumference was measured midway between the lower rib and the iliac crest using a measuring tape to the nearest 0.1 cm (16).

The OGTT with venous blood sampling at three intervals (fasting, and at 30 min, and 2-h after 75 g anhydrous glucose) was done at baseline and annual visits. Plasma glucose (sodium fluoride as a preservative, hexokinase method, a coefficient of variation (CV)<3% at 180 mg/dl and 360 mg/dl) was measured at each visit.  Serum fasting triglycerides (GPO‐PAP method, CV <2.0%; detection range: 4.42–1000 mg/dl), total cholesterol (CHOD‐PAP method, CV <2.0%; detection range: 3.1–801 mg/dl) and HDL cholesterol (HDL plus‐third generation; CV <3.0%; detection range: 3.1–120 mg/dl) were measured annually with an appropriate quality controls in an automated autoanalyzer. Plasma insulin estimation was carried out on the Elecsys® platform (CV <3%; detection range: 1.39–6,945 pmol/L; ELICA; Roche Diagnostics, Germany).


The Institutional Review Boards (IRB-00016503) of Emory University (Atlanta, USA) and the Madras Diabetes Research Foundation (Chennai, India) reviewed and approved the study procedures and materials. Written informed consent was obtained from each participant before the screening, baseline testing, and randomization. The trial was registered on clinicaltrials.gov (NCT01283308; last updated: October 31, 2016).

Statistical analyses

Categorical and continuous variables stratified by the 30-min-PG groups are described as proportions (percentages) and the median (interquartile range), respectively. Because the primary objective of this post hoc analysis was to assess the predictive power of 30-min PG on incident diabetes, we considered both control and intervention groups as a single cohort for this analysis. Unadjusted and adjusted (age, sex, parental history of diabetes, and study allocation group) differences in baseline characteristics were estimated with quantile regression models. In addition to that, we also stratified the analyses based on 30-min PG values in individuals with different prediabetes categories. Multivariate Cox regression models with a natural cubic spline analysis were used to determine the potential nonlinear association between baseline 30-min PG levels (as a continuous variable) and the risk of incident T2DM. Follow-up time was calculated as the time from study entry until diabetes diagnosis or last examination. The proportional hazard assumption was tested for all predictors and covariates in a multivariate model, using the Schoenfeld residuals regressed against follow-up time; no violation of proportionality was observed. The “surv_cutpoint” command of the “survminer”(17), R package was used to split 30-min PG concentrations into high- or low-level groups based on the maximally selected log-rank statistics (maxstat). Moreover, we set the “minprop” parameter of the “surv_cutpoint” function (referring to the minimal proportion of observations per group) to 30% to reduce the occurrence of too few individuals in a certain group.

 Hazard ratios (HRs) and 95% CIs for the relationship between 30-min PG categories and the incidence of T2DM were generated with Cox proportional hazards models. Model-1 was adjusted for baseline age, gender, and allocation group; Model 2 was adjusted for the variables in Model-1 + parental history of diabetes, and baseline body mass index (BMI); Model-3 was adjusted for the baseline systolic blood pressure (SBP), HDL cholesterol, and triglycerides concentrations in addition to the variables in Model-2; Model-4 was adjusted for the variables in Model 3 plus baseline levels of FPG and 2-h-PG. Since the glycemic variables are highly correlated, we tested the collinearity for each of the covariates in the Cox models using the collinearity diagnostics. We observed no evidence of multicollinearity between covariates for any of the fitted models (variance inflation factor < 2 for all independent variables).

 The ability of 30-min PG to enhance prognostication in addition to IFG and IGT models was tested with the deviance analysis, the area under the receiver operating characteristic (ROC) curve, net reclassification index (NRI), and integrated discrimination improvement (IDI).  For this analysis, the following models were tested: model-1) traditional diabetes risk factor model containing IFG, age, gender, parental history of diabetes, HDL cholesterol, triglycerides, BMI, and SBP; model-2) model-1 and IGT; model-3) model-1 and 30-min PG or model-4) model-2+ 30-min PG. Enhancements in the predictive performance of model 2, 3, or 4 were compared with model 1 (IFG model) and tested using continuous NRI. The continuous NRI is advantageous as that it does not depend on the choice of specific risk categories and any change in predicted risk in the correct direction is considered appropriate. The goodness-of-fit was assessed by the Hosmer-Lemeshow chi-square test. All estimates are reported with 95% CI. All analyses and visualizations were conducted in R version 3.5.1 (R foundation of statistical computing, Vienna, Austria) using the TableOne (version: 0.10.0) (18), survival (version: 2.37-7) (19), survminer (version: 0.3.1) (17), and predictABEL (version: 1.2-2) packages.


Baseline Characteristics of Study Population

The mean age was 44.6 years, with a moderate representation of females (n=206, 37.5%) and a mean BMI of 27.9 kg/m2 at baseline. Glucose levels at 30-min and FPG (r= 0.408) were moderately correlated, whereas there was a weaker association between 30-min PG values and 2-hr postload glucose (r= 0.126). Table-1 depicts demographic, anthropometric, and metabolic characteristics for the entire cohort stratified by the presence or absence of incident diabetes after follow-up. Participants who progressed to T2DM had a higher prevalence of a positive family history of diabetes, higher BMI, waist circumference, diastolic blood pressure, fasting, 30-min, and 2-h PG, and increased fasting plasma insulin. Moreover, the levels of HDL-cholesterol and 30-min plasma insulin levels were lower in T2DM group compared with those who did not progress to diabetes.


Higher30-min-PG levels are Associated with Increased Type 2 Diabetes Incidence


During a median follow-up time of 2.5 (IQR: 2.1-3.1) years, 167 (30.4%) individuals with prediabetes at baseline developed type 2 diabetes. Table-2 present Cox proportional hazards models testing the effects of baseline 30-min PG groups hazard ratios by intervention group. Based on natural cubic splines analysis, there was no evident deviation from linearity of baseline 30-min-PG concentrations in the Cox’s proportional hazards model (P=0.27; supplemental figure-1). Overall, higher min-30-PG levels at baseline were associated with an increase in incident diabetes (adjusted hazard ratio (aHR): 1.21 per 1 SD increase [95%CI: 1.02 -1.44]). Based on the maximally selected rank statistics function, the optimal cut point for 30-min-PG to predicts incident diabetes was 182 mg/dl (supplemental figure-2).

 Subsequently,  in cumulative event curves and Cox proportional hazards models with 30-min-PG< 182 mg/dl as a reference category, the cumulative event probabilities for incident diabetes were higher in the min-30-PG>182 mg/dl group compared with 30-min PG< 182 mg/dl group (Fig-1).The median time to develop diabetes was 4.2 years in the 30-min-PG< 182 mg/dl group compared with 3.5 years in the 30-min-PG>182 mg/dl group. We noted the impact of 30-min-PG>182 mg/dl group on the incidence of diabetes was early in the study (at 6 months), compared with 30-min-PG< 182 mg/dl showing that their effect was rapid. As shown in Table 2, compared with the reference group, individuals with 30-min-PG concentrations >182 mg/dl had 85% higher risks of experiencing T2DM (aHR: 1.85 [95% CI: 1.32, 2.59]) after adjustment for baseline age, gender, parental history of diabetes, allocation group, BMI, HDL-cholesterol, TG levels, SBP, FPG, and 2-h PG levels (model 4).

Subgroup analysis

The robustness of the association between 30-min-PG and incident diabetes was tested in analyses stratified by different subgroups. For this analysis, we divided participants into two groups based on min-30-PG cutoff values (above and less than or equal to 182 mg/dl). We did not observe any differences in the demographic, anthropometric, and clinical characteristics between the two groups except FPG values. Associations of elevated 30-min-PG levels with incident diabetes were stronger among obese patients (BMI>27.0 kg/m2; aHR: 2.07 [95%CI: 1.37-3.12]), presence of the family history of diabetes (aHR: 2.30 [1.49-3.55], in the standard care group (aHR: 1.89 [95%CI: 1.20-2.98]), and the presence of combined glucose intolerance (aHR: 2.40 [95%CI: 1.56-3.70]) and iIGT (aHR: 2.36 [95%CI: 1.11-5.02]) at baseline (Fig-2).

Predictive performance of 30-min-PG values in addition to FPG, and 2-h PG

The model including 30-min PG (deviance =636.61) showed a significantly improved model fit compared with the traditional diabetes risk factor modeling including IFG (deviance = 654.2) (difference in deviance: -17.6, P < 0.0001). Importantly, the addition of 30-min PG to the extended risk model including IFG and IGT (deviance =625.6) significantly improved the model fit (deviance =610.3; the difference in deviance: -15.4; P<0.0001). Furthermore, the addition of min-30-PG to the IFG and IGT model significantly improves the area under the ROC curve for the incidence of diabetes (Table-3; ESM Fig 3-4), and re-classification (NRI: 0.51 [0.33- 0.69]; p-value: <0.0001; IDI: 0.080 [0.058 – 0.103]; p-value: <0.0001). Moreover, the Hosmer-Lemeshow goodness-of fit test showed adequate calibration for all the models (all >P 0.05; ESM-Fig:5).


There are two significant findings in this analysis. Our study showed that 30-min-PG is a reliable and independent predictor of incident diabetes in individuals with prediabetes. Besides, when 30-min-PG was added together with IFG or IGT, it improves the risk prediction of prediabetes progression to T2DM.

Although large scale epidemiological studies showed an association between 1-h PG and diabetes (4), only a handful of studies showed an association between 30-min-PG and incidence of diabetes (8; 20). Abdul-Ghani et al. (8)demonstrated that both elevated 30-min-PG and 1-h PG measurements were better predictors of future diabetes than fasting glucose in the San Antonio study. Additionally, a randomized controlled study among Asian Indians showed that 30-min-PG in upper tertile was independently increased risk of developing diabetes in 3 years compared with lower tertile (20). Our findings, however, contribute to new knowledge by identifying the optimal cut-point for 30-min PG in predicting incident diabetes. Specifically, the results from our study suggest that elevated 30-min PG of greater than 182 mg/dl during an OGTT may identify a subgroup of individuals at increased risk for developing T2DM, likely due to decreased insulin sensitivity and pancreatic beta-cell dysfunction. The sub-group analysis revealed that the association of 30-min-PG with incident diabetes were stronger among obese patients, in individuals with a positive family history of diabetes, in the standard care group, and the presence of combined glucose intolerance and iIGT groups at baseline.

A better understanding of the etiology and pathophysiology of the prediabetic states might give a basis for the development of individualized prevention and treatment strategies for T2DM. Previously, using the retrospective samples from a real-world clinical setting, the STOP DIABETES study (21) showed the treatment benefit of 1-h-PG among individuals with NGT at baseline. As the peak glucose absorption occurs mostly at 30–60 min after ingesting a mixed meal, this period potentially represents an optimal period to detect the earliest evidence of metabolic dysfunction. Hence, glucose measurements at 30-min-PG provide an additional clinical advantage in further stratifying high-risk individuals even when applied as a filter within traditional glycaemic categories. Further investigations are warranted to validate our findings and explore the potential mechanisms.

Secondly, to our knowledge, this is the first study to explore the clinical utility of 30-min plasma glucose with incident diabetes in South Asians with prediabetes. Our finding demonstrates that the all the indices of discrimination and reclassification (i.e., deviance, ROC, continuous NRI and IDI) suggested a significant and valuable gain in model fit and classification accuracy in when 30-min-PG was added to a risk model including IFG and IGT in predicting diabetes. In clinical practice, neither IFG nor IGT are recognized clinical entities per se but rather risk categories for T2DM and cardiovascular diseases progression (22). Our study shows for the first time, the ability of elevated 30-min-PG to separate high-risk from low-risk prediabetes categories. Although the potential contribution of 30 min-PG was previously appreciated, as shown by the inclusion of 30-min-PG in the 1979 National Diabetes Data Group criteria for classifying IGT, it was later deemed redundant in criteria set out by the WHO, which included 2 h-plasma glucose as the only post-challenge time point required for IGT classification. As only one-third of the individuals with the current prediabetic criteria eventually convert to diabetes over 5-7 years, the additional glucose measurement at 30-minutes may proffer a basis for the further stratification of prediabetic individuals who are at extreme risk of developing diabetes.

This study has some limitations, which in our opinion, does not critically influence the outcomes or their interpretation. The OGTT glucose at each time point was carried out once, which may have limited reproducibility, albeit, represents real-world clinical setting. As the D-CLIP study only included overweight or obese South Asians; the findings of this study may not apply to those individuals with dysglycemia and healthy BMI.  Although we controlled for the group allocation (control vs. intervention) in all the models, we did not additionally adjust for metformin, since majority of the individuals allocated to the intervention prescribed metformin (72%), teasing out its effects if hard. Furthermore, it should not alter the relationship between 30-min PG and incidence of diabetes. In summary, our data suggest that the 30-min-PG may represent an index of metabolic impairment, and useful in clinical practice to identify individuals at high risk of developing T2DM. Predictive utility of glycemic thresholds at intermediary time points other than the traditional glycemic measures (fasting, 2-h plasma glucose) values should, therefore, be considered in clinical settings.


  1. care ADAJD. 2. Classification and diagnosis of diabetes: standards of medical care in diabetes—2018. 2018;41:S13-S27
  2. Unwin N, Shaw J, Zimmet P, Alberti KG. Impaired glucose tolerance and impaired fasting glycaemia: the current status on definition and intervention. Diabetic medicine : a journal of the British Diabetic Association 2002;19:708-723
  3. Pareek M, Bhatt DL, Nielsen ML, Jagannathan R, Eriksson K-F, Nilsson PM, Bergman M, Olsen MH. Enhanced Predictive Capability of a 1-Hour Oral Glucose Tolerance Test: A Prospective Population-Based Cohort Study. 2018;41:171-177
  4. Bergman M, Manco M, Sesti G, Dankner R, Pareek M, Jagannathan R, Chetrit A, Abdul-Ghani M, Buysschaert M, Olsen MH, Nilsson PM, Medina JL, Roth J, Groop L, Del Prato S, Raz I, Ceriello A. Petition to replace current OGTT criteria for diagnosing prediabetes with the 1-hour post-load plasma glucose>/=155mg/dl (8.6mmol/L). Diabetes research and clinical practice 2018;146:18-33
  5. Abdul‐Ghani MA, Lyssenko V, Tuomi T, DeFronzo RA, Groop LJDmr, reviews. The shape of plasma glucose concentration curve during OGTT predicts future risk of type 2 diabetes. 2010;26:280-286
  6. 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;58:87-97
  7. Chamukuttan S, Ram J, Nanditha A, Shetty AS, Sevick MA, Bergman M, Johnston DG, Ramachandran A. Baseline level of 30-min plasma glucose is an independent predictor of incident diabetes among Asian Indians: analysis of two diabetes prevention programmes. Diabetes/metabolism research and reviews 2016;32:762-767
  8. 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;31:1650-1655
  9. Luzi L, DeFronzo RA. Effect of loss of first-phase insulin secretion on hepatic glucose production and tissue glucose disposal in humans. The American journal of physiology 1989;257:E241-246
  10. Del Prato S, Marchetti P, Bonadonna RC. Phasic insulin release and metabolic regulation in type 2 diabetes. Diabetes 2002;51 Suppl 1:S109-116
  11. Unnikrishnan R, Anjana RM, Mohan VJD. Diabetes in South Asians: is the phenotype different? 2014;63:53-55
  12. Hulman A, Simmons RK, Brunner EJ, Witte DR, Faerch K, Vistisen D, Ikehara S, Kivimaki M, Tabak AG. Trajectories of glycaemia, insulin sensitivity and insulin secretion in South Asian and white individuals before diagnosis of type 2 diabetes: a longitudinal analysis from the Whitehall II cohort study. Diabetologia 2017;60:1252-1260
  13. Weber MB, Ranjani H, Meyers GC, Mohan V, Narayan KM. A model of translational research for diabetes prevention in low and middle-income countries: The Diabetes Community Lifestyle Improvement Program (D-CLIP) trial. Prim Care Diabetes 2012;6:3-9
  14. Weber MB, Ranjani H, Staimez LR, Anjana RM, Ali MK, Narayan KM, Mohan V. The Stepwise Approach to Diabetes Prevention: Results From the D-CLIP Randomized Controlled Trial. Diabetes Care 2016;39:1760-1767
  15. Knowler WC, Barrett-Connor E, Fowler SE, Hamman RF, Lachin JM, Walker EA, Nathan DM. Reduction in the incidence of type 2 diabetes with lifestyle intervention or metformin. N Engl J Med 2002;346:393-403
  16. Physical status: the use and interpretation of anthropometry. Report of a WHO Expert Committee. World Health Organization technical report series 1995;854:1-452
  17. Kassambara A, Kosinski M, Biecek P, Fabian S. survminer: Drawing Survival Curves using’ggplot2′. R package version 0.3. 1. 2017
  18. Yoshida K. Package ‘tableone’. 2019;
  19. Therneau TM, Lumley T. Package ‘survival’.
  20. Chamukuttan S, Ram J, Nanditha A, Shetty AS, Sevick MA, Bergman M, Johnston DG, Ramachandran AJDmr, reviews. Baseline level of 30‐min plasma glucose is an independent predictor of incident diabetes among Asian Indians: analysis of two diabetes prevention programmes. 2016;32:762-767
  21. Armato JP, DeFronzo RA, Abdul-Ghani M, Ruby RJJTLD, Endocrinology. Successful treatment of prediabetes in clinical practice using physiological assessment (STOP DIABETES). 2018;6:781-789
  22. Tabák AG, Herder C, Rathmann W, Brunner EJ, Kivimäki M. Prediabetes: a high-risk state for diabetes development. Lancet (London, England) 2012;379:2279-2290

Table 1: Baseline Demographics, anthropometric, and metabolic characteristics of participants based on diabetes status during follow-up 

For the continuous variables, the median (Q1–Q3) is presented in the first row followed by the median (95% CI) adjusted (age, gender, and allocation group) differences in the second row. Categorical variables are expressed as n (%).

Table-2: Cox Proportional Hazard model for incident type 2 diabetes, stratified according to elevated 30-minutes plasma glucose levels

30-min PG as a continuous variable

30-min PG < 182 mg/dl

n= 355

30-min PG > 182 mg/dl

n= 193

Incidence of Type 2 Diabetes (n, %)

167 (30.4)

84 (23.7)

83 (43.0)


1.42 [1.21, 1.67]


2.18 [1.60, 2.97]


1.43 [1.22, 1.68]


2.19 [1.60, 2.98]


1.42 [1.21, 1.67]


2.18 [1.60, 2.98]


1.23 [1.03, 1.46]


1.85 [1.32, 2.59]

All the participants were prediabetes at baseline. Incidence of Type 2 Diabetes presented as number of events (percentage) during the follow-up. Model-1 was adjusted for baseline age, gender, and allocation group; Model 2 was adjusted for the variables in Model-1 + parental history of diabetes, and baseline body mass index; Model-3 was adjusted for the baseline systolic blood pressure, HDL cholesterol and triglycerides concentrations in addition to the variables in Model-2; Model-4 was adjusted for the variables in Model 3 plus baseline levels of fasting plasma glucose, 2-h postload glucose, and HbA1c levels.

Table-3:  Performance of adding 2-h PG and 30-min PG concentrations to the FPG PG for the prediction of incident diabetes


Net reclassification improvement – continuous

Integrated discrimination improvement



P value

Mean (95%CI)

P value

Mean (95%CI)

P value

Model: 1 (IFG)

0.62 [0.58-0.68]





Model: 2 (IFG and IGT)

0.69 [0.64-0.74]



0.48 [0.31 – 0.65]


0.056 [ 0.037 – 0.075]


Model: 3 (IFG and 30-min PG > 182mg/dl)

0.66 [0.61-0.71]



0.40 [0.22 – 0.578]


0.029 [0.015 – 0.042]


Model: 4 (IFG+30-min-PG+IGT)

0.71 [0.66-0.76]



0.51 [0.33- 0.69]


0.080 [0.058 – 0.103]


Model-1: includes baseline model (age, sex, parental history of diabetes, allocation group, systolic blood pressure and total cholesterol) with IFG

Model-2: Baseline model + IFG + IGT

Model-3: Baseline model + IFG + 30-min-PG > 182 mg/dl

Model-4: Baseline model + FPG + 30-min-PG > 182 mg/dl+ IGT.

Model performances were tested using logistic regression, where the “Model-1” was used as the reference

Fig-2: Subgroup analyses for effect modification by risk factors on risk for incident type 2 diabetes in participants with 30-min plasma glucose > 182 mg/dl as compared to those with <182 mg/dl. Hazard ratio and 95% CI were obtained from multivariable Cox regression models adjusted for age, gender, BMI, SBP, TG, HDL, FPG, and 2-h PG.  Subgroup variable was excluded from the model. CI, confidence interval.


Cite This Work

To export a reference to this article please select a referencing stye below:

Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.
Reference Copied to Clipboard.

Related Services

View all

DMCA / Removal Request

If you are the original writer of this essay and no longer wish to have your work published on UKEssays.com then please: