Predicting glycated hemoglobin levels in the non-diabetic general population: Development and validation of the DIRECTDETECT prediction model-a DIRECT study

Simone P. Rauh, Martijn W. Heymans, Anitra D.M. Koopman, Giel Nijpels, Coen D. Stehouwer, Barbara Thorand, Wolfgang Rathmann, Christa Meisinger, Annette Peters, Tonia De Las Heras Gala, Charlotte Glümer, Oluf Pedersen, Henna Cederberg, Johanna Kuusisto, Markku Laakso, Ewan R. Pearson, Paul W. Franks, Femke Rutters, Jacqueline M. Dekker

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Aims/hypothesis To develop a prediction model that can predict HbA1c levels after six years in the non-diabetic general population, including previously used readily available predictors. Methods Data from 5,762 initially non-diabetic subjects from three population-based cohorts (Hoorn Study, Inter99, KORA S4/F4) were combined to predict HbA1c levels at six year follow-up. Using backward selection, age, BMI, waist circumference, use of anti-hypertensive medication, current smoking and parental history of diabetes remained in sex-specific linear regression models. To minimize overfitting of coefficients, we performed internal validation using bootstrapping techniques. Explained variance, discrimination and calibration were assessed using R2, classification tables (comparing highest/lowest 50% HbA1c levels) and calibration graphs. The model was externally validated in 2,765 non-diabetic subjects of the population-based cohort METSIM. Results At baseline, mean HbA1c level was 5.6% (38 mmol/mol). After a mean follow-up of six years, mean HbA1c level was 5.7% (39 mmol/mol). Calibration graphs showed that predicted HbA1c levels were somewhat underestimated in the Inter99 cohort and overestimated in the Hoorn and KORA cohorts, indicating that the model's intercept should be adjusted for each cohort to improve predictions. Sensitivity and specificity (95% CI) were 55.7% (53.9, 57.5) and 56.9% (55.1, 58.7) respectively, for women, and 54.6% (52.7, 56.5) and 54.3% (52.4, 56.2) for men. External validation showed similar performance in the METSIM cohort. Conclusions/interpretation In the non-diabetic population, our DIRECT-DETECT prediction model, including readily available predictors, has a relatively low explained variance and moderate discriminative performance, but can help to distinguish between future highest and lowest HbA1c levels. Absolute HbA1c values are cohort-dependent.

Original languageEnglish
Article numbere0171816
JournalPLoS ONE
Volume12
Issue number2
DOIs
Publication statusPublished - 1 Feb 2017

Cite this

Rauh, Simone P. ; Heymans, Martijn W. ; Koopman, Anitra D.M. ; Nijpels, Giel ; Stehouwer, Coen D. ; Thorand, Barbara ; Rathmann, Wolfgang ; Meisinger, Christa ; Peters, Annette ; De Las Heras Gala, Tonia ; Glümer, Charlotte ; Pedersen, Oluf ; Cederberg, Henna ; Kuusisto, Johanna ; Laakso, Markku ; Pearson, Ewan R. ; Franks, Paul W. ; Rutters, Femke ; Dekker, Jacqueline M. / Predicting glycated hemoglobin levels in the non-diabetic general population : Development and validation of the DIRECTDETECT prediction model-a DIRECT study. In: PLoS ONE. 2017 ; Vol. 12, No. 2.
@article{13175f6465fe44c99e4df1dda03af0ec,
title = "Predicting glycated hemoglobin levels in the non-diabetic general population: Development and validation of the DIRECTDETECT prediction model-a DIRECT study",
abstract = "Aims/hypothesis To develop a prediction model that can predict HbA1c levels after six years in the non-diabetic general population, including previously used readily available predictors. Methods Data from 5,762 initially non-diabetic subjects from three population-based cohorts (Hoorn Study, Inter99, KORA S4/F4) were combined to predict HbA1c levels at six year follow-up. Using backward selection, age, BMI, waist circumference, use of anti-hypertensive medication, current smoking and parental history of diabetes remained in sex-specific linear regression models. To minimize overfitting of coefficients, we performed internal validation using bootstrapping techniques. Explained variance, discrimination and calibration were assessed using R2, classification tables (comparing highest/lowest 50{\%} HbA1c levels) and calibration graphs. The model was externally validated in 2,765 non-diabetic subjects of the population-based cohort METSIM. Results At baseline, mean HbA1c level was 5.6{\%} (38 mmol/mol). After a mean follow-up of six years, mean HbA1c level was 5.7{\%} (39 mmol/mol). Calibration graphs showed that predicted HbA1c levels were somewhat underestimated in the Inter99 cohort and overestimated in the Hoorn and KORA cohorts, indicating that the model's intercept should be adjusted for each cohort to improve predictions. Sensitivity and specificity (95{\%} CI) were 55.7{\%} (53.9, 57.5) and 56.9{\%} (55.1, 58.7) respectively, for women, and 54.6{\%} (52.7, 56.5) and 54.3{\%} (52.4, 56.2) for men. External validation showed similar performance in the METSIM cohort. Conclusions/interpretation In the non-diabetic population, our DIRECT-DETECT prediction model, including readily available predictors, has a relatively low explained variance and moderate discriminative performance, but can help to distinguish between future highest and lowest HbA1c levels. Absolute HbA1c values are cohort-dependent.",
author = "Rauh, {Simone P.} and Heymans, {Martijn W.} and Koopman, {Anitra D.M.} and Giel Nijpels and Stehouwer, {Coen D.} and Barbara Thorand and Wolfgang Rathmann and Christa Meisinger and Annette Peters and {De Las Heras Gala}, Tonia and Charlotte Gl{\"u}mer and Oluf Pedersen and Henna Cederberg and Johanna Kuusisto and Markku Laakso and Pearson, {Ewan R.} and Franks, {Paul W.} and Femke Rutters and Dekker, {Jacqueline M.}",
year = "2017",
month = "2",
day = "1",
doi = "10.1371/journal.pone.0171816",
language = "English",
volume = "12",
journal = "PLoS ONE",
issn = "1932-6203",
publisher = "Public Library of Science",
number = "2",

}

Rauh, SP, Heymans, MW, Koopman, ADM, Nijpels, G, Stehouwer, CD, Thorand, B, Rathmann, W, Meisinger, C, Peters, A, De Las Heras Gala, T, Glümer, C, Pedersen, O, Cederberg, H, Kuusisto, J, Laakso, M, Pearson, ER, Franks, PW, Rutters, F & Dekker, JM 2017, 'Predicting glycated hemoglobin levels in the non-diabetic general population: Development and validation of the DIRECTDETECT prediction model-a DIRECT study' PLoS ONE, vol. 12, no. 2, e0171816. https://doi.org/10.1371/journal.pone.0171816

Predicting glycated hemoglobin levels in the non-diabetic general population : Development and validation of the DIRECTDETECT prediction model-a DIRECT study. / Rauh, Simone P.; Heymans, Martijn W.; Koopman, Anitra D.M.; Nijpels, Giel; Stehouwer, Coen D.; Thorand, Barbara; Rathmann, Wolfgang; Meisinger, Christa; Peters, Annette; De Las Heras Gala, Tonia; Glümer, Charlotte; Pedersen, Oluf; Cederberg, Henna; Kuusisto, Johanna; Laakso, Markku; Pearson, Ewan R.; Franks, Paul W.; Rutters, Femke; Dekker, Jacqueline M.

In: PLoS ONE, Vol. 12, No. 2, e0171816, 01.02.2017.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Predicting glycated hemoglobin levels in the non-diabetic general population

T2 - Development and validation of the DIRECTDETECT prediction model-a DIRECT study

AU - Rauh, Simone P.

AU - Heymans, Martijn W.

AU - Koopman, Anitra D.M.

AU - Nijpels, Giel

AU - Stehouwer, Coen D.

AU - Thorand, Barbara

AU - Rathmann, Wolfgang

AU - Meisinger, Christa

AU - Peters, Annette

AU - De Las Heras Gala, Tonia

AU - Glümer, Charlotte

AU - Pedersen, Oluf

AU - Cederberg, Henna

AU - Kuusisto, Johanna

AU - Laakso, Markku

AU - Pearson, Ewan R.

AU - Franks, Paul W.

AU - Rutters, Femke

AU - Dekker, Jacqueline M.

PY - 2017/2/1

Y1 - 2017/2/1

N2 - Aims/hypothesis To develop a prediction model that can predict HbA1c levels after six years in the non-diabetic general population, including previously used readily available predictors. Methods Data from 5,762 initially non-diabetic subjects from three population-based cohorts (Hoorn Study, Inter99, KORA S4/F4) were combined to predict HbA1c levels at six year follow-up. Using backward selection, age, BMI, waist circumference, use of anti-hypertensive medication, current smoking and parental history of diabetes remained in sex-specific linear regression models. To minimize overfitting of coefficients, we performed internal validation using bootstrapping techniques. Explained variance, discrimination and calibration were assessed using R2, classification tables (comparing highest/lowest 50% HbA1c levels) and calibration graphs. The model was externally validated in 2,765 non-diabetic subjects of the population-based cohort METSIM. Results At baseline, mean HbA1c level was 5.6% (38 mmol/mol). After a mean follow-up of six years, mean HbA1c level was 5.7% (39 mmol/mol). Calibration graphs showed that predicted HbA1c levels were somewhat underestimated in the Inter99 cohort and overestimated in the Hoorn and KORA cohorts, indicating that the model's intercept should be adjusted for each cohort to improve predictions. Sensitivity and specificity (95% CI) were 55.7% (53.9, 57.5) and 56.9% (55.1, 58.7) respectively, for women, and 54.6% (52.7, 56.5) and 54.3% (52.4, 56.2) for men. External validation showed similar performance in the METSIM cohort. Conclusions/interpretation In the non-diabetic population, our DIRECT-DETECT prediction model, including readily available predictors, has a relatively low explained variance and moderate discriminative performance, but can help to distinguish between future highest and lowest HbA1c levels. Absolute HbA1c values are cohort-dependent.

AB - Aims/hypothesis To develop a prediction model that can predict HbA1c levels after six years in the non-diabetic general population, including previously used readily available predictors. Methods Data from 5,762 initially non-diabetic subjects from three population-based cohorts (Hoorn Study, Inter99, KORA S4/F4) were combined to predict HbA1c levels at six year follow-up. Using backward selection, age, BMI, waist circumference, use of anti-hypertensive medication, current smoking and parental history of diabetes remained in sex-specific linear regression models. To minimize overfitting of coefficients, we performed internal validation using bootstrapping techniques. Explained variance, discrimination and calibration were assessed using R2, classification tables (comparing highest/lowest 50% HbA1c levels) and calibration graphs. The model was externally validated in 2,765 non-diabetic subjects of the population-based cohort METSIM. Results At baseline, mean HbA1c level was 5.6% (38 mmol/mol). After a mean follow-up of six years, mean HbA1c level was 5.7% (39 mmol/mol). Calibration graphs showed that predicted HbA1c levels were somewhat underestimated in the Inter99 cohort and overestimated in the Hoorn and KORA cohorts, indicating that the model's intercept should be adjusted for each cohort to improve predictions. Sensitivity and specificity (95% CI) were 55.7% (53.9, 57.5) and 56.9% (55.1, 58.7) respectively, for women, and 54.6% (52.7, 56.5) and 54.3% (52.4, 56.2) for men. External validation showed similar performance in the METSIM cohort. Conclusions/interpretation In the non-diabetic population, our DIRECT-DETECT prediction model, including readily available predictors, has a relatively low explained variance and moderate discriminative performance, but can help to distinguish between future highest and lowest HbA1c levels. Absolute HbA1c values are cohort-dependent.

UR - http://www.scopus.com/inward/record.url?scp=85012240340&partnerID=8YFLogxK

U2 - 10.1371/journal.pone.0171816

DO - 10.1371/journal.pone.0171816

M3 - Article

VL - 12

JO - PLoS ONE

JF - PLoS ONE

SN - 1932-6203

IS - 2

M1 - e0171816

ER -