Previous studies on trends in the prevalence of prediabetes among U.S. adults were based on either fasting plasma glucose (FPG) and 2-h plasma glucose (2-h PG) definition (1) or FPG and hemoglobin A1c (HbA1c) definition (2) and observed inconsistent results. The trend in prevalence of prediabetes among U.S. adults based on all three glycemic measures is unknown.
To comprehensively examine the trends in prevalence and awareness of prediabetes among U.S. adults and evaluate the extent of underestimation of prevalence according to various glycemic measures, we analyzed the nationally representative data of the National Health and Nutrition Examination Survey (NHANES) from 2005 to March 2020.
The NHANES is a serial, cross-sectional survey with a stratified, multistage probability design to monitor the health status of the U.S. noninstitutionalized civilian population (3). A subsample of participants was randomly selected to take an oral glucose tolerance test (OGTT) in the mobile examination center. Since the field operations were suspended in mid-March 2020 due to the coronavirus 2019 pandemic, the partial data of the 2019–2020 cycle were combined with those of the previous 2017–2018 cycle (denoted as the 2017–March 2020 period here) to generate nationally representative estimates (4). Sample weights were constructed to account for nonresponse and oversampling. The NHANES was approved by the research ethics review board of the National Center for Health Statistics. Written informed consent was obtained from each participant.
All nonpregnant adults aged 20 years or older who attended medical examinations in the morning session and met the 8- to less than 24-h fasting criteria from 2005–2006 to 2017–March 2020 periods were eligible for this study. The analyses in the fasting subsample included 17,405 participants who had measurements of FPG and HbA1c. Among them, 2-h PG levels were further measured in 10,803 adults from the 2005–2006 to 2015–2016 periods. Plasma glucose results were calibrated to early survey cycles by using the recommended backward equations (3).
To generate nationally representative estimates, we used fasting subsample weight for the FPG/HbA1c definition and OGTT subsample weight for the FPG/HbA1c/2-h PG definition separately. Estimates were age adjusted to the civilian noninstitutionalized population for 2017–March 2020 (3). All confidence intervals (CIs) were estimated using Taylor series linearization to account for the complex sample design. Analyses were performed using SAS, version 9.4 (SAS Institute).
In the fasting subsample, the age-adjusted prevalence of prediabetes based on FPG/HbA1c definition increased from 32.1% (95% CI 29.8–34.5%) in 2005–2006 to 39.6% (37.1–42.0%) in 2007–2008 and then plateaued to 38.6% (36.5–40.7%) in 2017–March 2020 without a significant linear trend (P = 0.06) or quadratic trend (P = 0.83) (Fig. 1A). In the OGTT subsample, the age-adjusted prevalence of prediabetes defined by FPG/HbA1c/2-h PG definition increased from 37.8% (35.4–40.2%) in 2005–2006 to 46.7% (44.4–49.0%) in 2007–2008 and then plateaued at 44.2% (40.4–48.0%) in 2015–2016 without a significant linear trend (P = 0.23) or quadratic trend (P = 0.20). The age-adjusted awareness of prediabetes increased from 7.5% (5.4–9.6%) in 2005–2006 to 19.1% (14.8–23.3%) in 2017–March 2020 based on FPG/HbA1c definition in the fasting subsample and increased from 6.7% (5.0–8.4%) in 2005–2006 to 15.0% (12.1–17.8%) in 2015–2016 based on FPG/HbA1c/2-h PG definition in the OGTT subsample (both P < 0.001 for linear trend) (Fig. 1B).
During the entire period of surveys from 2005 to 2016, there were 89.8% (95% CI 88.6–91.1%) adults with prediabetes (FPG/HbA1c/2-h PG definition) detected by the combination of FPG and HbA1c, higher than 81.6% (80.2–83.0%) for the combination of FPG and 2-h PG, 65.5% (63.5–67.5%) for the combination of HbA1c and 2-h PG, and any single glycemic measures (Fig. 1C).
A large-scale meta-analysis showed that each glycemic measure alone was associated with an increased risk of cardiovascular disease and all-cause mortality (5). Therefore, using just one or two glycemic measures would certainly underestimate the prediabetes prevalence and lead to unavoidable misdiagnosis of some high-risk individuals. We found that the combination of FPG and HbA1c could detect the vast majority of total prediabetes cases based on the full American Diabetes Association definition (6). Considering that the administration of an OGTT is time-consuming and burdensome for participants, the FPG/HbA1c definition could be used as an acceptable alternative definition for surveillance of prediabetes when OGTT data were unavailable.
Several limitations should be considered. First, the glycemic levels were measured one time, which might lead to misclassification of prediabetes. Second, the OGTT data were unavailable in the period 2017–March 2020.
From 2005 to March 2020, there was no changing trend in the prevalence of prediabetes, whereas the awareness of prediabetes increased among adults in the U.S. The FPG/HbA1c definition could be used as an acceptable alternative definition for surveillance of prediabetes.
The content is solely the responsibility of the authors and does not necessarily represent the official position of the Huazhong University of Science and Technology.
Article Information
Funding. A.P. was supported by the National Nature Science Foundation of China (81930124 and 82021005), the Shanghai Municipal Science and Technology Major Project (grant no. 2017SHZDZX01), and the Fundamental Research Funds for the Central Universities (2021GCRC 075). G.L. was supported by the National Natural Science Foundation of China (82073554), National Nutrition Science Research Grant (CNS-NNSRG2021-10), and the Fundamental Research Funds for the Central Universities (2021GCRC 076). T.-T.G. was supported by the China Postdoctoral Science Foundation (2021M691129).
The funding agencies play no role in the design or conduct of the study, collection, management, analysis, or interpretation of the data, preparation, review, approval of the manuscript, or decision to submit the manuscript for publication.
Duality of Interest. No potential conflicts of interest relevant to this article were reported.
Author Contributions. P.-F.X. contributed to the study concept and design, statistical analysis, interpretation of data, and drafting of the manuscript. Y.-X.T. contributed to the interpretation of data and drafting of the manuscript. T.-T.G. contributed to the interpretation of data and drafting of the manuscript. Y.L. contributed to the statistical analysis. A.P. contributed to the study concept and design, drafting of the manuscript, funding, and study supervision. All authors were involved in the critical revision of the manuscript and approved the final version of this article. P.-F.X. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.