2025-08-08 バッファロー大学(UB)
<関連情報>
- https://www.buffalo.edu/news/releases/2025/08/Prediabetes-mortality-link-younger-adults.html
- https://jamanetwork.com/journals/jamanetworkopen/fullarticle/2837340
人口統計、生活習慣、合併症、前糖尿病、および死亡率 Demographics, Lifestyle, Comorbidities, Prediabetes, and Mortality
Obinna Ekwunife, PhD; Xuemeng Wang, MSc; Raphael Fraser, PhD; et al
JAMA Network Open Published:August 7, 2025
DOI:10.1001/jamanetworkopen.2025.26219
Introduction
In addition to increasing the risk of developing type 2 diabetes,1,2 prediabetes is also linked to increased cardiovascular disease risk3 and elevated all-cause and cause-specific mortality.4,5 However, the association between prediabetes and mortality remains complex, particularly when accounting for factors such as demographics, lifestyle, and comorbidities. To better understand how these factors affect the association, this study examines their modifying effect in a US adult population with prediabetes. Associations were assessed by adjusting for 3 levels of factors: demographic factors alone, demographic and lifestyle factors, and demographic, lifestyle, and comorbidity factors.
Methods
Following STROBE guidelines, this cohort study used data from the National Center for Health Statistics linked to the National Death Index mortality follow-up for individuals who participated in the National Health and Nutrition Examination Survey (NHANES) from 1999 through 2018.6 Adults 20 years or older who completed both the interview and physical examination, had valid mortality data, and participated in survey cycles from 2005 to 2018 were included. Prediabetes was defined by self-report or hemoglobin A1c level (5.7%-6.4%) using NHANES data. Covariates included demographics, lifestyle, and comorbidities. Race and ethnicity were self-reported during NHANES interviews, and participants were categorized as non-Hispanic Black, non-Hispanic White, or other (due to limited sample size). Multivariable Cox proportional hazards models were used to assess associations and adjust for potential confounders. We also conducted stratified analyses by age group and race and ethnicity to examine possible modification effect. Analyses used weighted data and R, version 4.4.1 (R Foundation for Statistical Computing), with significance set at P < .05. This retrospective analysis of deidentified data did not require institutional review board approval or patient consent, in accordance with 45 CFR §46.102(e).
Results
Of the 38 093 respondents, 9971 (26.2%), representing more than 51 million US adults, had prediabetes. Table 1 shows weighted demographics of these individuals. Most were female and aged 20 to 54 years. Prediabetes was initially associated with mortality (hazard ratio [HR], 1.58; 95% CI, 1.43-1.74), but lost significance in the fully adjusted model (HR, 1.04; 95% CI, 0.92-1.18) (Table 2). Significant interactions were observed between prediabetes and age group and race and ethnicity. Stratified Cox models found that prediabetes was statistically significantly associated with mortality only among adults aged 20 to 54 years (HR, 1.64; 95% CI, 1.24-2.17) (Table 2). No significant associations were found among racial and ethnic groups.
| Characteristic | All respondents, No. (weighted %) (N = 38 093)a | With prediabetes, No. (weighted %) (n = 9971)a | Deaths among those with prediabetes, No.a | Without prediabetes, No. (weighted %) (n = 28 122)a | Deaths among those without prediabetes, No.a |
|---|---|---|---|---|---|
| Demographic | |||||
| Age, yb | |||||
| 20-54 | 22 492 (65.1) | 4374 (46.8) | 126 | 18 118 (70.6) | 416 |
| 55-74 | 11 426 (27.1) | 4076 (40.6) | 437 | 7350 (23.1) | 1223 |
| ≥75 | 4175 (7.8) | 1521 (12.6) | 666 | 2654 (6.3) | 1343 |
| Sex | |||||
| Male | 18 427 (48.1) | 4801 (46.3) | NA | 13 626 (48.7) | NA |
| Female | 19 666 (51.9) | 5170 (53.7) | NA | 14 496 (51.3) | NA |
| Race and ethnicityb | |||||
| Non-Hispanic Black | 8295 (11.4) | 2559 (14.6) | 227 | 5736 (10.4) | 677 |
| Non-Hispanic White | 15 891 (66.7) | 3755 (62.7) | 796 | 12 136 (67.9) | 1745 |
| Otherc | 13 907 (21.9) | 3657 (22.7) | 206 | 10 250 (21.6) | 560 |
| Marital status | |||||
| Not married | 18 526 (44.9) | 4611 (41.9) | NA | 13 915 (45.8) | NA |
| Married | 19 542 (55.1) | 5353 (58.1) | NA | 14 189 (54.2) | NA |
| Lifestyle | |||||
| Smoking status | |||||
| Non-smoker | 21 127 (55.2) | 5253 (51.7) | NA | 15 874 (56.3) | NA |
| Former smoker | 9115 (24.5) | 2651 (27.3) | NA | 6464 (23.6) | NA |
| Current smoker | 7822 (20.3) | 2057 (21.0) | NA | 5765 (20.1) | NA |
| ≥12 Drinks in past y | 22 933 (82.3) | 5703 (77.2) | NA | 17 230 (83.9) | NA |
| Comorbities | |||||
| Diabetes | 4942 (9.6) | 0 | NA | 4942 (12.5) | NA |
| Hypertension | 13 669 (31.8) | 4500 (43.4) | NA | 9169 (28.3) | NA |
| Heart disease | 3291 (6.9) | 1062 (10.0) | NA | 2229 (5.9) | NA |
| Stroke | 1505 (2.9) | 436 (3.7) | NA | 1069 (2.7) | NA |
| Cancer | 3620 (10.1) | 1141 (13.2) | NA | 2479 (9.1) | NA |
| Body mass index, mean (SD)d | 29 (7) | 29 (7) | NA | 31 (7) | NA |
| Mortality status | |||||
| Alive | 33 882 (91.9) | 8742 (89.6) | NA | 25 140 (92.6) | NA |
| Deceased | 4211 (8.1) | 1229 (10.4) | 1229 | 2982 (7.4) | 2982 |
(opens in new tab)| Modela | HR (95% CI) | P value |
|---|---|---|
| Overall | ||
| Unadjusted | 1.58 (1.43-1.74) | <.001 |
| Adjusted for demographics | 0.88 (0.80-0.98) | .02 |
| Adjusted for demographics and lifestyle factors | 0.92 (0.82-1.04) | .17 |
| Adjusted for demographics, lifestyle factors, and comorbidities (fully adjusted) | 1.05 (0.92-1.19) | .47 |
| Stratified by age group | ||
| 20-54 y | 1.68 (1.25-2.20) | <.001 |
| 55-74 y | 0.87 (0.69-1.09) | .22 |
| ≥75 y | 0.97 (0.83-1.12) | .66 |
| Stratified by race and ethnicity | ||
| Non-Hispanic Black | 1.02 (0.80-1.30) | .86 |
| Non-Hispanic White | 1.06 (0.91-1.23) | .45 |
| Otherb | 0.81 (0.56-1.19) | .29 |
Discussion
Using NHANES data linked to mortality records, this cohort study found that while unadjusted models showed a significant association as reported by several studies,4 ,5 adjusting for demographic, lifestyle factors, and comorbidities attenuated this finding (see Supplement 1). Stratified analyses revealed that prediabetes was significantly associated with mortality only among younger adults (age 20-54 years), highlighting the importance of age-specific interventions. Lifestyle behaviors, limited health care access, and life stage challenges may contribute to the increased mortality risk in younger adults. Early-onset health problems in this group may also reflect stronger genetic predispositions, leading to more rapid disease progression and more severe health outcomes. These findings underscore the need for tailored diabetes prevention programs targeting young adults—such as flexible, virtual, and peer-led options—to increase accessibility and engagement. Routine screening and timely referrals to age-appropriate programs are essential. Although the use of nationally representative NHANES data strengthens the study, limitations include its cross-sectional design, potential self-report bias, lack of longitudinal tracking, and the inability to infer causality due to its observational nature. Future research should focus on longitudinal studies and targeted interventions to reduce mortality among young adults with prediabetes. Early intervention is key to preventing disease progression and improving long-term health outcomes.
