Depression Risk in Type 1 Versus Type 2 Diabetes: Cross-Sectional Analysis of Body Mass Index (BMI) in a Nationally Diverse Cohort - PubMed (original) (raw)

Comparative Study

Depression Risk in Type 1 Versus Type 2 Diabetes: Cross-Sectional Analysis of Body Mass Index (BMI) in a Nationally Diverse Cohort

Natalia Cruz-Vespa et al. Endocrinol Diabetes Metab. 2026 Mar.

Abstract

Introduction: Major depressive disorder (MDD) commonly co-occurs with diabetes, but comparative risk across type 1 diabetes (DM1), type 2 diabetes (DM2) and non-diabetic groups-and the role of body mass index (BMI)-remains uncertain.

Methods: Using All of Us Research Program data, adults were classified as DM1, DM2 or non-diabetic. Multivariable logistic regression estimated odds of MDD adjusting for age, sex at birth, race and ethnicity; BMI was added in secondary models. Effect modification by sex and race was tested. Structural equation modelling (SEM) assessed whether BMI statistically explained group differences.

Results: In models excluding BMI, both DM1 and non-diabetic participants had higher odds of MDD than DM2 (DM1 vs. DM2: OR = 1.53, 95% CI 1.17-1.99; non-diabetic vs. DM2: OR = 1.20, 95% CI 1.16-1.25). Interactions by sex and race were significant; contrasts were stronger among females and heterogeneous across race strata. Adding BMI yielded directionally consistent group estimates and confirmed an independent association of higher BMI with higher MDD odds. SEM indicated statistical suppression for the non-diabetic vs. DM2 contrast: non-diabetic status related to lower BMI, while higher BMI related to higher MDD, producing a small indirect effect (~8%). The indirect path for DM1 vs. DM2 was non-significant.

Conclusions: Compared with DM2, both DM1 and non-diabetic groups show higher adjusted odds of MDD. BMI is independently related to MDD but only modestly-and partly suppressively-accounts for the non-diabetic vs. DM2 contrast. Findings support subgroup-aware screening and the need for longitudinal data to clarify mechanisms.

Keywords: body mass index; depression; type 1 diabetes; type 2 diabetes.

© 2026 The Author(s). Endocrinology, Diabetes & Metabolism published by John Wiley & Sons Ltd. This article has been contributed to by U.S. Government employees and their work is in the public domain in the USA.

PubMed Disclaimer

Conflict of interest statement

The authors declare no conflicts of interest.

Figures

FIGURE 1

FIGURE 1

Participant inclusion and exclusion flowchart. Construction of analytic cohorts in the All of Us Researcher Workbench. Participants were identified using the Cohort Builder tool with parallel inclusion logic for each group. Type 1 Diabetes and Type 2 Diabetes cohorts were mutually exclusive, with cross‐exclusion of the alternate diabetes type. The control cohort included participants of all ages without diabetes, excluding any Type 1 or Type 2 Diabetes diagnoses. Final counts were 808 (Type 1 DM), 67,640 (Type 2 DM) and 431,810 (Controls), yielding a total analytic registry sample of 500,258.

FIGURE 2

FIGURE 2

Mediation model testing whether BMI mediates the association between diabetes group (DM1 vs. DM2) and major depressive disorder (MDD). The indirect effect (a × b) was small but statistically detectable (+0.00184, p ≈ 0.031), accounting for ≈8% of the total effect. The BMI → MDD path was near zero in magnitude (b ≈ −0.00025), indicating only a modest statistical (inconsistent mediation/suppression) pattern rather than a clinically meaningful mediated pathway.

FIGURE 3

FIGURE 3

Mediation model testing whether BMI mediates the association between diabetes group (Non‐diabetic vs. DM2) and major depressive disorder (MDD). The path from group to BMI was negative and significant (a = −4.689, p < 0.001) and the path from BMI to MDD was positive and significant (b = +0.00073, p < 0.001), yielding a small but significant negative indirect effect (a × b = −0.00340, p < 0.001). The direct effect remained positive and significant (_c_′ = +0.0114, p < 0.001), with an overall total effect of +0.0080. Thus, BMI acted as a suppressor, offsetting ~43% of the total association (−0.00340/0.0080), that is, statistical suppression rather than explanatory mediation.

References

    1. Egede L. E. and Ellis C., “Diabetes and Depression: Global Perspectives,” Diabetes Research and Clinical Practice 87, no. 3 (2010): 302–312, 10.1016/j.diabres.2010.01.024. -DOI -PubMed
    1. Fanelli G., Raschi E., Hafez G., et al., “The Interface of Depression and Diabetes: Treatment Considerations,” Translational Psychiatry 15, no. 1 (2025): 22, 10.1038/s41398-025-03234-5. -DOI -PMC -PubMed
    1. Chireh B., Li M., and D'Arcy C., “Diabetes Increases the Risk of Depression: A Systematic Review, Meta‐Analysis and Estimates of Population Attributable Fractions Based on Prospective Studies,” Preventive Medical Reports 14 (2019): 100822, 10.1016/j.pmedr.2019.100822. -DOI -PMC -PubMed
    1. Khubchandani J., Banerjee S., Gonzales‐Lagos R., and Szirony G. M., “Depression Increases the Risk of Mortality Among People Living With Diabetes: Results From National Health and Nutrition Examination Survey, USA,” Diabetes and Metabolic Syndrome: Clinical Research and Reviews 17, no. 11 (2023): 102892, 10.1016/j.dsx.2023.102892. -DOI -PubMed
    1. Wojujutari Ajele K. and Sunday Idemudia E., “The Role of Depression and Diabetes Distress in Glycemic Control: A Meta‐Analysis,” Diabetes Research and Clinical Practice 221 (2025): 112014, 10.1016/j.diabres.2025.112014. -DOI -PubMed

Publication types

MeSH terms

LinkOut - more resources