Symptoms of Anxiety and Depression Are Independently Associated With Impaired Awareness of Hypoglycemia in Type 1 Diabetes

Impaired awareness of hypoglycemia (IAH) affects 18–30% of people with type 1 diabetes (1). The reduction of hypoglycemic symptoms and counterregulatory hormonal responses in IAH increases the risk of developing severe hypoglycemia sixfold (2). Repeated exposure to hypoglycemia creates further impairments in awareness. Interventions to reduce hypoglycemia and restore awareness continue to be needed (3).

Compared with healthy individuals, people with type 1 diabetes are at increased risk of mental disorders, with depression being twice as likely and prevalence of anxiety 14% compared with 3–4% in healthy individuals (4,5). Mental disorders are associated with poorer outcomes, reduced quality of life, mortality, diabetes distress, and eating disorders (4,6).

Although IAH and mental disorders are common in type 1 diabetes, little is known about how they may interact. With growing evidence for emotional and cognitive involvement in IAH (2), we hypothesized that symptoms of anxiety and depression would be higher in IAH. We sought potential relationships between these factors in a large cohort of people with type 1 diabetes.

The study was conducted between January 2016 and April 2018 in a large, urban, multicenter clinic for adults with type 1 diabetes aged ≥18 years and diagnosed for ≥12 months. Data were collected during routine clinic visits and as part of the Integrating Mental and Physical Healthcare: Research, Training and Services (IMPARTS) program (7), which collects mental health questionnaire data in clinic settings. Ethics approval was gained from the National Research Ethics Service Research Database Committee (ref. no. 18/SC/0039). A Clarke score of ≥4 indicated IAH (8). Criteria for probable generalized anxiety disorder (pGAD) were met if the Generalized Anxiety Disorder-7 score was ≥10. Probable major depressive disorder (pMDD) was identified if the respondent reported at least one of low mood or anhedonia on the Patient Health Questionnaire-9 and at least five of the nine items for more than half the days in the past 2 weeks. Further information regarding criteria for assessment of IAH, pMDD, pGAD, and other variables are detailed in Supplementary Table 1.

Statistical analyses were performed using Stata 16 software (significance at α = 0.05). The psychological, sociodemographic, and clinical characteristics of respondents with and without IAH were compared using χ² and Mann-Whitney U tests. Logistic regression models were used to examine the relationship between the independent variables (pMDD and pGAD) and dependent variable (IAH), with sociodemographic and clinical variables added as covariates. To assess how selection of adjustment variables affect the relationship between mental health disorders and IAH, we computed the following three models: model 1, unadjusted; model 2, adjusted for sociodemographic variables; and model 3, additionally adjusted for clinical variables. Further logistic regression analyses were performed with anxiety and depression in separate models. Regression models were also conducted to assess whether the severity of common mental disorders affected the association with IAH; 95% CIs are reported.

Data from the study are available upon request from the corresponding authors.

A total of 1,009 patients completed IMPARTS during the study. Fifty-nine respondents were excluded because of lack of clinical data, leaving 950 included in the study sample (∼43% of the clinic population). Table 1 lists the sample characteristics.

Prevalence of IAH, pGAD, and pMDD were 22.6% (95% CI 20.0–25.4), 9.4% (7.6–11.4), and 9.8% (8.0–11.9), respectively. Fifty-nine respondents (6.2% [4.8–7.9]) met criteria for both pGAD and pMDD (Table 1).

A significantly larger percentage of respondents with IAH met criteria for pGAD or pMDD than those without IAH (14.9 vs. 7.8% [P = 0.002] and 18.1 vs. 7.4% [P < 0.001], respectively); 11.2% of respondents with IAH vs. 6.8% without met criteria for pMDD and pGAD (P = 0.001). Those with IAH had longer diabetes duration (z = −2.31, P = 0.021); were older (z = −2.11, P = 0.034) and more likely female (56.3 vs. 47.1%, P = 0.018); had an HbA_1c of ≤58 mmol/mol (7.5%) (45.6 vs. 33.3%, P = 0.002), a BMI <25 kg/m² (52.1% vs. 45.9%, P = 0.029), and urine albumin-to-creatinine ratio ≥3 mg/g (78.4% vs. 84.5%, P = 0.036); and were using continuous subcutaneous insulin infusion (42.3 vs. 34.2%, P = 0.028) (Table 1). In contrast, raised HbA_1c levels and obesity were more common with pMDD or pGAD (Supplementary Tables 2 and 3).

In logistic regression (Table 2), the odds of having IAH were significantly greater for respondents with pMDD than for those without, including after adjustment for potential confounders (odds ratio [OR] 2.62 [95% CI 1.47–4.66]). In model 2, adjusted for age, sex, ethnicity, and Index of Multiple Deprivation (IMD), the relationship between pMDD and IAH remained unaltered (OR 2.65 [1.46–4.80]). In model 3, the addition of clinical variables, including indicators of disease severity and complications, did not attenuate the effect and strengthened the relationship (OR 3.27 [1.74–6.15]). However, when anxiety was included in the same regression model with depression, the odds of having IAH were not significantly affected by having pGAD (OR 1.21 [0.62–2.34]). Additional data from logistic regression are listed in Supplementary Table 4.

Because depression and anxiety were related and often coexisted (66.3% with pGAD and pMDD, P < 0.001), regression analysis with each mental disorder in a separate model was performed. The likelihood of having IAH was significantly greater for patients with pGAD (OR 2.46 [95% CI 1.46–4.14]) and pMDD (OR 3.64 [2.19–6.04]), robust to adjustment (Table 2). A further regression model with severity of anxiety and depression and IAH suggested a dose-response relationship between increased severity of probable mental disorder and increased odds of having IAH that were statistically significant after adjustment (pMDD severity: OR 3.87 [2.32–6.47, P for linear trend < 0.001], pGAD severity: OR 2.55 [1.51–4.30, P for linear trend < 0.001]) (Table 2).

This study revealed a robust independent relationship between symptoms of mental disorders, in particular depression, and IAH that has not previously been reported. The literature is limited on the relationship between common mental disorders and IAH in type 1 diabetes, although increased diabetes distress has been described (9).

Our study design does not establish causality. Common mental disorders could lead to altered behavior or motivations in diabetes self-care, increasing the risk for hypoglycemia and, consequently, IAH (10). A reduced ability to discriminate between anxiety and hypoglycemia symptoms may delay hypoglycemia treatment, contributing to IAH (11). Equally, recurrent hypoglycemia and/or IAH may have a negative impact on quality of life and emotions. It is also plausible that shared central nervous system pathways between the homeostatic glucose-sensing and mesocorticolimbic circuits are associated with central nervous system neurotransmitter release related to hypoglycemia and emotional regulation (12).

The strengths of our study are that the analysis was based on a large cohort and that our data on prevalence of IAH, pGAD, and pMDD (1,5,13) and demography of IAH (older age and long diabetes duration) (14) are concordant with previous studies. The study is limited by its cross-sectional design, meaning that questions of causality cannot be assessed. To define probable mental disorders and IAH, we used validated self-reported screening tools rather than clinical interviews. The effect sizes we report for the main analyses were large and clinically significant. Our use of recognized and clinically valid categorical cutoff values instead of dimensions would have reduced statistical power to detect more modest associations.

To conclude, we demonstrate a significant, independent, and robust relationship between probable depression and anxiety and IAH in adults with type 1 diabetes, independent of socioeconomic factors and disease severity. Our data support a recent report that measuring mental health and providing psychological interventions alongside education could achieve overall best outcomes in IAH (15). Future research may provide further insights into the mechanisms underlying the relationship and whether integrated mental health programs in people with type 1 diabetes and IAH may help to reverse IAH and/or improve mental health.

Acknowledgments. The authors thank Guy’s and St. Thomas’ Charity and Stephanie Singham (Department of Diabetes, Guy’s and St. Thomas’ NHS Foundation Trust) for help with implementation of the IMPARTS program within the diabetes service. The authors also thank the clinical diabetes team members at Guy’s and St. Thomas’ Hospitals and participants of the study. The authors thank Prof. Stephanie Amiel (Department of Diabetes, King’s College, London) for the helpful comments and feedback on the manuscript. The authors dedicate this work in memory of Faith Ekperuoh, whose energy in helping to implement the IMPARTS program at Guy’s hospital was vital.

Funding. Support for this study was partially provided by the National Institute for Health Research Biomedical Research Centre at South London and Maudsley NHS Foundation Trust and King’s College London.

The views expressed are those of the authors and not necessarily those of the National Health Service, the National Institute of Health Research, or the Department of Health and Social Care.

Duality of Interest. No potential conflicts of interest relevant to this article were reported.

Authors Contributions. B.A.P. and G.A.I.B.-C coauthored the manuscript. B.A.P and J.C. conducted the clinical data extraction. B.A.P. and S.H. designed the study. G.A.I.B.-C conducted the analysis. A.S. and A.B. supported the implementation of the IMPARTS program in the diabetes clinic. M.H. designed, gained funding for, and implemented the IMPARTS program. J.K., M.H., and S.H. supervised the study design and data analysis and assisted with writing the manuscript. All authors critically reviewed and approved the final version of the manuscript. S.H. is the guarantor of the 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.

Prior Presentation. Parts of this study were presented in poster form at the 81st Virtual Scientific Sessions of the American Diabetes Association, 25–29 June 2021.