Skip to main content
Download PDF

Is alcohol use disorder associated with higher rates of depression and anxiety among people with new onset type 2 diabetes? A cohort study using linked primary care data in England

BMC Primary Care volume 25, Article number: 386 (2024) Cite this article

Abstract

Introduction

Depression and alcohol use disorder (AUD) in people living with Type 2 diabetes mellitus (T2DM) are associated with worse health outcomes. AUD is strongly associated with depression and anxiety, but it is not known how these conditions cluster in people with T2DM. We investigated rates of new episodes of depression and anxiety following T2DM diagnosis in people with and without prior AUD among an English primary care population.

Methods

The study population was people diagnosed with T2DM between 2004 and 2019. We used the Clinical Practice Research Datalink (CPRD) Aurum database and linked Hospital Episode Statistics Admitted Patient Care (HES APC) and Office for National Statistics (ONS) mortality data. We examined incidence of new episodes of anxiety or depression in people with T2DM with and without AUD. AUD was defined as any of i) clinical diagnosis; ii) alcohol withdrawal; or iii) chronic alcohol-related harm (physical or mental) using SNOMED-CT or ICD-10 codes. People were excluded if they had codes for depression/anxiety 12 months prior to T2DM diagnosis. Poisson regression models were fitted adjusting sequentially for a) age, gender, calendar time; b) region, Index of Multiple Deprivation, ethnicity, body mass index, smoking status, Charlson co-morbidity index; and c) history of a mental health condition.

Results

Our study population was 479,447 people of whom 10,983 (2.3%) had an AUD code prior to T2DM diagnosis.

After adjusting for all measured confounders except history of a mental health condition, IRR for depression was 2.00 (95% CI 1.93, 2.06) for people with AUD compared to without AUD. This reduced to 1.45 (95% CI 1.41, 1.50) after further adjustment for history of a mental health condition.

Findings for anxiety were substantially similar to those for depression (adjusted for all measured confounders except history of a mental health condition, IRR 2.08 95% CI 1.99, 2.18 fully adjusted IRR 1.48 95% CI 1.41, 1.55).

Conclusions

People with AUD have over double the rates of depression and anxiety following T2DM diagnosis than those without AUD. This was only partially explained by pre-existing diagnoses of mental health conditions. A holistic approach incorporating mental health support is needed to improve health outcomes for people with AUD who develop T2DM.

Trial registration

Not applicable.

Peer Review reports

Background

Diabetes is a serious public health concern ranked as the eighth leading cause of death and disability worldwide in 2019 [1]. Global prevalence of diabetes, particularly type 2 diabetes (T2DM), is increasing [2] with an estimated 529 million people worldwide living with diabetes in 2021 [2]. High levels of alcohol consumption (estimated as > 63 g/day from a meta-analysis from 2015 [3]) have been consistently associated with increased risk of developing type 2 diabetes [3,4,5,6].

If blood glucose is not well-controlled, T2DM can have severe consequences including sight loss, amputation, chronic kidney disease, myocardial infarction, and stroke. Management of T2DM is crucial for preventing adverse health outcomes. Management is complex, involving adherence to medication, a healthy diet, physical activity, and attending regular check-ups within primary care. This can put a high burden on patients. Co-occurring mental health conditions including both depression and anxiety are common in T2DM [7,8,9,10,11,12] and can make management even more challenging.

Depression, anxiety and alcohol use disorder (AUD), an umbrella term for alcohol dependence and alcohol-related harm, are all associated with poorer outcomes from T2DM [13,14,15,16,17,18,19,20,21]. There is evidence that co-occurring depression and anxiety in people with diabetes are associated with glycemic variability [12, 19, 22], lower treatment adherence [23, 24], poorer quality of life [25], higher health care costs [26], higher risk of complications [13, 15, 21] and mortality [14, 16, 20].

There are high rates of co-occurrence of AUD and common mental disorders (depression and anxiety) in the general population [27,28,29] but there is little evidence looking at how these conditions cluster together in people with T2DM. Where several conditions are present, this could amplify the effects of each other and increase the risk of adverse outcomes. To our knowledge, no studies have previously investigated whether AUD is associated with an increased risk of developing depression and anxiety following T2DM diagnosis. To target interventions effectively, it is important to quantify to what extent people with AUD are at higher risk of developing depression and anxiety following a diagnosis of T2DM.

The aim of this paper was to investigate if people with AUD were more likely to develop a new episode of depression or anxiety following T2DM diagnosis compared to those without AUD among a primary care population in England.

Methods

Study sample and data source

We used data from the Clinical Practice Research Datalink (CPRD) Aurum [30, 31]. CPRD Aurum contains electronic health record (EHR) data from a sample of primary care practices within the UK that use EMIS Web® general practice patient management software, collected during routine healthcare. This includes information on demographic characteristics, diagnoses and symptoms, medication prescriptions and laboratory tests. The CPRD Aurum population is representative of the UK population in terms of age, gender, geographic location and deprivation [30]. The May 2022 version of the database used for this study covered 19.83% of the UK population [31].

CPRD Aurum data were linked with Hospital Episode Statistics Admitted Patient Care Data (HES APC) [32], and Index of Multiple Deprivation (IMD) [33].

HES APC contains data on patients admitted to National Health Service (NHS) hospitals in England. As eligibility for linkage with HES data was a requirement for this study, we restricted our study population to people resident in England rather than the whole UK population.

Cohort construction and case definition

The target population was people with incident T2DM between 1st January 2004 and 1st January 2020. We restricted our study period to pre 2020 to avoid biasing incidence estimates from health service disruption during the COVID-19 pandemic [34].

We defined date of T2DM incidence as the first date for a T2DM diagnosis or a T2DM -related code from a broader code list of diabetes SNOMED CT codes (https://github.com/NHLI-Respiratory-Epi/diabetes_alcohol_use_disorder/blob/main/diabetes_broad_list_vs3.xls).

We excluded people if they were aged less than 18 at the time of diagnosis, their gender, region or IMD was missing, they were not a fully registered patient; were not eligible for linkage to HES APC and IMD datasets, or were registered in a GP practice outside of England. To capture incident (not prevalent) T2DM, we also excluded people if they had less than 1 year of registration prior to their first T2DM code; their first code for diabetes was not diagnostic for T2DM and was more than a year from the time of first T2DM code. We also excluded people if they did not have any codes specific to T2DM (only codes for diabetes in general without specifying type) or if they had more type 1 diabetes (T1DM) codes than T2DM codes (Fig. 1). Codes for gestational diabetes were not classed as incident T2DM but if someone had codes for gestational diabetes and codes for T2DM then their first T2DM codes was classed as the date of T2DM diagnosis.

Fig. 1
figure 1

Selection of People with a diagnosis of type 2 diabetes between 2004 and 2019

Full size image

Alcohol use disorder

Alcohol use disorder was defined from SNOMED-CT codes entered by health care practitioners within primary care or ICD-10 codes from hospital admitted patient care records.

For the SNOMED-CT codes we adapted the International Classification of Disease 10 (ICD-10) definition of AUD to capture diagnoses of alcohol dependence syndrome and harmful alcohol use as defined in Chapter F10.1-F10.9 Mental and Behavioural disorders due to Alcohol [35]. The codes used were categorised as AUD independently by at least two of the co-authors (SC reviewed all codes and TB, SG, ALN, AM, RP (all practicing general practitioners within primary care in England) who each reviewed a subset of codes (100–200 codes each)). Divergent codes were discussed by at least three of the co-authors and a consensus reached.

In line with this definition, AUD was defined as presence of a SNOMED-CT code in the primary care record which indicated i) an AUD diagnosis in line with ICD-10 definitions or ii) alcohol withdrawal as a symptom of alcohol dependence syndrome or iii) chronic alcohol-related harm to health (physical or mental) in line with an ICD-10 diagnosis of harmful alcohol use; or any ICD-10 code from F10.1 – F10.9 from within hospital inpatient records prior to the date of T2DM diagnosis.

Outcome definition

The outcomes were new episodes of depression or anxiety following diagnosis of T2DM. Depression and anxiety were defined from SNOMED-CT codes entered by health care practitioners within primary care or ICD-10 codes from hospital admitted patient care records.

New episode depression was classified as the earliest of i) any SNOMED-CT code indicating definite or probable depression or ii) a SNOMED-CT code indicating possible depression plus a prescription for an antidepressant medication within 3 months or iii) any ICD 10 code from hospital inpatient records indicating depression (Any codes within chapter F32(depressive episode); F33 (recurrent depressive episode); F34 (persistent mood disorders); F41.2 (mixed anxiety and depressive disorder) and F92.0 Depressive conduct disorder).

Definite/probable depression refers to SNOMED-CT codes which were more specific for depression, for example a diagnostic code. Possible depression refers to codes which were less specific, for example depression symptoms or use of screening tools for depression. SNOMED-CT codes related to the following diagnoses were included: major depressive disorder, dysthymia, recurrent depressive disorder, seasonal affective disorder, mixed anxiety and depression, disruptive mood dysregulation disorder, depression in dementia (or other condition), trauma and stress-related disorders with depressed mood, including adjustment disorders with depressed mood.

The same definition was used for anxiety but for equivalent codes for anxiety and anxiolytic medication. For anxiety from hospital episode statistics data we used the following ICD-10 codes: any codes within F40 (phobic anxiety disorders); F41 (other anxiety disorders); F42 (obsessive–compulsive disorder); F43 (reaction to severe stress and adjustment disorders); F44 (Dissociative [conversion] disorder) and F48 (other neurotic disorders). SNOMED-CT codes related to the following diagnoses were included: generalized anxiety disorder, panic disorder, mixed anxiety and depression, obsessive compulsive disorders, trauma- and stress-related disorders with anxiety, anxiety disorder NOS, agoraphobia, and social anxiety disorder.

People with codes for depression or anxiety in the 12 months prior to T2DM diagnosis using the same definitions as for the study outcome were excluded from analyses to prevent existing episodes of depression/anxiety at the time of diagnosis being treated as new episodes.

All codelists for depression and anxiety are available on Github (https://github.com/NHLI-Respiratory-Epi/diabetes_alcohol_use_disorder).

Confounding variables

Confounding variables included are listed in Table 1.

Table 1 Confounding factors included in the model
Full size table

Prior history of diagnosed mental health conditions (depression for anxiety analyses, anxiety for depression analyses, severe mental illness (SMI)) was considered in analyses as a potential confounder, but was also considered as potential mediator, and models were included with and without adjustment for prior diagnoses of mental health conditions.

SMI was defined as either a SNOMED-CT code within primary care or any ICD 10 code in inpatient hospital records at any time point prior to T2DM diagnosis for schizophrenia, other psychotic disorders or bipolar disorder.

Ethnicity, smoking and BMI were defined using previously described algorithms [37, 38].

All other variables were defined using primary care data using codelists available on Github (https://github.com/NHLI-Respiratory-Epi/diabetes_alcohol_use_disorder).

Statistical analysis

The baseline characteristics of people with and without AUD codes recorded at the time of T2DM diagnosis were compared using proportions for categorical variables and mean and standard deviation or median and interquartile range for continuous variables. Kaplan–Meier curves were plotted showing rates of new episode depression and anxiety from the time of T2DM diagnosis stratified by history of past depression and anxiety.

Poisson regression models were fitted separately for outcomes a) new episode depression and b) new episode anxiety with AUD at the time of T2DM diagnosis as the exposure adjusting sequentially for i) age at T2DM diagnosis, gender, and year of T2DM diagnosis, current age and calendar time, ii) region, IMD, ethnicity, smoking status, body mass index, and physical co-morbidities at the time of T2DM diagnosis measured using the Charlson Co-morbidity Index and iii) past history of depression, anxiety and SMI. Sequential adjustment was conducted to explore first the association with minimal adjustment (model 1), with adjustment for all measured confounders (model 2) and thirdly with additional adjustment for past mental health history to understand to what extent past mental health history explained any observed associations (model 3). Study follow up for each person started from date of T2DM diagnosis and ended at the earliest of first episode of the outcome (depression or anxiety respectively), death, end of GP registration, last collection date of the GP practice or 1st January 2020. Data were split into ten year age bands and 4 year calendar periods and current age and calendar period were adjusted for within all models with the assumption that incidence rates were constant within these age and time bands.

Interactions investigated were i) calendar period to test whether effect estimates were constant over calendar time and ii) history of past depression (where depression was the outcome) and iii) history of past anxiety (where anxiety was the outcome) to test whether there was effect modification by previous history of depression/anxiety. Interactions were investigated by fitting the same Poisson models as for main analyses with and without interaction terms between the interaction variable of interest (calendar period/history of past depression/history of past anxiety respectively) and AUD and comparing models with and without interaction terms using likelihood ratio tests.

We hypothesized rates would be different for people who were currently drinking compared to people who were not currently drinking at the time of their T2DM diagnosis. Therefore, we also investigated the impact of dividing the AUD group into people who were more likely to be drinking at the time of their diagnosis of T2DM and people with evidence to suggest they were not currently drinking at the time of the T2DM diagnosis. The “not currently drinking” group was defined from 1) a SNOMED-CT code within primary care for non-drinking more recent than their most recent code for AUD or 2) most recent AUD SNOMED-CT code in primary care indicated remission or 3) no AUD codes in primary or secondary care in the past 5 years.

Complete case analysis was used for regression modelling restricting models to people with complete data on all variables.

Analyses were conducted using Stata 17.

Results

There were 479,447 people in the sample with an incident diagnosis of T2DM between 1st January 2004 and 31st December 2019 and no codes for depression or anxiety in the year before T2DM diagnosis. Of these, 10,983 (2.3%) were coded as having AUD prior to diagnosis of T2DM.

The characteristics of the study population at time of T2DM diagnosis are shown in Table 2. People with a code for AUD prior to T2DM diagnosis had a higher prevalence of pre-existing mental health conditions compared to people without AUD codes prior to T2DM diagnosis (depression 43.8% vs 20.1%; anxiety 28.8% vs 13.0% and SMI 9.2% vs 2.0%; p < 0.001 for all three conditions) (Table 2).

Table 2 Characteristics of people with and without a code for Alcohol Use Disorder (AUD) prior to Type 2 Diabetes diagnosis at time of type 2 diabetes (T2DM) diagnosisa
Full size table

Association between AUD and incident depression following diagnosis of T2DM

For depression analyses follow up time was 2,256,586 person years for people without codes for AUD and 34,973 person years for people with codes for AUD. Median follow up time was 3.7 years (IQR 1.40–7.29 years).The rate of incident depression was 123.9 (95% CI 120.3, 127.6) per 1000 person years in people with a code for AUD compared to 51.6 (95% CI 51.3, 51.9) per 1000 person years in people with no codes for AUD, a rate difference of 72.3 per 1000 person years (Table 3).

Table 3 Rates of new episode depression following type 2 diabetes (T2DM) diagnosis by whether people had a code for alcohol use disorder (AUD) at the time of T2DM diagnosis and history of previous depression
Full size table

After minimal adjustment (Model 1), people with AUD had an incidence rate ratio (IRR) of depression 2.44 times higher than people without AUD (95% CI 2.37, 2.52). After adjusting for measured confounders except past history of mental health conditions (Model 2) this reduced to 2.00 (95% CI 1.93, 2.06). Further adjustment for past history of mental health conditions partially explained this association (Model 3), however people with AUD still had almost 50% higher rate of depression (IRR 1.45 95% CI 1.41, 1.50) even after adjusting for past history of depression, anxiety and SMI (Table 4).

Table 4 Incidence rate ratio for incident depression following diagnosis of Type 2 Diabetes (T2DM) by whether people have an Alcohol Use Disorder (AUD) code in their medical record prior to diagnosis
Full size table

The absolute rate of incident depression was highest in people with both a history of depression and AUD (Fig. 2 and Table 4). However, while the rate difference in rates of depression by whether people had AUD or not was higher in people with a past history of depression (106.4 per 1000) compared to those without a past history of depression (26.8 per 1000) (Table 3), after adjusting for all measured confounders including history of anxiety and SMI (model 3), there was a statistical evidence for interaction in the opposite direction i.e. the IRR was higher in people with no prior codes for depression (IRR 1.56 95% CI 1.48, 1.64) compared to people with codes for a history of depression (IRR 1.40 95% CI 1.34, 1.46) (test for interaction p = 0.002) (Table 4 and Fig. 3).

Fig. 2
figure 2

Kaplan–Meier graph of time to new episode of depression post Type 2 Diabetes (T2DM) diagnosis by Alcohol Use Disorder (AUD) and history of depression

Full size image
Fig. 3
figure 3

Incidence rate ratio for incident depression following diagnosis of Type 2 Diabetes (T2DM) by whether people have an Alcohol Use Disorder (AUD) code in their medical record prior to diagnosis Figure Legend: *Model 1 adjusted for sex, age and year of T2D diagnosis, current age and calendar time. **Model 2 = Model 1 + region, IMD quintile, ethnicity, BMI, smoking status, number of co-morbidities from Charlson co-morbidity index. ***Model 3 = Model 2 + past history of depression, past history of anxiety, past history of severe mental illness

Full size image

When the AUD groups were stratified by likely drinking status at T2DM diagnosis, the fully adjusted IRR (Model 3) was slightly higher for people AUD who were likely to be currently drinking (IRR 1.51 95% CI 1.45, 1.57) than for people with AUD who were likely not currently drinking (IRR 1.36 95% CI 1.29, 1.44) p value for a difference = 0.003 (Table 4 and Fig. 3).

There was no evidence that the association between AUD and depression was changing with calendar period (p = 0.91) (Table 4).

Association between AUD and incident anxiety following diagnosis of T2DM

For anxiety analyses, follow up time was 2,660,611 person years for people without codes for AUD and 47,140 person years for people with codes for AUD. Median follow up time was 4.7 years (IQR 1.99–8.69 years).The rate of incident anxiety was 45.3 (95% CI 43.3, 47.3) cases per 1000 person years in people with a code for AUD and 26.4 (95% CI 24.8, 25.2) cases per 1000 person years in people with no code for AUD, a rate difference of 26.4 per 1000 person years (Table 5).

Table 5 Rates of new episode anxiety following type 2 diabetes (T2DM) diagnosis by whether people had a code for alcohol use disorder (AUD) at the time of T2DM diagnosis and history of previous anxiety
Full size table

Substantive findings for the association between AUD and incident anxiety were very similar to those for depression.

After minimal adjustment (Model 1), people with AUD had an incidence rate ratio (IRR) of anxiety 2.56 times higher than people without AUD (95% CI 2.45, 2.68). After adjusting for measured confounders except past history of mental health conditions (Model 2), this reduced to 2.08 (95% CI 1.99, 2.18). As for depression, further adjustment for past history of mental health conditions partially explained this association (Model 3), however people with AUD still had almost 50% higher rates of anxiety (IRR 1.48 95% CI 1.41, 1.55) even after adjusting for past history of depression, anxiety and SMI (Table 6).

Table 6 Incidence rate ratio for incident anxiety following diagnosis of Type 2 Diabetes (T2DM) by whether people have an Alcohol Use Disorder (AUD) code in their medical record prior to diagnosis
Full size table

The absolute rate of incident anxiety was highest in people with both a history of anxiety and AUD (Fig. 4 and Table 5). However, while the rate difference in rates of anxiety by whether people had AUD or not was higher in people with a past history of anxiety (42.9 per 1000) compared to those without a past history of anxiety (12.9 per 1000) (Table 3), after adjusting for all measured confounders (model 3) there was a statistical evidence that the IRR was higher in people with no prior codes for anxiety(IRR 1.60 95% CI 1.50, 1.72) compared to people with codes for a history of anxiety (IRR 1.38 95% CI 1.30, 1.47) (p = 0.001) (Table 6 and Fig. 5).

Fig. 4
figure 4

Kaplan–Meier graph of time to new episode anxiety post Type 2 Diabetes (T2DM) diagnosis by Alcohol Use Disorder (AUD) and past anxiety

Full size image
Fig. 5
figure 5

Incidence rate ratio for incident anxiety following diagnosis of Type 2 Diabetes (T2DM) by whether people have an Alcohol Use Disorder (AUD) code in their medical record prior to diagnosis. Figure Legend: *Model 1 adjusted for sex, age and year of T2D diagnosis, current age and calendar time. **Model 2 = Model 1 + region, IMD quintile, ethnicity, BMI, smoking status, number of co-morbidities from Charlson co-morbidity index. ***Model 3 = Model 2 + past history of depression, past history of anxiety, past history of severe mental illness

Full size image

When the AUD groups were stratified by likely drinking status at T2DM diagnosis, the fully adjusted IRR (Model 3) was slightly higher for people with AUD who were likely to be currently drinking (IRR 1.56 95% CI 1.48, 1.65) than for people with AUD who were likely not currently drinking (IRR 1.33 95% CI 1.24, 1.44) p value for a difference p = 0.001 (Table 6 and Fig. 5).

There was no evidence that that association between AUD and anxiety changed over calendar period (p = 0.38) (Table 6 and Fig. 5).

Discussion

In this study of people with T2DM and AUD identified from electronic health care records, people with AUD had approximately two and a half times higher rates of developing a new episode of depression or anxiety. While a history of mental health problems at the time of T2DM diagnosis was a strong factor explaining this effect, adjusting for this did not explain this completely. People with AUD were still approximately 50% more likely to be diagnosed with depression or anxiety after a diagnosis of T2DM, even after adjusting for measured confounders including history of depression, anxiety and SMI. Our findings were consistent over calendar time and applied to people with and without a prior history of depression/anxiety.

There was strong evidence that incidence rate ratios for depression and anxiety were higher in people with AUD who were likely to be currently drinking compared to not currently drinking, after accounting for all measured confounders, including past medical history. However, those who were categorized as likely to be not currently drinking still had 36% higher rate of depression and 33% higher rates of anxiety than people without AUD, suggesting ever having an AUD if associated with longer-term effects on mental health even if someone is not currently drinking. Our AUD definition included chronic alcohol-related harms to physical and mental health, which may contribute to the continued impact of AUD on mental health even among people who were likely to be not drinking at the time of their T2DM diagnosis.

Current UK guidance from National Institute of Health and Clinical Excellence (NICE) is to treat AUD first before treating depression or anxiety in the presence of co-occurring AUD and depression or anxiety symptoms [39]. This may mean our findings underestimate the true effect of AUD on incidence of depression and anxiety, particularly in those who are currently drinking, as GPs may be less likely to diagnose depression or anxiety in people with AUD who are currently drinking.

Despite this possible under-ascertainment, we found people with a history of AUD were at substantially higher risk of developing both anxiety and depression following diagnosis of T2DM than people without a history of AUD. The increased rates of incident depression in people with AUDfound here is important both for the negative impact on quality of life [25] but also for the potential negative impact on diabetes prognosis. Co-morbid depression and anxiety in people with diabetes are associated with glycemic variability [12, 19, 22], lower treatment adherence [23, 24], higher risk of complications [13, 15, 21] and death [14, 16, 20]. From a clinical perspective, our study highlights the importance of support for people with a history of AUD who develop T2DM, both to stop drinking if they are currently drinking, and also with their mental health more generally.

Comparison with previous literature

A previous study using EHR data from the US found that among 16,243 people with prevalent T2DM and substance use disorder had 2.22 times higher odds of a prevalent mood disorder and 1.87 times higher odds of anxiety than people without substance use disorder after adjusting for demographic factors and other physical co-morbidities. However, this study was cross-sectional in nature and did not distinguish between alcohol use from tobacco and other drug use [40]. To our knowledge, our study is the first to investigate longitudinally whether AUD is associated with higher incidence of depression and anxiety following diagnosis of T2DM.

A systematic review and meta-analysis found AUD was associated with a risk ratio of 1.57 (with a range from 1.03–2.43) for incident depression among study samples not restricted to people with T2DM [29]. While it is hard to make a direct comparison, given the 16 studies included in the meta-analysis adjusted for different confounders to each other and to our study, the highest effect estimate in the meta-analysis (2.43) is from a population-based cohort study of 3481 adults in East Baltimore in the US which adjusted for age, sex and years of education only [41]. This is very similar to our minimally adjusted IRR (2.44), while the pooled risk ratio from the meta-analysis was close in size to our fully adjusted IRR (1.45). Overall, this suggests that the association of AUD with incident depression was similar in our sample of people with T2DM, with similar increased rates of depression in people with AUD compared to without AUD to studies from the general population, and the co-occurrence of T2DM does not seem to amplifying further the association between AUD and depression. Effect estimates from a longitudinal study population-based study from the Netherlands investigating the association between AUD and anxiety disorder at three year intervals among 6,646 participants were also of a similar effect size to our study after adjusting for socio-demographic factors, smoking and clinical factors (odds ratio for prediction of prevalent anxiety disorder 1.69 95% CI 1.11–2.43; odds ratio for first episode anxiety 2.03 95% CI 1.17–3.51). However the study design (data collected at three year intervals) and effect estimates from this study are not the same (odds ratios compared to incident rate ratio) limiting direct comparisons between the studies [42]. Further work comparing our findings in comparable study populations with and without T2DM is needed to confirm this.

Strengths and limitations

CPRD Aurum is a large nationally representative database. Although AUD is a relatively rare condition, we were able to identify 10,983 people meeting the study criteria for AUD with incident T2DM, providing sufficient power to investigate the research question of interest.

EHR data are routinely collected data and are not collected for research purposes. Measurement error is always a consideration when using these data. Completeness and quality of primary care data is dependent both on the coding practices by GPs when recording what takes place during a consultation and on people being in contact with health services. There will be variation between individual GPs and GP practices in how coding is conducted. There may have been misclassification of the exposure and outcome. In our study population AUD was determined from the presence of codes within primary care and hospital records. This will not include people who did not have a diagnosis of AUD recorded within primary care or in their hospital records due to lack of health service contact or under-recognition. Similarly, people who do not seek help for depression and anxiety or who did not have this coded will be misclassified as not developing the outcome. This may have led to bias in the findings if people with AUD are more or less likely to seek medical help for their depression and anxiety symptoms or to receive a diagnosis/code, for example clinicians may be less likely to give a diagnosis of depression or anxiety in people with AUD who are currently drinking due to current clinical guidance on treating the AUD first [39]. This may have led to an underestimate of the true rates of depression and anxiety in people with AUD. While we have used codelists which were reviewed by clinicians to mitigate measurement error it was beyond the scope of this study to formally validate these and there is the potential for misclassification of AUD, depression and anxiety. Furthermore only information recorded in the form of SNOMED-CT or ICD-10 codes was available for research purposes—information recorded as free text was not available for data protection purposes. This increases the potential for misclassification if GPs recorded information on mental health or alcohol use in free text but did not record an associated code in the EHR record.

Here we have considered AUD as a binary exposure at the time of T2DM diagnosis and have not considered alcohol use overall or variations in drinking pattern. We have also not captured changes in drinking over time. While we tried to capture whether people with AUD were likely to be currently drinking or not at the time of T2DM diagnosis, this was defined pragmatically based on what information was captured within the EHR data and misclassification for some people is likely. Residual confounding by unmeasured or incompletely recorded confounders is also a limitation when using these data. For example, there were missing data for some of the confounders (BMI, smoking status and ethnicity) and we were only able to adjust for area level and not individual socio-economic status. The same issues around potential misclassification and measurement are true for confounders as for depression, anxiety and AUD and even where present, smoking or BMI data may not be contemporary or accurate. In addition only smoking status and not a measure of amount smoked for example pack years was included. There are also many other factors like education, employment status, social networks and traumatic life events which may influence both AUD and depression and anxiety. However, these are not recorded in EHRs and therefore could not be adjusted for.

Public health relevance and future work

We have identified here that people with AUD are at higher risk of depression and anxiety following a diagnosis of T2DM than people without AUD. As well as the strong negative impact on people’s quality of life in terms of the distress caused by depression and anxiety, this may also lead to worse physical health outcomes, given existing evidence base showing depression and anxiety in diabetes are associated with poorer outcomes.

People with AUD are at high risk of developing multiple long term conditions [43] with negative impacts on their health and well-being. The magnitude of effect size for the association between AUD and depression was similar in our study to studies from the general population high-lighting the need for mental health support for all people with AUD regardless of whether they have T2DM. However given the additional complexities of living with T2DM future work is needed to understand whether more targeted interventions would better support the mental health needs of people with AUD and T2DM. This should include strategies for increasing integrated and holistic care which consider the specific issues for people living with both AUD and T2DM. Options should be explored for how best to do this within the range of services supporting people with AUD and T2DM including but not limited to primary care, alcohol recovery services, and peer support groups for both conditions. This study has shown in particular the importance for future work in this area of including mental health support as an integral component of care for people with AUD and T2DM.

Conclusions

Among people with T2DM, people with AUD have over double the rates of depression and anxiety following their diabetes diagnosis compared to those without AUD. This was only partially explained by pre-existing diagnoses of mental health conditions. A holistic approach incorporating mental health support is needed to improve health outcomes for people with AUD who develop T2DM.

Data availability

Data are available on request from the CPRD. Their provision requires the purchase of a license, and this license does not permit the authors to make them publicly available to all. This work used data from the version collected in May 2022 and have clearly specified the data selected within each Methods section. To allow identical data to be obtained by others, via the purchase of a license, the code lists will be provided upon request. Licenses are available from the CPRD (http://www.cprd.com): The Clinical Practice Research Datalink Group, The Medicines and Healthcare products Regulatory Agency, 10 South Colonnade, Canary Wharf, London E14 4PU.

Abbreviations

AUD:

Alcohol use disorder

BMI:

Body mass index

CI:

Confidence interval

CPRD:

Clinical practice research datalink

HER:

Electronic health record

GP:

General practice/general practitioner

HES APC:

Hospital episode statistics admitted patient care

ICD-10:

International classification of disease 10

IMD:

Index of multiple deprivation

IRR:

Incidence rate ratio

NHS:

National Health Service

NICE:

National Institute of Health and Clinical Excellence

ONS:

Office for National Statistics

SMI:

Severe mental illness

T2DM:

Type 2 Diabetes Mellitus

UK:

United Kingdom

US:

United States

References

  1. GBD 2019 Diseases and Injuries Collaborators. Global burden of 369 diseases and injuries in 204 countries and territories, 1990–2019: a systematic analysis for the Global Burden of Disease Study 2019. Lancet (London, England). 2020;396(10258):1204–22.

    Article  Google Scholar 

  2. Ong KL, Stafford LK, McLaughlin SA, Boyko EJ, Vollset SE, Smith AE, et al. Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: a systematic analysis for the Global Burden of Disease Study 2021. Lancet. 2023;402(10397):203–34.

    Article  Google Scholar 

  3. Knott C, Bell S, Britton A. Alcohol Consumption and the risk of type 2 diabetes: a systematic review and dose-response meta-analysis of more than 1.9 million individuals from 38 observational studies. Diabetes Care. 2015;38(9):1804–12.

    Article  CAS  PubMed  Google Scholar 

  4. Kao WHL, Puddey IB, Boland LL, Watson RL, Brancati FL. Alcohol consumption and the risk of type 2 diabetes mellitus: atherosclerosis risk in communities study. Am J Epidemiol. 2001;154(8):748–57.

    Article  CAS  PubMed  Google Scholar 

  5. Cao C, Wei C, Han Y, Luo J, Xi P, Chen J, et al. Association between excessive alcohol consumption and incident diabetes mellitus among Japanese based on propensity score matching. Sci Rep. 2024;14(1):17274.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  6. Baliunas DO, Taylor BJ, Irving H, Roerecke M, Patra J, Mohapatra S, et al. Alcohol as a risk factor for type 2 diabetes: a systematic review and meta-analysis. Diabetes Care. 2009;32(11):2123–32.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  7. Anderson RJ, Freedland KE, Clouse RE, Lustman PJ. The prevalence of comorbid depression in adults with diabetes: a meta-analysis. Diabetes Care. 2001;24(6):1069–78.

    Article  CAS  PubMed  Google Scholar 

  8. Mersha AG, Tollosa DN, Bagade T, Eftekhari P. A bidirectional relationship between diabetes mellitus and anxiety: a systematic review and meta-analysis. J Psychosom Res. 2022;162:110991.

    Article  PubMed  Google Scholar 

  9. Rathmann W, Kuß O, Anderson D, Busch S, Hahn M, Engelhard J, et al. Increased depression symptom score in newly diagnosed type 2 diabetes patients. Psychiatry Res. 2018;261:259–63.

    Article  PubMed  Google Scholar 

  10. Lin EHB, Korff MV. Mental disorders among persons with diabetes—results from the world mental health surveys. J Psychosom Res. 2008;65(6):571–80.

    Article  PubMed  PubMed Central  Google Scholar 

  11. Mezuk B, Chen Y, Yu C, Guo Y, Bian Z, Collins R, et al. Depression, anxiety, and prevalent diabetes in the Chinese population: findings from the China Kadoorie Biobank of 0.5million people. J Psychosom Res. 2013;75(6):511–7.

    Article  PubMed  PubMed Central  Google Scholar 

  12. Lloyd CE, Dyer PH, Barnett AH. Prevalence of symptoms of depression and anxiety in a diabetes clinic population. Diabet Med. 2000;17(3):198–202.

    Article  CAS  PubMed  Google Scholar 

  13. Lin EHB, Rutter CM, Katon W, Heckbert SR, Ciechanowski P, Oliver MM, et al. Depression and advanced complications of diabetes: a prospective cohort study. Diabetes Care. 2009;33(2):264–9.

    Article  PubMed  PubMed Central  Google Scholar 

  14. Iversen MM, Nefs G, Tell GS, Espehaug B, Midthjell K, Graue M, et al. Anxiety and depressive symptoms as predictors of all-cause mortality among people with insulin-naïve type 2 diabetes: 17-year follow-up of the second nord-trøndelag health Survey (HUNT2), Norway. PLoS One. 2016;11(8):e0160861.

    Article  PubMed  PubMed Central  Google Scholar 

  15. de Groot M, Anderson R, Freedland KE, Clouse RE, Lustman PJ. Association of depression and diabetes complications: a meta-analysis. Psychosom Med. 2001;63(4):619–30.

    Article  PubMed  Google Scholar 

  16. Naicker K, Johnson JA, Skogen JC, Manuel D, Øverland S, Sivertsen B, et al. Type 2 diabetes and comorbid symptoms of depression and anxiety: longitudinal associations with mortality Risk. Diabetes Care. 2017;40(3):352–8.

    Article  PubMed  Google Scholar 

  17. Winhusen T, Theobald J, Kaelber DC, Lewis D. Medical complications associated with substance use disorders in patients with type 2 diabetes and hypertension: electronic health record findings. Addiction. 2019;114(8):1462–70.

    Article  PubMed  Google Scholar 

  18. Leung G, Zhang J, Lin WC, Clark RE. Behavioral disorders and diabetes-related outcomes among Massachusetts Medicare and Medicaid beneficiaries. Psychiatric services (Washington, DC). 2011;62(6):659–65.

    Article  Google Scholar 

  19. Whitworth SR, Bruce DG, Starkstein SE, Davis WA, Davis TME, Bucks RS. Lifetime depression and anxiety increase prevalent psychological symptoms and worsen glycemic control in type 2 diabetes: The Fremantle Diabetes Study Phase II. Diabetes Res Clin Pract. 2016;122:190–7.

    Article  CAS  PubMed  Google Scholar 

  20. Shalaeva E, Bano A, Kasimov U, Atakov S, Mirakhmedova K, Dadabaeva N, et al. Depression and anxiety symptoms are underestimated risk factors for postoperative prognosis in patients with Type 2 diabetes and peripheral artery disease undergoing partial foot amputation: results from a prospective cohort study. J Psychosom Res. 2024;183:111824.

    Article  PubMed  Google Scholar 

  21. Han B, Wang L, Zhang Y, Gu L, Yuan W, Cao W. Baseline anxiety disorders are associated with progression of diabetic kidney disease in type 2 diabetes. Ren Fail. 2023;45(1):2159431.

    Article  PubMed  PubMed Central  Google Scholar 

  22. Cherrington A, Wallston KA, Rothman RL. Exploring the relationship between diabetes self-efficacy, depressive symptoms, and glycemic control among men and women with type 2 diabetes. J Behav Med. 2010;33(1):81–9.

    Article  PubMed  PubMed Central  Google Scholar 

  23. Gonzalez JS, Peyrot M, McCarl LA, Collins EM, Serpa L, Mimiaga MJ, et al. Depression and diabetes treatment nonadherence: a meta-analysis. Diabetes Care. 2008;31(12):2398–403.

    Article  PubMed  PubMed Central  Google Scholar 

  24. Ciechanowski PS, Katon WJ, Russo JE. Depression and diabetes: impact of depressive symptoms on adherence, function, and costs. Arch Intern Med. 2000;160(21):3278–85.

    Article  CAS  PubMed  Google Scholar 

  25. Liu X, Haagsma J, Sijbrands E, Buijks H, Boogaard L, Mackenbach JP, et al. Anxiety and depression in diabetes care: longitudinal associations with health-related quality of life. Sci Rep. 2020;10(1):8307.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  26. Iturralde E, Chi FW, Grant RW, Weisner C, Van Dyke L, Pruzansky A, et al. Association of anxiety with high-cost health care use among individuals with type 2 diabetes. Diabetes Care. 2019;42(9):1669–74.

    Article  PubMed  PubMed Central  Google Scholar 

  27. Lai HM, Cleary M, Sitharthan T, Hunt GE. Prevalence of comorbid substance use, anxiety and mood disorders in epidemiological surveys, 1990–2014: a systematic review and meta-analysis. Drug Alcohol Depend. 2015;154:1–13.

    Article  PubMed  Google Scholar 

  28. Boden JM, Fergusson DM. Alcohol and depression. Addiction. 2011;106(5):906–14.

    Article  PubMed  Google Scholar 

  29. Li J, Wang H, Li M, Shen Q, Li X, Zhang Y, et al. Effect of alcohol use disorders and alcohol intake on the risk of subsequent depressive symptoms: a systematic review and meta-analysis of cohort studies. Addiction. 2020;115(7):1224–43.

    Article  PubMed  Google Scholar 

  30. Wolf A, Dedman D, Campbell J, Booth H, Lunn D, Chapman J, et al. Data resource profile: Clinical Practice Research Datalink (CPRD) Aurum. Int J Epidemiol. 2019;48(6):1740.

    Article  PubMed  PubMed Central  Google Scholar 

  31. Clinical Practice Research Datalink. CPRD Aurum May 2022 (Version 2022.05.001). In: Clinical Practice Research Datalink, editor. https://doiorg.publicaciones.saludcastillayleon.es/10.48329/t89s-kf122022.

  32. Clinical Practice Research Datalink. CPRD Aurum HES APC January 2022 (Version 2022.01.001). In: Clinical Practice Research Datalink, editor. https://doiorg.publicaciones.saludcastillayleon.es/10.48329/VAGX-9D962022.

  33. Clinical Practice Research Datalink. CPRD Aurum Small Area data (Patient) January 2022 (Version 2022.01.001). In: Clinical Practice Research Datalink, editor. https://doiorg.publicaciones.saludcastillayleon.es/10.48329/AYTT-H2222022.

  34. Mansfield KE, Mathur R, Tazare J, Henderson AD, Mulick AR, Carreira H, et al. Indirect acute effects of the COVID-19 pandemic on physical and mental health in the UK: a population-based study. Lancet Digit Health. 2021;3(4):e217–30.

    Article  CAS  PubMed  PubMed Central  Google Scholar 

  35. World Health Organization. International statistical classification of diseases and related health problems. 10th revision, Fifth edition, 2016 ed. Geneva: World Health Organization; 2015 2015.

  36. Metcalfe D, Masters J, Delmestri A, Judge A, Perry D, Zogg C, et al. Coding algorithms for defining Charlson and Elixhauser co-morbidities in Read-coded databases. BMC Med Res Methodol. 2019;19(1):115.

    Article  PubMed  PubMed Central  Google Scholar 

  37. Mathur R, Bhaskaran K, Chaturvedi N, Leon DA, vanStaa T, Grundy E, et al. Completeness and usability of ethnicity data in UK-based primary care and hospital databases. J Public Health (Oxf). 2014;36(4):684–92.

    Article  PubMed  Google Scholar 

  38. Hatam S, Scully ST, Cook S, Evans HT, Hume A, Kallis C, et al. Harmonised Approach to curating research-ready datasets for asthma, Chronic Obstructive Pulmonary Disease (COPD) and Interstitial Lung Disease (ILD) in England, Wales and Scotland Using Clinical Practice Research Datalink (CPRD), Secure Anonymised Information Linkage (SAIL) Databank and DataLoch. Clin Epidemiol. 2024;16:235–47.

    Article  PubMed  Google Scholar 

  39. National Institute for Health and Care Excellence (NICE). Alcohol-use disorders: diagnosis, assessment and management of harmful drinking (high-risk drinking) and alcohol dependence https://www.nice.org.uk/guidance/cg1152014.

  40. Wu L-T, Ghitza UE, Batch BC, Pencina MJ, Rojas LF, Goldstein BA, et al. Substance use and mental diagnoses among adults with and without type 2 diabetes: results from electronic health records data. Drug Alcohol Depend. 2015;156:162–9.

    Article  PubMed  PubMed Central  Google Scholar 

  41. Eaton WW, Shao H, Nestadt G, Lee HB, Bienvenu OJ, Zandi P. Population-based study of first onset and chronicity in major depressive disorder. Arch Gen Psychiatry. 2008;65(5):513–20.

    Article  PubMed  PubMed Central  Google Scholar 

  42. Ummels SA, Seldenrijk A, Bos EH, de Graaf R, Batelaan NM, ten Have M. The bidirectional relationship between anxiety disorders and alcohol use disorders in adults: findings from a longitudinal population-based study. J Affect Disord. 2022;314:126–32.

    Article  PubMed  Google Scholar 

  43. Rehm J, Gmel GE Sr, Gmel G, Hasan OSM, Imtiaz S, Popova S, et al. The relationship between different dimensions of alcohol use and the burden of disease—an update. Addiction. 2017;112(6):968–1001.

    Article  PubMed  PubMed Central  Google Scholar 

Download references

Acknowledgements

We would like to thank Melinda King for support with facilitating public involvement work informing this publication and all the public contributors who took part in shaping this research.

Funding

SC was funded by an NIHR Three Research Schools Mental Health Fellowship for this work. This Fellowship (MH055) was funded as part of the Three NIHR Research Schools Mental Health Programme. SS is funded by the National Institute for Health and Care Research (NIHR) Senior Investigator Award, NIHR School for Public Health Research (SPHR) (grant number: NIHR 204000), NIHR Northwest London Applied Research Collaboration, and Imperial NIHR Biomedical Research Centre. The NIHR SPHR is a partnership between the Universities of Bristol, Cambridge, and Sheffield; Imperial; University College London; the London School of Hygiene and Tropical Medicine; LiLaC — a collaboration between the Universities of Liverpool and Lancaster; and Fuse — the Centre for Translational Research in Public Health, a collaboration between Newcastle, Durham, Northumbria, Sunderland, and Teesside Universities. RM is supported by Barts Charity (MGU0504). TB is supported by a grant from the Wellcome Trust. SG is funded by the NIHR School for Public Health Research and NIHR Northwest London Applied Research Collaboration. ALN is funded by the NIHR Northwest London Applied Research Collaboration and NIHR Northwest London Patient Safety Research Collaboration. Infrastructure for this research was supported by the NIHR Imperial Biomedical Research Centre (BRC). The views expressed are those of the author(s) and not necessarily those of the NIHR or the Department of Health and Social Care.

Author information

Authors and Affiliations

  1. School of Public Health, Imperial College London, London, UK

    Sarah Cook, Ravi Parekh, Arti Maini, Ana Luisa Neves, Shamini Gnani, Thomas Beaney, Sonia Saxena & Jennifer K. Quint

  2. Division of Psychiatry, University College London, London, UK

    David Osborn

  3. Camden and Islington NHS Foundation Trust, London, UK

    David Osborn

  4. Wolfson Institute of Population Health, Queen Mary University of London, London, UK

    Rohini Mathur

  5. Department of Non-Communicable Disease Epidemiology, Faculty of Epidemiology and Population Health, London School of Hygiene and Tropical Medicine, London, UK

    Harriet Forbes

  6. Primary Care and Population Health, University College London, London, UK

    Kate Walters

Authors
  1. Sarah Cook
    View author publications

    You can also search for this author inPubMed Google Scholar

  2. David Osborn
    View author publications

    You can also search for this author inPubMed Google Scholar

  3. Rohini Mathur
    View author publications

    You can also search for this author inPubMed Google Scholar

  4. Harriet Forbes
    View author publications

    You can also search for this author inPubMed Google Scholar

  5. Ravi Parekh
    View author publications

    You can also search for this author inPubMed Google Scholar

  6. Arti Maini
    View author publications

    You can also search for this author inPubMed Google Scholar

  7. Ana Luisa Neves
    View author publications

    You can also search for this author inPubMed Google Scholar

  8. Shamini Gnani
    View author publications

    You can also search for this author inPubMed Google Scholar

  9. Thomas Beaney
    View author publications

    You can also search for this author inPubMed Google Scholar

  10. Kate Walters
    View author publications

    You can also search for this author inPubMed Google Scholar

  11. Sonia Saxena
    View author publications

    You can also search for this author inPubMed Google Scholar

  12. Jennifer K. Quint
    View author publications

    You can also search for this author inPubMed Google Scholar

Contributions

Conception and design of the study (SC, DO, KW, SS, JKQ); Statistical analysis (SC); Methodology (SC; RM, HF, DO, SS, JKQ, RP, AM, ALN, SG, TB). SC drafted the initial manuscript to which all authors contributed with substantial revision. All authors read and approved the final manuscript.

Corresponding author

Correspondence to Sarah Cook.

Ethics declarations

Ethics approval and consent to participate

CPRD has NHS Health Research Authority (HRA) Research Ethics Committee (REC) approval to allow the collection and release of anonymised primary care data for observational research [NHS HRA REC reference number: 05/MRE04/87]. Each year CPRD obtains Sect. 251 regulatory support through the HRA Confidentiality Advisory Group (CAG), to enable patient identifiers, without accompanying clinical data, to flow from CPRD contributing GP practices in England to NHS Digital, for the purposes of data linkage [CAG reference number: 21/CAG/0008]. The protocol for this research was approved by CPRD’s Research Data Governance (RDG) Process (protocol number: 22_002317) and the approved protocol is available upon request. Generic ethical approval for observational research using CPRD with approval from RDG was granted by a Health Research Authority (HRA) Research Ethics Committee (East Midlands Derby, REC reference number: 05/MRE04/87) and informed consent was waived. All data analysis was performed in accordance with the Declaration of Helsinki. Linked pseudonymised data was provided for this study by CPRD. Data is linked by NHS Digital, the statutory trusted third party for linking data, using identifiable data held only by NHS Digital. Select general practices consent to this process at a practice level with individual patients having the right to opt-out.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Cook, S., Osborn, D., Mathur, R. et al. Is alcohol use disorder associated with higher rates of depression and anxiety among people with new onset type 2 diabetes? A cohort study using linked primary care data in England. BMC Prim. Care 25, 386 (2024). https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12875-024-02628-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doiorg.publicaciones.saludcastillayleon.es/10.1186/s12875-024-02628-6

Keywords

BMC Primary Care

ISSN: 2731-4553

Contact us