Study population

The sample in this study were Bhutanese adult patients with AUD admitted for alcohol detoxification in the psychiatric ward, fulfilling the following inclusion and exclusion criteria.

Inclusion criteria

Patient diagnosed as AUD and admitted for detoxification were included. Patients with AUD were included only once during the study period to avoid redundancy associated with multiple hospitalizations of an individual patient.

Exclusion criteria

Excluded were those patients with existing co-morbid psychiatric illnesses apart from MDD. This is because patients with comorbid schizophrenia and other psychotic illnesses, bipolar affective disorder, other substance use disorders apart from alcohol, personality disorders, and intellectual disability disorders could present with symptoms overlapping major depressive disorder, or their presentations could mask the symptoms of major depressive disorder.

Sample size

The sample size in this study was calculated using the Krejcie and Morgan formula¹⁹ to determine the sample size for a finite population. After substituting the values into the formula, the required sample size for this study was 153 participants. A total of 155 participants were recruited over the study period.

Sampling method and study procedure

All the patients fulfilling the criteria from the Diagnostic and Statistical Manual of Mental Disorders, fifth edition (DSM-5) AUD, who were admitted for alcohol detoxification during the specified time frame, were recruited for the study after obtaining informed written consent.

The principal investigator assessed patients admitted for alcohol detoxification through a detailed psychiatric history and selected only those patients fulfilling the inclusion criteria. Eligible participants were interviewed by the principal investigator after 5th detoxification day and administered the PHQ-9 and sociodemographic questionnaire. This process avoided confounding assessment factors of acute intoxication, delirium, sedation and transient cognitive impairment from withdrawal or medication burden.

Research instruments

The research instruments used in this study were a sociodemographic questionnaire and Patient Health Questionnaire (PHQ)-9.

Sociodemographic questionnaire

A self-designed questionnaire including basic demographic variables such as age, gender, marital status, education level, employment status, annual income, family history of psychiatric illness and family history of alcohol and substance use disorder, was employed in gathering the demographic profile of the participants.

The Patient Health Questionnaire (PHQ)-9

The PHQ-9 is a self-administered questionnaire designed to screen for major depressive disorder and its severity. Its 9-item questions coincide with the 9 symptoms of major depressive disorder described in the DSM-5 (Supplementary file).

A summary score ranging from 0–27 was calculated. Each of the 9 items is scored on a scale of 0 to 3: Not at all=0, Several days=1, More than half the days=2, and Nearly every day=3.

The severity of major depressive disorder is measured as follows: total score of 0–4=normal, 5–9=mild major depressive disorder, 10–14=moderate major depressive disorder, 15–19= moderately severe major depressive disorder, and 20–27=severe major depressive disorder.

Participants with PHQ-9 scores of ≥10 were considered to meet the criteria for MDD, as the sensitivity and specificity for PHQ-9 scores ≥10 is 88% for the diagnosis of MDD²⁰. Participants with mild depression (PHQ-9 score of 5–9 were not included in the assessment of the overall prevalence.

Data management and analysis

Data were entered, double entered and validated using EpiData Entry version 3.1. The data were analyzed using statistical software (SPSS version 21). Descriptive statistics (frequency, percentage, mean and standard deviation (SD) were used to present the sociodemographic characteristics. Multivariate binary logistic regression was carried out to determine factors associated with depression using Model 1 and Model 2. All independent variables were computed for multivariate logistic regression in Model 1. For Model 2, variables with p<0.2 from Model 1 (gender, family history of psychiatric illness and family history of alcohol and drug abuse) were included. Variables with p>0.2 from Model 1 (age, marital status, education level, employment status, and income per year) were excluded. Akaike information criterion (AIC) score was used to construct our final multivariable model.

Sociodemographic characteristics

Table 1 describes the sociodemographic characteristics of the participants. The majority of the participants were male (76.1%) aged 30–39 years (mean=38.07; SD=9.71). More than half (57.4%) of the participants were married, and 75.5% of them had attended formal education. Most of the participants were employed (76.8%), and their annual income was >100000 BTN (1000 Bhutan Ngultrum about 12 US$); 7.7% of the participants had a family history of psychiatric illness, and 53.5% reported having a family history of alcohol and drug abuse.

Prevalence and severity of MDD in those with AUD

The overall prevalence of MDD (PHQ-9 score ≥10) among the participants was 38% (59/155). According to the PHQ-9 score, 29.7% of the participants had mild depression (PHQ-9 score = 5–9), 23.3% of the participants had moderate depression (PHQ-9 score = 10–14), 10.3% of the participants had moderately severe depression (PHQ-9 score = 15–19) and 4.5% of the participants had severe depression (PHQ-9 score = 20–27). The cut-off score for major depressive disorder is ≥10, although scores from 5–9 are considered as mild depression. Mild depression does not require active interventions, but they can be followed up and monitored for an increase in their subsequent PHQ-9 scores, so participants with mild depression are not included in the overall prevalence.

As shown in Table 1, female (54.05%) participants had a higher prevalence of depression than male (33.05%) (p=0.02). The prevalence of depression was significantly higher among participants with a family history of psychiatric illness (p=0.01) and participants with a family history of alcohol and substance use (p=0.03) compared to their counterparts.

Risk factors associated with depression

In unadjusted analysis (univariate binary logistic regression), female participants (OR=2.38; 95% CI: 1.12–5.05), participants with a family history of psychiatric illness (OR=5.58; 95% CI: 1.44–21.54) and those with a family history of alcohol and drug abuse (OR=2.05; 95% CI: 1.05–3.99) were significantly associated with depression compared to their counterparts as shown in Table 2.

Table 3 shows factors associated with depression among the study participants through multivariate logistic regression analysis. All independent variables were computed for multivariate logistic regression in Model 1. For Model 2, only those variables with p<0.2 from Model 1 were included. After adjusting for confounding factors, the findings show that females had a more than two times (AOR=2.19; 95% CI: 1.01–4.75) likelihood of developing depression than male patients with AUD. In addition, participants with a family history of psychiatric illness had nearly five times (AOR=4.63; 95% CI: 1.17–18.44) the likelihood of developing of depression compared to those without a family history of psychiatric illness.

Prevalence of MDD in those with AUD

The association between depression and AUD is well established, and each disorder is a risk factor for the other. This study shows the occurrence of depression in AUD in the studied population to have an overall prevalence of 38%. This prevalence rate is consistent with other studies^6,7,21-23. Studies conducted by Becker et al.^17,24,25, Cho et al.^17,24,25 in Korea, and Bott et al.^17,24,25 in Germany, however, reported lower prevalence rates. These differences could be due to different study designs and sample size, psychometric tools used in the study and different timing of the administration of the study tools. Studies conducted by Luitel et al.^22,26 in Nepal and Charnsil and Aroonrattanapong²⁶ in Thailand using the same study tool (PHQ-9) demonstrated that the prevalence of MDD in AUD was around 40% and 32.2% respectively. Mchugh and Weiss⁶ reported that AUD patients are 3.7 times more likely to have MDD whereas Becker et al.¹⁷ reported that the presence of alcohol dependence increases the risk for depression by 5 times (OR=4.7).

Alcohol is a central nervous system depressant, and prolonged alcohol use could precipitate symptoms of depression. Patients with depression may be self-medicating with alcohol to alleviate their symptoms of depression. Maladaptive coping skills in depressed patients could result in using alcohol as a coping strategy. Furthermore, social and environmental stressors could trigger both AUD and MDD simultaneously. MDD could also result from the consequences of alcohol through its effect on an individual’s social, occupational and financial life. Finally, both disorders could be independent of each other and yet may present as comorbidity¹⁰.

According to the literature, AUD and MDD share bidirectional causality^7,14. In Bhutan, the prevalence of AUD was reported to be high³. Hence, a possible reason for the high prevalence of MDD among our participants could be linked to a high prevalence of AUD among the Bhutanese population. Further, high prevalence of AUD in Bhutan is attributed to easy access of alcohol, its cultural acceptance in routine religious rituals and social gatherings, and its relative inexpensiveness²⁷.

The validity of PHQ-9 was not checked in the Bhutanese context; however, similar studies using PHQ-9 have been done in India and Nepal. The translated and validated version of PHQ-9 used in Nepal had a sensitivity of 94% and specificity of 88% in diagnosing depression with a PHQ-9 score >10^28,22. It has also been used in India for similar studies^29,30. Bhutan shares its geographical boundaries with the two countries, and there are also some sociocultural, ethnic and religious similarities between these countries.

The co-occurrence infers that there are two deadly co-occurring psychiatric problems with multiple detrimental consequences. Some of the consequences, as explained in many studies, are a higher risk of delayed diagnosis, more severe psychopathological symptoms, less compliance with treatment, poorer effects of treatment, more impairment of social functioning, increased admissions to emergency department, and higher prevalence of physical comorbidity^10,17,18. In addition, it can also cause a significant economic burden to society due to high levels of healthcare consumption, inadequate treatment outcomes, high work absenteeism, and lost productivity¹⁸.

Gender differences in MDD in those with AUD

The majority of the participants in this study were males (76.1%). AUD is more common in males than females³¹. Males are more readily exposed to alcohol than females. They tend to drink more frequently and in greater amounts and this behavior is partly enforced by cultural factors, stigma against women who are drinking alcohol or abusing substances, and underrepresentation of females with AUD in research^31,32. However, evidence suggests that the prevalence of AUD is increasing in females and is associated with greater severity of consequences of alcoholism^31,32.

In the current study female participants were more than two times likely to develop MDD (AOR=2.19; 95% CI: 1.01–4.75). A similar finding was demonstrated in Thailand by Suttajit et al.³³ where females with AUD had increased risk for MDD (OR=4.09; 95% CI: 2.31–7.26) compared to their male counterparts (OR=2.49; 95% CI: 1.76–3.53). There are epidemiological studies supported by genetic studies which suggest that the comorbidity of depression and AUD is more prevalent among females³⁴. Twin studies have also shown a strong association between depression and AUD in females³⁴.

Family history as a risk factor for MDD in those with AUD

According to the present study, participants with a positive family history of psychiatric illness were nearly 5 times more likely to develop MDD (AOR=4.63; 95%CI: 1.17–18.44). This occurrence could be possibly due to genetics. According to Edwards et al.³⁵, genetic hereditability and the occurrence of comorbidity of depression and AUD have been established by family, twin and adoption studies, and the genetic influences account for at least 50% of their phenotypic association. It is associated with earlier onset and recurrent depression, which is greater in severity⁹. People with alcohol use problems who attempt suicide also tend to have a family history of alcoholism³².

Considering the high prevalence of MDD among AUD patients in our study, healthcare workers (HCWs) across the country are recommended to routinely screen for MDD when they are providing alcohol detoxification to AUD patients. Identified patients with MDD should be administered proper treatment or referrals to higher health centers when indicated. Further, HCWs should focus more on females and people with a family history of psychiatric illness as per our findings.

Strengths and limitations

This study is the first study known to date to examine the prevalence of MDD in AUD patients in Bhutan. The findings of the study could be used as a baseline prevalence for future studies. Owing to the study design, we could not determine the temporal nature and could not establish the causality of the two disorders. Moreover, we recommend conducting further studies to determine the temporal nature and to find out the association between the risk factors identified in this study.