The Relationship between Selected Factors (Temperament, Bipolar Traits, Sleep Quality, Severity of Addiction) and Depressive Symptoms in Alcohol-Dependent Men

Clinical and epidemiological studies have demonstrated a relationship between alcohol addiction and mood disorders. Alcohol-dependent patients with depression tend to demonstrate clinically more severe manic symptoms, which complicates the process of diagnosis and therapy. However, the predictors indicating the risk of mood disorders in addicted patients remain unclear. The aim of the study was to examine the relationship between personal dispositions, bipolar traits, depth of addiction, quality of sleep, and depressive symptoms in alcohol-dependent men. The study group comprised 70 men (age M = 46.06, SD = 11.29) diagnosed with alcohol addiction. The participants completed a battery of questionnaires: BDI, HCL-32, PSQI, EPQ-R and MAST. The results were tested using Pearson’s correlation quotient and general linear model. The findings indicate that some of the studied patients are likely to have mood disorders of clinically significant severity. High neuroticism and poor sleep quality are independent predictors of depressive symptoms in alcohol-dependent patients. Among the components of sleep quality, problems with falling asleep and waking up at night appear to be most strongly associated with depressive symptoms. The intensity of depressive symptoms may relate to the intensity of certain bipolar features, such as risk-taking activity and irritability. High neuroticism and poor sleep quality are independent predictors of depressive symptoms in the studied group.


Introduction
Affective disorders and alcohol use disorder (AUD) are significant comorbidities [1][2][3][4]. Among people with AUD, as much as 50% meet the criteria of mood disorders [5,6]. Patients with such dual diagnosis are often described as clinically more severe, with complex courses of both disorders; as such, treatment can be more difficult than for patients with only one medical diagnosis [7,8]. The clinician must not only treat the addiction, typically the primary problem faced by the patient, but also the accompanying depression, especially when it is severe. In most cases, the patient is treated for one disorder before receiving treatment for the other [9]. This approach follows from the long-standing clinical The research was carried out among consenting patients who had maintained alcohol abstinence for at least three weeks: this criterion negated the influence of alcohol abstinence syndrome symptoms on the study results. The exclusion criteria comprised the following: addictions to psychoactive substances other than alcohol and nicotine; serious mental disorders such as schizophrenia, depression, bipolar affective disorder, delirium tremens, alcohol-related psychosis, insomnia, or neurological disorders; somatic diseases such as malignant tumors, diabetes, hepatic cirrhosis, or ischemic heart disease, which may result in symptom overlap or measurement bias. M-mean value, SD-standard deviation, Min-minimal value, Max-maximal value.
All patients were investigated during the third week of psychotherapy. Informed consent was obtained from the participants after the nature of the procedures had been fully explained. The raters were psychologists trained to use standardized questions and specific instructions. A total number of 63 patients entered the study: three patients were rejected from the statistical analysis because of incomplete questionnaires, and another four because of the co-occurrence of other addictions and mental illness.

Methods
The study uses Polish adaptations of the battery of well-known psychometric questionnaires, all of which are commonly used in clinical practice, psychological diagnosis, and for the purpose of scientific research.
The Polish version of the Eysenck Personality Questionnaire-Revised (EPQ-R), adapted by Brzozowski and Drwal [44], was used to assess personality in three main dimensions, according to Eysenck's PEN theory: Psychoticism, Extraversion, and Neuroticism. The model also includes an additional Lie scale, measuring self-deception and impression adjustment. Although a newer version of the EPQ-R is present on the Polish market [45], the older version was chosen for the study due to its known suitable psychometric values for scientific research (Cronbach's α was N-0.84, E-0.83, P-0.67, K-0.75) [44].
The severity of alcohol-related problems was assessed using the Michigan Alcoholism Screening Test (MAST) diagnostic questionnaire by Selzer well described in scientific works [46], adapted into Polish by Falicki et al. [47].
Depressive symptoms over the month prior to examination were assessed using a self-reported method-Beck Depression Index (BDI) [48,49]. The Polish adaptation was prepared by Zawadzki et al. [50], and Cronbach's α was 0.95; in accordance with their suggestion, a cut-off value of 17 points was considered suggestive of clinically significant depressive symptoms for the Polish population.
To assess bipolar features, the Polish version of the Hypomania Checklist 32 (HCL-32) by [51], validated by Łojko et al. [52], including subscale assessed bipolar features connected with irritation and risk-taking-Subscale irr/ri-ta HCL-32 and subscale act/ela HCL-32bipolar features connected with activity/elated and elevated mood was used. Cronbach's alfa of Polish validation was 0.93, and a score of 14 or higher was considered as indicating bipolar features [52].
Sleep quality for the four weeks prior to the study was assessed using the Pittsburgh Sleep Quality Index (PSQI) which includes components: Subjective Sleep Quality, Sleep Duration, Sleep Latency, Habitual Sleep Efficiency, Use of Sleeping Medication, Sleep Disturbances (interruptions), and Daytime Dysfunction. The score is calculated on the basis of the respondent's answers and each question is scored from 0 to 3 points, with Cronbach's alpha of 0.82 [53]. A Polish translation is available via the website of the Centre of Sleep Medicine at Institute of Psychiatry and Neurology in Warsaw [54]. To the best of our knowledge, no validation has been published for the Polish population. The cutoff for poor sleep quality was regarded as five points, in accordance with the original version. Additionally, the questionnaire considering sociodemographic status designed by the authors.
Operationalization of Variables is Presented in Table 2. Table 2. Presentation of dependent and independent variables included in the analysis, together with their indicators.

Variable Index
Self-assessed depressive symptoms intensity BDI score Self-assessed bipolar features intensity HCL-32 score Self-assessed bipolar features connected with irritation and risk-taking Subscale irr/ri-ta HCL-32 Self-assessed bipolar features connected with activity/elated and elevated mood Subscale act/ela HCL-32 Depth of alcohol dependence (the severity of alcohol-related problems) MAST score

Data Analyses
The data were analyzed with STATISTICA 12 PL (Kraków, Poland and Tulsa, OK, USA). Student's t-test was used for comparisons between two independent groups. Pearson's correlation coefficient was used to assess the relationship between two continuous variables, and general linear models were used to assess the effect of independent variables on the severity of depressive symptoms. The Holm-Bonferroni correction for multiple testing was utilized. Level of significance was adopted for alpha = 0.05. A post-hoc power analysis was performed with G*Power 3.1 software (Supplementary Materials).

The Prevalence of Self-Reported Depressive Symptoms, Bipolar Features, and Poor Sleep Quality in a Study Group of Addict Men
It was found that 28.6% of the patients (N = 18) achieved a BDI score suggestive of clinically significant depressive symptoms (i.e., at least 17 points). In addition, 83% of the investigated group (N = 52) fulfilled the bipolarity criterion, based on an HCL-32 score of 14 points or more, and 47.6% of the patients (N = 30) reported poor sleep quality (at least 5 points in PSQI).

Comparison between Groups-Depression vs. No Depression
A result of 17 or more on the BDI was considered as an indicator of clinicallyrelevant self-described depressive symptoms; this was the cut-off point proposed by Zawadzki et al. [50] for the Polish population. In the study group, the mean BDI score was 11.94 (±8.97). Eighteen participants were classified as demonstrating depression based on their BDI score. The patients with BDI scores more than 17 achieved higher results in the EPQ-R Neuroticism scale (raw results and stens) and the PSQI Questionnaire and its dimensions (subjective sleep quality, sleep latency, sleep interruptions).
No statistically significant differences in HCL-32 score or subscale act/ela were found between patients with BDI < 17 and those with BDI > 17; however, significantly higher subscale irr/ri-ta scores were noted in patients with depressive symptoms (i.e., those with BDI > 17) (see Table 3). Table 3. Comparison of means (M) and standard deviations (SD) in EPQ-P, MAST, HCL-32 and PSQI questionnaires between patients with AD and depression (N = 18) and without depression (N = 45). This division was based on a BDI cut-off value of 17.

Correlations
Selected correlation quotients confirm the findings given above. More precisely, statistically significant correlations were observed between depressive symptom intensity (BDI score) and neuroticism intensity (Neuroticism scale of EPQ-R score) and between BDI score and bipolar feature intensity (HCL-32), especially with regard to irr/ri-ta. No statistically significant correlation was found between BDI and MAST scores. More detailed results are shown in Table 4.
A significant positive correlation was found between HCL-32 irr/ri-ta score and the intensity of neuroticism assessed by the EPQ-R.
The intensity of depressive symptoms was also found to be associated with the intensity of bipolar features, such as risk activity taking and irritability, in the studied group; however, no such relationship was demonstrated for the "increased activity and elevated mood" component.
Of all the assessed components of sleep quality, it was found that problems falling asleep and waking up at night were most closely related to depressive symptoms. Hence, from a clinical perspective, these could be valuable topics for specialists to raise with patients with alcohol addiction to assess the risk of mood disorders.

Linear Regression Models Predicting the Severity of Depressive Symptoms
First, a logistic regression model was constructed to assess the influence of selected independent variables (age, years of addiction, EPQ-R subscales score, MAST score, HCl-32 subscales scores, PSQI score) on the intensity of depressive symptoms (BDI score) in the studied group. The results indicate that the only independent predictors were neuroticism and sleep quality, with a stronger relationship being demonstrated for sleep quality.
The coefficient of determination (R 2 ) for Model 1 is 0.526: this model explained 53% of variance of the BDI score in the studied group. The calculated post-hoc power of the model was 1.0. More detailed results of the hierarchical model are given in Table 5.
Since sleep quality is such a relatively strong and independent predictor, it is possible to check the effect of the PSQI and its individual components on the severity of depressive symptoms independent of the other effects. All PSQI dimensions correlated with BDI scores in both the high BDI (>17) and the low BDI (<17) groups and in correlation analysis. However, when taking into account the simultaneous effect of all on BDI score, the only predictor of an increase in BDI score was found to be an increase in sleep disturbances (interruptions in sleep, i.e., waking up at night due to different factors).
The coefficient of determination (R 2 ) for Model 2 is 0.361: this model explained 36% of the variance of the BDI score in the studied group. Calculated post-hoc power of the model was 0.999. More detailed results of the hierarchical model are given in Table 6.

Discussion
Our findings indicate that some of the alcohol-dependent patients (28.6%) are likely to experience clinically-significant depression. However, a large number of previous studies indicate a more significant likelihood of comorbidity of AUD and depression, reaching as high as 68% [5,55,56]. The disparity between our findings and those of previous studies may be due to our exclusion of patients with a diagnosis of mood disorders, and of those who failed to maintain alcohol abstinence for at least three weeks before the study.
The nature of depression may differ between individuals with alcohol dependence and those without; therefore, categorizing depression as independent (ID), i.e., in the course of mood disorders, or substance-induced (SID) could be a useful model for predicting the course of depression in alcohol use disorder (AUD) [57,58]. Substance-induced depression is thought to account for a substantial proportion of major depressive episodes among patients with AUD [59] and usually improves after two to three weeks of abstinence [60]; in contrast, independent depression does not necessarily remit with abstinence. Therefore, a period of abstinence of at least three weeks in patients without independent depression may result in total or partial withdrawal of depressive symptoms: this could explain the relatively low ratio of co-occurrence in the studied group.
Our findings also indicate that as many as 75% of patients with AUD fulfil the criterion of bipolarity. However, very little data has been gathered for the assessment of bipolarity in samples of alcohol-dependent patients. One of the few studies completed to date suggests that 6.5% of men with AUD have also suffered from at least one episode of mania [61]. Most of the previous research in this area has examined co-occurring alcohol abuse as a consequence of bipolar disorder [1,11]. However, alcohol-dependent patients currently in abstinence without diagnosis of bipolar disorder have been found to demonstrate high clinical severity of bipolar traits. This suggests that AUD withdrawal may prompt bipolarity symptoms. As such, this pathway of relationship requires further studies.
In the studied group, the intensity of depressive symptoms may be connected with the intensity of certain bipolar features, such as risk-taking activity and irritability. No such relationship has been demonstrated for the increased activity and elevated mood component. However, it should be noted that risk activity taking, and irritability are not independent of the structure of temperament, depth of addiction, duration of addiction, and quality of sleep. This subject therefore requires further study.
Our data also indicates that high neuroticism and poor sleep quality are independent predictors of depressive symptoms in alcohol-dependent patients. This is in line with previous findings that both of these indicators represent risk factors for depression [35,[62][63][64]. More interestingly, however, from the point of view of our study, they are predictors of depression, regardless of the depth and duration of addiction; furthermore, they have been found to have no effect on the severity of depressive symptoms, independently of neuroticism and PSQI.
Insomnia is considered a risk factor of and a symptom of major depressive disorder, with the latter being reflected in the content of the diagnostic criteria [65,66]. Thus, the correlation between PSQI score and BDI score is not surprising in the studied group. However, testing the association between PSQI components and BDI score provided an answer to the question of which specific characteristics of sleep symptoms (including insomnia) might play the most important role in its link to depression among patients with addiction to alcohol. Among the studied components of sleep quality, problems with falling asleep and waking up at night were found to be most strongly associated with depressive symptoms in the study population. Hence, from a clinical perspective, these could be valuable topics for specialists to raise with patients with alcohol addiction to assess the risk of mood disorders.
There are some limitations inherent in the study. The lack of gender and racial diversity or relatively small sample size means that the sample may not be a truly representative, and the results of the study cannot be generalized across the entire alcohol addict men population. When analyzing the results obtained, further research is needed to establish to what extent results related to bipolar traits are really bipolar related or are more general personality traits (e.g., irritability or impulsivity). The same applies to traits that have previously been linked to alcohol (ab)use and sleep.
A limitation of this study was also the sleep quality assessment: this was based on selfreport only, and as such, may include some subjectivity bias. However, previous studies have demonstrated that self-reported data is reliable [67]. In addition, the PSQI itself has good psychometric properties. In future studies, it may be worth considering expanding the scope of research by using actigraphy to validate physical activity and sleep/wake measurements. Due to the relatively rapid withdrawal of depressive symptoms after the onset of alcohol abstinence, its severity should be assessed both during an episode of drinking and during abstinence. Additionally, a comparison group of non-addicted individuals was not included. Yet, it should be considered the next research step to further answer the question whether risk factors delineated in this study are specific for the study group or reflect a link in a general population.

Conclusions
High neuroticism and poor sleep quality are independent predictors of depressive symptoms in alcohol-dependent patients. Among the components of sleep quality, problems with falling asleep and waking up at night appear to be most strongly associated with depressive symptoms. The intensity of depressive symptoms may relate to the intensity of certain bipolar features, such as risk-taking activity and irritability. High neuroticism and poor sleep quality are independent predictors of depressive symptoms in the studied group. This is the first study to investigate the correlations between personal dispositions, mood disorders, and sleep quality among people with alcohol addiction. Although results are still preliminary and derived from a small population, they nevertheless have certain practical implications. Personality and sleep quality assessments may allow more accurate and faster identification of patients with alcohol dependency who may be at higher risk of mood disorders.