The Mutual Relationship Between Men’s Drinking and Depression: A 4-Year Longitudinal Analysis

The Mutual Relationship Between Men’s Drinking and Depression: A 4-Year Longitudinal Analysis Abstract Aims The purpose of the current study was to examine the longitudinal reciprocal relationship between depression and drinking among male adults from the general population. Methods This study used a panel dataset from the Korean Welfare Panel (from 2011 to 2014). The subjects were 2511 male adults aged between 20 and 65 years. Based on the Korean Version of the Alcohol Use Disorders Identification Test (AUDIT-K) scores, 2191 subjects were categorized as the control group (AUDIT-K < 12) and 320 subjects were categorized as the problem drinking group (AUDIT-K ≥ 12). An autoregressive cross-lagged modelling analysis was performed to investigate the mutual relationship between problem drinking and depression measured consecutively over time. Results The results indicated that alcohol drinking and depression were stable over time. In the control group, there was no significant causal relationship between problem drinking and depression while in the problem drinking group, drinking in the previous year significantly influenced depression in the following second, third and fourth years. Conclusion This study compared normal versus problem drinkers and showed a 4-year mutual causal relationship between depression and drinking. No longitudinal interaction between drinking and depression occurred in normal drinkers, while drinking intensified depression over time in problem drinkers. Short summary This study found that problem drinking was a risk factor for development of depression. Therefore, more attention should be given to problem alcohol use in the general population and evaluation of past alcohol use history in patients with depressive disorders. INTRODUCTION Alcohol use disorder (AUD) is a brain disease that impairs the structure and function of the brain (Bühler and Mann, 2011), and its lifetime prevalence is reported to be 13.4% (men 20.7%, women 6.1%) (Ministry of Health and Welfare, 2012). AUD is commonly comorbid with depression, bipolar disorder, panic disorder and antisocial personality disorder (Kumar et al., 2010; Klimkiewicz et al., 2015). The National Epidemiologic Survey on Alcohol and Related Conditions III studied 36,309 US citizens on alcohol related dynamics during 2012–2013 and found an important relationship between adjusted odds ratios of 12-month and Lifetime DSM-5 AUD and other psychiatric disorders (Grant et al., 2015), which were 1.2–1.4 times in major depression disorders following the severity of AUD. Also, according to the Korean Epidemiologic Survey, the zero-order odds ratio of comorbidity of alcohol dependence and major depressive disorder was 3.9 (2.2–7.0) (Ministry of Health and Welfare, 2012). Park et al. (2015) reported that half (N = 402) of patients with depression drank more than one bottle of ‘Soju’ (the most popular distilled liquor in Korea; one bottle contains 72 g of alcohol), showing risky levels of alcohol consumption. Coryell et al. (1992) noted that since alcoholism and major depression are relatively common and related to high chronicity and morbidity of the two conditions, their mutual relationship is more important than that of other disorders. Some researchers have reported that a high degree of comorbidity negatively affects clinical progress (Grant and Harford, 1995; Gilman and Abraham, 2001). For example, AUD prolongs the course of depression (Mueller et al., 1994), as persistent depressive symptoms during abstinence act as a risk factor in relapse drinking (Greenfield et al., 1998; Hasin et al., 2002), and AUD increases the risk of suicide (Berglund, 1984; Cornelius et al., 1995; Grant and Hasin, 1999) and aggravates other drug dependence such as nicotine (Rounsaville et al., 1987; Kranzler et al., 1996). There is an ongoing debate regarding the cause of the high prevalence of comorbid AUD and mood disorder (Coryell et al., 1992). There are three possible explanations for the comorbidity mechanism of MDD and AUD. First, depression is induced by the pharmacodynamic effect of alcohol. Second, alcoholism results from self-treatment of depression. Third, job loss because of alcoholism causes major depression (Kessler and Price, 1993; Merikangas et al., 1994; Compton et al., 2000; Sadock and Sadock, 2011). Biological hypotheses suggest the high comorbidity of AUD and depression could be mediated by dysfunction of the hypothalamic–pituitary–adrenal (HPA) axis, which is often found in patients with AUD or depression (Sadock et al., 2015). Chronic alcohol consumption and hypercortisolaemia are both neurotoxic (Sapolsky, 1986, 1993, 2000), causing dysregulation of the HPA axis negative feedback system. Hypercortisolaemia can trigger changes in norepinephrine and serotonin neurotransmission, which are neurotransmitters implicated in the pathogenesis of depression(Watson and Mackin, 2006). Thus, hypercortisolaemia, common in both chronic alcohol consumption and depression, could explain the high comorbidity. Furthermore, there could be another explanation, that the relationship between AUD and depression may be overestimated, because mental illnesses are more prevalent in clinical populations than the general population (Allen and Francis, 1986). Thus, it is necessary to identify the causal relationship between alcohol use and depression in the general population. This research separated normal drinkers (the control group) and those with drinking problems (the problem drinkers group) based on longitudinal data from the general population and aimed to examine aspects of the causal relationship between depression and drinking of each group over time. METHODS Subjects The data used in this study were extracted from the nationally representative Korean Welfare Panel Study. The Korea Welfare Panel was established in 2006 and involves sampling from 90% of the population and a housing census (except islands and special facilities), allowing representation of the whole country including Jeju island. This study used 4-year data from 2011 to 2014. The subjects were 2511 adult males from 20 to 65 years old. The control group and the problem drinking group were separated based on a cut-off point of 12 from the results of the Korean version of the Alcohol Use Disorders Identification Test (AUDIT-K) answered in 2011, the beginning point of the research data. Hence, a normal drinking group of 2191 and a problem drinking group of 320 people were analysed. Measures Korean version of the AUDIT-K To measure problem drinking levels in the panel data, the AUDIT was used, which consists of 10 items rated on a 5-point Likert scale developed by the World Health Organization (WHO) (Babor et al., 2001). The total score ranges from 0 to 40 points and higher scores indicate higher levels of problem drinking. The standardized score of ‘drinking with risk’ from the AUDIT, developed by the WHO, is more than 8 points. On the other hand, when the Korean version was translated, standardized, and verified for its validity and reliability by Kim et al. (1999), cut-off scores of 12 or more were recommended as ‘problem drinking’. Therefore, Kim et al. (1999) recommended that a cut-off point of more than 12 be used for standardization if the purpose is to screen for ‘problem drinking’ within biological–psychological–social concepts, considering the high prevalence of AUD in Korea. Moreover, the Korean Ministry of Health and Welfare modified the AUDIT for the Korean environment in 2009 to develop the AUDIT-K and examined its reliability and validity. Based on this test, the Ministry recommended a cut-off of 12 points for ‘problem drinking’. Therefore, this study used a cut-off point of 12 for ‘problem drinking’. Centre for Epidemiological Studies Depression Scale (CESD-11) Depression was assessed using the 11-item Centre for Epidemiological Studies Depression Scale (CESD-11) (Radloff, 1977) with a 4-point Likert scale. Higher scores indicate higher levels of depression. Research questions The following were the research questions considered in both the problem drinking and control groups. [Q1] Is drinking stable over time? [Q2] Is depression stable over time? [Q3] Is the impact of drinking on depression significant? [Q4] Is the impact of depression on drinking significant? [Q5] What are the differences between the two groups? Statistical analysis This study applied autoregressive cross-lagged (ARCL) modelling analysis to investigate the mutual and causal relationship between alcohol drinking and depression measured consecutively over time. ARCL combines an autoregressive model that describes the value at the time of [t] with the value at the time of [t−1], with a cross-lagged effect model that can detect the mutual effect of both variables considering the time lag (Fig. 1). Since ARCL is a model used to study the non-recursive relationship longitudinally, it is useful to analyse the causal relationship between two factors (Hong, 2008) (Fig. 1). Fig. 1. View largeDownload slide Example of research model (autoregressive cross-lagged model). Fig. 1. View largeDownload slide Example of research model (autoregressive cross-lagged model). To analyse the data, the order of analysis should be as follows: analysing the homogeneity of measurements → the homogeneity of paths → the homogeneity of error covariance. The homogeneity test of error covariance verifies whether the error covariance is stable each time it is measured (Kim et al., 2009). To measure the homogeneity of measurements, paths, and the error covariance, six competitive models were set and compared. Moreover, to evaluate the model fitness, χ2 tests, incremental fit indices such as the comparative fit index (CFI), normed fit index (NFI), and Tucker–Lewis Index (TLI), and absolute indices such as the goodness-of-fit index (GFI) and root mean square error of approximation (RMSEA) were used. Incremental fit indices larger than 0.9 indicate good model fit. Additionally, a RMSEA value under 0.05 indicates good model fit (Hu and Bentler, 1999). For the data analysis, SPSS 19.0 was used for descriptive statistics and AMOS 21.0 was used for the main analysis, ARCL. However, due to the characteristics of the panel data, there were some missing values which could have possibly caused errors in the results. Considering these missing values, full information maximum likelihood (FIML), which estimates parameters, was used. The FIML estimation method includes missing data in the analysis and is more accurate than other missing data approaches (Arbuckle, 1996). Model comparisons Before performing analysis in accordance with the ARCL model, the following were the competing models set to verify the homogeneity of paths and error covariance (Supplementary Fig. S1). Model 1: No constraint to the base model Model 2: Homogeneity constraint model to the autoregressive coefficient (A) of drinking Model 3: Homogeneity constraint model to the autoregressive coefficient (B) of depression Model 4: Homogeneity constraint model to the cross-regression coefficient (C) of depression by drinking Model 5: Homogeneity constraint model to the cross-regression coefficient (D) of drinking by depression Model 6: Homogeneity constraint model to the error covariance (E) between drinking and depression To find the optimal model among the six mentioned above, the models were sequentially compared. Based on these comparison results, this study selected the basic model with no constraint (Model 1), which had the highest fitness as the final research model, and examined how the relationship between drinking and depression varied over time (Supplementary Table S1). RESULTS Subjects’ general characteristics and correlations between variables The average age of the normal drinking group was 46.9 years (SD = 0.22) whereas that of the problem drinking group was 49.9 (SD = 0.47). The normal drinking group’s education level was higher than that of the problem drinking group; 51% (1118) had graduated from university, whereas 49.1% (157) of the problem drinking group were high school graduates. For both groups, more than 50% were married or cohabiting. More than 90% of both groups did not have any disability. Regarding religion, 41.5% (909) of the normal drinking group had a religion compared with 36.9% (118) of the problem drinking group. The mean AUDIT-K score of the normal drinking group was 5.16 whereas that of the problem drinking group was 14.92 (Table 1). Table 1. General characteristics Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) View Large Table 1. General characteristics Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) View Large All correlations between the variables ranged from 0.041 to 0.583 and were statistically significant (Supplementary Table S2). Autoregressive cross-lagged relationships between drinking and depression To evaluate the mutual influence of drinking and depression, ARCL modelling was applied to the data measured at four different times. Compared to other models, Model 1 (basic model with no constraint) was chosen as the final model for this study. ARCL analysis was applied to the selected model and the model was divided into a problem drinking group and a control group. The following are the specific results of the problem drinking group’s and control group’s path coefficient estimation (Table 2). First, in the control group, the paths from one drinking timepoint to another and the longitudinal autoregressive coefficient from one depression timepoint to another were significant. This indicates that previous drinking had a meaningful influence on later drinking and the same applied for depression. Especially, drinking (ß = 0.465–0.550, P < 0.001) had a greater influence on the auto-regression than depression (ß = 0.237–0.345, P < 0.001). However, the cross-regression coefficient from drinking to depression was not statistically significant in all four years of analysis. The cross-over regression coefficient from depression to alcohol was not statistically significant at any timepoints except from second-year depression to third-year drinking (ß = −0.071, P < 0.001) (Fig. 2). Fig. 2. View largeDownload slide Control group (AUDIT 12). Fig. 2. View largeDownload slide Control group (AUDIT 12). Second, in the problem drinking group, similar to the control group, both drinking and depression’s auto-regression coefficients showed that the previous time (t − 1) significantly affected the later time (t) (Fig. 3), which indicates that drinking and depression had a significant effect on later drinking and depression, respectively, over time. Furthermore, the problem drinking group showed a higher auto-regression coefficient for depression than the control group. Fig. 3. View largeDownload slide Problem drinking group (AUDIT ≥ 12). Fig. 3. View largeDownload slide Problem drinking group (AUDIT ≥ 12). In contrast to the results in the control group, drinking of the previous time (t − 1) in the problem drinking group had a statistically significant influence on depression of a later time (ß = 0.170–0.232, P < 0.001). Additionally, previous drinking (t − 1, t, t+ 1) had a significant effect on the cross-regression coefficients of depression at later periods (t, t+ 1, t + 2). Thus, previous drinking of problem drinkers showed an intensifying effect on depression in the subsequent year and this continued to occur in all four years of analysis. However, the cross-regression coefficient from depression to drinking was statistically significant only in the first year of depression to the second year of drinking (ß = 0.230, P < 0.001) and appeared to be non-significant in the subsequent years. Table 2. Autoregressive cross-lagged relationships between drinking and depression Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 B, non-standardized coefficients; β, standardized coefficients; CR, critical ratio. *P < 0.05, **P < 0.01, ***P < 0.001. View Large Table 2. Autoregressive cross-lagged relationships between drinking and depression Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 B, non-standardized coefficients; β, standardized coefficients; CR, critical ratio. *P < 0.05, **P < 0.01, ***P < 0.001. View Large DISCUSSION This study aimed to investigate the mutual and causal relationship between drinking and depression in a control group (AUDIT-K < 12) and a problem drinking group (AUDIT-K ≥ 12) using longitudinal data of men who were 20–65 years old in Korea. Results showed the following. First, drinking alcohol showed a positive autoregressive longitudinal relationship in both the control and problem drinking groups. Second, depression showed a positive autoregressive longitudinal relationship in both control and problem drinking groups. Third, the effect of drinking on depression differed between the control and problem drinking groups. In the control group, the effect of previous drinking on later depression was not significant. Conversely, in the problem drinking group, there was a constant significant positive relationship between previous drinking and later depression. Fourth, in the control and problem drinking groups, depression had a partial impact on later drinking, which was not significant. Finally, there was a difference between the control group and problem drinking group in ARCL in terms of alcohol use and depression in Korean men. Research and clinical implications of these results are as follows. Drinking and depression each appeared to have a positive fixed relationship over time, such that drinking in the first year increased drinking in the following year, and depression in the first year increased depression in the following year. As many prior studies have shown, AUD and depression can be seen as chronic diseases that worsen over time. In the problem drinking group, previous drinking increased depression in later years and this intensified over time. In contrast, the path from depression to drinking had a significant positive relationship only at the first and second time points of depression and did not have a significant influence afterwards. According to a 1-year longitudinal study that investigated the causal relationship between drinking and depression in a sample of 742 people in Los Angeles (Aneshenel and Huba, 1983), drinking led to a decrease in depression at the beginning (4 or 8 months); however, in the long term (12 months), the level of depression increased as drinking increased. This is consistent with our finding that drinking predicted depression over time. We speculate that people may begin drinking to self-medicate or as a countermeasure to relieve depression, but over time the drinking exacerbates the level of depression. Additionally, a meta-analysis of longitudinal research on the relationship between drinking and depression in the general population of the US, UK and Canada (Hartka et al., 1991) reported that previous drinking and depression predicted amounts of drinking and levels of depression at later time periods. These findings are consistent with the results of the present study. In this study, there was no longitudinal interaction between drinking and depression in the control group, which reflects that the control group consisted of normal or low-risk drinkers in the general population. As such, there may not be a significant relationship between alcohol and depression in the general population. However, in the high-risk drinking group, continued drinking led to worsening of depression over time, which implies that problem drinking itself could be a risk factor for development of depression. However, because this study did not include a clinical sample, the results may be limited in their generalizability to AUD patients. There are several limitations to the current study. First, because it focused on the general population rather than a clinical population, no medical and diagnostic procedures were used to separate the normal drinking and high-risk drinking groups. Thus, the results should be interpreted with caution. Second, it is necessary to control various confounding variables to clarify the relationship between drinking and depression (Paschall et al., 2005). The study has a limitation in that it examined relationships only using drinking and depression as variables without considering other confounding variables. Third, the study sample only included males, which limits its conclusions to Korean men. It is vital to study differences according to gender, and future studies need to include women. Despite these limitations, this study is significant in that it explored what should be prioritized between alcohol drinking and depression and revealed what is irrelevant. Especially, the present study is meaningful since most existing studies are cross-sectional or partially longitudinal, only investigating the rate of change of one variable in one period affecting other variables over time. The present study rises above the limits of prior studies by examining how the relationship between the previous point and the next point of two separate variables changes. CONCLUSION We did not find a longitudinal mutual causal relationship between depression and drinking in the normal drinking group; however, a consistently significant influence of drinking on depression was found in the problem drinking group. Individuals may initially use alcohol to self-medicate for depression; however, over time, increased alcohol use may worsen the intensity of depression. Therefore, for patients with either or both conditions, that is problematic alcohol drinking and mood symptoms, it is necessary to explore the detailed history of each condition. SUPPLEMENTARY MATERIAL Supplementary data are available at Alcohol And Alcoholism online. Conflict of Interest Statement None declared. REFERENCES Allen MH , Francis RJ . ( 1986 ) Varieties of psychopathology found in patients with affective disorders: a review. In Meyers RE (ed) . Psychopathology and Addictive Disorders . New York : Guilford Press , 86 – 102 . Aneshenel CS , Huba GJ . ( 1983 ) Depression, alcohol use and smoking over one year: a four-wave longitudinal causal model . J Abnorm Psychol 92 : 134 – 50 . Google Scholar CrossRef Search ADS PubMed Arbuckle JL . ( 1996 ) Full information estimation in the presence of incomplete data. In Marcoulides GA , Schumacker RE (eds) . Advanced Structural Equation Modeling: Issues and Techniques . Mahwah, NJ : Lawrence Erlbaum Associates , 243 – 78 . Babor TF , Higgins-Biddle JC , Saunders JB , et al. . ( 2001 ) The Alcohol Use Disorders Identification Test: Guidelines for Use in Primary Care , 2nd edn . Geneva: World Health Organization . Berglund M . ( 1984 ). Suicide in alcoholism . Anxiety 15 : 14 . Bühler M , Mann K . ( 2011 ) Alcohol and the human brain: a systematic review of different neuroimaging methods . Alcohol Clin Exp Res 35 : 1771 – 93 . Google Scholar CrossRef Search ADS PubMed Compton WM , Cottler LB , Ben Abdallah A , et al. . ( 2000 ) Substance dependence and other psychiatric disorders among drug dependent subjects: race and gender correlates . Am J Addict 9 : 113 – 25 . Google Scholar CrossRef Search ADS PubMed Cornelius JR , Salloum IM , Mezzich J , et al. . ( 1995 ) Disproportionate suicidality in patients with comorbid major depression and alcoholism . Am J Psychiatry 152 : 358 – 64 . Google Scholar CrossRef Search ADS PubMed Coryell W , Winokur G , Keller M , et al. . ( 1992 ) Alcoholism and primary major depression: a family study approach to co-existing disorders . J Affect Disord 24 : 93 – 9 . Google Scholar CrossRef Search ADS PubMed Gilman SE , Abraham HD . ( 2001 ) A longitudinal study of the order of onset of alcohol dependence and major depression . Drug Alcohol Depend 63 : 277 – 86 . Google Scholar CrossRef Search ADS PubMed Grant BF , Harford TC . ( 1995 ) Comorbidity between DSM-IV alcohol use disorders and major depression: results of a national survey . Drug Alcohol Depend 39 : 197 – 206 . Google Scholar CrossRef Search ADS PubMed Grant BF , Hasin DS . ( 1999 ) Suicidal ideation among the United States drinking population: results from the National Longitudinal Alcohol Epidemiologic survey . J Stud Alcohol 60 : 422 – 9 . Google Scholar CrossRef Search ADS PubMed Grant BF , Goldstein RB , Saha TD , et al. . ( 2015 ) Results from the National Epidemiologic Survey on Alcohol and Related Conditions III . JAMA Psychiatry 72 : 757 – 66 . Google Scholar CrossRef Search ADS PubMed Greenfield SF , Weiss RD , Muenz LR , et al. . ( 1998 ) The effect of depression on return to drinking: a prospective study . Arch Gen Psychiatry 55 : 259 – 265 . Google Scholar CrossRef Search ADS PubMed Hartka E , Johnstone B , Leino EV , et al. . ( 1991 ) A meta-analysis of depressive symptomatology a consumption over time . Br J Addict 86 : 1283 – 1298 . Google Scholar CrossRef Search ADS PubMed Hasin D , Liu X , Nunes E , et al. . ( 2002 ) Effects of major depression on remission and relapse of substance dependence . Arch Gen Psychiatry 59 : 375 – 80 . Google Scholar CrossRef Search ADS PubMed Hong SH . ( 2008 ) Longitudinal study methodology using potential growth model. Workshop Material. Younsei University Social Welfare Dept. Hu LT , Bentler P . ( 1999 ) Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives . Struct Equ Modeling 6 : 1 – 55 . Google Scholar CrossRef Search ADS Kessler RC , Price RH . ( 1993 ) Primary prevention of secondary disorders: a proposal and agenda . Am J Commun Psychol 21 : 607 – 33 . Google Scholar CrossRef Search ADS Kim JH , Kim MG , Hong SH . ( 2009 ) Writing Scholarly Papers with Structural Equation Modeling . Seoul : Communication Books . Kim JS , Oh MK , Park BK , et al. . ( 1999 ) Screening criteria of alcoholism by alcohol use disorders identification (AUDIT) in Korea . J Korean Acad Fam Med 20 : 1152 – 59 . Klimkiewicz A , Klimkiewicz J , Jakubczyk A , et al. . ( 2015 ) Comorbidity alcohol dependence with other psychiatric disorders. Part 1. Epidemiology of dual diagnosis . Psychiatr Pol 49 : 265 – 75 . Google Scholar CrossRef Search ADS PubMed Kranzler HR , Del Boca FK , Rounsaville BJ . ( 1996 ) Comorbid psychiatric diagnosis predicts three-year outcomes in alcoholics: a post-treatment natural history study . J Stud Alcohol 57 : 619 – 26 . Google Scholar CrossRef Search ADS PubMed Kumar V , Dalal PK , Trivedi JK , et al. . ( 2010 ) A study of psychiatric comorbidity in alcohol dependence . Delhi Psychiatr J 13 : 291 – 3 . Merikangas KR , Risch NJ , Weissman MM . ( 1994 ) Comorbidity and co-transmission of alcoholism, anxiety and depression . Psychol Med 24 : 69 – 80 . Google Scholar CrossRef Search ADS PubMed Ministry of Health and Welfare . ( 2012 ) The Epidemiological Survey on Mental Disorder in Korea . Seoul : Ministry of Health and Welfare, Seoul National University , 2011: 143 . Mueller TI , Lavori PW , Keller MB , et al. . ( 1994 ) Prognostic effect of the variable course of alcoholism on the 10-year course of depression . Am J Psychiatry 151 : 701 – 6 . Google Scholar CrossRef Search ADS PubMed Park SC , Lee SK , Oh HS , et al. . ( 2015 ) Hazardous drinking-related characteristics of depressive disorders in Korea: the CRESCEND study . Korean Med Sci 30 : 74 – 81 . Google Scholar CrossRef Search ADS Paschall MJ , Freisthler B , Lipton RI . ( 2005 ) Moderate alcohol use and depression in young adults: findings from a national longitudinal study . Am J Public Health 95 : 453 – 7 . Google Scholar CrossRef Search ADS PubMed Radloff LS . ( 1977 ) The CES-D scale a self-report depression scale for research in the general population . Appl Psychol Meas 1 : 385 – 401 . Google Scholar CrossRef Search ADS Rounsaville BJ , Dolinsky ZS , Babor TF , et al. . ( 1987 ) Psychopathology as a predictor of treatment outcome in alcoholics . Arch Gen Psychiatry 44 : 505 – 13 . Google Scholar CrossRef Search ADS PubMed Sadock BJ , Sadock VA . ( 2011 ) Kaplan and Sadock’s Synopsis of Psychiatry: Behavioral Sciences/Clinical Psychiatry . Philadelphia, PA : Lippincott Williams & Wilkins , 625 . Sadock BJ , Sadock VA , Ruiz P . ( 2015 ) Kaplan & Sadock’s Synopsis of Psychiatry Behavioral Sciences/Clinical Psychiatry , 11th edn . Philadelphia, PA : Wolters Kluwer , 350 . Sapolsky RM . ( 1986 ) The neuroendocrinology of stress and aging: the glucocorticoid cascade hypothesis . Endocrinol Rev 7 : 284 – 301 . Google Scholar CrossRef Search ADS Sapolsky RM . ( 1993 ) Potential behavioral modification of glucocorticoid damage to the hippocampus . Behav Brain Res 57 : 175 – 82 . Google Scholar CrossRef Search ADS PubMed Sapolsky RM . ( 2000 ) Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders . Arch Gen Psychiatry 57 : 925 – 35 . Google Scholar CrossRef Search ADS PubMed Watson S , Mackin P . ( 2006 ) HPA axis function in mood disorders . Psychiatry 5 : 16 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Medical Council on Alcohol and Oxford University Press. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Alcohol and Alcoholism Oxford University Press

The Mutual Relationship Between Men’s Drinking and Depression: A 4-Year Longitudinal Analysis

Loading next page...
 
/lp/ou_press/the-mutual-relationship-between-men-s-drinking-and-depression-a-4-year-Z2SHGgieBz
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Medical Council on Alcohol and Oxford University Press. All rights reserved.
ISSN
0735-0414
eISSN
1464-3502
D.O.I.
10.1093/alcalc/agy003
Publisher site
See Article on Publisher Site

Abstract

Abstract Aims The purpose of the current study was to examine the longitudinal reciprocal relationship between depression and drinking among male adults from the general population. Methods This study used a panel dataset from the Korean Welfare Panel (from 2011 to 2014). The subjects were 2511 male adults aged between 20 and 65 years. Based on the Korean Version of the Alcohol Use Disorders Identification Test (AUDIT-K) scores, 2191 subjects were categorized as the control group (AUDIT-K < 12) and 320 subjects were categorized as the problem drinking group (AUDIT-K ≥ 12). An autoregressive cross-lagged modelling analysis was performed to investigate the mutual relationship between problem drinking and depression measured consecutively over time. Results The results indicated that alcohol drinking and depression were stable over time. In the control group, there was no significant causal relationship between problem drinking and depression while in the problem drinking group, drinking in the previous year significantly influenced depression in the following second, third and fourth years. Conclusion This study compared normal versus problem drinkers and showed a 4-year mutual causal relationship between depression and drinking. No longitudinal interaction between drinking and depression occurred in normal drinkers, while drinking intensified depression over time in problem drinkers. Short summary This study found that problem drinking was a risk factor for development of depression. Therefore, more attention should be given to problem alcohol use in the general population and evaluation of past alcohol use history in patients with depressive disorders. INTRODUCTION Alcohol use disorder (AUD) is a brain disease that impairs the structure and function of the brain (Bühler and Mann, 2011), and its lifetime prevalence is reported to be 13.4% (men 20.7%, women 6.1%) (Ministry of Health and Welfare, 2012). AUD is commonly comorbid with depression, bipolar disorder, panic disorder and antisocial personality disorder (Kumar et al., 2010; Klimkiewicz et al., 2015). The National Epidemiologic Survey on Alcohol and Related Conditions III studied 36,309 US citizens on alcohol related dynamics during 2012–2013 and found an important relationship between adjusted odds ratios of 12-month and Lifetime DSM-5 AUD and other psychiatric disorders (Grant et al., 2015), which were 1.2–1.4 times in major depression disorders following the severity of AUD. Also, according to the Korean Epidemiologic Survey, the zero-order odds ratio of comorbidity of alcohol dependence and major depressive disorder was 3.9 (2.2–7.0) (Ministry of Health and Welfare, 2012). Park et al. (2015) reported that half (N = 402) of patients with depression drank more than one bottle of ‘Soju’ (the most popular distilled liquor in Korea; one bottle contains 72 g of alcohol), showing risky levels of alcohol consumption. Coryell et al. (1992) noted that since alcoholism and major depression are relatively common and related to high chronicity and morbidity of the two conditions, their mutual relationship is more important than that of other disorders. Some researchers have reported that a high degree of comorbidity negatively affects clinical progress (Grant and Harford, 1995; Gilman and Abraham, 2001). For example, AUD prolongs the course of depression (Mueller et al., 1994), as persistent depressive symptoms during abstinence act as a risk factor in relapse drinking (Greenfield et al., 1998; Hasin et al., 2002), and AUD increases the risk of suicide (Berglund, 1984; Cornelius et al., 1995; Grant and Hasin, 1999) and aggravates other drug dependence such as nicotine (Rounsaville et al., 1987; Kranzler et al., 1996). There is an ongoing debate regarding the cause of the high prevalence of comorbid AUD and mood disorder (Coryell et al., 1992). There are three possible explanations for the comorbidity mechanism of MDD and AUD. First, depression is induced by the pharmacodynamic effect of alcohol. Second, alcoholism results from self-treatment of depression. Third, job loss because of alcoholism causes major depression (Kessler and Price, 1993; Merikangas et al., 1994; Compton et al., 2000; Sadock and Sadock, 2011). Biological hypotheses suggest the high comorbidity of AUD and depression could be mediated by dysfunction of the hypothalamic–pituitary–adrenal (HPA) axis, which is often found in patients with AUD or depression (Sadock et al., 2015). Chronic alcohol consumption and hypercortisolaemia are both neurotoxic (Sapolsky, 1986, 1993, 2000), causing dysregulation of the HPA axis negative feedback system. Hypercortisolaemia can trigger changes in norepinephrine and serotonin neurotransmission, which are neurotransmitters implicated in the pathogenesis of depression(Watson and Mackin, 2006). Thus, hypercortisolaemia, common in both chronic alcohol consumption and depression, could explain the high comorbidity. Furthermore, there could be another explanation, that the relationship between AUD and depression may be overestimated, because mental illnesses are more prevalent in clinical populations than the general population (Allen and Francis, 1986). Thus, it is necessary to identify the causal relationship between alcohol use and depression in the general population. This research separated normal drinkers (the control group) and those with drinking problems (the problem drinkers group) based on longitudinal data from the general population and aimed to examine aspects of the causal relationship between depression and drinking of each group over time. METHODS Subjects The data used in this study were extracted from the nationally representative Korean Welfare Panel Study. The Korea Welfare Panel was established in 2006 and involves sampling from 90% of the population and a housing census (except islands and special facilities), allowing representation of the whole country including Jeju island. This study used 4-year data from 2011 to 2014. The subjects were 2511 adult males from 20 to 65 years old. The control group and the problem drinking group were separated based on a cut-off point of 12 from the results of the Korean version of the Alcohol Use Disorders Identification Test (AUDIT-K) answered in 2011, the beginning point of the research data. Hence, a normal drinking group of 2191 and a problem drinking group of 320 people were analysed. Measures Korean version of the AUDIT-K To measure problem drinking levels in the panel data, the AUDIT was used, which consists of 10 items rated on a 5-point Likert scale developed by the World Health Organization (WHO) (Babor et al., 2001). The total score ranges from 0 to 40 points and higher scores indicate higher levels of problem drinking. The standardized score of ‘drinking with risk’ from the AUDIT, developed by the WHO, is more than 8 points. On the other hand, when the Korean version was translated, standardized, and verified for its validity and reliability by Kim et al. (1999), cut-off scores of 12 or more were recommended as ‘problem drinking’. Therefore, Kim et al. (1999) recommended that a cut-off point of more than 12 be used for standardization if the purpose is to screen for ‘problem drinking’ within biological–psychological–social concepts, considering the high prevalence of AUD in Korea. Moreover, the Korean Ministry of Health and Welfare modified the AUDIT for the Korean environment in 2009 to develop the AUDIT-K and examined its reliability and validity. Based on this test, the Ministry recommended a cut-off of 12 points for ‘problem drinking’. Therefore, this study used a cut-off point of 12 for ‘problem drinking’. Centre for Epidemiological Studies Depression Scale (CESD-11) Depression was assessed using the 11-item Centre for Epidemiological Studies Depression Scale (CESD-11) (Radloff, 1977) with a 4-point Likert scale. Higher scores indicate higher levels of depression. Research questions The following were the research questions considered in both the problem drinking and control groups. [Q1] Is drinking stable over time? [Q2] Is depression stable over time? [Q3] Is the impact of drinking on depression significant? [Q4] Is the impact of depression on drinking significant? [Q5] What are the differences between the two groups? Statistical analysis This study applied autoregressive cross-lagged (ARCL) modelling analysis to investigate the mutual and causal relationship between alcohol drinking and depression measured consecutively over time. ARCL combines an autoregressive model that describes the value at the time of [t] with the value at the time of [t−1], with a cross-lagged effect model that can detect the mutual effect of both variables considering the time lag (Fig. 1). Since ARCL is a model used to study the non-recursive relationship longitudinally, it is useful to analyse the causal relationship between two factors (Hong, 2008) (Fig. 1). Fig. 1. View largeDownload slide Example of research model (autoregressive cross-lagged model). Fig. 1. View largeDownload slide Example of research model (autoregressive cross-lagged model). To analyse the data, the order of analysis should be as follows: analysing the homogeneity of measurements → the homogeneity of paths → the homogeneity of error covariance. The homogeneity test of error covariance verifies whether the error covariance is stable each time it is measured (Kim et al., 2009). To measure the homogeneity of measurements, paths, and the error covariance, six competitive models were set and compared. Moreover, to evaluate the model fitness, χ2 tests, incremental fit indices such as the comparative fit index (CFI), normed fit index (NFI), and Tucker–Lewis Index (TLI), and absolute indices such as the goodness-of-fit index (GFI) and root mean square error of approximation (RMSEA) were used. Incremental fit indices larger than 0.9 indicate good model fit. Additionally, a RMSEA value under 0.05 indicates good model fit (Hu and Bentler, 1999). For the data analysis, SPSS 19.0 was used for descriptive statistics and AMOS 21.0 was used for the main analysis, ARCL. However, due to the characteristics of the panel data, there were some missing values which could have possibly caused errors in the results. Considering these missing values, full information maximum likelihood (FIML), which estimates parameters, was used. The FIML estimation method includes missing data in the analysis and is more accurate than other missing data approaches (Arbuckle, 1996). Model comparisons Before performing analysis in accordance with the ARCL model, the following were the competing models set to verify the homogeneity of paths and error covariance (Supplementary Fig. S1). Model 1: No constraint to the base model Model 2: Homogeneity constraint model to the autoregressive coefficient (A) of drinking Model 3: Homogeneity constraint model to the autoregressive coefficient (B) of depression Model 4: Homogeneity constraint model to the cross-regression coefficient (C) of depression by drinking Model 5: Homogeneity constraint model to the cross-regression coefficient (D) of drinking by depression Model 6: Homogeneity constraint model to the error covariance (E) between drinking and depression To find the optimal model among the six mentioned above, the models were sequentially compared. Based on these comparison results, this study selected the basic model with no constraint (Model 1), which had the highest fitness as the final research model, and examined how the relationship between drinking and depression varied over time (Supplementary Table S1). RESULTS Subjects’ general characteristics and correlations between variables The average age of the normal drinking group was 46.9 years (SD = 0.22) whereas that of the problem drinking group was 49.9 (SD = 0.47). The normal drinking group’s education level was higher than that of the problem drinking group; 51% (1118) had graduated from university, whereas 49.1% (157) of the problem drinking group were high school graduates. For both groups, more than 50% were married or cohabiting. More than 90% of both groups did not have any disability. Regarding religion, 41.5% (909) of the normal drinking group had a religion compared with 36.9% (118) of the problem drinking group. The mean AUDIT-K score of the normal drinking group was 5.16 whereas that of the problem drinking group was 14.92 (Table 1). Table 1. General characteristics Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) View Large Table 1. General characteristics Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) Characteristics Less than 12 More than 12 % (N) % (N) Age  20s 4.2 (93) 0  30s 22.0 (483) 12.8 (41)  40s 32.6 (714) 37.5 (120)  50s 28.0 (614) 34.1 (109)  60–65 13.1 (287) 15.6 (50) Mean (SD) 46.9 (0.218) 49.9 (0.466) Education  Less than high school 11.6 (254) 15.9 (51)  High school 37.4 (819) 49.1 (157)  Some college of higher 51.0 (1118) 35.0 (112) Marital status  Married/cohabiting 65.2 (1429) 76.3 (244)  Widowed/separated/divorced 5.8 (126) 8.4 (27)  Never been married 29.0 (636) 15.3 (49) Disability  Yes 7.3 (160) 8.8 (28)  No 92.7 (2031) 91.3 (292) Religion  Yes 41.5 (909) 36.9 (118)  No 58.5 (1282) 63.1 (202) AUDIT-K (mean) 5.16 14.92  Total 100.0 (2191) 100.0 (320) View Large All correlations between the variables ranged from 0.041 to 0.583 and were statistically significant (Supplementary Table S2). Autoregressive cross-lagged relationships between drinking and depression To evaluate the mutual influence of drinking and depression, ARCL modelling was applied to the data measured at four different times. Compared to other models, Model 1 (basic model with no constraint) was chosen as the final model for this study. ARCL analysis was applied to the selected model and the model was divided into a problem drinking group and a control group. The following are the specific results of the problem drinking group’s and control group’s path coefficient estimation (Table 2). First, in the control group, the paths from one drinking timepoint to another and the longitudinal autoregressive coefficient from one depression timepoint to another were significant. This indicates that previous drinking had a meaningful influence on later drinking and the same applied for depression. Especially, drinking (ß = 0.465–0.550, P < 0.001) had a greater influence on the auto-regression than depression (ß = 0.237–0.345, P < 0.001). However, the cross-regression coefficient from drinking to depression was not statistically significant in all four years of analysis. The cross-over regression coefficient from depression to alcohol was not statistically significant at any timepoints except from second-year depression to third-year drinking (ß = −0.071, P < 0.001) (Fig. 2). Fig. 2. View largeDownload slide Control group (AUDIT 12). Fig. 2. View largeDownload slide Control group (AUDIT 12). Second, in the problem drinking group, similar to the control group, both drinking and depression’s auto-regression coefficients showed that the previous time (t − 1) significantly affected the later time (t) (Fig. 3), which indicates that drinking and depression had a significant effect on later drinking and depression, respectively, over time. Furthermore, the problem drinking group showed a higher auto-regression coefficient for depression than the control group. Fig. 3. View largeDownload slide Problem drinking group (AUDIT ≥ 12). Fig. 3. View largeDownload slide Problem drinking group (AUDIT ≥ 12). In contrast to the results in the control group, drinking of the previous time (t − 1) in the problem drinking group had a statistically significant influence on depression of a later time (ß = 0.170–0.232, P < 0.001). Additionally, previous drinking (t − 1, t, t+ 1) had a significant effect on the cross-regression coefficients of depression at later periods (t, t+ 1, t + 2). Thus, previous drinking of problem drinkers showed an intensifying effect on depression in the subsequent year and this continued to occur in all four years of analysis. However, the cross-regression coefficient from depression to drinking was statistically significant only in the first year of depression to the second year of drinking (ß = 0.230, P < 0.001) and appeared to be non-significant in the subsequent years. Table 2. Autoregressive cross-lagged relationships between drinking and depression Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 B, non-standardized coefficients; β, standardized coefficients; CR, critical ratio. *P < 0.05, **P < 0.01, ***P < 0.001. View Large Table 2. Autoregressive cross-lagged relationships between drinking and depression Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 Control group (<12 point) Problem drinking group (≥12 point) B S.E. C.R. B B S.E. C.R. B Drink1 → drink2 0.465*** 0.797 0.036 22.436 0.234*** 0.438 0.121 3.636 Drink2 → drink3 0.550*** 0.354 0.013 26.377 0.519*** 0.332 0.035 9.381 Drink3 → drink4 0.484*** 0.768 0.036 21.517 0.311*** 0.469 0.096 4.895 Depression1 → depression2 0.345*** 0.333 0.021 15.588 0.408*** 0.331 0.049 6.775 Depression2 → depression3 0.343*** 0.372 0.026 14.462 0.384*** 0.444 0.067 6.666 Depression3 → depression4 0.237*** 0.227 0.024 9.367 0.319*** 0.321 0.061 5.289 Drink1 → depression2 −0.020 −0.022 0.023 −0.918 0.170** 0.195 0.069 2.824 Drink2 → depression3 0.002 0.001 0.016 0.09 0.191*** 0.135 0.041 3.323 Drink3 → depression4 −0.020 −0.02 0.025 −0.79 0.232*** 0.258 0.067 3.846 Depression1 → drink2 −0.018 −0.029 0.032 −0.889 0.230*** 0.306 0.085 3.576 Depression2 → drink3 −0.071*** −0.073 0.022 −3.341 0.059 0.061 0.058 1.060 Depression3 → drink4 −0.031 −0.047 0.035 −1.353 0.063 0.086 0.087 0.994 B, non-standardized coefficients; β, standardized coefficients; CR, critical ratio. *P < 0.05, **P < 0.01, ***P < 0.001. View Large DISCUSSION This study aimed to investigate the mutual and causal relationship between drinking and depression in a control group (AUDIT-K < 12) and a problem drinking group (AUDIT-K ≥ 12) using longitudinal data of men who were 20–65 years old in Korea. Results showed the following. First, drinking alcohol showed a positive autoregressive longitudinal relationship in both the control and problem drinking groups. Second, depression showed a positive autoregressive longitudinal relationship in both control and problem drinking groups. Third, the effect of drinking on depression differed between the control and problem drinking groups. In the control group, the effect of previous drinking on later depression was not significant. Conversely, in the problem drinking group, there was a constant significant positive relationship between previous drinking and later depression. Fourth, in the control and problem drinking groups, depression had a partial impact on later drinking, which was not significant. Finally, there was a difference between the control group and problem drinking group in ARCL in terms of alcohol use and depression in Korean men. Research and clinical implications of these results are as follows. Drinking and depression each appeared to have a positive fixed relationship over time, such that drinking in the first year increased drinking in the following year, and depression in the first year increased depression in the following year. As many prior studies have shown, AUD and depression can be seen as chronic diseases that worsen over time. In the problem drinking group, previous drinking increased depression in later years and this intensified over time. In contrast, the path from depression to drinking had a significant positive relationship only at the first and second time points of depression and did not have a significant influence afterwards. According to a 1-year longitudinal study that investigated the causal relationship between drinking and depression in a sample of 742 people in Los Angeles (Aneshenel and Huba, 1983), drinking led to a decrease in depression at the beginning (4 or 8 months); however, in the long term (12 months), the level of depression increased as drinking increased. This is consistent with our finding that drinking predicted depression over time. We speculate that people may begin drinking to self-medicate or as a countermeasure to relieve depression, but over time the drinking exacerbates the level of depression. Additionally, a meta-analysis of longitudinal research on the relationship between drinking and depression in the general population of the US, UK and Canada (Hartka et al., 1991) reported that previous drinking and depression predicted amounts of drinking and levels of depression at later time periods. These findings are consistent with the results of the present study. In this study, there was no longitudinal interaction between drinking and depression in the control group, which reflects that the control group consisted of normal or low-risk drinkers in the general population. As such, there may not be a significant relationship between alcohol and depression in the general population. However, in the high-risk drinking group, continued drinking led to worsening of depression over time, which implies that problem drinking itself could be a risk factor for development of depression. However, because this study did not include a clinical sample, the results may be limited in their generalizability to AUD patients. There are several limitations to the current study. First, because it focused on the general population rather than a clinical population, no medical and diagnostic procedures were used to separate the normal drinking and high-risk drinking groups. Thus, the results should be interpreted with caution. Second, it is necessary to control various confounding variables to clarify the relationship between drinking and depression (Paschall et al., 2005). The study has a limitation in that it examined relationships only using drinking and depression as variables without considering other confounding variables. Third, the study sample only included males, which limits its conclusions to Korean men. It is vital to study differences according to gender, and future studies need to include women. Despite these limitations, this study is significant in that it explored what should be prioritized between alcohol drinking and depression and revealed what is irrelevant. Especially, the present study is meaningful since most existing studies are cross-sectional or partially longitudinal, only investigating the rate of change of one variable in one period affecting other variables over time. The present study rises above the limits of prior studies by examining how the relationship between the previous point and the next point of two separate variables changes. CONCLUSION We did not find a longitudinal mutual causal relationship between depression and drinking in the normal drinking group; however, a consistently significant influence of drinking on depression was found in the problem drinking group. Individuals may initially use alcohol to self-medicate for depression; however, over time, increased alcohol use may worsen the intensity of depression. Therefore, for patients with either or both conditions, that is problematic alcohol drinking and mood symptoms, it is necessary to explore the detailed history of each condition. SUPPLEMENTARY MATERIAL Supplementary data are available at Alcohol And Alcoholism online. Conflict of Interest Statement None declared. REFERENCES Allen MH , Francis RJ . ( 1986 ) Varieties of psychopathology found in patients with affective disorders: a review. In Meyers RE (ed) . Psychopathology and Addictive Disorders . New York : Guilford Press , 86 – 102 . Aneshenel CS , Huba GJ . ( 1983 ) Depression, alcohol use and smoking over one year: a four-wave longitudinal causal model . J Abnorm Psychol 92 : 134 – 50 . Google Scholar CrossRef Search ADS PubMed Arbuckle JL . ( 1996 ) Full information estimation in the presence of incomplete data. In Marcoulides GA , Schumacker RE (eds) . Advanced Structural Equation Modeling: Issues and Techniques . Mahwah, NJ : Lawrence Erlbaum Associates , 243 – 78 . Babor TF , Higgins-Biddle JC , Saunders JB , et al. . ( 2001 ) The Alcohol Use Disorders Identification Test: Guidelines for Use in Primary Care , 2nd edn . Geneva: World Health Organization . Berglund M . ( 1984 ). Suicide in alcoholism . Anxiety 15 : 14 . Bühler M , Mann K . ( 2011 ) Alcohol and the human brain: a systematic review of different neuroimaging methods . Alcohol Clin Exp Res 35 : 1771 – 93 . Google Scholar CrossRef Search ADS PubMed Compton WM , Cottler LB , Ben Abdallah A , et al. . ( 2000 ) Substance dependence and other psychiatric disorders among drug dependent subjects: race and gender correlates . Am J Addict 9 : 113 – 25 . Google Scholar CrossRef Search ADS PubMed Cornelius JR , Salloum IM , Mezzich J , et al. . ( 1995 ) Disproportionate suicidality in patients with comorbid major depression and alcoholism . Am J Psychiatry 152 : 358 – 64 . Google Scholar CrossRef Search ADS PubMed Coryell W , Winokur G , Keller M , et al. . ( 1992 ) Alcoholism and primary major depression: a family study approach to co-existing disorders . J Affect Disord 24 : 93 – 9 . Google Scholar CrossRef Search ADS PubMed Gilman SE , Abraham HD . ( 2001 ) A longitudinal study of the order of onset of alcohol dependence and major depression . Drug Alcohol Depend 63 : 277 – 86 . Google Scholar CrossRef Search ADS PubMed Grant BF , Harford TC . ( 1995 ) Comorbidity between DSM-IV alcohol use disorders and major depression: results of a national survey . Drug Alcohol Depend 39 : 197 – 206 . Google Scholar CrossRef Search ADS PubMed Grant BF , Hasin DS . ( 1999 ) Suicidal ideation among the United States drinking population: results from the National Longitudinal Alcohol Epidemiologic survey . J Stud Alcohol 60 : 422 – 9 . Google Scholar CrossRef Search ADS PubMed Grant BF , Goldstein RB , Saha TD , et al. . ( 2015 ) Results from the National Epidemiologic Survey on Alcohol and Related Conditions III . JAMA Psychiatry 72 : 757 – 66 . Google Scholar CrossRef Search ADS PubMed Greenfield SF , Weiss RD , Muenz LR , et al. . ( 1998 ) The effect of depression on return to drinking: a prospective study . Arch Gen Psychiatry 55 : 259 – 265 . Google Scholar CrossRef Search ADS PubMed Hartka E , Johnstone B , Leino EV , et al. . ( 1991 ) A meta-analysis of depressive symptomatology a consumption over time . Br J Addict 86 : 1283 – 1298 . Google Scholar CrossRef Search ADS PubMed Hasin D , Liu X , Nunes E , et al. . ( 2002 ) Effects of major depression on remission and relapse of substance dependence . Arch Gen Psychiatry 59 : 375 – 80 . Google Scholar CrossRef Search ADS PubMed Hong SH . ( 2008 ) Longitudinal study methodology using potential growth model. Workshop Material. Younsei University Social Welfare Dept. Hu LT , Bentler P . ( 1999 ) Cutoff criteria for fit indexes in covariance structure analysis: conventional criteria versus new alternatives . Struct Equ Modeling 6 : 1 – 55 . Google Scholar CrossRef Search ADS Kessler RC , Price RH . ( 1993 ) Primary prevention of secondary disorders: a proposal and agenda . Am J Commun Psychol 21 : 607 – 33 . Google Scholar CrossRef Search ADS Kim JH , Kim MG , Hong SH . ( 2009 ) Writing Scholarly Papers with Structural Equation Modeling . Seoul : Communication Books . Kim JS , Oh MK , Park BK , et al. . ( 1999 ) Screening criteria of alcoholism by alcohol use disorders identification (AUDIT) in Korea . J Korean Acad Fam Med 20 : 1152 – 59 . Klimkiewicz A , Klimkiewicz J , Jakubczyk A , et al. . ( 2015 ) Comorbidity alcohol dependence with other psychiatric disorders. Part 1. Epidemiology of dual diagnosis . Psychiatr Pol 49 : 265 – 75 . Google Scholar CrossRef Search ADS PubMed Kranzler HR , Del Boca FK , Rounsaville BJ . ( 1996 ) Comorbid psychiatric diagnosis predicts three-year outcomes in alcoholics: a post-treatment natural history study . J Stud Alcohol 57 : 619 – 26 . Google Scholar CrossRef Search ADS PubMed Kumar V , Dalal PK , Trivedi JK , et al. . ( 2010 ) A study of psychiatric comorbidity in alcohol dependence . Delhi Psychiatr J 13 : 291 – 3 . Merikangas KR , Risch NJ , Weissman MM . ( 1994 ) Comorbidity and co-transmission of alcoholism, anxiety and depression . Psychol Med 24 : 69 – 80 . Google Scholar CrossRef Search ADS PubMed Ministry of Health and Welfare . ( 2012 ) The Epidemiological Survey on Mental Disorder in Korea . Seoul : Ministry of Health and Welfare, Seoul National University , 2011: 143 . Mueller TI , Lavori PW , Keller MB , et al. . ( 1994 ) Prognostic effect of the variable course of alcoholism on the 10-year course of depression . Am J Psychiatry 151 : 701 – 6 . Google Scholar CrossRef Search ADS PubMed Park SC , Lee SK , Oh HS , et al. . ( 2015 ) Hazardous drinking-related characteristics of depressive disorders in Korea: the CRESCEND study . Korean Med Sci 30 : 74 – 81 . Google Scholar CrossRef Search ADS Paschall MJ , Freisthler B , Lipton RI . ( 2005 ) Moderate alcohol use and depression in young adults: findings from a national longitudinal study . Am J Public Health 95 : 453 – 7 . Google Scholar CrossRef Search ADS PubMed Radloff LS . ( 1977 ) The CES-D scale a self-report depression scale for research in the general population . Appl Psychol Meas 1 : 385 – 401 . Google Scholar CrossRef Search ADS Rounsaville BJ , Dolinsky ZS , Babor TF , et al. . ( 1987 ) Psychopathology as a predictor of treatment outcome in alcoholics . Arch Gen Psychiatry 44 : 505 – 13 . Google Scholar CrossRef Search ADS PubMed Sadock BJ , Sadock VA . ( 2011 ) Kaplan and Sadock’s Synopsis of Psychiatry: Behavioral Sciences/Clinical Psychiatry . Philadelphia, PA : Lippincott Williams & Wilkins , 625 . Sadock BJ , Sadock VA , Ruiz P . ( 2015 ) Kaplan & Sadock’s Synopsis of Psychiatry Behavioral Sciences/Clinical Psychiatry , 11th edn . Philadelphia, PA : Wolters Kluwer , 350 . Sapolsky RM . ( 1986 ) The neuroendocrinology of stress and aging: the glucocorticoid cascade hypothesis . Endocrinol Rev 7 : 284 – 301 . Google Scholar CrossRef Search ADS Sapolsky RM . ( 1993 ) Potential behavioral modification of glucocorticoid damage to the hippocampus . Behav Brain Res 57 : 175 – 82 . Google Scholar CrossRef Search ADS PubMed Sapolsky RM . ( 2000 ) Glucocorticoids and hippocampal atrophy in neuropsychiatric disorders . Arch Gen Psychiatry 57 : 925 – 35 . Google Scholar CrossRef Search ADS PubMed Watson S , Mackin P . ( 2006 ) HPA axis function in mood disorders . Psychiatry 5 : 16 . Google Scholar CrossRef Search ADS © The Author(s) 2018. Medical Council on Alcohol and Oxford University Press. All rights reserved. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Alcohol and AlcoholismOxford University Press

Published: Mar 17, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off