Abstract
Aims
To assess whether health changes affect cessation of alcohol consumption and to compare the health status of former drinkers and abstainers.
Methods
Cohort data from 9001 Korean participants aged 40–69 years old were analyzed. Alcohol consumption was assessed every 2 years for 10 years. Participant age, sex, marital status, education level, employment status, smoking, chronic disease, perceived health and changes in these variables were analyzed to identify factors associated with quitting alcohol drinking. The number of diseases and perceived health of former drinkers and people who at baseline were lifetime abstainers were compared.
Results
Among 4037 drinkers at baseline, 673 (16.7%) were classed as quitters and 3364 (83.3%) were classed as non-quitters. Sex, age and worsened perception of health were significantly associated with cessation of drinking. Women and individuals >60 years were more likely to cease drinking. There was a significant association between disease onset or treatment and alcohol cessation for cancer cases, but not for cardiovascular disease or chronic disease cases. There was no significant difference in number of diseases or perceived health between former drinkers and people who at baseline were lifetime abstainers.
Conclusions
The effect of disease onset or treatment on alcohol consumption cessation depended on disease type. Former drinkers did not show significantly worse health than people who at baseline were lifetime abstainers. Further studies of alcohol consumption and its effects on health are needed to consider disease occurrence and changes in alcohol consumption.
Short summary
Disease onset or treatment significantly affected alcohol consumption cessation for cancer cases, but not for cardiovascular disease or other chronic disease cases. There was no significant difference in health status between former drinkers and lifetime abstainers.
INTRODUCTION
Although alcohol consumption can lead to various adverse conditions and diseases, recent studies have described the beneficial effect of mild alcohol consumption on mortality (Rehm et al., 2003; Di Castelnuovo et al., 2006; Ronksley et al., 2011; Jin et al., 2013; Zheng et al., 2015). Some research findings suggest that the beneficial effect of mild alcohol drinking might be partly attributable to inappropriate selection of the reference group and weak adjustment for confounders (Knott et al., 2015). Several studies have highlighted the bias of the ‘sick quitter effect’; that is, the inclusion of former drinkers in the reference group (Wannamethee and Shaper, 1988). Some findings indicate that the use of non-drinkers as a reference group may lead to bias, as many people stop drinking because of health concerns or aging (Stockwell et al., 2012). The poor overall health of former drinkers (Fillmore et al., 1998) may result in overestimation of the protective effect of alcohol consumption (Fillmore et al., 2006).
However, findings from studies assessing the sick quitter effect are conflicting. Although several studies have shown that the beneficial effect of mild alcohol consumption is no longer significant after excluding former drinkers from the non-drinker group (Fillmore et al., 2006; Stockwell et al., 2016), other studies have reported that former drinker misclassification has no effect on alcohol-related health risk (Zeisser et al., 2014; Gémes et al., 2016).
The sick quitter effect is based on the assumption that disease onset and changes in health condition lead to cessation of alcohol consumption, so the risk for former drinkers is higher than that for abstainers. Although several studies have reported no relationship between changes in alcohol consumption and health condition (Zins et al., 1999; Brennan et al., 2010), many studies have reported an association between disease status and reduced drinking (Walton et al., 2000; Perreira and Sloan, 2001; Molander et al., 2010). Zins et al. (1999) found that self-perception of poor health and taking sleeping pills were associated with a cessation of alcohol consumption. Similarly, Perreira and Sloan (2001) found that hospitalization and chronic conditions were associated with reduced drinking. Another study found that major medical diagnosis (such as diabetes, stroke, myocardial infarction, cancer and hospitalization during the follow-up period) were associated with fewer drinks per month (Molander et al., 2010).
Previous studies assessing risk among former drinkers have also shown conflicting results. Although some studies have reported high risk among former drinkers (Lazarus et al., 1991; Liang and Chikritzhs, 2013; Chan et al., 2015), others have shown a lower risk for former drinkers than for abstainers (Andrews-Chavez et al., 2015). To confirm the sick quitter effect, both investigation of the relationship between health status changes and quitting alcohol drinking and assessment of the risk to former drinkers are needed.
In addition, to assess the effect of health condition on alcohol drinking, other factors that affect changes in alcohol consumption should be considered. The stability of alcohol consumption can vary according to participant characteristics such as sex, age and health status.
Sex
Previous studies indicate that men and women exhibit different patterns of change in their alcohol consumption. Molander et al. (2010) found that the number of drinks per month was stable among women, but that this measure increased among men. Furthermore, women are more likely to decrease or stop drinking (Perreira and Sloan, 2001). In Park's study, female drinkers were more likely to quit drinking during follow-up monitoring and were also more likely to keep abstaining than men (Park et al., 2009). However, another study reported that male sex was associated with a faster age-related reduction in alcohol consumption (Moore et al., 2005). Brennan et al. (2010) conducted a 10-year study of drinking patterns among 55–65 year olds, and found correlations with age of 0.31–0.54 for men and 0.61–0.76 for women.
Age
Previous studies indicate an association between older age and lower alcohol consumption (Kemm, 2003). For example, Eigenbrodt et al.’s (2001) cross-sectional study revealed an inverse association between drinking prevalence and age, and a reduction in drinking prevalence during the 6-year follow-up. Longitudinal studies of alcohol consumption also generally show age-related changes (Moore et al., 2005), which often involve reduced consumption or a transition to non-drinker status. However, several studies have reported stable alcohol consumption over multi-year evaluation periods (Benzies et al., 2008; Brennan et al., 2010).
Marriage
Becoming a widow or widower is associated with increased alcohol consumption, but the effect of getting married and divorced is not consistent; these factors are associated with both decreases and increases in alcohol consumption (Perreira and Sloan, 2001). Karlamangla et al. (2006) found that getting married was associated with a reduction in the frequency of heavy drinking. Another study reported that predictors of attempts to quit drinking and factors leading to successful quitting varied among individuals with alcohol abuse or dependence; for example, being single increased the likelihood of attempting to quit, while being married increased the odds of success (Chiappetta et al., 2014).
Employment
A study of the effect of employment status on alcohol consumption revealed that retirement increased the likelihood of periodic heavy drinking, compared with working individuals (Bacharach et al., 2004). Other studies show that retirement is associated with increased (Perreira and Sloan, 2001) or frequent alcohol consumption (Molander et al., 2010). Furthermore, employment type can affect alcohol consumption; Syden et al. (2014) found that self-employed women exhibited more stable alcohol consumption than women who worked as manual laborers.
Education level
Moore et al. (2005) reported that higher education levels predicted greater alcohol consumption at baseline and slower age-related decreases in drinking. Similarly, another study found that higher education levels predicted increased daily drinking (Molander et al., 2010).
Smoking
Smoking status is associated with greater alcohol consumption at baseline (Moore et al., 2005), and Zins et al. (1999) found that smoking status is associated with increased alcohol consumption. Other research indicates that quitting smoking is related to a reduction in the incidence of heavy drinking (Karlamangla et al., 2006; Berg et al., 2015) or alcohol use disorder (Cavazos-Rehg et al., 2014). However, Kahler et al. (2010) reported that quitting smoking did not lead to significant change in alcohol use.
The aim of this study was (1) to determine whether health changes affect quitting alcohol drinking, and (2) to assess health risks for former drinkers and lifetime abstainers.
MATERIALS AND METHODS
This study used data from KoGES, a community-based cohort study conducted by the Korean Center for Disease Control and Prevention. A total of 9001 participants aged 40–69 years were recruited and followed for 10 years. Demographic characteristics, including age, sex, marital status, education level and employment status, were assessed using a standardized questionnaire. Information on health behaviors, including alcohol consumption and smoking, were obtained. Medical history was also obtained to analyze the onset and type of disease. All data were collected by trained investigators using standardized methods. The study was approved by the institutional review board of Seoul National University.
Analysis of factors associated with quitting drinking
Alcohol consumption was assessed every 2 years. To identify factors associated with quitting alcohol drinking, the drinkers were divided into quitters and non-quitters. Only participants whose alcohol consumption was zero at both 8-year and 10-year assessment were considered to have quit drinking. Participant demographics at baseline and changes in those variables were compared. Analysis of change involved comparison of values at baseline and 8 years later. For participants with missing data at 8 years, the last follow-up data were used to assess variable changes. The baseline covariates were sex, age group, marital status, educational level, employment status, baseline presence of chronic disease and baseline perception of health. In addition, changes in marital status, employment status, smoking and perceived health status were assessed from baseline to the 10-year follow-up.
To analyze the impact of health changes on quitting alcohol consumption, disease onset and disease treatment were evaluated, and diseases were categorized into three groups: cardiovascular disease (CVD), cancer and other chronic diseases. The CVD group included hypertension, myocardial infarction, heart failure, coronary artery disease, hyperlipidemia, peripheral artery disease, stroke, transient ischemic heart disease and arrhythmia. The cancer group included various types of cancer, such as lung cancer, stomach cancer, liver cancer, colon cancer, pancreatic cancer, uterine cancer and breast cancer. The other chronic diseases included diabetes, gastritis, hepatitis, kidney disease, tuberculosis and pulmonary disease.
Only if disease onset occurred before quitting drinking could the disease be considered to have affected the decision to quit. To determine factors associated with quitting alcohol drinking, multiple logistic analysis was conducted.
Comparison of health status among abstainers, former drinkers
To analyze health conditions, the number of diseases for abstainers, former drinkers and current drinkers were compared. Excluding participants with no follow-up data or drinking status data at baseline, 8923 cases were included.
Among the abstainers at baseline, people who did not drink any alcohol for the 10-year follow-up period were considered the abstainer group. For non-abstainers at baseline, those whose alcohol consumption was zero at both the 8-year and 10-year follow-up were considered former drinkers; the rest were considered current drinkers. The number of CVD, cancer and chronic disease occurrences in each group were compared using analysis of variance, and categorical variables were evaluated using the χ2 test. All analyses were performed using R software (version 3.1), and P values of <0.05 were considered statistically significant.
RESULTS
Factors associated with cessation of alcohol consumption
Of the 9001 participants, 8923 (abstainers: 4116; former drinkers: 572; current drinkers: 4235) reported their drinking status at baseline. Of the current drinkers, 198 reported their alcohol consumption as zero. After these were excluded, only 4037 were considered drinkers at baseline. Of the 4037 drinkers at baseline, 673 were considered quitters by this definition, and 3364 were considered non-quitters.
The analysis of participants who continued drinking or quit drinking revealed that significantly more men continued drinking than quit drinking (P < 0.001). Participants who were 40–49 years old were prominent in both groups, although there were significantly more individuals of this age in the group that continued drinking (P < 0.001). Single people were more likely to quit drinking (P < 0.001), but the proportion of participants who were unemployed or who had a chronic disease was not significantly different. Many participants in the quit-drinking group reported poor self-perceived health status, whereas the group that continued drinking had a higher proportion of participants with good and average self-perceived health status (P < 0.001). The group that quit drinking contained a higher percentage of people who became separated or divorced (P = 0.02). The prevalence of losing a job or quitting smoking was not significantly different between participants who continued drinking and those who quit drinking.
The group that quit drinking showed a significantly lower prevalence of worsened perception of health status (P < 0.001), and a higher prevalence of CVD (P = 0.85), cancer (P < 0.001) and other chronic disease onset (P < 0.51). The prevalence of CVD (P = 0.18), cancer (P = 0.002) and other chronic disease treatment (P = 0.14) was higher in the group that had quit drinking; however, only cancer onset and cancer treatment were statistically significant (Table 1).
Table 1.
Comparison of demographic data between participants who continued drinking and those who quit drinking
| Drinking status, N (%) | Total (n = 4037) | P value | ||
|---|---|---|---|---|
| Kept drinking (n = 3364) | Quit drinking (n = 673) | |||
| Baseline | ||||
| Male | 2595 (77.1) | 315 (46.8) | 2910 (72.1) | <0.001 |
| Age | ||||
| 40–49 | 1928 (57.3) | 295 (43.8) | 2223 (55.1) | <0.001 |
| 50–59 | 805 (23.9) | 196 (29.1) | 1001 (24.8) | |
| 60–69 | 631 (18.8) | 182 (27.0) | 813 (20.1) | |
| Separated or divorced | 185 (5.5) | 65 (9.7) | 250 (6.2) | <0.001 |
| Low education level | 1482 (44.2) | 378 (56.3) | 1860 (46.2) | <0.001 |
| Unemployed | 604 (18.0) | 108 (16.1) | 712 (17.7) | 0.26 |
| Chronic disease | 1526 (45.4) | 309 (45.9) | 1835 (45.5) | 0.83 |
| Perceived health | ||||
| Poor | 802 (23.9) | 231 (34.4) | 1033 (25.6) | <0.001 |
| Average | 1319 (39.3) | 246 (36.6) | 1565 (38.8) | |
| good | 1237 (36.8) | 195 (29.0) | 1432 (35.5) | |
| Change of baseline | ||||
| Separated or divorced | 124 (3.7) | 39 (5.8) | 163 (4.0) | 0.02 |
| Became unemployed | 167 (5.0) | 46 (6.9) | 213 (5.3) | 0.06 |
| Smoking | ||||
| Maintained | 2887 (86.1) | 584 (87.3) | 3471 (86.3) | 0.43 |
| Quit | 468 (13.9) | 85 (12.7) | 553 (13.7) | |
| Worsened perceived health | 1548 (46.0) | 259 (38.5) | 1807 (44.8) | <0.001 |
| Disease onset | ||||
| Cardiovascular | 676 (21.8) | 142 (22.2) | 818 (21.9) | 0.85 |
| Cancer | 39 (1.3) | 33 (5.2) | 72 (1.9) | <0.001 |
| Other chronic disease | 565 (18.2) | 124 (19.4) | 689 (18.4) | 0.51 |
| Disease treatment | ||||
| Cardiovascular | 209 (21.2) | 59 (25.5) | 268 (22.1) | 0.18 |
| Cancer | 9 (0.9) | 9 (4.0) | 18 (1.5) | 0.002 |
| Other chronic disease | 88 (12.2) | 24 (17.3) | 112 (13.0) | 0.14 |
Table 2 shows the factors that were associated with cessation of alcohol consumption. Model 1 analyzed the effect of baseline characteristics on cessation of alcohol consumption. Age and sex affected cessation of alcohol consumption; women were more likely to quit drinking (odds ratio [OR]: 3.95; 95% confidence interval [CI]: 3.19, 4.91) and older people (OR: 1.7 for 50–59 years; OR: 2.08 for 60–69 years) were more likely to quit drinking. Marital status, educational level, employment status, current smoking and having a chronic disease did not significantly affect the likelihood of quitting drinking. Participants with good perceived health were less likely to quit drinking than participants with poor perceived health (P = 0.002). In Model 2, which included the significant baseline characteristics and changes in baseline characteristics, becoming single was not significant. Individuals who lost a job (OR: 1.47; 95% CI: 1.01, 2.12) or quit smoking (OR: 1.3; 95% CI: 0.99, 1.71) had greater odds of quitting drinking, whereas worsening perceptions of health (OR: 0.77; 95% CI: 0.63, 0.93) had the opposite effect. Those with cancer onset (OR: 3.9; 95% CI: 2.34, 6.46) had greater odds of quitting drinking, although this was not significant for those with CVD (OR: 0.88; 95% CI: 0.7, 1.1) or other chronic disease (OR: 1.05; 95% CI: 0.83, 1.33). Model 3 analyzed the effect of receiving disease treatment on cessation of alcohol consumption. The likelihood of quitting drinking was significantly higher in individuals receiving cancer treatment (OR: 3.48; 95% CI: 1.07, 10.63) than in those not receiving treatment, although there was no significant effect for CVD (OR: 0.99; 95% CI: 0.61, 1.58) and other chronic disease (OR: 1.63; 95% CI: 0.93, 2.8) (Table 2).
Table 2.
Multiple logistic regression analyses of factors associated with cessation of alcohol consumption
| Cessation of alcohol consumption | ||||||
|---|---|---|---|---|---|---|
| Model 1a | Model 2b | Model 3c | ||||
| Exp (B) (CI) | P value | Exp (B) (CI) | P value | Exp (B) (CI) | P value | |
| Baseline character | ||||||
| Sex (ref: male) | 3.95 (3.19, 4.91) | <0.001 | 4.45 (3.67, 5.41) | <0.001 | 4.27 (2.78, 6.64) | <0.001 |
| Age, years (ref: 40–49) | ||||||
| 50–59 | 1.7 (1.36, 2.12) | <0.001 | 1.71 (1.37, 2.13) | <0.001 | 1.87 (1.17, 2.98) | <0.01 |
| 60–69 | 2.08 (1.62, 2.68) | <0.001 | 2.0 (1.59, 2.53) | <0.001 | 1.26 (0.65, 2.35) | 0.49 |
| Marriage status (ref: single) | 1.02 (0.74, 1.44) | 0.89 | – | – | – | |
| Education level (ref: low) | 1.06 (0.86, 1.3) | 0.58 | – | – | – | |
| Employment status (ref: employed) | 1.15 (0.91, 1.46) | 0.24 | – | – | – | |
| Current smoking | 0.87 (0.7, 1.08) | 0.19 | – | – | – | |
| Having chronic disease | 0.96 (0.8, 1.15) | 0.64 | 1.02 (0.85, 1.23) | 0.82 | 1.02 (0.68, 1.55) | 0.92 |
| Perceived health (ref: poor) | ||||||
| Average | 0.81 (0.65, 1.0) | 0.05 | 0.85 (0.68, 1.07) | 0.17 | 0.78 (0.49, 1.24) | 0.29 |
| Good | 0.7 (0.55, 0.88) | 0.002 | 0.79 (0.62, 1.02) | 0.07 | 0.83 (0.48, 1.42) | 0.49 |
| Change from baseline | ||||||
| Separated/divorced/widowed | – | 0.97 (0.63, 1.45) | 0.89 | 0.84 (0.29, 2.11) | 0.72 | |
| Become unemployed | – | 1.47 (1.01, 2.12) | 0.04 | 0.8 (0.26, 1.99) | 0.65 | |
| Quit smoking (ref: maintained) | – | 1.3 (0.99, 1.71) | 0.056 | 0.89 (0.44, 1.7) | 0.73 | |
| Worsened perceived health | – | 0.77 (0.63, 0.93) | <0.01 | 0.38 (0.25, 0.57) | <0.001 | |
| Disease onset (ref: none) | ||||||
| Cardiovascular | – | 0.88 (0.7, 1.1) | 0.26 | – | – | |
| Cancer | – | 3.9 (2.34, 6.46) | <0.001 | – | – | |
| Other chronic | – | 1.05 (0.83, 1.33) | 0.66 | – | – | |
| Disease treatment (ref: none) | ||||||
| Cardiovascular | – | – | 0.99 (0.61, 1.58) | 0.97 | ||
| Cancer | – | – | 3.48 (1.07, 10.63) | 0.03 | ||
| Other chronic | – | – | 1.63 (0.93, 2.8) | 0.08 | ||
| Intercept | 0.03 (0.01, 0.06) | <0.001 | 0.02 (0.02, 0.03) | <0.001 | 0.03 (0.01, 0.08) | <0.001 |
| Pseudo R2 | 0.1569 | 0.237 | 0.8185 | |||
| Akaike information criteria | 3270.3 | 3072.2 | 688.29 | |||
aModel 1 included only the baseline variables.
bModel 2 included changes in variables and disease onset.
cModel 3 included changes in variables and disease treatment.
Comparison of health status among abstainers, former drinkers and current drinkers
In the comparison of health status, the 8923 participants (excluding those without drink status data) were categorized as abstainers, former drinkers or current drinkers based on their drinking status at baseline and their alcohol consumption during the follow-up period. Among the 4116 abstainers at baseline, 3139 remained non-drinkers during the follow-up period and were considered abstainers in this analysis. Taking into account the baseline and follow-up data, 3139 participants were abstainers, 1470 were former drinkers and 4314 were drinkers (Table 3).
Table 3.
Changes in drinking status based on follow-up data
| Drinking status at baseline | Change of drinking | Drinking status at 8–10 years’ follow-up | N (%) |
|---|---|---|---|
| Abstainer (n = 4116) | Keeping non-drinking | Abstainer | 3139 (76.2) |
| Start drinking and quitting again | Former drinker | 411 (10) | |
| Start drinking | Current drinker | 566 (13.8) | |
| Former drinker (n = 572) | Keeping non-drinking | Former drinker | 270 (47.2) |
| Start drinking and quitting again | Former drinker | 66 (11.5) | |
| Start drinking | Current drinker | 236 (41.3) | |
| Current drinker (n = 4235) | Quit drinking | Former drinker | 723 (17) |
| Keeping drinking | Current drinker | 3512 (83) |
Although there were fewer cases of CVD (0.59 former drinkers, 0.61 abstainers) and chronic disease (0.67 former drinkers, 0.68 abstainers) in former drinkers than in abstainers, the number of cancers was significantly higher in former drinkers (0.076 former drinkers, 0.047 abstainers). The difference in the number of disease variables between abstainers and former drinkers was not significant for CVD (P = 0.74), chronic disease (P = 0.92) or total disease (P = 0.99), but was for cancer (P < 0.001). Individuals reporting good perceived health were most prevalent in current drinkers, whereas individuals reporting poor perceived health were most prevalent in abstainers. Poor perceived health tended to be commoner in abstainers than former drinkers (P = 0.051) (Table 4).
Table 4.
Number of disease variables for abstainers, former drinkers and current drinkers at 10-year follow-up
| Abstainer (n = 3139) | Former (n = 1470) | Current (n = 4314) | P value between groups | ||
|---|---|---|---|---|---|
| Number of CVD | 0.61 ± 0.9 | 0.59 ± 0.89 | 0.52 ± 0.87 | Abstainer–former | 0.74 |
| Abstainer–current | <0.001 | ||||
| Former–current | 0.035 | ||||
| Number of cancer | 0.047 ± 0.23 | 0.076 ± 0.29 | 0.029 ± 0.18 | Abstainer–former | <0.001 |
| Abstainer–current | 0.0012 | ||||
| Former–current | <0.001 | ||||
| Number of chronic disease | 0.68 ± 0.92 | 0.67 ± 0.91 | 0.66 ± 0.91 | Abstainer–former | 0.92 |
| Abstainer–current | 0.59 | ||||
| Former–current | 0.94 | ||||
| Number of total disease | 1.33 ± 1.41 | 1.33 ± 1.38 | 1.21 ± 1.35 | Abstainer–former | 0.99 |
| Abstainer–current | <0.001 | ||||
| Former–current | 0.01 | ||||
| Perceived health | |||||
| Poor | 1148 (36.6) | 496 (33.8) | 972 (22.5) | Abstainer–former | 0.051 |
| Average | 1149 (36.6) | 591 (40.2) | 1862 (43.2) | Abstainer–current | <0.001 |
| Good | 842 (26.8) | 382 (26.0) | 1480 (34.3) | Former–current | <0.001 |
DISCUSSION
Previous studies have verified the sick quitter effect, which attributes cessation of alcohol drinking to health problems. Although Brennan et al. (2010) have reported that baseline health condition is not related to changes in drinking, several studies have found that health status and disease at baseline are associated with cessation of alcohol consumption (Zins et al., 1999; Perreira and Sloan, 2001). Furthermore, a medical diagnosis during follow-up has also been associated with decreased alcohol consumption (Molander et al., 2010). The present results indicate that disease onset significantly increased the likelihood of quitting drinking; however, the effect of this factor varied according to disease type. Cancer significantly increased the likelihood of quitting drinking, but CVD and other chronic diseases (e.g. diabetes or pulmonary disease) did not have a significant effect. Although one study has shown that colorectal cancer survivors are more likely than a matched population group to be heavy drinkers (Hawkes et al., 2008), the present study showed that cancer onset and cancer treatment were positively associated with quitting drinking. Therefore, future studies of disease effects on health behavior need to consider the type of disease, rather than simply the presence of disease.
The present study did not demonstrate a higher health risk among former drinkers compared with lifetime abstainers. Several studies have investigated the purported high risk among former drinkers, which is another factor in the sick quitter effect. In a study comparing self-reported health among abstainers and former and current drinkers, health status was worst in former drinkers, followed by abstainers and best in current drinkers (Liang and Chikritzhs, 2013). Other studies have reported that poor self-reported health is positively associated with former drinking (Chan et al., 2015), and former drinkers had a higher prevalence of poor health among abstainers, former drinkers and current drinkers (Liang and Chikritzhs, 2013). However, in Andrews-Chavez et al.’s (2015) study, fair/poor perceived health status was more prevalent in former drinkers (58%) than in moderate drinkers (55%), but lower than in lifetime abstainers (83%). Similarly, in this study, prevalence of poor perceived health was greatest in abstainers, and the difference between abstainers and former drinkers was not significant. In addition, we found that the number of disease variables was greater in abstainers than in former drinkers, which poses problems for assumptions that the health status of former drinkers is worse than that of abstainers. There is a need for further studies assessing the health status of former drinkers and abstainers using objective indexes.
In this study, only 76.2% of abstainers at baseline remained non-drinkers throughout the follow-up period. Analysis of the effect of alcohol consumption based only on the amount of consumed alcohol at baseline may bias estimates of the risk of alcohol consumption. It is therefore important to properly define groups of drinkers based on follow-up data rather than one-time measurement and to consider changes in consumption when analyzing the effect of alcohol consumption.
CONCLUSION
These findings do not support the assumption that former drinkers have a greater health risk than abstainers. Only cancer was associated with cessation of alcohol consumption, and there was no significant difference in health status between former drinkers and lifetime abstainers. Further studies of alcohol consumption and its effects on health need to consider disease occurrence and change of alcohol consumption.
FUNDING
This work was supported by a KIOM Grant (K16093) from Korea Institute of Oriental Medicine.
CONFLICT OF INTEREST STATEMENT
None declared.
REFERENCES
- Andrews-Chavez JY, Lee CS, Houser RF, et al. (2015) Factors associated with alcohol consumption patterns in a Puerto Rican urban cohort. Public Health Nutr 18:464–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Bacharach SB, Bamberger PA, Sonnenstuhl WJ, et al. (2004) Retirement, risky alcohol consumption and drinking problems among blue-collar workers. J Stud Alcohol 65:537–45. [DOI] [PubMed] [Google Scholar]
- Benzies KM, Wangby M, Bergman LR (2008) Stability and change in health-related behaviors of midlife Swedish women. Health Care Women Int 29:997–1018. [DOI] [PubMed] [Google Scholar]
- Berg KM, Piper ME, Smith SS, et al. (2015) Defining and predicting short-term alcohol use changes during a smoking cessation attempt. Addict Behav 48:52–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Brennan PL, Schutte KK, Moos RH (2010) Patterns and predictors of late-life drinking trajectories: a 10-year longitudinal study. Psychol Addict Behav 24:254–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Cavazos-Rehg PA, Breslau N, Hatsukami D, et al. (2014) Smoking cessation is associated with lower rates of mood/anxiety and alcohol use disorders. Psychol Med 44:2523–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chan YY, Teh CH, Lim KK, et al. (2015) Lifestyle, chronic diseases and self-rated health among Malaysian adults: results from the 2011 National Health and Morbidity Survey (NHMS). BMC Public Health 15:754. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Chiappetta V, Garcia-Rodriguez O, Jin CJ, et al. (2014) Predictors of quit attempts and successful quit attempts among individuals with alcohol use disorders in a nationally representative sample. Drug Alcohol Depend 141:138–44. [DOI] [PubMed] [Google Scholar]
- Di Castelnuovo A, Costanzo S, Bagnardi V, et al. (2006) Alcohol dosing and total mortality in men and women: an updated meta-analysis of 34 prospective studies. Arch Intern Med 166:2437–45. [DOI] [PubMed] [Google Scholar]
- Eigenbrodt ML, Mosley TH Jr, Hutchinson RG, et al. (2001) Alcohol consumption with age: a cross-sectional and longitudinal study of the Atherosclerosis Risk in Communities (ARIC) study, 1987–1995. Am J Epidemiol 153:1102–11. [DOI] [PubMed] [Google Scholar]
- Fillmore KM, Golding JM, Graves KL, et al. (1998) Alcohol consumption and mortality. I. Characteristics of drinking groups. Addiction 93:183–203. [DOI] [PubMed] [Google Scholar]
- Fillmore MT, Kerr WC, Stockwell TR, et al. (2006) Moderate alcohol use and reduced mortality risk: systematic error in prospective studies. Addict Res Theory 14:101–32. [DOI] [PubMed] [Google Scholar]
- Gémes K, Janszky I, Laugsand LE, et al. (2016) Alcohol consumption is associated with a lower incidence of acute myocardial infarction: results from a large prospective population-based study in Norway. J Intern Med 279:365–75. [DOI] [PubMed] [Google Scholar]
- Hawkes AL, Lynch BM, Youlden DR, et al. (2008) Health behaviors of Australian colorectal cancer survivors, compared with noncancer population controls. Support Care Cancer 16:1097–104. [DOI] [PubMed] [Google Scholar]
- Jin M, Cai S, Guo J, et al. (2013) Alcohol drinking and all cancer mortality: a meta-analysis. Ann Oncol 24:807–16. [DOI] [PubMed] [Google Scholar]
- Kahler CW, Borland R, Hyland A, et al. (2010) Quitting smoking and change in alcohol consumption in the International Tobacco Control (ITC) Four Country Survey. Drug Alcohol Depend 110:101–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Karlamangla A, Zhou K, Reuben D, et al. (2006) Longitudinal trajectories of heavy drinking in adults in the United States of America. Addiction 101:91–9. [DOI] [PubMed] [Google Scholar]
- Kemm J. (2003) An analysis by birth cohort of alcohol consumption by adults in Great Britain 1978–1998. Alcohol Alcohol 38:142–7. [DOI] [PubMed] [Google Scholar]
- Knott CS, Coombs N, Stamatakis E, et al. (2015) All cause mortality and the case for age specific alcohol consumption guidelines: pooled analyses of up to 10 population based cohorts. BMJ 350:h384. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Lazarus NB, Kaplan GA, Cohen RD, et al. (1991) Change in alcohol consumption and risk of death from all causes and from ischaemic heart disease. BMJ 303:553–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Liang W, Chikritzhs T (2013) The association between alcohol exposure and self-reported health status: the effect of separating former and current drinkers. PLoS One 8:e55881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Molander RC, Yonker JA, Krahn DD (2010) Age-related changes in drinking patterns from mid- to older age: results from the Wisconsin longitudinal study. Alcohol Clin Exp Res 34:1182–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Moore AA, Gould R, Reuben DB, et al. (2005) Longitudinal patterns and predictors of alcohol consumption in the United States. Am J Public Health 95:458–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Park P, Im J, Yin J, et al. (2009) All cause mortality and the case for age specific alcohol consumption guidelines: pooled analyses of up to 10 population based cohorts. J Korean Alcohol Sci 10:101–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Perreira KM, Sloan FA (2001) Life events and alcohol consumption among mature adults: a longitudinal analysis. J Stud Alcohol 62:501–8. [DOI] [PubMed] [Google Scholar]
- Rehm J, Gmel G, Sempos CT, et al. (2003) Alcohol-related morbidity and mortality. Alcohol Res Health 27:39–51. [PMC free article] [PubMed] [Google Scholar]
- Ronksley PE, Brien SE, Turner BJ, et al. (2011) Association of alcohol consumption with selected cardiovascular disease outcomes: a systematic review and meta-analysis. BMJ 342:d671. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Stockwell T, Greer A, Fillmore K, et al. (2012) How good is the science. BMJ 344:e2276. ; author reply e94. [DOI] [PubMed] [Google Scholar]
- Stockwell T, Zhao J, Panwar S, et al. (2016) Do ‘moderate’ drinkers have reduced mortality risk? A systematic review and meta-analysis of alcohol consumption and all-cause mortality. J Stud Alcohol Drugs 77:185–98. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Syden L, Wennberg P, Forsell Y, et al. (2014) Stability and change in alcohol habits of different socio-demographic subgroups—a cohort study. BMC Public Health 14:525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Walton MA, Mudd SA, Blow FC, et al. (2000) Stability in the drinking habits of older problem-drinkers recruited from nontreatment settings. J Subst Abuse Treat 18:169–77. [DOI] [PubMed] [Google Scholar]
- Wannamethee G, Shaper AG (1988) Men who do not drink: a report from the British Regional Heart Study. Int J Epidemiol 17:307–16. [DOI] [PubMed] [Google Scholar]
- Zeisser C, Stockwell TR, Chikritzhs T (2014) Methodological biases in estimating the relationship between alcohol consumption and breast cancer: the role of drinker misclassification errors in meta-analytic results. Alcohol Clin Exp Res 38:2297–306. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zheng YL, Lian F, Shi Q, et al. (2015) Alcohol intake and associated risk of major cardiovascular outcomes in women compared with men: a systematic review and meta-analysis of prospective observational studies. BMC Public Health 15:773. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Zins M, Carle F, Bugel I, et al. (1999) Predictors of change in alcohol consumption among Frenchmen of the GAZEL study cohort. Addiction 94:385–95. [DOI] [PubMed] [Google Scholar]