Risk of harm from alcohol use and heavy alcohol consumption: Its association with other NCD risk factors in Thailand

While there is an abundance of literature examining the relation between quantity of alcohol consumption and risk factors for non-communicable diseases (NCD), there is less evidence on whether the risk of harm from alcohol use would have a similar relationship with NCD risk factors. The study aims to determine the association between level of harm from alcohol use and NCD risk factors. A cross-sectional survey was conducted among health care workers in Thailand in 2013. The Alcohol, Smoking and Substance Involvement Screening Test (ASSIST) was used to assessed risk of harm from alcohol use. The results suggest that higher risk of harm from alcohol use was associated with two of the eight NCD risk factors among women (higher blood pressure and higher triglyceride level) and five of the eight NCD risk factors among men (smoking, physical inactivity, higher blood pressure, higher blood glucose and higher triglyceride level). For men, assessing risk of harm could be incorporated as part of NCD programs as practitioners do not have to worry about the accuracy of the alcohol quantification and conversion to standard drinks. However, among women, quantifying volume may still be needed.

Measures of alcohol consumption and harm. Risk of harm from alcohol use: The ASSIST was used to categorize the level of harm in the past three months into three categories: (1) Lifetime abstainers, (2) low risk of   13 .
Pattern of alcohol use: The WHO STEPS questionnaire was used to categorized pattern of alcohol consumption within the past month into three groups: (1) lifetime abstainer (2) non-binge drinker (<4 standard drinks per occasion for women and <5 standard drinks per occasion for men) and (3) binge drinker (≥4 standard drinks per occasion for women and ≥5 standard drinks per occasion for men). other risk factors for ncDs. The WHO STEPS questionnaire was used to quantify the following behavioral risk factors for NCD consisting of: Tobacco use: Patients were categorized as either current smoker or non-smoker. Insufficient physical activity: WHO recommends that to achieve sufficient physical activity for health, any combination of moderate-and/or vigorous-intensity activities should achieve a minimum of at least 600 metabolic equivalent of task (MET)-minutes per week. Those failing to achieve at least 600 MET minutes per week, were considered as having inadequate physical activity.
Insufficient fruit and vegetable consumption: Participants eating less than five servings of fruit and/or vegetables per day is categorized eating insufficient amount of fruit and vegetable consumption.
Other physiological risk factors for NCDs include high body mass index (BMI), high blood pressure (BP), high fasting blood sugar level (FBS) and high cholesterol specifically high low density lipoprotein (LDL) cholesterol and high triglycerides (TG).
other factors of interests. From a review of literature, key potential confounding factors between alcohol use and other NCD risk factors were included, such as age, sex, education, income and occupation. All these factors were measured though the self-administered online questionnaire. In addition, biological markers of liver injury and alcohol use, serum alanine aminotransferase (ALT) and aspartate transaminase (AST) were also measured in the physical exam and venous sampling. These two markers usually rise among those with alcoholic hepatitis and heavy alcohol use. Statistical analysis. Descriptive statistics were used to describe basic demographic data. Alcohol consumption and patterns of alcohol consumption were compared between men and women using t-test, Wilcoxon rank sum test or chi-square. To explore the associations between harmful alcohol use and other risk factors for NCDs, linear regression was used for continuous outcomes and logistic regression were used for binary outcomes. Age and sex were considered a priori confounders and all analyses were stratified by sex and adjusted for age. Additional adjustments for other potential confounders including education, income, and occupation were also done using regression analyses. The p-value from the overall likelihood ratio test for the association between the main exposure (alcohol consumption or harm) and that NCD risk factor were obtained.

Results
In total, there were 3,204 participants who completed the survey, representing a 60% response rate from the total eligible population. Among the participants, 1,394 (43.5%) were lifetime abstainers from alcohol use. Over ninety percent of abstainers were female. Among drinkers, 80.7% drank alcohol at low risk of harm and 19.3% were moderate to high risk. Three-fourths of low risk drinkers were female. Demographic data categorized by risk of harm from alcohol use were described in Table 1. Drinkers at moderate to high risk of harm from alcohol use tended to be younger with lower education and income (Table 1).
Relationship between risk of harm from alcohol use and heavy alcohol drinking. The average number of standard drinks per day for drinkers at low risk of harm was 1.03 for women and 2.98 drinks for men. The average number of standard drinks per day for those at moderate to high risk of harm was 3.56 for women and 7.05 for men. Almost 10% of women at low risk of harm from alcohol were binge drinkers. About 44% of women at high risk of harm from alcohol use were binge drinkers while over 80% of men at high risk of harm from alcohol use were binge drinkers ( Table 2).
Risk of harm from alcohol use and risk factors for ncD. In the multivariable model adjusting for age, income, education and occupation, higher risk of harm from alcohol use was not significantly associated with www.nature.com/scientificreports www.nature.com/scientificreports/ current smoking, inadequate fruit and vegetable consumption, physical activity and BMI among women (Fig. 1). For men, higher risk of harm from alcohol use was associated with current smoking and inadequate physical activity but not associated with inadequate fruit/vegetable consumption or BMI (Fig. 2).
For physiological risk factors among women, higher risk of harm from alcohol use was associated with higher levels of triglycerides and marginally associated with higher levels of systolic blood pressure. Increased risk of harm from alcohol use was not significantly associated with higher levels of blood glucose or LDL cholesterol (Fig. 3). While for men, higher risk of harm from alcohol use was associated with higher levels of systolic blood pressure, blood glucose and TG, but not associated with LDL cholesterol (Fig. 4).

Heavy alcohol use and risk factors for ncD.
In the adjusted model, among women, binge drinking was associated with current smoking and higher levels of BMI but not associated with inadequate fruit/vegetable consumption and inadequate physical activity (Fig. 1). For men, binge drinking was associated with current smoking, inadequate physical activity and higher BMI but not associated with inadequate fruit/vegetable consumption (Fig. 2).
For physiological risk factors among women, binge drinking was associated with higher levels of systolic blood pressure, triglycerides but not associated with blood glucose, LDL cholesterol (Fig. 3). For men, binge drinking was associated with increased blood pressure, increased TG and weakly associated with increased LDL cholesterol but not associated with blood glucose (Fig. 4).

Discussion
Using ASSIST as a screening tool, higher risk of harm from alcohol use was associated with two of the eight NCD risk factors among women and five of the eight NCD risk factors among men. The same result in both sexes found in systolic blood pressure and TG. Using WHO STEPS as a screening tool, heavy alcohol consumption was associated with four of the eight NCD risk factors among women and six of the eight NCD risk factors for men.
WHO STEPS detected frequency and pattern of alcohol use through measurements of standard drink that reflect volume of alcohol use. Binge drinking has been shown to be associated with other NCD risk factors 14,15 . Previous literature in Thailand also found that drinking four or more glasses per occasion was associated with elevated risk of NCD 6 . ASSIST, however, detected risk of harm from alcohol use by assessing the impact of alcohol  (N = 2,472). p-value obtained from overall likelihood ratio test between exposure of interests (alcohol consumption or risk of harm) and NCD risk factor using linear or logistic regression. Error bars represent 95% Confidence intervals. (2019) 9:16343 | https://doi.org/10.1038/s41598-019-52754-w www.nature.com/scientificreports www.nature.com/scientificreports/ consumption on psychosocial aspects such as social problems and concern of peers 10 . The different aspects of alcohol use detected by the two instruments can help explain the difference in associations with other NCD risk factors.
The study found that the association between risk of harm and other NCD risk factors differed by gender. The risk of harm from alcohol use correlated with higher volume of alcohol use among men, which potentially translates to higher levels of NCD risk factors. However, higher risk of harm from alcohol use in women did not correlate with volume and thus did not correlate with as many NCD risk factors as men. There are multiple factors which may contribute to these findings.
Firstly, biology and alcohol metabolism are different between men and women. Women are at higher risk of breast cancer than men 16 . Moreover, women have less total body water than men who have the same body weight. Therefore, given the same amount of alcohol, blood alcohol concentration is higher in women, potentially putting women at high risk from alcohol related diseases such as alcoholic liver diseases [17][18][19] For women, small volumes of alcohol may result in intoxication. ASSIST is sensitive to detect risk of harm from alcohol use such as intoxication and its effect. This is supported by the literature that women are at higher risk of alcohol dependence than men, given similar pattern/volume of use 20 . In addition, women eliminate alcohol faster than men 18 , thus literature has shown that the risk of alcohol-related NCD such as stroke, cirrhosis or some cancer are less among women [21][22][23][24] . Thus, while there was high risk of harm from alcohol use, its longer-term effects in terms of other NCD were less likely to be observed among women.
The social norm in Thailand may also play a role why women may be at higher risk of harm from alcohol use despite low volume. Thai society tends to stigmatize women consuming alcohol 25,26 . In addition, alcohol use among Thai men is considered normal when socializing 27 . Women, however, are more likely to use alcohol when they have negative moods or emotions 17,28 . All these factors contribute to why women may be at higher risk of harm from alcohol use despite low volume comsumption.
The study is not without limitations. Due to the nature of the cross sectional design, we could not establish causality between alcohol consumption and NCD risk factors. The outcomes of our study rely on self-reported behavioural risk factors for NCD. However, the results also seemed to translate to other physiological markers of NCDs. Quantification of alcohol use may be affected by measurement bias and social desirability bias but the study used third party interviewers who were trained on the questionnaires to quantify volume of alcohol use. www.nature.com/scientificreports www.nature.com/scientificreports/ This help increased the accuracy of conversion to standard drinks and help minimize such bias. However, recruiting our sample from a single center might limit the generalizability of our results.
Because harmful use of alcohol is a significant contributor to the global burden of NCD 29 , WHO suggested a target of an at least 10% reduction in the harmful use of alcohol within a national context 30 . Since our study found evidence that both volume and risk of harm from alcohol consumption correlated with NCD risk factors, there are some practical implications. Firstly, for men, ASSIST may be an easier screening tool to be incorporated as part of NCD programs as practitioners do not have to worry about the accuracy of the alcohol quantification and conversion to standard drinks. However, among women, using tools to quantify volume may still be needed. Secondly, the study supports the literature that when healthcare providers encounter patients with various NCDs risk factors, alcohol consumption should be assessed 31,32 . When assessing volume of alcohol use, measures of harm related to alcohol should also be assessed as these two measurements of harmful alcohol use did not correlate well among women.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.