Fit and Tipsy?

The Interrelationship Between Cardiorespiratory Fitness and Alcohol Consumption and Dependence

Kerem Shuval; David Leonard; Karen G. Chartier; Carolyn E. Barlow; Bob M. Fennis; David L. Katz; Katelyn Abel; Stephen W. Farrell; Andjelka Pavlovic; Laura F. Defina

Disclosures

Med Sci Sports Exerc. 2022;54(1):113-119. 

In This Article

Methods

The cross-sectional relationship of fitness (primary independent variable) with alcohol consumption and dependence (primary dependent variables) was examined among adults enrolled in the Cooper Center Longitudinal Study (CCLS). The CCLS aims to explore lifestyle behaviors (e.g., physical activity, fitness, diet) as they relate to chronic disease prevention.[28] Cooper Center Longitudinal Study participants consist of patients who come to the Cooper Clinic (Dallas, TX) for preventive medicine examinations, which include fitness and laboratory testing, as well as an extensive medical questionnaire with items including alcohol consumption and physical activity. Patients interested in the CCLS opt into the study and provide written informed consent. Participants are primarily non-Hispanic White and well educated.[28,29] The CCLS is reviewed and approved annually by The Cooper Institute Institutional Review Board, and this research received approval from the University of Haifa Institutional Review Board.

The present study sample began with 55,082 participants 20 yr or older who came to the Cooper Clinic (1988–2019), responded positively to a question on the medical history survey: "Do you drink alcoholic beverages (yes/no)" (i.e., current drinkers) and had complete information on all study variables. Of these, participants were excluded if they were pregnant (n = 50) or excluded if not apparently healthy; that is, had abnormal electrocardiogram (n = 4419); reported a personal history of myocardial infarction, stroke, or diabetes (n = 8450); were underweight (body mass index [BMI] < 18.5 kg·m−2) (n = 2751); or did not reach 85% of maximal heart rate during the treadmill examination (n = 759). These exclusion criteria resulted in 38,653 apparently healthy participants in the analytic sample.

Independent and Dependent Variables

The primary independent variable, fitness, was based on a maximal treadmill test during a clinical examination while adhering to the modified Balke protocol,[28] as previously described.[30] Based on the final treadmill speed and grade, which is correlated highly with maximal oxygen uptake,[30–32] we computed maximal metabolic equivalent (METs), where 1 MET = 3.5 mL O2 uptake/kg body weight/minute.[29,33] In accordance with a standardized CCLS approach,[30,34] participants' were categorized into age- and sex-specific quintiles and then grouped into the following three categories: 1) quintile 1, low fitness; 2) quintiles 2 and 3, moderate fitness; and 3) quintiles 4 and 5, high fitness. For analyses, the low fitness group was regarded the reference group. In addition, physical activity was based on questions pertaining to the frequency (sessions per week) and duration (on average) of activities in the three previous months.[35] These included aerobic activities, such as walking, jogging, or running.[36] The reported frequency and duration of activity were converted to minutes of activity per week and multiplied by an estimated MET value based on the Compendium of Physical Activities.[37] This resulted in metabolic equivalent of task (MET) minutes per week (MET·min·wk−1) for each participant. Based on the Health and Human Services Physical Activity Guidelines for Americans (1), three categories were constructed: 1) not meeting guidelines (<500 MET·min·wk−1); 2) meeting guidelines (500–1000 MET·min·wk−1); and 3) exceeding guidelines (>1000 MET·min·wk−1).

The dependent variables consisted of current alcohol consumption and dependence (i.e., clinically relevant alcohol problems). In the medical history questionnaire, participants indicating that they consumed alcoholic beverages (i.e., current drinkers), were asked to specify the number of drinks per week of beer (12 oz), wine (5 oz), and hard liquor (1.5 oz) they consumed.[28] Consistent with previous research,[9,28] current drinking was grouped into three categories for participants age 18 to 64 yr: 1) light drinking, three or less drinks per week; 2) moderate drinking, more than three to seven drinks per week (women) and more than three to 14 drinks per week (men); and 3) heavy drinking, more than seven drinks per week (women) and more than 14 drinks per week (men). Among study participants 65 yr or older, to adhere to American Geriatric Society and the National Institute for Alcohol Abuse and Alcoholism recommendations,[38] moderate drinking was regarded as more than three to seven weekly drinks for both women and men, whereas heavy drinking was considered more than seven drinks a week for both sexes. Those participants at the heavy drinking level exceeded recommended weekly drinking guidelines.[39]

In addition, alcohol dependence was assessed via participants' responses to the Cut down, Annoyed, Guilty, Eye opener (CAGE) questionnaire.[40–42] This questionnaire, aimed at screening for clinically relevant alcohol problems, is designed for a clinical setting because of its brevity and focus on behavioral aspects of drinking to facilitate clinician–patient discussions.[40] The CAGE questionnaire consists of four questions inquiring whether patients ever: 1) felt they needed to cut down on drinking; 2) felt annoyed by criticism pertaining to their drinking; 3) felt guilty about drinking; and 4) drank first thing in the morning (eye opener).[40,41] Total scores on the questionnaire range from 0 to 4, with a threshold of 2, indicative of suggested alcohol dependence.[41] Subsequently, in the analyses, the total CAGE score was dichotomized into suggested dependence (score, ≥2) or no suggested dependence (score, <2).[43]

Covariates

The covariates, based on the literature,[44–46] adjusted for in multivariable analyses, included the following: age, sex, marital status (married: yes/no), and BMI. In addition, based on reported age and date of examination, participants were placed into four birth year cohorts, because a cohort effect independent of age has been observed in the literature pertaining to drinking habits.[47] Specifically, participants were placed into the following birth cohorts: Silent (born ≤1942), Baby Boomers (born 1943–1960), Generation X (born 1961–1981), and Millennials (born ≥1982).[48,49]

Statistical Analysis

Characteristics of participants were summarized in total and by sex. The prevalence of light, moderate, and heavy drinking across low, moderate, and high fitness categories was summarized by sex and the unadjusted trends between the two ordinal variables was tested using Jonckheere–Terpstra statistics. Adjusted drinking level odds ratios (OR) for moderate and high versus low fitness were estimated using ordinal logistic regression. Specifically, partial proportional odds models were fit to higher versus lower levels of drinking. Consequently, there were separate logits for 1) light versus moderate and heavy drinking and 2) light and moderate versus heavy drinking, with the latter comparing those participants who adhered to versus exceeded weekly drinking recommendations. Odds ratios were adjusted for age, birth year cohort, marital status, and BMI. Adjusting for the decade of examination year (instead of birth year cohort) had little bearing on the primary estimates and was therefore not included in the models presented. Moreover, in a parallel analysis, self-reported physical activity replaced fitness as the exposure of interest (see Table, Supplemental Digital Content, Physical Activity and Alcohol Consumption among Women and Men, http://links.lww.com/MSS/C421). In addition, the prevalence of total CAGE score ≥2 across low, moderate, and high fitness categories was summarized by sex, and adjusted trends were tested using multiple logistic regression of the binary outcome. All analyses were programmed in SAS/STAT®, version 9.4 (SAS Institute, Inc., Cary, NC).

processing....