ORIGINAL CONTRIBUTION Association of Alcohol Consumption With Brain Volume in the Framingham Study Carol Ann Paul, MS; Rhoda Au, PhD; Lisa Fredman, PhD; Joseph M. Massaro, PhD; Sudha Seshadri, MD; Charles DeCarli, MD; Philip A. Wolf, MD Background: While adults who drink low to moderate Participants: A total of 1839 subjects from the Framing- amounts of alcohol have lower rates of cardiovascular dis- ham Offspring Study who had magnetic resonance ease than other adults, the effect of alcohol on the brain imaging of the brain between 1999 and 2001. is less clear. There is evidence that drinking large amounts of alcohol is related to brain atrophy. It is uncertain what Results: Most participants reported low alcohol con- the effects of low to moderate consumption might be. sumption, and men were more likely than women to be moderate or heavy drinkers. There was a significant nega- Objective: To determine whether consumption of tive linear relationship between alcohol consumption and smaller amounts of alcohol negatively affects brain vol- TCBV (r=−0.25; PϽ.001). This relationship was modi- ume or is protective in reducing the well-documented fied by sex, with alcohol consumption having a stronger age-related decline in brain volume. association with TCBV in women than in men (r=−0.29 vs −0.20). Design: Total cerebral brain volume (TCBV) was com- puted, correcting for head size. Multivariate linear re- Conclusions: In contrast to studies on cardiovascular gression models were used to evaluate the association be- disease, this study found that moderate alcohol con- tween 5 categories of alcohol consumption (abstainers, former drinkers, low, moderate, high) and TCBV, ad- sumption was not protective against normal age-related justing for age, sex, education, height, body mass index differences in total brain volume. Rather, the more alco- (calculated as weight in kilograms divided by height in hol consumed, the smaller the total brain volume. meters squared), and the Framingham Stroke Risk Pro- file. Pairwise comparisons were also conducted be- tween the alcohol consumption groups. Arch Neurol. 2008;65(10):1363-1367 ANY STUDIES HAVE Excess alcohol consumption has often considered the costs been correlated with decline in cognition Author Affiliations: and benefits of alco- and can lead to Korsakoff syndrome.11 Yet Neuroscience Program, hol consumption in di- moderate alcohol consumption has been Wellesley College, Wellesley, verse populations. Massachusetts (Ms Paul); associated with improved cognitive func- Moderate alcohol consumption has fre- 12-14 Departments of Neurology M tioning, a lower risk of Alzheimer dis- quently been reported to have a benefi- 15 16,17 (Drs Au, Seshadri, and Wolf), ease (AD), and less severe WMH, al- cial effect on cardiovascular disease Epidemiology (Ms Paul and 1-3 though one study found no protective Dr Fredman), and Biostatistics, (CVD). Because the brain is perfused by effect of moderate alcohol intake on ce- the cardiovascular system, moderate al- Boston University School of rebral infarction.18 cohol consumption may attenuate age- Public Health, Boston, The effect of alcohol on brain volume Massachusetts (Dr Massaro); related decline in brain volume. Department of Mathematics & It is recognized that brain volume de- and WMH, however, has not been exam- Statistics, Boston University, clines with age4 at an estimated rate of 1.9% ined in a community-based sample that is Boston, Massachusetts per decade,5 while white matter lesions free of clinically evident neurological dis- (Dr Massaro); Framingham (WMH) increase with age.6 In addition, de- ease. This cross-sectional study tested the Heart Study, Framingham, cline in brain volume and increase in hypothesis that low or moderate alcohol Massachusetts (Drs Au, WMH accompany the progression of de- consumption was associated with larger Massaro, Seshadri, and Wolf); 7,8 9 and Department of Neurology mentia and cognitive deficits. Lower brain volume and less WMH when com- and Center for Neuroscience, brain volumes and larger WMH are also pared with no drinking or high levels of University of California, Davis, observed in persons with higher cardio- alcohol consumption in a sample of com- California (Dr DeCarli). vascular risk.5,10 munity-dwelling adults. (REPRINTED) ARCH NEUROL / VOL 65 (NO. 10), OCT 2008 WWW.ARCHNEUROL.COM 1363 Downloaded from www.archneurol.com at NAVARRA UNIV, on October 15, 2008 ©2008 American Medical Association. All rights reserved. through seventh grade, [3] high school or some college, and Table 1. Distribution of Alcohol Consumption in the Total [4] college graduate), height, body mass index (BMI; calcu- Sample, Men, and Womena lated as weight in kilograms divided by height in meters squared), and Framingham Stroke Risk Profile (FSRP, an estimate of the No. (%) 10-year risk of stroke on a scale of 0 to 1 based on age, sex, Total, systolic blood pressure, antihypertensive therapy, diabetes, No. Men Women smoking status, history of CVD, and the presence of atrial fi- Participants 1839 861 (46.81) 978 (53.18) brillation and left ventricular hypertrophy19). Alcohol consumption Abstainers 71 23 (2.67) 48 (4.91) Former drinkers 545 217 (25.20) 328 (33.54) STATISTICAL ANALYSIS Lowb 762 326 (37.86) 436 (44.58) Moderatec 276 163 (18.93) 113 (11.55) d The mean TCBV and WMHV were calculated for each of the 5 High 185 132 (15.33) 53 (5.42) alcohol consumption groups and the means were compared using an analysis of covariance with pairwise comparisons a Sample size is 1839 participants from the Framingham Offspring Cohort. (␣=.0125 for each pairwise comparison). Means were tested b Defined as 1 to 7 drinks per week. c for a linear trend across alcohol consumption groups at a level Defined as 8 to 14 drinks per week. ␣ d Defined as more than 14 drinks per week. of significance of =0.5. The initial model included all covar- iates; those covariates that did not contribute significantly to the model were eliminated 1 at a time, provided that their ex- METHODS clusion did not alter the ␤ coefficient for the alcohol consump- tion term and that they were not clinically important. Sex was SUBJECTS assessed as an effect modifier to determine whether the asso- ciation between amount of alcohol consumption and TCBV and The Framingham Offspring cohort, begun in 1971, included WMH differed between men and women. Analyses were per- children from the original Framingham Heart Study cohort and formed using SAS version 8.2 (SAS Inc, Cary, North Caro- their spouses (n=5124) aged 33 to 88 years during the study. lina). Study participants had had 7 health examinations every 4 to 8 years for the past 30 years. Participants at examination 7 RESULTS (n=3539; 1999-2001) were invited to have a magnetic reso- nance image (MRI) of the brain; 1886 participants were im- aged through August 2001. Major reasons for not being im- The sample included 1839 participants aged 33 to 88 years aged were residence out of state, medical contraindications for (mean [SD] age,60.64 [9.42] years). There were 861 men MRI such as pacemakers, and claustrophobia. An additional 47 and 978 women. As shown in Table 1, most partici- participants were excluded for prevalent stroke, dementia, or pants reported consuming 1 to 7 drinks per week, fol- other neurological disorder, resulting in a sample of 1839 re- lowed by former drinkers. While women were more likely spondents without stroke or dementia. Unpublished data find than men to be abstainers, former drinkers, or those who that those who did not have a brain MRI were older and had a drink 1 to 7 drinks per week, men were more likely to higher risk of cardiovascular disease, suggesting that our find- report moderate or high alcohol consumption. ings are a conservative estimate of the relationship between al- For the total sample, the mean (SD) TCBV was 77.40% cohol consumption and brain volume. (3.41%) for men (range, 64.89%-85.52%) and 78.35% (2.94%) (range, 67.84%-85.85%) for women. MEASUREMENT OF TOTAL BRAIN VOLUME The covariates for each group were evaluated sepa- rately for women and men (Table 2). Among women, The total (parenchymal) brain volume (TBV) and WMH vol- umes were corrected for total cranial volume (TCV); the re- amount of alcohol consumption was significantly asso- sulting measures are TCBV=(TBV/TCV)ϫ100 and WMHV= ciated with each covariate (age, sex, BMI, and FSRP score), (WMH/TCV)ϫ100, as described by DeCarli et al,4 where TCBV but among men, alcohol consumption was associated only indicates total cerebral brain volume; WMHV, white matter le- with FSRP score. Notably, women who consumed mod- sion volume. erate amounts of alcohol had the lowest BMI, most edu- cation, and lowest FSRP scores, while men who drank MEASUREMENT OF ALCOHOL CONSUMPTION low amounts had the lowest FSRP scores. Participants reported at examination 7 the number of alco- ASSOCIATION BETWEEN ALCOHOL holic drinks per week (beer, white wine, red wine, or liquor) CONSUMPTION AND TCBV AND WMHV they consumed during the past month. Alcohol consumption IN THE TOTAL SAMPLE was recorded as a continuous variable and participants were classified into 1 of 5 categories that have been used in other In unadjusted analyses, participants who had low alco- studies14,17,18: abstainers, former drinkers (drinkers based on their drinking status at earlier examinations), low (1-7 drinks per hol consumption had slightly larger TCBV compared with week), moderate (8-14 drinks per week), and high (Ͼ14 drinks the other groups. However, when adjusted for covari- per week). ates (age, sex, education, BMI, and FSRP score), there was a significant negative linear association between the ␤ COVARIATES amount of alcohol consumed and TCBV ( coeffi- cient,−0.25; PϽ.01) (Figure 1). This slope of −0.25 is Covariates obtained at examination 7 included age (in years), slightly larger than the average decline in brain volume sex, education (categorized as [1] fourth grade or less, [2] fifth per year with aging.