The effect of alcohol and red wine consumption on clinical and MRI outcomes in multiple sclerosis

The effect of alcohol and red wine consumption on clinical and MRI outcomes in multiple sclerosis

Multiple Sclerosis and Related Disorders 17 (2017) 47–53 Contents lists available at ScienceDirect Multiple Sclerosis and Related Disorders journal ...

313KB Sizes 4 Downloads 17 Views

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

Contents lists available at ScienceDirect

Multiple Sclerosis and Related Disorders journal homepage: www.elsevier.com/locate/msard

The effect of alcohol and red wine consumption on clinical and MRI outcomes in multiple sclerosis

MARK

Camilo Diaz-Cruza, Alicia S. Chuaa, Muhammad Taimur Malika, Tamara Kaplana, Bonnie I. Glanza,b, Svetlana Egorovaa,b, Charles R.G. Guttmanna,d, Rohit Bakshia,b,d, ⁎ Howard L. Weinera,b, Brian C. Healya,b,c, Tanuja Chitnisa,b,e, a

Partners Multiple Sclerosis Center, Brigham and Women's Hospital, Boston, MA, USA Department of Neurology, Harvard Medical School, Boston, MA, USA c Biostatistics Center, Massachusetts General Hospital, Boston, MA, USA d Department of Radiology, Brigham and Women's Hospital, Boston, MA, USA e Partners Pediatric Multiple Sclerosis Center, Massachusetts General Hospital, Boston, MA, USA b

A R T I C L E I N F O

A B S T R A C T

Keywords: Multiple sclerosis Alcohol EDSS MSSS MRI

Background: Alcohol and in particular red wine have both immunomodulatory and neuroprotective properties, and may exert an effect on the disease course of multiple sclerosis (MS). Objective: To assess the association between alcohol and red wine consumption and MS course. Methods: MS patients enrolled in the Comprehensive Longitudinal Investigation of Multiple Sclerosis at the Brigham and Women's Hospital (CLIMB) who completed a self-administered questionnaire about their past year drinking habits at a single time point were included in the study. Alcohol and red wine consumption were measured as servings/week. The primary outcome was the Expanded Disability Status Scale (EDSS) at the time of the questionnaire. Secondary clinical outcomes were the Multiple Sclerosis Severity Score (MSSS) and number of relapses in the year before the questionnaire. Secondary MRI outcomes included brain parenchymal fraction and T2 hyperintense lesion volume (T2LV). Appropriate regression models were used to test the association of alcohol and red wine intake on clinical and MRI outcomes. All analyses were controlled for sex, age, body mass index, disease phenotype (relapsing vs. progressive), the proportion of time on disease modifying therapy during the previous year, smoking exposure, and disease duration. In the models for the MRI outcomes, analyses were also adjusted for acquisition protocol. Results: 923 patients (74% females, mean age 47 ± 11 years, mean disease duration 14 ± 9 years) were included in the analysis. Compared to abstainers, patients drinking more than 4 drinks per week had a higher likelihood of a lower EDSS score (OR, 0.41; p = 0.0001) and lower MSSS (mean difference, − 1.753; p = 0.002) at the time of the questionnaire. Similarly, patients drinking more than 3 glasses of red wine per week had greater odds of a lower EDSS (OR, 0.49; p = 0.0005) and lower MSSS (mean difference, − 0.705; p = 0.0007) compared to nondrinkers. However, a faster increase in T2LV was observed in patients consuming 1–3 glasses of red wine per week compared to nondrinkers. Conclusions: Higher total alcohol and red wine intake were associated with a lower cross-sectional level of neurologic disability in MS patients but increased T2LV accumulation. Further studies should explore a potential cause-effect neuroprotective relationship, as well as the underlying biological mechanisms.

1. Background Several observational studies have linked moderate alcohol intake with a lower risk of developing autoimmune diseases such as systemic lupus erythematosus, rheumatoid arthritis, and autoimmune hypothyroidism (Carle et al., 2012; Liao et al., 2009; Wang et al., 2008).



Additionally, recent epidemiological studies investigating the association between alcohol consumption and the risk of multiple sclerosis (MS) have shown conflicting results. For instance, data from two large population-based case-control studies in Sweden showed that the risk of MS was significantly lower in males and females reporting high alcohol consumption (> 112 g/week in women and > 168 g/week for men)

Correspondence to: Partners Multiple Sclerosis Center, 60 Fenwood Road, Office 9002L, Boston, MA 02115-6128, USA. E-mail address: [email protected] (T. Chitnis).

http://dx.doi.org/10.1016/j.msard.2017.06.011 Received 9 March 2017; Received in revised form 25 May 2017; Accepted 21 June 2017 2211-0348/ © 2017 Published by Elsevier B.V.

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

C. Diaz-Cruz et al.

and the type of alcoholic beverage most frequently consumed (Supplementary Appendix). A standard drink key was provided to all patients indicating that one drink is defined as a 5-ounce glass of wine, 12-ounce bottle of beer or one shot (1.4 oz) of 80-proof spirits (vodka, rum, whiskey, tequila) (NIAAA). A single questionnaire was completed by each subject so longitudinal changes in alcohol use could not be determined.

when compared to abstainers. Furthermore, participants drinking more than 3 glasses of wine per week had a 30% lower risk of MS (Hedstrom et al., 2014). On the other hand, in the Nurse's Health Study (NHS) cohort, no association was found between alcohol intake and the incidence of MS (Massa et al., 2013). Furthermore, no significant associations were found when beer, liquor, and wine were analyzed independently. There are limited studies investigating the role of alcohol consumption on already established MS. Using a cross-sectional design, D’hooghe and collaborators demonstrated that in relapsing-remitting MS patients, regular alcohol consumption was associated with a decreased risk for reaching EDSS 6; however, this was not the case for patients with progressive-onset MS (D'Hooghe M et al., 2012). Another cross-sectional study showed a U-shaped effect of alcohol consumption on EDSS scores. Using a non-linear regression model, Foster, et al. found that mean EDSS was lower in patients who consumed alcohol for less than 15 years compared to those who did not drink or who had consumed it for longer than 15 years (Foster et al., 2012). These studies are limited by their cross-sectional designs. Red wine contains significant levels of resveratrol, a polyphenol which has been shown to possess pleiotropic effects including antineoplastic, anti-oxidant and anti-inflammatory actions, (Baur and Sinclair, 2006; Magrone and Jirillo, 2010), and its effects have not been explored in MS. The aim of this study was to assess the effects of alcohol and red wine consumption on clinical and MRI outcomes in a cohort of MS patients.

2.3. Clinical and neuroimaging outcomes Our primary outcome was the EDSS at the time of the questionnaire. Secondary clinical outcomes of interest were the number of relapses in the year prior to the questionnaire and the Multiple Sclerosis Severity Score (MSSS) at the time of the questionnaire. The MSSS corrects EDSS for disease duration by comparing an individual's disability with the distribution of scores in patients having similar disease duration (Roxburgh et al., 2005). The MSSS was obtained for every patient using the software available at: https://www-gene.cimr.cam.ac.uk/MSgenetics/GAMES/MSSS/ Readme.html. We used brain MRI scans obtained within ± 90 days of the questionnaire to measure the following secondary MRI outcomes: brain parenchymal fraction (BPF) and T2 hyperintense lesion volume (T2LV). 2.4. Image acquisition and processing Of the 923 subjects, 418 subjects had both an MRI scan within ± 90 days of the clinic visit when the alcohol questionnaire was completed and an MRI scan within ± 90 days of the previous year's clinic visit. 211 of these subjects had at least one scan that was not usable for the following reasons: technical processing failure in the automated segmentation pipeline, major deviation from the CLIMB MRI acquisition protocol, missing the previous year's MRI scan for subjects who had their first alcohol questionnaire at enrollment, and unsynchronized MRI scans for CLIMB study visits (every 6 months) that do not require an MRI (which is only performed every 12 months). Therefore, 207 subjects had MRI scans for the required time points that contributed to our analysis. Brain MRI scans were acquired on the 1.5 T Signa GE family scanners at the Brigham and Women's Hospital. MRI protocols included axial dual-echo conventional spin-echo images: TR = 2350–3167 msec, TE1/TE2 30/80 msec, slice thickness 3 mm, with no inter-slice gaps, and pixel size 0.7812–0.9375 mm2. Quantitative image analysis was performed using an automated segmentation pipeline (TDS +) (Wei et al., 2002). After the pipeline is completed, MRI scans undergo manual correction to adjust the segmentation of T2 lesions, brain parenchyma, and CSF by an expert reader using 3D Slicer software (http:// slicer.org/). Brain parenchymal fraction (BPF) was calculated by the following formula: BPF = (GM + WM + T2LV)/ICC, (Wei et al., 2004) where GM is the global gray matter volume, WM is the global white matter volume, T2LV is the total cerebral hyperintense lesion volume, and ICC is the volume of the intracranial cavity serving as a normalization for individual head size (Kikinis et al., 1992).

2. Methods 2.1. Subjects A subset of 923 patients enrolled in the Comprehensive Longitudinal Investigation of Multiple Sclerosis at the Brigham and Women's Hospital (CLIMB) study at the Partners MS Center, (Gauthier et al., 2006) who completed a self-administered questionnaire about their past year drinking habits, were included in this study. Table 1 provides the demographic and clinical characteristics of these subjects broken down based on alcohol consumption, and Supplementary Table 1 compares their demographic and clinical characteristics with those of subjects enrolled in the CLIMB study who did not complete the questionnaire. The CLIMB is an ongoing prospective cohort study initiated in 2000. Information about attacks, visits, and treatment is prospectively entered into a relational database. Patients included had a diagnosis of MS according to the 2010 McDonald criteria (Polman et al., 2011). Patients were excluded if they had a diagnosis of clinically isolated syndrome, neuromyelitis optica, or acute disseminated encephalomyelitis. As part of the CLIMB study, demographic characteristics, estimated date of MS symptoms onset and previous MS treatments are recorded for every patient at the enrollment visit. Patients then have a complete neurological examination every six months including a measurement of the Expanded Disability Status Scale score (EDSS) (Kurtzke, 1983). Symptoms suggestive of a new relapse are recorded along with their date of onset. Additionally, patients undergo annual brain MRI scans at the Brigham and Women's Hospital. This study complied with the Declaration of Helsinki and was approved by the Partners Institutional Review Board. All participants provided written informed consent.

2.5. Statistical analysis The association between alcohol consumption and EDSS at the time of the questionnaire was assessed using a proportional odds cumulative logit model. For MSSS, BPF and T2LV at the time of the questionnaire, the association with alcohol consumption was evaluated using multiple linear regression models. Due to the skewness observed in the distribution of T2LV, a log transformation was applied to T2LV prior to all analyses. In the primary analysis, the number of drinks per week was added to the model as a categorical variable using four categories (0, 0–1, 1–4, > 4 drinks per week). In a second model to investigate the potential for a U-shaped relationship between alcohol consumption and each outcome, a model with both a linear and quadratic effect of

2.2. Alcohol history During their visit at the Partners MS Center, patients completed a self-reported alcohol questionnaire about their drinking habits in the previous year. The questionnaire was based on and modified from the questions in alcohol research recommended by the National Institute on Alcohol Abuse and Alcoholism (NIAAA), and it was used to estimate average weekly alcohol and red wine intake during the previous year 48

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

C. Diaz-Cruz et al.

Table 1 Demographic characteristics of patients by alcohol consumption frequency. Alcohol consumption (servings/week) Characteristic

Never drinkers

≤1

> 2–4

>4

No. of patients Age, mean (SD), y Female sex, No. (%) Race, No. (%) American Indian/Alaska Native Asian Black More than 1 race White Unknown/unreported EDSS score, median (IQR; range) MSSS, mean (SD) Disease duration, mean (SD), y Disease category, No. (%) Relapsing-remitting Secondary progressive Primary progressive Progressive relapsing Season-adjusted 25(OH)D level, mean (SD), ng/ml BMI, mean (SD), kg/m2 Proportion of time on DMT in the year prior to cross-sectional questionnaire date, % (SD) Ever smoker, No. (%) Most frequently consumed - type of alcohol, No. (%) Red wine White wine Beer Malt liquor 80-proof spirits Unknown/unreported

114 50.83 (11.05) 88 (77.19)

253 45.30 (10.53) 204 (80.63)

426 46.70 (10.53) 316 (74.18)

130 49.65 (12.80) 77 (59.23)

0 (0.00) 1 (0.88) 9 (7.89) 4 (3.51) 95 (83.33) 5 (4.39) 2.5 (4.5; 0.0–8.0) 3.35 (2.72) 15.58 (9.76)

1 (0.40) 2 (0.79) 13 (5.14) 7 (2.77) 228 (90.12) 2 (0.79) 2.0 (2.0; 0.0–8.5) 2.64 (2.37) 13.09 (8.10)

0 (0.00) 2 (0.47) 10 (2.35) 4 (0.94) 408 (95.77) 2 (0.47) 1.5 (3.0; 0.0–8.0) 2.13 (2.16) 13.55 (8.86)

0 (0.00) 0 (0.00) 0 (0.00) 1 (0.77) 128 (98.46) 1 (0.77) 1.5 (2.5; 0.0–7.5) 2.00 (2.09) 15.35 (11.04)

75 (65.79) 30 (26.32) 7 (6.14) 2 (1.75) 37.85 (15.01) 28.31 (7.03) 71.79 (4.47) 57 (50.00)

209 (82.61) 39 (15.42) 4 (1.58) 1 (0.40) 36.02 (14.64) 28.90 (6.67) 76.54 (4.01) 124 (49.01)

364 (85.45) 43 (10.09) 17 (3.99) 2 (0.47) 39.85 (15.12) 26.65 (5.51) 80.38 (3.71) 204 (47.89)

106 (81.54) 17 (13.08) 6 (4.62) 1 (0.77) 41.96 (17.54) 26.20 (5.17) 79.47 (3.87) 77 (59.23)

N/A N/A N/A N/A N/A N/A

63 (24.90) 72 (28.46) 64 (25.30) 5 (7.94) 22 (8.70) 27 (10.67)

146 (34.27) 101 (23.71) 114 (26.76) 2 (0.47) 49 (11.50) 14 (3.29)

48 (36.92) 25 (19.23) 39 (30.00) 2 (1.54) 15 (11.54) 1 (0.77)

Legend: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); DMT, disease modifying therapy; EDSS, Expanded Disability Status Scale; IQR, interquartile range; MSSS, Multiple Sclerosis Severity Score; SD, standard deviation; 25(OH)D, 25-hydroxyvitamin D.

number of drinks per week was fit. All models were controlled for sex, age, body mass index (BMI), disease phenotype (relapsing vs. progressive), proportion of time on disease modifying therapy (DMT) during the previous year/exposed period, and smoking exposure (never or ever smoker). For all outcomes except MSSS, disease duration was also included in the model; disease duration was not included in the model for MSSS since it is involved in the calculation of this outcome. In the models for the MRI measures, analyses were also adjusted for acquisition protocol. To account for potential confounding due to vitamin D level, all of the previous analyses were run with a further adjustment for seasonally-adjusted vitamin D (Bove et al., 2014; Munger et al., 2006) in the subset of patients who had vitamin D levels measured within ± 90 days of the clinic visit when the alcohol questionnaire was completed (n = 413). Finally, all of the previous models were also fit using red wine consumption as the exposure of interest. For red wine consumption, the categories for the categorical variable analysis were 0, 0–1, 1–3, and > 3 drinks per week. Since our questionnaire assessed the amount of alcohol consumed in the previous year, we also investigated the association between alcohol consumption over the past year and change in each of the outcome measures over the same period. A subset of subjects (n = 648) had complete information allowing us to evaluate changes in clinical outcomes, and a subset of 207 patients had both baseline and previous year MRI data allowing us to assess changes in the BPF and T2LV. The impact of alcohol consumption on the change in the EDSS was investigated using a mixed effects proportional odds cumulative logit model. For the change in MSSS, BPF, and T2LV, a linear mixed effects model was used. Models for BPF and T2LV were adjusted for acquisition protocol (Chua et al., 2015). The primary analysis included the number of drinks per week in the model as a categorical variable using the categories described above. For every categorical model in the crosssectional and longitudinal analyses, the overall group effect was tested

using either an F-test or a likelihood ratio test. For the number of relapses in the year before the questionnaire, a negative binomial model was used to assess the association with alcohol consumption. As with the cross-sectional analysis, all of the previous models were also fit using red wine consumption as the exposure of interest. All analyses were performed using the Statistical Analysis System (SAS) 9.3 (Cary, NC) or Stata/IC version 14 (StataCorp, College Station TX). 3. Results 923 patients were included in the main cross-sectional analysis. Demographic and clinical characteristics of subjects grouped by weekly alcohol consumption levels are provided in Table 1. Demographic and clinical characteristics of subjects who contributed to cross-sectional, longitudinal and MRI outcomes analyses are presented in Supplementary Table 2. 3.1. Alcohol consumption We first performed a cross-sectional analysis, with alcohol consumption treated as a categorical variable, based on number of drinks per week. This showed a significant difference between the groups (Table 2). Furthermore, an increase in the number of drinks per week was associated with an improvement on the EDSS or MSSS, and the subjects in the highest alcohol consumption categories had the lowest EDSS and MSSS scores. For instance, Table 2 shows that the estimated odds that a patient drinking more than four drinks per week have a higher EDSS score is 0.41 times the odds for abstainers. In other words, they are less likely to have a score at the higher end of the EDSS. When we included linear and quadratic terms in the model, a significant quadratic effect was observed (p < 0.05) for the analyses for both EDSS and MSSS (Fig. 1). The quadratic models showed that the estimated 49

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

C. Diaz-Cruz et al.

Table 2 Estimated effects and odds ratio of weekly alcohol consumption on MS outcome measures. Cross-sectional outcome Expanded disability status scale score

Multiple sclerosis severity score

Brain parenchymal fraction

T2 hyperintense lesion volume

Longitudinal outcome Expanded disability status scale score

Multiple sclerosis severity score

Brain parenchymal fraction

T2 hyperintense lesion volume

Relapse count in the previous year

Category Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Category Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week

Non-drinker ≤ 1 drink per week > 1–4 drinks per week > 4 drinks per week

Adjusted odds ratio Reference 0.87 0.57 0.41 Adjusted difference Reference − 0.18 − 0.50 − 1.75 Reference − 0.002 − 0.009 − 0.016 Reference − 0.09 0.02 0.17 Adjusted regression coefficient Reference 0.54 0.44 0.59 Adjusted difference Reference 0.38 0.34 0.43 Reference −0.00003 −0.002 −0.001 Reference 0.17 0.19 0.23 Adjusted rate ratio

95% Confidence interval

p-value

0.58–1.29) 0.39–0.83) 0.26–0.64) 95% Confidence interval

0.478 0.003 < 0.001 p-value

(− 0.59 to 0.23) (− 0.88 to − 0.11) (− 1.22 to − 0.28)

0.393 0.012 0.002

(− 0.026 to 0.022) (− 0.032 to 0.014) (− 0.042 to 0.011)

0.873 0.450 0.242

(− 0.66 to 0.49) (− 0.54 to 0.57) (− 0.45 to 0.79) 95% Confidence interval

0.768 0.950 0.587 p-value

(−0.18 to 1.26) (−0.19 to 1.06) (−0.18 to 1.37) 95% Confidence interval

0.141 0.172 0.132 p-value

0.03–0.74) 0.03–0.64) 0.05–0.80)

0.034 0.034 0.027

(−0.005 to 0.005) (−0.007 to 0.003) (−0.007 to 0.005)

0.991 0.474 0.754

0.01–0.33) 0.04–0.34) 0.06–0.41) 95% Confidence interval

0.033 0.015 0.010 p-value

0.57–2.43) 0.53–2.07) 0.35–2.03)

0.665 0.888 0.710

Global F-test p-value < 0.0001

Global F-test p-value 0.003

0.451

0.711

Global F-test p-value 0.409

Global F-test p-value 0.104

0.526

0.066

Reference 1.17 1.05 0.85

Global Likelihood ratio test pvalue 0.850

Legend: All models were controlled for sex, age, body mass index (BMI), disease phenotype (relapsing vs. progressive), proportion of time on disease modifying therapy (DMT) during the previous year/exposed period, and smoking exposure (never or ever smoker). For all outcomes except MSSS, disease duration was also included in the model; disease duration was not included in the model for MSSS since it is involved in the calculation of this outcome. In the models for the MRI measures, analyses were also adjusted for the acquisition protocol.

mean EDSS and MSSS decreased until approximately 10 drinks per week. This departure from linearity was not observed in the categorical model because the highest category was > 4 drinks/week. Interestingly, this quadratic model also demonstrated that an alcohol intake greater than 10 drinks per week was associated with an increase in the estimated mean EDSS and MSSS values (Fig. 1). No significant association was observed between weekly alcohol consumption and BPF or T2LV in the primary analysis. When season-adjusted vitamin D level was included as a covariate, the results were similar, but the association with MSSS was not statistically significant in this smaller sample (data not shown). In the longitudinal analysis, we assessed the association between alcohol consumption and change in each outcome measure over the previous year. No significant association was observed (Table 2, longitudinal analyses). These results did not change after controlling for season-adjusted vitamin D levels (data not shown).

3.2. Red wine consumption Fig. 1. Scatter plot showing MSSS distribution according to alcohol consumption with quadratic fit. Dashed lines represent 95% confidence intervals.

When we assessed the association between weekly red wine consumption and all outcomes at the time of the questionnaire (crosssectional analysis), increased red wine consumption was associated with lower EDSS score and lower MSSS (Table 3). No significant 50

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

C. Diaz-Cruz et al.

Table 3 Estimated effects and odds ratio of weekly red wine consumption on MS outcome measures. Cross-sectional outcome Expanded disability status scale score

Multiple sclerosis severity score

Brain parenchymal fraction

T2 hyperintense lesion volume

Category Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week

Longitudinal outcome Expanded disability status scale score

Multiple sclerosis severity score

Brain parenchymal fraction

T2 hyperintense lesion volume

Relapse count in the previous year

Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week

Non-drinker ≤ 1 drink per week > 1–3 drinks per week > 3 drinks per week

Adjusted odds ratio Reference 0.83 0.73 0.49 Adjusted difference Reference − 0.26 − 0.30 − 0.71 Reference − 0.007 − 0.005 − 0.006 Reference 0.30 0.37 − 0.01 Adjusted regression coefficient Reference 0.53 0.12 0.24 Adjusted difference Reference 0.21 0.14 0.12 Reference 0.002 0.001 − 0.004 Reference 0.17 0.13 0.11 Adjusted rate ratio

95% Confidence interval

p-value

0.63–1.09) 0.53–1.02) 0.33–0.73) 95% Confidence interval

0.176 0.063 0.0005 p-value

(− 0.55 to 0.02) (− 0.64 to 0.03) (− 1.11 to − 0.30)

0.071 0.077 0.0007

(− 0.020 to 0.007) (− 0.020 to 0.011) (− 0.025 to 0.014)

0.354 0.535 0.557

(− 0.02 to 0.63) 0.01–0.73) (− 0.46 to 0.44) 95% Confidence interval

0.063 0.045 0.957 p-value

0.03–1.04) (− 0.43 to 0.67) (− 0.43 to 0.91) 95% Confidence interval

0.039 0.667 0.481 p-value

(− 0.04 to 0.45) (− 0.13 to 0.40) (− 0.21 to 0.44)

0.098 0.311 0.482

(− 0.001 to 0.005) (− 0.003 to 0.004) (− 0.008 to 0.0002)

0.225 0.724 0.059

0.07–0.26) 0.03–0.24) (− 0.03 to 0.25) 95% Confidence interval

0.001 0.013 0.111 p-value

0.76–1.99) 0.65–1.96) 0.44–2.07)

0.407 0.671 0.894

Global F-test p-value 0.004

Global F-test p-value 0.006

0.807

0.097

Global F-test p-value 0.218

Global F-test p-value 0.403

0.052

0.004

Reference 1.23 1.13 0.95

Global Likelihood ratio test pvalue 0.833

Legend: All models were controlled for sex, age, body mass index (BMI), disease phenotype (relapsing vs. progressive), proportion of time on disease modifying therapy (DMT) during the previous year/exposed period, and smoking exposure (never or ever smoker). For all outcomes except MSSS, disease duration was also included in the model; disease duration was not included in the model for MSSS since it is involved in the calculation of this outcome. In the models for the MRI measures, analyses were also adjusted for the acquisition protocol.

case for patients with progressive-onset MS (D'Hooghe M et al., 2012). Previous evidence has pointed out that moderate alcohol intake might have a beneficial impact on the immune system compared to alcohol abuse or abstinence (Romeo et al., 2007). Foster and colleagues showed in a cross-sectional study a U-shaped effect of alcohol consumption on EDSS scores (Foster et al., 2012). Considering these findings, we included a quadratic term for alcohol intake in our regression model for EDSS, and it showed a statistically significant quadratic effect. Therefore, we consider it reasonable to hypothesize that alcohol intake could have a U-shaped effect on EDSS score and other outcomes, thus suggesting there could be an optimal dose to suppress the immune system. Beyond that dose, the intrinsic neurotoxicity of ethanol overrides its immunosuppressive properties, thus increasing neurological disability. Our MRI findings yielded conflicting results. Red wine consumption showed an increase in the T2LV in moderate drinkers versus nondrinkers in longitudinal analysis. This could be explained by the direct neurotoxic effects of ethanol leading to the formation of T2 hyperintense lesions (Noble and Weimer, 2014; Zahr, 2014). We did not find an effect of either alcohol or red wine on brain atrophy cross-sectionally or longitudinally. Conversely, Foster and colleagues showed an inverted U-shaped relationship in which moderate duration of alcohol intake was associated with a greater normalized gray matter volume (Foster

quadratic effect was observed in either of these analyses. No significant association was found between red wine intake and any of the cross-sectional MRI outcomes. When season adjusted vitamin D level was included as a covariate, the results were similar (data not shown). In the longitudinal analysis, where the change over the previous year in MRI metrics was assessed, the results differed based on the outcome. For brain atrophy, increased red wine consumption was not associated with longitudinal brain atrophy (Table 3). For T2 lesion volume, comparing to nondrinkers, only patients consuming 3 drinks or fewer per week had a greater increase in longitudinal T2LV (Table 3).

4. Discussion This observational study provides further evidence of an inverse association between alcohol consumption and neurological disability scores in a cohort of patients with MS, after controlling for multiple potential confounders. Additionally, we found the same inverse association between red wine consumption and EDSS. However, we did not find an effect of alcohol or red wine consumption on longitudinal EDSS change. Our results are consistent with findings from other observational studies. D’hooghe and collaborators demonstrated that in relapsing-remitting MS patients, regular alcohol consumption was associated with a decreased risk for reaching EDSS 6; however, this was not the 51

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

C. Diaz-Cruz et al.

more than 1 drink per day after MS onset compared to 12.7% drinking more than 1 drink/day before MS (Foster et al., 2012). Additionally, the single time point measurement of alcohol consumption performed in this study limits our ability to assess reverse causality. Future studies collecting multiple measurements of alcohol consumption and EDSS might be able to determine their relationship. Although we adjusted all regression models for several potential confounders, we acknowledge that unmeasured confounding could be an explanation for our findings. Several other variables such as socioeconomic status, physical activity and the type of DMT could have confounded the relationship between alcohol consumption and disease measures in MS. For example, the differential effect of DMTs on disease activity or progression would constitute a classic example of confounding. Patients using DMTs daily or every other day (e.g. injections or oral medications), may not drink alcohol, because of concomitant medication use. However, those receiving intermittent intravenous infusions of DMTs may drink alcohol more easily, because they do not take daily medications. Finally, it is important to point out two additional limitations of this study. Firstly, the questionnaire allowed us to calculate the association between total alcohol/red wine intake and clinical/MRI outcomes. However, this tool was not designed to gather specific information about the quantity and frequency of consumption of other alcohol types (e.g. beer, white wine, 80-proof spirits). Therefore, we could not assess the potential effects of these other types. Lastly, the results obtained in the longitudinal analyses could have been affected by the short followup period of 1 year. Since all the patients included in this study are enrolled in the CLIMB cohort, we are collecting data at additional time points with the aim of preparing a follow-up manuscript describing long-term associations.

et al., 2012). However, we did not have information about this specific parameter to test in our sample. Thus, the association between moderate alcohol intake and EDSS may thus be mediated by a protective effect in the cerebral cortex, as opposed to T2LV. Likewise, a protective effect on gray matter atrophy in the spinal cord would have been undetected in our study. Altogether, alcohol may have a complex relationship on MRI in MS because of separate and opposing immune effects and other direct effects on the central nervous system. Furthermore, we cannot exclude the possibility that this finding was the consequence of a limited ability to control for other longitudinal determinants of T2 lesion formation such as the amount of cumulative tobacco smoking, diet, and physical activity. Our results raise the question of the mechanisms by which alcohol can potentially exert a protective effect in MS disability accrual. Previous studies have indicated alcohol impairs the ability of white blood cells to migrate to sites of injury and infection, to induce functional abnormalities in lymphocytes, natural killer cells, and monocytes/macrophages, and to alter cytokine production (Romeo et al., 2007). Mandrekar et al. administered vodka in a dose of 2 ml/kg body weight to 19 men and women between 21 and 60 years. They found out this single dose of alcohol attenuated monocyte inflammatory responses through inhibition of nuclear factor kappa β (NF-κB) and induction of interleukin 10, an anti-inflammatory cytokine (Mandrekar et al., 2006). Interestingly, these same effects on the immune system have been described in the mechanism of action of dimethyl fumarate, a new oral agent approved for the treatment of relapsing-remitting MS (Phillips and Fox, 2013). Besides ethanol, additional compounds such as flavonoids and resveratrol are present in red wine (Magrone and Jirillo, 2010) and may explain the inverse association we found between red wine intake and MSSS/EDSS scores. Flavonoids have been shown to increase the activity of the nuclear factor-kappa B inhibitor (IκBα), therefore reducing the activity of the pro-inflammatory NF-κB pathway (Comalada et al., 2005; Magrone and Jirillo, 2010). Moreover, resveratrol decreases neuroinflammation through the attenuation of microglial activation. This phenomenon has been explained by three main mechanisms: 1. Deactivation of reactive oxygen species synthesized by microglial cells, 2. Inhibition of mitogen-activated protein kinase, thus reducing the expression of multiple pro-inflammatory cytokines and 3. Activation of Sirtuin 1, a protein involved in the inhibition of the NF-kB pathway (Langley and Sauve, 2013; Zhang et al., 2010). In contrast to our results and the studies mentioned above, Pakpoor et al. demonstrated an increased risk of MS in patients with alcohol use, dependence or abuse as compared to the general population in England (Pakpoor et al., 2014). However, they acknowledge that the very high alcohol consumption reported in their exposure group may be positively associated with other MS factors not measured in their study. In fact, in a prospective study investigating the association between alcohol and MS risk, Massa et al. reported that vitamin D consumption was reduced in those women reporting the highest amount of alcohol intake (Massa et al., 2013). Since higher levels of vitamin D are associated with a reduced risk of developing MS and with decreased disease activity and progression (Ascherio et al., 2014; Fitzgerald et al., 2015; Soilu-Hanninen et al., 2012), the findings reported by Pakpoor et al. could be confounded by this inverse association between alcohol intake and vitamin D consumption. The results presented here have to be interpreted with caution because reverse causality may have influenced them. In particular, the socalled “sick quitter hypothesis” could be the cause of a reverse causation phenomenon where patients with higher EDSS scores tend to drink less alcohol because they feel sicker and are more physically and mentally disabled (Fillmore et al., 2007; Nova et al., 2012). In fact, in the study by Foster et. al., authors reported that the percentage of patients who did not consume alcohol after MS diagnosis was higher than before MS. Furthermore, 13.1% of patients indicated that they stopped drinking after their MS diagnosis and 3.0% reported consumption of

5. Conclusion In conclusion, our findings add to the increasing amount of literature describing an inverse association between moderate alcohol intake and neurological disability in MS patients and additionally show similar effects with red wine consumption. Considering that our results are also consistent with published studies describing an alcohol-mediated inhibition of molecular pathways involved in the activation of the immune system, we propose that the effect of alcohol consumption should be investigated in an animal model of MS. This subsequent step would help to further characterize a potential cause-effect relationship, as well as the underlying biological mechanisms. Conflict of interest The authors declare no conflict of interest with the above listed funding sources. Role of funding source This work was supported by the National Multiple Sclerosis Society RG-4256A4/2 (TC). The CLIMB Study has received support from Merck Serono and the Nancy Davis Center Without Walls. Acknowledgements The authors wish to thank the following colleagues at the Brigham and Women's Hospital: Mariann Polgar- Turcsanyi, MS and Mark Anderson, MS for their role in managing the Partners MS Center research database, as well as Taylor Saraceno, BA for her research assistance. Appendix A. Supplementary material Supplementary data associated with this article can be found in the 52

Multiple Sclerosis and Related Disorders 17 (2017) 47–53

C. Diaz-Cruz et al.

humans attenuates monocyte inflammatory responses: inhibition of nuclear regulatory factor kappa B and induction of interleukin 10. Alcohol. Clin. Exp. Res. 30 (1), 135–139. Massa, J., O'Reilly, E.J., Munger, K.L., Ascherio, A., 2013. Caffeine and alcohol intakes have no association with risk of multiple sclerosis. Mult. Scler. 19 (1), 53–58. Munger, K.L., Levin, L.I., Hollis, B.W., Howard, N.S., Ascherio, A., 2006. Serum 25-hydroxyvitamin D levels and risk of multiple sclerosis. JAMA: J. Am. Med. Assoc. 296 (23), 2832–2838. NIAAA, National Institute on Alcohol Abuse and Alcoholism. Recommended Alcohol Questions. 〈https://www.niaaa.nih.gov/research/guidelines-and-resources/ recommended-alcohol-questions〉 (Accessed 8 January 2016). NIAAA, National Institute on Alcohol Abuse and Alcoholism. What is a standard drink? 〈http://pubs.niaaa.nih.gov/publications/practitioner/pocketguide/pocket_guide2. htm〉 (Accessed 8 January 2016). Noble, J.M., Weimer, L.H., 2014. Neurologic complications of alcoholism. Continuum (Minneapolis, Minn.) 20(3 Neurology of Systemic Disease), pp. 624–641. Nova, E., Baccan, G.C., Veses, A., Zapatera, B., Marcos, A., 2012. Potential health benefits of moderate alcohol consumption: current perspectives in research. In: Proceedings of the Nutrition Society, England, pp. 307–315. Pakpoor, J., Goldacre, R., Disanto, G., Giovannoni, G., Goldacre, M.J., 2014. Alcohol misuse disorders and multiple sclerosis risk. JAMA Neurol. 71 (9), 1188–1189. Phillips, J.T., Fox, R.J., 2013. BG-12 in multiple sclerosis. Semin. Neurol. 33 (1), 56–65. Polman, C.H., Reingold, S.C., Banwell, B., Clanet, M., Cohen, J.A., Filippi, M., Fujihara, K., Havrdova, E., Hutchinson, M., Kappos, L., Lublin, F.D., Montalban, X., O'Connor, P., Sandberg-Wollheim, M., Thompson, A.J., Waubant, E., Weinshenker, B., Wolinsky, J.S., 2011. Diagnostic criteria for multiple sclerosis: 2010 Revisions to the McDonald criteria. Ann. Neurol. 69 (2), 292–302. Romeo, J., Warnberg, J., Nova, E., Diaz, L.E., Gomez-Martinez, S., Marcos, A., 2007. Moderate alcohol consumption and the immune system: a review. Br. J. Nutr. 98 (Suppl 1), S111–S115. Roxburgh, R.H., Seaman, S.R., Masterman, T., Hensiek, A.E., Sawcer, S.J., Vukusic, S., Achiti, I., Confavreux, C., Coustans, M., le Page, E., Edan, G., McDonnell, G.V., Hawkins, S., Trojano, M., Liguori, M., Cocco, E., Marrosu, M.G., Tesser, F., Leone, M.A., Weber, A., Zipp, F., Miterski, B., Epplen, J.T., Oturai, A., Sorensen, P.S., Celius, E.G., Lara, N.T., Montalban, X., Villoslada, P., Silva, A.M., Marta, M., Leite, I., Dubois, B., Rubio, J., Butzkueven, H., Kilpatrick, T., Mycko, M.P., Selmaj, K.W., Rio, M.E., Sa, M., Salemi, G., Savettieri, G., Hillert, J., Compston, D.A., 2005. Multiple sclerosis severity score: using disability and disease duration to rate disease severity. Neurology 64 (7), 1144–1151. Soilu-Hanninen, M., Aivo, J., Lindstrom, B.M., Elovaara, I., Sumelahti, M.L., Farkkila, M., Tienari, P., Atula, S., Sarasoja, T., Herrala, L., Keskinarkaus, I., Kruger, J., Kallio, T., Rocca, M.A., Filippi, M., 2012. A randomised, double blind, placebo controlled trial with vitamin D3 as an add on treatment to interferon beta-1b in patients with multiple sclerosis. J. Neurol. Neurosurg. Psychiatry 83 (5), 565–571. Wang, J., Pan, H.F., Ye, D.Q., Su, H., Li, X.P., 2008. Moderate alcohol drinking might be protective for systemic lupus erythematosus: a systematic review and meta-analysis. Clin. Rheumatol. 27 (12), 1557–1563. Wei, X., Guttmann, C.R., Warfield, S.K., Eliasziw, M., Mitchell, J.R., 2004. Has your patient's multiple sclerosis lesion burden or brain atrophy actually changed? Mult. Scler. 10 (4), 402–406. Wei, X., Warfield, S.K., Zou, K.H., Wu, Y., Li, X., Guimond, A., Mugler 3rd, J.P., Benson, R.R., Wolfson, L., Weiner, H.L., Guttmann, C.R., 2002. Quantitative analysis of MRI signal abnormalities of brain white matter with high reproducibility and accuracy. J. Magn. Reson. Imaging: JMRI 15 (2), 203–209. Zahr, N.M., 2014. Structural and microstructral imaging of the brain in alcohol use disorders. In: Handbook of Clinical Neurology, vol. 125, pp. 275–290. Zhang, F., Liu, J., Shi, J.S., 2010. Anti-inflammatory activities of resveratrol in the brain: role of resveratrol in microglial activation. Eur. J. Pharmacol. 636 (1–3), 1–7.

online version at http://dx.doi.org/10.1016/j.msard.2017.06.011. References Ascherio, A., Munger, K.L., White, R., Kochert, K., Simon, K.C., Polman, C.H., Freedman, M.S., Hartung, H.P., Miller, D.H., Montalban, X., Edan, G., Barkhof, F., Pleimes, D., Radu, E.W., Sandbrink, R., Kappos, L., Pohl, C., 2014. Vitamin D as an early predictor of multiple sclerosis activity and progression. JAMA Neurol. 71 (3), 306–314. Baur, J.A., Sinclair, D.A., 2006. Therapeutic potential of resveratrol: the in vivo evidence. Nat. Rev. Drug Discov. 5 (6), 493–506. Bove, R., Musallam, A., Healy, B., Raghavan, K., Glanz, B., Bakshi, R., Weiner, H., De Jager, P., Miller, K., Chitnis, T., 2014. Low testosterone is associated with disability in men with multiple sclerosis. Mult. Scler. Carle, A., Pedersen, I.B., Knudsen, N., Perrild, H., Ovesen, L., Rasmussen, L.B., Jorgensen, T., Laurberg, P., 2012. Moderate alcohol consumption may protect against overt autoimmune hypothyroidism: a population-based case-control study. Eur. J. Endocrinol./Eur. Fed. Endocr. Soc. 167 (4), 483–490. Chua, A.S., Egorova, S., Anderson, M.C., Polgar-Turcsanyi, M., Chitnis, T., Weiner, H.L., Guttmann, C.R., Bakshi, R., Healy, B.C., 2015. Handling changes in MRI acquisition parameters in modeling whole brain lesion volume and atrophy data in multiple sclerosis subjects: Comparison of linear mixed-effect models. NeuroImage Clin. 8, 606–610. Comalada, M., Camuesco, D., Sierra, S., Ballester, I., Xaus, J., Galvez, J., Zarzuelo, A., 2005. In vivo quercitrin anti-inflammatory effect involves release of quercetin, which inhibits inflammation through down-regulation of the NF-kappaB pathway. Eur. J. Immunol. 35 (2), 584–592. D'Hooghe M, B., Haentjens, P., Nagels, G., De Keyser, J., 2012. Alcohol, coffee, fish, smoking and disease progression in multiple sclerosis. Eur. J. Neurol. 19 (4), 616–624. Fillmore, K.M., Stockwell, T., Chikritzhs, T., Bostrom, A., Kerr, W., 2007. Moderate alcohol use and reduced mortality risk: systematic error in prospective studies and new hypotheses. Ann. epidemiol. S16–S23. Fitzgerald, K.C., Munger, K.L., Kochert, K., Arnason, B.G., Comi, G., Cook, S., Goodin, D.S., Filippi, M., Hartung, H.P., Jeffery, D.R., O'Connor, P., Suarez, G., Sandbrink, R., Kappos, L., Pohl, C., Ascherio, A., 2015. Association of Vitamin D levels with multiple sclerosis activity and progression in patients receiving interferon Beta-1b. JAMA Neurol. 72 (12), 1458–1465. Foster, M., Zivadinov, R., Weinstock-Guttman, B., Tamano-Blanco, M., Badgett, D., Carl, E., Ramanathan, M., 2012. Associations of moderate alcohol consumption with clinical and MRI measures in multiple sclerosis. J. Neuroimmunol. 243 (1–2), 61–68. Gauthier, S.A., Glanz, B.I., Mandel, M., Weiner, H.L., 2006. A model for the comprehensive investigation of a chronic autoimmune disease: the multiple sclerosis climb study. Autoimmun. Rev. 5 (8), 532–536. Hedstrom, A.K., Hillert, J., Olsson, T., Alfredsson, L., 2014. Alcohol as a modifiable lifestyle factor affecting multiple sclerosis risk. JAMA Neurol. 71 (3), 300–305. Kikinis, R., Shenton, M.E., Gerig, G., Martin, J., Anderson, M., Metcalf, D., Guttmann, C.R., McCarley, R.W., Lorensen, W., Cline, H., et al., 1992. Routine quantitative analysis of brain and cerebrospinal fluid spaces with MR imaging. J. Magn. Reson. Imaging: JMRI 2 (6), 619–629. Kurtzke, J.F., 1983. Rating neurologic impairment in multiple sclerosis: an expanded disability status scale (EDSS). Neurology 33 (11), 1444–1452. Langley, B., Sauve, A., 2013. Sirtuin deacetylases as therapeutic targets in the nervous system. Neurother.: J. Am. Soc. Exp. NeuroTher. 10 (4), 605–620. Liao, K.P., Alfredsson, L., Karlson, E.W., 2009. Environmental influences on risk for rheumatoid arthritis. Curr. Opin. Rheumatol. 21 (3), 279–283. Magrone, T., Jirillo, E., 2010. Polyphenols from red wine are potent modulators of innate and adaptive immune responsiveness. Proc. Nutr. Soc. 69 (3), 279–285. Mandrekar, P., Catalano, D., White, B., Szabo, G., 2006. Moderate alcohol intake in

53