Abstract
Alcohol, coffee, and tobacco consumption was assessed on 151 FTD outpatients and 151 matched controls in a multicenter retrospective case-control design. No association was found for smoking and coffee intake. The risk of FTD was decreased by alcohol consumption (adj. OR 0.30, 95% CI 0.14–0.63); risk reduction was significant in current alcohol consumers (adj. OR 0.22, 95% CI 0.10–0.51). The risk of FTD inversely correlated with the duration of exposure (adj. OR 0.88, 95% CI 0.81–0.95, for every 5 years of exposure increase). Retrospective information and the unknown amount of consumed alcohol are limits of the present work.
INTRODUCTION
The influence of environmental factors in frontotemporal dementia (FTD) has not been extensively studied [1]. Even genetically-determined cases have striking variations in phenotypic expression, implying environmental and epigenetic effects. In particular, the impact of lifestyle is still largely unclear, with very few exceptions [2]. In particular, diet and recreational habits are of special relevance, since they imply public health policies potentially able to reduce the disease burden. Exploring cognate research fields, coffee consumption was reported as a protective factor for amyotrophic lateral sclerosis (ALS) [3], a disease that shares with FTD a TDP-43 neuropathology and mixed clinical expressions [4]. Furthermore, coffee consumption was previously associated with reduced risk of Alzheimer’s dementia, albeit the issue is still controversial [5]. Besides caffeine, antioxidants contained in coffee drinks have been claimed to exert possible protective effects [6]. Analogously, mild-to moderate alcohol consumption may exert protective effects, possibly due to the presence of polyphenols [7]. Smoking, on the other hand, has been reported as a risk factor for dementia [8].
To fill this gap on the role of environmental risk factors in FTD, the aim of this study was to explore coffee, alcohol, and tobacco consumption by using a retrospective case-control design. Traumatic brain injury (TBI) was recorded as well, as a known independent risk factor for FTD, in order to control for possible confounding factors [9].
MATERIALS AND METHODS
Recruited subjects
Following Ethics Committee approval and informed consent, 151 consecutive FTD outpatients (85/66 M/F, age 69.6±8.5 y, age of onset 64.7±9.1 y, disease duration 5.3±3.2, FTLD-modified CDR sum of boxes 8.3±3.0, mean±SD) and 151 age- (±2 y; age 69.8±8.4 y), center-, and sex-matched healthy controls were recruited from four centers in the Lombardy Region, Italy. Patients were recruited according to current clinical criteria [10, 11]: 110 behavioral variant FTD (bvFTD), 24 agrammatic variant FTD (avFTD), 11 semantic variant FTD (svFTD), 6 FTD with motor neuron disease (FTD/MND). A causative mutation was known in 20 patients (13%, including n = 11 GRN del. ex 8, n = 5 C9ORF72 expansions, n = 4 TARDBP mutations, namely, M359V, IVS-IV +45C/T, Y214Y, c. -122G/A). Patients were taking neuroleptics (n = 54, 36%), antidepressants (n = 30, 20%), and mood stabilizers (n = 18, 12%); n = 66 (44%) were not under psychoactive medications. Controls were recruited from other neurological patients’ spouses and siblings; subjects on neuroleptics or antidepressant were excluded. Cognitive impairment was excluded by clinical interview, Mini-Mental State Examination (MMSE) score >26, and a corrected frontal assessment battery (FAB) score >13.4. An ad hoc structured form was used to collect data regarding coffee, tobacco, and alcohol intake, assessed as daily consumption for at least 6 months during the participant’s entire lifetime (ever versus never), duration of exposure (in years), and history of consumption (never, former, or current) [3]. Information was primarily collected from the recruited subjects and then corroborated by the main caregiver in 92% of the cases. The few patients without surrogate informer were all affected by mild forms of the disease. Age (in years), education (in years), and history of major TBI (defined as any head trauma leading to hospitalization with or without loss of consciousness) were collected as well.
Power calculation
Based on a previous study [3], coffee is expected to reduce the risk of ALS, with an OR of 0.4 for lifetime exposure (ever versus never). Assuming that 90% of controls is exposed to coffee, a total of 151 patients and 151 controls is needed to have 80% power to detect an OR for FTD of 0.4. Assuming a proportion of exposed individuals among controls of 80%, the number of patients needed is 95 per group.
Statistical analysis
Cases and controls were described using counts, percentages, medians, and interquartile range (IQR), and compared using the Chi-square test or the Wilcoxon-Mann-Whitney test, as appropriate. Disease duration and age of onset were compared between patients with and without a lifetime exposure to coffee, alcohol, and smoking, using the Wilcoxon-Mann-Whitney test. The association between each substance intake and FTD was assessed using univariable and multivariable conditional logistic regression models. Different models were used for: lifetime exposure (model I); history of exposure (model II); duration of exposure (model III). In each model, data were adjusted for TBI, age, and education. The effect of TBI on the occurrence of FTD was also assessed, in both univariable and multivariable analyses, adjusting for age, education, and lifetime exposure to coffee, alcohol, and smoking (model IV). Results are reported as odds ratios (OR) and adjusted OR (adj. OR) with 95% confidence intervals. Interactions of each substance intake and TBI with gender and clinical center were also tested. In order to understand if part of the observed associations could be due to reverse causality, we performed a sensitivity analysis, censoring all exposures at the estimated year of onset and, for controls, the estimated year of onset of the matched case.
The same multivariable models were assessed, separately, in bvFTD patients and in those without known mutations. All tests were two-tailed and the significance level was set to 5%. Missing data were handled using the listwise deletion method. All analyses were carried out using the SAS statistical analysis system version 9.2 (SAS Institute Inc., Cary, NC, USA).
RESULTS
Table 1 shows the recruited sample. The median number of years of education were 5 (IQR 5–8) for cases and 8 (IQR 5–13) for controls. TBI was more frequent in cases than in controls (17.2% versus 5.8%).
Characteristics of the sample of FTD patients and age- and sex-matched controls
IQR, interquartile range.
The proportion of coffee drinkers was similar for cases and controls (82.8% versus 82.1%); 67.1% of FTD patients and 73.2% of controls were current coffee drinkers and the median exposure was 46 years for cases (IQR 39–54) and 47 years for controls (IQR 32–53). Alcohol consumption was less frequent in FTD patients (52.3% versus 65.6%); the proportion of current alcohol consumers was 30.0% in FTD patients and 56.7% in controls and median alcohol intake was 24 years for the FTD group and 42 years for the controls. Smokers were evenly distributed between cases and controls (43.1% versus 45.7%), with 11.3% and 10.6% of current smokers, respectively; the duration of smoking was similar (median 0, IQR 0–36 for cases and 0–26 for controls).
Disease duration, age of onset, and drug exposure did not differ between patients with and without a lifetime exposure to coffee, alcohol, and smoking (data not shown).
Table 2 shows the results of univariable and multivariable analyses. The risk of FTD was decreased by alcohol consumption (adj. OR 0.30, 95% CI 0.14–0.63). The reduction was significant in current alcohol consumers (adj. OR 0.22, 95% CI 0.10–0.51), but not in former consumers (adj. OR 0.64, 95% CI 0.25–1.68). The risk of FTD also inversely correlated with the duration of exposure (adj. OR 0.88, 95% CI 0.81–0.95, for every 5 years of exposure increase). No significant associations were found for tobacco and coffee. Subjects who had a TBI in their history showed a more than 3-fold increase in the odds of FTD (adj. OR 3.16, 95% CI 1.20–8.33). No significant interactions with gender or clinical center were found. When censoring exposures at the year of onset, all the estimated risks remained virtually unchanged: no significant associations were observed for coffee and tobacco (data not shown), while a significant risk reduction was found for current alcohol consumers (adj. OR 0.27, 95% CI 0.12–0.60) and for increasing duration of exposure (adj. OR 0.88, 95% CI 0.81–0.96, for every 5 years of exposure).
Odds ratios and adjusted odds ratios for FTD by coffee/alcohol/tobacco consumption and TBI categories
OR, odds ratio; adj. OR, adjusted odds ratio (model I, II, and III were adjusted for traumatic brain injury, age, and education; model IV was adjusted for age, education, and lifetime exposure to coffee, alcohol and smoking); CI, confidence interval.
Finally, the results did not change substantially when analyzing separately bvFTD patients (n = 110) and those without known mutations (n = 131) (Supplementary Table 1).
DISCUSSION
The aim of this work was to explore recreational habits in FTD, namely coffee, alcohol, and tobacco intake, as possible risk modifiers. Coffee and tobacco consumption was not different between cases and controls. On the other hand, alcohol consumers were more frequent in the group of controls versus FTD patients, with an estimated risk reduction of 70%. A similar effect was shown when considering current alcohol consumers, although it is reasonable that caregivers may forbid alcohol consumption to FTD patients, for its possible interaction with psychoactive drugs. Nevertheless, in support of a potential protective role, the duration of alcohol consumption was significantly associated to lower risk of FTD. The population of the present study was heterogeneous and included different clinical syndromes and, plausibly, different genetic profiles. We therefore separately analyzed bvFTD since they were the most common clinical phenotype (72.8%) and potentially presented a different history of substance use as compared to primary progressive aphasias [12]. The results did not change. Analogously, the exclusion of the minority of patients for whom a causative mutation was known, was uneventful. However, the possibility that a specific mutation could interact with these environmental exposures cannot be excluded, as patients with inherited FTD have considerable phenotypic variability [13].
Alcohol misuse is known to be associated to cognitive impairment; in spite of this, mild-to-moderate consumption exert protective effects on cardiovascular risk factors, promoting anti-inflammatory/anti-oxidative processes that eventually might delay cognitive impairment [7].
The lack of data regarding the total amount of alcohol and the type of beverages represent one limit of the present study. Other limits are the retrospective design and the lack of standardized scales. In addition, reverse causality cannot be excluded; although we repeated the analyses excluding exposures documented after the year of onset of symptoms, the biological onset of FTD might have occurred several years before the first clinical manifestations. Only prospective designs might avoid this bias and those generated by surrogate informers, but this is unfeasible because large numbers of individuals would be required to be followed for prolonged time. Overall, all these issues confer a limited clinical transferability to our current results. Finally, the possibility also exists that the association is due to other (unknown) factors that act as confounders. In fact, alcohol intake failed to be associated to FTD in a previous study from the Netherlands, performed with a different design on 80 patients and 124 controls [9]. In that study, alcohol consumption was measured in drinks per day and considered mild to moderate if individuals had taken between one and three drinks a day, and severe if they had taken more than three drinks a day, for at least 10 years [9]. Further studies are thus needed to draw firm conclusions on this issue.
DISCLOSURE STATEMENT
Authors’ disclosures available online (http://j-alz.com/manuscript-disclosures/17-0260r1).
