Bell’s palsy and obesity, alcohol consumption and smoking: A nested case-control study using a national health screening cohort

The aim of this study was to investigate the association of body mass index (BMI), alcohol consumption, and smoking status with the occurrence of Bell’s palsy. The Korean National Health Insurance Service-Health Screening Cohort of a ≥ 40-year-old population from 2000–2003 was used. A total of 5,632 Bell’s palsy participants were matched with 22,528 control participants in terms of age, sex, income, region of residence, and past medical histories of hypertension, diabetes, and dyslipidemia. Bell’s palsy was classified by a history of ≥2 diagnoses with ICD-10 code (G510) and steroid treatment. BMI (kg/m2) was classified as <18.5 (underweight), ≥18.5 to <23 (normal), ≥23 to <25 (overweight), ≥25 to <30 (obese I), and ≥30 (obese II). Alcohol consumption was divided into non-drinkers and those who drank 2–3 times a month, 1–2 times a week, and ≥3 times a week. Smoking status was categorized as current smokers, past smokers, and non-smokers. The odds of obesity, alcohol consumption, and smoking with Bell’s palsy were analyzed using logistic regression analysis. BMI showed proportionally positive associations with Bell’s palsy (adjusted OR [95% CI] = 0.61 [0.47–0.79] for underweight, 1.16 [1.08–1.26] for normal, 1.24 [1.15–1.33] for obese I, and 1.61 [1.38–1.88] for obese II, P < 0.001). The odds of alcohol consumption with Bell’s palsy were 0.90 (95% confidence interval [CI] = 0.82–0.99) for 2–3 times a month, 0.77 (95% CI = 0.69–0.85) for 1–2 times a week, and 0.79 (95% CI = 0.71–0.88) for ≥3 times a week compared to nondrinkers (P < 0.001). Smoking did not show a relationship with the occurrence of Bell’s palsy. Obesity was related to the risk of Bell’s palsy in the population over 40 years old. On the other hand, alcohol consumption was negatively associated with the occurrence of Bell’s palsy.


Results
The BMI distributions were different between the Bell's palsy group and the control group (P < 0.001, Table 1), as 9.8% (550/4,466) of Bell's palsy and 11.0% (2,478/17,202) of control participants drink alcohol ≥3 times a week (P < 0.001). The rates of smoking were comparable between groups (P = 0.862). Age, sex, income, region of residence, and past medical histories of hypertension, diabetes, and dyslipidemia were matched between the Bell's palsy and control groups (all P = 1.00).

Discussion
Higher BMI was related to higher odds for Bell's palsy in the present study. Alcohol consumption ≥1 time per week was associated with lower odds for Bell's palsy. These results were consistent in subgroups according to age and sex. To the best of our knowledge, this study is the first to investigate the relationship between alcohol consumption and Bell's palsy.
The adverse cardiovascular and metabolic effects of obesity might increase the risk of Bell's palsy. Obesity is associated with numerous cardiovascular and metabolic diseases, including hypertension, diabetes, and metabolic syndrome. Because these cardiovascular and metabolic disorders were suggested as risk factors for Bell's palsy, obesity could indirectly influence the risk of Bell's palsy due to these cardiovascular and metabolic comorbidities. Several previous studies have reported that the risk of Bell's palsy is associated with hypertension, diabetes, and metabolic syndromes 4,7,10 . Hypertension was related to the onset of Bell's palsy in ≥40-year-old patients 4 . Although the pathophysiologic mechanisms underlying the relationship between hypertension and Bell's palsy remain elusive and controversial, fluid retention or edema of the vasa nervorum and increased susceptibility to viral infection due to higher cortisol levels could explain the elevated risk of Bell's palsy in hypertension patients, as in pregnancy 11 . Diabetes has also been proposed as a risk factor for Bell's palsy 7,10 . A retrospective study reported that as many as 31.5% and 29% of Bell's palsy patients had diabetes mellitus and hypertension, respectively 10 . A prospective study demonstrated that abnormally elevated serum glycosylated hemoglobin (HbA1c) levels were related to the severity grade of Bell's palsy (OR = 4.7, 95% CI = 1.4-15.2, P = 0.008) 7 . The elevated levels of inflammatory cells and inflammatory cytokines in diabetic patients could induce diabetic peripheral facial neuropathy 10 . No study has shown a relationship between BMI and the onset of Bell's palsy. However, retrospective studies demonstrated that subjects with obesity demonstrated a lower recovery rate of Bell's palsy than normal subjects 9,12 . In addition, patients with metabolic syndrome showed a lower recovery rate from Bell's palsy than control subjects 13 .
Contrary to our expectation, alcohol consumption was negatively associated with the risk of Bell's palsy in this study. It is possible that the biased control group may have affected the relationship between alcohol and Bell's palsy. To minimize the potential selection bias of the control group, selection of the control group was repeated with a novel random number in this study. Results consistent with the low odds of Bell's palsy with alcohol consumption were still observed when the new control group was used. The neuroprotective and cardioprotective www.nature.com/scientificreports www.nature.com/scientificreports/ effects of mild to moderate alcohol consumption may be linked to a reduced risk of Bell's palsy 14 . A meta-analysis showed that multiple cardiovascular outcomes of coronary heart disease and stroke were decreased in moderate alcohol consumption groups 15 . Moderate alcohol consumption has been suggested to induce anti-inflammatory responses related to adenosine receptors, protein kinase C, nitric oxide synthase, and heat shock proteins, thereby resulting in cardioprotective effects 14,16 . In addition, alcohol exerted direct neuroprotective effects against www.nature.com/scientificreports www.nature.com/scientificreports/ ischemic, endotoxin, beta-amyloid, and neuroinflammatory proteins, probably by activating reactive oxygen species, a number of protein kinases, and heat shock proteins 14,17 . There has been some evidence supporting the neuroprotective roles of alcohol consumption in epidemiologic studies. For example, cognitive function showed a proportionally positive association with the amount of alcohol consumption in a cross-sectional study 18 . In summary, these cardiovascular and neural effects of alcohol could have led to the protective effects against Bell's palsy observed in this study. Because only 21% of participants consumed alcohol ≥1 time per week in our cohort, the relatively moderate frequency of alcohol consumption of ≥1 time per week was used as the reference criterion for alcohol consumption. Thus, the negative effects of higher-frequency alcohol consumption could be masked due to the low threshold for alcohol consumption used in this study.
In the subgroup analysis according to age and sex, obesity consistently increased the odds of Bell's palsy. On the other hand, alcohol consumption decreased the odds of Bell's palsy in men. However, no relationship between alcohol consumption and Bell's palsy was found in women. The small number of participants with alcohol consumption ≥1 time per week may have weakened the statistical power in the women subgroup (Supplementary Table S1). Women are presumed to be more susceptible to the effects of alcohol consumption, such as alcohol-related liver diseases, cardiomyopathy, and breast cancer 19 . Thus, further studies that include a sufficient number of participants who consume alcohol are warranted.
A large, nationwide, population-based cohort was used in the present study. The large number of participants enabled the selection of a sufficient number of control participants matched for age, sex, income, region of residence, and past medical histories. Because this study was based on health claim codes, matching of socioeconomic factors was crucial because these factors determine medical accessibility. In addition, the fidelity of the sample cohort was verified in a previous study 20 . No missing participants are expected because the national health insurance system of Korea legally registers and manages all medical records of Korean citizens. Moreover, the relationship between lifestyle factors, including alcohol consumption and smoking, and the onset of Bell's palsy were evaluated using the most recent health check-up data before the index date (onset of Bell's palsy) in this study. The participants who did not undergo a health check-up before the index date were excluded.
Nevertheless, a few limitations should be considered when interpreting the present results. On the one hand, because the diagnosis of Bell's palsy was based on the ICD-10 codes and treatment histories, subclinical or spontaneous recovery cases were excluded from this study. On the other hand, acute facial palsy other than Bell's palsy could be included in this study due to incorrect ICD-10 coding. In addition, the severity and treatment responses of Bell's palsy could not be assessed. Regarding lifestyle factors, recall bias is possible for alcohol consumption and smoking because they were surveyed based on a questionnaire. In addition, the total amount of alcohol consumption and smoking could not be calculated because a categorical questionnaire was used. Finally, although this study considered the lifestyle factors of obesity, smoking, and alcohol consumption, possible confounders such as caffeine consumption still existed. Because caffeine consumption might be associated with alcohol reward or alcohol-seeking behavior, alcohol consumption could be correlated with caffeine consumption 21 . Thus, the possible protective effect of caffeine consumption on the risk of Bell's palsy could mediate the negative association between Bell's palsy and alcohol consumption in this study. www.nature.com/scientificreports www.nature.com/scientificreports/ In conclusion, obesity was associated with higher odds of Bell's palsy in the ≥40-year-old population. Alcohol consumption ≥1 time per week was related to lower odds of Bell's palsy than alcohol consumption <1 time per week.

Materials and Methods
Study population and data collection. The Ethics Committee of Hallym University (2017-I102) approved the use of these data. The study was exempted from the need for written informed consent by the Institutional Review Board. All analyses adhered to the guidelines and regulations of the Ethics Committee of Hallym University.  20,22 . This was a nested case-control study. The participants with Bell's palsy were included in this study (n = 6,071). The control subjects who were never diagnosed with Bell's palsy were randomly selected from the entire population (n = 508,815) and 1:4 matched with the Bell's palsy participants. The matching variables were age group, sex, income group, region of residence, hypertension, diabetes, and dyslipidemia. The index date was defined as the date of diagnosis with Bell's palsy. The index date of control participants was defined as the identical date of their matched Bell's palsy participants. Bell's palsy participants without health check-up data before index data were excluded (n = 439). Resultantly, 5,632 Bell's palsy participants and 22,528 control participants were enrolled. The latest health check-up data before the index date were analyzed for the Bell's palsy and control groups (Fig. 1).
Variables. Independent variable. Tobacco smoking was defined by current smoking state, duration of smoking, and amount of cigarette smoking (Supplementary Table S2). Smoking status was categorized into 3 groups: current smoker, past smoker, and nonsmoker.
Covariate analysis. Age groups were classified using 5-year age intervals. The income groups were divided into 5 classes (class 1 [lowest income]-5 [highest income]). The region of residence was grouped into urban and rural areas.
Dependent variable. Bell's palsy was defined using ICD-10 code (G510). We only included participants who were treated for Bell's palsy ≥1 time with steroid treatment.

Statistical analyses.
Chi-square tests were used to compare the general characteristics between Bell's palsy and control groups.
To analyze the odds ratio (OR) of smoking, drinking alcohol, and obesity with Bell's palsy, conditional logistic regression analysis was used. Models were stratified by age, sex, income, region of residence, hypertension, diabetes, and dyslipidemia. In this analysis, a crude (simple), adjusted model (adjusted model for obesity, smoking state, and frequency of drinking alcohol) was used. The 95% confidence interval (CI) was calculated. www.nature.com/scientificreports www.nature.com/scientificreports/ For the subgroup analyses, we divided the participants according to age and sex (<60 years old and ≥60 years old, men and women). The age threshold for age groups was determined as the median value of the total number of participants.
Two-tailed analyses were conducted, and P values less than 0.05 were considered to indicate significance. The results were statistically analyzed using SPSS v. 22.0 (IBM, Armonk, NY, USA).

Figure 1.
A schematic illustration of the participant selection process used in the present study. Of a total of 514,866 possible participants, 5,632 Bell's palsy patients were matched with 22,528 control participants for age group, sex, income group, region of residence, and past medical histories.