Relationship between alcohol consumption and Rheumatoid factor(RF) with alcohol-induced facial flushing response

Journal Pre-proof Relationship between alcohol consumption and Rheumatoid factor(RF) with alcoholinduced facial flushing response Jihan KIM, Chan Keol PARK, Jong Sung KIM, Sami LEE PII:

S0741-8329(18)30278-7

DOI:

https://doi.org/10.1016/j.alcohol.2019.10.004

Reference:

ALC 6952

To appear in:

Alcohol

Received Date: 28 September 2018 Revised Date:

4 October 2019

Accepted Date: 7 October 2019

Please cite this article as: KIM J., Keol PARK C., Sung KIM J. & LEE S., Relationship between alcohol consumption and Rheumatoid factor(RF) with alcohol-induced facial flushing response, Alcohol (2019), doi: https://doi.org/10.1016/j.alcohol.2019.10.004. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Inc.

Relationship between alcohol consumption and Rheumatoid factor(RF) with alcohol-induced facial flushing response

Jihan KIM1, Chan Keol PARK2, Jong Sung KIM1*, Sami LEE1

1

Department of Family Medicine, Research Institute for Medical Science

Chungnam National University School of Medicine Daejeon, Korea 2Division

of Rheumatology, Department of Internal Medicine

Chungnam National University School of Medicine Daejeon, Korea

*Corresponding author : Jong Sung KIM, MD, PhD Postal address : Department of Family medicine, Chungnam National University Hospital, 282 Munhwa-Ro, Jung-Gu, Daejeon 35015, South Korea Cellular phone : +82-10-5407-1249 Fax : +82-42-280-7879 E-mail address: [email protected]

1 2

Abstract This study investigated the relationship between alcohol consumption with alcohol induced

3

facial flushing response and rheumatoid factor (RF) in adult men. The cohort comprised 1675 men

4

who underwent a general medical check-up between July 2016 and June 2017, including 355 non-

5

drinkers, 498 flushers, and 822 non-flushers. One drink was defined as 14 grams of alcohol. RF

6

was considered negative if ≤18 IU/mL and positive if 18< IU/mL. Logistic regression analyses were

7

used. Compared to non-drinkers, the odds ratio for a positive RF among non-flushers was 0.92

8

(95% confidence interval [CI], 0.37 2.29) for those with an average alcohol consumption of ≤4

9

drinks per week, 1.64 (95% CI, 0.67 3.98) for those consuming 4<,≤8 drinks per week, and 1.17

10

(95% CI, 0.55 2.50) for those consuming 8< drinks per week; the differences were not statistically

11

significant. Compared to non-drinkers, flushers also had a non-significant odds ratio for positive

12

RF of 1.26 (95% CI, 0.54 2.90) among those with an average alcohol consumption of ≤4 drinks

13

per week. However, flushers showed significantly higher odds ratio for positive RF of 3.12 (95% CI,

14

1.18 8.24) among those consuming 4<,≤8 drinks per week and 3.27 (95% CI, 1.42 7.52) among

15

those consuming 8< drinks per week. Additionally, flushers consuming 8< drinks per week were

16

associated with significantly higher rates of positive RF than non-flushers (odds ratio, 2.38; 95% CI,

17

1.05 5.17). Our study revealed that flushers consuming 4< drinks per week showed a higher

18

probability of positive RF than non-drinkers. Furthermore, flushers consuming 8< drinks per week 1

19

had a higher probability of positive RF than non-flushers. Our results strongly indicate that the

20

average weekly alcohol consumption level and the presence or absence of flushing should be

21

considered when interpreting the results of RF examinations in healthy men.

22 23

Keywords: Alcohol drinking, Flushing, Rheumatoid factor, Acetaldehyde

24

2

25 26

Introduction Rheumatoid factor (RF) is an autoantibody against the Fc region of immunoglobulin G (IgG),

27

and its presence is an important diagnostic criteria for rheumatoid arthritis (RA). RF testing in

28

patients with RA has a sensitivity of 60 90% and a specificity of 85%. High levels of RF can also

29

be found in patients with systemic lupus erythematosus and Sjögren's syndrome. In most cases,

30

RF is associated with the presence of a rheumatoid disease (Nishimura et al., 2007; Nell et al.,

31

2005). However, positive RF can sometimes be also observed in individuals without rheumatoid

32

disease, such as those with chronic viral hepatitis, liver cirrhosis, or malignant tumors (Ingegnoli,

33

Castelli, & Gualtierotti, 2013). Among lifestyle habits, smoking is known to increase the odds of RF

34

(Krol et al., 2015), and has also been shown to be a risk factor for RA. (Sugiyama et al., 2010).

35

Many previous studies have also investigated the relationship between alcohol consumption and

36

autoimmune reactions. While some data suggest an increase in autoimmune reactions with

37

alcohol consumption (Thiele et al., 2010), other data indicate a decrease in the risk of RA with

38

alcohol consumption (Jin, Xiang, Cai, Wei, & He, 2014). Therefore, the relationship between

39

alcohol consumption and RF remains unclear.

40

The degree of alcohol metabolism in the body varies between individuals because of the

41

different levels of aldehyde dehydrogenase (ALDH) activity. This enzyme is encoded by ALDH1 and

42

ALDH2; a mutation in the ALDH2 gene in the form of a lysine-to-glutamine substitution results in 3

43

an inactive enzyme product. This in turn causes individuals with this mutation to experience facial

44

redness or flushing, even when consuming only 1 2 drinks, owing to acetaldehyde buildup

45

(Agarwal, Harada, & Goedde, 1981). Previous studies have shown that these flushers have an

46

increased risk of metabolic syndrome or hypertension with low levels of alcohol consumption

47

compared to non-flushers (Jung, Kim, Kim, Oh, & Yoon, 2014; Kim et al., 2012). Studies have also

48

shown that the risks of cancers, such as esophageal and bladder malignancies, are higher in

49

flushers than in non-flushers (Masaoka et al., 2017; Zhang et al., 2017). Given this background, it is

50

prudent to consider flushing due to alcohol consumption when also investigating autoimmune

51

diseases or their related factors. Therefore, we performed this study to investigate the relationship

52

between the presence of flushing due to alcohol consumption and RF elevation.

53 54

Method

55

Research Subjects

56

The study cohort comprised 1881 adult male subjects who underwent a general medical check-

57

up at a university hospital health examination center in Daejeon between July 2016 and June 2017.

58

Men with factors that could influence RF such as a history of rheumatoid disease, chronic viral

59

hepatitis, history of cancer, chronic kidney disease, chronic heart failure, liver cirrhosis and HIV

60

infection were excluded. Moreover, those who failed to respond I m not sure to the survey 4

61

question asking about the presence of facial flushing response were also excluded. Ultimately,

62

1675 subjects comprising 355 non-drinkers, 498 flushers and 822 non-flushers were included in

63

this study (Fig. 1). This is a retrospective cross-sectional study. It was performed in compliance

64

with the Declaration of Helsinki and was approved waiver of informed consent by the Ethics

65

Committee of Chungnam National University Hospital (Institutional Review Board Number: 2018-

66

08-050).

67 68

Research Methods

69

Lifestyle habits and basic data pertaining to the study subjects, such as past/current medical

70

history and drug/alcohol use, were obtained through medical records prepared during general

71

checkups. Height and weight were measured, and the body mass index (BMI) was calculated by

72

dividing the weight (kg) by the square of the height ( ).

73

The evaluation of facial flushing response during alcohol consumption was conducted via a

74

questionnaire administered during the physical examination. Respondents were asked to choose 1

75

of the 4 answers ( always true, sometimes true, not true, or I m not sure ) to the question In

76

the past or present, have you experienced facial flushing response after one drink? This was

77

based on previous research showing that investigation using a survey is as effective as using an

78

ethanol patch to evaluate inactive ALDH2 (Yokoyama et al., 1997). Subjects who responded 5

79

always true or sometimes true were categorized as flushers, while those who responded not

80

true were categorized as non-flushers. These categories were based on Yokoyama et al. s data

81

showing that the sensitivity and specificity of ALDH2 mutation identification using the

82

questionnaire were 96.1% and 79.0%, respectively, when those who responded sometimes true

83

were considered flushers.

84

The frequency of alcohol consumption per week, average number of drinks per occasion, and

85

highest number of drinks per occasion were evaluated to measure alcohol consumption levels.

86

Based on the metrics provided by the National Institute on Alcohol Abuse and Alcoholism, 14 g

87

of alcohol was considered 1 drink (Willenbring, Massey, & Gardner, 2009). One drink is equivalent

88

to 90 mL of 20% soju (1/4 bottle of soju), 12 oz of beer (1 can of beer, 355 mL), 45 mL of liquor

89

(1 shot), 150 mL of wine (1 wine glass), or 300 mL of rice wine (1 bowl). The calculated number of

90

drinks was then multiplied by the frequency of drinking per week to evaluate the weekly average

91

alcohol consumption level.

92

The average weekly alcohol consumption for flushers and non-flushers was categorized into

93

three groups as ≤4, 4<,≤8 and 8< drinks based on the moderate alcohol consumption guidelines

94

for Koreans (Lee S et al., 2019).

95 96

Non-smokers were defined as those who had never smoked as of the date of their medical checkup. Subjects who had smoked in the past but not within 1 month of the physical 6

97

examination were categorized as past smokers, while those who had smoked within 1 month of

98

the examination were categorized as current smokers.

99

Blood tests were conducted during the health examination after subjects had fasted for 12

100

hours to evaluate RF as well as erythrocyte sedimentation rate (ESR), C-reactive protein (CRP), and

101

uric acid(UA).

102

An immunoturbidimetric assay was used for RF testing. Following the guidelines of the

103

American College of Rheumatology and European League Against Rheumatism (Aletaha et al.,

104

2010), up to 3-fold the normal upper limit was defined as low-positive, whereas levels exceeding

105

this limit were considered high-positive. Based on the inspection protocol and instrument

106

documentation, TBA-2000FR (TOSHIBA medical systems corporation, Tochigi, Japan), we defined a

107

negative RF as ≤18 IU/mL, while a positive RF was defined as 18< IU/mL. RF values of 18<,≤54

108

IU/mL were considered low-positive, while those 54< IU/mL were considered high-positive.

109 110 111

Statistical Analyses General characteristics, body measurements, and blood test results were compared among non-

112

drinkers, flushers, and non-flushers. One-way distributed analysis of variance was used for

113

continuous variables, while the chi-square test was used for categorical variables (such as smoking

114

status). 7

115

The proportions of flushers and non-flushers who had a high level of RF based on alcohol

116

consumption level were compared to non-drinkers using the chi-square test and binary logistic

117

regression analyses. Multiple binary logistic regression analysis models were used: Model 1 used

118

no corrections; Model 2 corrected for age and BMI; and Model 3 corrected for Model 2 variables

119

in addition to ESR, CRP, uric acid and smoking status. Statistical significance level was set to p <

120

0.05, and all statistical processing was performed using IBM SPSS version 21.0 (IBM Corp, Armonk.

121

NY, USA). We performed a post-hoc analysis using G*Power 3.1.9.4 (Franz Faul, Universität Kiel,

122

Germany) to calculate the statistical power.

123 124

Results

125

Subject characteristics

126

The average differences in the ages of the non-drinkers (56.7 years), flushers (52.2 years), and

127

non-flushers (50.5 years) were significantly different (p < 0.001). The average alcohol consumption

128

level per week was significantly higher among non-flushers (14.3 drinks) than flushers (7.6 drinks)

129

(p < 0.001). And the average of GGT in non-flushers (61.0 U/L) as weekly alcohol consumption

130

level was significantly higher than non-drinkers (32.7 U/L; p < 0.001), also higher than flushers

131

(46.9 U/L; p = 0.001). However, flushers (9.7 IU/mL) had a higher RF average than non-drinkers

132

(7.0 IU/mL) and non-flushers (7.5 IU/mL), although RF averages were not significantly different (p 8

133

= 0.139). Moreover, the proportions of RF-positive flushers (6.0%) were higher than non-drinkers

134

(3.7%) and non-flushers (4.4%), although the proportions were not significantly different (p =

135

0.204). The percentage of non-smokers was higher among non-drinkers than among flushers and

136

non-flushers, while the proportions of past and current smokers were higher among flushers and

137

non-flushers than among non-drinkers (p < 0.001). There were no significant differences in CRP

138

and ESR (Table 1).

139 140

Differences in the proportions of RF positivity based on the average alcohol consumption level

141

per week

142

We investigated whether the proportion of subjects positive for RF was affected by whether the

143

subjects are drinkers (flushers and non-flushers) or non-drinkers. The proportion of RF positivity

144

among non-drinkers was 3.7%. Among drinkers, RF was positive in 3.6% of those who consumed

145

≤4 drinks per week, 6.8% of those who consumed 4<,≤8drinks per week, and 5.3% of those who

146

consumed 8< drinks per week. No statistically significant differences were found between drinkers

147

and non-drinkers (p = 0.169, p for trend = 0.109) (Fig.A.1).

148

However, categorizing drinkers into flushers and non-flushers produced more notable results.

149

First, 3.4% of non-flushers who consumed ≤4 drinks, 6.1% of those who consumed 4<,≤8 drinks,

150

and 4.3% of those who consumed 8< drinks per week were RF-positive; these differences were 9

151

not statistically significant (p = 0.653, p for trend = 0.541). In contrast, 4% of flushers who

152

consumed ≤4 drinks, 9.8% of those who consumed 4<,≤8 drinks, and 9.6% of those who

153

consumed 8< drinks per week were RF-positive. The percentage of flushers with an alcohol

154

consumption level of 4< drinks who were RF-positive was significantly higher that of RF-positive

155

non-drinkers (4<,≤8 drinks per week: p = 0.043; 8<: p = 0.021, p for trend = 0.008). Importantly, a

156

higher proportion of flushers with an alcohol consumption level of 8< drinks per week than non-

157

flusher was RF-positive (p = 0.039) (Fig. 2).

158 159 160

Analysis of flushers and non-flushers using multiple binary logistic regression The odds ratios for positive RF was calculated for flushers and non-flushers using non-drinkers

161

as a baseline using logistic regression analysis. When using Model 1 (no corrections), the odds

162

ratio of flushers with a weekly average of 8< drinks was high at 2.53 (95% confidence interval [CI],

163

1.12 5.68). When using Model 2, the odds ratio of flushers who consumed 4<,≤8 drinks weekly

164

was higher than that of non-drinkers at 2.92 (95% CI, 1.11 7.65). Furthermore, the odds ratio of

165

flushers with a weekly average of 8< drinks was higher than when Model 1 was used, at 3.08 (95%

166

CI, 1.34 7.05). The odds ratios were even higher when employing Model 3, where the odds ratio

167

in flushers who consumed 4<,≤8 drinks weekly was 3.12 (95% CI, 1.18 8.24) while that of flushers

168

who consumed 8< drinks weekly was 3.27 (95% CI, 1.42 7.52). The power (1- err prob) was 0.89 10

169

and 0.95, respectively. The odds ratio in flushers with an alcohol consumption level of 8< drinks

170

weekly was observed to be higher than among flushers who consumed 4<,≤8 drinks weekly.

171

In contrast, the odds ratio of flushers with an alcohol consumption level of ≤4 drinks weekly

172

was higher than in non-drinkers, at 1.26 (95% CI, 0.54 2.90) in model 3, although not significantly.

173

In case of non-flushers, the odds ratio with an alcohol consumption level of ≤4 drinks weekly was

174

comparatively lower, but also non-significantly so, at 0.92 (95% CI, 0.37 2.29) in model 3. The

175

odds ratio of non-flushers who consumed 4<,≤8 drinks weekly was higher than of those who

176

consumed ≤4 drinks weekly, at 1.64 (95% CI, 0.67 3.98) in model 3, albeit not significantly. The

177

odds ratio of non-flushers who consumed 8< drinks weekly was also high at 1.17 (95% CI, 0.55

178

2.50) in model 3, again, not significantly (Table 2).

179

The odds ratio of RF positivity was compared between flushers and non-flushers after correcting

180

for age, BMI, ESR, CRP, uric acid and smoking status. The odds ratio of flushers who consumed ≤4

181

drinks of alcohol weekly compared to non-flushers was higher (1.20 [95% CI, 0.46 3.10]).

182

Furthermore, the odds ratio of flushers who consumed 4<,≤8 drinks of alcohol weekly was 1.39

183

(95% CI, 0.48 4.06), which was higher than those who consumed ≤4 drink weekly, but not

184

significantly. However, the odds ratio of flushers with an alcohol consumption level of 8< drinks

185

weekly was 2.38 (95% CI, 1.05 5.17), which was significantly higher than that of non-flushers and

186

the power (1- err prob) was 0.76 (Table 3). 11

187 188 189

Discussion We explored the relationship between RF positivity and the presence or absence of facial

190

flushing response during alcohol consumption. Our data indicate that the odds ratio of positive

191

RF in flushers who consumed 4<,≤8 drinks weekly was 3.12 compared to non-drinkers. We also

192

found that the odds ratio of RF-positive flushers who consumed 8< drinks weekly was 3.27 as

193

high as in non-drinkers and more than twice as high as in non-flushers.

194

The most significant point of our result was that the positive rate of RF was higher only in the

195

flushers when the average weekly drinking amount was 4< drinks. The relationship between

196

alcohol consumption and rheumatoid arthritis has been confirmed by previous studies, but few

197

studies have studied the relationship between alcohol consumption and RF. Therefore, it is difficult

198

to analyze the alcohol consumption and RF relationship compared to the previous studies.

199

However, the results of this study can be significant in several points.

200

First, it can be considered that certain features which appear distinctively in the flushers, such as

201

increased acetaldehyde in the body, may affect the rate of RF positivity. Flushers are believed to

202

have inactive ALDH2, so acetaldehyde accumulates in the body 6 to 19 times more than the non-

203

flushers (Mizoi Y et al., 1994). This accumulation of acetaldehyde has been shown to promote the

204

ROS reactions (Ohta, Ohsawa, Kamino, Ando, & Shimokata, 2004), and these can lead to various 12

205

metabolic processes in the body. One of the most interesting findings is that acetaldehyde causes

206

structural and functional modification of IgG (Waris S et al., 2018). It is also reported that antibody

207

formation which binds to this modified IgG is also increased (Waris S et al., 2019). Since RF is an

208

autoimmune antibody directed against the Fc portion of IgG, the high RF positive rate only in the

209

flushers may suggest that the increased acetaldehyde in the body causes the Ig modification,

210

leading to the corresponding increase in RF.

211

In contrast, the acetaldehyde accumulation in the non-flushers was relatively less likely than the

212

flushers, resulting in statistically non-significant increase in RF positivity, as shown in our results.

213

As this study, however, did not confirm the modification of IgG through experiments but only

214

assumed its possibility based on the clinical findings, further studies are required to support this

215

causality.

216

Second, this study revealed that increased alcohol consumption in flushers is associated with a

217

higher probability of positive RF. This appears to be consistent with the results of experiments that

218

found a correlation between alcohol consumption levels and IgM RF levels in a mouse model

219

(r=0.65, p=0.006) (Nowak, Gill, Skamene, and Newkirk, 2007). In Nowak et al.s study, which

220

included the duration of alcohol consumption for analysis, mice injected with 7.9 ± 2.5 g/kg/day

221

for 60 days had greater levels of RF than those injected for 5 days. On the other hand, another

222

study found that there is no significant differences between the drinkers and non-drinkers in 13

223

terms of the risk of RF positive RA in accordance to the alcohol consumption amount (Heliövaara

224

et al., 2000). However, a direct comparison is inappropriate because Heliövaara et al.s study

225

analyzed the relationship between alcohol consumption and RF positive RA instead of RF positivity.

226

There are some limitations to this study that may limit is generalizability, including 1) the

227

research cohort did not include women, and 2) the study was performed at a single institution.

228

Despite these limitations, the increase in the odds ratio of positive RF in flushers compared to

229

non-flushers and non-drinkers is a notable finding. This is especially meaningful considering that

230

the current study adjusted for common clinical factors when investigating the relationship of

231

flushing with RF.

232

Taken together, flushers who consume 4< drinks weekly on average appear to experience

233

changes in RF positivity and/or autoimmune response. Therefore, even in individuals without RA,

234

positivity for RF may still be associated with flushing during alcohol consumption as well as heavy

235

drinking. Moreover, other previous studies found an inverse relationship between alcohol

236

consumption and RA (Källber et al., 2009; Scott et al., 2013). If it is found that the likelihood of RA

237

is higher among flushers with positive RF, more research ought to be considered to establish any

238

relationship between alcohol consumption and RA.

239 240

Appendices 14

241 242

Fig.A.1, The difference of percentage for RF (㎡) between drinkers and non-drinkers according to alcohol consumption by chi-square test.

243 244 245

Acknowledgements:

The first two authors, Jihan Kim and Chan Keol Park, contributed equally to this work.

246 247 248 249

Funding sources: This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

250 251 252

Declarations of interest: None.

253 254 255

References

256

Agarwal, D.P., Harada, S., & Goedde, H.W. (1981). Racial differences in biological sensitivity to

257

ethanol: the role of alcohol dehydrogenase and aldehyde dehydrogenase isozymes. Alcoholism

258

Clinical and Experimental Research, 5(1), 12 16. 15

259

Aletaha, D., Neogi, T., Silman, A.J., Funovits, J., Felson, D.T., Bingham, C.O. 3rd, et al. (2010). 2010

260

rheumatoid arthritis classification criteria: an American College of Rheumatology/European

261

League Against Rheumatism collaborative initiative. Arthritis & Rheumatology, 62(9), 2569 2581.

262

Heliövaara, M., Aho, K., Knekt, P., Impivaara, O., Reunanen, A., & Aromaa, A. (2000). Coffee

263

consumption, rheumatoid factor, and the risk of rheumatoid arthritis. Annals of Rheumatic

264

Diseases, 59(8), 631 635.

265 266 267

Ingegnoli, F., Castelli, R., & Gualtierotti, R. (2013). Rheumatoid factors: Clinical applications. Disease

Markers, 35(6), 727 734. Jin, Z., Xiang, C., Cai, Q., Wei, X., & He, J. (2014). Alcohol consumption as a preventive factor for

268

developing rheumatoid arthritis: a dose-response meta-analysis of prospective studies. Annals

269

of Rheumatic Diseases, 73(11), 1962 1967.

270

Jung, J.G., Kim, J.S., Kim, Y.S., Oh, M.K., & Yoon, S.J. (2014). Hypertension associated with alcohol

271

consumption based on the facial flushing reaction to drinking. Alcoholism Clinical and

272

Experimental Research, 38, 1020 1025.

273

Källberg, H., Jacobsen, S., Bengtsson, C., Pedersen, M., Padyukov, L., Garred, P., et al. (2009). Alcohol

274

consumption is associated with decreased risk of rheumatoid arthritis: results from two

275

Scandinavian case-control studies. Annals of Rheumatic Diseases, 68(2), 222 227.

16

276

Kim, M.Y., Kim, S.S., Kim, J.S., Jung, J.G., Kwon, B.R., & Ryou, Y.I. (2012). Relationship between

277

alcohol consumption and metabolic syndrome according to facial flushing in Korean males.

278

Korean Journal of Family Medicine, 33(4), 211 218.

279

Krol, A., Garred, P., Heegaard, N.H., Christensen, A.F., Hetland, M.L., Stengaard-Pedersen K., et al.

280

(2015). Interactions between smoking, increased serum levels of anti-CCP antibodies,

281

rheumatoid factors, and erosive joint disease in patients with early, untreated rheumatoid

282

arthritis. Scandinavian Journal of Rheumatology, 44(1), 8 12.

283 284 285

Lee, S., Kim, J.S., Jung, J.G., Oh, M.K., Chung, T.H., & Kim, J. (2019). Korean alcohol guidelines for moderate drinking based on facial flushing. Korean J Fam Med, 40(4), 204-211. Masaoka, H., Matsuo, K., Sawada, N., Yamaji, T., Goto, A., Shimazu, T., et al. (2017). Alcohol

286

consumption and bladder cancer risk with or without the flushing response: The Japan Public

287

Health Center-based Prospective Study. International Journal of Cancer, 141(12), 2480 2488.

288

Mizoi Y, Yamamoto K, Ueno Y, et al. (1994). Involvement of genetic polymorphism of alcohol and

289

aldehyde dehydrogenases in individual variation of alcohol metabolism. Alcohol Alcohol,

290

29(6):707 710.

291

Nell, V.P., Machold, K.P., Stamm, T.A., Eberl, G., Heinzl, H., Uffmann, M., et al. (2005). Autoantibody

292

profiling as early diagnostic and prognostic tool for rheumatoid arthritis. Annals of Rheumatic

293

Diseases, 64(12), 1731 1736. 17

294

Nishimura, K., Sugiyama, D., Kogata, Y., Tsuji, G., Nakazawa, T., Kawano S, et al. (2007). Meta-

295

analysis: diagnostic accuracy of anti-cyclic citrullinated peptide antibody and rheumatoid factor

296

for rheumatoid arthritis. Annals of Internal Medicine, 146(11), 797 808.

297 298 299 300 301

Nowak, U., Gill, K., Skamene, E., & Newkirk, M.M. (2007). Rheumatoid factor induction in murine models of liver injury. Clinical and Experimental Immunology, 147(2), 324 329. Ohta, S., Ohsawa, I., Kamino, K., Ando, F., & Shimokata, H. (2004). Mitochondrial ALDH2 deficiency as an oxidative stress. Annals of the New York Academy of Sciences, 1011, 36 44. Scott, I.C., Tan, R., Stahl, D., Steer, S., Lewis, C.M., Cope, A.P. (2013). The protective effect of alcohol

302

on developing rheumatoid arthritis: a systematic review and meta-analysis. Rheumatology

303

(Oxford), 52(5), 856 867.

304

Sugiyama, D., Nishimura, K., Tamaki, K., Tsuji, G., Nakazawa, T., Morinobu, A., et al. (2010). Impact

305

of smoking as a risk factor for developing rheumatoid arthritis: a meta-analysis of observational

306

studies. Annals of Rheumatic Diseases, 69(1), 70 81.

307

Thiele, G.M., Duryee, M.J., Willis, M.S., Tuma, D.J., Radio, S.J., Hunter, C.D., et al. (2010). Autoimmune

308

hepatitis induced by syngeneic liver cytosolic proteins biotransformed by alcohol metabolites.

309

Alcoholism Clinical & Experimental Research, 34(12), 2126 2136.

18

310

Waris S, Habib S, Khan S, Kausar T, Naeem SM, Siddiqui SA et al. (2019) Molecular docking

311

explores heightened immunogenicity and structural dynamics of acetaldehyde human

312

immunoglobulin G adduct. IUBMB Life, 71(10):1522-1536.

313

Waris S, Habib S, Tantry IQ, Khan RH, Mahmood R & Ali A. (2018). Acetaldehyde-induced

314

structural and conformational alterations in human immunoglobulin G: A physicochemical and

315

multi-spectroscopic study. Int J Biol Macromol, 113:701-710.

316

Willenbring, M.L., Massey, S.H., Gardner, M.B. (2009). Helping patients who drink too much: an

317

evidence-based guide for primary care clinicians. American Family Physician, 80(1), 44 50.

318

Yokoyama, A., Muramatsu, T., Ohmori, T., Kumagai, Y., Higuchi, S., & Ishii, H. (1997). Reliability of a

319

flushing questionnaire and the ethanol patch test in screening for inactive aldehyde

320

dehydrogenase-2 and alcohol-related cancer risk. Cancer Epidemiology, Biomarkers &

321

Prevention, 6(12), 1105 1107.

322 323

Zhang, J., Zhang, S., Song, Y., Ma, G., Meng, Y., Ye, Z., et al. (2017). Facial flushing after alcohol consumption and the risk of cancer. Medicine (Baltimore), 96(13), e6506.

324 325

Figure captions

326

Fig.1. Study population

19

327

Fig.2. The percentage of RF (㎡) between subjects according to alcohol consumption. F; Flushers,

328

NF; Non-flushers, *; p value <0.05 compared to non-drinkers, †; p value <0.05 compared to

329

non-flushers by chi-square test.

20

Non-drinkers (n=355)

Drinkers

Age (year)

56.7

± 11.5

Flushers (n=489) 52.2 ± 11.7***

Non-flushers (n=822) 50.5 ± 11.5***†

BMI (kg/ ) SBP (mmHg)

25.0

± 3.2

24.8

± 3.0

24.6

± 3.0

127.6

± 14.1

127.8

± 14.1

127.3

± 13.6

DBP (mmHg)

79.2

± 9.9

79.9

± 10.7

79.9

± 10.4

AST (U/L)

26.5

± 12.3

27.3

± 15.5

30.7

± 52.6

ALT (U/L)

30.4

± 18.2

29.3

± 21.5

33.7

± 100.1

GGT (U/L)

32.7

± 23.8

46.9

± 67.2**

61.0

± 81.9***†

103.0

± 20.7

103.6

Glucose (mg/dL)

± 22.4

105.7

± 26.2

CRP (mg/dL)

0.2

± 0.3

0.2

± 0.3

0.2

± 0.3

ESR (mm/h)

9.5

± 8.2

8.8

± 6.8

8.5

± 7.6

UA (mg/dL)

5.6

± 1.4

5.8

± 1.3

5.8

± 1.3*

RF (IU/mL)

7.0

± 12.1

9.7

± 31.9

7.5

± 18.9

RF positive

13 (3.7)

30 (6.0)

35 (4.4)

Low positive

7 (2.0)

21 (4.2)

23 (2.8)

High positive

6 (1.7)

9 (1.8)

12 (1.5)

Smoking status° Non-smoker

140 (39.4)

125 (25.1)

205 (24.9)

Ex-smoker

137 (38.6)

224 (45.0)

310 (37.7)

Current smoker

78 (22.0) 0

Drinking amount (drink/week) ¶ Table1. Characteristics of subjects

149 (29.9) 7.6 ± 10.3***

307 (37.3) 14.3 ± 14.0***‡

Values are presented as mean ± SD or number (%). * <0.05, ** < 0.01, *** <0.001 compared to non-drinkers by ANOVA with Bonferroni posthoc test † < 0.05, ‡ < 0.001 compared to flushers by ANOVA with Bonferroni posthoc test ° <0.001 compared to each groups by chi-square test ¶ 1 drink = 14g alcohol BMI, body mass index;SBP, systolic blood pressure;DBP, diastolic blood pressure;AST, aspartate transaminase;ALT, alanine transferase;GGT, gamma-glutamyl transferase;CRP, c-reactive protein; ESR, erythrocyte sedimentation rate;UA, uric acid;RF, rheumatoid factor

Table2. Odds ratio(OR) of RF positive by alcohol consumption compared to non-drinkers OR (95% CI) Non-drinkers ≤4 (drinks/week) Model 1

Flushers

Non-flushers

1

1

1.06 (0.47-2.41)

0.92 (0.37-2.24)

Model 2

1.27 (0.55-2.91)

0.92 (0.37-2.29)

Model 3

1.26 (0.54-2.90)

0.92 (0.37-2.29)

Model 1 Model 2

2.59 (1.00-6.73) 2.92 (1.11-7.65)

1.61 (0.67-3.85) 1.65 (0.68-4.00)

Model 3

3.12 (1.18-8.24) *

1.64 (0.67-3.98)

2.53 (1.12-5.68)

1.16 (0.56-2.39)

4<, ≤8 (drinks/week)

8< (drinks/week) Model 1

Model 2 3.08 (1.34-7.05) Model 3 3.27 (1.42-7.52) ** Model 1: Crude Model 2: Adjusted for Age and BMI Model 3: Adjusted for Age, BMI, CRP, ESR, Uric acid and Smoking status * and ** presented statistical power (1-

1.18 (0.56-2.50) 1.17 (0.55-2.50)

err prob) which was 0.89 and 0.95, respectively.

Table3. Odds ratio(OR) of RF positive by alcohol consumption compared to non-flushers in Model 3 OR (95% CI) Flushers Non-flushers

1

(drinks/week) ≤4

1.20 (0.46-3.10)

4<,≤8

1.39 (0.48-4.06)

8<

2.38 (1.05-5.17) *

* presented statistical power (1-

err prob) which was 0.75.

Highlights


Rheumatoid factor (RF) is associated with rheumatoid arthritis




We investigate the link between alcohol consumption-related flushing and RF elevation




More flushers consuming 4< drinks weekly had positive RF than non-drinkers




More flushers with 8< drinks weekly had positive RF than non-flushing counterparts




RF readings should consider the subject’s alcohol consumption levels and with or without facial flushing