Modifying Effect of N-Acetyltransferase 2 Genotype on the Association Between Systemic Lupus Erythematosus and Consumption of Alcohol and Caffeine-Rich Beverages




N-acetyltransferase 2 (NAT2) is involved in the metabolism of various environmental substances, both with and without carcinogenic potential. Alcoholic and nonalcoholic caffeine-rich beverages may be associated with markers of inflammation. Systemic lupus erythematosus (SLE) is a chronic, multifaceted inflammatory disease. We investigated the effects of alcoholic and nonalcoholic caffeine-rich beverages on risk of SLE and determined whether the effects were modified by NAT2 status.


The NAT2 polymorphism was genotyped in 152 SLE cases and 427 healthy controls, all women and Japanese. We assessed effect modification by testing an interaction term for the NAT2 polymorphism and consumption of beverages.


Consumption of black tea (odds ratio [OR] 1.88, 95% confidence interval [95% CI] 1.03–3.41) and coffee (OR 1.57, 95% CI 0.95–2.61), but not green tea, was associated with an increased risk of SLE, while alcohol use (OR 0.33, 95% CI 0.20–0.55) was associated with a decreased risk of SLE. There were significant interactions between the NAT2 polymorphism and either alcohol use (Pinteraction = 0.026) or consumption of black tea (Pinteraction = 0.048).


The NAT2 polymorphism significantly modified the effects of alcohol use and black tea consumption on SLE, emphasizing the importance of incorporating genetic and metabolic information in studies on management of SLE. Additional studies are warranted to confirm the findings suggested in this study.


Despite intensive research, the etiology of systemic lupus erythematosus (SLE) remains unclear. Many environmental exposures, including smoking, ultraviolet light, medications, infectious agents, hair dyes, and dietary factors, have all been hypothesized to be associated with the development of SLE ([1-4]), although the strength of the evidence implicating each of these factors varies.

Studies of twin concordance are commonly used in epidemiology to estimate the role of genetics and the influence of environmental factors on disease susceptibility. Disease concordance is much higher in monozygotic twins (24–57%) than in dizygotic twins (2–5%), suggesting a genetic component to SLE ([5, 6]). However, identification of these genetic factors has been slow. The genetic basis of SLE is very complex; it has been estimated that more than 100 genes may be involved in SLE susceptibility ([7]), but it is difficult to predict how many genes contribute to SLE susceptibility.

SLE, like other common multifactorial diseases such as cancers, diabetes mellitus, asthma, obesity, and cardiovascular disease, results from a complex interplay of genetic and environmental risk factors. However, triggering events for SLE may include many environmental factors ([8]). A meta-analysis suggested that moderate alcohol consumption, compared with no consumption, was significantly associated with decreased SLE risk (summary odds ratio [OR] 0.66, 95% confidence interval [95% CI] 0.49–0.89) based on 5 studies ([9]). Similarly, our recent study found that light to moderate alcohol consumption was inversely associated with SLE risk, irrespective of the type of alcoholic beverage ([10]). Women who consumed >200 ml of coffee/day had increased inflammation markers, such as interleukin-6 (IL-6) and tumor necrosis factor α, compared with coffee nondrinkers ([11]). Therefore, caffeine-rich beverages such as coffee may be associated with an increased risk of SLE.

As N-acetyltransferase 2 (NAT2) is an important xenobiotic-metabolizing enzyme ([12]), impaired ability to remove reactive substances from the body (the accumulation of the nonacetylated xenobiotics) may play a role in the etiology of autoimmune diseases such as SLE. Therefore, the genetic polymorphism of NAT2 may play a role in susceptibility to SLE. The first study reported a predominance of individuals with slow acetylation activity (slow acetylators) among patients with hydralazine-induced lupus ([13]). Furthermore, procainamide-induced lupus appeared to be more common and to develop more rapidly after a smaller cumulative dose in slow acetylators than in rapid acetylators ([14]). The observation that xenobiotics can cause drug-induced SLE, especially in slow acetylators, suggests that the nonacetylated xenobiotics may accumulate and convert into reactive metabolites. N-acetylation is generally accepted as a detoxificative reaction because acetylation indirectly blocks the oxidation of arylamines ([12]). Hydralazine and procainamide are arylamine drugs. Toxic intermediate metabolites of smoking-related arylamines are detoxified by NAT2. Therefore, the slow acetylator status is associated with a diminished N-acetylation ability to detoxify toxic compounds, thereby increasing SLE risk. The genetic polymorphism of hepatic NAT2 enzyme causes interindividual variation in the response to a variety of amine drugs and potential carcinogens ([15, 16]). Different haplotypes are encoded by at least 7 single nucleotide polymorphisms (G191A, C282T, T341C, C481T, G590A, A803G, and G857A) within the single 870-bp exon of NAT2 ([17]). The most common mutations in the Japanese population are at positions C481T, G590A, and G857A of NAT2 ([18-20]). The major alleles that led to a reduction in NAT2 activity are *6A and *7B, which contain the G590A and G857A substitutions, respectively. NAT2*5B contains the T341C, C481T, and A803G substitutions ([17]). Identification of mutations at positions 481, 590, 803, and 857 will be sufficient to determine mutated alleles as the remaining mutations ([17]). Our previous study found that the NAT2 slow acetylator status may be a determinant in susceptibility to SLE ([21]).

There are conflicting studies on the association between consumption of caffeine-rich beverages and the risk of rheumatoid arthritis ([22-24]), while there are no reports that have addressed the risk of SLE and consumption of caffeine-rich beverages, except for an abstract for a scientific meeting ([25]). Caffeine plays an important role in the regulation of cytokines ([26]), which have been identified as important players in SLE risk ([27]). Ethanol or its metabolites, rather than specific substances in alcoholic beverages, may modulate cytokine release, which in turn will decrease SLE risk. Because caffeine consumption is very common, we investigated whether consumption of caffeine-rich beverages, such as coffee, black, or green tea, affects the risk of SLE. Caffeine is most commonly used as a probe drug for NAT2 phenotype determinations ([28]) and excellent NAT2 genotype-phenotype association has been reported ([16]). Furthermore, we investigated risk modification by the NAT2 polymorphisms in the association of alcohol use, coffee, and other caffeine-rich beverages and SLE risk in Japanese women.

Box 1. Significance & Innovations

  • Consumption of black tea and coffee was associated with increased systemic lupus erythematosus (SLE) risk.
  • The N-acetyltransferase 2 polymorphism modified the association between alcohol and black tea consumption and SLE risk.


Study subjects

The Kyushu Sapporo SLE study was a case–control study to evaluate risk factors for SLE among women. SLE patients (n = 129) were recruited from outpatients of Kyushu University Hospital, Saga University Hospital, and their collaborating hospitals in Kyushu from 2002–2005, while 51 SLE patients were recruited from outpatients of Sapporo Medical University Hospital and its collaborating hospital in Hokkaido from 2004–2005. All patients (n = 180) fulfilled the American College of Rheumatology 1982 revised criteria for SLE ([29]). The mean ± SD duration of SLE was 11.9 ± 8.55 years. An antinuclear antibody (ANA) test was ordered as a routine screening test if there was a reasonable suspicion of SLE from family history and/or physical findings. Therefore, we performed an ANA test for all SLE patients and almost all of the patients had a positive ANA test result. The rheumatologists in charge asked eligible SLE patients to take part in this study and obtained written informed consent from them. SLE patients with cognitive dysfunction were not included in this study.

Controls were not, individually or in larger groups, matched to cases. Controls (n = 268) were recruited from nursing college students and care workers in nursing homes (n = 57) in Kyushu, while in Hokkaido controls (n = 188) were recruited from participants at a health clinic in a local town.

In analysis, 18 subjects (8 cases and 10 controls) were excluded because of male sex. A portion of the participants agreed to donate blood samples, which were stored until use for DNA extraction and genotyping of the candidate genes of SLE. Only women who agreed to donate blood samples were included in this study (152 cases and 427 healthy controls).

All SLE patients and controls provided written informed consent for cooperation in the study. The present study was approved by the institutional review boards of Kyushu University Graduate School of Medical Sciences, Sapporo Medical University, St. Mary's College, and the other institutions involved.

Questionnaire survey

Cases were asked to complete a self-administered questionnaire about their lifestyles before the diagnosis of SLE, while controls completed the questionnaire about their current lifestyles. Subjects were considered current smokers if they smoked or had stopped smoking less than 1 year before either the date of diagnosis (SLE patients) or the date of completion of the questionnaire (controls). The relevant ages would be age at diagnosis (SLE patients) and age at time of questionnaire (controls). Nonsmokers were defined as those who had never smoked in their lifetime. Former smokers were those who had stopped smoking 1 year or more before either the date of diagnosis (SLE patients) or the date of completion of the questionnaires (controls).

Similarly, subjects were considered current drinkers if they consumed alcohol before either the date of diagnosis (SLE patients) or completion of the questionnaire (controls). Nondrinkers were defined as those who had never consumed alcohol in their lifetime. Frequency of consumption of nonalcoholic beverage items was measured on a scale of 5 categories (never, 1 cup/day, 2–3 cups/day, 4–6 cups/day, 7–9 cups/day, and >9 cups/day) at the relevant age. Study subjects were also asked about educational background as a surrogate for socioeconomic status (junior high school, high school, junior college/vocational college, and university/postgraduate school). All subjects were asked about their medical history and family history of selected diseases (such as SLE, rheumatoid arthritis, cancer, diabetes mellitus, stroke, etc.). There were no controls with self-reported SLE and with a family history of SLE. Details of the health examination and the self-administered questionnaire have been documented elsewhere ([30, 31]).

Genetic analysis

Genomic DNA was extracted from buffy coat stored at −80°C using the QIAamp blood kit (QIAGEN). The most common mutations in the Japanese population at positions C481T, G590A, and G857A of NAT2 were analyzed using Kpn I, Taq I, and BamH I by the polymerase chain reaction–restriction fragment length polymorphism (PCR-RFLP) method, as described elsewhere ([18-20]). According to the nomenclature of NAT2, wild-type and 3 variant alleles were defined as NAT2*4 and *5B, *6A, *7B. Subjects were classified by this genotyping into 3 groups: homozygous for the major allele *4/*4 (rapid acetylator), heterozygous for the major and minor alleles *4/*5B, *4/*6A, and *4/*7B (intermediate acetylator), and homozygous for the minor alleles *5B/*5B, *5B/*6A, *5B/*7B, *6A/*6A, *6A/*7B, and *7B/*7B (slow acetylator). For genotyping quality control, we retyped randomly selected samples (10% of previously typed samples) with the PCR-RFLP method and confirmed the complete agreement of genotyping.

Statistical analysis

We used chi-square statistics for homogeneity to test for case–control differences in the distribution of several covariates and the NAT2 genotypes. The distribution of the NAT2 genotypes in controls was compared with that expected from Hardy-Weinberg equilibrium by the chi-square test. Unconditional logistic regression was used to compute the ORs and their 95% CIs with adjustments for several covariates (age, region of residence, smoking status, alcohol intake, and educational background). Age was treated as a continuous variable. The remaining covariates were treated as categorical variables. Region of residence fell into 2 categories (Kyushu and Hokkaido), as did smoking status (current and former smokers combined and nonsmokers), alcohol drinking status (current and former drinkers combined and nondrinkers), and the NAT2 status (slow and intermediated acetylators combined and rapid acetylators). Consumption of nonalcoholic beverages was classified into 4 categories (0, 1, 2–3, and >3 cups/day) due to the small number in the highest 3 consumption categories.

The trend was assessed by assigning ordinal values for categorical variables. The interaction between NAT genotypes and either alcohol use or consumption of nonalcoholic caffeine-rich beverages on the risk of SLE was statistically evaluated based on the likelihood ratio test, comparing the logistic models with and without (multiplicative scale) terms reflecting the product of the genotype and consumption status for interaction ([32]). In a logistic regression model, interaction refers to a departure from multiplicativity.

All statistical analyses were performed using the computer program STATA, version 12.1. P values were 2-sided, with those less than 0.05 considered statistically significant. Because of the low power of the test for interaction, P values of less than 0.1 were used for statistical significance ([33]).


There were 152 women with SLE and 427 healthy women enrolled in this study. As shown in Table 1, the age (mean, 95% CI) of patients with SLE (41.2 years, 39.2–43.3) was significantly higher than that of controls (31.9 years, 30.5–33.2) (P < 0.0001). From the questionnaire, the mean age (95% CI) at the time of diagnosis of SLE was 29.1 years (27.3–31.0) (data not shown). There was also a significant difference between the age at diagnosis (SLE) and age at completion of the questionnaire (controls) (P = 0.04; data not shown). Compared with controls, cases were more likely to report a history of smoking (P = 0.001) and a higher educational background (P < 0.0001). On the other hand, controls tended to drink alcohol more frequently than SLE patients (P < 0.0001). The distribution of the NAT2 genotypes was significantly different between cases and controls (P = 0.001).

Table 1. Selected characteristics of SLE cases and controls*
CharacteristicsCases (n = 152)Controls (n = 427)P
  1. Values are the number (percentage) unless indicated otherwise. SLE = systemic lupus erythematosus; 95% CI = 95% confidence interval; NAT2 = N-acetyltransferase 2.
  2. aSeveral observations with missing values.
  3. bRapid: *4/*4; intermediate: *4/*5B, *4/*6A, *4/*7B; slow: *5B/*5B, *5B/*6A, *5B/*7B, *6A/*6A*6A/*7B, *7B/*7B.
Age, mean (95% CI) years41.2 (39.2–43.3)31.9 (30.5–33.2)< 0.0001
Region of residence   
Hokkaido51 (33.6)176 (36.4)0.10
Kyushu101 (66.5)251 (63.6) 
Cigarette smoking statusa   
Nonsmoker98 (64.9)339 (79.6)0.001
Former smoker7 (4.64)18 (4.23) 
Current smoker46 (30.5)69 (16.2) 
Alcohol drinking statusa   
Nondrinker68 (45.0)120 (28.1)< 0.0001
Former drinker0 (0.00)0 (0.00) 
Current drinker83 (55.0)303 (71.6) 
Educational backgrounda   
Junior high school14 (9.21)26 (6.10)< 0.0001
High school77 (50.7)291 (68.3) 
Junior college/vocational college45 (29.6)103 (24.2) 
University/postgraduate school16 (10.5)6 (1.41) 
NAT2 polymorphismb   
Rapid acetylator genotype23 (15.1)130 (30.4)0.001
Intermediate acetylator genotype89 (58.6)207 (48.5) 
Slow acetylator genotype40 (26.3)90 (21.1) 

Table 2 shows the association between consumption of alcoholic and nonalcoholic caffeine-rich beverages and SLE risk. After adjustment for age, region, smoking status, and educational background, alcohol drinkers had significantly decreased SLE risk (OR 0.33, 95% CI 0.20–0.55) compared with alcohol nondrinkers. Green tea drinking was not significantly associated with decreased SLE risk (OR 0.69, 95% CI 0.42–1.13). In contrast, black tea drinking was significantly associated with increased SLE risk (adjusted OR 1.88, 95% CI 1.03–3.41). Similarly, coffee drinking was marginally associated with increased SLE risk (OR 1.57, 95% CI 0.95–2.61). A dose-dependent relationship (Ptrend = 0.048) was revealed between number of cups of coffee consumed per day and SLE risk.

Table 2. Association between consumption of alcoholic and nonalcoholic caffeine-rich beverages and SLE risk*
 No. cases/ controlsOR (95% CI)
  1. SLE = systemic lupus erythematosus; OR = odds ratio; 95% CI = 95% confidence interval.
  2. aAdjusted for age, region, smoking status educational background and, where appropriate, for drinking status.
  3. bSeveral observations with missing values.
Alcohol useb   
Drinkers83/3030.48 (0.33–0.71)0.33 (0.20–0.55)
Green teab   
0 cups/day48/1361.0 (reference)1.0 (reference)
1 cup/day21/720.83 (0.46–1.49)0.70 (0.35–1.43)
2–3 cups/day43/1280.95 (0.59–1.53)0.63 (0.35–1.14)
≥4 cups/day31/671.31 (0.77–2.25)0.78 (0.39–1.56)
Ptrend 0.2430.561
Green tea drinkers95/2671.01 (0.67–1.51)0.69 (0.42–1.13)
Black teab   
0 cups/day92/2941.0 (reference)1.0 (reference)
1 cup/day19/481.26 (0.71–2.26)1.67 (0.84–3.32)
2–3 cups/day8/122.13 (0.85–5.37)3.37 (1.10–10.4)
≥4 cups/day3/51.92 (0.45–8.18)1.11 (0.16–7.69)
Ptrend 0.1140.192
Black tea drinkers30/651.48 (0.90–2.41)1.88 (1.03–3.41)
0 cups/day42/1761.0 (reference)1.0 (reference)
1 cup/day30/951.32 (0.78–2.25)1.13 (0.60–2.12)
2–3 cups/day56/912.58 (1.61–4.14)1.98 (1.10–3.58)
≥4 cups/day15/252.51 (1.22–5.18)1.97 (0.79–4.89)
Ptrend 0.0010.048
Coffee drinkers101/2112.01 (1.33–3.03)1.57 (0.95–2.61)

Table 3 shows the interaction between the NAT2 polymorphism and consumption of alcohol or nonalcoholic caffeine-rich beverages with regard to SLE risk. After adjustment for age, region, smoking status, and educational background, the slow acetylator and intermediate acetylator genotypes were associated with increased SLE risk (OR 2.15, 95% CI 1.20–3.88 and OR 2.26, 95% CI 1.15–4.47, respectively; data not shown). Since ORs of the intermediate acetylator and slow acetylator genotypes were similar in our Japanese population, they were combined in subsequent analyses. The combined adjusted OR of the intermediate acetylator and slow acetylator genotypes was 2.19 (95% CI 1.24–3.85) compared with the rapid acetylator genotype (data not shown). Alcohol drinkers with the rapid acetylator genotype (OR 0.08, 95% CI 0.03–0.28) had lower SLE risk than those with the nonrapid acetylator genotype (OR 0.28, 95% CI 0.11–0.72), relative to nonalcohol drinkers with the rapid acetylators genotype (reference). The OR for nonalcohol drinkers with the nonrapid acetylator genotype was 0.76 (95% CI 0.28–2.01). The interaction between the NAT2 polymorphism and alcohol drinking was statistically significant (Pinteraction = 0.026).

Table 3. Interaction between NAT2 polymorphism and alcohol or nonalcoholic beverage with regard to SLE risk in Japanese women*
BeverageNAT2 status, rapid acetylator genotypeNAT2 status, nonrapid acetylator genotypePinteractive
No. cases/controlsAdjusted OR (95% CI)aNo. cases/controlsAdjusted OR (95% CI)a
  1. NAT2 = N-acetyltransferase 2; SLE = systemic lupus erythematosus; OR = odds ratio; 95% CI = 95% confidence interval.
  2. aAdjusted for age, region, smoking status, educational background and, where appropriate, for drinking status, consumption of green tea, black tea, and coffee.
Never13/331.0 (reference)55/870.76 (0.28–2.01) 
Ever10/960.08 (0.03–0.28)73/2070.28 (0.11–0.72)0.026
Green tea     
Never13/361.0 (reference)35/1001.16 (0.47–2.82) 
Ever9/880.29 (0.09–0.88)86/1790.81 (0.35–1.91)0.158
Black tea     
Never20/851.0 (reference)72/2091.38 (0.71–2.66) 
Ever2/230.41 (0.07–2.27)28/423.60 (1.57–8.28)0.048
Never10/481.0 (reference)32/1281.06 (0.43–2.61) 
Ever12/680.71 (0.25–2.03)89/1432.06 (0.86–4.95)0.094

Consumption of green tea with the rapid acetylator genotype was significantly associated with decreased SLE risk (adjusted OR 0.29, 95% CI 0.09–0.88) compared with nonconsumption of green tea with the rapid acetylator genotype (reference). The interaction between the NAT2 polymorphism and consumption of green tea failed to reach significance. Among consumers of black tea, individuals with the rapid acetylator genotype (OR 0.41, 95% CI 0.07–2.27) presented lower SLE risk than those with the nonrapid acetylator genotype (OR 3.60, 95% CI 1.57–8.28), relative to nonconsumers of black tea with the rapid acetylator genotype (reference). Evidence of interaction between the NAT2 polymorphism and consumption of black tea was observed (P = 0.048). Consumers of coffee with the nonacetylator genotype were nonsignificantly associated with increased SLE risk (OR 2.06, 95% CI 0.86–4.95). An interaction between the NAT2 polymorphism and consumption of coffee was also observed (P = 0.094).


We examined the impact of alcoholic and nonalcoholic caffeine-rich beverages on SLE risk alone or in combination with the NAT2 polymorphism among 152 SLE cases and 427 controls in Japanese women. To the best of our knowledge, this is the first study showing that consumption of black tea and coffee was associated with an increased risk of SLE, and that there are significant interactions between the NAT2 polymorphism and alcohol use, consumption of black tea, or consumption of coffee in relation to SLE risk.

Green tea and coffee are the most popular nonalcoholic beverages in Japan. Black tea is the third most popular beverage in Japan. Caffeine is an important component of each of these beverages and is widely used in other foods and medications. The complex pharmacogenetic and physiologic effects of caffeine have prompted a great deal of investigation into the health consequences of caffeine ingestion ([34]). According to Standard Tables of Food Composition in Japan (2010) ([35]), caffeine content (mg/100 ml) for coffee, black tea, and green tea is 60 mg, 30 mg, and 20 mg, respectively.

Consumption of coffee was nonsignificantly associated with an increased risk of SLE, while consumption of black tea was significantly associated with an increased risk of SLE (Table 2). Caffeine is the common and major constituent present in coffee and black tea. Coffee drinkers had increased inflammation markers compared with coffee nondrinkers ([11]). Therefore, it is plausible that caffeine may be associated with increased SLE risk. Chlorogenic acid, a major phenolic acid in coffee, is a well-known potent antioxidant and may moderately attenuate the deleterious effects of caffeine. Consumption of black tea and green tea had opposite effects on the risk of SLE (Table 2). Although black tea is made from the same plant leaves used to make green tea, green tea is not fermented and contains more catechins than black tea. The biologic mechanisms whereby consumption of green tea may not affect SLE remain speculative, and several hypotheses have been considered. One hypothesis is that epigallocatechin gallate (EGCG), one of the active ingredients of green tea, stimulates the immune system and attenuates the effect of caffeine. Laboratory studies showed that the EGCG has marked modulating effects on cytokine production by immune cells ([36, 37]). EGCG is also a powerful antioxidant. Black tea contains much lower concentrations of catechins such as EGCG than green tea ([38]). It has been reported that green and black teas contained total phenols equal to 165 and 124 mg gallic acid, respectively. They also found that the antioxidant capacity per serving of green tea (436 mg vitamin C equivalents) was much higher than that of black tea (239 mg). Therefore, they concluded that green tea has more health benefits than an equal volume of black tea in terms of antioxidant activity ([39]). Another hypothesis is that L-theanine (γ-glutamylethylamide, theanine), an amino acid found in green tea, has an antagonistic effect on caffeine's stimulatory action. Several studies reported reduction of the stimulatory effects of caffeine by theanine administration ([40-42]). Furthermore, green tea has been reported to have the opposite effect of black tea ([38]). Animal studies have demonstrated that green tea consumption may reduce the severity of some autoimmune disorders ([43, 44]), but the mechanism is unclear. Ingredients in green tea may mitigate the adverse effect caused by caffeine or another component(s) rich in black tea.

A recent meta-analysis reported that light to moderate alcohol consumption had a significant protective effect on SLE risk (summary OR 0.72, 95% CI 0.55–0.95) when limited to patients treated for <10 years ([9]), although a recent study showed that alcohol consumption before SLE diagnosis was not associated with the risk of SLE ([45]). There is a U-shaped relationship between alcohol consumption and mortality from all causes ([46]). Our subjects (many of whom were light to moderate drinkers) ([10]) may have lower SLE risk than nondrinkers. The biologic mechanisms whereby alcohol may affect SLE remain speculative. First, alcoholic beverages potentially attenuate the risk of inflammatory disease such as SLE ([47]). It has been suggested that the overproduction of IL-6 in SLE patients may lead to the pathogenesis of the disease ([48]). Moderate alcohol consumption inhibits production of IL-6 ([49]). It has been suggested that ethanol or its metabolites, rather than specific substances in alcoholic beverages, may modulate cytokine release, which, in turn, will decrease SLE risk ([10]). Beer is the most popular alcoholic beverage in our study population, followed by wine, sake (Japanese rice wine), and shochu (Japanese distilled spirit) (data not shown). Beer is a rich source of niacin (vitamin B3), with a 350 ml serving of regular beer providing approximately 2.8 mg according to the food composition table ([50]). As niacin possesses strong antioxidant and antiinflammatory properties ([51]), niacin may be beneficial to the development of SLE. Antioxidants such as resveratrol or humulones contained in wine or beer have also been shown to influence cytokine cascades in vitro ([52]). Therefore, it is biologically plausible that appropriate alcohol drinking is associated with decreased SLE risk.

It is widely accepted that SLE development requires environmental factors acting on a genetically predisposed individual. Studying gene–environment interactions in relation to SLE risk may be valuable, as positive findings would clearly implicate the substrates with which the gene interacts as disease-causing exposures, clarifying SLE etiology and pointing to environmental modifications for disease prevention. As metabolism of toxic xenobiotics is controlled by individual genetics, we analyzed the interaction between alcohol use and the NAT2 polymorphism (Table 3). A significant effect of alcohol on SLE risk was observed in this study without accounting for the NAT2 polymorphism, although there is presently no evidence that the NAT2 enzyme could directly metabolize ethanol. Other studies have also reported an interaction between the NAT2 polymorphism and alcohol use with respect to various outcomes ([53-56]). Therefore, the effect modification by alcohol use could be a reflection of the effect modification of the substances correlated with alcohol use, or the alcohol adding to total body burden of toxicants, so that the NAT2 enzyme and other enzymes cannot metabolize toxicants as efficiently.

As NAT2 is a key enzyme in the metabolism of caffeine, we evaluated whether interactions existed between consumption of nonalcoholic caffeine-rich beverages and the NAT2 polymorphism (Table 3). There was also a significant interaction between the NAT2 polymorphism and consumption of either black tea or coffee. Namely, the effects of the drinking status of black tea or coffee (the NAT2 acetylator status) on SLE risk significantly varied depending on the NAT2 acetylator status (the drinking status of black tea or coffee). The results suggest that the accumulation of nonacetylated metabolite(s) of the black tea or coffee component(s) may be associated with an increased risk of SLE, while the accumulation of the acetylated metabolite(s) of those may have no significant effects on SLE risk. Further research is needed to fully understand the interaction between environmental and genetic factors.

Several limitations of this study warrant mention. Our study may have included a bias due to the self-reporting of alcohol use and consumption of nonalcoholic caffeine-rich beverages (misclassification bias). However, the validity of self-reports on alcohol use is generally high ([57, 58]). Similarly, the validity of consumption of coffee and tea using a self-administered questionnaire is relatively high ([59, 60]). Recall bias, which occurs when cases and controls recall exposures differently, is also a well-recognized potential problem in case–control studies. SLE patients may be more likely to report their prior exposures than healthy controls because they think they might be related to their disease. The possibility of recall bias in reporting consumption of alcoholic and nonalcoholic beverages may be minimized because SLE patients are unlikely to be aware that these habits may be associated with SLE risk. In addition, it has been reported that the report of remote (3–20 years ago) diet correlated more closely with original dietary report than did the report of current diet ([61-63]). The observations suggest that if diet from several years past is thought to be relevant to SLE risk, we may generate a more reliable estimate of the past beverage consumption by questioning subjects directly about past beverage consumption rather than current beverage consumption. The average interval between the age at diagnosis and age at enrollment was 12.1 (95% CI 10.7–13.4). When the cases were limited within 10 years of the onset of SLE, the similar but nonsignificant associations were observed due to the limited sample size. The adjusted OR (95% CI) for consumption of alcohol, green tea, black tea, and coffee was 0.58 (0.24–1.42), 0.56 (0.25–1.26), 1.81 (0.62–5.28), and 2.03 (0.81–5.09), respectively. Furthermore, the alcohol and coffee/black tea had an opposite effect on SLE risk. Inaccuracies in recall and reporting were possible and, as they were likely nondifferential, could cause dilution of a true association. Population-based case–control studies may have underestimated slightly the true association due to recall bias ([64]). Finally, reproducibility of beverage consumption has been reported to be higher than that of food intake ([65]).

Case–control studies tend to be susceptible to selection bias, particularly in the control group. Selection bias may occur if the decision to participate is affected by exposure status. In many cases, selection bias is not extreme enough to have an impact on inference and conclusions ([66]). As the possibility of recall and selection biases could not be completely excluded in case–control studies, our findings should be interpreted with caution. A fundamental conceptual issue, selection of controls, is whether the controls should be similar to the cases in all respects other than status of the disease in question. As controls were not selected to match SLE patients on confounding factors, there were significant differences between them, such as age and educational background. Although matching is one approach to control for confounding bias in the design of the study, the confounding bias can also be controlled for by using a statistical modeling approach in the analysis, as was carried out in our study.

In conclusion, our data indicate that consumption of black tea, and possibly coffee, may increase SLE risk, although we cannot provide any prompt explanation in regard to underlying biologic mechanisms. We observed significant effects of alcohol use and consumption of black tea and coffee on SLE patients who were nonrapid NAT2 acetylators. Findings from gene–environment interaction analyses must be interpreted with caution due to reduced numbers of observations in the subgroups. Replication of findings is very important before any causal inference can be drawn. Testing replication in different populations is an important step. Future studies involving larger control and case populations are warranted to corroborate the association among Japanese samples suggested in the present study.


All authors were involved in drafting the article or revising it critically for important intellectual content, and all authors approved the final version to be submitted for publication. Dr. Kiyohara had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study conception and design. Kiyohara, Washio, Horiuchi, Asami, Ide, Atsumi, Kobashi, Takahashi, Tada.

Acquisition of data. Kiyohara, Washio, Horiuchi, Asami, Ide, Atsumi, Kobashi, Takahashi, Tada.

Analysis and interpretation of data. Kiyohara, Washio, Horiuchi, Asami, Ide, Atsumi, Kobashi, Takahashi, Tada.


The authors thank Takasu Town, Hokkaido, and its town people for their kind cooperation.


Members of the Kyushu Sapporo SLE (KYSS) Study Group listed in alphabetical order for each affiliation are as follows: Saburo Ide, Hiroko Kodama, Masakazu Washio (principal investigator) (St. Mary's College); Koichi Akashi, Mine Harada, Takahiko Horiuchi (co-principal investigator) (Department of Internal Medicine, Kyushu University Beppu Hospital); Chikako Kiyohara (co-principal investigator), Hiroaki Niiro, Hiroshi Tsukamoto (Graduate School of Medical Sciences, Kyushu University); Toyoko Asami, Takao Hotokebuchi, Kohei Nagasawa, Yoshifumi Tada, Osamu Ushiyama (Faculty of Medicine, Saga University); Mitsuru Mori, Asae Oura, Yasuhisa Sinomura, Hiromu Suzuki, Hiroki Takahashi, Motohisa Yamamoto (Sapporo Medical University School of Medicine); Gen Kobashi (Research Center for Charged Particle Therapy, National Institute of Radiological Science); Tatsuya Atsumi, Tetsuya Horita, Takao Koike (Hokkaido University Graduate School of Medicine); Takashi Abe (Kushiro City General Hospital); Hisato Tanaka (Tanaka Hospital); Norihiko Nogami (Wakakusuryouikuen Hospital); Kazushi Okamoto (Aichi Prefectural College of Nursing and Health); Naomasa Sakamoto (Hyogo College of Medicine); Satoshi Sasaki (School of Public Health, the University of Tokyo); Yoshihiro Miyake (Faculty of Medicine, Fukuoka University); Tetsuji Yokoyama (National Institute of Public Health); Yoshio Hirota (Faculty of Medicine, Osaka City University); Yutaka Inaba (Juntendo University School of Medicine); Masaki Nagai (Saitama Medical School).