Abstract
Objectives:
Moderate alcohol consumption has been associated with decreased systemic lupus erythematosus (SLE) risk, but the biologic basis for this association is unknown. We aimed to determine whether moderate alcohol consumption was associated with lower concentrations of SLE-associated chemokines/cytokines in an ongoing cohort of female nurses without SLE, and whether the association was modified by presence of SLE-related autoantibodies.
Methods:
About 25% of participants from the Nurses’ Health Study (NHS, n=121,700 women) and NHS2 (n=116,429) donated a blood sample; of these, 1177 women were without SLE at time of donation. Cumulative average and current (within 4 years) intakes of beer, wine or liquor were assessed from pre-blood draw questionnaires. Chemokine/cytokine concentrations (stem cell factor [SCF], B-lymphocyte stimulator, interferon-inducible protein-10, interferon-alpha, interleukin-10) and antibodies against double-stranded DNA and extractable nuclear antigens were obtained using ELISAs. Antinuclear antibodies were detected by indirect immunofluorescence on HEp-2 cells.
Results:
At blood draw, women’s mean age was 56 years and 22% were ANA positive; 36% were African-American. About half (46%) reported consuming > 0 - <5g/day of alcohol. SCF levels were 0.5% lower (p<0.0001) for every g/day increase in cumulative average alcohol consumption. Women who consumed >5g/day had mean SCF levels 7% lower (p= 0.002) than non-drinkers. Other cytokines were not significantly associated with alcohol intake. Autoantibody status did not modify observed associations.
Conclusion:
In this study of female nurses, moderate alcohol consumption was associated with lower SCF levels, suggesting a plausible mechanism through which alcohol may lower SLE risk might be by decreasing circulating SCF.
Keywords: lupus, alcohol, risk factor, biomarker, cytokine, chemokine, SCF, epidemiology
Introduction
Systemic lupus erythematosus (SLE) is a complex autoimmune disease likely arising from a combination of genetic factors and environmental exposures. Because irreversible organ damage has often occurred by the time of clinical diagnosis, it is important to understand the preclinical stages of SLE, and how environmental factors may influence individuals’ risks, to develop more effective prevention and treatment strategies.
Clinical diagnosis of SLE is often preceded by years of asymptomatic autoimmunity, with appearance and accumulation of autoantibodies and upregulation of specific inflammatory cytokines and chemokines. A study of 55 subjects’ samples from the Department of Defense Serum Repository detected increased anti-nuclear autoantibodies (ANA), interferon-alpha (IFN-α) activity, and elevated levels of IFN-gamma-inducible protein-10 (IP-10) and B lymphocyte stimulator (BLyS) years to months before SLE diagnosis1. Another prospective study following participants for 6.4 years found that those with relatives who had SLE and who themselves later developed SLE had elevated baseline plasma levels of inflammatory mediators including BLyS, stem cell factor (SCF), interferon-associated chemokines including IP-10 (p ≤ 0.02), and lower levels of immune regulatory mediators, including interleukin-10 (IL-10)2. Although ANA accumulate prior to SLE diagnosis3, their presence is not prognostic as 14% or more of healthy people in the US are reported to be ANA-positive but do not develop an autoimmune rheumatic disease4. However, a case-control study suggested that elevated levels of BLyS, SCF, and type I interferons (including IFN-α) in ANA+ individuals might be involved in progression from asymptomatic autoimmunity to clinical SLE5. Thus, environmental factors that alter levels of these biomarkers in ANA+ individuals might influence the risk of developing SLE.
Several studies have reported that individuals who consumed moderate versus lower amounts of alcohol had significantly decreased risk of developing SLE6, 7. One recent two-sample Mendelian randomization analysis, using several SNPs as a genetic instrumental variable for alcohol consumption, did not find evidence of a strong causal inverse association between alcohol consumption genes and SLE8, and potential biologic mechanisms are unknown. One such pathway is through anti-inflammatory effects. Moderate alcohol consumption has been associated with lower levels of C-reactive protein (CRP) and fibrinogen in observational studies9 and a randomized trial10. Reduction of SLE-associated systemic inflammation might be one pathway by which moderate alcohol consumption decreases the risk of developing SLE, particularly in ANA+ women.
This study examined whether alcohol consumption in women without SLE was associated with peripheral blood concentrations of SLE-related cytokines and chemokines, and if this association differed by ANA status. We hypothesized that both cumulative average amounts of alcohol consumption and recent alcohol consumption within the past 4 years would be associated with lower concentrations of SCF, BLyS, IP-10, and IFN-α, and increased concentration of IL-10; and that these associations would be stronger for women who were ANA+. We considered covariates from the most recent questionnaire prior to the blood draw based on prior studies in NHS cohorts or literature demonstrating an association with alcohol consumption, SLE-related biomarkers, or SLE6.
Methods
Study design and population.
The NHS enrolled 121,700 US nurses aged 30 to 55 years in 1976; the NHS2 enrolled 116,670 nurses aged 25 to 42 years in 1989. Participants completed a baseline questionnaire and are followed biennially(7). In 1988-1990 (NHS) and 1996-1999 (NHSII), ~25% of participants in each cohort donated a blood sample. We identified 697 women who had previously had ANA, anti-double stranded DNA (dsDNA) and extractable nuclear antigen antibodies (ENA, including anti-Sm, anti-RNP,anti-Ro/SSA and anti-La/SSB) tested in the Brigham and Women’s Hospital (BWH) Clinical Immunology lab11, and excluded 33 with confirmed SLE at the time of blood draw. We included all African-American women with blood samples (n=420) to increase minority representation, and 44 women who developed SLE after blood draw, giving a study sample of 1177 women. The Partners Healthcare institutional review board approved this study.
Alcohol Consumption.
Alcohol consumption was assessed with a semiquantitative food frequency questionnaire (FFQ), administered every four years starting in 1980 for NHS and 1989 for NHS212. Respondents reported average frequency of consumption during the previous 12 months. Alcohol consumption (grams/day) was calculated as the sum of daily number of drinks multiplied by average alcohol content per type of alcoholic beverage (12.8g of alcohol per 12-oz serving of beer, 11.0g per 4-oz serving of wine, and 14.0g per serving of liquor). This measure was validated in a sample of NHS participants by comparison with intake calculated from detailed food diaries (Spearman’s r=0.90)12.
Cumulative average alcohol intake was calculated by averaging alcohol use from baseline until the individual’s date of blood draw. If a participant was missing data for one of the assessment years, we took an average of the available measurements. We assigned current/recent alcohol consumption as the most recent response on the FFQ questionnaire, from zero to four years prior to blood draw. In our sample, 92% of NHS women and about half of NHS2 women reported alcohol consumption within one year of blood draw.
Biomarkers.
Women provided plasma samples in heparinized tubes in 1989-1990 (NHS) and in 1996-1999 (NHS2). Collection and storage procedures were similar for the two cohorts and have been described elsewhere11. Plasma BLyS, SCF, IFN-a, IL-10 and IP-10 were measured by individual ELISA assays at the Oklahoma Medical Research Foundation Human Phenotyping Core Laboratory (all coefficients of variation [CVs] <10%). Autoantibody assays passed NHS quality control using blinded splits (Kappas 1.0 for positive vs. negative results). ANA (by indirect immunofluorescence on HEp-2 cells), and anti-dsDNA and ENA antibodies (by ELISA with established clinical cutpoints (Biorad)), were quantified from plasma samples (Brigham and Women’s Hospital Clinical Immunology Laboratory). We defined the ANA-positive cut-off as a serum dilution titer of ≥1:4011. We considered individuals autoantibody-positive if they tested positive either for ANA, dsDNA or ENA.
Covariates.
Demographics included age (continuous), race/ethnicity (African-American vs. non-African-American), and US-Census-tract-based median household income. Health factors were taken from questionnaire closest to blood draw, and included body mass index (BMI, kg/m2; continuous) derived from self-report weight and height13 ; oral contraceptive use (ever/never) and menopausal status/hormone use (pre-menopause, post-menopause/never used, post-menopause/past use, or post-menopause/current use). Oral corticosteroid use (yes/no), which could potentially influence cytokine levels, was ascertained by questionnaire at the time of blood draw. Participants reported smoking status (never/past/current) at baseline, with updates on subsequent biennial questionnaires. Based on prior work, we classified smoking status as current smoker or quit within 4 years before blood draw vs. quit >4 years before blood draw or never smoked(7) .
Statistical methods.
Biomarker concentrations were natural-log-transformed to improve normality. Using restricted cubic splines and likelihood ratio tests, we found no evidence to support a non-linear relationship of average alcohol consumption with cytokine/chemokine outcomes14. Thus, we used general linear regression models to assess associations with BlyS and SCF. We used tobit regression models (SAS PROC QLIM) to assess associations with IP-10, IL-10, and IFN-α as 165 (14%), 846 (72%), and 754 (64%), respectively, of samples fell below the assays’ detection limits. We additionally ran all models using categorical variables for cumulative average alcohol consumption (0 [reference], >0 -<5 gm/day, and ≥5 gm/day), and for current/recent alcohol consumption. Following prior work in NHS and NHS26, 15, analyses were conducted using pooled samples.
We fit two models for each combination of alcohol and biomarker. Model 1 adjusted for age, race, household income, and cohort. Model 2 further adjusted for BMI, corticosteroid use, oral contraceptive use, menopause/hormone status, and recent smoking. To increase interpretability of regression coefficients, since biomarker concentrations were log-transformed, we exponentiated coefficients to obtain differences in the ratio of the expected geometric means of biomarkers. We ran models including an interaction term for autoantibody status with alcohol consumption and reported the p-value for the interaction term.
Results
Age-adjusted characteristics of the 1177 women are shown in Table 1. 287 women (24%) reported cumulative average alcohol consumption of 5g (about half a drink) or more per day; 545 women (46%) reported less than 5g/day; and 29% reported not consuming alcohol. Mean age, income, oral contraceptive use, current use of hormones, and oral corticosteroid use at blood draw all increased with increasing levels of alcohol consumption. BMI decreased with increasing alcohol consumption, as did the percent of pre-menopausal and African-American women. Since the time of blood collection, 44 women reported new onset doctor-diagnosed SLE, which were confirmed by medical record review. Of those, 5 had 3/11 ACR criteria; the rest had ≥4.
Table 1.
Age-standardized participant characteristics, at blood draw, by alcohol consumption, in the NHS/NHS2 study sample (N=1177).
| Cumulative Average Alcohol Consumption (g/day) | |||
|---|---|---|---|
| None | >0 - <5 | ≥5 | |
| Characteristic | (n=345) | (n=545) | (n=287) |
| Age at blood draw, mean (sdb)a | 54.9 (10.6) | 55.2 (9.8) | 59.2 (9.1) |
| Pack years, smokers, median (iqrb) | 0 (0, 4.0) | 0 (0, 15.0) | 7.0 (0, 22.0) |
| Alcohol, g/day, median (iqr) | 0 | 1.2 (0.6, 2.6) | 10.2 (7.4, 17.6) |
| Median tract income, $, mean (sd) | 55,947 (20,948) | 60,820 (23, 387) | 65,459 (25560) |
| BMI, kg/m2, mean (sd) | 28.0 (6.9) | 26.6 (5.6) | 24.8 (5.0) |
| Race, n(%) | |||
| Non-black | 175 (53.9) | 346 (64.9) | 236 (79.0) |
| Black | 170 (46.12) | 199 (35.1) | 51 (21.0) |
| Census region of residence, n(%) | |||
| New England | 23 (7.2) | 43 (8.4) | 22 (8.3) |
| Mid-Atlantic | 102 (30.2) | 195 (35.7) | 106 (35.7) |
| Midwest | 85 (24.2) | 128 (23.1) | 50 (18.8) |
| South | 46 (12.1) | 48 (8.5) | 28 (9.8) |
| West | 89 (26.4) | 131 (24.3) | 81 (27.4) |
| Ever oral contraceptive use, n(%) | 215 (61.2) | 383 (68.6) | 192 (70.9) |
| Menopause/Hormone use, n(%) | |||
| Pre-menopause | 128 (37.1) | 160 (29.4) | 51 (17.8) |
| Post-menopause/never | 93 (27.0) | 176 (32.3) | 86 (30.0) |
| Post-menopause/past | 36 (10.4) | 56 (10.3) | 25 (8.7) |
| Post-menopause/current | 88 (25.5) | 153 (28.1) | 125 (43.6) |
| Steroid use at blood draw, n(%) | 11 (3.3) | 22 (4.1) | 19 (6.4) |
| BLyS (pg/ml), median (iqr) | 1049.9 (843.0, 1276.3) |
1077.1 (869.8, 1315.6) |
1021.7 (832.7, 1201.5) |
| SCF (pg/ml), median (iqr) | 1083.4 (944.5, 1278.2) |
1053.9 (892.6, 1233.2) |
993.9 (855.7, 1169.8) |
| IFN-α (pg/ml), median (iqr)d | 1.7 (0.6, 2.8) |
1.6 (0.7, 3.0) |
2.0 (0.7, 3.1) |
| IP-10 (pg/ml), median (iqr)d | 59.7 (40.4, 102.0) |
58.2 (38.3, 89.6) |
59.8 (38.8, 91.4) |
| IL-10 (pg/ml), median (iqr)d | 14.1 (3.4,101.7) |
9.8 (2.1, 52.1) |
18.7 (2.8, 90.5) |
| ANA ≥ 1:40 | 70 (20.1) | 135 (24.5) | 60 (20.1) |
| ANA ≥ 1:80 | 18 (5.2) | 47 (8.6) | 15 (5.2) |
| ENAc | 12 (3.7) | 36 (6.6) | 21 (6.5) |
| dsDNA | 2 (.6) | 6 (1.1) | 2 (.6) |
Value is not age adjusted.
sd: standard deviation; iqr: inter-quartile range.
Anti- dsDNA, Sm, Ro/SSA, La/SSB, RNP.
Percent below limit of detection: 62.3% (IFN-a), 13.8% (IP-10), 71.9% (IL-10).
In minimally adjusted models, we observed a 0.6% decrease in SCF concentration per 1 gm/day increase in alcohol consumption (p<0.0001; Bonferroni correction for multiple testing p=0.005). The highest category of cumulative average alcohol drinkers had levels of SCF 9% lower (p<0.001) than non-drinkers, with similar results for current/recent alcohol consumption (8% lower, p<0.001). IFN-α and IL-10 levels were lower in higher categories of alcohol consumption compared to non-drinkers, but these differences were not statistically significant. Associations of IP-10 and BLyS with alcohol consumption were not statistically significant.
Adjusting for BMI, corticosteroid use, oral contraceptive use, menopause hormone status, and smoking did not substantially change results (Table 2). We next examined whether associations of alcohol with biomarkers differed according to women’s ANA status, and found no evidence of effect modification for any of the biomarkers (data not shown).
Table 2.
Difference in cytokine/chemokine concentrations with alcohol consumption in the NHS/NHS2 study sample (N=1177).a
| Minimally Adjusted Modelb | Fully Adjusted Modelc | ||||||
|---|---|---|---|---|---|---|---|
| Alcohol Consumptiona | GMR (SD) | % Diff | P | GMR (SD) | %Diff | P | |
| SCFd | Cumulative average alcohol (g/day) | 0.99 (1.00) | −0.60 | <.0001 | 1.00 (1.00) | −0.50 | <.0001 |
| Cumulative average alcohol, categorical | |||||||
| >0-<5g/day | 0.98 (1.02) | −1.88 | 0.3174 | 0.99 (1.02) | −1.29 | 0.506 | |
| ≥5g/day | 0.91 (1.02) | −8.79 | <.0001 | 0.93 (1.02) | −7.13 | 0.0012 | |
| Average alcohol, past 4 years | |||||||
| >0-<5g/day | 0.96 (1.02) | −4.40 | 0.0194 | 0.96 (1.02) | −3.82 | 0.042 | |
| ≥5g/day | 0.92 (1.02) | −8.06 | 0.0003 | 0.93 (1.02) | −7.04 | 0.0015 | |
| IP-10e | Cumulative average alcohol (g/day) | 0.99 (1.01) | −0.80 | 0.195 | 0.99 (1.01) | −0.80 | 0.212 |
| Cumulative average alcohol, categorical | |||||||
| >0-<5g/day | 1.01 (1.12) | 0.90 | 0.9334 | 1.00 (1.12) | −0.30 | 0.9794 | |
| ≥5g/day | 0.92 (1.14) | −8.24 | 0.5137 | 0.93 (1.14) | −6.76 | 0.6002 | |
| Average alcohol, past 4 years | |||||||
| >0-<5g/day | 0.82 (1.12) | −17.96 | 0.0762 | 0.84 (1.12) | −16.39 | 0.1075 | |
| ≥5g/day | 0.88 (1.14) | −12.10 | 0.3328 | 0.91 (1.14) | −8.61 | 0.5034 | |
| IFNαe | Cumulative average alcohol (g/day) | 0.98 (1.03) | −2.18 | 0.3898 | 0.98 (1.03) | 0.4107 | |
| Cumulative average alcohol, categorical | |||||||
| >0-<5g/day | 0.51 (1.58) | −49.14 | 0.1373 | 0.53 (1.57) | −47.11 | 0.1585 | |
| ≥5g/day | 0.77 (1.71) | −23.36 | 0.6182 | 0.82 (1.72) | −18.21 | 0.7093 | |
| Average alcohol, past 4 years | |||||||
| >0-<5g/day | 0.53 (1.59) | −47.22 | 0.1669 | 0.61 (1.58) | −39.47 | 0.2717 | |
| ≥5g/day | 0.68 (1.72) | −31.55 | 0.4833 | 0.81 (1.72) | −18.78 | 0.7001 | |
| BLySd | Cumulative average alcohol (g/day) | 1.00 (1.00) | −0.20 | 0.0775 | 1.00 (1.00) | −0.20 | 0.1486 |
| Cumulative average alcohol, categorical | |||||||
| >0-<5g/day | 1.03 (1.02) | 3.15 | 0.1571 | 1.03 (1.02) | 2.63 | 0.2401 | |
| ≥5g/day | 0.97 (1.03) | −2.86 | 0.2745 | 0.98 (1.03) | −2.27 | 0.3851 | |
| Average alcohol, past 4 years | |||||||
| >0-<5g/day | 1.00 (1.02) | −0.20 | 0.9219 | 1.00 (1.02) | −0.10 | 0.9631 | |
| ≥5g/day | 0.97 (1.03) | −2.96 | 0.2689 | 0.97 (1.03) | −2.66 | 0.3247 | |
| IL-10e | Cumulative average alcohol (g/day) | 0.99 (1.01) | −1.49 | 0.1262 | 0.99 (1.01) | −1.39 | 0.1663 |
| Cumulative average alcohol, categorical | |||||||
| >0-<5g/day | 0.76 (1.20) | −24.27 | 0.1215 | 0.80 (1.20) | −20.47 | 0.2061 | |
| ≥5g/day | 0.81 (1.24) | −18.62 | 0.3368 | 0.87 (1.24) | −13.06 | 0.5229 | |
| Average alcohol, past 4 years | |||||||
| >0-<5g/day | 1.01 (1.20) | 1.01 | 0.9561 | 1.03 (1.20) | 2.53 | 0.8885 | |
| ≥5g/day | 0.70 (1.24) | −30.16 | 0.0977 | 0.72 (1.24) | −28.04 | 0.1326 | |
Differences expressed as geometric mean ratio (GMR) and percent difference (% Diff). Reference category for categorical variables is 0 grams/day.
Model 1 adjusted for age/5yrs (continuous); Race (African-American vs. other) SES: US Census tract-based household income (continuous); cohort (NHS1 or NHS2)
Model 2 additionally adjusted for steroid use at time of blood draw; ever used oral contraceptives; menopause hormone status: (1) pre (ref), (2) post never, (3) post past use, (4) post current; BMI (continuous); current or recent smoking (within past four years).
Models run using linear regression.
Models run using tobit regression to account for proportion of data below the limit of detection of the assay.
Discussion
In this cross-sectional study of 1177 female nurses without SLE, women who reported drinking ≥5 g/day of alcohol, either cumulatively or within the past four years, had lower plasma levels of the SLE-related cytokine SCF. This association did not differ by women’s ANA status. Levels of IFN-α and IL-10 were lower in drinkers compared to non-drinkers, but these results did not reach statistical significance.
In this cohort, with 244 incident cases of SLE among 204,055 nurses, cumulative average alcohol consumption ≥5gm/day was associated with decreased risk of SLE (HR 0.61, 95% CI 0.41-0.89) compared to no alcohol consumption6. Similarly, in the Black Women’s Health Study (127 incident SLE cases among 59,000 women), current drinking (≥4 drinks/week) was associated with decreased risk of SLE (HR 0.71, 95% CI 0.45-1.12) compared to never drinking7. In contrast, a recent study using Mendelian randomization analysis reports no evidence for causal inverse association between alcohol intake and SLE8. Currently identified genetic variants tend to explain only a very small proportion of the variance in a given exposure, (i.e. are weak instrumental variables), and very large numbers of cases are required to detect a causal relationship of an exposure modestly associated with an outcome (such as would be expected with alcohol consumption and SLE). The small number of SNPs used as instrumental variables may have further limited power to detect a true association. Identifying a potential pathway for the observed association of moderate alcohol consumption with decreased risk of SLE would help clarify a causal association.
SCF plays a central role in many biological pathways, including hematopoiesis and immunity, through binding to and activating the receptor tyrosine kinase c-Kit. Immune system mast cells and dendritic cells express high levels of c-Kit and depend on c-Kit for proliferation, survival and function; SCF-c-Kit signaling increases systemic inflammation16. SCF has been implicated in SLE pathogenesis. In a prospective study of 409 healthy relatives of SLE patients, relatives who went on to develop SLE had higher baseline levels of inflammatory mediators, including SCF2. A study comparing ANA+ to ANA- individuals and SLE patients found that SCF was among a handful of pro-inflammatory cytokines that were elevated only among SLE patients, and not in healthy individuals or in ANA+ individuals without SLE. This led authors to propose a model in which increased levels of SCF in ANA+ individuals play a causal role in the transition from asymptomatic autoimmunity to clinical disease5.
Ethanol is known to affect the immune system, but little has been published concerning ethanol’s effects on SCF; and to our knowledge, ours is the first study to examine whether effects differ by ANA status. One study, examining the association between peripheral blood concentrations of SCF and risk of cardiovascular events in 4742 individuals from the Malmö Diet and Cancer Study, incidentally reported that higher alcohol consumption was associated with lower levels of SCF (linear p-for-trend 0.001)17. This is consistent with our own findings of lower SCF concentrations in both ANA- and ANA+ women without SLE who consumed moderate amounts of alcohol. As our study was cross-sectional, however, we cannot infer that alcohol consumption caused altered SCF levels. We could not clearly identify an association between alcohol and IFN-α or IL-10, in part because a high proportion of our assay samples fell below limits of detection so our statistical power was low. Because only 44 of the women in our sample later developed SLE, we lacked power to conduct analyses to predict future SLE onset.
Conclusion
Our finding that moderate alcohol consumption, both long-term and current/recent, is associated with lower SCF is consistent with a model suggesting that elevated SCF plays a role in the etiology of SLE, and that moderate alcohol consumption may decrease risk of SLE by lowering circulating levels of SCF. This might decrease the heightened risk for ANA+ women of transitioning from benign autoimmunity to active disease. Valuable next steps would be to test whether alcohol might act on a pathway through SCF to reduce risk of SLE; and at what consumption levels alcohol is beneficial.
Acknowledgements
The authors would like to acknowledge the Channing Division of Network Medicine and participants in the NHS and NHSII.
References
- 1.Munroe ME, Lu R, Zhao YD, et al. Altered type II interferon precedes autoantibody accrual and elevated type I interferon activity prior to systemic lupus erythematosus classification. Annals of the Rheumatic Diseases. 2016; 75: 2014–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Munroe ME, Young KA, Kamen DL, et al. Soluble mediators and clinical features discern risk of transitioning to classified disease in relatives of systemic lupus erythematosus patients. Arthritis & rheumatology (Hoboken, NJ) 2017; 69: 630. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Arbuckle MR, McClain MT, Rubertone MV, et al. Development of autoantibodies before the clinical onset of systemic lupus erythematosus. New England Journal of Medicine. 2003; 349: 1526–33. [DOI] [PubMed] [Google Scholar]
- 4.Satoh M, Chan EK, Ho LA, et al. Prevalence and sociodemographic correlates of antinuclear antibodies in the United States. Arthritis & Rheumatism. 2012; 64: 2319–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Slight-Webb S, Lu R, Ritterhouse LL, et al. Autoantibody-Positive Healthy Individuals Display Unique Immune Profiles That May Regulate Autoimmunity. Arthritis & rheumatology (Hoboken, NJ). 2016; 68: 2492–502. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Barbhaiya M, Lu B, Sparks JA, et al. Influence of alcohol consumption on the risk of systemic lupus erythematosus among women in the Nurses’ Health Study Cohorts. 2017; 69: 384–92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Cozier YC, Barbhaiya M, Castro-Webb N, et al. Relationship of cigarette smoking and alcohol consumption to incidence of systemic lupus erythematosus in the Black Women’s Health Study. Arthritis care & research. 2018. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Bae S and Lee YH. Alcohol intake and risk of systemic lupus erythematosus: a Mendelian randomization study. Lupus. 2019; 28: 174–80. [DOI] [PubMed] [Google Scholar]
- 9.Imhof A, Froehlich M, Brenner H, Boeing H, Pepys MB and Koenig W. Effect of alcohol consumption on systemic markers of inflammation. The Lancet. 2001; 357: 763–7. [DOI] [PubMed] [Google Scholar]
- 10.Sierksma A, Van Der Gaag M, Kluft C and Hendriks H. Moderate alcohol consumption reduces plasma C-reactive protein and fibrinogen levels; a randomized, diet-controlled intervention study. European journal of clinical nutrition. 2002; 56: 1130. [DOI] [PubMed] [Google Scholar]
- 11.Karlson EW, Hankinson SE, Liang MH, et al. Association of silicone breast implants with immunologic abnormalities: a prospective study 1. 1999; 106: 11–9. [DOI] [PubMed] [Google Scholar]
- 12.Mostofsky E, Mukamal KJ, Giovannucci EL, Stampfer MJ and Rimm EBJ, Ajoph. Key findings on alcohol consumption and a variety of health outcomes from the Nurses’ Health Study. 2016; 106: 1586–91. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Rimm EB, Stampfer MJ, Colditz GA, Chute CG, Litin LB and Willett WCJE. Validity of self-reported waist and hip circumferences in men and women. 1990: 466–73. [DOI] [PubMed] [Google Scholar]
- 14.Durrleman S and Simon R. Flexible regression models with cubic splines. Statistics in medicine. 1989; 8: 551–61. [DOI] [PubMed] [Google Scholar]
- 15.Barbhaiya M, Tedeschi SK, Lu B, et al. Cigarette smoking and the risk of systemic lupus erythematosus, overall and by anti-double stranded DNA antibody subtype, in the Nurses’ Health Study cohorts. 2018; 77: 196–202. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Lennartsson J and Rönnstrand L. Stem cell factor receptor/c-Kit: from basic science to clinical implications. Physiological reviews. 2012; 92: 1619–49. [DOI] [PubMed] [Google Scholar]
- 17.Björkbacka H, Yao Mattisson I, Wigren M, et al. Plasma stem cell factor levels are associated with risk of cardiovascular disease and death. Journal of internal medicine. 2017; 282: 508–21. [DOI] [PubMed] [Google Scholar]