March 18th, 2003
Curr Opin Rheumatol 15(2):99-103, 2003.
M. A. Dooley, MD, MPH; S. L. Hogan, PhD, MPH, Division of Rheumatology and Immunology, University of North Carolina at Chapel Hill, Chapel Hill, North Carolina, USA
Abstract and Introduction
It has long been recognized that environmental influences play an important role in the risk of developing chronic rheumatic disease. Defining specific pathogenic environmental mediators that may trigger the development or progression of autoimmune disease remains a focus of increasing investigative effort. Factors promoting disease may not be identical to factors that influence the severity or progression of the disorder. Human monozygotic twin studies, animal studies, and genetic models demonstrate that genetic influences strongly determine whether one will develop autoimmunity, however, genes affecting the metabolism of exogenous agents that may trigger disease expression have only recently drawn attention. In this article the authors review recent reports that advance our understanding of previously recognized environmental risk factors and challenge accepted beliefs that increased estrogenic exposures predate the incidence of autoimmune disorders, systemic lupus erythematosus in particular.
The cause(s) of autoimmune disorders remain largely unknown. Considerable evidence supports a role for environmental agents in inducing autoimmune disorders.[1, 2*] An estimated 3% of the population in the United States is affected by a tissue-specific or systemic autoimmune disorder. Most autoimmune disorders affect women more frequently, including 85% of patients with systemic lupus erythematosus (SLE), scleroderma, or thyroiditis. Autoimmune diseases thus contribute disproportionately to morbidity and mortality among young to middle-aged women. Rather than disease-specific genes for individual autoimmune disorders, there may be "autoimmunity genes" that increase the risk for development of autoimmune disorders in families. Autoimmune disorders may result from multiple interactions of genes and environmental factors.
The evaluation of genetic contributions in human autoimmunity has frequently relied on family studies, and particularly, twin studies. Twin studies have been criticized for including a higher proportion of monozygotic, female, and disease concordant volunteers.[7-9] Previous twin studies have reported significantly higher concordance rates of 24 to 50% in autoimmune diseases such as rheumatoid arthritis (RA) and SLE among monozygotic twins.[7, 10-12] However, recent population-based twin studies in RA detect at most a 12 to 15% concordance rate among monozygotic twins, though are criticized for their lack of validated classification criteria for disease confirmation.[13-15]
A recent nationwide study among twins in Denmark reports only a minor contribution of genetic factors to the risk of developing RA.[16*] Subjects were ascertained by questionnaire; those reporting RA provided a detailed clinical profile and a confirmatory phone interview. Record linkage with the Danish discharge registry was employed to locate additional subjects. Both twins were invited to a clinical examination, structured interview, and blood sampling for zygosity, rheumatoid factor, and HLA typing. The response rate was 75% in older and 86% in younger twins. A total of 13 monozygotic and 36 dizygotic RA concordant twin pairs participated. No significant differences between monozygotic and dizygotic twins were found in sex, age at onset of RA, mean discordance time, presence of rheumatoid factor, bony erosions, or shared epitopes. Given the small number of twins pairs included, the study is not sufficiently powered to conclude that genetics contributes little to the risk of developing RA. A recent population-based study of risks of autoimmune thyroid disease stresses the increasing importance of environmental contributions to the development of autoimmunity.[17*]
Case-Controlled Population Based Studies
In the past year, results from two large, well designed, population-based case-control studies have contributed substantial insights into risk factors for developing RA[16*] and SLE.[18*, 19*, 20*, 21*] The case-control study design offers a unique opportunity to simultaneously evaluate a number of potential causal factors. These studies include genetic typing and a detailed description of environmental exposures. The details of these case control studies are summarized in Table 1. The RA study describes a positive association with smoking as well as with a short fertile period in women and inverse relationships with length of breast-feeding, history of atopic allergies, psychosocial stress, and level of education,[16*] consistent with previous studies. This study also suggests increased risk for RA with the use of private well water, a history of self-reported thyroid conditions, and use of insulin. These factors have received little prior attention but may provide new insights and should be explored more fully.
Table 1. Summary of two recent case-control studies
The Carolina Lupus Study (CLU), a population-based case-control study based in the southeastern United States explored a number of risk factors for the development of SLE.[18*, 19*, 20*, 21*] Risk factors evaluated in this study included smoking, use of hair treatments, infectious exposures, hormonal and reproductive factors, and occupational exposure to crystalline silica. Each of these exposures is reviewed in more detail below.
Cigarette smoking was associated with an increased risk of developing RA.[23, 24] Three US-based studies report no association between smoking and the onset of SLE,[17*, 25*, 26] while three case control studies from Japan, the United Kingdom, and Sweden detected statistically significant increased risks of developing SLE among current smokers.[23, 27, 28*] Cigarette smoking has been identified as a risk factor leading to the development of lupus nephritis with poor outcome.[29-31] The apparent contradictions among studies may reflect differences in exposures. Information on the composition of cigarettes is proprietary and does not require disclosure in many countries, including the United States. There may be considerable differences in nicotine and other cigarette components among geographic regions. Differences in cigarette products with respect to pesticides or fertilizers applied, curing practices, ingredients, filters, additives, papers, and other components could account for these differences in the associations been between smoking and SLE from country to country.
The use of permanent hair dyes in women was associated with a borderline increase in the risk of developing SLE, with higher risk among those with a longer duration of dye use in the CLU study.[18*] Although one previous study reported a strong association between the use of hair dyes and connective tissue disease, other studies that have evaluated the association of this exposure specifically with the onset of SLE have found no association.[33-35] As with smoking, there are large variations in the chemical content of hair dyes (International Agency for Research on Cancer 1993). The weak association appreciated in the CLU study[18*] makes it difficult to discount the potential contribution of some ingredients within permanent dyes as possible promoters or contributing causal components in conjunction with other exposures to the development of SLE. However, the difficulty posed by recall of specific products used, and changes in products over time makes this a particularly complex exposure to study.
Infection as a possible trigger for autoimmune disorders has long been proposed. Viral infections have received particular attention in SLE studies, with findings of virus-like inclusions in renal biopsy tissue. Previous studies have noted associations of SLE with history of herpes zoster and Epstein-Barr infections. Vaccination against hepatitis B has been noted to predate the diagnosis of SLE in some instances. A recent model linking environmental exposures and autoimmunity invokes interferon- as a superantigen.[40*]
Evaluations of infectious exposures in studies of RA[16*] are inconsistent and difficult to interpret. Urinary tract infections and rubella were found to be associated with decreased risk and pneumonia with an increased risk of RA in women, but no type of infection was associated with the onset of RA in men. Also, no associations were observed between vaccinations, exposure to pets, previous operations, or other kinds of trauma and the onset of RA in this study.[16*] The CLU study included assessments of infectious exposures in SLE but the findings are in press.[41*]
A number of hormonal and reproductive factors were found not to be associated with the onset of SLE in the CLU study. The unassociated factors include early menarche, early natural menopause, number of pregnancies or live births, and current use or duration of use of hormone replacement therapy or oral contraceptives.[19*] Hormonal variability has been evaluated in the setting of established SLE in both mice and humans as risk factors for disease severity and progression. No prior human studies have evaluated these specific factors in terms of the risk of onset of SLE. Studies evaluating a posible association between the use of oral contraceptives and the onset of SLE[42-44] were negative; one study reported only a weak association. A variety of hormonal factors have been associated with severity and progression of SLE but few associations have been observed with the onset of SLE.[19*] However, a strong association has been noted in observational studies between the use of hormone replacement therapy and the onset of SLE,[45-47] not observed in the CLU study.[19*]
The lack of expected hormonal associations observed in the CLU study should be confirmed by exploration of other study samples. Differences in control of potential confounding measures, age at menopause, and in the types of replacement therapies available for evaluation may account for the inconsistency in results. Furthermore, the CLU study had limited sample size in specific subgroups by type of hormone replacement therapy, duration of its use, and among former users. A large sample that focuses on post-menopausal women and collects detailed and validated use of hormone replacement therapies will be needed. Such a study of women with established SLE (SELENA study) has recently completed enrollment.
The strongest hormonal-related associations seen in the CLU study were that a history of breast-feeding was associated with a decreased risk of developing SLE, while a history of pre-eclampsia was associated with an increased risk of developing SLE.[19*] These provocative results have not been previously observed in SLE, but have been reported in RA.
Exposure to occupational crystalline silica has been identified as a strong risk factor for developing autoimmune disease[2*] and was a strong risk factor associated with SLE in the CLU study.[20*] The CLU study provided rigorous attention to the definitions and assignments of silica exposure. Occupational silica exposure is most frequently associated with dusty, nonfarming related jobs. This study included an assessment of these types of occupational exposures by industrial hygienists, but also included a detailed assessment of farm-related silica exposures using both a dust-exposure matrix for specific farming tasks and soil system maps for specific farm locations. In previous studies of silica exposure and autoimmune disease, mostly men have been studied, and farming exposures have not been considered. Farming exposures were of particular importance in this eastern North Carolina and South Carolina study location. Furthermore, the farming exposures accounted for much of the silica exposure among women, who often worked in part-time or seasonal farming jobs. In a disease such as SLE that predominantly affects females, it is critical to pay careful attention to variations in opportunities for occupational exposures. The CLU study has set a new precedent for careful attention to issues that are critical to understanding risk factors for women in any disease setting.
It has previously been observed that stress from major life events may contribute to the onset of RA. The recent case-control study of RA provides several new insights broadening the scope of psychosocial stress that may contribute to the onset of RA.[16*] In women, "matrimonial quarrels" during the 5-year period preceding the RA symptoms and the 5 years preceding the diagnosis of RA were significantly associated with the onset of RA compared with controls. In men "problems at work" preceding the onset of symptoms and "economic problems" within 5 years of diagnosis of RA were also associated with the onset of RA.
Disparate social environmental factors and medical status have been associated with the onset of autoimmune disorders such as type I diabetes. Psychological mechanisms are directly linked by hormonal and nervous system signals, influencing the need for insulin. Stress has also been shown to modulate immune responses. A population-based study in Sweden recruited 78% of all prospective parents over a 2-year period with 97% completing 1-year follow up data on 4337 children. Evaluation of the potential association of stressors in this population with subsequent development of diabetes in the offspring awaits long-term follow up.
There is currently considerable interest in the increasing number of pharmaceutical agents, including newer biologic agents, which have been associated with the development of autoantibodies and drug-related SLE. Two recent epidemics of scleroderma-like illness, one in Spain to a contaminant of rapeseed oil and one in the United States to a contaminant of l-tryptophan, have raised interest in potential environmental triggers of autoimmunity. Many patients receiving procainamide (PA) develop autoantibodies that may persist after discontinuation of the drug; most of these patients do not develop drug-induced SLE. Slow acetylator status correlated with IgG antibodies to H2A-2B but was not a risk factor for developing PA-related lupus. Case reports increasingly note drug-induced SLE as an adverse event after therapy with TNF inhibitors, g interferon, and other biologics employed as therapeutic agents.
Risk Factors for Disease Progression: Developing Lupus Nephritis in SLE
The LUMINA study, a multiethnic study designed to evaluate outcomes among SLE patients, has recently evaluated predictors for developing lupus nephritis.[52, 53] The high morbidity and mortality associated with lupus nephritis and complications from the use of immunosuppressive therapies make this a critical area to study. The study is strengthened by the inclusion of a broad range of potential predictive variables including sociodemographic information, health behavioral information, clinical information, immunologic and immunogenetic measures, and adequate representation of African American patients (42%) and Hispanic patients (27%).
Both African American and Hispanic patients were more likely to develop lupus nephritis compared with Caucasian patients, even when controlling for socioeconomic and other differences among the races. Increased incidence, prevalence, and severity of lupus nephritis between African Americans and Caucasians is in agreement with other reports;[54-60*] the study confirms increased prevalence and severity of lupus nephritis among Hispanics. The LUMINA study is the first to document a higher incidence of lupus nephritis among Hispanics compared with Caucasians in the United States. Additionally, patients who were not married or living together were also more likely to develop lupus nephritis. Measures of race and other sociodemographics are important in identifying high-risk groups, which is particularly imperative for focusing resources toward adequate screening and intervention. These measures may also be surrogate markers for unidentified settings and exposures that contribute to the development of lupus nephritis. Research identifying components contributing to the differences observed by race and sociodemographics will provide further insights into causes of renal progression.
Other features predictive of lupus nephritis among LUMINA patients included higher baseline activity scores across all organ systems (SLAM score) and the presence of anti-dsDNA and anti-RNP antibodies. Overall, these findings are consistent with the previous literature in this area, though measures previously associated with the development of lupus nephritis (history of fetal losses, baseline presence of hypertension, thrombocytopenia, leukopenia, anti-RNP antibodies, and anti-phospholipid antibodies) were not increased.
As more information is published on both risk factors for the onset of autoimmunity and for the development of progressive disease, it is becoming apparent that factors that initiate or promote the disease may not be the factors that influence the severity or progression of the disease. Many potential exposures have not been thoroughly or consistently studied for their impact on both disease onset and disease progression. Currently, the impact of silica and other exposures that were found to be important in the onset of lupus are being studied as risk factors for the development of lupus nephritis in a follow-up study of the CLU study. Hopefully, more of these types of studies will identify and resolve different risk factors for onset and progression of SLE. Better understanding of these risk factors will likely lead to a better understanding of mechanisms for the onset and progression of SLE as well as to more successful interventions targeted to specific high-risk groups.
Correspondence to M.A. Dooley, MD, MPH, Division of Rheumatology and Immunology, University of North Carolina at Chapel Hill, CG#7280 3330 Thurston Boulevard, Chapel Hill, North Carolina, 27599, USA
CLUC, Carolina Lupus Study; RA, rheumatoid arthritis; SLE, systemic lupus erythematosus
Papers of particular interest, published within the annual period of
review, have been highlighted as:
* Of special interest
** Of outstanding interest
© 2003 Lippincott Williams & Wilkins