The environment and Systemic Autoimmune Rheumatic Diseases (SARDS): What do we know after more than 15 years of research?

Systemic autoimmune rheumatic diseases (SARDs) are a group of systematic inflammatory conditions characterized by self-directed inflammation and a host of rheumatic manifestations.¹ Although prevalence estimates vary widely due to case definitions and the specific outcomes included in the estimates, some studies have suggested that over 5 million individuals in the United States alone may live with SARDs.² Primary prevention at a population level is crucial if we are to reduce the high medical and quality of life burden of these diseases.

The etiology of these heterogeneous conditions remains mostly unknown. For over fifteen years,³ there has been intensifying interest in examining environmental exposures, especially air pollution, as population-level risk factors for SARDs, based on observed geographic differences in risk and previously identified associations between SARDs and cigarette smoking. Early on, the majority of these studies focused on associations between air pollution and rheumatoid arthritis, but the field rapidly expanded to include studies of other SARDs.

In this issue of Arthritis & Rheumatology, Geslin et al report the results of a large Canadian study exploring the association between particulate matter (PM) air pollution and incidence of SARDs in a dataset of over 7.4 million adults in Quebec.⁴ They used the Quebec Integrated Chronic Disease Surveillance System (QICDSS) to assemble an open cohort of 7,482,397 residents, followed for over 98 million person-years from 2000 to 2019. New cases of SARDs were identified based on two or more physician claims more than eight weeks apart but within two years or a single hospitalization and included International Classification of Diseases (ICD-10) codes for systemic lupus erythematosus, systemic scleroderma, Sjögren syndrome, polymyositis, dermatomyositis, and undifferentiated connective tissue disease. Time-varying annual average exposures were based on the residential postal code of each cohort member throughout follow-up. Annual average concentrations of particulate matter less than 2.5 microns in diameter (PM_2.5), based on a 1 km grid and seven major PM_2.5 components (black carbon, sulfate, ammonium, nitrate, organic matter, sea salt, and mineral dust) were assigned to each residential postal code to define exposures. For participants who moved during the study period, annual exposure was weighted by time living at each postal code. Although only age and sex were available from the QICDSS, the authors adjusted for area level socioeconomic status at a small scale and Local Service Network region as a measure of local health service availability. The authors used standard Cox proportional hazards models to assess the risk of SARDs associated with each exposure, then assessed deviations of linearity, and used quantile g-computation to assess the impact of each decile increase in all seven major PM_2.5 components combined.

Over 55,000 cases of SARDs were identified during follow-up. In adjusted single-exposure models, each interquartile range increase in PM_2.5 was associated with a Hazard Ratio (HR) for SARDs of 1.036 (95% Confidence Interval (CI): 1.002, 1.070), with positive associations also observed for nitrate and organic matter. In models assessing the association with a decile increase in all components, the HR was 1.012 (95%CI: 1.002, 1.021), with the largest contributions from black carbon, ammonium, sulfate, and dust. For the two most common outcomes, systemic lupus erythematosus and Sjögren’s syndrome, the associations with PM_2.5 were 1.082 (95%CI: 0.986, 1.189) and 1.414 (95%CI: 1.003, 1.298), models for the components or component mixture were not presented.

Overall, the analyses presented in this manuscript make a number of important advancements. First, the use of a comprehensive population-level dataset ensures that the results from this study are representative of the full population of Quebec, avoiding concerns of generalizability that may arise from more selected study populations. Additionally, by assessing the impact of individual PM_2.5 components, in addition to the commonly examined PM_2.5 mass, the authors have provided more refined information on the potential sources of air pollution that may underlie the observed associations. This is a first step in also identifying targets where reductions could be the most beneficial for reducing the burden of SARDs in the population. However, there are also limitations to this analysis, and many others like it in the field. Without more detailed information on health behaviors, it is challenging to be confident that there is no residual confounding in the presented estimates by factors such as time spent outside, smoking, vaping, exercising and healthy diet, for example, and effect modification to identify population groups at the greatest risk could not be conducted.

Although great strides have been made worldwide in reducing air pollution, there is still substantial work to be done. In a recent analysis, the World Health Organization estimated that more than 99% of the world’s population lived in areas that did not meet their recommendations for target levels of PM_2.5 (5μg/m³).⁵ There is likely no level of PM_2.5 that is safe for human health, as illustrated in the linear associations presented in the splines from this paper, because Canada has very low levels of PM_2.5 (average 8.02, max 15.1 μg/m³) compared to most areas of the world. These decades long improvements in exposures are under attack in countries such as the United States, where rollbacks to environmental regulations responsible for decreasing exposure levels are taking effect. Moreover, the sources of air pollution that are hazardous to human health and related to risk of SARDs, including fire smoke, are growing.⁶ As this article and the literature suggest, we anticipate that these rollbacks will lead to increases in the incidence of SARDs, at a great cost to society and population health.

More optimistically, however, this work points to how environmental epidemiology can provide insights into population-level risk factors for SARDs. In addition to air pollution, there are a large number of environmental exposures that could have plausible relationships with risk of SARDs. Using the mixture methods utilized here, and other related methods being developed at a rapid pace,⁷ we should be able to more comprehensively investigate the impacts of multiple environmental exposures, referred to now collectively as the exposome, in influencing the risk of SARDs.⁸ This is vital as there are likely unexplored interactions between exposures that may be synergistic or antagonistic, and the relationships between these exposures may be in flux due to climate change.

Now, at more than fifteen years into the study of environmental factors in relation to the risk of SARDs, we look into our crystal (the traditional fifteenth anniversary gift) balls to see what findings the next fifteen years may yield. Given the strong foundation of current science, it is time to expand past the standard air pollutants and examine other exposures and their interactions with each other and other known risk factors. What we cannot yet quantify is how much of the risk of SARDs may be explained by environmental exposures, which of the exposures may interact to pose the greatest risk, and which subgroups of the population may bear the greatest risks from these exposures. Most importantly, we must remain firm against those who would call into question these findings. We must also stand with those scientists who will be targeted for doing this work, as we cannot afford to slide backwards and allow PM levels to grow to dangerous levels and other potentially important population-level risk factors to go unnoticed.