The association between silica exposure and development of ANCA-associated vasculitis: systematic review and meta-analysis

Abstract

Background: Crystalline silica is among the environmental exposures associated with increased risk of autoimmune diseases, including rheumatoid arthritis, systemic sclerosis and systemic lupus erythematosus. Silica exposure has also been related to the development of ANCA-associated vasculitides (AAV), but past studies appear to conflict as to the presence and magnitude of the associated risks of disease. We aimed to conduct a systematic review of the existing studies and meta-analysis of their results.

Methods: We searched EMBASE, MEDLINE and international scientific conference abstract databases for studies examining the association of silica exposure with AAV. Studies in English, French, or Spanish were included and those examining the association of silica with ANCA-positivity alone were excluded. We assessed study quality using the Newcastle-Ottawa Scale. We meta-analyzed the results using random effects models and tested for heterogeneity. We performed sensitivity and subgroup analyses, examining studies that adjusted for smoking and occupational risk factors as well as studies that analyzed by subtypes of AAV.

Results: We identified 158 potential manuscripts and 3 abstracts related to silica exposure and risk of AAV. 147 were excluded after abstract review and 14 underwent detailed evaluation of full manuscript/abstract. After further application of exclusion criteria, 6 studies (all cases-controls) remained. The studies had moderate heterogeneity in selection of cases and controls, exposure assessment, disease definition and controlling for potential confounders. We found an overall significant summary effect estimate of silica "ever exposure" with development of AAV (summary OR 2.56, 95% CI 1.51-4.36), with moderate heterogeneity (I(2)=48.40%). ORs were similar for studies examining only MPA (OR 3.95, CI 95% 1.89-8.24), compared to those only studying GPA (OR 3.56, CI 95% 1.85-6.82).

Conclusion: Despite moderate heterogeneity among studies, the totality of the evidence after meta-analysis points to an association between silica exposure and risk for developing AAV.

Keywords: ANCA-associated vasculitides; Antineutrophil cytoplasmic antibodies; Meta-analysis; Silica.

Figure 1

Overview of Literature Search

Figure 2

Random effects analysis of all studies for the association between ANCA-associated vasculitis with prior exposure to silica. Point (square) and overall (diamond) estimates are given as odds ratios with 95% confidence interval (CI) (horizontal bar).

Figure 3A

Random effects analysis of all studies for the association between ANCA-associated vasculitis with prior exposure to silica grouped according to studies focused on renal involvement vs studies without differentiation between renal and other organ involvement (label as systemic). Point (square) and overall (diamond) estimates are given as odds ratios with 95% confidence interval (CI) (horizontal bar).

Figure 3B

Random effects analysis of all studies for the association between ANCA-associated vasculitis with prior exposure to silica grouped according to studies focused in MPA and GPA. Point (square) and overall (diamond) estimates are given as odds ratios with 95% confidence interval (CI) (horizontal bar).

Figure 3C

Random effects analysis of all studies for the association between ANCA-associated vasculitis with prior exposure to silica grouped according to studies that are adjusted and unadjusted. Point (square) and overall (diamond) estimates are given as odds ratios with 95% confidence interval (CI) (horizontal bar).

Figure 4

Begg's funnel plot for detection of publication bias. The y-axis (standard error) measures precision of the results of the studies. The x-axis (log-odds ratio) measures the odds ratio of each study included in the meta-analysis. Therefore, smaller, less precise studies will likely have higher standard error and likely be more skewed from the center where larger, more precise studies lie. Therefore, asymmetry of the funnel plot suggests publication bias [33], [34],[35].

Appendix Figure 1

Random effects analysis of all studies for the association between ANCA-associated vasculitis with prior exposure to silica with jack knife sensitivity analysis. Point (square) and overall (diamond) estimates are given as odds ratios with 95% confidence interval (CI) (horizontal bar).