Causal relationship between spondylarthritis and stroke in a European population: a two sample Mendelian randomization study

Abstract

Background: Observational studies have found an increased risk of stroke in patients with spondyloarthritis, but the results are susceptible to reverse causality and confounders. Therefore, the study aimed to further explore the association between spondyloarthritis and different subtypes of stroke by using a two sample Mendelian randomization (MR) analysis.

Methods: Genetic instrumental variables for spondyloarthritis were identified using summary level data from a genome-wide association study involving 201,581 people. Summary statistics from the Multiancestry Genome-wide Association Study of Stroke Consortium were used to obtain genetic data on stroke. There was no sample overlap between the exposure and outcome datasets. Inverse-variance weighted was considered the primary MR method for causal analysis. Heterogeneity, pleiotropy and sensitivity analyses were performed to ensure robustness, and single nucleotide polymorphism (SNP) with potential confounders was further screened in the PhenoScanner database to better evaluate the stability of our study.

Results: One SNP (rs1065045) was excluded due to schizophrenia. After excluding SNP (rs1065045), results of the second MR analysis were slightly different from the first, which were considered as the final result: a significant positive causality between spondyloarthritis and cardioembolic stroke (OR=1.296, 95% CI:1.094-1.534, p=0.003); a possible positive causality between spondyloarthritis and any stroke (OR=1.082, 95% CI:1.016-1.152, p=0.013)/any ischemic stroke (OR=1.086, 95% CI:1.013-1.163, p=0.020); no significant/possible causality between spondyloarthritis and small vessel stroke (OR=1.168, 95% CI:0.993-1.375, p=0.061). Insufficient power may be one possible reason why a causality was not observed between spondyloarthritis in our study.

Conclusions: This study suggests that the possible causative effects of spondyloarthritis predicted by genetics on stroke may be limited to any stroke, any ischemic stroke, and cardioembolic stroke, especially the last.

Keywords: Mendelian randomization; cardioembolic stroke; causality; spondyloarthritis; stroke.

Figure 1

Directed acyclic graphs illustrating causal pathway from genetic variants to outcomes. Genetic variants influence the exposure, which has downstream effect on a related variable which does not affect the outcome. Related variables may be known or unknown, with known related variables referring to variables that have been reported to be associated with exposure but do not affect the outcome, and unknown related variables referring to variables that have not been reported to be associated with exposure but do not affect the outcome. There is no alternative pathway from the genetic variants to the outcome, the instrumental variable assumptions are satisfied.

Figure 2

Scatter plots of the causality between exposure and outcome by five MR analyses. (A) Scatter plots of the causality between spondyloarthritis and AS by five MR analyses; (B) Scatter plots of the causality between spondyloarthritis and AIS by five MR analyses; (C) Scatter plots of the causality between spondyloarthritis and LAS by five MR analyses; (D) Scatter plots of the causality between spondyloarthritis and SVS by five MR analyses; (E) Scatter plots of the causality between spondyloarthritis and CES by five MR analyses.

Figure 3

The leave-one-out sensitivity plots between exposure and outcomes. (A) The leave-one-out sensitivity plot between spondyloarthritis and AS; (B) The leave-one-out sensitivity plot between spondyloarthritis and AIS; (C) The leave-one-out sensitivity plot between spondyloarthritis and LAS; (D) The leave-one-out sensitivity plot between spondyloarthritis and SVS; (E) The leave-one-out sensitivity plot between spondyloarthritis and CES.