Multiple long-term conditions in people with psoriasis: a latent class and bidirectional Mendelian randomization analysis

Abstract

Abstract Background Coexisting long-term conditions (LTCs) in psoriasis and their potential causal associations with the disease are not well ­established. Objectives To determine distinct clusters of LTCs in people with psoriasis and the potential bidirectional causal association between these LTCs and psoriasis. Methods Using latent class analysis, cross-sectional data from people with psoriasis from the UK Biobank were analysed to identify distinct psoriasis-related comorbidity profiles. Linkage disequilibrium score regression (LDSR) was applied to compute the genetic correlation between psoriasis and LTCs. Two-sample bidirectional Mendelian randomization (MR) analysis assessed the potential causal direction using independent genetic variants that reached genome-wide significance (P < 5 × 10–8). Results Five comorbidity clusters were identified in a population of 10 873 people with psoriasis. LDSR revealed that psoriasis was positively genetically correlated with heart failure [genetic correlation (rg) = 0.23, P = 8.8 × 10−8], depression (rg = 0.12, P = 2.7 × 10−5), coronary artery disease (CAD; rg = 0.15, P = 2 × 10−4) and type 2 diabetes (rg = 0.19, P = 3 × 10−3). Genetic liability to CAD was associated with an increased risk of psoriasis [inverse variance weighted (IVW) odds ratio (ORIVW) 1.159, 95% confidence interval (CI) 1.055–1.274; P = 2 × 10−3]. The MR pleiotropy residual sum and outlier (MR-PRESSO; ORMR-PRESSO 1.13, 95% CI 1.042–1.228; P = 6 × 10−3) and the MR-robust adjusted profile score (RAPS) (ORMR-RAPS 1.149, 95% CI 1.062–1.242; P = 5 × 10−4) approaches corroborate the IVW findings. The weighted median (WM) generated similar and consistent effect estimates but was not statistically significant (ORWM 1.076, 95% CI 0.949–1.221; P = 0.25). Evidence for a suggestive increased risk was detected for CAD (ORIVW 1.031, 95% CI 1.003–1.059; P = 0.03) and heart failure (ORIVW 1.019, 95% CI 1.005–1.033; P = 9 × 10−3) in those with a genetic liability to psoriasis; however, MR sensitivity analyses did not reach statistical significance. Conclusions Five distinct clusters of psoriasis comorbidities were observed with these findings to offer opportunities for an integrated approach to comorbidity prevention and treatment. Coexisting LTCs share with psoriasis common genetic and nongenetic risk factors, and aggressive lifestyle modification in these people is anticipated to have an impact beyond psoriasis risk. Genetically predicted CAD is possibly associated with an increased risk of psoriasis, altering our prior knowledge.