Psoriasis is a systemic disease. Whether clearing the skin protects the heart is unsettled

Psoriasis is not confined to the skin. About a fifth of people with it develop psoriatic arthritis, and the disease travels with an elevated burden of cardiovascular disease that has been documented for years [s4] [s2]. What remains genuinely unresolved is the next step — whether suppressing the inflammation with modern biologic drugs reduces cardiovascular events. Three separate 2025 analyses of largely the same drugs reached three different answers, and the reason they disagree is instructive.

The joint disease

A meta-analysis of 266 studies covering 976,408 patients with psoriasis found a pooled prevalence of psoriatic arthritis of 19.7% (95% confidence interval 18.5% to 20.9%) [s4]. Prevalence varied by region — 22.7% in European patients, 21.5% in South American, 19.5% in North American, 14.0% in Asian — and reached 23.8% in the studies that applied the formal classification criteria [s4]. In children and adolescents under 18 the pooled prevalence was 3.3% [s4]. Incidence among people with psoriasis ranged from 0.27 to 2.7 per 100 person-years [s4].

A discrepancy in the source: the paper's stated pooled estimate is 19.7%, roughly one in five, while its conclusion states that one in four patients with psoriasis have psoriatic arthritis. The one-in-four figure matches the 23.8% subgroup analysed under the classification criteria rather than the overall pooled estimate.

The prevalence of psoriasis itself is reported inconsistently even between two papers in the same journal in the same year: one gives 0.14% to 1.99% globally [s3], the other approximately 2% to 3% of the world population [s2].

Three answers to the cardiovascular question

A matched cohort says biologics look protective. A retrospective cohort study using a global federated health-records network compared psoriasis patients newly prescribed biologics with patients newly starting oral anti-psoriatic drugs, propensity-matched on age, sex, race, comorbidities, body mass index, lipids and inflammatory markers [s1]. Each arm held 12,732 patients, about half female, mean age 57 [s1]. The five-year cumulative incidence of any cardiovascular disease was 10.68% in the biologic cohort against 16.17% in the non-biologic cohort [s1]. Hazard ratios favoured the biologic cohort across nearly every outcome measured: any cardiovascular disease 0.621 (0.571 to 0.676), major adverse cardiac events 0.697 (0.614 to 0.792), ischaemic heart disease 0.579, heart failure 0.637, peripheral arterial occlusive disease 0.501 [s1]. By class, anti-TNF (0.886), anti-IL-17 (0.724) and anti-IL-23 (0.739) were each associated with reduced risk, while anti-IL-12/23 alone was not (0.915, 0.742 to 1.128) [s1].

The authors name the limitation themselves: the design is observational and can establish association, not causation [s1]. Patients started on biologics differ from patients kept on oral therapy in ways propensity matching cannot fully capture — including, plausibly, in how closely they are monitored.

A meta-analysis of randomised trials says nothing is happening. Pooling 43 randomised controlled trials reported across 36 papers, an updated review found no statistically significant difference in the risk of major adverse cardiovascular events between biologic therapies and placebo, a Peto odds ratio of 1.26 (95% CI 0.53 to 3.01, p = 0.59) [s2]. Nor did any class separate: TNF-alpha inhibitors 1.13 (0.29 to 4.32), IL-17 inhibitors 0.60 (0.16 to 2.25), IL-12/23 inhibition 3.80 (0.37 to 39.44), IL-23 inhibitors 1.75 (0.25 to 12.43) [s2]. Every one of those intervals is enormous, which is the finding: short placebo-controlled trials in a relatively young population accumulate too few cardiac events to detect anything [s2]. The authors conclude that longer studies and post-marketing surveillance are needed [s2].

A network meta-analysis flags one drug in the opposite direction. Including 68 randomised trials with 34,414 patients, a network meta-analysis of cardiovascular and kidney outcomes found bimekizumab both the top-ranked treatment for skin clearance (odds ratio for 75% improvement 101.12, 95% CI 34.26 to 301.46, high certainty) and associated with reduced total cardiovascular events (0.06, 0 to 0.80, moderate certainty) [s3]. Ixekizumab, by contrast, showed strong skin efficacy (odds ratio for 90% improvement 86.92, 39.06 to 199.66, high certainty) but was associated with increased major adverse cardiovascular events against placebo (3.26, 1.26 to 9.31, high certainty) and against bimekizumab (31.92, with a credible interval running from 2.01 to 1123.25) [s3]. Renal outcomes were similar across all groups [s3].

That last interval — 2.01 to 1123.25 — is the honest measure of how much information sits behind the comparison. A point estimate flanked by a bound above one thousand is a statement that the number of events was very small.

Why the three disagree

They are not measuring the same thing. The cohort study observes real-world patients over five years and captures many events, but cannot rule out that healthier or better-managed patients are the ones who get biologics. The trial meta-analyses observe randomised comparisons that eliminate that confounding but run for months rather than years and record too few events to be conclusive in either direction. The network meta-analysis adds indirect comparisons between drugs never tested against each other, which widens intervals further.

Two of the three analyses are by the same journal, published three months apart, and reach opposite conclusions about IL-17 inhibitors as a class — no signal in one [s2], a specific high-certainty harm signal for one member of the class in the other [s3]. Neither is obviously wrong. They are different questions asked of overlapping data.

What this leaves a reader with

Established: psoriasis carries psoriatic arthritis in about a fifth of patients, and is associated with cardiovascular disease [s4] [s2]. Unestablished: whether treating it with any particular drug class changes cardiovascular outcomes. The strongest-looking evidence for benefit is observational; the strongest-looking evidence for harm from one agent rests on very few events; and the randomised evidence that would settle it, pooled across 43 trials, is too imprecise to be informative, its interval running from 0.53 to 3.01 [s2].

Anyone told that a psoriasis biologic will protect their heart is being told something the evidence does not yet support, in either direction.

This article is informational and is not medical advice.

Sources

• [s1] Cardiovascular disease risk in patients with psoriasis receiving biologics targeting TNF-α, IL-12/23, IL-17, and IL-23: A population-based cohort study — PLOS Medicine, 17 April 2025. https://doi.org/10.1371/journal.pmed.1004591
• [s2] Biologic Therapies and Major Cardiovascular Events in Psoriasis: Updated Systematic Review and Meta-analysis — Dermatology and Therapy, 27 October 2025. https://doi.org/10.1007/s13555-025-01529-5
• [s3] Cardiovascular and Kidney Outcomes After Systemic Treatment for Plaque Psoriasis: A Systematic Review and Network Meta-analysis — Dermatology and Therapy, 5 July 2025. https://doi.org/10.1007/s13555-025-01472-5
• [s4] Prevalence of psoriatic arthritis in patients with psoriasis: A systematic review and meta-analysis of observational and clinical studies — Journal of the American Academy of Dermatology, 19 June 2018. https://doi.org/10.1016/j.jaad.2018.06.027