Safety Profile of Ustekinumab

Safety considerations for ustekinumab, including infection risk and malignancy signals.

Overview

Surveyed ustekinumab safety literature and extracted findings from 62 sources. Across pooled trial data, registries, and pharmacovigilance, ustekinumab's safety profile looks favorable relative to other biologics — the load-bearing signals are infections (mostly unremarkable, with rare opportunistic cases) and a contested cardiovascular signal in patients initiating treatment at high baseline risk. Malignancy data are reassuring in trials but limited in duration.

Infections — generally low, with caveats

The strongest dataset is the final pooled IBD safety analysis (2,575 ustekinumab-treated patients, 4,826 patient-years, through 5 years in Crohn's and 4 years in UC), which reports rates of serious infections, opportunistic infections including TB, and malignancies that were similar to or lower than placebo (Ghosh et al. 2024). In PSOLAR, the largest real-world psoriasis registry (12,093 patients, 40,388 patient-years), unadjusted serious infection rates were 0.93/100 PY for ustekinumab vs 2.91 for infliximab and 1.91 for other biologics; exposure to biologics other than ustekinumab was associated with serious infection (HR 1.96, p<0.001), while ustekinumab was not (Papp et al. 2015).

For tuberculosis and other opportunistic infections specifically, the cross-indication pooled analysis (13,807 patient-years) found OI rates of 0.10/100 PY on ustekinumab vs 0.40 on placebo, with only one active TB case on drug — most OIs occurred in IBD patients on concomitant immunosuppression (Long et al. 2022). Rare severe opportunistic cases do occur: there's a published case of disseminated Nocardia farcinica infection (bacteremia, renal abscess, cerebral lesions) attributed to ustekinumab (Couture-Cossette et al. 2019), which is mechanistically plausible since IL-12/23p40 blockade overlaps with the pathway implicated in Mendelian Susceptibility to Mycobacterial Disease (Couture-Cossette et al. 2019).

Malignancy — no clear signal in trials, watch real-world data

A meta-analysis of 15 RCTs (5,835 patients) found no significant difference in malignant tumors (POR 0.88, 95% CI 0.37–2.11) or non-melanoma skin cancer (POR 0.77, 95% CI 0.23–2.61) vs placebo (Huang et al. 2022). PSOLAR similarly showed no biologic — including ustekinumab — was associated with increased malignancy risk (Papp et al. 2015). The pooled IBD analysis reported no lymphomas through 5 years (Ghosh et al. 2024).

The main caveat is duration: most RCT follow-up is short (the meta-analysis median was 16 weeks), trial populations are selected, and a FAERS disproportionality analysis has flagged signals for several cancers — but FAERS cannot establish causality and the signal is most plausibly confounded by indication and prior immunosuppression.

Cardiovascular — the most contested signal

A French case-time-control study using national health insurance data (9,290 ustekinumab-exposed patients, 179 SCEs) found that among patients with high baseline cardiovascular risk, ustekinumab initiation was associated with severe cardiovascular events (OR 4.17, 95% CI 1.19–14.59), with no signal in low-risk patients (OR 0.30, 0.03–3.13) (Poizeau et al. 2020). A more recent Swedish national-register cohort (15,502 patients, 852 on ustekinumab) found no difference in MACE versus adalimumab, etanercept, or secukinumab (HRs all near 1.0) (Banefelt et al. 2025). The pooled IBD analysis likewise reports MACE rates similar to placebo (Ghosh et al. 2024). The most defensible reading is that ustekinumab does not raise CV risk in the general treated population, but the early-initiation high-risk signal hasn't been fully explained away.

What's missing

Comparative head-to-head safety vs IL-23-specific inhibitors (risankizumab, guselkumab) is thin — one small propensity-matched IBD comparison showed no significant 5-year differences in cancer, thromboembolism, or hospitalization. Pediatric data are limited but a US claims study in children with psoriasis showed lower outpatient infection rates vs etanercept (ARR 0.58) and methotrexate (0.66). Pregnancy outcome data remain sparse.