Design: Meta-analysis of 8 observational studies with reconstructed time-to-event data
Population: 4516 patients on omalizumab for chronic urticaria
Primary outcome: Drug survival and reasons for discontinuation
Effect: Median drug survival 3.1 years (95% CI 2.8 to 3.4); autoimmunity HR 2.03 for stopping due to inefficacy

A meta-analysis in JAMA Dermatology reconstructed patient-level data from Kaplan–Meier curves in eight observational studies, covering 4516 people on omalizumab for chronic urticaria.

Median drug survival was 3.1 years. Survival at seven years was higher in chronic inducible urticaria (43–49%) than in chronic spontaneous urticaria alone (30%). Early stopping was mainly because the disease was well controlled; stopping for adverse events was uncommon. Autoimmune comorbidity, mostly thyroid disease, was associated with about twice the risk of stopping for lack of effect, whereas an atopic background was associated with stopping because the urticaria had settled.

In practice this supports trying to step down omalizumab when control is good, and suggests a lower threshold for considering next-line treatment in patients with autoimmune features who respond poorly. In India, cost is often the deciding factor in how long omalizumab is continued, which these largely European and American data do not address.

Check thyroid function and thyroid antibodies in chronic spontaneous urticaria; autoimmune features predicted poorer omalizumab response.
Reassess control regularly on omalizumab and consider stepping down or stopping when urticaria is well controlled.
Consider earlier escalation to next-line options in patients with autoimmune comorbidity who respond poorly.
Reassure patients that stopping for side effects was uncommon in these studies.

Why it matters

Knowing who is likely to stop for failure helps decide when to move to the newer agents now licensed for urticaria.

The statistics, in plain English

A hazard ratio of 2.03 (95% CI 1.20 to 3.41) means patients with autoimmune comorbidity stopped for lack of effect about twice as quickly. The data were reconstructed from published curves, which is validated but less precise than original patient records.