Need for longitudinal studies to assess the real-world effectiveness of allergy immunotherapy in patients with allergic rhinitis and asthma.

Using a large administrative claims database, Fritzsching et al have evaluated the real-world effectiveness of allergen-specific immunotherapy (AIT) and concluded that treatment was associated with reductions in allergic rhinitis (AR) and asthma prescriptions and a decrease in healthcare utilization compared with a matched cohort of non-AIT controls. However, the data source, study design, and statistical analyses used limit the strength of conclusions and magnitude of effects presented.

A ‘fit for purpose’ database is essential for conducting a valid real-world effectiveness analysis. Consistent with AR management guidelines and as part of standard of care, patients use a wide array of allergy medications prior to and after AIT initiation. Over the past decade, a vast majority of these medications have become available over-the-counter (OTC). The annual number of anti-histamine and intranasal corticosteroid fills in the 12-months prior to AIT initiation may be lower than expected and the marked reductions in AR fills among control patients may reflect an overall transition to OTC medications. Therefore, an evaluation of AR prescription use in a claims database while ignoring OTC medication use is somewhat biased, leaving the real-world impact of AIT on the reduction of AR medication use unanswered.

Diagnosis of allergic disease using ICD-10 codes without the associated allergic clinical history or sensitization status poses challenges (eg. vasomotor rhinitis [J30.0; non-allergic rhinitis] is reported with the highest frequency among rhinitis causes). Patients with AR may be treated by primary-care physicians (PCP) whereas AIT is administered by specialists with training in allergy immunology requiring frequent office visits. Authors provide an argument that AIT subjects were more likely to get AR prescriptions in Year 1 but a consistent decrease thereafter. However, systemic differences in the care provided by a specialist compared with a PCP (ie. close follow-up, optimization of care-pathways, adherence, etc.) can also improve outcomes.

AIT is typically indicated for patients with moderate-severe uncontrolled AR. While propensity score matching, to identify matched controls, adjusts for several variables captured in the database, the lack of data on allergen sensitization, symptoms, and disease severity/control may lead to unbalanced cases and controls. An ideal comparison group would be patients with comparable disease characteristics and severity who have not initiated or completed AIT; however, this was not feasible and the study conclusions do not consider these limitations.

Large sample sizes in each cohort (n = 25,488) may drive statistical significance and no multiplicity control is described. For example, while statistically significant, the reduction in AR fills from baseline to Year 4 in the AIT case group (0.526 fills/year) and control group (0.418 fills/year) suggests that the difference in AR fills between the AIT case and control groups is 0.526-0.418 = 0.1 AR fills per year, which may not be clinically meaningful; similar concerns are present with the low effect sizes for asthma outcomes.

The analysis method does not entirely account for differences in follow-up and the authors report outcomes separately for each year, leading to sample size/study cohort changes in each of these analyses. Reporting the difference in outcomes between two time points, ie. by comparing data from Year 4 with the year prior to AIT initiation, is not an optimal method of conducting longitudinal data analysis​; alternative methods such as a negative binomial or Poisson regression accounting for the variable follow-up duration of individual patients may be more appropriate.

The study results using a claims-based dataset to evaluate outcomes in AR may be inherently biased to evaluate disease-related medication use. Regardless, AIT is an important disease modifying therapeutic option for moderate-severe AR/allergic rhinoconjunctivitis and longitudinal studies with better methodologies are required for a broader exploration of the long-term real-world impact.

Contributors

All authors contributed equally. All authors contributed on conceptualization, data interpretation, writing of the letter, and approved the letter for submission.

Declaration of interests

All authors are employees of and stockholders in Regeneron Pharmaceuticals, Inc.