A meta-analysis of sublingual allergen immunotherapy and pharmacotherapy in pollen-induced seasonal allergic rhinoconjunctivitis

Abstract

Background: The capacity of sublingual allergen immunotherapy (SLIT) to provide effective symptom relief in pollen-induced seasonal allergic rhinitis is often questioned, despite evidence of clinical efficacy from meta-analyses and well-powered, double-blind, placebo-controlled randomized clinical trials. In the absence of direct, head-to-head, comparative trials of SLIT and symptomatic medication, only indirect comparisons are possible.

Methods: We performed a meta-analysis of classes of products (second-generation H1-antihistamines, nasal corticosteroids and grass pollen SLIT tablet formulations) and single products (the azelastine-fluticasone combination MP29-02, and the leukotriene receptor antagonist montelukast) for the treatment of seasonal allergic rhinitis in adults, adolescents and/or children. We searched the literature for large (n >100 in the smallest treatment arm) double-blind, placebo-controlled randomized clinical trials. For each drug or drug class, we performed a meta-analysis of the effect on symptom scores. For each selected trial, we calculated the relative clinical impact (according to a previously published method) on the basis of the reported post-treatment or season-long nasal or total symptom scores: 100 × (scorePlacebo - scoreActive)/scorePlacebo.

Results: Twenty-eight publications on symptomatic medication trials and ten on SLIT trials met our selection criteria (total number of patients: n = 21,223). The Hedges' g values from the meta-analyses confirmed the presence of a treatment effect for all drug classes. In an indirect comparison, the weighted mean (range) relative clinical impacts were -29.6% (-23% to -37%) for five-grass pollen SLIT tablets, -19.2% (-6% to -29%) for timothy pollen SLIT tablets, -23.5% (-7% to -54%) for nasal corticosteroids, -17.1% (-15% to -20%) for MP29-02, -15.0% (-3% to -26%) for H1-antihistamines and -6.5% (-3% to -10%) for montelukast.

Conclusions: In an indirect comparison, grass pollen SLIT tablets had a greater mean relative clinical impact than second-generation antihistamines and montelukast and much the same mean relative clinical impact as nasal corticosteroids. This result was obtained despite the presence of methodological factors that mask the clinical efficacy of SLIT for the treatment of seasonal allergic rhinitis.

Figure 1

RCI and meta-analysis of efficacy for H1-antihistamines (on the basis of symptom scores). N Act: number of subjects in the active treatment group; Mean Act: mean score in the active treatment group; SD Act; standard deviation for the score in the active treatment group; N Plac: number of subjects in the placebo group; Mean Plac: mean score in the placebo group; SD Plac: standard deviation for the score in the placebo group; RCI: relative clinical impact; hg: Hedges' g; ci-: lower confidence interval; ci+: upper confidence interval; z: z score: p: p-value; w: weighting; Des: desloratadine; Bila: bilastine; Lor: loratadine; Fex: fexofenadine; Cet: cetirizine. RCI, relative clinical impact.

Figure 2

RCI and meta-analysis of efficacy for nasal corticosteroids (on the basis of symptom scores). N Act: number of subjects in the active treatment group; Mean Act: mean score in the active treatment group; SD Act; standard deviation for the score in the active treatment group; N Plac: number of subjects in the placebo group; Mean Plac: mean score in the placebo group; SD Plac: standard deviation for the score in the placebo group; RCI: relative clinical impact; hg: Hedges' g; ci-: lower confidence interval; ci+: upper confidence interval; z: z score: p: p-value; w: weighting; Beclo: beclomethasone. Mom: mometasone. RCI, relative clinical impact.

Figure 3

RCI and meta-analysis of efficacy for montelukast (on the basis of symptom scores). N Act: number of subjects in the active treatment group; Mean Act: mean score in the active treatment group; SD Act; standard deviation for the score in the active treatment group; N Plac: number of subjects in the placebo group; Mean Plac: mean score in the placebo group; SD Plac: standard deviation for the score in the placebo group; RCI: relative clinical impact; hg: Hedges' g; ci-: lower confidence interval; ci+: upper confidence interval; z: z score: p: p-value; w: weighting. RCI, relative clinical impact.

Figure 4

RCI and meta-analysis of efficacy for an azelastine-fluticasone combination (on the basis of symptom scores). N Act: number of subjects in the active treatment group; Mean Act: mean score in the active treatment group; SD Act; standard deviation for the score in the active treatment group; N Plac: number of subjects in the placebo group; Mean Plac: mean score in the placebo group; SD Plac: standard deviation for the score in the placebo group; RCI: relative clinical impact; hg: Hedges' g; ci-: lower confidence interval; ci+: upper confidence interval; z: z score: p: p-value; w: weighting. RCI, relative clinical impact.

Figure 5

RCI and meta-analysis of efficacy (symptom scores) for five-grass pollen SLIT tablets and timothy pollen SLIT tablets. N Act: number of subjects in the active treatment group; Mean Act: mean score in the active treatment group; SD Act; standard deviation for the score in the active treatment group; N Plac: number of subjects in the placebo group; Mean Plac: mean score in the placebo group; SD Plac: standard deviation for the score in the placebo group; RCI: relative clinical impact; hg: Hedges' g; ci-: lower confidence interval; ci+: upper confidence interval; z: z score: p: p-value; w: weighting. RCI, relative clinical impact.