In a meta-analysis of 88 trials of several different kinds of allergy immunotherapies in asthma patients, Michael Abramson, MD, of Monash University in Melbourne, Australia, and colleagues found that treating only three to four patients would prevent one from having a symptom flare.
Pooled data from placebo-controlled studies that reported effects on asthma symptom scores, covering 727 patients receiving allergy desensitization and 557 assigned to placebo, showed a reduction in symptom severity of 0.59 points (95% CI 0.35 to 0.83) on a 20-point standardized scale.
But in their report published online in the Cochrane Database of Systematic Reviews — updating an earlier meta-analysis published in 1995 — the researchers cautioned that the included studies varied substantially in methodology, and that treatment allocations may not have been concealed adequately in most of the 72 studies billed as double- or single-blind.
Abramson and colleagues also noted that the numbers needed to harm for localized or systemic reactions to the immunotherapies were not a lot larger than the numbers needed to treat for efficacy endpoints.
For every 16 patients treated, one localized reaction was noted. And systemic reactions such as hives or breathing difficulties occurred in one patient for every nine treated, although fatal anaphylaxis was extremely rare.
“The evidence assembled in this review confirms the efficacy of immunotherapy in terms of a reduction in asthma symptoms and use of asthma medication,” Abramson and colleagues wrote.
“The data show that immunotherapy can be considered when asthma is extrinsic and an unavoidable clinically relevant allergen can be identified,” they recommended.
On the other hand, the limitations of the individual studies and the meta-analysis process meant that the degree of benefit could not be readily compared with other asthma therapies. Abramson and colleagues indicated that immunotherapy “is possibly comparable to inhaled steroids,” but it’s equally possible that the benefit is much smaller.
They also suggested that, in view of the risk of reactions, patients receiving allergen desensitizations be kept under observation long enough to detect and treat serious reactions.
To be included in the meta-analysis, trials had to be randomized and controlled using some type of allergen-specific immunotherapy and reporting at least one clinical outcome related to asthma.
Abramson and colleagues identified 88 that met these basic criteria. House mite allergens were used in 42 studies, pollen allergens in 27 trials, and animal dander proteins in 10. Other trials examined mold, latex, or multiple allergens.
Nearly all the control treatments were placebo. Sixty-three of the trials were double-blind, nine were single-blind, 13 were open-label, and three had mixed or unclear designs.
But the researchers determined that only 16 of the blinded studies had “clearly adequate” strategies to assure that participants did not know their treatment assignments.
“The adequacy or otherwise of 57 studies could not be determined from the details published in the papers … and further information was sought from the authors, but rarely forthcoming,” Abramson and colleagues remarked.
Despite this limitation, all 88 trials were included in the pooled assessments.
Twenty-one of these reported quantitative information on medication use. As with symptom scores, Abramson and colleagues pooled the data and reported the relative change between immunotherapy and control treatment on a 20-point standardized scale. They found that the interventions reduced medication use by a mean of 0.53 points (95% CI 0.27 to 0.80).
When these trials were combined with 17 others that simply reported whether medication use increased, decreased, or stayed the same, Abramson and colleagues determined that treating five patients with immunotherapy (95% CI 4 to 7) would prevent one from needing an increase in medication.
Wide variations in study methodologies precluded firm conclusions regarding lung function or airway sensitivity to irritant compounds, the researchers found.
The pooled data indicated no significant effect on FEV1 values but some benefit on bronchial hyperreactivity (0.35-point reduction, 95% CI 0.11 to 0.59, from 18 studies). For the latter endpoint, Abramson and colleagues found that results varied significantly depending on the specific challenge agent used, such as methacholine versus histamine or cold air.
The researchers recommended additional research in several areas, including clinical identification of patients most likely to respond to allergen desensitization and the allergens best targeted in asthma, as well as ways to reduce adverse reactions.