RESUMO
BACKGROUND: Five published meta-analyses (MAs) seem to prove the efficacy of sublingual immunotherapy in allergic asthma and rhinoconjunctivitis. OBJECTIVE: We aimed to assess the consistency, magnitude, and robustness of the results of these MAs. METHODS: The data reported in the MAs were checked with the data reported in the original studies. Funnel plots were performed to test for potential publication bias, and the trim-and-fill method was used to assess and correct the estimate of the effects if asymmetry was present. RESULTS: The 5 MAs included 43 studies; 17 were used in more than one MA. There were discrepancies among the MAs in the data reported from the same original studies: the MAs reported different estimates for the same outcome or the same estimates for different outcomes in 16 of those 17 studies. The MAs evaluated 15 main outcomes, 10 of which showed benefits that reached statistical significance. Funnel plots showed asymmetry in 7 outcomes, and correction by using the trim-and-fill method led to a decrease in their effect estimates and even to a loss of statistical significance in 4 of the previously significant outcomes. There was inconsistency among the MAs in the benefits when considering age, disease, allergen, or symptoms and medication use. CONCLUSION: Because of discrepancies, inconsistencies, and lack of robustness, the MAs on sublingual immunotherapy do not provide enough evidence to support its current routine management in patients with allergic asthma or rhinoconjunctivitis. Sensitivity to potential publication bias should be tested and reported in all MAs.
