Allergen-specific immunotherapy ("hyposensitization" or "desensitization") has long been a controversial treatment for asthma. Although beneficial effects upon clinically relevant outcomes have been demonstrated in randomized controlled trials, there remains a risk of severe and sometimes fatal anaphylaxis. The recommendations of professional bodies have ranged from cautious acceptance (1) to outright dismissal (2). A recent WHO position paper, which has been endorsed by eight other international and national bodies, concluded that allergen immunotherapy was an effective treatment for patients with allergic asthma (3). We have previously conducted a meta-analysis of 20 randomized controlled trials of allergen immunotherapy for asthma published between 1954 and 1990 (4). We subsequently conducted a systematic review for the Cochrane Collaboration, including a further 34 trials published between 1957 and 1997 (5). Both reviews concluded that subjects randomized to immunotherapy reported significantly fewer asthma symptoms, required significantly less asthma medication, and demonstrated both reduced nonspecific and reduced allergen-specific bronchial hyperreactivity (BHR) compared to those randomized to placebo. During recent years, there has been increasing interest in new allergen vaccines and new methods of delivery including oral, sublingual, and inhaled immunotherapy. Recombinant peptides containing the relevant epitopes, but lacking the ability to cross-link IgE bound to mast cells, have been evaluated in clinical trials. Finally, the Cochrane Collaboration has standardized protocols for systematic reviews and improved the statistical software for performing meta-analysis. Thus, it was again opportune to update our systematic review of allergen-specific immunotherapy for asthma. The objectives were as follows: to identify as many as possible of the published ran-domized controlled trials of allergen-specific immuno-therapy for allergic asthma to assess the methodological quality of these randomized controlled trials to estimate the overall efficacy of allergen-specific immunotherapy upon asthma symptoms, medication requirements, lung function, nonspecific BHR, and allergen-specific BHR to compare the efficacy in asthma of mite, pollen, and animal dander immunotherapy with extracts of other allergens. This review was restricted to randomized controlled trials (RCTs), which provide the best evidence for the efficacy of any medical treatment (6). Only studies which focused upon asthma were included. Studies of immunotherapy for hay fever were considered only if the results for subjects with asthma were separately identified. For the purposes of this review, allergen-specific immunotherapy included the administration of extracts of house-dust mites, pollens, animal danders, or moulds; chemically modified allergoids; or antigen–antibody complexes. Only the subcutaneous route of administration was considered. Placebo-controlled trials are methodologically stronger, although studies which have administered house dust or other relatively inactive antigenic preparations to the control group were also considered. Double-blind trials were preferred, but single-blind and open studies were also reviewed for possible inclusion. At least one of the following clinical outcomes had to be reported: asthma symptoms (including symptom scores) asthma medication requirements (including medication scores) lung function (including peak expiratory flow, FEV1, and thoracic gas volume) nonspecific BHR (to histamine, methacholine, or acetylcholine) allergen-specific BHR. Studies which reported only nonclinical outcomes such as allergen-specific IgE (by RAST or skin prick test) or other in vitro results were excluded. Studies published in languages other than English were considered if the translated abstract indicated that the study was an RCT of allergen immunotherapy for asthma, and a translator was sought (see Acknowledgments). MEDLINE searches were undertaken for 1966–98 for studies of immunotherapy and asthma. Review articles identified in this process were surveyed for additional and earlier citations. Dissertation abstracts and Current Contents were also searched to identify more recently published and unpublished studies. A search was performed of the Asthma database maintained by the Cochrane Airways Group at St George's Hospital Medical School, London, UK. This database included all studies with the key words "Asthma" or "Wheez*" from the MEDLINE, EMBASE, and Cinahl databases, together with other studies identified by hand searching. A total of over 900 nonunique citations were identified with the key words "Immunotherap*", "Hyposensiti*", or "Desensiti*" and were examined for possible inclusion. Inclusion of studies in the review was decided by a simple majority of all three reviewers, who independently read the Methods sections of papers identified by the search strategy and applied the stated criteria. Quality assessment was performed by two reviewers, who independently assessed the concealment of allocation by the guidelines of the Cochrane Collaboration and applied the scoring system of Jadad et al. (7). The planned comparisons were as follows: allergen immunotherapy vs placebo allergen immunotherapy vs antigenically inactive control house-dust extracts vs placebo allergen immunotherapy vs untreated control allergen immunotherapy vs inhaled steroid. These comparisons were performed separately for each outcome, namely, asthma symptoms, medication, lung function, nonspecific BHR, and allergen-specific BHR, whenever the results were reported. Outcome data were extracted and entered into RevMan, version 3.1.1 (Update Software, Oxford, 1998), for statistical analysis (8). Categorical outcomes were analysed as odds ratios (OR) and 95% confidence intervals (95% CI) calculated by Peto's method for individual studies. The odds ratio is simply the ratio between the immunotherapy and control groups for the odds of subjects being the same or worse to subjects being better after treatment. The convention of the Cochrane Collaboration is to consider odds ratios greater than 1.0 as indicating clinically undesirable outcomes (in this review, worse symptoms, increased medication, deteriorating lung function, and increased nonspecific and allergen-specific BHR). Continuous outcomes (symptom and medication scores, lung-function parameters, and indices of nonspecific and allergen-specific BHR) were also extracted from tables of results. Where the results were presented only in graphs, these were digitized and then converted to numbers with Un-Scan-It v4.0 (Silk Scientific Corporation, Orem UT, 1996). Continuous outcomes were analysed as standardized mean differences (SMD). The SMD is a statistic which expresses the difference in means between immunotherapy and control groups in units of the pooled standard deviation. Although the Cochrane Collaboration currently favours the weighted mean difference, we chose the SMD because many studies measured the same outcomes on different scales. Fixed effects models were generally used to obtain summary statistics for the overall efficacy of allergen immunotherapy upon both categorical and continuous outcomes, and chi-square tests were performed to assess heterogeneity between studies. In this context, a P value of <0.05 indicates significant differences between studies and raises the question of whether the results can be meaningfully combined. When significant heterogeneity was found, the overall efficacy was recalculated by random effects models, which provide more realistic estimates of the confidence intervals under these circumstances (9). Sixty-two randomized controlled trials published in 67 papers between 1954 and 1998 were identified which satisfied the inclusion criteria. The methods, participants, allergens administered, and outcomes of the included studies are listed in Table 1. There were 28 studies report-ing immunotherapy for mite allergy, generally Dermatophagoides pteronyssinus or D. farinae, although Armentia Medina et al. (15) used extracts of the storage mite Lepidoglyphus destructor. Three early studies (10, 16, 75) used extracts of house dust, which has since generally been considered antigenically inactive. There were 15 studies of immunotherapy for pollen allergy, utilizing extracts of Bermuda grass, orchard grass/cocksfoot, timothy grass, velvet grass, sweet vernal grass, perennial ryegrass, or ragweed pollen. Nine studies reported immunotherapy for animal dander allergy, particularly to cats and dogs. Two studies utilized extracts of the mould Cladosporium, and five attempted simultaneous immunotherapy for multiple aeroallergens. Four studies from one group of investigators (43–46) used allergen–antibody complexes for immuno-therapy. A further 125 studies were excluded from this review (77–201). The most common but not mutually exclusive reasons for exclusion were as follows: 37 studies were not randomized (84, 86, 91, 93, 97, 98, 103, 105, 108, 114, 116, 120, 122, 123, 126, 130, 132, 133, 135, 140, 145, 146, 149, 151, 157, 159, 163, 164, 166, 176, 177, 180, 185, 189, 193, 194, 198), 25 studies were not controlled (87–89, 99, 101, 107, 115, 121, 128, 131, 134, 137, 140, 148–150, 155, 169, 172, 181, 184, 189, 195, 199, 201), 29 studies solely or predominantly involved subjects who had allergic rhinitis/conjunctivitis rather than asthma (77, 79, 80, 82, 85, 90, 94, 96, 100, 108–113, 117–119, 127, 130, 138, 139, 153, 158, 160, 161, 174, 176, 192), 13 studies did not report clinically relevant outcomes (81, 129, 134, 144, 152, 155, 165, 166, 173, 179, 190, 196, 200), and one study was not analysed by intention to treat (171). We also identified 14 clinical trials of oral immunotherapy (92, 106, 124, 125, 139, 156, 158, 167, 170, 175, 186–188, 191), seven trials of inhaled or intranasal immunotherapy (79, 80, 95, 142, 143, 147, 183), nine trials of sublingual immunotherapy (78, 90, 102, 110, 154, 158, 162, 182, 197), and one trial of homoeopathic immunotherapy (168) which were beyond the scope of the present review. Finally, six abstracts (83, 104, 119, 136, 141, 178) contained insufficient information for further assessment. Two studies included in our previous review were excluded after closer examination, as they did not meet all the inclusion criteria. The untreated controls in Murray et al. (149) were not randomly allocated and the "placebo" controls were contaminated by concurrent administration of pollen immunotherapy. The allocation of subjects to high- or low-dose immunotherapy by Tuchinda & Chai (189) was not randomized. A further 13 studies published before or during 1998 were considered for possible inclusion. Clarification was sought from the authors as to whether they had conducted randomized controlled trials, but no replies were received by the time of submission. A key criterion for assessing the quality of randomized controlled trials for which there is evidence of a relationship to the potential for bias in the results is concealment of allocation after randomization to active treatment or control (6). For this review, concealment of allocation was assessed (Table 1) as clearly adequate (category A) in only 13 studies (10, 16, 21, 23, 24, 27, 29, 42, 51, 55, 56, 62, 70). The adequacy or otherwise of 40 studies could not be determined from the details published in the papers (category B) and further information was sought from the authors. Only two studies (36, 45) used a clearly inadequate method for concealment of allocation (category C). A further seven open studies did not use any concealment of allocation (category D). The quality score developed by Jadad et al. (7) awarded points for randomization, double blinding, whether these methods were well described and appropriate, and whether withdrawals and dropouts were described. Only three studies (39, 40, 56) achieved the maximum possible score of five points. Eleven studies received four points, 24 received three points, 15 received two points, and nine received only one point for methodological quality (Table 1). There was a high level of agreement between reviewers, and differences were resolved by consensus. Fifty-three studies were placebo-controlled trials of allergen-specific immunotherapy, two compared immuno-therapy with an antigenically inactive control (28, 50), three compared house dust with placebo (10, 16, 75), and three utilized untreated controls (19, 22, 41). Only one study directly compared immunotherapy with inhaled steroids (64), although Torres Costa et al. (69) administered inhaled beclomethasone to both the active treatment and control groups for the first 18 months. The results of the individual placebo-controlled trials and combined effects for each outcome are presented in Figs. 1-10. Symptom scores were reported by 22 studies, although neither Dreborg et al. (32) nor Olsen et al. (56) published the standard deviations (SDs), thus preventing the calculation of the standardized mean difference (SMD) for these studies. The combined SMD for symptom scores after mite immunotherapy was –0.71 with a 95% confidence interval of –1.37 to –0.05, which excluded 0, thus indicating a significant reduction in asthma symptoms (Fig. 1). The combined SMD after pollen immunotherapy was –0.72 (95% CI –1.14 to –0.31) also indicating significant symptomatic improvement. However, there was no significant improvement after immunotherapy with cat, dog, or multiple allergen extracts. For all allergens combined, the SMD was –0.73 (95% CI –1.07 to –0.39), but there was significant heterogeneity between studies (χ2=86.2, P<0.0005), with three studies (11, 25, 37) not finding any reduction in symptoms. Meta-analysis of asthma symptom scores from placebo-controlled trials of allergen immunotherapy for asthma – standardized mean differences (with 95% confidence intervals) for each study, studies of mite, pollen, and other immunotherapy and all studies combined. Meta-analysis of symptomatic deterioration from placebo-controlled trials – odds ratios (with 95% confidence intervals) for each study, studies of mite, pollen, animal dander, and other immunotherapy and all studies combined. Meta-analysis of asthma medication scores from placebo-controlled trials – standardized mean differences (with 95% confidence intervals) for each study, studies of mite, pollen, and other immunotherapy and all studies combined. Meta-analysis of increased asthma medication requirements from placebo-controlled trials – odds ratios (with 95% confidence intervals) for each study and all studies combined. Meta-analysis of lung function parameters from placebo-controlled trials – standardized mean differences (with 95% confidence intervals) for each study, peak expiratory flow, FEV1, and thoracic gas volume. Meta-analysis of deterioration in lung function from placebo-controlled trials – odds ratios (with 95% confidence intervals) for each study and all studies combined. Meta-analysis of indices of nonspecific bronchial hyperreactivity from placebo-controlled trials – standardized mean differences (with 95% confidence intervals) for each study, logPD20 to methacholine, logPC20 to histamine, logPC35 to acetylcholine, and all studies combined. Meta-analysis of increased nonspecific bronchial hyperreactivity from placebo-controlled trials – odds ratios (with 95% confidence intervals) for each study and all studies combined. Meta-analysis of indices of allergen-specific bronchial hyperreactivity from placebo-controlled trials – standardized mean differences (with 95% confidence intervals) for each study, studies of mite, pollen, animal dander, and other immunotherapy and all studies combined. Meta-analysis of increased allergen-specific bronchial hyperreactivity from placebo-controlled trials – odds ratios (with 95% confidence intervals) for each study and all studies combined. Symptoms were simply reported as worse, the same, or improved in another 22 studies. The combined odds ratio (OR) was 0.26, with a 95% CI from 0.17 to 0.41, which excluded 1, again indicating that patients randomized to immunotherapy were significantly less likely to report a deterioration in asthma symptoms than those randomized to placebo (Fig. 2). There were significant homogeneous improvements after immunotherapy with extracts of pollen (OR 0.13, 95% CI 0.04–0.42), animal dander (OR 0.20, 95% CI 0.05–0.87), and other allergens (OR 0.17, 95% CI 0.09–0.32). Less improvement was seen after mite immunotherapy (OR 0.39, 95% CI 0.21–0.75), and there was significant heterogeneity between these studies (χ2=21.3, P<0.05). Although the results of studies of children were relatively homogeneous (data not shown), there was significant heterogeneity between the results of adult studies (χ2=40, P<0.001), two studies (17, 26) actually finding symptoms to be more likely in treated patients. Asthma medication scores were reported by 13 studies, with Cantani et al. (27), Dreborg et al. (32), and Olsen et al. (56) failing to publish the SDs. The combined SMD was –0.71 (95% CI –1.09 to –0.32), indicating a significant reduction in medication after immunotherapy (Fig. 3). There was a large reduction after mite immunotherapy, although this was accompanied by significant heterogeneity (χ2=16.1, P<0.005). The reduction after pollen immunotherapy achieved statistical significance in a fixed effects model (SMD=–0.54, 95% CI –0.85 to –0.23). Medication requirements were simply reported as increased, unchanged, or decreased in 16 studies. The combined OR was 0.32 (95% CI 0.22–0.46), indicating that patients randomized to immunotherapy were significantly less likely to require increased medication than those randomized to placebo (Fig. 4). Although there was significant heterogeneity between the studies reporting medication scores (χ2=28.7, P<0.001), there was substantial homogeneity between the latter 15 studies. Lung-function results were reported by 14 studies, with several studies failing to provide SDs for peak expiratory flow (32), FEV1 (35, 46), or thoracic gas volume (61, 76). There was no overall improvement in lung-function parameters after immunotherapy (Fig. 5), and there was marked heterogeneity in peak expiratory flow between studies (χ2=27.6, P<0.0005). Indeed, there was even a suggestion that FEV1 deteriorated after immunotherapy in a small study by Paranos & Petrovic (58), in which the baseline lung function of patients randomized to mite immunotherapy and placebo was poorly matched. Lung function was simply reported as worse, the same, or improved in seven studies. Although there was homogeneity between these studies and an overall trend against a deterioration in lung function after immunotherapy, this did not achieve statistical significance (Fig. 6). Indices of nonspecific BHR were reported by 12 studies. There were significant improvements in PD20 FEV1 to methacholine challenge (SMD –0.30, 95% CI –0.54 to–0.05) and in PC35 sGaw to acetylcholine after immunotherapy with allergen–antibody complexes (43). The improvement in PC20 FEV1 to histamine challenge failed to achieve statistical significance in a random effects model. Although there was significant heterogeneity between the results of these studies (χ2=22.7, P<0.025), there was an overall reduction in nonspecific BHR after immunotherapy (SMD –0.48, 95% CI –0.81 to –0.14) (Fig. 7). Nonspecific BHR was simply reported as increased, unchanged, or reduced in five small studies. There was homogeneity between these studies, and the combined OR of 0.22 (95% CI 0.10–0.48) indicated that patients randomized to immunotherapy were significantly less likely to develop increased nonspecific BHR than those randomized to placebo (Fig. 8). Indices of allergen-specific BHR (such as PD20 FEV1 to allergen challenge) were reported by 14 studies. There was homogeneity between these studies with an overall SMD of–0.70 (95% CI –0.91 to –0.48), indicating a significant reduction in allergen-specific BHR after immunotherapy (Fig. 9). The effect was most marked for mite immunotherapy (SMD –1.14, 95% CI –1.62 to –0.65), and similar for pollen (SMD –0.69, 95% CI –1.09 to –0.30) and animal dander (SMD –0.71, 95% CI to but not significant for other allergens. Allergen-specific BHR was simply reported as increased, unchanged, or reduced by 16 studies. There was homogeneity between studies, and the combined OR of (95% CI indicated that patients randomized to immunotherapy were significantly less likely to develop increased allergen-specific BHR (Fig. of the small and outcomes reported by the randomized trials, only meta-analysis could be did not a significant reduction in asthma symptoms immunotherapy with house dust 95% to However, et al. did that patients randomized to were significantly less likely to report asthma symptoms than those randomized to house dust 95% CI did not any significant reduction in asthma after immunotherapy 95% CI The three studies (10, 16, 75) which compared house-dust extracts with placebo results for most randomized to house dust were less likely to report worse asthma symptoms, but the combined effect did not achieve statistical significance 95% CI In the study, those randomized to house dust had reduction in asthma medication 95% CI and were less likely to deterioration in lung function 95% CI However, that children randomized to house dust had significantly reduced allergen-specific BHR 95% CI the other studies with untreated controls reported results of allergen immunotherapy. Both et al. and et al. that patients randomized to immunotherapy demonstrated a significant reduction in asthma symptoms 95% CI to and a significant improvement in lung function 95% CI to et al. also reported significantly reduced medication requirements in the immunotherapy group 95% CI to Finally, et al. that those randomized to immunotherapy were significantly less likely to develop increased allergen-specific BHR 95% CI The only trial to compare allergen immunotherapy directly with inhaled steroids that symptom scores and FEV1 improved more in the group randomized to However, as no standard deviations were it was not possible to standardized mean In any there were no other trials with which these results could be meaningfully combined. This systematic review of allergen immunotherapy for asthma has identified an additional randomized controlled trials which were not by our previous review of studies published to 1997 (5). of the studies not previously included were published during or after Two trials which were included in the previous review were excluded on this after closer that they did not meet all the inclusion criteria. A total of 125 studies were excluded from the review, because they were not not did not report clinically relevant outcomes, or involved only subjects with allergic the bias for the of studies, it is possible that additional studies have been but that results have not been published in the medical This could be in the by of all clinical trials of allergen immunotherapy, as has been attempted in other clinical the concealment of allocation after randomization could not be clearly determined from the information presented in most of the We authors of papers to the guidelines and report such However, there was no relationship between concealment or methodological quality and the of individual of immunotherapy. This meta-analysis has that allergen-specific immunotherapy can significantly asthma symptoms and medication The heterogeneity of symptom and medication scores is likely to have from the different scoring used in different studies. The heterogeneity of symptom scores significant after for the allergen but one study combined both symptomatic and into a We have previously this as in requirements could be by in inhaled of asthma symptom and medication scores is required for of studies. However, we not that the heterogeneity of the results means that the studies not be combined at all or that it the Medication requirements which were reported as in to the analysis medication scores, significant We that this finding can into a clinically outcome, as one of the of allergen immuno-therapy is to medication inhaled remains the of asthma any reduction in this of treatment asthma control be There was no effect of immunotherapy upon lung The marked heterogeneity have from to such of lung function as FEV1, peak expiratory flow, and thoracic gas volume. These parameters different of the of asthma, such as reduced and However, significant heterogeneity after the analysis for each In our first meta-analysis did a improvement in lung However, in this from the exclusion of one study in to a randomized to immunotherapy were significantly less likely to develop increased nonspecific BHR, and there were improvement indices of nonspecific BHR. was achieved by the meta-analysis for each of nonspecific PD20 FEV1 to methacholine, PC20 FEV1 to histamine, and PC35 sGaw to indices of BHR are generally considered to a the means and standard deviations of logPD20 were entered into to the immunotherapy significantly reduced allergen-specific BHR. was achieved by the meta-analysis for the allergen administered and the results as be for studies to use a standardized for bronchial allergen and to report the results in a more it that allergen immunotherapy has a greater effect upon allergen-specific BHR than upon nonspecific BHR. The finding that allergen immunotherapy significantly and allergen-specific bronchial BHR is clinically with asthma are at risk of deterioration to increased of an to which they are which the risk of an of asthma under these circumstances be clinically the of allergen-specific BHR is the only method of assessing such a Less be to the results of the relatively small of trials. There is evidence that extracts asthma symptoms and allergen-specific BHR compared to house-dust extracts. study even that house-dust extracts could allergen-specific BHR, the that such extracts are antigenically inactive. However, these results are with the greater efficacy of extracts compared with placebo. Studies with untreated controls are particularly to bias and the of treatment. The effects of immunotherapy upon symptoms, medication, and lung function in such studies are beyond the confidence for placebo-controlled studies. The results of this systematic review are with our previous 5), the have not been by the inclusion of additional randomized controlled trials. the combined effects of allergen immunotherapy upon symptom and medication scores are greater than previously In this review, greater control of heterogeneity was achieved with by allergen and of combined effects by random rather than fixed effects The effects on all other outcomes were the there are to the present review. of the studies included any data to asthma or patients are included in clinical trials, those of allergen immunotherapy. The of immunotherapy, particularly the of beyond the scope of the present review. are not assessed in of clinical trials and are better by reporting A review by identified fatal over a during which to allergen were administered in the The mean of was This report the efficacy of in we the clinically outcomes of decreased medication requirements and improved allergen-specific BHR. only did these two outcomes significant but also the results were We that a reduction in medication and decreased allergen-specific BHR are clinically and can to improved asthma These results no the clinical of allergen immunotherapy. We have previously stated the that we in immunotherapy in asthma (4). These are more in a position of the of and and the of and and the WHO position (3). We these position papers and that to However, to be are the most of the clinical of an is the of the better there is a of skin or is it as effective in immunotherapy better than the use of multiple allergens to which the is the of treatment and the best of the of immunotherapy compare to new for of these be in the studies currently but require new and better randomized controlled trials. We and for in studies published in languages other than The was used to papers published in to the data from randomized controlled trial of pollen immunotherapy in children with asthma. unpublished methodological details the trials conducted by The Cochrane Airways Group database was searched by and and from the Cochrane and on the use of and on a previous version of this review.
No takes yet. Share an insight, caveat, or question.
Abramson et al. (1999) studied this question.
Synapse has enriched 2 closely related papers on similar clinical questions. Consider them for comparative context: