Effects of Antihistamines in Adult Asthma: a Meta-Analysis of Clinical Trials
Total Page:16
File Type:pdf, Size:1020Kb
Eur Respir J 1997; 10: 2216–2224 Copyright ERS Journals Ltd 1997 DOI: 10.1183/09031936.97.10102216 European Respiratory Journal Printed in UK - all rights reserved ISSN 0903 - 1936 Effects of antihistamines in adult asthma: a meta-analysis of clinical trials E. Van Ganse*, L. Kaufman**, M.P. Derde**, J.C. Yernault+, L. Delaunois++, W. Vincken* Effects of antihistamines in adult asthma: a meta-analysis of clinical trials. E. Van *Respiratory Division, Academic Hospital Ganse, L. Kaufman, M.P. Derde, J.C. Yernault, L. Delaunois, W. Vincken. ©ERS Journals University of Brussels (AZ VUB), Brussels. Ltd 1997. **Dept of Biostatistics, University of + ABSTRACT: A meta-analysis of clinical trials of antihistamines was performed to Brussels (AZ VUB), Brussels. Dept of Pneumology, University of Brussels (ULB assess the risk-benefit ratio of this therapeutic class in asthma. Erasme), Brussels. ++Dept of Pneumology, Double-blind randomized placebo-controlled trials assessing lung function chan- University of Louvain (UCL Mont-Godinne), ges under repeated use of antihistamine in adult asthma were selected, and the Belgium. quality of studies was scored. Morning peak expiratory flow rate (PEFR) was the Correspondence: E. Van Ganse, Dept of primary outcome: an effect size was computed for each study, with a 95% confi- Clinical Pharmacology, University Claude- dence interval (95% CI), and a mean effect size was computed, combining all stu- Bernard, 162, avenue Lacassagne, BP 3041, dies. Effect sizes were also determined for secondary outcomes: evening PEFR, F-69394 Lyon Cedex 03, France. forced expiratory volume in one second (FEV1) and daily use of inhaled β-ago- Keywords: Asthma drug therapy nists. histamine H1-receptor blockers Nineteen studies were included in the meta-analysis. Mean quality score of stud- meta-analysis ies was 59.4%; asthma was generally uncontrolled at study inclusion. Altogether, risk-benefit ratio 582 antihistamine-treated and 557 placebo-treated asthma patients were evalu- Received: December 3 1996 able. Antihistamines had little effect on airway calibre (mean increase in morning Accepted after revision June 28 1997 PEFR: 13 L·min-1; 95 CI: 8–18 L·min-1) and on use of inhaled β-agonists (mean This study was supported by the Dept of reduction in daily use: 0.4 doses; 95% CI: 0–0.8 doses). Sedation occurred more Pneumology and the Dept of Biostatistics often with antihistamines than with placebo (p<0.001); additional side-effects were of the Brussels University Hospital (AZ mentioned, including weight gain, altered taste, headache and dry mouth. VUB) and, in part, by the EC BIOMED Respiratory and systemic effects observed after repeated use of antihistamines grant No. BMH4 CT965033. Parts of the results have been presented at the XVth do not support the use of these medications in the treatment of asthma; better Congress of the European Academy of designed studies could affect this appraisal. Allergology and Clinical Immunology Eur Respir J 1997; 10: 2216–2224. (Paris, May 1992). Asthma is a common disease, affecting up to 5% with improved risk-benefit ratios. Research has focused of the population in Western countries. Management on inhaled and oral agents that might inhibit the effects of asthma should start with patient education, informa- of pro-inflammatory mediators [3]. tion on risk factors and preventive measures. The next Disodium cromoglycate, a nonsteroidal inhaled agent, stage includes administration of medicines, and a step- was produced as an alternative to corticosteroids. Des- wise approach is recommended in recent guidelines pite encouraging results from early studies, more recent [1]. data tend to question its effects in asthma [2]. The first Pharmacological treatment of asthma includes two cat- oral nonsteroidal compound that has been developed egories of drugs: bronchodilators and anti-inflammatory for the treatment of inflammation in asthma is ketotifen, medications. Anti-inflammatory drugs target the inflam- an antihistamine [4]. Ketotifen was tested in asthma matory process that develops in asthmatic bronchi. This following initial reports of mast cell stabilizing effects therapeutic class is represented by corticosteroids, ad- and is now widely used in some countries [5]. Histamine ministered via local or systemic route. Corticosteroids exhibits numerous actions of relevance to asthma, such have demonstrated their efficacy in asthma, reducing as bronchoconstriction, enhanced mucus secretion and in- severity and decreasing the need for symptomatic med- creased vascular permeability: these actions are partly ications. Present recommendations favour early use of H1-receptor mediated [6]. New antihistamines are consi- inhaled corticosteroids, since beneficial effects may be dered to be less sedating than older compounds, lead- expected against lung function losses and disease dete- ing to the use of higher doses [7]. This property, combined rioration [2]. with early reports that new antihistamines may have Widespread use of corticosteroids is, however, still specific anti-allergic properties in addition to H1-block- advised with some caution, since side-effects may occur, ing activity, has led to development in asthma therapy with both oral and inhaled formulations, as a conseq- [8–10]. To date, clinical trials have delivered contradic- uence of low specificity of action. This has stimulated tory results, and the matter is still the subject of debate considerable efforts to produce anti-inflammatory agents [11, 12]. EFFECTS OF ANTIHISTAMINES IN ASTHMA: A META-ANALYSIS 2217 Meta-analysis is the statistical analysis of a large publications published since January 1991 that were rel- collection of results from individual studies for the pur- evant to the present study. A computerized search was pose of integrating independent research results, and performed in the Medline and Embase databases, look- the methods are clearly defined [13–15]. Meta-analyses ing for articles published since January 1991 and cor- have been performed in several fields, including pneu- responding to the criteria defined in the first part of the mology, and have contributed to providing a basis for study. This was completed by a manual search in Index rational interventions [16]. This study was performed Medicus and in major pneumology and allergology jour- to investigate the positive and negative effects of anti- nals. The articles retrieved were also submitted to a histamines as a therapeutic class in asthma, by retriev- quality assessment, but there were no secondary exclu- ing and combining all relevant studies published since sions of articles by the assessors. 1980. Data abstraction and choice of primary and secondary outcomes. The following data were abstracted from each Materials and methods eligible report, using a standard data extraction form: morning and evening peak expiratory flow rate (PEFR); forced expiratory volume in one second (FEV1) values; Literature review daily use of inhaled β-agonists; incidence of sedation, fatigue and drowsiness; and incidence of other adverse Studies published between 1980 and 1990. A comput- events. Morning PEFR was chosen as a primary out- erized search was performed in the Medline, Ringdoc come; evening PEFR, FEV1 and daily use of inhaled and Emdrugs databases, looking for articles published bronchodilators were used as secondary outcomes. in English between 1980 and 1990 that investigated the In some studies, in addition to placebo, antihistamines effects on humans of oral antihistamines in asthma. were also compared to other asthma medications: data Bibliographies of retrieved articles were reviewed to from this latter group were not used. When data were identify additional articles not listed in the computer- not tabulated, they were extracted from figures, where ized search. This was completed by a manual search of possible. When data were available for different treat- Index Medicus and in major pneumology and allergy ment durations, we used the data corresponding to the journals. This part of the study was completed in Sep- longest duration of exposure. When different doses of tember 1991. an antihistamine were studied, we included only the The following criteria were applied for including arti- data corresponding to the most effective dose, accord- cles: randomized placebo-controlled trials; studies pub- ing to the authors' claim, when available. In studies lished between January 1980 and December 1990; studies where results were separated into distinct groups, (e.g. with more than 50% of the patients older than 15 yrs; co-therapy with inhaled corticosteroids: yes/no), we results published in English; evidence of peer-review recomputed the data to the whole study group using the process; and patients treated for more than 2 weeks. following computation: Studies which were excluded at this stage (primary exclu- sion) contained only pharmacodynamic and/or pharma- <xs> = n1.x1 + n2.x2 / n1 + n2 cokinetic data; were published only as abstracts; or lacked quantified assessment of lung function (in tables or fig- where n1 the number of patients in group 1, n2 the num- ures). ber of patients in group 2, <x > is the mean global score All asthma definitions given by the authors were ac- s in the study, x1 the mean global score in group 1 and cepted. Degrees of asthma severity or concomitant the- x the mean global score in group 2. rapy were not used as criteria to reject a study, but we 2 attempted to group the studies according to disease se- verity. Therefore, priority was