Thomas M. Ormiston

Meta-Analysis: Effect of Long-Acting -Agonists on Severe Asthma Exacerbations and Asthma-Related Deaths

Context Long-acting -agonists may help improve asthma symptoms, but they also may increase risks for adverse outcomes. Contribution This meta-analysis summarizes data from 19 randomized, placebo-controlled trials involving 33826 participants with asthma. Compared with placebo, long-acting -agonists increased severe exacerbations requiring hospitalization (odds ratio, 2.6 [95% CI, 1.6 to 4.3]), life-threatening exacerbations (odds ratio, 1.8 [CI, 1.1 to 2.9]), and asthma-related deaths (odds ratio, 3.5 [CI, 1.3 to 9.3]; risk difference, 0.07%). Risks were similar for salmeterol and formoterol and in children and adults. Cautions Several trials did not report information about potential harms, and the number of reported deaths was small. The Editors Long-acting -agonists produce bronchodilation and improve asthma symptoms, effects that are maintained with regular use over time (1). However, monotherapy with -agonists has consistently been shown to be inferior to the use of inhaled corticosteroids, which can reduce the underlying inflammation associated with the disease (2-5). For this reason, inhaled corticosteroids are recommended as first-line maintenance treatment for asthma (6). For patients whose condition is inadequately controlled with inhaled corticosteroids alone, the addition of a long-acting -agonist is recommended (7). Studies have shown that adding a long-acting -agonist can improve symptoms and reduce the risk for an asthma exacerbation, defined as clinical worsening of disease or an asymptomatic decrease in peak flows (8, 9). The conclusion from these studies was that long-acting -agonists improve asthma control (1, 7). Another possible explanation is that they improve peak flows and symptoms while worsening disease control. Much controversy has surrounded the use of -agonists in patients with asthma ever since their introduction over 50 years ago (10-14). Regular -agonist use is associated with tolerance to the drugs effects and a worsening of disease control (15-20). This effect results from a tight negative feedback mechanism that is an adaptive response to continued adrenergic stimulation (21). After the U.S. Food and Drug Administration (FDA) received postmarketing reports of several asthma-related deaths associated with the long-acting -agonist salmeterol, the Salmeterol Multi-center Asthma Research Trial (SMART) was performed. This study followed more than 26000 participants for 6 months and found a 4-fold increased risk for asthma-related deaths (22, 23). In July 2005, an advisory panel to the FDA met to examine whether long-acting -agonists should be taken off the market (24). The panel concluded that strong warnings of increased risk should be placed on the labeling of all long-acting -agonists, with recommendations that they be used only after other asthma drugs have failed (25, 26). Life-threatening and fatal asthma attacks are relatively rare outcomes, even in large trials. For example, SMART found approximately 2 asthma-related deaths in 1000 patient-years of salmeterol use (23). The effect of long-acting -agonists on these events can be more precisely estimated by pooling the results of many trials. The objective of our study was to assess the effect of long-acting -agonists on severe asthma exacerbations requiring hospitalization, life-threatening asthma attacks, and asthma-related deaths. We used subgroup analyses to compare results for salmeterol and formoterol and for children and adults. Methods Search Strategy We searched the MEDLINE, EMBASE, CINAHL, and Cochrane databases to identify randomized, controlled trials on long-acting -agonist use in patients with asthma that were published between 1966 and December 2005. The search used the terms bronchodilator, sympathomimetic, adrenergic beta-agonist, formoterol, eformoterol or salmeterol and asthma, bronchial hyperreactivity, wheeze, respiratory hypersensitivity, obstructive lung disease, and obstructive airway disease or obstructive pulmonary disease. Trials were not excluded on the basis of language. The search was augmented by scanning references of identified reviews, as well as relevant files from the FDA Web site (www.fda.gov). Study Selection and Assessment of Validity We included studies if they were randomized, controlled trials of long-acting -agonists compared with placebo and lasted at least 3 months. Two reviewers assessed the methodologic quality of each trial according to the following factors: 1) Was the randomization procedure adequate and was allocation concealment described? 2) Were patients and providers blinded to the interventions? 3) Were dropouts and withdrawals reported and was analysis performed by the intention-to-treat principle? Each of these quality domains was scored on a 3-point scale. Trials received an A score when all quality criteria for the domain were met, a B score when the criteria were partially met, and a C score when the criteria were not met. The quality assessment was used for a sensitivity analysis (27, 28). Data Extraction and Synthesis Two reviewers independently extracted data from the selected articles, reconciling differences by consensus. Outcomes assessed were severe asthma exacerbations requiring hospitalization, life-threatening asthma exacerbations requiring intubation and ventilation, and asthma-related deaths. Asthma deaths were those thought to be related to asthma as the underlying cause. Asthma deaths were also included as life-threatening exacerbations. We attempted to contact investigators to obtain additional information on asthma exacerbations and deaths. The proportions of patients with severe exacerbations or asthma-related deaths to patients without those events from each trial were pooled by using the fixed-effects method expressed as a Peto odds ratio (OR) with corresponding 95% CIs (29, 30). We considered this method appropriate because we noted low event rates and minimal heterogeneity in the analyses. Evidence of interstudy heterogeneity was evaluated, with statistical significance set at an value of 0.1. The analysis was performed by using Cochrane Review Manager 4.2 (Cochrane Library Software, Oxford, United Kingdom). Only trials that reported at least 1 event, such as hospitalization or death, could be used in the estimation of ORs. If more than 1 event occurred in the same patient, only the first event was counted. Risk differences and exact 95% CIs were calculated for the difference between 2 independent binomial proportions (StatXact 7, Cytel Software, Cambridge, Massachusetts). The results for each trial were pooled by using the fixed-effects method (29). Trials that reported no events were included in the analysis of risk difference. Subgroup analyses, chosen a priori, were performed to evaluate the difference in ORs between trials of salmeterol versus formoterol and trials in children (<12 years of age) versus adults. The results of the subgroups were compared with each other by using the test of interaction (31). Role of the Funding Source This analysis was funded by salary support from Santa Clara Valley Medical Center for Drs. Salpeter and Ormiston. The institution had no role in the design, conduct, or reporting of the study. All investigators had complete access to the data, and no sponsorship from the institution or the pharmaceutical industry was provided to conduct this analysis. Data Synthesis Search Results Figure 1 shows the results of the search for articles. Through the MEDLINE search, we identified approximately 5000 articles, of which 133 were potentially relevant trials of long-acting -agonist use in patients with asthma. After scanning references from selected articles and the FDA Web site, we identified an additional 6 trials. The EMBASE and Cochrane databases provided no additional trials. Of these 139 trials, 47 met the inclusion criteria (4, 23, 32-76). Thirty of these trials did not report outcomes of interest. After we attempted to contact investigators, 2 responded with unpublished information on deaths (36, 39). The 28 trials that did not provide adequate information on exacerbations or asthma-related deaths (4, 50-76) were not included in the primary pooled analysis but were used in a sensitivity analysis to estimate the lower limit of risk difference by assuming that no deaths occurred in them. Figure 1. Study flow diagram. Trials were excluded for the following reasons: One trial was not randomized, 1 trial was on asthma and chronic obstructive pulmonary disease (COPD), 30 trials did not compare a long-acting -agonist with placebo, 51 trials lasted less than 3 months, and 9 trials provided duplicate data on participants from other trials. Trial Characteristics The primary analysis included 19 trials, with a total of 33826 participants followed for 16848 patient-years (Table). The mean trial duration was 6.0 months (range, 3 to 12 months), with a mean sample size of 1780 participants (range, 110 to 26353 participants). The mean age of participants at baseline was 37.2 years (SD, 5.7) in the -agonist group and 37.7 years (SD, 4.7) in the placebo group. The percentage of men in the 2 groups was 51.2% and 50.2%, respectively. Approximately 15% of the participants were African American. The dropout rate was 20.3% in the -agonist group and 22.6% in the placebo group. Table. Studies Included in Primary Analysis* The long-acting -agonists used in the studies were salmeterol, formoterol, and eformoterol. During the trials, concomitant inhaled corticosteroids were used in 53.9% of participants in the -agonist group and 53.2% of those in the placebo group. Of note, all trials except 2 (36, 48) were sponsored by a pharmaceutical company that manufactures a long-acting -agonist, and all allowed the use of as-needed short-acting -agonists. All trials were randomized, double-blind, placebo-controlled trials that performed analysis according to the intention-to-treat principle and described withdrawals. Nine trials described the metho

Meta-analysis: effect of hormone-replacement therapy on components of the metabolic syndrome in postmenopausal women

Aim:  To quantify the effects of hormone‐replacement therapy (HRT) on components of the metabolic syndrome in postmenopausal women.

Cardioselective β-Blockers in Patients with Reactive Airway Disease: A Meta-Analysis

Context Although -blockers improve clinical outcomes in many patients with cardiovascular disease, clinicians sometimes avoid these agents in patients with concomitant lung disease because they fear precipitation of acute bronchospasm. Contribution This meta-analysis of 29 randomized trials shows that cardioselective-blockers (1-blockers), given for a few days to a few weeks, do not significantly worsen pulmonary function or respiratory symptoms and do not lead to increased use of inhalers in patients with mild to moderate reactive (reversible) airway disease. Cautions The studies in this meta-analysis were short, evaluated only cardioselective -blockers, and did not include patients with severe or irreversible airway disease. The Editors -Adrenergic blocking agents, or -blockers, are indicated in the management of angina pectoris, myocardial infarction, hypertension, congestive heart failure, cardiac arrhythmia, and thyrotoxicosis and are given to reduce perioperative complications (1-13). Despite clear evidence of the effectiveness and mortality benefit of these drugs, clinicians are often hesitant to administer them in patients with some common conditions for fear of adverse reactions (14-17). Many patients with reactive airway disease, with or without a chronic obstructive component, have concomitant conditions such as hypertension or cardiac arrhythmias, which necessitate the use of -blockers. However, review articles and practice guidelines usually list asthma and chronic obstructive pulmonary disease (COPD) as contraindications to -blocker use, citing cases of acute bronchospasm during administration of noncardioselective -blockers (6, 10, 18-22). Cardioselective -blockers, or 1-blockers, have greater than 20 times more affinity for 1 receptors than for 2 receptors and in theory should pose much less risk for bronchoconstriction (23). We used data from randomized, blinded, placebo-controlled trials to evaluate the effect of cardioselective 1-blockers on respiratory function in patients with reactive airway disease (defined as asthma or COPD with a reversible obstructive component). We also sought to evaluate the respiratory response to 2-agonists administered after 1-blockers or after placebo in the same participants. This analysis has already been published as a review in the Cochrane Library (24). Methods Patients We chose to evaluate only patients with documented reactive airway disease because these patients are thought to be particularly susceptible to the adverse respiratory effects of -blockers. Patients with COPD are generally at greater risk for ischemic heart disease than are patients with asthma and thus may benefit more from the use of -blockers. This study evaluates a subgroup of patients with a documented chronic obstructive component of disease but was not designed to make recommendations about patients with COPD. A recent meta-analysis evaluated the use of cardioselective -blockers in patients with COPD, given as a single dose or as continued treatment (25). Pooled data from 19 trials demonstrated no adverse effect on FEV1 or respiratory symptoms for 1-blockers compared to placebo, even in patients with severe chronic airway obstruction. Search Strategy A search was performed to identify all relevant published clinical trials that addressed the effects of cardioselective -blockers on airway function in patients with reactive airway disease. Two investigators jointly developed strategies with the help of an information service librarian and the Cochrane Airways Group Trial Search Coordinator. The EMBASE, MEDLINE, and CINAHL databases were searched comprehensively to identify all relevant clinical trials in humans published between 1966 and May 2001. The search was performed by using the Cochrane Airways Group registry to identify randomized, blinded, placebo-controlled trials of reactive airways disease. Terms used in the search were asthma*, bronchial hyperreactivity*, respiratory sounds*, wheez*, obstructive lung disease* and obstructive airway disease*. Trials of -blockers were sought by using the terms adrenergic antagonist*, sympatholytic* and adrenergic receptor block*. Trials were not excluded on the basis of language. The search was further augmented by scanning references of identified articles, reviews, and abstracts at clinical symposia. Study Selection Two investigators independently evaluated studies for inclusion. In choosing articles, investigators were blinded to results but not to journal, author, or institution of studies. The observed interrater agreement for the assessment of inclusion was calculated as a percentage. For all clinical trials identified from the search, investigators determined whether the -blocker used was cardioselective and whether it was considered to have intrinsic sympathomimetic activity (1, 26-36). Studies were evaluated if intravenous or oral cardioselective -blockers were administered as a single dose or as continued treatment lasting 3 days or longer. Single-dose trials were included if 1) FEV1 at rest was reported, either as liters or as a percentage of the normal predicted value at baseline and at follow-up; 2) 2-agonists were withheld for at least 8 hours before initial FEV1 measurement; 3) patients were not selected on the basis of previous response to -blockers; 4) the study was randomized, placebo-controlled, and single- or double-blinded; and 5) only patients with documented reactive airway disease were included. Reactive airway disease was demonstrated by a mean increase of at least 15% in FEV1 in response to 2-agonist, response to methacholine challenge, or presence of asthma as defined by the American Thoracic Society (37). Crossover trials were included if different interventions were administered in random order. We decided a priori that inclusion criteria 3, 4, and 5 would be applied to trials of continued treatment. Studies of continued treatment were included if they did not report FEV1 but instead evaluated the amount of 2-agonist use and respiratory symptoms compared with placebo. Trials were also included if 2-agonists were not withheld during the trial. Assessment of Validity The methodologic quality of each trial was assessed according to the following factors: 1) Was the study randomized? If so, was the randomization procedure adequate, and was allocation concealed? 2) Were the patients and people administering the treatment blinded to the intervention? 3) Were withdrawals and dropouts described, and was the analysis performed on an intention-to-treat basis? On the basis of these criteria, studies were broadly subdivided as all quality criteria met (A), one or more quality criteria only partially met (B), or one or more criteria not met (C). Clinical trials that did not meet criteria for inclusion but gave information on FEV1 response to cardioselective -blockers in patients with reactive airway disease were analyzed separately and used in a sensitivity analysis. These included studies that were not placebo-controlled; did not document asthma criteria; did not give baseline FEV1 data; or, for single-dose studies, did not withhold 2-agonists for 8 hours before measurements. Study Characteristics The main intervention of interest was intravenous or oral cardioselective -blockers versus placebo, given as a single dose or as continued treatment. Administration of a 2-agonist, intravenously or by inhalation, after the study medication or after placebo was also studied. Each 1-blocker used was classified into one of two categories: 1-blockers without intrinsic sympathomimetic activity, and 1-blockers with intrinsic sympathomimetic activity. Data Extraction Two investigators independently extracted data on change in mean group FEV1 in response to placebo or study drug; response of FEV1 to 2-agonist administered after placebo or study drug; symptoms reported during the trial, such as wheezing, dyspnea, or exacerbation of asthma; and, for trials of continued treatment, weekly use of inhaled short-acting 2-agonists. Data Synthesis The ratio of the lowest group FEV1 value after administration of study drug to baseline FEV1 was measured for placebo and active treatment and was recorded as the percentage change from baseline. The placebo response was then subtracted from the treatment response to obtain the net treatment effect, reported as a percentage of the baseline FEV1 value. For response to 2-agonists given after treatment or placebo, the new baseline value was the mean group FEV1 value obtained after study drug but before 2-agonist administration. The net treatment effect was estimated by calculating the ratio of FEV1 measured after agonist administration to the new baseline value for both placebo and active treatment and then subtracting the placeboagonist response from the treatmentagonist response. Whenever possible, the SD for the net treatment effect was calculated from individual-patient data or P values and was then used to derive the SDs for the analysis. Some trials provided SDs for treatment response and placebo response separately. For trials that reported no information on SDs, the average SD was obtained from trials that provided such data, calculated separately for placebo, treatment, and -agonist responses. Sensitivity analyses were performed to evaluate the effect of including these trials by using the lowest and highest available SD in place of the pooled SD and also by excluding these trials from the analysis. The Appendix Table shows the method used to obtain SDs for each trial. The mean treatment effects were pooled to obtain a weighted average of the study means using the fixed-effects model for continuous outcomes (38, 39). Confidence intervals with 95% significance were obtained for the pooled study means. The analysis was performed by using Meta View 4.1 (Cochrane Library software [Update Software, Oxford, United Kingdom]). Results for respiratory symptoms were measured as a risk difference by subtractin

Mortality Associated with Hormone Replacement Therapy in Younger and Older Women: A Meta-analysis

OBJECTIVE: To assess mortality associated with hormone replacement in younger and older postmenopausal women.DESIGN: A comprehensive search of medline, cinahl, and embase databases was performed to identify randomized controlled trials of hormone replacement therapy from 1966 to September 2002. The search was augmented by scanning selected journals through April 2003 and references of identified articles. Randomized trials of greater than 6 months’ duration were included if they compared hormone replacement with placebo or no treatment, and reported at least 1 death.MEASUREMENTS: Outcomes measured were total deaths and deaths due to cardiovascular disease, cancer, or other causes. Odds ratios (OR) for total and cause-specific mortality were reported separately for trials with mean age of participants less than and greater than 60 years at baseline.MAIN RESULTS: Pooled data from 30 trials with 26,708 participants showed that the OR for total mortality associated with hormone replacement was 0.98 (95% confidence interval [CI], 0.87 to 1.12). Hormone replacement reduced mortality in the younger age group (OR, 0.61; CI, 0.39 to 0.95), but not in the older age group (OR, 1.03; CI, 0.90 to 1.18). For all ages combined, treatment did not significantly affect the risk for cardiovascular or cancer mortality, but reduced mortality from other causes (OR, 0.67; CI, 0.51 to 0.88).CONCLUSIONS: Hormone replacement therapy reduced total mortality in trials with mean age of participants under 60 years. No change in mortality was seen in trials with mean age over 60 years.

Cardioselective beta-blockers for chronic obstructive pulmonary disease: a meta-analysis.

Beta-blocker therapy has a mortality benefit in patients with hypertension, heart failure and coronary artery disease, as well as during the perioperative period. These drugs have traditionally been considered contraindicated in patients with chronic obstructive pulmonary disease (COPD). The objective of this study was to assess the effect of cardioselective beta-blockers on respiratory function of patients with COPD. Comprehensive searches were performed of the EMBASE, MEDLINE and CINAHL databases from 1966 to May 2001, and identified articles and related reviews were scanned. Randomised, blinded, controlled trials that studied the effects of cardioselective beta-blockers on the forced expiratory volume in 1 s (FEV1) or symptoms in patients with COPD were included in the analysis. Interventions studied were the administration of beta-blocker, given either as a single dose or for longer duration, and the use of beta2-agonist given after the study drug. Outcomes measured were the change in FEV1 from baseline and the number of patients with respiratory symptoms. Eleven studies of single-dose treatment and 8 of continued treatment were included. Cardioselective beta-blockers produced no significant change in FEV1 or respiratory symptoms compared to placebo, given as a single dose (-2.05% [95% CI, -6.05% to 1.96%]) or for longer duration (-2.55% [CI, -5.94% to 0.84]), and did not significantly affect the FEV1 treatment response to beta2-agonists. Subgroup analyses revealed no significant change in results for those participants with severe chronic airways obstruction or for those with a reversible obstructive component. In conclusion, cardioselective beta-blockers given to patients with COPD do not produce a significant reduction in airway function or increase the incidence of COPD exacerbations. Given their demonstrated benefit in conditions such as heart failure, coronary artery disease and hypertension, cardioselective beta-blockers should be considered for patients with COPD.

Meta-Analysis: Respiratory Tolerance to Regular β2-Agonist Use in Patients with Asthma

Context Although 2-agonists are an important component of asthma therapy, studies indicate that tachyphylaxis may make regular use of these agents counterproductive. Contribution Pooled data from 22 trials of regular 2-agonist use versus placebo showed that regular use leads to decreased responsiveness to both the bronchodilator and nonbronchodilator effects of subsequent 2-agonist administration and increased airway inflammation. Implications Patients with asthma who use 2-agonists regularly may not respond to subsequent doses of 2-agonists as well as do patients who had not previously been using these agents, and they may have increased airway inflammation compared with those not taking the medicine at all. The Editors Asthma is a chronic illness characterized by inflammation and reversible airflow obstruction (1). Bronchodilators, such as 2-adrenergic agonists, can produce acute improvements in FEV1, an effect that lasts for a few to several hours (2). However, the duration of response following multiple doses is considerably shorter (2). There is evidence that 2-agonist therapy acutely protects against challenges from allergens and other bronchoconstrictive stimuli, but tachyphylaxis to its bronchoprotective effect is demonstrated after only a few doses (2). Data from randomized, controlled trials have indicated that regular use of 2-agonists results in tolerance to bronchodilating and nonbronchodilating effects, increased airway inflammation, and an increase in asthma exacerbations (3-5). Tachyphylaxis to the beneficial effects of 2-agonists may contribute to the association between use of these agents and increased asthma mortality rates (6, 7). The objective of this analysis was to pool data from randomized, placebo-controlled trials to evaluate the effect of regular 2-agonist use on respiratory function and 2-receptor function in patients with asthma. Methods Search Strategy We performed a comprehensive search of the EMBASE, MEDLINE, and CINAHL databases to identify all relevant clinical trials that were published between 1966 and June 2003 and examined 2-agonist use in patients with reactive airway disease. The search was performed by using the terms bronchodilator, sympathomimetic, adrenergic beta-agonist, albuterol, salbutamol, bitolterol, isoetharine, metaproterenol, salmeterol, terbutaline, fenoterol, formoterol, procaterol, isoproterenol, reproterol, eformoterol, or bambuterol, and asthma *, bronchial hyperreactivity, respiratory sounds, wheez *, respiratory hypersensitivity, obstructive lung disease, obstructive airway disease, obstructive pulmonary disease, or COPD. Trials were not excluded on the basis of language. The search was further augmented by scanning references of identified articles and reviews. Study Selection Two investigators independently evaluated studies for inclusion. The observed percentage agreement between the investigators for the assessment of inclusion was calculated by using the statistic (8). Trials were included in the review if they were randomized, placebo-controlled trials in which researchers administered regular inhaled or oral 2-agonists for at least 1 week to patients with asthma and did not allow as-needed 2-agonist use in the placebo group. Assessment of Validity The methodologic quality of each trial included in the meta-analysis was assessed for each of the following individual quality domains: 1) Was the randomization procedure adequate, and was allocation concealment described? 2) Were the patients and people administering the treatment blind to the intervention? 3) Were dropouts and withdrawals reported, and was analysis performed by intention to treat? Trials were characterized for each of the 3 domains as meeting all quality criteria (A), only partially meeting one or more quality criteria (B), or not meeting one or more criteria (C) (9, 10). Two reviewers independently assessed the trials for each of the quality scores, and interrater agreement was calculated by using the statistic. Study Characteristics The main intervention was inhaled or oral 2-agonists given regularly over a course of 1 to 6 weeks compared with placebo. The trials were all crossover in design and had at least 2 study periods. Each trial had a run-in period or a washout period of 1 to 4 weeks, without any 2-agonists, before administration of active treatment or placebo. Inhaled ipratropium bromide was allowed as rescue medication throughout the trials. Baseline FEV1 was measured before each study period, and post-treatment FEV1 was measured at least 6 hours after administration of the last dose of the study drug. Another intervention studied was the administration of a short-acting 2-agonist, given repeatedly in increasing doses, after the course of study drug or placebo. Researchers measured FEV1 several times from 1 to 4 hours after 2-agonist administration. A third intervention was administration of bronchoconstrictive challenges with methacholine, adenosine monophosphate (AMP), or allergen after the course of study drug or placebo and subsequent measurement of FEV1. Data Extraction Two independent reviewers extracted data from the selected articles, reconciling differences by consensus. In addition, we attempted to contact the investigators of included studies to obtain more information for data extraction. The outcomes measured were change in mean FEV1 after a course of placebo or study drug; acute FEV1 response to subsequent 2-agonist administered after the study periods, measured as peak change in FEV1 or area under a doseresponse curve; provocative concentration of methacholine, AMP, or allergen challenge that caused a 20% reduction in FEV1 (PC20); and in vitro assessments of leukocyte 2-receptor density, binding affinity, or AMP response to isoproterenol. We chose to evaluate these outcomes because they were the most common measurements used to assess bronchodilation, bronchial provocation, and 2-receptor function. Data Synthesis To evaluate the effect of regular 2-agonist treatment on FEV1, we determined the mean FEV1 after the course of treatment and the net change in FEV1 from baseline for both the placebo and active treatment groups. We subtracted the placebo results from the treatment results to obtain the net treatment effect. For the subsequent acute response to 2-agonist administered after the course of treatment or placebo, the new baseline used was the mean FEV1 measured after administration of the study drug but before subsequent 2-agonist administration. We measured the net peak change in FEV1 for the treatment and placebo groups and reported it as a percentage of the placebo results. The placeboagonist response was then subtracted from the treatmentagonist response to obtain the net treatment effect, reported as a percentage of the placebo response. For the 2-agonist doseresponse curve, the area under the curve was reported as a percentage of the placebo results, and the placebo response was subtracted from the treatment response. For the analyses that calculated a treatment effect from baseline, the investigators of each trial provided the SDs of the distribution separately for baseline and post-treatment values for both the active treatment and placebo groups. The baseline and post-treatment SDs were then combined for the analysis, assuming independence. To assess the bronchial provocation response seen after administration of study drug or placebo, the PC20 of methacholine, AMP, or allergen was recorded. The results were then log10-transformed for analysis, with the assumption of a normal distribution for the log-PC20. For each study, we obtained the mean values and SDs of the log-PC20 separately for the treatment and placebo groups. The log-PC20 for placebo was subtracted from the treatment response to obtain the net treatment effect for each study. In addition, we assessed tolerance to the bronchoprotective effect of the 2-agonist by subtracting the log-PC20 for the first dose of active treatment from the log-PC20 measured after the last dose. For in vitro studies of 2-receptor function, the mean values and SDs for receptor density, receptor affinity, and AMP response to isoproterenol were obtained separately for the treatment and placebo groups. We recorded the treatment results as a percentage of the placebo results to obtain the net treatment effect for each study. The net treatment effects for each analysis were pooled to obtain a weighted average of the study means by using the random-effects model for continuous outcomes (11). The random-effects model was used because it accounts for the possibility of significant interstudy heterogeneity. We obtained 95% CIs for the summary study means. The analyses were performed by using MetaView 4.1 (Cochrane Library software, Update Software, Oxford, United Kingdom). To test for interstudy variability, we calculated the chi-square value for the assumption of homogeneity; a P value less than 0.1 indicated statistical significance. A subgroup analysis, chosen a priori, was performed to evaluate the effect of concomitant inhaled corticosteroids on the response to regular 2-agonist use. For trials that provided information on treatment with and without concomitant corticosteroids, we analyzed the data for each subgroup separately. The subgroups of patients with and without concomitant inhaled corticosteroids were compared by evaluating the P value for the difference in the results. Role of the Funding Source The funding for this analysis came from salary support for Dr. Shelley R. Salpeter and Dr. Ormiston through Santa Clara Valley Medical Center. The hospital had no role in the design, conduct, or reporting of the study. Data Synthesis Search Results The Appendix Figure shows the results of the search for articles. The electronic database search identified approximately 5000 articles, 192 of which were potentially relevant trials on 2-agonist use in patients with reactive airway disease. After scanning selected

Encephalopathy and Hyperammonemia Associated with Valproic Acid

Objective: To discuss the case of a patient who experienced an altered level of consciousness secondary to hyperammonemia associated with valproic acid who had normal drug concentrations and liver function tests. Case Summary: A 48-year-old man with a history of seizure disorder treated with valproic acid presented with an altered level of consciousness. An extensive workup revealed only an increased ammonia level. His medications were withheld. Within 24 hours, the ammonia level returned to normal and he was able to be aroused. His antiseizure medication was changed to gabapentin. All of his chronic medications, other than valproic acid, were restarted. The patient has had no further increases in ammonia levels or episodes of altered mental status. Discussion: Encephalopathy due to hyperammonemia is generally due to severe liver dysfunction. In our case, however, hyperammonemia occurred in a patient with normal liver function who was also taking valproic acid. An objective causality assessment revealed that the adverse drug event was probably related to valproic acid. A review of the literature indicates that the increased ammonia levels were probably secondary to the effect of valproic acid on the kidneys or the liver. Withdrawal of the offending agent resulted in normalization of ammonia levels and resolution of encephalopathy. Conclusions: With a widening of its clinical indications, the use of valproic acid has been increasing. Even though valproic acid concentrations may be therapeutic/nontoxic and liver function tests normal, practitioners are encouraged to check the ammonia level. Hyperammonemia, once recognized, is easily treated.