Weight loss interventions in asthma: EAACI Evidence-Based Clinical Practice Guideline (Part I)

Authors


  • Edited by: Thomas Bieber

Correspondence

André Moreira, Faculty of Medicine, University of Porto and Hospital São João, Porto, Portugal.

Tel.: +351 225026470

Fax: +351 225025766

E-mail: andremoreira@med.up.pt

Abstract

Background

Asthma and obesity are chronic multifactorial conditions that are associated with gene–environment interaction and immune function. Although the data are not fully consistent, it seems that obesity increases the risk of asthma and compromises asthma control.

Objective

To investigate the impact that weight changes have on asthma.

Methods

We carried out a systematic review of three large biomedical databases. Studies were scrutinized and critically appraised according to agreed exclusion and inclusion criteria. Quality assessment of eligible papers was conducted using the GRADE method. Meta-analyses of comparable studies were carried out.

Results

Thirty studies met the eligibility criteria of the review. Interventions were limited to dietary manipulation in three studies, one of which also used anti-obesity drugs, and bariatric surgery in four. All the other studies reported observational data. Becoming obese increased the odds for incident asthma by 1.82 (95% CI 1.47, 2.25) in adults and 1.98 (95% CI 0.71, 5.52) in children. Weight loss was associated with significant improvement in mean scores for symptoms, rescue medication score, and asthma exacerbations in the only randomized controlled trial. Similarly, evidence gathered from observational studies, with follow-up ranging between 8 weeks to 1 year, and from changes 1 year after bariatric surgery showed improvements in all asthma control-related outcomes. Changes in lung function were reported in one randomized controlled and eight observational studies of asthmatic subjects, with conflicting results. Either improvement after weight loss, decline with weight gain, or no effects at all were reported. Changes in airway inflammation and responsiveness were reported only by observational studies.

Conclusion

Weight increases above the obesity threshold significantly increase the risk of asthma. The available studies show weak evidence of benefits from weight reduction on asthma outcomes.

Background

Obesity is a chronic disease that is complex and multifactorial in its nature. Its development is dependent upon the interaction between genetics and environment. According to most recent data, obesity and overweight have reached epidemic proportions in Westernized countries. In the United States, obese or overweight subjects represent more than two-thirds of the adults, and in Europe, prevalence of obesity has risen threefold or more since the 1980s, even in countries with traditionally low rates [1]. The importance of obesity as a risk factor for diseases including type 2 diabetes, hypertension, and atherosclerosis has been recognized for long time. In the last decade, it has also been increasingly understood as a risk factor for asthma [2].

Evidence suggests that obesity both increases the risk of incident asthma and changes prevalent asthma toward a more difficult-to-control phenotype. A recent meta-analysis showed that being overweight or obese increased the odds of incident asthma in a dose-dependent manner [3]. Evidence from animal experimental models suggests that increased serum leptin and decreased serum adiponectin levels, which are observed in obese patients, may exacerbate some features of the allergic airway inflammation [4]. Furthermore, obesity impacts on the response to asthma treatment [5]. Taken together, the evidence supports an obese–asthma endotype with increased severity of illness and a variable response to therapy most likely related to a specific underlying pathophysiology of the disease.

The association between obesity and asthma satisfies a few criteria of the causal relationship often observed between a disease and a risk factor such as consistency, dose–response relationship, biological plausibility, and a correct temporal order. Reversibility would presuppose that if overweight antedates asthma, then weight loss would improve asthma-related outcomes. Studies reporting the effects of weight loss on asthma are mainly observational noncontrolled reports after surgical or dietary interventions. Although the suggestion that weight loss may be associated with improvement in lung function, medication use, symptoms, morbidity, quality of life, and health status, evidence supporting a distinct therapeutic approach in the obese–asthma phenotype is currently lacking.

Objective

We aimed to assess the evidence of the impact of weight changes on selected outcomes of asthma.

Methods

Criteria for eligibility

Type of studies

Published reports of randomized, controlled trials and reports of observational studies, including case series, with either a cohort design (i.e., with concurrent selection of controls) or a case–control design, were included.

Participants

Individuals of all ages with a clinical diagnosis of asthma, physician-diagnosed, self-reported physician-diagnosed, or defined as exercise-induced bronchoconstriction, of any gender and age group, unrestricted by disease severity, previous or current treatment.

Type of interventions

Studies comparing or assessing the effect of weight changes, naturally or induced, on asthma-related outcomes.

Type of outcome measures

The primary outcome was the number of asthma exacerbations requiring hospitalization or oral steroids. Secondary outcomes included measures reflecting chronic asthma control, such as changes in: (i) symptoms, quality of life, and use of rescue short-acting ß2-agonists; (ii) lung function tests; (iii) airway responsiveness; and (iv) airway inflammation, adverse events, study withdrawal, and incident asthma.

Search methods for identification of studies

Electronic searches

In December 2011, electronic searches were undertaken in these databases: MEDLINE, ISI Web of Science, LILAC, and The Cochrane Library. In the absence of a gold standard for the definition of asthma, we used the following keywords (first group): ‘asthma’, ‘wheez*’, ‘broncho*’ coupled with (second group) ‘obese’, ‘obesity’, ‘fat’, ‘adiposity’, ‘body fat’, ‘body composition’, ‘body weight’, ‘weight gain’, ‘overweight’, ‘weight loss’, ‘body mass index’ and (third group) ‘change’, ‘difference’, ‘incidence’, ‘intervention’, and ‘effect’. A priori inclusion criteria limited retrieved articles to those that assessed weight loss, weight changes, or weight control strategies on asthma-related outcomes, and subsequently, each study was evaluated to determine whether it met the entry criteria. Only manuscripts in English were included in the final review.

Searching other resources

Hand searches were performed of the reference lists of all pertinent reviews and studies examined. Abstracts from relevant conferences were searched.

Data collection and analysis

Selection of studies

Following electronic literature search, two review authors (AM and JF) independently selected articles on the basis of title or abstract or both for full-text scrutiny. The authors agreed a list of articles that were retrieved, and they subsequently assessed each study to determine whether it met the entry criteria of the guidelines. Figure 1 shows the results of the systematic literature search.

Figure 1.

Results of systematic literature search.

Data extraction and management

Aspects from each study were assessed including the following: design (description of randomization, blinding, number of study centers and location, number of study withdrawals); participants (sample size, mean age, age range of the study); intervention (type and dose, study duration); outcomes (type of outcome analysis, outcomes analyzed).

Quality assessment of published evidence (GRADE system)

Analyses within this review were based on the outcome measure. For studies that were comparable with respect to the outcome, results were pooled using a random effects meta-analysis. This was only possible for incident asthma because no study reported significant baseline differences between cases and control subjects. Additionally, publication bias was tested by inspecting the funnel plot. For these analyses, Review Manager software (version 5.0) was used [6].

We present the overall quality of the evidence using the GRADE approach as recommended by the Cochrane Handbook for Systematic Reviews of Interventions [7]. That is, for each specific outcome, the quality of the evidence was assessed for five factors: (i) limitations of the study design or the potential for bias across all studies that measure that particular outcome, (ii) consistency of results, (iii) directness (generalizability), (iv) precision (sufficient data), and (v) the potential for publication bias. The overall quality was considered to be higher if multiple randomized controlled trials with a low risk of bias provided consistent, generalizable results for the outcome. The quality of evidence was downgraded by one level if one of the factors described above was not met. Likewise, if two or three factors were not met, we downgraded the level of evidence by two or three levels, respectively. Thus, the GRADE approach resulted in four levels of quality of evidence: high, moderate, low, and very low. In the case of only one study measuring an outcome, the data were considered to be ‘sparse,’ and subsequently, the evidence was labeled as ‘low-quality evidence.’ If only one study was present for a given comparison, the results are described in the text. GRADE profiler software (version 3.2) was used [8].

Results

A summary of findings (SoF) table is presented in Table 3.

Results of the search

The search for relevant studies was carried out in December 2011 and identified 2112 potentially relevant publications (1982 PubMed, 69 ISI web of science, 18 Cochrane Library, 41 LILACS, 2 abstracts). From these, 1952 papers were excluded on the basis of title and abstract, 113 were duplicates, and 46 were retrieved as full articles for detailed evaluation and assessed in duplicate. From these, 30 met the eligibility criteria of the review and were included (Table 1). Excluded studies [5, 9-23] and reason for are presented in Table 2. See Data S3 for a description of studies.

Table 1. Studies on the systematic review of the effects of weight changes on asthma
ReferenceStudy design and participantsOutcomesAssociations/effects observed
Camargo et al. [24]Prospective cohort (Nurses' Health Study II): 85 911 adults, aged 26–46 years of age followed for previous 4 yearsIncidence of self-report of physician-diagnosed asthma with use of asthma medicationRelative risk of asthma increased by 1.5 (95% CI, 1.1–2.1), 1.5 (95% CI, 1.0–2.3), and 3.8 (95% CI, 2.9–5.0) for subjects with baseline BMI between 25.0 and 27.4, 27.5 and 29.9, and over 30, respectively
Hakala et al. [36]Observational, 14 obese–asthmatic patients (11 women, aged 25–62 years), before and after 8 weeks of a very-low-calorie dietLung volumes, airways resistance and conductanceAs patients decreased, BMI, FEV1, FEF25–75, FVC, functional residual capacity, and expiratory reserve volume increased, while airway resistance and resting minute ventilation decreased
Stenius-Aarniala et al. [29]One-year randomized controlled trial; 38 obese–asthmatic patients; weight reduction program vs standard careAsthma symptoms, control and quality of life, and lung functionSignificant differences were observed in changes for: symptom scores −12 (range −1 to −22, P = 0.04); total asthma-related scores −10 (−18 to −1, P = 0.02); asthma exacerbation (1 (0−4) vs 4 (0−7) in control group, P = 0.001); lung function FEV1 7.6% (1.5–13.8%, P = 0.02), and 7.6% for FVC (3.5–1.8%, P = 0.001)
Beckett et al. [48]Prospective cohort; 4547 adults aged 18–30 years, followed for 10 yearsIncidence of self-report of physician-diagnosed asthmaRelative risk of asthma increased in the highest quintile of BMI change by 1.72 (95% CI: 1.25–2.38) in women but not in men
Castro-Rodriguez et al. [25]Prospective cohort (Tucson Children's Respiratory Study) children followed between 5 and 11 years of ageIncidence of asthma defined by positive response to albuterolFemales who became overweight or obese were more likely to develop bronchodilator responsiveness.
Chen et al. [49]Prospective cohort (Canadian National Population Health Survey); 9149 subjects, aged 20–64, followed for 2 yearsIncidence of self-report of physician-diagnosed asthmaRelative risk of asthma increased in obese women by 1.9 (95% CI: 1.1, 3.4) but not in men 1.1 (95% CI: 0.3, 3.6)
Gold et al. [26]Prospective cohort; 9828 children aged 6–14 years, followed for 5 yearsIncidence of parent report of physician-diagnosed asthmaRelative risk of asthma decreased among boys in the middle ranges of annual changes in BMI [RR 0.36 (95% CI: 0.18–0.74 for quintile], while in girls increased in the 5th quintile [RR of 2.20 (95% CI 1.13–4.28)]
Huovinen et al. [50]Prospective cohort (Finnish Twin Cohort); 10 597 adult twins, followed for 9 yearsIncidence of self-report of physician-diagnosed asthmaNo significant effect of previous weight change (between 1975 and 1981) on the risk of incident asthma: respectively, OR of 1.04 (0.99–1.09) and 1.00 (0.94–1.05) for men and women
Romieu et al. [51]Prospective cohort (E3N Cohort Study); 67 229 women aged 40–65 followed for 3 yearsIncidence of self-report of physician-diagnosed asthmaNonsignificant risk of asthma increase [1.17 (95% CI, 0.91–1.50) and 1.37 (95% CI, 0.99–1.89), respectively, for those who gain more than 5 kg or loss between 1 to 5 kg during the follow-up]
Oddy et al. [27]Prospective birth cohort (Western Australian Pregnancy Cohort Study); 2602 children followed for 6 yearsIncidence of self-report of physician-diagnosed asthmaOdds ratio for asthma increased by 1.56 (95% CI = 1.02, 2.38; P = 0.038) per unit increase in BMI
Aaron et al. [38]Observational study, 58 obese women 24 of whom had asthma, 6-month weight loss programLung function, airway responsiveness, and disease-specific health statusFor every 10% relative loss of weight, the FVC improved by 92 ml (P < 0.05) and FEV1 improved by 73 ml (P = 0.04); airway reactivity did not change
Ford et al. [52]Prospective cohort (NHANES I); 9456 participants aged 25–74 followed between 1971–1975 to 1982–1984Incidence of self-report of physician-diagnosed asthmaBMI was not significantly associated with asthma incidence (OR 1.52, 95% CI 0.62–3.77)
Gunnbjornsdottir et al. [53]Prospective cohort (ECRHS); 16 191 participants followed between 5–10 yearsIncidence of self-report of physician-diagnosed asthmaRelative risk of asthma increased by 1.67 (1.24–2.25) obese subjects.
Nystad et al. [54]Prospective cohort; 135 000 Norwegians aged 14–60 years followed for 21 yearsIncidence of self-report of physician-diagnosed asthmaRelative risk of asthma increased by 1.27 (95% CI: 1.13–1.43) for men and 1.30 (95% CI: 1.17–1.45) for women; and by 1.78 (95% CI: 1.35–2.34) for men and 1.99 (95% CI: 1.67–2.37) for women, respectively, overweight or obese
Hasler et al. [55]Prospective cohort (Zurich Cohort Study); 591 participants followed between ages 20 and 40 years oldIncidence of self-report of physician-diagnosed asthmaObesity did not increased risk of future asthma
Beuther and Sutherland [3]Systematic review with meta-analysis of studies assessing the relationship between categories of BMI and incident asthma in adultsOdds for incident asthmaOR for incident asthma was 1.38 (95% CI, 1.17–1.62) and 1.92 (95% CI, 1.43–2.59), respectively, for overweight and obesity compared with normal weight
Mamun et al. [28]Prospective cohort (Mater University Study of Pregnancy); 2911 participants followed between 5 and 14 years of ageIncidence of self-reported asthmaA greater increase in BMI z-score from the age of 5 was associated with the development of asthma at 14-yr follow-up
Johnson et al. [30]Observational, 10 obese–asthmatic patients, 8 weeks on alternate day calorie restriction dietary regimenAsthma control and quality of life, lung functionSignificant improvement in all outcomes within 2 weeks of diet initiation
Maniscalco et al. [37]Prospective cohort; 12 female subjects with physician-diagnosed asthma followed 1 year after bariatric surgeryAsthma control questionnaire score, lung function, airway inflammation (exhaled nitric oxide)Weight reduction was associated with improvement of respiratory symptoms, rescue medication use, and lung function
Jartti et al. [56]Retrospective cohort (Cardiovacular Risk in Young Finns Study); 2624 subjects aged 24–39 years followed for previous 21 yearsIncidence of self-report of physician-diagnosed asthmaIncrease in was BMI associated with incident asthma during adulthood (OR: 1.08; 95% CI, 1.01–1.17) but not in adolescence/young adulthood; risk of asthma increased 25% when BMI increased from 25 to 30 kg/m2
Jamrozik et al. [57]Case–control study (Busselton Health Study); 1554 adults aged 20–69 years followed for previous 15 yearsIncidence of self-report of physician-diagnosed asthmaIncrease in BMI associated with incident asthma (OR: 1.09; 95% CI, 1.02–1.18)
Coogan et al. [58]Prospective cohort (Black Women's Health Study); 46 435 adult women, aged 21–69 years, followed for previous 10 yearsIncidence of self-report of physician-diagnosed asthma with use of asthma medicationRelative risk increased by 1.26 (95% CI, 1.05–1.51), 1.62 (95% CI, 1.32–1.98), 2.24 (95% CI, 1.76–2.84), or 2.85 (95% CI, 2.19–3.72) for subjects with baseline BMI between 25–29, 30–34, 35–39, and over 40, respectively
Marcon et al. [41]Prospective cohort within European Community Respiratory Health Survey (1991–1993) of 638 subjects with asthma aged 20–44 years, followed up from 1998 to 2002FEV1 declineFEV1 decline was associated with weight gain in men (20; 95% CI, 10–30, ml/year/kg) and in women (6; 95% CI, 1–11, ml/year/kg)
Sutherland et al. [39]Post hoc analysis of data from 10 studies of the Asthma Clinical Research Network; 1265 adult asthmaticsTreatment response, lung function, airway inflammation, airway responsiveness, and asthma clinical impairmentNo differences between the lean vs overweight/obese with regard to lung function (FEV1 and FEV1/FVC ratio), airway hyper-responsiveness, airway inflammation (fraction of exhaled nitric oxide), or clinical status (quality of life, symptoms, and rescue beta-agonist use)
Sikka et al. [32]Retrospective cohort; 40 patients with physician-diagnosed asthma followed 1 year before and 1 year after bariatric surgeryAsthma control assessed by medication useThe mean difference (SD) decrease in the postsurgical year was of −3.2 (6.0) respiratory medication prescription fills annually (P < 0.002). Compared with the year before surgery, 72.5% of patients filled fewer prescriptions, 17.5% filled the same number of prescriptions, and 10% filled more respiratory prescriptions in the postsurgical period.
Reddy et al. [40]Prospective cohort; 257 patients with physician-diagnosed asthma followed 1 year after bariatric surgeryAsthma control assessed by medication useThe number of patients regularly using oral corticosteroids, daily inhaled corticosteroids, or inhaled bronchodilators to control their asthma all decreased significantly at 1-year follow-up. One hundred and one (40%) of these patients became medication-free. There were no significant changes in patients using antileukotrienes.
Dixon et al. [35]Prospective cohort; 23 patients with physician-diagnosed asthma followed 1 year after bariatric surgeryAsthma control and quality of life questionnaire scores, lung function, airway responsivenessAsthma control and quality of life improved at 1-year follow-up. Airway responsiveness improved in patients with normal IgE levels, suggesting atopic status affects this association
Fida et al. [34]Cross-sectional; 3737 women (participants Omega study), mean age 32.6 ± 4.5 years oldIncidence of self-reported asthmaAdult diagnosed asthma was positively associated with adult weight change higher than 20 kg
Dias-Junior et al. [31]6 months, controlled trial; 33 obese subjects with difficult-to-treat asthma; low-energy diet and anti-obesity drugs (orlistat, sibutramine) targeted to a weight reduction of at least 10%Asthma control questionnaire scores, use of rescue medication and FEV1Subjects in the treatment group who achieved the targeted weight loss had a significant reduction in asthma control scores and in doses of rescue medication. No other significant changes were observed
Lombardi et al. [33]Prospective cohort; 14 patients with physician-diagnosed asthma followed 1 year after bariatric surgeryAsthma symptoms, lung function, and exhaled nitric oxideOnly exhaled nitric oxide decreased during the follow-up
Table 2. Excluded studies and reasons for so
ReferenceDesign and participantsOutcomesResultsReasons for exclusions
Chinn and Rona [9]Cross-sectional; children participating in the British National Study of Health and GrowthTrends in the obesity–asthma associationTrends in overweight and obesity do not explain the increase in asthmaThe study does not evaluate the impact of weight changes on asthma
Gilliland et al. [10]Prospective cohort (Children's Health Study); 3792 children followed for 5 yearsIncidence of parent report of physician-diagnosed asthmaBeing overweight or obese was associated with increased risk of incident asthmaThe study does not evaluate the impact of weight changes on asthma
Sin et al. [11]Prospective cohort for 10 years of all neonates born at term between 1985 and 1988; n = 83 595Risk of emergency visits for asthmaIncrease of 10% (95% CI, 2–19%) for every increment of 0.10 kg over a birth weight of 4.5 kgThe study does not evaluate the impact of weight changes on asthma
Hendler et al. [12]Prospective cohort (Asthma During Pregnancy Study); 1699 asthmatic and 867 controls followed during pregnancyAsthma control (prevalence of asthma exacerbations, hospitalizations, use of steroids, and change in severity)No differences in the rates of asthma improvement or deterioration between obese and nonobese. Obesity was associated with an increase in asthma exacerbations (OR 1.3, 95% CI 1.1–1.7)The study does not evaluate the impact of weight changes on asthma
Mannino et al. [59]Prospective cohort (National Longitudinal Survey of Youth); 4393 asthma-free children followed for up to 14 yearsIncidence of self-reported asthmaBMI above 85th percentile at age 2–3 increased risk factor subsequent asthma development in boys [hazard ratio (HR) 1.6, 95% CI: 1.1, 2.4] but not in girls (HR 0.8, 95% CI 0.5, 1.4)The study does not evaluate the impact of weight changes on asthma
Peters-Golden et al. [5]Post hoc analysis of four randomized controlled trials; 3073 adult asthmaticsAsthma control daysNormal-weight subjects had a higher percentage of asthma-controlled days than overweight or obeseThe study does not evaluate the impact of weight changes on asthma
Saint-Pierre et al. [13]Prospective cohort; 406 adult asthmatics followed for 6 monthsAsthma controlOverweight subjects had a lower risk of transiting from the unacceptable to the acceptable health stateThe study does not evaluate the impact of weight changes on asthma
Burgess et al. [14]Prospective cohort (Tasmanian Asthma Survey); 1494 subjects, aged 32 years old, followed for previous 25 yearIncidence of self-reported asthmaIn women, adiposity at 7 years of age was associated with current asthma at 32 yearsThe study does not evaluate the impact of weight changes on asthma
Rodrigo and Plaza [15]Cross-sectional; 426 patients with severe asthma exacerbationEmergency department stay and rate of hospitalizationOverweight or obese subjects had significant increases in length of emergency stay (2.3 vs 1.9 h, P < 0.01) and rate of hospitalization (13.7% vs 6.8%, P = 0.02)The study does not evaluate the impact of weight changes on asthma
Belle et al. [16]Retrospective cohort (The Longitudinal Assessment of Bariatric Surgery study) of 2559 patients with a BMI of at least 40 kg/m2 older than 18 years, who underwent a bariatric surgical procedureSelf-report of physician-diagnosed asthmaOdds ratios for asthma was 1.36 (95% CI: 1.11–1.68) in subjects with BMI over 50The study does not evaluate the impact of weight changes on asthma
Haldar et al. [17]Post hoc analysis of data from 3 clinical trials; 371 adults with asthmaAsthma clinical phenotypesFour clusters were defined: early-onset atopic; obese, noneosinophilic; early-onset symptom predominant; and late-onset inflammation predominantThe study does not evaluate the impact of weight changes on asthma
Vortmann and Eisner [18]Cross-sectional; 843 adults with severe asthmaQuality of life and asthma controlAsthma-specific quality of life was worse in the underweight and obese groups. Obese had a higher number of restricted activity daysThe study does not evaluate the impact of weight changes on asthma
Taveras et al. [19]Retrospective cohort (Project Viva) of 932 children followed from 6 months to 3 yearsRecurrent wheezing defined as parents report of wheezing between 2 and 3 years of age plus wheezing in either year 1 or 2 of lifeEach 1-unit increment in 6-month weight-for-length z-score was associated with recurrent wheezing (OR, 1.46; 95% CI, 1.11−1.91)The study does not refer to asthma but recurrent wheezing
Clerisme-Beaty et al. [20]Cross-sectional; 292 subjects with asthma, mean age of 47 yearsAsthma ControlNo association between obesity and asthma controlThe study does not evaluate the impact of weight changes on asthma
Ma et al. [21]Randomized clinical trial of the efficacy of a 12-month weight loss intervention in the management of asthma in obese adultsAsthma control, lung function, asthma-specific quality of life, asthma medication use, and asthma-related healthcare utilizationNot availableNo results
Scott et al. (abstract) [22]Ten weeks, randomized controlled trial; 30 obese–asthmatic subjects; calorie-restricted diet (n = 11), physical activity intervention (n = 9), or combination of these (n = 10)Asthma quality of life questionnaire scores, expiratory reserve volumeImprovement in ERV after caloric restrictionAlthough stated on the aim and methods, the results presented on the abstract do not give data on any outcomes of the current review
Holguin et al. [23]Cross-sectional; 1049 subjects from the Severe Asthma Research Program, aged 18–79 yearsInteraction between obesity and age of asthma onset (before or after 12 years of age)In subjects with early-onset asthma but not in subjects with late-onset asthma, there was a significant association between increasing BMI and duration of asthmaThe study does not evaluate the impact of weight changes on asthma

Findings

Asthma incidence

Incident asthma was the primary outcome in 16 of the 29 scrutinized studies (Table 1) with a combined total of 306 646 participants. The majority of papers defined incident asthma as a self-report of a medical diagnosis of asthma not present at baseline. However, in two studies, it was defined additionally by the use of asthma medication [24] or by a positive response to albuterol [25]. Four studies reported data on children aged below 14 who were followed for a period between 5 and 9 years [25-28]. All the other papers referred to adults with a follow-up length between 2 and 21 years. Data were presented in a way that allowed extraction in 10 studies facilitating meta-analysis. Obesity increased the odds for incident asthma by 1.82 (95% CI 1.47, 2.25) in adults and 1.98 (95% CI 0.71, 5.52) in children (Fig. 2). In absolute terms, obesity increased the risk of incident asthma as much as 20 more subjects per 1000 (from 12 more to 31 more). Although the incidence of asthma in obese subjects was considered a critical outcome, the quality of the evidence gathered was rated according with the GRADE initiative parameters as very low. All studies were observational and different in terms of populations, definition of cases, and length of follow-up. In terms of dose–response gradient, it was observed that the highest quintiles of body weight were more strongly associated with asthma.

Figure 2.

Forest plot showing meta-analysis of incident asthma risk comparing obese vs normal-weight subjects. Studies are ordered by year of publication, the center of each square represents studyspecific ORs, and horizontal lines represent 95% CIs; vertical vertex of the diamond represents the OR summary estimate, whereas the ends of the diamond (width) correspond to the 95% CI; vertical line indicates an OR of 1 (no difference).

A funnel plot of the effects of weight change on incident asthma suggested a possible shortage of smaller studies (Fig. 3).

Figure 3.

Funnel plot of precision estimates [calculated by using the inverse of the standard error (SE) of log (OR); that is, the higher the estimate, the more precise the study] from studies that explored the association between obesity and incident asthma. The vertical dashed line is the summary estimate of the odds ratio for all studies.

Weight loss interventions

Interventions were limited to dietary manipulation in three studies [29-31] and bariatric surgery in four [32-35]. In the randomized controlled trial by Stenius-Aarniala and coworkers, the treatment group took part in a weight reduction program that included 12 group sessions and lasted 14 weeks, including 8 weeks in which participants took a very-low-energy dietary preparation with 1760 kJ of energy and daily allowances of all essential nutrients [29]. One additional study reported a follow-up period of a 2-month alternate day caloric restriction dietary regimen, in which women were instructed to consume 320 calories and men 380 calories of a commercially available canned meal replacement shake (Atkins Advantage or Carb Solutions) [30]. On the other day subjects were told to eat ad libitum whatever they normally ate to the point of satisfaction but not to intentionally overeat. In the Dias-Junior et al. [31] study, subjects were given a prescription of the anti-obesity drugs orlistat and sibutramine over a 6-month period in addition to the supervised weight reduction program involving low-energy diet.

Asthma control-related measures

In the only randomized controlled trial, reporting 1 year of follow-up after a weight reduction program, there was a significant improvement in mean scores for symptoms, rescue medication score and asthma exacerbations in the intervention group [36]. The quality of this evidence was rated the highest; restricted only by the limited number of study participants.

Four observational studies, reporting data on 80 subjects, followed for a period ranging between 8 weeks to 1 year, showed improvement of respiratory symptoms [37], rescue medication use [37], asthma control [30] and quality of life scores [30, 38]. Evidence drawn from these studies was rated as very low. The study by Aaron et al. which enrolled subjects in an intensive 6-month weight loss program had no control group and also reported data on nonasthmatic subjects. Significant improvements in asthma control within 2 weeks of diet initiation was identified in one of the two studies that used the Asthma Control Questionnaire score as an outcome [30].

In the Dias-Junior et al. [31] study, the authors did not report the comparison between study arms but differences between subjects who achieved the targeted 10% weight loss independently of the intervention. With this significant selection bias, the observational controlled report on the effect of weight loss showed a significant reduction in symptom scores as assessed by the Asthma Control Questionnaire [31].

In a post hoc analysis of data from 10 studies by the Asthma Clinical Research Network referring to 1265 subjects, being overweight or obese was associated with a reduced effect of steroids on exhaled nitric oxide [39]. However, a longitudinal study of subjects allocated to the placebo arm lasting between 8 and 48 weeks in duration, found no substantial differences between BMI categories with regard to clinical outcomes, suggesting that overweight and obese subjects with mild-to-moderate asthma did not necessarily fare worse over time than their lean counterparts.

Four studies reported on asthma control assessed by medication use [32, 40], asthma control, and quality of life questionnaire scores [35] and asthma symptoms [33] 1 year after bariatric surgery. Improvements were seen in all outcomes; however this evidence was limited by study design and the lack of adequate controls.

Lung function

Changes in lung function were reported in 10 studies. In the clinical trial of Stenius-Aarniala et al. [29] lung function measured by FEV1 improved 7.6% (1.5–13.8%, P = 0.02) and by 7.6% for forced vital capacity (FVC) (3.5–1.8%, P = 0.001). Similar findings were reported by the same group in their follow-up of 14 obese–asthmatic patients, before and after 8 weeks of a very-low-calorie diet [36].

Data from the largest prospective cohort within the European Community Respiratory Health Survey of 638 adult subjects with asthma followed up for 4 years showed FEV1 decline was associated with weight gain in men (20; 95% CI, 10–30, ml/year/kg) and in women (6; 95% CI, 1–11, ml/year/kg) [41]. Weight loss benefits were reported in shorter follow-up even in a limited number of subjects [30, 38]. However, data from the Asthma Clinical Research Network failed to support these findings as no differences between the lean vs overweight/obese were seen [39]. This report was a post hoc analysis of existing clinical trial data, and in none of the studies, all including mild-to-moderate asthmatics was BMI an a priori stratification variable.

Two of the three studies reporting on the effects of bariatric surgery on lung function after the 1-year follow-up period [35, 37] showed modest improvements in clinical outcomes. The Lombardi et al. study [33] observed improvements. However, these improvements failed to reach statistical significance. Again, evidence from these studies was limited by their design, and sample size not large enough to detect differences; altogether these studies reported data obtained from 49 subjects.

Airway inflammation and responsiveness

Changes in levels of exhaled nitric oxide were reported by three observational studies. The largest showed no benefit [39], while conflicting results appeared with follow-ups after bariatric surgery [33, 37].

Two of the three studies reporting data on airway responsiveness, one referring to a post hoc analysis of 10 studies with a total of 1265 adult asthmatics, failed to show any benefit from weight reduction [38, 39]. In the prospective cohort of 23 patients with physician-diagnosed asthma followed for 1 year after bariatric surgery, airway responsiveness improved in patients with normal IgE levels, suggesting interference with the atopic status [35].

Discussion

In common with other systematic reviews, our findings are limited by the data provided by the included studies. Meta-analysis showed that weight gain, as much as required to become obese, almost doubled the odds of incident asthma. In relation with disease-related outcomes, the heterogeneity of the studies precluded a quantitative synthesis. High-quality evidence supported by randomized controlled trials came only from one study that showed benefits of losing weight on asthma control and lung function. Conflicting data appeared from the several observational studies that provided limited and low-quality evidence of the beneficial effects of losing weight on asthma-related outcomes. Nevertheless, taken together, these observations support the recommendation of weight management and tackling obesity as a part of asthma treatment.

Similar findings have been recently reported by a Cochrane review aiming to assess the effect of various interventions for weight loss on measures of asthma control and weight loss among overweight or obese patients with chronic asthma [42]. From the four studies conducted among adults, only one found a significant reduction in symptoms scores and a reduction in doses of rescue medication and improvement in lung function. Our review has some methodological differences. Using the GRADE approach, we also included observational studies aiming to assess the effect of weight change on incident asthma, not solely on chronic disease. Furthermore, due to our inclusion criteria, we failed to include the abstract by Scott et al. [22] as no results were given in relation to the quality of life–related scores nor do we considered changes in expiratory reserve volume as one asthma-related outcome. Also, the Spanish manuscript of Hernandez Romero [43] was not included due to language restrictions. In this trial, 96 obese adults, aged 18–71 with moderately persistent asthma, were randomized to receive either of the two different low-calorie diets, each providing 1200–1500 kcal/day, over a 40-day period. An improvement was seen, although the study did not report the measure of effect nor the associated statistical tests. Our study was compliant with the GRADE approach that had two main steps: evaluating the quality of evidence gathered (Data S2), specifically addressing issues such as study design, consistency, and precision of the results, and applicability of the evidence with respect to the populations, interventions, and settings where the proposed intervention may be used and the likelihood of publication bias. We conclude most of the studies were of relatively poor methodological quality with a high risk of performance and detection bias. Furthermore, we built up a SoF table (Table 3) that presents data for each outcome more clearly showing that only some studies contributed with information for a specific outcome.

Table 3. Summary of findings
Do weight changes impact on asthma?
Patient or population: asthmatic overweight or obese subjects Settings: No restrictions were made Intervention: natural or induced weight change
OutcomesIllustrative comparative risksa (95% CI)Relative effect (95% CI)No of Participants (studies)Quality of the evidence (GRADE)Comments
Assumed riskCorresponding risk
Standard careWeight change strategies
  1. a

    The basis for the assumed risk (e.g., the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI); CI, confidence interval; OR, odds ratio; FEV1, forced expiratory volume in the first second vital capacity.

  2. b

    Data pooling was possible in 10 studies.

  3. c

    Studies were different in terms of populations, definition of cases, and length of follow-up.

  4. d

    Highest quintiles of body weight increase are strongly associated with incident asthma.

  5. e

    Because studies were heterogeneous, we refrained from pooling and restrict the analysis to a qualitative overview.

  6. f

    Failure of accurate measurement of all known prognostic factors.

  7. g

    Inadequately short follow-up.

  8. h

    Two of the three studies had less than minimum sample size to detect minimal important differences, either for benefit or for harm.

  9. i

    One of the studies included 638 subjects (ECRHS).

  10. j

    Failure to adequately control confounding.

  11. k

    Studies were different in terms of populations, interventions, and outcomes.

  12. l

    No or insufficient information on sample size necessary to detect.

Incidence

Follow-up: 2–21 years

Study population OR 1.82 (1.48 to 2.25)306 646 (16 studiesb)

⊕⊝⊝⊝

very lowc,d

Observational studies
26 per 100046 per 1000 (38–57)
Medium-risk population
24 per 100043 per 1000 (35–52)

Asthma control

asthma exacerbations, symptoms, rescue medication

Follow-up: 1 year

Follow-up: 8–24 weeks

The median (range) number of asthma exacerbations was 1 (0–7); no changes in rescue medication score occurredThe median (range) number of asthma exacerbations was 1 (0–4; P = 0.001); Overall reduction in rescue medication score was 0.5 (P = 0.002); mean difference in symptom scores was −12 (95% CI −22 to −1; P = 0.04) 23 (1 study)

⊕⊕⊕⊕

high

Randomized controlled trial
See commenteSee commenteNot estimable447 (8 studies)

⊕⊝⊝⊝

very lowc,f,g,h

Observational studies

Lung function

mean percentual predicted change in FEV1

Follow-up: 1 year

4.9 (−0.5 to 10.3)7.9 (3.4–12.4)Not available38 (1 study)

⊕⊕⊕⊕

high

Randomized controlled trial
See commenteSee commenteNot estimable4050 (9 studies)i

⊕⊝⊝⊝

very lowj,k,l

Observational studies

Airway inflammation

exhaled nitric oxide

Follow-up: 1 year

See commenteSee commenteNot estimable1291 (3 studies)

⊕⊝⊝⊝

very lowj,k,l

Observational studies
Airway responsiveness – follow-up: 24 weeks to 1 yearNot estimableNo effect: for every 10% relative loss of weight, the log2 change in PC20 improved by 0.19; P = 0.66Not estimable1312 (3 study)

⊕⊝⊝⊝

very lowj,l

Observational studies

Obesity seems to increase both asthma severity and the chance of developing a more difficult-to-control asthma. A post hoc analysis of data from four double-blind, placebo-controlled studies randomizing moderate asthmatic adults showed asthma-controlled days were higher for patients with normal weight than for those who were overweight or obese, even after adjustment for asthma severity [5]. In a longitudinal study, aimed to assess the effect of excess weight on asthma control, overweight patients who were initially poorly controlled were more likely to remain so, in spite of pharmacological treatment [13]. Data from the only randomized controlled trial assessing the impact of 14-week dietary intervention for weight loss showed improvement in lung function, symptoms, morbidity, and health status even 1-year post intervention [29]. These studies do not exclude comorbidities of obesity, which were not assessed, such as atopy, gastroesophageal reflux, or sleep-disordered breathing, as modifiers of the obesity–asthma link.

A relationship between obesity and asthma has been traditionally explained by inflammatory and mechanical hypothesis [44]. Inflammatory mediators produced by adipose tissue may modulate the immune responses in the lung, while chronic low-grade inflammation of the obese may influence the susceptibility to airway obstruction. Obesity causes a reduction in respiratory system compliance, lung volumes, and peripheral airway diameter, as well as an increase in airway responsiveness, alteration in pulmonary blood volume, and a ventilation–perfusion mismatch [45]. Changes in lung functional residual and tidal volumes result in reduced smooth muscle stretch and in the change from the normal rapid- to the slow-cycling actin–myosin cross-bridges of the airway smooth muscle. This results in a latch state that may in turn lead to chronic bronchoconstriction [46]. In spite of the evidence that obesity precedes asthma development and increases its severity, the mechanistic basis for this relationship remains unclear. Most likely there are at least two phenotypes of asthma in obese patients, which may reflect different underlying mechanisms: (i) one group with nonatopic late-onset asthma with airway hyper-responsiveness and a variable association with atopic markers that will improve with weight loss and (ii) another characterized by early-onset childhood asthma, atopy, and signs of eosinophilic airway inflammation. The former likely have asthma caused by obesity, the latter likely have early-onset atopic asthma, which is complicated by the coexistence of obesity. Future studies and interventions for obese–asthmatic patients need to treat the subgroup of obese–asthmatic patients with little evidence of allergic inflammation as having a distinct phenotype of asthma.

Obesity and asthma may also influence each other's phenotype due to shared environmental or genetic risk factors. Common genetic determinants contribute substantially to the covariation between asthma and obesity [47]. Furthermore, the hypothesis that reversible effects of dietary and lifestyle factors on asthma and obesity are influenced by gene–environment interaction have not been assessed at all. One additional challenge for the efficacy of weight change interventions in asthma is the timing of the disease process. Not only elevated body mass index per se but weight gain during adult life may differently influence the association with asthma. Moreover, data on the impact of weight change on asthma during childhood have been poorly accessed. Four studies in children showed increased incident asthma in children who became obese [25-28], but no intervention studies have yet accessed the effect of weight loss in asthma in children.

Obesity can affect the response of patients to standard asthma medication. Recently, it has been shown that overweight or obese adults with severe asthma exacerbations required longer durations of treatment in the emergency room and were admitted more frequently than normal or underweight normal asthmatics [15]. It is also possible that obesity changes the responsiveness to asthma medications independently from genetically determined differences in expression of drug target molecules. It has been shown that in a post hoc analysis of pooled data from four double-blind, placebo-controlled randomized studies, increasing BMI was associated with decreased response to inhaled corticosteroid, whereas the response to the leukotriene antagonist remained stable [5].

The control of asthma drives asthma management, and as such, these observations may have consequences for defining new asthma management strategies in overweight patients. Obesity is strongly dependent on the balance between dietary intake and physical activity. Currently the Global Initiative for Asthma Guidelines do not include recommendations for exercise as part of the treatment for patients with asthma. Exercise is a recognized trigger for asthma symptoms and maybe for this reason it is seldom prescribed as part of asthma treatment.

More research is needed to provide the basis for the development of an easily administered therapeutic healthy lifestyle intervention, which could be used alongside current treatment plans for asthma. These types of study have been generally underfunded at the expense of large-scale multicenter pharmacological studies. Researchers and funding agencies now need to develop and support larger and better designed randomized controlled trials assessing combined nutritional and physical activity approaches, applying evidence-based recommendations for reduction in multiple diet and physical activity risk factors, including smoking cessation, physical inactivity, weight management, and nutrition in combination with monitoring of treatment programs.

Conclusions

Meta-analysis showed that weight gain above and beyond the threshold required to be defined as obese almost doubled the odds of incident asthma. In relation to disease-related outcomes, the heterogeneity of the studies precluded a quantitative synthesis. High-quality evidence supported by randomized controlled trials came only from one study, which showed benefits of losing weight on asthma control and lung function. Conflicting data from several observational studies provided limited and low-quality evidence of the beneficial effects of losing weight on asthma-related outcomes. Nevertheless, taken together, these observations support the recommendation of tackling obesity and weight management as part of asthma treatment.

Acknowledgment

The authors thank David Charles for English revision of the manuscript.

Funding

The EAACI Evidence-Based Clinical Practice Guideline Task Force on ‘Lifestyle interventions in allergy and asthma’ was supported by unrestricted grants of the European Academy of Allergy and Clinical Immunology.

Authors' contributions

This report is part 1, of 3, and is a result of the implementation and development of the EAACI Evidence-Based Clinical Practice Guideline Task Force on ‘Lifestyle interventions in allergy and asthma’ chaired by AM, MB, and VGL. AM extracted data and drafted the manuscript. AM, MB, VGL, SB, SG, IA, JF, NP, KHC, LD, and TH conceived the idea and cowrote the protocol, analyzed, and wrote the review. All authors read and approved the final manuscript.

Lay summary

See Data S1.

Conflict of interest

None known.

Ancillary