The Association between Body Mass Index, Abdominal fatness, Weight Change and the risk of Adult Asthma: A Systematic Review and Meta-Analysis of Cohort Studies

doi:10.21203/rs.3.rs-1608906/v1

PDF

Article

The Association between Body Mass Index, Abdominal fatness, Weight Change and the risk of Adult Asthma: A Systematic Review and Meta-Analysis of Cohort Studies

https://doi.org/10.21203/rs.3.rs-1608906/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

Obesity correlate with increasing risk of adult asthma. However, not all studies have found a clear association between overweight and the incidence of asthma, and data on other adiposity measures have been limited. Hence, we aimed to summarize evidence on association between adiposity and adult asthma. Relevant studies were retrieved through searches conducted in PubMed, and EMBASE up to March 2021. A total of fifteen studies (63,952 cases and 1,161,169 participants) were eligible for the quantitative synthesis. The pooled RR per 5 kg/m² increase in BMI was 1.32 (95% CI 1.21–1.44, I² = 94.6%, p_{heterogeneity}<0.0001, n = 13), 1.26 (95% CI: 1.09–1.46, I² = 88.6%, p_{heterogeneity}<0.0001, n = 5) per 10 cm increase in waist circumference and 1.33 (95% CI: 1.221.44, I² = 62.3%, p_{heterogeneity}=0.05, n = 4) per 10 kg increase in weight gain. There was a significant on test of nonlinearity for BMI (p_nonlinearity<0.00001), weight change (p_nonlinearity=0.002), and waist circumference (p_nonlinearity=0.02). The magnitude of the associations and the consistency of the results across studies and adiposity measures provide strong evidence that overweight and obesity, waist circumference and weight gain increases asthma risk. These findings support policies to curb the global epidemic of excess weight.

Asthma

obesity

overweight

adiposity

weight gain

Asthma and obesity are considered major public health problems in worldwide. Bronchial asthma is a non-communicable, chronic medical condition that causes the airways in the lungs to become narrowed and swollen that induce asthmatic symptoms.¹ The greatest risk of asthma is observed at age of 18 to 45 years. Each year, there was 7% increases in the risk of severe asthma until the asthmatics reach 45 years of age.² Obese adults with asthma getting frequent respiratory symptoms and severe persistent asthma than non-obese adults with asthma.³

Overweight (body mass index, BMI 25-<30) or obesity (BMI ≥ 30) is described as a weight that is higher than what is considered as a healthy weight for a given height. Globally, the age-standardized mean BMI increased from 21.7 kg/m² to 24.2 kg/m² between 1975 and 2015 among men, and it increased from 22.1 kg/m² to 24.4 kg/m² among women in the same period. Currently, the global prevalence of obesity is 10.8% in men and 14.9% in women. It has been projected that if this trend continues, the prevalence of obesity will be 18% in men and 21% in women in 2025.⁴

According the World Health Organization (WHO), more than 339 million people fall ill with asthma, 417,918 deaths occurred due to asthma, and 24.8 million disability-adjusted life-years (DALYs) were happened due to asthma globally in 2016. Asthma-related deaths occurs more than 80% in low and lower-middle-income countries. The estimated asthma prevelance based on the World Health Survey (2002–2003). Overall, 4.3% of the younger adults (18 to 45 years) reported doctor diagnosed asthma, 4.5% reported either a doctor’s diagnosis or that they were taking treatment for asthma, and 8.6% reported that they had experienced attacks of wheezing or whistling breath (symptoms of asthma) in the preceding 12 months.⁵ Risk factors for asthma include genetic susceptibility, allergic sensitization, environmental tobacco smoke, air pollution, mould and damp, exposure to animals, use of medications (antibiotics and acetaminophen), and occupational exposures such as woodworking, farming, paint containing isocyanates, diet (fast food). Thus, environmental factors are much more likely to cause asthma than genetic factor.⁶

Although evidence from cohort studies suggests that there is a positive association between obesity and increased incidence of asthma,^7–19 studies have been less consistent with regards to whether overweight is associated with increased risk as some studies reported an increased risk ^{8–12,14−17,19} and others found no association.^7,18 Most of the available studies have reported a dose-response relation between increasing BMI and asthma incidence but have differed somewhat with regard to what BMI level risk of asthma starts to increase.^{7,8,17−22,9−16} Fewer studies have been published on other measures of adiposity and risk of asthma, but three studies reported increased asthma risk with increasing waist circumference,^{8,12,14,21,22} while two studies reported an increased risk with greater weight gain,^11,17 and a third study found no clear association.¹⁹ In contrast studies have generally reported no association between weight loss and asthma.^11,19,23 For these reasons, we conducted a systematic review to summarize the association between adiposity and the risk of developing asthma among adults from the available evidence of prospective cohort studies, with particular aims of clarifying the shape and strength of the dose-response pattern between adiposity and asthma and to clarify potential confounding by smoking and other established risk factors.

Search strategy

This systematic review was registered in PROSPERO (CRD42020167990). We adopted a detailed search strategy with search terms as presented in Supplementary Table 1. With the search strategy, we systematically searched in PubMed and Embase databases for cohort studies that investigated the association between adiposity and asthma. The search was carried out in mentioned databases by using both Medical Subject Headings (MeSH) terms related to our exposure and outcome and free-text words with appropriate truncations, wildcards, and proximity. Appropriate Boolean operators (“OR” and “AND”) were used to combine the individual search results. The search was limited to articles written in English.

Eligibility criteria

We have included the studies with following study design: Prospective cohort studies, retrospective cohort studies, case-cohort studies, nested case-control studies, and retrospective cohort studies and we included studies reporting adjusted estimates (relative risk (RR)) with 95% confidence intervals (CIs) for the association between any measure of adiposity and the risk of incident bronchial asthma in adults aged ≥ 18 years. Retrospective case-control studies, cross-sectional studies, and case-only studies as well as studies in pregnant women were excluded. Only full-text articles were included, while conference abstracts, unpublished data, and grey literature were excluded.

Study selection process

We adopted three stages in the study selection process

Stage 1 - Primary screening: All retrieved studies from data bases were imported to the Mendeley and the list of seleceted studies to be screened was prepared by removing the duplicates. Two independent investigators (LA and RN) screened the title, abstract, and keywords of the identified citations. Full-text articles were retrieved from these shortlisted studies.

Phase 2 - Secondary screening: Based on the eligibility criteria of our review, full-text were screened on selected article from the primary screening by the same two reviewers (LA and RN). We excluded the studies which not met the eligibility criteria and exclusion reasons were noted.

Phase 3 - Final study selection: During screening process, conflicts on a selection of the studies between two investigators were resolved and the final studies was selected by consensus with the third reviewer (PG). The list of excluded studies with reasons are provided in Supplementary Table 2.

Data extraction

Data extraction from the selected studies was carried out independently by two reviewers (LA and RN) using a pre-designed data extraction form. When extractions differed between the reviewers, they were resolved by consensus. We extracted the name of the first author, publication year, geographic location, name of the study, follow-up period, sample size, age, gender, number of cases, the method for assessment of adiposity (self-report vs. measured), adiposity variable, quantity of the adiposity exposure, relative risk estimates (e.g. hazard ratios, risk ratios, odds ratios) and corresponding 95% confidence intervals (CIs), and confounders adjusted for in the analysis from each study.

Study quality assessment

We adopted the Newcastle-Ottawa (NO) scale for cohort studies for assessing the quality of all included stides.²⁴ It was performed by two independent reviewers (LA and RN). Study quality was assessed on the following domains; selection, exposure, comparability, and outcome. We modified the Newcastle Ottawa scale for this review by excluding the point about representability as this is not directly a measure of study quality, and by modifying the score for the questions on confounding factors, as the original scale gave 1 point for each of two confounding factors adjusted for, however, a study with adjustment for age and sex would receive a maximum score, but could still be confounded by other risk factors. A score of 0.25 was given for each confounding factor that the analyses were adjusted for, up to a maximum of two points for eight or more confounders that were adjusted for. The score for record linkage was modified to 0.5 instead of 1, and a 1 point was given if the outcome assessment had been validated. This gave a total score of between 0–8 which was further grouped as 0–3 (low quality), > 3–6 (medium quality) and > 6–8 (high quality).

Evidence grading

We used World Cancer Research Fund criteria for evaluating evidence of causality of the observed associations between adiposity and risk of adult asthma according to predefined criteria.²⁵ The system grades evidence ranging from substantial effect on risk unlikely, limited - no conclusion, limited-suggestive, probable, to convincing.

Statistical analysis

We used random effects model to estimate the pooled RRs and 95% CIs for the association between BMI, waist circumference, and weight change (the three exposures with a sufficient number of studies for statistical analyses) and asthma. We adopted the method of Greenland and Longnecker to assess study-specific linear trends across categories of BMI, waist circumference, and weight change.²⁶ For studies reporting median or mean BMI, waist circumference, or weight change per category these values were used as reported, while for studies reporting a range of BMI, waist circumference, or weight change, the average of the lower and upper cut-off points was used to estimate the midpoint. When extreme ranges were open-ended, we estimated an upper or lower cut-off value based on the width of the adjacent interval. When the reference category was not the lowest, we used Hamling's approach to recalculate the risk estimations so that the lowest category became the reference category.²⁷ We used a BMI of 15 as a lower bound for underweight and 18.5 as a lower bound for normal weight where studies provided analyses by WHO categories of overweight and obesity. We used fractional polynomial models for examining the nonlinear dose-response relationships between BMI, waist circumference, weight change and asthma.²⁸ Also the model with the lowest deviation was found to be the best fit second-order fractional polynomial regression model. A likelihood ratio test was performed to assess differences between the linear and nonlinear models in order to test for nonlinearity.

Heterogeneity was evaluated by chi-square of heterogeneity and I² statistic.²⁹ We considered a p-value < 0.10 in chi-square test as indication of significant heterogeneity, while I² value was used to quantify the degree of heterogeneity, ranging from 0-100%. We tested for publication bias using Egger’s test,³⁰ Begg’s test,³¹ and by visually inspecting the funnel plots. A p-value < 0.10 with Egger’s test was considered as an indication of potential publication bias. Sensitivity analyses were conducted to test whether the overall association was explained by a study with an intense result. We also reported the summary estimates by excluding the studies with the most positive and negative impact in the results section. We calculated E-values for the association between adiposity variables and asthma, to assess the potential impact of unmeasured or uncontrolled confounding. The E-value is defined as the lowest strength that an uncontrolled or unmeasured confounder would have with both the exposure and the outcome to fully explain away the observed association.³² Meta-analyses were conducted by using Stata 13.1 (StataCorp, College Station, TX, USA).

We found 1651 studies through the systematic literature search of PubMed (n = 1048) and EMBASE (n = 603)[Figure 1]. After removing the duplicates, 1411 studies were screened during the primary screening stage. We deemed 72 of those studies relevant for the full-text retrieval. During the second screening stage, we excluded 56 studies which did not meet the eligibility criteria. Totally, fifteen studies were included in our review.^{7,8,17–19,22,23,9−16}

Characteristics of the included studies

Fifteen included studies characteristics are presented in Table 1. Of these, three studies were conducted in Norway, three from the USA, two from France, one from Australia. In terms of gender inclusion, three studies were conducted among females and twelve studies were conducted among both males and females. The number of study participants from our included studies was 4532 to 246,361 and the duration of follow-up was from 2 years to 25 years.

Table 1

Characteristics of included studies (N = 15)
Author name, publication year, country	Study name	Follow-up period	Study size	Sex	Age	Number of cases	Assessment of height and weight	Exposure and subgroup	Comparison	RR (95% CI)	Adjustment for confounders
Camargo CA et al., 1999, USA	The Nurses’ Health Study II	1991–1995, 4 years follow up	85911	Women	26–46 years	1596	Self-reported (validated)	BMI Weight change since age 18 years	< 20 kg/m² 20.0-22.4 22.5–24.9 25.0-27.4 27.5–29.9 ≥ 30.0 <-5 kg -5 to -2.1 -2 to 2 2.1 to 5 5.1 to 10 10.1 to 20 20.1 to 25 > 25	0.9 (0.7–1.1) 1.0 1.1 (1.0-1.3) 1.6 (1.3–1.9) 1.7 (1.4-2.0) 2.7 (2.3–3.1) 0.8 (0.6–1.1) 0.8 (0.6–1.1) 1.00 0.9 (0.8–1.2) 1.1 (0.9–1.3) 1.4 (1.2–1.7) 2.0 (1.6–2.5) 2.5 (2.0-3.1)	Age, race, US region, smoking status, physical activity, total energy intake, hysterectomy status, birth weight, and duration of breastfeeding
Beckett WS et al., 2001, California	CARDIA study	1986–1996, 10 years follow-up	4532	Men and women	18–30	434	Measured	BMI percentiles	1–19 20–39 40–59 60–79 80–100	1.09 (0.78–1.51) 1.00 0.99 (0.71–1.37) 1.10 (0.80–1.53) 1.21 (0.88–1.68)	Age, race, sex, centre, and maximal education
Chen Y et al, 2002, Canada	Canadian National Population Health Surveys	1994 - 1996, 2 years follow-up	4883	Women	20–64 years	127	Self-reported	BMI	< 20.0 kg/m² 20.0-24.9 25.0-29.9 ≥ 30.0	0.62 (0.24–1.62) 1.00 1.25 (0.72–2.18 1.92 (1.09–3.41)	Age, smoking, pet(s) at home, immigrant status, history of allergy, income adequacy, and alcohol drinking
Huovinen E et al., 2003, Finland	The Finnish Twin Cohort	1982–1990, 9 years follow-up	10597	Men and women	25–52 years	149	Self-reported	BMI, men BMI, women Weight change, men Weight change, women	< 20.0 kg/m² 20.0-<25 ≤ 25–30 ≥ 30.0 < 20.0 kg/m² 20.0-<25 ≤ 25–30 ≥ 30.0 <-2 kg < 2 2 to 10 > 10 <-2 kg < 2 2 to 10 > 10	0.41 (0.06–3.06) 1.00 1.07 (0.62–1.85) 3.47 (1.56–7.72) 1.56 (0.87–2.78) 1.00 1.60 (0.91–2.80) 2.30 (0.87–6.08) 0.72 (0.31–1.63) 1.00 0.98 (0.56–1.74) 1.41 (0.50–3.98) 1.12 (0.57–2.21) 1.00 0.95 (0.57–1.59) 1.45 (0.54–3.90)	Age, atopy and respiratory symptoms
Romieu I et al., 2003, France	E3N Cohort Study	1990–1993, 3 years follow-up	67229	Women	40–60 years	372	Self-reported (not validated)	BMI	< 20.20 kg/m² 20.20-21.41 21.42–22.68 22.69–24.61 24.62–26.99 ≥ 27	0.93 (0.66–1.31) 1.00 1.22 (0.88–1.69) 1.18 (0.84–1.64) 1.46 (1.01–2.12) 2.02 (1.38–2.98)	Age, total caloric intake, physical activity, smoking status, and menopausal status.
Ford ES et al., 2004, USA	First National Health and Nutrition Examination Survey	1971–1974–1982–1984, 10 years follow-up	14407	Men and women	25–74 years	317	Measured	BMI, all BMI, men BMI, women	18.5-<25.0 kg/m² 25.0-<30.0 30.0-<35.0 ≥ 35.0 18.5-<25.0 kg/m² 25.0-<30.0 30.0-<35.0 ≥ 35.0 18.5-<25.0 kg/m² 25.0-<30.0 30.0-<35.0 ≥ 35.0	1.00 0.95 (0.68–1.33) 1.28 (0.83–1.96) 1.87 (1.12–3.13) 1.00 0.79 (0.47–1.34) 1.52 (0.82–2.82) 1.98 (0.60–6.54) 1.00 1.13 (0.73–1.75) 1.10 (0.64–1.89) 1.87 (1.05–3.33)	Age, sex (except sex-specific models), race or ethnicity, education, smoking status, recreational physical activity and nonrecreational activity
Coogan PF et al, 2009, USA	Black Women’s Health study	1995–2005, 10 years follow-up	46435	Women	21–69 years	1068	Measured	BMI Weight change since age 18 years	< 20 kg/m² 20–24 25–29 30–34 35–39 ≥ 40 < 0 kg 0–9 10–14 15–19 20–24 ≥ 25	0.84 (0.56–1.26) 1.00 1.26 (1.05–1.51) 1.62 (1.32–1.98) 2.24 (1.76–2.84) 2.85 (2.19–3.72) 1.01 (0.72–1.41) 1.00 1.07 (0.84–1.37) 1.29 (1.02–1.63) 1.35 (1.07–1.71) 1.88 (1.54–2.28)	Adjusted for age, time period; current smoker, pack-years of smoking, second hand smoke, parental history of asthma, menopausal status, marital status, income, menopausal female hormone use, metabolic equivalents, BMI at age 18 years, energy intake
Hjellvik V et al., 2010, Norway	Norwegian Health Survey	2004–2007, 3 years follow up	118723	Men and women	40 years	4049	Measured	BMI	< 20 kg/m² 20-24.9 25-29.9 30-34.9 ≥ 35	1.0 (0.9–1.2) 1.0 1.3 (1.2–1.4) 1.8 (1.6–2.0) 2.4 (2.0–2.9)	Age, sex, education, physical activity, disability pension, urban/rural status, smoking status
Korda RJ et al., 2012, Australia	45 and Up Study	2006–2009, 2.3 years follow up	246361	Men and women	45-≥80 years	275	Self-reported data (validated)	BMI, age 45–64 years BMI, age 65–79 years BMI, age ≥ 80 years	18.5-<25 kg/m² 25.0-<30 ≥ 30 18.5-<25 kg/m² 25.0-<30 ≥ 30 18.5-<25 kg/m² 25.0-<30 ≥ 30	1.00 2.30 (1.41–3.75) 3.97 (2.48–6.36) 1.00 0.84 (0.50–1.41) 1.64 (0.98–2.73) 1.00 1.69 (0.80–3.56) 3.06 (1.35–6.94)	Age, sex, area of residence, education, pre-tax annual household income, smoking, alcohol intake, private health insurance
Brumpton B et al., 2012, Norway	HUNT study	1997–2008, 11 years follow-up	23245	Men and women	19–55 years	818	Measured	BMI, women BMI, men WC, women WC, men	25.0 kg/m² 25.0-29.9 ≥ 30.0 Per 5 kg/m² 25.0 kg/m² 25.0-29.9 ≥ 30.0 Per 5 kg/m² < 80.0 cm 80-87.9 ≥ 88 Per 10 cm < 80.0 cm 80-87.9 ≥ 88 Per 10 cm	1.00 1.44 (1.18–1.76) 1.96 (1.52–2.52) 1.32 (1.19–1.45) 1.00 1.21 (0.93–1.58) 1.84 (1.30–2.59) 1.38 (1.17–1.62) 1.00 1.30 (1.04–1.62) 1.88 (1.51–2.34) 1.28 (1.18–1.38) 1.00 1.15 (0.87–1.52) 1.55 (1.08–2.21) 1.23 (1.08–1.41)	Age, smoking, physical activity, education, family history
Leone N et al., 2012, France	The 3C study	NA-NA, 4 years follow-up	6267	Men and women	> 65 years	67	Measured	WC	< 94/80 cm 94/80–102/88 ≥ 102/88	1.00 2.69 (1.21–5.98) 3.84 (1.55–9.49)	Age, centre, educational level, smoking status, physical ability, cardiovascular disease history, dyspnoea, and chronic bronchitis symptoms.
Assad N et al, 2013, USA	CARDIA study	1985–2010, 25 years follow-up	4619	Men and women	Mean age (24.9 ± 6 3.6)	602	Measured	BMI, all BMI, men BMI, women	Per 5 kg/m² Per 5 kg/m² Per 5 kg/m²	1.14 (1.06–1.22) 1.08 (0.92–1.27) 1.15 (1.06–1.26)	Age, race, smoking, and physical activity
Tomita Y et al, 2018, Japan	Health insurance claims and health checkups	2012–2015, ~ 3 years follow-up	9888	Men and women	40–64 years	424	Measured	BMI, men BMI, women WC, men WC, women	< 18.4 kg/m² 18.5–24.9 25-29.9 ≥ 30.0 < 18.4 kg/m² 18.5–24.9 25-29.9 ≥ 30.0 ≤ 84.9 cm 85-89.9 ≥ 90 ≤ 84.9 cm 85-89.9 ≥ 90	0.64 (0.20–2.04) 1.00 1.10 (0.76–1.59) 0.94 (0.44–1.98) 1.03 (0.66–1.59) 1.00 2.02 (1.38–2.96) 2.65 (1.26–5.54) 1.00 0.97 (0.59–1.60) 0.89 (0.52–1.53) 1.00 1.34 (0.87–2.07) 2.33 (1.37–3.96)	Age, smoking status, allergic rhinitis, and metabolic syndrome
Park S et al., 2019, Korea	National Health Insurance Service-Health Screening Cohort	2004–2013, 9 years follow-up	459529	Men and women	≥ 40 years	53371	Measured	BMI, men BMI, women WC, men WC, women	< 18.5 kg/m² 18.5-<23 23-<25 25-<30 ≥ 30.0 < 18.5 kg/m² 18.5-<23 23-<25 25-<30 ≥ 30.0 < 80 cm 80-<90 90-<100 ≥ 100 < 75 cm 75-<85 85-<95 ≥ 95	1.32 (1.23–1.43) 1.00 1.00 (0.96–1.03) 1.00 (0.96–1.03) 1.23 (1.13–1.34) 0.94 (0.87–1.03) 1.00 1.11 (1.07–1.14) 1.24 (1.21–1.28) 1.40 (1.32–1.48) 1.00 1.05 (0.99–1.12) 1.13 (1.05–1.21) 1.34 (1.16–1.57) 1.00 1.14 (1.08–1.21) 1.23 (1.14–1.32) 1.19 (1.03–1.37)	Age, insurance type, household income, smoking, drinking, physical activity
Wang H et al, 2020, China	Shandong Multi-Center Health Check-up Longitudinal Study	1999–2000–2015–2016, 12 years follow-up	42304	Men and women	≥ 20 years	90	Measured	BMI, men BMI, women	< 25 kg/m² 25-<30 ≥ 30 < 25 kg/m² 25-<30 ≥ 30	1.00 0.78 (0.44–1.39) 0.77 (0.27–2.21) 1.00 2.71 (1.31–5.61) 6.41 (2.34–17.54)	Age, smoking and drinking status
Wang T et al., 2021, USA	National Health and Nutrition Examination Survey	1999–2000–2015–2016, 10 years follow-up	20771	Men and women	40–74 years	627	Measured height, self-reported recalled weight	Weight change since age 25 years	<-2.5 kg -2.5 to < 2.5 2.5 to < 10 10 to < 20 ≥ 20	0.91 (0.53–1.56) 1.00 0.99 (0.73–1.34) 1.19 (0.88–1.60) 1.53 (1.15–2.03)	Age, sex, race/ethnicity, education, family income-poverty level, smoking status, family history of asthma

Assessment of study quality:

The quality of included study was assessed using the modified Newcastle Ottawa scale and is presented in Supplementary Table 3. Overall, the mean (median) study quality was 5.9 (6.0).

BMI and adult asthma

A total of thirteen studies including 63,258 cases and 1,134,131 participants were included in the high vs. low and linear dose-response analysis of BMI and risk of adult asthma.^{7,8,17−19,9−16} The pooled RR was 2.13 (95% CI: 1.66–2.73, I² = 87.9%, p_{heterogeneity}<0.0001) for high vs. low BMI (Supplementary Fig. 1) and 1.32 (95% CI: 1.21–1.44, I² = 95.3%, p_{heterogeneity} <0.0001) per 5 kg/m² increase (Fig. 2A). In sensitivity analyses, excluding one study at a time, the pooled RR ranged from 1.29 (95% CI: 1.19–1.41) when removing The Nurses’ Health Study II¹¹ to 1.34 (95% CI: 1.27–1.43) when removing the CARDIA study¹³ (Supplementary Fig. 2). There was indication of publication bias significant with Egger’s test (p = 0.01) and by investigation of the funnel plot (Supplementary Fig. 3), but not significant indication with Begg’s test (p = 0.36), however, there was no indication of bias when excluding a large Korean study with Egger’s test (p = 0.63) was no longer significant and there was no longer asymmetry in the funnel plot (Supplementary Fig. 4), which found a much weaker association than the remaining studies, however, the summary RR was not substantially altered (Supplementary Fig. 2). We observed a nonlinear association between BMI and risk of adult asthma (p_nonlinearity<0.0001), and the non-linear curve was flat between on BMI value of 19 and 21 and increased at above this range (Fig. 2B, Supplementary Table 4).

Waist circumference and adult asthma

Four studies including 14,081 cases and 282,138 participants were included in the quantitative synthesis of waist circumference and asthma.^8,12,14,22 The pooled RR was 1.56 (95% CI: 1.22–2.01, I² = 77.8%, p_{heterogeneity}=0.004) for high vs. low waist circumference (Supplementary Fig. 5) and 1.26 (95% CI: 1.09–1.46, I² = 88.6%, p_{heterogeneity} = < 0.0001) per 10 cm increase in waist circumference (Fig. 3A). The pooled RR ranged from 1.20 (95% CI: 1.04–1.39) when removing the study by Leone et al, 2012 ²² to 1.35 (95% CI: 1.16–1.57) when removing the study by Park et al, 2019 ⁸(Supplementary Fig. 6). Although the test of nonlinearity between waist circumference and asthma risk was significant (p_nonlinearity=0.02), the association was dose-dependent and nearly linear (Fig. 3B, Supplementary Table 5).

Weight change and adult asthma

Four studies including 3440 cases and 163,714 participants were included in the quantitative synthesis of weight gain and weight loss and asthma risk.^11,17,19,23 The pooled RR was 1.93 (95% CI: 1.53–2.44, I² = 62.3%, p_{heterogeneity}=0.05) for high vs. low weight gain (Supplementary Fig. 7) and 1.33 (95% CI: 1.22–1.44, I² = 62.3%, p_{heterogeneity}=0.047) per 10 kg increase in weight gain (Fig. 4A). The pooled RR was 1.28 (95% CI: 1.18–1.38) when excluding the study by Camargo et al, 1999¹¹ to 1.39 (95% CI: 1.33–1.46) when excluding the study by Wang et al, 2021²³ (Supplementary Fig. 8). There was significant evidence of a nonlinear association between weight gain and asthma risk (p_nonlinearity=0.002), with a minor J-shaped curve, however, the association was approximately linear from a weight gain of around 2 kg (Fig. 4B, Supplementary Table 6). We couldn’t perform dose-response analyses of weight loss and asthma risk, but a high vs. low comparison gave a pooled RR of 0.90 (95% CI: 0.74–1.09, I² = 0%, p_{heterogeneity}=0.783) for weight loss in relation to asthma risk based on data from 4 cohort studies^11,17,19,23 (Fig. 5) and in sensitivity analyses that removed one study at a time, there was minimal variation in the summary estimate. (Supplementary Fig. 9).

Subgroup and sensitivity analyses

We perfromed subgroup analyses based on duration of follow-up, sex, geographic location, method for assessment of weight and height, number of cases, study quality as well as adjustment for a range of confounding factors, the positive association between BMI and asthma was noticed across nearly all the subgroup and also we performed meta-regression analyses, we found little evidence of heterogeneity between subgroups (Supplementary Table 7). The only subgroup analyses with significant heterogeneity between subgroups were the subgroup analyses stratified by method of assessment of weight and height, where there was a weaker association among studies using measured vs. self-reported measures of weight and height and a stronger association among studies with a high study quality score vs. studies with a medium study quality score. Although heterogeneity within subgroups were in general high to very high, we observed heterogeneity was less among the subgroups of European studies, studies with self-reported weight and height, lower number of cases, and with adjustment for allergy (Supplementary Table 7).

Evidence grading

Using World Cancer Research Fund/American Institute for Cancer Research (WCRF/AICR) criteria for evaluating strength of evidence (evaluating causality) (Supplementary Table 8), we considered evidence for the association between BMI and asthma to be convincing, and there was limited-suggestive evidence that greater waist circumference and weight gain increases the risk of asthma (Supplementary Table 9, 10). There was limited evidence and no conclusion could be drawn with regard to the impact of weight loss on asthma risk (Supplementary Table 9, 10).

In this dose-response quantitative syntheses of multiple measures of adiposity and the risk of asthma, study concluded that 5 units increase in BMI there was a 32% increase in the RR per 5 units increase in BMI, a 26% increase in RR per 10 cm increase in waist circumference, and a 33% increase in RR per 10 kg increase in weight gain. Weight loss was not significantly associated with reduced risk, but statistical power may have been limited to detect a weak association. There was a strong positive dose-response relationship between increasing BMI and risk of asthma, with RRs of 1.23, 1.69, 2.45 and 3.64 at BMI values of 25, 30, 35, and 40 compared with the reference BMI of 20, and there was some indication of increased risk even within the high range of the normal weight category. There was also a strong positive dose-response relationship between greater waist circumference and weight gain and the risk of asthma.

Our meta-findings are similar with two earlier meta-analyses on BMI³³ and abdominal obesity³⁴ and asthma risk. After all, the present meta-analysis includes a much greater number of studies on BMI (13 vs. 7 studies) and was restricted to cohort studies to avoid and reduce potential recall and selection bias and reverse causation, while a previous meta-analysis also included case-control and cross-sectional studies³⁴. In addition, we also assessed the dose-response analysis to find out relationship between BMI, waist circumference and weight gain and risk of asthma. Other studies have reported that higher BMI is associated with narrowing of the airways,³⁵ reduced lung volume³⁶ and lung function³⁷ and one study found that greater duration of obesity was associated with reduced lung function³⁸. Therefore, obesity is considered an important determinant of impaired pulmonary function.³⁹ Further support for this is found in a trial conducted among obese patients with severe asthma and in a review which found a benefit of weight loss interventions in relation to asthma control.^40,41 Although we did not detect a significant inverse association for weight loss in relation to asthma risk in the current analysis, the summary estimate was in the direction of reduced risk and it is possible that the weight loss observed in the studies too small to detect association and/or that an association could have been more clearly demonstrated for weight loss if the analyses performed to subjects who were overweight or obese initially. Another issue that potentially could have affected the results is whether the observed weight loss was intentional or not.

The rigorous literature search was the major strength of our review and methodology as well as detailed subgroup and sensitivity analyses, which showed that the results were consistent across major subgroups, robust to the influence of single studies as well as potential confounders. The results were also consistent across geographic regions with different patterns of confounding factors. Moreover, to our knowledge this is the first meta-analysis to clarify the dose-response relationship between different measures of adiposity and asthma risk through both linear and nonlinear dose-response analyses. The current meta-analysis was based on data from cohort studies, and excluded data from cross-sectional and case-control studies. Therefore our results are less likely to have been affected by recall bias, selection bias and reverse causation. We also discovered that the studies included in our review were of medium to higher quality.

Our review also has some limitations. We can't rule out the likelihood that confounding from additional risk variables contributed to the observed relationships, as with any observational study. However, this seems less likely as the observed association between BMI and asthma was consistent across most subgroups. In addition, the associations with the different adiposity variables were relatively strong, thus we consider it less likely that confounding could fully explain the observed associations. Based on the estimated RRs and the E-values we found that an unadjusted confounder might have to be very strongly associated with both exposure (adiposity) and with outcome (asthma) to fully explain away the observed associations. For example RRs for an unadjusted confounder would have to be in the size of 2–3 for overweight and obesity, and 4–9 for higher levels of BMI, waist circumference, and weight gain to explain away the observed associations fully. Despite the fact that there was a lot of variability among the studies included, it appeared to be due to differences in the strength of the link rather than differences in the direction of the association, since all the studies reported that risk estimates were all in the direction of higher risk. This is much less issue than when there is heterogeneity in the direction of the risk estimates. In addition, there was lower heterogeneity in some subgroups including European studies, studies with self-reported weight and height (both validated and not validated), studies conducted with a lesser number of participants and cases and among studies with adjustment for allergy. Publication bias can alter the meta-findings of published studies. Although we did find some indication of publication bias in the primary analysis, the asymmetry in the funnel plot was driven by a large Korean study which showed a much lower risk estimate compared to the remaining studies, and when excluded from the analysis the test for publication bias and the asymmetry in the funnel plot did not persist, while the summary estimate was only slightly modified.

In adults, obesity produces considerable alterations in normal lung physiology, including the accumulation of fat in the thoracic and abdominal cavities, resulting in lung compression and a reduction in lung volume. Increased adipose tissue inflammation causes airway hyperresponsiveness (a distinguishing feature of asthma) that leads to increased risk of asthma in adults.⁴² Obesity is associated with a chronic inflammatory response. It has been shown that adipose tissue in obese individuals infiltrates with macrophages that promote and secrete the systemic pro-inflammatory mediators such as tumor necrosis factor-α, leptin, and adiponectin cause inflammation in the airway which leads to airway hyperresponsiveness which may contribute to asthma in obese.⁴³ Obesity also alters lung mechanics which contribute to symptoms of dyspnoea that occurs in absence of the airway inflammatory changes which lead to an increased risk of asthma.⁴⁴ The hormone leptin partly plays a role as indirect mediator of the association between high body adiposity and persistent asthma over time.⁴⁵ Also leptin involved in pathophysiological process of asthma through improved phagocytosis, activation, proliferation, and alteration of macrophages, and induced production of proinflammatory cytokines.⁴⁶ In obesity, adipokines and adiponectin hormone level decreased which lowers the pulmonary function test, airway hyperresponsiveness, and airway inflammation.⁴⁷ Visceral fat has more inflammatory cells than subcutaneous cells and visceral fat is the main source of proinflammatory cytokines that may promote airway inflammation or may influence the structures of airways such as smooth muscle cells, blood vessels, and nerves which leads to airway hyperresponsiveness (AHR) in asthma.^48,49

This meta-analysis found evidence that different measures of adiposity including higher BMI, weight gain, and waist circumference were strongly associated with increased risk of asthma in adults. These findings provide strong suggestion and reinforce for public health policy recommendations and interventions (e.g. to target dietary patterns and physical activity) to reduce the burden of excess weight in the general population for the prevention of asthma.

ACKNOWLEDGEMENT:

None

CONFLICT OF INTEREST:

None

DATA AVAILABILITY:

The datasets used and/or analysed during the current study available from the corresponding author on reasonable request.

Mims JW. Asthma: Definitions and pathophysiology. Int Forum Allergy Rhinol. 2015;5(April):S2-S6. doi:10.1002/alr.21609
Zein JG, Dweik RA, Comhair SA, et al. Asthma is more severe in older adults. PLoS One. 2015;10(7):1-13. doi:10.1371/journal.pone.0133490
Taylor B, Mannino D, Brown C, Crocker D, Twum-Baah N, Holguin F. Body mass index and asthma severity in the National Asthma Survey. Thorax. 2008;63(1):14-20. doi:10.1136/thx.2007.082784
Di Cesare M, Bentham J, Stevens GA, et al. Trends in adult body-mass index in 200 countries from 1975 to 2014: A pooled analysis of 1698 population-based measurement studies with 19.2 million participants. Lancet. 2016;387(10026):1377-1396. doi:10.1016/S0140-6736(16)30054-X
Moores G, Boulet L-P, Gershon AS, et al. Global asthma prevalence in adults: findings from the cross-sectional world health survey. BMC Public Health. 2012;12(1):204. http://www.biomedcentral.com/1471-2458/12/204%5Cnfile:///Files/53/53d34e04-9fa9-4bc6-a820-4dbf3af7810e.pdf%5Cnpapers3://publication/doi/10.1186/1471-2458-12-204
Global Asthma Network. The Global Asthma Report Asthma Affects.; 2018.
Wang H, Bai C, Yi M, et al. Metabolic syndrome and incident asthma in chinese adults: An open cohort study. Diabetes, Metab Syndr Obes Targets Ther. 2020;13:3411-3420. doi:10.2147/DMSO.S274159
Park S, Jung S-Y, Kwon J-W. Sex differences in the association between asthma incidence and modifiable risk factors in Korean middle-aged and older adults: NHIS-HEALS 10-year cohort. BMC Pulm Med. 2019;19(1):248. doi:10.1186/s12890-019-1023-3
Romieu I, Avenel V, Leynaert B, Kauffmann F, Clavel-Chapelon F. Body mass index, change in body silhouette, and risk of asthma in the E3N cohort study. Am J Epidemiol. 2003;158(2):165-174. doi:10.1093/aje/kwg131
Chen Y, Dales R, Tang M, Krewski D. Obesity may increase the incidence of asthma in women but not in men: Longitudinal observations from the Canadian National Population Health Surveys. Am J Epidemiol. 2002;155(3):191-197. doi:10.1093/aje/155.3.191
Dumas O, Boggs KM, Cozier YC, Stampfer MJ, Camargo CA. Prospective study of body mass index and risk of sarcoidosis in US women. Eur Respir J. 2017;50(4). doi:10.1183/13993003.01397-2017
Tomita Y, Fukutomi Y, Irie M, et al. Obesity, but not metabolic syndrome, as a risk factor for late-onset asthma in Japanese women. Allergol Int. 2019;68(2):240-246. doi:10.1016/j.alit.2018.10.003
Assad N, Qualls C, Smith LJ, et al. Body mass index is a stronger predictor than the metabolic syndrome for future asthma in women. The longitudinal CARDIA study. Am J Respir Crit Care Med. 2013;188(3):319-326. doi:10.1164/rccm.201303-0457OC
Brumpton B, Langhammer A, Romundstad P, Chen Y, Mai X-M. General and abdominal obesity and incident asthma in adults: the HUNT study. Eur Respir J. 2013;41(2):323-329. doi:10.1183/09031936.00012112
Korda RJ, Liu B, Clements MS, et al. Prospective cohort study of body mass index and the risk of hospitalisation: Findings from 246 361 participants in the 45 and Up Study. Int J Obes. 2013;37(6):790-799. doi:10.1038/ijo.2012.155
Hjellvik V, Tverdal A, Furu K. Body mass index as predictor for asthma: a cohort study of 118,723 males and females. Eur Respir J. 2010;35(6):1235-1242. doi:10.1183/09031936.00192408
Coogan PF, Palmer JR, O’Connor GT, Rosenberg L. Body mass index and asthma incidence in the Black Women’s Health Study. J Allergy Clin Immunol. 2009;123(1):89-95. doi:10.1016/j.jaci.2008.09.040
Ford ES, Mannino DM, Redd SC, Mokdad AH, Mott JA. Body mass index and asthma incidence among USA adults. Eur Respir J. 2004;24(5):740-744. doi:10.1183/09031936.04.00088003
Huovinen E, Kaprio J, Koskenvuo M. Factors associated to lifestyle and risk of adult onset asthma. Respir Med. 2003;97(3):273-280. doi:10.1053/rmed.2003.1419
Guerra S, Sherrill DL, Bobadilla A, Martinez FD, Barbee RA. The relation of body mass index to asthma, chronic bronchitis, and emphysema. Chest. 2002;122(4):1256-1263. doi:10.1378/chest.122.4.1256
Kronander UN, Falkenberg M, Olle Z. Prevalence and incidence of asthma related to waist circumference and BMI in a Swedish community sample. Respir Med. 2004;98(11):1108-1116. doi:10.1016/j.rmed.2004.03.022
Leone N, Courbon D, Berr C, et al. Abdominal obesity and late-onset asthma: Cross-sectional and longitudinal results: The 3C study. Obesity. 2012;20(3):628-635. doi:10.1038/oby.2011.308
Wang T, Zhou Y, Kong N, Zhang J, Cheng G, Zheng Y. Weight gain from early to middle adulthood increases the risk of incident asthma later in life in the United States: a retrospective cohort study. Respir Res. 2021;22(1):1-9. doi:10.1186/s12931-021-01735-7
Wells G, Shea B, Robertson J, Peterson J, Welch V, Losos M. The Newcastle-Ottawa Scale ( NOS ) for Assessing the Quality of Nonrandomized Studies in Meta- Analysis Bias and Confounding Newcastle-Ottowa Scale. Ottawa Hosp Res Inst. Published online 2012. http://www.evidencebasedpublichealth.de/download/Newcastle_Ottowa_Scale_Pope_Bruce.pdf
World Cancer Research Fund International. Diet, Nutrition, Physical Activity and Cancer: A Global Perspective.; 2018. http://dietandcancerreport.org
Longnecker MP. Methods for Trend Estimation from Summarized Dose- Response Data , with Applications to Meta-Analysis. 1992;135(11):1301-1309.
Hamling J, Lee P, Weitkunat R, Amb M. Facilitating meta-analyses by deriving relative effect and precision estimates for alternative comparisons from a set of estimates presented by exposure level or disease category. 2008;(February 2007):954-970. doi:10.1002/sim
Bagnardi V, Zambon A, Quatto P, Corrao G. Flexible Meta-Regression Functions for Modeling Aggregate Dose-Response Data , with an Application to Alcohol and Mortality. 2004;159(11):1077-1086. doi:10.1093/aje/kwh142
Higgins JPT, Thompson SG. Quantifying heterogeneity in a meta-analysis. 2002;1558:1539-1558. doi:10.1002/sim.1186
M E, G DS, M S, C M. Bias in meta-analysis detected by a simple, graphical test. BMJ. 1997;315(7109):629-634. doi:10.1136/BMJ.315.7109.629
Begg CB, Mazumdar M. Operating characteristics of a rank correlation test for publication bias. Biometrics. 1994;50(4):1088. doi:10.2307/2533446
Van Der Weele TJ, Ding P. Sensitivity analysis in observational research: Introducing the E-Value. Ann Intern Med. 2017;167(4):268-274. doi:10.7326/M16-2607
Beuther DA, Sutherland ER. Overweight, obesity, and incident asthma: a meta-analysis of prospective epidemiologic studies. Am J Respir Crit Care Med. 2007;175(7):661-666. doi:10.1164/rccm.200611-1717OC
Jiang D, Wang L, Bai C, Chen O. Association between abdominal obesity and asthma: A meta-analysis. Allergy, Asthma Clin Immunol. 2019;15(1). doi:10.1186/s13223-019-0333-6
Martin SE, Mathur R, Marshall I, Douglas NJ. The effect of age, sex, obesity and posture on upper airway size. Eur Respir J. 1997;10(9):2087-2090. doi:10.1183/09031936.97.10092087
King GG, Brown NJ, Diba C, et al. The effects of body weight on airway calibre. Eur Respir J. 2005;25(5):896-901. doi:10.1183/09031936.05.00104504
Spathopoulos D, Paraskakis E, Trypsianis G, et al. The effect of obesity on pulmonary lung function of school aged children in Greece. Pediatr Pulmonol. 2009;44(3):273-280. doi:10.1002/ppul.20995
Santamaria F, Montella S, Greco L, et al. Obesity duration is associated to pulmonary function impairment in obese subjects. Obesity (Silver Spring). 2011;19(8):1623-1628. doi:10.1038/OBY.2011.1
Santamaria F, Montella S, Greco L, et al. Obesity duration is associated to pulmonary function impairment in obese subjects. Obesity (Silver Spring). 2011;19(8):1623-1628. doi:10.1038/OBY.2011.1
Dias-Júnior SA, Reis M, De Carvalho-Pinto RM, Stelmach R, Halpern A, Cukier A. Effects of weight loss on asthma control in obese patients with severe asthma. Eur Respir J. 2014;43(5):1368-1377. doi:10.1183/09031936.00053413
Eneli IU, Skybo T, Camargo CA. Weight loss and asthma: A systematic review. Thorax. 2008;63(8):671-676. doi:10.1136/thx.2007.086470
Peters U, Dixon AE, Forno E. Obesity and asthma. J Allergy Clin Immunol. 2019;141(4):1169-1179. doi:10.1016/j.jaci.2018.02.004
Shore SA. Obesity, airway hyperresponsiveness, and inflammation. J Appl Physiol. 2010;108(3):735-743. doi:10.1152/JAPPLPHYSIOL.00749.2009
Sutherland ER. Linking obesity and asthma. Ann N Y Acad Sci. 2014;1311(1):31-41. doi:10.1111/nyas.12357
Li Z, Leynaert B, Dumas O, et al. Role of Leptin in the Association Between Body Adiposity and Persistent Asthma: A Longitudinal Study. Obesity. 2019;27(6):894-898. doi:10.1002/oby.22466
Tilg H, Moschen AR. Adipocytokines: Mediators linking adipose tissue, inflammation and immunity. Nat Rev Immunol. 2006;6(10):772-783. doi:10.1038/nri1937
Shore SA. Obesity and asthma: Possible mechanisms. J Allergy Clin Immunol. 2008;121(5):1087-1093. doi:10.1016/j.jaci.2008.03.004
Bruun JM, Lihn AS, Pedersen SB, Richelsen B. Monocyte chemoattractant protein-1 release is higher in visceral than subcutaneous human adipose tissue (AT): Implication of macrophages resident in the AT. J Clin Endocrinol Metab. 2005;90(4):2282-2289. doi:10.1210/jc.2004-1696
Lommatzsch M. Airway hyperresponsiveness: New insights into the pathogenesis. Semin Respir Crit Care Med. 2012;33(6):579-587. doi:10.1055/s-0032-1325617

No competing interests reported.

Supplementarytablesandfigures.pdf

PDF

Editor invited by journal
12 May, 2022
Submission checks completed at journal
12 May, 2022
First submitted to journal
29 Apr, 2022

You are reading this latest preprint version

The Association between Body Mass Index, Abdominal fatness, Weight Change and the risk of Adult Asthma: A Systematic Review and Meta-Analysis of Cohort Studies

Status:

Version 1

Abstract

Figures

Introduction

Methods

Search strategy

Eligibility criteria

Study selection process

Data extraction

Study quality assessment

Evidence grading

Statistical analysis

Results

Characteristics of the included studies

Assessment of study quality:

BMI and adult asthma

Waist circumference and adult asthma

Weight change and adult asthma

Subgroup and sensitivity analyses

Evidence grading

Discussion

Declarations

References

Competing Interests

Supplementary Files

Status:

Version 1