Metabolic syndrome and incidence of asthma in adults: the HUNT study

Introduction

The prevalence of asthma has been increasing worldwide during the past half century [1]. This increase has been particularly evident in countries such as the UK and the USA [2, 3]. The increase in asthma prevalence coincided with an increasing prevalence of obesity and, since the late 1990s, researchers have focused their investigations on a possible causal link between these two diseases [4]. Over the past decade, numerous studies have confirmed that obesity is a risk factor for incident asthma in adults [5], with emerging evidence of a distinct asthma phenotype linked to obesity [6, 7].

Obesity is a major component of metabolic syndrome, a cluster of metabolic components that indicate a significantly increased risk of cardiovascular disease [8]. The other components of metabolic syndrome include elevated triglycerides, reduced high-density lipoprotein (HDL) cholesterol, high blood pressure, and elevated glucose or diabetes. To date, few studies have investigated the association of metabolic syndrome or its components with asthma development [9–14], with the notable exception of obesity [5]. Of the limited studies looking at metabolic syndrome or its nonobesity components, one found that metabolic syndrome was associated with asthma-like symptoms [9], while mixed results have been observed for the associations between the other nonobesity metabolic components and asthma [9–14]. Many of these studies, however, cannot assign temporality to the association, due to their cross-sectional design. Research also suggests that the associations between components of metabolic syndrome and asthma may vary by atopic status [11, 15]. We aimed to investigate the prospective associations between metabolic syndrome and its components with incidence of asthma in a large, population-based cohort of adults.

Results

Characteristics of the participants with (n = 2971) and without (n = 20 220) metabolic syndrome are shown in table 1. Participants with metabolic syndrome were generally older, less physically active, were less educated and were more likely to receive social benefits and to report economic difficulties (sex-specific characteristics of the cohort can be found in online supplementary table S1).

The association between metabolic syndrome and incident asthma is shown in table 2. Metabolic syndrome was associated with incident asthma (OR 1.69, 95% CI 1.42–2.01 in model 1). After adjustment for all covariates (model 2), the odds ratio was not substantially attenuated (OR 1.57, 95% CI 1.31–1.87). In sex-specific analyses, the odds ratios for females and males were 1.69 (95% CI 1.35–2.13) and 1.41 (95% CI 1.07 to 1.87) in model 2, respectively (online supplementary tables S2 and S3). In our subgroup analyses by allergic rhinitis, we observed similar associations between metabolic syndrome and cumulative incidence of asthma in participants with allergic rhinitis (OR 1.44, 95% CI 1.08–1.94) and participants without allergic rhinitis (OR 1.67, 95% CI 1.27–2.19) in model 2 (online supplementary table S4). In subgroup analyses stratified by heartburn/acid regurgitation (as a proxy for GERD), we also observed similar associations in subgroups with or without heartburn/acid regurgitation (data not shown). The sensitivity analysis with a stricter definition of incident asthma showed a moderate association consistent with the above findings: OR 1.42 (95% CI 1.13–1.79) in model 2 (table 3). Further exclusion of those reporting COPD at follow-up did not substantially change our findings (data not shown).

The associations of the components of metabolic syndrome with incident asthma are shown in table 2. Most components showed a weak-to-moderate association with incident asthma in model 1. However, after adjustment for all covariates and the other metabolic components (model 3), most associations were substantially attenuated. Only the associations for high waist circumference (OR 1.62, 95% 1.36–1.94) and elevated glucose or diabetes (OR 1.43, 95% CI 1.01–2.04) remained relatively stable after mutual adjustment for all covariates and the other metabolic components. Similar associations were observed in the sensitivity analysis using the stricter asthma definition (table 3). The odds ratio for high waist circumference was 1.55 (95% CI 1.23–1.95) and for elevated glucose or diabetes the odds ratio was 1.64 (95% CI 1.07–2.52) in model 3. Further excluding of those reporting COPD at follow-up did not change our results (data not shown). By contrast, the other components of metabolic syndrome were not significantly associated with incident asthma.

Discussion

In this large population-based study, we found that metabolic syndrome was associated with incident asthma in adults. This association was consistent in females and males, and for participants with or without allergic rhinitis. Our study also suggested that high waist circumference and elevated glucose or diabetes were the metabolic components driving this association.

The associations we observed between metabolic syndrome and asthma were consistent with one previous study, which found a moderate association between metabolic syndrome and asthma-like symptoms [9]. In this cross-sectional study of 10 038 Korean adults aged 40 to 69 years, participants with metabolic syndrome were more likely to report symptoms of wheeze, resting dyspnoea and post-exercise dyspnoea than their counterparts without metabolic syndrome. Our findings expand on this earlier work by investigating the association prospectively and in subgroups by allergic rhinitis (as a proxy measure for allergy status). Several distinct forms of asthma are recognised (allergic and nonallergic) and the underlying mechanisms can vary between these forms [15]. For example, mechanisms of allergic asthma include adaptive immunity pathways while mechanisms of nonallergic asthma include innate pathways. These forms can also coexist and interact [15]. Our observations suggest that the association between metabolic syndrome and incident asthma in adults does not differ by allergy status.

We found that after mutual adjustment for other metabolic components, the associations between high waist circumference or elevated glucose or diabetes and incident asthma remained. The association between high waist circumference and asthma is consistent with prior studies [19, 20], and it is generally accepted that obesity is a risk factor for adult asthma [5, 21]. The associations of both abdominal and general obesity with incident asthma have previously been investigated in the HUNT 2 cohort, in which incident asthma was associated with both abdominal and general obesity in females and mostly with by general obesity in males [17]. While this previous study attempts to separate the associations between general obesity and abdominal obesity with incident asthma, the present study uses abdominal obesity as a general marker of adiposity. Therefore, the association between high waist circumference and incident asthma in the present analysis reflects both underlying associations, as we did not adjust for body mass index in the analysis.

The association between elevated glucose or diabetes and incident asthma remained after mutual adjustment for the other metabolic components including waist circumference. This association is described less frequently in the literature. Two cross-sectional studies found no association between insulin resistance or elevated glucose and asthma [9, 11]. By contrast, one cross-sectional study and one prospective study, both from Denmark, reported an association [10, 12]. In the prospective study, Thuesen et al. [10] observed that insulin resistance was associated with incident wheezing and asthma-like symptoms after a 5-year follow-up [10]. The Danish population-based study (n = 3441) estimated insulin sensitivity using the homeostasis model assessment of insulin resistance (HOMA-IR). The HOMA-IR closely mirrors the glucose clamp technique and has been reported to be a reliable measure of insulin sensitivity for epidemiological studies [22]. However, the outcome variables were based on self-reported symptoms, which might have caused some misclassification of asthma. The inconsistency between the null findings in the former two cross-sectional studies [9, 11] and the positive findings in the studies from Denmark [10, 12], and our current study may be due to the differences in study designs and populations. For example, the prevalence of abdominal obesity was quite different across all five studies, being ∼54% and 7% in the former two cross-sectional studies [9, 11] compared with ∼20% in the studies from Denmark [10, 12] and 13% in the current analysis.

In our study, we found little evidence that the other metabolic components (i.e. elevated triglycerides, reduced HDL cholesterol, and elevated blood pressure or anti-hypertension medication use) were associated with incident asthma, after mutual adjustment for the other metabolic components. Our results are consistent with some previous studies, including one that did not find an association between elevated triglycerides and asthma [9], and two other studies that found no association between HDL cholesterol and asthma [9, 13]. However, our study was not consistent with the results from the Korean Health and Genome Study, which found that elevated blood pressure was associated with asthma-like symptoms [9]. This discrepancy may be due to differences in study design (cross-sectional in the Korean Health and Genome Study and prospective in the current study) and study populations, such as the higher mean age in the Korean study (mean age 55 years for subjects with metabolic syndrome) [9], versus the current analysis (mean age 43 years for subjects with metabolic syndrome). Adjusted covariates also differed between the analysis models in the two studies. Further research is needed to clarify the associations between the components of metabolic syndrome and asthma in prospective studies.

There are several possible mechanisms connecting metabolic syndrome and incident asthma. The associations we observed may be driven by obesity, which is thought to change pulmonary physiology, including changes in lung function [23]. Obesity-associated changes in cytokines, chemokines and energy regulating hormones have been suggested to promote local and systemic inflammation and may also play a role [23–25]. The obesity–asthma association may also be due to common genetics shared by both obesity and asthma [25, 26]. Several potentially confounding comorbidities have also been suggested, including GERD and sleep disordered breathing, which may explain this association [25, 26]. However, our results remained unchanged in subgroup analysis for participants with and without heartburn/acid regurgitation (as a proxy for GERD). Leaving aside obesity for a moment, the associations between elevated glucose or diabetes and incident asthma is intriguing. It has been hypothesised that systemic low-grade inflammation associated with insulin resistance may be one of the underlying mechanisms of asthma development [10–12], and the association did not seem to be explained by obesity in our study. Although our results support a moderate association between elevated glucose or diabetes and incident asthma, and support earlier findings that linked insulin resistance to asthma and asthma-like symptoms [10, 12], our study was somewhat limited in the investigation of this component.

A weakness of many epidemiological studies of asthma is the definition of asthma. In this study, asthma was defined by a survey questionnaire and further restricted to those who reported asthma, asthma medication use at follow-up and were free from wheeze and asthma at baseline. We were also able to exclude participants who reported COPD at follow-up. Although the diagnosis of asthma has not been confirmed by objective means, self-reported asthma has been evaluated as a reliable case finding method for epidemiology studies [27, 28]. Studies have also observed similar rates of asthma misdiagnosis in obese and nonobese people [29, 30]. In addition, asthma diagnosis is verified by a physician when asthma medication is prescribed in Norway. Our definition of metabolic syndrome was also limited. The blood serum measures were nonfasting, and both triglycerides and glucose are sensitive to eating, while HDL cholesterol is less affected [31]. To be stringent, we defined elevated glucose levels only when time between last meal and venipuncture was ≥4 h. Participants with elevated and <4 h since their last meal and no reported diabetes were included in the reference group, which could have underestimated the observed association.

We need to be cautious when interpreting the results for the individual components, as the categorisation of these components in the definition of metabolic syndrome are dichotomous and do not allow for accurate adjustment [8]. Further studies should address the individual components in more detail, for example using continuous fasting blood serum levels repeatedly measured over time and incidence of asthma. This would help to remove residual confounding and aid in separating the associations of the individual components of metabolic syndrome with incident asthma. Regarding medication, we only had information on antihypertensive medication use. Medication use may influence two aspects of the results. On the one hand, it may improve the study precision by identifying more cases of metabolic syndrome; on the other hand, adequate medication use may attenuate the associations if the underlying associations are due to elevated serum levels of metabolic components. Despite these study limitations, our study is the largest and longest prospective study to be conducted on the association of metabolic syndrome and its components with incident asthma in adults. The prospective study design allows us to assign temporality.

In this population-based study, metabolic syndrome and two of its components (high waist circumference and elevated glucose or diabetes) were associated with an increased risk of incident asthma in adults. This observation of a link between metabolic syndrome and asthma suggests that clinicians could consider metabolic syndrome as a marker of future risk of asthma. The findings of this study warrant further research, particularly addressing the individual metabolic components and the underlying mechanisms for these associations. Such work could improve the understanding of asthma pathogenesis and thereby assist ongoing efforts to prevent asthma.