Impact of Insulin Concentrations and Insulin Resistance on Lung Function and Morbidity in Asthma

To the Editor:

Prior work has shown insulin resistance (IR) modifies the association between obesity and asthma¹ and is associated with decreased lung function in adolescents and children with asthma in cross-sectional national surveys,² providing a potential mechanistic pathway driving obesity-associated asthma. IR has been shown to be associated with decreased lung function and increased lung function decline in the Severe Asthma Research Program (SARP-3) cohort.³ Hyperinsulinemia has been shown to lead to bronchoconstriction in animal models but remains an unstudied mechanistic pathway in humans with asthma.⁴ We sought to replicate the findings in SARP-3 in those with mild-to-moderate asthma and to evaluate the impact of longitudinal changes in insulin concentrations on lung function and asthma outcomes.

Study Design and Methods

We used data from participants with asthma diagnosed by bronchodilator responsiveness or methacholine challenge who were undergoing research bronchoscopy at the University of Colorado and data from participants at a baseline visit from a longitudinal cohort of patients with asthma recruited from primary care and pulmonary clinics in Aurora, Colorado, and New York City, New York, which has been described elsewhere.⁵ Fasting glucose, insulin concentrations, and spirometry were obtained at the same visit for both cohorts. In the longitudinal study, fasting blood insulin concentrations, spirometry, asthma control (ACQ⁶), and asthma quality of life (AQLQ⁷) were obtained every 6 months for 18 months. IR was measured using the homeostatic model assessment for insulin resistance (HOMA-IR) using the following formula: fasting insulin (mU/L) × fasting glucose (mg/dL)/405. This is a validated surrogate for the criterion standard euglycemic hyperglycemic clamp test.^8,9 In the cross-sectional analysis, Spearman correlation and ordinal logistic regression was used to evaluate association between HOMA-IR and spirometry with adjustment for BMI. For longitudinal analysis, we used (1) linear mixed effect models to determine the relationship between insulin concentrations and spirometry, ACQ, and AQLQ; and (2) Pearson correlation of within-patient differences to assess how changes in insulin concentration between visits relates to changes in spirometry, ACQ, and AQLQ. Analyses were performed in R version 4.4 (R Foundation for Statistical Computing).

Results

In the cross-sectional analysis, there were 177 participants: 50 from the bronchoscopy cohort and 127 from the baseline visit of the longitudinal cohort. Participants from the bronchoscopy cohort had a median age of 34 years (interquartile range [IQR], 28–49), included 78% women, included 84% who self-identified as White race, had a median BMI of 35 kg/m² (IQR, 29–41), and had a median baseline fasting glucose of 87 mg/dL (IQR, 81–07). The median HOMA-IR was 3.04 (IQR, 1.73–5.13). The median FEV₁ percent predicted was 94% (IQR, 87%−104%) and median FVC percent predicted was 99% (IQR, 90%−110%). In the longitudinal cohort, fasting serum insulin concentrations were available for 124 participants at baseline, with a total of 329 observations over 18 months. The median age was 43 years (IQR, 29–56), 74% of participants were women, 68% self-identified as White, and the median BMI was 28 kg/m² (IQR, 24–33), which did not significantly change within participants over the follow-up period. Fourteen participants (11.2%) reported having comorbid diabetes. The median baseline fasting glucose (n = 102) was 88 mg/dL (IQR, 80–94) and median HOMA-IR was 0.79 (IQR, 0.41–1.31). The median FEV₁ percent predicted was 92% (IQR, 77%−102%) and median FVC percent predicted was 106% (IQR, 93%−118%). At baseline, HOMA-IR was inversely associated with FVC percent predicted (R = −0.26, P = .0012) (Fig 1); this result remained statistically significant after adjusting for BMI (P = .04). HOMA-IR did not have a statistically significant association with percent predicted FEV₁ in liters or percent predicted FVC in liters.

Figure 1 –

Cross-sectional analysis of weight-adjusted HOMA and percent predicted FVC among 2 independent asthma cohorts. The 95% CIs are depicted by the gray shading. HOMA = homeostatic measure of insulin resistance.

The intraclass correlation coefficient (ICC) for insulin over the follow-up period was 0.439, indicating variability over time within participants. Insulin variability did not meaningfully vary when stratified by early vs late asthma onset, presence of type 2 inflammation, obesity, or sex. Those with baseline asthma control by ACQ score had higher variability (ICC, 0.304) than those without asthma control (ICC, 0.598). In multivariable linear mixed effect models adjusting for BMI, age, race, sex, and presence of type 2 inflammation, insulin concentration was inversely associated with worse AQLQ (effect estimate for the expected change in AQLQ for a 1-SD increase in log insulin, −0.13; 95% CI, −0.22 to −0.03); however, there was no evidence of associations with FEV₁ or ACQ. In differenced models, insulin change was associated with (1) worse asthma control (ACQ: R = 0.17, P = .018; adjusting for BMI: P = .019) (Fig 2), and (2) reduced asthma quality of life (AQLQ: R = −0.19, P = .0078; adjusting for BMI: P = .011) (Fig 2). We found no association between change in insulin concentrations and change in FEV₁ (R = −0.0019, P = .98) or change in FVC (R = −0.025, P = .74).

Figure 2 –

A-B, Longitudinal analysis of the change in plasma insulin concentrations from baseline with changes in asthma control and asthma quality of life: (A) relationship between fasting plasma insulin concentrations on asthma control, and (B) relationship between fasting plasma insulin concentrations and asthma quality of life. The 95% CIs are depicted by the gray shading.

Discussion

To our knowledge, this is the first study to demonstrate that insulin concentration is negatively associated with asthma quality of life and that, independently of BMI, increased insulin concentrations over time relate to worsening asthma control and reduced quality of life. A recent prospective study found that baseline IR was associated with an increased probability of asthma exacerbation over 1 year¹⁰; however, like the SARP-3 study, this study used a single measurement of IR to predict exacerbation risk¹⁰ or lung function decline.³ Our work leverages longitudinal measurements of insulin and can demonstrate change in fasting insulin within patients over time. Why increasing insulin concentrations adversely impact asthma control and quality of life is unknown but could be related to changes in airway epithelial cellular carbohydrate metabolism, which has been associated with increased incidence of wheezing.¹⁰ Glucose was not measured in longitudinal samples, limiting our ability to test the impact of IR and change in IR with asthma outcomes, and data on glucose altering medications are not available in this cohort. The lack of longitudinal associations with lung function is not surprising given the length of follow-up in a mild-to-moderate asthma cohort; however, prior work has shown that hyperinsulinemia, which is often a compensatory mechanism in those with IR, may increase vagally mediated bronchoconstriction and impact lung function in animal models.⁴ These findings are exploratory in nature and additional work is needed to understand how longitudinal changes in IR influence asthma outcomes independent of BMI.

Cross sectionally, we demonstrated that IR is negatively associated with FVC, independent of BMI. This is concordant with findings from Peters et al³; however, notably their study was in patients with severe asthma with more severe IR than the current cohort. Typically, low FVC is attributed to extrathoracic restriction from obesity; however, in our study and the Peters et al³ study, this finding appears to be independent of BMI. The mechanism driving decreased FVC is unknown; however, IR may drive airway inflammation, drive airway smooth muscle proliferation, and induce fibroblast collagen deposition.¹¹ Although there is a biologic basis for this finding, additional work is needed to better understand these mechanisms.

A deeper understanding of the impact of IR on asthma outcomes is needed including future work evaluating longitudinal changes in IR on asthma outcomes, and mechanistic studies are needed to validate our findings. This future work is particularly important because existing therapies to treat IR exist and have potential to impact this difficult-to-treat asthma phenotype.