Results
Ninety-nine children were screened and found eligible for inclusion during the study period. Two subjects were excluded in view of refusal to provide consent and inability to follow instructions for spirometry on account of developmental delay. Fourteen children were further excluded as they did not report for sampling. Eighty-three participants were included in the study for analysis. Table 1 provides the baseline characteristics of the 83 children. Nine (10.8%) children were overweight where as none were obese.
No significant difference was observed in the average body mass index (BMI), BMI ‘z’ score, waist circumference, hip circumference, or waist-hip ratio between the three groups based on asthma symptom control.
HOMA-IR index did not show a normal distribution in the population. The median (IQR) HOMA-IR was 1.65 (1.06, 2.39). Thirty-five children were found to have IR taking a HOMA-IR cut-off at 1.77; thus, the prevalence of IR was 42.3% (95% CI: 31.7- 52.9%). The HOMA-IR index showed a significant positive correlation with weight, weight ‘z’ score, BMI, BMI ‘z’ score, waist circumference, hip circumference, and waist-hip ratio (E-table 1).
Seventy-two out of 83 children were found to have at least one lipid abnormality. The prevalence of dyslipidemia was estimated at 86.7% (79.4-94.0%). Number of children with elevated TG, TC, and LDL-C was 4 (4.8%), 4 (4.8%) and 5 (6%), respectively, whereas 67 (80.7%) children had low HDL-C. Only one subject was found to fulfil the criteria for MS.
The median fasting serum insulin level and HOMA-IR index were compared between the asthma symptom control groups. An increasing trend was observed from controlled to partly controlled and uncontrolled, which was statistically significant in case of insulin level but not HOMA-IR. A similar statistically significant increasing trend was observed for median TG and LDL-C levels. There was no significant difference in the median TC and HDL-C levels. A significant difference was not seen for prevalence of IR (HOMA-IR ≥1.77). These comparisons are summarized in table 2, figures 1A, B and C.
We used an ordinal logistic regression model to adjust for BMI, separately for each significant variable, including fasting insulin, serum TG and LDL-C. Except for LDL-C, the rest of the variables showed a significant positive correlation with poorer asthma symptom control, after adjusting for BMI.
We correlated individual spirometry indices, including observed and percentage predicted values of FEV1, FVC, FEV1/FVC, PEFR, FEF25, FEF50 and FEF75 to HOMA-IR index, fasting serum insulin and lipid levels. Interestingly, HOMA-IR index showed a statistically significant but weak positive correlation with FEV1 and FEV1 percentage predicted, and serum insulin level showed a similar correlation only with FEV1 percentage predicted. These correlations, however, were not significant on adjusting for height. Also, serum TG, TC, LDL-C, and HDL-C all showed a statistically significant weak negative correlation with PEFR, which were again not significant on adjusting for height. These results are summarized in table 3.
A HOMA-IR cut off of 2.5 has been shown to be more sensitive and specific in Indian adolescents 10 to 17 years of age [23]. On taking this higher cut-off at 2.5, 17 children were found to have IR [prevalence 20.5% (11.8-29.2%)] in our study population. The prevalence of IR (HOMA-IR ≥2.5) also showed a statistically significant increase, from controlled to partly controlled, and further to uncontrolled asthma, after adjusting for BMI.