Search for a command to run...
A major cause of death in young adults with sickle cell disease (SCD) is cardiopulmonary disease.1 Decreased lung function, defined as low forced expiratory volume in 1 second (FEV1)% predicted, is associated with earlier death in the general population, young adults with cystic fibrosis, and young adults with sickle cell anemia (SCA). In a prospective cohort of 430 adults with SCA, a baseline FEV1% predicted less than 70% was associated with early death.2 In a retrospective cohort study, a convenience sample (n = 413) of children with SCD (8-18 years of age) demonstrated a longitudinal decline in FEV1% predicted (approximately 2.5% per year) in children with SCA.3 In one of the few studies evaluating body mass index (BMI) and its relationship to longitudinal lung function in children with SCD, Koumbourlis et al., 4 did not reveal BMI as a significant risk factor for decline in FEV1% predicted (relative risk = 0.11; 95% confidence interval [CI], −0.29 to 0.50; P = 0.6).4 However, this study did not adjust for expected BMI percentile in children with SCA. Despite extensive research demonstrating the relationship between low BMI and low FEV1% predicted in children with cystic fibrosis, and the relationship between low FEV1% predicted and death in young adults with SCA, minimal research has been conducted to study the effects of low BMI on lung function in children with SCA. Our prior analysis in the Sleep and Asthma Cohort (SAC) study, an unselected cohort of children with SCA (defined here as HbSS or HbSβ0 thalassemia), demonstrated that age was the only predictor of FEV1% predicted, with a decline of 0.3% for every additional year of age (−0.30; 95% CI, −0.56 to −0.05; P = 0.020).5 Sex, asthma history, hemoglobin level, reticulocyte count, white blood cell count, incidence rate of severe acute pain requiring hospitalization, incidence rate of acute chest syndrome episodes, and hydroxyurea therapy were not associated with a decline in FEV1% predicted. Extending our analysis in the SAC cohort, we tested the hypothesis that BMI percentile was an independent predictor of FEV1% predicted in children with SCA. To calculate the BMI percentile, we used quantile regression to construct percentile growth curves using the Silent Cerebral Infarct Multi-Center Clinical Trial (SIT) as the reference population6 for BMI z-scores. We did not use percentiles from the World Health Organization (WHO), which are based on mean anthropomorphic measurements for children in the general population. The calculation of BMI percentile is specific to the SCA population in the age range of 5 to 18 years. In this study, we tested the hypothesis that there will be a curvilinear (nonlinear) effect of BMI percentile on FEV1% predicted, such that: (a) there is generally a positive association between low BMI percentile and FEV1% predicted; (b) the strength of the association decreases as BMI percentile increases; and (c) at high values of BMI percentile, the association reverses and becomes negative, and FEV1% predicted declines because in the general pediatric population, with increasing total body fat percent, a negative slope is expected with decreasing FEV1% predicted values, after for adjusting age and sex.7 A total of 182 participants in the SAC study had at least two spirometry evaluations. The median age at first testing was 10.5 years (ranging 5.0-17.4 years), with an average follow-up of 3.4 years (ranging 0.5-6.5 years) from baseline to the second test. In the current multivariable mixed model, we used the same set of variables, adding only the SCA-specific BMI percentile with both linear and quadratic terms to allow for the postulated curvilinear effect. In this modified model, age and BMI percentile were significant. BMI percentile was found to be positively associated with FEV1% predicted. The linear effect of BMI percentile is positive and significant (0.21; 95% CI, 0.06 to 0.35; P = 0.006) while the quadratic term is negative and significant (−0.001, 95% CI –0.003 to 0.000, P = 0.046). Substantively, the positive linear term means that FEV1% predicted increases with increasing BMI percentile. However, the negative quadratic term means that the magnitude of the association between BMI percentile and FEV1% predicted decreases as BMI increases, leveling off around 75%. At the high end of BMI percentile (in approximately the upper quartile), FEV1% predicted begins to decline as BMI percentile increases. To illustrate the relationship, Figure 1 shows the effect of BMI percentile on FEV1% predicted at two different ages, while holding all other covariates constant. As in the previous study, age remained significantly associated with FEV1% predicted, with a 0.6% decrease in FEV1% predicted for every additional year of age (−0.59; 95% CI, −0.91 to −0.27; P < 0.001). Similar to the prior multivariate analysis5 without BMI percentile included, the following covariates were not associated with a decline in FEV1% predicted: asthma history (−2.97; 95% CI ‑6.62 to 0.69, P = 0.11), hydroxyurea therapy (1.43; 95% CI, −0.95 to 3.82; P = 0.24), hemoglobin level (0.56; 95% CI, −0.84 to 1.97; P = 0.43), reticulocyte count (0.10; 95% CI, −0.21 to 0.42; P = 0.51), white blood cell count (0.20; 95% CI –0.26 to 0.67, P = 0.39), acute chest syndrome episode rate (−0.38; 95% CI, −1.25 to 0.49; P = 0.39), and incidence rate of severe acute pain requiring hospitalization (0.10; 95% CI, −0.30 to 0.50; P = 0.62). Our results provide compelling support that in children with SCA and low BMI percentile, based on children with SCA, and not the general population of children, improving overall nutrition to increase BMI toward the average may increase their FEV1% predicted over time. In contrast, for overweight or obese children, increasing BMI may decrease FEV1% predicted. Nothing to report.