RESULTS
There were 45 participants enrolled in this study, of which 28 met criteria for BPD in infancy, and 17 did not. Among several differences in the two groups of children, those with BPD had a lower gestational age at delivery, lower average birth weight, a greater percentage of males and lower height percentile at the time of evaluation (Table 1).
Of the 45 participants, 37 children had valid pedometer data and were included in the pedometer analyses. Five children were excluded due to insufficient days of recording, 2 children did not return their pedometer, and 1 had their data accidentally deleted prior to upload. An additional 4 participants of the 37 were missing data on MVPA due to pedometer data uploading issues at a single site.
Comparing children with and without history of BPD, there were no clear differences between the average number of daily steps, MVPA, or PAQ score (Table 2). Of the 37 participants with valid step count data, 27 had also had reliable PFT and MRI data. A Spearman correlation matrix was conducted to assess the correlation between daily step count, daily MVPA, PAQ score, PFT measures, and MRI measures (Table 3). While PAQ score was weakly correlated with both pedometer outcomes, there was a very strong correlation between average daily step count and average daily MVPA (r=0.90). A higher average daily step count was moderately correlated with higher FVC % predicted, lower LCI, and greater proton density at full expiration. A higher average daily MVPA was moderately correlated with higher FVC % predicted, higher TLC % predicted, lower LCI, and greater proton density at full expiration. All other pulmonary function measure associations were weak or negligible.
A linear regression analysis demonstrated an association between FEV1 % predicted and daily step count, after adjusting for MRI total proton density and BPD status. MRI total proton density and BPD status were not associated with average daily step count (Table 4a). A separate linear regression analysis with average daily MVPA as the outcome found similar results, with FEV1 % predicted marginally associated with average daily MVPA, after adjusting for MRI total proton density and BPD status (Table 4b). Sensitivity analyses examining the impact of site found that while it resulted in wider confidence intervals, the direction of association between FEV1 % predicted and both pedometer outcomes remained the same.
Of the 45 participants, 42 had medication data available. Of the 42 participants with available medication data, 6 were taking ADHD medications – 1 in the non-BPD group, and 5 in the BPD group. When we compared the average daily step count data by ADHD medication status, there were no differences between groups (Figure 1). Among children with BPD, the mean (standard deviation) daily step count of the 3/5 children who were taking ADHD medication for whom a valid step count was available, was 15772 (5054) compared to 16416 (4288) in the 21 children who were not taking ADHD medication (Figure 2).