Multiple Breath Washout Testing to Identify Pulmonary Chronic Graft Versus Host Disease in Children After Hematopoietic Stem Cell Transplantation.

Pulmonary chronic graft versus host disease (p-cGvHD) is a highly morbid, late complication of hematopoietic stem cell transplantation (HSCT). The 2014 National Institutes of Health cGvHD consensus criteria require a tissue biopsy or a drop in spirometry (with other features) to establish the diagnosis of p-cGvHD. Unfortunately, children are often incapable of performing spirometry, which can delay the diagnosis of this condition. Multiple breath washout testing (MBW) can detect abnormal pulmonary physiology in older children and adults after HSCT, but its feasibility and utility have not been assessed in younger children and in those who cannot perform spirometry. In this study, we assess the feasibility and sensitivity of MBW to detect p-cGvHD in children as young as 3 years of age after HSCT. We performed a cross-sectional analysis of children age 3 to 18 years, between 100 days and 5 years after allogenic HSCT. Participants were recruited from the HSCT population at BC Children's Hospital (Vancouver, Canada). All participants attempted nitrogen MBW, and children age 6 years and over attempted spirometry. Nonparametric statistical techniques were used; descriptive statistics used median (interquartile range [IQR]) and group medians were compared using Wilcoxon rank-sum test. Twenty-six children, median age 11.0 (range 3.6-18.5) years, were recruited a median of 26.4 (IQR 15.7, 51.8) months after HSCT. Six of the 26 children (23%) had a clinical diagnosis of p-cGvHD. MBW was successful in all (26/26, 100%) participants. The lung clearance index (LCI; the primary outcome of MBW) was higher in those with a history of p-cGvHD (median 11.8 [IQR 9.6, 18.7]) than in those with no history of cGvHD (median 7.7 [IQR 7.1, 8.0]; P = .001) or a history of extrapulmonary cGvHD (median 7.5 [IQR 6.9, 7.6], P = .007). A threshold LCI = 9 resulted in a sensitivity of 100% and specificity of 90% for the correct identification of clinically diagnosed p-cGvHD using MBW (area under the receiver operator characteristic curve is 0.97 [95% confidence interval 0.80, 0.99]). Spirometry was successful in most (17/26, 65%) participants. Similar to LCI, forced expiratory volume in 1 second (FEV1)/ forced vital capacity could distinguish between p-cGvHD and no cGvHD (P = .02) and extrapulmonary cGvHD (P = .01). FEV1 alone could not distinguish between either of these groups (P = .87, P = .24 respectively). MBW is feasible in young children after HSCT and in those who cannot perform spirometry. LCI has high discriminative power for correctly identifying p-cGvHD, but these preliminary results require confirmation in a larger validation cohort.