Birgitta Versluijs

138 Patients and methods Pulmonary follow-up We perform routinely PFT at day 100, and at 5 and 10 years after HCT according to our national Dutch Childhood Oncology Group late effect screening protocol. During follow up of patients with Allo-LS, PFT is repeated more often. Outcomes The main outcome of interest was “Allo-LS free survival” defined as time from Allo-LS to last follow up whereby death or “refractory or progressive” Allo-LS were regarded as events. Other outcomes of interest: Non Relapse Mortality (NRM) defined as death from another cause than relapse of disease. Therapy response defined as 1) Poor response/ therapy failure: refractory disease leading to a switch in first line therapy and/or progres- sive lung disease leading to death or chronic severe pulmonary disease, including patient requiring lung transplantation or chronic ventilatory support. 2) Good response: impro- vement of respiratory situation (e.g. improvement of PFT, reducing oxygen dependency) leading to recovery of lung disease (while receiving the MP-pulse therapy). Exploratory outcomes of interest were: Effect of baseline PFT and early response (after 1st MP pulse) on main outcome of interest was analysed. In addition, we were interested in the evolution of Pulmonary Function tests (PFT) in Allo-LS free survivors. Statistical analysis Differences in patient characteristics between Allo-LS and non Allo-LS patients were tested using Pearson’s chi-square. Results with a p value < 0.05 were considered statis- tically significant. The duration of follow up was the time to death, or to stable chronic Allo-LS or the last assessment for survivors. Patients were censored at the date of last contact. Factors considered to influence outcome included patient variables (gender, RV status pre HCT, age at HCT, primary disease), transplantation variables (year of trans- plant, conditioning regimen (chemotherapy or TBI based), donor factors (stem cell sour- ce, HLA disparity, (un)related)) and factors at diagnosis of Allo-LS (time between HCT and Allo-LS, time between symptoms Allo-LS and start therapy, RV status at time-point Allo-LS, HRCT scores, extra oxygen requirement, Intensive Care Unit (ICU) admission because of mechanical ventilation, systemic GVHD treatment prior to Allo-LS, viral re- activation prior to HCT). Allo-LS was analyzed as a total group, IPS and BOS were also analyzed separately. Only in BOS patients (because data were lacking in majority of IPS patients), we analyzed the effect of baseline PFT (FEV1%, FEV1/FVC, RV/TLC at diagno- sis of Allo-LS), and early after initiation of treatment (FEV1% after first MP-pulse, change in FEV1% after first MP-pulse) on the outcomes. For the outcomes of interest, we used Cox proportional hazard models; univariable predictors of outcome with a p-value ≤ 0.10 in the total group were used for multivariable analysis. 8