METHODS
This was a retrospective study where data was collected from chart review of patients who underwent their first hematopoietic stem cell transplant at Riley Hospital for Children. A database maintained by the Pediatric Stem Cell Transplant Program was used and contained information on patients who underwent stem cell transplant from January 1, 2007 to April 30th, 2017. The database provided information such as demographics, date of diagnosis, and type of oncologic diagnosis. Additional inclusion criteria for the study were as follows: patients who performed pre-transplant DLCO that met American Thoracic Society (ATS) criteria, and a hemoglobin recorded within 7 days of their pulmonary function testing. This study was approved by the Indiana University Institutional Review Board.
Information on patients’ demographics, primary disease, pre-transplant pulmonary function tests, and chest imaging prior to transplant was compiled in an Indiana School of Medicine RedCap database. Post-transplant data collected included additional transplants, relapse status, graft versus host disease (GVHD) status, all post-transplant PFT data, mortality, and pulmonary complications. Post-transplant data was collected, including mortality data at one and three years. Pulmonary complications were defined as documented and treated pulmonary infections as well as clinically significant abnormal chest imaging. Mortality was defined as death by any cause.
Paired t-tests and ANOVA models were used to define any differences between the two equations at baseline and when stratifying by Hgb level. Logistic regression models were used to determine associations between the Dinakara and Cotes equation with mortality at one- and three-years post-transplant. Results were reported as odds ratios with 95% confidence intervals and p-values. Two sample t-test statistical analysis were then used to compare PFT’s and hemoglobin correction equations with and without pulmonary complications. Spirometry data were reported as z-scores while DLCOHgb values were reported as percent predicted values and p-values produced from student’s t- test were used for comparison. The Global Lungs Initiative reference equations for used for spirometry, and data from Kim et al. were used for DLCO reference equations. All analyses were performed using SAS v9.4 and all analytic assumptions were verified.