Relationship of physical function to outcomes in autologous hematopoietic stem cell transplantation.

Abstract

116 Background: Physical function is an important consideration for autologous hematopoietic stem cell transplantation (auto-HSCT) but it is often judged subjectively rather than tested. This study examined the ability of functional test results to predict key transplant outcomes versus subjective rating. Methods: Patients receiving outpatient auto-HSCT performed a pre-transplant 2-minute walk test (2MWT), isometric handgrip testing, 5-times sit-to-stand test (5XSST) and timed up and go test (TUG). Karnofsky Performance Status (KPS) score was assigned by the treating oncologist. At day 100 following auto-HSCT, charts were extracted for age, gender, diagnosis, total number of hospital days, and days of platelet and neutrophil recovery. Predictions were determined through stepwise regression analysis. Results: Data from 37 patients (24 M, 13 F) were analyzed. The most common diagnosis was multiple myeloma. One patient expired prior to day 100. KPS was weakly correlated with all functional tests (spearman rho values 0.15 to 0.27, p values 0.07 to 0.32). KPS did not predict # hospital days (p = 0.53), neurophil recovery (p = 0.22) or platelet recovery (p = 0.51). In contrast, there was a trend for 5XSTS to predict # hospital days (p = 0.10), although predictive ability was low (R2 = 0.07). TUG was the best predictor of platelet recovery (p = .021), followed by 2MWT (p = 0.23) (R2 = 0.10). KPS did not predict any outcome measure, and none of the variables predicted neutrophil recovery. Conclusions: KPS does not correlate well with physical function testing nor does it predict key auto-HSCT short-term outcomes. While not a major contributor to hospitalization days or blood count recovery time, 5XSST and TUG appear to have potential to predict transplant outcomes. Future testing with larger samples is needed to determine the impact of physical function on long-term auto-HSCT outcomes.