Individual quality assessment of autografting by probability estimation for clinical endpoints: a prospective validation study from the European group for blood and marrow transplantation

The aim of supportive autografting is to reduce the side effects from stem cell transplantation and avoid procedure-related health disadvantages for patients at the lowest possible cost and resource expenditure. Economic evaluation of health care is becoming increasingly important. We report clinical and laboratory data collected from 397 consecutive adult patients (173 non-Hodgkin lymphoma, 30 Hodgkin lymphoma, 160 multiple myeloma, 7 autoimmune diseases, and 28 acute leukemia) who underwent their first autologous peripheral blood stem cell transplantation (PBSCT). We considered primary endpoints evaluating health economic efficacy (eg, antibiotic administration, transfusion of blood components, and time in hospital), secondary endpoints evaluating toxicity (in accordance with Common Toxicity Criteria), and tertiary endpoints evaluating safety (ie, the risk of regimen-related death or disease progression within the first year after PBSCT). A time-dependent grading of efficacy is proposed with day 21 for multiple myeloma and day 25 for the other disease categories (depending on the length of the conditioning regimen) as the acceptable maximum time in hospital, which together with antibiotics, antifungal, or transfusion therapy delineates four groups: favorable (≤7 days on antibiotics and no transfusions; ≤21 [25] days in hospital), intermediate (from 7 to 10 days on antibiotics and <3 transfusions, ≤21 to 25 days in hospital or ≥7 days on antibiotics and no transfusions; from 21 to 30 days [25 to 34] in hospital), unfavorable (>7 days on antibiotics, >3 but <6 transfusions; >30/34 days in hospital after transplantation), and very unfavorable (>10 days on antibiotics, >6 transfusions; >30 to 34 days in hospital). The multivariate analysis showed that (1) PBSC harvests of ≥4 × 106/kg CD34 + cells in 1 apheresis procedure were associated with a favorable outcome in all patient categories except acute myelogenous leukemia and acute lymphoblastic leukemia (P = .001), (2) ≥5 × 106/kg CD34 + cells infused predicted better transplantation outcome in all patient categories (P < .0001) except acute myelogenous leukemia and acute lymphoblastic leukemia, (3) 1 or 2 aphereses (P = .001) predicted good outcome, (4) toxicity increased with higher graft volume reinfused (>500 mL) (P = .002), and (5) patients with a central venous catheter during both collection and infusion of PBSC had a more favorable outcome post-PBSCT than peripheral access (P = .007). The type of mobilization regimen did not affect the outcome of auto-PBSCT. The present study identified predictive variables, which may be useful in future individual pretransplantation probability evaluations with the goal to improve supportive care.