Summary:We have performed univariate and multivariate analysis to determine the factors that affect the kinetics of neutrophil and platelet recovery in 546 recipients of T cell-depleted (TCD) marrow allografts. All patients received marrow depleted of mature CD3 ؉ T cells by complement-mediated lysis using T 10 B 9 -1A3 (n = 489) or Muromonab-Orthoclone OKT3 (n = 57) monoclonal antibodies. Neutrophil engraftment to 0.5 ؋ 10 9 /1 and platelet engraftment to 20 ؋ 10 9 /l were assessed as endpoints. Factors significantly affecting neutrophil or platelet engraftment in the univariate analysis included patient age, T cell dose, anti-thymocyte globulin use, gender, diagnosis at transplant, CMV serostatus, HLA mismatch, CD34 cell dose (n = 249), and growth factor use and type. These variables were included in the multivariate Cox proportional hazards regression model. The results showed that a faster rate of neutrophil engraftment was independently associated with CD34 + cell dose у5 ؋ 10 6 /kg and most strongly with growth factor administration. Faster platelet engraftment was associated with transplantation for chronic leukemia, CD34 + cell dose Ն2 ؋ 10 6 /kg, an HLA matched related donor, and the absence of growth factor use. G-CSF had a higher relative risk (RR) of enhancing neutrophil engraftment than GM-CSF and significantly delayed platelet engraftment. The combined use of G-CSF ؉ GM-CSF was similar to G-CSF alone. The enhancing effect of G-CSF for neutrophil recovery was most striking for patients who engrafted to 0.5 ؋ 10 9 /1 at or before day 12 (RR = 9.5, P Ͻ 0.0001) compared to patients who received no growth factor. Conversely, the delaying effect of G-CSF on platelet engraftment was strongest for patients engrafting on or before day 25 (RR = 0.4, P = 0.0004). Of the independent variables affecting engraftment kinetics in recipients of TCD marrow allografts only growth factor, and to a limited extent, CD34 ؉ cell dose can be controlled by the