ERRATUM Predicting survival for myeloid leukemia after HLA-identical sibling donor allogeneic stem cell transplantation D Gallardo, R De la Ca ´mara, A Torres, S Brunet, A Jime ´nez, JC Vallejo, G Sanz, D Serrano, E Carreras, C Martı ´n, C Sanz-Rodrı ´guez, J Sierra, I Espigado, D Caballero and JJ Berlanga Leukemia (2004) 18, 1165–1167. doi:10.1038/sj.leu.2403397 Correction to: Leukemia (2004) 18, 1031–1034. doi:10.1038/sj.leu.2403344 Due to a typesetting error, the above article was published in print containing a Table 1 on page 1033, which is not part of this article. The correct version of the paper is reproduced below. The Publisher would like to apologise for any inconvenience this may have caused. CORRESPONDENCE Predicting survival for myeloid leukemia after HLA-identical sibling donor allogeneic stem cell transplantation Leukemia (2004) 18, 1031–1034. doi:10.1038/sj.leu.2403344 Published online 18 March 2004 TO THE EDITOR Several variables have been previously identified as adverse risk factors for survival after HLA-identical sibling donor allogeneic stem cell transplantation (SCT). Indeed, advanced phase of disease is associated with increased relapse rate and death due to leukemia progression. On the other hand, factors influencing the incidence and severity of graft-versus-host disease (GvHD), such as sex mismatch (female donor and male recipient), older recipient age or positive CMV serology, are related to an increased transplant-related mortality. 1,2 However, it is some- times difficult to know the real impact of these risk factors on transplant outcome when considering a single patient, because favorable and adverse conditions may be present at the same time. In this work, we present a clinical score for predicting the survival of allogeneic SCT from HLA-identical sibling donors for myeloid malignancies on the basis of the evaluation and grading of risk factors available pretransplant. We retrospectively analysed the pretransplant data and the clinical post-transplant outcome for 319 adult patients diag- nosed of acute myeloid leukemia (AML) (n ¼ 150) or CML (n ¼ 169) who received an allogeneic SCT from an HLA- identical sibling donor from 1991 to 2002, performed in 14 centers. Patients with AML were assigned to the following FAB groups: M0: 3%; M1: 15%; M2: 24%; M3: 6%; M4: 25%; M5: 18%; M6: 3%; M7: 1%; post-MDS: 5%. The median follow-up for living patients was 2.8 years (range: 3 months–13 years). All the patients were older than 15 years and conditioned with myeloablative regimens. Patients with AML beyond first complete remission and CML beyond first chronic phase were considered as with an advanced phase of disease. We considered sex mismatch as a female donor and a male recipient. GvHD was diagnosed and classified according to the Seattle’s criteria. Transplant-related mortality (TRM) was defined as a death without disease relapse. Univariate analysis for overall survival (OS) and disease-free survival (DFS), as well as the actuarial probabilities for grades III– IV acute GvHD, TRM and relapse were calculated by the Kaplan and Meier method. Comparison of curves was performed using the log-rank test. Multivariate analysis for OS was performed using the stepwise proportional hazard regression model. The actuarial OS at 8 years was 59.2% for the whole cohort of patients. The multivariate analysis detected only three indepen- dent risk factors for poor survival: Age beyond 30 years (P ¼ 0.003, OR: 2.40; 95% CI: 1.34–4.30), advanced phase of disease (Po0.001, OR: 3.03; 95% CI: 1.92–4.78) and female donor for a male recipient (P ¼ 0.023, OR: 1.71; 95% CI: 1.07– 2.74). Based on the results obtained in the multivariate analysis, patients were divided in four groups (or scores) depending on the following criteria: score 1: no risk factors (age o30 years, early phase of disease and absence of sex mismatch), 58 cases; score 2: early phase of disease with 1 or 2 risk factors (age 430 years7sex mismatch), 182 cases; score 3: advanced disea- se7age beyond 30 years or sex mismatch, 54 cases; and score 4: all three risk factors, 22 cases. Figure 1 shows the actuarial OS and DFS for each score. The actuarial OS was 94.2, 59.1, 45.7 and 8.3% for patients with scores 1–4, respectively. These differences were statistically significant (score 1 vs score 2: Po0.001; score 2 vs score 3: P ¼ 0.001; score 3 vs score 4: P ¼ 0.010). Actuarial DFS was 72.9, 51.6, 29.9 and 0% for each group. The comparison between each score was also statistically significant (score 1 vs Received 3 December 2003; accepted 16 January 2004; Published online 18 March 2004 Correspondence: Dr D Gallardo, Clinical Hematology Department, Institut Catala ` d’Oncologia, Hospital Dura ´n i Reynals. Avda, Gran vı ´a s/n, km 2.7 08907 L’Hospitalet, Barcelona, Spain; Fax: þ 34932607797; E-mail: 27532dgg@comb.es On behalf of the Spanish Group of Hemopoietic Transplant (GETH). Leukemia (2004) 18, 1165–1167 & 2004 Nature Publishing Group All rights reserved 0887-6924/04 $30.00 www.nature.com/leu