Skin Immune Stimulation By Infections and Drug Toxicities during Induction and/or Consolidations Increases Incidence of Skin Acute Graft Versus Host Disease in Acute Myeloid Leukemia: A Study on Behalf of SFGM-TC and ALFA

Introduction: 60-70% of AML patients have an indication of allogeneic hematopoietic stem cell transplantation (allo-HSCT) during their treatment. Among those who undergo allo-HSCT, prognosis and quality of life depended on presence or absence of graft versus host disease (GvHD). Immune stimulation supports the principle of GvHD and graft versus leukemia (GvL) after allo-HSCT. The impact of immune activation prior to allo-HSCT on the pathogenesis of GvHD has never been evaluated. The aim of this study was to determine whether immune stimulation induced by infection or drug toxicity before transplantation increased the incidence of acute GvHD (aGvHD) in AML patients.

Materials and methods: 345 AML patients were transplanted in first complete remission (CR) in 21 French centers from 2009 to 2013 after prospectively enrollment in the ALFA-0702 trial (patients aged 18-60y, de novo AML, favorable-risk AML excluded). Clinical data (skin, gut and liver infections or drug toxicities) before allo-HSCT were collected from the ALFA database and clinical data (skin, gut and liver aGvHD) after allo-HSCT were collected from the SFGM-TC database (ProMise). GvHD grading was defined according to Glucksberg criteria. Infection and drug toxicity grading was defined according to CTCAE v4.0. Mann-Withney and Kruskall-Wallis tests were used for continuous variables. Chi-square test was used for non-continuous variables. Overall survival (OS) and progression free survival (PFS) were assessed by Kaplan Meier method.

Results: Median age at transplant and M/F sex ratio were 45 years (range, 19-61) and 195/150, respectively. Cytogenetics was intermediate-1, intermediate-2 and unfavorable in 24, 164, 144 patients according to ELN classification, respectively. 82.5% of the patients had reached CR after one cycle of induction. 193 (53%) patients presented infections during induction and 46 (17%) during consolidations. Moreover, 110 (30%) patients suffered from drug toxicity during induction and 36 (10%) during consolidations. Allo-HSCT was performed in all patients after one to three cycles of consolidation. 132 (36%) patients received reduced intensity conditioning. Sex matching was female in male for 74 (21%) patients. ABO matching was matched for 187 and mismatched for 158 patients. HLA matching was related for 138 and unrelated for 198 patients. Source of graft was bone marrow, peripheral stem cell and cord blood in 108, 217 and 12 patients, respectively. 181 (47%) patients underwent aGvHD, most frequently skin aGvHD (stage 1-2: 27.5%; stage 3-4: 9.5%).

First, we observed a significant increase of incidence of aGvHD (all grades) if skin toxicities occurred during induction (45/57% for no toxicity and toxicities all grades, respectively p=0.07) or consolidations (46/70% for no toxicity and toxicities all grades, respectively p=0.04). Secondly, we observed a significant increase of skin aGvHD in cases of skin toxicities during induction [stage 0/1-2/3-4 skin aGvHD: 66/27/7% and 47/32/21% for no toxicity and toxicities all grades, respectively (p=0.001)] or consolidations [stage 0/1-2/3-4 skin aGvHD: 64/28/8% and 35/35/30% for no toxicity and toxicities all grades, respectively (p<0.003)]. Thirdly, we observed a correlation between infections and skin aGvHD during consolidations [stage 0/1-2/3-4 skin aGVHD: 64/28/8% vs 49/27/24% for no infection and infections, respectively (p=0.002)], and more particularly between skin infections and skin aGVHD [stage 0/1-2/3-4 skin aGVHD: 63/29/8% vs 44/23/33% for no infection and infections, respectively (p=0.003)]. No correlation was found between others types of infections or toxicities and aGvHD. Finally, we observed no impact of infections and/or toxicities on OS and PFS.

Conclusion: Skin immune stimulation induced either by infections during consolidations or by drug toxicity during induction and/or consolidations significantly increased the incidence of skin aGvHD. Nevertheless, we found no impact on OS and PFS in our cohort. Our results confirm the participation of inflammatory process in the physiopathology of GvHD. These data should be confirmed in a larger study to determine whether patients with prior infection and/or drug toxicity to allo-HSCT should receive different GvHD prophylaxis strategies.