ALBERT

Description: Abstract 4490 Background: Imatinib is currently standard therapy for patients with success in Chronic Myeloid Leukemia (CML). Several instances of cardiac adverse events have been reported for patients while on therapy with imatinib. In many instances, these events occur in patients with pre-existing cardiac conditions. The aim of our study was to determine the occurrence of cardiac events in patients with CML treated with Imatinib, and the impact that cardiac risk factors and pre-existing cardiac conditions had on the risk of developing cardiac adverse effects. Methods: We reviewed the medical records of 51 patients with chronic phase CML who were treated with imatinib after failing prior therapies. For each patient we collected cardiac risk factors, pre-existing cardiac disease, pre-treatment EKG and echocardiogram (ECHO) readings, as well as post-treatment changes in EKG and ECHO findings. Results: Pre-existing cardiac conditions were found in 14 (27%) patients, including congestive heart failure in 2 (4%), myocardial infarction in 4 (8%), atrial fibrillation in 1 (2%), benign arrhythmias in 1(2%), aortic regurgitation in 1(2%), mitral valve prolapse in 1 (2%), mitral regurgitation in 1(2%), pericarditis in 1(2%), bradycardia in1(2%) and benign arrhythmia in 1(2%). Cardiac risk factors were present in 26 patients (51%), including smoking in 10 patients (20%), hypertension in 17 (33%), diabetes mellitus in 9 (18%), obesity in 2 (4%), hyperlipidemia in 5 (10%), stress (self-reported by patient or on anxiolytic therapy) in 5 (10%), alcohol in 11 (22%), atherosclerosis in 3 (6%), and positive family history for cardiac disease in 5 patients (9.8%). Cardiac events were noted in 19 patients (37%) of whom 11 (58%) had pre-existing cardiac conditions prior to initiating imatinib treatment and 14 patients (27%) had at least one cardiac risk factor. Congestive heart failure with clinical manifestations was seen in 9 patients (18%) with documentation of decreased ejection fraction on echocardiogram seen in 3 patients (6%) who had a decrease in LVEF from a median of 55% (range 50% to 72%) to a median of 45% (range 25% to 60%). Out of the patients who developed CHF while on treatment with imitanib, 3 patients (6%) had history of cardiac conditions (atrial fibrillation in 1 (2%), congestive heart failure in 2 (4%)). Myocardial infarction was documented in 3 patients (6%), one of which had prior myocardial infarction and pacemaker, another had history of mitral valve prolapse and hypertension; one patient had hypertension, diabetes mellitus and positive family history but no prior history of heart disease. Arrhythmia was seen in 3 patients (6%). Post-treatment EKG changes occurred in 14 patients (27%) including bradycardia, PAC's, PVC's, ST-T wave changes, tachycardia and other rhythm abnormalities. These changes were usually asymptomatic. Gated cardiac study done after a median of 63 months (range 29 to 83 months) after initiation of imatinib treatment showed EF below 60% in 9 patients (18%) with a median of 55% (range 36% to 59%). None of the patients died of cardiac conditions and none discontinued imatinib therapy because of cardiac events. Conclusion: Although cardiac events occur in some patients treated with imatinib, these are much more common among patients with pre-existing cardiac conditions and/or cardiac risk factors. These patients need to be monitored closely to minimize their risk and intervene early when new cardiac events arise. Disclosures: Cortes: novartis: Research Funding; BMS: Research Funding; Pfizer: Consultancy, Research Funding. Kantarjian:novartis: Research Funding; BMS: Research Funding. O'Brien:Novartis: Research Funding.

Description: Abstract 1486 Background: Corticosteroid therapy is a well-known cause of osteoporosis. Studies have shown that the use of bisphosphonates such as risedronate can prevent bone loss in patients receiving moderate to high doses of corticosteroids. Patients with acute lymphoblastic leukemia and lymphoblastic lymphoma (ALL/LL) are at high risk of bone loss due to both leukemia and high dose steroids used in chemotherapy regimens. Objectives: The primary objective was to determine whether the use of risedronate decreases the amount of bone loss in adult patients with ALL/LL receiving chemotherapy with high dose corticosteroids when compared to placebo. Methods: Adult patients with newly diagnosed ALL/LL receiving Hyper-CVAD or BFM were randomly assigned to receive either risedronate 35 mg weekly or placebo. Both patients and physicians were blinded to the treatment assignment. Patients in both arms were given vitamin D and calcium. Bone mineral density (BMD) using DEXA scan was measured at baseline and at 6 months. Patients who lost more than 10% BMD when compared to baseline measurements were taken off the study and unblinded. We excluded patients with osteoporosis. Results: 36 patients were randomized to each treatment arm. Median age was 29 years in the risedronate arm and 42 years in the placebo arm (p=0.06). 22 patients had an available 6 month DEXA scan in the risedronate arm and 29 in the placebo arm. Patients on the placebo arm had a greater decrease in BMD at the left hip (mean −0.12) and right hip (mean −0.11) when compared to baseline than patients in the risedronate arm (left hip mean −0.08; right hip mean −0.078) and this was statistically significant (p=0.03 and p=0.04, respectively). There were no significant differences observed in the lumbar spine between the two groups. 11/22 (50%) patients in the risedronate arm were taken off study for significant % decrease in BMD at any site, and 18/29 (62%) patients in the placebo arm (p=0.28). Two patients, 1 in each arm, experienced a vertebral fracture during the first 6 months of chemotherapy. Conclusions: Adult ALL/LL patients receiving high dose corticosteroids with HyperCVAD have profound bone loss and should be monitored for this while receiving chemotherapy. Risedronate significantly reduced bone loss at the hips, but did not have a significant effect at the lumbar spine. Despite treatment with risedronate, 50% of patients had profound bone loss during the first 6 months of chemotherapy, thus a more potent bisphosphonate may be necessary to preserve bone mass, particularly at the lumbar spine. Whether a potent bisphosphonate or an agent such as a monoclonal antibody to RANK ligand will reduce fractures and improve quality of life should be studied in future clinical trials. Disclosures: Hu: Amgen: Research Funding, Speakers Bureau. Cabanillas:Procter and Gamble: Research Funding.

Description: Abstract 1013 Poster Board I-35 AKI is a common complication in pts with hematologic malignancies. However, its precise incidence, predisposing factors and the clinical outcome associated with AKI in these pts have not been studied. The RIFLE criteria (R= risk, I = injury; F= failure, L= loss of kidney function, and E = end-stage kidney disease), is a multilevel classification system for AKI. The RIFLE system was proposed with the objective to define and classify AKI and to predict mortality. Objectives: a) to study the incidence of AKI in pts with AML and HR-MDS undergoing intensive chemotherapy (IC); 2) to describe the risk factors for the development of AKI 3) to assess the RIFLE criteria to predict mortality. Methods: Retrospective review of pts with AML or HR-MDS who underwent IC from 1999 to 2007. Pts with chronic kidney disease (baseline serum creatinine [Cr] 〉 1.5 mg/dL) were excluded. Backward stepwise logistic regression was used to determine predictors of AKI, defined as at least a 50% rise in Cr from baseline during IC. Severity of AKI was further defined by the RIFLE criteria (Risk = 〉50% rise in Cr, Injury = 〉100% rise in Cr, and Failure = 〉200% rise in Cr or requiring dialysis). Kaplan-Meier estimates for 8-week cumulative mortality were determined for each RIFLE category. Results: 537 pts were included. Median age was 55 yo (range 17-84), 52% were male, 88% were treated for AML, 97% had good performance status. All the pts received high-dose cytarabine (HD-AC) containing regimens (46% in combination with idarubicin; 18% with daunorubicin; 17 with fludarabine, 10% with clofarabine and 9% with cytoxan + topotecan). Fifty-two percent of the pts received vancomycin at any time from day 1 of CHEMO to the day of maximum Cr, 7% received aminoglycosides, 44% amphotericin B lipid formulation, 59% diuretics, and 8% each required vasopressors or mechanical ventilation. Mortality at 8-weeks among pts in the RIFLE category was significantly higher than in pts without AKI (p=0.000). Additionally, there was a progressive increase in 8-week cumulative mortality with worsening RIFLE class (p =0.000) (Table 1). Univariate analysis for risk factors for AKF revealed that older age, baseline bilirubin, ICU admission, mechanical ventilation, use of vasopressors, intravenous contrast, vancomycin, aminoglycosides, amphotericin B lipid formulation, diuretics, presence of infection at start chemo, and treatment with HD-ARA C + anything other than fludarabine were associated with higher risk for AKI. The following variables remained significant in the multivariate analysis: older age (OR:1.03, p

Description: Background: IFI is the most frequent cause of mortality in pts with AML and MDS undergoing chemotherapy. None of the antifungal prophylactic regimens used since 1992 at MD Anderson Cancer Center appeared to be significantly superior in the prevention of IFI. Study Aims: To compare the efficacy and safety of IV-VORI versus IV-ITRA as antifungal prophylaxis in AML and MDS pts receiving chemotherapy Patients and Methods: Pts older than 18 years old receiving induction or salvage chemotherapy, without documentation of prior IFI were eligible. Pts were randomized on day 1 of chemotherapy to receive IV-VORI 400 mg q12 h x 2 days followed by 300 MG IV twice per day or IV-ITRA 200 mg q12 h x 2 days followed by 200 mg IV once per day. Prophylaxis continue until recovery from neutropenia, developed possible/proven IFI, complete remission, declared resistant or up to 35 days for induction pts and up to 42 days for salvage pts. Results: 114 pts were evaluable (49 on IV-ITRA, 65 on IV-VORI). Baseline characteristics were similar in both groups. 102 were induction pts and 12 were on first salvage. Median time on prophylaxis was 21 (induction) and 17 days (salvage) for both groups. 45% pts on IV-ITRA and 48% pts on IV-VORI completed prophylaxis without modification (p=ns). Two pts on IV-ITRA developed IFI (1 disseminated C. glabrata and 1 disseminated Fusarium) as oppose to none on the IV-VORI arm (p=0.101). No significant differences were seen in the number of pts that required empiric antifungal therapy due to persistent fever or possible fungal pneumonia (14% on IV-VORI, 18% on IV-ITRA). 15/49 pts on IV-VORI (23%) versus 4/49 (8%) on IV-ITRA discontinued prophylaxis due to side effects (p=0.036). Reversible increase in the liver function tests (9 on IV-VORI, 4 on IV-ITRA) and hallucinations (5 on IV-VORI) were the most frequent adverse events. Response to induction chemotherapy, overall induction mortality and survival were similar in both groups. Conclusions: 1) IV-VORI appears to be more efficacious than IV-ITRA in preventing IFI in AML and MDS pts receiving chemotherapy. More pts are needed to confirm this finding. 2) IV-VORI tends to be more toxic than IV-ITRA. The usage of high dose IV-VORI may explain the incidence of side effects.

Description: Background: As part of our standard of care, patients (pts) with acute myelogenous leukemia (AML) and high risk myelodysplastic syndrome (HR-MDS) older than 50 years undergoing induction remission therapy (IC) are offered to be admitted to rooms in the “Protected Environment” (PE) unit during their therapy. The objective of this measure is to decrease the incidence of infections. Since Nov 2007 several changes were made to the PE rooms due to construction at our institution: Old PE unit (OPE) New PE unit (NPE) Use of shoe covers Shoe covers should be used at the entrance of the unit and should be kept on until leaving the unit. Shoe covers should be used upon entrance to the unit and changed when entering each room. Visitations Visitors are not allowed in the room. Visitors use a separate room for visitation. Visitors are allowed inside the room and must use gown, gloves and mask. Water filtration Water supply for OPE was independent from main hospital water supply. Special filtration x3 was plumbed into the unit. The ice machine was plumbed with filtered water. Water from main hospital lines. Ice machine/water from main hospital lines and plumbed with a filtration system. Window Blinds Single blind built inside the windows with an electronic switch for patient operation. Double shades inside the room operated by the patient. Both units (OPE and NPE) continue with the same HEPA filtration system with air exchanged 15 times per hour. Objective: To evaluate the impact of these changes in the incidence of infections and mortality of pts admitted to the NPE. Methods: Admission to the NPE became effective Nov 2007. We retrospectively reviewed medical records of pts who were admitted to the OPE from May 2007 to October 2007 and compared to the pts who were admitted to the NPE from Nov 2007 to May 2008. Results: 42 patients were admitted to the OPE, and 32 to the NPE. No differences were found in baseline characteristics between the 2 groups. The majority had AML [OPE=90% and NPE=84%, p=0.430] and Zubrod performance status ≥ 2 [83% and 91 %, OPE and NPE respectively, p= 0.363]. Median age was 64 yo (r: 31–89) for OPE pts and 60 (r: 18–79) for NPE. Ninety-five percent of the pts in OPE and 97% of the pts in the NPE received chemotherapy with cytarabine containing regimens; 43% and 47% of the pts had ANC

Description: Background: Acute pulmonary hemorrhage is an established complication during remission induction in patients with AML. Its pathogenesis is not well known. Injury to the alveolar capillary endothelium and release of inflammatory cytokines may play a role. The standard management is usually not effective. Aim: To analyze the course of patients with newly diagnosed AML who develop pulmonary hemorrhage during induction chemotherapy. Study Group: 1765 patient with AML received frontline therapy from 1998–2005. We analyzed 53 patients (3%) who were admitted to medical intensive care unit due to acute pulmonary hemorrhage. Results: The incidence of pulmonary hemorrhage was 3%. Outcome by patients characteristic was: parameters category Total No. Pulmonary Hemorrhage (%) P value Age (years) 〉60
 ≤60 998
 767 32
 21 3%
 3% 0.569 Diagnosis M3
 M4-M5
 others 109
 319
 1337 812
 33 7%
 4%
 2% 0.011 Platelet (×109/L) ≥20
 10
 ≤10 653
 1112 32
 21 5%
 2% 0.000 Performance status 0–2
 3–4 1662
 103 44
 9 3%
 9% 0.000 Serum CR (mg/dl) 1.5 1622
 140 46
 7 3%
 5% 0.150 HGB (g/dl) 8.0 872
 893 32
 21 4%
 2% 0.105 Infection No
 Yes 1218
 494 32
 21 3%
 3% 0.089 The predictive factors for pulmonary hemorrhage were: leukemia morphology (M3, M4–5), performance status and WBC count. The 8week survival of patients who develop pulmonary hemorrhage was 32%, (median survival days from pulmonary hemorrhage to death is 16 days). Conclusion: pulmonary hemorrhage is a serious complication of induction therapy in AML. Modalities that reduce its incidence are needed.

Description: Background: IFI remain an important cause of morbidity and mortality in pts with acute myelogenous leukemia or high risk myelodysplastic syndrome (AML/HR-MDS). We have previously shown that voriconazole (VORI) or LIPO AB 3 mg/kg/day TIW were effective prophylactic regimens in AML/HR-MDS. LIPO AB given once every week achieves tissue levels expected to prevent IFI, decreases the risk for infusion-related adverse events (IRE) and would simplify prophylaxis. Materials and Methods: We conducted a 3-arm randomized trial comparing LIPO AB, 3 mg/kg/d TIW (LIPO AB 3); versus LIPO AB 9 mg/kg/d, 1/week (LIPO AB 9); versus VORI 200 mg PO 2/day among pts with AML/MDS undergoing induction or salvage chemotherapy (CHEMO). Pts were stratified by age and disease status and randomized to receive any of the 3 regimens 24 hours after completion of CHEMO. Serum Galactomannan Index (GMI) was obtained 2/week while CT scan of chest (CT) was performed for persistent fever after 3 days of broad spectrum antibiotics. Proven and probable IFI were defined according to EORTC/MSG criteria. The results of the first 59 pts (of 150 planned) enrolled between Dec 06 -July 07 are presented. Results: Pts characteristics and response are shown in Table 1. All pts had Zubrod performance status ≤ 2 and most underwent remission induction chemo (90% in each group). No significant differences were observed on key baseline characteristics. Three of the 59 pts did not receive study drug (AMBI 9=2; VORI=1) and were excluded from efficacy and safety analysis. There were no proven IFI; 3 pts developed probable pulmonary Aspergillosis [GMI(+); CT (+), cultures (−) ], while 10 pts received additional empirical antifungal therapy (AFT) because of FUO [ 7 pts; GMI (−), CT (−), cultures (−)] or pneumonia [ 3 pts; GMI (−), cultures (−)]. None of these 10 pts developed proven/probable IFI. Two pts in each study arm developed reversible side effects that lead to drug discontinuation [AMBI 3: Grade 2 hyperbilirubinemia (1); Grade 3 infusion related events (1); AMBI 9: Grade 3 infusion related event (2); VORI: Grade 2 hyperbilirubinemia (1), visual hallucinations (1)]. Overall mortality was 5% (1 pt/arm). There were no IFI-related deaths. Conclusion: Intermittent LIPO AB (3 mg/kg/d TIW or 9 mg/kg/d, 1/week) and VORI 200 mg PO 2/day prophylaxis appear to be effective and well-tolerated regimens. Enrollment of additional pts is ongoing. Table 1 LIPO AB 3 (n=20) LIPO AB 9 (n=20) VORI (n=19) *p=ns; **p=0.061 Median age* (range) 60 (40–79) 60 (23–69) 58 (31–77) Pts in protected environment* (%) 80 65 79 Diabetes mellitus**, n(%) 1 (5) 1 (5) 5 (26) Median days on prophylaxis* (range) 17 (1–34) 14 (1–37) 17 (1–37) Efficacy and adverse events LIPO AB 3 (n=20) LIPO AB 9 (n=18) VORI (n=18) No IFI*, n(%) 14 (70) 14 (78) 15 (83) Proven/Probable IFI*, n(%) 2 (10) 1 (5) 0 Empiric AFT*, n(%) 4 (10) 3 (17) 3 (17) Adverse events*, n 4 4 3 All drug-related*, n(%) 2 (10) 2 (11) 2 (11)

Description: Abstract 993 Poster Board I-15 Background: Although it is well known that CINV is frequent in pts with hematologic malignancies, the extent of the problem has been weakly investigated as well as the therapeutic options. Study Objectives: 1) To compare if PALO given in alternate doses or in multiple-day dosing is better than ondansetron (ONDA) given daily in the prevention of CINV in pts with AML/HR-MDS receiving chemotherapy (CHEMO) with high-dose cytarabine (HDAC)-containing regimens (Ara C =1.5 or 2gm/m2 for 3-5 days); 2) To evaluate the incidence of CINV in patients with AML; 3) to identify the risk factors for CINV. Methods: Eligible pts were randomized to receive 30 minutes before CHEMO: ONDA: 8 mg IV bolus, then 24mg IV, continuous infusion on d1 up to d5 and for 12 hours after Ara C infusion ends [ONDA 1-5]; or PALO 0.25mg IV bolus over 30 seconds, on d1 up to d5 of Ara C infusion, [PALO 1-5]; or PALO 0.25mg IV bolus over 30 seconds, on day 1,3, and 5 of Ara C infusion [PALO 1,3,5]. All patients received methylprednisolone 40 mg IV before Ara C infusion daily. Complete response (CR) was defined as no emesis and no use of rescue medication (RM) during CHEMO and up to 7 days; Complete control (CC) was define as CR + pts with ≤ 1 episode of emesis within 24 hours and no more than moderate nausea with no need of RM for either case. Failure (F) was defined as the needs to use of RM. Pts were followed for 7 days. A diary to evaluate impact of CINV on daily activities was filled by the pts during the 7 days. Results: 150 pts were register in the study, 143 were evaluable. The 3 group of pts were comparable in regards to baseline characteristics. Response rates are shown in the table below. The proportion of patients without nausea on day 1 was similar in the three study arms, (p = 0.50) (Graphic 1). There was an evident decline in the proportion of patients without nausea from day 2 through day 5 in the 3 study arms. However, on study days 6 and 7, [PALO 1-5] significantly reduced the incidence of nausea (p = 0.0001 and 0.024, respectively). Similarly, less number of pts on [PALO 1-5] required RM on days 6 and 7 (p =0.034 and 0.0350, respectively). Severity of nausea, (none, mild, moderate or severe) and the impact of CINV in daily activities were reported by the pts in a diary. Significantly more pts on [PALO 1-5] reported to have none/mild nausea on days 6 and 7 of the study (p = 0.039 and p = 0.219, respectively) and more pts in the same group reported a minimal impact (not at all or little effect) of CINV in their activities on study day 6 (p= 0.0229). Univariate analysis showed that female pts, younger age, good performance status, HDAC + idarubicin and the use of prophylactic antibiotics were significant factors for developing CIN during IC. Only younger age (OR: 1.05, p =0.02); CHEMO with HDAC+ idarubicin (OR: 3.27, p=0.04) and the use of prophylactic antibiotics (OR: 2.86, p=0.009) maintained significance in the multivariate analysis. Three pts, one in each arm, developed NCI G3 adverse events, two pts had headache (one on [ONDA 1-5] and the other on [PALO 1-5]); another pt who received [PALO 1,3,5] had diarrhea. No cardiac adverse events possible or probable related to study drug were reported. Conclusions: 1) No significantly difference was detected between ONDA and PALO in the prevention of acute CINV in ALM and HR-MDS pts 2) PALO given daily for up to 5 days is significantly better than Ondansetron in the prevention of delayed CINV; 2) Young patients, receiving prophylactic antibiotics and HDAC + idarubicin are at higher risk for CIN during IC; 3) Nausea after day 1 of CHEMO in this multi-day regimen remains sub-optimally controlled with any type of 5HT3 antagonist. New strategies need to be incorporated to improve the prevention of nausea during these days. Disclosures: Mattiuzzi: Eisai, Inc.: Research Funding, Speakers Bureau. Borthakur:Eisai, Inc.: Research Funding. Kantarjian:Eisai: Research Funding.

Description: Background Patients with high risk (HR) myelodysplastic syndrome (MDS) and acute myeloid leukemia (AML) have significant toxicities such as mucositis, protracted neutropenia and severe infections when treated with standard chemotherapy. This had led to the development of ‘less intense’ chemotherapy (targeted therapy, TT). These treatments are expected to produce less toxicities, especially less immunosuppression. Antibiotic and antifungal prophylaxis are routinely given to patients undergoing intensive chemotherapy. It is not clear if the same strategy should be used for patients receiving less intensive chemotherapy. The objective of this study is to evaluate the outcome of patients receiving TT according to the use of antimicrobial prophylaxis. Methods We retrospectively reviewed the medical records of patients with AML and HR MDS that received TT as induction therapy from January 2000 to July 2007 at our institution. Baseline characteristics and antibiotic usage was recorded. All courses of TT received from start of therapy until outcome (response or failure) was assessed were evaluated, and infections or death occurring during any of these courses constituted an event. Results 225 patients received TT [decitabine or azacitidine n = 137 (61%); miscellaneous (tipifarnib, PKC412, imatinib, SAHA, and others) n=88 (39%)] for a total of 583 courses (median course per patient = 2). Median age was 72 years (range 13–89), 60% were male, 95% had Zubroad performance status ≤ 2 and 28% were neutropenic at the start of TT. None of the patients were placed in HEPA-filtered rooms (‘protected environment’) at any time. Each course of therapy was grouped into 1 of 4 groups based on the strategy use for infectious prophylaxis (table 1). Clinically documented infections and FUO were the most frequent type of infection reported in all the groups, followed by bacterial infections. Fungal infections were infrequent (total 5; group 1 = 1; group 2 = 2, group 3 = 2). There was no significant difference in the number of infectious episodes per course between the groups that received both antibacterial and antifungal prophylaxis vs. those who received no prophylaxis (p= 0.984). However, mortality was significantly higher during courses of TT administered without prophylaxis (p= 0.005). Conclusions As opposed to standard chemotherapy, fungal infections are infrequent in the patients treated with TT. Mortality is significantly higher in patients who did not receive any anti-microbial prophylaxis. The use of antibacterial and antifungal prophylaxis should be considered in patients receiving TT. Table 1: Groups based on antimicrobial strategy Strategy Strategy No. of courses No. of infectious episodes (%) No. of death (%) * p=0.984; # p=0.005 No prophylaxis 202 45 (22%) * 12 (5.94%) # Both bacterial and fungal prophylaxis 171 38 (18%) * 1 (0.58%) # Only bacterial prophylaxis 206 31 (15%) 6 (2.91%) Only fungal prophylaxis 4 0 (0%) 0 (0%)

Description: Abstract 1611 Poster Board I-637 Excess iron adversely affects the phagocytic, chemotactic and bactericidal activities of neutrophils and monocytes; inhibits the activity of natural killer cells and macrophages, and enhances the ability of microbes to grow in various host tissues. Iron overload has also been described to be an important risk factor for the development of mucormycosis. Serum ferritin is a surrogate marker of iron overload. Studies in recipients of bone marrow transplantation showed that elevated pre-transplant serum ferritin levels (as markers of iron overload) were associated with a high incidence of infection-related mortality. Objective 1) To evaluate iron overload as a predictor of infection in patients with acute myelogenous leukemia (AML) or high risk myelodysplastic syndrome (HR-MDS) undergoing chemotherapy (CHEMO); 2) to compare baseline bone marrow iron stores (BMIS) versus baseline serum ferritin as predictor of infections. Methods Patients with AML or HR-MDS who underwent induction or first salvage CHEMO were included in this prospective observational study. BMIS was analyzed at baseline (prior CHEMO) in bone marrow aspirate smears by standard staining techniques. BMIS was evaluated using a light microscope and graded on a scale of 0 to 6, with a score of 1 to 3 considered normal. Serum ferritin levels were measured at baseline. Invasive fungal infections (IFI) were defined as probable or proven according to the EORTC-MSG criteria. Bacterial infections were defined by the presence of fever and positive culture from a sterile site. Ferritin values were considered normal if they fell in the range of 10ng/mL – 291ng/mL. Results 112 patients were evaluated. Median age was 69 years (23-84 years) and 49% were female; 97% received induction CHEMO for AML. Forty–six patients developed at least one infectious episode from day 1 of chemotherapy until chemotherapy response was assessed (median days to response = 28). Nineteen patients developed bacterial infections, 9 patients developed fungal [2 candidiasis; 3 proven and 4 probable aspergillosis] and 18 patients had clinically but not microbiologically documented infections (cellulitis, pneumonia, neutropenic colitis). Baseline BMIS and serum ferritin were evaluable in 83 and 112 patients, respectively (Table 1). Median baseline ferritin was significantly higher in patients who developed any infection compare to that of patients without infections (p = 0.018). In addition, patients who developed IFI had higher median baseline ferritin to those without IFI (p=0.039). There were no significant differences in the baseline BMIS between patients without infection and those who developed any infection (p= 0.94) or an IFI (p= 0.71). Conclusions: Disclosures Mattiuzzi: Novartis: Research Funding.