Clinical Lymphoma, Myeloma & Leukemia, Vol.20, Suppl.1 - September 2020

Abstracts Clinical Lymphoma, Myeloma & Leukemia September 2020 S200 Reinhold Munker 3 , Gerhard Hildebrandt 3 , Maxwell Krem 3 * 1 University of Louisville School of Medicine, Louisville, KY, USA, 2 Department of Public Health and Markey Cancer Center, University of Kentucky College of Medicine, Lexington, KY, USA, 3 Division of Hematology and Blood & Marrow Transplant, Markey Cancer Center, University of Kentucky College of Medicine, Lexington, KY, USA Context: Infections cause substantial morbidity and mortality in acute leukemia patients. Neutropenic fever (NF) frequently complicates leukemia chemotherapy, but the mortality rate of NF and predictors of mortality are not fully defined for adult leukemia patients. We examined factors that influence mortality and type of infection in NF patients admitted to the hematologic malignancy service of a referral hospital. Design: Retrospective chart review of NF patients admitted to University of Louisville from April 2016 to July 2018 who had AML, ALL, CML, CLL, or MDS, excluding active transplant patients. We used logistic regression and Cox proportional hazard analysis to evaluate mortality predictors and the Kaplan-Meier method for survival. Median follow-up was 16.6 months. Results: Sixty-two patients had 101 episodes of NF. Median age was 56. Frequent primary diagnoses were AML (70%), ALL (24%), CML (2%), and CLL (2%). Infections and bacteremia occurred in 72% and 54% of episodes, respectively. Mean length of stay was 19 +/- 12 days. Refractory/relapsed (R/R) disease (RR 1.53, p=0.034) and ANC nadir of 0 (RR 2.64, p=0.001) correlated with bacteremia. Thirty-day mortality was 14% and increased for R/R disease (OR 3.13, CI=0.99–9.92), bacteremia (OR 3.58, CI=0.94- 13.7), and ICU-level care (OR 15.24, CI=4.11-56.5) in univariate analysis. ICU-level care (OR 29.1, CI=5.42–156) and R/R disease (OR 7.30, CI=1.43–37.4) retained significance in multivariate logistic regression. Gram +/- status, resistance pattern, neutropenia severity, age, and neutropenia duration did not influence short-term mortality. Eighty-six episodes had a high-risk MASCC score (<21); none of the 15 low-risk episodes resulted in ICU-level care or early mortality; p = 0.036. In MVA age > 60 (HR 2.16, CI=1.22–3.84), R/R disease (HR 5.58, CI=2.74–11.4) and ICU-level care (HR 4.96, CI=2.14–11.5) predicted long-term mortality. High-risk MASCC approached significance for OS; median OS 7.8 months (high-risk) versus not reached (low-risk, p = 0.076). Conclusions: Bacteremia occurred in the majority of NF episodes we recorded. Thirty-day NF mortality was 14%. Mortality increases for patients with age > 60, R/R disease, and ICU-level care. Identification of high-risk factors in NF may guide the study of interventions such as myeloid growth factor support in appropriate populations. Keywords: leukemia, AML, ALL, neutropenic fever, bacteremia, infection, mortality AML-303 Sys t ema t ic Review o f Frail t y Assessmen t T oo ls f o r E lderly Pa t ien t s wi t h Acu t e Myel o id Leukemia Dina Khalaf 1 *, Tobias Berg 1 , Chris Hillis 1 , Tom Kouroukis 1 , Brian Leber 1 , Justin Lee 2 1 Division of Malignant Hematology, Department of Oncology, McMaster University, Hamilton, Ontario, Canada Hamilton Health Sciences, Hamilton, Ontario, Canada, 2 Division of Geriatric Medicine, Department of Medicine, McMaster University, Hamilton, Ontario, Canada Hamilton Health Sciences, Hamilton, Ontario, Canada Context: Treating elderly AML patients (>65 years) with intensive remission-induction chemotherapy without proper frailty assessment results in poor survival. We hypothesize a lack of health status measurement tools specific for elderly AML patients that inform risk stratification and treatment decisions. Objective: To determine the availability of frailty assessment tools for elderly AML patients. Methods: We searched MEDLINE, EMBASE, PsycINFO and Cochrane CENTRAL databases from inception through March 5, 2020, for case-control, cohort, randomized controlled and validation studies. We used MeSH and free text terms for acute myeloid leukemia, frailty and the elderly. We included all languages and conference abstracts covered in the databases. Two reviewers independently screened references against inclusion/exclusion criteria and resolved differences through discussion. Results: Out of 4115 retrieved references, 29 studies (3,526 patients) met eligibility criteria.We did not find any frailty assessment tools specific for elderly AML patients. Our search revealed 2 main categories of functional status measurement tools used in elderly AML patients. The first is disease-specific and used for functional status evaluation of cancer patients of any age. They included the Karnofsky performance scale (5 studies, 1,136 patients) and the Eastern Cooperative Oncology Group (ECOG) performance scale (6 studies, 458 patients). The second is age-specific and used for functional status assessment of elderly patients. These were Comprehensive Geriatric Assessment (CGA) tools (8 studies, 870 patients) including the Clinical Frailty Scale, instrumental activities of daily living functional score, Grip strength, and Timed Up and Go Test (TUGT). Both Karnofsky, ECOG performance scales, and most of the CGA tools were not significantly predictive of poor overall survival, except for a TUGT of more than 10 seconds [n = 96, HR 3.4 (95% CI 1.4 – 8.0), P = 0.005] and an abnormal CRP level [n = 96, HR 4.30 (95% CI 1.60 - 11.50), P = 0.003]. Conclusions: Our systematic review demonstrates a lack of specific frailty assessment tools for elderly AML patients. There is a need to develop a novel, practical, time and cost-efficient health status measurement tool that will assess frailty in elderly AML patients to support hematologists in risk stratification and decision making. Keywords: acute myeloid leukemia, AML, frailty, elderly, systematic review