Impact of Initial Body Mass Index (BMI) on Survival Outcome of Patients with Acute Leukemia (AL): Myth or Reality?

Obesity is associated with an increased risk of mortality from cardiovascular causes and from malignancies while meta-analyses have evidenced a significant association between an elevated body mass index (BMI) and the risk to develop certain hematological malignancies including acute myeloid (AML) and acute lymphoid leukemia (ALL). Adipocytes are present in the bone marrow were it promote leukemic cells survival and resistance to treatment through the production of amino acids, free fatty acids, pro-inflammatory adipokines and cytokines. Obesity modifies the pharmacokinetic of numerous drugs while obese patients may be undertreated due to dosage capping based on ideal body weight. In addition, obesity increases the risk of treatment-related complication. Obesity adversely affects outcome in pediatric ALL (Butturini et al, J Clin Oncol 2007) and AML (Inaba et al, Cancer 2012). In contrast two retrospective studies based on the analysis of 63 (Lin et al, Leuk Lymphoma 2012) and 329 cases (Lee et al, Ann Hematol 2012) did not found any significant association between obesity and disease outcome in adult patients treated for AML. Here we have investigated whether the body mass index, determined at the time of diagnosis could impact disease outcome in adult patients treated for acute leukemia (AL).

A total of 531 consecutive AL patients entered into clinical trials in our Institution between 1994 and 2012 were analyzed retrospectively. They included 343 AML, 146 ALL (102 B-lineage ALL, 41 T-lineage ALL and 3 undifferentiated cases) and 42 acute promyelocytic (APL) leukemias. Based on data collected at the time of diagnosis, BMI was calculated as weight (kg) divided by the square of height (m). Patients were stratified according to the World Health Organization (WHO) BMI classification: underweight (BMI < 18.5), normal weight (BMI 18.5 to <25), overweight (BMI 25 to <30) and obese (BMI ≥ 30). Baseline clinical and biological features and clinical outcome were compared according to this stratification.

Overall, median BMI was 25.7 (16.5 - 46). Two hundred and thirty-three (54.1%) patients were under/normal weight, whereas 197 (45.9%) patients were overweight/obese. An increased BMI (>25) was associated with age > 60 (p<0.001), APL diagnosis (p=0.02), and circulating blasts percentage > 33% (p=0.049). The CR rate after 1^stinduction course in underweight, normal, overweight and obese were 78.9%, 79.5%, 80.3%, and 90.1%, respectively. These differences were not statistically significant. The same results were observed when the comparisons were restricted to ALL, AML, in younger or older patients. With a 4.7 years median follow-up (95%CI: 4.0-5.1), median disease-free survival (DFS) and median overall survival (OS) were 13.3 and 22.1 months in AML and 20.8 and 36.5 months in ALL respectively. Medians were not reached in APL. Median OS and DFS (with 95% CI) were 23.1 (13.8-31.6 ) and 29.5 (18-34.1), 17.6 (13.7-22.8) and 32.1 (25.1-36.9), 17.9 (12.1-24.1) and 26.5 (18.6-41.9), and 20.3 (13.4-27.5) and 23.5 (20.3-32.4) months in underweight, normal, overweight and obese, respectively. These differences were not significant. There was no significant impact of BMI on complete response rate, DFS and OS in patients from the whole cohort when considering a cut-off for BMI at 25, and when analyzing according to age, AL subtypes, or to cytogenetics. Only BMI in patients with T- cell lineage ALL showed a significant impact on survival. In patients with BMI >25, median DFS was not reached with a 3-year DFS at 76%, while median DFS was 16.1 months with 3-year DFS at 13% for those with BMI <= 25 (p=0.005). Similarly, median OS was not reached in T-cell ALL patients with BMI > 25 versus 28.3 months in those with BMI <= 25 (3-year OS: 78% versus 41%, p= 0.04). In a model including factors of prognostic value in univariate analysis, multivariate analyses confirmed the prognostic value of BMI in T cell lineage-ALL when comparing overweight/obese patients with under/normal weight patients, but only in terms of DFS (HR=0.25; 95% CI (0.05 – 0.87); p=0.037) and not of OS.

These findings provide further evidence that initial body size may have a potential prognostic impact in some subset of leukemia, and more specifically in T cell lineage-ALL. In this AL subtype, validation in larger prospective studies is however warranted.