Abstract
Background: A wide landscape of factors exists to stratify the prognosis of patients with newly diagnosed AML including age, race, cytogenetics, and functional status. However, weight loss and malnutrition have yet to be fully evaluated as possible prognostic indicators for the patient with AML.
Aim: This study aims to 1) investigate how weight loss and nutritional markers (BMI and albumin) change over the course of treatment and 2) their prognostic significance in terms of overall survival (OS) and complete remission (CR) in patients with newly diagnosed AML.
Methods: We retrospectively analyzed patients from January 2000 to June 2012 diagnosed with AML over 18 years of age, who were treated at OUHSC with induction chemotherapy. 215 AML patients were identified. BMI and albumin levels were examined using mean plots over 3 time points. Patients treated before 2005 were eventually excluded due to the poor documentation of the height and weight. A total of 121 patients had BMI measurements for the first 2 time periods, while 89 had BMI measures at all 3 time periods. Simple descriptive statistics were created for all covariates [mean and SD for continuous covariates; n (%) for categorical variables]. For aim 1, to evaluate changes in BMI and albumin over time, mixed models were used including age (<60 years and ≥ 60 year) and gender.To examine aim 2, a Cox proportional hazards model was used to assess the univariate association of each covariate with survival. A final model was then created using all covariates from the univariate models with p-values <0.25. All two-way interactions with BMI group were explored. Binary logistic regression models were created to examine the association of each covariate with CR. BMI was categorized into two groups according to increase or decrease in BMI over time period.
Results: Among the 121 patients mean age was 53 years (range 16 to 89 years). BMI (mean 29.85, 28.75, and 27.9 across the three time points) and albumin (mean 3.30, 3.21, and 2.55 across the three time points) both decrease over time and these changes were not different by age or gender. For OS, a significant interaction between cytogenetic risk group and BMI group was identified; therefore, patients were stratified by these risk groups. Age (p=0.0012) was found to be a more significant prognostic indicator than BMI group in the unfavorable risk group. Among patients in the intermediate cytogenetic risk group, BMI group (p=0.0210) was significantly associated with OS; those with increased BMI have 2.38 times the hazard of death relative to those with a BMI decrease, adjusting for age group. BMI group is not significantly associated with OS in any other model. Albumin was not associated with OS in any model.
Conclusion: Age was found to be a more significant prognostic indicator than BMI in the unfavorable risk group, and BMI increase was found to have an increased hazard of death in the intermediate risk group. Further studies are needed to examine the prognostic significance of weight loss during induction chemotherapy.
No relevant conflicts of interest to declare.
Author notes
Asterisk with author names denotes non-ASH members.