Effect of Early Non-Hematologic Treatment Toxicity on Complete Remission (CR) and Time to CR in Acute Myeloid Leukemia Patients Undergoing Remission Induction Chemotherapy

Background: Acute Myeloid Leukemia (AML) patients (pts) being screened for clinical trials are often excluded based on laboratory values outside the normal range because of concerns of early treatment-related toxicities that can lead to adverse outcomes including death or delayed time to complete remission (CR). It has been shown that the time to achieve CR (Tc) has greater prognostic value in predicting survival than achieving CR per se (Estey et al., Blood 2002). In this study we assessed correlations between pretreatment risk factors and early (within 30 days of induction) treatment-related non-hematologic toxicities (TNHT) following IC, and between these toxicities and Tc and the probability of achieving CR.

Methods: Adult pts diagnosed with AML (excluding acute promyelocytic leukemia) at the Cleveland Clinic from 10/08 - 11/12 who underwent IC with 7+3 (cytarabine and anthracycline) and had complete toxicity data were included. Variables including demographics, AML disease characteristics, abnormal laboratory measurements at diagnosis and during treatment, treatment response, and ICU stay, were assessed. The outcome of interest were development of non-fatal early TNHT and its effect on CR and Tc. TNHT were mapped as linear distance beyond the normal laboratory range and were set to zero if within limits. This was done for serum sodium (Na), potassium (K), bicarbonate, liver enzymes (AST, ALT), total bilirubin, INR (international normalized ratio), serum creatinine and albumin. These metrics were analyzed as a function of covariates at diagnosis using standard linear regression. They were then treated as covariates in logistic regression models of CR and in linear models of Tc. Akaike’s Information Criterion (AIC) was used in stepwise logistic regression model selection. Due to multiple testing, parameters are reported as significant only if p<0.01.

Results: Of 91 pts, 58% (n=53) were female, 80% (n=73) were Caucasian, 64% (n=58) were younger than 60 years, the median age at treatment was 53.4 years (range 22.2 to 77.2), 13.2% (n=12) received 2 cycles of induction therapy and 30% of pts (n=27) required ICU care. Disease characteristics (per WHO classification) were – AML with recurrent cytogenetic abnormalities 31% (n=28), secondary AML 27% (n=24), therapy-related AML 6% (n=5), AML not otherwise specified 34% (n=31), myeloid sarcoma 1% (n=1) and unknown 1% (n=1). Cytogenetic risk groups per CALGB 8461 were – favorable 15% (n=14), intermediate 54% (n=49), unfavorable 23% (n=21), monosomal karyotype 7% (n=12) and unknown cytogenetics (n=7). BMI groups at diagnosis included normal (18.5-24.9, 28%), overweight (25-29.9, 29%), moderately obese (30-39.9, 29%), severely obese (>40, 12%) and one was underweight. All patients developed TNHT following IC: 98% (n=89) had electrolyte abnormalities, 26% (n=24) had elevated serum creatinine levels and 99% (n=89) had abnormal liver function indicated by AST, ALT, bilirubin, albumin and INR. Overall, 68% pts achieved CR, 9% (n=8) had complete response with incomplete recovery of counts, 10% (n=9) had persistent disease and in 12, determination of CR was not done due to early death or severe debilitation. Of all pretreatment variables, advancing age correlated with worsening hypoalbuminemia (p=0.0076); higher WBC with worsening hyperkalemia (p=0.001); absolute neutrophil counts with high K (p=.0002); and LDH with high K levels (p<< 0.0001) and elevations in ALT (p=0.0004) and AST (p=0.0002). Among early TNHT, worsening hyperkalemia (p=0.004) and acidemia (low bicarbonate levels) (p=0.008) correlated with less likelihood of achieving CR. In a multivariate analysis using stepwise logistic regression and AICs, none of the covariates in the final model for CR retained significant P values. None of the covariates independently predicted for Tc in the best fitted stepwise regression linear model.

Conclusion: In pts considered fit to undergo 7+3 IC with potential curative intent, none of the baseline pretreatment variables independently predicted for early TNHT and occurrence of TNHT did not affect the probability of CR or Tc. Our data suggest that decisions regarding clinical trial eligibility or intensive therapy should be based on predicted risk of early treatment-related death rather than arbitrary age cut-offs and or abnormal laboratory values.