Experience suggests that many physicians regard AML as a disease requiring urgent treatment, which should begin within 1–3 days of diagnosis. This practice was lent support by an ECOG trial in pts age above 55, which suggested that a delay of 3–5 days in initiating therapy resulted in a lower CR rate (38% in 235 pts vs. 50% in 113 pts, p=.03,

Blood 2004; 103: 479–485
). The delay was due to logistical factors inherent in a blinded randomization to priming with GM-CSF vs. placebo in the 235 pts; in contrast priming, which ultimately had no effect on CR rate, was not studied in the initial 113 pts. All 348 pts were randomized among 7 days of ara-C plus 3 days of idarubicin, daunorubicin, or mitoxantrone, with CR rate independent of which was given. While the paper noted that risk factors in pts in the delay and the no delay groups were similar, it is unclear whether a regression analysis of pt covariate effects was done or whether risk factors other than cytogenetics, WBC count and age (< vs > 70) were analyzed. The ECOG paper was of particular interest to us because, in the 80% of our pts presenting with WBC < 50,000, at least several days typically elapse between MD Anderson (MDA) diagnosis and start of treatment. For example, among 197 such pts given IA (idarubicin 12 mg/m2 daily days 1–3 + ara-C 1.5 gm/m2 daily x 3–4 days by continuous infusion) from 2001–2004, our most commonly used induction regimen during these years, a median of 4 days separated MDA diagnosis and treatment dates, with a wide variation: thus 25 % of pts received IA within 2 days of MDA diagnosis while another 25% only received IA after 9 or more days. Since reasons for the delay were generally unclear, we performed a regression analysis to assess whether, after accounting for other covariates, the number of days from MDA diagnosis to administration of IA affected CR rate, which was 53% in all 197 pts. We examined: Zubrod performance status < 3 vs > 2, (8% > 2), cytogenetics (26% unfavorable, as commonly defined, 2 % favorable), treatment in a HEPA-filtered or laminar air flow room (80%), and as numerical variables, age (median 64 years, minimum 19, 28% > 70), beta 2 microglobulin (b2M, median 2.9), albumin, bilirubin, creatinine, wbc (median 5.2) and days from MDA diagnosis to IA. The independent predictors of CR were age, or alternatively b2M, which was highly correlated with age, and unfavorable cytogenetics, but not days from MDA diagnosis to IA (p = 0.22, with age in the model, p = 0.29, with b2M in the model). Because the ECOG paper dealt only with pts age > 55 we also evaluated the effect on CR rate of interactions between days from MDA diagnosis to IA and either age < vs > 55 or age as a numerical variable; these interactions were not significant. Thus our results, unlike the ECOG’s, suggest that delay in initiating induction therapy is unlikely to be harmful. It is plausible that a delay may be useful if it allows knowledge about cytogenetics and FLT3 status to become available. Given that the benefit from standard induction therapy in older pts with unfavorable cytogenetics or FLT3 mutations is not necessarily commensurate with a 20–30% risk of induction mortality, it might be appropriate to give such pts investigational regimens as their initial treatment, whereas such a strategy might be more difficult to justify in pts with normal cytogenetics and no FLT3 aberrations.

Topics:
neoadjuvant therapy, brachial plexus neuritis, ms-like tyrosine kinase 3, cytarabine, idarubicin, albumins, beta 2-microglobulin, bilirubin, creatinine, daunorubicin

Author notes

Corresponding author

2005, The American Society of Hematology
2004
Sign in via your Institution