Daunorubicin for newly diagnosed AML: Does more equate to better outcomes?

In acute myeloid leukaemia (AML) patients who received the 7+3* induction, a higher daunorubicin dose or a double induction did not seem to yield better outcomes, findings from the two-part phase III SAL Dauno-Double trial suggest.

“[We] set up this trial to answer two fundamental questions in standard induction,” said Dr Christoph Röllig from Universitätsklinikum Carl Gustav Carus, Dresden, Germany, at ASH 2022. “[We sought to determine whether] daunorubicin 90 mg is superior to 60 mg in 7+3 induction [and] whether a double induction is necessary in patients with good** early response after the first induction.”

The study enrolled 864 participants with newly diagnosed AML (median age 52 years) who were fit for intensive induction. They received the 7+3 regimen and were randomized 1:1 according to daunorubicin dose – 60 or 90 mg/m². [ASH 2022, abstract 217]

 

Part 1: High vs low dose

In the interim analysis, the percentages of good early responders were similar between the 60- and 90-mg arms (43 percent vs 48 percent; p=0.29). “The between-arm difference was <5 percent, which was neither statistically significant nor considered clinically relevant,” said Röllig. This finding led to the closure of the 90-mg arm. All participants subsequently received daunorubicin 60 mg.

At end of enrolment, the proportions of good early responders remained similar between the 60- and 90-mg arms (44 percent vs 48 percent), again yielding a nonsignificant difference (p=0.93).

“None of the subgroups had a significant benefit from either dose. In the FLT3-high arm, there was a small trend for benefit in the 90-mg arm, but that was not significant either,” added Röllig.

Complete remission (CR) rates were also similar following standard induction (75 percent vs 82 percent; p=0.16). There were also no significant differences between the 60- and 90-mg doses, be it among good responders (91 percent vs 99 percent; p=0.08) or moderate/nonresponders (62 percent vs 66 percent; p=0.57).

After a median follow-up of 44 months, relapse-free survival (RFS) was similar between the two daunorubicin arms (hazard ratio [HR], 1.08; p=0.56), as was overall survival (OS; HR, 1.19; p=0.19).

The 90-mg arm had slightly higher rates of grade ≥3 adverse events (AEs; 20 percent vs 18 percent) and early mortality (5 percent vs 2 percent) compared with the 60-mg arm.

 

Part 2: Is two better than one?

A total of 376 participants had a good response following the first induction. They were re-randomized to receive no further induction or a second one using daunorubicin 45 or 60 mg.

In the per-protocol (PP) population, the difference in CR rates was 4.7 percent in favour of double induction. However, this did not achieve significance nor was it low enough to formally show noninferiority. The difference was even smaller in the intention-to-treat cohort (ITT; 0.4 percent).

Three-year RFS was slightly better with double vs single induction in the ITT population (60 percent vs 51 percent; HR, 1.35; p=0.074), more so in the PP cohort (60 percent vs 52 percent; HR, 1.43; p=0.05). Although the PP findings seem borderline significant on univariate analysis, this trend disappeared on multivariate analysis after adjusting for potential differences (HR, 1.39; p=0.10), noted Röllig.

OS at 3 years was identical, both in the ITT (77 percent [single] vs 75 percent [double]; HR, 1.02; p=0.91) and PP cohorts (77 percent vs 76 percent; HR, 1.12; p=0.63).

More AEs were seen in the double vs the single arm (median per patient 7 vs 5), as these have accumulated throughout two cycles of induction. “The incidence of serious AEs is slightly higher in the double arm, but this again was far from statistical significance,” said Röllig.

 

Higher, or more, does not mean better

Taken together, in the context of the 7+3 induction, the findings showed that daunorubicin 90 mg did not lead to higher remission rates or longer survival than the 60-mg dose, noted Röllig. Among good responders, a second induction only had a limited impact on RFS, with not much benefit in terms of CR rates and OS.