Do Early Events Excluding Patients with Acute Promyelocytic Leukemia (APL) From Trial Enrollment Modify Treatment Result Evaluation? Real-Life Management of 100 Patients Referred to the University Hospital Saint-Louis Between 2000 and 2010.

Since combining differentiating agents and chemotherapy, acute promyelocytic leukemia (APL) is associated with a high cure rate. One remaining issue is the significant rates of early death and relapse observed in patients with high count APL (initial white blood cell count [WBC] ≥ 10.10⁹/L). Early death rate might be underestimated in clinical trials, due to an unknown proportion of patients not registered because of initial severity. For this reason, we reviewed individual histories of all patients with APL referred to our institution during the last 10 years (09/2000-06/2010), with a special focus on admission in intensive care unit (ICU) and inclusion or non-inclusion in recruiting APL trials (European group APL-2000 and APL-2006), as well as long-term follow-up.

A total of 100 patients with newly-diagnosed previously untreated APL, including 8 children, were admitted during this time period. Diagnosis was based on morphology and subsequently confirmed by the presence of the t(15;17) translocation and/or PML/RARA fusion transcript.

The rate of patients not enrolled within recruiting trials was 29% (n= 29). This rate was higher in children (n= 5/8, 62.5%) than in adults (n= 24/92, 26%) and remained stable during the two protocol periods (n=17/62, 27% for APL-2000; n= 12/38, 32% for APL-2006). Reasons for non-enrollment were inability to give informed consent in 10 patients (mechanical ventilation or neurological deficiency), physician's decision in 5 patients (2 very high leucocytosis, 1 severe infection, 2 severe liver dysfunction), and concomitant disease in 4 patients (2 HIV patients, 2 other cancers), refusal in 5 patients (including 1 Jehovah witness who eventually survived), and various administrative reasons in 5 patients. Non-enrolled patients had similar sex ratio (F/M=15/14 vs 35/36; p=.99), median age (40.5 [range, 4–79] vs 46 years [4-81]; p=.97), and frequency of additional chromosomal abnormalities (24% vs 28%; p=.80) than enrolled patients. Conversely, they had a higher rate of WBC ≥ 10.10⁹/L (n=15/29 vs 22/71; p=.07) or ≥ 50.10⁹/L (n=8/29 vs 5/71; p=.01), a lower rate of platelet count < 40.10⁹/L (28/29 vs 46/71; p=.001), and a higher frequency of microgranular M3-variant subtype (11/29 vs 8/71; p=.004) and BCR3 PML-RARA isoform (14/25 vs 24/70; p=.09). Among the 29 non-enrolled patients, 24 nevertheless received the whole planned standard induction therapy, 2 received arsenic trioxide-based induction, and 3 early died before or the day after chemotherapy initiation. Ninety-nine patients were evaluable for response to induction (1 patient ongoing). Due to a higher early death rate (21% vs 3%; p=.007), the complete remission (CR) rate was lower in non-enrolled patients (79% vs 97%; p=.007). At 5 years, event-free survival (EFS) was estimated at 62% (95%CI, 37–79) vs 84% (95%CI, 72–91) (p=.02) and overall survival (OS) at 63% (95%CI, 36–81) vs 85% (95%CI, 72–93) (p=.03) in the non-enrolled and enrolled group, respectively. Once CR had been reached, non-enrolled patients displayed, however, a good post-CR outcome with 5-year remission duration at 78% and OS from CR at 80%. Of note, only one patient from this cohort died in first CR from a second neoplasia. Twenty-six patients (26%) were admitted in ICU for or during induction (13, 8, and 4 of them requiring mechanical ventilation, amine therapy, and dialysis, respectively). Chemotherapy was initiated in ICU in 19 of them. The rate of trial enrollment was 54% (n= 14/26) in ICU patients compared to 77 % (n= 57/74) in non-ICU patients (p=.04). Again, CR rate (p<.001), EFS (p=.004), and OS (p=.002) were significantly lower in ICU patients, but remission duration and OS from CR were very satisfactory in these patients despite their admission in ICU for or during induction (91% and 90% at 5 years, respectively).

Even if this study only reports the experience of a large single center with potential patient selection, the observation that initial APL severity and/or need for ICU are associated with a lower trial enrollment rate suggests that early mortality might be underestimated in multicenter APL trials. Interestingly, patients who survive after early intensive care, including mechanical ventilation, may nevertheless receive an optimal induction and post-remission therapy and display the expected good outcome associated with APL.

No relevant conflicts of interest to declare.