Standard dose anthracycline plus all-trans retinoic acid and arsenic trioxide as induction chemotherapy significantly reduces early death and relapse for high-risk acute promyelocytic leukemia: a single-center real-world analysis

Abstract

Background:

All-trans retinoic acid (ATRA) and arsenic trioxide (ATO) have revolutionized the treatment of acute promyelocytic leukemia (APL). However, the management of high-risk APL has not been conclusively established. The optimal dosage of anthracycline in the induction has long been debated when ATO is added.

Objectives:

To explore the management of high-risk APL regarding the optimal dosage of anthracycline in the induction and the predicators of prognosis.

Design:

This was a retrospective study in the real-world setting.

Methods:

High-risk APL patients defined as white blood cell (WBC) greater than 10 × 10⁹/L who received ATO-based induction regimens were included. Data on clinical characteristics, treatment regimens, and prognosis including early death (ED) and overall survival (OS) were collected from medical records. Risk factors of ED and OS were analyzed.

Results:

This research included a total of 130 participants. Fifty (38.5%) patients received ATO+ATRA dual induction plus standard-dose anthracycline (ATO + ATRA + stDNR). Fifty-nine (45.4%) patients received ATO + ATRA with consecutive low-dose anthracycline (ATO + ATRA + ldDNR). Twenty-one (16.2%) patients were treated with ATO and various chemotherapies (ATO + others). Compared with the other two groups, the ATO + ATRA + stDNR group had the lowest ED rate of 4.0% (10.2% and 52.4%, respectively; p < 0.001). Multivariate analysis revealed that age ⩾60 years (odds ratio (OR) = 8.888, 95% confidence interval (CI): 1.126–70.129), prothrombin time (PT) ⩾18 s (OR = 4.749, 95% CI: 1.252–18.007) and WBC ⩾100 × 10⁹/L (OR = 10.591, 95% CI: 1.995–56.232) were independent risk factors for ED. The 5-year OS rates of the three induction groups were 96%, 80%, and 31%, respectively. None of the 48 patients who underwent ATO + ATRA + stDNR induction relapsed, whereas 9.4% (5/53) patients in ATO + ATRA + ldDNR group relapsed, and the relapse rate was 30.0% (3/10) in ATO + others group (p = 0.003). The survival advantage of ATO + ATRA + stDNR was demonstrated by a Cox regression (hazard ratio (HR) = 5.079, 95% CI: 1.071–24.079). WBC ⩾100 × 10⁹/L was correlated with an inferior OS (HR = 3.402, 95% CI: 1.359–8.518).

Conclusion:

Compared with low-dose anthracycline, standard-dose anthracycline combined with ATO and ATRA dual induction resulted in excellent outcome for high-risk APL patients.

Plain language summary

Standard dose anthracycline plus all-trans retinoic acid and arsenic trioxide as induction chemotherapy brought in excellent outcome for high-risk APL patients

All-trans retinoic acid (ATRA) and arsenic trioxide (ATO) have revolutionized the treatment of acute promyelocytic leukaemia (APL). However, optimal management of high-risk APL patients have not been conclusively established especially for the role of cytoreductive anthracycline chemotherapy in the dual induction era. This study reviewed 130 high-risk APL patients treated with an ATO-based induction regimen in a single center in China. Rather than reducing anthracyclien doses and giving it in a prolonged course, using standard dose anthracycline in the dual induction brought in excellent outcome for high-risk APL patients which had lower early mortality and relapse rate. Thus, standard dose anthracycline is still indisposable for high-risk APL even though the voice of chemofree induction regimen with ATRA and ATO is getting louder in many centers.

Introduction

Acute promyelocytic leukemia (APL) is a distinct acute myeloid leukemia subtype characterized by balanced reciprocal translocation t(15;17) involving the promyelocytic leukemia (PML) gene on chromosome 15 and the retinoic acid receptor alpha (RARA) gene on chromosome 17. ¹ All-trans retinoic acid (ATRA) and arsenic trioxide (ATO) have revolutionized its treatment, resulting in cures in more than 90% of APL patients. ² For non-high-risk APL patients, ATRA plus ATO or ATRA plus anthracycline chemotherapy (in countries where ATO has not yet been approved or has a cost consideration) is universally recommended as front-line induction therapy. ³ Nevertheless, high-risk APL, defined as a white blood cell (WBC) count greater than 10 × 10⁹/L according to the Sanz risk model ⁴ remains a challenge due to early death (ED).

Since most phase III trials included low proportions of high-risk APL patients, the optimal management of high-risk APL patients has not been conclusively established. More recently, the first results of the phase III, open-label, randomized APOLLO trial presented at the European Hematology Association 2024 Congress revealed that ATRA+ATO regimen with two initial doses of idarubicin on days 1 and 3 as induction significantly improved event-free survival (EFS) compared with standard ATRA and anthracycline-based chemotherapy in patients with newly diagnosed high-risk APL. ⁵ Since ATO was first introduced by China’s Shanghai group, ⁶ ATRA plus ATO combined with cytoreductive chemotherapy such as anthracyclines (idarubicin or danuorubicin) has become a routine regimen for high-risk APL patients in China. However, the optimal dose of anthracycline has long been debated. In this study, we analyzed high-risk APL patients treated with an ATO-based induction regimen at West China Hospital, Sichuan University (WCHSCU), to explore the optimal method of using cytoreductive anthracycline and the prognostic factors for high-risk APL patients.

Methods

Patients

This was a retrospective study from a tertiary hospital in China. The inclusion criteria were as follows: (1) newly diagnosed high-risk APL patients hospitalized at WCHSCU from January 2009 to December 2023 were included; (2) the diagnosis of APL was based on cytomorphology and confirmed by the presence of PML::RARA fusion gene by molecular assay in accordance with the Chinese guidelines for the diagnosis and treatment of APL ⁷ ; (3) high-risk APL was defined as WBC count over 10 × 10⁹/L upon diagnosis; (4) all patients received ATO-based induction regimens. Patients aged less than 14 years were excluded. Variant APL cases involving the fusion of RARA gene with other non-PML partners were also excluded. Clinical data including demographic data, clinical manifestations, laboratory parameters, imaging, and outcome information were collected from the electronic medical records.

Treatment

All patients received ATO-based induction therapy at the discretion of the attending physicians and were categorized into three treatment groups: (1) ATO + ATRA + standard-dose anthracycline (stDNR); (2) ATO + ATRA + low-dose anthracycline (ldDNR); and (3) ATO + others. In all groups, ATO (0.16 mg/kg/day) was administered intravenously until complete remission (CR) was achieved. ATRA was given orally at 25 mg/m²/day. For the ATO + ATRA + stDNR group, the patients received stDNR (idarubucin 8 mg/m² or daunorubicin 45 mg/m² for three consecutive days beginning on the second day) plus ATO and ATRA dual induction. In the ATO + ATRA + ldDNR group, the patients were given low-dose daunorubicin at 20 mg/day continuously until the WBC count decreased to less than 10 × 10⁹/L. In group 3, patients received ATO plus cytoreductive chemotherapy involving daunorubicin, cytarabine, or hydroxyurea at various dosages or combinations without ATRA or with ATRA added during the myelosuppressive period afterward. Consolidation and maintenance were conducted following the 2014 and 2018 Chinese APL guidelines.^7,8 For consolidation, three cycles of the DA (daunorubicin 45 mg/m² on days 1–3 and cytarabine 100 mg/m² day 1–5) or IDA (idarubicin 8 mg/m² on days 1–3 and cytarabine 100 mg/m² on days 1–5) regimens were used. Afterward, two cycles of ATRA and one cycle of oral Realgar-Indigo naturalis formula (RIF) or intravenous ATO were considered as one round of maintenance, and the patients were supposed to receive a total of eight rounds of maintenance therapy for a 2-year course. The choice of ATO or RIF in the maintenance period was determined mainly by the patient’s willingness.

Statistical analysis

All the statistical analyses were performed with SPSS version 29 (IBM Corp., Armonk, NY, USA). ED was defined as death from any cause within 30 days of diagnosis. Overall survival (OS) was defined as the time from diagnosis to the date of death from any cause or the last follow-up. Nonnormal continuous variables are expressed as the median (range). Categorical variables are presented as frequencies and percentages. Comparisons between categorical variables were performed via the chi-square test or Fisher’s exact test, and the Kruskal–Wallis H test was used for continuous variables. A binary logistic regression model (forward: LR method) was used to identify the risk factors for ED. OS was analyzed via the Kaplan–Meier method with the log-rank test. A time-dependent Cox proportional hazard model (forward: LR method) was used for multivariate analysis. Variables with p value <0.2 in the univariate analysis were included in the multivariate analysis. A two-tailed p < 0.05 was considered significant.

Results

Baseline characteristics

A total of 130 patients were included, and 71 (54.6%) patients were male. The median age was 38 years (range, 14–71 years). Fifty (38.5%) patients received ATO + ATRA dual induction plus standard-dose anthracycline (ATO + ATRA + stDNR). Fifty-nine (45.4%) patients received dual induction therapy with low-dose anthracycline chemotherapy (ATO + ATRA + ldDNR). Twenty-one (16.2%) patients were treated with ATO and chemotherapy (ATO + others). The clinical characteristics of the patients in the three treatment groups were described in Table 1. The baseline characteristics were not significantly different among the three treatment groups.

Table 1.

Baseline characteristics of the total patients and the three induction regimen groups.

Characteristics	Total patients (n = 130)	Induction regimen			p-Value
		ATO + ATRA + stDNR (n = 50)	ATO + ATRA + ldDNR (n = 59)	ATO + others (n = 21)
Male gender, n (%)	71 (54.6)	31 (62.0)	31 (52.5)	9 (42.9)	0.305
Age (years), median(range)	38 (14–71)	35 (16–68)	43 (14–71)	33 (18–64)	0.145
Hemoglobin (g/L), median(range)	77 (34–167)	81 (34–142)	78 (39–167)	72 (45–145)	0.093
Platelet (×109/L), median(range)	25 (2–146)	26 (5–146)	23 (2–93)	26 (6–64)	0.731
WBC (×109/L), median(range)	36.66 (10.16–274.52)	26.44 (11.06–261.96)	34.00 (10.16–218.92)	35.28 (10.35–274.52)	0.412
PT (s), median (range)	15.4 (9.8–120)	14.8 (11.2–120.0)	15.5 (9.8–120.0)	18.2 (9.9–120)	0.098
APTT (s), median (range)	18.8 (13.8–32.3)	18.5 (14.8–37.3)	19.0 (14.8–29.4)	18.2 (13.80–34.6)	0.454
Fib (g/L), median (range)	1.15(0.20–4.84)	1.06 (0.30–4.61)	1.22 (0.20–4.81)	1.08 (0.50–2.48)	0.873
D-dimer (mg/L), median (range)	13.85 (0.50–38.00)	13.69 (0.50–38.00)	14.45 (0.50–38.00)	9.45 (0.97–38.00)	0.255
Mucocutaneous bleeding, n (%)	91 (70.0)	37 (74.0)	39 (66.1)	15 (71.4)	0.661
Pulmonary infiltration on computed tomography, n (%)	51 (39.2)	14 (28.0)	26 (44.1)	11 (52.4)	0.093

ATRA, all-trans retinoic acid; ATO + ATRA + ldDNR, ATO + ATRA with consecutive low-dose anthracycline; ATO + ATRA + stDNR, ATO + ATRA dual induction plus standard-dose anthracycline; WBC, white blood cell; PT, prothrombin time; APTT, activated partial thromboplastin time; Fib, fibrinogen.

ED and risk factors

A total of 19 ED cases were documented (Table S1), resulting in an ED rate of 14.6%. The causes of ED were intracranial hemorrhage (10/19), pulmonary hemorrhage (5/19), renal failure (2/19), gastrointestinal bleeding (1/19), and sepsis (1/19). For ED patients, the median survival after diagnosis was 7 days (range, 1–23 days). The ATO + others group had the highest ED rate, followed by the ATO + ATRA + ldDNR group and the ATO + ATRA + stDNR group (52.4% vs 10.2% vs 4.0%, p < 0.001). We stratified patients into three WBC range groups based on initial WBC count: WBC < 50 × 10⁹/L (WBC ~50), 50 × 10⁹/L ⩽WBC < 100 × 10⁹/L (WBC 50–100) and WBC ⩾100 × 10⁹/L (WBC ~100). The ED rate significantly increased with increasing WBC count (6.3% vs 13.8% vs 47.6%, p < 0.001). Moreover, patients with PT ⩾18 s had a higher incidence of ED (37.1% vs 6.3%, p < 0.001). Compared with high-risk APL patients diagnosed before 2017, the ED rate was not significantly lower after 2017 (12.5% vs 18.5%, p = 0.388). The ED rates in the different risk groups are summarized in Table 2. In the multivariate analysis, age ⩾60 years (odds ratio (OR) = 8.888, 95% confidence interval (CI): 1.126–70.129; p = 0.038), PT ⩾18 s (OR = 4.749, 95% CI: 1.252–18.007; p = 0.022) and WBC ⩾100 × 10⁹/L (OR = 10.591, 95% CI: 1.995–56.232; p = 0.006) were found to be independent risk factors of ED (Table 3). However, after further stratification of the 21 patients with a WBC over 100 × 10⁹/L by induction regimens, the ATRA + ATO + stDNR subgroup still had the lowest ED rate of 11.1%, followed by the ATRA + ATO + ldDNR subgroup at 33.3% and the ATO + others subgroup at 88.9% (p = 0.004; Table S2).

Table 2.

Early death rate in each risk group.

Characteristics	Early death rate, % (n)	χ2	p-Value
Gender
Male	16.9 (12/71)	0.655	0.418
Female	11.9 (7/59)
Age
⩾60 Years	30.0 (3/10)	2.055	0.163
<60 Years	13.3 (16/120)
Hemoglobin
⩾60 g/L	13.2 (14/106)	0.912	0.345
<60 g/L	20.8 (5/24)
Platelet (×109/L)
⩾40	6.3 (2/32)	2.38	0.156
<40	17.3 (17/98)
WBC (×109/L)
WBC ~50	6.3 (5/79)	22.546	<0.001
WBC 50–100	13.8 (4/29)
WBC ~100	47.6 (10/21)
PT
⩾18 s	37.1 (13/35)	19.477	<0.001
<18 s	6.3 (6/95)
APTT
⩾20 s	20.4 (11/54)	2.451	0.136
<20 s	10.5 (8/76)
Fib
⩾1.0 g/L	11.7 (9/77)	1.297	0.315
<1.0 g/L	18.9 (10/53)
D-dimer
⩾20.0 mg/L	11.8 (6/51)	0.547	0.460
<20.0 mg/L	16.5 (13/79)
Mucocutaneous bleeding
Yes	18.7 (17/91)	4.018	0.057
No	5.1 (2/39)
Pulmonary infiltration on CT
Yes	17.6 (9/51)	0.618	0.432
No	12.7 (10/79)
Induction regimens
ATO + ATRA + stDNR	4.0 (2/50)	29.450	<0.001
ATO + ATRA + ldDNR	10.2 (6/59)
ATO + others	52.4 (11/21)
Era
Before 2017	18 (9/50)	0.746	0.388
After 2017	12.5 (10/80)
Differentiation syndrome
Yes	20.3 (13/64)	3.279	0.085
No	9.1 (6/66)

Significant p-value is in boldface.

ATO, arsenic trioxide; ATRA, all-trans retinoic acid; ATO + ATRA + ldDNR, ATO + ATRA with consecutive low-dose anthracycline; ATO + ATRA + stDNR, ATO + ATRA dual induction plus standard-dose anthracycline; WBC, white blood cell; PT, prothrombin time; APTT, activated partial thromboplastin time; Fib, fibrinogen.

Table 3.

Risk factors of ED identified by multivariate analysis.

Variables	Coefficient	Standard error	Walds	p-Value	OR	95% CI of OR
ATO + ATRA + ldDNR vs ATO + ATRA + stDNR	1.194	0.976	1.498	0.221	3.301	0.488–22.355
ATO + others vs ATO + ATRA + stDNR	3.231	0.99	10.65	0.001	25.313	3.635–176.271
WBC 50–100 vs WBC ~50	2.881	0.9	1.206	0.272	2.687	0.461–15.672
WBC ~100 vs WBC ~50	1.765	0.852	7.677	0.006	10.591	1.995–56.232
Age ⩾60 Years	0.988	1.054	4.297	0.038	8.888	1.126–70.129
PT ⩾18 s	2.36	0.68	5.248	0.022	4.749	1.252–18.007

Significant p-value is in boldface.

ATO, arsenic trioxide; ATO + ATRA + ldDNR, ATO + ATRA with consecutive low-dose anthracycline; ATO + ATRA + stDNR, ATO + ATRA dual induction plus standard-dose anthracycline; PT, prothrombin time; WBC, white blood cell.

OS and risk factors

Apart from 19 ED patients, 109 patients achieved molecular CR at the postinduction evaluation. Eight patients experienced relapse during the follow-up. None of the 48 patients in the ATO + ATRA + stDNR group had relapse, while 9.4% (5/53) patients in the ATO + ATRA + ldDNR group relapsed, and the relapse rate was the highest at 30.0% (3/10) in the ATO + others group. The difference was significant (p = 0.003). After a median follow-up of 65 months (range, 1–185 months), the 5-year OS (5y-OS) of the total patients was 79.2% (Figure 1). Compared with the 5y-OS at 80% in the ATO + ATRA + ldDNR group, the ATO + ATRA + stDNR group had the highest 5y-OS at 96% (p = 0.039). However, the 5y-OS was 31% in the ATO + others group, which was significantly inferior to the previous two groups (p < 0.001; Figure 2(a); Table S3). The 5y-OS difference in WBC ~50 and WBC 50–100 groups was insignificant (83% vs 81%, p = 0.738). However, for patients with WBC 100~, the 5y-OS was 46%, which was significantly inferior to the previous WBC groups (p < 0.001 and p = 0.008, respectively; Figure 2(b); Table S4). Further stratification by WBC count (Table S5) also revealed a trend toward an OS advantage in the ATO + ATRA + stDNR subgroup over the ATO + ATRA + ldDNR subgroup (88% vs 67%, p = 0.04) in the WBC ⩾100 × 10⁹/L group, and ATO + others subgroup had the poorest OS (p < 0.001; Figure 2(c)). In the WBC < 100 × 10⁹/L group, the ATO + ATRA + stDNR subgroup had no statistical difference in the OS compared with the ATO + ATRA + ldDNR subgroup (97% vs 81%, p = 0.392), while ATO + others still had the most inferior 5y-OS (55%) compared with the previous two groups (p < 0.001 and p = 0.02, respectively; Figure 2(d)). In addition, 5y-OS of patients with PT ⩾18 s was significantly poorer than patients with PT < 18 s (88% vs 46%, p < 0.001; Figure 2(e)). Compared with patients after 2017, patients diagnosed before 2017 had a poorer 5-year OS (68% vs 85%, p = 0.039; Figure 2(f)). The 5y-OS rates of different risk groups were listed in Table 4. By multivariate analysis (Table 5) ATO + ATRA + stDNR group was still associated with a better OS compared with ATO + ATRA + ldDNR (hazard ratio (HR) = 5.079, 95% CI: 1.071–24.079; p = 0.041) and ATO+ others (HR = 16.173, 95% CI: 3.611–72.438; p < 0.001). Moreover, no significant difference was found between WBC 50–100 and WBC ~50 (HR = 1.283, 95% CI: 0.414–3.979; p = 0.666). Only WBC ⩾100 × 10⁹/L was associated with inferior OS (HR = 3.402, 95% CI: 1.359–8.518; p = 0.009).

Figure 1.

The Kaplan–Meier curve of the total high-risk APL patients.

APL, acute promyelocytic leukemia.

Figure 2.

The Kaplan–Meier curves of the OS of high-risk APL patients stratified by (a) induction regimens (ATO + ATRA + stDNR vs ATO + ATRA + ldDNR vs ATO + others), (b)WBC count(WBC ⩾100 × 109/L vs WBC 50–100 × 109/L vs WBC < 50 × 109/L), (c) induction regimens (ATO + ATRA + stDNR vs ATO + ATRA + ldDNR vs ATO + others) with WBC < 100 × 109/L, (d) induction regimens (ATO + ATRA + stDNR vs ATO + ATRA + ldDNR vs ATO + others) with WBC ⩾100 × 109/L, (e) PT < 18 s vs PT⩾18 s, and (f) era before 2017 vs era after 2017.

APL, acute promyelocytic leukemia; ATO, arsenic trioxide; ATO+ATRA+ldDNR, ATO+ATRA with consecutive low-dose anthracycline; ATO+ATRA+stDNR, ATO+ATRA dual induction plus standard-dose anthracycline; PT, prothrombin time; WBC, white blood cell.

Table 4.

Survival of different risk groups.

Risk factors	5-Year OS	χ2	p-Value
Gender
Male	0.7	2.338	0.126
Female	0.84
Age
⩾60 Years	0.52	2.785	0.095
<60 Years	0.79
Hb
⩾60 g/L	0.8	3.379	0.066
<60 g/L	0.59
Plt (×109/L)
⩾40	0.77	0.123	0.726
<40	0.77
WBC (×109/L)
WBC <50	0.83	17.018	<0.001
50 ⩽ WBC < 100	0.81
WBC ⩾100	0.46
PT
⩾18 s	0.46	24.145	<0.001
<18 s	0.88
APTT
⩾20 s	0.75	0.035	0.851
<20 s	0.77
Fib
⩾1.0 g/L	0.77	0.221	0.638
<1.0 g/L	0.76
D-dimer
⩾20.0 mg/L FEU	0.80	0.326	0.568
<20.0 mg/L FEU	0.75
Mucocutaneous bleeding
Yes	0.74	1.53	0.216
No	0.83
Pulmonary infiltration on CT
Yes	0.69	1.877	0.171
No	0.82
Induction regimens
ATO+ATRA+stDNR	0.96	30.106	<0.001
ATO+ATRA+ldDNR	0.8
ATO+others	0.31
Era
Before 2017	0.68	4.267	0.039
After 2017	0.85
Differentiation syndrome
Yes	0.73	1.374	0.241
No	0.8

Significant p-value is in boldface.

ATO, arsenic trioxide; ATO + ATRA + ldDNR, ATO + ATRA with consecutive low-dose anthracycline; ATO + ATRA + stDNR, ATO + ATRA dual induction plus standard-dose anthracycline; OS, overall survival; PT, prothrombin time; WBC, white blood cell.

Table 5.

Risk factors of OS identified by multivariate analysis.

Variables	Coefficient	Standard error	Walds	p-Value	HR	95% CI of HR
ATO+ATRA+stDNR vs ATO+ATRA+ldDNR	1.625	0.794	4.188	0.041	5.079	1.071–24.079
ATO+ATRA+stDNR vs ATO+others	2.783	0.765	13.237	<0.001	16.173	3.611–72.438
WBC ~50 vs WBC 50–100	0.249	0.578	0.186	0.666	1.283	0.414–3.979
WBC ~50 vs WBC 100~	1.224	0.468	6.838	0.009	3.402	1.359–8.518

Significant p-value is in boldface.

ATO, arsenic trioxide; ATO + ATRA + ldDNR, ATO + ATRA with consecutive low-dose anthracycline; ATO + ATRA + stDNR, ATO + ATRA dual induction plus standard-dose anthracycline; OS, overall survival; WBC, white blood cell.

Discussion

In this study, we retrospectively analyzed the prognosis and risk factors of high-risk APL patients treated ATO-based induction regimens in a real-world setting. Earlier trials focusing on the AIDA (ATRA plus idarubicin) regimen, which involves the administration of anthracycline to the first week of induction with the omission of cytarabine, have shown reduced rates of differentiation syndrome and made early incorporation of anthracycline into induction the standard of care.^9,10 In 2015, the APML4 trial studied an induction strategy utilizing ATRA, ATO, and idarubicin, with an improved 5-year disease-free survival of 83% and OS of 87% in high-risk APL patients, ¹¹ further establishing ATRA/ATO/chemotherapy as the standard of care. A recent report by Zhu et al. ¹² also supported the addition of the strong cytoreductive chemotherapy anthracycline/cytarabine combination, which improved the 5-year relapse-free survival in lower-risk APL patients. For high-risk APL, strong cytoreduction certainly constrains leukocytosis more effectively and may reduce the risk of severe DS. On the other hand, the worsening of disseminated intravascular coagulation (DIC) is worrisome, especially for those with hyperleukocytosis and severe disruption of DIC parameters. Thus, a prolonged course of anthracycline at a reduced dosage is also common in real-world practice. APOLLO trial was the first phase III trial comparing ATO + ATRA plus idarubicin as induction therapy followed with four cycles of consolidation therapy by ATO and ATRA (ATRA-ATO arm) with standard AIDA induction followed with three cycles of chemotherapy (CHP) based consolidation (ATRA-CHP arm) for high-risk APL. After a median follow-up of 31 months, no statistically significant differences were observed in the ED rates (7% vs 10%), or OS (93% vs 87%; p = 0.33) between the ATRA-ATO and ATRA-CHT arms. However, a significant difference in the 2-year EFS (88% vs 70%; p = 0.02) was observed. Moreover, the cumulative incidence of relapse rate was significantly higher in the ATRA-CHT arm than in the ATRA-ATO arm (14% vs 1.6%; p = 0.011). ⁵ Gemtuzumab ozogamicin (GO), an anti-CD33 monoclonal antibody conjugated to calicheamicin, is another drug that has shown benefits when added to ATRA plus ATO-based induction regimens for high risk APL patients. GO allows long-term OS rates close to 90%.^13,14 Nevertheless, GO has a relatively high price and has not yet been approved in China. Still there has been no head-to-head comparison between idarubicin and GO in the dual induction era.

In our study, the ATO + ATRA + stDNR induction group, which received standard dose of anthracycline, had the lowest ED rate of 4.0%. This result was superior to those treated with long-duration and low-dose anthracycline, whose ED rate was 10.2%. Although this improvement in ED was not statistically significant by multivariate analysis, the OS advantage of the ATO + ATRA + stDNR group with a 5-year OS rate of 96% was demonstrated by Cox regression. Further stratification by a WBC cutoff of 100×10⁹/L also revealed a statistically significant improvement of OS in the ATO + ATRA + stDNR group compared with the ATO + ATRA + ldDNR group in patients with WBC over 100 × 10⁹/L (p = 0.04). Moreover, none of the 48 patients who received ATO + ATRA + stDNR induction experienced relapse. Thus, we could conclude that adequate cytoreductive chemotherapy with stDNR was superior to low-dose anthracycline when combined with dual induction. As ATRA rapidly improved coagulopathy, early intervention with adequate strength of cytoreductive chemotherapy in high-risk patients might reduce severe DS and subsequent disastrous multiorgan dysfunction. Mucocutaneous bleeding signs were present in 70% of the patients in our study. Thus, administering anthracycline on the second day after the initiation of dual induction is reasonable for patients with DIC. Moreover, the dismal outcome in the ATO + others group confirmed the indispensable role of instantaneous ATRA in high-risk APL patients.

Historically, the prevalence of ED in APL of all risk groups was reported to be up to 30%,^{15 –17} but the proportion of patients with high-risk APL was uncertain. More recent studies reported ED rates less than 10%.^14,18,19 The lower ED rates observed in clinical trials could be explained by selection bias and aggressive hemostatic support in the trial setting. Nevertheless, in some population-based studies, the ED rate has not improved with treatment advances. ²⁰ Similarly, despite obvious improvements in supportive care and socioeconomic conditions in China, ED rate in our study was not significantly reduced after 2017 (18.5% vs 12.5%, p = 0.388). However, the 5y-OS of patients with high-risk APL was improved from 68% to 85% (p = 0.039). The impact of the era on OS was primarily due to differences in the ED rate. However, another important factor is the nationwide use of RIF in China. RIF has been shown to be an effective and convenient replacement for intravenous ATO during both the induction and maintenance periods. ²¹ RIF plus ATRA is used in a home-based manner for APL treatment during postremission therapy. This could dramatically improve patient amenability, decrease medical costs in the maintenance period, and largely prevent the relapse of inadequate maintenance treatment or treatment discontinuation.

Hyperleukocytosis in high-risk APL was a more formidable existence with a much higher ED rate. In an earlier retrospective cohort study, Daver et al. ²² reviewed 242 consecutive APL patients from MD Anderson Cancer Center and found that only 29 patients (12%) had hyperleukocytosis defined as WBC ⩾50 × 10⁹/L at presentation and that patients with hyperleukocytosis had a higher ED rate (24% vs 9%; p = 0.018). More recently, Iland et al. ²³ reported the outcomes of 37 APL patients with extreme hyperleukocytosis (defined as WBC ⩾100 × 10⁹/L) with data obtained from hematologists in Australia, the United Kingdom, and North America. The ED rate was 19%, and 5y-OS rate was 76%. Stratification with treatment showed that adding ATO into ATRA-based regimen significantly improved the prognosis. In our study, 21 patients with extreme hyperleukocytosis were included, and the ED rate was as high as 47.6%. Relapse was in fact uncommon for survivors of ED in the total cohort, making the 5y-OS for patients with WBC < 100 × 10⁹/L still greater than 80%. Thus, reducing ED was the still the key of improving the prognosis of high-risk APL especially for patients with extreme hyperleukocytosis. Nevertheless, the fact that in our study the ATO + ATRA + stDNR subgroup in patients with hyperleukocytosis had the lowest ED of only 11.1% (1/8), and a record of no relapse still made stDNR a preferable choice.

Similar to previous studies,^24,25 the main cause of ED is still DIC-associated catastrophic bleeding. Thus, the presence of coagulopathy is undoubtedly a strong predictor of ED.^26,27 Among those parameters, PT ⩾18 s was shown to be an independent risk factor of ED in the multivariate analysis. Thus, we believe that PT prolongation indicates the consumptive depletion stage of DIC, while low fibrinogen levels and elevated D-dimer levels without prolonged PT are more often observed in the earlier hyperfibrinolytic stage. Thus, supportive transfusion with cryoprecipitate or plasma should be more aggressive for high-risk patients with PT prolongation. The fact that a high proportion of patients at baseline were already presented with mucocutaneous bleeding, low platelet count, low fibrinogen levels, and elevated D-dimer levels might make these parameters less significant for high-risk APL patients.

Age over 60 years was another important risk factor of ED. However, unlike previous reports,^28,29 the median age in our study seemed to be much younger at only 38 years (range, 14–71 years). This might be explained by admission bias, that is, the Berkson’s bias, because younger patients were more often transferred to a tertiary medical center like us, whereas elderly patients were more willing to receive treatment in local hospitals. This was determined by considerations such as economic burden and a traditional culture of staying locally for the elderly. Another explanation could be the increased prehospital mortality of the elderly APL patients due to delayed clinical visits and diagnosis. Therefore, how to treat this elderly subgroup has yet to be elucidated.

This study had several limitations. First, as a retrospective study at a single center, some clinical details were not intact and the treatment regimens were broadly categorized. This could restrict the number of predictors that could be included in the multivariable model. Second, in the real-world setting the patients enrolled had a prolonged time span and the treatment choice might be influenced by physician preference, which could lead to certain biases and confounders. Thus, multicenter randomized controlled trials are still needed in the future to fully validate the results of this study.

Conclusion

In conclusion, dual induction with ATO and ATRA plus a cytoreductive anthracycline was the most commonly used regimen for high-risk APL patients in this single-center real-world retrospective study. Both ATO and ATRA are indispensable for high-risk APL patients. Compared with the consecutive use of low-dose anthracycline, stDNR was much more preferable, with a reduced ED rate and excellent OS. Although a high WBC count was associated with ED risk, the prognosis of patients with a WBC count less than 100 × 10⁹/L was largely improved, with a 5y-OS greater than 80%. In addition to age and WBC count, PT prolongation was an independent risk factor of ED. More work is required to reduce the ED of APL patients with a WBC count greater than 100 × 10⁹/L.