A retrospective chart review was performed of all the 141 patients with APL registered at the Hemope Foundation between January 2007 and December 2018. Hemope is a state reference blood center for diagnosing and treating hematologic diseases in Recife, PE, Brazil. A total of 112 patients were included in the study for analysis whether they were eligible for the IC-APL 2006 protocol (n = 97) or not (n = 15). The main reasons for exclusion from the protocol were treatment in another hospital (n = 10), no molecular diagnosis (n = 9), younger than 18 years old (n = 3), loss to follow-up (n = 2), and mislaid charts (n = 2).

The IC-APL 2006 protocol adapted from PETHEMA LPA 2005 was used, however treatment with idarubicin was replaced by the more cost-effective daunorubicin. 19 Patients were selected for the IC-APL treatment protocol based on eligibility criteria: age 18–75 years old at diagnosis; Eastern Cooperative Oncology Group (ECOG) performance status ≤3; morphological diagnosis of APL or promyelocytic leukemia-retinoic acid receptor-α (PML-RARα) rearrangement. The PML-RARα rearrangement was confirmed by reverse transcription polymerase chain reaction (RT-PCR). Patients who met these criteria but had one of the following characteristics received alternative treatment protocol adaptations accordingly: associated psychiatric or neoplastic disease, seropositivity for the human immunodeficiency virus, serum creatinine ≥3.5 mg/dL or clinical contraindication for anthracycline.

Treatment was initiated promptly upon diagnostic suspicion. The induction therapy consisted of ATRA and daunorubicin, and the consolidation therapy consisted of three courses of treatment as described by Rego et al., 19 transfusion support aimed at maintaining a platelet count >30–50×109/L and a fibrinogen level >15 g/L. Differentiation syndrome prophylaxis was made with dexamethasone for 15 days in patients who presented WBC counts >5.0 × 109/L within the first two weeks.

This study was approved by the Hemope Ethics Committee and was carried out in accordance with the Declaration of Helsinki. For this study, the signing of informed consent forms was waived, as patients had already given their consent for the main IC-APL project.

Data from medical records were analyzed using SPSS 13.0. Measures of central tendency and dispersion presented numerical variables. The Student's t-test was used to compare standard distribution variables and Mann-Whitney for non-normal data. Fisher's exact test and Pearson chi-square test were used for crossing of categorical variables. The multivariate analysis used logistical regression (ENTRE method) and included variables with p-values ≤0.20 in the bivariate analysis. All the tests were applied with 95% confidence intervals.

A total of 112 patients were analyzed (Table 1); they were predominantly young, with a median age of 36 (range: 18–81) years, as reported by earlier studies. 11,13,19–21 Interestingly, more recent studies have reported older ages at diagnosis, with median ages varying from 44 to 54 years. 7–9,17,18,22 This divergence might be attributed to the prevalence of database studies in more recent literature, unlike the previous clinical trial studies that tend to exclude factors that could heighten patient risk, such as older age. Consistent with existing literature, no significant sex differences were found.5,7–9,17,22 Most records analyzed belonged to patients eligible for the IC-APL protocol. Among the 15 patients not eligible for the protocol, most exhibited clinical features that rendered them unqualified with the main reason for protocol ineligibility being ECOG performance status of 4.

All patients had anemia and low platelet counts at diagnosis; WBC count varied from leukopenia to severe leukocytosis. This classic APL presentation is well-established, though a higher WBC count is considered a poor prognostic marker.2,7,11,13–17 Patients were categorized as low, intermediate, and high risk as per the risk model of Sanz et al.23 Symptom-diagnosis time spanned 1–90 days (interquartile range: 7–15 days).

The overall ED rate was 22.3%. Among patients eligible for the IC-APL protocol, the ED rate was 16.5%, whereas non-eligible patients exhibited a much higher rate of 60%. Studies on single-drug treatment protocols10–15,19,21,24,25 and database studies7–9,22,26 reported 12–32.6% ED rates (Table 2). Notably, centers using the ATRA and ATO combo in multiagent therapy achieve 4.0–7.9% ED rates.2,4,5,7,13 The association between both drugs has also been reported as a factor for better overall and disease-free survival.2,4,5,7,13

Beyond multi-drug availability, socioeconomic factors also influence ED rates, as evidenced by worse outcomes in developing nations21,25 and disadvantaged areas in developed countries. 8,9,17,22,26 Real-life studies also tend to present higher ED rates, probably due to the restrictive conditions of clinical trials and the selective nature of patient recruitment. 8,9,17,18,22 In this study, all patients ineligible for the IC-APL protocol were included, and their worse clinical status might explain the higher ED rate.

ED was caused mainly by hemorrhage and infection (Table 3). CNS was the primary bleeding site with two patients presenting concomitant intracranial hypertension. The lungs were the second major site for bleeding and the first for infection. ED was also related to respiratory and kidney failures; only one patient died from differentiation syndrome. The high incidence of hemorrhagic in EDs is expected because of the classical clinical presentation of the disease with coagulopathy. Lethal hemorrhage within the first 30 days is often reported as the main cause of death, with CNS bleeding accounting for 50–92.3% of ED rates.2,5,7,10–14,21,25

The percentages of hemorrhagic and infectious causes of death varies based on many factors. Jin et al. found that the proportion of infectious ED was higher in older patients (>50 years old) compared to younger patients (30.3% vs. 3.1%), while the hemorrhagic ED rate was significantly lower (48.5% vs. 87.5%; p-value = 0.0006), possibly because of the worse clinical status of the elderly group, especially regarding the immune system. 11,19 Time of death also matters, as most hemorrhagic deaths occur within the first seven days of diagnosis and infectious EDs happen between the 7th and 30th day with a higher incidence of sepsis-related death described in the consolidation stage of treatment. 11,21,25 This might be related to a low tolerance to treatment protocols, as hemorrhagic deaths are attributed to the disease itself, whereas infectious deaths are also associated with immune impairment from chemotherapy side effects. This study took place at a blood center, which is known to offer hematology and blood transfusion services in the same building; it even has a specialized intensive care unit (ICU). This might have reduced the incidence of ED due to coagulopathy and differentiation syndrome.

Socioeconomic factors might play a role in infectious ED rates. Silva et al. found that all patients treated at an academic center in Brazil experienced febrile neutropenia during induction, with a high incidence of positive blood cultures and evidence of fungal infections. 21 Developed countries and China usually show lower infectious death rates (0–17.4%) than developing nations (33–52%) such as Brazil and South Africa.5,7,12,21,25 This might be due to population characteristics, epidemiological distribution of resistant microorganisms and the less common use of multi-drug treatment, such as ATO, which is not widely available in these countries.