Welcome international healthcare professionals

This site is no longer supported and will not be updated with new content. You are welcome to browse and download all content already included in the site. Please note you will have to register your email address to access the site.

You are here

Delay in the administration of all-trans retinoic acid and its effects on early mortality in acute promyelocytic leukemia: Final results of a multicentric study in the United States

Leukemia Research

Highlights

 

  • Early death continues to occur in 17% of patients with APL.
  • ATRA is often not administered promptly upon first suspicion of APL.
  • Prompt (versus delayed) ATRA administration did not decrease overall early death in our series.

Abstract

Early death (ED) occurs in 10–30% of patients with acute promyelocytic leukemia (APL). Is all-trans retinoic acid (ATRA) promptly given and does it decrease overall early mortality? ATRA was administered within 24 h of morphological suspicion in only 44% of the 120 consecutive patients treated in the four collaborating centers. Absence of disseminated intravascular coagulation (p = 0.012) and admission to a non-university-affiliated hospital (p = 0.032) were independent predictors of ATRA delay. ED occurred in 17% of patients, and was independently correlated only with ICU admission (p = 0.002). Our results do not demonstrate that prompt (versus delayed) ATRA administration decreases overall early death.

Keywords: ATRA, Death, Leukemia, Promyelocytic.

1. Introduction

Approximately 600–800 new cases of acute promyelocytic leukemia (APL) are diagnosed each year in the United States [1] . When untreated, APL is the most rapidly fatal acute myeloid leukemia (AML). However, with the introduction of all-trans retinoic acid, and recently arsenic trioxide (minimizing requirements for use of cytotoxic chemotherapy), APL has become the most curable acute myeloid leukemia, with complete remission and cure rates exceeding 90% and 70%, respectively, in those who do not succumb to early death [2], [3], [4], [5], and [6]. Several major practice guidelines recommend starting all-trans retinoic acid (ATRA) as soon as APL is morphologically and clinically suspected without waiting for cytogenetic confirmation [7], [8], [9], [10], and [11].

This recommendation is supported by the following facts: ATRA is known to correct biologic signs of APL coagulopathy, is only rarely associated with clinically significant side effects, and, as shown recently, may decrease the relative risk of early hemorrhagic death [11] and [12].

During the last few years, we have noticed that prompt initiation of ATRA upon suspicion of APL is suboptimal at least in some hospitals in the United States, and early mortality rates are frequently higher than expected. These observations prompted us to conduct a retrospective multicentric study in an attempt to answer the following two questions: (i) Are oncologists optimally adherent to the strategy of prompt ATRA administration, and if not, what factors are associated with suboptimal adherence? (ii) Does prompt ATRA administration independently decrease overall early mortality? In other words, in reducing the high mortality of the “sick” APL patient, what is the independent role of prompt ATRA administration relative to factors such as aggressive transfusion support and intensive care unit (ICU) level of care? We have recently reported a preliminary report of our findings as related to early outcome effects of ATRA [13] . The results of the now completed study enable us to assess clinical practice attitudes with regards to timing of ATRA administration.

2. Materials and methods

The study protocol was approved by the institutional review boards of all collaborating centers (Sentara Hospitals, Norfolk, VA; University of Virginia, Charlottesville, VA; West Virginia University, Morgantown, WV, and Penn State Milton S. Hershey Medical Center, Hershey, PA). Data for all patients who were diagnosed with APL and treated with ATRA (with or without cytotoxic chemotherapy) in one of the collaborating centers between January 1996 and March 2013 were retrospectively collected. All data were pooled, anonymized, and analyzed by one of the authors (AR). The diagnosis of APL required cytogenetic confirmation of t(15;17) and/or presence of PML-RARα on reverse-transcriptase polymerase chain reaction. Disseminated intravascular coagulation (DIC) was defined as described elsewhere [14] . The Sanz's risk scoring system was used for risk stratification [15] , with high-risk disease defined as white blood cell count (WBC) > 10,000/μL and low-risk disease as WBC≤10,000/μL. Delayed ATRA was defined as administration more than 24 h following the first suspicion of APL. The time of first suspicion of APL was determined by reviewing the oncologists’ notes as well as the results of peripheral blood smears (and when they were relayed to the oncologists). Early death was defined as death occurring within the first 30 days from the date of admission to the hospital. The criteria for ICU admission or transfer was according to the physician's discretion but general medical criteria such as hemodynamic instability, presence of DIC, intensive monitoring of vital signs, and aggressive transfusion support were applied to determine the need for critical care.

2.1. Statistical analysis

Data are presented as mean ± standard deviation or frequency (%). Intergroup analysis was performed using Chi-squared test and Fisher's exact test for frequencies, Student's t-test for means of variables with a normal distribution, and Mann–Whitney U-test for medians of variables with a skewed distribution. Univariate analysis was first performed to identify variables with significant difference between the two groups. These variables were then entered as potential predictors in a multivariate binary logistic regression model with a categorical (binary) variable of interest being the dependent variable. The effects of significant predictors in multivariate models were assessed by odds ratio (OR). All analyses were done using the Statistical Software for Social Sciences (SPSS 21.0) and a p-value of smaller than 0.05 was considered statistically significant throughout analyses.

3. Results

A total of 120 patients were included. The mean ± standard deviation age of patients was 49 ± 16 years, and 55% were female. Bleeding (48%), symptoms of anemia (25%), and infections (20%) were the most common presenting symptoms (n = 93). According to Sanz's risk stratification, 74 (64%), and 41 (36%) patients were low- and high-risk, respectively (n = 115). A total of 63/115 (55%) patients presented with or later developed DIC, and 26/85 (31%) patients had additional cytogenetic abnormalities, consistent with previous reports [16] . A total of 105/120 (88%) patients presented directly and were admitted to a university-affiliated hospital. A total of 39/112 (35%) patients required intensive care unit (ICU) admission or transfer during the course of their hospitalization, and 28/111 (25%) patients developed the differentiation syndrome. The exact time of ATRA administration was available in 102 patients. Among these patients, ATRA was administered within 24 h of the time APL was morphologically suspected in 45 (44%) patients, 24–48 h later in 30 (29%) patients, and with a longer delay in the remainder (27%). The median (5–95th percentile) delay between the first suspicion of APL and ATRA administration was 1 (0–3.9) days. ATRA was readily available in the pharmacy in all cases, with no limitations. Early death occurred in 20/120 (17%) patients, 11 (55%) of which was due to catastrophic intracranial or intrapulmonary hemorrhage.

We first divided our patients into two groups ( Table 1 ): A (early death; n = 20) and B (no early death, n = 100). There was no significant difference between the groups with regards to hospital type (university-affiliated vs. community; p = 1.0), age (p = 0.19), sex (p = 0.34), presence of additional cytogenetic abnormalities (p = 1.0), development of the differentiation syndrome (p = 0.57), and whether or not ATRA was delayed (p = 0.08). However, DIC on admission and ICU admission/transfer were significantly more common among patients in group A than B (p = 0.003 and <0.001, respectively). Also, there were more high-risk patients in group A than in group B (p = 0.001). In a multivariate binary logistic regression model with group (A vs. B) as the dependent variable and DIC (0 vs. 1), ICU admission/transfer (0 vs. 1), and Sanz's risk score (L vs. H) as potential categorical predictors, only ICU admission/transfer emerged as a significant independent predictor of early mortality (OR = 10, p = 0.002; R2 for model: 0.39).

Table 1 Characteristics of the entire study population and comparison between groups A and B.

  All A (early death) B (no early death) p-value
n 120 20 100
Age (years) 49 ± 16 53 ± 16 (n = 20) 48 ± 15 (n = 100) 0.19
Male 54/120 (45%) 11/20 (55%) 43/100 (43%) 0.34
University-affiliated hospital 105/120 (88%) 17/20 (85%) 88/100 (88%) 0.71
Additional cytogenetic abnormalities 26/85 (31) 3/11 (27) 23/74 (31) 1.00
Delayed ATRA 71/120 (59%) 8/20 (40%) 63/100 (63%) 0.08
Differentiation syndrome 28/115 (24%) 3/19 (16%) 25/96 (26%) 0.57
DIC 63/115 (55%) 17/20 (85%) 46/95 (48%) 0.003 **
ICU admission/transfer 39/112 (35%) 17/20 (85%) 22/92 (24%) <0.001 **
Sanz's risk score (L:H) 74:41 6:14 68:27 0.001 **

** p < 0.01.

ATRA: all-trans retinoic acid; DIC: disseminated intravascular dissemination; ICU: intensive care unit.

Next, we restricted the analysis to patients who were low-risk according to Sanz's risk score. As shown in Table 2 , ICU admission/transfer was the only variable with significant difference between the groups, being more frequent in group A (p = 0.016). Next, we restricted the analysis to patients who were high-risk ( Table 3 ). Again, ICU admission/transfer was the only significant correlate of early mortality (OR = 1.8, p = 0.01). Next, we restricted the analysis to patients who had DIC ( Table 4 ). ICU admission/transfer (p = 0.001) and Sanz's risk score (p = 0.011) were the only variables with significant difference between the groups, with higher frequency of ICU admission/transfer and higher risk scores in group A. In a multivariate binary logistic regression model with group (A vs. B) as the dependent variable and Sanz's risk score (L vs. H) and ICU admission/transfer (0 vs. 1) as potential categorical predictors, only ICU admission/transfer emerged as a significant independent predictor of early mortality (OR = 8, p = 0.015; R2 for the model: 0.31). Finally, we restricted the analysis to patients who were admitted/transferred to the ICU. There was no significant difference between groups A and B in any of the studies variables ( Table 5 ).

Table 2 Comparison between groups A and B restricted to low-risk patients according to Sanz's risk score.

  A (early death) B (no early death) p-value
n 6 68
Age (years) 51 ± 21 49 ± 15 0.77
Male 2/6 (33%) 31/68 (46%) 0.69
University-affiliated hospital 4/6 (67%) 57/68 (84%) 0.28
Additional cytogenetic abnormalities 1/4 (25) 15/52 (29) 1.00
Delayed ATRA 3/6 (50%) 46/68 (68%) 0.40
Differentiation syndrome 0/6 (0%) 15/68 (22%) 0.34
DIC 4/6 (67%) 28/68 (41%) 0.39
ICU admission/transfer 4/6 (67%) 11/65 (17%) 0.016 *

* p < 0.05.

ATRA: all-trans retinoic acid; DIC: disseminated intravascular dissemination; ICU: intensive care unit.

Table 3 Comparison between groups A and B restricted to high-risk patients according to Sanz's risk score.

  A (early death) B (no early death) p-value
n 14 27
Age (years) 54 ± 14 (n = 14) 48 ± 15 (n = 27) 0.18
Male 9/14 (64%) 10/27 (37%) 0.12
University-affiliated hospital 13/14 (93%) 26/27 (96%) 1.00
Additional cytogenetic abnormalities 2/7 (29) 8/22 (36) 1.00
Delayed ATRA 5/14 (36%) 14/27 (52%) 0.51
Differentiation syndrome 3/13 (23%) 10/27 (37%) 0.48
DIC 13/14 (93%) 18/27 (67%) 0.12
ICU admission/transfer 13/14 (93%) 11/26 (42%) 0.002 **

** p < 0.01.

ATRA: all-trans retinoic acid; DIC: disseminated intravascular dissemination; ICU: intensive care unit.

Table 4 Comparison between groups A and B restricted to patients with DIC.

  A (early death) B (no early death) p-value
n 17 46
Age (years) 53 ± 17 (n = 17) 47 ± 15 (n = 46) 0.22
Male 10/17 (59%) 19/46 (41%) 0.26
University-affiliated hospital 14/17 (82%) 40/46 (87%) 0.69
Additional cytogenetic abnormalities 2/9 (22) 9/40 (23) 1.00
Delayed ATRA 6/17 (35%) 24/46 (52%) 0.27
Differentiation syndrome 2/16 (13%) 15/46 (33%) 0.19
ICU admission/transfer 15/17 (88%) 18/43 (42%) 0.001 **
Sanz's risk score (L:H) 4:13 28:18 0.011 *

* p < 0.05.

** p < 0.01.

ATRA: all-trans retinoic acid; DIC: disseminated intravascular dissemination; ICU: intensive care unit.

Table 5 Comparison between groups A and B restricted to patients with ICU admission/transfer.

  A (early death) B (no early death) p-value
n 17 22
Age (years) 52 ± 17 (n = 17) 47 ± 16 (n = 22) 0.28
Male 9/17 (53%) 8/22 (36%) 0.35
University-affiliated hospital 15/17 (88%) 19/22 (86%) 1.00
Additional cytogenetic abnormalities 1/8 (13) 6/20 (30) 0.63
Delayed ATRA 8/17 (47%) 13/22 (59%) 0.53
Differentiation syndrome 3/16 (19%) 9/22 (41%) 0.18
DIC 2/17 (12%) 4/22 (18%) 0.68
Sanz's risk score (L:H) 4:13 11:11 0.11

ATRA: all-trans retinoic acid; DIC: disseminated intravascular dissemination; ICU: intensive care unit.

These results collectively demonstrate that, among the studied variables, ICU admission/transfer is the only significant independent predictor of early mortality. To determine what factors may have contributed to the clinician's decision to admit/transfer the patient to the ICU, we divided the study population to two categories: Group C (patients admitted/transferred to the ICU) and Group D (patients not admitted/transferred to the ICU). As shown in Table 6 , DIC and high-risk disease were significantly more common among patients admitted/transferred to the ICU (p < 0.001 for both). In a multivariate binary logistic regression model (R2 = 0.35) with group (C vs. D) as the dependent variable and Sanz's risk score (L vs. H) and DIC (0 vs. 1) as potential categorical predictors, both DIC (OR = 7, p < 0.001) and high Sanz's risk score (OR = 3.9, p = 0.004) emerged as significant independent predictors of ICU admission/transfer. Although there was a significant association between DIC and Sanz's risk score (p = 0.001, Spearman's correlation coefficient: 0.31), each had a significant independent contribution to the decision on ICU admission/transfer.

Table 6 Comparison between groups C and D.

  C (ICU admission/transfer) D (no ICU admission/transfer) p-value
n 39 73
Age (years) 49 ± 16 (n = 39) 49 ± 14 (n = 73) 0.97
Male 17/39 (44%) 34/73 (47%) 0.84
University-affiliated hospital 34/39 (87%) 64/73 (88%) 1.00
Additional cytogenetic abnormalities 7/28 (25) 18/54 (33) 0.61
Delayed ATRA 21/39 (54%) 43/73 (59%) 0.69
Differentiation syndrome 12/38 (32%) 15/72 (21%) 0.25
DIC 33/39 (85%) 27/72 (38%) <0.001 **
Sanz's risk score (L:H) 15:24 56:16 <0.001 **

** p < 0.01.

ATRA: all-trans retinoic acid; DIC: disseminated intravascular dissemination; ICU: intensive care unit.

There was a significant association between ATRA delay and a low risk score (p = 0.036, Spearman's correlation coefficient: 0.20). Interestingly, the absence of DIC was significantly associated with ATRA delay (p = 0.005, Spearman's correlation coefficient: 0.26). Finally, ATRA was delayed significantly (p = 0.024) more frequently in non-university-affiliated hospitals (13/15; 87%) than university-affiliated hospitals (58/105; 55%). A multivariate binary logistic regression model with ATRA delay (0 vs. 1) as the dependent variable and DIC (0 vs. 1), Sanz's risk score (L vs. H), and hospital type (university-affiliated vs. non-university-affiliated) as potential categorical predictors, the absence of DIC (p = 0.012) and admission to a non-university-affiliated hospital (p = 0.032) were significant independent predictors of ATRA delay (R2 = 0.17). As mentioned above, hospital type was not associated with early death ( Table 1 ) or ICU admission/transfer ( Table 6 ). Similarly, we did not find a significant association between hospital type and DIC (p = 0.78).

4. Discussion

Several national and international practice guidelines recommend explicitly that ATRA be promptly administered as soon as APL is clinically and morphologically suspected, without waiting for cytogenetic confirmation [7], [8], [9], [10], and [11]. This recommendation is based mainly on the favorable side effect profile of ATRA (which would be an important factor if the suspected diagnosis was later found to be incorrect), its dramatic effects on biologic signs of APL coagulopathy, and as shown recently, decreasing the relative proportion of patients who die with hemorrhagic complications of APL [12] . Our results from the present study shed light on two relevant questions: (i) Are oncologists following the prompt ATRA recommendation, and if not, why? (ii) Is delayed ATRA administration an independent predictor of increased early overall mortality?

Are oncologists following the prompt ATRA recommendation? The answer is clearly no. In a large recent study, ATRA was delayed in 69% of patients, with a median (5–95th percentile) delay of 1 (0–6.5) day [12] . These rates were somewhat better in our study, with a delay occurring in 59% of patients and a median (5–95th percentile) delay of 1 (0–3.9) day. While we cannot confidently generalize this result to the whole country, suboptimal adherence to recommended practice guidelines are concerning. Among all studied variables, only absence of DIC and admission to a non-university-affiliated hospital emerged as significant correlates of delayed ATRA. There was a significant association between ATRA delay and a low risk score (p = 0.036). Importantly, ATRA was readily available in all cases should the clinician have chosen to administer it without delay. In multivariate analysis, both the absence of DIC and admission to a non-university-affiliated hospital independently and significantly predicted delayed ATRA administration. However, these two variables only accounted for 17% of the variation in the dependent variable (i.e. delayed ATRA) in a linear regression model, suggesting that there are other factors impacting the clinician in their decision making regarding ATRA administration. Transfer of patients with suspected APL from non-university-affiliated hospitals to university-affiliated hospitals before administration of ATRA may explain, at least partly, why initial admission to non-university-affiliated hospitals was associated with ATRA delay. These results suggest that, especially in non-university-affiliated hospitals, clinicians tend to delay ATRA administration in the absence of DIC, even when morphological/clinical suspicion of APL exists. Several factors may underlie this tendency, including but not limited to: (i) concerns about prescribing ATRA to a non-APL AML patient who might get affected by some side effect of ATRA, (ii) concerns about re-imbursement from insurance companies if the diagnosis turns out to be non-APL AML after empiric ATRA administration, and (iii) issues related to not working in active academic centers, particularities of shift schedules, cross coverage, intent to transfer patient to university-affiliated hospitals, reticence in giving ATRA to patients who do not appear to be “very sick”, etc. Further studies are needed to evaluate the generalizability of these results and also to identify the factors that affect the decision making process regarding prompt ATRA administration.

Early death continues to occur in 10–30% of all patients diagnosed with APL, and has not declined in the last two decades despite widespread use of ATRA [12], [17], [18], and [19]. A retrospective review of the data from 13 population-based cancer registries (n = 1400) in the United States estimated the incidence of early death to be 17.3% in the study period (1999–2007). Patients diagnosed between 1996 and 2001 had a significantly lower early death rate than those diagnosed between 1992 and 1995 (OR = 0.6, p = 0.012). No improvement in early death rates was observed between 2002 and 2007 compared to early years 1992–1995 [18] . The authors attributed the clinically modest improvement in early death rates, despite widespread ATRA availability, to non-prompt initiation of ATRA and delayed referrals to more experienced centers. In other studies, early death was estimated to occur in 26% of APL patients in the pre-ATRA era [20] and [21], and average early mortality between 1977 and 2007 has been 20% [17] .

The natural question here is: “does early administration of ATRA affect overall early mortality?” The International Consortium on Acute Promyelocytic Leukemia (IC-APL), established in 2005, led to creation of a network of collaborating institutions in developing countries with the common purpose of improving APL outcomes. A report of the results of this consortium on 183 patients diagnosed between 2006 and 2010 estimated an early death rate of 15% [22] . A comparison of the results with those in a previous Brazilian study showed a 50% reduction in early death rates [23] . A combination of factors was thought to have contributed to this success, including wider ATRA availability, implementation of educational measures for clinicians, early referral to more experienced centers, and finally, early initiation of ATRA. Since independent contribution of each factor was not assessed during analysis, it cannot be determined whether prompt ATRA initiation per se reduced early mortality.

Our results from the present study do not demonstrate that prompt ATRA initiation per se decreases overall early mortality. We are, to the best of our knowledge, the fourth group to demonstrate this: two multicentric studies and one single-center study made the same surprising observation [12], [17], and [19]. However, caution needs to be practiced in interpreting our results. The group of patients that received ATRA promptly was very different from the group that experienced a delay in ATRA administration. As explained above, DIC was significantly less common in the latter group. Additionally, patients in the latter group were treated significantly more frequently in non-university-affiliated hospitals. Did patients who received ATRA with a delay do as well (in terms of early overall mortality) simply because they were “less sick”? The answer to this question using our data can be sought only via control of the differences between the two groups during analysis. Various modifications of analysis (e.g. restricting the analysis to patients with low Sanz's risk scores, those with high risk scores, those with ICU admission/transfer, and those with DIC) did not show any significant mortality role for delayed ATRA. A study with two otherwise similar groups, one receiving prompt ATRA and the other receiving delayed ATRA, would be an ideal setting to directly evaluate the role of early ATRA in early overall mortality. While conducting such a study in a prospective manner does not seem feasible, a similar retrospective study may be possible in a setting where delayed ATRA is purely due to availability limitations, and not due to patient characteristics or the clinician's intuitive judgment of who might need ATRA more urgently. Finally, early death in a recent study was not associated with a lengthier antecedent symptomatic period, suggesting that patients with early death did not die early simply because they presented late [17] .

Our results do not address potential clinical roles of ATRA on APL-related coagulopathy. In fact, a large recent study showed that although prompt ATRA administration delay was not associated with decreased early mortality, it significantly decreased the percentage of death due to hemorrhage [12] . To explain this observation, we speculate that the group that received ATRA without delay enjoyed less DIC due to receiving ATRA sooner, but suffered from more non-DIC related complications of their disease (e.g. infections, late presentation) because it was again this “sicker” group that prompted the clinicians to start ATRA early. ICU admission/transfer emerged in our study as the only significant independent predictor of early mortality in the entire dataset and in all subset analyses. DIC and high-risk disease were significant independent predictors of ICU admission/transfer. The absence of explicit data on transfusion support in our study is limiting us in determining the independent effect of transfusion support on early death.

In conclusion, we showed in a large multicentric retrospective study, that adherence to practice guidelines regarding prompt administration of ATRA in potential APL patients is suboptimal. Our results regarding suboptimal adherence to ATRA administration guidelines are consistent with a recent previous multicentric study [12] , and it appears, from our results, that oncologists, particularly in non-university-affiliated hospitals, tend to delay ATRA in the absence of DIC and in patients with lower Sanz's risk scores. While poor adherence to the available guidelines for the most curable AML is concerning, and requires more in-depth analysis, we and others [12], [17], and [19] have demonstrated that prompt ATRA administration does not appear to decrease early overall mortality (although it likely decreases the percentage of early deaths due to bleeding [12] ). It is likely, however, that a specific subset of patients enjoys lower early mortality when treated with ATRA promptly, and identification of such patients (e.g. new scoring systems) would be an interesting topic for future research. For the time being, early administration of ATRA seems a reasonable strategy, not to decrease overall early mortality in all patients, but to decrease hemorrhagic complications and probably also early mortality in a particular yet to-be-identified subset of patients.

Our results from the present study suggest that patients with DIC and/or high-risk disease tend to be admitted/transferred to the ICU, which while indicating their critical condition, predicts their high early mortality. With early death occurring in only 20 patients in our study, it is hard to know whether some of our statistically non-significant results were due to low statistical power. Another limitation of the present study is its retrospective nature, which does not allow for randomization. Finally, although we are not the first group to report some of these findings, lack of a national database for APL does not allow us to generalize our results to settings other than those of the present study, and future collaborative studies are needed to confirm our results on a national, and perhaps also international, scale. A collaborative initiative by large academic cancer centers can facilitate the formation of a national registry of APL patients.

Conflict of interest statement

None.

Acknowledgements

All authors contributed to data collection, critically reviewed the manuscript, and approved the final draft. AR analyzed the data. AR and SIF wrote the manuscript. The study had no source of funding.

References

  • [1] D. Douer. The epidemiology of acute promyelocytic leukaemia. Best Pract Res Clin Haematol. 2003;16:357-367 Crossref
  • [2] P. Fenaux, S. Chevret, A. Guerci, et al. Long-term follow-up confirms the benefit of all-trans retinoic acid in acute promyelocytic leukemia. European APL group. Leukemia. 2000;14:1371-1377 Crossref
  • [3] M.S. Tallman, J.W. Andersen, C.A. Schiffer, et al. All-trans-retinoic acid in acute promyelocytic leukemia. N Engl J Med. 1997;337:1021-1028 Crossref
  • [4] M.S. Tallman, J.W. Andersen, C.A. Schiffer, et al. All-trans retinoic acid in acute promyelocytic leukemia: long-term outcome and prognostic factor analysis from the North American Intergroup protocol. Blood. 2002;100:4298-4302 Crossref
  • [5] A. Kanamaru, Y. Takemoto, M. Tanimoto, et al. All-trans retinoic acid for the treatment of newly diagnosed acute promyelocytic leukemia. Japan Adult Leukemia Study Group. Blood. 1995;85:1202-1206
  • [6] M.S. Tallman, J.K. Altman. How I treat acute promyelocytic leukemia. Blood. 2009;114:5126-5135 Crossref
  • [7] M.A. Sanz, M.S. Tallman, F. Lo-Coco. Tricks of the trade for the appropriate management of newly diagnosed acute promyelocytic leukemia. Blood. 2005;105:3019-3025 Crossref
  • [8] R. Ohno, N. Asou, K. Ohnishi. Treatment of acute promyelocytic leukemia: strategy toward further increase of cure rate. Leukemia. 2003;17:1454-1463 Crossref
  • [9] British Committee for Standards in Haematology, D.W. Milligan, D. Grimwade, J.O. Cullis, et al. Guidelines on the management of acute myeloid leukaemia in adults. Br J Haematol. 2006;135:450-474 Crossref
  • [10] National Comprehensive Cancer Network. NCCN Clinical Practice Guidelines in Oncology: Acute Myeloid Leukemia, v. 2. (, 2012) http://www.nccn.org
  • [11] M.A. Sanz, D. Grimwade, M.S. Tallman, et al. Management of acute promyelocytic leukemia: recommendations from an expert panel on behalf of the European LeukemiaNet. Blood. 2009;113:1875-1891 Crossref
  • [12] J.K. Altman, A. Rademaker, E. Cull, et al. Administration of ATRA to newly diagnosed patients with acute promyelocytic leukemia is delayed contributing to early hemorrhagic death. Leuk Res. 2013;37:1004-1009 Crossref
  • [13] A. Rashidi, R.K. Goudar, F. Sayedian, et al. All-trans retinoic acid and early mortality in acute promyelocytic leukemia. Leuk Res. 2013;37:1391-1392 Crossref
  • [14] F.B. Taylor Jr., C.H. Toh, W.K. Hoots, H. Wada, M. Levi. Scientific Subcommittee on Disseminated Intravascular Coagulation (DIC) of the International Society on Thrombosis and Haemostasis (ISTH). Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost. 2001;86:1327-1330
  • [15] M.A. Sanz, F. Lo Coco, G. Martín, et al. Definition of relapse risk and role of nonanthracycline drugs for consolidation in patients with acute promyelocytic leukemia: a joint study of the PETHEMA and GIMEMA cooperative groups. Blood. 2000;96:1247-1253
  • [16] D. Grimwade, K. Howe, S. Langabeer, et al. Establishing the presence of the t(15;17) in suspected acute promyelocytic leukaemia: cytogenetic, molecular and PML immunofluorescence assessment of patients entered into the M.R.C. ATRA trial. M.R.C. Adult Leukaemia Working Party. Br J Haematol. 1996;94:557-573
  • [17] J.S. McClellan, H.E. Kohrt, S. Coutre, et al. Treatment advances have not improved the early death rate in acute promyelocytic leukemia. Haematologica. 2012;97:133-136 Crossref
  • [18] J.H. Park, B. Qiao, K.S. Panageas, et al. Early death rate in acute promyelocytic leukemia remains high despite all-trans retinoic acid. Blood. 2011;118:1248-1254 Crossref
  • [19] S. Lehmann, A. Ravn, L. Carlsson, et al. Continuing high early death rate in acute promyelocytic leukemia: a population-based report from the Swedish Adult Acute Leukemia Registry. Leukemia. 2011;25:1128-1134 Crossref
  • [20] C. Cordonnier, J.P. Vernant, B. Brun, et al. Acute promyelocytic leukemia in 57 previously untreated patients. Cancer. 1985;55:18-25 Crossref
  • [21] E. Di Bona, G. Avvisati, G. Castaman, et al. Early haemorrhagic morbidity and mortality during remission induction with or without all-trans retinoic acid in acute promyelocytic leukaemia. Br J Haematol. 2000;108:689-695 Crossref
  • [22] R.H. Jácomo, R.A. Melo, F.R. Souto, et al. Clinical features and outcomes of 134 Brazilians with acute promyelocytic leukemia who received ATRA and anthracyclines. Haematologica. 2007;92:1431-1432
  • [23] E.M. Rego, H.T. Kim, G.J. Ruiz-Argüelles, et al. Improving acute promyelocytic leukemia (APL) outcome in developing countries through networking, results of the International Consortium on APL. Blood. 2013;121:1935-1943 Crossref

Footnotes

a Department of Internal Medicine, Eastern Virginia Medical School, Norfolk, VA, United States

b Department of Pathology, Penn State Milton S. Hershey Medical Center, Hershey, PA, United States

c Department of Pathology, University of Virginia, Charlottesville, VA, United States

d Department of Pathology, West Virginia University, Morgantown, WV, United States

e Virginia Oncology Associates, Norfolk, VA, United States

f Pathology Sciences Medical Group/Sentara Laboratory Services, Norfolk, VA, United States

lowast Corresponding author at: Division of Oncology, Washington University School of Medicine, 660 S. Euclid Ave., Campus Box 8056, St. Louis, MO 63110, United States. Tel.: +1 314 747 8479.