Clinical networking results in continuous improvement of the outcome of patients with acute promyelocytic leukemia Free

Acute promyelocytic leukemia (APL) was once considered 1 of the most life-threatening subtypes of acute myeloid leukemia (AML).¹ However, since the 1990s, clinical trials based on the administration of all-trans retinoic acid (ATRA) with anthracycline-based chemotherapy for the treatment of newly diagnosed APL reported complete remission (CR) rates of 90% to 95% and cure rates of >80%.^2-7 Nevertheless, reports from real-world databases have indicated that outcomes are considerably worse than those in clinical trials, and this is especially true for countries with lower human development indexes in which costs, limited infrastructure, and a paucity of trained health care professionals strongly affect outcomes.^8-11 Indeed, Jácomo et al showed a death rate of 32% during induction treatment and a 2-year overall survival (OS) rate of <60% among Brazilian patients with APL treated between 2003 and 2006.¹² Similar outcomes were observed in other Latin American countries.¹¹ To improve treatment outcomes and to foster clinical and investigational collaborative efforts among the region and with well-recognized groups of the United States and Europe, the International Consortium on Acute Promyelocytic Leukemia was created in 2004 by the International Members Committee of the American Society of Hematology.¹³ Later, the group was renamed the International Consortium on Acute Leukemia (ICAL), with its activities expanded to include other types of AML. One of the basic concepts of the ICAL is working with national networks rather than with multiple independent institutions, and protocols and manuals for the diagnosis, management, supportive care, and specific treatment were adapted to the local or regional capacity.¹⁴ 

Since its conception, ICAL has crafted educational material in collaboration with local experts (see supplemental Material, available on the Blood website) and fostered educational activities on daiagnosis, risk assessment, and management of APL through special sessions at national meetings and regional workshops, and, through the American Society of Hematology’s International Programs, has supported the training of young investigators, clinicians, and laboratory human resources in American and European institutions as well as in the participating countries.¹⁴ The national central laboratories of ICAL were enrolled in an external quality control program (United Kingdom External Quality Assessment Services) and samples have been exchanged among laboratories to assure reproducibility of results. The collaboration among laboratories also allowed the investigation of genetic factors associated with prognosis and pathogenesis of the disease.^15-17 In addition to training and expanding local infrastructure, ICAL supported members in their plea to local authorities and pharmaceutical companies to ensure drug availability and reinforce awareness about the disease. Here, we describe the laboratory and clinical features of the largest cohort of patients with APL studied in Latin America and report the long-term outcomes of the International Consortium on Acute Promyelocytic Leukemia (ICAPL) study.

To participate in the ICAPL, countries had to fulfill minimum requirements. These included the availability of the antileukemic drugs specified in the ICAPL study protocol, transfusion medicine services enabling the adequate availability of platelets and other blood components, hematology laboratory facilities capable of performing basic fluorescence microscopy, ability to report data and participate in meetings using web-based tools, and identification of 1 national coordinator and 1 responsible for molecular studies. Thus, 5 Latin American countries were selected: Brazil, Chile, Paraguay, Peru, and Uruguay.

The definitions of complete hematological response (CHR), molecular remission, molecular resistance, and molecular relapse were as previously reported¹³ and are aligned with the European LeukemiaNet 2019 recommendations.¹⁸ The death during induction was defined as death due to disease or treatment complications occurring from the first dose of ATRA up to the first day of the consolidation. Differentiation syndrome (DS) was diagnosed according to the presence of the following signs or symptoms: unexplained fever, dyspnea, pleural and/or pericardial effusion, pulmonary infiltrates, renal failure, hypotension, and unexplained weight gain of >5 kg.¹⁹ The incidence of therapy-related AML (t-AML) was determined by considering only patients who completed consolidation treatment and were adequately followed-up as the at-risk population.

Patients aged 15 to 75 years, in whom the diagnosis of APL was suspected, underwent bone marrow (BM) aspiration. ATRA treatment was initiated immediately in all cases in which the diagnosis of APL was suspected based on morphology. Cytogenetic analysis for t(15;17) and/or polymerase chain reaction (PCR) analyses for rearrangements between the promyelocytic leukemia (PML) and retinoic acid receptor-α (RARA) genes (PML/RARA rearrangements) were performed on BM samples; the detection of 1 or the other was required for confirmation of the genetic diagnosis of APL. Patients with de novo APL, normal hepatic and renal function, no cardiac contraindications to anthracycline chemotherapy, and an Eastern Cooperative Oncology Group (ECOG) performance status score of <4 were eligible. Informed consent was obtained from all patients. According to the Declaration of Helsinki, the research ethics board of each participating hospital approved the protocol.

Patients who died before starting ATRA were not included (death before treatment). Other ineligibility criteria were pregnancy; previous chemotherapy or radiotherapy and chronic infections such as HIV, hepatitis C virus, and hepatitis B virus infection; impairment of renal or liver function; or left ventricle ejection fraction of <50%. In the Chilean cohort, the value of creatinine was not captured in the database, but renal function was categorized qualitatively as either normal or abnormal by the clinicians based on this parameter.

Induction therapy was adopted from the Programa Español de Tratamiento en Hematologia/Dutch-Belgian Hemato-Oncology Cooperative Group Leucemia Promielocítica Aguda (PETHEMA/HOVON LPA) 2005 trial,⁵ except that idarubicin (12 mg/m² per day) was replaced by daunorubicin (60 mg/m² per day) because daunorubicin was more readily available and less expensive in the participating countries. Oral ATRA (45 mg/m² per day) was administered in 2 daily divided doses until CHR was achieved. Daunorubicin was given as an intravenous bolus on days 2, 4, 6, and 8, except for patients aged >70 years who received only the 3 first doses of daunorubicin. For patients aged <20 years, the ATRA dose was adjusted to 25 mg/m² per day.

Patients who achieved CHR received 3 monthly consolidation courses with ATRA and anthracycline-based chemotherapy that was based on the PETHEMA/HOVON LPA2005 trial.⁵ Consolidation therapy followed a risk-adapted strategy.⁶ For low-risk patients, ATRA was combined with daunorubicin (25 mg/m² on days 1-4) in the first cycle, mitoxantrone (10 mg/m² on days 1-3) in the second cycle, and daunorubicin (60 mg/m² on day 1) in the third cycle. For intermediate-risk patients, consolidation was intensified by increasing the dose of daunorubicin to 35 mg/m² in the first cycle and by repeating the infusion of 60 mg/m² on the second day of the third cycle. For high-risk patients, consolidation was further intensified by adding cytarabine in cycles 1 (1000 mg/m² IV over 6 hours on days 1-4) and 3 (150 mg/m² every 8 hours on days 1-4). In addition, mitoxantrone (10 mg/m² per day) was administered for 5 days in the second cycle; anthracycline dosages in cycles 1 and 3 were as described for low-risk patients. High-risk patients aged >60 years did not receive cytarabine and were treated as intermediate-risk patients. Maintenance therapy was identical to that used in the LPA2005 trial and it was given to patients in molecular remission. Patients received ATRA for 2 years (45 mg/m² per day divided into 2 daily doses for 2 weeks, every 3 months) combined with intramuscular or oral methotrexate (15 mg/m² per week) and oral mercaptopurine (50 mg/m² per day) during the ATRA pause. Central nervous system (CNS) prophylaxis was not given.

Treatment response was confirmed by morphological analysis and by qualitative reverse transcription PCR (RT-PCR) at the end of the 3 consolidation cycles. Minimal residual disease was monitored at national reference laboratories by RT-PCR for PML/RARA in BM samples obtained every 3 months during maintenance and for 2 years after completion of treatment. RT-PCR assays were performed with a sensitivity level of 1 leukemic cell in 10 000 cells (10⁻⁴).¹³ 

The European LeukemiaNet 2019 guidelines were followed for supportive measures for patients with APL.¹⁸ Platelets were transfused to maintain the platelet count >30 000 per μL to 50 000 per μL, and cryoprecipitate was administered to maintain the fibrinogen level above 150 mg/dL. Treatment with dexamethasone at a dose of 10 mg twice daily by IV injection was started at the earliest symptom or sign of DS and maintained until the resolution of the syndrome. ATRA was temporarily discontinued only in the setting of severe DS that led to respiratory or renal dysfunction requiring admission to the intensive care unit. It is important to stress that ATRA treatment was initiated based on the suspicion of APL before genetic confirmation. To allow this early administration in ICAPL, ATRA was available in satellite pharmacies in the emergency rooms of participating institutions. Febrile neutropenia and infections were treated by local protocols.

Primary end points of interest included CHR, OS, disease-free survival (DFS), and rate of death during induction. Frequency of CHR was reported descriptively and compared using a χ² test. The association between baseline characteristics and the likelihood of achieving CHR was assessed using univariable and multivariable logistic regression analysis. OS was defined as the time from diagnosis to death from any cause; those alive or lost to follow-up were censored at the date last known alive. DFS was defined as the time from CHR to disease relapse or death from any cause, whichever occurred first. Patients who were alive without disease relapse were censored at the time last seen alive and disease free. OS and DFS were estimated using the Kaplan-Meier method and compared using the log-rank test. Cumulative incidence of nonrelapse mortality (NRM) and relapse were constructed, reflecting time to relapse and time to NRM as competing risks, respectively. Cumulative incidence of NRM and relapse were compared using the Gray test.²⁰ Univariable and multivariable proportional hazards (PH) regression analysis for OS and DFS, and Fine and Gray regression analysis²¹ for NRM and relapse were performed for potential prognostic factors. Factors examined in regression analysis included age at diagnosis, patient sex, white blood cell (WBC) counts, platelet counts, hemoglobin levels, PML breakpoint, creatinine, albumin, uric acid, fibrinogen, BM blasts, relapse risk, and morphology. Factors with P value <.1 from univariable analysis were included in the multivariable analysis. Before regression analysis, collinearity among characteristics was assessed using correlation analysis and unsupervised hierarchical clustering analysis. Before PH modeling, PH assumption for each variable of interest was tested. Linearity assumption for all continuous variables was examined in both logistic and PH models using restricted cubic spline estimates of the relationship between the continuous variable and log relative hazard/risk.²² The country effect was assessed using a frailty model. All P values are 2 sided at the significance level of .05. Multiple comparisons were not considered. All calculations were performed using SAS 9.4 (SAS Institute, Cary, NC) and R 3.5.2 (The CRAN project).

The research ethics board of each participating hospital approved the protocol.

Based on the cytomorphological analysis of peripheral blood and BM aspirate smears, 1004 patients with a suspected diagnosis of APL were enrolled in the ICAPL registry between March 2005 and March 2020. Of 1004 patients, 806 patients had the diagnosis of APL confirmed genetically and were considered eligible. Of 198 patients excluded, 71 (35.9%) were enrolled on different protocols, 38 (19.2%) had insufficient genetic confirmation, 21 (10.6%) were aged <15 years or >75 years, and 3 (1.5%) had died before receiving the first dose of ATRA (supplemental Table 1).

The main clinical and laboratory characteristics of the 806 eligible patients are shown in Table 1. The difference in the number of patients enrolled in each country reflects the size of the population and accrual period. Only 87 patients (10.8%) were classified as being at low risk of relapse. Bleeding was the most frequently reported symptom at diagnosis, which occurred in the genitourinary (GU), gastrointestinal (GI), and CNS in 99 (12.3%), 52 (6.5%), and 65 patients (8.1%), respectively. Thrombosis was diagnosed in 36 (4.5%) patients (Table 1). Significant differences in clinical and laboratory features according to nationality were apparent regarding ECOG performance status scores, relapse-risk groups, blast morphologic subtypes, percentage of blasts infiltrating the BM, and incidence of thrombosis, bleeding in the GI, GU, and respiratory tract, but not in the CNS. Patients from Paraguay presented with lower ECOG performance status scores and WBC counts; consequently, fewer were assigned to the high-risk group. In the cohorts from Paraguay and Uruguay, the frequencies of the bcr3 PML/RARA isoform were less than in the remaining countries. Bleeding in the GI, GU, and respiratory tract was less frequently reported in the cohort from Chile, and thrombosis was more frequently diagnosed in the Brazilian cohort (Table 1).

Information about the time interval between symptom onset and diagnosis was available for 732 patients (Table 1). Notably, the diagnosis was established >10 days from symptom onset in 44.7% of cases. Nevertheless, most patients (84.1%) in Chile were diagnosed within 3 days of presenting symptoms, whereas the time interval in other countries was ≥4 days for >75% of the cases.

Considering the entire cohort, the CHR rate was 85.4% (687/804 patients; Table 2). In 2 patients, remission could not be assessed because of the loss of follow-up and cardiotoxicity causing treatment discontinuation. There were no significant differences in CHR rates nor in the induction death rates among countries (because of the small sample size, Uruguay was excluded from the comparison; Table 2). Overall, 117 patients (14.6%) died during induction treatment, and the causes of death are shown in Table 3.

DS was diagnosed in 251 of 774 patients (32.4%), and its frequency was higher in Brazil (41.5%) than in the other 4 countries (Table 2). Data to evaluate DS severity were not available. The incidence of liver or kidney toxicity in grades 3 or 4 during induction was 7.7% and 5%, respectively. The frequency of pseudotumor cerebri was 2.3%, and the frequency of hemorrhage grade ≥2 was 19.6% (Table 2).

The multivariable logistic regression analysis for induction death revealed that age of ≥40 years, ECOG performance status score of 3, high-risk of relapse category, bcr3 PML/RARA isoform, a time interval from symptom onset to diagnosis of >48 hours, and CNS and respiratory tract bleeding were significantly associated with higher risks of death during induction (Table 4).

The induction death rates in Brazil, Chile, and Peru were pooled for the analysis of the temporal changes. The 3 countries were selected because of their substantial accrual, and the period from 2008 to 2019 was divided into 3-year intervals to minimize the effect of a small number of events. Little change was observed over the 12-year period, with induction death rates ranging from 17.1% for the period 2008 to 2010, to 14.1% for the period 2017 to 2019 (supplemental Figure 1). The 3 countries exhibited a similar pattern of temporal changes (data not shown).

Of 687 patients who achieved CHR, 650 received all courses of chemotherapy planned as consolidation, and 646 tested RT-PCR negative for PML/RARA at the end of consolidation. There were 4 patients with molecular persistence (0.6%) who were censored at the time point and subsequentially received treatment in alignment with the institutional protocol. Twenty-three (3.3%) patients abandoned protocol treatment during consolidation, and 1 is still undergoing consolidation treatment. Thirteen patients died during consolidation, and the main cause of death was infection (12 patients), with 1 death due to cardiac failure (Table 3). There were no differences in the response and reported toxicities during consolidation among the participating countries (data not shown).

All 646 patients achieving complete molecular response proceeded to maintenance therapy. Of 646 patients, 22 were lost to follow-up, 3 were censored (1 because of pregnancy, 1 presented a refractory pseudotumor cerebri, and 1 because of severe hepatotoxicity), and 62 were still undergoing maintenance treatment at the time of analysis. There were 11 deaths during maintenance and their causes are shown in Table 3. Four patients developed t-AML after completion of consolidation (Table 3).

Overall, 94 patients relapsed: 51 during maintenance and 43 in follow-up. There were 47 hematological relapses (2 of them concomitant with CNS relapse, and 1 of them concomitant with testicular granulocytic sarcoma), 42 molecular relapses (3 of them concomitant with CNS relapse; and 2 of them concomitant with granulocytic sarcoma, 1 testicular and 1 cranial), 4 CNS relapses, and 1 unknown. There was no correlation regarding risk categories and relapse (P = .99). In a multivariable regression analysis, CNS bleeding (P = .024) and time to diagnosis of >4 days (P = .043) demonstrated a significant association with relapse.

The median follow-up of the surviving patients is 53 months (range, 0.2-157.6) from diagnosis. Four-year OS from time of diagnosis for the entire cohort was 81% (95% confidence interval [95% CI], 748-84), 4-year DFS from the time of CHR was 80% (95% CI, 77-83), 4-year cumulative incidence of NRM from the time of CHR was 4.8% (95% CI, 3.3-6.7), and the 4-year cumulative incidence of relapse (CIR) from the time of CHR was 15% (95% CI, 12-18; Table 5).

Figure 1 shows the outcomes according to risk group. OS was significantly lower for high-risk patients (4-year OS: 92% for low risk, 89% for intermediate risk, and 68% for high risk; P < .0001) and NRM was significantly higher for high-risk patients (4-year cumulative incidence of NRM: 2.7% for low risk, 2.8% for intermediate risk, and 7.8% for high risk; P = .014), but no differences were noted for DFS and CIR. The main determinant of the worse outcome in the higher-risk APL subsets was a greater death rate during induction.

The multivariable regression analysis for survival outcomes identified M3v morphological subtype, a period between symptom onset and diagnosis of ≥4 days, and lower values of hemoglobin as associated with a higher risk of NRM, whereas only the detection of bleeding in CNS was associated with a higher risk of relapse (supplemental Table 3).

The present report represents a unique demonstration of the impact of an international platform based on medical and scientific exchange and expert guidance according to standardized therapeutic management and diagnostic procedures on disease outcomes in 5 countries: Brazil, Chile, Paraguay, Peru, and Uruguay. APL was chosen as an ideal disease to evaluate this platform because it is rapidly fatal if left untreated yet highly curable if access to rapid diagnosis and appropriate and specialized care is available. In addition, therapy is based on relatively inexpensive drugs. Before the establishment of the ICAL, the mortality rate during induction was 32%, and the 2-year OS was 52% in Brazil¹²; and in Chile, the 5-year event-free survival was 64%.²³ Data were not available for the remaining countries participating in the ICAL. The relative decrease of ∼50% in the induction death rate reported in the first interim analysis¹³ persisted through a 12-year span. The long-term OS and DFS in the ICAPL study were 81% and 80%, respectively, and are similar to that reported by the PETHEMA LPA99,⁶ PETHEMA/HOVON LPA2005,⁵ and GIMEMA/AIDA2000 trials³ (Table 5), which were conducted in high-resource settings. Nevertheless, the induction death rate of the present study was higher than those of the aforementioned trials, suggesting that additional efforts to increase awareness about APL diagnosis among general practitioners and accelerate the transfer of patients from primary care to hospitals with hematology services should be considered. In approximately one-third of the cases only, the period between the onset of symptoms and APL diagnosis was <72 hours, and this variable was associated with the outcome (Table 4). Real-world data on APL show higher induction death rates compared with that of clinical trials both in high-income and low- and middle-income countries,^9-11,24-26 and the figures reported here may reflect the nature of the ICAPL study, which included all APL cases that received at least 1 dose of ATRA. In addition, in our cohort, the frequency of patients with leukocyte counts of >10 000 per μL and those with ECOG performance status scores of 3 was greater than in the European and American trials.^3,5,6,27 

Bleeding remained the main cause of death in induction in our analysis. In particular, the occurrence of CNS and pulmonary bleeding was identified as a significant risk factor for death during induction, thus reinforcing the relevance of considering APL as a medical emergency requiring prompt administration of ATRA and aggressive transfusions to control APL coagulopathy. In a retrospective analysis of 204 patients with APL treated in 4 centers in the United States, the early death rate (considered to be death occurring within the first 30 days of presentation) was 11%, and hemorrhage accounted for 61% of those.²⁸ The authors showed that the percentage of early death secondary to bleeding increased from 33% to 70% in the group of patients to whom ATRA was given ≥1 days after APL was suspected, in comparison with those receiving the drug on the first day.²⁸ Jillella et al developed a strategy to reduce early death and facilitate prompt start of ATRA that was focused on oncologists working in community centers in Georgia, North Carolina, and Florida. After training professionals and establishing a regular channel of communication between expert and treating physician, the authors reported early death rate of 8.1% and 9.1% for patients treated in community and academic centers, respectively.²⁹ 

Osterroos et al presented a scoring system for identifying patients with APL at high risk of early death. WBC count, platelet count, and age were identified as the most significant variables for early death prediction.³⁰ The score identified low-, high- and very high-risk patients with early death risks of 4.8%, 20.2%, and 50.9%, respectively, in the training cohort, and with 6.7%, 25.0%, and 36.0%, respectively, as corresponding values for the validation cohort.³⁰ Mantha et al analyzed 995 patients enrolled in 5 trials conducted by 4 cooperative groups.³¹ Their research further corroborated that high WBC count serves as the main predictor of early hemorrhagic death among patients with APL.³¹ The relevance of WBC count at diagnosis in identifying high-risk patients was also demonstrated previously in a mathematical model based on an ecological paradigm to describe the first 30 days of the induction, which could capture the dynamics of leukemic and normal leukocytes in peripheral blood.³² 

Delving into country-specific nuances, it is imperative to recognize that although all participating countries are from Latin America, this does not imply a homogeneous population. The existing health care disparities in the region may have influenced disease characteristics and outcomes. For instance, patients in Peru often faced a prolonged time from symptom onset to diagnosis (in 74.4% of the patients this period was >10 days), which may have contributed to a higher frequency of patients with poorer ECOG performance status scores (13.5% with ECOG performance status scores of 3). Patients from Paraguay and Uruguay presented more favorable features than those from Peru. Therefore, the country effect was, to some extent, reflected via baseline characteristics in regression analysis. However, the stratified regression analysis adjusted for the country effect, in which data from Uruguay and Paraguay were combined because of the small size of cohorts, showed no significant differences in the outcomes.

The ICAL protocol was identical to that adopted in the PETHEMA/HOVON LPA 2005 trial, except that idarubicin was replaced by daunorubicin. In a previous report, 350 patients from the PETHEMA/HOVON APL cohort were matched with 175 patients in the ICAPL cohort.³³ Lower overall and event-free survivals were observed in the ICAPL cohort, which was mainly because of a higher death rate during induction, but patients who achieved CR had comparable CIR and DFS rates.³³ Recently, Jaime-Pérez et al conducted a retrospective analysis of 61 patients with APL treated with ATRA plus mitoxantrone or doxorubicin.³⁴ No significant differences were found in CR rates, survival, hospitalization days, or frequency of adverse events. Taken together, these data suggest that there are no significant differences in the efficacy and safety of different anthracyclines in the treatment of APL.

The 4-year OS rates for patients in the low- and intermediate-risk groups were ∼90% and significantly higher than in the high-risk group, whereas no difference was detected in DFS and CIR values, indicating that the worse outcome was a consequence of higher mortality during induction and higher NRM rates. Infectious complications were the major cause of death during consolidation, therefore the treatment intensification according to risk improved the CIR and DFS in the high-risk group at the cost of increasing the risk of infection. In the ICAPL study protocol, there was a general recommendation for the management and prophylaxis of infections in patients with neutropenia, but because of challenges such as diverse microbial profiles, varying levels of health care infrastructure, and drug availability in the 5 countries, the guidelines were adapted according to local (hospital-based) protocols. More recently, the publication of guidelines by the national societies helped to standardize the management and prophylaxis in the region.^35,36 

Four patients developed t-AML, corresponding to 0.65% of the patients who completed consolidation. Sobas et al³⁷ reported 58 t-AML/MDS cases among 2670 patients with APL who were treated according to PETHEMA “chemotherapy based” (APL1996, 1999, 2005, 2012, and 2017 trials for high-risk patients) and “chemotherapy-free” regimen (2017 for low/intermediate risk) between 1996 and 2021. Of note, in the APL2005 study, which was a twin of ICAPL, the authors reported 29 cases of t-AML/MDS among 1080 patients (2.7%). The lower frequency reported here probably reflects the fact that we did not include t-MDS cases because of insufficient information for a definitive diagnosis.

Currently, the majority of patients with APL diagnosed in the United States and Europe are treated with chemotherapy-free regimens in which arsenic trioxide is associated with ATRA.^38-40 The practice is based on the excellent results obtained in the Italian-German APL0406,⁴¹ the Australasian Leukaemia and Lymphoma Group APML4,⁴² National Cancer Research Institute Acute Myeloid Leukaemia Working Group AML17,⁴³ and North American Leukemia Intergroup Study C9710²⁷ trials. However, arsenic trioxide is not available for the treatment of newly diagnosed APL in most countries of Latin America and in many low/middle income countries worldwide because of cost and/or unavailability in the market.^11,44,45 It is conceivable that a chemotherapy-free regimen would further improve the OS and NRM by reducing the deaths due to infections, which represented 34.7% of all deaths in this study and occurred both before and after CR was achieved. Moreover, it may also improve the outcome by decreasing CIR. Therefore, the ICAL is committed to evaluating in future studies the effect of a chemotherapy-free regimen in newly diagnosed patients with APL treated in middle-income countries.

The authors thank Ana Silvia G. Lima for her technical support.

This study was supported by the ASH Foundation and by a grant from Fundação de Apoio à Pesquisa do Estado de São Paulo, Brazil (grant no. 2013/08135-2).

Contribution: L.C.d.A.K. acquired and interpreted data; H.T.K. performed the statistical analyses; M.S.U., J.R.N.-C., V.S., P. Muxi, and E.M.R. were the national coordinators of ICAL in Chile, Peru, Paraguay, Uruguay, and Brazil, respectively, and contributed to the development of the clinical network, the oversight of laboratory and management procedures, and data acquisition and interpretation; R.A.M.M., A.B.G., K.P., E.C.N., R.I.B., N.R., S.Q., A.A.-L., A.C.O., L.F.-P., F.T., F.M., A.P.M.-A., E.M.F., B.K.L.D., P.O., and J.U. were responsible for centers performing the diagnosis and treatment of patients and acquired data; M.T., R.R., A.G., R.D., P.J.M.V., M.S., B.L., N.B., H.T.K., and E.M.R. contributed to the conception and design of the work, interpreted data, and drafted the manuscript; and all authors have critically reviewed and approved the manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Eduardo M. Rego, Laboratory of Medical Investigation 31 (LIM-31), Faculdade de Medicina, University of São Paulo, Av Enéas Carvalho de Aguiar 255, Hemocentro, Prédio dos Ambulatórios, 1°Andar, CEP 05403-000, São Paulo, Brazil; email: eduardo.rego@fm.usp.br.