Management of hemostatic complications in acute leukemia: guidance from the SSC of the ISTH

Abstract

Patients with acute leukemia frequently develop thrombocytopenia and hemostatic complications caused by coagulopathy. Coagulopathy complicates the management of these patients and can lead to significant morbidity and mortality. This guidance document aims to review and provide guidance on the management of hemostatic complications in adult patients with acute leukemia, addressing four main issues, including platelet transfusion, disseminated intravascular coagulation, L-asparaginase-related hypofibrinogenemia, and the use of antifibrinolytic agents.

Scope and methodology

Patients with acute leukemia commonly develop thrombocytopenia and hemostatic complications from factors that include coagulopathy due to the underlying disease, anti-neoplastic therapy, and potential complications (e.g., sepsis) that contribute to morbidity and mortality. This guidance document represents a collaboration between two Scientific and Standardization Committees (SSC) of the International Society on Thrombosis and Haemostasis (ISTH), the Hemostasis & Malignancy Subcommittee and the Perioperative & Critical Care Thrombosis and Hemostasis Subcommittee, with the objective to provide clinical guidance on the management of hemostatic complications in patients with acute leukemia. Recommendations reflect strong guidance statements supported by high-quality evidence from clinical trials. Suggestions reflect weaker guidance statements based on low-quality evidence or expert opinion. Given the complexity of coagulopathy in this population, we will focus on four main issues related to hemostatic complications in adult patients with acute leukemia, including platelet transfusion, disseminated intravascular coagulation (DIC), L-asparaginase-related hypofibrinogenemia, and the use of antifibrinolytic agents. For each of the topics, guidance statements were generated following review of published literature and are based on consensus opinion of the authors. Thrombotic complications applicable to acute leukemia have been addressed in previous SSC guidance documents that include the management of thrombosis in cancer patients with thrombocytopenia [1] and those associated with L-asparaginase [2].

Definitions

1. Hemostatic complications. The target conditions are acquired complications related to hemostasis due to the underlying acute leukemia or its treatment.

2. Acute leukemia. This guidance document includes reference to acute myeloid leukemia (AML) and acute lymphoblastic leukemia (ALL).

Background

Hemostatic complications due to coagulopathy is one of the leading causes of mortality in patients with acute leukemia [3]. Routine coagulation tests may reveal DIC with prolongation of prothrombin time (PT), variable activated partial thromboplastin time (aPTT), low fibrinogen levels and platelet counts, and elevated D-dimer and fibrin-degradation products. Majority of patients with acute promyelocytic leukemia (APL) (70–80%) develop DIC during the course of diagnosis and treatment (Table 1). Although modern APL therapeutic regimens incorporating all-trans retinoic acid (ATRA) or arsenic trioxide (ATO) are associated with high likelihood of long-term remission and disease-free survival, hemorrhage remains the most common cause of death during induction therapy [3]. Similarly, DIC has been reported in 10–20% of non-APL acute leukemia on diagnosis, with an even higher risk (60–70%) after the initiation of induction chemotherapy (Table 1). This guidance document addresses the management of hemostatic complications in adult patients with acute leukemia, encompassing issues related to platelet transfusion, DIC, L-asparaginase-related hypofibrinogenemia, and the use of antifibrinolytic agents.

Table 1.

Summary of studies on disseminated intravascular coagulopathy in acute leukemia

Study	N	Type of leukemia	DIC diagnostic criteria	Incidence of DIC (%)	Transfusion threshold/other tx	Outcomes
Tornebohm et al. 199224 Retrospective	57	All acute leukemia	Clinician discretion	Overall: 12.3 AML: 10.6 ALL: 20	Plt < 10	Significantly lower protein C and antiplasmin levels in DIC patients.
Sarris A. et al. 199225 Retrospective	153	ALL	Fibrinogen ≤ 1.6 g/L, prolonged PT with rapidly decreasing fibrinogen and/or positive FSP	Presentation: 12 Induction: 78	Plt < 20, FFP low dose heparin may be added	34% with DIC had hemorrhagic or thrombotic complications
Tornebohm et al. 199326 Retrospective	259	All acute leukemia	Clinician discretion	7	Plt < 10, Hgb < 8 mg/dL, FFP if bleeding and suspect DIC	Hemorrhagic mortality: 23%, leukocyte > 20 and plt <20 are risk factors for hemorrhagic complications
Nur et al. 199527 Prospective	104	All acute leukemia	Any of the 2: fibrinogen < 1.5 g/L, DFP > 10 μg/ml, prolonged PT/PTT/TT	Overall: 13.4 AML: 18.4 ALL: 9.1	Not reported	In AML, M3 and M5 are associated with DIC, AML with DIC has higher leukocyte count
Sletnes et al. 199528 Retrospective	71	All acute leukemia	Fibrinogen ≤ 1.6 g/L or rapid filling fibrinogen and elevated d-dimer or a positive ethanol gelation test	Overall: 20 AML: 17.5 ALL: 33	Plt ≤ 20, anemia or clinical bleeding Most also received low dose heparin and tranexamic acid	DIC patients had increased risk of WHO grade ≥2 bleed but no hemorrhagic fatalities
Sarris A. et al. 199629 Retrospective	125	ALL	Fibrinogen < 2 g/L with elevated FDP	Presentation: 10 Induction: 67	FFP or cryopricipitate until fibrinogen is normal, plt transfusion for plt <20	Patients with fibrinogen < 1 g/L had significantly higher risk of bleeding or thrombotic complications
Chojnowski et al. 199930 Retrospective	70	All acute leukemia	Any of the 2: fibrinogen ≤ 1.5 g/L, DFP > 10 μg/ml, prolonged PT (PT ratio <70%)	Overall: 18.6 AML: 16.3 ALL: 23.8	Not reported	83% of patients had coagulation activation by laboratory finding on diagnosis, which did not predict remission.
Higuchi et al. 200531 Retrospective	51	ALL	JMHW criteria	Presentation: 22 Induction: 42	Not reported	Hemorrhagic symptoms are more frequent in DIC patients compared to patients without DIC
Yanada et al. 200622 Retrospective	125	AML	Expert opinion	29	plt < 30 or fibrinogen < 1 g/L	FDP is the only factor associated with DIC (cut-off of 15 ug/mL)
Dixit et al. 200732 Retrospective	67	All acute leukemia	ISTH DIC score ≥ 5	Overall: 14.9 AML: 16.6 ALL: 13.9	FFP, no specified threshold	The presence of DIC is associated with severe bleeding in AML but not ALL
Uchiumi et al. 200733 Retrospective	161	AML (non-APL)	JMHW criteria (with thrombocytopenia) ≥ 4	32	Not reported	Factors associated with DIC: high CRP, leukocyte count, negative expression of CD13 and HLA-DR, cytogenetics of normal of 11q23 DIC is not associated with CR, OS, DFS
Chang et al. 201234 Retrospective	116	APL	ISTH DIC score ≥ 5	77.6	Plt <20, FFP for prolonged PT/PTT, cryoprecipitate for fibrinogen < 1 g/L	DIC was not associated with bleeding. Predictors for bleeding: prolonged PT, PTT, and high WBC
Mitrovic et al. 201335 Retrospective	56	APL	ISTH DIC score ≥ 6	84	Plt <30, FFP for fibrinogen < 1 g/L	Hemorrhagic early death rate 14.3%, with ISTH DIC score ≥ 6 being the most significant predictor. Most significant predictor for bleeding: leukocytosis
Libourel et al. 201636 Prospective	404	AML (non-APL)	ISTH DIC score ≥ 5	Derivation: 8.5 Validation: 6.3	Not reported	ISTH DIC score ≥ 5 is a strong predictor of thrombosis but not major bleeding
Shahmarvand et al. 201737 Retrospective	149	All acute leukemia	ISTH DIC score ≥ 5	APL: 75 Others: 6.4	Not reported	Elevated d-dimer ≥ 19,000 ng/mL fibrinogen equivalent units is a sensitive indication for APL DIC is not an indicator of poor prognosis

Abbreviations: ALL - acute lymphocytic leukemia; AML - acute myeloid leukemia; APL- acute promyelocytic leukemia; CRP - c-reactive protein; CR - complete remission; DFS - disease free survival; DIC- disseminated intravascular coagulopathy; FDP - fibrin degradation product; FFP - fresh frozen plasma; ISTH - International Society on Thrombosis and Haemostasis; OS - overall survival; plt- platelet count; WHO - World Health Organization

Unit of leukocyte, platelet and blast count: x 10⁹/L

Platelet transfusion

Thrombocytopenia is common in patients with acute leukemia. Several randomized trials in patients with hematologic malignancy or chemotherapy-related thrombocytopenia have established that severe bleeding events were no more frequent in patients transfused prophylactically for a platelet count <10 × 10⁹/L compared with <20 or 30 × 10⁹/L [4–7]. Prophylactic platelet transfusion for a platelet count <10 × 10⁹/L prolonged the time to the first grade 2 or higher bleed (as defined by World Health Organization) and reduced the number of days of hemorrhage as compared to therapeutic platelet transfusion (transfusion only for symptoms). Nevertheless, a reduction in bleeding risk was not associated with a survival benefit [8, 9]. Furthermore, a benefit from prophylactic platelet transfusion is not evident in patients undergoing autologous stem cell transplant [10], and the dose of platelet transfusion does not appear to affect bleeding outcomes [11].

As patients with leukemia frequently receive platelet transfusions, platelet refractoriness, defined as the inability to produce a satisfactory increment in the post-transfusion platelet count, is common and challenging to manage. Etiologies for platelet refractoriness include non-immune factors often due to increased consumption that account for over 80% of the cases (e.g., splenomegaly, bleeding, fever, infection, DIC, and medications) [12], and immune-related factors (e.g., alloimmunization). In patients refractory to platelet transfusion, a strategy of therapeutic platelet transfusion for bleeding and prophylactic transfusion for periprocedural management may be more appropriate. Little evidence suggests a benefit from continuous platelet infusion, intravenous immunoglobulin, splenectomy, or plasma exchange [13].

Among the minority of patients in whom clinical factors are insufficient to explain refractoriness to platelet transfusion, the primary immune cause is alloimmunization to human leukocyte antigens (HLA). Alloimmunization results from previous pregnancy, blood transfusion, or organ transplantation. As platelets express only class I HLA-A and HLA-B, alloantibodies to these antigens are most commonly implicated in alloimmune refractoriness. The Trial to Reduce Alloimmunization to Platelets (TRAP) demonstrated the efficacy of leukoreduction in reducing both the risk of alloimmunization and the development of platelet refractoriness in patients with AML receiving induction chemotherapy [14]. Among patients who are HLA-alloimmunized, three approaches can be used to find superior compatible platelet products: 1) platelet crossmatching [15]; 2) HLA matching [16]; and 3) HLA avoidance [17]. Platelet crossmatching involves incubating patients’ serum with donor platelets to assess compatibility. HLA matching involves matching class I HLA-A and B antigens between the donor and patient. HLA avoidance requires the identification of the HLA alloantibodies made by the patient. The platelet products are then selected based upon the absence of antigens to which the patient has identified antibodies. The optimal strategy should be decided after consultation with local transfusion departments.

Guidance statements

1. We recommend empiric platelet transfusion, with one standard adult unit, for a daily platelet count < 10 × 10⁹/L in patients with acute leukemia without active bleeding or coagulopathy. In patients with active bleeding, a higher transfusion threshold should be considered such as 20 × 10⁹/L or greater as determined by clinical factors that include the degree of hemorrhage, site of hemorrhage, and responsiveness to platelet transfusion.

2. We recommend leukoreduction to reduce the risk of alloimmunization from platelet transfusion.

3. In patients who are refractory to platelet transfusion, we recommend consideration of both non-immune and immune etiologies.

4. In patients who are refractory to platelet transfusion and HLA-alloimmunized, we suggest one of the three approaches: 1) platelet crossmatching; 2) HLA matching; or 3) HLA avoidance.

Disseminated intravascular coagulopathy (DIC)

Definition and criteria

DIC is characterized by systemic intravascular activation of the coagulation system from various causes that can result in multiorgan failure, thrombosis, and/or excessive bleeding. The diagnosis of DIC is challenging due to the complex underlying medical conditions of leukemia that can lead to variable presentations. Three main diagnostic scores have been proposed, assigning 0–3 points to different levels of laboratory results, including platelet count, fibrin-related markers, fibrinogen, and prothrombin time (PT) (Supplemental Table 1) [18–20]. A prospective study showed no significant differences in the prognostic outcomes of DIC among these three scoring systems [21]. However, in acute leukemia, thrombocytopenia is highly prevalent due to bone marrow involvement by the disease or chemotherapy. The Japanese Ministry of Health and Welfare (JMHW) modified criteria for severe thrombocytopenia (Supplemental Table 1), which excluded platelet count or bleeding symptoms, showed superior sensitivity and negative predictive values than the ISTH criteria in the diagnosis of acute leukemia-associated DIC in one study [22]. More recently, another scoring system, the Chinese Society of Thrombosis and Hemostasis scoring system for DIC (CDSS) were proposed and validated prospectively in patients with and without hematological malignancy and found to have favorable diagnostic and prognostic utility, but it has not been widely utilized [23].

Incidence and risk factors

The incidence of DIC varies widely in published studies, although APL has the highest risk with a reported incidence of 70–80%. Table 1 summarizes studies of DIC in acute leukemia [22, 24–37]. The majority of the studies were small, single-center and retrospective, and often without a standardized protocol for evaluation of DIC. Various DIC criteria were utilized, and different types of leukemia were included, resulting in a highly variable incidence of DIC. In general, the incidence of DIC was higher in patients with AML (as compared to ALL) and during induction therapy (Table 1). Patients with DIC had more hemorrhagic or thrombotic complications as compared to those without DIC and overt DIC by ISTH criteria was shown to be a strong predictor of thrombosis but not major hemorrhage [36] (Table 1).

APL is associated with a particularly high risk of hemorrhagic complications, regardless of the presence of DIC, and bleeding remains to be the main cause of early mortality [3]. The main treatment goal in this population has been to reduce early mortality. Several retrospective studies investigated the incidence and risk factors for severe hemorrhage or hemorrhagic death (Supplemental Table 2) [3, 38–48]. The hemorrhagic death rate varied from 5 to 20%, with a severe bleeding rate of 30%. Intracranial bleeding was the cause of hemorrhagic death in a high percentage of patients (33–90% of cases, Supplemental Table 2). The most common risk factors for hemorrhage included a high white blood cell count, blast count, or coagulopathy.

Management

The mainstay in the management of DIC is the treatment of the underlying disease. In patients with acute leukemia, antineoplastic therapy is often accompanied by a rapid resolution of DIC. This is typically the case with APL, where DIC typically abates within one week following ATRA initiation [49]. Below we reviewed other treatment options for DIC in patients with acute leukemia.

Transfusion support

Leukemia therapy can exacerbate DIC and bleeding risks in the short term. Transfusion support is often required, but transfusion frequency and threshold of different blood products are based on expert consensus without randomized trials available to guide the transfusion regimens. Several guidelines recommend similar thresholds for patients with acute leukemia-associated DIC, including platelet transfusion for a platelet count < 20–30 × 10⁹/L in patients without bleeding and for a platelet count < 50 × 10⁹/L in patients with bleeding [50]. Fibrinogen levels < 1 g/L have been associated with hemorrhagic and/or thrombotic complications in patients with ALL and APL, and a transfusion goal to achieve a fibrinogen level > 1.5–2 g/L is widely used, particularly in patients with active bleeding [29, 40] (Table 1 and Supplemental Table 2). Comorbidities such as liver disease or vitamin K deficiency can further impact hemostasis in patients with acute leukemia and should be considered, to allow appropriate treatments such as vitamin K supplementation for its deficiency.

Antifibrinolytic agents

Antifibrinolytic agents are not recommended in patients with active DIC due to concern of exacerbation of fibrin deposition [51]. (The use of antifibrinolytic agents in acute leukemia without DIC is described below).

Other agents

There is only low-quality evidence (i.e. case series) describing the use of unfractionated heparin or recombinant activated factor VII (rFVIIa) for the treatment or prevention of DIC in patients with APL and should not be routinely considered [52, 53]. In patients with sepsis-associated DIC, sub-analyses of clinical trials suggest that antithrombin (AT) repletion may improve outcomes [54], but the utility of AT repletion in leukemia-associated DIC remains unknown. In Japan, recombinant soluble thrombomodulin (rTM) is approved for the treatment of DIC based on a phase III trial that demonstrated the efficacy and safety of rTM over heparin in patients with hematological malignancies [55].

Guidance statements

1. We recommend a high level of vigilance for DIC in patients with acute leukemia, with systemic evaluation for coagulopathy (including platelet counts and routine coagulation parameters such as PT, aPTT, fibrinogen) at least daily until normalization and utilization at least one of the standardized scores to diagnose DIC.

2. We recommend that the presence of DIC should not preclude or delay antineoplastic therapy in patients with acute leukemia such as ATRA or ATO in APL.

3. We recommend transfusion support in patients with DIC associated with acute leukemia.

   1. In non-APL patients, we suggest transfusion for platelet count < 20 × 10⁹/L in patients without active bleeding and higher platelet count for patients with clinically significant active bleeding.
   2. In APL patients, given the high risk of hemorrhagic mortality during early induction, we suggest a higher transfusion threshold of platelet count < 30 × 10⁹/L.
   3. In patients with acute leukemia-associated DIC, particularly APL, we suggest transfusion support for fibrinogen level <1.5 g/L using either fibrinogen concentrates or cryoprecipitate.
   4. We recommend supportive transfusion to be continued during induction therapy until normalization of clinical and laboratory signs of coagulopathy.

4. We suggest against routine use of heparin, LMWH, antifibrinolytic agents, or rFVIIa to treat DIC associated with acute leukemia.

Hypofibrinogenemia associated with L-asparaginase

L-asparaginase is commonly included as a crucial component of induction therapy for ALL but can cause multiple coagulation abnormalities. Depletion of asparagine impairs hepatic protein synthesis of AT, fibrinogen and other coagulation proteins, predisposing to hemorrhagic and thrombotic outcomes. The prevention and management of thrombotic outcomes associated with L-asparaginase, including the role of antithrombin, have been reviewed in a recent ISTH guidance document [2]. In that guidance document, it was suggested that AT concentrates be administered for levels < 50–60% to a target level of 80–120%. LMWH thromboprophylaxis was suggested throughout induction therapy with L-asparaginase in patients without severe thrombocytopenia (ie, platelet count < 30 × 10⁹/L) or concern for hemorrhage [2].

Studies have consistently shown a reduction of fibrinogen levels with L-asparaginase treatment [56, 57]. Hypofibrinogenemia can result in abnormal coagulation assays, but its effects on hemorrhagic or thrombotic outcomes are conflicting. One study of 91 pediatric patients with ALL on L-asparaginase revealed that a fibrinogen level < 0.5 g/L was a risk factor for thrombosis but not hemorrhage [58]. By contrast, a study that included 214 adult patients with ALL on L-asparaginase failed to show the correlation of venous thromboembolism (VTE) or hemorrhage with fibrinogen levels < 0.5 g/L [57]. Another recent study in 784 adult patients treated on the GRAALL-2005 protocol (containing L-asparaginase), 70 (9%) received fibrinogen concentrates for a fibrinogen level < 0.5 g/L per protocol, although none had grade 3–4 bleeding events despite severe hypofibrinogenemia. Fibrinogen concentrate administration was associated with a 2-fold increased risk of VTE compared to those who did not receive fibrinogen concentrates [59]. The currently available literature was limited in that most were retrospective analyses and highly susceptible to confounding. In addition, all reports adopted the practice of transfusing fibrinogen concentrates or cryoprecipitate for a fibrinogen level <0.5 or 1 g/L [57–59], so the risk of hemorrhage is unknown without transfusion.

Guidance statements

1. We suggest monitoring and repletion of AT following L-asparaginase therapy per ISTH Guidance [2].

2. In patients undergoing L-asparaginase treatment, we suggest replacement of fibrinogen for a level < 0.5 g/L. In patients with active bleeding, we suggest targeting a higher fibrinogen level.

Antifibrinolytic agents

Given the high risk of hemorrhage, particularly in the setting of thrombocytopenia in patients with acute leukemia, antifibrinolytic agents such as aminocaproic acid or tranexamic acid have been investigated as an adjunctive treatment modality. In large randomized trials, antifibrinolytic agents significantly reduced hemorrhage-associated mortality in post-partum and trauma patients and reduced bleeding in surgical patients, without an increased risk of thrombosis [60]. However, evidence supporting the use of antifibrinolytic agents in acute leukemia patients remains limited and is largely based on single-center retrospective studies and pre-1990 small randomized trials (Table 2) [3, 61–70]. Most studies were small, single-arm studies without a control group, employing various dosing strategies, and with highly variable outcome reporting (Table 2). The GIMEMA and PETHEMA studies both did not demonstrate a therapeutic benefit of antifibrinolytic agents to reduce hemorrhagic mortality in patients with APL [3, 71]. A 2016 Cochrane meta-analysis conclude that the available evidence of using antifibrinolytic agents in patients with hematological diseases is limited and further studies are required [72].

Table 2.

Summary of studies of antifibrinolytic use in patients with acute leukemia

Study	Type of study	N, total	N, acute leukemia	Type and dose of antifibrinolytic	Bleeding outcome/Rate of hemorrhage	Thrombotic outcomes/Rate of thrombosis
Gallardo et al. 198361	Prospective RCT	19	19 (AML and ALL)	EACA (vs placebo)	Major bleeding: EACA 15%, placebo 19%	0
Avvisati et al. 198962	Prospective RCT	12	12 (APL)	TA 2g IV q8 (vs placebo) x 6 days	TA significantly reduced the hemorrhagic score, RBC and platelet transfusion	0
Shpilberg et al. 199563	Prospective RCT	38	38 (AML)	TA 1g q6 (vs placebo)	TA significantly reduced bleeding events and platelet transfusion during consolidation but not induction	0
Ben-Bassat et al. 199064	Prospective cohort	54*	54 (AML)*	TA 1g PO/IV q6	4.6%	0
Brown et al. 200065	Retrospective cohort	31	31 (APL)	TA 1–2 g IV for 6 days (19 patients)	N/A	3 out of 4 early death showed wide spread thrombosis in the microvasculature
Kalmadi et al. 200666	Retrospective cohort	77	37 (AML and ALL)	EACA, median daily dose 6 g for median of 8 days	EACA significantly reduced platelet transfusion, reduced RBC transfusion by >50% in 62% patients	0
De la Serna et al. 20083	Prospective cohort	561	561 (APL)	TA IV 100 mg/kg per day	Hemorrhagic mortality 5%	6%
Wassenaar et al. 200867	Retrospective cohort	30	30 (APL)	EACA 6 g IV loading, then 1 g/hr continuous (17 patients)	57%	N/A
Antun et al. 201368	Retrospective cohort	44	15 (AML and ALL)	EACA 1 g twice daily for median of 47 days	16% major bleeding (25% minor bleeding)	0
Marshall et al. 201469	Retrospective cohort	54	24 (AML and ALL)	EACA, various dosing, median of 6 days	N/A	5.7%
Juhl et al. 201770	Retrospective cohort	109	76 (AML and ALL)	EACA, median daily dose 3.58 g for mean of 9.26 days	N/A	4.6%

^*131 induction or consolidation courses

Abbreviations: ALL - acute lymphoblastic leukemia; AML - acute myeloid leukemia; APL - acute promyelocytic leukemia; EACA - epsilon aminocaproic acid; g - gram; IV - intravenous; N/A - not available; RBC - red blood cells; RCT - randomized controlled trial; TA - tranexamic acid; tx - treatment

Note: In studies that included patients other than acute leukemia, the rates of hemorrhage and thrombosis are those for the entire study population

Acknowledging the need for better evidence, four ongoing randomized trials are investigating the utility of antifibrinolytic agents in patients with hematologic malignancies and thrombocytopenia (Supplemental Table 3). Two studies (TREATT and A-TREAT) aim to investigate the addition of tranexamic acid to a standard prophylactic platelet transfusion protocol, while the other two studies (PROBLEMA and PATH) target the use of antifibrinolytic agents to replace prophylactic platelet transfusion. These highly anticipated results could inform best practices for these challenging scenarios.

Guidance statement

1. In patients with hematological malignancies, in the absence of DIC and where clinical trial participation is not an option, we suggest consideration of antifibrinolytic agents as adjuvant therapy for patients with severe thrombocytopenia refractory to platelet transfusion.

Addendum

T-F. Wang contributed to the writing of the manuscript, development and review of guidance document, and final approval of the submitted version. R.S. Makar contributed to the development and review of the guidance document, writing of the manuscript, and final approval of the submitted version. D. Antic contributed to the development and review of the guidance document and final approval of the submitted version. J.H. Levy contributed to the development and review of the guidance document and final approval of the submitted version. J.D. Douketis contributed to the development and review of the guidance document and final approval of the submitted version. J. M. Connors contributed to the development and review of the guidance document and final approval of the submitted version. M. Carrier contributed to the development and review of the guidance document and final approval of the submitted version. J.I. Zwicker contributed to the development and review of the guidance document, writing of the manuscript, and final approval of the submitted version.