Morbidity and mortality are not improved by preemptive ICU transfer of acute myeloid leukemia patients presenting a high risk of early complications

doi:10.21203/rs.3.rs-2985491/v1

Download PDF

Research Article

Morbidity and mortality are not improved by preemptive ICU transfer of acute myeloid leukemia patients presenting a high risk of early complications

https://doi.org/10.21203/rs.3.rs-2985491/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

Background: Acute myeloid leukemia (AML) is associated with a high rate of life-threatening early complications. Patients presenting with hyperleukocytosis >50x10⁹/L and/or promyelocytic leukemia at the time of AML diagnosis can be considered at high risk of early complications (HReC) and thus at high risk of mortality. At our institution, we propose preemptive ICU admission to HReC patients. In so doing, our goal is to prevent complication occurrence, or, failing that, to provide rapid life-sustaining treatment (LST). In the present retrospective study, we sought to determine whether preemptive ICU admission improves survival for patients newly diagnosed with AML.

Results: We analyzed a total study population of 634 patients within a ten-year period. Of that population, 24.7% (n=157) was admitted to the ICU, 20.5% (n=130) due to complications and 4.2% (n=27) preemptively. Delays to ICU admission were 12.9 hours, 7 days and 181 days for respectively the preemptive, early complications and late complications groups, p<0.001. The preemptive group showed significantly better survival at one month post ICU admission (66.7%) compared to the early (37.7%) and late (46.4%) complications groups, p=0.039. Furthermore, LST recourse in the ICU was significantly lower in the preemptive group (41% (n=11)) compared to the early (73% (n=56)) and late (73% (n=41)) complications groups, p<0.001. However, when considering survival at one-month post AML diagnosis, no significant differences were observed between patients admitted preemptively and those admitted following complications. Additionally, Cox regression identified preemptive ICU admission as a predictive factor for mortality at one month post AML diagnosis (hazard ratio 2.72, p<0.013). These contradictory data may be explained by selection and chronological biases when the analysis is limited to only ICU patients.

Conclusion: Preemptive ICU admission does not improve survival for patients with AML. Suggested benefits for preemptive ICU admission may reflect analysis biases.

Critical Care & Emergency Medicine

Acute myeloid leukemia

intensive care unit

preemptive admission

life-sustaining treatment

prognosis

bias

Acute myeloid leukemia (AML) is a heterogenous blood cancer characterized by myeloblast proliferation and myeloid differentiation blockade. Although progress has been made over the past 30 years for the treatment of adult AML, its prognosis remains nonetheless predominantly poor. Life threatening complications are frequent in the malignancy, resulting from those revealing the pathology (leukostasis, disseminated intravascular coagulation) or appearing secondarily to chemotherapy-induced aplasia (tumor lysis syndrome or sepsis). In the literature, early admissions to the intensive care unit (ICU) for AML cases varies from 9 to 28%, illustrating the disparity of admissions criteria across institutions [1–3]. Indeed, AML incidence increases with age and the states of some patients make them unable to undergo trying induction chemotherapies and even more so benefit from ICU admission.

The "golden hour" concept developed in severe trauma has expanded out to numerous other pathologies. For example, the guidelines of the Surviving Sepsis Campaign underline the need for rapid antibiotic therapy and restoration of tissue perfusion in septic shock [4]. A recent meta-analysis suggested that early admission at the start of organ dysfunction, as opposed to delayed admission obligated by the need for life-sustaining treatment (LST), lowers ICU mortality generally and especially for critically-ill cancer patients [5, 6].

At AML diagnosis, the presence of hyperleukocytosis and severe coagulation disorders identifies patients with high risks of early complications (HReC). Indeed, hyperleukocytic leukemias (leukocyte (or blast) count > 100,000/µL) are associated with an estimated 20 to 40% risk of early death [7–9], and coagulation disorders, particularly in promyelocytic leukemia, present significant risks for fatal hemorrhagic complications [10]. Numerous studies have shown that the appearance of a complication obligating transfer to the ICU is prognostically deleterious in hemopathies [2]. Preemptive ICU admission, i.e., ICU admission of patients presenting early-complication risk factors in the hours following AML diagnosis and upstream of any organ dysfunction, has been proposed with the goal of preventing complications via increased clinical and laboratory monitoring or providing very rapid support should organ dysfunction occur. Although interesting in principle, few clinical studies have explored this still-controversial approach [11, 12].

At the University Hospital Center of Nice (France), our ICU team has been working closely with the hematology department for many years now. Over the past ten years, the number of preemptive admissions for HReC AML cases has grown and brought with it challenges concerning the use of limited resources. We present here a retrospective analysis of these practices. Our objective for this work was to assess the ability of preemptive ICU admission to reduce early complications and improve the prognoses of patients with AML.

The present single-center (University Hospital Center of Nice, France) retrospective study considers the ten-year period starting on 1 January 2011 and ending 31 December 2020, this latter chosen because of the COVID-19 pandemic. The following ICD-10 codes were searched in the center's database (Clinicom software): C92.0 (acute myeloblastic leukemia), C92.2 (atypical chronic myeloid leukemia), C92.3 (myeloid sarcoma), C92.4 (acute promyelocytic leukemia), C92.5 (acute myelomonocytic leukemia), C92.6 (acute myeloid leukemia with 11q23-abnormality), C92.7 (other myeloid leukemia), C92.8 (acute myeloid leukemia with multilineage dysplasia), C92.9 (myeloid leukemia, unspecified), C93.0 (acute monoblastic/monocytic leukemia), C94.7 (other specified leukemias), C95.0 (acute leukemia of unspecified cell type), C95.7 (other leukemia of unspecified cell type), C95.9 (leukemia, unspecified). Search results involving coding errors (erroneous diagnoses, chronic myeloid leukemia, acute lymphocytic leukemia), diagnoses predating the study period, patients lost to follow-up and patients aged less than 18 years at diagnosis were excluded.

The disease start date was that of the bone marrow biopsy confirming the diagnosis or, if absent, that of the detection of peripheral blood blasts greater than 10%.

"Early" defines here the first hospitalization following AML diagnosis and any complications occurring therein (before the date of discharge). "Late" describes hospitalizations and complications occurring after the first hospitalization.

Patients were assigned to the HReC group when they presented hyperleukocytosis ≥ 50x10⁹/L and/or acute promyelocytic leukemia (APL) [8–10, 13].

AML cases were defined as presenting complications i) when one or more extra-hematological organ dysfunction was detected (tachypnea > 30 cycles/min, PaO2 > 60 mmHg or spO2 > 90% in ambient air, persistent arterial hypotension (mean < 65mmHg) or need for vasopressors, acute renal failure KDIGO > 2, Glasgow score < 12), some necessitating LSTs (vasopressors, noninvasive (NIV ou high flow oxygen therapy) or invasive ventilation, renal replacement therapies (RRTs)), or ii) when complex therapeutic interventions (surgery, endoscopy, interventional radiology) were needed [4]. The development of an early complication was considered a theoretical indication for ICU admission.

Preemptive ICU administration was defined as ICU admission within the 24-hour window encompassing the AML diagnosis with no complications present as previously described [11, 12]. Preemptive admission criteria were HReC patient status or laboratory results indicating the presence of severe coagulation disorders (platelets < 20x10⁹/L and/or fibrinogen < 1 g/L).

At preemptive admission, the care protocol included the maintenance of platelets at > 20x10⁹/L in the absence of signs of bleeding and at > 50x10⁹/L in their presence [14], and fibrinogen levels at > 1 g/L. For APL cases, the platelets threshold was set at 30x10⁹/L and the fibrinogen level at > 1.5 g/L [10]. Coagulation disorders were treated by the administration of platelet concentrate, fibrinogen concentrate or fresh frozen plasma. Cytoreduction with hydroxyurea (50 mg/kg daily) was performed, accompanied by dexamethasone (10 mg every 12 hours until neutropenia) corticosteroid therapy in the presence of hyperleukocytosis ≥ 50x10⁹/L.

Specific hematological therapeutics were schematized into four categories: induction chemotherapy (usually cytarabine in continuous infusion for seven days associated with an anthracycline (daunorubicin or idarubicin) for three days); hypomethylating agents (azacitidine and decitabine); palliative treatment (corticosteroid therapy, transfusions of blood products, isolated hydroxyurea administration); or early death when occurring before any initiation of therapy.

Baseline assessment and data collection

The following data were collected for all patients diagnosed with AML: age; sex; Charlson score [15]; leukocyte, platelet and lactate dehydrogenase (LDH) levels; FAB classification of AML subtype; cytogenetic prognosis (poor, intermediate, favorable); and therapeutic strategy.

Furthermore, the following data were collected for patients admitted to the ICU: leukocyte, platelet, LDH, creatinine and hematocrit levels; and SOFA [16] and SAPS II [17] severity scores for the first 24 hours. The various intensive care techniques (vasopressors, invasive or noninvasive ventilation, RRTs) and causes of death in the ICU were also noted.

Statistical analysis

Statistical analyses were performed with pvalue.io, a graphic user interface to the R statistical analysis software for scientific medical publications (Medistica, 2021, available at https://www.pvalue.io/fr). Statistical significance was set at p < 0.05.

Continuous variables were expressed as medians and inter-quartile ranges (IQR). Categorical variables were reported as percentages and numbers. Continuous variables were compared between groups with the Kruskal-Wallis test and categorical variables with the chi-square test or Fisher’s exact test.

Survival was analyzed with one-year Kaplan-Meier curves and compared with the log-rank test at 30 days, six months and one year.

Survival in the preemptive group was compared via three analytical models:

Model A: Survival in three subgroups admitted to the ICU: for early complications; for late complications; preemptively. This model corresponds to that used in earlier publications [16].

Model B: Survival in three subgroups of the total hospitalized AML population: patients never admitted to the ICU; patients admitted to the ICU for complications; patients admitted to the ICU preemptively.

Model C: Survival in three subgroups according to HReC risk: patients without HReC; patients with HReC; patients admitted to the ICU preemptively. This model corresponds to that used in recent publication of Desprez and al. [18].

Survival analysis was performed using a Cox model, with survival at one month, six months and one year as the outcome variables, and type of admission, age, Charlson score, HReC and leukocytes count as the explanatory variables. The candidate covariates were selected from the set of collected variables in such a way as to ensure that there were less than 20% of patients with missing data or less than 5% of variables with missing values. The candidate covariates were included in a least absolute shrinkage and selection operation (LASSO) penalized regression model. The penalty coefficient (lambda) was chosen to provide an estimation error lower than one standard deviation of the minimum error obtained by 10-fold cross-validation, while being as parsimonious as possible. No variable had a coefficient different from 0 with this lambda coefficient.

Judgement criteria

The primary and secondary judgement criteria were respectively improved survival in the preemptive ICU admission group and reduced recourse to LST resulting from preemptive ICU admission.

Ethics approval

The study was approved by the Société de Réanimation de Langue Française Ethics Committee (n° CE SRLF 22 − 017).

Accordance with guidelines and regulations

His study did not require individual patient consent because it involved research on a previously approved database by the French Informatics and Liberty Commission (CNIL) in accordance with French legislation on noninterventional studies.

Mining of the hospital database identified 16,116 hospitalizations corresponding to 929 patients, 295 of whom were excluded for diverse reasons (Flowchart), leaving a total study population of 634 patients. Of this population, 24.7% (n = 157) was admitted to the ICU, 20.5% (n = 130) due to complications and 4.2% (n = 27) preemptively. Concerning the complication-associated ICU admissions, 58% (n = 77) and 42% (n = 56) presented early and late complications respectively (2 patients in early admission and 1 patient in preemptive admission were hospitalized a second time for late complications).

ICU-admitted population (Table 1)

The median delay between AML diagnosis and ICU admission was 12.9 hours [IQR 0–16.9] for the preemptive, 7 days [0.8–20.4] for the early complication and 181 days [79.8–327] for the late complication groups, p < 0.001. Preemptively admitted patients presented hyperleukocytosis (97.2x10⁹/L [52.8–203]) more frequently than the early (7.9x10⁹/L [0.3–46.7] and late (0.4x10⁹/L [0–5.65]) complications groups did, p < 0.001. Severity scores at admission were significantly different between the preemptive (SOFA: 2 points [1–5]; SAPS II: 34 points [29.5–38]), early complication (SOFA: 6.5 points [4–10]; SAPS II: 49.5 points [41.8–67.2]) and late complication (SOFA: 6.0 points [4–10]; SAPS II: 50 points [41–69.5]) groups p < 0.001. LST recourse was significantly lower in the preemptive group (41% (n = 11)) compared to the early (73% (n = 56)) and late (73% (n = 41) complications groups, p < 0.001. ICU survival was significantly better for patients not requiring LST (91.3% [84.4–98.9]) compared to those requiring it (45.6% [37.2–55.8]), p < 0.001.

Table 1

Characteristics of patients admitted to the ICU.
		Early complication		Late complication		Preemptive Admission		p
		%, median [IQR]	n	%, median [IQR]	n	%, median [IQR]	n
Age		65.1 [53.3–71.4]	77	64.8 [50–71.1]	56	58.1 [46.4–67.8]	27	0.23
Sex (M/F), %		68/32	52/25	66/34	37/19	44/56	12/15	0.086
Delay from diagnosis (days)		7 [0.8–20.4]	75	181 [79.8–327]	56	0.54 [0–0.705]	27	< 0.001
WBC counts at diagnosis (x10⁹/L)		31.1 [7.35–91.3]	71	10.7 [2.18–46.2]	44	100 [50.3–213]	27	< 0.001
WBC counts at ICU admission ( x10⁹/L)		7.9 [0.3–46.7]	76	0.4 [0–5.65]	56	97.2 [52.8–203]	27	< 0.001
Platelets ICU		29.5 [14.8–61.8]	69	22.5 [11–55.5]	56	54 [41.5–61.5]	27	< 0.01
Creatinine		121 [76–214]	75	104 [61.8–162]	56	90.0 [69.5–118]	27	0.036
Hematocrit		24.0 [22–27]	76	25.8 [22.8– 29]	56	25.0 [22.5–27.5]	27	0.33
FAB AML subtype
AML 4–5		26	20	16	9	41	11
ALM 3		10	8	5.4	3	7.4	2	0.1
others		64	49	79	44	52	14
Cytogenetics
Poor		36	28	36	21	15	4	0.079
Intermediate		23	18	36	20	48	13	0.042
Favorable		15	12	14	8	22	6	0.66
Unkown		25	19	13	7	15	4	0.13
Allograft complication		0	0	27	15	0	0	< 0.001
HReC group		40	31	25	14	78	21	< 0.001
SAPS2		50 [41–69.5]	77	49.5 [41.8–67.2]	56	34.0 [29.5–38]	27	< 0.001
SOFA		6.5 [4–10]	77	6.0 [4–10]	62	2 [1–5]	27	< 0.001
ICU techniques		73	56	73	41	41	11	< 0.01
Mechanical ventilation		38	29	52	29	26	7	0.061
NIV		14	11	7.1	4	3.7	1	0.24
Vasopressors		50	38	57	32	19	5	< 0.01
RRT		26	20	16	9	15	4	0.27
Length of stay ICU		2.8 [1.06–5.84]	77	2.1 [0.925–6.33]	56	4.75 [3.01–5.86]	27	0.087
Cause of ICU–death
Septic shock – MOF		39	13	50	13	14	1	0.19
ARDS – MOF		24	8	27	7	14	1
Cerebral hemorrhage		18	6	19	5	72	5
Hemorrhagic shock		9.1	3	3.8	1	0	0
Cardiac arrest		9.1	3	0	0	0	0
ICU Survival	56.0% [45.8–68.59]		44	52.7% [41.0–67.8]	30	73.9% [59.0–92.6]	20	0.16
At 1 month	37.7.0% [28.3–50.2]		29	46.4% [35.0–61.5]	26	66.7% [51.1–87]	18	0.039
At 6 months	32.4% [23.5–44.8]		25	30.4% [20.4–45.1]	17	51.9% [36.1–74.6]	14	0.082
At 1 year	27.3% [18.9–39.3]		21	23.6% [14.7–38.0]	13	37.0% [22.6–60.6]	10	0.16
Abbreviations: ARDS, acute respiratory distress syndrome ; HReC, high risks of early complications ; ICU, intensive care unit; IQR, interquartile range ; FAB, French-American-British ; M/F, male/female ; MOF, multiple organ failure ; NIV, non invasive ventilation; RRT, renal replacement therapy; SAPS2, simplified acute physiology score; SOFA, sequential organ failure assessment score; WBC, white blood cells.

Survival according to delay to ICU admission (Fig. 2A)

Post ICU discharge survival was significantly better at one month for the preemptive group (66.7% [51.1–87]) compared to the early (37.7% [28.3–50.2]) and late (46.4% [36.1–74.6]) complications groups, p = 0.039. However, no significant differences in survival were observed at six months or one year after ICU discharge.

Non-ICU-admitted population (Table 2)

Of the 634-patient total study population, 75.2% (n = 477) was not admitted to the ICU. These patients, compared to those admitted to the ICU, had greater age (71.7 years [60.9–80.7] vs 65.1 years [52.2–71.8], p < 0.001), more comorbidities (Charlson score 6 [5–7] vs 6 [4–7], p < 0.001), less hyperleukocytosis (5.8x10⁹/L [2.2–27.4] vs 24.0x10⁹/L [2.98–71.7], p < 0.001) and less frequent induction chemotherapy (47% (n = 218) vs 71% (n = 85), p < 0.01). At AML diagnosis, 8.9% (n = 57) of non-ICU-admitted patients presented aspects theoretically indicative of ICU admission but nonetheless were not transferred there. This subgroup, compared to the ICU-admitted group, was significantly older (81.2 years [72.2–85.5] vs 64.7 years [50.6–71.9], p < 0.001), more affected by comorbidities (Charlson score 7 points [6–9] vs 6 points [6–9], p < 0.001) and more frequently in palliative care (61% (n = 31) vs 9.7% (n = 7), p < 0.001). Their survival, again compared to ICU-admitted patients, was lower: 24.6% [15.6–38.7] vs 46.8% [37–59.2], p < 0.01, at one month; 12.3% [6.14–24.6] vs 34.2% [25.2–46.4], p < 0.01 at six months; and 12.3% [6.14–24.6] vs 27.8% [19.5–39.7], p < 0.01 at one year.

Table 2

Characteristics of all population with newly diagnosed ALM
	No ICU		ICU		Preemptive		p
	%, mediane [IQR]	n	%, mediane [IQR]	n	%, mediane [IQR]	n	p
Age (years)	71.7 [60.9–80.7]	477	65.1 [52.2–71.8]	130	58.1 [46.4–67.8]	27	< 0.001
Sexe (M/F)	55/45	260/217	68/32	88/41	44/56	12/15	< 0.01
WBC counts at diagnosis (x10⁹/L)	5.8 [2.2–27.4]	425	24.4 [2.98–71.7]	116	103 [50.8–209]	27	< 0.001
> 50 (x10⁹/L)	15	66	31	36	74	20	< 0.001
> 100 (x10⁹/L)	5	23	17	20	52	14	< 0.001
Platelets (x10⁹/L)	53 [29–108]	393	51 [25.2–118]	102	58 [34.0–62.0]	25	0.92
LDH (UI/L)	618 [398–1061]	301	882 [508–1704]	78	1395 [764–2268]	19	< 0.001
Charlson Score (Points)	6 [5–7]	477	6 [4–7]	129	4 [3–5.5]	27	< 0.001
FAB subtype
AML 4–5	11	52	22	28	41	11	< 0.001
AML 3	5	23	8	10	7	2	0.27
Others	84	402	71	92	52	14	< 0.001
Cytogenetics
Poor	49	232	37	48	15	4	< 0.001
Intermediate	24	116	28	37	44	12	0.098
Favorable	13	61	15	19	26	7	0.27
Unknow	14	68	20	26	15	4	0.31
Theorical indication to ICU during first admission	12	57	54	68	44	12	< 0.001
Risk group	19	89	35	46	81	21	< 0.001
Treatment
Induction	47	218	71	85	74	20	< 0.001
Hypomethylating agents	33	154	17	20	0	0	< 0.001
Palliative care	18	83	4.2	5	11	3	< 0.001
Early deaths	1.5	7	7.6	9	15	4	< 0.001
Allograft	26	124	23	29	22	6	0.37
Micro-allograft	0.63	3	0.78	1	3.7	1
Survival after AML diagnosis							logrank
1 month	83.4% [80.1–86.8]	393	67.9% [60.2–76.5]	89	66.7% [51.1- 87.0]	18	< 0.001
6 months	59.2% (55–63.8]	279	53.1% [45.1–62.5]	68	55.6% [39.6–77.8]	15	0.18
1 year	44.8% (40.5–49.6]	210	35.7% (28.2–45.1]	46	36.7% [22.2–60.4]	10	0.07
Abbreviations: IQR, interquartile range; FAB, French-American-British; LDH, lactate deshydrogenase; M/F, male/female; WBC, white blood cells.

Survival from AML diagnosis according to ICU stay (Fig. 2B)

Survival at one month post AML diagnosis for patients not admitted to the ICU (83.4% [80.1–86.8]) was significantly better than that for patients admitted to the ICU because of complications (early and late; 67.9% [60.2–76.5]) and that for patients admitted preemptively (66.7% [51.1–87.0]), p < 0.001, but that endpoint did not differ between these two latter groups. At six months and one year post AML diagnosis, no differences between the three groups remained significant.

High risk of early complications group

HReC patients represented 25% (n = 156) of the population. Subgroup distributions were 57% (n = 89) of patients not admitted to the ICU, 29% (n = 46) of patients admitted to the ICU for early or late complications and 13% (n = 21) of patients admitted to the ICU preemptively. The patients in the preemptive admission group had significantly lower age (58.1 years [46.4–67.8] vs 67.5 years [50.1–79.5], p < 0.015) and less morbidity (Charlson score 4 points [3–5.5] vs 6 points [4–7]), p < 0.01. Hyperleukocytosis was more marked in the preemptive group as well (103x10⁹/L [50.8–209] vs 75.4x10⁹/L [51.4–130]), p = 0.0502.

Early theoretical indication for ICU transfer motivated by the development of organ dysfunction was found in 42% (n = 57) of HReC patients vs 41% (n = 11) of preemptive admission patients, p = 0.89.

Survival from AML diagnosis according to risk group (Fig. 2C)

One-month post AML-diagnosis survival for non-HReC patients (83.3% [80–86.8]) was significantly better than those for HReC patients (71.2% [63.9–79.4]) and preemptive ICU admission patients (66.7% [51.1–87.0]), p < 0.001. At six months and one year, no differences between the three groups remained significant. Survival for the preemptive ICU admission patients appeared comparable to that of HReC patients at all analysis times, notably at four days post AML diagnosis (median length of preemptive ICU stay), the delay at which 15% (n = 4) of preemptive admission patients had died vs. 14% (n = 19) of HReC patients, p = non-significant.

Factors predictive of mortality at AML diagnosis (Table 3)

Using Cox regression in the total study population (n = 634), the factors predictive of post AML-diagnosis mortality were ICU admission for complications (hazard ratio (HR) 2.4, p < 0.001 at 1 month; HR 1.6, p < 0.001 at 6 months; HR 1.554, p < 0.01 at 1 year), age (HR 1.04, p < 0.001 at 1 month; HR 1.04, p < 0.001 at 6 months; HR 1.034, p < 0.001 at 1 year), Charlson score (HR 1.15, p < 0.0495 at 1 month; HR 1.19, p < 0.001 at 6 months; HR 1.18, p < 0.001 at 1 year) and leukocytes (HR 1.06, p < 0.001 at 1 month; HR 1.07, p < 0.001 at six months; HR 1.07, p < 0.001 at 1 year). HReC status was not identified as a mortality risk factor. In contrast, preemptive ICU admission was predictive of one-month mortality (HR 2.72, p < 0.013).

Table 3

Factors predictive of mortality at one month, six months and one year post AML diagnosis.
	1–month mortality			6–months mortality			1–year mortality
	HR [95% CI]		p–Value	HR [95% CI]		p–Value	HR [95% CI]		p–Value
Type admission
ICU vs no ICU	2.41	[1.57–3.69]	< 0.001	1.67	[1.22–2.30]	< 0.01	1.59	[1.22–2.09]	< 0.001
Preemptive vs no ICU	2.72	[1.24–5.98]	0.013	1.69	[0.886–3.22]	0.11	1.53	[0.872–2.69]	0.14
GHR	1.05	[0.597–1.74]	0.86	0.782	[0.522–1.17]	0.23	0.723	[0.512–1.02]	0.066
Age	1.04	[1.02–1.06]	< 0.01	1.04	[1.03–1.06]	< 0.001	1.03	[1.02–1.05]	< 0.001
Charlson	1.15	[0.998–1.31]	0.0495	1.19	[1.08–1.3]	< 0.001	1.18	[1.09–1.27]	< 0.001
Hyperleucocytosis	1.06	[1.03–1.09]	< 0.001	1.07	[1.05–1.10]	< 0.001	1.07	[1.05–1.09]	< 0.001
HR Hazard–Ratio; CI Confidence Interval.

For patients diagnosed with AML and presenting laboratory markers of potentially life-threatening complications, preemptive ICU admission should, in principle, enable the prevention of those complications or at least improved prognosis via the rapid deployment of LSTs should they occur. In our study, preemptive admission occurred very rapidly—12.9 hours on average—and concerned mainly patients identified as having a high risk of early complications. The studies on preemptive ICU admission in the literature do not inspire great confidence. Lengliné et al reported a tendency toward improved survival with preemptive admission (ICU survival of 79% for preemptive ICU admission vs. 65% for late ICU admission, p = 0.12) [11]. That team also observed less LST recourse with preemptive admission (nearly 2-fold decrease in mechanical ventilation and nearly 4-fold decrease in vasoactive drugs). Mottal et al found similar results, reporting greater use of RRT, amines and invasive ventilation in late ICU admission and reduced morbidity with preemptive admission [12]. But these studies compare survival in ICU between a preemptively admitted group and a group admitted for early complication (identic to our Model A). In this model, we too observed improved post-ICU survival at one month and significantly reduced LST recourse in preemptive ICU admission. But comparing the preemptive group to groups admitted to ICU for complications introduces an obvious bias. Indeed, we observed significant differences between the severity scores of patients admitted preemptively and those admitted following complications. Those differences could be interpreted as inherent to the selection process of the preemptive group. These patients were chosen before the development of complications and thus in better states of health; the reduced use of LSTs would logically be the result of preventive treatments warding off the development of those complications. However, that interpretation would require, in the absence of specific interventions, a probability of complications in the preemptive group close to 100%, similar to that of the group admitted for complications. It was therefore essential to know the incidence of complications in the HReC group from which the preemptive group was selected. We found that the real incidence of complications in the HReC group was 42%, which was similar to that of the preemptive group. These findings suggests that more than half of HReC patients and more than half of preemptively admitted patients will never present a complication imposing transfer to the ICU. Therefore, the potentiality of exacerbation is largely different between patients admitted to the ICU preemptively and those admitted there because of the development of complications, and any interpretations on this model will show a selection bias. Furthermore, there is a chronological bias. In patients admitted preemptively, complication development is limited to a short, median four-day window at the start of hospitalization. In patients admitted following the development of a complication, that chronological window is much wider. To take that time effect into account, it is more logical to compare survival in the whole population from the date of the AML diagnosis (Model B). Here again, in our study, we were unable to demonstrate an improvement in survival. A best methodology is probably to compare the survival of the preemptive group to the survival of the HReC group (model C) because patients in preemptive admissions are essentially patients with risks of early complications. Recently, Desprez et al, found that preemptive admission did not provide a survival benefit in a cohort of newly diagnosed ALM with hyperleukocytosis ≥ 50x10⁹/L at diagnosis [18]. In our study, preemptive admission does not show an improvement in early or one-year overall survival of the HReC group and also does not appear to increase the number of patients receiving bone marrow transplants.

A possible explanation for the lack of benefit of preemptive admission is the brutality and intensity of the observed complications, despite the preventive measures deployed to avoid them. The main cause of early death is hemorrhage, usually intraparenchymal cerebral or pulmonary, that rapidly becomes uncontrollable in a setting of severe coagulation disorders [7]. Leukostasis, i.e., the obstruction of the microvasculature by aggregations of circulating blasts, seems to be at the root of the problem. Leukostasis sets off complex endothelial-cell activation mechanisms involving such cytokines and inflammatory factors as IL-1β and TNF-α and promoting the expression of cell adhesion molecules like ICAM-1 and VCAM-1 [19]. Although hemorrhagic events are correlated with hyperleukocytosis, the rapid reduction of circulating blasts is not sufficient for reducing mortality. Techniques thereto like rapidly efficacious leukapheresis or slower cytoreduction with hydroxyurea have not been shown to provide a significant improvement in survival [20, 21]. It may be possible that during cytoreduction, certain patients will have accesses of inflammation, triggering endothelial activation and the resulting hemorrhage. Building upon this pathophysiological hypothesis, the use of dexamethasone has been proposed, but any benefits of it remain controversial and demonstrated only for lung injury and in only a small study population [22].

Our study is limited by its retrospective and single-center nature and by the small number of patients in the preemptive group that does not allow for a propensity analysis. Survival analyses at different times can also lead to confusion but it was important to choose a model previously used to demonstrate that identical results could be falsely obtained. The HReC group selection criteria, i.e., the threshold of 50x10⁹/L to define hyperleukocytosis and the inclusion of patients with promyelocytic leukemia, can also be criticized. Those choices would tend to diminish severity in the HReC group and may explain why HReC status was not a factor predictive of mortality in regression analysis. However, that limitation should have favored the preemptive group. Hyperleukocytosis in contrast is a better predictive factor of mortality. The results of our study may also raise questions as to the validity of deployed preventive measures, notably for the management of coagulation disorders. Having followed current treatment guidelines in our study, the results of this latter may suggest that treatment thresholds are underestimated for this subgroup with high risks of early complications. There is no consensus on hydroxyurea and dexamethasone in the AML setting, but their use is more so a question of futility than an explanation of our negative results. Only a prospective, randomized study could perhaps bring clarity to this question by discarding the selection bais. However, considering our discouraging results, doing so would surely result in a high number needed to treat and immobilize precious human and technical resources.

We were not able to demonstrate a beneficial effect of preemptive ICU admission for AML patients at high risk of early complications. Targeting this population is judicious considering the number of complications occurring in it, but intensive care specialists remain frustratingly poorly armed in the battle against acute myeloid leukemia. In our opinion, to improve the prognosis of hyperleukocytic leukemia, it may be preferable to aim for a better understanding of the leukostasis mechanisms underlying hemorrhagic complications and the development of specific hemostatic therapeutics rather than focusing on early transfer to the intensive care unit.

AML: Acute myeloid leukemia; CNIL: French Informatics and Liberty Commission; FAB: French-American-British; HReC: High Risk of early Complications; ICD: International Classification of Diseases; ICU: Intensive care unit; KDIGO: Kidney Disease: Improving Global Outcome; LASSO: Least Absolute Shrinkage and Selection Operation; LST: Life-Sustaining Treatment; RRTs: Renal Replacement Therapies; SOFA: Sequential Organ Failure Assessment; SAPS: Simplified Acute Physiology Score; UO: Urine output.

Author contributions

HH designed the study, HH, BM assisted in data collection and analysis, ML conducted the statistical analysis and HH, BM drafted and DD, JM, DJ revised the manuscript. All authors read and approved the final manuscript.

Funding

Not applicable.

Availability of data and materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Consent for publication

All authors consent for publication.

Competing interests

The authors declare that they have no competing interests.

Acknowledgments

The authors would like to sincerely thank Mr. Kevin ERWIN for proofreading the English manuscript.

Schellongowski P, Staudinger T, Kundi M, Laczika K, Locker GJ, Bojic A, et al. Prognostic factors for intensive care unit admission, intensive care outcome, and post-intensive care survival in patients with de o acute myeloid leukemia: a single center experience. Haematologica. 2011;96(2):231‑7.
Jackson K, Mollee P, Morris K, Butler J, Jackson D, Kruger P, et al. Outcomes and prognostic factors for patients with acute myeloid leukemia admitted to the intensive care unit. Leuk Lymphoma. 2014;55(1):97‑104.
Atallah E, Cortes J, O’Brien S, Pierce S, Rios MB, Estey E, et al. Establishment of baseline toxicity expectations with standard frontline chemotherapy in acute myelogenous leukemia. Blood. 15 2007;110(10):3547‑51.
Evans L, Rhodes A, Alhazzani W, Antonelli M, Coopersmith CM, French C, et al. Surviving sepsis campaign: international guidelines for management of sepsis and septic shock 2021. Intensive Care Med. 2021;47(11):1181‑247.
Hourmant Y, Mailloux A, Valade S, Lemiale V, Azoulay E, Darmon M. Impact of early ICU admission on outcome of critically ill and critically ill cancer patients: A systematic review and meta-analysis. J Crit Care. 2021;61:82‑8.
Mokart D, Lambert J, Schnell D, Fouché L, Rabbat A, Kouatchet A, et al. Delayed intensive care unit admission is associated with increased mortality in patients with cancer with acute respiratory failure. Leuk Lymphoma. 2013;54(8):1724‑9.
Porcu P, Danielson CF, Orazi A, Heerema NA, Gabig TG, McCarthy LJ. Therapeutic leukapheresis in hyperleucocytic leukaemias: lack of correlation between degree of cytoreduction and early mortality rate. Br J Haematol. 1997;98(2):433‑6.
Dutcher JP, Schiffer CA, Wiernik PH. Hyperleukocytosis in adult acute nonlymphocytic leukemia: impact on remission rate and duration, and survival. J Clin Oncol. 1987;5(9):1364‑72.
Slovak ML, Kopecky KJ, Wolman SR, Henslee-Downey JP, Appelbaum FR, Forman SJ, et al. Cytogenetic correlation with disease status and treatment outcome in advanced stage leukemia post bone marrow transplantation: A Southwest Oncology Group Study (SWOG-8612). Leuk Res. 1995;19(6):381‑8.
Yanada M, Matsushita T, Asou N, Kishimoto Y, Tsuzuki M, Maeda Y, et al. Severe hemorrhagic complications during remission induction therapy for acute promyelocytic leukemia: incidence, risk factors, and influence on outcome. Eur J Haematol. 2007;78(3):213‑9.
Lengliné E, Raffoux E, Lemiale V, Darmon M, Canet E, Boissel N, et al. Intensive care unit management of patients with newly diagnosed acute myeloid leukemia with no organ failure. Leuk Lymphoma. 2012;53(7):1352‑9.
Mottal N, Issa N, Dumas PY, Camou F, Sauvezie M, Gros FX, et al. Reduce Mortality and Morbidity in Acute Myeloid Leukemia With Hyperleukocytosis With Early Admission in Intensive Care Unit: A Retrospective Analysis. J Hematol. 2020;9(4):109‑15.
Porcu P, Cripe LD, Ng EW, Bhatia S, Danielson CM, Orazi A, et al. Hyperleukocytic Leukemias and Leukostasis: A Review of Pathophysiology, Clinical Presentation and Management. Leuk Lymphoma. 2000;39(1‑2):1‑18.
Döhner H, Estey EH, Amadori S, Appelbaum FR, Büchner T, Burnett AK, et al. Diagnosis and management of acute myeloid leukemia in adults: recommendations from an international expert panel, on behalf of the European LeukemiaNet. Blood. 2010;115(3):453‑74.
Charlson M, Szatrowski TP, Peterson J, Gold J. Validation of a combined comorbidity index. J Clin Epidemiol. 1994;47(11):1245‑51.
Vincent JL, Moreno R, Takala J, Willatts S, De Mendonça A, Bruining H, et al. The SOFA (Sepsis-related Organ Failure Assessment) score to describe organ dysfunction/failure: On behalf of the Working Group on Sepsis-Related Problems of the European Society of Intensive Care Medicine (see contributors to the project in the appendix). Intensive Care Med. 1996;22(7):707‑10.
Le Gall JR. A new Simplified Acute Physiology Score (SAPS II) based on a European/North American multicenter study. JAMA J Am Med Assoc. 1993;270(24):2957‑63.
Desprez C, Kouatchet A, Marchand T, Mear JB, Tadié JM, Peterlin P, et al. Outcome of patients with newly diagnosed AML admitted to the ICU, including preemptive admission – a multi-center study. Ann Hematol. 2023;102(6):1383‑93.
Stucki A, Rivier AS, Gikic M, Monai N, Schapira M, Spertini O. Endothelial cell activation by myeloblasts: molecular mechanisms of leukostasis and leukemic cell dissemination. Blood. 2001;97(7):2121‑9.
Bewersdorf JP, Giri S, Tallman MS, Zeidan AM, Stahl M. Leukapheresis for the management of hyperleukocytosis in acute myeloid leukemia—A systematic review and meta‐analysis. Transfusion (Paris). 2020;60(10):2360‑9.
Stahl M, Shallis RM, Wei W, Montesinos P, Lengline E, Neukirchen J, et al. Management of hyperleukocytosis and impact of leukapheresis among patients with acute myeloid leukemia (AML) on short- and long-term clinical outcomes: a large, retrospective, multicenter, international study. Leukemia. 2020;34(12):3149‑60.
Azoulay E, Canet E, Raffoux E, Lengline E, Lemiale V, Vincent F, et al. Dexamethasone in patients with acute lung injury from acute monocytic leukaemia. Eur Respir J. 2012;39(3):648‑53.

Download PDF

Version 1

posted

You are reading this latest preprint version

Morbidity and mortality are not improved by preemptive ICU transfer of acute myeloid leukemia patients presenting a high risk of early complications

Status:

Version 1

Abstract

Figures

Background

Methods

Baseline assessment and data collection

Statistical analysis

Judgement criteria

Ethics approval

Accordance with guidelines and regulations

Results

ICU-admitted population (Table 1)

Survival according to delay to ICU admission (Fig. 2A)

Non-ICU-admitted population (Table 2)

Survival from AML diagnosis according to ICU stay (Fig. 2B)

High risk of early complications group

Survival from AML diagnosis according to risk group (Fig. 2C)

Factors predictive of mortality at AML diagnosis (Table 3)

Discussion

Conclusion

Abbreviations

Declarations

References

Status:

Version 1