report is the first to describe the use of blinatumomab in the treat-

ment of CD19+ MPAL and to establish its efficacy in controlling this

rare type of acute leukemia.

Although the survival of patients with MPAL has improved in

recent years, primarly due to incorporation of allo-HSCT into treat-

ment, the overall prognosis remains poor. The risk of death from

MPAL is 59% and 26% higher than for ALL and AML, respectively.6 In

the absence of prospective clinical trials, MPAL treatment is guided by

retrospective studies and is based on ALL regimens. Given high

relapse rates with chemotherapy, targeted approaches may improve

outcomes.

B/myeloid MPAL blasts express CD19. In this report, we aimed to

study the efficacy of blinatumomab in the treatment of two patients

with CD19+ MPAL. The clinical hypothesis was that treatment of

these patients with CD19+ MPAL with blinatumomab, along with a

tyrosine kinase inhibitor for Ph(+) disease in the first patient, would

result in the achievement of sustained MRD-negative CR and mainte-

nance of MRD-negative CR in the first and second patient, respec-

tively. Our patients had short periods of myelosuppression and few

adverse events, which led to overall improved functional status after

therapy. Both patients were able to proceed to allo-HSCT, thus far

resulting in sustained MRD-negative clinical remissions. The poor out-

comes of R/R MPAL with chemotherapy underscore the need for a

prospective clinical study of targeted therapy in this patient popula-

tion. We suggest that blinatumomab is an excellent candidate for this

purpose.

CONFLICT OF INTEREST

A.E. is a Global Oncology Advisory Board Member for Amgen and has

served as a consultant to Amgen.

ORCID

Vu H. Duong https://orcid.org/0000-0003-0278-7727

Firas El Chaer1,2

Omer M. Ali1

Edward A. Sausville1,2

Jennie Y. Law1,2

Seung Tae Lee1,2

Vu H. Duong1,2

Maria R. Baer1

Rima Koka3

Zeba N. Singh3

Jerry Wong4

Jean A. Yared1,2

Elizabeth A. Griffiths5

Ashkan Emadi1,2

1Departments of Medicine, University of Maryland School of Medicine,

Baltimore, Maryland2University of Maryland Greenebaum Comprehensive Cancer Center

Baltimore, Baltimore, Maryland3Departments of Pathology, University of Maryland School of Medicine,

Baltimore, Maryland

4Department of Pathology, Roswell Park Comprehensive Cancer Center,

Buffalo, New York5Department of Medicine, Roswell Park Comprehensive Cancer Center,

Buffalo, New York

Correspondence

Ashkan Emadi, Medicine and Pharmacology, Associate Director for

Clinical Research, Marlene & Stewart Greenebaum Comprehensive

Cancer Center, Director, Hematology & Medical Oncology Fellowship

Training Program, University of Maryland School of Medicine, 22 South

Greene Street, Room N9E24, Baltimore, MD 21201.

Email: aemadi@umm.edu

Firas El Chaer and Omer M. Ali contributed equally to this study.

REFERENCES

1. Yan L, Ping N, Zhu M, et al. Clinical, immunophenotypic, cytogenetic,

and molecular genetic features in 117 adult patients with

mixed-phenotype acute leukemia defined by WHO-2008 classification.

Haematologica. 2012;97(11):1708-1712.2. Arber DA, Orazi A, Hasserjian R, et al. The 2016 revision to the World

Health Organization classification of myeloid neoplasms and acute leu-

kemia. Blood. 2016;127(20):2391-2405.3. Maruffi M, Sposto R, Oberley MJ, Kysh L, Orgel E. Therapy for children

and adults with mixed phenotype acute leukemia: a systematic review

and meta-analysis. Leukemia. 2018;32(7):1515-1528.4. Wolach O, Stone RM. How I treat mixed-phenotype acute leukemia.

Blood. 2015;125(16):2477-2485.5. Tian H, Xu Y, Liu L, et al. Comparison of outcomes in mixed phenotype

acute leukemia patients treated with chemotherapy and stem cell trans-

plantation versus chemotherapy alone. Leuk Res. 2016;45:40-46.6. Shi R, Munker R. Survival of patients with mixed phenotype acute

leukemias: a large population-based study. Leuk Res. 2015;39(6):

606-616.

Received: 27 August 2018 Revised: 12 October2018

Accepted: 15 October2018

DOI: 10.1002/ajh.25321

Platelet sequestration andconsumption in the placentalintervillous space contributeto lower platelet countsduring pregnancy

To the editor:

Platelet counts of all women with an uncomplicated pregnancy gradu-

ally decrease at approximately the same rate throughout their preg-

nancy.1 Dilution of platelets by plasma volume expansion is an

apparent cause for lower platelet counts, but multiple additional phys-

iologic changes during pregnancy may also contribute to the lower

platelet counts.

E8 CORRESPONDENCE

Blood circulation through the placenta is similar to blood circula-

tion through the spleen. In the spleen, some blood flows directly from

arterioles to venules, without intervening capillaries, whereas other

blood is shunted into the low pressure pools in the sinusoids. Plasma

preferentially flows directly to the venules, described as “plasma skim-

ming”, while the blood cells are sequestered in the sinusoids.2 Platelet

counts in splenic blood are 7-fold greater than platelet counts in

peripheral blood.3 In normal adults, one-third of all circulating platelets

TABLE 1 Comparison of peripheral and placental intervillous space blood counts in women with scheduled cesarean sections

Median (range)

Blood cells Peripheral blood Placental blood (correcteda) Absolute difference between medians P valueb

All 40 women with uncomplicated deliveries and acceptable placental blood samples

Hemoglobin (g/dL) 11.9 (8.0-14.4)

White blood cells (103/μL) 9.50 (5.60-12.99) 9.94 (1.64-23.16) +0.44 .0285

Neutrophils (103/μL) 6.50 (3.46-9.51) 6.40 (0.11-17.61) −0.10 .2231

Lymphocytes (103/μL) 2.05 (1.21-4.81) 2.44 (0.04-6.69) +0.39 .0012

Platelets (103/μL) 233 (98-340) 164 (18-728) −69 .0285

The 20 women with placental blood samples that had fetal blood contamination ≤ 13.5%c

Hemoglobin (g/dL) 11.6 (9.1-14.4)

White blood cells (103/μL) 9.56 (5.6-12.99) 11.52 (5.92-23.16) +1.96 0.0002

Neutrophils (103/μL) 6.65 (3.75-9.51) 7.69 (3.55-17.61) +1.04 0.0032

Lymphocytes (103/μL) 1.99 (1.29-4.81) 2.54 (1.42-5.07) +0.55 0.0020

Platelets (103/μL) 233 (98-324) 179 (73-728) −54 0.1992

a White blood cell and platelet counts in blood from the placental intervillous space were corrected for the amount of dilution by the EDTA anticoagulantsolution, based on the relative hemoglobin concentrations in the peripheral blood and placental intravillous space blood. The median correction ratio forall 40 placental blood samples (peripheral hemoglobin/placental hemoglobin) was 1.9 (range, 0.69-4.31).

b P values were obtained from the Wilcoxon signed-rank test.c The median fetal blood contamination of the placental intervillous blood samples was 13.5%. To determine the effect of fetal blood contamination onwhite blood cell and platelet counts, data for the 20 placental blood samples with ≤13.5% contamination by fetal blood are presented.

FIGURE 1 Placental morphology and immunohistochemical staining of sections of a placenta obtained at delivery and included in this study.

Placental sections were stained with hematoxylin and eosin (H&E) to demonstrate the structure of the villi and perivillous fibrinoid, and alsowith a rabbit anti-CD61 (platelet membrane glycoprotein IIIa) monoclonal antibody by ProPath referral pathology service, Dallas, TX. A,Lowmagnification view of the placental parenchyma illustrating mature terminal villi. Perivillous fibrinoid filling gaps in the syncytiotrophoblastsurface of the villi are identified by the homogeneous pink stain. Representative examples of the perivillous fibrinoid are indicated by the arrows.(H&E, 10× objective) B, An adjacent location in the same placental section stained with anti-CD61 (brown color) demonstrating platelets withinthe perivillous fibrinoid, in a linear pattern along the villous surfaces throughout this section. Representative examples of platelet staining areindicated by the arrows (CD61 immunostaining, 10× objective) C, high-magnification of the placental parenchyma illustrates a mature terminalvillus. The arrow indicates perivillous fibrinoid sealing a gap in the syncytiotrophoblast surface. (H&E, 100× objective) D, An adjacent locationstained with antiCD61 (brown color) demonstrating platelets within the perivillous fibrinoid, indicated by the arrow. (CD61 immunostaining,100× objective)

CORRESPONDENCE E9

are transiently sequestered within the splenic sinusoids. Increased

spleen size results in greater splenic sequestration of platelets and

lower peripheral blood platelet counts.3

In the placenta some blood also flows directly from arterioles to

venules, without intervening capillaries, while other blood is shunted

into the low pressure pools in the intervillous spaces. These similar pat-

terns of circulation suggest that there may be increased sequestration

of blood cells in the placental intervillous spaces, similar to the splenic

sinusoids. Although platelets sequestered within splenic sinusoids

return to the circulation,3 platelets sequestered within the intervillous

spaces may be consumed by incorporation into the perivillous fibrinoid.

Perivillous fibrinoid is present in all term placentas. It is a product of

coagulation that fills gaps in the syncytiotrophoblast cells lining the sur-

face of mature villi and maintains the integrity of the villous surface.4,5

Platelet counts of women with twin pregnancies are lower than

platelet counts of women with singleton pregnancies at each trimester

and at delivery.1 Women with twin pregnancies have larger placentas, or

may have 2 placentas. This suggests that larger placentas in women with

twin pregnancies may contribute to their greater platelet count decrease.

Based on these observations, we postulated that blood cells are

sequestered within the intervillous spaces of the placenta and that plate-

lets are then consumed by incorporation into perivillous fibrinoid. To test

this hypothesis, we compared blood cell counts in maternal peripheral

blood to blood obtained from the placental intervillous space.

We collected placentas from 40 women with uncomplicated

singleton pregnancies who had a scheduled cesarean delivery for

routine indications at the Oklahoma University Medical Center. The

mother’s complete blood count (CBC) from the day of delivery was

recorded. At delivery, placentas were immediately immersed in a

solution of 4.5 mmol ethylenediaminetetraacetic acid (EDTA) in

0.85% sodium chloride to prevent blood coagulation within the

intervillous space. Maternal placental blood was collected from the

intervillous space6 and placental tissue samples for microscopy

were also collected at the same time. Contamination of maternal

blood with fetal blood was determined by the percent of fetal red

blood cells in each sample. Then a CBC was performed on the pla-

cental blood sample with the lowest percentage of fetal blood.

A stained smear of the placental blood was examined for platelet

clumping; if clumping occurred, the sample was not used. To cor-

rect for dilution of placental blood by the EDTA solution, we used

the ratio of hemoglobin concentration in the peripheral blood to

placental blood to adjust the placental white blood cell and platelet

counts. To determine if contamination of maternal blood from the

intervillous space by fetal blood influenced the blood counts, we

separately analyzed blood counts on the 20 women with placental

blood samples that had fetal blood contamination less than the

median value. This study was approved by the University of Okla-

homa Institutional Review Board (approval #4156).

The median time between delivery and placental blood sample col-

lection was 18 minutes (range, 13-82 minutes). The placental weight

(median 501 g, range 364-938) was more than 2-fold the weight of a

normal adult spleen (150-200 g). The median total white blood cell

count, neutrophil count, and lymphocyte count, corrected for dilution

by the EDTA solution, were all higher in placental blood than in the

peripheral venous blood. These differences were significant when the

placental blood of the 20 women with less fetal blood contamination

was analyzed (Table 1). The median corrected platelet count in placental

blood of all 40 women was lower than the platelet count in peripheral

venous blood. The median corrected platelet count in placental blood

of the 20 women with less fetal blood contamination was also lower

than the platelet count in peripheral venous blood; however, the differ-

ence was not significant, possibly due to the smaller sample size and

the variation among the platelet counts.

Microscopic examination of all 40 placentas documented that they

were normal. None of the placentas had excessive subchorionic or perivil-

lous fibrinoid, intervillous or subchorionic thrombi, or inflammatory infil-

trates. Immunohistochemical stains using a rabbit monoclonal antibody to

the platelet membrane glycoprotein IIIa (CD61) identified platelets within

the perivillous fibrinoid (Figure 1). These images suggest that platelets are

consumed in the process of blood coagulation and fibrinoid generation to

seal the gaps in the syncytiotrophoblastic surface of the terminal villi.

The higher white blood cell counts in placental blood compared

to peripheral venous blood are consistent with the sequestration of

blood cells within the intervillous space, similar to the sequestration

of blood cells in splenic sinusoids.3 The lower or equivalent platelet

counts in placental blood compared to peripheral venous blood and

the immunohistochemistry identification of platelets within the peri-

villous fibrinoid are consistent with the hypothesis that platelets are

not only sequestered within the placental intervillous space but also

consumed by incorporation into the perivillous fibrinoid, which main-

tains integrity of the syncytiotrophoblast surface. These data support

the hypothesis that the sequestration and consumption of maternal

platelets within the placental intervillous space contribute to lower

platelet counts during pregnancy.

ACKNOWLEDGMENT

This project was supported in part by 1K01HL135466-01 from the

National Heart, Lung and Blood Institute (Terrell).

ETHICS STATEMENT

This study was approved by the University of Oklahoma Health Sci-

ences Center Institutional Review Board (approval #4156).

CONFLICT OF INTEREST

Nothing to report.

ORCID

Deirdra R. Terrell https://orcid.org/0000-0003-0794-5851

James N. George https://orcid.org/0000-0002-4243-2691

Jessica A. Reese1,2

Jennifer D. Peck1

Zhongxin Yu3

Teresa A. Scordino3

David R. Deschamps4

Jennifer J. McIntosh4

E10 CORRESPONDENCE

Deirdra R. Terrell1

Sara K. Vesely1

James N. George1,2

1Department of Biostatistics & Epidemiology, College of Public Health,

University of Oklahoma Health Sciences Center

Oklahoma City, Oklahoma2Hematology-Oncology Section, Department of Medicine, College of

Medicine, University of Oklahoma Health Sciences Center

Oklahoma City, Oklahoma3Department of Pathology, College of Medicine, University of Oklahoma

Health Sciences Center

Oklahoma City, Oklahoma4Department of Obstetrics and Gynecology, College of Medicine,

University of Oklahoma Health Sciences Center

Oklahoma City, Oklahoma

Correspondence

James N. George, The University of Oklahoma Health Sciences Center,

Hematology-Oncology Section, Room CHB 237, P.O. Box 26901,

Oklahoma City, OK 73190.

Email: james-george@ouhsc.edu

REFERENCES

1. Reese JA, Peck JD, Deschamps DR, et al. Platelet counts during preg-

nancy. N Eng J Med. 2018;379:32-43.2. Jonsson V, Bock JE, Nielsen JB. Significance of plasma skimming and

plasma volume expansion. J Appl Physiol. 1992;72:2047-2051.3. Aster RH. Pooling of platelets in the spleen: role in the pathogenesis of

"hypersplenic" thrombocytopenia. J Clin Invest. 1966;45:645-657.4. Kaufmann P, Huppertz B, Frank H-G. The fibrinoids of the human pla-

centa: origin, composition and functional relevance. Ann Anatomy.

1996;178:485-501.5. Benirschke K, Burton GJ, Baergen RA. Pathology of the Human Placenta.

Heidelberg. Germany: Springer-Verlag; 2012.6. Othoro C, Moore JM, Wassenuehler K, et al. Evaluation of various

methods of maternal placental blood collection for immunology studies.

Clin Vacc Immunol. 2006;13:568-574.

Received: 5 July 2018 Revised: 12 October2018

Accepted: 15 October2018

DOI: 10.1002/ajh.25320

Effect of novel agents on therisk of early death of newlydiagnosed symptomaticmultiplemyeloma patients: A singleCentre retrospective analysis

To the Editor:

Incorporation of novel agents (ie, proteasome inhibitors, bortezomib

and carfilzomib, and immunomodulatory drugs, thalidomide and

lenalidomide) in the first line treatment of patients suffering from mul-

tiple myeloma (MM) has led to an unprecedented improvement in clin-

ical outcomes1. Nonetheless, a substantial proportion of patients still

dies early in the course of treatment, mostly due to infections. Male

gender, hypercalcemia, increased LDH serum level, decreased albumin

serum level, renal failure, low body mass index, high Charlson comor-

bidity index, thrombocytopenia and plasma cell leukemia have been

identified as predictors of an increased risk of early death (ED).2 It is

worth noting, that some of these risk factors exclude patients from

enrollment in clinical trials, thus making it difficult to extrapolate their

results to daily practice. In reality, if regulatory Agencies make novel

agents available outside of clinical trials, more and more elderly and/or

fragile patients are likely to receive them. Whether these patients also

benefit from programs containing one or more novel agents has not

been clearly established yet. We hereby analyzed the clinical out-

comes of 991 unselected newly diagnosed MM patients, consecu-

tively treated before and after the introduction of novel agents-

containing programs. We focused on their risk of ED, defined as the

death occurring within 90 days from diagnosis of symptomatic MM.

Details on source of data, treatments of MM and statistical analysis

are reported in Supplemental Material. The consort diagram of MM

patients diagnosed between 1997 and 2015 is shown in Supporting Infor-

mation Figure S1. The main demographic data of the patients who started

their medical treatment at our Institution is shown in Supplemental Table.

During the 18 years of analysis, 708 patients (71.4%) died. Of them,

92 patients died ≤90 days (ED cohort) (9.3% of the entire patients' popula-

tion and 13% of all deaths) from starting treatment (Supporting Informa-

tion Figure S1). By univariate and multivariate analysis, older age, lambda

light chain, stage B, and plasma cell leukemia were significantly more com-

mon in the ED group patients, whereas incorporation of at least a novel

agent in the first line treatment resulted in a significant reduction of ED

(Table 1). This beneficial effect was only seen in patients not enrolled in

high-dose treatment with autotransplantation (ASCT) programs, in whom

the incorporation of novel drugs reduced the prevalence of ED from 22%

to 9% (adjusted odds ratio 0.09, 95% CI 0.03-0.22, P < .0001). Conversely,

the prevalence of ED of patients enrolled in ASCT programs was 7%

irrespective of inclusion of novel agents in induction. Information about

the cause of ED was available in 65 (71%) patients: progression of MM

(n = 34), infections (n = 16), cardio-vascular complications (n = 13), hep-

atotoxicity (n = 1), other cause while in partial remission (n = 1).

ED has been defined quite differently in the studies. We set its cut-

off at 90 days from starting treatment, in order to focus on the role of

the first line treatment and to avoid the confounding effect of further

therapies on this clinical outcome. Indeed, among the ED cohort only one

patient received a second line treatment. The prevalence of ED in our

cohort closely resembles that reported by a tertiary care Korean Institu-

tion.3 These Authors found pneumonia to be the leading cause of ED of a

large group of newly diagnosed MM patients. Bacterial and viral infec-

tions show a bimodal incidence in patients with MM, the first peak occur-

ring during the induction phase and the second one at the time of

relapse/recurrence. This coincides with moments in which MM requires

treatment and is not yet under proper control. Although we were unable

to analyze in detail the role of infections as cause of ED in our cohort,

their contributing role is highly likely, since infections were documented

in approximately one quarter of our MM patients at the time of death.

CORRESPONDENCE E11