CORRESPONDENCE
TABLE I. Prevalence of HFE Mutations in Patients With Iron Overload Measured by MRI-T2*, by Affected Organ
Group (n 5 159) Liver
No HFE mutation (n 5 50) 12/48 (25%)
C282Y mutation (n 5 16) 12/16 (75%)
C282Y—homozigous/heterozygous 5/5 (100%)/7/11 (63.6%)
H63D mutation (n 5 83) 31/83 (37.3%)
H63D—homozigous/heterozygous 9/14 (64.3%)/22/69 (31.9%) 1/48 (2.1%)/1/11 (9.1%)
C282Y/H63D mutation (n 5 8) 5/8 (62.5%)
S65C mutation (n 5 2) 1/2 (50%)
with the aim of describing MRI-T2* findings in a large sample of patients, and correlat-
ing these with HFE genetic profiles. The local Ethics Committee approved the protocol.
We analyzed data from all the 159 patients admitted in the study period with sus-
pected iron overload based on high TS (above 55% in men and 45% in women) and/or
SF (> 322 ng/mL), who had undergone MRI-T2* for heart, liver, spleen, and/or pancreas
iron overload and had been screened for the presence of HFE mutations by allele-specific
PCR (polymerase chain reaction). The calculations of liver iron concentration (LIC) val-
ues were based on liver MRI-T2* measurements, using the Thalassemia-Tools software
(Cardiovascular Imaging Solutions, London, UK).
Mutations in the HFE gene were identified in 109/159 (68.6%) patients. The most
common mutation in our sample was H63D, present in 91 patients (57.2%): 14 (8.8%)
were homozygous, 69 (43.4%) heterozygous, and 8 (5%) compound heterozygous for
C282Y/H63D. For the C282Y mutation, in contrast, only 5 patients (3.1%) were homozy-
gous and 11 (6.9%) were heterozygous. The S65C mutation was detected in heterozygous
state in 2 (2.5%) cases.
All 159 patients underwent abdominal MRI-T2* and 126 underwent cardiac MRI-T2*
too. Only 3 out of 126 cardiac MRIs had a positive T2* result, mild cardiac overload
(T2*: 18.98, 19.14, and 19.8 ms). Of these, two patients had the H63D mutation (1
homozygous and 1 heterozygous) and one patient did not have any of the mutations
studied. In the liver, 61 (38.4%) patients had iron overload (T2*: < 11.4 ms and
LIC > 2.0 mg/g) of which 57 (35.8%) were light (T2*: 3.83–11.4 ms and LIC: 2.01–
6.86 mg/g), and four (2.5%) moderate (T2*: 2.0–3.8 ms and LIC: 7.06–13.56 mg/g). Of
these patients with liver overload, 27.9% were C282Y carriers (8.2% homozygous, 11.5%
heterozygous, and 8.2% compound heterozygous C282Y/H63D), and 50.8% carried the
H63D mutation (14.8% in homozygosis and 36.1% in heterozygosis). Only 12 (19.7%)
patients with liver overload did not have the HFE mutation.
The presence of C282Y mutation (in either homo or heterozygosis), compound heter-
ozygous (C282/H63D), and H63D in homozygosis was significantly associated with a
higher frequency of iron overload in the liver as measured by T2* (P 5 0.001). However,
this was not true in patients with H63D in heterozygosis or absence of mutation
(P 5 0.42), in which overload frequency was 68.4% and 29.1%, respectively.
Pancreatic overload was diagnosed in 33 patients (21%), and 56 patients (35.7%) had
splenic overload (Table I). The presence of the C282Y was associated with an overall
higher frequency of iron overload. There was also a relatively high frequency (37.3%) of
abnormal T2* values in H63D mutants both in the liver and in the spleen, and the fre-
quency of splenic iron overload in H63D mutants was similar to that associated with the
C282Y mutation.
SF results were available for 152 patients. Median SF was 647 ng/mL (72–13,625), and in
138 patients (90.8%) SF was abnormally high. Overall, in 28 patients (18.2%) serum levels
were higher than 1,000 ng/mL, in 80 patients (54%) they varied from 501 to 1,000 ng/mL and
in 30 (20.3%) they ranged from 324 to 500 ng/mL. Serum TS was obtained from 94 patients,
with a median of 42% (31–57) and elevated results in 32 (34%) of the tests.
Considering MRI findings as standard, SF was the most sensitive test (sensitivity
94.7%, specificity 11.8%), whereas TS was the most specific (sensitivity 34%, specificity
65.4%), indicating that they might be complementary. Sensitivity and specificity values
were similar in patients with and without HFE mutation.
Iron overload prevalence was different according to the affected organ or the type of
HFE mutation; over 50% of patients with liver iron overload carried the H63D mutation,
and two out of three patients who had cardiac iron overload were also H63D carriers. A
total of 38% of the H63D carriers presented with iron overload in the liver and spleen
and 22% in the pancreas, showing that this mutation alone might correlate with iron
overload. It is worth noticing, however, that the absence of HFE mutations does not rule
out the presence of other mutations associated with hereditary hemochromatosis.
Our study demonstrates, in a large sample of Brazilian patients, that MRI-T2* is a
non-invasive, accurate, and sensitive technique for the detection of low levels of iron
overload in patients with HH type 1. Excessive iron stores in the liver were detected in
39% of patients, showing that iron accumulation begins in the liver [6]. Given the obser-
doi:10.1002/ajh.24193
Heart Spleen Pancreas
1/46 (2.2%) 18/48 (37.3%) 9/48 (18%)
0 6/16 (37.5%) 5/16 (31.5%)
– 3/5 (60%)/3/11 (27.3%) 1/5 (20%)/4/11 (36.4%)
2/59 (3.4%) 31/83 (37.3%) 19/83 (22.9%)
8/14 (57.1%)/23/69 (33.3%) 2/14 (14.3%)/17/69 (24.6%)
0 0 0
0 0 0
vation that MRI is an accurate and safe tool to measure iron stores in these organs, we
believe that this technology should be incorporated in the investigation of suspected
cases of hemochromatosis and contribute to guide therapeutic decisions such as
phlebotomy.
REIJÂNE A. ASSIS,1 FERNANDO U. KAY,2 FABIANA M. CONTI,1 PAULO V. CAMPREGHER,1
GILBERTO SZARF,2 MICHELLI S. DINIZ,3 MORGANI RODRIGUES,1 RICARDO HELMAN,1
MARCELO B.G. FUNARI,2JOHN WOOD,4 NELSON HAMERSCHLAK1*
1
Department of Hematology, Hospital Israelita Albert Einstein, S~ ao Paulo, Brazil; 2Imaging
Department, Hospital Israelita Albert Einstein, S~ ao Paulo, Brazil; 3Clinical Research Insti-
tute, Hospital Israelita Albert Einstein, S~ao Paulo, Brazil; 4Division of Cardiology, Univer-
sity of Southern California, Children’s Hospital, Los Angeles, California
Conflict of interest: The authors declare that they have no conflicts of interest for this sub-
mission.
This is an open access article under the terms of the Creative Commons Attribution-Non-
Commercial-NoDerivs License, which permits use and distribution in any medium, pro-
vided the original work is properly cited, the use is non-commercial and no modifications
or adaptations are made.
*Correspondence to: Nelson Hamerschlak, Centro de Pesquisa Clınica, Instituto Israelita de
Ensino e Pesquisa Albert Einstein, Av. Albert Einstein, 627/520, S~ao Paulo (SP), Brazil,
CEP 05256-900. E-mail: hamer@einstein.br
Received for publication: 26 July 2015; Revised: 4 September 2015; Accepted: 9 September 2015
Published online: 11 September 2015 in Wiley Online Library
(wileyonlinelibrary.com)
DOI: 10.1002/ajh.24189
䊏 References
1. Brandhagen DJ, Fairbanks VF, Baldus W. Recognition and management of hereditary hemo-
chromatosis. Am Fam Physician 2002;65:853–860.
2. Adams PC. Population screening for hemochromatosis. Hepatology 1999; 29:1324–1327.
3. Brissot P, Troadec MB, Bardou-Jacquet E, et al. Current approach to hemochromatosis.
Blood Rev 2008;22:195–210.
4. Fleming RE, Ponka P. Iron overload in human disease. N Engl J Med 2012;366:348–359.
5. Moyer TP, Highsmith WE, Smyrk TC, et al. Hereditary hemochromatosis: Laboratory evalua-
tion. Clin Chim Acta 2011;412:1485–1492.
6. van Bokhoven MA, van Deursen CT, Swinkels DW. Diagnosis and management of hereditary
haemochromatosis. Br Med J 2011;342:c7251.
Autoimmune neutropenia of infancy: Data from the Italian
neutropenia registry
To the Editor: Neutropenia is characterized by a reduced Absolute Neutrophil Count
(ANC). Among Caucasians the lower normal limit of ANC in children up to the age of 1
year is 1.0 3 109/L, whereas from greater than 1 year to adulthood this limit is 1.5 3
109/L; neutropenia is defined as mild if ANC is between 1.0 and 1.5 3 109/L, moderate if
between 0.5 and 1.0 3 109/L, and severe if less than 0.5 3 109/L [1]. Autoimmune Neu-
tropenia of Infancy (AIN) is due to auto-antibodies against Human Neutrophil Antigens
(HNA). HNA-1 (FcgIIIb) is the most frequently involved, mainly in its isoforms HNA-1a
and HNA-1b. In this report, that is the second largest study ever conducted on AIN [2],
we presented the clinical characteristics of the disease of 157 AIN patients from the Ital-
ian Neutropenia Registry of the Associazione Italiana di Onco-Ematologia Pediatrica
(A.I.E.O.P.) and assessed the sensitivity of the indirect anti-neutrophil antibody test. The
methods are described in the on-line Supporting Information 1.
Historically the reported prevalence of AIN is 1/100,000 children under 10-year-old
[3], but this is probably an underestimation; in the analysis restricted to patients of the
present cohort diagnosed in western Sicily the incidence was 1 out of 6,300 live births.
Some characteristics of the study population are shown in Table I. The female/male
ratio was 3.6/6.4. The median age at onset was 0.70 year (range 0.005–4.59): in 82% of
the cases neutropenia appeared at up to 18 months of age. In three patients (1.9%) the
onset was at under 1 month of age: alloimmune neonatal neutropenia was excluded on
American Journal of Hematology, Vol. 90, No. 12, December 2015 E221
 CORRESPONDENCE

the basis of a negative cross match between maternal serum and paternal neutrophils. It Fifty-one AIN children underwent an extensive autoimmunity investigation: no positive
has been reported [2] that the onset of AIN is not possible at less than 1 month of age. case was observed.
Actually there have been three published exceptions [4,5] and three patients in the pres- At the time of the analysis 10/157 patients were lost at follow-up and 131 out of 147
ent series, so we can say that the appearance at this age is unusual but not impossible. (89.1%) recovered from neutropenia. The median age at resolution was 2.14 years with a
The frequency of children who were born preterm and then developed AIN was median length of disease of 1.30 years: the Kaplan Meier recovery curve is shown in Fig.
13.2%. Interestingly this figure is significantly higher (P 5 0.016) than that observed 1. In the group of recovered children 65.1% had a disease duration of 24 months and
(6.9%) in a cohort of 487 children consecutively hospitalized for various reasons during 89.9% recovered at 5 years of age; the remaining 10.1% had a spontaneous remission at
2014 in a pediatric center of Sicily. This finding is in keeping with one possible patho- 5.1–11.1 years of age.
genic hypothesis: that AIN is due to immaturity of the suppressor systems and that spon- When we analyzed the modality of recovery we found that in 67.4% there was a sud-
taneous recovery corresponds to the full development of suppressor T-cell function. At den resolution (stable maintenance of normal ANC after neutropenia period), whereas in
onset, median ANC was 0.45 (range 0.0–1.45 3 109/L), and the neutropenia was severe 32.5% there was a transient intermittent neutropenia phase (alternation of normal and
in 56.0%, moderate in 38.2% and mild in 5.7%. In 29.3% of the patients neutropenia was subnormal ANCs) lasting up to 24 months: in this group median time of definite recov-
diagnosed by chance: this is in line with the figure of 27% in another report [6], but ery from the first normal WBC was 0.65 years, with a range of 0.15–2.11 (standard error
higher than the classic study by Bux et al. (8%) [2]. Median WBC at onset was 6.1 3 of 0.48).
109/L. Leucopenia for age and monocytosis were present in 41.7% and 19.3% of cases, Overall 44.2% of the children were hospitalized for fever or ascertained infections but
respectively. only 9.6% suffered from severe infections without any long-term consequences. Antibiotic
Bone Marrow (BM) examination was done in 54 patients (34%), and in all but 2 it prophylaxis was never administered. Granulocyte Colony Stimulating Factor (GCSF) was
showed normal or increased cellularity with or without a relative paucity of the more administered (always “on demand,” for short periods and mainly in the case of severe
mature stages of granulocyte development: in 2 children a moderate decrease of myeloid infections) to 7.1% of the patients.
cellularity was observed. Among the parameters analyzed in the Cox model only early age at onset
The prevalence of selected IgA deficiency (a condition predisposing to autoimmunity) (P 5 0.0001873) and absence of monocytosis (P 5 0.009641) were associated with a signif-
was 3% and was significantly higher than that observed in a group of 470 laboratory con- icantly earlier recovery. A lower age at presentation was also associated with a reduced
trols (0.21%, P 5 0.001679). Increased Immunoglobulin G level, a probable consequence length of disease (P 5 0.028).
of augmented immune stimulation, was observed in 6% of the 133 evaluable patients. Diagnosis of AIN is frequently troublesome. The direct test for anti-neutrophil
auto-antibody detection has an elevated rate of false positives, and the indirect test has
TABLE I. Clinical Characteristics of the Patients elevated false negatives [2]. We assessed the sensitivity of the Granulocyte Immunofluo-
rescence Test (GIFT), after 1, 2, 3, and 4 assays: detection of circulating anti-
granulocyte antibodies was always performed with unselected donors (non-genotyped
Autoimmune for HNA). As shown in on-line Table I in Supporting Information the sensitivity of the
neutropenia of
test increased from 61.8% at the first determination to 81.8% at the fourth, thus sup-
infancy (157)
porting the hypothesis that repeated testing may remarkably improve the power of this
Male 64.3% diagnostic tool.
Median age at onset (years) 0.70 In conclusion our report after nearly two decades provides an update of the clinical
Median age at diagnosis (years) 1.06 presentation and outcome of AIN, which appears as a substantially benign and self-
Median age at resolution 2.14 limiting condition although remission may occur even after long-term course in late pedi-
Median duration (years) 1.30 atric age. Furthermore this study offers some new clinical insights such as a higher inci-
Recovery 89.1%
dence in preterms and frequent recovery pattern consisting of alternating normal and
Median WBC at onset (3109/L) 6.1
neutropenic values.
Median ANC at onset (3109/L) 0.45
Leucopenia at onset 41.7%
Monocytosis at onset 19.3% 䊏 Acknowledgments
Increased IgG at onseta 6.0%
Selected IgA deficiencya 3% ERG spa, Rimorchiatori Riuniti, Cambiaso & Risso, SAAR Depositi Oleari Portuali,
Severe infections 9.6% UC Sampdoria are acknowledged for supporting the work of the Clinical and Experimen-
tal Hematology Unit of G. Gaslini Institute. No specific funding was received for this
a study. The Sicilian Primary Immunodeficiency Association is acknowledged for support-
Data available on 133/157 patients.
ing the work of the Pediatric Oncology-Hematology Unit of A.R.N.A.S. Civico Hospital.
No specific funding was received for this study.

PIERO FARRUGGIA,1 FRANCESCA FIOREDDA,2 GIUSEPPE PUCCIO,3 LAURA PORRETTI,4

TIZIANA LANZA,2 UGO RAMENGHI,5 FRANCESCA FERRO,5 ALESSANDRA MACALUSO,1
ANGELICA BARONE,6 SONIA BONANOMI,7 SILVIA CARUSO,8 GABRIELLA CASAZZA,9
MIRELLA DAVITTO,5 ROBERTA GHILARDI,10 SAVERIO LADOGANA,11 ROSALBA MANDAGLIO,12
NICOLETTA MARRA,13 BALDASSARE MARTIRE,14 ELENA MASTRODICASA,15
LUCIA DORA NOTARANGELO,16 DANIELA ONOFRILLO,17 GIUSEPPE ROBUSTELLI,7
GIOVANNA RUSSO,18 ANGELA TRIZZINO,1 FABIO TUCCI,19
MARTA PILLON,20 AND CARLO DUFOUR2
1
Pediatric Hematology and Oncology Unit, A.R.N.A.S. Ospedale Civico, Palermo, Italy;
2
Clinical and Experimental Hematology Unit, G. Gaslini Children’s Hospital, Genova, Italy;
3
Department of Sciences for Health Promotion, University of Palermo, Palermo, Italy;
4
Flow Cytometry Service, Laboratory of Clinical Chemistry and Microbiology, IRCCS “Ca’
Granda” Foundation, Maggiore Hospital Policlinico, Milan, Italy; 5Ematologia del
Dipartimento Di Scienze Pediatriche, Ospedale Infantile Regina Margherita, Torino, Italy;
6
Department of Pediatric Onco-Hematology, University Hospital, Parma, Italy;
7
Fondazione MBBM, Clinica Pediatrica, Universit a Di Milano - Bicocca, Monza, Italy;
8
Servizio Di Epidemiologia E Biostatistica-Istituto, G. Gaslini Children’s Hospital,
9
Genova, Italy; Paediatric Hematology Oncology, Bone Marrow Transplant, Azienda
Ospedaliero Universitaria Pisana, Ospedale S. Chiara, Pisa, Italy; 10Department of
Pediatrics, IRCCS “Ca’ Granda” Foundation, Maggiore Hospital Policlinico, Milan, Italy;
11
Department of Hematology, IRCCS Casa Sollievo Della Sofferenza,
San Giovanni Rotondo, Italy; 12Pugliese-Ciaccio Hospital, Catanzaro, Italy; 13A.O.R.N.
Santobono Pausillipon, Naples, Italy; 14Dipartimento Di Scienze E Chirurgia Pediatriche,
U. O. Oncoematologia Pediatrica, Ospedale Policlinico- Giovanni XXIII, Bari, Italy;
15
Figure 1. Kaplan Meier recovery curve. Pediatric Oncology Hematology Unit, S. Maria Della Misericordia Hospital,

E222 American Journal of Hematology, Vol. 90, No. 12, December 2015 doi:10.1002/ajh.24193
CORRESPONDENCE
Perugia, Italy; 16Pediatric Oncology-Hematology and BMT Unit, Children’ Hospital,
Spedali Civili, Brescia, Italy; 17Hematology Unit, Hospital of Pescara, Pescara, Italy;
18
Pediatric Hematology and Oncology Unit, Policlinico Hospital,
University of Catania, Catania, Italy; 19Department of Pediatric Oncology-Hematology,
Meyer Children’s Hospital, Florence, Italy; 20Dipartimento Di Oncoematologia Pediatrica,
Universita Di Padova, Padova, Italy
Additional Supporting Information may be found in the online version of this article.
Conflict of interest: Nothing to report.
*Correspondence to: Piero Farruggia; Oncoematologia Pediatrica, A.R.N.A.S. Ospedali
Civico, Di Cristina e Benfratelli, Piazza Nicola Leotta 4, 90127 Palermo, Italy.
E-mail: piero.farruggia@arnascivico.it
Received for publication: 6 September 2015; Accepted: 9 September 2015
Published online: 11 September 2015 in Wiley Online Library
(wileyonlinelibrary.com)
DOI: 10.1002/ajh.24187
䊏 References
1. Dinauer MC. The phagocyte system and disorders of granulopoiesis and granulocyte func-
tion. In: Nathan DG, Orkin SH, editors. Nathan and Oshi’s Hematology of Infancy and
Childhood. Philadelphia: WBS Saunders Company; 2003. pp 923–1010.
2. Bux J, Behrens G, Jaeger G, Welte K. Diagnosis and clinical course of autoimmune neutrope-
nia in infancy: Analysis of 240 cases. Blood 1998;91:181–186.
3. Lyall EG, Lucas GF, Eden OB. Autoimmune neutropenia of infancy. J Clin Pathol 1992;45:431–434.
4. Calhoun DA, Rimsza LM, Burchfield DJ, et al. Congenital autoimmune neutropenia in two
premature neonates. Pediatrics 2001;108:181–184.
5. Lejkowski M, Maheshwari A, Calhoun DA, et al. Persistent perianal abscess in early infancy
as a presentation of autoimmune neutropenia. J Perinatol 2003;23:428–430.
6. Audrain M, Martin J, Fromont P, et al. Autoimmune neutropenia in children: Analysis of
116 cases. Pediatr Allergy Immunol 2011;22:494–496.
Pulmonary endarterectomy as treatment for chronic thromboembolic
pulmonary hypertension in sickle cell disease
To the Editor: Pulmonary hypertension (PHT) is defined as a resting mean pulmonary
arterial pressure (PAP)  25 mmHg by right heart catheterization (RHC). It is a common
complication of sickle cell disease (SCD), with an estimated prevalence of 6–11% [1]. The
pathogenesis of PHT in SCD is multifactorial with contributions from hemolysis, inflam-
mation, endothelial dysfunction, hypercoagulability, and thrombosis. Effective strategies
for the treatment of SCD-associated PHT have not yet been determined. Here, we report
the successful treatment of two cases of chronic thromboembolic pulmonary hypertension
(CTEPH) in SCD with pulmonary endarterectomy.
A 26-year-old male with HbSS complicated by frequent pain episodes, acute chest syn-
drome, and recurrent pulmonary thrombosis developed progressive dyspnea on exertion
Figure 1. Pulmonary angiogram demonstrating generalized hypoperfusion of the bilateral upper lobes and lateral basal segments with central vessel trunca-
tion (e.g., left lateral basal lobe branch indicated by red arrow) and filling defects (e.g., right upper lobe branch indicated by green arrow).
doi:10.1002/ajh.24193
with NYHA class IV symptoms by age 25. Pre-operative echocardiography showed a tricus-
pid regurgitant jet velocity (TRV) of 3.9 m/sec. Six-minute walk distance (6MWD) was
373 m (47% predicted). Ventilation-perfusion (V/Q) scan revealed multiple, bilateral
wedge-shaped perfusion defects. A pulmonary angiogram showed bilateral hypoperfusion
in the upper lobes and basal segments, and chronic occlusive changes (Fig. 1). RHC showed
a mean PAP of 41 mmHg and pulmonary artery occlusion pressure of 13 mmHg. The
patient had a RBC exchange transfusion followed by bilateral pulmonary endarterectomy
without complications. Nine months after surgery, a repeat echocardiogram showed a
decreased TRV of 2.2 m/sec and his 6MWD improved to 461 m (58% predicted). One year
after surgery, his symptoms have improved to NYHA Class II.
A 38-year-old female with HbSS complicated by frequent pain episodes, acute chest
syndrome, and recurrent pulmonary thrombosis developed progressive shortness of breath
with NYHA class IV symptoms despite chronic anticoagulation. Pre-operative echocardiog-
raphy showed a TRV of 4.3 m/sec. Her 6MWD was 263 m (43% predicted) and a V/Q
scan showed multiple, bilateral perfusion defects. A RHC revealed a mean PAP of 40
mmHg, with a pulmonary artery occlusion pressure of 11 mmHg. The patient underwent
RBC exchange transfusion pre-operatively followed by bilateral pulmonary endarterectomy
without complication. Two months after the procedure a repeat echocardiogram showed
an improved TRV of 2.5 m/s. Repeat V/Q scan also showed improvement in all old perfu-
sion defects. However, the scan also showed a new defect in the right upper lobe, likely rep-
resenting a new subacute thrombosis, despite the presence of an IVC filter. Her symptoms
improved, from NYHA Class IV to Class III, 6 months post-operatively.
Multiple studies in SCD patients show that nearly every aspect of hemostasis is altered in the
direction of a procoagulant phenotype. Although an area of ongoing research, several factors,
including hemolysis, ischemia-reperfusion injury, and RBC phosphatidylserine exposure are
reported to play a role in the pathogenesis of coagulation activation in SCD. Heme, an inflam-
matory mediator and a product of intravascular hemolysis, induces tissue factor (TF) expression
in cultured macrovascular and microvascular endothelial cells in a concentration-dependent
manner [2]. In addition, increased bioavailability of nitric oxide (NO) by pharmacologic and
genetic techniques has been shown to reduce endothelial TF expression in mouse models of
SCD [3].
While the contribution of coagulation activation to SCD-related complications has
not been extensively studied, pulmonary thrombosis may be a major risk factor for PHT
in patients with SCD. No significant associations were observed between plasma markers
of coagulation activation and echocardiography-derived TRV [4]. However, autopsy stud-
ies have demonstrated an increased incidence of pulmonary thrombosis in patients with
SCD. National and statewide hospital discharge data have also shown an increased inci-
dence of pulmonary embolism amongst SCD patients, but interestingly not an increased
incidence of deep vein thrombosis suggesting a high rate of in situ pulmonary thrombo-
sis. More recently, retrospective studies have confirmed the increased risk of venous
American Journal of Hematology, Vol. 90, No. 12, December 2015 E223