European Journal of Radiology 95 (2017) 141146

Contents lists available at ScienceDirect

European Journal of Radiology

journal homepage: www.elsevier.com/locate/ejrad

Research article

Partial anomalous pulmonary venous return in Turner syndrome MARK

a,b a,b c d
Allard T. van den Hoven , Raluca G. Chelu , Anthonie L. Duijnhouwer , Laurent Demulier ,
Daniel Devosd, Koen Niemanb, Maarten Witsenburga, Annemien E. van den Boscha,

Bart L. loeyse,f, Iris M. van Hagena, Jolien W. Roos-Hesselinka,
a
Department of Congenital Cardiology, Erasmus MC, Rotterdam, the Netherlands
b
Department of Radiology, Erasmus MC, Rotterdam, the Netherlands
c
Department of Cardiology, UMC Radboud University Medical Center, Nijmegen, the Netherlands
d
Department of Cardiology, UZ Gent, Gent, Belgium
e
Center for Medical Genetics, University of Antwerp/Antwerp University Hospital, Antwerp, Belgium
f
Department of Human Genetics, Radboud University Medical Center, Nijmegen, the Netherlands

A R T I C L E I N F O A B S T R A C T

Keywords: Purpose: The aim of this study is to describe the prevalence, anatomy, associations and clinical impact of partial
PAPVR anomalous pulmonary venous return in patients with Turner syndrome.
Turner syndrome Methods and results: All Turner patients who presented at our Turner clinic, between January 2007 and October
Computed tomography 2015 were included in this study and underwent ECG, echocardiography and advanced imaging such as cardiac
PAPVC
magnetic resonance or computed tomography as part of their regular clinical workup. All imaging was re-
evaluated and detailed anatomy was described. Partial anomalous pulmonary venous return was diagnosed in 24
(25%) out of 96 Turner patients included and 14 (58%) of these 24 partial anomalous pulmonary venous return
had not been reported previously. Right atrial or ventricular dilatation was present in 11 (46%) of 24 partial
anomalous pulmonary venous return patients.
Conclusion: When studied with advanced imaging modalities and looked for with specic attention, PAPVR is
found in 1 out of 4 Turner patients. Half of these patients had right atrial and/or ventricular dilatation.
Evaluation of pulmonary venous return should be included in the standard protocol in all Turner patients.

1. Introduction recognition of the syndrome [69]. However, PAPVR can cause an

hemodynamically signicant left-to-right shunt and subsequently, right
Turner syndrome (TS) is a partial or complete monosomy of the X- chamber dilatation, arrhythmias and even pulmonary hypertension
chromosome, and was originally described by Henry Turner in 1938. It [10]. This necessitates early diagnosis and, in case of a large shunt,
occurs in 1 per 2500 live born females [13]. Turner syndrome is treatment. In our tertiary center for Turner syndrome all adult patients
nowadays often diagnosed prenatally or during early childhood and undergo a CT scan, while children and adolescents will have an MRI
patients present with a broad variety of disorders, including short sta- [11].
ture, estrogen deciency and cardiovascular abnormalities [4]. Aside The aim of this study is to describe the prevalence, anatomy and
from a higher risk prole for ischemic heart disease, often congenital clinical signicance of PAPVR, in patients with Turner syndrome.
cardiac defects, located mainly on the left side, can be found in these Furthermore, factors, possibly related with PAPVR were studied.
patients. The most common lesions are: elongation of the aortic arch
(49%), bicuspid aortic valve (1430%), coarctation (718%), persistent 2. Methods
left superior vena cava (13%), and atrial and ventricular septal defect
(08%) [2,59]. PAPVR has been described to be more prevalent in TS Every Turner patient is enrolled into a prospective surveillance
patients, however knowledge about detailed anatomy and precise pre- program according to a local standardized clinical protocol. This pro-
valence is scarce, partly because it is not routinely looked for at rst tocol includes clinical assessment by a cardiologist, ECG,

Corresponding author at: Erasmus University Medical Centre, Department of Cardiology, Room Ba-583a, P.O. Box 2040, 3000, CA, Rotterdam, the Netherlands.
E-mail addresses: a.vandenhoven@erasmsusmc.nl (A.T. van den Hoven), r.saru@erasmusmc.nl (R.G. Chelu), Toon.Duijnhouwer@radboudumc.nl (A.L. Duijnhouwer),
Laurent.demulier@uzgent.be (L. Demulier), Daniel.devos@uzgent.be (D. Devos), k.nieman@erasmusmc.nl (K. Nieman), m.witsenburg@erasmusmc.nl (M. Witsenburg),
a.e.vandenbosch@erasmusmc.nl (A.E. van den Bosch), bart.loeys@uantwerp.be (B.L. loeys), i.vanhagen@erasmusmc.nl (I.M. van Hagen), j.roos@erasmusmc.nl (J.W. Roos-Hesselink).

http://dx.doi.org/10.1016/j.ejrad.2017.07.024
Received 21 February 2017; Received in revised form 19 June 2017; Accepted 29 July 2017
0720-048X/ 2017 Published by Elsevier Ireland Ltd.
A.T. van den Hoven et al. European Journal of Radiology 95 (2017) 141146

Fig. 1. Measurement of PAPVR. Panel A) 3D volume rendered reconstruction showing a PAPVR (*), the Aorta (Ao), Pulmonary Artery (PA), innominate vein (IV). Panel B and C) Sagittal
and coronal oblique views of how the connecting site area of the PAPVR (*) was measured. D) Cross section of the PAPVR (*) perpendicular to the vessel.

echocardiography and cardiac CT (adults) or MR imaging (children and which it originated; the site of origin. Right atrial and right ventricular
adolescents). For this study all adult patients, who were evaluated at dilatation was visually assessed on echocardiography images by two
the department of congenital cardiology between January 2007 and observers (A.T. and J.R.), blinded to CT/MR data, on apical 4-chamber
October 2015 with genetically proven Turner syndrome were included. view, parasternal long-axis and parasternal short-axis according to the
We excluded patients with insucient CT or MR imaging quality. The latest guidelines [12] (Fig. 2). To assess global RV systolic function the
study was approved by the medical ethical committee of our center. right ventricular fractional area change (RV FAC) was calculated, where
Informed consent was waived. a RV FAC < 32% was considered impaired [12].

2.1. Imaging 2.2. Computer tomography angiography (CTA)

For this study all CT and MR-images were re-evaluated by two in- For this retrospective study, most patients (n = 32, 33%) were
dependent investigators (A.H. and R.C.), blinded to each other. When scanned according to a now standard protocol where a retrospectively
more than one scan (CT or MR) was available per patient, the most ECG-triggered CTA was obtained on a dual source scanner (Force or
recent CT scan was used. Evaluation was done by multi planar re- Flash, Siemens, Germany). Sixty-ve ml of iodinated contrast material
construction, using the AquariusNet (TeraRecon, Inc, San Mateo, CA) (Iodixanol, GE, USA) was administered through an antecubital vein
software. The connecting site was dened as the location where the followed by a 40 ml 70/30% saline/contrast chaser, both at 5 ml/
anomalous pulmonary vein drained into the systemic venous system. At second ow rates. Images were reconstructed at each 2% for 040% of
this location vessel diameter and area were measured; perpendicular to the RR interval with 1.0 mm/0.6 mm thickness and interval. Also,
the anomalous vein, as proximal to the vessel it feeds into as possible images at each 5% of the entire RR interval were reconstructed at
but where it was still discernable as a separate vessel (Fig. 1.) The 1.5 mm with a 1.0 mm interval. Maximal aortic diameters in CTA were
anomalous vessel was then traced back to the lung lobe or lobes from generated using the double-oblique short-axis technique. Another 32

142
A.T. van den Hoven et al. European Journal of Radiology 95 (2017) 141146

Fig. 2. PAPVR as seen on CT. Panel A) Right upper pulmonary vein connecting (RUPV) to superior vena cava (*SVC); and the corresponding volume rendered image (Panel B) of the same
patient. Right inferior pulmonary vein connecting (RIPV). Panel C) Coronal oblique view of the left upper pulmonary vein connecting (LUPV) connecting to the innominate vein (*); and
the corresponding volume rendered image of the same patient (Panel D). Panel E) Sagittal oblique image showing the right upper pulmonary vein connecting to the superior vena cava (*)
and the corresponding volume rendered image of the same patient (Panel F).

143
A.T. van den Hoven et al. European Journal of Radiology 95 (2017) 141146

patients (33%) were scanned according to a ash protocol. Table 1

Baseline characteristics, *saturation was available in 49 patients.
2.3. Magnetic resonance imaging (MRI)
Patient characteristics (N = 96) Mean sd

The MR scans (n = 17) were made on a 3T (Discovery MR750, GE Age, y 35 13

Healthcare, USA, n = 4) or on a 1.5 T (Discovery MR450 scanner, Signa Height, cm 155 14
Weight, kg 65 17
HDxt or Signa Excite, GE Healthcare, USA, n = 13).
Systolic BP, mmHg 124 19
The spatial resolution varied amongst protocols; matrix size varied Diastolic BP, mmHg 76 13
from 160 160 to 320 192, and pixel size from 1.03 1.72 to Body Surface Area, m2 1.6 0.25
2.82 1.96 with a slice thickness from 1.6 mm to 2.2 mm. Body Mass Index, kg/m2 27 6
For the MR protocol a double dose Gadolinium-based agent was Saturation*,%O2 (N = 49) 99 1.4

used for the majority of patients, this was Gadovist (Bayern, Berlin,
Germany).
Table 2
Echocardiographic measurements. Presented as N = (%) and mean sd.
2.4. Electrocardiography
Measurement PAPVR N = 24 Non-PAPVR Total N = 96
ECGs were scored for signs of right atrial or ventricular abnorm- (%) N = 72 (%) (%)

alities. Right atrial dilatation was dened as: a P-wave in lead Right Atrial Dilatation P = 0.004
II > 0.25 mV. Right ventricular hypertrophy was dened as: an R/S No 13 (54.2) 61 (84.7) 74 (77.1)
ratio in V1 > 1, R in V1 > 7 mm, or a right bundle branch block Yes 11 (45.8) 11 (15.3) 22 (22.9)
RV1 > 15 mm. Right axis deviation was dened as a QRS-axis of more Right Ventricular P = 0.039
Dilatation
than 90.
No 17 (70.8) 65 (90.3) 82 (85.4)
Yes 7 (29.2) 7 (9.7) 14 (14.6)
2.5. Karyotype analysis Right Ventricular N = 22 N = 51 p-value
measurements
EDA 22.5 6.3 16.4 6.8 < 0.001
For purpose of statistical analysis karyotypes were divided into one
ESA 12.6 4.2 9.2 4.1 0.001
of seven dierent categories; monosomy (e.g. 45,X0), mosaicism for RV FAC 43.7 9.7 44.1 9.1 0.86
monosomy (e.g. 45,X/46,XX [9/41]), full or mosaic isochromosomes
(e.g. 45,X0/46,i(Xq)), polyploidy (e.g. 45,X0/47,XXX (100/2)), ring EDA: End diastolic Area, ESA: END systolic Area, RV FAC: Right ventricle fractional area
chromosomes (e.g. 45,X0/46X,+r(X) [36/14]) and karyotypes with Y- change, RA: Right Atrial: RV: Right ventricle.
material (e.g. 45,X0/46,XY/46,XX/47,XY [8/3/2/37]).
of 72 in the non-PAPVR group (29% vs 10%, p = 0.039). Adequate
2.6. Statistical analysis measurement of the RV FAC was possible in 73 patients and was con-
sidered impaired in 13 patients (19%). RA and RV area were measured
Continuous variables with a normal distribution were reported as in 75 patients, from which RV fractional area change was calculated
mean standard deviation. We reported a median and range in case (Table 2). Both the RV end-diastolic area (EDA) and the end-systolic
of non-normal distribution, checked using the Shapiro-Wilk test. area (ESA) were enlarged in the PAPVR group (p < 0.001 and
Categorical variables were summarized as frequencies and percentages. p = 0.001). The RV fractional area change (RV FAC) did not dier
The Chi squared test was used to detect associations between patient signicantly between the groups.
characteristics and imaging data, in case of an expected value < 5 a The ECGs showed right axis deviation (6%) in 6 patients, all 6 had a
Fischers exact test was conducted. Independent sample t-tests and one- PAPVR (p < 0.001). Two ECGs showed signs of right atrial enlarge-
way ANOVA were used to compare means. A p-value of less than 0.05 ment, one of the patients had a PAPVR, no signs of right ventricular
was considered statistically signicant. All statistical analyses were hypertrophy or dilatation were found.
performed using the Statistical Package for Social Sciences (Released
2012. IBM SPSS Statistics for Windows, Version 21.0. Armonk, NY: IBM 3.1. Anatomy
Corp).
Details of the anatomy of the PAPVRs are presented in Table 3. The
3. Results lung lobe of origin and the site where the PAPVR connects to the cir-
culation could be determined in all 24 PAPVR patients. The most
Between January 2007 and October 2015, 104 patients with TS common site of origin was the left upper lung lobe (n = 14, 58%) and
were referred to the outpatient clinic of our center for routine cardiac
screening. In eight patients the image quality of the CT/MR was in- Table 3
sucient and these patients were excluded from further analysis. No PAPVR, anatomy. IVC: Inferior Vena Cava, IV: Innominate Vein, SCV: superior Vena Cava,
dierences in baseline characteristics were found between included and PLVCS: Persistent Left Vena Cava.
excluded patients. The baseline characteristics of the 96 patients in-
Sites of Origin Connecting sites
cluded in our study are presented in Table 1. At least one PAPVR was
found in 24 of the 96 patients (25%) and had not been reported before (Lung segments) SVC IV Left Subclavian Vein PLVCS Total
in 14 (58%). In the majority of the patients a CT scan (n = 79, 82%) Right Upper Lobe 2 2
was performed, PAPVR was detected on 21 of these (27%). In the re- Right Middle Lobe 2 2 4
Right Lower Lobe 0 0
maining 17 patients, an MR was performed and in 3 (18%) a PAPVR Right Middle + Lower Lobe 2 2
was found. Right Upper + Middle Lobe 1 1 2
Echocardiographic imaging was of sucient or good quality in all
96 patients. Right atrial (RA) and ventricular (RV) dimensions were Left Upper Lobe 10 1 1 12
Left Lower Lobe 1 1
measured and are reported in Table 2. RA dilatation was found in 22 of
Left Upper + Lower Lobe 1 1
the 96 patients (23%), and associated with the presence of PAPVR. In 7 Total 7 14 1 2 24
of the 24 PAPVR patients the right ventricle was dilated compared to 7

144
A.T. van den Hoven et al. European Journal of Radiology 95 (2017) 141146

Table 4 but it is remarkable that only PAPVR of the supra cardiac type (i.e.
Karyotypes. pulmonary veins draining into the superior caval vein or innominate
vein) were observed, in concurrence with another study in a TS popu-
Karyotype PAPVR (N=, Non-PAPVR Total (N=, p-value
%) (N=, %) %) lation [2]. This is in contrast to what may be expected based on ndings
in non-Turner patients [16]. It seems that the supracardiac type of
(Mosaic) Monosomy X 19 (79.2) 36 (50) 55 (57.3) 0.008 PAPVR is an inherent part of the Turner cardiac phenotype, where in
Isochromosomes 0 (0) 17 (23.6) 17 (17.7) 0.005
non-Turner cohorts this supracardiac type is relatively uncommon [17].
Deletions 1 (4.2) 3 (4.2) 4 (4.2) 1
Polyploidy 1 (4.2) 4 (5.6) 5 (5.2) 1 Another remarkable observation is the absence of atrial septal defects,
Ring chromosomes 1 (4.2) 3 (4.2) 4 (4.2) 1 which are supposedly associated with PAPVR in the non-TS population,
Y material 0 (0) 3 (4.2) 3 (3.1) 0.570 predominantly with right-sided PAPVR [17,18].
Missing 2 (8.3) 6 (8.3) 8 (8.3) A single underlying cause for the wide spectrum of left-sided cardiac
Total 24 72 96 (100.0)
defects in TS, genetic or otherwise, has not yet been established. There
are however several theories; cardiac abnormalities could be caused by
the most common connecting site was the innominate vein (n = 14, either haemodynamical changes in utero or could have a multifactorial
58%). All PAPVRs were of the supra cardiac type, i.e. none of the or genetic origin. It is suggested that either a jugular lymphatic ob-
PAPVRs connected to the coronary sinus or inferior caval vein. The struction may lead to compression of outow structures [19], or that a
median area of PAPVR was 38 mm2 (inter quartile range: 83, range; left-sided blockage of the fetal circulation forms the explanation for the
3391 mm2) with a median diameter of 8 mm (inter quartile range: 8.7, defects [20]. The association between TS genotype and PAPVR phe-
range; 225 mm). The vessel area correlated with the RV EDA notype needs further study and knowledge about the various cardiac
(r = 0.49, p = 0.014). BAV was found in 18 (18.8%) patients and CoA abnormalities and their interrelation might play a meaningful role in
in 10 patients (10.4%). An arteria lusoria and a common origin of in- unravelling the etiology of the Turner cardiac phenotype.
nominate and left carotid artery were found in 4 (4.2%) and in 1 (1%)
patient respectively. BAV occurred more frequently in combination 4.1. Limitations
with PAPVR, (p = 0.023). For CoA there was no clear association with
PAPVR (p = 0.115). A possible inclusion bias may have been in place as only TS patients
from a tertiary center were included. However, we feel that this eect is
likely to be very small as all TS patients are referred for follow-up ir-
3.2. Genetics respective of complaints or symptoms and follow-up is done according
to a standard protocol.
The information on the exact karyotype was available in 88 (92%)
patients and shown in Table 4. The most prevalent karyotype was a 4.2. Conclusion
monosomy of the X-chromosome (45,X0), followed by a mosaic
monosomy pattern. PAPVR was observed more frequently in patients PAPVR is present in one quarter of patients with Turner syndrome,
with a 45,X0 monosomy or mosaic monosomy (19 in 55) when com- and often missed at rst diagnosis of the syndrome. Medical specialists
pared to the other karyotypes (3 in 33; 35% vs. 9%, p = 0.008). Re- involved in the care for TS patients, should include attention for PAPVR
markably, PAPVR was not seen in any of the seventeen patients with an in their standard work-up.
isochromosome (0% vs. 31%, p = 0.005). Of the isochromosome kar-
yotypes in this cohort all had deletion of the shorter p-arm (iso- Conict of interest declaration
chromosome-q), 11 were a mosaic form (e.g. 45,X0/46-X,i(X)(q10)
[24/5]), the other were of the non-mosaic type (e.g. 46,X,i(Xq)). None of the authors have a have nancial association that might
pose a conict of interest in connection with this article at the time of
4. Discussion submission. Especially no relationship with industry as meant in the
Relationship with industry Policy.
In this large group of Turner syndrome patients a surprisingly high
prevalence of PAPVR (25%) was found. This percentage is markedly Institutional review board approval
higher than in previous studies, where PAPVR prevalence, when re-
ported at all, is found to be between 2.9 and 18% [2,6,7,9,13]. The The study was approved by the medical ethical committee of the
higher incidence may be explained by the careful attention in our study, Erasmus medical center. Informed consent was waived.
but also by dierences in imaging techniques. Echocardiography is
probably less sensitive to pulmonary vein abnormalities than CT or MR, Acknowledgements
explaining a possible underestimation of PAPVR prevalence in earlier
studies. Contrast-enhanced CT and contrast-enhanced 3D MR angio- This study was funded by a grant of the Dutch Heart Foundation
graphy are both adequate techniques to assess the presence of a PAPVR (The Hague, The Netherlands, grant number: 2013T093).
[14] where CT of course carries the disadvantage of radiation.
In this cohort none of the newly diagnosed PAPVRs needed im- References
mediate surgical intervention, but adequate follow-up of these patients
is warranted to assess changes in right heart dimensions and pressures [1] H. Turner, A syndrome of infantilism, congenital webbed neck and cubitus valgus,
over time. PAPVR can become clinically relevant as adult-onset pul- Endocrinology 23 (1938) 556574.
[2] V.B. Ho, V.K. Bakalov, M. Cooley, et al., Major vascular anomalies in Turner syn-
monary hypertension has been reported in PAPVR patients [10,15]. We drome: prevalence and magnetic resonance angiographic features, Circulation 110
found some degree of right heart dilatation in half of the PAPVR pa- (12) (2004) 16941700.
tients. Therefore, we would like to stress the importance of including [3] K. Stochholm, S. Juul, K. Juel, et al., Prevalence, incidence, diagnostic delay, and
mortality in Turner syndrome, J. Clin. Endocrinol. Metab. 91 (10) (2006)
advanced imaging in the clinical protocol of Turner syndrome patients 38973902.
with specic attention for possible PAPVR and right heart dilatation. [4] C.A. Bondy, L.A. Matura, N. Wooten, et al., The physical phenotype of girls and
In addition to the clinical implications of having PAPVR, also some women with Turner syndrome is not X-imprinted, Hum. Genet. 121 (3-4) (2007)
469474.
pathophysiological aspects warrant further investigation. This study [5] C.A. Bondy, Congenital cardiovascular disease in Turner syndrome, Congenit Heart
was not designed to examine the PAPVR from a developmental aspect,

145
A.T. van den Hoven et al. European Journal of Radiology 95 (2017) 141146

Dis. 3 (1) (2008) 215. [13] D. Prandstraller, L. Mazzanti, F.M. Picchio, et al., Turner's syndrome: cardiologic
[6] L. Mazzanti, E. Cacciari, Congenital heart disease in patients with turner's syn- prole according to the dierent chromosomal patterns and long-term clinical
drome. italian study group for turner syndrome (ISGTS), J. Pediatr. 133 (5) (1998) follow-Up of 136 nonpreselected patients, Pediatr. Cardiol. 20 (2) (1999) 108112.
688692. [14] S.K. Prasad, N. Soukias, T. Hornung, et al., Role of magnetic resonance angiography
[7] H.K. Kim, W. Gottliebson, K. Hor, et al., Cardiovascular anomalies in Turner syn- in the diagnosis of major aortopulmonary collateral arteries and partial anomalous
drome: spectrum, prevalence, and cardiac MRI ndings in a pediatric and young pulmonary venous drainage, Circulation 109 (2) (2004) 207214.
adult population, AJR Am. J. Roentgenol. 196 (2) (2011) 454460. [15] S. Pauwaa, A. Farzaneh-Far, Isolated partial anomalous pulmonary venous return
[8] K.L. Dawson-Falk, A.M. Wright, B. Bakker, et al., Cardiovascular evaluation in with intact atrial septum: a rare but treatable cause of pulmonary hypertension in
Turner syndrome: utility of MR imaging, Australas. Radiol. 36 (3) (1992) 204209. adults, Eur. Heart J. Cardiovasc. Imaging 15 (7) (2014) 830.
[9] I. Gutmark-Little, K.N. Hor, J. Cnota, et al., Partial anomalous pulmonary venous [16] N.M. Ammash, J.B. Seward, C.A. Warnes, et al., Partial anomalous pulmonary ve-
return is common in Turner syndrome, J. Pediatr. Endocrinol. Metab. 25 (5-6) nous connection: diagnosis by transesophageal echocardiography, J. Am. Coll.
(2012) 435440. Cardiol. 29 (6) (1997) 13511358.
[10] E.H. Sears, J.M. Aliotta, J.R. Klinger, Partial anomalous pulmonary venous return [17] F. Senocak, S. Ozme, A. Bilgic, et al., Partial anomalous pulmonary venous return.
presenting with adult-onset pulmonary hypertension, Pulm. Circ. 2 (2) (2012) Evaluation of 51 cases, Jpn. Heart J. 35 (1) (1994) 4350.
250255. [18] S. Kivisto, H. Hanninen, M. Holmstrom, Partial anomalous pulmonary venous re-
[11] Akker ELTvd. De richtlijn Syndroom van Turner is ontwikkeld op initiatief van het turn and atrial septal defect in adult patients detected with 128-slice multidetector
Nederlands-Vlaamse multidisciplinaire netwerk voor Turner syndroom. Oktober computed tomography, J. Cardiothorac. Surg. 6 (2011) 126.
2012. [19] E.B. Clark, Neck web and congenital heart defects: a pathogenic association in 45 X-
[12] R.M. Lang, L.P. Badano, V. Mor-Avi, et al., Recommendations for cardiac chamber O Turner syndrome? Teratology 29 (3) (1984) 355361.
quantication by echocardiography in adults: an update from the American Society [20] E.A. Shinebourne, A.M. Elseed, Relation between fetal ow patterns, coarctation of
of Echocardiography and the European Association of Cardiovascular Imaging, Eur. the aorta, and pulmonary blood ow, Br. Heart J. 36 (5) (1974) 492498.
Heart J. Cardiovasc. Imaging 16 (3) (2015) 233270.

146