Semin Neonatol 2001; 6: 49–61

doi:10.1053/siny.2000.0037, available online at http://www.idealibrary.com on

Diagnosis of patent ductus arteriosus

Jon Skinner

Department of Paediatrics, Green Lane This review shows that clinical signs cannot be relied on to reveal left-to-right ductal
Hospital, Auckland, New Zealand shunting in the preterm. Echocardiography can define many of the haemodynamic
features, including occult ventricular dysfunction, and categorize shunts into small,
moderate or large, without defining ‘significance’ which is multifactorial. Large shunts
have an elevated left atrial:aortic root ratio (>1.3:1 or >1.5:1, depending on fluid and
diuretic policy) a ductal diameter >1.4 mm and retrograde diastolic flow in the
descending aorta exceeding 30% of the antegrade flow.
Key words: ductus arteriosus, The neonatal clinician and echocardiographer is reminded to remain wary of occult
diagnosis, echocardiography, congenital heart disease, particularly before closing the ductus.
Doppler, prematurity  2001 Harcourt Publishers Ltd

Introduction gradient across the duct, between the aorta and

the pulmonary artery. The classical teaching was
that since the pressures are balanced, the pressure
Echocardiography has made the detection of a
gradient across the ductus is low, so that a murmur
patent ductus arteriosus (PDA) very straight-
is unlikely to be generated even by a large ductus.
forward. However, deciding when a PDA is patho-
Then later, as pulmonary vascular resistance falls,
logical or harmful (or ‘significant’) is more difficult.
a left-to-right shunt develops across the ductus
This article addresses the problem of defining
(generating a murmur) which may have potentially
when a left-to-right ductal shunt has reached clini-
detrimental haemodynamic [8,9] or respiratory [10]
cal or haemodynamic significance, and gives some
consequences.
guidelines. It also describes pitfalls in clinical and
We now know that this fall in pulmonary
echocardiographic diagnosis.
vascular resistance can occur very early—even in
the first few hours—especially in ventilated infants
below 1000 g. This important observation is dis-
Background cussed further by Kluckow and Evans in chapter 7
in this journal. It was formerly believed that once a
Ductal patency is normal in healthy term and left-to-right shunt was large, clinical signs (in par-
preterm infants up to the end of the third day [1,2]. ticular a murmur) would be associated with it and
However, in about half of preterm infants with reveal its presence to the attentive clinician [11].
respiratory failure the ductus remains patent after However, it is now clear that some infants with a
the third day, commonly until the end of the first large left-to-right ductal shunt have no clinical
week [3–6]. Compared to healthy infants, pulmon- signs of ductal shunting, and in particular no
ary arterial pressure falls more slowly in those with murmur [3,12–15].
hyaline membrane disease [5,7] and aortic pressure
rises more slowly [5] so that there is little pressure
The ‘winking ductus’
Correspondence to: Dr Jon Skinner MD, MRCP (UK), DCH, FRCPCH,
Paediatric Cardiologist, Green Lane Hospital, Green Lane West, Auckland,
New Zealand. Tel.: +64 9 6389909; Fax +64 9 631 0785; E-mail: During my training in neonatology, I often heard
[email protected] the expression that a ‘ductus has opened up’ when
1084–2756/01/010049+13 $35.00/0 © 2001 Harcourt Publishers Ltd
50 J. Skinner

an infant presented with a systolic murmur, typi- rhage (IVH) if the left-to-right shunt develops in
cally in association with cardiomegally and pro- the first few hours [8,20].
longed ventilation requirements. In fact it has been
known for a long time that, in these infants, the
ductus had usually remained patent from birth rather Defining ‘haemodynamic significance’
than opened up after closing. McGrath et al. had Accepting then that a PDA is the norm in the first
used aortography for their study in 1978 (‘the week or so—when is it abnormal to have a PDA
silent Ductus’ [3]), but more recent echocardio- and when does it matter? What is meant by a
graphic studies have confirmed that the ductus ‘haemodynamically significant duct’?
does not open and close periodically—once a duct
shuts spontaneously, it almost invariably stays shut
Significance from the clinician’s point of view
[15]. The ‘winking ductus’ which opens and closes
is, in fact, most unusual—though it can occur rarely A problem here is that the significance of a PDA
in cases of extreme prematurity (less than 25 means different things to different people. Is a
weeks, or under 700 g) following pharmacological left-to-right ductal shunt significant because it
ductal closure [15]. may be associated with a statistically significant
Echocardiographic studies in ventilated infants increased risk of IVH, or because it is known
of less than 28 weeks’ gestation have suggested to reduce respiratory compliance? Is it only sig-
that pulmonary vascular resistance tends to falls nificant when it causes overt cardiac failure, or
more rapidly than in the more mature infants [5], is it significant when certain echocardiographic
manifesting with early development of a large criteria are met—such as the size of the left atrium
left-to-right ductal shunt, mostly within the first 2 [24]?
days [4,6,16,17], and initially without overt clinical The management of PDAs varies on the defini-
signs. tion of significance. Some neonatal units try to
close all ducts prophylactically in extreme
prematurity—to prevent the development of a
The ‘silent dangerous ductus’ left-to-right ductal shunt in any infant. Others
screen for a large left-to-right ductal shunt at a
Echocardiographic features of a large left-to-right certain age and base treatment on the echo-
ductal shunt usually predate clinical signs by cardiogram. Others only close those that present
between one and seven days in infants of 27–34 clinically with signs of a murmur or with cardiac
weeks’ gestational age [6,16–18]. One study [12] failure.
found that such infants can collapse due to the ‘Significance’ is a matter of opinion, which not
large shunt following extubation, without having only varies between neonatologists, but also varies
detectable clinical signs of ductal shunting. The between patients. Some infants with a large shunt
term ‘silent dangerous ductus’ was therefore on echocardiography (classified in many papers
coined. The increased pulmonary blood flow is as haemodynamically significant PDA [HsPDA])
associated with decreased pulmonary compliance seem to do just fine, never present clinically with
[10], such that the loss of continuous positive physical signs, and the duct closes spontaneously.
airway pressure (CPAP) to maintain small airway Others do very badly, with early signs of low
patency results in rapid deterioration following cardiac output, respiratory failure or IVH.
extubation. There is also increasing evidence that Many studies have described the echocardio-
some infants have inadequate systemic cardiac graphic features of clinically diagnosed left-to-right
output with a clinically silent left-to-right ductal ductal shunting in preterm infants. Others have
shunt [9,19,20,22]. Much of the left ventricular established at what age the presence of these
output is diverted into the pulmonary circulation echocardiographic features predicts later presenta-
(‘stolen’ from the systemic circulation) and failure tion with an overtly symptomatic ductal shunt
to increase the left ventricle (LV) output to com- [4,6,7,16,17]. Some of these echocardiographic
pensate for this results in low systemic cardiac features are discussed here, but I will avoid trying
output. If myocardial failure, relatively common in to apply ‘significance’ to any set of criteria. You,
the very preterm [19,23], is combined with this the reader, can do this later having decided which
ductal steel, then there may be serious adverse type of duct in which type of infant is causing
consequences, including intraventricular haemor- harm and, therefore, worthy of treatment, in your
Diagnosis of patent ductus arteriosus 51

opinion. Chapters 6 and 7 in this publication individual physical signs had a sensitivity greater
discuss this in more detail. than 50%. The presence of a heart murmur or an
overactive precordium were, however, relatively
specific when present (up to 87% [26]), but the
incidence of heart murmur with a large left-to-right
Detecting a left-to-right ductal shunt varied from 0–20% in the first 2 days of life
shunt [25,26] to 80% at 5–7 days when taking care to
switch off the ventilator before listening intently
Clinical signs—throw away the [15]. The most reliable combination of signs was
stethoscope? an overactive precordium and a murmur with a
positive predictive value at best of 77% [26].
In general, echocardiographic signs of a ductal
Those who have performed echocardiography in shunt preceded the development of overt clinical
the preterm will have no doubts about the lack of signs by an average of 3 days and up to 1 week.
reliable agreement between the size of a left-to- In summary, the presence of classical physical
right ductal shunt determined echocardiographi- signs means that there is usually (but not always)
cally with that determined clinically. For those who a big left-to-right shunt (and we should not
are not convinced, it may be useful to review the throw away the stethoscope), but their absence
research into the sensitivity and specificity of does not exclude a large shunt, especially in the
clinical symptoms, signs and radiography in detect- first 2 days.
ing a large left-to-right ductal shunt, and to try to
explain the apparent disparity between echocardio-
graphic findings and clinical signs. Why are clinical signs unreliable?

Clinical signs and symptoms of a large

Looking at the individual clinical signs one by one,
left-to-right ductal shunt there are some plausible physiological explanations
why each may be absent with a large left-to-right
Overt clinical presentation is that of cardiac failure, ductal shunt.
occurring typically during recovery from the acute
phase of hyaline membrane disease, after day 3 Heart murmur from a PDA
with persistent ventilatory or oxygen requirements.
The classical physical signs are: The typical ductal murmur is systolic (in about
75%), but can be continuous (in 25%) [15] and is
1. a murmur usually best heard at the upper left sternal edge
2. tachycardia under the clavicle. It is not completely clear why a
3. wide pulse pressure presenting as bounding big shunt may have no murmur but a large shunt
pulses may be evident on echocardiography for some
4. an overactive precordium days before a murmur develops [16,18].
5. a mid-diastolic murmur at the apex or a gallop One study compared Doppler echocardiograms
rhythm from preterm infants with a large shunt and a
6. hepatomegaly. murmur (n=55), to those with a large shunt and no
A chest radiograph reveals pulmonary plethora, murmur (n=62) [27]. The main difference between
cardiomegaly and a wide angle between the left the two groups, who had similar left atrial dimen-
and right main bronchi due to left atrial dilatation. sions, was that those infants with a murmur tended
There is frequently a mild metabolic acidosis. to have a higher left ventricular output (mean 501
vs 352 ml/kg/min) and higher left ventricular stroke
Comparing clinical signs with echocardiography volume (3.3 vs 2.3 ml/kg). The pattern of flow
through the duct was typically highest in late
Several studies looked at the sensitivity and spe- systole and very low in late diastole (ref Fig. 4C2).
cificity of clinical signs using echocardiography as A murmur was less likely in the presence of
the gold standard [12–15,25,26]. In general, these continuous high velocity left-to-right flow. This is
showed that clinical signs were unreliable with presumably because the ductus is effectively
sensitivity varying from 37% [14], to 72% [12] and restricting flow across it as it constricts and the
radiography contributed little [26]. None of the volume of flow is insufficient to generate a noise
52 J. Skinner

Following indomethacin or Ibuprofen a systolic

heart murmur may persist—leading to the pre-
sumption that the duct remains patent. Careful
examination may reveal that the murmur goes
through to the back more than before, and that
the chest radiograph seems to show less pul-
monary plethora. This is classical of branch
pulmonary artery stenosis—readily detected on
echocardiography—and a common physiological
occurrence in the preterm after ductal closure [28];
a further course of indomethacin is not indicated!
A pansystolic murmur may be uncommonly be
due to mitral regurgitation, which is usually a
transient functional phenomenon related to myo-
cardial dysfunction before and/or after ductal
closure.
In summary, a murmur is a late sign of a large
left-to-right ductal shunt and although relatively
specific, may represent pathology other than a
patent ductus.1

Tachycardia
I was taught that tachycardia is a useful sign of
ductal shunting. Animal studies of cardiac output
Figure 1. (A) An echocardiographic view from the regulation showed that newborns can alter stroke
suprasternal notch in a newborn infant with aortic volume less than adults. In 1983 Rudolph stated
coarctation and and patent ductus. The acceleration in flow
in the descending aorta, beyong the coarctation (and the ‘During the early neonatal period, the cardiac
ductus) is shown as a mosaic, turbulent pattern with colour output is heart rate dependent’ [29]. Since cardiac
Doppler. (B) In the same infant, a continuous wave Doppler failure in infants is usually associated with tachy-
tracing in the descending aorta reveals continuous forward cardia, and since infants were thought not to be
flow in diastole. Such continuous forward flow would also able to raise stroke volume to increase cardiac
be found using pulsed Doppler in the descending aorta.
output, it was believed that preterm neonates could
only respond to left-to-right ductal shunting by
(even though the velocity is higher). Such continu- increasing heart rate. This was a misconception
ous high velocity flow is the norm prior to ductal because the haemodynamic situation during those
closure in healthy infants (see Fig. 1—ductal flow, animal studies was different entirely; preload and
high velocity continuous vs high flow type). afterload to the left ventricle had not been altered
Perhaps infants with a murmur are those who are because the ductus was usually tied before the
able to maintain a good left ventricular output experimentation began.
and, therefore, quantitatively have more blood In the human preterm with a big left-to-right
traversing the ductus. Paradoxically, therefore, the ductal shunt, the preload is greatly increased due to
absence of a murmur may prove to be a more increased pulmonary venous return, and afterload
sinister sign than its presence. is greatly reduced; blood leaving the left ventricle
pours into the low resistance pulmonary circulation
Heart murmurs from causes other than a PDA via the ductus.
In a serial study of preterm infants during
It needs to be remembered that a systolic murmur therapeutic ductal closure, there was barely a dis-
can arise from a number of different causes. It is cernible difference between heart rates (Fig. 2). The
dangerous to assume it comes from a ductus. There left ventricular stroke volume, on the other hand,
may be congenital heart disease such as pulmonary fell by over half. The increased left ventricular
or aortic stenosis for example, or an outflow output due to the ductal shunt is in fact achieved
murmur from cardiac hypertrophy secondary to 1
Persistence of a murmur after antiprostaglandin therapy may be related
maternal diabetes or steroid therapy. to branch pulmonary artery stenosis or mitral regurgitation.
Diagnosis of patent ductus arteriosus 53

diastolic aortic pressure is reduced. However, ven-

tilated preterm infants less than 1500 g or 30
weeks’ gestation are less able to maintain the
systolic pressure with a large left-to-right ductal
shunt [9,21]. Both diastolic and systolic pressure are
reduced due to failure of the left ventricle to
adequately elevate LV stroke volume [9] and thus
the pulse pressure remains unchanged. Systolic
pressure only returns to normal after ductal closure,
and diastolic pressure rises with it.
Hence, wide pulse pressure is not a reliable sign
of a left-to-right ductal shunt in very low-birth-
weight infants.

Hyperactive precordium and mid-diastolic murmur

A hyperactive precordium and a mid-diastolic flow

murmur (from increased flow across the mitral
valve) are typical with any large left-to-right shunt.
Tuning in to a mid-diastolic murmur with a stetho-
scope at heart rates of over 160 beats per min in a
noisy incubator can be very difficult, but precordial
hyperactivity is easily seen.
There are many other things which can cause the
precordium to be hyperactive- particularly sepsis,
CO2 retention (through vasodilation) and inotropes
and vasodilators. It is hyperactive because there is
an increased stroke volume. Since the precordium is
not always overactive with a large ductal shunt in
Figure 2. Change in heart rate and stroke volume in 10 the preterm one might guess it is because the
preterm infants (26–32 weeks, 1250–2100 g) before
indomethacin and during tight constriction after
cardiac output is not so elevated. Perhaps these
indomethacin. Heart rate changes little in most infants, physical signs, like the murmur, tend to occur with
rising in some and falling in some, whereas stroke volume a ductus when there is a healthy myocardial
falls in every infant. response to the increased demand for blood flow,
with high cardiac output, and tend to be absent
by an increase in stroke volume, not heart rate. when there is an element of myocardial failure.
Many other studies have shown this phenomenon; Despite these problems, a hyperactive pre-
LV stroke volume can double or even triple with a cordium is one of the more specific, though not
large shunt and a healthy myocardium [4,30–32]. sensitive, signs of a large left-to-right ductal shunt
In summary, change in cardiac output in and a mid-diastolic murmur, or a gallop rhythm, is
response to changing ductal shunting is mostly an important clinical sign of a large shunt when
mediated by a change in stroke volume, not heart present.
rate. Since we also know that tachycardia is also a
Hepatomegally
feature of hypovolaemia, stress of various types,
and most inotropic drugs, it is safe to conclude that The size of the liver is difficult to assess accurately
changes in heart rate do not reliably reflect changes in a tiny preterm neonate and may be pushed down
in left-to-right ductal shunting. from positive pressure ventilation. However, the
true size of the liver reflects both right atrial pres-
Bounding pulses sure and the state of hydration. Fluid restriction not
A bounding pulse is caused by a wide pulse only reduces left atrial size on echocardiography,
pressure, i.e. a big difference between the systolic but also reduces the size of the liver, and body
and diastolic aortic pressure. In a large infant with weight in general. The duct can be huge, but if
a PDA the systolic pressure is maintained whilst blood volume is depleted enough, the left and right
54 J. Skinner

atrial pressure will be low, and the liver will be a. good foot pulses should be palpable
small. b. the aortic arch should be seen in its entirety
True hepatomegally is difficult to assess but can c. there should be pure left-to-right ductal flow
be a useful sign of left or right cardiac failure and on Doppler
fluid overload. However, its absence does not indi- 2. Pulmonary atresia, or critical pulmonary valve
cate absence of a large left-to-right ductal shunt. stenosis can present with high systemic satura-
tions, well into the 90 s when there is a large
duct, and the pulmonary valve can look remark-
Echocardiography ably normal. In the routine be sure therefore
that:
There is no better way to assess ductal shunting a. 100% systemic arterial oxygen saturation is
than with echocardiography. However, this is not a easily achieved
skill which can be learnt overnight, and someone on b. the pulmonary valve opens normally
each neonatal unit needs to specialize and develop 3. The left ventricle is usually hyperdynamic with
expertise in echocardiographic haemodynamic a left-to-right ductal shunt (fractional shortening
assessment under the guidance of a paediatric of the left ventricle greater than 40%). If it is
cardiologist. not, be sure the cause for ventricular dysfunc-
tion is not aortic stenosis, aortic coarctation or
systemic hypertension (usually from a renal
Missing or mis-diagnosing congenital cause).2
heart disease

Many units may not have the luxury of a paediatric Assessing ductal shunting
cardiologist to do all of their PDA echocardio-
grams. Amongst preterm neonates on neonatal Once the heart is confirmed to be structurally
units, however, congenital heart disease (CHD) is normal, a logical approach to assessing ductal
rare, whereas PDAs are the norm. Important CHD shunting is to:
is likely to present with overt clinical signs— 1. Establish ductal patency
cyanosis, heart failure, abnormal pulses etc.—and a 2. Determine the direction of the shunt and the
good cardiovascular examination (along with blood pattern of flow during the cardiac cycle
gases, chest X ray and echocardiography [ECG]) 3. Assess the internal diameter of the ductus
will reveal most cases which can be referred on to 4. Evaluate the volume loading on the heart,
the cardiologist. Occasionally, however, clinically noting the size of the foramen ovale, ventricular
unsuspected CHD is detected first in a routine function and cardiac output
haemodynamic study. 5. Evaluate the peripheral circulatory effects
After a thorough clinical examination, the 6. Put the evidence together to assign
echocardiogram should be done in a logical and a. The size of the left-to-right shunt (small,
thorough sequence, identifying venous drainage, moderate or large)
each chamber, septums, the outlets and the valves b. Left ventricular function (normal, moderately
[33]. Structural normality of the heart must be impaired, severely impaired)
proven, not assumed. In particular, duct dependent
lesions must not be missed since pharmacological Is the duct patent?
ductal closure can be fatal. Most of these have
clinical features different from the preterm with It is usually possible to obtain a view of the entire
a suspected left-to-right ductal shunt, not least duct (see Fig. 3) using a parasternal view roughly
cyanosis or weak rather than prominent pulses. half way between a standard ‘short axis’ view, and
Most also have obvious echocardiographic abnor- a full view of the aortic arch. At birth the duct has
malities, but some can be more subtle. There are no constriction and is as wide as the descending
some useful pointers: aorta, just as in the fetus. Ductal constriction
typically begins at the pulmonary end or in the
1. Early aortic coarctation can be difficult to detect,
middle of the duct [34].
(see Fig. 1) and can worsen as the duct closes. As 2
Do not assume absence of congenital heart disease. In particular, foot
a routine therefore prior to antiprostanglandin pulses should be strong and cyanotic heart disease must not be suspected
therapy: clinically (and 100% SaO2 achievable) before ductal closure.
Diagnosis of patent ductus arteriosus 55

systole, left-to-right in diastole) gradually changes

to pure left-to-right flow, with the left-to-right
velocity being high throughout the cardiac cycle
(see C1 in Fig. 4) [1]. If the ductus remains large as
pulmonary arterial pressure falls, the flow pattern
typically changes to pure left-to-right flow but
with most left-to-right flow occurring in systole,
and very little during diastole (see, C2 in Fig. 4).
The low velocity flow in diastole demonstrates that
the aortic and pulmonary arterial pressures are
almost equal in diastole. This is because left atrial
pressure is now very high, (so that diastolic PA
pressure is high) and aortic diastolic pressure is low,
due to run off into the ductus. This pattern never
occurs in healthy infants.
Pure right-to-left flow indicates that there is
likely to be elevated pulmonary vascular resistance
and anti-prostaglandins should be avoided in this
situation.

Assess the internal diameter of the ductus

Measuring the internal diameter of the duct can be

difficult even when clear images are obtained and is
Figure 3. (A) This echocardiograph, from the high left helped by measuring the width of the colour
parasternum, shows the ductus in its entirety. (B) Colour Doppler jet within the duct. Using this method in
Doppler shows the left-to-right flow as an orange flare, healthy term infants, minimal internal diameter of
seen as the lighter shade in this black and white photo, and the duct was 4.20.6 mm at 2 hr of age and
is very large (3.3 mm) in this 2 kg infant. 2.30.5 mm at 12 hr [37]. A study in preterms
less than 1500 g suggested that a minimal colour
Colour Doppler makes visualization of the duct map diameter of greater than 1.5 mm was neces-
easier. Typically a bright flare is seen along the sary before an important left-to-right ductal shunt
anterior portion of the main pulmonary artery (see could occur [16], median size at clinical diagnosis
Fig. 3), creating turbulence which is also detectable was between 2.5 and 3 mm. At a mean age of
with either pulsed or continuous wave Doppler. 19 hr, a ductal diameter of >1.4 mm predicted
A crude grading of the size of a left-to-right subsequent overt clinical presentation with 81%
shunt can also be made just from the width and sensitivity and 85% specificity [6].
length of this flare. Similar turbulence can rarely be It needs to be remembered though that the
due to aorto-pulmonary collateral arteries in infants ductus may have different measurements in differ-
with chronic lung disease [36] so it is important to ent planes, and the variability of colour Doppler
see the duct.3 maps and gain settings can cause inaccuracies in
measurement. Furthermore, the size of the shunt
Direction of flow through the duct
depends not only on the size of the duct, but also
on the balance of resistances on each side of it;
The direction and pattern of flow through the duct there is a huge duct at birth in term infants, but
is determined with Doppler techniques (see Fig. 4). very little left-to-right shunt.
If the post-natal ductus constricts normally as
pulmonary arterial pressure falls and aortic pressure Evaluate the volume loading on the heart,
rises, then bidirectional ductal flow (right-to-left in noting the size of the foramen ovale, ventricular
function and cardiac output
3
Diastolic turbulence in the main pulmonary artery can occasionally arise
from aorto-pulmonary collateral arteries in babies with chronic lung
disease, or more rarely from aorto-pulmonary window or anomalous The most striking feature of the cross-sectional
coronary artery origin. Thus, it is important to image the duct. echocardiogram with a big left-to-right ductal shunt
56 J. Skinner

is left atrial and left ventricular dilatation. The left

ventricle becomes globular and the interatrial
septum bows to the right (see Fig. 5).
Left atrial dilatation can be semi-quantified by
comparing its diameter with the diameter of the

Figure 5. (A) Subcostal four chamber echocardiographic

view to show the volume loaded appearance of the left
heart with a large left-to-right ductal shunt. Both the left
atrium and ventricle are dilated and note how the atrial
septum bulges to the right. (B) Parasternal long axis view
to show the dilated left atrium and ventricle, diastole on
the left and systole ion the right. Note that the left atrial
diameter is about twice the diameter of the aortic root in
systole, and note the good LV contraction.

aortic root; this allows comparison of values

between infants of different size. A well-evaluated
method is to use m-mode echocardiography from
the left parasternum (see Fig. 6). In 1974 Silverman
et al. [24], using m-mode echocardiography, found

Figure 4. Diagrammatic representation of changes in

pattern of ductal flow, determined by pulsed Doppler, over
the first few days of life in a healthy preterm infant and in
one with respiratory failure in whom the ductus does not
close. (A) Bidirectional flow, occurs immediately after birth.
(B) After a few hours as pulmonary arterial pressure falls in
relationship to aortic pressure. (C1) In the healthy infant
the pulmonary arterial pressure continues to fall as the duct
constricts, and the left-to-right velocity is high throughout
the cardiac cycle. (C2) If the duct fails to constrict, high
volume left-to-right flow develops, with left to-right flow
restricted mostly to systole. This demonstrates why the
murmur is restricted to systole in most preterm infants with
a large ductal shunt.
Diagnosis of patent ductus arteriosus 57

shunt [40]. LV function should appear hyper-

dynamic, with an elevated ejection fraction due to
the large preload (high left atrial pressure) and low
afterload (run off into the pulmonary arteries), and
increased stroke volume.
A ‘normal’ value for ejection fraction (or normal
appearance) is, therefore, too low in this situation.
In one study, only 29% of preterms had LV
fractional shortening (LVEDD-LVESD/LVEDD
100), in the normal range prior to ductal ligation,
the rest were higher [19]. As a rule of thumb, LV
fractional shortening, normally between 28 and
45% in infants without a ductal shunt, should be
Figure 6. M-mode measurement of the left atrial:aortic root over 40% in infants with a large left-to-right ductal
ratio. LA dimension (B) is 2.5 cm, Aortic root dimension shunt. Values below 30% demonstrate severe LV
(A) is 1.09 cm, the La:Ao ratio is therefore 2.3:1. dysfunction and inotropic support is indicated.
Since it is possible to measure ventricular out-
that ten preterm infants requiring ductal ligation puts with echocardiography [20,41,42], one can
had a mean LA:Ao ratio of 1.38:1 (all were above measure right and left ventricular outputs and
1.15:1) compared to a mean of 0.86:1 in controls calculate the difference to truly quantify a ductal
with a closed duct. Mellander et al. [17] found shunt. Using this technique, Phillipos et al. [42]
that an LA:Ao ratio >1.3:1 on day 3 predicted found that a 2:1 ductal shunt was associated with
subsequent symptomatic PDA with an error rate of severe IVH or periventricular leucomalacia. How-
only 8%. ever, the output ratio is complicated by the co-
However, the degree of left atrial enlargement existence of an interatrial shunt (of variable size
depends on a number of factors other than just the [20,38]), and by difficulty in getting good flow
size of the ductal shunt, including the size of the signals in the pulmonary artery when there is
foramen ovale, state of hydration, and presence of turbulence due to the flow from the duct. LV stroke
mitral regurgitation. volume can be shown reliably to be markedly
elevated, however. An early increase of 60%
The size of the foramen ovale has been shown to predict the development of
Left atrial size is reduced if the foramen ovale is symptomatic ductal shunting [4].
moderate (2–4 mm) or large (>4 mm), allowing it However, an ominous sign for the infant is a
to decompress into the right atrium. This left-to- normal or low LV output with all the other echo-
right atrial shunt may also be of haemodynamic cardiographic features suggesting a large shunt.
importance; acting like a true atrial septal defect
and increasing pulmonary blood flow further [38].
Evaluate the peripheral circulatory
Intravascular volume and the level of hydration effects
In units using a relatively frugal fluid regime and/or
diuretics, the LA:Ao ratio signifying a large left-to- Diastolic aortic pressure is low with a large left-to-
right ductal shunt is >1.2:1 [12] whereas in more right ductal shunt due to ‘ductal steal’. Blood
liberal fluid regimes values it is >1.5:1 [39]. passing down the descending aorta during systole
goes backwards up the ductus and into the pulmon-
Mitral regurgitation and left ventricular dysfunction ary arteries during diastole. Large ductal shunts can
LV dysfunction and/or mitral regurgitation cause cause over 50% of flow to go backwards up the
left atrial dilation. Both can occur after ductal aorta [42] resulting in relative under-perfusion of all
closure with indomethacin, so that the LA:Ao ratio of the systemic arteries including those supplying
may remain unchanged initially. Clearly the LA:Ao the cerebral hemispheres and the gut (see Fig. 7D).
ratio should not be used as the sole indicator of Absent or retrograde diastolic cerebral blood flow
ductal shunting. is said to be present at all times in babies requiring
An LVEDD:Ao ratio over 2.1:1 demonstrates ductal ligation [13], and rare in babies without a
significant LV dilation and is usual with a large ductus. Similar findings in brachial and femoral
58 J. Skinner

Figure 7. Pulsed Doppler recordings from an infant with a large left-to-right ductal shunt, taken from. (A) The main
pulmonary artery; note the turbulent flow pattern. (B) The left pulmonary artery; note the continuous forward flow in
diastole with peak velocity just over 50 cm/sec. (C) The transverse aortic arch; note continuous forward flow in diastole as
the blood runs around the arch and into the ductus. (D) The descending aorta; retrograde flow is seen in diastole
representing diastolic steal. Retrograde flow velocity integral (VTI) is 3.8 cm, over 80% of forward flow, which has a very
low VTI of 4.5 cm. This infant was hypotensive and acidotic and had poor ventricular function associated with the big
left-to-right ductal shunt.
Diagnosis of patent ductus arteriosus 59

Table 1. Echocardiographic features differentiating large, moderate and small left-to-right ductal shunt in infants less than
1500 g

Echocardiographic Small left-to-right Moderate left-to-right Large left-to-right

feature shunt shunt shunt

Ductal size on colour <1.5 mm 1.5–2.0 mm >2 mm

LA:Ao ratio* <1.4:1 1.4:1–1.6:1 1.6:1
Diastolic flow in desc Ao Mostly antegrade, Zero or modest reversal Reversal throughout diastole
down to zero (<30% of forward flow) (>30% forward flow)
Antegrade peak diastolic <30 cm/s 30–50 cm/s >50 cm/s
flow velocity in LPA

*Expect lower values when the atrial septal defect (foramen ovale) is wider than 2 mm [38], and with fluid restriction (which is the norm in many units).

Table 2. Echocardiographic signs of left ventricular failure in the presence of a large left-to-right ductal shunt

No/minimal Moderate
Clinical/echocardiographic Severe
myocardial myocardial
feature myocardial failure
failure failure

LV fractional shortening (%) >40% 30–40% <30%

LV stroke volume (ml/kg) >2.5 1.5–2.5 <1.5
LV output (ml/kg/min) Markedly elevated (>350) Mildly elevated (200–350) Not elevated (<200)

arteries are less specific; they commonly occur in 2. Left atrium enlarged; LA:Ao ratio>1.4:1—long
babies with a closed ductus [13]. axis view.
In the branch pulmonary arteries there is abnor- 3. Colour Doppler—continuous flare in the main
mally high antegrade diastolic flow as the continu- pulmonary artery from arterial duct
ous stream of blood from the aorta pours into them 4. Pulsed wave Doppler—turbulent flow in the
via the duct [Fig. 7(B)]. A peak velocity in diastole main pulmonary artery, continuous antegrade
over 50 cm/sec suggests a large shunt, and less flow in diastole in left pulmonary artery and arch
than 30 cm/sec a small shunt [44].4 of aorta, retrograde diastolic flow in descending
aorta, cerebral and gut blood vessels
Put the evidence together to assign the size of 5. Continuous wave Doppler—continuous left-to-
the shunt and the degree of circulatory failure right flow in main pulmonary artery from the
ductus, with low velocity (<1 m/sec) at end
diastole
The size of the shunt: small, moderate or large 6. Raised left ventricular stroke volume
It is useful to list the findings and categorize them
as shown in the Table 1, applying to infants less Assess degree of myocardial failure
than 1500 g. Nature being what it is, not all infants The ‘significance’ of a large left-to-right shunt will
have features in the same column. depend, amongst other factors, upon the degree of
Typical echocardiographic features of a left-to- myocardial reserve. This will tend to be reduced in
right ductal shunt are: infants with a history of perinatal asphyxia, and in
1. Bowing of the interatrial septum to the right extreme prematurity. When these two are com-
with enlarged left atrium and left ventricle— bined there should be a high index of suspicion that
four chamber views myocardial failure is likely, and early echocardiog-
raphy is important. Absolute blood pressure bears
little relationship to the presence or absence of
4
Abnormal retrograde diastolic flow in the descending aorta is also seen
with severe aortic incompetence and aorto-pulmonary window. However,
with a PDA, diastolic flow is antegrade in the transverse aortic arch, circulatory or myocardial failure, though it usually
whereas it is retrograde with the other diagnoses. rises as the ductus closes.
60 J. Skinner

Table 2 may be helpful as a ‘rule of thumb’ 2 Reller MD, Ziegler ML, Rice MJ, Solin RC, McDonald
guide, in infants who have other features of a large RW. Duration of ductal shunting in healthy preterm
infants: An echocardiographic color flow Doppler study.
left-to-right shunt. The classical features of cardiac J Pediatr 1988; 112: 441–446.
failure (hepatomegally, pulmonary oedema etc.) are 3 McGrath RL, McGuinness GA, Way GL, Wolfe RR, Nora
a useful guide in larger infants and those more than JJ, Simons MA. The silent ductus arteriosus. J Pediatr 1978;
3–4 days old, but are much less reliable in the first 93: 110–113.
48 h, especially in extreme prematurity. 4 Walther F, Kim D, Ebrahimi M, Siassi B. Pulsed Doppler
A new technique which may be useful to detect measurement of left ventricular output as early predictor
of symptomatic patent ductus arteriosus in very preterm
early circulatory failure is flow velocity in the infants. Biology of the Neonate 1989; 56: 121–128.
superior vena cava; this is discussed in more detail 5 Skinner JR, Boys RJ, Hunter S, Hey EN. Pulmonary and
in chapter 7 of this journal. Other important clues systemic arterial pressure in hyaline membrane disease.
clinically are increasing values of base deficit, Arch Dis Child 1992; 67: 366–373.
arterial lactate and serum potassium. 6 Kluckow M, Evans N. Early echocardiographic prediction
of symptomatic patent ductus arteriosus in preterm
infants undergoing mechanical ventilation. J Pediatr 1995;
Conclusion 127: 774–779.
7 Seppanen M, Kaapa PO, Kero PO. Hemodynamic predic-
Clinical signs cannot be relied on to reveal left-to- tion of complications in neonatal respiratory distress
right ductal shunting in the preterm and since syndrome. J Pediatr 1995; 127: 780–785.
8 Evans N, Kluckow M. Early ductal shunting and intra-
prolonged ductal patency is the norm in ventilated ventricular haemorrhage in ventilated preterm infants.
very preterm neonates, echocardiographic screen- Arch Dis Child 1996; 75: F183–F186.
ing should take place, particularly to identify myo- 9 Mellander M, Larrson LE. Effects of left-to-right ductus
cardial failure. Reasonably accurate prediction of shunting on left ventricular output and cerebral blood
the development of later clinically overt shunting flow velocity in 3 day old preterm infants with and
can be done on day 3 by identifying an elevated without severe lung disease. J Pediatr 1988; 113: 101–
109.
LA:Ao ratio (>1.3:1 or >1.5:1, depending on fluid 10 Heldt GP, Pesonen E, Merritt TA, Elias W, Sahn D.
and diuretic policy) and a ductal diameter >1.4 mm. Closure of the ductus arteriosus and mechanics of breath-
Doppler echocardiographic techniques are suffi- ing in preterm infants after surfactant therapy. Pediatr Res
ciently advanced to allow reliable combined 1989; 25: 305–310.
assessment of shunting at atrial and ductal levels, 11 Archer LNJ, Glass EJ, Godman MJ. The silent ductus
systemic circulatory effects, and myocardial per- arteriosus in idiopathic respiratory distress syndrome.
Acta Paediatr Scand 1984; 73: 652–656.
formance, which may appear normal when it is not. 12 Valdez-Cruz LM, Dudell GG. Specificity and accuracy of
No one single measurement indicates ‘significance’ echocardiographic diagnosis and clinical criteria for diag-
of a shunt; this is best defined by a number of nosis of patent ductus arteriosus in fluid-restricted infants.
clinical and echocardiographic features. J Pediatr 1981; 98: 298–305.
The neonatal echocardiographer screening for a 13 Kupfershmid C, Lang D, Pohlandt F. Sensitivity, specifi-
PDA must never assume that the heart is structur- city and predictive value of clinical findings, m-mode
echocardiography and continuous-wave Doppler sonog-
ally normal, even in the absence of clinical signs to raphy in the diagnosis of symptomatic patent ductus
suggest congenital heart disease. This process is arteriosus in preterm infants. Eur J Pediatr 1988; 147:
helped by a carefully ordered clinical examination 279–282.
and scanning routine and a low threshold to seek 14 Hirsimaki H, Kero P, Wanne O. Doppler ultrasound and
expert cardiological advice [45]. clinical evaluation in detection and grading of patent
ductus arteriosus in neonates. Cr Care Med 1990; 18:
490–493.
15 Roberson DA, Silverman NH. Color Doppler flow map-
Acknowledgement ping of the patent ductus in very low birth weight
neonates: echocardiographic and clinical findings. Pediatr
My grateful thanks to my wife Karen for holding the toddlers Cardiol 1994; 15: 219–224.
at bay whilst I wrote this. 16 Evans N, Iyer P. Longitudinal changes in the diameter of
the ductus arteriosus in ventilated preterm infants: corre-
lation with respiratory outcomes. Arch Dis Child 1995; 72:
References F156–F161.
17 Mellander M, Larsson LE, Ekstrom-Jodal B, Sabel KG.
1 Skinner JR, Boys RJ, Hunter S, Hey EN. Non-invasive Prediction of symptomatic patent ductus arteriosus in
determination of pulmonary arterial pressure in healthy preterm infants using Doppler and m-mode echocardiog-
neonates. Arch Dis Child 1991; 66: 386–390. raphy. Acta Paediatr Scand 1987; 76: 553–559.
Diagnosis of patent ductus arteriosus 61

18 Skinner JR, Hunter S, Hey EN. Cardiorespiratory collapse 33 Hunter S. A logical approach to cardiac scanning. In:
and the silent ductus. Klinische Padiatrie 1991; 203: 52 Skinner JR, Alverson D, Hunter S (eds) Echocardiography
(Abstr). For The Neonatologist. London: Churchill Livingstone,
19 Baylen B, Meyer RA, Korfhagen J, Benzing G, Bubb ME, 2000; 8: 171–180.
Kaplan S. Left ventricular performance in the critically ill 34 Smallhorn JF, Gow R, Olley PM, Freedom RM, Swyer PR,
premature infant with patent ductus arteriosus and pul- Perlman M, Rowe RD. Combined noninvasive assessment
monary disease. Circulation 1977; 55: 182–188. of the ductus arteriosus in the preterm infant before
20 Evans N, Kluckow M. Early determinants of right and left and after Indomethacin treatment. American Journal of
ventricular output in ventilated preterm infants. Arch Dis Cardiology 1984; 54: 1300–1304.
Child 1996; 74: F88–F94. 35 Skinner JR, Hunter S, Hey EN. A new method to assess
21 Evans N, Moorcraft J. Effect of patency of the ductus ductal shunting. Paediatr Rev Comm 1991; 6: 58–59
arteriosus on blood pressure in very preterm infants. Arch (Abstr).
Dis Child 1992; 67: 1169–1173. 36 Skinner JR, Silove ED. Aorto-pulmonary collateral arteries
22 Evans N. Atrial shunting. In: Skinner JR, Alverson D, mimicking symptomatic ductal shunting in a preterm
Hunter S (eds) Echocardiography For The Neonatologist. infant. Br Heart Journal 1995; 74: 93–94.
London: Churchill Livingstone, 2000; 3: 101–112. 37 Harada K, Shiota T, Takahashi Y, Tamura M, Toyono M,
23 Gill AB, Weindling AM. Echocardiographic assessment of Takada G. Doppler echocardiographic evaluation of left
cardiac function in shocked very low birth weight infants. ventricular output and left ventricular diastolic filling
Arch Dis Child 1993; 68: 17–21. changes in the first day of life. Pediatr Res 1994; 35:
24 Silverman NH, Lewis AL, Heyman MA, Rudolph AM. 506–509.
Echocardiographic assessment of ductus arteriosus shunt
38 Evans N, Iyer P. Assessment of ductus arteriosus shunt in
in premature infants. Circulation 1974; 50: 821–825.
preterm infants supported by mechanical ventilation:
25 Skelton R, Evans N, Smyth J. A blinded comparison of
effects of interatrial shunting. J Pediatr 1994; 125: 778–
clinical and echocardiographic evaluation of the preterm
785.
infant for patent ductus arteriosus. J Pediatr Child Health
1994; 30: 406–411. 39 Johnson GL, Breart GL, Gewitz MH, Brenner JI, Lang P,
26 Davis P, Turner-Gomes S, Cunningham K, Way C, Dooley KJ, Ellison RC. Echocardiographic characteristics
Roberts R, Schmidt B. Precision and accuracy of clinical of premature infants with patent ductus arteriosus.
and radiological signs in premature infants at risk of Pediatrics 1983; 72: 864–871.
patent ductus arteriosus. Arch Pediatr Adolesc Med 1995; 40 Iyer P, Evans N. Re-evaluation of the left atrial to aortic
149: 1136–1141. root ratio as a marker of patent ductus arteriosus. Arch Dis
27 Skinner JR. Why do some babies with a large ductal shunt Child 1994; 70: F112–117.
have a murmur when others do not? In: Skinner JR (ed.) 41 Alverson D. Assessment of cardiac output. In: Skinner JR,
Non-invasive determination of pulmonary arterial pressure in Alverson D, Hunter S (eds) Echocardiography For The
the newborn. MD thesis 1993. University of Leicester, 14: Neonatologist, London: Churchill Livingstone, 2000; 5:
234–249. 121–131.
28 Maroto E, Fouron J-C, Ake E, van Doesburg NH, 42 Phillipos EZ, Robertson MA, Byrne PJ. Serial assessment
Cartwright D, Teyssier G, Bard H. Closure of the ductus of ductus arteriosus hemodynamics in hyaline membrane
arteriosus: determinant factor in the appearance of tran- disease. Pediatrics 1996; 98: 1149–1153.
sient peripheral pulmonary stenosis. J Pediatr 1991; 119: 43 Serwer GA, Armstrong BE, Anderson PAW. Continuous
955–999. wave Doppler ultrasonographic quantitation of patent
29 Rudolph AM. Circulatory changes during the perinatal ductus arteriosus flow. J Pediatr 1982; 100: 297–299.
period. Pediatr Cardiol 1983; 4(suppl 2): 17–20. 44 Hiraishi S, Horiguchi Y, Misawa H, Oguchi K, Kadoi N,
30 Alverson D, Eldridge M, Johnson J, Burstein R, Papile L, Fujino N, Yashiro K. Noninvasive Doppler echocardio-
Dillon T, Yabek S, Berman W. Effect of patent ductus graphic evaluation of shunt flow haemodynamics of the
arteriosus on left ventricular output in premature infants. ductus arteriosus. Circulation 1987; 75: 1146–1153.
J Pediatr 1983; 102: 754–757. 45 Skinner JR. Echocardiography on the neonatal unit: A job
31 Skinner JR, Hunter S, Hey EN. Regulation of cardiac for the neonatologist or the cardiologist? Arch Dis Child
output in the premature neonate: stroke volume or heart 1998; 78: 401–407.
rate? Pediatr Cardiol 1991; 12(4): 258 (Abstr).
32 Lindner W, Seidel M, Versmold H, Dohlemann C, Riegel
K. Stroke volume and left ventricular output in preterm
infants with patent ductus arteriosus. Pediatr Res 1990; 27:
278–281.