Article Text

Peak systolic to end diastolic flow velocity ratio is associated with ductal patency in infants below 32 weeks of gestation
  1. A Smith1,
  2. M Maguire2,
  3. V Livingstone1,3,
  4. E M Dempsey1,3
  1. 1Department of Paediatrics and Child Health, Neonatal Intensive Care Unit, University College Cork, Cork, Ireland
  2. 2Department of Cardiology, Cork University Hospital, Cork, Ireland
  3. 3Infant Centre, University College Cork, Cork, Ireland
  1. Correspondence to Dr Eugene M Dempsey, Department of Paediatrics and Child Health, Neonatal Intensive Care Unit, University College Cork, Cork, Ireland; gene.dempsey{at}


Background Early diagnosis and effective treatment of the patent ductus arteriosus (PDA) in infants less than 32 weeks gestation remains contentious.

Objective To determine which clinical and echocardiographic parameters are associated with PDA patency in preterm infants less than 32 weeks gestation.

Design/Methods This was a prospective cohort study. An echocardiography (echo) was performed within 12–48 h of birth and a follow-up echo at 1 month of life. Parental consent was obtained.

Results 55 babies were enrolled. Median (range) gestation was 28 (24–31) weeks and birth weight 1090 g (470–1800 g). ECHO 1 demonstrated that 50 babies had a PDA present within 48 h of birth, of which 19 were large (≥2 mm) (36%) and 31 were small (59%) on colour Doppler assessment of duct diameter. Three babies died before 1 month. At 1 month 30 babies still had a PDA (58%), 10 of which were large (19%) and 19 were small (36%). Parameters significantly associated with large PDAs versus no PDA at 1 month were gestational age (26 weeks vs 30 weeks, p=0.002), birth weight (860 g vs 1290 g, p=0.007) and ventilator support at 48 h (80% vs 17%, p=0.001). Echo parameters revealed that ductal size on colour Doppler (2.5 mm vs 1.5 mm, p=0.003), end diastolic flow velocity (57 m/s vs 147 m/s, p<0.001) and peak systolic to end diastolic flow velocity ratio (2.29 vs 1.23, p=0.001) at 48 h were associated with large PDAs at 1 month.

Conclusions For infants less than 32 weeks gestation a peak systolic to end diastolic flow velocity ratio>2 within 48 h of birth is associated with a persistent large PDA at 1 month of age.

  • Neonatology
  • Cardiology

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

What is already known on this topic

  • Controversy surrounds the diagnosis and treatment of a patent ductus arteriosus.

  • Gestational age predicts closure.

What this study adds

  • Interrogation of ductus should include quantification of the systolic and diastolic flow velocities.

  • The peak systolic to end diastolic flow velocity ratio was the best echo parameter to predict ductal patency in the preterm infant.


The diagnosis and subsequent management of a patent ductus arteriosus (PDA) in the very preterm infant remains contentious.1 Many different diagnosic criteria have been proposed including various echocardiographic,2–5 biochemical6–8 and clinical parameters9 or combinations of these. Various interventions have been advocated including prophylactic medical treatment with indomethacin10 or ibuprofen,11 ,12 early surgical closure,13 an early targeted echocardiography (echo) approach,14 a symptomatic approach and finally a more conservative approach incorporating respiratory support, fluid restriction and diuretic therapy.15

The Trial of Indomethacin Prophylaxis in Preterms (TIPP) trial10 has shown a reduction in the number of infants who required PDA ligation, a reduction in severe intraventricular haemorrhage (IVH) and pulmonary haemorrhage but no dfference in long-term outcome. There have been two large ibuprofen prophylactic trials.11 ,12 While there was a reduction in the need for surgical ligation in the Gournay trial, there was no difference in mortality and the trial was stopped early because of safety concerns in the treatment arm.11 Van Overmeire et al12 found no difference in the primary outcome of IVH but did identify a statistically significant reduction in the incidence of PDA and a non-significant reduction in the need for surgical ligation. Each of these prophylactic studies had placebo arms: incidence of spontaneous PDA closure in TIPP was 50%, in the Gournay study 45% and in the Van Overmeire study 60%. Much of the variation relates to the gestational age, Koch having noted a spontaneous closure rate of approximately 34% in extremely low gestational age newborns in the 1st week of life.16

The management of the ductus remains unresolved.17 Various algorithms have been developed in an attempt to identify newborns with spontaneous closure and thereby avoid unnecessary medical intervention. These often incorporate clinical, biochemical and echocardigraphic parameters.3 ,7 ,18–20 Early treatment is more likely to result in a higher closure rate and hence the need to identify patients early if one is to adopt a targeted approach. Therefore the aim of this study was to identify clincial and ductal paramters within the first 48 h associated with subsequent ductal patency at 1 month of age.


A prospective, observational study was conducted at the Neonatology Department of Cork University Maternity Hospital, Cork, Ireland over a 12 month period from December 2011 to November 2012. All newborns less than 32 weeks old were eligible for inclusion in this study. Patients with complex cardiac malformations and/or gross congenital abnormalities were excluded. Parental consent was obtained. The Cork Research and Ethics Committee of the Cork Teaching Hospitals approved this study.

All eligible infants underwent echo by a consultant neonatologist within the first 48 h of life (ECHO 1). At 1 month of life all babies had a second echo (ECHO 2) performed by an echo technician who was unaware of the original study findings. Babies may have had additional echos in the interim period as clinically indicated by the attending clinician who was unaware of the original echocardiographic findings. The Philips HE11XE ultrasound scanner with curved array transducer (5–8 MHz) or sector array transducer (4–12 MHz) incorporating colour flow and pulsed wave Doppler with adaptive Doppler technology was used. The scan was recorded on the hard disk of the ultrasound machine and these were subsequently electronically exported for later measurements. Structural normality of the heart was confirmed and the following standard echo parameters were recorded:

  • PDA diameter on colour Doppler;

  • PDA diameter on non-colour echo;

  • Left atrium to aortic ratio;

  • Peak systolic flow velocity across ductus;

  • End diastolic flow velocity across ductus;

  • Peak systolic to end diastolic flow ratio (ductal pattern);

  • Diastolic flow velocity pattern in pulmonary artery.

The following clinical and physiological data were recorded: Gestation, birth weight, respiratory support, and vital signs including heart rate and blood pressure and biochemical parameters (lactate and base excess) were recorded within the first 48 h of life. At 1 month babies were categorised into three groups based on ductal presence and size21: large PDA (≥2 mm), small PDA (<2 mm) and those with no PDA. We chose 2 mm as a cut-off to determine a large PDA as this value has been considered as representative of a haemodynamically significant ductus.21 ,22

Categorical data was described using frequency (percentage) and continuous data using median (IQR). Differences between the three PDA groups were investigated using the Kruskal-Wallis test and the χ2 test. ECHO parameters were compared between babies with large PDAs and babies with no PDA or a small PDA using receiver operator curve (ROC) curves, and the optimal cut-off value for each parameter was determined using the Youden index. The diagnostic accuracy of each ECHO parameter using the optimal cut-off was described using sensitivity, specificity, positive predictive value and negative predictive value. All statistical analysis was performed using Stata V.13.0 (StataCorp LP, College Station, Texas, USA). All tests were two-sided and a p value <0.05 was considered to be statistically significant.


Fifty-five babies were enrolled in this study. Three babies were excluded from the analysis, one baby died from transfusion related acute gut injury on day 10 of life, a second from respiratory failure within 12 h secondary to pulmonary hypoplasia and a third infant diagnosed with trisomy 21 was excluded. Therefore analysis was carried out on 52 babies. The initial echo performed on the patient with pulmonary hypoplasia revealed severe pulmonary hypertension with pure right to left ductal shunting. The patient who developed transfusion related acute gut had a small PDA on initial echo. Within the first 48 h of birth echo 1 demonstrated that 50 babies have a PDA (96%), of which 19 were ≥2 mm on colour Doppler (36%) and 31 were less than 2 mm (60%). At 1 month 30 babies still had a PDA (58%), 10 of which were large (19%) and 19 were small (36%). The PDA was not present in 23 (44%) babies. The results are presented according to this allocation in tables 1 and 2. Only one baby received ibuprofen at 5 days of age and ultimately had a PDA ligation performed at 1 month of age. There were 26 infants with a gestational age up to 28+6 weeks.

Table 1

Comparison of clinical characteristics by PDA group

Table 2

Comparison of echocardiography characteristics by PDA group

Tables 1 and 2 display the clinical and echo data at the time of the original echo performed within first 48 h of life. Clinical characteristics that were associated with PDA group were gestational age, birth weight, mean airway pressure (MAP), lactate at 12 h, 1 min Apgar value and ventilation status. No patient had a pulmonary haemorrhage. Pairwise comparisons revealed that the differences were only between the no PDA group and the large PDA group for birth weight and lactate at 12 h, between the small and no PDA group for MAP, while the differences were between the no PDA group and the large PDA group and the no PDA group and the small PDA group for gestational age and ventilation status.

Echo characteristics that were associated with the PDA group were duct size on colour assessment, systolic and diastolic flow velocity and their ratio. In the pairwise comparisons, differences were found between the large PDA group and the no PDA group and the small PDA group for duct size and ratio while for systolic flow velocity, the differences were between the no PDA group and the large PDA and small PDA groups. Diastolic flow velocity was significantly different between the large PDA group and the small PDA group. Ratio was the best performing ECHO parameter in discriminating between those with a large PDA and those with a small or no PDA with an area under the ROC curve of 0.89 and positive predictive value of 72.7% (table 3). Figure 1 displays the ROC for ratio. When we analysed babies less than 29 weeks the area under the curve (AUC) for ratio was similar (0.87), the specificity increased to 93.3% and the positive predictive value (PPV) increased to 85.7% (see online supplementary table S4), a reflection of the increased prevalence of PDA in the more immature babies.

Table 3

Diagnostic accuracy of echocardiography characteristics, n=41*

Figure 1

Receiver operator curve for ratio of peak systolic to end diastolic flow velocity with ratio >1.92.


We have identified a number of clincial and echocardiographic findings within the first 48 h of life that are associated with the presence of a large PDA at approximately 1 month of age. We determined, a priori, a ductal diameter of 2 mm or more would represent a large patent duct at 1 month of age. The clinical parameters identifed in the first 48 h were gestational age, birth weight and need for respiratory support. Gestational age and birth weight parameters are consistent with the natural history of the patent ductus, namely an increasing incidence of PDA with decreasing gestational age and birth weight. The need for ventilator support would be consistent with the concept of a potentially haemodynamically significant duct, although some babies less than 28 weeks may be ventilated within the first 48 h of life. In a recent survey of practice in French neonatal intensive care units the need for respiratory support was considered a criteria consistent with a hamodynamically significant dutcus,23 however the exact timing of this support was not clear. The number of infants in our cohort is too small to allow one to make any firm conclusions in relation to ventilator support at this time. We did not include clinical signs of a ductus within this time period, due to their poor correlation with echo parameters2 and the subjective nature of many of these findings, such as bounding pulses or a murmur. There was no difference in any blood pressure parameter recorded.

The echo findings consisted of the size of duct identified on colour flow doppler and the ratio of the peak systolic to end diastolic flow velocity ratio (PSEDR). The non-colour assessment of the ductus was not significant and consistently overestimated the actual size of the ductus. We believe this finding reaffirms the importance of appropriate colour flow interrogation of the duct, and although colour interrogation is not without its potential errors, we believe it allows one to adequately assess the functional diameter of the ductus. Colour assessment of the ductus should be incorporated and, as can be seen from our findings, large ducts identified on colour doppler within the first 48 h were associated with patency. Kwinta et al3 evaluated a group of 60 babies, 16 of whom went on to have PDA ligation and found that the ductal diameter was 2.6 mm compared with 0.9 mm for the group that did not require surgery. In our cohort the left atrium to aorta (LA : Ao) ratio was not larger in the group that maintained patency, this finding was not statistically significant and the median ratio was lower than that described previously in a number of studies.7 ,19 This is not surprising considering that all our echocardiograms were performed within the first 48 h, a time frame in which it was unlikely that the left atrium would become overdistended resulting in a larger diameter detected on M-mode assessment. Diastolic disturbance in the pulmonary artery was not associated with patency, and again this may reflect the time period in which the echocardiograms were perfomed.

The peak systolic velocity of the ductus may also be an independent indicator of closure. The narrower the ductus the greater the systolic velocity across the ductus. Our study identified that a lower peak systolic velocity was associated with patency, however this finding was not statistically significant between the large and small PDA groups. A low end diastolic flow velocity within the ductus was significantly associated with a large patent ductus at 1 month of age. We believe this is the first study to quantify restrictive and pulsative flows in the PDA with predictive value of these observations.

However a notable finding was the systolic to diastolic flow velocity ratio within the ductus was different across the three groups. This finding is consistent with patterns of ductal flow first described by Su et al4 ,5 and used susequently in a number of randomised trials.24 ,25 They described a number of patterns including a pulmonary hypertension type pattern, a growing pattern, a pulsatile pattern and a restrictive pattern. We believe that our findings are representative of a pulsatile pattern and provide objective measurement of this, rather than a subjective visual assessment. We propose defining a pulsatile pattern as a ratio of systolic to diastolic flow velocity of greater than 2. We believe that this is a relatively easy assessment to perform at the bedside and is also readily calculable. The finding of a large ductus with a pulsatile flow pattern is consistent with a retrospective review perfomed by Condo et al.18 We did not include babies with evidence of elevated pulmonary pressures as the bidirectional blood flow precluded us from calculating a ratio, and generally we would not advocate treatment of a bidirectional flow pattern initially but rather recommend a repeat echo in the next 24 h prior to considering treatment.

We did not routinely evaluate other echo parameters such as the peak flow velocity in the left pulmonary artery (LPA), LPA : PDA ratio, left ventricular output (LVO) or LVO : superior vena cava (SVC) ratio. These measurements have been used previously as markers of a ‘haemodynamically significant’ PDA and as predictors of ‘symptomatic’ PDA, spontaneous closure or the necessity for medical intervention. For example, Thankavel reported that ductal diameter, LA : Ao ratio and LPA diastolic flow velocity were independent predictors of spontaneous closure, but that the ratio of the PDA : LPA diameters was the only significant predictor of spontaneous closure.26 In a retrospective study Sehgal and Menahem27 report that the LVO : SVC ratio correlated significantly with a PDA 3 mm in size or larger with a specificity of 0.83 and a sensitivity of 0.95. However, considerable variability in the cut-offs used and the context in which such measurements are recorded is evident across many studies. A number of studies have suggested a role for certain biomarkers, either urinary or plasma derived, as a useful indicator of a significant ductus or as an adjunct to the overall assessement of a patent ductus.6 ,8 ,28 We do not routinely evaluate biomarkers in our intensive care unit in this setting and so these parameters were not assessed. Recent developments in PDA evaluation also include the use of MRI to quantify ductal shunt volume and the effect of PDA on systemic blood flow volume in neonates as per Groves et al.29

There are a number of other limitations. The absolute number of babies was small (55 newborns). However this number compares favourably with other studies that have assessed ductal parameters in the preterm infant, and also encompasses a gestational age range similar to other trials conducted in this area. While the gestational age did include babies up to 31+6 weeks, the vast majority of these more mature babies were on some form of respiratory support. The total number of babies less than 28 completed weeks was low (27 infants). When these were evaluated independently we found that the ductal size and the ratio were the two most significant factors. The timing of the first scan was variable as this was performed by a single neonatologist, but the median time of scan was 34 h (12–48 h). Notwithstanding these limitations the findings of this study may help to inform future trials that may target ductal closure in the first few days. While we do not necessarily advocate this approach, we believe that interrogation of the ductus with colour Doppler to assess size and determination of flow patterns across the ductus are the key parameters that should be used, and may help to identify those newborns who will potentially benefit from early administration of a non-steroidal medication and avoid unnecessary exposure to medical therapies.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors AS had primary responsibility for overall content and manuscript preparation. She was involved with study design, research collection and research interpretation. MM performed the second echocardiograms and was involved with study design. VL was involved with the statistical analysis and manuscript preparation. EMD was the supervisor for this study. He was involved with overall content, study design, manuscript preparation, research collection and research interpretation.

  • Funding This study was supported by the European Commission within the 7th Framework Programme (EU FP7/2007-2013) under grant agreement no. 260777 (The HIP Trial).

  • Competing interests None.

  • Patient consent Parental consent obtained.

  • Ethics approval Cork Research and Ethics Committee, Cork, Ireland.

  • Provenance and peer review Not commissioned; externally peer reviewed.