Article Text
PDF
Abstract
Background Patent ductus arteriosus (PDA) in preterm infants is often assessed with echocardiographic parameters, especially colour Doppler ductal diameter and pulsed Doppler flow pattern. Clinical algorithms have been proposed in which PDA treatment is indicated by either large diameter or a particular flow pattern, however it is unknown whether ductal diameter and flow pattern provide equivalent stratification of infants.
Aim Retrospectively assess both parameters in 197 echocardiograms from 104 infants (gestational age <31 weeks).
Methods Echocardiograms were independently reviewed and the internal colour Doppler diameter of the PDA and the pulsed Doppler flow pattern were characterised for each study (169 records had both parameters recorded).
Results Diameter varied widely within each group but was significantly associated with flow pattern: mean diameter was greatest in the pulmonary hypertension (PH) group (2.6 mm), progressively narrowed across growing and pulsatile groups, and was smallest in the closing group (1.3 mm). When echocardiograms were categorised using previously published diameters, 82.4% of the PH group had diameters >2.0 mm, large diameters predominated in the growing and pulsatile groups but to a progressively smaller extent, and 98.1% of closing group had diameters <2.0 mm.
Conclusion Ductal diameter and flow patterns are significantly associated, consistent with a narrowing of the ductus until closure. Overall, the two parameters are in good agreement but will result in different treatment decisions in some cases. Clinicians might consider using both methods as a cross check against each other, to assist in the management of preterm infants with a clinically detectable PDA.
Statistics from Altmetric.com
Introduction
In the preterm infant the ductus arteriosus commonly fails to close in the early postnatal period, causing shunting of blood in a predominantly left-to-right direction, often with adverse haemodynamic consequences.1,–,3 Early diagnosis of patent ductus arteriosus (PDA) based on clinical signs has low sensitivity,4 5 so echocardiography is the diagnostic method of choice.6 7 Echocardiographic signs proposed for characterising a PDA include ductal diameter measured at the point of maximum constriction8 and qualitative assessment of PDA by assessing the pattern of blood flow through the ductus. Su et al9 identified four patterns of ductal flow based on the Doppler velocity waveform. The growing and pulsatile patterns were more often observed in infants who developed a significant PDA.9 Determination of these and other parameters10 11 allows prediction of spontaneous closure of a PDA and facilitates triaging of infants referred for PDA ligation using a PDA staging system.12
What is already known on this topic
▶ Functional echocardiography permits bedside assessment of a patent ductus arteriosus (PDA) in the newborn infant.
▶ Both ductal diameter and flow pattern are used to assess PDA clinical significance.
▶ Use of clinical and echocardiographic algorithms to determine treatment of PDA can result in different decisions.
What this study adds
▶ There is a strong correlation between early ductal diameter and flow patterns.
▶ The addition of flow pattern assessment to ductal diameter measures may further enhance the clinical predictive capacity of echocardiography.
Two recent clinical trials showed that PDA treatment decisions can be made with clinical efficacy and safety (fewer drug doses) when guided by the assessment of ductal diameter13 or flow pattern.14 It is currently unknown how the two parameters compare clinically. Therefore, we retrospectively evaluated both parameters on echocardiographic recordings from a cohort of preterm neonates to determine how ductal diameter varies with flow pattern and how this may impact treatment decisions. We did not investigate the effect of these measurements on outcome.
Materials and methods
In the neonatal intensive care units of two Sydney hospitals all echocardiograms were digitally archived from January until March 2008 and studied retrospectively. The study protocol was approved by both ethics committees, which waived the need to obtain informed consent as the data had been previously collected for other approved purposes.
Echocardiography
Echocardiography was routinely performed by neonatologists trained in functional echocardiography within the first 24 h after birth in all infants of <28 weeks gestation and in infants between 28 and 31.6 weeks who were ventilated and required surfactant or had other cardiovascular concerns. Infants were rescanned at varying intervals depending on clinical signs up to the point of ductal closure or hospital discharge. We used 10S Sector probes on Vivid 7 Dimension ultrasonography systems (GE Healthcare, GE Medical Systems, NSW, Australia) through a high parasternal window. A video clip of the colour flow and of the pulsed wave Doppler image with the sample gate positioned on the pulmonary end of the PDA was saved and archived using ECHOPAC software (GE Healthcare).
Data collection and image analysis
Clinical records were reviewed to obtain additional information regarding gender, birth weight and age at the time of examination. Echocardiographic recordings were analysed by a single operator (MC). Ductal diameter was measured at the narrowest point of the colour mapping at the pulmonary end of the duct.8 The Doppler velocity waveform traces were classified as pulmonary hypertension (PH), growing, pulsatile or closing pattern, according to Su et al.9 These analyses were blinded to clinical data and outcomes.
Statistical analyses
For each flow pattern, ductal diameter was described with descriptive statistics as well as with frequency counts, using as cut-offs 1.58 and 2.0 mm.2 13 The association between ductal diameter and gestational age was tested with Pearson's correlation. The Kruskal–Wallis test was used to compare distributions of gestational age and ductal diameter across the four flow pattern groups. The Mann–Whitney test was used to compare ductal diameter between pairs of flow pattern groups, while the χ2 test was used to assess this association when ductal diameter was categorised. Statistical analyses were performed using SPSS v 13 (SPSS, Chicago, Illinois, USA).
Results
We retrospectively studied the echocardiographic recordings of 104 preterm infants (68 boys). The study group included 97 infants (93.3%) of very low birth weight (<1500 g). Median (range) gestational age was 28.0 (23.0–30.4) weeks and birth weight was 1043 (498–1800) g. A single echocardiographic study was performed in 59 infants (56.7%), while other infants had two to six studies, for a total of 197 recordings available for analysis. All traces had a clearly measurable ductal diameter, which ranged from 0.6 to 4.0 mm (median, 1.9 mm). There was no significant correlation between ductal diameter and gestational age (Pearson's r=−0.02, p=0.711).
Of the 197 echocardiographic studies, 169 (85.8%) had a ductal flow pattern with a clearly identifiable Doppler velocity waveform (table 1). More studies were classified as pulsatile (n=55; 32.5%) or closing (n=52; 30.8%) patterns than either PH (n=34; 20.1%) or growing (n=28; 16.6%) patterns (χ2 goodness of fit, p=0.006). Studies were carried out in infants on the first 2 days of life in 99 cases (58.6%) and between 3 and 30 days of life in 68 cases (40.2%); two examinations were carried out on infants in the second month of life. Age distributions for the four groups were significantly different (Kruskal–Wallis, p<0.001), and infants in whom a PDA remained beyond 2 days of life were more likely to have a pulsatile or closing pattern. An additional 28 studies (14.2%) had a waveform pattern that could not be unambiguously classified.
Classification of neonatal echocardiographic recordings by Doppler flow velocity waveform pattern and postnatal age at examination
The relationship between ductal diameter and flow pattern was assessed for the 169 studies in which both parameters were determined. Median ductal diameter was highest in the PH pattern group (2.6 mm) and progressively decreased across the groups, reaching 1.3 mm in the closing pattern group (figure 1). Despite the wide variability in ductal diameter within each group, these values were significantly different (Kruskal–Wallis, p<0.001). Ductal diameter in the closing pattern group was significantly smaller than those in all other groups, and there also was a significant difference between the PH and pulsatile groups (Mann–Whitney, p<0.001 for each comparison). When ductal diameter values were categorised into three groups (figure 2), a significant trend emerged with 82.4% of the PH pattern group having values >2.0 mm and 98.1% of the closing pattern group having values <2.0 mm (χ2 test, p<0.001). Large ductal diameter values also predominated in the growing and pulsatile patterns, but to a progressively smaller extent. Together, these results demonstrate that ductal diameter and flow patterns are significantly associated, consistent with a narrowing of the ductus until closure. Restricting analysis to measurements taken within the first 2 days of life led to similar results; trends of association were similar in infants both younger (Kruskal–Wallis, p<0.001) and older than 2 days of life (Kruskal–Wallis, p<0.01).
Relationship between ductal diameter and flow pattern of the patent ductus arteriosus in 169 echocardiographic traces from 104 preterm infants: Tukey's box and whisker plot shows median (horizontal bar), IQR (box), minimum and maximum values of ductal diameter. p<0.001 for all four groups (Kruskal–Wallis test); p<0.001 pulmonary hypertension versus closing, p<0.001 pulmonary hypertension versus pulsatile (Mann–Whitney test).
Relationship between ductal diameter and flow pattern of a patent ductus arteriosus in 169 echocardiographic traces from 104 preterm infants: frequency counts of ductal diameter in three size classes (χ2 test, p<0.001).
In this study, 28 echocardiographic studies had a flow pattern that was not clearly classifiable. These Doppler traces all appeared intermediate between the pulsatile (type 3) and closing (type 4) patterns. Median ductal diameter in this group was 1.8 mm (range 0.8–2.6 mm). These values were similar to those of the pulsatile group (Mann–Whitney, p=0.106) but significantly different from those of the closing group (Mann–Whitney, p<0.001).
Discussion
The study documented a significant association between two echocardiographic parameters commonly used to predict the likelihood that a PDA will close and, thus, to guide treatment decisions. The four ductal flow pattern groups had significantly different ductal diameters, with widest diameters in the PH pattern group, characteristic of the first hours of life, and narrowest diameters having a closing pattern. These results are consistent with the physiological narrowing of the ductus arteriosus after birth until closure. Ductal diameter values were widely distributed in all groups except the closing group, which had a narrow range of values. When ductal diameter values were categorised according to clinically used cut-off values, a significant and marked pattern emerged with the PH group mostly having ductal diameter values >2.0 mm and the closing group mostly having values <2.0 mm. The PH pattern is consistent with physiologically raised pulmonary pressures in the first hours after birth and also the increase in flow through a larger PDA which will also inevitably increase pressure in the pulmonary circulation.
Understanding how ductal diameter varies with flow pattern provides new insight into the clinical application of echocardiographically based treatment algorithms. For example, if a ductal diameter >2.0 mm is used to select infants for treatment, as Carmo et al13 did for infants in the first 12 h of life in a recent randomised controlled trial (RCT), the likelihood of unnecessarily treating an infant with a closing pattern is small: as shown here, only one of 52 echocardiographic traces classified as having a closing pattern had a ductal diameter >2.0 mm. If, instead, treatment is indicated by a pulsatile or growing pattern, as was done in another RCT,14 a substantial proportion of infants may be treated despite having a ductal diameter <2.0 mm: as found here, 40 of the 83 echocardiographic traces classified as growing or pulsatile had a ductal diameter <2.0 mm. These numbers are only indicative, however, because this retrospective study included echocardiographic examinations performed on preterm infants irrespective of postnatal age, an important factor in any treatment algorithm. There is currently no clinical information to suggest which of the two classification schemes is better.
Limitations of the current study include the fact that a significant portion (14.2%) of echocardiographic traces could not be clearly classified into one of the four ductal flow patterns. The patterns were characterised by only one person, but the original strict definitions were used.9 In this retrospective study, we cannot exclude a technical or operational error, such as inaccurate angling of the ultrasound probe. Additional limitations are the inclusion of examinations performed on infants of any postnatal age, in order to have a substantial number of echocardiographic traces for analysis, although analyses before and after 2 days of age showed similar findings to the total group.
In conclusion, this retrospective study demonstrates good agreement between two commonly used echocardiographic systems for identifying preterm infants needing treatment for PDA, but at the same time illustrates how implementation of one or the other parameter in treatment algorithms will result in different treatment decisions for some infants. Which of the echocardiographic parameters can better guide treatment decisions, or if superior results can be achieved with a combination of parameters is unknown. The accuracy of echocardiographic measures is variable and since it is simple to simultaneously measure ductal diameter and classify ductal flow pattern, both measures may improve the clinical utility of echocardiographic prediction of ductal significance. This concept could be tested in future clinical trials.
Acknowledgments
Valerie Matarese provided editorial advice and scientific writing assistance.
References
Footnotes
Presented in abstract form at the European Society for Paediatric Research meeting (Nice 2008) and at the American Pediatric Society/Society for Pediatric Research meeting (Baltimore 2009).
-
Funding The North Shore Heart Research Foundation provided funding for this study.
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.