Archives ofDisease in Childhood 1993; 68: 58-61

CURRENT TOPIC

Diagnosis of patent ductus arteriosus in thepreterm newborn

Nick Evans

Persisting patency of the ductus arteriosus(PDA) remains an important complication forthe very preterm infant undergoing intensivecare. There are two separate, but related,issues with respect to the diagnosis of PDA inthe preterm, firstly there is the diagnosis ofthe physical patency of the duct and secondly,and much more controversially, there is theassessment of the haemodynamic and clinicalsignificance of the shunt through that patentduct. Eighty per cent of preterm infants withhyaline membrane disease will have a ductthat is physically patent during the first fourpostnatal days,' however only about a third ofthese infants will develop shunts large enoughto cause symptoms. This article will considerthese two issues, initially reviewing the currentstatus of clinical and echocardiographic diag-nosis of physical patency and then looking atthe available non-invasive methods for assess-ing the size of the shunt through the duct.

Department ofPerinatal Medicine,King George VHospital,Missenden Road,Camperdown,Sydney, NSW 2050,Australia

Correspondence to:Dr Evans.

Clinical diagnosisA hyperdynamic precordial impulse, fullpulses, widened pulse pressure, hepatomegaly,and a high parastemal systolic murmur havebeen described as the classical physical signsof the PDA. These signs usually appear fromabout day 5 onwards and, together with evi-dence of interrupted improvement of or wors-

ening respiratory status, have been establishedas the clinical criteria of haemodynamic signif-icance.2 The accuracy of these signs in diag-nosing a PDA is not high; Kupferschmid et alcompared clinical and echocardiographic find-ings in infants with symptomatic PDA andfound that bounding pulses and a murmur

were absent in, respectively, 15% and 20% ofthe patients.3 These authors suggested thatthe most sensitive diagnostic sign was a hyper-dynamic precordium, present in 95%. In theearly postnatal period these signs are even lesssensitive, with silent ductal shunting being thenorm until the latter half of the first postnatalweek.4 Wide pulse pressure has also not stoodup to critical analysis as a physical sign. Twostudies have now compared pulse pressure inpreterm infants with and without PDA,5 6 inboth studies no difference was found duringthe first postnatal week. A PDA seems to beassociated with a similar reduction in bothsystolic and diastolic pressure that in thesmallest infants may produce significanthypotension.5 False positive clinical findings

also occur, not all murmurs in preterm infantsare due to PDA, pulmonary stenosis or pul-monary flow murmurs being common reasonsfor misdiagnosis.The electrocardiogram (ECG) is usually

normal, particularly early on, though the pres-ence of abnormalities on the ECG may sug-gest that the condition is not simply a PDAand that further cardiac investigation is indi-cated. The chest radiograph may show somecardiomegaly and pulmonary plethora, how-ever the presence of parenchymal lung diseaseoften makes interpretation of both cardiac sizeand pulmonary vascularity difficult; 22% ofinfants with a symptomatic PDA in a study byHiggins et al showed no increase in radiologi-cal heart size.7

Echocardiographic diagnosisM mode and more recently Doppler and twodimensional echocardiography have given us awindow on the ductus arteriosus that hasallowed for more accurate diagnosis8 9 andalso more detailed study of the natural historyof ductal shunting.' 10 l There are basicallythree categories of ultrasound method for thediagnosis of PDA.The first are the M mode methods for

assessing the degree of left sided cardiac over-load. The ratio of the diameter of the leftatrium to that of the aortic root (the LA:Aoratio) is probably the most widely used ofthese methods. This ratio depends on the factthat left to right ductal shunting increases thevolume load on the left side of the heart so theleft atrium will dilate relative to the aortic rootwhich will be unaffected by the volume load(fig 1). Published mean normal ratios varybetween 086:1 and 1.3:1,12 13 and the ratiothat is said to indicate the presence of a PDAvaries between 1*15:1 and 1*4:1.12 1 Certainlyusing a cut off towards the upper end of thisrange will increase the specificity of this tech-nique. Used in isolation the LA:Ao is not veryspecific. False positives can result from left toright shunting through a ventricular septaldefect, any cause of left ventricular dysfunc-tion,'4 and mitral valve abnormalities. If the Mmode beam transverses the atrium at anoblique angle, this will also increase the ratio.On the other hand false negatives have beendescribed in infants who had large ductalshunts on clinical grounds and a normalLA:Ao. One study hypothesised that the

58

on June 5, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.68.1_Spec_N

o.58 on 1 January 1993. Dow

nloaded from

Diagnosis ofpatent ductus arteniosus in the preterm newborn

Figure 1 This shows theM mode imagefrom which the LA.Ao ratio is derived, the dottedline on the Duplex two dimensional image shows the position of theM mode beam throughthe aortic root. The Ao diameter is measured, at the end of diastole, between the anterioredges of the anterior and posterior aortic walls. The LA diameter is measured, at the end ofsystole, from the anterior edge of the posterior aortic wall to the endocardial surface of theposterior atrial wall. In this example the LA.Ao was 1 3:1 and the duct was closed.

atrium may be enlarging in a lateral rather thanan anteroposterior direction,'5 another that thismay be a reflection of more severe fluid restric-tion.'6 However the LA:Ao ratio remains a use-

ful method because of its simplicity and it pro-vides some index of the degree of left to rightshunting, once the presence of the PDA hasbeen confirmed by the other methods.The second category of diagnostic method

is the indirect use of pulsed Doppler. PulsedDoppler analysis of flow in the main pul-monary artery allows demonstration of thecharacteristic diastolic turbulence associatedwith a PDA (fig 2). This method is both sen-

sitive and quite specific.'7 18 There are onlytwo conditions that can mimic this flow pat-tern, aortopulmonary window and a coronary

Figure 2(A) Represents the normal pulsed Doppler flow pattern in the main pulmonaryartery with minimal diastolic turbulence. (B) Shows the pattern of main pulmonary arteryflow seen with a PDA with marked diastolic turbulence and a degree of systolic turbulence.

artery fistula to the pulmonary artery, both arevery rare. However, in isolation it does notgive any assessment of the size of the duct orthe shunt through it. Other authors have usedDoppler assessment of descending aorticflow,'9 looking for the reversed diastolic flowassociated with a large ductal shunt. Thismethod will pick up the largest and probablythe clinically important ductal shunts, but itwill not detect the small to moderate sizedPDA which will not necessarily produce thispattern of flow.The third and most accurate method of

echocardiographic diagnosis is direct imagingof the duct with direct pulsed wave or colourflow Doppler analysis of the shunt.9 20 Theduct can be imaged from the high leftparasternal position with the beam cutting atrue saggital section through the body. If theoperator angles the beam to the left throughthe long axis of the main pulmonary artery,the pulmonary end of the duct can be seenjust superior to the root of the left pulmonaryartery as it emerges posteroinferiorly (fig 3). Ifthe pulsed Doppler sample is placed in theduct then the shunt pattern can be assessed.With colour Doppler, the flow through thePDA is seen as an orange jet streaming backup the anterior wall of the main pulmonaryartery (fig 3). It is also possible to assessdirectly the ductal shunt with continuouswave Doppler, this allows analysis of a greaterrange of shunt velocities than pulsed waveDoppler. While assessment of the size of thelumen of the ductus arteriosus and thestrength and direction of the Doppler signalfrom the shunt will give a qualitative measureof the size of the shunt, this method does notquantify the degree of shunting. This tech-nique does require more echocardiographicexpertise than the M mode or indirectDoppler methods. In the early postnatal daysclear images of the whole length of the ductcan be obtained, however towards the end ofthe first week imaging sometimes becomesmore difficult, particularly when the infant isventilated, as the left lung hyperinflates overthe parasternal ultrasound windows.A good approach to echocardiographic diag-

nosis is to start with direct imaging andDoppler, only if that is not possible then to useDoppler in the main pulmonary artery. Then,as an assessment of shunt size, to measure theLA:Ao ratio and the descending aortic flowpattern, with retrograde diastolic aortic flowand an LA:Ao >1 4 indicating a PDA likely tobe of haemodynamic significance.

Diagnosis of the haemodynamically sig-nificant PDAThe fundamental problem that affects moststudies of the PDA is how haemodynamicsignificance should be defined. Should it bedefined by clinical criteria of when a PDAbecomes symptomatic or should it be definedby echocardiographic criteria of the size of theshunt through the duct? Much of the work onthis has involved applying echocardiographyto infants who were felt to have symptomatic

59

on June 5, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.68.1_Spec_N

o.58 on 1 January 1993. Dow

nloaded from

60

-:_Figure 3 (A) Is the direct view of the PDA obtainedfrom the high right parasternal win-dow. A small PDA is seen running between the main pulmonary artery (MPA) and thedescending aorta just supenior to the root of the left pulmonary artery (LPA). (B) Is thesame image with the colourflow Doppler mapping superimposed. The orange/redflowthrough the PDA is seen streaming up the anterior wall of the MPA whose fonvardflow isrepresented by the blue.

ducts and then defining echocardiographiccriteria for haemodynamic significance fromthis .2 12 13 15

If clinical criteria are used, which one accu-

rately reflects the haemodynamic status of thePDA?3 The important symptoms of the PDAare respiratory and, because these infantsalmost always have associated acute and oftenchronic respiratory disease, there is scope formisinterpreting the influence of the duct onthe respiratory status. In the early acutephase the clinical signs and influence of thePDA are usually not seen until the hyalinemembrane disease starts to improve.Traditionally this has been interpreted as

reflecting the increased left to right shuntthrough the duct resulting from falling pul-monary vascular resistance at this time. Whilethis is certainly true, echocardiographic studiesof ductal shunting in hyaline membrane dis-ease have shown that even in the first threepostnatal days that the usual direction ofshunting is left to right' and that shunts thatare going to later cause clinical symptoms are

usually significant, by echocardiographiccriteria, by 48 hours after delivery." 21 Thelater chronic respiratory phase, in affectedinfants, is characterised by fluctuations in res-

piratory status and oxygenation, these fluctua-tions are seen whether the duct is patent or

not though often these changes are ascribed tothe duct when it is patent.Of the accepted physical signs of PDA, the

lack of sensitivity of full pulses, wide pulsepressure, and the murmur have been men-tioned above. And while an hyperdynamicprecordium is a sensitive sign when symptomsdevelop3 it has not been assessed as a methodfor early recognition of PDA. Likewise hep-atomegaly is often found in the chronic respi-ratory phase; this can be caused by hyper-inflated lungs and or elevated pulmonaryartery pressures as well as ductal shunts. Thuswhile clinical symptoms and signs are clearly

important in the assessment of ductal status,they are open to misinterpretation and it isdifficult to argue that they should be the 'goldstandard' of haemodynamic significance.

In physiological terms the measure ofhaemodynamic significance is the size of theshunt relative to the baseline cardiac output.Accurate measurement of this requires car-diac catheterisation and is clearly not practi-cal or ethical in preterm infants. So whichechocardiographic criteria have the best cor-relation with these invasive measures? The Mmode measures of left sided volume over-load, that is left atrial diameter (LAD) andleft ventricular end diastolic diameter(LVEDD), were the first to be studied in thisfield. Left atrial volume and left ventricularend diastolic volume, measured by angio-graphic techniques, have been correlatedwith ventricular septal defect and PDA shuntsize.22 Left atrial diameter, measuredechocardiographically, has been shown tocorrelate well with left atrial volume23 andalso with shunt size through ventricular sep-tal defects.24 Because cardiac catheterisationis rarely indicated in the preterm newborn,the available data is derived from the studyof older subjects. When applied to the studyof the newborn, both LAD and LVEDDwere related to aortic root diameter to con-trol for the range of body size.'2 In the new-born these measures have been shown to beincreased in infants with symptomatic PDA'213 and to fall promptly after ductal ligation.'3Direct correlation with invasive measures ofshunt size has not been performed in thepreterm. There do seem to be good reasonsfor suggesting that the relationship describedin older subjects24 holds for the neonatalperiod and, allowing for specificity problemsmentioned earlier, these methods remain thesimplest means, if not necessarily the mostaccurate, of assessing the degree of shuntthrough a duct.Two methods of assessing shunt size have

evolved from Doppler studies of flow in thedescending aorta. One method involvedpulsed Doppler assessment of flow volume inthe aortic arch before and after the ductalinsertion.25 This method has not been vali-dated and because Doppler flow assessmentsdepend on an accurate peak velocity, itdepends on being able to insonate at an angleof less than 200, not always easy in thedescending aorta. Despite these limitationsthese authors derived estimates of shunt sizewhich were in line with what has beendescribed using other methods. A simplermethod was described by Serwer et al whoshowed in infants outside the neonatal periodthat there was a close correlation between theratio of the velocity time integrals (that is thearea under Doppler frequency shift/timecurve) of the retrograde diastolic and systolicantegrade flow and the size of the shuntthrough a PDA measured by a radioisotopemethod.'9 This method has two advantages,firstly it does not depend on the angle ofinsonation and secondly it has been validatedagainst invasive measures of shunt size. If

Evans

on June 5, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.68.1_Spec_N

o.58 on 1 January 1993. Dow

nloaded from

Diagnosis ofpatent ductus arteriosus in the preterm newborn

there is diastolic retrograde flow in thedescending aorta then it is highly likely thatthe shunt through the PDA is significant.

Non-invasive Doppler methods of assessingcardiac output have been well studied and val-idated in the newborn.2627 The product of theDoppler velocity time integral from eitherventricular outflow tract and the area of thatoutflow tract gives a good estimate of strokevolume. The product of stroke volume andthe heart rate gives the cardiac output. Left toright shunting at ductal level produces anincreased volume load to the left side of theheart and the left ventricle has to increase out-put to cope. In theory the left ventricular out-put minus the right ventricular output shouldbe equal to the ductal shunt, but in practicethis may not be completely accurate in thenewborn because there is always a degree ofincompetence of the foramen ovale and so aleft to right atrial shunt. While this shunt isoften more obvious in infants with PDAbecause of the left atrial dilatation, it has notbeen thought to be of great haemodynamicsignificance.28 While it has been shown thatthere is an increase in left ventricular outputin association with PDA in the preterm new-born which falls after duct closure, the degreeof this increase has not yet been directly corre-lated with shunt size.21When considering the relative merits of

clinical and echocardiographic criteria of duc-tal patency we have to consider not only accu-racy, but also the method which allows theearliest definition of the duct which is causingor will cause problems. Echocardiography isbetter on both these counts but particularlyon the latter. Using M mode criteria orincreasing left ventricular output measured byDoppler, it is possible by day 3 of life to pre-dict the patent ducts which will later becomesymptomatic." 21 Mellander et al showed thatusing M mode and Doppler diagnosis on day3 of life, one could predict with 100% sensi-tivity and 85% specificity those PDAs thatwould later become symptomatic." Similarly,Walther et al showed that an increase of leftventricular output of more than 60 ml/kg/minconsistently preceded the development ofsymptomatology by at least 24 hours.2' This isnot dissimilar to another area of cardiology,that of valvular stenosis, where the aim ofechocardiography is to define haemodynamicsignificance so that clinical symptoms can bepredicted and so avoided. It seems rationalthat our approach to PDA in the preterminfant should be the same; by the time clinicalsymptoms appear it may well be too late. Thisis speculative, and whether treatment of thepatent duct on echocardiographic rather thanclinical criteria will improve the outlook forpreterm infants remains an open question thatneeds to be answered by a randomised con-trolled trial.

1 Evans NJ, Archer LNJ. Doppler assessment of pulmonaryartery pressure and extrapulmonary shunting in the acute

phase of hyaline membrane disease. Arch Dis Child199 1;66:6-1 1.

2 Gersony W, Peckham GL, Lang P, et al. Effects ofindomethacin in premature infants with patent ductusarteriosus: results of a collaborative study. J Pediatr1983;102:895-906.

3 Kupferschmid Ch, Lang D, Pohlandt F. Sensitivity, speci-ficity and predictive value of clinical findings, M-modeechocardiography and continuous wave Doppler sonogra-phy in the diagnosis of symptomatic patent ductus arte-riosus in preterm infants. EurJ Pediatr 1988;147:279-82.

4 Hammerman C, Strates E, Valaitis S. The silent ductus: Itsprecursors and aftermath. Pediatr Cardiol 1986;7:121-7.

5 Evans NJ, Moorcraft J. Effect of patency of the ductusarteriosus on blood pressure in very preterm infants. ArchDis Child 1992;67:1169-73.

6 Ratner I, Perelmuter B, Toews W, Whitfield J. Associationof low systolic and diastolic blood pressure with signifi-cant patent ductus arteriosus in the very low birth weightinfant. Crit Care Med 1985;13:497-500.

7 Higgins CB, Rausch J, Friedman WF, et al. Patent ductusarteriosus in preterm infants with idiopathic respiratorydistress syndrome. Radiology 1977;124:189-95.

8 Huhta JC, Cohen M, Gutgesell HP. Patency of the ductusarteriosus in normal neonates: two dimensional echocar-diography versus Doppler assessment. J Am Coll Cardiol1984;4:561-4.

9 Vick GW, Huta Gutgesell HP. Assessment of the ductusarteriosus in preterm infants utilising suprasternal twodimensional/Doppler echocardiography. J Am Coll Cardiol1985;5:973-7.

10 Evans NJ, Archer LNJ. Postnatal circulatory adaptation inhealthy term and preterm newborns. Arch Dis Child1990;65:24-6.

11 Mellander M, Larsson LE, Ekstrom-Jodal B, Sabel KG.Prediction of symptomatic patent ductus arteriosus inpreterm infants using Doppler and M-mode echocardio-graphy. Acta Paediatr Scand 1987;76:553-9.

12 Silverman NH, Lewis AB, Heymann MA, Rudolph AM.Echocardiographic assessment of ductus arteriosus shuntin premature infants. Circulation 1974;50:821-5.

13 Johnson GL, Breart GL, Gewitz MH, et al. Echo-cardiographic characteristics of premature infants withpatent ductus arteriosus. Pediatrics 1983;72:864-7 1.

14 Sahn DJ, Vaucher Y, Williams DE, et al. Echocardiographicdetection of large left to right shunts and cardiomyopathiesin infants and children. Am Jf Cardiol 1976;38:73-9.

15 Hirschlau MJ, Disessa TG, Higgins CB, Friedman WF.Echocardiographic diagnosis: pitfalls in the prematureinfant with a large patent ductus arteriosus. I Pediatr1978;92:474-7.

16 Valdes-Cruz LM, Dudell GG. Specificity and accuracy ofechocardiographic and clinical criteria for diagnosis ofpatent ductus arteriosus in fluid restricted infants.I Pediatr 1981;98:298-305.

17 Stevenson JG, Kawabori I, Guntheroth WG. PulsedDoppler echocardiographic diagnosis of patent ductusaarteriosus. sensitivity, specificity, limitations and techni-cal features. Cathet Cardiovasc Diagn 1980;6:255-63.

18 Gentile R, Stevenson G, Dooley T, et al. Pulsed Dopplerechocardiographic determination of the time of ductalclosure in normal newborn infants. _J Pediatr 1981;98:443-8.

19 Serwer GA, Armstrong BE, Anderson PAW. Continuouswave Doppler ultrasonographic quantitation of patentductus arteriosus flow. Y Pediatr 1982;100:297-9.

20 Reller MD, Ziegler ML, Rice MJ, Solin RC, McDonaldRW. Duration of ductal shunting in healthy preterminfants: an echocardiographic color flow Doppler study.YPediatr 1988;112:441-6.

21 Walther I;, Kim DH, Ebrahimi M, Siassi B. PulsedDoppler measurement of left ventricular output as anearly predictor of symptomatic patent ductus arteriosusin very preterm infants. Biol Neonate 1989;56:121-8.

22 Jarmakani MM, Graham TP, Canent RV, Spach MS,Capp MP. Effect of site of shunt on left heart volumecharacteristics in children with ventricular septal defectand patent ductus arteriosus. Circulation 1969;40:411-8.

23 Hirata T, Wolfe SB, Popp RL, Helman CH, FeigenbaumH. Estimation of left atrial size using echocardiography.Am Heartj 1969;78:43-52.

24 Carter WH, Bowman CR. Estimation of shunt flow in iso-lated ventricular septal defect by echocardiogram.Circulation 1973;48:(suppl IV):IV-64.

25 Drayton MR, Skidmore R. Ductus arteriosus blood flowduring the first 48 hours of life. Arch Dis Child1987;62: 1030-4.

26 Mellander M, Sabel K, Caidal K, Solymar L, Eriksson B.Doppler determination of cardiac output in infants andchildren: comparison with simultaneous thermodilution.Pediatr Cardiol 1987;8:241-6.

27 Mandelbaum-Isken VH, Linderkamp 0. Pulsed Dopplerevaluation of cardiac output in neonates using an apicalwindow. Pediatr Cardiol 1991;12:13-6.

28 Zhou TF, Guntheroth WG. Valve incompetent foramenovale in premature infants with ductus arteriosus: aDoppler echocardiographic study. J Am CoU Cardiol1987;10:193-9.

61

on June 5, 2020 by guest. Protected by copyright.

http://adc.bmj.com

/A

rch Dis C

hild: first published as 10.1136/adc.68.1_Spec_N

o.58 on 1 January 1993. Dow

nloaded from