Journal of Perinatology (2007) 27, 291–296 r 2007 Nature Publishing Group All rights reserved. 0743-8346/07 $30 www.nature.com/jpORIGINAL ARTICLEDiagnosis of patent ductus arteriosus by a neonatologistwith a compact, portable ultrasound machineHC Lee1, N Silverman2 and SR Hintz31 Department of Pediatrics, Division of Neonatal and Developmental Medicine, Stanford University, Palo Alto, CA, USA; 2Departmentof Pediatrics, Division of Pediatric Cardiology, Stanford University, Palo Alto, CA, USA and 3Department of Pediatrics, Divisionof Neonatal and Developmental Medicine, Stanford University, Palo Alto, CA, USA hemorrhage, bronchopulmonary dysplasia and pulmonary Objectives: To conduct a pilot study assessing a neonatologist’s accuracy hemorrhage.1–5 In a review of periventricular leukomalacia in in diagnosing patent ductus arteriosus (PDA) using compact, portable preterm infants, PDA was associated with decreased cerebellar ultrasound after limited training. volume, area of the vermis and diameter of the pons.6 The Study design: Prospective study of premature infants scheduled for frequency of PDA is high in premature infants, ranging from echocardiography for suspected PDA. A neonatologist with limited training 53% in infants born before 34-weeks gestation, up to 65% in performed study exams before scheduled exams. Sensitivity and speciﬁcity infants born before 26-weeks gestation.7,8 PDA in preterm were calculated, compared to the scheduled echocardiogram interpreted infants often requires pharmacologic or surgical closure. by a cardiologist. Treatment of PDA has been shown to improve pulmonary function, including increase in dynamic compliance, tidal volume and Results: There were 24 exams. Compared to the scheduled exam, the minute ventilation.9 Early pharmacological treatment has also neonatologist’s exam had sensitivity 69% (95% conﬁdence interval (CI), been shown to reverse low renal and splanchnic blood ﬂow in 41 to 89%) and speciﬁcity 88% (95% CI, 47 to 99%). When a cardiologist infants with PDA.10 interpreted the study exams, the sensitivity was 87% (95% CI, 60 to 98%) Although there are clinical signs for diagnosis of PDA including and speciﬁcity 71% (95% CI, 29 to 96%). auscultation of a cardiac murmur, bounding pulses and wide pulse Conclusion: A neonatologist with limited training was able to detect PDA pressure, studies have shown that physical exam is inadequate in with moderate success. A more rigorous training process or real-time detecting signiﬁcant PDAs in preterm infants.11–13 The current transmission with cardiologist interpretation may substantially improve standard of care for diagnosis of PDA is an echocardiogram, yet accuracy. Institutions with experienced technicians and on-site pediatric echocardiography is not routinely taught to neonatologists. cardiologists may not gain from intensive training of neonatologists, but Nevertheless, it has been suggested that patient care could be hospitals where diagnosis and treatment of PDA would be delayed may improved in the neonatal intensive care unit (NICU) if beneﬁt from such processes. neonatologists were trained in echocardiography.14 Journal of Perinatology (2007) 27, 291–296. doi:10.1038/sj.jp.7211693; Advances in technology are allowing echocardiography to be published online 15 March 2007 performed by nontraditional personnel in certain settings. Ultrasound technology has become more compact, allowing use as Keywords: patent ductus arteriosus; echocardiography; neonatologist; diagnosis a point-of-care diagnostic tool. A handheld ultrasound was found to be effective in the diagnosis of congenital heart disease, including PDA.15 In this series, the interpreters were experienced echocardiographers. Such devices are being utilized in settingsIntroduction such as the emergency department and surgical ICUs where aPatent ductus arteriosus (PDA) in premature infants is an quick, focused evaluation could beneﬁt the patient.16 Neonatalassociated factor in serious neonatal morbidities, associated units may also beneﬁt from small handheld ultrasound devices forwith increased risk of necrotizing enterocolitis, intraventricular assessment of umbilical line placement and cerebral blood ﬂow.Correspondence: Dr SR Hintz, Department of Pediatrics, Division of Neonatal and We performed a pilot study to assess the potential utility of aDevelopmental Medicine, Stanford University, 750 Welch Road, Suite 315, Palo Alto, limited training program to diagnose PDA for a neonatologist withCA 94304, USA. no cardiology background. We used a portable ultrasoundE-mail: firstname.lastname@example.orgReceived 5 September 2006; revised 18 December 2006; accepted 10 January 2007; published machine, one which could potentially be used in small NICUsonline 15 March 2007 without ready access to a pediatric cardiologist.
Diagnosis of PDA by a Neonatologist HC Lee et al292Methods the treating physician decided that the patient should not beThis was a prospective, masked pilot study of the accuracy of PDA enrolled for any reason, the scheduled echocardiogram would bediagnosis by neonatologist-performed compact ultrasound exam performed before the study exam could be completed, or if thecompared with routine echocardiography. This pilot study was trained neonatologist was unavailable to perform the study exam.approved by the Stanford University Institutional Review Board. Informed consent was obtained for all patients who participated in The ‘routine echocardiogram’ was the exam ordered by the the study.medical team taking care of the patient for clinical indications. On some occasions, an infant received more than oneThis exam was performed by an experienced technologist or echocardiogram to follow-up on the status of PDA. For thesepediatric cardiologist using the Acuson Sequoia (Siemens USA, infants, the study exam could be repeated for a maximum of twoMalvern, PA, USA), the primary device used by the pediatric times, before each routine echocardiogram.cardiology service in our institution. The ‘study exam’ was performed by a neonatologist who Procedureunderwent training in echocardiography, focusing on diagnosis of Enrolled patients underwent the study exam consisting of aPDA; no other neonatologists were trained. This exam was compact, portable ultrasound exam (Acuson Cypress, Siemens USA)performed with the Acuson Cypress (Siemens USA), a portable by the study neonatologist before the routine echocardiogramultrasound machine, approximately the size of a briefcase, with performed by pediatric cardiology. Using the information asfewer advanced capabilities than the Sequoia, but with features outlined previously, the neonatologist made a determination ofsuch as Doppler and M-mode. The standard Cypress neonatal probe patent ductus if color Doppler views indicated shunting acrosswas used at a frequency of 7.5 MHz. The training process included the ductus. A subjective determination of ductal size was made:instruction on the working of the ultrasound machine, application small, moderate or large. Results of the study exam were notof the transducer to the various sites on the patient, addition of revealed to caregivers, cardiologists or families and did notDoppler techniques, and pulsed continuous wave to denote inﬂuence treatment decisions. No clinical actions were based onmagnitude of shunting. Five factors were considered as criteria for these study results. These results were noted on a dated and timeddiagnosis of a patent ductus: (1) left atrial size assessed from the conﬁdential study form immediately after the study exam, and sentaortic root and the left atrial aortic root relation, (2) left to two different conﬁdential electronic email accounts, to assureventricular size and function, (3) Color Doppler estimate of ductus they would not be amended. The neonatologist recorded thesize at the point of the vena contracta (the narrowest portion of the patency of the ductus, and if patent, an interpretation of the sizeﬂow jet), (4) the velocity and character of the ductus Doppler of the ductus.signal, and (5) the amount of retrograde abdominal ﬂow. The The primary outcome variable to be measured was the accuracytraining included visualization of the parasternal long axis view, of the diagnosis of PDA by the neonatologist using the compact,the parasternal short axis view to visualize pulmonary arteries, portable ultrasound machine. The routine exam performed by theapical four-chamber view, subcostal view and suprasternal views, pediatric cardiology service was considered the gold standard forincluding the ductus cut.17 Using this information, the diagnosis. In general, this exam was performed by a skilledneonatologist made a determination of patent ductus if the color pediatric echocardiography technician or member of the pediatricDoppler views indicated shunting across the ductus. Pulse wave cardiology staff or faculty, and subsequently interpreted by aDoppler in the descending aorta was used to augment the pediatric cardiologist. None of these practitioners knew the resultsdiagnosis. A subjective determination of ductal size was made of the study exam.(small, moderate or large) based on this information. The study exams were recorded on electronic media and later The total training experience consisted of 2 h of lecture reviewed in a masked fashion by a cardiologist who was notincluding recorded tapes, observation of eight exams by familiar with the patients in the study. This cardiologist knewexperienced technicians, and three practice exams with guidance neither the study exam interpretation by the neonatologist, nor theon real patients. Approximately 50% of these evaluations had PDA. routine exam interpretation by the pediatric cardiologist. The studyThe total training time was approximately 8 h. Midway through the cardiologist interpreted the study exams and also graded the qualitystudy, the study cardiologist reviewed the study exams already done of the exams in a subjective fashion. For some selected study examswith the neonatologist for 1 h. in which there were discrepancies between the study exam results and the routine exam, the study cardiologist also reviewed theSubjects routine exam.Patients in the NICU who were undergoing evaluation only forsuspected PDA, with birth weight 401 to 2000 g or <34-weeks Data analysisgestational age were eligible for this pilot study. The period of Sensitivities and speciﬁcities with 95% conﬁdence intervals of theenrollment was from January to July 2005. Infants were excluded if neonatologist’s study exam compared with the routineJournal of Perinatology
Diagnosis of PDA by a Neonatologist HC Lee et al 293echocardiogram were calculated. After the primary analysis, we also within six h of the cardiology exam for 20 of the studies. Two studylooked at the diagnosis of moderate or large PDAs as determined by exams occurred 9 h before (exam no. 2) and 23 h before (examthe routine exam. no. 10) the cardiology exam. The exact timing of the cardiology We also calculated sensitivity and speciﬁcity of the study exam for two of the studies was unable to be determined (examscardiologist’s interpretation of the study exam performed by the nos. 19 and 21). The study cardiologist did not routinely go overneonatologist compared with the routine echocardiogram results, the routine exams by the pediatric cardiology service that correlatedagain with the routine echocardiogram considered as the gold with the study exams. However, on one occasion, when thestandard. The study cardiologist was masked to both the cardiologist saw a PDA in the study exam for which the cardiologyneonatologist’s interpretation of the study exam, and the formal service had reported no PDA, he reviewed both studies and hisinterpretation of the routine echocardiogram performed by the interpretation of both the study exam and the routine exam waspediatric cardiology service. the presence of a small PDA (exam no. 8). Compared to the routine echocardiogram, the study exam as interpreted by the neonatologist had 69% sensitivity (95% CI, 44 toResults 86%) and 88% speciﬁcity (95% CI, 53 to 98%) (Table 2). TheThere were 24 exams performed on a total of 14 patients (Table 1). positive predictive value was 92%, whereas the negative predictiveAll study exams were performed before the routine echocardiogram value was 58%. In the ﬁve cases where the neonatologist interpretedperformed by the cardiology service. Study exams were performed the study exam as negative when the routine cardiology evaluationTable 1 Interpretation of study exams and routine examsExam no. Birth weight Gestational age Age at time Study exam Routine exam (grams) (weeks) of study Neo interpretation Cardiologist interpretation PDA Y/N Size PDA Y/N Size PDA Y/N Size 1 860 30 78 h Y Small Y Tiny Y Small-moderate 2 1390 32 67 h Y Small Y Tiny Y Moderate 3 1390 32 5d Y Moderate Y Small Y Moderate 4 647 24 13 d Y Small N N 5 1197 28 7d N N Y Very small 6 1053 29 55 h Y Large Y Big Y Large 7 1053 29 101 h Y Moderate Y Moderate Y Moderate 8 820 24 7d Y Small Y 1 mm Ya Smalla 9 1146 30 5d N Y Tiny Y Tiny10 820 24 13 d N N N11 968 29 6d N N N12 968 29 16 d Y Small N Y Small-moderate13 866 25 42 h N Y Large Y Large14 969 26 78 h Y Small Y Small Y Large15 969 26 5d Y Small Y Small Y Small b16 866 25 8d N N17 969 26 8d N Y Small N18 740 26 11 d N N N19 740 26 17 d N Y Tiny N20 1125 28 5d N N N21 1470 30 106 h N Y Large Y Moderate22 1743 31 5d Y Moderate Y Small Y Small23 1743 31 7d Y Small Y Tiny Y Small b24 1743 31 14 d N Y TinyAbbreviations: d, days; N, no; PDA, patent ductus arteriosus; Y, yes.a Exam was initially noted to be negative by routine exam cardiologist, however upon secondary review, there was noted to be a small PDA.b Exam was judged to be inadequate by study cardiologist. Journal of Perinatology
Diagnosis of PDA by a Neonatologist HC Lee et al294Table 2 Results of routine exams and study exams interpreted by Table 3 Results of routine exams and study exams interpreted byneonatologist neonatologist, excluding small PDAs Routine Routine Routine Routine ECHO: PDA ECHO: no PDA ECHO: PDA ECHO: no PDAStudy exam: PDA 11 1 12 Study exam: PDA 7 1 8Study exam: no PDA 5 7 12 Study exam: no PDA 2 7 9 16 8 24 9 8 17Abbreviation: PDA, patent ductus arteriosus. Abbreviation: PDA, patent ductus arteriosus.Sensitivity ¼ 11/16 ¼ 0.69; positive predictive value ¼ 0.92. Sensitivity ¼ 7/9 ¼ 0.78; positive predictive value ¼ 0.88.Speciﬁcity ¼ 7/8 ¼ 0.88; negative predictive value ¼ 0.58. Speciﬁcity ¼ 7/8 ¼ 0.88; negative predictive value ¼ 0.78.was positive for PDA, three of those ﬁve cases had PDAs which were Table 4 Results of routine exams and study exams interpreted by studyconsidered small or tiny (Table 1). cardiologist Six PDAs were considered to be small or tiny by echocardiogram Routine Routineperformed by the cardiology service; if these exams were excluded, ECHO: PDA ECHO: no PDAthe sensitivity was 78% (45 to 94%) and speciﬁcity 88% (53 to 98%)(Table 3). The sensitivity and speciﬁcity of the masked cardiologist Study exam: PDA 13 2 15interpretation of the study exam compared with the routine Study exam: no PDA 2 5 7echocardiogram were also determined (Table 4). Two of the studies 15 7 22were deemed uninterpretable owing to poor quality. The study Abbreviation: PDA, patent ductus arteriosus.cardiologist interpretation of the remaining 22 studies had Sensitivity ¼ 13/15 ¼ 0.87; positive predictive value ¼ 0.87. Speciﬁcity ¼ 5/7 ¼ 0.71; negative predictive value ¼ 0.71.sensitivity of 87% (62 to 96%) and speciﬁcity 71% (36 to 92%). increasingly by nontraditional practitioners, such as nurses inDiscussion obstetrics, trauma surgeons and emergency physicians.20–23In our pilot study, we found that a neonatologist with very limited By using telephone lines, real-time transmission oftraining in echocardiography was able to detect PDA in preterm echocardiography can be performed. Some NICUs are takinginfants with moderate sensitivitiy and speciﬁcity using a portable, advantage of this technology, in which a cardiologist guides acompact ultrasound device; accuracy improved further when technician to perform the study, then interprets the study from aconsidering only moderate to large PDAs. A cardiologist interpreting distance.24 In that series, 182 of the 500 exams performed were forthe study exams had slightly better success than the neonatologist. suspected PDA. There may be practicing neonatologists who doThese preliminary ﬁndings in a small number of patients suggest perform echocardiography in their NICUs for the diagnosis of PDA,that a more intensive training program, perhaps with integration particularly in situations when a cardiologist is not available.of real-time evaluation techniques, could result in further However, we are not aware of any previous studies reportingdiagnostic improvement. Although such a program may not be training regimens, accuracy or reliability of neonatologist-necessary in hospitals with pediatric cardiologists and technicians, performed exams using a portable bedside device.it may be an appropriate and feasible approach for institutions There were several limitations to this study. The methodologywithout this consistent on-site availability. included several factors, any or all of which may have effected the This is the ﬁrst study describing the training of a neonatologist study results: (1) the use of a smaller portable ultrasound device asto interpret ultrasound for the diagnosis of PDA. A previous study compared to the more sophisticated machine used by thelooked at echocardiography by a neonatologist of infants suspected cardiology service, (2) the fact that only one neonatologistto have congenital heart disease. In that study, the cardiologist performed the study exams, and (3) the training process. Theinterpretation was not considered a true gold standard, and a images on the portable device were not as clear as the usualCohen’s kappa of 0.84 of overall agreement was reported.18 If the machine. However, it did feature the ability to perform Dopplercardiologist interpretation had been considered the gold standard, ultrasound and the large majority of exams were consideredthe sensitivity of the neonatologist’s exam would have been 75%. sufﬁcient by the study cardiologist. We may have seen different Technological advancements have made ultrasound devices results had more than one neonatologist performed the studymore readily available and portable. A recent study found that a exams. However, the device was available for use on a limitedhandheld ultrasound device had utility in diagnosing valvular basis and it was not practical for more than one person toregurgitation in adults.19 Ultrasounds are being performed perform the studies. It may be the case that another neonatologistJournal of Perinatology
Diagnosis of PDA by a Neonatologist HC Lee et al 295with the same training and similar patients may have performed 3 Tortorolo L, Vento G, Matassa PG, Zecca E, Romagnoli C. Early changesdifferently. of pulmonary mechanics to predict the severity of bronchopulmonary We reviewed the two cases (nos. 13 and 21) in which the study dysplasia in ventilated preterm infants. J Matern Fetal Neonatal Med 2002;neonatologist missed the diagnosis of a large PDA. Both study 12: 332–337.exams by the neonatologist were technically competent with clear 4 Redline RW, Wilson-Costello D, Hack M. Placental and other perinatal risk factors for chronic lung disease in very low birth weight infants. Pediatr Resviews of the PDA on several views, retrograde abdominal aorta ﬂow. 2002; 52: 713–719.The study cardiologist (and in retrospect, the study neonatologist) 5 Lin TW, Su BH, Lin HC, Hu PS, Peng CT, Tsai CH et al. Risk factors ofvisualized a large PDA in both studies. We attribute the missed pulmonary hemorrhage in very-low-birth-weight infants: a two-yeardiagnoses to lack of experience and conﬁdence on the part of a retrospective study. Acta Paediatr Taiwan 2000; 41: 255–258.novice interpreting the exams. 6 Argyropoulou MI, Xydis V, Drougia A, Argyropoulou PI, Tzouﬁ M, Bassounas A longer training program may also have allowed for improved A et al. MRI measurements of the pons and cerebellum in children borndiagnosis of PDA. We tried to simulate a relatively short, 1–2 days preterm; associations with the severity of periventricular leukomalacia andtraining program that would allow participation by a busy perinatal risk factors. Neuroradiology 2003; 45: 730–734.neonatologist, in a setting where direct supervision would likely be 7 Costeloe K, Hennessy E, Gibson AT, Marlow N, Wilkinson AR. The EPICureunavailable after the training program. The training involved study: outcomes to discharge from hospital for infants born at the thresholdeducation in learning how to use the ultrasound device itself, of viability. Pediatrics 2000; 106: 659–671.learning the various views, and the nuances of the diagnosis of 8 Hammoud MS, Elsori HA, Hanaﬁ EA, Shalabi AA, Fouda IA, Devarajan LV. Incidence and risk factors associated with the patency of ductus arteriosus inPDA in a very small infant. We found that the neonatologist had preterm infants with respiratory distress syndrome in Kuwait. Saudi Med Jroom for improvement in both the technical skills of 2003; 24: 982–985.echocardiography as well as interpretation of exams to diagnose 9 Szymankiewicz M, Hodgman JE, Siassi B, Gadzinowski J. MechanicsPDA. of breathing after surgical ligation of patent ductus arteriosus in There can also be some subjectivity in the interpretation of newborns with respiratory distress syndrome. Biol Neonate 2004; 85:echocardiograms. In our study, we reviewed one particular patient 32–36.who had a PDA seen in the study exam by both the neonatologist 10 Shimada S, Kasai T, Hoshi A, Murata A, Chida S. Cardiocirculatoryand study cardiologist (exam no. 8). Although the practicing effects of patent ductus arteriosus in extremely low-birth-weightcardiologist’s interpretation of the routine exam in that infants with respiratory distress syndrome. Pediatr Int 2003; 45:circumstance had been no PDA, the study cardiologist interpreted 255–262.the routine exam as positive for PDA. However, practically 11 Davis P, Turner-Gomes S, Cunningham K, Way C, Roberts R, Schmidt B.speaking, most clinicians would consider that echocardiography is Precision and accuracy of clinical and radiological signs in premature infants at risk of patent ductus arteriosus. Arch Pediatr Adolesc Med 1995;the gold standard for diagnosis of PDA. 149: 1136–1141. In summary, we found that, even with extremely limited 12 Urquhart DS, Nicholl RM. How good is clinical examination at detecting atraining, a neonatologist was able to detect PDA with moderate signiﬁcant patent ductus arteriosus in the preterm neonate? Arch Dis Childsuccess. A more rigorous training process or real-time transmission 2003; 88: 85–86.with cardiologist interpretation and guidance could improve 13 Skelton R, Evans N, Smythe J. A blinded comparison of clinical andaccuracy and thus avoid inappropriate patient treatment. echocardiographic evaluation of the preterm infant for patent ductus arteriosus. J Paediatr Child Health 1994; 30: 406–411. 14 Skinner JR. Echocardiography on the neonatal unit: a job for theAcknowledgments neonatologist or the cardiologist? Arch Dis Child 1998; 78: 401–402.Siemens Medical Solutions USA provided equipment for this study. Henry Lee is 15 Li X, Mack GK, Rusk RA, Dai XN, El-Sedfy GO, Davies CH et al. Will athe recipient of a Fellowship Grant from Discovery Labs. Neither institution handheld ultrasound scanner be applicable for screening for heartparticipated in the design of the study, training of the study staff, or analysis of the abnormalities in newborns and children? J Am Soc Echocardiogr 2003; 16:results. 1007–1014. 16 Duvall WL, Croft LB, Goldman ME. Can hand-carried ultrasound devices be extended for use by the noncardiology medical community?References Echocardiography 2003; 20: 471–476. 1 Wang YH, Su BH, Wu SF, Chen AC, Lin TW, Lin HC et al. Clinical analysis of 17 Smallhorn JF, Huhta JC, Anderson RH, Macartney FJ. Suprasternal necrotizing enterocolitis with intestinal perforation in premature infants. cross-sectional echocardiography in assessment of patent ductus arteriosus. Acta Paediatr Taiwan 2002; 43: 199–203. Br Heart J 1982; 48: 321–330. 2 Evans N, Kluckow M. Early ductal shunting and intraventricular 18 Samson GR, Kumar SR. A study of congenital cardiac disease in a neonatal haemorrhage in ventilated preterm infants. Arch Dis Child Fetal Neonatal population – the validity of echocardiography undertaken by a Ed 1996; 75: F183–F186. neonatologist. Cardiol Young 2004; 14: 585–593. Journal of Perinatology
Diagnosis of PDA by a Neonatologist HC Lee et al29619 Kobal SL, Tolstrup K, Luo H, Neuman Y, Miyamoto T, Mirocha J et al. 22 Kirkpatrick AW, Simons RK, Brown R, Nicolaou S, Dulchavsky S. The Usefulness of a hand-carried cardiac ultrasound device to detect clinically hand-held FAST: experience with hand-held trauma sonography in a level-I signiﬁcant valvular regurgitation in hospitalized patients. Am J Cardiol urban trauma center. Injury 2002; 33: 303–308. 2004; 93: 1069–1072. 23 Holmes JF, Brant WE, Bond WF, Sokolove PE, Kuppermann N. Emergency20 Gegor CL, Paine LL, Costigan K, Johnson TR. Interpretation of biophysical department ultrasonography in the evaluation of hypotensive and proﬁles by nurses and physicians. J Obstet Gynecol Neonatal Nurs 1994; normotensive children with blunt abdominal trauma. J Pediatr Surg 2001; 23: 405–410. 36: 968–973.21 Brooks A, Davies B, Smethhurst M, Connolly J. Prospective evaluation 24 Sable CA, Cummings SD, Pearson GD, Schratz LM, Cross RC, Quivers ES of non-radiologist performed emergency abdominal ultrasound for et al. Impact of telemedicine on the practice of pediatric cardiology in haemoperitoneum. Emerg Med J 2004; 21: e5. community hospitals. Pediatrics 2002; 109: E3.Journal of Perinatology