Pediatr Cardiol DOI 10.1007/s00246-014-0877-7
ORIGINAL ARTICLE
Who Should Be Referred? An Evaluation of Referral Indications for Fetal Echocardiography in the Detection of Structural Congenital Heart Disease Lydia Wright • Nanci Stauffer • Cyrus Samai Matthew Oster
•
Received: 1 October 2013 / Accepted: 28 January 2014 Ó Springer Science+Business Media New York 2014
Abstract Although fetal echocardiography is a useful tool for earlier detection of congenital heart disease (CHD), no clear, evidence-based appropriate use criteria exist. This study aimed to determine those referral indications most predictive of CHD. A retrospective cohort study of fetal echocardiograms was conducted at a single tertiary care center. Binomial proportion tests were used to compare likelihoods of structural CHD detection based on primary indication with the general population prevalence of 0.8 %. Fetal echocardiograms of 2,380 pregnancies were performed, with CHD detected in 29.3 % of these pregnancies. Patients referred for suspected CHD on screening ultrasound had the highest prevalence of CHD. Of the women without a history of cardiac anomalies shown on screening ultrasound, those with extracardiac anomalies or known chromosomal abnormalities had the highest prevalence of CHD. Screening ultrasound currently is the most useful method for determining who should be referred for fetal echocardiography, with a high prevalence of CHD found. For women with normal cardiac screening ultrasound results, fetal factors such as extracardiac anomalies
L. Wright (&) N. Stauffer C. Samai M. Oster Emory University School of Medicine, Atlanta, GA, USA e-mail:
[email protected] M. Oster e-mail:
[email protected] L. Wright N. Stauffer C. Samai M. Oster Division of Pediatric Cardiology, Children’s Healthcare of Atlanta, 1405 Clifton Road NE, Atlanta, GA 30322, USA L. Wright Pediatrics Residency, University of Washington, Seattle, WA, USA
and chromosomal abnormalities are more predictive of CHD than maternal or familial factors. Keywords Prenatal diagnosis Fetal echocardiography Congenital heart disease Referral indications
Introduction Congenital heart disease (CHD) is the most common congenital abnormality, with a reported prevalence of approximately 8 per 1,000 live births [19]. Many of these cardiac defects have a substantial impact on morbidity and mortality, especially in the perinatal period. Between 20 and 32 % of perinatal deaths can be attributed to CHD [1]. Because of this sizeable disease burden, early detection of possible cardiac anomalies is an important component of prenatal care, especially in at-risk pregnancies. Fetal echocardiography is a useful tool for the early detection of CHD, especially when used for properly selected patients. Several studies have reported sensitivity for detecting CHD to be as high as 85 % [13, 23]. The current guidelines from the American Institute of Ultrasound in Medicine recommend performing fetal echocardiograms for pregnancies with fetal, maternal, or familial indications, although they do not provide an exhaustive list of those indications or evidence for their inclusion [3]. Common referral indications include abnormal cardiac screening ultrasound, extracardiac anomalies shown on screening ultrasound, maternal diabetes, and family history of CHD [22]. Despite its widespread use, no clear, evidence-based, appropriate use criteria for fetal echocardiography exist. This study aimed to determine those referral indications most predictive of CHD, which in turn might inform appropriate use criteria.
123
Pediatr Cardiol Table 1 Prevalence of congenital heart disease (CHD) found for primary referral indication compared with a population prevalence of 0.8 %
Indication
CHD detected
Total with given indication
CHD prevalence (%)
95 % CIa (%)
p valueb
High CHD prevalence (20–100 %) Suspected CHD on screening ultrasound
554
925
59.9
56.7–63.1
\0.01
Known chromosomal abnormality
14
33
42.4
25.6–59.3
\0.01
Incomplete screening exam
8
24
33.3
14.5–52.2
\0.01
Twin-to-twin transfusion syndrome
4
14
28.6
4.9–52.2
\0.01
42
198
21.2
15.5–26.9
\0.01
Extracardiac anomalies on screening ultrasound
Moderate CHD prevalence (10–20 %) Abnormal cardiac axis
3
18
16.7
3.6–41.4
\0.01
Pericardial effusion on ultrasound
4
31
12.9
1.1–24.7
\0.01
Echogenic focus on ultrasound
9
78
11.5
4.5–18.6
0.01
Teratogen exposure
5
45
11.1
1.9–20.3
\0.01
Increased nuchal translucency
2
18
11.1
1.4–34.7
\0.01
Low CHD prevalence (0–10 %) 21
305
6.9
4.0–9.7
\0.01
4
83
4.8
1.3–11.9
\0.01
a
Increased trisomy risk by serum screen
b
Advanced maternal age Fetal arrhythmia
4 8
83 186
4.8 4.3
1.3–11.9 1.4–7.2
\0.01 \0.01
11
285
3.9
1.6–6.1
\0.01
0
19
0.0
–
698
2,380
29.3
Family history of CHD CI confidence interval Using the binomial proportions test
Compared with the CHD rate of 0.8 % in the general population
c
Includes rare indications with fewer than 10 patients each (35 patients, total of 5 with CHD)
Maternal diabetes Maternal connective tissue disorder Totalc
Materials and Methods We conducted a retrospective cohort study of all fetal echocardiograms performed at one pediatric cardiology practice from August 2006 until December 2011. The data collected included maternal age, gestational age, up to three referral indications for echocardiography as documented by the referring obstetrician, anticipated delivery hospital, and estimated date of delivery. Patients were excluded from the study if they had received a prior fetal echocardiogram by an outside pediatric cardiologist (because these patients were referred for surgical planning, not diagnosis) or if a viable fetus was not found at examination. All other patients were included in the study. Prenatal cardiac diagnoses were recorded for each fetus. In cases of multiple echocardiograms during one pregnancy, the diagnosis from the final echocardiogram was used. The primary outcome of interest was the suspected presence or absence of structural CHD on the prenatal
123
–
echocardiogram. The final prenatal diagnoses of isolated arrhythmias, cardiomyopathies, twin-to-twin transfusion syndrome, and premature ductal constriction without associated structural defects were categorized as ‘‘structurally normal heart.’’ Similarly, variants of normal cardiac structures with no functional significance (e.g., dextrocardia, echogenic foci, and bilateral superior vena cava) were categorized as ‘‘structurally normal heart.’’ Binomial proportion tests were used to compare the proportion of CHD found for each primary indication with the reported population birth prevalence of 0.8 % (based on data from metropolitan Atlanta, where the study practice is based) [19]. Exact 95 % confidence intervals were calculated for each indication. Chi square analyses (or Fisher’s exact test, where appropriate) and relative risk calculations were used to analyze the effects of additional referral indications on the likelihood of finding CHD. For all statistical analyses, SAS version 9.3 (SAS, Cary, NC, USA) was used.
Pediatr Cardiol Suspected CHD on screening ultrasound
59.9%
Family history of CHD
6.9%
Maternal diabetes
3.9%
Extracardiac anomalies on screening ultrasound
21.2%
Fetal arrhythmia
4.3%
Advanced maternal age
4.8%
Increased trisomy risk by serum screen
4.8%
Echogenic focus
11.5%
Teratogen exposure
CHD Detected No CHD Detected
11.1%
Known chromosomal abnormality
42.4%
Pericardial effusion
12.9%
Incomplete screening exam
33.3%
Maternal connective tissue disorder
0%
Increased nuchal translucency
11.1%
Abnormal cardiac axis
16.7%
Twin-to-twin transfusion syndrome
28.6%
0
200
400
600
800
1000
Total Patients Fig. 1 Primary referral indications for fetal echocardiography and the rate of congenital heart disease detected
Results During the study period, 3,354 fetal echocardiograms were performed for 2,380 pregnancies. Among all pregnancies, CHD was confirmed or probable in 698 cases (29.3 %). A small but statistically significant difference in the gestational age of the fetus at the time of the first echocardiogram was found between patients with a diagnosis of CHD and those with no CHD diagnosis (26.9 vs 27.8 weeks, p \ 0.01). Maternal age was similar in the two groups (30 years for mothers of patients with a CHD diagnosis vs 30.1 years for mothers of patients with no diagnosis of CHD). The primary referral indications included fetal, maternal, and familial factors (Fig. 1; Table 1). A total of 925 patients (38.2 % of all referred patients) were referred due to suspected CHD on screening ultrasound. Other common reasons for referral were a family history of CHD, maternal diabetes, and extracardiac anomalies seen on screening ultrasound (Fig. 1). All indications other than maternal connective tissue disease were associated with a prevalence of CHD significantly higher than the 0.8 % seen in the general population (p \ 0.01, Table 1). However, the prevalence of CHD differed in our cohort depending on the primary referral indication. Patients referred for suspected CHD on screening ultrasound had a much higher percentage of CHD detected on echocardiogram than those referred for other indications (60 % for those with suspected CHD on screening ultrasound versus
9.8 % for those referred for other indications). A high prevalence of CHD ([20 %) was found among patients referred for suspected CHD on screening ultrasound, a known chromosomal abnormality in the fetus, an incomplete screening exam, known twin-to-twin transfusion syndrome, or extracardiac anomalies observed on screening ultrasound (Table 1). Of the 2,380 patients referred, 717 had either two or three referral indications listed, whereas the remaining 1,663 patients had only one reason for referral. The prevalence of CHD increased significantly for each additional indication (26.3 % for one, 34.12 % for two, and 53.6 % for three indications; p \ 0.01).
Discussion In this study, abnormalities discovered during routine prenatal screening conferred the greatest risk of finding CHD on subsequent fetal echocardiographies. Screening cardiac ultrasound currently is the most useful method for determining whom to refer, with a high prevalence of CHD found. In our population, 60 % of those patients referred after abnormal cardiac imaging on screening ultrasound were found to have CHD, a much higher proportion than found in any of the other indication categories. Previous studies also have found the highest percentages of CHD in patients with suspected CHD on screening ultrasound, with the percentages ranging widely from 10 to 80 % [6, 17, 20, 21].
123
Pediatr Cardiol
In our study, abnormal cardiac screening ultrasound was by far the most common indication for referral to fetal echocardiography. It also was the most likely predictor of CHD. This pattern of referrals, with many patients referred after abnormal cardiac imaging, has been consistently reported in recent series [20, 21]. This contrasts with earlier studies, in which family history of CHD was the most common referral indication, with few patients referred for abnormal cardiac screening ultrasound [4, 7]. The changing pattern of referrals during the last 20 years coincides with technological advances that have made imaging of the fetal heart on routine ultrasound much more accurate and sensitive [23]. It also has resulted in greater detection of CHD on fetal echocardiography at the tertiary level (29 % with CHD in our study vs 5 % in a 1995 study) [7]. Continued improvement in screening ultrasound technology and operator ability may be important in reducing unnecessary fetal echocardiograms while minimizing missed CHD diagnoses. A high prevalence of CHD also was found among fetuses with known chromosomal abnormalities or extracardiac anomalies, as well as among those with incomplete screening exams. Congenital heart disease was detected in nearly half of fetuses with chromosomal abnormalities and normal cardiac screening exams, confirming the importance of fetal echocardiography for this population. Those patients with chromosomal abnormalities in addition to other referral indications had an even higher percentage of CHD (79 %). Of the patients referred because of extracardiac anomalies, 21 % had CHD, consistent with the percentage found by other investigators in a smaller study in China [10]. A variety of different fetal anomalies were present in our referred patients, but by far the most common encountered was a single umbilical artery (nearly 25 % of all the referrals were made for this indication category). Findings have shown this fetal anomaly to be associated with renal and cardiac anomalies, often prompting referral for fetal echocardiography [9]. Some authors, however, argue that an isolated single umbilical artery may not increase the risk for CHD sufficiently to warrant specialized cardiac imaging for all patients [15, 18]. Our population was heterogeneous, including patients with both multiple and isolated anomalies. However, the high prevalence of CHD found ([20 % of those referred for all extracardiac anomalies) supports the appropriateness of referring these patients, including those with a single umbilical artery, for fetal echocardiography. Although screening ultrasound is an excellent tool for identifying at-risk pregnancies through the identification of both cardiac and extracardiac anomalies, it usually is performed in the second trimester. Current guidelines from the American College of Obstetrics and Gynecology recommend that women without specific indications for earlier
123
studies should receive a detailed anatomic ultrasound at 18–20 weeks gestation [2]. In our population, suspicion of a cardiac abnormality on routine prenatal ultrasound prompts referral to a maternalfetal medicine specialist. If CHD continues to be a concern, the patient is referred for fetal echocardiography by pediatric cardiologists and for possible consultation with cardiac surgeons. This process likely has reduced some unnecessary fetal echocardiograms, but the time delay may have led to a high average gestational age at the first echocardiogram (*27 weeks in our cohort). Gestational age at diagnosis is an important consideration because options available to parents narrow as the pregnancy progresses. It thus is important to look at other high-yield indications that may allow an earlier diagnosis. The use of increased nuchal translucency measurements as a potential referral indication shows promise. Prior studies have shown an association between increased nuchal translucency measurements and CHD, even in chromosomally normal patients [12]. After the institution of a national screening program in the Netherlands, Clur et al. [6] found an increase in referrals for fetal echocardiography, with the most common indication being increased nuchal translucency. Among those patients primarily referred for this indication, 17 % had some sort of cardiac pathology. The median gestational age at diagnosis among those referred for increased nuchal translucency was 14, 9 weeks earlier than the median for those referred due to suspected cardiac anomaly on screening ultrasound. An American study showed an increasing incidence of major CHD with increasing nuchal translucency and recommended using the measure as a referral indication for fetal echocardiography [21]. In our cohort, only 18 patients were referred with a primary indication of increased nuchal translucency, for a CHD prevalence of 11 %, suggesting that this method of screening for cardiac anomalies had not been widely adopted among obstetricians in our area during the study period. An additional 12 patients were noted to have increased nuchal translucency as additional indications, with 67 % of the patients having CHD. Nuchal translucency measurements typically are performed at 13–14 weeks of gestational age, versus the 18–20 weeks recommended for screening ultrasound. Several studies have shown the feasibility and accuracy of fetal echocardiograms performed at less than 17 weeks [5, 11, 14]. A greater focus on nuchal translucency as a screening tool for CHD, with referral for fetal echocardiography when increased, may help to provide accurate diagnoses and appropriate counseling earlier in an affected pregnancy. The role of maternal and familial factors as referral indicators continues to be unclear, although comparison of primary referral indications with additional referral
Pediatr Cardiol
indications can provide some insight. When used as primary referral indications in our study, these risk factors were less predictive of CHD than most fetal indications, with CHD prevalence estimates lower than 10 %. However, patients referred with maternal diabetes in addition to another indication (often suspected CHD on screening ultrasound) were significantly more likely to have CHD detected than patients without that indication. This points to a wide variation of referral practiced by area obstetricians. Some refer all patients with diabetes (and other maternal risk factors) for fetal echocardiography to our group, whereas others refer only after observing suspicious findings on ultrasound. Because we did not have a control group of pregnancies with no risk factors, we could not determine the absolute risk conferred by the aforementioned risk factors. A previous study using the indications of ‘‘self-referral’’ or ‘‘maternal anxiety’’ for control groups did not find a significantly higher prevalence of CHD in those referred for maternal or familial risk factors compared with the control subjects [17]. In one review of published data, Davey et al. [8] found a low absolute risk conferred by gestational diabetes and advanced maternal age. In a cost-effectiveness analysis among patients with diabetes, Odibo et al. [16] recommended fetal echocardiography only after abnormal screening ultrasound. Our data support the aforementioned recommendations, showing a higher prevalence of CHD only in conjunction with other indications. Further larger studies are needed to determine in what situations these risk factors merit consultation with a pediatric cardiologist, and in what situations they can be managed by high-risk obstetricians. The major limitation of this study was that it was a retrospective, single-institution study. Consequently, the results may not be widely generalizable. An additional limitation was the lack of a control other than the baseline population prevalence of CHD. Comparison with women who have no risk factors would be helpful in determining the utility of imaging lower-risk women. We also did not examine the risk for nonstructural cardiac pathology, such as cardiomyopathy and arrhythmias, in patients referred for fetal echocardiography (those findings were categorized as denoting structurally normal hearts in our cohort). Although this makes the risk of structural heart disease more clear, recommendations as to who should be offered fetal echocardiography may be affected by these findings, especially among patients with diabetes, who face a higher risk of having a child with hypertrophic cardiomyopathy.
Conclusion Screening ultrasound currently is the most effective tool for identifying at-risk fetuses that would benefit from fetal
echocardiography. Screening ultrasound allows for the identification of suspected CHD through examinations of the fetal heart and identification of extracardiac anomalies that could be associated with CHD that has been missed. The role of maternal and familial risk factors remains unclear. Further studies are necessary to determine definitively which women with these risk factors, if any, should be referred to a pediatric cardiologist for fetal echocardiography.
References 1. Achiron R, Glaser J, Gelernter I et al (1992) Extended fetal echocardiographic examination for detecting cardiac malformations in low risk pregnancies. BMJ 304:671–674 2. American College of Obstetricians and Gynecologists (ACOG) (2009) Ultrasonography in pregnancy. (ACOG); February, (ACOG practice bulletin; no. 101) American College of Obstetricians and Gynecologists, Washington DC 3. American Institute for Ultrasound in Medicine (AIUM) (2010) AIUM practice guideline for the performance of fetal echocardiography. American Institute for Ultrasound in Medicine, Laurel 4. Callan NA, Maggio M, Steger S et al (1991) Fetal echocardiography: indications for referral, prenatal diagnoses, and outcomes. Am J Perinatol 8:390–394 5. Carvalho JS, Moscoso G, Tekay A, Campbell S, Thilaganathan B, Shinebourne EA (2004) Clinical impact of first and early second trimester fetal echocardiography on high-risk pregnancies. Heart 90:921–926 6. Clur SA, Van Brussel PM, Mathijssen IB et al (2011) Audit of 10 years of referrals for fetal echocardiography. Prenat Diagn Dec 31:1134–1140. doi:10.1002/pd.2847 7. Cooper MJ, Enderlein MA, Dyson DC et al (1995) Fetal echocardiography: retrospective review of clinical experience and an evaluation of indications. Obstet Gynecol 86(4 Pt 1):577–582 8. Davey BT, Seubert DE, Phoon CK (2009) Indications for fetal echocardiography high referral, low yield? Obstet Gynecol Surv 64:405–415 9. Hua M, Odibo AO, Macones GA, et al (2010) Single umbilical artery and its associated findings. Obstet Gynecol May 115:930–934 10. Li M, Wang W, Yang X et al (2008) Evaluation of referral indications for fetal echocardiography in Beijing. J Ultrasound Med 27:1291–1296 11. Lombardi CM, Bellotti M, Fesslova V, Cappellini A (2007) Fetal echocardiography at the time of the nuchal translucency scan. Ultrasound Obstet Gynecol 29:249–257 12. McAuliffe FM, Hornberger LK, Winsor S, Chitayat D, Chong K, Johnson JA (2004) Fetal cardiac defects and increased nuchal translucency thickness: a prospective study. Am J Obstet Gynecol 191:1486–1490 13. Meyer-Wittkopf M, Cooper S, Sholler G (2001) Correlation between fetal cardiac diagnosis by obstetric and pediatric cardiologist sonographers and comparison with postnatal findings. Ultrasound Obstet Gynecol 17:391–397 14. Moon-Grady A, Shahanavaz S, Brook M, Rodriguez H, Hornberger LK (2012) Can a complete fetal echocardiogram be performed at 12 to 16 weeks’ gestation? J Am Soc Echocardiogr 25:1342–1352 15. Murphy-Kaulbeck L, Dodds L, Joseph KS et al (2010) Single umbilical artery risk factors and pregnancy outcomes. Obstet Gynecol 116:843–850
123
Pediatr Cardiol 16. Odibo AO, Coassolo KM, Stamilio DM et al (2006) Should all pregnant diabetic women undergo a fetal echocardiography? A cost-effectiveness analysis comparing four screening strategies. Prenat Diagn 26:39–44 17. Perri T, Cohen-Sacher B, Hod M et al (2005) Risk factors for cardiac malformations detected by fetal echocardiography in a tertiary center. J Matern Fetal Neonatal Med 17:123–128 18. Prefumo F, Gu¨ven MA, Carvalho JS (2010) Single umbilical artery and congenital heart disease in selected and unselected populations. Ultrasound Obstet Gynecol 35:552–555 19. Reller MD, Strickland MJ, Riehle-Colarusso T et al (2008) Prevalence of congenital heart defects in metropolitan Atlanta, 1998–2005. J Pediatr 153(6):807–813
123
20. Sharony R, Fejgin MD, Giron-Shental T et al (2009) Who should be offered fetal echocardiography? One center’s experience with 3,965 cases. IMAJ 11:542–545 21. Simpson LL (2004) Indications for fetal echocardiography from a tertiary-care obstetric sonography practice. J Clin Ultrasound 32:123–128 22. Small M, Copel JA (2004) Indications for fetal echocardiography. Pediatr Cardiol 25:210–222 23. Tegnander E, Eik-Nes SH (2006) The examiner’s ultrasound experience has a significant impact on the detection rate of congenital heart defects at the second-trimester fetal examination. Ultrasound Obstet Gynecol 28:8–14