Research
ajog.org
OBSTETRICS
Risk of critical congenital heart defects by nuchal translucency norms Laura L. Jelliffe-Pawlowski, PhD; Mary E. Norton, MD; Gary M. Shaw, DrPH; Rebecca J. Baer, MPH; Monica C. Flessel, PhD; Sara Goldman, MPH; Robert J. Currier, PhD OBJECTIVE: The purpose of this study was to compare the perfor-
mance of first-trimester nuchal translucency (NT) cutoff of 3.5 mm with NT percentiles that were calculated for crown-rump length to identify fetuses with critical congenital heart defects (CCHDs). STUDY DESIGN: This was a population-level study of singleton preg-
nancies in California with NT measurements performed between 11 and 14 weeks of gestation. Eligible cases were those that resulted in live births from 2009-2010 and had information about the presence or absence of CCHDs available in the hospital discharge records through age 1 year (n ¼ 76,089). Logistic binomial regression methods were used to compare the rate of CCHDs by an NT percentile for crownrump length and millimeter cutpoints. RESULTS: Compared with fetuses with an NT measurement of 5 times as
likely to have a CCHD (1.3% vs 0.2%; relative risk, 5.66; 95% confidence interval, 3.19e10.04) and fetuses with an NT measurement 3.5 mm were >12 times as likely to have a CCHD (2.8% vs 0.2%; relative risk, 12.28; 95% confidence interval, 5.11e29.51). NT 99th percentile had a sensitivity of 5.8% and a specificity of 98.9% for the detection of CCHDs compared with 2.6% and 99.8% for NT 3.5 mm. CONCLUSION: Results show that NT measurements of 99th percentile and 3.5 mm are not equivalent and that substantial risk for CCHD extends to the less restrictive 99th percentile cutpoint. Data suggest that the use of this cutpoint compared with the current standard could double the number of CCHDs that are identified based on NT risk.
Key words: critical congenital heart defect, nuchal translucency, prenatal screening
Cite this article as: Jelliffe-Pawlowski LL, Norton ME, Shaw GM, et al. Risk of critical congenital heart defects by nuchal translucency norms. Am J Obstet Gynecol 2015;212:518.e1-10.
C
ongenital heart defects affect approximately 1% of newborn infants.1-3 Critical congenital heart defects (CCHDs), which are the most serious of these, typically require surgical repair within the first year of life to avoid death or permanent morbidity. Prenatal detection is therefore particularly important for these newborn infants.4-8 Over the past decade, first-trimester nuchal translucency (NT) measurement has been used increasingly for prenatal detection of chromosomal abnormalities and has been associated with CCHDs in pregnancies with and without aneuploidy.9-12 Although an
enlarged NT has been associated with CCHDs (defined by either an NT measurement percentile for crownrump-length [CRL] or by a specific NT measurement in millimeters),9-11,13 having an NT measurement of 3.5 mm is the most commonly applied cutoff at which fetal echocardiography is recommended, based on guidelines of the American College of Obstetricians and Gynecologists.14,15 Here we evaluate how well an NT cutoff of 3.5 mm performs in comparison with NT percentile by CRL for identifying pregnancies carrying fetuses with CCHDs in a large, diverse population-based sample of live-born
From the Genetic Disease Screening Program, California Department of Public Health, Richmond (Drs Jelliffe-Pawlowski, Flessel, and Currier, Ms Baer, and Ms Goldman), and the Departments of Epidemiology and Biostatistics (Dr Jelliffe-Pawlowski) and Obstetrics and Gynecology (Dr Norton), University of California, San Francisco, School of Medicine, San Francisco, and Department of Pediatrics, Stanford University School of Medicine, Stanford (Dr Shaw), CA. Received June 20, 2014; revised Sept. 11, 2014; accepted Oct. 28, 2014. The authors report no conflict of interest. Corresponding author: Laura L. Jelliffe-Pawlowski, PhD. [email protected] 0002-9378/$36.00 ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2014.10.1102
518.e1 American Journal of Obstetrics & Gynecology APRIL 2015
singleton pregnancies (n ¼ 76,089) with complete CCHD ascertainment.
M ATERIALS
AND
M ETHODS
All pregnancies included in the study were participants in the California Prenatal Screening Program that was administered by the Genetic Disease Screening Program within the California Department of Public Health. Included pregnancies had CRLs from 45.0-84.0 mm and had NT measurements performed as part of routine screening for chromosomal defects. Study pregnancies were drawn from a larger sample of all singleton pregnancies with NT results and estimated dates of delivery between July 2009 and December 2010 based on CRL (n ¼ 163,899). The study was restricted to pregnancies with linked newborn infant screening records (n ¼ 127,259) and linked birth certificate and hospital discharge records (n ¼ 117,717). We further restricted the study sample to pregnancies for which NT measurements were done by clinicians who
Obstetrics
ajog.org had practitioner-specific NT medians (n ¼ 76,286) and to pregnancies without chromosomal defects (n ¼ 76,089; Figure 1). Details regarding how practitioner-specific medians are calculated as part of the California Prenatal Screening Program are described elsewhere.16 All California NT practitioners are credentialed by the Nuchal Translucency Quality Review Program17 or the Fetal Medicine Foundation.18 Clinicians who had practitioner-specific medians are those who had a minimum of 75 examinations for whom the slope of their NT measurements increased by at least 11% across gestational weeks. In California, practitionerspecific medians have been shown to help control for less experienced practitioners who were tending towards smaller NT measurements.16,19 Data that were used to exclude newborn infants from the analyses based on the presence of a chromosomal
abnormality were obtained by a search of the Genetic Disease Screening Program Chromosome Registry. The registry collects information on chromosomal abnormalities from multiple ascertainment sources (physicians, laboratories, hospitals, and prenatal diagnostic centers) on all births in the state. All sources are mandated by California law to report diagnosed chromosomal abnormalities (whether diagnosed by karyotype or microarray) to the program.20-24 Information on CCHDs was collected by the linkage of screening records with all hospital discharge records through 1 year of age for each study infant (wherein records were linked with the use of multiple identifiers [baby and mother date of birth; baby first and last name; mother first, last, and maiden name, address, phone number, and hospital of baby birth]). These records include all inpatient discharge information that is submitted to the state each
FIGURE 1
Singleton pregnancies with first trimester nuchal translucency
The flow chart delineates the measurement of singleton pregnancies with first-trimester nuchal translucency that were included in our study. NT, nuchal translucency. Jelliffe-Pawlowski. Risk of CCHDs by nuchal translucency norms. Am J Obstet Gynecol 2015.
Research
time a patient is treated in a licensed general acute care hospital in California and includes all diagnoses present at the time of discharge based on 4- or 5-digit codes from the Ninth Revision of the International Classification of Diseases.25 For this study, CCHDs were considered to be those associated with substantive morbidity and mortality rates that included “common truncus,” “transposition of great vessels,” “Tetralogy of Fallot,” “common ventricle,” “endocardial cushion defects,” “anomalies of pulmonary valve,” “tricuspid atresia and stenosis,” “Ebstein’s anomaly,” “congenital stenosis of aortic valve,” “hypoplastic left heart syndrome,” “coarctation of the aorta,” and “anomalies of the great veins.” Hospital discharge records were considered the reference standard (gold standard) for the presence of a CCHD in our analyses. All NT measurements are submitted to the California Prenatal Screening Program as part of routine first- and secondtrimester screening for aneuploidies. Screening requires submission of a single valid NT measurement that then generates an interpretation for aneuploidy risk when accompanied by first- and/or second-trimester serum analyte results. All requisitions for screening include information on maternal race/ethnicity, weight, smoking status, and self-reported presence of pregestational diabetes mellitus. We used the prenatal screening records that were associated with the NT measurement to identify these sociodemographic factors. To compare performance of an NT cutoff of 3.5 mm to NT percentile by CRL for the identification of fetuses with CCHDs, we first categorized each pregnancy as having an NT measurement of 3.5 mm and/or an NT measurement
ajog.org
OBSTETRICS
Risk of critical congenital heart defects by nuchal translucency norms Laura L. Jelliffe-Pawlowski, PhD; Mary E. Norton, MD; Gary M. Shaw, DrPH; Rebecca J. Baer, MPH; Monica C. Flessel, PhD; Sara Goldman, MPH; Robert J. Currier, PhD OBJECTIVE: The purpose of this study was to compare the perfor-
mance of first-trimester nuchal translucency (NT) cutoff of 3.5 mm with NT percentiles that were calculated for crown-rump length to identify fetuses with critical congenital heart defects (CCHDs). STUDY DESIGN: This was a population-level study of singleton preg-
nancies in California with NT measurements performed between 11 and 14 weeks of gestation. Eligible cases were those that resulted in live births from 2009-2010 and had information about the presence or absence of CCHDs available in the hospital discharge records through age 1 year (n ¼ 76,089). Logistic binomial regression methods were used to compare the rate of CCHDs by an NT percentile for crownrump length and millimeter cutpoints. RESULTS: Compared with fetuses with an NT measurement of 5 times as
likely to have a CCHD (1.3% vs 0.2%; relative risk, 5.66; 95% confidence interval, 3.19e10.04) and fetuses with an NT measurement 3.5 mm were >12 times as likely to have a CCHD (2.8% vs 0.2%; relative risk, 12.28; 95% confidence interval, 5.11e29.51). NT 99th percentile had a sensitivity of 5.8% and a specificity of 98.9% for the detection of CCHDs compared with 2.6% and 99.8% for NT 3.5 mm. CONCLUSION: Results show that NT measurements of 99th percentile and 3.5 mm are not equivalent and that substantial risk for CCHD extends to the less restrictive 99th percentile cutpoint. Data suggest that the use of this cutpoint compared with the current standard could double the number of CCHDs that are identified based on NT risk.
Key words: critical congenital heart defect, nuchal translucency, prenatal screening
Cite this article as: Jelliffe-Pawlowski LL, Norton ME, Shaw GM, et al. Risk of critical congenital heart defects by nuchal translucency norms. Am J Obstet Gynecol 2015;212:518.e1-10.
C
ongenital heart defects affect approximately 1% of newborn infants.1-3 Critical congenital heart defects (CCHDs), which are the most serious of these, typically require surgical repair within the first year of life to avoid death or permanent morbidity. Prenatal detection is therefore particularly important for these newborn infants.4-8 Over the past decade, first-trimester nuchal translucency (NT) measurement has been used increasingly for prenatal detection of chromosomal abnormalities and has been associated with CCHDs in pregnancies with and without aneuploidy.9-12 Although an
enlarged NT has been associated with CCHDs (defined by either an NT measurement percentile for crownrump-length [CRL] or by a specific NT measurement in millimeters),9-11,13 having an NT measurement of 3.5 mm is the most commonly applied cutoff at which fetal echocardiography is recommended, based on guidelines of the American College of Obstetricians and Gynecologists.14,15 Here we evaluate how well an NT cutoff of 3.5 mm performs in comparison with NT percentile by CRL for identifying pregnancies carrying fetuses with CCHDs in a large, diverse population-based sample of live-born
From the Genetic Disease Screening Program, California Department of Public Health, Richmond (Drs Jelliffe-Pawlowski, Flessel, and Currier, Ms Baer, and Ms Goldman), and the Departments of Epidemiology and Biostatistics (Dr Jelliffe-Pawlowski) and Obstetrics and Gynecology (Dr Norton), University of California, San Francisco, School of Medicine, San Francisco, and Department of Pediatrics, Stanford University School of Medicine, Stanford (Dr Shaw), CA. Received June 20, 2014; revised Sept. 11, 2014; accepted Oct. 28, 2014. The authors report no conflict of interest. Corresponding author: Laura L. Jelliffe-Pawlowski, PhD. [email protected] 0002-9378/$36.00 ª 2015 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2014.10.1102
518.e1 American Journal of Obstetrics & Gynecology APRIL 2015
singleton pregnancies (n ¼ 76,089) with complete CCHD ascertainment.
M ATERIALS
AND
M ETHODS
All pregnancies included in the study were participants in the California Prenatal Screening Program that was administered by the Genetic Disease Screening Program within the California Department of Public Health. Included pregnancies had CRLs from 45.0-84.0 mm and had NT measurements performed as part of routine screening for chromosomal defects. Study pregnancies were drawn from a larger sample of all singleton pregnancies with NT results and estimated dates of delivery between July 2009 and December 2010 based on CRL (n ¼ 163,899). The study was restricted to pregnancies with linked newborn infant screening records (n ¼ 127,259) and linked birth certificate and hospital discharge records (n ¼ 117,717). We further restricted the study sample to pregnancies for which NT measurements were done by clinicians who
Obstetrics
ajog.org had practitioner-specific NT medians (n ¼ 76,286) and to pregnancies without chromosomal defects (n ¼ 76,089; Figure 1). Details regarding how practitioner-specific medians are calculated as part of the California Prenatal Screening Program are described elsewhere.16 All California NT practitioners are credentialed by the Nuchal Translucency Quality Review Program17 or the Fetal Medicine Foundation.18 Clinicians who had practitioner-specific medians are those who had a minimum of 75 examinations for whom the slope of their NT measurements increased by at least 11% across gestational weeks. In California, practitionerspecific medians have been shown to help control for less experienced practitioners who were tending towards smaller NT measurements.16,19 Data that were used to exclude newborn infants from the analyses based on the presence of a chromosomal
abnormality were obtained by a search of the Genetic Disease Screening Program Chromosome Registry. The registry collects information on chromosomal abnormalities from multiple ascertainment sources (physicians, laboratories, hospitals, and prenatal diagnostic centers) on all births in the state. All sources are mandated by California law to report diagnosed chromosomal abnormalities (whether diagnosed by karyotype or microarray) to the program.20-24 Information on CCHDs was collected by the linkage of screening records with all hospital discharge records through 1 year of age for each study infant (wherein records were linked with the use of multiple identifiers [baby and mother date of birth; baby first and last name; mother first, last, and maiden name, address, phone number, and hospital of baby birth]). These records include all inpatient discharge information that is submitted to the state each
FIGURE 1
Singleton pregnancies with first trimester nuchal translucency
The flow chart delineates the measurement of singleton pregnancies with first-trimester nuchal translucency that were included in our study. NT, nuchal translucency. Jelliffe-Pawlowski. Risk of CCHDs by nuchal translucency norms. Am J Obstet Gynecol 2015.
Research
time a patient is treated in a licensed general acute care hospital in California and includes all diagnoses present at the time of discharge based on 4- or 5-digit codes from the Ninth Revision of the International Classification of Diseases.25 For this study, CCHDs were considered to be those associated with substantive morbidity and mortality rates that included “common truncus,” “transposition of great vessels,” “Tetralogy of Fallot,” “common ventricle,” “endocardial cushion defects,” “anomalies of pulmonary valve,” “tricuspid atresia and stenosis,” “Ebstein’s anomaly,” “congenital stenosis of aortic valve,” “hypoplastic left heart syndrome,” “coarctation of the aorta,” and “anomalies of the great veins.” Hospital discharge records were considered the reference standard (gold standard) for the presence of a CCHD in our analyses. All NT measurements are submitted to the California Prenatal Screening Program as part of routine first- and secondtrimester screening for aneuploidies. Screening requires submission of a single valid NT measurement that then generates an interpretation for aneuploidy risk when accompanied by first- and/or second-trimester serum analyte results. All requisitions for screening include information on maternal race/ethnicity, weight, smoking status, and self-reported presence of pregestational diabetes mellitus. We used the prenatal screening records that were associated with the NT measurement to identify these sociodemographic factors. To compare performance of an NT cutoff of 3.5 mm to NT percentile by CRL for the identification of fetuses with CCHDs, we first categorized each pregnancy as having an NT measurement of 3.5 mm and/or an NT measurement
