- Email: [email protected]
Pregnancy interruption after second trimester diagnosis of fetal structural anomalies: The New Jersey Fetal Abnormalities Registry
American Journal of Obstetrics and Gynecology (2005) 193, 1492–7
www.ajog.org
Pregnancy interruption after second trimester diagnosis of fetal structural anomalies: The New Jersey Fetal Abnormalities Registry Eden R. Rauch, MD, MPH,a,b John C. Smulian, MD, MPH,a Kristin DePrince, MS, CGC,a Cande V. Ananth, PhD, MPH,a Stephen W. Marcella, MD, MPH,b for the New Jersey Fetal Abnormalities Registry Departments of Obstetrics, Gynecology and Reproductive Sciences,a and Environmental and Community Medicine,b University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School/Robert Wood Johnson University Hospital, New Brunswick, NJ Received for publication October 18, 2004; revised January 7, 2005; accepted February 17, 2005
KEY WORDS Fetal abnormality Anomaly Birth defect Ultrasound examination
Objective: The purpose of this study was to identify factors that predict a decision to interrupt a pregnancy in which there are fetal anomalies in the second trimester. Study design: The New Jersey Fetal Abnormalities Registry prospectively recruits and collects information on pregnancies (R15 weeks of gestation) from New Jersey residents in whom a fetal structural anomaly has been suspected by maternal-fetal medicine specialists. Enrolled pregnancies that have major fetal structural abnormalities identified from 15 to 23 weeks of gestation were included. Outcomes were classified as either elective interruption or a natural pregnancy course, which might include a spontaneous fetal death or live birth. Predictors of elective interruption of pregnancy were examined with univariable and multivariable logistic regression analyses. Results: Of the 97 cases, 33% of the women (n = 32) interrupted the pregnancy. Significant variables in the regression model that were associated with a decision to interrupt a pregnancy were earlier identification of fetal anomalies (19.0 G 2 weeks of gestation vs 20.5 G 2 weeks of gestation; P = .003), the presence of multiple anomalies (78% [25/32] vs 52% [33/63]; P = .01], and a presumption of lethality (56% [18/32] vs 14% [9/65]; P = .0001). These variables corresponded to an odds ratio for pregnancy interruption of 4.2 (95% CI, 1.0, 17.0) for multiple anomalies, 0.8 (95% CI, 0.7, 1.0) for each week of advancing gestational age, and 36.1 (95% CI, 2.9, 450.7) for presumed lethal abnormalities. Conclusion: Early diagnosis, the identification of multiple abnormalities, and an assessment of likely lethality of fetal anomalies are important factors for the optimization of parental autonomy in deciding pregnancy management. Ó 2005 Mosby, Inc. All rights reserved.
Supported in part by National Institutes of Health grant R01-HD038902 (C.V.A.; J.C.S); the New Jersey Fetal Abnormalities Registry (Principle Investigator, J.C.S.) was supported by a cooperative agreement between the University of Medicine and Dentistry of New Jersey- Robert Wood Johnson Medical School and the New Jersey Department of Health and Senior Services through the National Birth Defects Prevention Study of the Centers for Disease Control and Prevention. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2005 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2005.02.099
Rauch et al Recent advances in prenatal diagnosis have significantly improved the detection of birth defects before birth. As a result, couples often have the options of expectant care, elective interruption, or, in rare instances, fetal surgery. Several studies have noted a decline in the live birth prevalence of certain birth defects over the last several decades, most notably anencephaly, Down syndrome, spina bifida, and cardiovascular malformations.1-3 Technologic advances in ultrasonography and prenatal diagnosis aid in the diagnosis of fetal anomalies and provide information that can influence the decision to continue or interrupt a pregnancy. Schectman et al4 showed that the severity of structural anomalies directly correlated with abortion rates of anomalous fetuses. Furthermore, at similar degrees of severity, central nervous system anomalies were more likely to be terminated electively. Other studies have reported that the specific organ system that is anomalous, the severity of structural anomalies, and the number of defects in other organs are associated with decisions to interrupt a pregnancy.5,6 However, there is limited information that is available to help understand whether other factors influence parental choices in the management of pregnancies that are prenatally diagnosed with structural anomalies. The New Jersey Fetal Abnormalities Registry (NJFAR) was established in 1999 at the University of Medicine and Dentistry of New Jersey (UMDNJ)Robert Wood Johnson Medical School by a cooperative agreement with the New Jersey Department of Health and Senior Services and the Centers for Disease Control and Prevention through the Centers for Birth Defects Research and Prevention. NJFAR collects information on pregnancies (R15 weeks of gestation) of New Jersey residents in which a fetal structural anomaly has been identified or suspected by maternal-fetal medicine specialists across New Jersey. This study was designed to use NJFAR-registered pregnancies to identify factors that predict parental decisions to interrupt pregnancies with fetuses for whom structural anomalies are identified in the second trimester.
Material and methods This study was approved by the Institutional Review Boards of the New Jersey Department of Health and Senior Services, the UMDNJ-Robert Wood Johnson Medical School, and all other affiliated sites (Acknowledgments). Enrollment of NJFAR cases began in January 2000 from maternal-fetal medicine practices that are linked with the UMDNJ Health Systems. These sites included prenatal diagnostic practices that are affiliated with the Robert Wood Johnson Medical School’s New Brunswick, Neptune, and Camden campuses and the New Jersey Medical School in Newark. These practices
1493 include approximately 20 maternal-fetal medicine specialists and provide prenatal diagnostic services to an estimated population of 31,000 women each year. All sites provide access to a full range of prenatal diagnostic options, genetic counseling, physician counseling, and discussions about pregnancy management. Pregnancy interruption services were available in each community that surrounds participating NJFAR practices. NJFAR is not a comprehensive population-based registry that covers the entire state. However, because a large portion of prenatally diagnosed fetal anomalies in New Jersey are seen by maternal-fetal medicine specialists at some point during their course of care, the practices are useful sources for capturing these cases. Cases that are eligible for NJFAR enrollment include fetuses with significant structural abnormalities that are reportable to the New Jersey Department of Health and Senior Services Birth Defects Registry. Cases with aneuploidy were included only if they are associated with structural abnormalities. Additional non-UMDNJ cases that are obtained from the New Jersey Maternal-Fetal Medicine Society member practices who volunteered to participate in NJFAR were not included for this specific study. All NJFAR registered cases with an estimated date of delivery between January 2000 and February 2002 were eligible for this study. Eligible cases had pregnancy information that was recorded at NJFAR enrollment including intake date, primary physician’s name and phone number, patient’s first and last initials, maternal date of birth, city and zip code of residence, gestational age at diagnosis, estimated date of delivery, number of fetuses in pregnancy, gravidity, parity, race, obstetrics clinic/offices that were providing care, fetal abnormalities that were identified, diagnostic procedures, referral information, preference for genetic counseling, and pregnancy continuation plans. A subset of these NJFAR-enrolled patients who provided informed consent underwent an interview to obtain pregnancy outcomes and extensive health behavior, nutritional behavior, and occupational exposure information. When interview information was available, this information was used as a data source for this study. For the patients who did not participate in the interview or whose pregnancy outcomes were not known, a questionnaire was mailed to the subject’s obstetrician that requested pregnancy outcome information. Nonresponding physicians were contacted directly to encourage completion of questionnaires. Only patients with anomalies that were identified before 24 weeks of gestation were included for study, because interruption of pregnancy at R24 weeks is not readily available in New Jersey. On the pregnancy outcome questionnaire, the patients’ primary referring physicians were provided with maternal initials, date of birth, estimated date of delivery, date of ultrasound diagnosis, and a summary
1494 Table I
Rauch et al Maternal characteristics, based on the decision whether to interrupt pregnancy: NJFAR
Characteristic
Pregnancy interruption
Natural course pregnancy
P value
N Age (y)* Gestational age at diagnosis (wk)* Gravidity (n)y Parity (n)y White race (n/N) Assisted reproductive technology (n/N) Genetic counseling (n/N)
32 32.9 G 5.6 19.0 G 2 2 (1-5) 1 (0-4) 22/29 (76%) 2/22 (9%) 21/23 (91%)
65 31.9 G 6.5 20.5 G 2 2 (1-11) 1 (0-10) 39/57 (68%) 6/56 (11%) 38/52 (73%)
.45 .003 .75 .36 .47 .83 .08
* Data are given as mean G SD. y Data are given as median (range).
of the suspected fetal abnormalities by which they could identify their own patients. The primary outcome of interest was the parental decision to interrupt the pregnancy. New Jersey statutes allow the interruption of pregnancy before viability without imposing specific gestational age or weight criteria. Outcomes were classified as either elective interruption (group 1) or natural pregnancy course (group 2), which might include a spontaneous fetal death or live birth. The variables that were recorded included maternal and fetal ages at diagnosis, gravidity, parity, race, specific body/organ system that was affected, aneuploidy, the number and type of organ/body systems that were affected, the use of genetic counseling, the presumed lethality of the anomaly, and the use of assisted reproductive or conception techniques. For the purposes of this study, cases were classified as presumed lethal on the basis of the specific spectrum of prenatally identified abnormalities. This was assessed with the information that was provided by the diagnosing maternal-fetal medicine specialist at the time of diagnosis and was independently reassessed by 1 of the authors (J.C.S.). Lethality was presumed in the presence of an abnormality or spectrum of abnormalities that were associated with either a likely in utero death or, if live born, an inability to survive independently even with surgical correction of the abnormality. Continuous variables were compared with the use of parametric and nonparametric tests, where appropriate. Categorical variables were compared with the use of contingency tables (chi-square and Fisher’s exact test). The unadjusted relative risk and 95% confidence interval (CI) for specific abnormalities that were associated with pregnancy interruption were derived. Similar analyses were performed after an adjustment was made for the gestational age at diagnosis. Variables that reached statistical significance with a probability value of %.05 were also examined for clinical significance. The potentially important predictors were used in forward selection logistic regression, with an entry criteria of a
probability value of !.10 to develop a model for the prediction of pregnancy interruption. Interactions were retained in the model after the statistical criteria of a probability value of %.10 was met. The logistic regression model was used to derive odds ratios (ORs) with 95% CI. The Hosmer-Lemeshow goodness-of-fit test with interactions that were included in the model was calculated to ensure the adequate fit of the logistic regression model. We reported the significant interactions by stratified analysis. All data were analyzed with SAS software (SAS Institute, Cary, NC).
Results One hundred thirty-eight questionnaires were mailed to physician’s offices, of which 78% (n = 108 questionnaires) were returned. An additional 11 cases had incomplete reporting of outcomes and were excluded from the analysis. Therefore, the study group comprised 97 cases with a fetal structural anomaly that was diagnosed between 15 and 24 weeks of gestation and estimated dates of delivery between January 2000 and February 2002. Twenty-three cases (23%) had information that was available from the detailed NJFAR interviews, which did not differ from the questionnaire information. Of the 97 cases, 33% (n = 32 cases) had a pregnancy interruption (group 1). Demographic and pregnancy information are presented in Table I. Group 1 had earlier gestational ages for identification of fetal anomalies by 1.5 weeks. The findings for each of the 27 fetuses with abnormalities that were presumed to be lethal are described in Table II. As shown in Table III, group 1 fetuses were more likely to have multiple anomalies, anomalies that were presumed to be lethal, head (intracranial) anomalies, neural tube anomalies, and skeletal anomalies. These associations were stronger after an adjustment was made for gestational age. Maternal age, race, and parity were not significant in the univariable analyses and therefore were not included for the adjustment. Because
Rauch et al
1495
Table II Description of 37 fetuses having anomalies identified in the second trimester that were presumed to be lethal: NJFAR Anomaly
N
Isolated acrania Isolated anencephaly Isolated bilateral renal agenesis Nonimmune hydrops With cystic hygroma With Ebstein’s anomaly With severe cardiomegaly Severe arthrogryposis sequence With generalized skin edema, small stomach/bladder, bowed femurs With scalp edema, cardiomegaly, hepatomegaly, very small stomach, severe hydronephrosis Bilateral severe polycystic dysplastic kidneys, ventriculomegaly, oligohydramnios Limb body wall defect Severe ventriculomegaly, large ventricular septal defect with complex disorder of great vessels, suspected tracheoesophageal fistula, abnormal facial structures, 46, XX with partial monosomy for chromosome 1 Large atrioventricular septal defect, fisted hands, trisomy 18 Omphalocele, ventricular septal defect, trisomy 18 Sirenomelia Large atrioventricular canal defect, suspected esophageal atresia, extremity contractures Severe hydrocephalus, hypoplastic right heart, hydronephrosis, partial duplication of chromosome 1 Cystic hygroma, Dandy-Walker malformation, pleural effusion, atrioventricular canal defect, hypoplastic right ventricle
4 3 3 6 4 1 1 2 1 1 1 1 1
1 1 1 1 1 1
information about genetic counseling was missing from 21 cases (23%), we were not able to use that variable for the adjustment. These variables and the interaction between multiple anomalies and presumed lethality were evaluated by logistic regression. The model confirmed that the best 3 predictors of a decision to interrupt a pregnancy were the presence of multiple anomalies (OR, 4.2; 95% CI, 1.0, 17.0; P = .05), earlier gestational age at diagnosis (OR, 0.8; 95% CI, 0.7, 1.0; P = .04) for each week of gestation, and anomalies that were presumed to be lethal (OR, 36.1; 95% CI, 2.9, 450.7; P = .005). Because an important interaction existed between lethality and the number of anomalies (P = .10), further analysis was stratified by multiple and single anomaly subgroups. When the logistic regression model was adjusted for gestational age, the stratified odds ratio for pregnancy interruption was 3.64 (95% CI, 1.2-11.4) for multiple anomalies that were presumed to be lethal and 45.2 (95% CI, 2.8, 744.8) for a single anomaly that were considered to be lethal. The Hosmer-
Lemeshow goodness-of-fit test with the interaction that was included in the model was not statistically significant (P O .5), which indicated a good model fit.
Comment The ability of ultrasound examination to calculate gestational age, to detect multiple gestations, and to evaluate fetal well-being by diagnosing growth, fluid, and structural abnormalities has had a major impact on the practice of obstetrics. A number of investigators have described benefits of prenatal diagnosis of fetal anomalies (such as improvement of management at delivery, maintenance of respect for patient autonomy, and promotion of the psychologic advantages and parental bonding that might be disrupted at the birth of a child with an unusual unsuspected anomaly).7-9 Our study identified a number of important components of prenatal diagnosis that appear to affect decisions to interrupt affected pregnancies with structural abnormalities that include presumed lethality, the presence of multiple anomalies, and earlier fetal age at diagnosis. The severity of structural anomalies has been recognized by many investigators to influence the likelihood of pregnancy interruption.4,5,10 However, various rating scales of severity are often so different that they are not comparable. In addition, data are usually sparse when multiple levels of severity are used.4 The difficulties of the assignment of correct prognoses (short of expected eventual lethality for abnormalities) led us to use only the assessment of presumed lethality as the clearest predictor of severity. However, some of our cases that were classified as non-lethal involved fetuses with anomalies that were believed to be associated with an anticipated poor prognosis for normal developmental outcome. Because of this, we believe that the effect of increased severity of abnormalities on parental decisions may be underestimated by our model. Previous studies have reported that the involvement of the central nervous system greatly increases the likelihood of pregnancy interruption.4,5 The results of our univariable analyses confirm these reports. Nevertheless, the multiple logistic regression analysis suggests that the decision to interrupt pregnancy is more dependent on whether those anomalies are associated with conditions that are perceived to be lethal or as part of complex anomaly clusters. This supports the importance of the provision of an accurate prenatal diagnosis by individuals with high skill levels and the knowledge base to provide accurate information. Overall, pregnancies with multiple anomalies were more likely to be interrupted electively than were pregnancies with single anomalies. This observation is not surprising and may be explained by the perception that there is a greater level of morbidity associated with
1496 Table III
Rauch et al Abnormalities of the fetus (organ system), placenta, and amniotic fluid that were identified prenatally: NJFAR
Affected organ system
Pregnancy interruption*
Natural course pregnancyy
Unadjusted relative risk (95% CI)
Adjusted relative risk (95% CI)z
Head (intracranial) Face Neck Chest /thorax Cardiac Abdominal Genitourinary tract Extremities Skeletal Neural tube Placenta Amniotic fluid Umbilical cord Other abnormalities Aneuploidy Presumed lethal Multiple anomalies
14 7 6 6 11 10 11 11 7 7 2 4 6 4 4 18 25
13 8 10 8 27 19 16 15 5 4 3 4 4 4 5 9 33
1.4 1.8 1.2 1.5 0.8 1.1 1.4 1.5 2.8 3.6 2.0 2.0 3.1 2.0 1.6 4.1 1.5
3.5 1.9 0.7 2.0 0.8 1.1 2.3 1.6 3.5 4.1 0.8 2.9 2.6 1.6 2.0 7.1 1.4
(44%) (22%) (19%) (19%) (34%) (31%) (34%) (34%) (22%) (22%) (6%) (13%) (19%) (13%) (13%) (56%) (78%)
(20%) (12%) (15%) (12%) (42%) (29%) (25%) (23%) (8%) (6%) (5%) (6%) (6%) (6%) (8%) (14%) (52%)
(1.0-2.0) (0.7-4.5) (0.5-3.1) (0.6-4.0) (0.5-1.5) (0.5-2.0) (0.7-2.7) (0.8-2.9) (1.0-8.3) (1.1-11.5) (0.1-31.4) (0.5-7.6) (0.9-10.0) (0.5-7.6) (0.5-5.6) (2.1-8.0) (1.1-2.1)
(1.3, 9.3) (0.6, 6.1) (0.2, 2.6) (0.6, 6.8) (0.3, 1.9) (0.4, 2.7) (0.8, 6.3) (0.6, 4.3) (1.0, 12.8) (1.1, 16.0) (0.04, 16.5) (0.6, 13.7) (0.7, 10.4) (0.3, 7.4) (0.5, 9.0) (2.5, 9.6) (1.1, 1.8)
Fetuses may have more than 1 involved organ system. * n = 32. y n = 65. z Adjusted for gestational age.
multiple anomalies. However, when the group of fetuses with anomalies that were presumed to be lethal was examined separately, we unexpectedly observed that the presence of a single lethal anomaly was more likely to result in pregnancy interruption than the presence of a cluster of multiple anomalies that were presumed to be lethal. It is possible that more information may be available about the prognosis and pregnancy course for an isolated lethal anomaly (such as acrania and aneuploidy), whereas a cluster of anomalies may have a less apparent specific diagnosis, despite a presumed lethal prognosis. The resultant uncertainty may make informed decisions for the management of the pregnancy that much more difficult for the parents. Nevertheless, other unrecognized factors that were not available in our data set may have influenced this finding. We were unable to determine the reasons that some of our subjects did not receive genetic counseling or the content of counseling for those who did. More than 85% of NJFAR subjects are offered genetic counseling, of whom approximately 25% decline (unpublished data). This is consistent with the current study cohort’s use of counseling. Presumably, the maternal-fetal medicine specialists also performed counseling. Although there was a somewhat higher rate of genetic counseling among pregnancy interruption cases, this did not reach significance in the univariable or logistic regression analyses. Therefore, in this group, genetic counseling did not appear to play a dominant role in the outcomes. Because some of the data on genetic counseling were
missing, we are hesitant to draw any strong conclusions about its influences on decisions to interrupt a pregnancy in our population. In 1994, Crane et al11 conducted a randomized trial in the United States on the impact of routine prenatal ultrasound screening. That study focused only on basic ultrasound examination screening and had an overall low detection rate of fetal anomalies (18%). One of the conclusions from that study was that ultrasonography screening in a low-risk pregnant population had no significant impact on the frequency of abortion for fetuses with anomalies that were diagnosed before 24 weeks of gestation. We have demonstrated in our cohort several important features of prenatal diagnosis for fetal anomalies, which include earlier diagnosis, multiple anomalies, and lethality that may significantly influence decisions to interrupt pregnancies with fetal anomalies. Some of this difference may be due to the high level of prenatal diagnosis skills of the maternal-fetal medicine practitioners of the NJFAR sites. Our findings are consistent with European studies, in which early diagnosis of fetal anomalies was associated with a tendency to interrupt the affected pregnancies.12,13 An earlier diagnosis likely affords parents more autonomy to make informed decisions and allows for more time to plan systematically and to prepare emotionally for elective pregnancy interruptions. Our study had several limitations. Because NJFAR is not a comprehensive population-based registry that covers the entire state, our distribution of fetal anomalies and rates of pregnancy interruptions may be
Rauch et al subject to some selection bias. Because cases were contributed from several maternal-fetal medicine groups, site- or practitioner-specific practices may have introduced some bias in the reporting or management for some cases. However, there were insufficient numbers of cases across the multiple practitioners and sites to be able to examine this issue in detail in our study. The sample size also limited our ability to assess weak associations, address large numbers of potential confounders, and explore associations with less frequent abnormalities. It is possible that our results may be subject to an ascertainment bias because outcomes of 30% of cases were not available. Nevertheless, cases were not reported selectively from individual practitioners. Physicians either sent responses for all cases or for none. Because there is no reason to suspect that the severity of cases had a biased distribution among certain practices, any ascertainment bias is limited mainly to the physician’s willingness to respond instead of to the abnormality severity or pregnancy outcome. Although we cannot be certain that our results can be generalized to other states or parts of the country, patients in the NJFAR were ascertained from a large and diverse population referral base that included many sites around the state of New Jersey. In addition, our findings are representative of actual clinical practices and parental decisionmaking because there are no specified protocols for pregnancy management for practices that participate in NJFAR. Our study suggests that the earlier diagnosis of fetal structural anomalies, a presumed lethal prognosis, and the presence of multiple anomalies significantly influence the likelihood to interrupt an anomalous pregnancy. Importantly, these results support the concept that early and accurate diagnosis is essential for the optimization of parental autonomy pregnancy management.
Acknowledgments NJFAR site investigators for UMDNJ include: John Smulian, MD, MPH (NJFAR Principal Investigator) for UMDNJ-Robert Wood Johnson Medical School (RWJMS) at Robert Wood Johnson University Hospital and Saint Peter’s University Hospital, New Brunswick, NJ; Joseph Canterino, MD, for UMDNJ-RWJMS at Jersey Shore University Medical Center, Neptune, NJ; Richard Fischer, MD, for UMDNJ-RWJMS at Cooper
1497 Hospital System, Camden, NJ; Joseph Appuzzio, MD, for UMDNJ-New Jersey Medical School at University Hospital, Newark, NJ; Marjorie Royle, PhD, and Leslie Beres-Sochka, MS, at New Jersey Department of Health and Senior Services, Early Identification Monitoring Program and New Jersey Center for Birth Defects Research and Prevention.
References 1. Limb CJ, Holmes LB. Anencephaly: changes in prenatal detection and birth status, 1972 through 1990. Am J Obstet Gynecol 1994;170:1333-8. 2. Forrester MB, Merz RD, Yoon PW. Impact of prenatal diagnosis and elective termination on the prevalence of selected birth defects in Hawaii. Am J Epidemiol 1998;148:1206-11. 3. Lin AE, Herring AH, Amstutz KS, Westgte MN, Lacro RV, Al-Jufan M, et al. Cardiovascular malformations: changes in prevalence and birth status, 1972-1990. Am J Med Genet 1990;4:102-10. 4. Schechtman KB, Gray DL, Baty JD, Rothman SM. Decisionmaking for termination of pregnancies with fetal anomalies: analysis of 53,000 pregnancies. Obstet Gynecol 2002;99:216-22. 5. Pryde PG, Isada NB, Hallak M, Johnson MP, Odgers AE, Evans MI. Determinants of parental decision to abort or continue after non-aneuploidy ultrasound-detected fetal anomalies. Obstet Gynecol 1992;80:52-6. 6. Mansfield C, Hopfer S, Marteau TM. Termination rates after prenatal diagnosis of Down syndrome, spina bifida, anenchaphly, and Turner and Klinefelter syndromes: a systemic literature review: European Concerted Action DADA (Decision-making after the Diagnosis of a fetal Anomaly). Prenat Diagn 1999;19: 808-12. 7. Luck CA. Value of routine ultrasound scanning at 19 weeks: a four-year study of 8,849 deliveries. BMJ 1992;304:1474-8. 8. Skupski DW, Chervenak FA, McCullough LB. Routine obstetric ultrasound. Int J Gynaecol Obstet 1995;50:233-42. 9. Persutte WH. Failure to address the psychosocial benefit of prenatal sonogram: another failing of the RADIUS study: routine antenatal diagnostic imaging with ultrasound. J Ultrasound Med 1995;14:795-6. 10. Grevengood C, Shulman LP, Dungan JS, Martens P, Phillips OP, Emerson DS, et al. Severity of anomaly influences decision to terminate pregnancies affected with fetal neural tube defects. Fetal Diagn Ther 1994;9:273-7. 11. Crane JP, LeFevre ML, Winborn RC, Evans JK, Ewigman BG, Raymond P, et al, for the RADIUS study group. A randomized trial of prenatal ultrasonographic screening: impact on the detection, management, and outcome of anomalous fetuses. Am J Obstet Gynecol 1994;171:392-9. 12. Grandjean H, Larroque D, Levi S. The performance of routine ultrasonographic screening of pregnancies in the Eurofetus study. Am J Obstet Gynecol 1999;181:446-54. 13. Chitty LS, Hunt GH, Moore J, Lobb MO. Effectiveness of routine ultrasonography in detecting fetal structural abnormalities in a low-risk population. BMJ 1991;303:1165-9.
www.ajog.org
Pregnancy interruption after second trimester diagnosis of fetal structural anomalies: The New Jersey Fetal Abnormalities Registry Eden R. Rauch, MD, MPH,a,b John C. Smulian, MD, MPH,a Kristin DePrince, MS, CGC,a Cande V. Ananth, PhD, MPH,a Stephen W. Marcella, MD, MPH,b for the New Jersey Fetal Abnormalities Registry Departments of Obstetrics, Gynecology and Reproductive Sciences,a and Environmental and Community Medicine,b University of Medicine and Dentistry of New Jersey-Robert Wood Johnson Medical School/Robert Wood Johnson University Hospital, New Brunswick, NJ Received for publication October 18, 2004; revised January 7, 2005; accepted February 17, 2005
KEY WORDS Fetal abnormality Anomaly Birth defect Ultrasound examination
Objective: The purpose of this study was to identify factors that predict a decision to interrupt a pregnancy in which there are fetal anomalies in the second trimester. Study design: The New Jersey Fetal Abnormalities Registry prospectively recruits and collects information on pregnancies (R15 weeks of gestation) from New Jersey residents in whom a fetal structural anomaly has been suspected by maternal-fetal medicine specialists. Enrolled pregnancies that have major fetal structural abnormalities identified from 15 to 23 weeks of gestation were included. Outcomes were classified as either elective interruption or a natural pregnancy course, which might include a spontaneous fetal death or live birth. Predictors of elective interruption of pregnancy were examined with univariable and multivariable logistic regression analyses. Results: Of the 97 cases, 33% of the women (n = 32) interrupted the pregnancy. Significant variables in the regression model that were associated with a decision to interrupt a pregnancy were earlier identification of fetal anomalies (19.0 G 2 weeks of gestation vs 20.5 G 2 weeks of gestation; P = .003), the presence of multiple anomalies (78% [25/32] vs 52% [33/63]; P = .01], and a presumption of lethality (56% [18/32] vs 14% [9/65]; P = .0001). These variables corresponded to an odds ratio for pregnancy interruption of 4.2 (95% CI, 1.0, 17.0) for multiple anomalies, 0.8 (95% CI, 0.7, 1.0) for each week of advancing gestational age, and 36.1 (95% CI, 2.9, 450.7) for presumed lethal abnormalities. Conclusion: Early diagnosis, the identification of multiple abnormalities, and an assessment of likely lethality of fetal anomalies are important factors for the optimization of parental autonomy in deciding pregnancy management. Ó 2005 Mosby, Inc. All rights reserved.
Supported in part by National Institutes of Health grant R01-HD038902 (C.V.A.; J.C.S); the New Jersey Fetal Abnormalities Registry (Principle Investigator, J.C.S.) was supported by a cooperative agreement between the University of Medicine and Dentistry of New Jersey- Robert Wood Johnson Medical School and the New Jersey Department of Health and Senior Services through the National Birth Defects Prevention Study of the Centers for Disease Control and Prevention. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2005 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2005.02.099
Rauch et al Recent advances in prenatal diagnosis have significantly improved the detection of birth defects before birth. As a result, couples often have the options of expectant care, elective interruption, or, in rare instances, fetal surgery. Several studies have noted a decline in the live birth prevalence of certain birth defects over the last several decades, most notably anencephaly, Down syndrome, spina bifida, and cardiovascular malformations.1-3 Technologic advances in ultrasonography and prenatal diagnosis aid in the diagnosis of fetal anomalies and provide information that can influence the decision to continue or interrupt a pregnancy. Schectman et al4 showed that the severity of structural anomalies directly correlated with abortion rates of anomalous fetuses. Furthermore, at similar degrees of severity, central nervous system anomalies were more likely to be terminated electively. Other studies have reported that the specific organ system that is anomalous, the severity of structural anomalies, and the number of defects in other organs are associated with decisions to interrupt a pregnancy.5,6 However, there is limited information that is available to help understand whether other factors influence parental choices in the management of pregnancies that are prenatally diagnosed with structural anomalies. The New Jersey Fetal Abnormalities Registry (NJFAR) was established in 1999 at the University of Medicine and Dentistry of New Jersey (UMDNJ)Robert Wood Johnson Medical School by a cooperative agreement with the New Jersey Department of Health and Senior Services and the Centers for Disease Control and Prevention through the Centers for Birth Defects Research and Prevention. NJFAR collects information on pregnancies (R15 weeks of gestation) of New Jersey residents in which a fetal structural anomaly has been identified or suspected by maternal-fetal medicine specialists across New Jersey. This study was designed to use NJFAR-registered pregnancies to identify factors that predict parental decisions to interrupt pregnancies with fetuses for whom structural anomalies are identified in the second trimester.
Material and methods This study was approved by the Institutional Review Boards of the New Jersey Department of Health and Senior Services, the UMDNJ-Robert Wood Johnson Medical School, and all other affiliated sites (Acknowledgments). Enrollment of NJFAR cases began in January 2000 from maternal-fetal medicine practices that are linked with the UMDNJ Health Systems. These sites included prenatal diagnostic practices that are affiliated with the Robert Wood Johnson Medical School’s New Brunswick, Neptune, and Camden campuses and the New Jersey Medical School in Newark. These practices
1493 include approximately 20 maternal-fetal medicine specialists and provide prenatal diagnostic services to an estimated population of 31,000 women each year. All sites provide access to a full range of prenatal diagnostic options, genetic counseling, physician counseling, and discussions about pregnancy management. Pregnancy interruption services were available in each community that surrounds participating NJFAR practices. NJFAR is not a comprehensive population-based registry that covers the entire state. However, because a large portion of prenatally diagnosed fetal anomalies in New Jersey are seen by maternal-fetal medicine specialists at some point during their course of care, the practices are useful sources for capturing these cases. Cases that are eligible for NJFAR enrollment include fetuses with significant structural abnormalities that are reportable to the New Jersey Department of Health and Senior Services Birth Defects Registry. Cases with aneuploidy were included only if they are associated with structural abnormalities. Additional non-UMDNJ cases that are obtained from the New Jersey Maternal-Fetal Medicine Society member practices who volunteered to participate in NJFAR were not included for this specific study. All NJFAR registered cases with an estimated date of delivery between January 2000 and February 2002 were eligible for this study. Eligible cases had pregnancy information that was recorded at NJFAR enrollment including intake date, primary physician’s name and phone number, patient’s first and last initials, maternal date of birth, city and zip code of residence, gestational age at diagnosis, estimated date of delivery, number of fetuses in pregnancy, gravidity, parity, race, obstetrics clinic/offices that were providing care, fetal abnormalities that were identified, diagnostic procedures, referral information, preference for genetic counseling, and pregnancy continuation plans. A subset of these NJFAR-enrolled patients who provided informed consent underwent an interview to obtain pregnancy outcomes and extensive health behavior, nutritional behavior, and occupational exposure information. When interview information was available, this information was used as a data source for this study. For the patients who did not participate in the interview or whose pregnancy outcomes were not known, a questionnaire was mailed to the subject’s obstetrician that requested pregnancy outcome information. Nonresponding physicians were contacted directly to encourage completion of questionnaires. Only patients with anomalies that were identified before 24 weeks of gestation were included for study, because interruption of pregnancy at R24 weeks is not readily available in New Jersey. On the pregnancy outcome questionnaire, the patients’ primary referring physicians were provided with maternal initials, date of birth, estimated date of delivery, date of ultrasound diagnosis, and a summary
1494 Table I
Rauch et al Maternal characteristics, based on the decision whether to interrupt pregnancy: NJFAR
Characteristic
Pregnancy interruption
Natural course pregnancy
P value
N Age (y)* Gestational age at diagnosis (wk)* Gravidity (n)y Parity (n)y White race (n/N) Assisted reproductive technology (n/N) Genetic counseling (n/N)
32 32.9 G 5.6 19.0 G 2 2 (1-5) 1 (0-4) 22/29 (76%) 2/22 (9%) 21/23 (91%)
65 31.9 G 6.5 20.5 G 2 2 (1-11) 1 (0-10) 39/57 (68%) 6/56 (11%) 38/52 (73%)
.45 .003 .75 .36 .47 .83 .08
* Data are given as mean G SD. y Data are given as median (range).
of the suspected fetal abnormalities by which they could identify their own patients. The primary outcome of interest was the parental decision to interrupt the pregnancy. New Jersey statutes allow the interruption of pregnancy before viability without imposing specific gestational age or weight criteria. Outcomes were classified as either elective interruption (group 1) or natural pregnancy course (group 2), which might include a spontaneous fetal death or live birth. The variables that were recorded included maternal and fetal ages at diagnosis, gravidity, parity, race, specific body/organ system that was affected, aneuploidy, the number and type of organ/body systems that were affected, the use of genetic counseling, the presumed lethality of the anomaly, and the use of assisted reproductive or conception techniques. For the purposes of this study, cases were classified as presumed lethal on the basis of the specific spectrum of prenatally identified abnormalities. This was assessed with the information that was provided by the diagnosing maternal-fetal medicine specialist at the time of diagnosis and was independently reassessed by 1 of the authors (J.C.S.). Lethality was presumed in the presence of an abnormality or spectrum of abnormalities that were associated with either a likely in utero death or, if live born, an inability to survive independently even with surgical correction of the abnormality. Continuous variables were compared with the use of parametric and nonparametric tests, where appropriate. Categorical variables were compared with the use of contingency tables (chi-square and Fisher’s exact test). The unadjusted relative risk and 95% confidence interval (CI) for specific abnormalities that were associated with pregnancy interruption were derived. Similar analyses were performed after an adjustment was made for the gestational age at diagnosis. Variables that reached statistical significance with a probability value of %.05 were also examined for clinical significance. The potentially important predictors were used in forward selection logistic regression, with an entry criteria of a
probability value of !.10 to develop a model for the prediction of pregnancy interruption. Interactions were retained in the model after the statistical criteria of a probability value of %.10 was met. The logistic regression model was used to derive odds ratios (ORs) with 95% CI. The Hosmer-Lemeshow goodness-of-fit test with interactions that were included in the model was calculated to ensure the adequate fit of the logistic regression model. We reported the significant interactions by stratified analysis. All data were analyzed with SAS software (SAS Institute, Cary, NC).
Results One hundred thirty-eight questionnaires were mailed to physician’s offices, of which 78% (n = 108 questionnaires) were returned. An additional 11 cases had incomplete reporting of outcomes and were excluded from the analysis. Therefore, the study group comprised 97 cases with a fetal structural anomaly that was diagnosed between 15 and 24 weeks of gestation and estimated dates of delivery between January 2000 and February 2002. Twenty-three cases (23%) had information that was available from the detailed NJFAR interviews, which did not differ from the questionnaire information. Of the 97 cases, 33% (n = 32 cases) had a pregnancy interruption (group 1). Demographic and pregnancy information are presented in Table I. Group 1 had earlier gestational ages for identification of fetal anomalies by 1.5 weeks. The findings for each of the 27 fetuses with abnormalities that were presumed to be lethal are described in Table II. As shown in Table III, group 1 fetuses were more likely to have multiple anomalies, anomalies that were presumed to be lethal, head (intracranial) anomalies, neural tube anomalies, and skeletal anomalies. These associations were stronger after an adjustment was made for gestational age. Maternal age, race, and parity were not significant in the univariable analyses and therefore were not included for the adjustment. Because
Rauch et al
1495
Table II Description of 37 fetuses having anomalies identified in the second trimester that were presumed to be lethal: NJFAR Anomaly
N
Isolated acrania Isolated anencephaly Isolated bilateral renal agenesis Nonimmune hydrops With cystic hygroma With Ebstein’s anomaly With severe cardiomegaly Severe arthrogryposis sequence With generalized skin edema, small stomach/bladder, bowed femurs With scalp edema, cardiomegaly, hepatomegaly, very small stomach, severe hydronephrosis Bilateral severe polycystic dysplastic kidneys, ventriculomegaly, oligohydramnios Limb body wall defect Severe ventriculomegaly, large ventricular septal defect with complex disorder of great vessels, suspected tracheoesophageal fistula, abnormal facial structures, 46, XX with partial monosomy for chromosome 1 Large atrioventricular septal defect, fisted hands, trisomy 18 Omphalocele, ventricular septal defect, trisomy 18 Sirenomelia Large atrioventricular canal defect, suspected esophageal atresia, extremity contractures Severe hydrocephalus, hypoplastic right heart, hydronephrosis, partial duplication of chromosome 1 Cystic hygroma, Dandy-Walker malformation, pleural effusion, atrioventricular canal defect, hypoplastic right ventricle
4 3 3 6 4 1 1 2 1 1 1 1 1
1 1 1 1 1 1
information about genetic counseling was missing from 21 cases (23%), we were not able to use that variable for the adjustment. These variables and the interaction between multiple anomalies and presumed lethality were evaluated by logistic regression. The model confirmed that the best 3 predictors of a decision to interrupt a pregnancy were the presence of multiple anomalies (OR, 4.2; 95% CI, 1.0, 17.0; P = .05), earlier gestational age at diagnosis (OR, 0.8; 95% CI, 0.7, 1.0; P = .04) for each week of gestation, and anomalies that were presumed to be lethal (OR, 36.1; 95% CI, 2.9, 450.7; P = .005). Because an important interaction existed between lethality and the number of anomalies (P = .10), further analysis was stratified by multiple and single anomaly subgroups. When the logistic regression model was adjusted for gestational age, the stratified odds ratio for pregnancy interruption was 3.64 (95% CI, 1.2-11.4) for multiple anomalies that were presumed to be lethal and 45.2 (95% CI, 2.8, 744.8) for a single anomaly that were considered to be lethal. The Hosmer-
Lemeshow goodness-of-fit test with the interaction that was included in the model was not statistically significant (P O .5), which indicated a good model fit.
Comment The ability of ultrasound examination to calculate gestational age, to detect multiple gestations, and to evaluate fetal well-being by diagnosing growth, fluid, and structural abnormalities has had a major impact on the practice of obstetrics. A number of investigators have described benefits of prenatal diagnosis of fetal anomalies (such as improvement of management at delivery, maintenance of respect for patient autonomy, and promotion of the psychologic advantages and parental bonding that might be disrupted at the birth of a child with an unusual unsuspected anomaly).7-9 Our study identified a number of important components of prenatal diagnosis that appear to affect decisions to interrupt affected pregnancies with structural abnormalities that include presumed lethality, the presence of multiple anomalies, and earlier fetal age at diagnosis. The severity of structural anomalies has been recognized by many investigators to influence the likelihood of pregnancy interruption.4,5,10 However, various rating scales of severity are often so different that they are not comparable. In addition, data are usually sparse when multiple levels of severity are used.4 The difficulties of the assignment of correct prognoses (short of expected eventual lethality for abnormalities) led us to use only the assessment of presumed lethality as the clearest predictor of severity. However, some of our cases that were classified as non-lethal involved fetuses with anomalies that were believed to be associated with an anticipated poor prognosis for normal developmental outcome. Because of this, we believe that the effect of increased severity of abnormalities on parental decisions may be underestimated by our model. Previous studies have reported that the involvement of the central nervous system greatly increases the likelihood of pregnancy interruption.4,5 The results of our univariable analyses confirm these reports. Nevertheless, the multiple logistic regression analysis suggests that the decision to interrupt pregnancy is more dependent on whether those anomalies are associated with conditions that are perceived to be lethal or as part of complex anomaly clusters. This supports the importance of the provision of an accurate prenatal diagnosis by individuals with high skill levels and the knowledge base to provide accurate information. Overall, pregnancies with multiple anomalies were more likely to be interrupted electively than were pregnancies with single anomalies. This observation is not surprising and may be explained by the perception that there is a greater level of morbidity associated with
1496 Table III
Rauch et al Abnormalities of the fetus (organ system), placenta, and amniotic fluid that were identified prenatally: NJFAR
Affected organ system
Pregnancy interruption*
Natural course pregnancyy
Unadjusted relative risk (95% CI)
Adjusted relative risk (95% CI)z
Head (intracranial) Face Neck Chest /thorax Cardiac Abdominal Genitourinary tract Extremities Skeletal Neural tube Placenta Amniotic fluid Umbilical cord Other abnormalities Aneuploidy Presumed lethal Multiple anomalies
14 7 6 6 11 10 11 11 7 7 2 4 6 4 4 18 25
13 8 10 8 27 19 16 15 5 4 3 4 4 4 5 9 33
1.4 1.8 1.2 1.5 0.8 1.1 1.4 1.5 2.8 3.6 2.0 2.0 3.1 2.0 1.6 4.1 1.5
3.5 1.9 0.7 2.0 0.8 1.1 2.3 1.6 3.5 4.1 0.8 2.9 2.6 1.6 2.0 7.1 1.4
(44%) (22%) (19%) (19%) (34%) (31%) (34%) (34%) (22%) (22%) (6%) (13%) (19%) (13%) (13%) (56%) (78%)
(20%) (12%) (15%) (12%) (42%) (29%) (25%) (23%) (8%) (6%) (5%) (6%) (6%) (6%) (8%) (14%) (52%)
(1.0-2.0) (0.7-4.5) (0.5-3.1) (0.6-4.0) (0.5-1.5) (0.5-2.0) (0.7-2.7) (0.8-2.9) (1.0-8.3) (1.1-11.5) (0.1-31.4) (0.5-7.6) (0.9-10.0) (0.5-7.6) (0.5-5.6) (2.1-8.0) (1.1-2.1)
(1.3, 9.3) (0.6, 6.1) (0.2, 2.6) (0.6, 6.8) (0.3, 1.9) (0.4, 2.7) (0.8, 6.3) (0.6, 4.3) (1.0, 12.8) (1.1, 16.0) (0.04, 16.5) (0.6, 13.7) (0.7, 10.4) (0.3, 7.4) (0.5, 9.0) (2.5, 9.6) (1.1, 1.8)
Fetuses may have more than 1 involved organ system. * n = 32. y n = 65. z Adjusted for gestational age.
multiple anomalies. However, when the group of fetuses with anomalies that were presumed to be lethal was examined separately, we unexpectedly observed that the presence of a single lethal anomaly was more likely to result in pregnancy interruption than the presence of a cluster of multiple anomalies that were presumed to be lethal. It is possible that more information may be available about the prognosis and pregnancy course for an isolated lethal anomaly (such as acrania and aneuploidy), whereas a cluster of anomalies may have a less apparent specific diagnosis, despite a presumed lethal prognosis. The resultant uncertainty may make informed decisions for the management of the pregnancy that much more difficult for the parents. Nevertheless, other unrecognized factors that were not available in our data set may have influenced this finding. We were unable to determine the reasons that some of our subjects did not receive genetic counseling or the content of counseling for those who did. More than 85% of NJFAR subjects are offered genetic counseling, of whom approximately 25% decline (unpublished data). This is consistent with the current study cohort’s use of counseling. Presumably, the maternal-fetal medicine specialists also performed counseling. Although there was a somewhat higher rate of genetic counseling among pregnancy interruption cases, this did not reach significance in the univariable or logistic regression analyses. Therefore, in this group, genetic counseling did not appear to play a dominant role in the outcomes. Because some of the data on genetic counseling were
missing, we are hesitant to draw any strong conclusions about its influences on decisions to interrupt a pregnancy in our population. In 1994, Crane et al11 conducted a randomized trial in the United States on the impact of routine prenatal ultrasound screening. That study focused only on basic ultrasound examination screening and had an overall low detection rate of fetal anomalies (18%). One of the conclusions from that study was that ultrasonography screening in a low-risk pregnant population had no significant impact on the frequency of abortion for fetuses with anomalies that were diagnosed before 24 weeks of gestation. We have demonstrated in our cohort several important features of prenatal diagnosis for fetal anomalies, which include earlier diagnosis, multiple anomalies, and lethality that may significantly influence decisions to interrupt pregnancies with fetal anomalies. Some of this difference may be due to the high level of prenatal diagnosis skills of the maternal-fetal medicine practitioners of the NJFAR sites. Our findings are consistent with European studies, in which early diagnosis of fetal anomalies was associated with a tendency to interrupt the affected pregnancies.12,13 An earlier diagnosis likely affords parents more autonomy to make informed decisions and allows for more time to plan systematically and to prepare emotionally for elective pregnancy interruptions. Our study had several limitations. Because NJFAR is not a comprehensive population-based registry that covers the entire state, our distribution of fetal anomalies and rates of pregnancy interruptions may be
Rauch et al subject to some selection bias. Because cases were contributed from several maternal-fetal medicine groups, site- or practitioner-specific practices may have introduced some bias in the reporting or management for some cases. However, there were insufficient numbers of cases across the multiple practitioners and sites to be able to examine this issue in detail in our study. The sample size also limited our ability to assess weak associations, address large numbers of potential confounders, and explore associations with less frequent abnormalities. It is possible that our results may be subject to an ascertainment bias because outcomes of 30% of cases were not available. Nevertheless, cases were not reported selectively from individual practitioners. Physicians either sent responses for all cases or for none. Because there is no reason to suspect that the severity of cases had a biased distribution among certain practices, any ascertainment bias is limited mainly to the physician’s willingness to respond instead of to the abnormality severity or pregnancy outcome. Although we cannot be certain that our results can be generalized to other states or parts of the country, patients in the NJFAR were ascertained from a large and diverse population referral base that included many sites around the state of New Jersey. In addition, our findings are representative of actual clinical practices and parental decisionmaking because there are no specified protocols for pregnancy management for practices that participate in NJFAR. Our study suggests that the earlier diagnosis of fetal structural anomalies, a presumed lethal prognosis, and the presence of multiple anomalies significantly influence the likelihood to interrupt an anomalous pregnancy. Importantly, these results support the concept that early and accurate diagnosis is essential for the optimization of parental autonomy pregnancy management.
Acknowledgments NJFAR site investigators for UMDNJ include: John Smulian, MD, MPH (NJFAR Principal Investigator) for UMDNJ-Robert Wood Johnson Medical School (RWJMS) at Robert Wood Johnson University Hospital and Saint Peter’s University Hospital, New Brunswick, NJ; Joseph Canterino, MD, for UMDNJ-RWJMS at Jersey Shore University Medical Center, Neptune, NJ; Richard Fischer, MD, for UMDNJ-RWJMS at Cooper
1497 Hospital System, Camden, NJ; Joseph Appuzzio, MD, for UMDNJ-New Jersey Medical School at University Hospital, Newark, NJ; Marjorie Royle, PhD, and Leslie Beres-Sochka, MS, at New Jersey Department of Health and Senior Services, Early Identification Monitoring Program and New Jersey Center for Birth Defects Research and Prevention.
References 1. Limb CJ, Holmes LB. Anencephaly: changes in prenatal detection and birth status, 1972 through 1990. Am J Obstet Gynecol 1994;170:1333-8. 2. Forrester MB, Merz RD, Yoon PW. Impact of prenatal diagnosis and elective termination on the prevalence of selected birth defects in Hawaii. Am J Epidemiol 1998;148:1206-11. 3. Lin AE, Herring AH, Amstutz KS, Westgte MN, Lacro RV, Al-Jufan M, et al. Cardiovascular malformations: changes in prevalence and birth status, 1972-1990. Am J Med Genet 1990;4:102-10. 4. Schechtman KB, Gray DL, Baty JD, Rothman SM. Decisionmaking for termination of pregnancies with fetal anomalies: analysis of 53,000 pregnancies. Obstet Gynecol 2002;99:216-22. 5. Pryde PG, Isada NB, Hallak M, Johnson MP, Odgers AE, Evans MI. Determinants of parental decision to abort or continue after non-aneuploidy ultrasound-detected fetal anomalies. Obstet Gynecol 1992;80:52-6. 6. Mansfield C, Hopfer S, Marteau TM. Termination rates after prenatal diagnosis of Down syndrome, spina bifida, anenchaphly, and Turner and Klinefelter syndromes: a systemic literature review: European Concerted Action DADA (Decision-making after the Diagnosis of a fetal Anomaly). Prenat Diagn 1999;19: 808-12. 7. Luck CA. Value of routine ultrasound scanning at 19 weeks: a four-year study of 8,849 deliveries. BMJ 1992;304:1474-8. 8. Skupski DW, Chervenak FA, McCullough LB. Routine obstetric ultrasound. Int J Gynaecol Obstet 1995;50:233-42. 9. Persutte WH. Failure to address the psychosocial benefit of prenatal sonogram: another failing of the RADIUS study: routine antenatal diagnostic imaging with ultrasound. J Ultrasound Med 1995;14:795-6. 10. Grevengood C, Shulman LP, Dungan JS, Martens P, Phillips OP, Emerson DS, et al. Severity of anomaly influences decision to terminate pregnancies affected with fetal neural tube defects. Fetal Diagn Ther 1994;9:273-7. 11. Crane JP, LeFevre ML, Winborn RC, Evans JK, Ewigman BG, Raymond P, et al, for the RADIUS study group. A randomized trial of prenatal ultrasonographic screening: impact on the detection, management, and outcome of anomalous fetuses. Am J Obstet Gynecol 1994;171:392-9. 12. Grandjean H, Larroque D, Levi S. The performance of routine ultrasonographic screening of pregnancies in the Eurofetus study. Am J Obstet Gynecol 1999;181:446-54. 13. Chitty LS, Hunt GH, Moore J, Lobb MO. Effectiveness of routine ultrasonography in detecting fetal structural abnormalities in a low-risk population. BMJ 1991;303:1165-9.