Alcohol abuse--a persistent preventable risk for congenital anomalies.

OBJECTIVE: To examine whether alcohol abuse (ALC) continued to be a health hazard to pregnant women in the 1990s. STUDY DESIGN: Analysis of a perinatal data base comprising 170,258 women with singleton pregnancies. Univariate cross table analysis and logistic regression were conducted to examine the association between alcohol abuse and congenital malformations coded according to the International Classification of Diseases (ICD). RESULTS: 14,727/170,258 mothers (8.6%) admitted to ALC during pregnancy and 36,705/170,258 (21.6%) to smoking. Anomaly rates for ALC (365/14,092, 4.3%) vs. Non-ALC (6187/149,344, 4.0%) differed significantly (p<0.001). The rates of specific anomalies varied between <0.1% and 1.1%. Odds ratios for 16 ICD 9 anomaly categories were >1 in 14 instances overall (Sign test, p=0.004), in 12 instances in women <30 years (p=0.08), and in 13 instances in women over 30 years (p=0.02). Congenital anomalies of the "respiratory system" (ICD9 748), of "genital organs" (ICD9 752.1), of the "integument" (ICD9 757), and "other anomalies of limbs/other musculoskeletal anomalies" (ICD 755/756) were statistically significantly associated with ALC, especially in women>30 years. CONCLUSION: ALC in pregnancy continued to be an important factor independently associated with an increased incidence of a broader range of congenital anomalies than previously recognized. Risk for anatomic anomalies was increased in offspring of ALC women over age 30, consistent with previous reports of increased risk of neurobehavioral abnormality in offspring of women over 30.

The effects of hypoxia and hyperoxia on fetal-placental vascular tone and inflammatory cytokine production.

OBJECTIVE: To determine the effects of fetal hypoxia and hyperoxia on placental vascular tone and production of interleukin-6 and tumor necrosis factor-alpha. STUDY DESIGN: The maternal and fetal circulation of 2 cotyledons from 5 human placentas were perfused for 4 hours. The fetal circulation of 1 cotyledon was perfused with hypoxic Hanks' balanced salt solution; the other was perfused with hyperoxic Hanks' balanced salt solution. Fetal vascular pressures were recorded every 10 minutes, and fetal vein effluents were collected hourly. RESULTS: Fetal-placental vascular perfusion pressure was reduced from baseline during hypoxic conditions. Cytokine concentrations were elevated during hyperoxic conditions compared with hypoxic conditions, with significant differences achieved at 2, 3, and 4 hours for interleukin-6 and at 4 hours for tumor necrosis factor-alpha. CONCLUSION: Fetal-placental vasodilation may be a compensatory mechanism to improve hypoxic conditions. Supraphysiologic oxygenation may contribute to the fetal inflammatory response syndrome and to the development of cerebral palsy.

Extreme elevation of maternal serum alpha-fetoprotein associated with mosaic trisomy 8 in a liveborn.

Constitutional mosaic trisomy 8 has been associated with syndromic dysmorphology, corneal opacities, leukemias, and trophoblastic disease. However, abnormal maternal serum alpha-fetoprotein (MSAFP) has not been reported in association with mosaic trisomy 8. Our case first presented for evaluation of an extremely elevated MSAFP with mild elevation of MShCG in an otherwise normal pregnancy: MSAFP 13.89 MoM, MShCG 3.57 MoM, and MSuE3 1.04 MoM. Fetal dysmorphism was limited to bilateral pyelectasis and a prominent third ventricle. Spontaneous labor at 38 weeks resulted in the birth of a 3,570-gram AGA male with APGARs 7(1)/8(5). The neonate had facial asymmetry, 5th finger clinodactyly, 2-3 toe syndactyly, undescended testicle, abnormal prepuce, and mild pyelectasis. CT scan revealed hypoplasia of the corpus callosum, while echocardiography demonstrated bicuspid aortic valve, and the neonatal karyotype (blood) returned 46,XY/47,XY+8. Evaluation at 3 months revealed more prominent facial asymmetry, plagiocephaly, plantar creases, descent of the testis, and mild developmental delay. Review of the literature does not include any previously reported maternal serum alpha-fetoprotein aberrations in mosaic trisomy 8.

Parental decision-making differences between patients in two healthcare systems for choroid plexus cysts.

OBJECTIVE: We evaluated the medical-sociological implications of parental perception of risk and decision-making choices for prenatally ascertained choroid plexus cysts (CPCs) between two obstetric populations with similar clinical situations. METHODS: The Wayne State University (WSU) Reproductive Genetics database and the Madigan Army Medical Center (MAMC) experience were reviewed to compare the rates of aneuploidy and invasive testing for cases with CPC. Aneuploidy rates were compared between those with isolated CPC, CPC with advanced maternal age (AMA), and CPC associated with multiple anomalies. RESULTS: 186 cases were identified in the WSU cohort, of whom 27 (15%) declined invasive fetal testing. In the remaining 159 cases, aneuploidy was present in 2/132 (1.5%) isolated CPCs, 3/11 (27%) CPCs with AMA, and 15/16 (93%) CPCs with multiple anomalies. 107 cases were identified in the MAMC cohort, of whom 99 (92%) declined invasive fetal testing. No cases of aneuploidy were found in the 3/12 AMA cases or 5/95 non-AMA cases who underwent amniocentesis. CONCLUSIONS: The 2 cases of aneuploidy with isolated CPC cannot be ignored, and provide an estimated attributable risk of at least 0.8%, a higher risk than 38 years of age. However, the parental sociologic context may be as important as the genetic-prognostic risk for decision-making.

Vasodilatory response of fetoplacental vasculature to adrenomedullin after constriction with the thromboxane sympathomimetic U46619.

OBJECTIVE: This study was undertaken to determine whether adrenomedullin, a hypotensive peptide, decreases vasomotor tone in fetoplacental vasculature that has been constricted with the thromboxane sympathomimetic U46619. STUDY DESIGN: The fetoplacental vascular beds of 20 perfused human placental cotyledons were vasoconstricted with a continuous infusion of U46619 (10(-8) mol/L). The vasculature was then sequentially injected with deionized water, 30 ng adrenomedullin, 300 ng adrenomedullin, and 3000 ng adrenomedullin. Any change in perfusion pressure was noted after each dose. In a separate experiment the fetoplacental vasculature in 2 perfused cotyledons from each of 10 placentas was vasoconstricted with U46619 (10(-8) mol/L). Adrenomedullin was infused continuously at either 200 ng/min (n = 5) or 2000 ng/min (n = 5) for 40 minutes. A corresponding control cotyledon from each placenta had isotonic sodium chloride solution added to its perfusion. Perfusion pressures were recorded every minute during the infusion and for 40 minutes afterward. Analysis of variance was used to compare pressure changes in the cotyledons that received bolus doses of adrenomedullin. Paired t tests of mean percentage pressure changes were used to compare the study and control groups that received the continuous infusions. RESULTS: In the cotyledons that received bolus doses of adrenomedullin, the mean (+/-SEM) percentage perfusion pressure changes from the baseline were -6.7 +/- 0.5 for 30 ng adrenomedullin (P =.0039), -8.5+/- 0.7 for 300 ng adrenomedullin (P <.0001), and -13.1 +/- 1.0 for 3000 ng adrenomedullin (P <.0001). With the continuous adrenomedullin infusion of 200 ng/min, there was no significant difference in the mean percentage pressure change from baseline between the study and control groups (-0.57%). At 2000 ng/min there was a significant difference (-15.34%; P <.0001). CONCLUSION: Adrenomedullin caused vasodilatation of fetoplacental vasculature previously constricted with the thromboxane sympathomimetic U46619 in the isolated perfused placental cotyledon. This vasodilatation occurred in a dose-dependent manner.

Localization of messenger ribonucleic acid for adrenomedullin and adrenomedullin receptor in the human placenta in normal pregnancies and pregnancies complicated by oligohydramnios.

OBJECTIVE: The purpose of this study was to identify the placental expression of adrenomedullin and adrenomedullin receptor messenger ribonucleic acid and compare them between placentas from pregnancies associated with oligohydramnios as a result of uteroplacental insufficiency and placentas from normal pregnancies. STUDY DESIGN: Total ribonucleic acid was extracted from the amnion, chorion, cotyledon, umbilical vein, and umbilical artery in 5 normal placentas and 3 placentas from pregnancies complicated by oligohydramnios. A cell line known to express messenger ribonucleic acid of adrenomedullin and its receptor was used to optimize the polymerase chain reaction and served as a positive control preparation in all experiments. Semiquantitative reverse transcriptase-polymerase chain reaction results for adrenomedullin and adrenomedullin receptor were compared between tissues as densitometric ratios of adrenomedullin or adrenomedullin receptor messenger ribonucleic acid to beta(2)-microglobulin messenger ribonucleic acid. Results were analyzed with a Kruskal-Wallis 1-way analysis of variance. Immunohistochemical staining with an antibody to human adrenomedullin was used to localize adrenomedullin in all tissue types. RESULTS: Messenger ribonucleic acid sequences for adrenomedullin and adrenomedullin receptor genes were identified in all tested placental tissue components. Within the normal placentas the expressions of adrenomedullin and adrenomedullin receptor messenger ribonucleic acid sequences did not differ statistically between the tissue components. Within placentas from patients with oligohydramnios the expressions of adrenomedullin and adrenomedullin receptor messenger ribonucleic acid did not differ statistically between the tissue components. When normal placentas were compared with placentas from pregnancies complicated by oligohydramnios, however, a 5-fold increase in adrenomedullin messenger ribonucleic acid and a 3-fold increase in adrenomedullin receptor messenger ribonucleic acid were seen in placentas from patients with oligohydramnios. Adrenomedullin immunoreactivity was present in all tissues studied. CONCLUSION: The expression of messenger ribonucleic acid for both adrenomedullin and its receptor in these tissue components implies that placental tissues function in both synthesis and action of adrenomedullin. The increased adrenomedullin messenger ribonucleic acid expression in the umbilical artery and the elevated adrenomedullin receptor messenger ribonucleic acid expression in the cotyledons of placentas from patients with oligohydramnios may represent a local fetoplacental physiologic adaptive response to vascular compromise.

Hypoperfusion causes increased production of interleukin 6 and tumor necrosis factor alpha in the isolated, dually perfused placental cotyledon.

OBJECTIVE: Our purpose was to determine whether exposure of the isolated, perfused human placental cotyledon to different fetal circuit perfusion rates, and to concomitant pressure differences, alters placental production of interleukin 6 and tumor necrosis factor alpha. STUDY DESIGN: The maternal and fetal circulations of 2 cotyledons from 5 placentas were perfused for 4 hours. The fetal circulation of 1 cotyledon was perfused at a low rate of 1 mL/min, and the other at a high rate of 10 mL/min. The maternal circulation of each cotyledon was perfused at 10 mL/min. Effluents from the fetal circulation were collected at hourly intervals, and concentrations of interleukin 6 and tumor necrosis factor alpha were determined by enzyme-linked immunosorbent assay. Concentrations of interleukin 6, obtained from a prior study with an estimated physiologic fetal circulation rate of 4 mL/min, were compared with the low and high perfusion rate results. RESULTS: Concentrations of interleukin 6 and tumor necrosis factor alpha were greater at the perfusion rate of 1 mL/min, in comparison with the perfusion rate of 10 mL/min, with statistically significant differences achieved at 2 and 4 hours for interleukin 6 and at 4 hours for tumor necrosis factor alpha. Concentrations of both cytokines increased exponentially with time. Placental perfusion pressures were significantly greater at the perfusion rate of 10 mL/min. CONCLUSION: Placental hypoperfusion results in an increased production of both interleukin 6 and tumor necrosis factor alpha. This finding links placental perfusion abnormalities to the myriad of disorders associated with elevated concentrations of inflammatory cytokines, including cerebral palsy.

Effects of lipopolysaccharide on interleukin-6 production in perfused human placental cotyledons.

OBJECTIVE: To determine if lipopolysaccharide (LPS) alters production of interleukin-6 (IL-6) or vascular tone in perfused placental cotyledons. METHODS: Control and study cotyledons from nine placentas were perfused for 3 h. Study cotyledons received LPS in concentrations of 0.01 mcg/ml (n = 3), 0.1 mcg/ml (n = 3), or 1.0 mcg/ml (n = 3). Effluents were collected at 30, 60, 120, and 180 min following infusion with LPS. IL-6 concentrations were measured by enzyme-linked immunosorbant assay. Perfusion pressures were recorded at 10-min intervals. Data were analyzed using ANOVA for repeated measures. RESULTS: IL-6 production significantly increased over time in both the study and control cotyledons (P = 0.002). LPS treatment did not affect IL-6 production (P = 0.85) and there were no observable dose effects (P = 0.13). Perfusion pressures did not differ (P = 0.16). CONCLUSIONS: The isolated perfused placental cotyledon produces IL-6 and concentrations increase over time. LPS does not alter production of IL-6 or fetoplacental vascular tone.

Parental decision-making differences between patients in two healthcare systems for choroid plexus cysts.

We evaluated the medical-sociological implications of parental perception of risk and decision-making choices for prenatally ascertained choroid plexus cysts (CPC) between two obstetric populations. The Wayne State University (WSU) Reproductive Genetics database and the Madigan Army Medical Center (MAMC) experience were reviewed to compare the rates of aneuploidy and invasive testing for cases with CPC. Aneuploidy rates were compared between those with isolated CPC, CPC with advanced maternal age (AMA), and CPC associated with multiple anomalies. In the WSU cohort 186 cases were identified, of whom 27 (15%) declined invasive fetal testing. In the remaining 159 cases, aneuploidy was present in 2/132 (1.5%) isolated CPC, 3/11 (27%) CPC with AMA, and 15/16 (93%) CPC with multiple anomalies. In the MAMC cohort 107 cases were identified, of whom 99 (92%) declined invasive fetal testing. No aneuploidy cases were found in the 3/12 AMA cases or 5/95 non-AMA cases that underwent amniocentesis. The two cases of aneuploidy with isolated CPC cannot be ignored, and provide an estimated attributable risk of at least 0.8%, a higher risk than 38 years of age. However, the parental sociologic context may be as important for decision-making as the genetic-prognostic risk.

Impact of prenatal care with reduced frequency of visits in a residency teaching program.

OBJECTIVE: To determine if decreasing the number of prenatal visits for routine obstetric patients affects pregnancy outcome. STUDY DESIGN: A historical control study was designed to include 734 deliveries from January 1 to December 31, 1991, in women who had prenatal care per American College of Obstetricians and Gynecologists Committee Opinion no. 79, January 1990, guidelines for uncomplicated obstetric care. A prospective study cohort of women with 711 deliveries from January 1 to December 31, 1994, underwent prenatal care with modified guidelines to include: first visit at 6-12 weeks to confirm dating and obtain initial laboratory data, second visit at 16-20 weeks to obtain maternal serum alpha-fetoprotein screening, third visit at 24-28 weeks for 28-week laboratory data, fourth visit at 32 weeks, fifth visit at 36 weeks, sixth visit at 38 weeks, seventh visit at 40 weeks and weekly thereafter. Pregnancy outcomes included estimated fetal weight, gestational age at delivery, preeclampsia, Apgar score at one and five minutes and delivery mode. Neonatal outcomes, including stillbirth rate, preterm delivery rate, intraventricular hemorrhage rate, bronchopulmonary dysplasia and neonatal mortality, were evaluated. RESULTS: There were no statistically significant differences in perinatal or neonatal outcomes with decreased prenatal visits from an average of 12 per pregnancy to 8. CONCLUSION: Prenatal visits can be decreased in a teaching hospital in women with uncomplicated pregnancies from the standard number, 12-14 visits, to an average of 7 or 8 per patient without adverse perinatal outcomes.

Adrenomedullin concentrations in umbilical cord plasma of uncomplicated term pregnancies.

OBJECTIVE: The study's objective was to determine whether there is a difference in the plasma concentration of adrenomedullin, a hypotensive peptide, between arterial and venous umbilical cord blood of uncomplicated gestations with vaginal delivery. STUDY DESIGN: Arterial and venous umbilical cord blood was obtained immediately after vaginal delivery of 44 term infants with uncomplicated antepartum and intrapartum courses. Radioimmunoassay was performed to assess adrenomedullin concentrations in the plasma. The paired t test was used to compare arterial and venous concentrations. Significance was set at P < .05. RESULTS: Mean +/- SE adrenomedullin concentrations were 178.7 +/- 4.7 pg/mL and 190.6 +/- 6.3 pg/mL for arterial and venous cord plasma, respectively. The difference between the 2 concentrations was not significant (11.8 pg/mL, P = .09). CONCLUSION: Arterial and venous umbilical plasma concentrations of adrenomedullin do not differ significantly in uncomplicated gestations terminating with uncomplicated vaginal deliveries. This suggests that in the normal state there is neither net production nor net clearance of adrenomedullin in the placenta.

Intracranial cavernous angioma initially presenting in pregnancy with new-onset seizures.

A case of an intracranial cavernous angioma, which presented with headaches and seizures in a pregnant patient, is described. Diagnosis was established with magnetic resonance imaging. A computer-assisted literature search uncovered no previously reported case of intracranial cavernous angioma initially presenting during pregnancy.

Integration of genetics and ultrasonography in prenatal diagnosis: just looking is not enough.

OBJECTIVE: There has been a gradual shift of the focus of prenatal diagnosis from genetics to ultrasonography. We assessed our primary genetics approach to determine what would be missed without the genetics component. STUDY DESIGN: We evaluated referral indications for patients with normal and abnormal prenatal findings from Jan. 1, 1990, to March 31, 1995, and categorized them according to type of fetal anomalies and genetic abnormalities found. Discordance among initial indication, identified risk factors, and observed abnormalities was assessed. RESULTS: The proportion of patients referred for very-high-risk indications increased over time; 13.5% of all patients (1992 of 14,725) had abnormalities. Abnormal outcomes were categorized as 26% chromosomal, 58% ultrasonographic dysmorphologic features, 11% biochemical or deoxyribonucleic acid disorders, 5% infectious, and 11% other. Of the cases of ultrasonographic dysmorphism (exclusive of the aneuploidies), 3.5% were ultimately determined to be syndromic and 2.5% to be discrepant, that is, having a different abnormality than the referred diagnosis. Including the whole spectrum of disorders seen, half of the abnormalities would not be detectable with even high-quality ultrasonography. CONCLUSION: A large number of abnormal findings were not consistent with initial indication for referral. Correct diagnosis depended on increased acuity provided by genetic pedigree analysis and recognition of syndromes. Diligence in the search for associated anomalies, aneuploidy, pedigree analysis, and syndromic abnormalities remain critical components in the differential diagnosis. The elucidation of unexpected findings suggests the advantages of early counseling and a genetics-based approach combined with tertiary rather than primary ultrasonography with counseling only when anomalies are detected.

Similarity of insulin-dependent diabetics' and non-insulin-dependent diabetics' levels of beta-hCG and unconjugated estriol with controls: no need to adjust as with alpha-fetoprotein.

OBJECTIVE: To determine if the same systematic alteration of serum values seen for alpha-fetoprotein (AFP) in diabetic patients is also seen for beta-hCG and unconjugated estriol (uE3). METHODS: Serum AFP, beta-hCG, and uE3 results were obtained in 18,639 patients for whom complete follow-up information was obtained. Patients were divided into euglycemic (n = 18,088), insulin-requiring diabetics (n = 104), and non-insulin-requiring diabetics (n = 437), as well as by race. RESULTS: The 20% adjustment used for AFP appropriately corrects serum values. No such systematic variation is seen for either beta-hCG or uE3 by glycemic status or race. CONCLUSION: No adjustment for beta-hCG or uE3 is necessary for diabetes for biochemical screening programs.

Aneuploidy among prenatally detected neural tube defects.

We have reported previously a 10% aneuploidy detection rate among 39 cases of fetal neural tube defects (NTD). Subsequently we amassed an additional experience of over 17,000 prenatal diagnosis cases over a 5-year period. During this period 106 cases of NTDs were identified; 44 with anencephaly, 62 with open spina bifida. The average maternal age of this population with NTDs was 29 years (15-40); 6 patients declined amniocentesis. Six of 100 cytogenetic studies were aneuploid; one anencephalic fetus had inherited a maternal marker chromosome, and 5 NTD cases had trisomy 18. The average maternal age of the aneuploid cases was 31 (19-40); 3 were 35 years or older. Four of 5 trisomy 18 cases had multiple congenital anomalies (MCA). The overall aneuploidy detection rate in our cohort was 5-6%, while aneuploidy occurred in 2% of the isolated NTD cases, and 24% of the MCA cases. Combining the earlier experience, 4/39 aneuploidy (2 trisomy 18, 4p+, del 13q) yields an aneuploidy detection frequency of 10/145 (7%), of which most (7/10) had trisomy 18. These data support fetal karyotyping for accurate diagnosis, prognosis, and recurrence-risk counseling.

Correction of hemodynamic abnormalities by vesicoamniotic shunting in familial congenital megacystis.

Vesicoamniotic shunting is typically reserved for treatment of fetal obstructive uropathy. We report a case of congenital megacystis without anatomic urinary obstruction in whom a vesicoamniotic shunt was used to improve fetal hemodynamics as measured by Doppler velocimetry.

Prenatal cytogenetic abnormalities: correlations of structural rearrangements and ultrasonographically detected fetal anomalies.

OBJECTIVE: Our purpose was to determine the distribution of karyotypic abnormalities detected at prenatal diagnosis, fetal anomalies, and ability for fluorescent in situ hybridization detection. STUDY DESIGN: Our cytogenetic database from January 1988 to April 1994 was categorized according to type and potential detection by current standard fluorescent in situ hybridization probes. Fetal anomalies and cytogenetic aberrations were compared. RESULTS: A total of 664 cases of abnormal fetal karyotypes were identified from 12,454 prenatal cytogenetic cases (7529 amniocenteses and 4925 chorionic villus sampling) and were classified as autosomal aneuploidy (331), sex aneuploidy (103), polyploidy (38), marker aneuploidy (19) and structural rearrangements (173). Standard fluorescent in situ hybridization probes would have missed 31% of the abnormal cases: 90 aneuploidy, 14 de novo marker aneuploidy, and 65 de novo structural aberrant cases. The 134 cases of structural chromosomal rearrangements with complete ultrasonographic records were further classified as polymorphism (42), familial (43), or de novo (49). Frequency of fetal anomaly detection by ultrasonography in de novo cases (22/49) was higher than other rearrangements (chi 2 7.4, p = 0.006). CONCLUSION: The contribution of unusual aneuploidies (16%) and structural chromosomal rearrangements (26%) in prenatal diagnostic practice is significant. Fetal anomalies were detected by ultrasonography in 45% of the de novo rearrangement cases. Fluorescent in situ hybridization would miss 31% of the abnormal cases.

Role of ultrasonography in pregnancies with marker chromosome aneuploidy.

The objective of this report was to evaluate the effect of ultrasonographic (US) findings on pregnancy management in patients with marker chromosome (MC) aneuploidy ascertained through prenatal diagnosis. From 1989 through June 1993, 15,522 prenatal diagnostic procedures were performed for accepted indications. Charts of patients with MC on amniocentesis or chorionic villus sampling (CVS) karyotype were evaluated with respect to US anomalies, pregnancy complications, and outcome. Nineteen cases of MC were identified. The prevalence of MC in our study was 0.12% (1:816 procedures). No significant difference between CVS and amniocentesis was found: 5/19 (26%) were CVS specimens, which is comparable to our CVS (3,259/15,522) case distribution. Three cases with incomplete records were excluded from the analysis. Four inherited MC cases were identified: 1 case had anencephaly. Of the 12 de novo MC cases 4 (33%) had abnormal US findings, and an additional 4 were found to have cytogenetic evidence for partial trisomy. Seven of these 8 abnormal de novo MC cases were terminated. MC aneuploidy is more common in pregnancies sampled for usual genetic indications than previously reported in pediatric series. High-resolution US may identify a major malformation not etiologically related to a MC inherited from a normal phenotypic parent. The association of the novo MC with US anomalies confers a poor prognosis, suggesting the expression of genetic imbalance from the accessory chromatin (partial trisomy). However, when US appears normal on initial and follow-up examinations, the chances for a normal-phenotypic newborn are high.

Histopathology of fetal diastrophic dysplasia.

We report on three cases of diastrophic dysplasia in second trimester fetuses and discuss the differential diagnosis and clinical, radiologic, and histopathologic findings. Manifestations of typical diastrophic dysplasia in infants and older patients include abnormal pinnae, scoliosis, and joint contractures; these were absent in the fetuses, in keeping with the tendency for the clinical and radiologic aspects of this disease to become more severe with age. The histopathologic characteristics of the cartilage appear to be similar in the fetus and older patient, and therefore may be useful in differentiating diastrophic dysplasia from other osteochondrodysplasias in the second trimester.

Risk of anomalies as a function of level of elevated maternal serum alpha-fetoprotein.

OBJECTIVE: Most neural tube defects risks are not actual but mathematical extrapolations. We sought to evaluate this risk and to compare actual performance. STUDY DESIGN: This was a retrospective study of a referral population with elevated maternal serum alpha-fetoprotein results between 1987 and 1992. Ultrasonography results, delivery records, and autopsy results were compared with entry levels of maternal serum alpha-fetoprotein, and the percentage of fetal anomalies detected in this study was evaluated. RESULTS: A total of 773 patients with elevated maternal serum alpha-fetoprotein levels were evaluated. There was a progressive increase in the incidence of anomalies as a direct function of the level of the maternal serum AFP, varying from 3.4% at a level of 2.5 to 40.3% at a level > 7.0. CONCLUSION: Data from this study support the correlation of maternal serum AFP levels with the risk of neural tube defect and ventral wall defects.