Elevated amniotic fluid ferritin levels are associated with inflammation-related pregnancy loss following mid-trimester amniocentesis.

OBJECTIVE: Occult infection accounts for up to 12% of pregnancy losses following genetic amniocentesis. Elevated serum and cervical fluid levels of ferritin, an acute-phase reactant, have been associated with spontaneous preterm delivery. We determined the association between amniotic fluid (AF) ferritin levels and post-amniocentesis pregnancy loss. METHODS: We performed a case-control study involving 66 women with a non-anomalous fetus who had a spontaneous pregnancy loss within 30 days following genetic amniocentesis and 66 term controls matched for maternal age, gestational age, time of test and indication for amniocentesis. Amniotic fluid ferritin and interleukin-6 (IL-6) levels were measured using commercially available kits. RESULTS: Mean (+/- SD) AF ferritin levels were similar between the cases (19.3 +/- 21.4 ng/ml) and the controls (19.8 +/- 22.7ng/ml) (p = 0.9). Mean (+/- SD) AF IL-6 levels were significantly higher in the women with post-amniocentesis pregnancy loss (4.0 +/- 13.1 ng/ml) than in controls (0.5 +/- 0.7 ng/ml) (p = 0.04). A significant proportion (12.1%, 8/66) of the women with post-amniocentesis pregnancy loss had elevated amniotic fluid IL-6 levels (> 3 SD, 2.5 ng/ml) indicating inflammation, as compared to none in the control group (p = 0.01). In this subgroup of women with pregnancy loss and elevated IL-6 levels, AF ferritin levels were significantly elevated (52.0 +/- 45.5 ng/ml) compared to the level in women who had a term delivery (19.8 +/- 22.7 ng/ml) (p = 0.002), and were strongly correlated with IL-6 levels among the cases (r = 0.67, p < 0.001). CONCLUSION: The strong correlation of AF ferritin with IL-6 levels, along with the high ferritin values in cases with high AF IL-6, indicates that ferritin is a marker of inflammation in asymptomatic women destined to have an early pregnancy loss.

Can prenatal ultrasound findings predict ambulatory status in fetuses with open spina bifida?

OBJECTIVE: To determine whether prenatal sonographic findings in fetuses with open spina bifida can predict ambulatory potential and the need for postnatal shunt placement. STUDY DESIGN: Ongoing pregnancies complicated by isolated open spina bifida from January 1996 to March 2000 were studied retrospectively. Static images and reports generated every 3-4 weeks from diagnosis until delivery were reviewed for lesion level and type, ventricular width, and lower extremity appearance. Operative summaries as well as neonatal and pediatric charts were reviewed. Ambulatory was defined in infants > or =2 years old as walking with or without appliances. In those <2 years of age, ambulatory was defined as at least 4/5 lower extremity muscle strength. RESULTS: Thirty-three cases of isolated open spina bifida were identified. Lower (more caudal) lesion levels and smaller ventricular size were associated with ambulatory status in univariate analyses (P <.001, P =.003, respectively). No infant with a thoracic lesion was ambulatory (n = 11); all had ventriculomegaly diagnosed prenatally and all required shunt placement. In contrast, all infants with L4-sacral lesions (n = 10) were ambulatory, and 60% had ventriculomegaly diagnosed prenatally. Of patients with L1-L3 lesions (n = 12), 50% were ambulatory. In this group, ambulatory potential could not be determined by the presence of ventriculomegaly, ventricular size, or the presence of club foot. In the entire cohort, no infant with a myeloschisis was ambulatory, and all infants except one with a sacral lesion required postnatal shunt placement. CONCLUSIONS: Sonographic determination of lesion level and type is useful in predicting the ambulatory potential of fetuses with open spina bifida.

Association of the C677T methylenetetrahydrofolate reductase mutation and elevated homocysteine levels with congenital cardiac malformations.

OBJECTIVE: The aim of this study was determine whether the cytosine-to-thymine mutation at base 677 of the gene for methylenetetrahydrofolate reductase (C677T MTHFR ), which has been associated with neural tube defects, is also associated with congenital cardiac malformations. STUDY DESIGN: Amniotic fluid homocysteine levels were measured and the presence or absence of the C677T MTHFR mutation in amniocytes was determined in stored amniotic fluid obtained from 26 pregnancies complicated by isolated (presumed multifactorial) fetal cardiac defects and from 116 normal pregnancies. RESULTS: The pregnancies affected by fetal cardiac defects had higher amniotic fluid homocysteine levels (1.7 +/- 1.7 vs 1.0 +/- 0.7 micromol/L; P =.07) and included more samples with homocysteine levels >90th percentile (27% vs 9%; P =.02) and more cases with the C677T MTHFR mutation (35% vs 13%; P =.01). Fifty percent of cases had either a high homocysteine level or the C677T MTHFR mutation (50% vs 20%; P =.003) and 12% had both (12% vs 0%; P =.0006). CONCLUSION: Fifty percent of these isolated congenital cardiac defects were associated with either the C677T MTHFR mutation or elevated amniotic fluid homocysteine levels, or both. This finding adds to what is already known about the multiple and complex biochemical and developmental functions of the homocysteine pathway.

Ultrasound findings and multiple marker screening in trisomy 18.

OBJECTIVE: To compare detection of trisomy 18 in the second trimester by ultrasound and multiple-marker testing. METHODS: A computerized genetics database was used to identify fetuses of 14-22 weeks' gestation who had comprehensive ultrasound examinations, multiple-marker screening tests (alpha-fetoprotein [AFP]), hCG, unconjugated estriol [E3], and trisomy 18 karyotype. A positive trisomy 18 screen was defined as AFP up to 0.75 multiples of the median (MoM), hCG up to 0.55 MoM, and unconjugated E3 up to 0.60 MoM. A risk of at least 1:190 defined a positive Down syndrome screen. Ultrasound abnormalities were diagnosed prospectively and were confirmed later by retrospective review of sonographic images. RESULTS: From 1988-1997, 30 trisomy 18 fetuses who had comprehensive ultrasounds and multiple-marker testing were identified. Twenty-one (70%) had abnormalities detected by ultrasound, of which the most common isolated finding was choroid plexus cyst. Eleven fetuses (37%) had positive trisomy 18 screens, and two had positive Down syndrome screens, for a total of 13 of 30 (43%) fetuses with positive multiple-marker screening tests. CONCLUSION: We found that ultrasound was more likely to be abnormal than multiple-marker screening tests in fetuses with trisomy 18 (70%) (95% confidence interval [CI] 54, 86 versus 43% CI 25, 61). However, combining the two testing methods yielded the highest detection rate (80% [CI 66%, 94%]).

Amniotic fluid homocysteine levels, 5,10-methylenetetrahydrafolate reductase genotypes, and neural tube closure sites.

A specific gene mutation leading to altered homocysteine metabolism has been identified in parents and fetuses with neural tube defects (NTDs). In addition, current animal and human data indicate that spine closure occurs simultaneously in five separate sites that then fuse. We sought to determine whether either this mutation or abnormal amniotic fluid homocysteine levels are associated with all five neural tube closure sites. We retrieved stored amniotic fluid from cases of isolated fetal neural tube defect diagnosed from 1988 to 1998 (n = 80) and from normal controls matched for race, month and year of amniocentesis, and maternal age. Cases were categorized according to defect site by using all available medical records. The presence or absence of the 677C-->T mutation of 5, 10-methylenetetrahydrafolate reductase (MTHFR) gene was determined, and homocysteine levels were measured; case and controls were compared. Significantly more cases than controls were heterozygous or homozygous for the 677C-->T MTHFR mutation (44% vs. 17%, P < or = 0. 001). Likewise, cases were significantly more likely than controls to have amniotic fluid homocysteine levels >90th centile (>1.85 micromol/L), 27% vs. 10%, P = 0.02. Most (83%) of control cases had both normal MTHFR alleles and normal amniotic fluid homocysteine levels (normal/normal), whereas only 56% of NTD case were normal/normal (P = 0.001). When evaluated by defect site, only defects involving the cervical-lumbar spine, lumbosacral spine, and occipital encephalocele were significantly less likely to be normal/normal than controls (P = 0.007, 0.0003, and 0.007, respectively), suggesting a strong association with the 677C-->T allele. In contrast, anencephaly, exencephaly, and defects confined to the sacrum included many cases that had both normal MTHFR alleles and normal homocysteine and were not significantly different from controls. The 677C-->T MTHFR mutation and elevated homocysteine levels appear to be disproportionately associated with defects spanning the cervical-lumbar spine, lumbosacral spine, and occipital encephalocele. In contrast, anencephaly, exencephaly, and defects confined to the sacrum may not be related to altered homocysteine metabolism.

Role of amniotic fluid homocysteine level and of fetal 5, 10-methylenetetrahydrafolate reductase genotype in the etiology of neural tube defects.

A mutation in the gene 5,10-methylenetetrahydrofolate reductase (MTHFR), leading to altered homocysteine metabolism, has been identified in parents and fetuses with fetal neural tube defects. We sought to determine which is of greater importance in fetal neural tube defect formation: the fetal MTHFR mutation or elevated amniotic fluid homocysteine level. We retrieved stored amniotic fluid from cases of isolated fetal neural tube defect diagnosed from 1988 to 1998 (n = 80), and from normal controls matched for race, month and year of amniocentesis, and maternal age. The presence or absence of the 677C-->T mutation of MTHFR was determined and homocysteine levels were measured; cases and controls were compared. Significantly more cases than controls were heterozygous or homozygous for the 677C-->T MTHFR mutation (44% vs 17%, P < or = 0.001). Cases were also significantly more likely than controls to have an amniotic fluid homocysteine level above the 90th centile (>1.85 micromol per liter); 27% vs 10%, P = 0.02. Thirty one cases and 12 controls had an abnormal genotype; however, amniotic fluid homocysteine levels were not significantly different between these two groups (6/31, or 19% of cases had an elevated homocysteine compared to 1/12, or 8% of controls; P = 0.65). In contrast, 40 cases and 60 controls had a normal genotype; the neural tube defect cases had significantly higher homocysteine levels (13/40, or 32% of cases had an elevated homocysteine level compared to only 6/60, or 10% of controls; P = 0.008). Although both abnormal fetal MTHFR genotype and abnormal amniotic fluid homocysteine concentration are significantly associated with neural tube defects, the association with amniotic fluid homocysteine concentration is significant regardless of the fetal MTHFR genotype. The relationship between maternal and fetal homocysteine metabolism is complex.

Hydramnios prediction of adverse perinatal outcome.

OBJECTIVE: To determine whether hydramnios is associated with an increased risk of adverse perinatal outcomes. METHODS: Computerized records of all ultrasound examinations done at the University of Alabama at Birmingham from 1986 to 1996 (n = 40,065) were reviewed to identify 370 women with singleton pregnancies beyond 20 weeks' gestation and hydramnios diagnosed sonographically by amniotic fluid index of 25 cm or more, largest vertical pocket of 8 cm or more, or subjective impression. Controls were all women with singleton gestations with normal amniotic fluid volumes (n = 36,426). Obstetric outcomes were determined by cross-reference to our database. Cases with hydramnios were compared with controls for perinatal death, anomaly rate, fetal growth restriction (FGR), cesarean delivery, fetal aneuploidy, and maternal diabetes. Cases were sorted according to diabetes status, after which perinatal death, anomaly rate, FGR, cesarean delivery, and fetal aneuploidy were compared again. RESULTS: The incidence of hydramnios was 1%. The perinatal mortality rate in all women with hydramnios was 49 per 1000 births, compared with 14 per 1000 births in the control group (P < .001). Women with hydramnios had 25 times more anomalies than controls (8.4% versus 0.3%; P < .001), although the prevalence of fetal aneuploidy was not significantly different (one in 370 versus one in 3643; P = .10). The cesarean rate was three times higher in women with hydramnios compared with controls (47.0% versus 16.4%; P < .001). When hydramnios cases were divided according to diabetes status, all of the increased risk was in nondiabetic women: Perinatal mortality was 60 per 1000 in nondiabetic women versus 0 per 1000 in diabetic women (P = .03); the anomaly rate was 10.4% versus 0%, respectively (P = .005). CONCLUSION: Hydramnios indicated an increased risk of adverse perinatal outcomes, especially if not associated with diabetes. A comprehensive fetal evaluation, a workup to rule out maternal factors, and fetal surveillance are warranted; amniocentesis for fetal karyotype analysis might not be necessary.

Value of perinatal autopsy.

OBJECTIVE: To determine how often a perinatal autopsy identified the cause of death, and how frequently this information changed recurrence risk estimates or altered parental counseling. METHODS: We reviewed all autopsies of fetal stillbirths and briefly viable neonates performed by one perinatal pathologist at the University of Alabama Hospital from 1992 to 1994. RESULTS: Four hundred sixteen fetal and early neonatal deaths occurred at our hospital from January 1, 1992, to June 1, 1994. Consent for an autopsy examination was granted for 139 of these (33%), and all autopsies were performed by a single perinatal pathologist. Abnormalities likely to be the cause of death were identified in 130 of 139 cases (94%). Ninety-one subjects did not have structural anomalies: In 14 cases autopsy revealed previously unsuspected pathology that altered parental counseling; in 68 cases autopsy findings were consistent with the clinical obstetrical diagnosis; and in nine cases the cause of death could not be identified. Forty-eight subjects were anomalous. Thirty-seven of these (79%) had been evaluated by antenatal ultrasonography, and autopsy identified additional abnormalities in 51% (19 of 37). In the 11 deaths evaluated neonatally, a previously unsuspected diagnosis likely to be the cause of death was identified in three. Overall, autopsy findings changed recurrence risk estimates and/or parental counseling in 36 of 139 cases (26%). CONCLUSION: The cause of fetal or perinatal death was determined by autopsy in 94% of cases in our series. Counseling and recurrence risk estimates were altered by autopsy findings in 26%.

A five-year experience with fragile X screening of high-risk gravid women.

OBJECTIVE: We sought to compare our 5-year program of fragile X screening of high-risk gravid women with our program of fragile X testing of affected individuals (probands). STUDY DESIGN: All women referred to the prenatal genetics clinic from 1994 to 1998 who had a family history of unspecified mental retardation or learning or behavioral disorders (known fragile X families excluded) were offered fragile X screening. Results were compared with those of probands with the same diagnoses who underwent fragile X testing during the same time period. RESULTS: We counseled 12,349 prenatal patients from 1994-1998, of whom 263 (2.1%) had a positive family history and underwent fragile X screening. No mutations or premutations were identified. In contrast, 31 (1.9%) of 1637 affected probands who underwent fragile X testing during the same time period had positive results, which was a significant difference (0/263 vs 31/1637; P <.05). CONCLUSIONS: Testing the affected proband is superior to screening the pregnant relative of the proband for identification of families at risk for fragile X syndrome.

Prospective evaluation of free beta-subunit of human chorionic gonadotropin and dimeric inhibin A for aneuploidy detection.

OBJECTIVE: Our goal was to prospectively evaluate the use of the free beta-subunit of human chorionic gonadotropin and dimeric inhibin A for the detection of fetal Down syndrome and other aneuploidies. STUDY DESIGN: Women who had a second-trimester multiple-marker screening test (alpha-fetoprotein, unconjugated estriol, human chorionic gonadotropin) and genetic amniocentesis from August 1996 to August 1998 were included. Serum was also analyzed for inhibin and the free beta-subunit of human chorionic gonadotropin. Detection and false-positive rates for 4 analyte combinations at 5 different screening risk cutoff points for Down syndrome were determined and compared. RESULTS: We evaluated 1256 patients, including 23 with aneuploidy (13 with Down syndrome, 10 others). The maternal age was 35.9 +/- 4.6 years (mean +/- SD). At the optimal risk cutoff point for Down syndrome detection (1:190; false-positive rate, 19%), the multiple-marker screening test plus inhibin was superior, detecting 85% of Down syndrome cases, in comparison with 69% when the multiple-marker screening test alone was used and 62% when the other 2 combinations were used. The multiple-marker screening test plus inhibin also detected 60% of the other aneuploidies. CONCLUSIONS: When evaluated prospectively in a high-risk population, the multiple-marker screening test plus inhibin was superior to the traditional multiple-marker screening test and 2 other analyte combinations, with a lower false-positive rate and increased detection of all aneuploidies in a high-risk population.

Terbutaline pump maintenance therapy for prevention of preterm delivery: a double-blind trial.

OBJECTIVE: This study's aim was to determine whether maintenance therapy with terbutaline administered by pump prolongs gestation in women after treatment with intravenous magnesium sulfate tocolysis for suspected preterm labor. STUDY DESIGN: Consenting women with a singleton gestation and intact membranes who had uterine contractions and >1 cm cervical dilation, 80% effacement, or progressive cervical change and whose contractions were successfully arrested with intravenous magnesium were randomly assigned to receive either terbutaline or normal saline solution placebo by subcutaneous infusion pump. Pump therapy was administered with a standardized protocol. Pump therapy was discontinued and parenteral magnesium was resumed if recurrent preterm labor developed while women were on the therapeutic regimen at <34 weeks' gestation and no contraindication for tocolysis existed. If recurrent labor was arrested, pump therapy was restarted according to the original treatment group. A sample size of 48 women was required to detect a 2-week intergroup difference in mean time to delivery. Analyses were based on intent to treat. RESULTS: Fifty-two women received terbutaline (n = 24) or placebo (n = 28). At random assignment the groups were similar with respect to age, race, parity, previous preterm delivery, gestational age, and cervical examination. Overall there was a 1-day difference in mean time to delivery between the groups (terbutaline 29 +/- 22 days and placebo 28 +/- 23 days, P = .78). There were no differences in the rates of preterm delivery at <34 and <37 weeks' gestation. Neonatal outcomes were similar. CONCLUSIONS: Maintenance terbutaline therapy administered by pump does not prolong gestation in women successfully treated for suspected preterm labor.

Maternal serum alpha-fetoprotein and dimeric inhibin A detect aneuploidies other than Down syndrome.

OBJECTIVE: Our purpose was to determine whether the combination of maternal serum alpha-fetoprotein, free human chorionic gonadotropin-beta, dimeric inhibin A, and maternal age detects aneuploidies other than Down syndrome. STUDY DESIGN: We retrieved stored serum from pregnancies complicated by aneuploidies other than Down syndrome from 1988 to 1997 (n = 55, mean maternal age 35.2 +/- 5.6 years). Alpha-fetoprotein levels were obtained from our database, and free human chorionic gonadotropin-beta and dimeric inhibin A levels were measured in the thawed serum with use of commercial assays. Analyte values were used in both 3-analyte and 2-analyte multiple-marker screening tests; detection rates were determined at several different Down syndrome risk-positive cutoff values. RESULTS: In the 3-analyte test 58% (32/55) of all aneuploidies were detected with use of both the Down syndrome protocol at a screen-positive risk cutoff value of 1:300 (false-positive rate 17%) and a novel trisomy 18 screening algorithm. However, 67% (37/55) detection was obtained with use of the 2-analyte combination of alpha-fetoprotein and dimeric inhibin A, with both the Down syndrome protocol (screen positive cutoff value 1:300) and the trisomy 18 algorithm: 12 of 13 trisomy 18 (92%), 9 of 17 Turner's syndrome (53%), 10 of 17 other sex chromosome aneuploidies (59%), 1 of 1 trisomy 22 (100%), and 5 of 7 trisomy 13 (71%). CONCLUSIONS: The combination of maternal serum alpha-fetoprotein, dimeric inhibin A, and maternal age detects autosomal trisomies other than Down syndrome at a rate superior to that of the traditional analyte combination.

Intrauterine viral infection at the time of second trimester genetic amniocentesis.

OBJECTIVE: To determine whether preexisting intrauterine viral infection is associated with postamniocentesis pregnancy loss. METHODS: We accessed our bank of second-trimester amniotic fluid (AF) samples obtained aseptically and stored at -20C from all 11,971 women who underwent genetic amniocentesis between 1988 and 1995. Samples were retrieved from every case of spontaneous pregnancy loss within 30 days of the amniocentesis (excluding aneuploidy and anomalies, n = 66). Sixty-six control samples were randomly chosen from subjects who delivered at term and were matched for year of test, gestational age, maternal age, and indication for amniocentesis. Investigators were blinded to the status of the samples, which were studied by polymerase chain reaction (PCR) for the presence of adenovirus, parvovirus, cytomegalovirus, Epstein-Barr virus, herpes simplex virus, enterovirus, influenza A virus, and beta-actin DNA. Results were compared with interleukin-6 (IL-6) levels previously measured by enzyme-linked immunosorbent assay in the same samples. RESULTS: Sixty-two study cases and 60 controls were sufficient for all PCR studies. Fourteen AF samples contained a single virus: five (8%) of 62 study cases and nine (15%) of 60 controls (P = .27). Adenovirus accounted for nine (64%) of 14 viruses identified: four of 62 cases and five of 60 controls (P = .74). Cytomegalovirus was not identified in any study cases but was found in three controls. The mean IL-6 levels in samples with and without virus were not significantly different (4.8+/-15.9 ng/mL with virus compared with 2.0+/-8.8 ng/mL without virus; P = .53). CONCLUSION: Presence of virus in second-trimester AF is not significantly associated with elevated IL-6 levels or with early postamniocentesis pregnancy loss.

Elevated second-trimester amniotic fluid interleukin-6 levels predict preterm delivery.

OBJECTIVE: Our purpose was to determine whether early second-trimester amniotic fluid interleukin-6 levels predict delivery before 34 weeks' gestation. STUDY DESIGN: We used stored second-trimester amniotic fluid samples obtained from women undergoing genetic amniocentesis from 1988 to 1996. Interleukin-6 levels were measured by enzyme-linked immunosorbent assay in samples from every case known to result in delivery from 20 to 34 weeks' gestation (n = 290), and 290 matched controls delivering at > or =37 weeks. Fetal aneuploidies, anomalies, and all cases delivering within 30 days of the amniocentesis (which were thought to be possibly procedure related) were excluded. RESULTS: Interleukin-6 levels were higher in cases than controls (1.9 +/- 5.2 vs 1.0 +/- 2.4 ng/ml, p = 0.004). Cases were grouped according to whether the preterm delivery was indicated or spontaneous: The mean interleukin-6 levels were significantly higher than controls in the spontaneous group (1.6 +/- 3.2 vs 0.8 +/- 1.2 ng/ml, p = 0.01) but not in the indicated group (1.4 +/- 4.0 vs 0.8 +/- 1.2 ng/ml, p = 0.12). In all samples the interleukin-6 level was negatively correlated with the gestational age at delivery (R = -0.11633, p = 0.007). CONCLUSION: Elevated early second-trimester amniotic fluid interleukin-6 levels are associated with preterm delivery, confirming that in some women this indicator of very early intrauterine inflammation predicts birth before 34 weeks' gestation.

Insulin-like growth factor binding protein-3 in the detection of fetal Down syndrome pregnancies.

OBJECTIVE: To study the usefulness of maternal serum insulin-like growth factor binding protein-3, a potential cell growth inhibitor, in second trimester prenatal screening for fetal Down syndrome. METHODS: Three hundred and forty-two samples from normal pregnancies and nine fetal Down syndrome pregnancies were analyzed for insulin-like growth factor binding protein-3 levels by radioimmunoassay. Data were converted to multiples of median (MoM) and analyzed statistically to compare the differences between control and Down syndrome pregnancies. RESULTS: The mean insulin-like growth factor binding protein-3 MoM of Down syndrome-affected pregnancies (1.09) was significantly higher than that of the normal pregnancies (1.00) (P < .01). Insulin-like growth factor binding protein-3, in combination with maternal serum alpha-fetoprotein (MSAFP), hCG, and maternal age, detected 89% of Down syndrome pregnancies at a screen positive rate of 2.1%. This compares favorably to the standard combination of MSAFP, hCG, and unconjugated estriol (E3), which had a 66.7% Down syndrome detection rate and a 4.1% screen positive rate in our study samples. CONCLUSION: This retrospective analysis suggested that the inclusion of insulin-like growth factor binding protein-3 into the triple screen program to replace unconjugated E3 might enhance the detection rate of fetal Down syndrome pregnancies. These data need to be confirmed by a larger prospective study.

Significance of a false-positive trisomy 18 multiple-marker screening test.

OBJECTIVE: To determine if a false-positive trisomy 18 multiple-marker screening test (all three analytes low: maternal serum alpha-fetoprotein [AFP] at most 0.75 multiples of the median [MoM], unconjugated estriol at most 0.60 MoM, and hCG at most 0.55 MoM) indicates increased risk for obstetric complications or is related to maternal weight. METHODS: We accessed our genetic database to obtain multiple-marker screening test results, fetal karyotypes, and pregnancy outcomes from all patients with a normal multiple-marker screening test (n = 3900) and from all patients with a positive trisomy 18 screening test (n = 103) seen in the prenatal diagnosis clinic from 1992 to 1996. During this period, only maternal serum AFP was adjusted for maternal weight. RESULTS: A positive trisomy 18 screen identified five of 12 trisomy 18 fetuses. Women with a false-positive trisomy 18 screen were heavier (175.6 +/- 43.8 lb versus 159.9 +/- 37.9 lb, P < .001) and younger (29.7 +/- 6.5 years versus 32.3 +/- 6.5 years, P < .001) than women with a normal multiple-marker screening test, but were not at increased risk for pregnancy complications. Weight-adjusting all three analytes reduced the false-positive trisomy 18 screen rate by 42% (from 1.9% to 1.1%) but did not change the trisomy 18 detection rate. CONCLUSION: A false-positive trisomy 18 screening test does not indicate increased risk to develop pregnancy complications and may be related to inadequate correction for increased maternal weight.

Alpha-fetoprotein, free beta-human chorionic gonadotropin, and dimeric inhibin A produce the best results in a three-analyte, multiple-marker screening test for fetal Down syndrome.

OBJECTIVE: The purpose of this study was to determine, among six second-trimester maternal serum analytes, the best three-analyte combination for fetal Down syndrome detection. STUDY DESIGN: With use of commercially available assay kits, medians for free beta-human chorionic gonadotropin, CA 125, and dimeric inhibin A were established in stored sera from 45 to 50 euploid pregnancies at each week of gestation from 14 to 22 weeks and from 33 Down syndrome pregnancies. Maternal serum alpha-fetoprotein, unconjugated estriol, and intact human chorionic gonadotropin levels measured in each sample before storage were retrieved. All 20 possible three-analyte combinations were evaluated in the multiple-marker screening test for Down syndrome. RESULTS: The mean maternal age of the study population was 35.6 +/- 5.3 years. The best three-analyte combination was maternal serum alpha-fetoprotein, free beta-human chorionic gonadotropin, and dimeric inhibin A: 97% of Down syndrome cases were detected at a false-positive rate of 16%. At a slightly higher false-positive rate (18%) maternal serum alpha-fetoprotein, estriol, and intact human chorionic gonadotropin detected only 79% of cases. CONCLUSIONS: Of six second-trimester maternal serum analytes, the best three-analyte combination for fetal Down syndrome detection was maternal serum alpha-fetoprotein, free beta-human chorionic gonadotropin, and dimeric inhibin A. This retrospective analysis should now be confirmed prospectively.