Search

Objectives To confirm that measuring pregnancy-associated plasma protein-A (PAPP-A) in both first- and second-trimester serum samples improves Down syndrome screening. Methods We selected paired first- and second-trimester stored serum samples from 34 Down syndrome pregnancies (cases) and 514 unaffected pregnancies (controls) and tested the second-trimester samples for PAPP-A and dimeric inhibin-A (DIA). First-trimester PAPP-A measurements were already available, as were second-trimester measurements of alpha-fetoprotein, unconjugated estriol (uE3), and human chorionic gonadotrophin (hCG). Results PAPP-A was lower among cases than controls (0.47 MoM) in the first trimester (at an average of 12.5 weeks); in the second trimester, it was not different (0.91 MoM). Using repeated measures of PAPP-A alone, 21 of 34 cases were detected (62%, 95%CI 44% to 78%) with 5% false positives. At an observed 2% false-positive rate, the detection rates (DR) for the quadruple (69%) and serum integrated (69%) tests were lower than for the repeated measures test (75%). Modelled performance at 12 weeks was similar to these observed findings (70, 75, and 82%, respectively). If the first-trimester samples were collected at 10 weeks, however, DR would be higher (70, 81, and 91%, respectively). Conclusions Adding a repeated measure of PAPP-A to existing serum markers improves Down syndrome screening to levels that are currently obtainable only by including ultrasound measurement of nuchal translucency (NT). Serum-based screening has the advantages of higher availability and reliability at a lower cost, resulting in a more effective screening strategy. A serum-based repeated measures test has a place in routine Down syndrome screening. Copyright © 2006 John Wiley & Sons, Ltd.

Published studies have reached varying conclusions as to the benefit of replacing human chorionic gonadotropin (hCG) measurements with the free β-subunit of hCG (the free β-subunit) for Down syndrome screening. One study reports 14 per cent higher detection for the free β-subunit, while another finds an actual loss in detection. To explore this issue further, we directly compared the screening performance of hCG and the free β-subunit, alone and in combination with other serum markers, using banked sera obtained prior to amniocentesis and karyotyping. Altogether, 52 Down syndrome and 5065 unaffected pregnancies were studied. Sera were thawed and assayed for hCG and the free β-subunit over 1 year. At a 5 per cent false-positive rate, the detection rate for hCG in combination with maternal age and alpha-fetoprotein was higher than when the free β-subunit was substituted (62 versus 57 per cent). Ultrasound dating and adding unconjugated oestriol both increased detection. The present findings, along with those from six case–control studies (our re-analysis), indicate that the screening performances of hCG and the free β-subunit are similar (median change in detection 0, range −8 to +3 per cent). Under optimal sample collection and transportation conditions, laboratories can expect to achieve similar screening performance using either hCG or the free β-subunit measurements. © 1998 John Wiley & Sons, Ltd.

This study examines the relationship between maternal weight and serum levels of alpha-fetoprotein, unconjugated oestriol, and human chorionic gonadotropin in a population of 47,585 women being provided with prenatal screening for Down's syndrome and open neural tube defects. The study population contains sufficient numbers of women at the extremes of weight to allow the determination that a reciprocal-linear equation more accurately describes the weight relationship for two of the three analytes than the currently used log-linear equations. The reciprocal-linear equations, while more appropriate, provide only a minimal advantage over the log-linear equations. A more important finding is that published weight equations may not be optimal for some screening programmes, due to differences in the mean weight of the populations being tested. Screening programmes are encouraged to calculate their own weight correction formulae, based on data from their own population, and to monitor the mean maternal weight to detect when modifications in the weight correction formulae might be indicated.

This study documents the screening performance, in practice, of a published protocol for interpreting second-trimester Down syndrome risk in twin pregnancies, using maternal serum biochemical markers. Within a cohort of 35,150 pregnancies, 410 twin pregnancies were identified. The rate of twinning was positively associated with maternal age. Of the 274 twin pregnancies known prior to screening, 15 (5.5 per cent) were classified as being screen-positive for Down syndrome. When maternal age and dating method were taken into account, the screen-positive rates in twin and singleton pregnancies did not differ significantly. Nine of the 14 screen-positive women with viable twin pregnancies chose amniocentesis [64 per cent, 95 per cent confidence interval (CI) 35-87]. No cases of Down syndrome were identified. Based on modelling, an estimated 73 per cent of monozygotic twin pregnancies and 43 per cent of dizygotic twin pregnancies with Down syndrome would be identified at a 5 per cent false-positive rate. The overall detection rate would be about 53 per cent. When laboratories offer multiple marker screening to women with twin pregnancies, the false-positive rate ought to be similar to that found in singleton pregnancies; the detection rate is, however, likely to be lower for dizygotic twins.

Nine centres collaborated to examine the feasibility of a screening method for trisomy 18 that was based on assigning individual risk, using a combination of maternal age and measurements of alpha-fetoprotein (AFP), unconjugated oestriol (uE3), and human chorionic gonadotropin (hCG). Second-trimester measurements of these analytes were obtained from 94 trisomy 18 pregnancies. In the 89 pregnancies without an associated open defect, the median levels for AFP, uE3, and hCG were 0.65, 0.43 and 0.36 multiples of the unaffected population median, respectively. The strongest individual predictor of risk for trisomy 18 was uE3, followed by hCG, AFP, and maternal age, in that order. Using a method of individual risk estimation that is based on the three markers and maternal age, 60 per cent of pregnancies associated with trisomy 18 would be detected at a risk cut-off level of 1:100, with a false-positive rate of about 0.2 per cent. One in nine pregnancies identified as being at increased risk for trisomy 18 would be expected to have an affected pregnancy. This risk-based screening method is more efficient than an existing method that is based on fixed analyte cut-off levels. Even though the birth prevalence of trisomy 18 is low, prenatal screening can be justified when performed in conjunction with Down syndrome screening and when a high proportion of women offered amniocentesis have an affected fetus.

Objectives Document patient choices and screening performance (false positive and detection rates) when three improved Down syndrome screening protocols were introduced coincidentally. Method Second-trimester ‘triple marker’ screening was expanded by adding second-trimester dimeric inhibin-A (four-marker), with or without first-trimester pregnancy-associated plasma protein-A (five-marker). Nuchal translucency (NT) measurements were included when available from accredited sonographers (six-marker). For assigning risk, two sets of marker distribution parameters were evaluated. Results Over 3.5 years, 8571 women enrolled (median age 30.6 years). Uptake of the four-, five- and six-marker protocols was 18%, 46% and 36%, respectively. Of those selecting an integrated test (five or six markers), 9.7% did not provide the second trimester serum sample. False positive rates decreased with added markers (5.2%, 5.1% and 2.5%, respectively) and varied between the two parameter sets, while detection remained high. Overall, 21 of 23 cases were detected (91%, 95% CI 73–98%) at a 4.2% false positive rate (95% CI 3.3–5.1%). Conclusions Integrated screening protocols were chosen 4.6 times more often than four-marker screening (82% vs. 18% uptake). Overall detection was higher and false positives lower, consistent with recent guidelines. Important performance factors include gestational dating method, risk cut-off, and the parameter set used to assign risk. Copyright © 2010 John Wiley & Sons, Ltd.

Two prenatal centres in New England, routinely using a screening protocol for fetal Down syndrome that included maternal serum alpha-fetoprotein (AFP), unconjugated oestriol (uE3), and human chorionic gonadotropin (hCG) measurements in combination with maternal age, adopted a separate screening protocol for trisomy 18. That protocol identified a pregnancy as being at high risk when AFP, uE3, and hCG measurements all fell at or below specified cut-offs (0.75, 0.60, and 0.55 multiples of the median, respectively), regardless of maternal age. Among the first 19,491 women screened, 98 (0.5 per cent) were found to have values which placed them in the high-risk category. Four of these women were subsequently found not to be pregnant. In two others, samples from non-pregnant individuals were found to have been incorrectly submitted for analysis in place of the samples from the pregnant women. All of the remaining 92 women were counselled and offered amniocentesis and fetal karyotyping. Eighty-eight (96 per cent) accepted. Karyotypes or birth outcomes were available on all 92 pregnancies. Six cases of trisomy 18 and one case of Turner syndrome were identified by karyotype. One case of trisomy 18 was identified for every 14 unaffected pregnancies offered amniocentesis. In the present prospective study, an estimated 85 per cent of the cases of trisomy 18 were identified. However, given the small number of cases (six), the 95 per cent confidence interval for the detection rate is broad (40-95 per cent).

Objective Cell-free DNA is a marker of cellular apoptosis and necrosis. We wished to determine if maternal smoking affects maternal and fetal serum cell-free DNA levels. Methods Case–control sets of stored second-trimester serum-screening samples from 27 smoking and 90 nonsmoking pregnant women were developed. Smoking status was confirmed by measuring serum cotinine levels. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and DYS1 levels were determined using real-time polymerase chain reaction (PCR) to measure total and fetal cell-free DNA, respectively. At delivery, medical records were reviewed to confirm gender and determine other factors that could affect DNA values. Results Smoking was associated with significantly elevated GAPDH levels compared with nonsmokers (median: 97 662 genome equivalents (GE)/mL vs 38 217 GE/mL; p = 0.018). DYS1 levels were not statistically significantly elevated in smokers (p = 0.29). Other factors that affected DYS1 levels included maternal age in nonsmokers only (r2 = 0.30, p = 0.013) and maternal Synthroid use (p = 0.0045) Conclusion Pregnant smokers have threefold higher levels of total cell-free DNA compared with pregnant nonsmokers. Maternal age and Synthroid exposure may also affect circulating cell-free fetal DNA levels. Copyright © 2008 John Wiley & Sons, Ltd.

Two series of pregnancies were studied to investigate the relationship between maternal smoking and the risk of fetal Down's syndrome. In the first series, ascertained in the 1960s, in which smoking habits were determined after the outcome of pregnancy was known, the proportion of smokers (47 per cent) among the 461 women whose pregnancies ended in the birth of an infant with Down's syndrome was similar to that in the 461 controls (46 per cent) who had pregnancies affected by other congenital disorders. In the second series, ascertained between 1973 and 1984, smoking habits were determined by measurement of cotinine in antenatal serum samples that were routinely collected and stored or, if a serum sample was not available, from information in the antenatal notes. In this series, the proportion of smokers (14 per cent) among the 91 women who had pregnancies associated with Down's syndrome was lower than that among 413 controls (19 per cent), though this was not statistically significant. Collectively, our results provide no evidence for an association between fetal Down's syndrome and smoking. Other published studies found a deficit of smokers among women who had pregnancies associated with Down's syndrome. This may be partly due to some studies not taking adequate account of maternal age (older women are more likely to have had a Down's syndrome pregnancy but are less likely to be smokers) and partly due to the greater tendency for positive findings to be published than negative ones.

To investigate whether statistical parameters used in Down syndrome screening between 15 and 22 weeks of pregnancy can be used at 14 weeks, we assayed alpha-fetoprotein (AFP), unconjugated oestriol (uE3), total human chorionic gonadotrophin (hCG), free alpha-hCG, free beta-hCG, and inhibin-A in 16 pregnancies with Down syndrome in the 14th week of pregnancy and expressed values in multiples of the normal median. The median and standard deviation values for these 16 pregnancies were not materially different from those published for 15-22 weeks. It is reasonable, therefore, to offer Down syndrome screening using these markers starting at 14 completed weeks of pregnancy instead of 15 weeks. It needs to be recognized, however, that serum AFP measurement for neural tube defect screening is less effective at this time than between 16 and 18 weeks of pregnancy.

Objectives Integrated serum screening for Down syndrome is potentially more effective than current second-trimester screening. We report results of an intervention trial of integrated serum screening that involved 229 primary prenatal care practitioners throughout Maine. Methods Women provided a first-trimester serum (for PAPP-A) followed by a second-trimester serum (for AFP, uE3, hCG, and DIA). These five marker measurements were used to calculate a Down syndrome risk in the second trimester. Screen-positive women (risk ≥ 1:100) were managed according to standard practice. Results During 24 months' enrollment, 11 159 women provided a first-trimester sample (61% of women receiving screening services). Nine thousand seven hundred twenty-three women also provided a second-trimester sample; 8773 women satisfied gestational age criteria for testing in both trimesters. Integrated serum screening detected 14 of 16 Down syndrome cases (87%) and 79% after adjustment for trimester-of-ascertainment bias. The initial false-positive rate was 3.2% and was 2.7% when restricted to ultrasound-dated pregnancies. Performance was better than any combination of second-trimester markers. Implementation challenges included initial samples being collected too early and sample matching. Conclusions Integrated serum screening for Down syndrome was successfully implemented in primary care settings; screening performance was consistent with predictions. It provides an accessible and acceptable alternative to screening protocols that require nuchal translucency measurements. Copyright © 2005 John Wiley & Sons, Ltd.

Objectives To estimate the prenatal screening performance of an integrated serum test for detecting trisomy 18, which combines measurements of first- and second-trimester markers with maternal age to assign patient-specific risks. Methods Published and new observations of maternal serum marker levels in trisomy 18 and unaffected pregnancies are used to derive population parameters. These parameters are then combined in a multivariate Gaussian model to assign patient-specific risks for trisomy 18. Results The best combination of serum markers includes pregnancy-associated plasma protein-A in the first trimester and alpha-fetoprotein, unconjugated estriol and human chorionic gonadotropin in the second trimester. At a second-trimester risk cutoff of 1 : 100, these 4 markers, in combination with maternal age, detect 90% of trisomy 18 pregnancies at a false-positive rate of 0.1%. The odds of a trisomy 18 pregnancy among screen-positive women are 1 : 4. Without the first-trimester marker, detection is reduced to 67% at about the same false-positive rate. Conclusion The algorithm described here is highly efficient for detecting trisomy 18 and should be considered by programs that offer serum-integrated screening for Down syndrome. Copyright © 2003 John Wiley & Sons, Ltd.

The potential of a new fluorescent in situ hybridization technique is discussed, which uses a complete set of telomeric probes to reveal cryptic chromosome rearrangements that remain undetected by standard cytogenetic analysis. We report the obstetric history of a patient who had a termination of pregnancy at 20 weeks for a fetus with multiple congenital anomalies but a normal male karyotype using conventional G-banding analysis on a mid-trimester placental biopsy. In a subsequent pregnancy, a diaphragmatic hernia and intra-uterine growth restriction were detected at 34 weeks' gestation and a fetal blood sample showed a normal female karotype. However, her child was born with dysmorphic features and additional severe abnormalities including microcephaly, anophthalmos and left fixed talipes. The child has shown marked developmental delay. In view of a strong family history of congenital abnormalities and recurrent miscarriage suggestive of a familial translocation, a fluorescent in situ hybridization technique using specific telomeric probes was performed on blood from the affected child and her parents. An unbalanced subtelomeric translocation was detected involving the long arms of chromosomes 2 and 7 in the child and a balanced translocation was detected in her father. Accurate genetic counselling and the opportunity for early prenatal diagnosis can now be offered to this family.

To investigate whether statistical parameters used in Down syndrome screening between 15 and 22 weeks of pregnancy can be used at 14 weeks, we assayed alpha-fetoprotein (AFP), unconjugated oestriol (uE3), total human chorionic gonadotrophin (hCG), free alpha-hCG, free beta-hCG, and inhibin-A in 16 pregnancies with Down syndrome in the 14th week of pregnancy and expressed values in multiples of the normal median. The median and standard deviation values for these 16 pregnancies were not materially different from those published for 15-22 weeks. It is reasonable, therefore, to offer Down syndrome screening using these markers starting at 14 completed weeks of pregnancy instead of 15 weeks. It needs to be recognized, however, that serum AFP measurement for neural tube defect screening is less effective at this time than between 16 and 18 weeks of pregnancy.

Published studies have reached varying conclusions as to the benefit of replacing human chorionic gonadotropin (hCG) measurements with the free beta-subunit of hCG (the free beta-subunit) for Down syndrome screening. One study reports 14 per cent higher detection for the free beta-subunit, while another finds an actual loss in detection. To explore this issue further, we directly compared the screening performance of hCG and the free beta-subunit, alone and in combination with other serum markers, using banked sera obtained prior to amniocentesis and karyotyping. Altogether, 52 Down syndrome and 5065 unaffected pregnancies were studied. Sera were thawed and assayed for hCG and the free beta-subunit over 1 year. At a 5 per cent false-positive rate, the detection rate for hCG in combination with maternal age and alpha-fetoprotein was higher than when the free beta-subunit was substituted (62 versus 57 per cent). Ultrasound dating and adding unconjugated oestriol both increased detection. The present findings, along with those from six case control studies (our re-analysis), indicate that the screening performances of hCG and the free beta-subunit are similar (median change in detection 0, range -8 to +3 per cent). Under optimal sample collection and transportation conditions, laboratories can expect to achieve similar screening performance using either hCG or the free beta-subunit measurements.

This study examines the relationship between maternal weight and serum levels of alpha-fetoprotein, unconjugated oestriol, and human chorionic gonadotropin in a population of 47,585 women being provided with prenatal screening for Down's syndrome and open neural tube defects. The study population contains sufficient numbers of women at the extremes of weight to allow the determination that a reciprocal-linear equation more accurately describes the weight relationship for two of the three analytes than the currently used log-linear equations. The reciprocal-linear equations, while more appropriate, provide only a minimal advantage over the log-linear equations. A more important finding is that published weight equations may not be optimal for some screening programmes, due to differences in the mean weight of the populations being tested. Screening programmes are encouraged to calculate their own weight correction formulae, based on data from their own population, and to monitor the mean maternal weight to detect when modifications in the weight correction formulae might be indicated.

This study documents the screening performance, in practice, of a published protocol for interpreting second-trimester Down syndrome risk in twin pregnancies, using maternal serum biochemical markers. Within a cohort of 35,150 pregnancies, 410 twin pregnancies were identified. The rate of twinning was positively associated with maternal age. Of the 274 twin pregnancies known prior to screening, 15 (5.5 per cent) were classified as being screen-positive for Down syndrome. When maternal age and dating method were taken into account, the screen-positive rates in twin and singleton pregnancies did not differ significantly. Nine of the 14 screen-positive women with viable twin pregnancies chose amniocentesis [64 per cent, 95 per cent confidence interval (CI) 35-87]. No cases of Down syndrome were identified. Based on modelling, an estimated 73 per cent of monozygotic twin pregnancies and 43 per cent of dizygotic twin pregnancies with Down syndrome would be identified at a 5 per cent false-positive rate. The overall detection rate would be about 53 per cent. When laboratories offer multiple marker screening to women with twin pregnancies, the false-positive rate ought to be similar to that found in singleton pregnancies; the detection rate is, however, likely to be lower for dizygotic twins.

Both a cross-sectional and a longitudinal study were performed to investigate whether or not the collection time should be taken into consideration when generating a patient's risk for fetal Down syndrome with multiple marker screening. Diurnal variations of third-trimester alpha-fetoprotein (AFP) levels and first-trimester human chorionic gonadotropin (hCG) levels have been previously reported. In addition, large episodic fluctuations of conjugated and unconjugated oestriol (uE3) as well as a diurnal variation have also been reported in the third trimester. If the levels of these analytes routinely fluctuate during the day, they could affect a patient's risk calculation for fetal Down syndrome. The longitudinal study evaluated ten non-diabetic women who underwent sequential sampling for AFP, hCG, and uE3. The cross-sectional study evaluated 1953 patients for these three markers whose time of sampling was recorded between 8·00 a.m. and 5·59 p.m. Using either study design, no significant effect was seen in the median MOM levels of the screening analytes as a function of the time of day.

This prospective study investigates the relationship between insulin-dependent diabetes and maternal serum levels of alpha-fetoprotein (AFP), unconjugated oestriol (uE3), and human chorionic gonadotropin (hCG). It also examines the potential impact on screening for Down syndrome. The population-based cohort included 20 321 pregnant women in Maine who underwent routine serum screening for Down syndrome in the second trimester. The cohort included 52 women with insulin-dependent diabetes. Maternal serum AFP levels are now routinely adjusted for insulin-dependent diabetes. These adjustments, therefore, were made routinely in the diabetic women, but no equivalent adjustments were made for uE3 and hCG values. The initial false-positive rate (using all three markers) among the women with diabetes was not significantly different from that in the non-diabetic population (7·7 and 5·4 per cent, respectively). Prior to adjustment for insulin-dependent diabetes, the median AFP level in the 52 women was 0·73 multiples of the median (MOM); the median levels of uE3 and hCG were 0·93 and 0·98 MOM, respectively. When the uE3 and hCG levels were adjusted, the initial false-positive rate was unchanged. Median serum levels of uE3 were significantly higher in the 33 women whose onset of diabetes was prior to 19 years of age (0·99 MOM) than in the 19 women whose onset of diabetes was at age 19 or older (0·84 MOM). This is the first population-based study to investigate the relationship between diabetes and serum levels of AFP, uE3, and hCG, and confirms earlier observations from a case—control study that found only slightly lower uE3 and hCG levels.

Unconjugated oestriol (uE3) and human chorionic gonadotropin (hCG) levels were determined in second-trimester maternal serum (MS) samples from 21 pregnancies associated with fetal anencephaly and 15 pregnancies associated with fetal open spina bifida. Each measurement was expressed as a multiple of the median (MoM) for unaffected pregnancies for each completed week of gestation. In pregnancies associated with anencephaly, the median value for MSuE3 was very low (0.17 MoM, range less than 0.12-0.33 MoM), suggesting a functional defect in the fetal adrenal prior to 20 weeks' gestation; the median value for MShCG was also low (0.73 MoM), although not to the same extent as for MSuE3. A biological explanation for the hCG result is not apparent. In pregnancies associated with open spina bifida, the MSuE3 and MShCG values were unremarkable, consistent with a lack of involvement of these open fetal defects in the synthesis and secretion of uE3 and hCG.

In this study we define for the first time the distribution of alpha-fetoprotein levels and acetylcholinesterase ratios in amniotic fluid samples associated with fetal gastroschisis. Elevated alpha-fetoprotein levels and positive acetylcholinesterase measurements are found in virtually all cases, indicating that these combined measurements are highly reliable in detecting this lesion.

In the present study, a recently reported immunochemical technique for measuring acetylcholinesterase (AChE) in amniotic fluid utilizing the 4F19 antibody was compared with the widely utilized polyacrylamide gel technique to determine whether the immunochemical assay provided an advantage in separating unaffected pregnancies from those associated with open spina bifida (OSB) and open ventral wall defects (OVWD). The study included (1) 73 amniotic fluid samples from unaffected pregnancies [alpha-fetoprotein (AFP) less than 2 MoM] with no visible gel AChE band, (2) nine bloodstained samples from unaffected pregnancies (AFP 2.2-4.0 MoM) with visible gel AChE bands, (3) 18 samples associated with OSB (AFP 2.2-7.0 MoM) with visible gel AChE bands, and (4) 20 samples associated with OVWD (AFP 3.2-53.5 MoM) with visible gel AChE bands. The immunochemical assay produced ranges of measurements in the four respective categories as follows: (1) 2-60 arbitrary units (AU): (2) 14-69 AU, (3) 61-593 AU, and (4) 22-476 AU. Eight of the nine unaffected pregnancies with visible gel AChE bands had immunochemical measurements below the highest measurement for the samples with no visible AChE band (60 AU), as did five out of 20 OVWD pregnancies. Two of the OSB cases had values of 61 and 62 AU. These data indicate that the 4F19 specific monoclonal antibody to AChE is capable of distinguishing unaffected from affected pregnancies with reasonable reliability but that more work needs to be done to establish the extent of overlap between the unaffected and affected populations.

Thyroid antibodies were measured in mid-trimester antenatal serum samples from 77 pregnancies affected by fetal Down's syndrome and 385 unaffected control pregnancies. Using a haemagglutination technique, thyroglobulin antibodies were detected in 5.2 per cent of cases (4) and 2.9 per cent of controls (11), and thyroid microsomal antibodies were detected in 22 per cent (17) and 15 per cent (59), respectively. Using an enzyme-linked immunosorbent assay (ELISA) for thyroglobulin antibodies and a cut-off level of 50 KIU/l, positive results were found in 25 per cent of cases (19) and 22 per cent of controls (84). Using an ELISA for thyroid microsomal antibodies and the same cut-off level, the proportions were 52 per cent (40) and 39 per cent (149), respectively. While not statistically significant, the differences were consistent with the previously reported increased levels of thyroid antibody found in non-pregnant women who had had pregnancies associated with Down's syndrome.

Very low maternal serum alpha-fetoprotein (MSAFP) levels (less than 10 ng/mL) are known to be associated with non-viable pregnancies, including conditions such as fetal death, molar pregnancies, and non-pregnancies. There has not been agreement, however, as to whether very low MSAFP levels indicate already existing fetal deaths or are actually predictive. We analysed 230 pregnancies with MSAFP levels less than 10 ng/mL from among 15,807 women (1.5 per cent) screened consecutively during a three-year period and identified 26 non-viable pregnancies, 22 of which were diagnosed sonographically as part of the screening process (17 missed abortions, 3 blighted ova, 2 non-pregnancies). Furthermore, 20 of these 22 pregnancies were associated with essentially undetectable MSAFP levels (less than 5 ng/mL). Our data indicate that pregnancies with MSAFP values less than 5 ng/mL are the group most strongly associated with fetal non-viability and that very low MSAFP values are not strongly predictive for fetal death.