Original Article
Second-Trimester Maternal Serum Alpha-Fetoprotein Levels and the Risk of Subsequent Fetal Death
N Engl J Med 1991; 325:6-10July 4, 1991
- Abstract
- Abstract
Background.
The finding of an elevated level of maternal serum alpha-fetoprotein during the second trimester of pregnancy may indicate that the fetus has died or is about to die. It is uncertain, however, whether the finding is associated with an increased risk of fetal death later in gestation independent of known causes of elevation, such as the presence of neural-tube defects or multiple gestation.
Methods.
To address this question, we performed a case–control study of 612 women whose pregnancies ended in fetal death and 2501 women who gave birth to live infants, using reports from California vital statistics for 1987. All the women had singleton pregnancies and alpha-fetoprotein screening in the second trimester.
Results.
Women with elevated levels of serum alpha-fetoprotein in the second trimester of pregnancy had an increased risk of fetal death, and the risk was increased until term. Women with the highest levels of serum alpha-fetoprotein — ≥3.0 times the median value — had a very high risk of fetal death (odds ratio, 10.4; 95 percent confidence interval, 4.9 to 22.0) as compared with women who had normal levels of alpha-fetoprotein. Maternal serum alpha-fetoprotein levels that were 2.0 to 2.9 times the median were also associated with an elevated risk of fetal death (odds ratio, 2.4; 95 percent confidence interval, 1.7 to 3.4). Elevated levels of alpha-fetoprotein were especially likely to be associated with fetal death in cases in which maternal hypertension or placental infarction was also present.
Conclusions.
An unexplained elevated level of maternal serum alpha-fetoprotein in the second trimester of pregnancy is associated with an increased risk of subsequent fetal death, up to four to five months after alpha-fetoprotein screening. (N Engl J Med 1991; 325:6–10.)
Media in This Article
Figure 1
Odds Ratios for Fetal Death, According to Maternal Serum Alpha-Fetoprotein Levels and Gestational Age at Death (in Completed Weeks).Table 1
Characteristics of Women Whose Pregnancies Ended in Fetal Death and of Those Who Gave Birth to Live Infants.Article Activity
- Article
FETAL deaths at ≥20 weeks of gestation now constitute a major proportion of perinatal deaths in the United States, and in some states the fetal death rate is reported to equal or exceed the neonatal death rate.1 For this reason, increased attention is being directed to the early detection of pregnancies in which there is a risk of fetal death and to the development of strategies to prevent fetal death in these pregnancies.2
This study examined the relation between secondtrimester measurements of maternal serum alpha-fetoprotein and subsequent fetal death. Screening of serum alpha-fetoprotein levels is offered to pregnant women primarily to identify those at risk of carrying a fetus with a neural-tube defect. Women with high alpha-fetoprotein levels have an elevated risk of having a child with a neural-tube defect and are referred for ultrasonography and other diagnostic procedures. Common explanations for elevated values of maternal serum alpha-fetoprotein include neural-tube defects, multiple gestation, and missed abortion.3 Elevated levels are also known to be associated with ventralwall defects3 and rare congenital disorders, such as congenital nephrosis.4 This study was designed to determine whether elevated levels of maternal serum alpha-fetoprotein independent of the above factors are associated with an increased risk of fetal death after alpha-fetoprotein testing. The study considered singleton pregnancies in which the fetus had no neural-tube defects or other birth defects related to alpha-fetoprotein and was viable at the time of alpha-fetoprotein testing.
Previous studies have reported a wide range of estimates of the risk of fetal death in such pregnancies.165 6 7 8 9 10 11 One purpose of this study was to provide accurate risk estimates for various high levels of alpha-fetoprotein. A second purpose was to investigate the possibility that the serum alpha-fetoprotein level can be used to predict which fetuses will die at relatively late gestational ages. This question is of clinical importance because some fetal deaths occurring after 28 weeks of gestation may be prevented by fetal monitoring and early delivery of infants with signs of distress.2
Methods
Assays were conducted in the laboratories of the California Maternal Serum Alpha-Fetoprotein Screening Program with the Abbott enzyme immunoassay for alpha-fetoprotein. The coefficient of variation was 3 to 5 percent throughout the study period.
Blood samples were collected between 15 and 19 weeks of gestation. On average, the concentration of maternal serum alpha-fetoprotein doubles between 15 and 19 weeks of gestation. The results were standardized for gestational age by dividing the actual value (in nanograms per milliliter) by the median value for the appropriate gestational day, according to the universal approach to analyses of alpha-fetoprotein.12 The resulting value is referred to as a multiple of the median. These values were then adjusted for maternal weight, maternal race, and the presence of insulin-dependent diabetes in the mother. Similar adjustments are used by many screening programs.13
Gestational age was determined by counting the number of completed days after the last menstrual period or, if this information was not available, on the basis of ultrasonographic findings (e.g., 15 weeks thus included any day from 15 weeks 0 day to 15 weeks 6 days). All women with alpha-fetoprotein values that were more than 2.5 times the median were examined for errors in gestational dating. In cases in which the gestational age was found to be inaccurate, the alpha-fetoprotein value was recalculated with the correct gestational age.
We linked a total of 774 fetal deaths reported to the state of California for 1987 as spontaneous and occurring at 20 weeks of gestation or later to the files of the California Maternal Serum Alpha-Fetoprotein Screening Program, using the mother's name, race and ethnic group, ZIP Code, and date of last menstrual period (or approximate date of conception). Live births occurring in 1987 were linked to maternal records of alpha-fetoprotein tests by the same method. A group of 2700 controls was randomly selected from the women who had live-born infants, so that the ratio of cases to controls was approximately 1:4.
In this and previous studies, the association between maternal serum alpha-fetoprotein levels and fetal death was examined after women with multiple pregnancies and those carrying fetuses with neural-tube defects or other birth defects related to alpha-fetoprotein had been excluded. It is well known that these conditions are associated with a high risk of fetal death.14 , 15 Interest has therefore focused on the risk of fetal death in pregnant women with unexplained elevations of serum alpha-fetoprotein levels.
We also wanted to exclude fetal deaths that occurred before alpha-fetoprotein screening. Approximately 12 to 15 percent of clinically recognized pregnancies end in a spontaneous abortion before 20 weeks of gestation,16 and a number of these miscarriages are retained in utero without vaginal bleeding or other evidence of fetal death.17 For this reason, an undetected fetal death, or missed abortion, will have occurred in some women who undergo alpha-fetoprotein screening. The date of death is usually reported as the date of expulsion of the fetus; thus, these losses may appear to have occurred after alpha-fetoprotein testing. It was important to exclude women with missed abortions from this study because such women often have high levels of alpha-fetoprotein.18 If they were included, the association between the maternal serum alpha-fetoprotein level and subsequent fetal death would be spuriously increased.
The majority of missed abortions do not appear sufficiently mature to warrant a fetal death certificate, but such certificates are required for deaths occurring after 20 weeks of gestation. However, in order to identify and exclude any missed abortions that might have been included in our sample, we excluded all 23 cases in which fetal death was reported to have occurred within three weeks of alpha-fetoprotein testing. It is unlikely that a missed abortion would remain undetected for more than three weeks at this point in gestation.
The women whose pregnancies ended in fetal death three weeks or more after testing and who were excluded from analysis included 85 with twin or triplet pregnancies, 14 with fetuses with neural-tube defects, 2 with fetuses with ventral-wall defects, 33 from whom blood specimens were drawn too early or too late in gestation for screening, 4 who had elective abortion because of other fetal anomalies (3 chromosomal and 1 structural), and 1 in whom fetal death was recognized at ultrasound examination 24 days after testing. Although they were reported to the department of vital statistics as spontaneous fetal deaths, 12 of the 14 fetuses with neural-tube defects were electively aborted because of these abnormalities. After these exclusions 612 women whose pregnancies ended in fetal death remained.
The same criteria for exclusion were applied to the controls (women whose pregnancies ended in the birth of live infants). We excluded 47 women who gave birth to twins or triplets, 1 whose infant had a neural-tube defect, and 151 from whom blood specimens were drawn too early or too late for alpha-fetoprotein screening. After these exclusions 2501 women whose pregnancies ended in live births remained.
Results
The characteristics of the two groups are shown in Table 1Table 1
Characteristics of Women Whose Pregnancies Ended in Fetal Death and of Those Who Gave Birth to Live Infants.. Fetal deaths were more likely if the mothers were black, were less than 20 years old, or were 35 years of age or older. Male fetuses were more likely to die than female fetuses.High levels of maternal alpha-fetoprotein were associated with an increased risk of subsequent fetal death, and levels that were at least 3.0 times the median were associated with a risk 10.4 times that associated with normal maternal alpha-fetoprotein values (Table 2Table 2
Risk of Fetal Death, According to the Level of Maternal Serum Alpha-Fetoprotein.). There was also an increased risk of fetal death with low maternal levels of alpha-fetoprotein—that is, levels less than 0.50 times the median — but this finding may be explained by chance (Table 2).In Figure 1Figure 1
Odds Ratios for Fetal Death, According to Maternal Serum Alpha-Fetoprotein Levels and Gestational Age at Death (in Completed Weeks). the odds ratios are shown separately for fetal deaths occurring at different gestational ages. Elevated maternal levels of alpha-fetoprotein were associated with an increased risk of fetal death throughout pregnancy, although the odds ratios were smaller for fetal deaths occurring at 38 weeks of gestation or later. The percentage of fetal deaths associated with maternal alpha-fetoprotein levels at least 2.0 times the median was 12.2 percent (17 of 139) between 20 and 23 weeks of gestation, 15.7 percent (16 of 102) between 24 and 27 weeks, 18.1 percent (15 of 83) between 28 and 31 weeks, 14.5 percent (24 of 166) between 32 and 37 weeks, and 4.9 percent (6 of 122) at 38 weeks or more.The risk of fetal death was computed with multiple logistic-regression analysis to adjust for the effects of other risk factors for fetal death (Table 3Table 3
Association of Fetal Death with Maternal Alpha-Fetoprotein Levels and Other Risk Factors.). The risks of fetal death associated with black race, previous fetal death in the mother, maternal age of 17 years or less, maternal age of 35 years or more, first birth, and male sex of the fetus were similar to the risks reported previously for these factors.19 The risks associated with either high or low levels of alpha-fetoprotein were unchanged by adjustment for these factors.The results were similar whether the gestational age was based on the date of the last menstrual period or on ultrasonographic findings. When the results were examined separately according to the week of gestation during which alpha-fetoprotein testing was done, they were also unchanged.
The results were examined according to the cause of death given on the fetal death certificate (Table 4Table 4
Risk of Fetal Death with Maternal Levels of Alpha-Fetoprotein at Least 2.0 Times the Median, According to Reported Cause of Death.). There was an association between elevated maternal levels of alpha-fetoprotein and most reported causes of fetal death. The association was particularly strong for deaths attributed to maternal hypertension, placenta previa, placental abruption, placental infarction, prolapsed cord, or premature rupture of the membranes.Discussion
Elevated maternal levels of serum alpha-fetoprotein were associated with an increased risk of fetal death, and the risk was highest with very high levels of alpha-fetoprotein, as compared with moderately elevated levels. Although the odds ratio was very high for the highest maternal alpha-fetoprotein levels, the great majority of fetal deaths (515 of 612; 84.2 percent) occurred in women with levels of alpha-fetoprotein within the generally accepted normal range of 0.5 to 1.9 multiples of the median. There were 78 fetal deaths in women with values of 2.0 multiples of the median or higher. Of these 78, 25 would be expected if the relative risk in this group was not elevated, leaving 53 cases — only 8.7 percent of all 612 fetal deaths — as the excess associated with elevated maternal serum alpha-fetoprotein levels.
The increased risk of fetal death that we found among women with singleton pregnancies and elevated levels of alpha-fetoprotein was similar to that reported by others. However, previous studies were based on small numbers of cases and reported conflicting results regarding the risk of fetal death at gestational ages of 28 weeks or more.6 , 7 , 11 We have examined a large number of cases of fetal death and have shown that the risk of fetal death among women with viable singleton pregnancies and elevated levels of serum alpha-fetoprotein is increased throughout pregnancy.
We cannot completely exclude the possibility that the increased risk of fetal death before 28 weeks of gestation is due to the inclusion of women in whom fetal death had occurred before alpha-fetoprotein testing. However, virtually all missed abortions are detected by the third trimester, and therefore the increased risk of fetal death after 28 weeks of gestation cannot be explained plausibly by missed abortions.
In cases of recent fetal death the maternal serum alpha-fetoprotein level is initially elevated, but over time the protein is cleared from the maternal circulation, resulting in extremely low values. Thus, fetal deaths that occur before alpha-fetoprotein screening may be associated with either high or low values.10 In this study there were no fetal deaths associated with maternal alpha-fetoprotein values of ≤0.25 times the median, and the relative risk of fetal death associated with low levels of alpha-fetoprotein (<0.5 times the median) was small and may be explained by chance (odds ratio, 1.6; 95 percent confidence interval, 0.9 to 2.7). Thus, our results concur with the suggestion of Simpson et al.20 that the very high risks of fetal death that were associated with low levels of maternal serum alpha-fetoprotein in some studies5 , 10 , 21 were due to the inclusion of women with missed abortions.
Pregnancies affected by autosomal trisomies have been shown to be associated with lower levels of maternal serum alpha-fetoprotein,22 and chromosomal abnormalities are more common among fetuses dying in utero than among live-born infants.23 This may explain our finding of a small increase in the relative risk of fetal death in the presence of low maternal alpha-fetoprotein values.
We found that the association between alpha-fetoprotein levels and fetal death could not be accounted for by differences between the cases and controls in maternal race or ethnic group, maternal age, fetal sex, maternal history of a previous fetal death, vaginal bleeding before alpha-fetoprotein screening, or parity. Vaginal bleeding is often associated with fetal death or impending fetal death.17 It is also known that women with vaginal bleeding during pregnancy have higher levels of alpha-fetoprotein.18 On the basis of these relations, it has been suggested that the association between maternal serum alpha-fetoprotein level and fetal death might be confined to fetal deaths preceded by vaginal bleeding. To the contrary, we found that elevated levels of alpha-fetoprotein were predictive of fetal death in pregnant women with or without vaginal bleeding.
Previous studies have found that elevated levels of maternal serum alpha-fetoprotein are associated with pregnancies complicated by preeclampsia, placental abruption, placental infarctions, chronic villositis, and low birth weight.7 , 24 , 25 It has been proposed that preeclampsia and placental abnormalities are associated with a breakdown in the placental barrier, resulting in an increase in the diffusion of alpha-fetoprotein into the maternal serum.26 This study agreed with previous reports in that women with elevated levels of alpha-fetoprotein had a very high risk of fetal death, especially if maternal hypertension, placental abruption, or placental infarction was also present. Such women also had a high risk of fetal death due to placenta previa or prolapsed cord. However, these findings are based on small numbers of cases and should be interpreted cautiously.
Almost all reported causes of fetal death showed some association with elevated maternal levels of serum alpha-fetoprotein. This may be explained by the tendency of the complications of pregnancy to occur together. For instance, maternal hypertension and placental complications are frequently found together and may result in prematurity or anoxia.27 Although several such conditions may actually have been present in the women whose pregnancies ended in fetal death, only the one cause of fetal death reported on the death certificate was available for analysis. Further research is needed to identify all conditions leading to fetal death that are related to elevated levels of maternal serum alpha-fetoprotein.
Supported by a grant from the March of Dimes Birth Defects Foundation.
We are indebted to John Arnopp and Lynn Palmer for technical assistance.
Source Information
From the School of Public Health (D.K.W., E.B.H.), University of California, Berkeley; the California Department of Health Services, Genetic Disease Branch, California Alpha-Fetoprotein Screening Program, Berkeley (L.S.L., G.C.C.); and the School of Medicine, Department of Obstetrics, Gynecology and Reproductive Sciences (M.S.G.) and Department of Pediatrics (E.B.H.), University of California, San Francisco. Address reprint requests to Dr. Hook at Warren Hall, University of California, Berkeley, CA 94720.
- References
References
1
Lammer
CrossRef | Web of Science | Medline2
Moore
Web of Science | Medline3
Maternal serum-alpha-fetoprotein measurement in antenatal screening for anencephaly and spina bifida in early pregnancy: report of U.K. Collaborative Study on Alpha-fetoprotein in Relation to Neural-tube Defects . Lancet 1977; 1:1323–32.
Web of Science | Medline4
Seppälä
CrossRef | Web of Science | Medline5
Burton
Web of Science | Medline6
Crandall
Medline7
Milunsky
Web of Science | Medline8
Nelson
Web of Science | Medline9
Brock
CrossRef | Web of Science | Medline10
Davenport
Web of Science | Medline11
Robinson
Web of Science | Medline12
Adams
Web of Science | Medline13
Evans
Web of Science | Medline14
Farooqui
CrossRef | Medline15
Byrne
CrossRef | Web of Science | Medline16
Warburton
Web of Science | Medline17
Simpson JL. Fetal wastage. In: Gabbe SG, Niebyl JR, Simpson JL, eds. Obstetrics: normal and problem pregnancies. New York: Churchill Livingstone, 1986:651–2.
18
Seppälä
CrossRef | Web of Science | Medline19
Bakketeig LS, Hoffman HJ, Titmuss Oakley AR. Perinatal mortality. In: Bracken MB, ed. Perinatal epidemiology. New York: Oxford University Press, 1984:99–151.
20
Simpson
Web of Science | Medline21
Bennett
Web of Science | Medline22
Merkatz
Web of Science | Medline23
Bauld
CrossRef | Web of Science | Medline24
Salafia
Web of Science | Medline25
Boyd
CrossRef | Web of Science | Medline26
Berkeley
Web of Science | Medline27
Anderson GD, Sibai BM. Hypertension in pregnancy. In: Gabbe SG, Niebyl JR, Simpson JL, eds. Obstetrics: normal and problem pregnancies. New York: Churchill Livingstone, 1986:819–64.
- Citing Articles (25)
Citing Articles
1
C. Clock, L. Pereira. (2011) Strip of the Month: March 2011. NeoReviews 12:3, e173-e182
CrossRef2
R. Salim, M. Okopnik, G. Garmi, Z. Nachum, N. Zafran, E. Shalev. (2010) Lack of Association between Unexplained Elevated Maternal Serum Alpha Fetoprotein and/or Human Chorionic Gonadotropin and the Occurrence of Placental Thrombotic Lesions. Placenta 31:4, 277-281
CrossRef3
Christine Brochet, Nathalie Morin, Mathieu Coudert, Danièle Vauthier-Brouzes, Nathalie Costedoat-Chalumeau, Maguy Bernard. (2009) Impact of antiphosholipid biology in maternal Down syndrome screening. Prenatal Diagnosis 29:5, 481-488
CrossRef4
S.L. Costa, L. Proctor, J.M. Dodd, M. Toal, N. Okun, J.-A. Johnson, R. Windrim, J.C.P. Kingdom. (2008) Screening for Placental Insufficiency in High-risk Pregnancies: Is Earlier Better?. Placenta 29:12, 1034-1040
CrossRef5
Meghana Toal, Cynthia Chan, Shafagh Fallah, Fawaz Alkazaleh, Vandana Chaddha, Rory C. Windrim, John C.P. Kingdom. (2007) Usefulness of a placental profile in high-risk pregnancies. American Journal of Obstetrics and Gynecology 196:4, 363.e1-363.e7
CrossRef6
Karl Einhäupl. (2007) ESPRIT study design and outcomes – a critical appraisal. Current Medical Research and Opinion 23:2, 271-273
CrossRef7
Fawaz Alkazaleh, Vandana Chaddha, Sandra Viero, Aasim Malik, Colleen Anastasiades, Hana Sroka, David Chitayat, Ants Toi, Rory C. Windrim, John C. Kingdom. (2006) Second-trimester prediction of severe placental complications in women with combined elevations in alpha-fetoprotein and human chorionic gonadotrophin. American Journal of Obstetrics and Gynecology 194:3, 821-827
CrossRef8
Helen A Lane, Jamie C Smith, J Stephen Davies. (2006) Noninvasive assessment of preclinical atherosclerosis. Vascular Health and Risk Management 2:1, 19-30
CrossRef9
N. J. Cowans, K. Spencer. (2005) First and second-trimester biochemical markers of chromosomal anomalies and their relationship to maternal haemoglobin levels. Prenatal Diagnosis 25:8, 693-699
CrossRef10
Simon Shenhav, Ofer Gemer, Essie Sassoon, Michael Volodarsky, Ronit Peled, Shmuel Segal. (2002) Mid-trimester triple test levels in early and late onset severe pre-eclampsia. Prenatal Diagnosis 22:7, 579-582
CrossRef11
A Antsaklis. (2001) False positive serum biochemical screening and subsequent fetal loss in women less than 35 years of age. British Journal of Obstetrics and Gynaecology 108:6, 589-593
CrossRef12
Aris J. Antsaklis, Nikolaos E. Papantoniou, George J. Daskalakis, Spyros A. Mesogitis, Stylianos J. Kitmirides, Stylianos S. Michalas. (2001) False positive serum biochemical screening and subsequent fetal loss in women less than 35 years of age. BJOG: An International Journal of Obstetrics and Gynaecology 108:6, 589-593
CrossRef13
Jill Mauldin. (2000) Prenatal diagnosis and fetal therapy-what lies in future?. The Indian Journal of Pediatrics 67:12, 899-905
CrossRef14
Tri huu Nguyen, Torben Larsen, Gerda Engholm, Henrik Moller. (2000) A discrepancy between gestational age estimated by last menstrual period and biparietal diameter may indicate an increased risk of fetal death and adverse pregnancy outcome. BJOG: An International Journal of Obstetrics and Gynaecology 107:9, 1122-1129
CrossRef15
Luthy, David A., . (1999) Maternal Markers and Complications of Pregnancy. New England Journal of Medicine 341:27, 2085-2087
Full Text16
Antoine Malek, Ruth Sager, Henning Schneider. (1998) Transport of Proteins Across the Human Placenta. American Journal of Reproductive Immunology 40:5, 347-351
CrossRef17
Catherine Y. Spong, Alessandro Ghidini, Charles N. Walker, Miguel Ossandon, John C. Pezzullo. (1997) Elevated maternal serum midtrimester α-fetoprotein levels are associated with fetoplacental ischemia. American Journal of Obstetrics and Gynecology 177:5, 1085-1087
CrossRef18
Deborah L. Yetman, William H. Kutteh, Robert Castorena, Charles Brown, Leland Baskin. (1997) Prevalence of elevated anticardiolipin antibodies in pregnant women with unexplained elevations of α-fetoprotein. Journal of Reproductive Immunology 33:1, 71-81
CrossRef19
Chandice Covington, Phyllis Gieleghem, Frances Board, Kimberly Madison, Daphne Nedd, Linda Miller. (1996) Family Care Related to Alpha-Fetoprotein Screening. Journal of Obstetric, Gynecologic, <html_ent glyph="@amp;" ascii="&"/> Neonatal Nursing 25:2, 125-130
CrossRef20
S. E. Auvil-Novak, R. D. Novak, N. El Sanadi. (1996) Twenty-Four-Hour Pattern in Emergency Department Presentation for Sickle Cell Vaso-Occlusive pain Crisis. Chronobiology International 13:6, 449-456
CrossRef21
Joe Leigh Simpson, Glenn E Palomaki, Brian Mercer, James E Haddow, Richard Andersen, Baha Sibai, Sherman Elias. (1995) Fetus-placenta-newborn. American Journal of Obstetrics and Gynecology 173:6, 1742-1748
CrossRef22
Manjeet Kaur, Ishwar C. Verma. (1995) Serum alpha feto-protein screening in high risk pregnancies. The Indian Journal of Pediatrics 62:1, 101-107
CrossRef23
Lyn S. Chitty, Martin Bobrow. (1994) Techniques in Current Use in Prenatal Diagnosis. International Journal of Technology Assessment in Health Care 10:04, 604
CrossRef24
Susan J. Gross, Owen P. Phillips, Lee P. Shulman, N. Lawrette Bright, Jeffrey S. Dungan, Joe Leigh Simpson, Sherman Elias. (1994) Adverse perinatal outcome in patients screen-positive for neural tube defects and fetal down syndrome. Prenatal Diagnosis 14:7, 609-613
CrossRef25
Cunningham, F. Gary, Gilstrap, Larry C., . (1991) Maternal Serum Alpha-Fetoprotein Screening. New England Journal of Medicine 325:1, 55-57
Full Text
