| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Department of Epidemiology Research, Danish Epidemiology Science Centre, Statens Serum Institut; and the Department of Clinical Biochemistry, Statens Serum Institut, Copenhagen, Denmark.
Address reprint requests to: Mads Melbye, MD, PhD Department of Epidemiology Research Statens Serum Institut Artillerivej nr. 5 DK-2300 Copenhagen S Denmark E-mail: mme{at}ssi.dk
 |
Abstract |
|---|
Top
 Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Methods: We followed 77,149 pregnant women and their infants from MSAFP screening in the 15th to 20th week of gestation until 1 year after birth. Information on pregnancy outcome was obtained from national registries. The relative risks (RRs) and 95% confidence intervals (CIs) for adverse pregnancy outcome were estimated according to the level of MSAFP, with adjustment for confounders.
Results: A total of 638 pregnancies resulted in spontaneous abortion, 289 in stillbirth, and 437 in infant death. Compared with women with MSAFP levels at 0.75–1.24 multiples of the median (MoM), those with MSAFP levels greater than or equal to 2.5 MoM had an increased risk of spontaneous abortion (RR 12.5; 95% CI 9.7, 16.1), preterm birth (RR 4.8; 95% CI 4.1, 5.5), small for gestational age (RR 2.8; 95% CI 2.4, 3.2), low birth weight (RR 5.8; 95% CI 5.0, 6.6), and infant death (RR 1.9; 95% CI 1.2, 2.8). Women with MSAFP levels below 0.25 MoM had an increased risk of spontaneous abortion (RR 15.1; 95% CI 9.3, 24.8), preterm birth (RR 2.2; 95% CI 1.3, 3.8), and stillbirth (RR 4.0; 95% CI 1.0, 16.0); those with levels less than 0.5 MoM had an increased risk of infant death (RR 1.9; 95% CI 1.2, 3.0). The increased risk of infant death remained after the subtraction of recognized conditions associated with extreme MSAFP values.
Conclusion: Pregnant women with extreme MSAFP values in the second trimester have an increased risk of fetal and infant deaths.
Unexplained high levels of maternal serum alpha-fetoprotein (MSAFP) have been associated with an increased risk of adverse pregnancy outcomes, such as fetal death before the 28th week, perinatal death, low birth weight (LBW), preterm birth, and other obstetric complications.1–19 Unexplained low levels of MSAFP have been associated primarily with an increased risk of fetal death, including spontaneous abortions and stillbirths,8,10,20–23 although it has not been established whether this increased risk is due to the well-known association between low MSAFP and chromosomal abnormalities. These findings suggest that MSAFP, besides its value in screening for neural tube defects and chromosomal abnormalities,24–26 might serve as a predictor of high-risk pregnancies.
Most previous studies on MSAFP and adverse pregnancy outcome have focused only on prenatal and perinatal pregnancy outcome, usually without adjustment for confounders. We examined whether second-trimester MSAFP levels could predict not only fetal but infant death and whether specific causes of infant death were associated with the MSAFP level. By using unique national registries, we followed an exceptionally large population-based cohort of pregnant women screened for MSAFP to determine pregnancy outcome up to 1 year after birth. This study design also enabled us to adjust for confounders such as maternal age, parity, and birth weight.
|
Materials and Methods |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
We obtained information on pregnancy outcome by linking the AFP database with national registries using the unique identification number given to all Danish residents. Information on vital status of the pregnant women and their offspring was obtained from the Danish Civil Registration System. Information on spontaneous abortion (International Classification of Diseases, 8th Revision [ICD-8], Danish version 643, 644, and 645), hydatidiform mole (ICD-8 645.0), and ectopic pregnancy (ICD-8 631) was obtained from the National Discharge Registry. Data on stillbirths, birth weight, length at birth, and gestational age at birth were obtained from the Danish National Birth Registry. Information on induced abortions was obtained from the National Register of Induced Abortions. Cause of death for stillbirths and infant deaths was obtained from the Cause-of-Death Registry. Deaths were coded according to the ICD-8 and ICD-10 international classification systems.
The study was restricted to pregnant women born in Denmark between April 1, 1935 and March 31, 1978 because complete information on parity was available from the civil registration system for these women.27 We considered pregnancies only if an MSAFP measurement had been performed after March 1, 1980, between the 15th and 20th weeks of gestation, and if the pregnancy ended before December 31, 1993. Pregnancies ending in multiple births were excluded (n = 943). Only the first MSAFP test taken during a pregnancy was used, but women could be included in the study with more than one sample if they had been taken during different pregnancies. In the analyses, we did not adjust for a possible individual effect for each woman. However, we did perform separate analyses in which we included only the first pregnancy with a recorded MSAFP measurement from each woman. The results from these analyses were similar to the general results.
The relative risk (RR) of spontaneous abortion between 15 and 28 weeks of gestation by the level of MSAFP was evaluated by log-linear Poisson regression.28 All pregnant women entered follow-up at the time of MSAFP sampling and were followed until spontaneous abortion, induced abortion, hydatidiform mole, ectopic pregnancy, delivery, the 29th week of gestation, emigration, or death of the woman, whichever came first. To investigate whether fetal death before sampling could account for the association with high levels of MSAFP and spontaneous abortion, we did an explanatory analysis and observed a decreasing RR in the first 3 weeks after the sample was taken, but the RR reached a plateau thereafter. In further analyses, follow-up was therefore divided into before and after the first 3 weeks after the sample was taken. In analyses of spontaneous abortion, we adjusted for maternal age at conception (1-year categories), calendar period at conception (5-year categories), interaction between previous birth and spontaneous abortion, weeks since the sample was taken, and gestational age at abortion.
The RR of delivering a child as preterm (before 37 weeks), small for gestational age (SGA), LBW (less than 2500 g), or as a stillbirth was estimated by log-linear binomial regression.29 We defined SGA as birth weight below the 10th percentile for gestational age.30
The RR of dying within the first year of life for live-born singletons was estimated by Cox regression. All live-born singletons entered follow-up on the day of birth and were followed until death, emigration, or their 1-year birthday, whichever came first. In this analysis, we adjusted for sex, maternal age at conception (1-year categories), parity at conception (0, 1, 2, 3, or 4+), calendar period at conception (5-year categories), and birth weight (250-g categories). Cause-of-death-specific analyses were performed using competing risks analysis in a Cox regression.31 In additional competing risks analyses we excluded infants dying due to conditions related to abnormal MSAFP from the definition of infant death using wide criteria for conditions previously suspected to be associated with abnormal MSAFP levels (neural tube defects including spina bifida, encephalocele, and anencephaly; gastrointestinal malformations including gastroschisis, omphalocele, and tracheoesophageal fistula; renal abnormalities; and chromosomal defects).
In all types of regression analyses, homogeneity and effect modification were evaluated with the Wald statistic.32 Effect modification was tested by evaluating the significance of an additional interaction in the regression analysis.
|
Results |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
The crude cumulative incidence of spontaneous abortion from 15 to 28 weeks of gestation was highest in women with MSAFP levels less than 0.25 and greater than or equal to 2.50 MoM. One percent of women with MSAFP in the range of 0.75–1.24 MoM had a spontaneous abortion during this interval, compared with 9.5% of women with a value greater than or equal to 2.5 MoM and 14% of women with a value less than 0.25 MoM. Table 1
shows the adjusted RR of spontaneous abortion at 15–28 weeks of gestation, overall, in the first 3 weeks after the sample was taken and more than 3 weeks after the sample was taken, by the level of MSAFP. The RR of spontaneous abortion showed a U-shaped association with MSAFP level. Compared with pregnancies with MSAFP levels between 0.75 and 1.24 MoM, those with levels less than 0.25 MoM and greater than or equal to 2.5 MoM had the highest risk of spontaneous abortion, and an increased risk at these levels persisted when more than 3 weeks passed from MSAFP sampling to abortion.
|
Among women with MSAFP levels greater than or equal to 2.0 MoM, 2010 women (58%) had amniocentesis and measurement of AF-AFP. Six percent of these women also had AF-AFP levels greater than 2.0 MoM. The association with spontaneous abortion more than 3 weeks after the AFP test was present regardless of the AF-AFP level. However, the association was strongest when AF-AFP also was greater than or equal to 2.0 MoM. For later adverse outcomes, we found no differences according to the level of AF-AFP.
A total of 76,138 pregnancies resulted in a live-born singleton or a stillbirth after 28 weeks’ gestation. Table 2 shows the RR of delivering a singleton as preterm, SGA, LBW, or as a stillbirth by the level of MSAFP. The RRs of preterm birth, SGA, and LBW among live-born singletons were associated with the level of MSAFP and tended to be highest in births with MSAFP values greater than or equal to 2.5 MoM and less than 0.25 MoM compared with births with MSAFP levels of 0.75–1.24 MoM, although the associations observed between LBW and SGA at the low levels of MSAFP were not statistically significant. The RR of stillbirth was highest at MSAFP less than 0.25 and greater than or equal to 2.5 MoM, but after adjustment for LBW, the RR at this level was markedly reduced and the associations first observed were no longer statistically significant. Among the 289 stillbirths, we excluded seven with conditions previously associated with abnormal levels of MSAFP, and this did not affect the risk estimates markedly.
|
shows the adjusted RR of dying in the first year of life, overall, during the first week after birth and from day 8 to day 365 by the level of MSAFP. The mortality rate was lowest in live-born singletons with an MSAFP level between 0.5 and 1.24 MoM; it tended to be higher with an MSAFP level less than 0.5 MoM and gradually increased with levels higher than or equal to 1.25 MoM. The high crude cumulative mortality rate in live-born singletons with MSAFP levels greater than or equal to 2.5 MoM was partly due to LBW in these children. After adjustment for birth weight, the RR of dying in the first year was reduced (from RR 5.1; 95% CI 3.4, 7.8 to RR 1.9; 95% CI 1.2, 2.8). Among the 437 infant deaths, we excluded 57 infants who died with conditions previously associated with abnormal MSAFP. This approach did not markedly affect the risk estimates for infant death.
|
2: P < .001). After excluding those who died with conditions known to be associated with high or low levels of MSAFP, we further evaluated the specific causes of death in live-born infants (ignoring prematurity) according to six general groups: congenital heart and blood vessel malformations (n = 91), other congenital malformations (n = 23), pregnancy and birth complications (n = 59), respiratory distress and asphyxia (n = 61), infections (n = 28), unknown causes (n = 91), and all other causes of death (n = 27). For each subgroup, an increased risk was associated with a high or low MSAFP, but the associations were significant only for the subgroup of pregnancy and birth complications (Wald test: P = .02). However, we could not identify any specific conditions or complications that accounted for this association. Including in the analyses only the 15,234 women who had ultrasound examinations yielded associations similar to those when gestational age was based on LMP.
|
Discussion |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Previous studies have found significant associations between elevated MSAFP and stillbirth or neonatal death.8,10,12,14 Our study had significant advantages when analyzing these outcomes, not only in its sample size and its long follow-up, but also in its ability to incorporate analyses of potential confounders. We found that birth weight is strongly associated with both MSAFP and stillbirth and infant death. However, although adjustment for LBW made the association between high MSAFP and stillbirth no longer significant in our study, the association with infant death remained significant for both high and low levels of MSAFP.
To explain these findings, we examined the cause-of-death records for live-born infants. Examining the diagnoses on the death certificates, we did not see patterns that identified specific conditions as new associations with either high or low MSAFP. Some conditions are well-established causes of high values, such as open neural tube defects and abdominal wall defects, where AFP leaks from the fetus into the AF. The only known association with low MSAFP involves chromosomal defects, although why this occurs is unclear. In our study, most of the pregnancies with these adverse outcomes were identified by follow-up of screening abnormalities and ended as induced abortions. However, we reviewed all death records and even after exclusion of some residual cases, the association remained unchanged.
Unexplained high levels of MSAFP have been associated with both ultrasonographic placental abnormalities (eg, placental sonolucencies, periplacental hemorrhage, increased placental thickness)33,34 and vascular and inflammatory placental lesions observed at delivery.35 Therefore, it has been proposed that placental pathology might permit more rapid diffusion of AFP from the fetus to the mother.36 We found that 6% of pregnancies with elevated levels of MSAFP also had a raised AF-AFP level, which could reflect fetal pathology with leakage of AFP into the AF. Compared with women with normal AF-AFP levels, these women had a higher risk of spontaneous abortion 3 weeks after blood sampling. However, a significantly increased risk of spontaneous abortion persisted in women with high MSAFP but normal AF-AFP, which could be explained by placental pathology in these pregnancies. Furthermore, the RR of other adverse pregnancy outcomes among women with elevated MSAFP was similar in women with elevated and normal AF-AFP levels. Together these findings suggest that unexplained high levels of MSAFP might reflect placental pathology, affecting the fetal viability during pregnancy as well as after birth. The reasons for the association between low levels of MSAFP and adverse pregnancy outcome are not clear. One possibility is that low MSAFP is a marker of fetal stress,37 which might also predict adverse pregnancy outcomes.
A major goal of antenatal care is to intervene in high-risk pregnancies. It has been suggested that MSAFP screening, apart from identifying fetuses with open neural tube defects and chromosomal abnormalities, could also identify pregnancies at high risk of adverse outcome. In our study, an MSAFP level of either less than 0.5 MoM or greater than or equal to 2.0 MoM (the most commonly used screening limits) would have identified 13% of those with adverse pregnancy outcomes, including spontaneous abortions, preterm births, SGA infants, LBW infants, stillbirths, and deaths within the first year after birth. This proportion excludes the 364 pregnancies (0.5%) ending in induced abortions. However, these screening limits also would have labeled as high risk 7% of pregnancies that resulted in birth at term of normal-weight infants, who were still alive 1 year after birth. Whether such surviving infants had abnormalities is unknown. The utility of a screening program to predict these pregnancy outcomes with such false-negative and false-positive results must be evaluated against its emotional and financial costs, which result from not only the screening itself but also the follow-up evaluation of abnormal values.
|
Footnotes |
|---|
Supported by the Danish National Research Foundation, Copenhagen, Denmark.
Received May 22, 2000. Received in revised form September 6, 2000. Accepted October 5, 2000.
|
References |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
2. Wald N, Cuckle H, Stirrat GM, Bennett MJ, Turnbull AC. Maternal serum-alpha-fetoprotein and low birth-weight. Lancet 1977;2:268–70.[Medline]
3. Brock DJ, Barron L, Duncan P, Scrimgeour JB, Watt M. Significance of elevated mid-trimester maternal plasma-alpha-fetoprotein values. Lancet 1979;1:1281–2.[Medline]
4. Smith ML. Raised maternal serum alpha-fetoprotein levels and low birth weight babies. Br J Obstet Gynaecol 1980;87:1099–102.[Medline]
5. Purdie DW, Young JL, Guthrie KA, Picton CE. Fetal growth achievement and elevated maternal serum alpha-fetoprotein. Br J Obstet Gynaecol 1983;90:433–6.[Medline]
6. Hamilton MP, Abdalla HI, Whitfield CR. Significance of raised maternal serum alpha-fetoprotein in singleton pregnancies with normally formed fetuses. Obstet Gynecol 1985;65:465–70.[Medline]
7. Nelson LH, Bensen J, Burton BK. Outcomes in patients with unusually high maternal serum alpha-fetoprotein levels. Am J Obstet Gynecol 1987;157:572–6.[Medline]
8. Burton BK. Outcome of pregnancy in patients with unexplained elevated or low levels of maternal serum alpha-fetoprotein. Obstet Gynecol 1988;72:709–13.
9. Robinson L, Grau P, Crandall BF. Pregnancy outcomes after increasing maternal serum alpha-fetoprotein levels. Obstet Gynecol 1989;74:17–20.
10. Milunsky A, Jick SS, Bruell CL, MacLaughlin DS, Tsung YK, Jick H, et al. Predictive values, relative risks, and overall benefits of high and low maternal serum alpha-fetoprotein screening in singleton pregnancies: New epidemiologic data. Am J Obstet Gynecol 1989; 161:291–7.[Medline]
11. Waller DK, Lustig LS, Cunningham GC, Golbus MS, Hook EB. Second-trimester maternal serum alpha-fetoprotein levels and the risk of subsequent fetal death. N Engl J Med 1991;325:6–10.[Abstract]
12. Crandall BF, Robinson L, Grau P. Risks associated with an elevated maternal serum alpha-fetoprotein level. Am J Obstet Gynecol 1991;165:581–6.[Medline]
13. Maher JE, Davis RO, Goldenberg RL, Boots LR, DuBard MB. Unexplained elevation in maternal serum alpha-fetoprotein and subsequent fetal loss. Obstet Gynecol 1994;83:138–41.
14. Brazerol WF, Grover S, Donnenfeld AE. Unexplained elevated maternal serum alpha-fetoprotein levels and perinatal outcome in an urban clinic population. Am J Obstet Gynecol 1994;171:1030–5.[Medline]
15. Simpson JL, Palomaki GE, Mercer B, Haddow JE, Andersen R, Sibai B, et al. Associations between adverse perinatal outcome and serially obtained second- and third trimester maternal serum alpha-fetoprotein measurements. Am J Obstet Gynecol 1995;173: 1742–8.[Medline]
16. Waller DK, Lustig LS, Cunningham GC, Feuchtbaum LB, Hook EB. The association between maternal serum alpha-fetoprotein and preterm birth, small for gestational age infants, preeclampsia, and placental complications. Obstet Gynecol 1996;88:816–22.[Abstract]
17. Cusick W, Rodis JF, Vintzileos AM, Albini SM, McMahon M, Campbell WA. Predicting pregnancy outcome from the degree of maternal serum alpha-fetoprotein elevation. J Reprod Med 1996; 41:327–32.[Medline]
18. Cho S, Durfee KK, Keel BA, Parks LH. Perinatal outcomes in a prospective matched pair study of pregnancy and unexplained elevated or low AFP screening. J Perinat Med 1997;25:476–83.[Medline]
19. Hsieh TT, Hung TH, Hsu JJ, Shau WY, Su CW, Hsieh FJ. Prediction of adverse perinatal outcome by maternal serum screening for Down syndrome in an Asian population. Obstet Gynecol 1997;89: 937–40.[Abstract]
20. Bennett MJ, Solymar M, Turnbull AC, Cuckle HS, Wald NJ. Pregnancies associated with low maternal serum alpha-fetoprotein concentrations. Am J Obstet Gynecol 1979;135:545–6.[Medline]
21. Davenport DM, Macri JN. The clinical significance of low maternal serum alpha-fetoprotein. Am J Obstet Gynecol 1983;146:657–61.[Medline]
22. Haddow JE, Hill LE, Palomaki GE, Knight GJ. Very low versus undetectable maternal serum alpha-fetoprotein values and fetal death. Prenat Diagn 1987;7:401–6.[Medline]
23. Simpson JL, Baum LD, Depp R, Elias S, Somes G, Marder R. Low maternal serum alpha-fetoprotein and perinatal outcome. Am J Obstet Gynecol 1987;156:852–62.[Medline]
24. Brock DJ, Sutcliffe RG. Alpha-fetoprotein in the antenatal diagnosis of anencephaly and spina bifida. Lancet 1972;2:197–9.[Medline]
25. Merkatz IR, Nitowsky HM, Macri JN, Johnson WE. An association between low maternal serum alpha-fetoprotein and fetal chromosomal abnormalities. Am J Obstet Gynecol 1984;148:886–94.[Medline]
26. Cuckle HS, Wald NJ, Lindenbaum RH. Maternal serum alpha-fetoprotein measurement: A screening test for Down syndrome. Lancet 1984;1:926–9.[Medline]
27. Westergaard T, Andersen PK, Pedersen JB, Olsen JH, Frisch M, Sorensen HT, et al. Birth characteristics, sibling patterns, and acute leukemia risk in childhood: A population-based cohort study. J Natl Cancer Inst 1997;89:939–47.
28. Breslow NE, Day NE. Statistical methods in cancer research. Vol. 2. The design and analysis of cohort studies. Lyon, France: International Agency for Research on Cancer, 1987.
29. Mccullagh P, Nelder JA. Generalized linear models. London: Chapman & Hall, 1989.
30. Secher NJ, Hansen PK, Lenstrup C, Pedersen-Bjergaard L, Eriksen PS, Thomsen BL, et al. Birthweight-for-gestational age charts based on early ultrasound estimation of gestational age. Br J Obstet Gynaecol 1986;93:128–34.[Medline]
31. Andersen PK, Borgan O, Gill RD, Keiding N. Statistical models based on counting processes. New York: Springer-Verlag, 1993.
32. Agresti A. Categorical data analysis. New York: Wiley, 1990:461.
33. Kelly RB, Nyberg DA, Mack LA, Fitzsimmons J, Uhrich S. Sonography of placental abnormalities and oligohydramnios in women with elevated alpha-fetoprotein levels: Comparison with control subjects. Am J Roentgenol 1989;153:815–9.
34. Bernstein IM, Barth RA, Miller R, Capeless EL. Elevated maternal serum alpha-fetoprotein: Association with placental sonolucencies, fetomaternal hemorrhage, vaginal bleeding, and pregnancy outcome in the absence of fetal anomalies. Obstet Gynecol 1992;79: 71–4.
35. Salafia CM, Silberman L, Herrera NE, Mahoney MJ. Placental pathology at term associated with elevated midtrimester maternal serum alpha-fetoprotein concentration. Am J Obstet Gynecol 1988; 158:1064–6.[Medline]
36. Thomas RL, Blakemore KJ. Evaluation of elevations in maternal serum alpha-fetoprotein: A review. Obstet Gynecol Surv 1990;45: 269–83.[Medline]
37. Seppala M, Ruoslahti E. Alpha fetoprotein in maternal serum: A new marker for detection of fetal distress and intrauterine death. Am J Obstet Gynecol 1973;115:48–52.[Medline]
This article has been cited by other articles:
|
|
 |
|
  L. Dugoff, J. C. Hobbins, F. D. Malone, J. Vidaver, L. Sullivan, J. A. Canick, G. M. Lambert-Messerlian, T. F. Porter, D. A. Luthy, C. H. Comstock, et al. Quad Screen as a Predictor of Adverse Pregnancy Outcome Obstet. Gynecol., August 1, 2005; 106(2): 260 - 267. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. H. Beall, F. Amidi, D. A. Gayle, S. Wang, R. Beloosesky, and M. G. Ross Placental and Fetal Membrane Nephrin and Neph1 Gene Expression: Response to Inflammation Reproductive Sciences, July 1, 2005; 12(5): 298 - 302. [Abstract] [PDF]
|
|
|
|
 |
|
 C. Cunniff and and Committee on Genetics Prenatal Screening and Diagnosis for Pediatricians Pediatrics, September 1, 2004; 114(3): 889 - 894. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
