Maternal serum alpha-fetoprotein (MSAFP) screening and ultrasonography have both been used for detecting neural tube defects (NTDs) prenatally.1–10 The American College of Obstetricians and Gynecologists (ACOG) recommends that all pregnant women be offered MSAFP screening in the second trimester of pregnancy, and this test has become standard care in the United States.1 The California Legislature has mandated that pregnant women be offered MSAFP screening per ACOG policy.11 ACOG does not recommend prenatal ultrasonography for routine screening2 and recommends ‘‘specialized ultrasound examination'’ for detecting NTDs only in women with elevated serum AFP levels.1 This position is supported by a 1993 randomized, controlled trial named RADIUS (Routine Antenatal Diagnostic Imaging Ultrasound Study), which concluded that routine ultrasonography did not improve pregnancy outcomes when compared with ultrasonography ordered only for medical indications.12,13
The purpose of this study was to compare ultrasonography with MSAFP screening in the prenatal diagnosis of NTDs in a large health care program, Kaiser Permanente in Northern California. Each year, about 35,000 pregnant women are seen at Kaiser Permanente's 40 prenatal facilities in Northern California. The Kaiser Permanente Perinatal Screening Program can track and determine outcomes for all prenatal patients. We used the large Kaiser Permanente database to study NTD outcomes retrospectively while comparing the effectiveness of serum screening (MSAFP) and ultrasonography for detecting fetal NTDs.
PATIENTS AND METHODS
Approval for conducting the study was obtained from the Kaiser Permanente Medical Care Program Institutional Review Board. Kaiser Permanente maintains a database containing prenatal testing results as well as information about pregnancy outcomes. The State of California Department of Health Services provides us with expanded alpha-fetoprotein (XAFP) results on all Kaiser Permanente patients. In addition, Kaiser Permanente has a separate, integrated Regional Genetics Database, which contains diagnostic information on all prenatal patients referred for genetic evaluation. We extracted data from all 3 systems and conducted medical chart review when appropriate. The Kaiser Permanente databases were used to identify types of testing used to diagnose NTDs among 219,000 consecutive pregnancy outcomes observed at Kaiser Permanente from August 1995 through July 2002. We also identified number, type, and maternal ethnicity of cases diagnosed with NTDs.
As mandated by the State of California, all prenatal patients were offered XAFP screening, performed at gestational age 15–20 weeks, which includes human chorionic gonadotropin (hCG) and unconjugated estriol in addition to MSAFP. The XAFP screening was accepted by 172,300 women (79%) during the study period. Although not universally recommended by ACOG as standard care, second-trimester ultrasonographic screening has been routine policy at Kaiser Permanente since 1995. At Kaiser Permanente, all pregnant women are offered second-trimester ultrasonography at 15 to 20 weeks of gestation to assess fetal anatomy. Some obstetric facilities chose to perform this test before (and some, after) MSAFP screening. In cases when an NTD is discovered by ultrasonography before MSAFP screening, State of California regulations recommend that the blood screen not be done (because a diagnostic test was already performed). All positive or suspicious results of routine ultrasonography are referred to the 4 Kaiser Permanente prenatal diagnostic centers (genetics department) for follow-up.
Ultrasonography was done using standardized equipment (Acuson machine; model 128 or higher; Acuson, Mountain View, CA) to examine patients at tertiary care centers (State of California–designated prenatal diagnostic centers) and nontertiary centers. Ultrasonography was performed by sonographers certified by the American Registry of Diagnostic Medical Sonographers, and scans were read by perinatologists or by radiologists.
XAFP screening conformed with standards established by the Genetic Disease Branch of the State of California14,15; data resided there and in Kaiser Permanente's XAFP database. For this screening, maternal serum was collected at between 15 and 20 weeks of gestation. MSAFP levels were reported as a multiple of the median; adjustments were made for gestational age, maternal weight and race, and presence of insulin-dependent diabetes mellitus before pregnancy. All MSAFP values ≥ 2.5 multiples of the median in singleton pregnancies were considered screen-positive for NTD. At Kaiser Permanente, ultrasonographic dating was used to improve accuracy of XAFP screening, and most centers routinely performed second-trimester ultrasonography before the MSAFP blood specimen was taken. All cases of NTDs in which XAFP screening was performed were reviewed by an XAFP coordinator or genetic counselor to ensure accuracy of the information used to interpret the result.
Table 1 describes the age, midpregnancy weight, and race/ethnicity of our maternal population. During the study, 189 cases of NTDs were identified (Table 2 and Fig. 1), yielding a prevalence of 1:1158. Of these 189 cases, 186 (98%) were detected prenatally; 3 (2%) were diagnosed after birth. Of the 189 cases with NTDs, 104 (55%) had anencephaly, 67 (35%) had open spina bifida (6 closed cases were excluded from the study), and 18 (10%) had encephalocele. Of the 189 NTD cases, 172 (91%) were isolated defects.
Of the 186 cases diagnosed prenatally, most (115, or 62%) were detected initially by routine second-trimester ultrasonography (Fig. 1). Of these 115 cases, 19 had negative XAFP results before ultrasonographic diagnosis, 14 had positive XAFP screens known after ultrasonographic diagnosis, and 82 never had XAFP screening. (The ultrasonographic diagnosis was made before scheduled serum screening, the women were already scheduled for maternal age-based amniocentesis, or they registered beyond the window for XAFP screening.) An additional 69 cases (37%) were discovered at our state-approved Prenatal Diagnosis Center when targeted ultrasonography was done because of an XAFP indication (ie, current positive screen result, NTD diagnosed during previous pregnancy, or certain history of anticonvulsant use). In 3 of these Prenatal Diagnosis Center cases, previous ultrasonography failed to diagnose the NTD; 2 were detected after an elevated amniotic fluid AFP level was found at amniocentesis done for anticonvulsant use (no serum AFP done); and the third case was diagnosed after a positive MSAFP result was obtained. Because the patients in all 3 of these cases were scheduled for detailed ultrasonographic examination later in the pregnancy (because the history included spina bifida or anticonvulsant drug use), the initial routine scan was brief and unfocused. In addition, 2 (1%) of our NTD cases were diagnosed by pathology examination after miscarriage at 12 weeks of gestation. Most cases of prenatal NTDs (181 of 184) in our study were discovered before 24 weeks of gestation, and the families had the option of pregnancy termination. Overall, 80% of families elected to terminate the pregnancy if an NTD was found.
Three of the 189 NTD cases had postpartum diagnoses. The mother of 1 neonate with spina bifida sought prenatal care late (initial care obtained in gestational week 31); third-trimester ultrasonography did not detect the lesion in this fetus. The mother of another neonate born with spina bifida had negative results of XAFP screening and did not receive second- or third-trimester ultrasonography. The mother of the infant born with encephalocele and intact dura had negative results of MSAFP screening and normal results of second-trimester ultrasonography.
Table 2 shows distribution of NTDs and associated results of MSAFP screening. Of the 102 cases in which XAFP screening was performed, 25% had a negative MSAFP result. MSAFP screening used alone would have failed to diagnose 38% of spina bifida cases, 8% of anencephaly cases, and 67% of encephalocele cases in the screened group.
Although almost all of the 186 NTD cases detected prenatally were ultimately diagnosed by ultrasonography (Fig. 1), we evaluated separately the subset of 130 women who had initial routine ultrasonography without any known MSAFP result (either the blood was not drawn for MSAFP screening or, if MSAFP screening was done, results were not known at ultrasonography). Thirty-two percent of women had the ultrasound examination performed in tertiary care centers; 68%, in nontertiary centers. This initial ultrasound examination detected defects in 125 (96%) of the 130 women and thus yielded a 4% false-negative rate. No women had false-positive NTD ultrasound diagnoses.
When elevated MSAFP levels were used as the screening measure, NTD detection rates of 75% to 80% for open spina bifida were reported.3,8,11,16,17 After ACOG recommended MSAFP screening,1 many U.S. states, including California,11 issued regulations mandating that all pregnant women be offered MSAFP screening.
In contrast, prenatal ultrasonography—used routinely by many obstetricians for diagnosing fetal anomalies—is neither mandated in California nor recommended by ACOG as a standard of care for diagnosing NTDs. Influential results of a randomized controlled study, the RADIUS trial,12 questioned the usefulness of routine ultrasonography. Although the RADIUS trial was strongly criticized,18 ACOG guidelines conclude that a positive cost-benefit for routine ultrasonography in pregnancy is not established and that “… a physician is not obligated to perform ultrasonography in a patient who is at low risk and has no indications” (p. 1454).2
The RADIUS trial has been criticized, particularly regarding the value of ultrasonography in fetal anomaly detection.13,18 The Crane analysis of the RADIUS trial data13 found an unusually low detection rate (35%) of anomalous fetuses and a low pregnancy termination rate (29% of 31 cases identified before 24 weeks of gestation). Although Crane et al13 did detect 6 of their 7 NTD cases, this number is too small for statistical comparison. Crane et al13 concluded that ultrasonographic screening for fetal anomaly detection cannot be justified given a lack of measurable benefit. By contrast, 96% of women in our study who carried a fetus with an NTD were identified before 24 weeks of gestation, and 80% chose to terminate the pregnancy.
Of the 172,300 women (79%) who had MSAFP screening during the study, almost all had ultrasonography. Our data consistently showed that ultrasonography was more predictive for diagnosing NTDs than was MSAFP screening alone (Table 2). Most clinically significant was the finding that 15 (38%) of 40 spina bifida cases screened by MSAFP were diagnosed in women with negative results; without ultrasonography, these 15 cases of spina bifida would have been missed. Although almost all cases of anencephaly screened (94%) were detected by MSAFP screening, anencephaly is a lethal condition that lacks the clinical effect of spina bifida, a lifelong condition requiring extensive, multidisciplinary care. Maternal serum alpha-fetoprotein screening failed to detect encephalocele in 67% of screened cases because this type of NTD is more likely to be closed. Further, because encephaloceles constituted only 10% of NTDs, their effect on the health care system is much less than the effect of spina bifida.
The poor predictability of MSAFP screening observed in our study might be explained by the State of California cutoff level for abnormal test results. This value, 2.5 multiples of the median, differs from the 2.0 multiples of the median cutoff used in other programs.3 When we recalculated our data using 2.0 multiples of the median as the cutoff level, the percentage of missed NTD cases fell from 25% to 17%, and for spina bifida cases in particular, fell from 38% to 22.5%. The overall rate of false-positive results tripled (increased from 0.9% to 2.7%) with this change in cutoff. The cutoff decision must be made administratively by considering benefits and disadvantages. For example, false-positive test results increase parental anxiety and cost of screening but have less clinical and economic effect than do false-negative test results. However, although a missed diagnosis of NTD is less likely if a lower MSAFP cutoff point is chosen, this likelihood is still much greater than if ultrasonographic screening is done. Our study identified 130 cases in which ultrasonography was done without knowing the MSAFP value, and only 5 (4%) of these 130 cases were not diagnosed by ultrasonography.
Our study has some weaknesses. Because we followed State of California protocol, bias may have been created by the lack of XAFP screening for NTD cases detected by routine ultrasonography done before XAFP. Women who had XAFP screening thus constituted a lower-risk group for whom NTD detection rates could have been underestimated by serum screening. With this possibility in mind, we reanalyzed serum screening detection of spina bifida (for which our detection rate, 62%, was lowest). Of our 67 cases of spina bifida, 27 had no serum screening. (The defect was detected before the planned XAFP blood test or the patient declined XAFP screening.) If we assumed the unlikely possibility that all 27 cases would have had serum that screened positive, the final detection rate would be only 78% (52 of 67)—still lower than the 96% detected by ultrasonography.
We did not address the issue of cost. Variables such as the added benefit of XAFP screening for other purposes (eg, for Down syndrome and for trisomy 18), as well as controversy regarding the best screening tools or combination thereof, would need to be considered in a cost-benefit analysis. The purpose of this study was to quantify the value of ultrasonography in NTD screening and to raise awareness among medical practitioners that the false-negative rate for MSAFP is a clinically important problem, especially in regard to spina bifida. Although still considered the accepted standard for screening, MSAFP screening alone can miss 25% of NTDs (including 38% of spina bifida cases). However, second-trimester ultrasonography is more likely to detect an NTD (only 4% of cases were missed). We recommend that MSAFP screening should not be the sole screening method for detecting NTDs and that routine second-trimester ultrasonography be added as part of standard prenatal care.
1. Clinical management guidelines for obstetricians–gynecologists. Neural tube defects. ACOG Practice Bulletin No. 44. (Replaces Committee Opinion Number 252, March 2001). Obstet Gynecol 2003 Jul;102:203–13.
2. Clinical management guidelines for obstetricians-gynecologists. Ultrasonography in pregnancy. ACOG Practice Bulletin No. 58. American College of Obstetricians and Gynecologists. Obstet Gynecol 2004 Dec;104:1449–58.
3. Palomaki GE, Williams JR, Haddow JE. Prenatal screening for open neural-tube defects in Maine. N Engl J Med 1999;340:1049–50.
4. Lennon CA, Gray DL. Sensitivity and specificity of ultrasound for the detection of neural tube and ventral wall defects in a high-risk population. Obstet Gynecol 1999;94:562–6.
5. Platt LD, Feuchtbaum L, Filly R, Lustig L, Simon M, Cunningham GC. The California Maternal Serum alpha-Fetoprotein Screening Program: the role of ultrasonography in the detection of spina bifida. Am J Obstet Gynecol 1992;166:1328–9.
6. Summers AM, Farrell SA, Huang T, Meier C, Wyatt PR. Maternal serum screening in Ontario using the triple marker test. J Med Screen 2003;10:107–11.
7. Burton BK. Elevated maternal serum alpha-fetoprotein (MSAFP): interpretation and follow-up. Clin Obstet Gynecol 1988 Jun;31:293–305.
8. Burton BK, Sowers SG, Nelson LH. Maternal serum alpha-fetoprotein screening in North Carolina: experience with more than twelve thousand pregnancies. Am J Obstet Gynecol 1983;146:439–44.
9. Velie EM, Shaw GM. Impact of prenatal diagnosis and elective termination on prevalence and risk estimates of neural tube defects in California, 1989–1991. Am J Epidemiol 1996;144:473–9.
10. Nicolaides KH, Campbell S, Gabbe SG, Guidetti R. Ultrasound screening for spina bifida: cranial and cerebellar signs. Lancet 1986;2:72–4.
11. Cunningham GC, Tompkinison DG. Cost and effectiveness of the California triple marker prenatal screening program. Genet Med 1999;1:199–206.
12. Ewigman BG, Crane JP, Frigoletto FD, LeFevre ML, Bain RP, McNellis D. Effect of prenatal ultrasound screening on perinatal outcome. RADIUS Study Group. N Engl J Med 1993;329:821–7.
13. Crane JP, LeFevre ML, Winborn RC, Evans JK, Ewigman BG, Bain RP, et al.. A randomized trial of prenatal ultrasonographic screening: impact on the detection, management, and outcome of anomalous fetuses. The RADIUS Study Group. Am J Obstet Gynecol 1994;171:392–9.
14. Walton DL, Norem CT, Schoen EJ, Ray GT, Colby CJ. Second-trimester serum chorionic gonadotropin concentrations and complications and outcome of pregnancy. N Engl J Med 1999;341:2033–8.
15. Schoen E, Norem C, O'Keefe J, Krieger R, Walton D, To TT. Maternal serum unconjugated estriol as a predictor for Smith-Lemli-Opitz syndrome and other fetal conditions. Obstet Gynecol 2003;102:167–72.
16. Wald NJ, Cuckle H, Brock JH, Peto R, Polani PE, Woodford FP. 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.
17. Cohen FL. Neural tube defects: epidemiology, detection, and prevention. J Obstet Gynecol Neonatal Nurs 1987;16:105–15.
18. Filly RA, Crane JP. Routine obstetric sonography. J Ultrasound Med 2002;21:713–8.