| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
ORIGINAL RESEARCH |
From the Division of Human Genetics, Department of Pediatrics; Division of Maternal Fetal Medicine, Department of Obstetrics and Gynecology; and the Department of Pediatric Dentistry, University of Connecticut Health Center, Farmington, Connecticut.
Address reprint requests to: Peter Benn, PhD University of Connecticut Health Center Division of Human Genetics 263 Farmington Avenue Farmington, CT 06030-6140 E-mail: benn{at}nso1.uchc.edu.
 |
Abstract |
|---|
Top
 Abstract Materials and MethodsResultsDiscussionReferences |
|---|
Methods: Second-trimester screening results for 41,565 women were reviewed to determine whether pregnancies could be considered to be at high risk for trisomy 18. The screening test was considered positive if either maternal serum concentrations of alpha-fetoprotein (MSAFP), hCG, and unconjugated estriol (E3) fell below defined levels, or the second-trimester patient-specific risk (based on maternal age and serum analytes) was greater than 1:100. Detection rates, false-positive rates, and pregnancy outcomes for the two protocols were compared.
Results: The fixed-cutoff method showed a 23% detection rate and a 0.19% false-positive rate for trisomy 18. These low rates were in close agreement with a theoretical expectation for fixed-cutoff trisomy 18 screening. The risk-based approach resulted in a 69% detection rate and a 0.45% false-positive rate. Both methods identified pregnancies with other fetal anomalies.
Conclusion: Overall, the risk-based method is more effective than the fixed-cutoff approach to trisomy 18 screening.
Lower-than-average concentrations of second-trimester maternal serum alpha-fetoprotein (MSAFP) and unconjugated estriol (E3), together with elevated hCG are found often in women whose fetuses are affected with Down syndrome.1–3 Prenatal screening for Down syndrome takes advantage of these associations by modifying each patient’s age-specific risk for Down syndrome by a factor, the likelihood ratio, that is determined by the analyte concentrations.4
In pregnancies in which trisomy 18 (Edwards syndrome) is present, the concentrations of MSAFP, hCG, and E3 are generally lower than normal.1,2,5 These findings can be applied to second-trimester screening by identifying a subgroup of screen-positive pregnancies on the basis of MSAFP, hCG, and E3 values falling below predefined levels.6 Concentrations of analytes are expressed as multiples of the median (MoM) for unaffected pregnancies. For example, using MSAFP 
0.75 MoM, hCG 0.60 MoM, and E3 0.55 MoM, it has been estimated that 60% of trisomy 18 fetuses could be identified with a false-positive rate of 0.4%.6 A disadvantage of this fixed-cutoff approach to trisomy 18 screening is that the well-established association between maternal age and the incidence of autosomal trisomies is not incorporated in the risk evaluation.
An alternative strategy is to calculate each pregnant woman’s risk for fetal trisomy 18 by modifying the age-specific risk with a likelihood ratio, similar to the approach that is used in Down syndrome screening.7,8 Screen-positive cases can include women with elevated MSAFP (when low hCG and E3 concentrations and/or advanced maternal age provide sufficient compensating risk), and the approach therefore has the potential to improve the detection of fetal trisomy 18 with omphalocele. Using a second-trimester risk of 1:100 as a cutoff, a 60% detection rate for trisomy 18 should be achieved for the pregnant population in the United States with a false-positive rate of 0.2%.9
In practice, for both protocols, the observed detection rates and false-positive rates may vary substantially from these theoretical expectations. In addition to differences in the performance of the assays at low concentrations, clinical factors such as the widespread use of ultrasound for gestational age assessment or the referral of high-risk patients for amniocentesis without utilization of maternal serum screening will affect screening efficiency. One prospective trial using the fixed-cutoff method showed close agreement with screening expectations.10 However, a second study showed a low detection rate, and the performance of this protocol has been questioned.11 No prospective data, and only limited retrospective data,9 are available currently to evaluate the patient-specific approach to trisomy 18 screening. In addition, little data are available to evaluate whether trisomy 18 screening is effective in identifying other fetal abnormalities.
We therefore chose to review outcomes of pregnancies that were screen-positive for trisomy 18 by using these two widely used protocols to determine if there is any particular advantage of one method over the other.
|
Materials and Methods |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Follow-up information on pregnancy outcome was collected by contacting the offices of referring physicians and requesting completion of a questionnaire for each pregnancy. Data also were collected from regional maternal-fetal medicine referral centers and the University of Connecticut cytogenetics laboratory. The collection of follow-up data was approved by the University of Connecticut Health Center Institutional Review Board. For the period covered by this study, follow-up was available for 93.7% of patients with trisomy 18 screen-positive results by the fixed-cutoff screening protocol and for 92.3% by the risk-based method. Overall, follow-up was available for 75.1% of referrals.
To compare the performance of the fixed-cutoff and risk-based methods for trisomy 18 screening, test results for the entire study period were re-evaluated to determine whether a result could be classified as screen-positive by either method. To minimize the effect of imprecise gestational age assessments, referrals to the laboratory with gestational age based on physical examination were excluded from the analysis (less than 2% of all referrals). Five patients with trisomy 18 screen-positive results (by both the fixed and risk-based approaches) were excluded on the basis of obvious explanations for the screening results (two nonpregnant patients, two with suspected sample mix-up prior to receipt by the screening laboratory, and one with a clerical error in referral information).
The 13 cases of trisomy 18 all were diagnosed during pregnancy, and none resulted in live births. Of these, four fetuses were affected with omphalocele, and none had a neural tube defect. All cases were ascertained initially through a combination of abnormal ultrasound findings, advanced maternal age, and screening results. Two patients were screen-positive for Down syndrome, two were screen-positive for an open neural tube defect, and three were screen-positive for trisomy 18 (fixed-cutoff method). Screening for trisomy 18 was in routine use, and ascertainment of affected cases may therefore have favored this approach.
To determine the expected detection rates and false-positive rates, 100,000 screening results for both trisomy 18 and normal pregnancies were simulated using the means, standard deviations, and analyte correlation coefficients determined by Palomaki et al9 in a computer program kindly provided by Dr. T. Reynolds.14 The maternal age distribution of all women screened was used in determining the overall detection rates and false-positive rates for the risk-based method.
Observed detection rates and false-positive rates refer to initial rates without subsequent corrections for gestational age errors. Statistical analyses were carried out using the SPSS (Statistical Package for Social Sciences, SPSS Inc., Chicago, IL). Comparisons of rates were carried out using Pearson 
2 or McNemar exact test, with a P value < .05 considered significant.
|
Results |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Table 1
summarizes the characteristics of the patients who were screen-positive for trisomy 18. There was no significant excess of older women with screen-positive results using the fixed-cutoff method, relative to the frequency for women with screen-negative results (2 = 0.77, P > .05). However, because the risk-based approach incorporated maternal age in the determination, this strategy did identify preferentially more women of advanced maternal age (2 = 151.8, P < .001). Race/ethnicity of the women screen-positive for trisomy 18 by the risk-based method also differed significantly from those women with negative results (2 = 12.26, P < .01). For both the fixed-cutoff and risk-based methods, there was no significant difference in the rate of women with screen-positive trisomy 18 when results were based on last menstrual period (LMP) dating compared with the rates seen when gestational age was determined by ultrasound evaluation.
|
). Excluding the cases with abnormalities or fetal death, the adjusted false-positive rate was 0.16%.
|
The risk-based method identified nine of the 13 cases of trisomy 18 (detection rate 69%). This detection rate was significantly better than that obtained with the fixed-cutoff approach (McNemar exact test, P < .05). Two of the four fetuses with omphalocele were identified. The false-positive rate for the risk-based method was 0.45%, including eight cases with other chromosome abnormalities (three cases of triploidy; two with trisomy 21 [both cases also screen-positive for Down syndrome]; one 45,X/46,XY mosaic; one 45,XX-15/46,XX/46,XXr(15); and one 46,XY,r(4)(p16q35)/45,XY, -r(4)(p16q35). Other reasons for a false-positive result included fetal death (20 cases of which 14 had elevated MSAFP), anencephaly (seven of the 15 present in the entire screened population), and other fetal abnormalities (one case with hydronephrosis; one ventriculomegaly; one X-linked ichthyosis; one ambiguous genitalia with dysmorphism). After excluding all abnormalities and fetal death, the adjusted false-positive rate was 0.35%.
Unlike the fixed-cutoff method, the risk-based approach identified many pregnancies that were either screen-positive for Down syndrome or screen-positive for an open neural tube defect (Table 1
). As an adjunct to Down syndrome and open neural tube defect screening, the risk-based method had an added trisomy 18 detection rate of 54% (seven of 13 cases) and increased the overall false-positive rate of screening by 0.36%. The combined detection rate for trisomy 18 using the risk-based method and Down syndrome and open neural tube defect screening together was 92% (12 of 13 cases), and all four cases with omphalocele were identified.
The computer simulation of analyte values in normal and trisomy 18 pregnancies indicated that 34% of trisomy 18 pregnancies and 0.18% of normal pregnancies could be expected to have MSAFP 0.75 MoM, hCG 0.55 MoM, and E3 0.60 MoM. These detection and false-positive rates were not improved substantially by relaxing the criteria for a positive test result. For example, using MSAFP 0.80 MoM, hCG 0.60 MoM, and E3 0.65 MoM, the detection rate was 39.6% with a false-positive rate of 0.46%, and for cutoffs MSAFP 0.85 MoM, hCG 0.65 MoM, and E3 0.70 MoM, the detection rate was 45.0% and false-positive rate 1.01%. Table 3 summarizes the expected performance of the risk-based trisomy 18 screening. For the 1:100 second-trimester cutoff used in this study, the expected detection rate was 69% and false-positive rate 0.23%. The data in Table 3 illustrate that the false-positive rate of this screening is strongly dependent on the maternal ages of the population screened. Expectations for the screening using alternative cutoffs also are presented.
|
|
Discussion |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
Recognizing these difficulties in evaluating the effectiveness of trisomy 18 screening, we have attempted to compare the fixed-cutoff and patient-specific risk methods. In this study, 13 cases of trisomy 18 were known to be present in the screened population of 41,565 women. Based on the 13 cases known to be present, the risk-based method had a significantly higher detection rate (69%) than the fixed-cutoff approach (23%). A total of 17 cases would be expected on the basis of the above population incidence, suggesting that there may have been some cases that were not referred for screening or not ascertained through incomplete follow-up. The detection rates for the two screening protocols should therefore be considered as maximum estimates.
The fixed-cutoff protocol was expected to identify 60% of trisomy 18 pregnancies with a 0.4% false-positive rate.6 Similar to the findings of Yankowitz et al,11 we observed a much lower detection rate for trisomy 18 (23%). A review11 of published analyte levels in trisomy 18 pregnancies indicates that in affected pregnancies the MSAFP concentration is higher than the 0.75 MoM cutoff in 37% of cases, the hCG limit of 0.55 MoM is exceeded in 26%, and E3 is higher than 0.60 MoM in 20% of cases. Because analyte levels are approximately independent of each other,9 a detection rate of approximately 37% is a more likely expectation based on the combination of results. An estimation of the detection rate and false-positive rate for the fixed-cutoff method can also be derived from population simulation data using the mean, standard deviation, and correlation coefficients of the analytes.9 Using this approach, we conclude that the detection rate for this protocol should in fact be approximately 34%, with a false-positive rate of 0.18%. It should be noted that the original estimate6 of a 60% detection rate with a 0.4% false-positive rate was based on only 10 cases of trisomy 18 and a relatively small number of normal pregnancies, and these may not have been sufficiently representative of the affected and unaffected groups.
In contrast, the observed detection rate for the risk-based protocol (69%) appeared to be in close agreement with expectations (65%) from our population simulation results. However, the observed false-positive rate (0.45%) was higher than that projected (0.23%), attributable, at least in part, to the high number of fetal deaths and anomalies in the screen-positive group. Simulated false-positive rates are based on the assumption that the trivariate Gaussian distribution model for analyte values will predict accurately the number of cases with the most extreme analyte values.
For both methods of trisomy 18 screening, the identification of a variety of other chromosome abnormalities added a considerable extra benefit (Table 2). The risk-based approach also identified many cases of fetal death, with approximately 10% of the positive results attributable to demise. These cases were often associated with high MSAFP and therefore most likely represented more recent death.17 Anencephaly, which is associated with elevated MSAFP and very low E3,18,19 was also a relatively common explanation for a positive result by the risk-based method. However, the detection rate (47%) was insufficient for this to be diagnostically useful.
Overall, we conclude that the patient-specific risk-based protocol for trisomy 18 screening is more effective than the fixed-cutoff approach. The second-trimester risk cutoff used (1:100) is not necessarily optimal; individual programs must evaluate carefully their false-positive rates and the maternal age distribution in the screened population. Viewed as an adjunct to Down syndrome and open neural tube defect screening, the risk-based protocol appears to be capable of providing high detection rates with acceptable false-positive rates. The risk-based method also has the considerable added advantage of providing information that is patient-specific. In this context, it is of interest to note that of the nine trisomy 18 pregnancies detected by this approach, six had risks of 1:10 or more.
|
Footnotes |
|---|
Received September 8, 1998. Received in revised form November 16, 1998. Accepted November 19, 1998.
|
References |
|---|
|
Top
AbstractMaterials and MethodsResultsDiscussionReferences |
|---|
2. Bogart MH, Pandian MR, Jones OW. Abnormal maternal serum chorionic gonadotrophin levels in pregnancies with fetal chromosome abnormalities. Prenat Diagn 1987;7:623–30.[Medline]
3. Canick JA, Knight GJ, Palomaki GE, Haddow JE, Cuckle HS, Wald NJ. Low second trimester maternal serum unconjugated oestriol in pregnancies with Down syndrome. Br J Obstet Gynaecol 1988;95: 330–3.[Medline]
4. Wald NJ, Cuckle HS, Densem JW, Nanchahal K, Royston P, Chard T, et al. Maternal serum screening for Down’s syndrome in early pregnancy. BMJ 1988;297:883–7.
5. Norgaard-Pederson B, Larsen SO, Arends J, Svenstrup B, Tabor A. Maternal serum markers in screening for Down syndrome. Clin Genet 1990;7:35–43.
6. Canick JA, Palomaki GE, Osthanondh R. Prenatal screening for trisomy 18 in the second trimester. Prenat Diagn 1990;10:546–8.[Medline]
7. Barkai G, Goldman B, Ries L, Chaki R, Zer T, Cuckle H. Expanding multiple marker screening for Down’s syndrome to include Edwards’ syndrome. Prenat Diagn 1993;13:843–50.[Medline]
8. Staples AJ, Robertson EF, Ranieri E, Ryall RG, Haan EA. A maternal serum screen for trisomy 18: An extension of maternal serum screening for Down syndrome. Am J Hum Genet 1991;49: 1025–33.[Medline]
9. Palomaki GE, Haddow JE, Knight GJ, Wald NJ, Kennard A, Canick JA, et al. Risk-based prenatal screening for trisomy 18 using alpha-fetoprotein, unconjugated oestriol and human chorionic gonadotropin. Prenat Diagn 1995;15:713–23.[Medline]
10. Palomaki GE, Knight GJ, Haddow JE, Canick JA, Saller DN, Panizza DS. Prospective intervention trial of a screening protocol to identify fetal trisomy 18 using maternal serum alpha-fetoprotein, unconjugated oestriol, and human chorionic gonadotropin. Prenat Diagn 1992;12:925–30.[Medline]
11. Yankowitz J, Fulton A, Williamson R, Grant SS, Budelier WT. Prospective evaluation of prenatal maternal serum screening for trisomy 18. Am J Obstet Gynecol 1998;178:446–50.[Medline]
12. Benn PA, Horne D, Briganti S, Greenstein RM. Prenatal diagnosis of diverse chromosome abnormalities in a population of patients identified by triple marker testing as screen positive for Down syndrome. Am J Obstet Gynecol 1995;173:496–501.[Medline]
13. Benn PA, Borgida A, Horne D, Briganti S, Collins R, Rodis JF. Down syndrome and neural tube defect screening: The value of using gestational age by ultrasonography. Am J Obstet Gynecol 1997;176:1056–61.[Medline]
14. Reynolds TM, Nix AB, Dunstan FD, Dawson AJ. Age-specific detection and false-positive rates: An aid to counseling in Down syndrome risk screening. Obstet Gynecol 1993;81:447–50.[Abstract]
15. Hook EB, Hamerton JL. The frequency of chromosome abnormalities detected in consecutive newborn studies-differences between studies-results by sex and by severity of phenotypic involvement In: Hook EB, Porter IH. eds. Population cytogenetics: Studies in humans. New York: Academic Press. 1977:63–79.
16. Hook EB, Topol BB, Cross PK. The natural history of cytogentically abnormal fetuses detected at midtrimester amniocentesis which are not terminated electively: New data and estimates of the excess and relative risk of later fetal death associated with 47, +21 and some other abnormal karyotypes. Am J Hum Genet 1989;45:855–61.[Medline]
17. Walters CP, McKinnon WC. Elevated maternal serum alpha-fetoprotein, very low unconjugated oestriol and very low human chorionic gonadotropin associated with recent fetal death. Prenat Diagn 1993;13:423–4.[Medline]
18. Canick JA, Knight GJ, Palomaki GE, Haddow JE. Second trimester levels of maternal serum unconjugated oestriol and human chorionic gonadotropin in pregnancies affected by fetal anencephaly and open spina bifida. Prenat Diagn 1990;10:733–7.[Medline]
19. Yaron Y, Hamly DD, O’Brien JE, Critchfield G, Leon J, Ayoub M, et al. Combination of elevated maternal serum alpha-fetoprotein (MSAFP) and low estriol is highly predictive of anencephaly. Am J Med Genet 1998;75:297–9.[Medline]
This article has been cited by other articles:
|
|
 |
|
  F. M. Breathnach, F. D. Malone, G. Lambert-Messerlian, H. S. Cuckle, T. F. Porter, D. A. Nyberg, C. H. Comstock, G. R. Saade, R. L. Berkowitz, S. Klugman, et al. First- and Second-Trimester Screening: Detection of Aneuploidies Other Than Down Syndrome Obstet. Gynecol., September 1, 2007; 110(3): 651 - 657. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. F. X. EGAN, P. BENN, A. F. BORGIDA, J. F. RODIS, W. A. CAMPBELL, and A. M. VINTZILEOS Efficacy of Screening for Fetal Down Syndrome in the United States From 1974 to 1997 Obstet. Gynecol., December 1, 2000; 96(6): 979 - 985. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
