Hazards of prenatal detection of neural*tube defects by screening maternal serum for afetoprotein In 1972 Brock and Sutcliffe1 noted the association betwen an increase in the concentration of a-fetoprotein (AFP) in the amniotic fluid and the presence of an open neural tube defect in the fetus. This observation has been amply confirmed and prenatal diagnosis of these defects in highrisk pregnancies is now a reasonably well established and widely practised component of prenatal care.2'3 Although the precise cause or causes of neural tube defects are unknown, the lesion is undoubtedly of multifactorial origin and genetic factors are presumed to be important. Empiric data indicate that once a woman has given birth to an infant with a neural tube defect the risk of her delivering another similarly affected infant increases from that for the general population (which varies from less than 1 per 1000 births to as high as 9 per 1000 births in some localized geographic areas) to approximately 1 in 20 births. After the birth of a second infant with a neural tube defect the risk of a third affected infant is about 1 in 10 births. Genetic and preamniocentesis counselling is usually straightforward because the parents have had to deal with the problem before. Nevertheless, difficulties have arisen with AFP determinations in the amniotic flu'd: a substantial number of conditions other than neural tube defects are also associated with increased concentrations of AFP; it is difficult to confirm the presence of a neural tube defect after a report of elevated concentrations of AFP in the amniotic fluid is received; and borderline elevations in concentrations of AFP in the amniotic fluid that may or may not be associated with neural tube defects do occur.2'3 Data began to appear in the literature suggesting that roughly parallel elevations in concentrations of AFP occurred in the mother's serum in the presence of a fetal neural tube defect, but the increase was much less than that in the amniotic fluid and was later in the pregnancy. Various centres then began measuring AEP in the maternal serum and tried to determine the efficacy of an
assay of AFP in the maternal serum as a method of screening for neural tube defects. Nineteen centres in the United Kingdom participated in a collaborative study.4 Their findings indicated that 69% of cases of open and closed spina bifida and 88% of cases of anencephaly would have been detectable prenatally between 16 and 18 weeks' gestation on the basis of serum concentrations of AFP equal to or greater than 2.5 times the median for unaffected pregnancies. The study concluded that "screening pregnant women by measuring the concentration of A.F.P. in their serum is an effective method of selecting women for ultrasonography and amniocentesis so that N.T.D.s can be diagnosed in utero." As a result, screening of all pregnancies was scheduled to begin in the United Kingdom in the spring of 1 978. and pressures are rising in Canada to follow suit. We and others5 are concerned about this approach for both technical and ethical reasons. There is variation between laboratories in the use of multiples of standard deviations, percentiles or multiples of the median to determine cut-off points.4 All these measures, however, by definition include some members of the normal population; for example, 1 % of the normal population will be classified as abnormal if the cut-off point is the 99th percentile. In tests where the values for normal and abnormal overlap, as for concentrations of AFP in both amniotic fluid and maternal serum, the price of moving the cut-off point closer to the 100th percentile is to decrease the proportion of abnormalities detected and thus increase the proportion of women with false-negative results. Therefore one must make a clinical and ethical judgement in deciding exactly where the cut-off point should be. In the case of maternal serum AFP tests one can determine the cutoff point by considering how many amniocenteses must be done in women whose fetus is normal compared with women whose fetus has a neural tube defect. We have calcul-
1186 CMA JOURNAL/MAY 20, 1978/VOL. 118
DlpNsone DIPROSONE (betamethasone dipropionate 0,05%) is a topical steroid. Chemically DIPROSONE is betamethasone- 17, 21 dipropionate. INDICATIONS AND CLINICAL USES: DIPROSONE provides anti-inflammatory, antipruritic and anti-allergic activity in the topical management of corticosteroid-responsive dermatoses. CONTRAINDICATIONS: Topical steroids are contraindicated in: 1. Untreated bacterial, tubercular and fungal infections involving the skin, and in certain viral diseases such as herpes simplex, chicken pox, and vaccinia. 2. Hypersensitivity to any of the components. WARNINGS: The safety of topical corticosteroids during pregnancy or lactation has not been established. The potential benefit of topical corticosteroids, if used during pregnancy or lactation, should be weighed against possible hazard to the fetus or the nursing infant. If used under an occlusive dressing, part icularly over extensive areas, sufficient absorption may take place to give rise to adrenal suppression and other systemic effects. Topical corticosteroids are not for ophthalmic use. PRECAUTIONS: Topical corticosteroids should be used with caution on lesions close to the eye. Although hypersensitivity reactions have been rare with topically applied steroids, the drug should be discontinued and appropriate therapy initiated if there are signs of sensitivity. In cases of bacterial infections of the skin, appropriate antibacterial agents should be used as primary therapy. If it is considered necessary, DIPROSONE may be used as an adjunct to control inflammation, erythema, and itching. If a symptomatic response is not noted within a few days to a week, the local applications of DIPROSONE should be discontinued until the infection is brought under control. Significant systemic absorption may occur when steroids are applied over large areas of the body, especially under occlusive dressings. To minimize this possibility, when longterm therapy is anticipated, interrupt treatment periodically or treat one area of the body at atime. Patients should be advised to inform subsequent physicians of the prior use of corticosteroids. Occlusive dressings should not be applied if there is an elevation of body temperature. ADVERSE REACTIONS: When occlusive dressings are used, pustules, miliaria, folliculitis and pyoderma may occur. The following adverse skin reactions have been reported with the use of topical steroids: dryness, itching, burning, local irritation, striae, skin atrophy, hypertrichosis, change in pigmentation, and secondary infection. Adrenal suppression has also been reported following topical corticosteroid therapy. Posterior subcapsular cataracts have been reported following systemic use of corticosteroids. DOSAGE AND ADMINISTRATION: A sufficient quantity of DIPROSONE should be applied to completely cover the affected area and should be massaged gently and thoroughly into the skin. The usual frequency of application is twice daily. For some patients adequate maintenance therapy may be achieved with less frequent application. AVAILABILITY: DIPROSONE Cream and Ointment are available in 20 g and 60 g tubes. Diprosone Lotion is available in 30m1 and 75m1 plastic squeeze bottles. 1 Medansky R. 5. and Handler R. M Analysis ol a New Highly Aclive conicosteroid. Illinois Medical Journal. July. 1976.
Complete information is available to health professionals on request from Schering Corporation Limited, Pointe Claire, Quebec H9R 1B4. MEMR
SCHERING EE9
ated this for women in Ontario for 1 year of screening. Although the birth rate for the province is approximately 125 000 per year, a proportion of pregnant women will fail to seek prenatal care before 18 weeks' gestation and others will refuse the test for personal reasons. We therefore arbitrarily based our calculations on 100 000 women. The frequency of spina bifida and anencephaly in Ontario is approximately 1.28 and 0.76 respectively per 1000 births.6 The cut-off point (2.5 times the median) was taken from the UK collaborative study.4 Of the 128 cases of spina bifida expected per year 40 would be missed, and of the 76 cases of anencephaly expected 9 would be missed; hence the detection rates would be 69% and 88% respectively. We determined that 3448 amniocenteses would be required to detect the 155 fetuses with neural tube defects; therefore about 21 amniocenteses would have to be done on women carrying normal fetuses to detect one fetus with a neural tube defect. In the United Kingdom the estimate was 10:1 with an average frequency of neural tube defects of 4.5 p. 1000 births. One of Wald and colleagues' tables revealed the dramatic effects of both the population frequency and the cut-off points on the number of amniocenteses that need to be done, and hence on the number of women with false-positive results.4 With a cut-off point 2.5 times the median and a hypothetical frequency of neural tube defects of 2 per 1000 births they predicted the 21:1 ratio that we calculated. If we had recalculated the ratio with the cut-off point at three times the median for Ontario, where the frequency of neural tube defects is 2 per 1000 births, the result would have been 10:1 with nearly 2000 fewer amniocenteses. This is the ratio predicted in Wald and colleagues' tables. The cost of obtaining this ratio would be the missing of 11 cases of spina bifida and 3 cases of anencephaly. Thus, each region must decide where its cut-off point should be on the basis of local frequency data and medical and ethical considerations. The concentration of AFP varies during pregnancy and one must be sure of the gestational age to interpret the test result correctly. One cannot assess the effect of misclassification of gestational age from the
UK collaborative study because the patients were included in the study only if there was no doubt about gestational age. If approximately 20% of women are unsure of the date of onset of their last menstrual period7 some of those carrying normal fetuses who underestimate the duration of gestation will have results of the initial serum screening test that are too high; however the correct duration should become apparent when ultrasonography is carried out before amniocentesis. On the other hand, in women who overestimate the duration of gestation the presence of a fetus with a neural tube defect may be missed by the screening test because the presence of these defects may not be detectable before 15 weeks' gestation.8 One way to avoid this would be to perform ultrasonography in all women unsure of their dates. Unless cheaper methods are developed for mass ultrasonography, this would be costly;9 we have calculated the cost to be nearly equal to all the other costs of screening combined.10 It is far from clear what the concentration of AFP in the amniotic fluid would be when women from the general population with high serum concentrations of AFP undergo amniocentesis. The UK collaborative study4 is not helpful here because much of their data were derived from women having amniocentesis at the time the blood sample was drawn. It is assumed that all women with a high serum concentration of AFP who are carrying a fetus with a neural tube defect would have a high concentration of AFP in the amniotic fluid. One would also expect high concentrations of AFP in the amniotic fluid in women whose fetuses have other abnormal conditions including duodenal atresia, omphalocele and Turner's syndrome.3 The most important unknown factor is what proportion of the women with normal fetuses would have a high concentration of AFP in the amniotic fluid. Milunsky" reported that three apparently normal fetuses were electively aborted because of elevated concentrations of AFP in the amniotic fluid and they estimate a falsepositive rate of approximately 0.15% for all tests performed.'2 Thus, in Ontario each year we can expect about five normal fetuses to be incorrectly labelled as abnormal and therefore be
1188 CMA JOURNAL/MAY 20, 1978/VOL. 118
at risk of abortion. By repeating the AFP test and using ancillary diagnostic procedures such as morphologic study of amniotic cells, ultrasonography, amniography and, perhaps in the future, fetoscopy, one may reduce the proportion of falsepositive results. The problem will be to decide which of the 160 cases with abnormal AFP test results (88 of spina bifida + 67 of anencephaly + 5 normal = 160) require the additional tests. Other factors concerning the serum AFP test require further clarification. No correlation (r = 0.17) has been found between serum concentrations and amniotic fluid concentrations of AFP in the nonelevated range.13 Shapiro, Skinner and Phillips'4 found that serum concentrations of AFP were significantly lower in pregnant Asian women than in Caucasian women, while others have shown no difference between black 'and Caucasian women.'5 Not all open neural tube defects cause serious debilitating handicaps. Concentrations of AFP in the amniotic fluid do not distinguish between meningocele and meningomyelocele, or between the mild and severe types of the latter. Neither ultrasonography nor fetoscopy has been developed sufficiently to allow safe and accurate prediction of the severity of the lesion. Given all of the above, we are deeply concerned by the often stated suggestion that such a test should become part of routine obstetric care. When a screening procedure in medicine becomes routine, communication between the screener, whether physician or nonprofessional, and the person screened tends to diminish. Some of the experiences with screening for sickle-cell disease in the United States serve as frightening examples.'6 In the case of neural tube defects we run the risk of providing couples with answers for which they had no question. As Bundey'7 pointed out 3½ years ago, "some of the couples who refuse amniocentesis with its implication that an affected fetus will be aborted will go on to have a child affected with spina bifida and will be reminded of their decision throughout the child's life." Other couples having very strong objections to abortion for any reason may, in the face of an elevated concentration of AFP from a test they did not request or fully understand
Atromid-S * (clofibrate) Indications ATROMID-S is indicated as an adjunct to diet and other measures for the reduction of elevated plasma lipids. Contraindications Pregnancy; lactation; hypersensitivity to ATROMID-S; clinically significant hepatic or renal dysfunction. Warnings Caution should be exercised when oral anticoag ulants are given in conjunction with A ROMID-S. The dosage of the anticoagulant should be reduced, usually by one-half (depending on the individual case), to maintain the prothrombin time at the desired level to prevent bleeding complications. Frequent prothrombin determinations are advisable until it has been definitely determined that the levels have been stabilized. ATROMID-S may displace other acidic drugs, such as phenytc.n and tolbutamide, from binding sites on plasma proteins. In diabetics on sulfonylurea type drugs, ATROMID-S may potentiate hypoglycemia. The safety of this drug in children has not yet been established. Precautions Patients with a history of jaundice or hepatic disease should be treated cautiously. Frequent liver function tests should be performed since the drug may produce functional liver abnormalities, which are usually reversible. If these tests are persistently abnormal, the drug should be discontinued. Complete blood counts should be carried out periodically because anemia, and more frequently, leukopenia have been reported in patients on ATROMID-S. Patients with low serum albumin levels should be treated cautiously. High levels of unbound drug may cause myalgia and raised creatine phosphokinase serum levels. This "flu like" syndrome, which usually disappears when the dose is reduced, has also been observed. Adverse Reactions Nausea may occur in 5% of patients. Less common effects are vomiting, loose stools, dyspepsia, flatulence, and abdominal distress. Headache, dizziness, fatigue, weakness, skin rash, urticaria, pruritus, stomatitis are reported occasionally. A slight transient rise in SGOT and/or SGPT has been observed in some patients without evidence of hepatotoxicity. A few cases of increased BSP retention have been reported. Muscle cramping, aching or weakness have also been reported. A Product Monograph is available on request. Dosage and administration For adults only-two capsules (1 g) twice daily with meals. Availability No.3243-Each capsule contains 500 mg clofibrate in bottles of 100 and 1000.
Ayerst . . 0 Industries, Limited.
R.G. DAVIDSON, MD, FRCP[C] Department of pediatrics L.J. SHEFFIELD, MB, B5, FRACP, DCH
Department of clinical epidemiology and biostatistics McMaster University Medical Centre Hamilton, Ont.
References 1. BROcK DJH, SUTCLTFFE RG: Alphafetoprotein in the antenatal diagnosis of anencephaly and spina bifida. Lancet 2: 197, 1972 2. MILUNSKY A, ALPERT E: Prenatal diagnosis of neural tube defects. I. Problems and pitfalls: analysis of 2495 cases using the alpha-fetoprotein assay. Obstet Gynecol 48: 1, 1976 3. Idem: Prenatal diagnosis of neural tube defects. II. Analysis of false positive and false negative aipha-fetoprotein results. Ibid, p 6 4. WALD NJ, CUCKLE H, BROCK DJH, et al: Maternal serum-aipha-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 1: 1323, 1977 5. CHECK W: Mass screening for open spina bifida needs careful consideration. JAMA 238: 1441, 1977 6. Congenital Anomaly Surveillance System, 1977 data, Ottawa, Health and Welfare Canada 7. BEAZLEY JM, UNDERHILL RA: Con-
finement date unknown. Nurs Times
AYER5T LABORATORIES Division of Ayerst, McKenna & Harrison Limited Montreal, canada Made in canada by arrangement with Imperial chemical
Reg'd
in the first place, submit to abortion and bear the psychologic consequences for as yet unknown periods. Aborting a normal fetus as a result of a false-positive result is a disaster that needs no further comment. In conclusion, we are not against screening for neural tube defects by the use of maternal serum concentrations of AFP. This caveat is intended to emphasize the need for carefully monitored pilot screening programs in several North American centres. Specific questions pertaining to the effect of miscalculation of gestational age, establishment of an appropriate cut-off point, the frequency of misclassification of normal fetuses as abnormal, educational programs for both patients and physicians, and the overall costs and benefits in financial, medical and emotional terms must be meticulously evaluated in prospective studies.
PAAB
.EEEEJ
67: 1414, 1971 8. WALD NJ: The detection of neural tube defects by screening maternal blood, in Prenatal Diagnosis; Cclloque, vol 61, Paris, Inserm, 1976, p 227
9. HAGARD 5, CARTER F, MILNE RG:
Screening for spina bifida cystica: a cost-benefit analysis. Br J Prey Soc Med 30: 40, 1976 10. SHEFFIELD U: A cost benefit and cost effective analysis of screening pregnant women in Ontario for neural tube defects. MSc thesis, 1977 11. MILUNSKY A: Prenatal detection of neural tube defects - false positive and negative results. Pediatrics 59: 782, 1977 12. KIMBALL ME, MILUNSKY A, ALPERT
E: Prenatal diagnosis of neural tube defects - III. A reevaluation of the alpha-fetoprotein assay. Obstet Gynecol 49: 532, 1977 13. BERG K, NOER G, MOLNE K, et al:
Maternal serum-amniotic fluid interrelationship: relevance to diagnosis of disorders in the fetus, in Prenatal Diagnosis; Colloque, op cit, p 245 14. SHAPIRO LM, SKINNER LG, PHILLIPS
HV, et al: Racial variation in maternal serum-alpha-fetoprotein (C). Lancet 2: 1142, 1975 15. MACRI JN, WEISs RR, ELLIGERS KW,
et al: Racial differences in maternal serum-alpha-fetoprotein (C). Lancet 1: 207, 1976 16. HAMPTON ML, ANDERSON J, LAvlzzo
BS, et al: Sickle cell "nondisease". Am J Dis Child 128: 58, 1974 17. BUNDEY 5: Screening for spina bifida (C). Lancet 2: 1141, 1974
BOOKS This list is an acknowledgement of books received. It does not preclude review at a later date. ADVANCES IN PEDIATRICS. Vol. 24, 1977. Edited by Lewis A. Barnes, Alfred M. Bongiovanni, Grant Morrow and others. 500 pp. lIlust. Year Book Medical Publishers, Inc., Chicago, 1977. $38.25. ISBN 0-8151-0496-0 CLINICAL PRACTICE AND ECONOMICS. Edited by C.l. Phillips and J.N. Wolfe. 216 pp. lIlust. Pitman Medical Publishing Co. Ltd., Tunbridge Wells, Kent, 1977. Price not stated, paperbound. ISBN 0-27279409-0 EVALUATION REPORT OF THE NUTRITION IN FAMILY PLANNING CENTRES. Demonstration Project, Toronto, Ontario, 1975-1977. Marion Powell and Charlotte Bonds. 206 pp, approx. Illust. University of Toronto, Dept. of Health Administration, Population Unit, Toronto, 1978. $8, spiral-bound EXHIBITIONISM: Facts, Fictions & Solutions. Ruth M. Bray and Alex Gigeroff. 90 pp. lIlust. Clarke Institute of Psychiatry, Toronto, 1978. $3.50, paperbound. Available only from' Dr. Bray, Toronto or Dr. Gigeroff, Yarmouth, NS continued on page 1235
CMA JOURNAL/MAY 20, 1978/VOL. 118
1191
assay of AFP in the maternal serum as a method of screening for neural tube defects. Nineteen centres in the United Kingdom participated in a collaborative study.4 Their findings indicated that 69% of cases of open and closed spina bifida and 88% of cases of anencephaly would have been detectable prenatally between 16 and 18 weeks' gestation on the basis of serum concentrations of AFP equal to or greater than 2.5 times the median for unaffected pregnancies. The study concluded that "screening pregnant women by measuring the concentration of A.F.P. in their serum is an effective method of selecting women for ultrasonography and amniocentesis so that N.T.D.s can be diagnosed in utero." As a result, screening of all pregnancies was scheduled to begin in the United Kingdom in the spring of 1 978. and pressures are rising in Canada to follow suit. We and others5 are concerned about this approach for both technical and ethical reasons. There is variation between laboratories in the use of multiples of standard deviations, percentiles or multiples of the median to determine cut-off points.4 All these measures, however, by definition include some members of the normal population; for example, 1 % of the normal population will be classified as abnormal if the cut-off point is the 99th percentile. In tests where the values for normal and abnormal overlap, as for concentrations of AFP in both amniotic fluid and maternal serum, the price of moving the cut-off point closer to the 100th percentile is to decrease the proportion of abnormalities detected and thus increase the proportion of women with false-negative results. Therefore one must make a clinical and ethical judgement in deciding exactly where the cut-off point should be. In the case of maternal serum AFP tests one can determine the cutoff point by considering how many amniocenteses must be done in women whose fetus is normal compared with women whose fetus has a neural tube defect. We have calcul-
1186 CMA JOURNAL/MAY 20, 1978/VOL. 118
DlpNsone DIPROSONE (betamethasone dipropionate 0,05%) is a topical steroid. Chemically DIPROSONE is betamethasone- 17, 21 dipropionate. INDICATIONS AND CLINICAL USES: DIPROSONE provides anti-inflammatory, antipruritic and anti-allergic activity in the topical management of corticosteroid-responsive dermatoses. CONTRAINDICATIONS: Topical steroids are contraindicated in: 1. Untreated bacterial, tubercular and fungal infections involving the skin, and in certain viral diseases such as herpes simplex, chicken pox, and vaccinia. 2. Hypersensitivity to any of the components. WARNINGS: The safety of topical corticosteroids during pregnancy or lactation has not been established. The potential benefit of topical corticosteroids, if used during pregnancy or lactation, should be weighed against possible hazard to the fetus or the nursing infant. If used under an occlusive dressing, part icularly over extensive areas, sufficient absorption may take place to give rise to adrenal suppression and other systemic effects. Topical corticosteroids are not for ophthalmic use. PRECAUTIONS: Topical corticosteroids should be used with caution on lesions close to the eye. Although hypersensitivity reactions have been rare with topically applied steroids, the drug should be discontinued and appropriate therapy initiated if there are signs of sensitivity. In cases of bacterial infections of the skin, appropriate antibacterial agents should be used as primary therapy. If it is considered necessary, DIPROSONE may be used as an adjunct to control inflammation, erythema, and itching. If a symptomatic response is not noted within a few days to a week, the local applications of DIPROSONE should be discontinued until the infection is brought under control. Significant systemic absorption may occur when steroids are applied over large areas of the body, especially under occlusive dressings. To minimize this possibility, when longterm therapy is anticipated, interrupt treatment periodically or treat one area of the body at atime. Patients should be advised to inform subsequent physicians of the prior use of corticosteroids. Occlusive dressings should not be applied if there is an elevation of body temperature. ADVERSE REACTIONS: When occlusive dressings are used, pustules, miliaria, folliculitis and pyoderma may occur. The following adverse skin reactions have been reported with the use of topical steroids: dryness, itching, burning, local irritation, striae, skin atrophy, hypertrichosis, change in pigmentation, and secondary infection. Adrenal suppression has also been reported following topical corticosteroid therapy. Posterior subcapsular cataracts have been reported following systemic use of corticosteroids. DOSAGE AND ADMINISTRATION: A sufficient quantity of DIPROSONE should be applied to completely cover the affected area and should be massaged gently and thoroughly into the skin. The usual frequency of application is twice daily. For some patients adequate maintenance therapy may be achieved with less frequent application. AVAILABILITY: DIPROSONE Cream and Ointment are available in 20 g and 60 g tubes. Diprosone Lotion is available in 30m1 and 75m1 plastic squeeze bottles. 1 Medansky R. 5. and Handler R. M Analysis ol a New Highly Aclive conicosteroid. Illinois Medical Journal. July. 1976.
Complete information is available to health professionals on request from Schering Corporation Limited, Pointe Claire, Quebec H9R 1B4. MEMR
SCHERING EE9
ated this for women in Ontario for 1 year of screening. Although the birth rate for the province is approximately 125 000 per year, a proportion of pregnant women will fail to seek prenatal care before 18 weeks' gestation and others will refuse the test for personal reasons. We therefore arbitrarily based our calculations on 100 000 women. The frequency of spina bifida and anencephaly in Ontario is approximately 1.28 and 0.76 respectively per 1000 births.6 The cut-off point (2.5 times the median) was taken from the UK collaborative study.4 Of the 128 cases of spina bifida expected per year 40 would be missed, and of the 76 cases of anencephaly expected 9 would be missed; hence the detection rates would be 69% and 88% respectively. We determined that 3448 amniocenteses would be required to detect the 155 fetuses with neural tube defects; therefore about 21 amniocenteses would have to be done on women carrying normal fetuses to detect one fetus with a neural tube defect. In the United Kingdom the estimate was 10:1 with an average frequency of neural tube defects of 4.5 p. 1000 births. One of Wald and colleagues' tables revealed the dramatic effects of both the population frequency and the cut-off points on the number of amniocenteses that need to be done, and hence on the number of women with false-positive results.4 With a cut-off point 2.5 times the median and a hypothetical frequency of neural tube defects of 2 per 1000 births they predicted the 21:1 ratio that we calculated. If we had recalculated the ratio with the cut-off point at three times the median for Ontario, where the frequency of neural tube defects is 2 per 1000 births, the result would have been 10:1 with nearly 2000 fewer amniocenteses. This is the ratio predicted in Wald and colleagues' tables. The cost of obtaining this ratio would be the missing of 11 cases of spina bifida and 3 cases of anencephaly. Thus, each region must decide where its cut-off point should be on the basis of local frequency data and medical and ethical considerations. The concentration of AFP varies during pregnancy and one must be sure of the gestational age to interpret the test result correctly. One cannot assess the effect of misclassification of gestational age from the
UK collaborative study because the patients were included in the study only if there was no doubt about gestational age. If approximately 20% of women are unsure of the date of onset of their last menstrual period7 some of those carrying normal fetuses who underestimate the duration of gestation will have results of the initial serum screening test that are too high; however the correct duration should become apparent when ultrasonography is carried out before amniocentesis. On the other hand, in women who overestimate the duration of gestation the presence of a fetus with a neural tube defect may be missed by the screening test because the presence of these defects may not be detectable before 15 weeks' gestation.8 One way to avoid this would be to perform ultrasonography in all women unsure of their dates. Unless cheaper methods are developed for mass ultrasonography, this would be costly;9 we have calculated the cost to be nearly equal to all the other costs of screening combined.10 It is far from clear what the concentration of AFP in the amniotic fluid would be when women from the general population with high serum concentrations of AFP undergo amniocentesis. The UK collaborative study4 is not helpful here because much of their data were derived from women having amniocentesis at the time the blood sample was drawn. It is assumed that all women with a high serum concentration of AFP who are carrying a fetus with a neural tube defect would have a high concentration of AFP in the amniotic fluid. One would also expect high concentrations of AFP in the amniotic fluid in women whose fetuses have other abnormal conditions including duodenal atresia, omphalocele and Turner's syndrome.3 The most important unknown factor is what proportion of the women with normal fetuses would have a high concentration of AFP in the amniotic fluid. Milunsky" reported that three apparently normal fetuses were electively aborted because of elevated concentrations of AFP in the amniotic fluid and they estimate a falsepositive rate of approximately 0.15% for all tests performed.'2 Thus, in Ontario each year we can expect about five normal fetuses to be incorrectly labelled as abnormal and therefore be
1188 CMA JOURNAL/MAY 20, 1978/VOL. 118
at risk of abortion. By repeating the AFP test and using ancillary diagnostic procedures such as morphologic study of amniotic cells, ultrasonography, amniography and, perhaps in the future, fetoscopy, one may reduce the proportion of falsepositive results. The problem will be to decide which of the 160 cases with abnormal AFP test results (88 of spina bifida + 67 of anencephaly + 5 normal = 160) require the additional tests. Other factors concerning the serum AFP test require further clarification. No correlation (r = 0.17) has been found between serum concentrations and amniotic fluid concentrations of AFP in the nonelevated range.13 Shapiro, Skinner and Phillips'4 found that serum concentrations of AFP were significantly lower in pregnant Asian women than in Caucasian women, while others have shown no difference between black 'and Caucasian women.'5 Not all open neural tube defects cause serious debilitating handicaps. Concentrations of AFP in the amniotic fluid do not distinguish between meningocele and meningomyelocele, or between the mild and severe types of the latter. Neither ultrasonography nor fetoscopy has been developed sufficiently to allow safe and accurate prediction of the severity of the lesion. Given all of the above, we are deeply concerned by the often stated suggestion that such a test should become part of routine obstetric care. When a screening procedure in medicine becomes routine, communication between the screener, whether physician or nonprofessional, and the person screened tends to diminish. Some of the experiences with screening for sickle-cell disease in the United States serve as frightening examples.'6 In the case of neural tube defects we run the risk of providing couples with answers for which they had no question. As Bundey'7 pointed out 3½ years ago, "some of the couples who refuse amniocentesis with its implication that an affected fetus will be aborted will go on to have a child affected with spina bifida and will be reminded of their decision throughout the child's life." Other couples having very strong objections to abortion for any reason may, in the face of an elevated concentration of AFP from a test they did not request or fully understand
Atromid-S * (clofibrate) Indications ATROMID-S is indicated as an adjunct to diet and other measures for the reduction of elevated plasma lipids. Contraindications Pregnancy; lactation; hypersensitivity to ATROMID-S; clinically significant hepatic or renal dysfunction. Warnings Caution should be exercised when oral anticoag ulants are given in conjunction with A ROMID-S. The dosage of the anticoagulant should be reduced, usually by one-half (depending on the individual case), to maintain the prothrombin time at the desired level to prevent bleeding complications. Frequent prothrombin determinations are advisable until it has been definitely determined that the levels have been stabilized. ATROMID-S may displace other acidic drugs, such as phenytc.n and tolbutamide, from binding sites on plasma proteins. In diabetics on sulfonylurea type drugs, ATROMID-S may potentiate hypoglycemia. The safety of this drug in children has not yet been established. Precautions Patients with a history of jaundice or hepatic disease should be treated cautiously. Frequent liver function tests should be performed since the drug may produce functional liver abnormalities, which are usually reversible. If these tests are persistently abnormal, the drug should be discontinued. Complete blood counts should be carried out periodically because anemia, and more frequently, leukopenia have been reported in patients on ATROMID-S. Patients with low serum albumin levels should be treated cautiously. High levels of unbound drug may cause myalgia and raised creatine phosphokinase serum levels. This "flu like" syndrome, which usually disappears when the dose is reduced, has also been observed. Adverse Reactions Nausea may occur in 5% of patients. Less common effects are vomiting, loose stools, dyspepsia, flatulence, and abdominal distress. Headache, dizziness, fatigue, weakness, skin rash, urticaria, pruritus, stomatitis are reported occasionally. A slight transient rise in SGOT and/or SGPT has been observed in some patients without evidence of hepatotoxicity. A few cases of increased BSP retention have been reported. Muscle cramping, aching or weakness have also been reported. A Product Monograph is available on request. Dosage and administration For adults only-two capsules (1 g) twice daily with meals. Availability No.3243-Each capsule contains 500 mg clofibrate in bottles of 100 and 1000.
Ayerst . . 0 Industries, Limited.
R.G. DAVIDSON, MD, FRCP[C] Department of pediatrics L.J. SHEFFIELD, MB, B5, FRACP, DCH
Department of clinical epidemiology and biostatistics McMaster University Medical Centre Hamilton, Ont.
References 1. BROcK DJH, SUTCLTFFE RG: Alphafetoprotein in the antenatal diagnosis of anencephaly and spina bifida. Lancet 2: 197, 1972 2. MILUNSKY A, ALPERT E: Prenatal diagnosis of neural tube defects. I. Problems and pitfalls: analysis of 2495 cases using the alpha-fetoprotein assay. Obstet Gynecol 48: 1, 1976 3. Idem: Prenatal diagnosis of neural tube defects. II. Analysis of false positive and false negative aipha-fetoprotein results. Ibid, p 6 4. WALD NJ, CUCKLE H, BROCK DJH, et al: Maternal serum-aipha-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 1: 1323, 1977 5. CHECK W: Mass screening for open spina bifida needs careful consideration. JAMA 238: 1441, 1977 6. Congenital Anomaly Surveillance System, 1977 data, Ottawa, Health and Welfare Canada 7. BEAZLEY JM, UNDERHILL RA: Con-
finement date unknown. Nurs Times
AYER5T LABORATORIES Division of Ayerst, McKenna & Harrison Limited Montreal, canada Made in canada by arrangement with Imperial chemical
Reg'd
in the first place, submit to abortion and bear the psychologic consequences for as yet unknown periods. Aborting a normal fetus as a result of a false-positive result is a disaster that needs no further comment. In conclusion, we are not against screening for neural tube defects by the use of maternal serum concentrations of AFP. This caveat is intended to emphasize the need for carefully monitored pilot screening programs in several North American centres. Specific questions pertaining to the effect of miscalculation of gestational age, establishment of an appropriate cut-off point, the frequency of misclassification of normal fetuses as abnormal, educational programs for both patients and physicians, and the overall costs and benefits in financial, medical and emotional terms must be meticulously evaluated in prospective studies.
PAAB
.EEEEJ
67: 1414, 1971 8. WALD NJ: The detection of neural tube defects by screening maternal blood, in Prenatal Diagnosis; Cclloque, vol 61, Paris, Inserm, 1976, p 227
9. HAGARD 5, CARTER F, MILNE RG:
Screening for spina bifida cystica: a cost-benefit analysis. Br J Prey Soc Med 30: 40, 1976 10. SHEFFIELD U: A cost benefit and cost effective analysis of screening pregnant women in Ontario for neural tube defects. MSc thesis, 1977 11. MILUNSKY A: Prenatal detection of neural tube defects - false positive and negative results. Pediatrics 59: 782, 1977 12. KIMBALL ME, MILUNSKY A, ALPERT
E: Prenatal diagnosis of neural tube defects - III. A reevaluation of the alpha-fetoprotein assay. Obstet Gynecol 49: 532, 1977 13. BERG K, NOER G, MOLNE K, et al:
Maternal serum-amniotic fluid interrelationship: relevance to diagnosis of disorders in the fetus, in Prenatal Diagnosis; Colloque, op cit, p 245 14. SHAPIRO LM, SKINNER LG, PHILLIPS
HV, et al: Racial variation in maternal serum-alpha-fetoprotein (C). Lancet 2: 1142, 1975 15. MACRI JN, WEISs RR, ELLIGERS KW,
et al: Racial differences in maternal serum-alpha-fetoprotein (C). Lancet 1: 207, 1976 16. HAMPTON ML, ANDERSON J, LAvlzzo
BS, et al: Sickle cell "nondisease". Am J Dis Child 128: 58, 1974 17. BUNDEY 5: Screening for spina bifida (C). Lancet 2: 1141, 1974
BOOKS This list is an acknowledgement of books received. It does not preclude review at a later date. ADVANCES IN PEDIATRICS. Vol. 24, 1977. Edited by Lewis A. Barnes, Alfred M. Bongiovanni, Grant Morrow and others. 500 pp. lIlust. Year Book Medical Publishers, Inc., Chicago, 1977. $38.25. ISBN 0-8151-0496-0 CLINICAL PRACTICE AND ECONOMICS. Edited by C.l. Phillips and J.N. Wolfe. 216 pp. lIlust. Pitman Medical Publishing Co. Ltd., Tunbridge Wells, Kent, 1977. Price not stated, paperbound. ISBN 0-27279409-0 EVALUATION REPORT OF THE NUTRITION IN FAMILY PLANNING CENTRES. Demonstration Project, Toronto, Ontario, 1975-1977. Marion Powell and Charlotte Bonds. 206 pp, approx. Illust. University of Toronto, Dept. of Health Administration, Population Unit, Toronto, 1978. $8, spiral-bound EXHIBITIONISM: Facts, Fictions & Solutions. Ruth M. Bray and Alex Gigeroff. 90 pp. lIlust. Clarke Institute of Psychiatry, Toronto, 1978. $3.50, paperbound. Available only from' Dr. Bray, Toronto or Dr. Gigeroff, Yarmouth, NS continued on page 1235
CMA JOURNAL/MAY 20, 1978/VOL. 118
1191
