α-Fetoprotein, human placental lactogen, and pregnancy-specific β1-glycoprotein in pregnant women who drink: Relation to fetal alcohol syndrome

a-Fetoprotein, human placental lactogen, and pregnancyspecific ~l-glycoprotein in pregnant women who drink: Relation to fetal alcohol syndrome E. Halmesmaki, M.D., I. Autti, M.D., M.-L. Granstrom, M.D., M. Ueikinheimo, M.D., Ph.D., K. O. Raivio, M.D., Ph.D., and O. Ylikorkala, M.D., Ph.D. Helsinki, Finland To determine the usefulness of the assays of maternal serum a-fetoprotein, human placental lactogen, and pregnancy-specific 13,-glycoprotein in prenatal diagnosis of fetal alcohol syndrome, these proteins were followed in 35 problem drinkers (>100 gm of ethanol weekly) and in 14 abstinent women throughout gestation. Thirteen women gave birth to infants with fetal alcohol syndrome, and these women had low levels of a-fetoprotein and pregnancy-specific 13,-glycoprotein but levels of human placental lactogen were normal. Low a-fetoprotein correctly predicted fetal alcohol syndrome in 59% with a relative risk of 2.46. Low pregnancy-specific 13,-glycoprotein predicted fetal alcohol syndrome in 56%, and a relative risk was 3.29. Low a-fetoprotein and pregnancy-specific 13,-glycoprotein may reflect primary or secondary effects of ethanol abuse in pregnancy and appear to be useful in predicting fetal alcohol syndrome. (AM J OBSTET GVNECOL 1986;155:598-601.)

Key words: a-Fetoprotein, human placental lactogen, pregnancy-specific 13,-glycoprotein, maternal drinking, fetal alcohol syndrome

No prenatal diagnostic measures exist for fetal alcohol syndrome (prenatal and/or postnatal growth retardation, facial characteristics, neurological aberrations), one of the most common fetal anomalies encountered today.' In this regard, a-fetoprotein may be of interest, because it is synthesized in fetal liver, which becomes exposed to ethanol in drinking women! Furthermore, it is well established that serum a-fetoprotein is often elevated in pregnant women with fetal death or anomalies (for example, intrauterine deaths, neural tube defects, intestinal atresias, renal malformations).' Moreover, placental proteins, such as human placental lactogen and pregnancy-specific 13,-glycoprotein, which are commonly employed for the biochemical assessment of placental function! might show abnormalities, particularly because ethanol may directly damage the placenta of drinking women. 5 Because no information is available about the concentrations of a-fetoprotein, human placental lactogen, and pregnancy-specific 13,glycoprotein in pregnant problem drinkers, we followed their concentrations throughout the pregnancy and related them to fetal outcome. From the First Department of Obstetrics and Gynecology and The Childrens' Hospital, University of Helsinki, and The Childrens' Castle Hospital. The study was supported by The Finnish Alcohol Research Foundation and Research Department of The Rinnekoti Foundation. Receivedfor publication March 18,1986; accepted May 25,1986. Reprint requests: Erja Halmesmaki, M.D., First Department of Obstetrics and Gynecology, University Central Hospital, Haartmaninkatu 2, 00290 Helsinki, Finland.

598

Material and methods Forty-two patients were referred by local midwives to a special outpatient department for drinking pregnant women between the ninth and twenty-fourth weeks of gestation. After the first alcohol counseling at registration and plea for abstinence during pregnancy, the drinkers were encouraged to visit the department at 2 to 4 weeks' intervals for additional counseling and control. However, 35 women continued to drink alcohol, as evident from their own reports (weekly consumption at least 100 gm of pure ethanol), from reports of nurses at the local antenatal care centers, and/or from determinations of ethanol in blood and urine at each visit, and they were included in the present work. We also recruited 14 abstinent women from the same region who were similarly followed at the same research center (Table I). The drinking and nondrinking groups were comparable in age, parity, and weight, but the drinkers smoked somewhat more commonly (77%) than controls (30%) (Table I). One of the drinkers was epileptic and used 300 mg of carbamazepine daily. In another one the right ovary hac! been removed because of a semimalignant ovarian tumor 1 year before pregnancy, but no recurrence or metastasis had occurred. No drinker or control developed hypertensive or cholestatic pregnancy complications. Thirteen drinkers (37%) gave birth to infants with fetal alcohol syndrome, one of whom was stillborn (Table I). The diagnosis was based on the following criteria: intrauterine growth retardation (birth weight

Fetal alcohol syndrome

Volume 155 Number 3

599

Table I. Clinical characteristics of the study populations (mean ± SD) Women who drank

Maternal age (yr) Parity Weight before pregnancy (kg) Smoking (%) Gestational age at delivery (wk) Birth weight (gm) Placental weight (gm)

Infant with fetal alcohol syndrome (n = 13)

Healthy infant (n = 22)

Control subjects (n = 14)

31.0 ± 6.3 1.8 ± 1.0 63.1 ± 14.1 70 38.7 ± 2.0 2702 ± 47l*t 529 ± 94*

29.4 ± 5.3 1.8 ± l.l 60.5 ± 9.0 80 39.8 ± 1.7 3203 ± 352* 572 ± 89

28.1 ± 5.4 1.9 ± 0.8 56.7 ± 7.2 30 39.3 ± 2.0 3420 ± 390t 643 ± 95*

*p < 0.05. tp < 0.005.

Table II. The sensitivity, specifity, predictive value, and relative risk of low a-fetoprotein, human placental lactogen, and pregnancy-specific 13,-glycoprotein in predicting fetal alcohol syndrome

Sensivity (%) Specifity (%) Predictive value (%) Relative risk

a-Fetoprotein

Human placental lactogen

43 85 59 2.46

93 33 1.06

below the tenth percentile; nine of 13 subjects), postnatal growth retardation (weight and/or height and/or head circumference <2 SD for age; eight of 12), neurological aberrations (10 of 12, with two having only hypotonia), and facial characteristics (12 of 12). In addition, two infants had bilateral clubfoot and one had an extra thumb. One of the infants of drinkers had esophageal atresia and one had unilateral clubfoot but no typical fetal alcohol syndrome; they were not included in the fetal alcohol syndrome group. All the infants born to the control women were healthy. The follow-up extended to at least 6 months for each infant. Venous blood samples were drawn at each follow-up visit. Serum was separated with centrifugation and stored frozen ( - 20° C) until assayed for a-fetoprotein and pregnancy-specific 13,-glycoprotein as described before,6.7 or for human placental lactogen, with use of a commercial radioimmunoassay kit (Nordiclab, Oulu, Finland). All the samples were assayed in the same batch. The interassay variation was between 13% to 15% and the intra-assay variation between 11 % to 14%. The normal ranges (between the tenth and ninetieth percentiles) for a-fetoprotein, human placental lactogen, and pregnancy-specific 13,-glycoprotein were calculated from the data of the abstinent women. Because a comparison with the controls (Mann-Whitney test) revealed that a-fetoprotein and pregnancy-specific 13 ,-glycoprotein were reduced in drinkers, the low levels of these markers (individual measures below the tenth percentile) were considered in relation to fetal alcohol syndrome as follows: true positive (TP) (test low, fetal

7

Pregnancy-specific f3rglycoprotein 60

77 56 3.29

alcohol syndrome positive), false positive (FP) (test low, fetal alcohol syndrome negative), true negative (TN) (test normal, fetal alcohol syndrome negative), and false negative (FN) (test normal, fetal alcohol syndrome positive). Sensitivity of the indices was TP/(TP + FN), specificity was TN/(TN + FP), predictive value was TP/ (TP + FP), and the relative risk factor was the ratio of TP/(TP + FP) to FN/(FN + TN). The significance of the differences between the percentages was studied by binomial t test. Results

Seventy-four percent of the a-fetoprotein concentrations in drinkers with healthy infants fell within the normal range (Fig. 1). The variation was larger if the infant had fetal alcohol syndrome, and the highest values were encountered in the woman with previous semi malignant ovarian tumor (Fig. 1). However, nine of the 13 subjects with fetal alcohol syndrome and nine of the 22 drinkers with healthy infants (69% and 41 %, respectively; p < 0.05) had one or fewer measures below the tenth percentile. The majority of the measures (43%) of the subjects with fetal alcohol syndrome were low (Table II). Low a-fetoprotein predicted fetal alcohol syndrome correctly in 59% with a relative risk of 2.46 (Table II). The levels of human placental lactogen were not affected by maternal drinking (Fig. 2), and occurrence of fetal alcohol syndrome could not be predicted by human placental lactogen assay (Table II). Maternal drinking tended to be accompanied by re-

600

Halmesmaki et al.

September 1986 Am J Obstet Gynecol

400

15.0

&T

300 200

10.0 8.0 6.0

t

150

4.0 100

3.0 2.0

50 40

1.5

::::

30

::::

Ol ::1...

Cl

E

1.0 0.8

..1'

20

11.

J: 1

0.:u.

:::E ~ 10.0 w 8.0 (/)

<

:::E 30 ::)

a: w en 200

6.0 4.0

150

3.0 2.0 1.5 50 40 30 20

16

20

24

28

32

36

40

12

16

20

24

28

32

36

40

GESTATIONAL WEEKS

GESTATIONAL WEEKS

Fig. 1. Concentrations of a-fetoprotein in drinking mothers. Shaded area indicates the normal tenth and ninetieth percentiles .• = Mothers of infants with fetal alcohol syndrome. o = Mothers who drink and have healthy infants. t = Fetal death. '" = Previous semimalignant ovarian tumor.

Fig. 2. Concentrations of human placental lactogen in drinking mothers. Shaded area indicates the normal tenth and ninetieth percentiles.• = Mothers of infants with fetal alcohol syndrome. 0 = Mothers who drink and have healthy infants. t = Fetal death.

duced pregnancy-specific 13,-glycoprotein levels from the twentieth gestational week onward (Fig. 3). This reduction was most marked if the infant had fetal alcohol syndrome (Table II), and in fact, 12 of these 13 infants had at least one measure below the tenth percentile. As only nine of the 22 drinkers with healthy infants had one or fewer pregnancy-specific 13,-glycoprotein measures below the tenth percentile, the difference between the groups was significant (p < 0.0005). The relative risk of fetal alcohol syndrome was 3.29 in drinking women with low pregnancy-specific 13,-glycoprotein, which had a predictive value of 56% (Table II). Three of four women with fetal alcohol syndrome, who had normal a-fetoprotein levels, had at least one subnormal level of pregnancy-specific 13,-glycoprotein. Likewise, the woman with normal pregnancy-specific 13,-glycoprotein levels through pregnancy also had nor-

mal a-fetoprotein levels. Thus the combined use of afetoprotein and pregnancy-specific 13,-glycoprotein revealed fetal alcohol syndrome in all cases except in one with previous semi malignant ovarian tumor.

Comment Although fetal alcohol syndrome and/or "fetal alcohol effects" are a common problem,' no previous efforts to predict them prenatally have been made. This may be partly due to difficulties in performing a prospective study on drinking women, who often totally neglect prenatal care. We have run a special outpatient department for pregnant problem drinkers since 1983, and this gave us an opportunity to evaluate the usefulness of a-fetoprotein, human placental lactogen, and pregnancy-specific 13,-glycoprotein in the prenatal diagnosis of fetal alcohol syndrome. All of our patients drank at least 100 gm of ethanol

Fetal alcohol syndrome 601

Volume 155 Number 3

weekly, and some of them consumed much more alcohol (up to the two bottles of vodka per day during the first half of pregnancy). Thirty-seven percent of these women delivered an infant with fetal alcohol syndrome, and they were characterized by low ex-fetoprotein and pregnancy-specific 131-glycoprotein but normal human placental lactogen. ex-Fetoprotein, which is synthesized in the fetal liver, is present in high concentrations in the fetal circulation and amniotic fluid, wherefrom it is transferred through the placenta and/or membranes to the maternal blood.' Basically, reduced ex-fetoprotein in the drinkers with fetal alcohol syndrome may result either from decreased synthesis and/or transport to the maternal side or from increased elimination of ex-fetoprotein from maternal blood. No data to support the latter possibility are available. Therefore we feel that ethanol and/or its metabolites may have decreased the synthesis and/or transport of ex-fetoprotein and thus led to reduced levels in maternal serum. The low ex-fetoprotein in women with fetal alcohol syndrome is not accounted for by the small size of the fetus or maternal smoking, as these factors are associated with elevated maternal ex-fetoprotein. s-IO Curiously, the woman with a history of semimalignant ovarian tumor had the highest ex-fetoprotein levels, although the infant had typical fetal alcohol syndrome. The high ex-fetoprotein in this woman is probably not connected with the previous tumor because no recurrence was seen during cesarean section, which was performed because of breech presentation. The placental syncytiotrophoblasts produce both human placental lactogen and pregnancy-specific 131-glycoprotein.< Therefore it was of interest that these proteins behaved differently in the drinking women with fetal alcohol syndrome, since only pregnancy-specific 131-glycoprotein levels were reduced. No data suggest an increased metabolism and/or excretion of pregnancy-specific 131-glycoprotein in drinking women, and therefore it is conceivable that pregnancy-specific 131glycoprotein synthesis was specifically reduced by ethanol and/or its metabolites in the heavy drinkers. It was also noteworthy that pregnancy-specific 131-glycoprotein levels in drinkers without fetal alcohol syndrome tended to be slightly reduced. Low maternal ex-fetoprotein and low pregnancy-specific 131-glycoprotein were about equally effective in the prediction of fetal alcohol syndrome. In view of the predictive values (59% to 56%) and relative risks (2.46 to 3.29), these tests can be recommended for routine use in the antenatal care of drinking women. If they are low and/or decreasing in serial determinations and strict abstinence cannot be achieved, the delivery of the infant should be considered as soon as fetal lung maturity can be confirmed. When, and if, the mechanisms

500 400 300 200 150 100 80 60 40 30

::::

Cl

E

..... a..

CI)

20 10 600

~

::l

a: 400 w

CI)

300 200 150 100 80 60 40 30 20 16

20

24

28

32

GESTATIONAL WEEKS

Fig. 3. Concentrations of pregnancy-specific 131-glycoprotein in drinking mothers. Shaded area indicates the normal tenth and ninetieth percentiles .• = Mothers of infants with fetal alcohol syndrome. 0 = Mothers who drink and have healthy infants. t = Fetal death.

by which ethanol damages the fetus are understood in the future, low a-fetoprotein and/or pregnancy-specific 131-glycoprotein in drinking women may lead to additional measures to prevent or ameliorate the development of fetal alcohol syndrome.

REFERENCES 1. Clarren SK, Smith DW. The fetal alcohol syndrome. N EnglJ Med 1978;298:1063-7. 2. Waltman R, Iniquez ES. Placental transfer of ethanol and its elimination at term. Obstet GynecoI1972;40:180-5. 3. Crandall BF. Alpha-fetoprotein: a review. Science 1981;15:127-85. 4. Varner MW, Hauser KS. Human placental lactogen and other placental proteins as indicators of fetal well-being. Clin Obstet Gynecol 1982;25:673-88. 5. Baldwin VJ, MacLeod PM, Benirschke K. Placental findings in alcohol abuse in pregnancy. Birth Defects 1982;18:89-94. 6. Ruoslahti E, Seppala M. Studies of carcino-fetal proteins. III. Development of a radioimmunoassay for alpha-feto-

Halmesmaki et al.

protein: demonstration of alpha-fetoprotein in serum of healthy human adults. 1m] Cancer 1971;8:374-83. 7. Heikinheimo M, Unnerus H.-A, Ranta T,]alanko H, Seppala M. Pregnancy specific beta-I-glycoprotein levels in cholestasis of pregnancy. Obstet Gynecol 1978;52:276-8. 8. Brock D]H, Barron L, Raab GM. The potential of midtrimester maternal plasma alpha-fetoprotein measurement in predicting infants of low birth weight. Br] Obstet Gynaecol 1980;87:582-5.

September 1986 Am J Obstet Gynecol

9. Purdie DW, Young ]L, Guthrie KA, Picton CEo Fetal growth achievement and elevated maternal serum alphafetoprotein. Br] Obstet Gynaecol 1983;90:433-6. 10. Thomsen SG, !sager-Sally L, Lange AP, Saubrey N, Schiolier V. Smoking habits and maternal serum alphafetoprotein levels during the second trimester of pregnancy. Br] Obstet Gynaecol 1983;90:716-7.

Lower genital tract infection with Chlamydia trachoma tis and Neisseria gonorrhoeae in Icelandic women with salpingitis Sigurdur S. Magnusson, M.D.,t Thordur Oskarsson, M.D., Reynir T. Geirsson, M.D., Benedikt Sveinsson, M.D., Olafur Steingrimsson, M.D., and Hannes Thorarinsson, M.D. Reykjavik, Iceland In a study of 225 women with acute salpingitis verified by laparoscopy or laparotomy we investigated the prevalence of gonococcal and chlamydial infection in the lower genital tract. Neisseria gonorrhoeae was isolated from 18.9% of the women and Chlamydia trachomatis from 38.5%. Women with positive cultures were significantly younger (p < 0.01) than those with negative cultures. A trend toward more severe inflammatory changes of the tubes was found in women with positive cultures compared with those with negative cultures. The majority of women with positive cultures stated they had only one sexual partner during the preceding 6 months. Single women had more partners (mean 1.9) than those cohabiting (mean 1.2). The ratio of single/multiple partners for women with Chlamydia was 2.5: 1, and for those with gonorrhea 1 : 1 (p < 0.05). Of the men, 60% could be examined and about 50% had positive cultures. Microbiologic investigation of both partners is necessary in order to prevent reinfection of the woman. (AM J OSSTETGVNECOL 1986;155:602-7.)

Key words: Salpingitis, gonorrhea, chlamydia infections, venereal diseases Acute salpingitis is a serious condition that often leads to permanent tubal damage and infertility. In a majority of cases salpingitis is associated with sexually transmitted organisms, primarily Neisseria gonorrhoeae and Chlamydia trachomatis. '·3 The relative ease with which most sexually transmitted diseases are now cured contrasts with the serious problems resulting from salpingitis such as infertility, ectopic pregnancy, chronic abdominal pain, dysmenorrhea, dyspareunia, and psychological morbidity.'·6 In the past it has been difficult to evaluate the role of sexually transmitted infections in the etiology of salpingitis, because the clinical diagnosis of salpingitis was frequently inaccurate 7. 9 and suitFrom the Department of Obstetrics and Gynecology, National Hospital, the Division of Bacteriology, Department of Pathology, University of Iceland, and the Venereal Diseases Clinic, Reykjavik Health Centre. Received for publication October 23, 1985; revised April 25, 1986; accepted May 12, 1986. Reprint requests: Dr. Reynir T. Geimon, Department of Obstetrics and Gynecology, National Hospital, Reykjavik, Iceland. tDr. MagnUsson died October 21, 1985.

602

able methods for the culture of sexually transmitted organisms have been lacking in many clinical laboratories. At the Department of Obstetrics and Gynecology, National Hospital, Reykjavik, Iceland, it has been policy for a number of years to verify all clinical diagnoses of salpingitis by laparoscopy.8 Improved culture techniques have also been instituted in an effort to obtain more accurate microbiologic identification. We report the findings from a study of the association between salpingitis and sexually transmitted microorganisms at our hospital. Material and methods Patients. The study population consisted of 225 consecutive cases of acute salpingitis verified by laparoscopy or laparotomy and treated in the Department of Obstetrics and Gynecology of the National Hospital, Reykjavik, Iceland, between January 1, 1982, and May 31, 1984. The criteria used for the diagnosis of salpingitis were those described by Westrom 6 and graded as follows: mild-tubes reddened, swollen, covered by pu-