TESTOSTERONE LEVELS IN MIDTRIMESTER MATERNAL AND FETAL PLASMA AND AMNIOTIC FLUID

Amniotic fluid testosterone was significantly higher in male than female fetuses but did not reliably predict fetal sex.

1 PRENATAL DAGNOSS, VOL. 5, (1985) TESTOSTERONE LEVELS N MDTRMESTER MATERNAL AND FETAL PLASMA AND AMNOTC FLUD c. H. RODECK, D. GLL, D. A. ROSENBERG* AND w. P. COLLNS HarrisBirthright Research Centre for Fetal Medicine & Department of Obstetrics and Gynaecology, King s College School of Medicine and Dentistry, London, S.E.S, U.K. SUMMARY Testosterone was measured in maternal plasma (58 samples), amniotic fluid (71 samples) and fetal plasma (55 samples) in 79 patients between 15 and 23 weeks gestation. Maternal plasma testosterone levels were unrelated to fetal sex. Amniotic fluid testosterone was significantly higher in male than female fetuses but did not reliably predict fetal sex. A correct diagnosis of fetal sex was made by testosterone assay of pure fetal plasma in 39 out of 4 males and in 15 out of 15 females using 1.7 nmol/l as the cutoff value. This investigation is not the method of choice for routine fetal sexing but may be of value in fetuses suspected of having certain endocrine disorders. KEY WORDS Fetal sex Fetal plasma testosterone Fetoscopy Fetal intersex NTRODUCTON n early intrauterine life the sexual differentiation of the hypothalamus is directly related to the hormonal environment produced by the gonad. The development of the male phenotype (except for the regression of the Mullerian duct system) depends on the gonadal secretion of testosterone (Jost, 1953). Testosterone is responsible for the differentiation of the Woman duct into epididymis, vas deferens and seminal vesicles (Wilson and Lasnitzki, 1973) and through its peripheral hydrolysation to dihydrotestosterone, the urogenital sinus and urogenital tubercle form the prostate and external genitalia. Several workers have detected raised arnniotic fluid testosterone in the midtrimester in association with a male fetus (Robinson et al., 1977; Nagamani et al., 1979) and it has been suggested (Judd et al., 1976; Zondek et al., 1977; Pirani et al., 1977) that a high amniotic fluid testosterone could be used as a quick way of sexing a fetus. However, others (Dawood and Saxena, 1977), while showing a significant difference in amniotic fluid testosterone (AFT) between male and female fetuses, found too great an overlap to allow accurate sex prediction. Fetal plasma testosterone (FPT) has been studied in cord blood obtained at delivery or hysterotomy (Reyes et al., 1974; Abramovitch et al., 1978), showing a significantly raised level in male fetuses but a fall in fetal plasma *Present address: Department of Obstetrics and Gynaecology, St. George s Hospital, London, S.W. 1, U.K /85/31757$ 1.OO 1985 by John Wiley & Sons, Ltd. Received 3 July 1984 Revised 22 October 984 Accepted 29 October 1984

2 ~ 176 C. H. RODECK ETAL. testosterone from about the 17th week of pregnancy. We decided to study the interrelationship of testosterone levels in fetal and maternal plasma and amniotic fluid obtained in vivo between 15 and 23 weeks and to see which gave the best sex prediction. PATENTS AND METHODS Patients were referred for prenatal diagnosis of Xlinked conditions such as haemophilia (hence the predominance of male fetuses) fi thalassaemia major and sickle cell disease. The pregnancies were dated by ultrasound measurement of the biparietal diameter (Campbell, 1968). Fetoscopy was performed as described previously (Rodeck, 198) between 15 and 23 weeks (mostly 18 to 2 weeks). Amniotic fluid was obtained at the time of fetoscopy, maternal blood samples were taken prior to the procedure and fetal blood was obtained from an umbilical cord vessel. Strict criteria for the purity ofthe fetal blood were observed (Rodeck and Campbell, 1978). Seventynine patients were studied, 53 with male and 26 with female fetuses. Table 1 shows details of the specimens analysed. Of these 79 patients, 44 delivered normal infants (31 male and 13 female). One patient had an intrauterine death of a male fetus three weeks after fetoscopy (normal male testosterone level), and 34 pregnancies (21 male and 13 female fetuses) were terminated. Fetal sex was confirmed at birth by observation or after abortion by observation and histology. Table 1. Testosterone estimations performed on maternal and fetal plasma and amniotic fluid Samples assayed Males Females Total Maternal plasma Amniotic fluid Fetal plasma Maternal plasma Amniotic fluid Amniotic fluid Fetal plasma Maternal plasma Fetal plasma Amniotic fluid Fetal plasma 1 1 Total Assay Each blood sample was collected into lithium heparin tubes, centrifuged immediately and the plasma that was not needed for diagnostic purposes stored at 2OOC. The concentrations of testosterone were measured by radioimmunoassay (Collins

3 ~~ ~ FETAL SEX HORMONES 177 et al., 1972) without knowledge of the fetal sex. Aliquots of 5 pl of plasma and 2 ml of amniotic fluid were used. Students f test was used to assess significance and correlation coefficients were calculated by regression analysis. RESULTS The results.are shown in Table.2 andinfigures 1,2 and 3.. Maternal plasma testosterone (MPT) There was no significant difference in maternal plasma testosterone between male and female fetuses (p <.35) nor did a significant change in mean values in either sex occur between 15 and 22 weeks' gestation (Figure 1). Amnioticfluid testosterone (AFT) AFT levels were significantly higher in male than female fetuses (p<.25). The mean k S.D. for females was.73 f 26 nmol/l and the range was.32 to 1.38 nmol/l, but there was considerable overlap. Again over the gestational range 15 to 22 weeks there was no significant change in mean values in either sex (Figure 2). Table 2. Midtrimester testosterone levels (nmol/l) in maternal plasma, amniotic fluid and fetal plasma Maternal Plasma Amniotic fluid Fetal Plasma SexMeanS.D. Range No. MeanS.D. Range No. MeanS.D. Range No. M 4.56' t ' $ F 4.3* *2$ * not significant. t p <.25. $p<.25. Table 3. Fetal sex prediction by amniotic fluid (AFT) and fetal plasma testosterone (FPT) Correct predictions (%) AFT Cut off.74 nm1/l FPT Cut off 17 XO/ MALES 38/49 (77.5) 39/4 (97.5) FEMALES 15/21 (75) 15/15 (1)

4 C. H. RODECK ETAL o = females 9 c 8 \ 7 E, 6 k t a b a M F M F M F M F M F M F M F M F M Gestation (weeks) Figure 1. Maternal plasma testosterone (MPT) according to fetal sex and gestational age.: males o = femalea. 7 74nmd// M F M F M F M F M F M F M F Gestation (weeks) 4 2 2: F Figure 2. Amniotic fluid testosterone (AFT) according to fetal sex and gestational age 1 P M

5 FETAL SEX HORMONES f \ 7 6 t e /.~nmon/ i L M F M F M F M F M F M Gestation (weeks ) M F M F M Figure 3. Fetal plasma testosterone (FPT) according to fetal sex and gestational age Fetalplasma testosterone (FPT) There was a highly significant difference (p <.25) in FPT levels between male and female fetuses. The mea& &S.D. were 5f5.33 and nmol/l respectively. Values ranged from 1.23 to 28.5 nmol/l in males and from.51 to 1.65 nmol/l in females. No female levels above 1.7 nmol/l were found (Figure 3). There was a tendency for male mean FPT to fall with gestational age and this became significant (p <.5) between 16 and 2 weeks. FPT levels in female fetuses did not change significantly between 16 and 22 weeks. No significant correlation was found between MPT and AFT, MPT and FPT or AFT and FPT. Fetal sex prediction MPT was of no value in discriminating between male and female fetuses. Using an AFT cutoff level of.75 nmol/l, 11 out of 49 males (22.5 per cent) would have been misdiagnosed and 38 out of 49 (77.5 per cent) correctly assigned. Six out of 21 females (25 per cent) would have been misdiagnosed and 15 out of 21 females (75 per cent) correctly predicted. At an FPT cutoff of 1.7Onmol/l, 1 out of 4 males (2.5 per cent) was misdiagnosed and 39 out of 4 (97.5 per cent) were correctly assigned. All female fetuses were correctly diagnosed. FPT was therefore found to be far better in predicting fetal sex. There was virtually no overlap at a cutoff of 17 nmol/l. Only one male fetus, with a gestational age of 22 weeks, had an FPT below this level (1.24 nmol/l).

6 18 C. H. RODECK ETAL. DSCUSSON Our results confirm the work of Warne et al. (1977), Nagamani et al. (1979) and Glass and Klein (1981) that there is no correlation between MPT and the sex of the fetus. Neither were we able to show any significant change in MPT between 15 and 22 weeks gestation. Female AFT appears to represent a basal level which alters little during gestation (Kuntzig et al., 1977). While Frazier et al. (1974) found no difference in AFT between male and female fetuses, most investigators have found significant differences with varying degrees of overlap. Judd etal. (1976) found significantly higher AFT levels in their 12 cases (72 females and 48 males). Zondek ef al. in 54 patients (31 male and 23 female) found no overlap between 14 weeks and 2 weeks. We, like others, found considerable overlap in AFT levels in male and female fetuses in mid pregnancy and the accuracy of sex prediction is not sufficiently reliable to be used in clinical practice (Dawood and Saxena, 1977 ; Kuntzig ef al., 1977). t may be significant that the studies claiming a high level of accuracy (Judd et al., 1976; Zondek et al., 1977) had a higher proportion of female fetuses than in ours and others (Dawood and Saxena, 1977). We found a highly significant difference (p <.25) in FPT levels between male and female fetuses in mid pregnancy with minimal overlap at a level of 1.7 nmol/l. This agrees with the findings of Abramovitch et al. (1978) and Reyes et al. (1974), both of whom measured FPT on aborted fetuses. However, unlike Reyes et al., only one of our male fetuses had an FPT in the female range, and no female fetus had an FPT in the male range, using 1*7Onmol/l as the cutoff. This may be because of our relatively low proportion of female fetuses (14 to 4 as opposed to 33 to 46). Our one incorrect prediction was a male fetus with a FPT of 1.24 nmol/l, which was delivered at term and was a normal male infant. A possible explanation is that this fetus was one of the most mature at the time of sampling (22 weeks) in our study group. Reyes et al. found that after approximately 17 weeks gestation, male FPT concentrations declined to levels indistinguishable from those of female fetuses and did not subsequently change with increasing gestation. This may be related to the falling maternal levels of HCG. At term, male and female FPT in cord blood is similar (Abramovitch, 1974). We were able to show a tendency for the mean FPT levels to fall from 16 to 2 weeks gestation (p c.5), but the number of our more mature fetuses is small. Although the higher male AFT is probably derived from fetal urine and therefore from fetal plasma, we were unable to show a significant correlation between AFT and FPT. While measurement of FPT is a quick and reliable means of determining fetal sex, fetoscopy is needed to obtain pure fetal blood and this carries a fetal mortality rate of two to three per cent (Rodeck, 198). t is therefore not the method of choice for routine fetal sexing. t can occasionally be of value when knowledge of the fetal sex is needed more quickly than is possible by karyotyping cultured amniotic fluid cells, or fetal lymphocytes, or the fetal genitalia cannot be seen by fetoscopy or ultrasound scanning. Another practical application is in the investigation of possible fetal intersex. We have recently done this in a fetus with androgen insensitivity, which had a 46XY chromosome complement, normal male FPT and

7 FETAL SEX HORMONES 181 normal female external genitalia. Another fetus was investigated because the mother had in error been given depot injections of androgens (unpublished observations). Measurement of testosterone and other sex steroids in fetal blood could also be useful in the prenatal diagnosis of congenital adrenal hyperplasia. REFERENCES Abramovitch, D.R. (1974). Human sexual differentiation in utero, J. Obstet. Gynaecol Brit. Cwlth., 81, Abramovitch, D.R., Towler, C.M., Bohn, H.H. (1978). The binding of sex steroids in human maternal and fetal blood at different stages of gestation, J. Steroid Bioch., 9, Campbell, S. (1968). An improved method of fetal cephalometry by ultrasound, J. Obstet, Gynaecol. Brit. Cwlth., 75, Collins, W.P., Mansfield, M.D., Allandina, N.S., Sommerville,.F. (1972). Radioimmunoassay of plasma testosterone, J. Steroid Bioch., 3, Dawood, N.Y., Saxena, B.B. (1977). Testosterone and dihydrotestosterone in maternal and cord blood and in amniotic fluid, Am. J. Obstet. Gynecol., 129,3742. Frazier, S.D., Weiss, B.A.M., Horton, R. (1974). Amniotic fluid testosterone: implications for the prenatal diagnosis of congenital adrenal hyperplasia, J. Pediatr., 84, Glass, A.R., Klein, R. (1981). Changes in maternal serum total and free androgen levels in early pregnancy: Lack of correlation with fetal sex, Am. J. Obstet. Gynecot., 14, Jost, A. (1953). Problems of fetal endocrinology. The gonadal hypophyseal hormones, Recent Prog. Horm. Res., 8,379. Judd, H.L., Robinson, J.D., Young, P.E., Jones, O.W. (1976). Amniotic fluid testosterone levels in mid pregnancy, Obstet. Gynecol., 48, Kuntzig, H.J., Meyer, U., SelmitsRoeckerath, B., Broer K.H. (1977). nfluence of fetal sex on the concentration of amniotic fluid testosterone : Antenatal sex determination? Arch. Gynak., 223,7584. Nagamani, M., McDonough, P.G., Ellegood, J.O., Mahesh, V.B. (1979). Maternal and amniotic fluid steroids throughout human pregnancy, Am. J. Obstet. Gynecol., 134, Pirani, B.B.K., Pairaudeau, N., Doran, A., Young, P.Y., Gardner, H.A. (1977). Amniotic fluid testosterone in prenatal determination of fetal sex, Am. J. Obstet. Gynecol., 129, Reyes, F.J. Boroditicky, R.S.M., Winter, J.S.D., Faimau, C. (1974). Studies on human sexual development. 11. Fetal and maternal serum gonadotrophin and sex steroid concentrations, J. Clin. Endocrinol. Metab., 38, Robinson, J.D., Judd, H.L., Young, P.E., Jones, O.W. Yen, S.C.C. (1977). Amnioticfluid androgens and estrogens in mid gestation, J. Clin. Endocrinol. Metab., 45, Rodeck, C.H. (198). Fetoscopy guided by realtime ultrasound for pure fetal blood samples, fetal skin samples and examination of the fetus in utero, Brit. J. Obstet. Gynaecol., 87, Rodeck, C.H., Campbell, S. (1978). Sampling pure fetal blood by fetoscopy in the second trimester of pregnancy, Brit. Med. J., 2, Warne, G.L., Fainau, C., Reyes, F.J., Winter, J.S.D. (1977). Studies on human sexual development. V. Concentrations of testosterone, 17hydroxyprogesterone and progesterone in human amniotic fluid throughout gestation, J. Clin. Endocrinol., #, Wilson, J.D., Lasnitzki,. (1973). Testosterone uptake by the urogenital tract of the rabbit embryo, Endocrinology, 92, Zondek, T., Mansfield, M.D., Zondek, L.H. (1977), Amniotic fluid testosterone and fetal sex determination in the first half of pregnancy, Brit. J. Obstet. Gynaecol., 84,