Elevated second-trimester amniotic fluid concentration of insulin-like growth factor binding protein-1 in fetal growth retardation

Elevated second-trimester amniotic fluid concentration of insulin-like growth factor binding protein-l in fetal growth retardation Tiina H. Hakala-Ala-Pietila, MD, Riitta A. Koistinen, PhD, Riitta K. Salonen, MD, and Markku T. Seppala, MD Helsinki, Finland OBJECTIVE: The aim was to study the levels of amniotic fluid and maternal serum insulin-like growth factor binding protein-1 levels at 15 to 16 weeks' gestation with regard to fetal growth in the third trimester. STUDY DESIGN: Amniotic fluid (n = 148) and maternal serum (n = 129) levels of insulin-like growth factor binding protein-1 were prospectively measured. RESULTS: Elevated amniotic fluid insulin-like growth factor binding protein-1 levels were related to fetal growth retardation. At a cutoff value of 3 multiples of the median a high-risk group comprising 16% of all pregnant women was identified. This group included 40% of all small-for-gestational age (birth weight < 10th percentile) and 55% of all very-small-for-gestational-age (birth weight < 5th percentile) infants. Maternal serum insulin-like growth factor binding protein-1 levels were not significantly elevated in these groups. CONCLUSION: In most cases of intrauterine growth retardation the disturbance may exist before the condition is detectable by ultrasonography. An elevated amniotic fluid insulin-like growth factor binding protein-1 level may be an early sign. (AM J OBSTET GVNECOL 1993;169:35-9.)

Key words: Amniotic fluid, growth factor binding protein, fetal growth retardation

Fetal growth retardation usually manifests itself during the third trimester. It would be important to find a marker for prediction of this condition before its clinical appearance to arrange for early surveillance and to learn about the underlying mechanisms. Insulin-like growth factors stimulate cell proliferation and differentiation in target cells through receptors on cell surfaces. I These factors are bound to specific binding proteins that carry them in the circulation. I So far, six different binding proteins have been well characterized. 2 Placental protein 12 has been shown to be the same as insulin-like growth factor binding protein-l (IGFBP-l).3, 4 The insulin-like growth factors are synthesized in a variety of normal tissues, growth hormone being the main regulator of the synthesis of insulin-like growth factor-I. The importance of growth hormone is

From Departments I and II of Obstetrics and Gynecology, Helsinki University Central Hospital. Supported by grants from the Sigrid Juselius Foundation, the Academy of Finland, the University of Helsinki, and the Nordisk Insulin Foundation Committee. Received for publication October 12, 1992; revised January 13, 1993; accepted January 30,1993. Reprint requests: Tiina Hakala-Ala-Pietila, MD, Department I of Obstetrics and Gynecology, Helsinki University Central Hospital, Haartmaninkatu 2, 00290 Helsinki, Finland. Copyright © 1993 by Mosby-Year Book, Inc. 0002-9378193 $1.00 + .20 6/1146063

likely to be less important for fetal growth. 5 Insulin is another regulator of insulin-like growth factor synthesis in man. 6 The synthesis of IGFBP-l is also insulinregulated but independent of growth hormone. 7 During pregnancy, maternal serum IGFBP-l levels rise partly as a result of increased synthesis by decidualized endometrium.8, 9 Both the insulin-like growth factors and the insulinlike growth factor binding proteins are likely to play a role in fetal growth. 10- 13 In normal pregnancy the IGFBP-l concentration in maternal serum and amniotic fluid is highest during the second trimester. 3 , 8 At term pregnancy both the maternal and fetal serum levels bear a negative correlation with birth weight. I I, 14 During pregnancy IGFBP-I is synthesized in the mother and the fetus. 14. 15 Whereas the mechanisms regulating fetal serum IGFBP-l concentrations are not yet known, in nonpregnant individuals elevated levels are related to hypoinsulinemia and disturbed glucose metab0Iism. 7 , 15 In decidua IGFBP-l inhibits binding of insulin-like growth factor-I to its receptor, thus inhibiting the insulin-like growth factor action. 16 There is no suitable clinical or laboratory test for prediction of fetal growth retardation at 16 weeks' pregnancy. The aim of the current study was to learn whether the inverse correlation in term pregnancy between fetal growth and the IGFBP-l level is present in

35

36

Hakala-Ala-Pietila et al.

the second trimester before growth retardation is clinically obvious. Material and methods

Subjects. The study included 189 consecutive pregnant women admitted at the I and II Departments of Obstetrics and Gynecology, Helsinki University Central Hospital, for fetal chromosome analysis because of advanced maternal age (~3 7 years) at 15 to 16 weeks' gestation. This prospective study was approved by the joint Ethical Committee of the Departments, and informed consent was obtained from all women. All the amniotic fluid examinations were carried out at the Helsinki University Central Hospital, but many women were subsequently delivered at other hospitals. All the deliveries took place between 35 and 42 weeks of gestation. Mter delivery all the patients were asked to return a questionnaire that included information on the child's date of birth, sex, weight, maternal disorders during pregnancy, and her smoking habits. Of 189 questionnaires, 153 were returned (81 %). Smoking history was available in 146 (95%) cases. Smoking of ~ 10 cigarettes a day was regarded as heavy smoking. The study group included four women with mild-tomoderate nonproteinuric preeclampsia (diastolic blood pressure 90 to 100 mm Hg), one with mild proteinuria and hypertension, one with White class D diabetes mellitus (well controlled), one with White class NB, and two with White class A diabetes mellitus. Their values did not differ from those of the rest of the women. Amniotic fluid was examined in 148 of these cases (97%), and a fasting maternal serum sample was taken before the amniotic fluid examination in 129 women (84%). Reference values for newborn weight were obtained from 75,061 children born in southern Finland.!7 The child was considered small for gestational age (SGA) when the sex and gestational age-adjusted birth weight was below the normal 10th percentile, and very SGA when the birth weight was below the normal 5th percentile. The child was considered large for gestational age (LGA) when the birth weight was above the normal 90th percentile and average for gestational age (AGA) when it was equal to or between the normal 10th and 90th percentiles. Assays. The IGFBP-l concentration was measured by a specific time-resolved immunofluorometric assay performed, essentially as described earlier,!8 with two monoclonal antibodies No. 6303 and No. 6301 from Medix Biochemica (Kauniainen, Finland).!9 The first antibody was bound to polystyrene microtiter wells, and the second antibody was labeled with europium chelate (Delfia Eu-labelling Kit, Wallac, Turku, Finland). Duplicate samples or standards (25 fLl) and 200 fLl of assay buffer were added to the coated microtiter wells and

July 1993 Am J Obstet Gynecol

incubated at 4° C overnight. The wells were washed, and 50 ng of europium-labeled antibody in 200 fLl of assay buffer was added. Mter 2-hour incubation the wells were washed five times, and 200 fLl of enhancement solution was added. The fluorescence in microtiter wells were measured with an Arcus fluorometer (LKB, Wallac). All the samples were assayed in the same assay. The sensitivity of the assay was 0.1 fLg/L, and the intraassay variation was 5.1 % at the level of 64.1 fLg/L, 6.3% at the level of 12.7 fLg/L, 12.3% at the level of 3.8 fLg/L, and 12% at the level of 0.6 fLg/L. Statistical methods. The results were analyzed by calculation of the median and multiples of the median and of the mean values and SDs after logarithmic transformation of the data. Differences between various groups were analyzed with the Student's t test. The confidence intervals were calculated with the appropriate p value from the standard error of the difference between the means of the two groups analyzed. Analysis of variance was performed by the Kruskal-Wallis test for comparison between IGFBP-l values in women with LGA, AGA, and growth-retarded infants. The X2 test was used for dichotomous (binomial) variables, and, where the minimum estimated expected value was < 5, Fisher's exact test was applied. Results

Amniotic fluid levels ofIGFBP-l. Because the mean IGFBP-l concentration did not significantly differ between 15 and 16 weeks' gestation (28.5 and 27.6 mg/L, respectively; p = 0.4604; one-tailed probability), the results were analyzed in one group comprising both. The 95% confidence interval of the difference between the means of the two groups was small, 0.56 to 1.91 mg/L. In the SGA group the mean amniotic fluid IGFBP-l concentration was 40.5 mg/L, whereas in the non-SGA group (including women with both AGA and LGA children) it was 27.4 mg/L (p = 0.09, I-tailed probability) (Tables I and II). In the very SGA group the mean concentration was 49.4 mg/L, whereas in the non-very SGA group it was 27.4 mg/L (p = 0.04, one-tailed probability) (Tables I and II). Amniotic fluid levels of IGFBP-l > 90th percentile were more frequent in SGA (27%) and very SGA (36%) than in non-SGA (8%) and non-very SGA groups (8%) (Table II, Fig. 1). With a cutoff value of 3 multiples of the median a high-risk group comprising 16% of pregnant women was identified. This group included 40% of all SGA children (with 87% specificity and 26% predictive value of a positive result), and 55% of all very SGA children (with 88% specificity and 26% predictive value of a positive result). Heavy smokers had an SGA child more often than did nonsmokers (31 % vs 9%, P = 0.029), but only one

Hakala-Ala-Pietila et aL

Volume 169, Number 1 Am J Obstet Gynecol

1000

0

0

E' Cl

§. .-

a.. I

37

I

I

0

0

0

100

al

u..

S2

"0

'5 ;;::: ,~

10

(5

'c

E

«

0

0

LGA

ALL

AGA

SGA

VSGA

Fig. 1. Amniotic fluid IGFBP-l levels in various groups and controls, LGA group includes cases with fetal weight> 90th percentile. AGA group includes cases with fetal weight ~ 10th percentile and :5 90th percentile. SGA group includes cases with fetal weight < 10th percentile, and the very SGA (VSGA) group includes cases with fetal weight < 5th percentile. Central boxes, 50% of values; box plus whiskers, 80% of values; open dots, extremes of range. Significant differences were found between following groups: p < 0.05 between SGA and non-SGA (AGA plus LGA) (Fisher's exact test), p < 0.05 between very SGA and non-very SGA (all minus very SGA) (Fisher's exact test), p < 0.05 between SGA, AGA, and LGA (Kruskal-Wallis test), and p < 0.05 between very SGA, AGA, and LGA (Kruskal-Wallis test).

Table I. Amniotic fluid IGFBP-l levels (milligrams per liter) in whole study group, SGA group, and very SGA group at 15 to 16 weeks' gestation Group

Geometric mean

95% Confidence interval of mean

Median

95% Confidence interval of median

All (N = 148) SGA (n = 15) Very SGA (n = 11)

28.5 40.5 49.4

23.9-34.1 22.7-72.2 22.7-107.8

27.6 41.6 88.1

23.5-36.6 16.4-99.4 7.8-146.5

Table II. Differences in amniotic fluid IGFBP-I levels between different groups at 15 to 16 weeks' gestation Group

Geometric mean (mg/L)

SGA

40.5

Non-SGA Very SGA

27.4 49.4

Non-very SGA Heavy smokers

27.4 54.6

Nonsmokers

26.6

95% Confidence interval*

Significance

0.38-1.20

P = 0.090

0.29-1.08 1.06-3.98

IGFBP-l value

P = 0.040

> 90th percentile (%) 27 8 36 8

Significance (Fisher's exact test)

P = 0.048 P = O.ot5

P = 0.034

*Confidence interval of the difference between the means of the two groups.

heavy smoker with an SGA child had an amniotic fluid IGFBP-I level > 90th percentile. However, the mean amniotic fluid level of IGFBP-I was higher in heavy smokers than in nonsmokers (p = 0.0337, two-tailed probability) (Table II). Amniotic fluid levels of IGFBP-I

did not correlate with the corresponding maternal serum levels (r = 0.15, P > 0.05). Maternal serum IGFBP·l levels. The mean serum IGFBP-I concentration did not significantly differ between 15 and 16 weeks (192.5 and 181.3 I1g/L, respec-

38

Hakala-Ala-Pietila et al.

Am J

July 1993 Obstet Gynecol

Table III. Maternal serum IGFBP-I levels (micrograms per liter) in whole study group, SGA group, and very SGA group at 15 to 16 weeks' gestation Group

Geometric mean

95% Confidence interval of mean

Median

95% Confidence interval of median

All (N = 129) SGA (n = 14) Very SGA (n = 10)

192.5 181.3 217.0

169.0-219.2 111.1-298.9 119.1-399.4

237.0 248.8 278.7

198.8-253.3 78.2-376.1 90.7-389.3

Table IV. The differences in maternal serum IGFBP-I levels between different groups at 15 to 16 weeks' gestation Group

Geometric mean (J.I{j/L)

SGA

181.3

Non-SGA Very SGA

192.5 217.0

Non-very SGA Heavy smokerst

188.7 149.9

Non smokers

196.4

IGFBP-l value

95% Confidence interval*

Significance (t test)

0.69-1.64

P = 0.784

0.53-1.44

P = 0.593

0.47-1.23

P = 0.273

> 90th percentile (%)

14 10 20 10

Significance (Fisher's exact test)

P = 0.467 P = 0.297

*Confidence interval of the difference between the means of the two groups. tNone of the heavy smokers had maternal serum IGFBP-l level> 90th percentile. tively, p = 0.4043, one-tailed probability). The 95% confidence interval of the difference between the means was small, 0.64-l.76 J-Lg/L. Therefore the results from the two weeks were combined and analyzed as one group. Unlike in amniotic fluid, no significant differences were observed in the corresponding maternal serum IGFBP-l levels (Tables III and IV). Comment

Previous studies have shown that serum IGFBP-l levels are negatively correlated with newborn weight after 35 weeks of pregnancy."· 20 At this time growth retardation is clinically obvious, and elevated levels are of academic interest with respect to pathophysiologic mechanisms. '2 At 15 to 16 weeks fetal growth retardation is usually not yet clinically manifest. Therefore it was felt necessary to study the potential of IGFBP-I measurement for early detection. The second-trimester amniotic fluid IGFBP-llevel appears to give some lead, but the 40% sensitivity and 87% specificity for an SGA fetus and the 55% sensitivity and 88% specificity for a very SGA fetus are not high enough for routine clinical use. Yet these figures are better than those in some other risk factor models introduced for prediction of low birth weight and perinatal death. 21 • 23 With a higher amniotic fluid IGFBP-l value (e.g., 3.58 multiples of the median) the sensitivity and specificity were similar to those presented by Howell et al. 11 for prediction of fetal growth retardation after 35 weeks by maternal

serum IGFBP-I levels. Although these figures are not efficient enough when used alone, the current study encourages the search for other markers, perhaps other binding proteins that, when used in combination with IGFBP-l, could be more effective. A similar approach has given promising results in screening for Down's syndrome. 24 Unlike the amniotic fluid levels, maternal serum IGFBP-l cannot be used for prediction of fetal growth retardation in the first half of pregnancy. Notwithstanding, the importance of our finding is that changes in IGFBP-I concentration take place at 15 to 16 weeks, before fetal growth retardation becomes clinically apparent, suggesting that insulin-like growth factors play a role in the pathogenesis of growth retardation. The mechanisms leading to high IGFBP-I levels in fetal growth retardation have been discussed. '4 Although the relative contributions of different sources have not been estimated, maternal blood, fetal urine, and maternal decidua obviously contribute to the amniotic fluid IGFBP-I pool. The local sources are likely to be more important than maternal serum, also supported by lack of any correlation between maternal serum and amniotic fluid IGFBP-I levels. An interesting observation was that heavy smokers had higher amniotic fluid IGFBP-l levels than nonsmokers. Although the overall incidence of SGA children was higher in that group, heavy smokers with high amniotic IGFBP-l levels were not delivered of growth-

Hakala-Ala-Pietila et al.

Volume 169, Number 1 Am] Obstet Gynecol

retarded babies. This suggests that smoking itself may be associated with changes in IGFBP-llevels by a so far unknown mechanism. REFERENCES 1. Daughaday WH, Rotwein P. Insulin-like growth factors I and II. Peptide, messenger ribonucleic acid and gene structures, serum, and tissue concentrations. Endocrinol Rev 1989;10:68-91. 2. Shimasaki S, Ling N. Identification and molecular characterization of insulin-like growth factor binding proteins (lGFBP-l, -2, -3, -4, -5, and -6). Prog Growth Factor Res 1991;3:243-66. 3. Koistinen R, Kalkkinen N, Huhtala ML, Seppala M, Bohn H, Rutanen EM. Placental protein 12 is a decidual protein that binds somatomedin and has an identical N-terminal amino acid sequence with somatomedin-binding protein from human amniotic fluid. Endocrinology 1986;118: 1375-8. 4. ]ulkunen M, Koistinen R, Aalto-Setala K, Seppala M, ]anne OA, Kontula K. Primary structure of human insulinlike growth factor-binding protein/placental protein 12 and tissue-specific expression of its mRNA. FEBS Lett 1988;236:295-302. 5. Hill D], Freemark M, Strain A], Handwerger S, Milner RDG. Placental lactogen and growth hormone receptors in human fetal tissues: relationships to fetal plasma hPL concentrations and fetal growth.] Clin Endocrinol Metab 1988;66: 1283-90. 6. Hill D], Milner RDG. Insulin as a growth factor. Pediatr Res 1985;19:879-86. 7. Suikkari AM, Koivisto VA, Rutanen EM, Yki-]arvinen H, Karonen SL, Seppala M. Insulin regulates the serum levels of low molecular weight insulin-like growth factor-binding protein.] Clin Endocrinol Metab 1988;66:266-72. 8. Rutanen EM, Bohn H, Seppala M. Radioimmunoassay of placental protein 12: levels in amniotic fluid, cord blood, and serum of healthy adults, pregnant women, and patients with trophoblastic disease. AM ] OBSTET GYNECOL 1982; 144:460-3. 9. Rutanen EM, Koistinen R, Wahlstrom T, Bohn H, Ranta T, Seppala M. Synthesis of placental protein 12 by human decidua. Endocrinology 1985; 116: 1304-9. 10. Bennett A, Wilson D, Liu F, Nagashima R, Rosenfeld R, Hintz R. Levels of insulin-like growth factors I and II in human cord blood.] Clin Endocrinol Metab 1983;57:60912. 11. Howell R, Perry L, Choglay N, Bohn H, Chard T. Placental protein 12: a new test for the prediction of the small-

12. 13.

14.

15.

16.

17. 18. 19.

20.

21. 22. 23. 24.

39

for-gestational-age infant. Br] Obstet Gynaecol 1985;92: 1141-4. lino K, Sjoberg], Seppala M. Elevated circulating levels of a decidual protein, placental protein 12, in preeclampsia. Obstet Gynecol 1986;68:58-60. D'Ercole A, Hill D, Strain A, Underwood L. Tissue and plasma somatomedin-C/insulin-like growth factor I concentrations in the human fetus during the first half of gestation. Pediatr Res 1986;20:253-5. Wang HS, Lim ], English ], Irvine L, Chard T. The concentration of insulin-like growth factor-I and insulinlike growth factor binding protein-l in human umbilical cord serum at delivery: relation to fetal weight. ] Endocrinol 1991; 129:459-64. Lewitt M, Baxter R. Regulation of growth hormone-independent insulin-like growth factor binding protein (BP28) in cultured human fetal liver explants. ] Clin Endocrinol Metab 1989;69:246-52. Rutanen EM, Pekonen F, Makinen T. Soluble 34K binding protein inhibits the binding of insulin-like growth factor I to its cell receptors in human secretory phase endometrium: evidence for autocrine/paracrine regulation of growth factor action. ] Clin Endocrinol Metab 1988;66: 173-80. Pihkala ], Hakala T, Voutilainen P, Raivio K. Uudet suomalaiset sikion kasvukayrat. Duodecim 1989;105: 1540-6. Koistinen R. Stenman U-H, Alftan H, Seppala M. Time resolved immunofluorometric assay of 34-kDa somatomedin-binding protein. Clin Chern 1987;33: 1126-8. Rutanen E-M, Karkkainen T, Lundqvist C, et al. Monoclonal antibodies to the 27-34K insulin-like growth factor binding protein. Biochem Biophys Res Commun 1988; 105:208-15. Seppala M, Riittinen L, ]ulkunen M, et al. Structural studies, localization in tissue and clinical aspects of human endometrial proteins.] Reprod Fen 1988;(suppl. 36): 12741. Nesbitt RE, Aubry RH. High-risk obstetrics II. Value of semiobjective grading system in identifYing the vulnerable group. AM J OBSTET GYNECOL 1969; 103:972-85. Hobel C], Youkeles L, Forsythe A. Prenatal and intrapartum high-risk screening. Risk factors reassessed. AM J OBSTFT GYNECOL 1979;135:1051-6. Halliday HL, Jones PK, Jones SL. Method of screening obstetric patients to prevent reproductive wastage. Obstet Gynecol 1980;55:565-9. Wald N], Cuckle HS, Densem JW, et al. Maternal serum screening for Down's syndrome in early pregnancy. BM] 1988;297:883-7.