Maternal serum activin A levels in association with intrauterine fetal growth restriction


* Dr E. M. Wallace, Centre for Women's Health Research, Department of Obstetrics and Gynaecology, Monash University, Monash Medical Centre, 246 Clayton Road, Clayton, Victoria 3168, Australia.


Objective To assess maternal serum activin A as a potential marker of fetal growth restriction.

Design A cohort study.

Setting A maternal–fetal medicine unit, university teaching hospital.

Population Fifty-seven women with a small fetus (less than 10th centile for gestation) referred for assessment of fetal size by ultrasound biometry.

Methods At the time of presentation for fetal biometry, maternal blood was collected for activin A measurement. The case records of each woman were independently reviewed after delivery and the pregnancy grouped into one of three groups: constitutionally small fetus, intrauterine growth restricted (IUGR) fetus or IUGR fetus and maternal pre-eclampsia (IUGR–pre-eclampsia). Activin A levels in the three groups were compared.

Main outcome measures Maternal serum activin A levels.

Results Sixteen of the 57 pregnancies were classified as constitutionally small, 17 as IUGR and 24 as IUGR–pre-eclampsia. Expressed as multiples of a normal median (MoMs), the median (95% CI) activin A level in the constitutionally small pregnancies was 1.12 (0.72–1.39) MoMs significantly lower than the level in both the IUGR pregnancies, 3.00 (1.84–4.11) MoMs, and the IUGR–pre-eclampsia pregnancies, 7.96 (5.73–10.62) MoMs (P= 0.002 and 0.0001 for IUGR vs constitutionally small and IUGR–pre-eclampsia vs constitutionally small, respectively).

Conclusions Maternal serum activin A may be useful in the assessment of the small for gestational age fetus.


Intrauterine fetal growth restriction (IUGR) is a concept defining the fetus that has failed to fulfil its programmed growth potential. IUGR is clinically important because the fetus with IUGR is at increased risk of perinatal morbidity and mortality1,2, risks that can be mitigated by appropriate fetal surveillance and timely delivery3. However, the challenge is in the efficient detection of fetuses with IUGR. While the majority of fetuses with IUGR are small for gestational age, some have birthweights above the 10th centile4 and so would not be identified as small. It is also widely known that many small fetuses remain undetected by clinical examination5–7 and that of those correctly identified as small for gestational age, the majority are constitutionally small, rather than growth restricted8,9. These fetuses are, by definition, healthy and do not normally require additional surveillance or early delivery. Unfortunately, there is no useful screening test for IUGR that can be applied to the general population. While Doppler assessment of the umbilical artery flow velocity waveform, an index of fetoplacental vascular resistance, discriminates well between the healthy small for gestational age fetus and the small at risk IUGR fetus in pregnancies where the fetus is known to be small3,10, it is not useful as a screening tool for an unselected obstetric population11 and may be less useful in pregnancies with a small for gestational age fetus after 35 weeks of gestation12,13. Given these limitations in the detection of the small fetus and the identification of the fetus with IUGR in late gestation, whether it is small or not, other methods of detecting fetal compromise would be of value.

Activin A is a glycoprotein belonging to the transforming growth factor β superfamily14. While initially isolated from gonadal tissue and characterised by its ability to modulate FSH secretion, activin has been identified in many other tissues including the pituitary gland, bone marrow, pancreas and adrenal glands15,16. During pregnancy, the fetoplacental unit is believed to be the major source of activin A in maternal serum15,16. Levels increase progressively throughout pregnancy such that peak levels are attained prior to delivery17,18 giving rise to the suggestion that activin may be a regulator of parturition, particularly preterm labour. This now appears not to be so (reviewed by Tong et al.16). Indeed, clear roles for activin in human pregnancy are yet to be defined. However, acute fetoplacental hypoxia has been shown to stimulate the secretion of activin from the fetoplacental unit in the sheep19, raising the possibility that activin may be a useful marker of fetoplacental compromise. In support of this, maternal serum levels of activin are significantly increased in pre-eclampsia20, a condition characterised by impaired placentation and strongly associated with fetal IUGR, and in a recent study of term deliveries, umbilical artery activin A levels were found to be significantly and inversely related to pH21.

We undertook this study, measuring activin A in the maternal circulation of women with a small for gestational age fetus, to explore whether the level of activin A may be able to distinguish between compromised IUGR and healthy small for gestational age fetuses.


Women with a singleton pregnancy, who were referred to the Maternal–Fetal Medicine Unit for ultrasound assessment of fetal size because of a clinical suspicion of fetal growth restriction and who were confirmed on ultrasound to have a fetus with biometry below the 10th centile for gestation, were recruited to the study. Each woman gave informed written consent and the study was approved by the Monash Medical Centre Human Research and Ethics Committee.

At the time of the initial ultrasound assessment, a single maternal blood sample was taken in a plain tube and the sample was centrifuged at 3500 ×g for 15 minutes at 4°C. Serum was stored at −20°C until assayed. After delivery, the case records of each woman were independently reviewed by two of the authors (AE, EMW), blinded to the activin results and the pregnancy was grouped into one of three groups: constitutionally small fetus, IUGR fetus or IUGR fetus and maternal pre-eclampsia (IUGR–pre-eclampsia). A fetus was determined as constitutionally small if subsequent fetal surveillance and pregnancy outcome was normal; as IUGR if there was evidence of impaired placentation or fetal compromise at the time of presentation or subsequently (including abnormalities in umbilical artery Doppler flow studies, subsequent fetal growth patterns and/or biophysical assessments of fetal wellbeing22) as IUGR–pre-eclampsia if the woman had or developed pre-eclampsia in association with a small for gestation fetus. Pre-eclampsia was defined according to internationally agreed criteria: hypertension (systolic BP ≥140 mmHg and/or diastolic BP [Korotkoff V] ≥90 mmHg) arising after 20 weeks of gestation, proteinuria (≥300 mg/24 hours with or without additional features such as renal insufficiency, elevated liver transaminases and haematological disturbances)23. All women who had a fetus with a structural or chromosomal abnormality were excluded.

Activin A was measured using an established two-site ELISA24 with some minor modifications25. All samples were assayed in one batch. The detection limit was 72 pg/mL and intra- and inter-assay coefficients of variation were 8% and 11%, respectively.

Because maternal serum activin A increases with advancing gestation17,18, levels were expressed as multiples of the normal median (MoM), correcting for gestation and allowing comparisons of levels between groups. The normal median levels were derived from 130 women with a normal singleton pregnancy, as previously reported18. The results for each group were expressed as median and 95% confidence intervals (CI) MoMs. Differences between the groups were analysed using Mann–Whitney U test. Differences were considered significant when P < 0.05.


Between January 1999 and June 2000, 57 women with a clinically small fetus and ultrasound biometry below the 10th centile for gestation were recruited. The median (range) gestation at recruitment for all 57 pregnancies was 33 (23–40) weeks of pregnancy. Of the 57 pregnancies, 16 were classified as constitutionally small, 17 as IUGR and 24 as IUGR–pre-eclampsia. Table 1 summarises the gestations at sampling and delivery, the birthweights and the caesarean section rates for the three groups separately. Both the constitutionally small and IUGR groups were sampled and delivered significantly later than the IUGR–pre-eclampsia group (P < 0.001) but there were no significant differences for gestation at sampling and delivery between the constitutionally small and IUGR groups (P= 0.08 and 0.09, respectively). All birthweights were below the 10th centile for gestation26. The constitutionally small neonates were significantly heavier at birth than the IUGR neonates (P= 0.0005) and the neonates of both of these groups were heavier than the IUGR–pre-eclampsia group (P < 0.0001).

Table 1.  Median (95% CI) gestations at sampling and delivery, birthweights and caesarean section rate for 57 small for gestational age pregnancies. (CS = constitutionally small, IUGR = intrauterine growth restriction, IUGR–PE = intrauterine growth restriction and pre-eclampsia).
Number of pregnancies161724
Gestation at sampling (weeks)37 (35.2–37.9)35.5 (33.5–37.0)27 (26.9–29.9)
Gestation at delivery (weeks)38.5 (37.4–38.7)37.0 (34.0–37.6)29 (27.9–31.3)
Birthweight (g)2496 (2393–2716)1968 (1245–2118)708 (644–1225)
Caesarean section rate (%)31 (8–54)53 (29–77)68 (50–86)

The median (95% CI) activin A level in the 130 normal control pregnancies was 1.00 (0.91–1.11) MoMs. The median (95% CI) activin A level in the pregnancies with a constitutionally small fetus, 1.12 (0.72–1.39) MoMs, did not differ significantly from the controls (P= 0.63). The levels in those pregnancies with IUGR, 3.00 (1.84–4.11) MoMs, and IUGR–pre-eclampsia, 7.96 (5.73–10.62) MoMs, were significantly higher than in both the control and the constitutionally small pregnancies (P= 0.002 and 0.0001 for IUGR vs constitutionally small and IUGR–pre-eclampsia vs constitutionally small, respectively). Figure 1 summarises these data.

Figure 1.

Maternal serum activin A, expressed as MoMs, in 57 pregnancies with a small for gestation fetus. (CS = constitutionally small, IUGR = intrauterine growth restriction, IUGR–PE = intrauterine growth restriction and pre-eclampsia).

Table 2 shows the number (%) of constitutionally small, IUGR and IUGR–pre-eclampsia pregnancies that would be identified at different arbitrary activin MoMs.

Table 2.  Number (percentage, %) of constitutionally small (CS), intrauterine growth restricted (IUGR) and IUGR–pre-eclampsia (IUGR–PE) pregnancies identified with a given activin A MoM above different arbitrary levels.
MoMCS pregnanciesIUGR pregnanciesIUGR–PE pregnancies
0.513 (81)17 (100)24 (100)
1.08 (50)15 (88)24 (100)
1.55 (31)13 (76)24 (100)
2.01 (5)11 (65)23 (96)
2.50 (0)10 (59)21 (88)
3.00 (0)8 (47)20 (83)
3.50 (0)6 (35)20 (83)
4.00 (0)5 (29)20 (83)
4.50 (0)4 (24)19 (79)


Others have previously shown that maternal serum activin A is significantly elevated in women with established pre-eclampsia20,27–29 and in asymptomatic women who subsequently develop pre-eclampsia30. The activin A levels observed in the women with IUGR–pre-eclampsia in our study confirm these reports and reflect the increased placental production of activin in association with pre-eclampsia31. However, in addition, our study has reported maternal serum activin A in pregnancies with isolated small for gestational age fetuses, attempting to differentiate between constitutionally small and IUGR fetuses. Our data suggest that activin A levels are normal in pregnancies with a healthy constitutionally small fetus but are increased threefold in association with fetal IUGR, offering the possibility of a novel test for fetoplacental compromise. These data contrast somewhat with a recent report of similar maternal serum activin A levels in 18 pregnancies with an idiopathic small for gestational age fetus and 22 gestation matched normal controls (median activin A: 6.8 and 8.0 ng/mL, respectively)29. However, in that study, the mean gestation at delivery of the small for gestational age pregnancies was 38.8 weeks and the mean birthweight was 2438 g, outcomes almost identical to those of the constitutionally small group described here (Table 1). It would therefore appear likely that the small for gestational age group described in that study were women with a normal healthy small baby, supporting our findings that in this group activin A levels are not increased. Indeed, Bobrow et al.32 recently reported that maternal serum activin A levels were significantly higher in women with a small fetus and abnormal umbilical artery Doppler studies than in women with a similarly sized fetus and normal Doppler studies. These data are consistent with our observations that activin A levels are increased in association with chronic fetal compromise.

The timely detection of the fetus with IUGR allows the institution of increased fetal surveillance and timely delivery with the aim of avoiding fetal loss and long term morbidity. However, as many as half of small for gestational age fetuses may be missed by traditional antenatal examination5–7, making it difficult to target that surveillance. Furthermore, because there are no uniform features of IUGR, it can be difficult to discriminate between the healthy small for gestational age fetus that is constitutionally small and the compromised IUGR fetus. While the small for gestational age fetus with abnormal umbilical artery Doppler studies is clearly growth restricted and has a high risk of mortality and morbidity1–3,10, it is also apparent that the small fetus with normal umbilical artery Doppler studies in late pregnancy may also be significantly compromised12,13,33. This is important because the relative risk of cerebral palsy in the growth-restricted fetus is higher for those at term than it is for those delivered preterm34. A test that can efficiently discriminate between the healthy and unhealthy small fetus in late pregnancy would therefore be of significant clinical value. Currently, complex additional ultrasound studies involving serial fetal biometry, uterine artery and fetal vessel Doppler waveforms, amniotic fluid volume measurements, placental grading and fetal behavioural studies are used to attempt to discriminate between constitutionally small and IUGR fetuses20. However, serial biometry has a significant false positive rate35 often leading to unnecessary intervention9 and the other tests require the facilities and expertise of tertiary referral centres. Accordingly, the observation here, and by others32, that activin A is significantly elevated, on average, in IUGR pregnancies but not in constitutionally small pregnancies is promising indeed, heralding the potential development of a simple non-invasive test of fetal compromise in those pregnancies with a confirmed small fetus. These data clearly require confirmation by more extensive evaluation but the potential applications are clear if confirmed.

Of course, given the difficulties in identifying the small fetus and that many fetuses with IUGR may not be ‘small’4, a screening test for IUGR/fetal compromise that could be applied to the entire pregnant population would be even more valuable. This study has not addressed whether activin A could be such a marker but the observation that maternal serum activin A is increased, albeit moderately, in early pregnancy in women who subsequently have a small for gestational age fetus29 or who develop pre-eclampsia29,30 offers further possibility that activin A may also be a useful predictor of future IUGR and pre-eclampsia rather than only as a test of established compromise as demonstrated here. This too merits further evaluation.

Why maternal serum activin A is increased in association with IUGR remains unclear but it is most likely due to increased placental production, as in pre-eclampsia31, possibly secondary to fetoplacental hypoxia19. The observation that umbilical artery activin A is inversely related to pH also raises the possibility of a fetal contribution21. However, fetal activin A levels are approximately a tenth of maternal levels21,36 making this less likely. Whatever the mechanism, the findings of this study suggest that activin A may be a novel marker of fetoplacental wellbeing in late pregnancy and offers the potential for a non-invasive test that may be applied to the pregnant population at large.


This study was funded in part by a project grant from the National Health and Medical Research Council (NHMRC) and by an Ella Macknight Research Scholarship from the Royal Australian and New Zealand College of Obstetricians and Gynaecologists to EMW.