Limitations of ultrasound in the diagnosis of fetomaternal haemorrhage


Correspondence: Mr K. Murphy, St Mary's Hospital, London W2 1NY, UK.

Case report

A 27 year old woman booked for confinement at 15 weeks of gestation. Her first pregnancy miscarried at 21 weeks, and her second pregnancy was complicated by fetal hydrocephalus for which no cause was found. In the index pregnancy her antenatal progress was satisfactory until 29 weeks of gestation when she was admitted to the hospital with a 20-hour history of reduced fetal movements. All of her earlier investigations had been normal: maternal serum alpha-fetoprotein 2.02 multiple of median and β-hCG 1.48 multiple of median; routine screening tests including Veneral Disease Reference Laboratory, haemoglobin electrophoresis, hepatitis B, hepatitis C and HIV were all negative and she was immune to rubella. Maternal blood group was B rhesus positive and no atypical blood group antibodies were detected in the booking sample. There was no significant medical history and she was not taking any medication.

The reduction in fetal activity began with the sudden onset of severe low back pain while the woman was lifting a heavy object. She described a strange sensation radiating upwards like a warm gush, followed by shivering and breathlessness. The admission cardiotocograph some 20 hours after this episode was not reassuring: it showed a baseline fetal heart rate of 160 bpm, an absence of accelerations but normal baseline variability (5–10 bpm). An urgent ultrasound scan carried out by the fetal medicine team showed the presence of gross body movements, breathing movements and normal tone, together with a normal amniotic fluid volume (biophysical score 8/10), and normal uterine and umbilical artery Doppler resistance. The woman was aware of fetal activity but felt that it was less vigorous than usual. She received betamethasone to promote fetal lung maturity and an urgent Kleihauer test was performed, which was negative: no fetal red blood cells were observed in the maternal sample. Following the scan, the features on the cardiotocogram deteriorated. Computerised fetal heart rate analysis (using Sonicaid 8002) revealed no episodes of high variation over a 60-minute interval, a baseline heart rate of 160 bpm, short term variation of 4.9 ms, no accelerations, persistent decelerations and an intermittent sinusoidal rhythm. Despite the negative Kleihauer result, we suspected a diagnosis of massive, silent fetomaternal haemorrhage. A normal rate of fetal growth had been demonstrated on ultrasonography during the pregnancy (performed because of her past obstetric history), and the uterus was soft and non-tender. There were no findings suggestive of fetal infection. An immediate delivery by caesarean section under general anaesthesia was undertaken. At operation a live female infant weighing 1300 g was delivered with Apgar scores of 2 and 7 at 1 and 5 minutes, respectively. The amniotic fluid was clear and the pH and blood gas results from umbilical arterial and venous blood were within the physiological range. The cord haemoglobin was 10.5 g/dL, and platelets 210 × 109/L. The infant had no physical signs of blood loss but was deemed to need a blood transfusion when the first neonatal haemoglobin level was recorded at 9.8 g/dL. As the mother was a Jehovah's Witness, a court order was necessary to administer this. Following the blood transfusion the baby's haemoglobin rose to 14.5 g/dL. She required ventilatory support and surfactant for mild respiratory distress syndrome. Erythropoietin and iron were commenced so as to avoid further blood transfusion. Haematological screening of the infant revealed no evidence of coagulopathy, haemoglobinopathy or haemolysis. The baby's blood group was Rhesus A positive with a negative direct antiglobulin test. No further blood transfusions were needed and she showed an excellent reticulocyte response to the erythropoietin. There was no evidence of neurological deficit either clinically or on cranial ultrasound scans, but careful long term follow up will be undertaken.

Maternal and neonatal blood was also taken to test for parvovirus and atypical antibodies, and these results were negative. The maternal serum alpha-feto-protein level, measured in blood taken on the day of delivery, was markedly elevated at 2410 kU/L (peak value should be < 375 kU/L)1. Examination of the placenta was undertaken and no histological abnormality was identified. In particular, there were no focal abnormalities or infarcts.


Spontaneous transplacental passage of fetal erythrocytes into the maternal circulation in the third trimester of pregnancy is common, but the amount of bleeding is normally less than 0.1 mL. Almeida et al.2 considered massive fetomaternal haemorrhage to be ≥ 80 mL of blood since at that level neonatal anaemia appeared in their population. It occurs in about one in 1000 pregnancies and can lead to a maternal transfusion reaction, fetal distress before or during labour, unexplained stillbirth and non-haemolytic neonatal anaemia. Transplacental haemorrhage may be acute or chronic, and acute haemorrhage may be superimposed on chronic bleeding. The clinical effects depend not only on the volume of the bleed, but also on the rate of loss. Experiments in fetal lambs have shown that the fetus is able to tolerate acute loss of up to 40% of blood volume3. The blood pressure is restored within 2 minutes, and the heart rate and 40% of the shed blood volume is restored within 30 minutes. Despite the ensuing increase in the pO2 of umbilical venous blood, the oxygen delivery to the tissues is reduced because of significant haemodilution, but the oxygen delivery to the heart and brain is maintained by increasing the proportion of blood flowing from the umbilical vein through the ductus venosus to the systemic circulation.

This case is important for a number of reasons. Firstly, standard ultrasonography may not be as helpful as cardiotocography in making the diagnosis of fetomaternal haemorrhage. Despite significant fetal anaemia the biophysical score was 8 out of 10 (2 points lost for a non-reactive cardiotocogram), giving false reassurance about fetal wellbeing. Obvious fetal breathing movements were observed as well as extension and flexion movements of the limbs. Hand tone also appeared normal. Middle cerebral and umbilical venous Doppler measurements, which were not done here, may have helped to diagnose fetal anaemia, but such detailed information is not usually available to the clinician managing fetomaternal haemorrhage as an emergency case. Delay in performing a scan may make matters worse in cases where there is continued bleeding. The obstetrician may make the diagnosis of massive, silent fetomaternal haemorrhage earlier by using the Sonicaid 8002 system for computerised fetal heart rate analysis to look for sinusoidal heart rate patterns. It is important to be aware that at least in the early stages of severe anaemia the sinusoidal pattern, as seen in the present case, can be intermittent rather than continuous, and can easily be missed by naked eye examination. It is important to differentiate a sinusoidal fetal heart rate trace from a pseudo sinusoidal trace, which may be completely innocuous4.

Likewise, maternal perception of reduced fetal activity can be a most useful diagnostic marker, especially when the change is dramatic or persistent. Fisher et al.5, treating a case of chronic massive fetomaternal haemorrhage with repeated fetal intravascular transfusions, observed that maternal perception of fetal movements correlated with fetal haemoglobin levels such that the timing of transfusion was guided by fetal movement evaluation.

Secondly, the mother clearly experienced a blood transfusion reaction secondary to massive fetomaternal haemorrhage. Symptoms of severe and sometimes near-fatal transfusion reactions in women suspected to have been exposed to large amounts of ABO incompatible fetal red cells have been reported previously6. These include chills, fever, vomiting, shock, diffuse intravascular coagulation and acute tubular necrosis resulting from rapid intravascular haemolysis with haemoglobinaemia and haemoglobinuria. Our patient described symptoms, which were indicative of a similar, albeit milder, transfusion reaction. It is important that the clinician checks for this reaction when taking a history from a patient complaining of a serious reduction in fetal activity.

Thirdly, the negative Kleihauer test was falsely reassuring, appearing to exclude significant fetomaternal haemorrhage. However, when maternal-fetal ABO incompatibility exists, as in our case, then fetal blood cells in the maternal circulation are rapidly destroyed, leading to the transfusion reaction described above. Finn et al.7 observed that neither the acid elution technique nor agglutination procedures will demonstrate massive fetomaternal haemorrhage in an ABO-incompatible pregnancy, owing to the rapid clearance of any fetal red cells entering the maternal circulation. While the Kleihauer test is a simple and relatively reliable test and the results can be obtained quickly, it is not positive in every case of fetomaternal haemorrhage. Elevated maternal serum alpha-fetoprotein in association with fetomaternal haemorrhage has been reported previously and several reports have suggested that maternal serum alpha-fetoprotein is more reliable and sensitive than the Kleihauer test in the detection of fetomaternal haemorrhage8, especially in the presence of maternal-fetal ABO incompatibility. The maternal serum alpha-fetoprotein level was very elevated in our case confirming the diagnosis of massive fetomaternal haemorrhage. Interestingly, fetal parvovirus infection is also associated with an elevated maternal serum alpha-fetoprotein level and can cause fetal anaemia by erythro-poietic suppression. Parvovirus infection was therefore excluded in both the mother and baby in this case.

In conclusion, massive, silent fetomaternal haemorrhage is a rare condition, but the diagnosis should at least be considered in the clinical setting of reduced or absent fetal movements with or without fetal heart rate changes. It is important to ask the mother about transfusion reaction symptoms. Standard ultrasonography may not help in making an early diagnosis, but computerised fetal heart rate analysis may be useful in identifying early, intermittent sinusoidal fetal heart rate patterns. Where the clinical suspicion is strong, other tests to detect fetal blood in the maternal circulation should be requested, especially if conservative management is being considered. The Kleihauer test may be negative in the presence of massive fetomaternal haemorrhage when there is maternal-fetal ABO incompatibility, and maternal serum alpha-fetoprotein measurement may be a useful backup test. Massive fetomaternal haemorrhage requires immediate delivery of the woman if gestation permits, with blood ready for early neonatal transfusion. Repeated in utero blood transfusion may be needed in a woman whose pregnancy has not attained a gestational age at which the fetus could survive outside the womb.