Determining the volume of blood required for the correction of foetal anaemia by intrauterine transfusion during pregnancies of Rh isoimmunised women

Abstract

Background

Severe anaemic foetuses of Rhesus (Rh) isoimmunised mothers are usually treated by intrauterine transfusion (IUT). It is helpful to determine the volume of blood necessary to raise the concentration of haemoglobin by 1.0 g/dL in response to intrauterine transfusions.

Methods

In this cross-sectional, observational study we evaluated 107 first IUT for the correction of anaemia caused by haemolysis triggered by maternal Rh immunisation. The concentration of foetal haemoglobin was determined in umbilical cord blood before and after the IUT. The variation in foetal concentration of haemoglobin after transfusion was compared between groups of hydropic and non-hydropic foetuses, between groups of foetuses with different degrees of anaemia and with groups of gestational age less than or more than 28 weeks. The t-test for averages and ANOVA were used to compare average differences among the groups. p values less than 0.05 were considered statistically significant.

Results

Fifty-five (61.4%) foetuses were found to be anaemic while hydrops was observed in 40 (44%) at the time of the IUT. The volume of red blood cell concentrate infused varied from 5 to 90 mL, with 11.2±1.5 mL being necessary to raise the circulating concentration of haemoglobin by 1.0 g/dL. The foetal response was not influenced significantly by either the degree of foetal anaemia (p=0.56) or the presence of hydrops (p=0.17). The foetuses with a gestational age of 28 weeks or less required a smaller volume of red blood cell concentrate than those with a gestational age of more than 28 weeks (9.3±5.4 mL and 13.4±4.8 mL, respectively; p<0.0001) in order to raise their concentration of circulating haemoglobin by 1.0 g/dL.

Conclusion

The volume of red blood cell concentrate necessary to correct anaemia in pregnancies complicated by Rh isoimmunisation must be considered carefully, since the response to the infusion of blood is peculiar in extremely premature infants. Hydrops and the degree of anaemia were not determinants of the change in the final concentration of circulating haemoglobin following the blood transfusion.

Introduction

Foetal anaemia has many causes which may or may not be related to an immunological process. Haemolytic disease secondary to maternal Rhesus (Rh) isoimmunisation is still the most common cause of foetal anaemia, despite the existence since 1968 of effective prophylaxis through the use of anti-Rh imunoglobin¹. Approximately 10% of isoimmunised women have a child affected in the uterus by severe haemolytic anaemia and proper monitoring of the pregnancy enables identification of the correct time for the treatment of severe anaemia and foetal hydrops².

Even in developed countries, foetal anaemia continues to be a substantial antenatal problem because of the inadequate use of anti-Rh immunoglobin, the presence of other immunising antigens (Kell, Kidd and Duffy) for which there is not yet effective immunoprophylaxis, and other causes of anaemia such as infections, foeto-maternal haemorrhages and foeto-foetal transfusion. For this reason, new methods of diagnosing and treating this disease are necessary^3,4.

Prior to the development of ultrasound technology, the foetal effects of Rh isoimmunisation were detected only at the moment of birth. Currently, sensitised Rh-negative pregnant women are monitored using specific investigations that enable early detection of signs of foetal haemolysis; in addition, the anaemia can be corrected through intrauterine transfusions (IUT) and the ideal moment for the delivery of the affected foetus can be determined^5–7.

As a pregnancy progresses, changes occur in the foetal circulation. The foetal cardiac output is calculated at 210 mL/min in the first half of the pregnancy, 960 mL/min in the 30^th week and 1,900 mL/min in the 38^th week of pregnancy⁸. Systolic blood pressure increases from 15–20 mmHg at 16 weeks to 30–40 mmHg at 28 weeks, while diastolic blood pressure increases from = 5 mmHg at 16–18 weeks to 5–15 mmHg at 19–26 weeks⁹. Venous umbilical pressure increases from 4.5 mmHg at 18 weeks to 6 mmHg at term¹⁰.

In the 10^th week of pregnancy, the average concentration of haemoglobin is 9 mg/dL. At around 22 to 24 weeks, the concentration reaches 14–15 mg/dL and, at the midpoint and at the end of the third trimester, levels in the umbilical cord are approximately the same as those of neonates carried to term, being, on average, 16.6 g/dL. The concentration of haemoglobin in the cord does not vary much in the last trimester of pregnancy, although prior to this it rises gradually¹¹.

Maternal Rh isoimmunisation promotes the passage of IgG type antibodies across the placenta; these antibodies reach the foetal circulation and cause progressive haemolysis. The ensuing anaemia in the foetus leads to heart failure and effusion of fluid into body cavities (foetal hydrops). If no measures are taken to correct the anaemia, progression to death is inevitable. IUT is the only treatment able to correct foetal anaemia, though it carries the risk of volume overload due to the infused blood. This is why it is important to determine the volume of blood to be transfused through an analysis of the degree of anaemia, gestational age and the presence of hydrops and, thereby, achieve an adequate increase in the concentration of haemoglobin for each foetus following the transfusion^12–14.

The aim of this study was to determine the volume of blood that must be infused into an anaemic foetus in order to raise the foetal concentration of haemoglobin by 1.0 g/dL. The roles of gestational age and hydrops in determining the volume that must be transfused were evaluated separately.

Methods

In this cross-sectional, observational study we evaluated 107 first IUT, selected from a cohort of 144 foetuses that underwent 274 cordocentesis procedures because of suspected anaemia. These transfusions took place between August, 1998 and June, 2009, at the Foetal Medicine Centre at the Hospital das Clínicas of the Universidade Federal de Minas Gerais (UFMG), a tertiary referral centre for pregnant, isoimmunised women. The patients were informed of the protocol and consented to participate in the study. The data were analysed retrospectively and no interventions to normal medical conduct were required. The study was approved by the UFMG Research Ethics Committee (ETIC 233/02).

The first evaluation of all isoimmunised pregnant women at our service involved a detailed obstetric history, an indirect Coombs’ test with measurement of the titres, identification - using a red cell panel - of the types of antibodies present, paternal blood typing and ultrasound examinations to detect the presence of hydrops. Following this, foetal anaemia was estimated from amniocentesis or the peak systolic velocity in the foetal middle cerebral artery (PSV-MCA). From 1999 to 2002 we used only amniocentesis to predict foetal anaemia. Amniocentesis was indicated, between 20 and 34 weeks gestation, every 14 to 21 days, for those pregnant women who had an obstetric history of haemolytic disease of the foetus or newborn or had, in the pregnancy being monitored, an indirect Coombs’ test showing a titre = 1:16 or rising titres in two or more dilutions. From 2002 to 2009 all pregnancies complicated by isoimmunisation were managed only on the results of assessment of PSV-MCA. This examination was repeated weekly from week 22 of gestation until delivery.

Cordocentesis was indicated, between 19 and 34 weeks of gestation, based on the findings of ultrasonographic evaluation (presence of hydrops), amniocentesis (amniotic-fluid ADO₄₅₀ values above the line between the middle and the upper third of zone 2 in the modified Liley’s chart) or Doppler examination of PVS-ACM (when this exceeded 1.5 MoM)¹⁵. Foetal haemoglobin concentration was initially measured using the Hemocue® photometric technique on drops of foetal blood taken immediately before and after finishing the transfusion. In addition, about 1 mL of foetal blood was sent for later confirmation of the haemoglobin concentration measured using conventional techniques at the Hospital das Clínicas laboratory. Sixteen cases were later excluded because there was a greater than 20% difference in the haemoglobin concentration measured by the conventional laboratory technique and the rapid test (Hemocue) and/or a difference of less than 1.0 g/dL between the pre- and post-transfusion haemoglobin levels. We also determined the mean corpuscular volume (MCV), with an MCV of over 108 fL being indicative of foetal blood. Gestational age was calculated in weeks, based on an ultrasound performed in the first half of the pregnancy.

IUT was performed when the foetal hemoglobin concentration deficit was 5 g/dL more or the haematocrit was <30%. An initial foetal haematocrit was determined at puncture of the umbilical vein. In our protocol, rather than making two simultaneous punctures (one for sample collection and another for the IUT), a cordocentesis was performed to determine the concentration of foetal haemoglobin, maintaining the cord puncture, with a slow infusion of concentrated red blood cells. Procedures in which the infused volumes were small were retained during the analysis provided that the volumes were greater than 5 mL and the difference between the haemoglobin found and the amount expected for that gestational age was greater than 1.0 g/dL. The IUT protocol was interrupted if the pre-transfusion haemoglobin level was found to be above 12.0 g/dL. In the cases in which IUT was not justifiable, a second sample for dosing of the haemoglobin concentration was obtained prior to immediate suspension of the procedure.

The total amount of red cells to transfuse depends on the initial foetal haematocrit, foeto-placental blood volume, an estimated volaemia of 101 mL/kg ± 13.6 mL¹⁹, haematocrit of the donor unit, the intermediate level of haemoglobin found during the procedure, and the desired final haematocrit. The foetal weight (in grams), estimated using ultrasonography, was multiplied by a factor of 0.02 to determine the volume of red cells to be transfused to achieve a increment in haematocrit of 10%¹⁶. The final target haematocrit in non-hydropic foetuses was 40–50%, whereas in the group of hydropic foetuses, we limited the change in post-transfusion haemotocrit to a 4-fold increase (or a haemotocrit close to 25%) to try to avoid post-transfusion cardiac overload. The red blood cells used for IUT were normally from a blood group O, RhD-negative donor and collected in the 72 hours preceding administration. The blood transfused had no antigens to which the mother was sensitised. Compatibility tests with fresh maternal serum were performed before each IUT. Cells were packed to a haemotocrit of 75–80%. The units underwent gamma irradiation and were leucodepleted by using millipore filters. The blood transfused was introduced slowly into the foetal circulation by injection into an umbilical vein, through a 22-gauge spinal needle.

The timing of subsequent transfusions depended on the rate of haemolysis of the antibodies involved in sensitisation, the volume of adult blood in the foetal circulation (determined by the Kleihauer-Betke test) and the presence of hydrops. The foetus was monitored carefully in order to decide the best time to give the next transfusion. Doppler ultrasonography of the PSV-MCA was performed once weekly between the first and subsequent IUT in non-hydropic foetuses, and twice weekly in those with hydrops. Transfusions were given from 19 to 34 weeks of gestation and Caesarean section was planned between 34 and 35 weeks of gestation. The antenatal use of steroids to promote lung maturity was considered when preterm delivery was necessary.

The variation in foetal haemoglobin in relation to the volume of transfused blood was determined and compared between groups of hydropic and non-hydropic foetuses, groups with different degrees of anaemia, and groups of different gestational ages (= 28 weeks and > 28 weeks). The anaemia was classified as mild (haemoglobin concentration from 0.84 to 0.65 times the median for gestational age), moderate (haemoglobin concentration from less than 0.65 to 0.55 times the median), and severe (haemoglobin concentration less than 0.55 times the median), as defined by Mari et al.¹⁵. Foetal hydrops was considered present when a fluid effusion was found in at least two body cavities or there was skin oedema.

The t-test for averages was employed in order to compare differences between the groups with or without hydrops and those with a gestational age above or below 28 weeks. The ANOVA test was used to compare groups classified according to the degree of anaemia (absent, mild, moderate and severe). MINITAB Inc, version 15 software was used and results were considered statistically significant when the p-value was less than 0.05.

Results

The ages of the 91 pregnant women selected varied from 19 to 41 years (mean, 29.3±5.6 years). The isoimmunisation was due to the D antigen alone in 64.5% of the cases, to D and C antigens in 18.7% of the cases and to associations of D, C, c, Lea, Kell, Fya and other antigens in 16.7% of the cases. In two pregnant women the cause of the sensitisation could not be identified. The gestational age at the time of foetal transfusion varied from 19 to 34 weeks (mean, 27.9±3.6 weeks); in 50 (55%) of cases, the gestational age was less than or equal to 28 weeks.

The gestational age at the first IUT was earlier among foetuses admitted at a gestational age of 28 weeks or less than in those admitted after 28 weeks (at 25.2±2.4 and 31.1±1.9 weeks of gestation, respectively, p<0.0001). IUT were more frequent among foetuses who were admitted before 28 weeks of gestation (3.1±1.4 procedures for those admitted before 28 weeks of gestation, and 1.6±0.7 for those admitted after 28 weeks of gestation, p<0.0001).

At the time of the first transfusion, foetal hydrops was present in 40 (44%) foetuses. Fifty-five (61.4%) foetuses were found to be anaemic: the anaemia was severe in 29 cases (31.9%), moderate in 4 cases (4.4%) and mild in 22 cases (24.2%). There were two cases (2.2%) of persistent, post-transfusion foetal bradycardia requiring emergency Caesarean section, with the outcomes being neonatal death in one case (1.1%) and intrauterine death due to thrombosis of the umbilical cord in the other (1.1%). The foetal mortality rate in this study was 13% (18 of 138 foetuses) and the overall perinatal mortality rate 18.1% (25/138). We could not determine perinatal outcome of four transfused foetuses, so these cases were excluded from this analysis. The incidence of foetal bradycardia in this series was 6.5% (9 out of 138 foetuses).

The volume of concentrated red blood cells infused varied from 5 to 90 mL, with an average foetal response to the infusion of 11.2±1.5 mL necessary to raise the circulating concentration of haemoglobin by 1.0 g/dL. The haemoglobin concentrations, assayed immediately before and after the IUT, and the volumes of blood infused are detailed in table I. The degree of foetal anaemia and the presence of hydrops did not significantly influence the foetal response, based on the volume of red blood cell concentrate required to raise the circulating concentration of haemoglobin by 1.0 g/dL (p=0.56 and p=0.17, respectively). The foetuses with gestational ages less than or equal to 28 weeks responded to the infusion in a distinctly different manner from those with gestational ages of more than 28 weeks. The extremely premature infants required smaller volumes of red blood cell concentrate than foetuses with gestational ages of more than 28 weeks (9.3±5.4 mL and 13.4±4.8 mL, respectively; p<0.0001) in order to raise their concentration of circulating haemoglobin by 1.0 g/dL.

Table I.

Haematological characteristics of the first intrauterine transfusion in 91 cases and the foetal response to the volume infused, divided by group studied

Group	N.	Hb pre-transfusion (g/dL)	Hb post-transfusion (g/dL)	Volume of RBC concentrate infused (mL)	Volume required to increase Hb by 1.0 g/dL	p-value
GA ≤ 28 weeks	50	8.3±3.7	13.4±3.2	40.0±19.3	9.3±5.4	<0.0001
GA > 28weeks	41	10.4±2.9	13.9±3.1	43.7±17.4	13.4±4.8
Hydrops present	40	7.5±3.6	12.4±3.3	45.0±20.3	10.3±4.5	0.17*
Hydrops absent	51	10.6±2.8	14.6±2.7	39.1±16.5	11.4±6.2
Anaemia absent	36	12.7±1.4	16.0±2.2	34.2±13.9	11.6±4.6	0.56**
Anaemia - mild	22	9.4±0.9	13.3±2.6	37.7±18.3	12.1±7.5
Anaemia - moderate	4	7.6±0.4	13.3±2.9	48.0±18.8	9.8±4.0
Anaemia - severe	29	5.0±1.5	11.1±2.4	53.2±18.3	10.1±5.1
Total	91	9.2±3.5	13.7±3.2	41.7±18.4	11.2±2.3

Difference between average volume required to increase the haemoglobin concentration by 1.0 g/dL:

^*t-test of averages,

^**ANOVA GA, gestational age; Hb, haemoglobin.

Discussion

This study shows that the volume of red blood cells required to correct haemoglobin levels differs significantly as a function of gestational age. The choice of using a gestational age of 28 weeks to divide the foetuses into two groups was made because this time is a point of foetal viability as it separates two groups whose statistical difference could be clearly determined and for whom the medical specialist, at the time of the transfusion procedure, could have an estimate of the volume of blood required to raise the concentration of foetal haemoglobin by 1.0 g/dL. The method of calculating the transfusion volume appropriate for the treatment of foetal anaemia in isoimmunised pregnant women has not yet been clearly established. Some attempts have been made to establish parameters for its quantification. In a few studies the volume of blood necessary to correct foetal anaemia was calculated using mathematical formulae based on projections from data such as initial haematocrit¹⁷ or even in an empirical manner^{18, 19}. In all of these studies, the pre-established volume consisted of an approximate value and was verified by measurements of haematocrit or haemoglobin during or at the end of the procedure, to determine whether the volume needed to be adjusted. That is to say, the pre-established calculation had the function of directing clinical conduct. The route of administration has also changed over the years. Transfusions were initially performed in the umbilical artery²⁰, a route that is no longer used.

With regards to the formulae, the parameters used included foeto-placental volume or estimated foetal weight^19–20 which, in addition to involving a considerable margin of error in their calculation, make these equations less practical. Although some authors have demonstrated the importance of foetal weight in determining the vascular bed and consequently volaemia, estimated at 101 mL/kg ± 13.6 mL for a foetus at term²¹, the calculation of foetal weight in clinical practice has limitations and includes a degree of imprecision in pregnancies with this type of complication. One of the parameters used to estimate weight based on ultrasound measurements is the abdominal circumference, which can be greatly altered by the disease itself since the foetal hematopoietic response to the haemolytic process includes extramedullary erythropoiesis in the liver and spleen, which increases abdominal volume significantly²².

When we prospectively studied the echographic evolution of the abdominal circumference of foetuses that had undergone IUT in our centre, we found that this measure is influenced by foetal ascites, by hepatosplenomegaly and by skin oedema²³. There are even cases in which ascites develops early in anaemic foetuses, which negatively affects the calculation of the foetal abdomen, calling the use of estimated foetal weight into question in these circumstances. The calculation of weight becomes imprecise, as it is influenced by this altered parameter, which is considered the main determinant for estimating weight in all the more commonly used formulae for calculating this variable. In the past 20 years we have, therefore, used gestational age and pre-transfusion haematocrit in the calculation of the volume of blood to be transfused. With a prenatal survival rate of 84% in 131 foetuses monitored in the period from 1999 to 2006, our results are comparable to those of large global foetal medicine centres^3,24.

Only the first IUT was considered in this analysis. After the first IUT, the physical characteristics of adult blood infused into the foetal circulation modify the foetal hemodynamic response to anaemia, and so may interfere with the determinations of the blood volume that should be transfused to correct foetal hypoxia. Further studies evaluating the foetal haemodynamic response to subsequent IUT are necessary.

Obviously, groups separated by a gestational age of 28 weeks are not homogeneous with regards to weight, and weight indirectly influences the volume required to correct foetal anaemia, although it does not have to be determined as done by other authors. The volume of concentrated red blood cells necessary to raise the circulating concentration of haemoglobin by 1.0 g/dL was 9.3 ± 5.4 mL in extremely premature infants, while a larger amount was required for foetuses with gestational ages of more than 28 weeks (13.4 ± 4.8mL). Our findings showed that other variables, such as hydrops and the severity of anaemia, had no influence on this foetal behaviour. Based on our results, we consider that it is possible to predict, based on the time of gestation, the volume of red blood cells that needs to be transfused in order to increase the haemoglobin in the foetus, taking into consideration the degree of foetal anaemia. Using a pre-established volume it is possible to improve planning of the treatment and to reduce the time of the invasive procedure. Obviously, this predicted value is for a specific volume, which will give an approximate increase, since the adaptation capacity of individual foetuses is different. This study showed that it is possible to calculate the volume of blood required to correct foetal anaemia based on easy-to-obtain and precise parameters: gestational age and pre-transfusion haemoglobin level. Although the ideal volume of red blood cell concentrate to be infused in an IUT can be estimated, this value should serve as an initial estimate, since there are limitations to the infusion of large volumes and other factors, such as foetal bradycardia during the procedure, need to be considered.