A new method to determine the feto-placental volume based on dilution of fetal haemoglobin and an estimation of plasma fluid loss after intrauterine intravascular transfusion

Authors

M. Hoogeveen,

Department of Obstetrics, Leiden University Medical Centre, The Netherlands

* Dr J. Egberts, Laboratory of Obstetrics, P3-P, Leiden University Medical Centre, Albinusdreef 2, 2333 ZA Leiden, The Netherlands.

Abstract

Objectives (1) To calculate the feto-placental volume (FPV), using the haematocrit (Ht) values and the percentages of fetal haemoglobin (HbF), before and after red blood cell transfusion. (2) To estimate the transfusion-induced loss of plasma fluid.

Design Retrospective analysis of data of 42 anaemic fetuses at the first transfusion [gestational age (GA) 19–36 weeks].

Setting Department of Obstetrics, Leiden University Medical Centre, The Netherlands.

Sample Fifteen hydropic and 27 non-hydropic fetuses.

Methods Donor blood volume (V_{donor}) and Ht (Ht_{donor}), fetal pre- and post-transfusion Ht values (Ht_{initial}, Ht_{final}) and percentages of HbF (HbF_{initial} and HbF_{final}) were used to calculate the FPV. The total red cell volume after transfusion (RCV_{final}) and Ht_{final} were used to estimate the plasma fluid loss.

Main outcome measures Feto-placental blood volume and loss of plasma fluid.

Results The equations that use Ht_{final} over-estimate the FPV when the formula does not account for the difference between donor and post-transfusion Ht (FPV_{Ht}= 21.36 * GA − 390; r= 0.89). FPV is under-estimated (FPV_{Ht}= 9.90 * GA − 172; r= 0.84) when the blood volume increases with a volume less than the added donor blood volume. The calculation of FPV, using HbF percentages and the initial fetal RCV, is independent of volume changes (FPV_{HbF}= 15.10 * GA − 279; r= 0.85). Comparing RCV_{final} and Ht_{final} values showed that 31.1 ± 14.5% of the transfused volume was lost. Results of the hydropic fetuses did not differ from those of the non-hydropic fetuses.

Conclusions FPV values based on Ht values are less reliable than those based on RCV and HbF findings. When, for practical reasons, Ht values have to be used, we propose an adapted equation for the calculation of the necessary volume of donor blood: V_{donor}= FPV_{HbF}* (Ht_{final}− Ht_{initial}) / (Ht_{donor}− 0.70 * Ht_{final}).

Intrauterine intravascular transfusion (IUT) has become a generally accepted treatment to restore the oxygen delivery capacity in anaemic fetuses. Calculations of the transfusion volume are based on nomograms that account for changes of the feto-placental volume (FPV) during pregnancy. These FPVs were calculated from the change of fetal haematocrit (Ht) values after packed red blood cell transfusions^{1–5}.

During transfusion, the FPV changes. However, the change does not equal the added volume of transfusion blood: Brace^{6} demonstrated in the chronically catheterised sheep fetus that approximately 50% of the added volume is lost 90 minutes after packed cell transfusions because fluid is shifted to the interstitium. If the same phenomenon occurs in the human fetus, the post-transfusion Ht becomes an unstable value. Consequently, the calculation of the FPV by using the post-transfusion Ht within a changing total blood volume leads to an incorrect and under-estimated feto-placental blood volume^{5}.

In infants needing blood transfusion, both post- as well as the pre-transfusion Ht values may be unreliable because of fluctuations in the intravascular plasma volume^{6,7}. Instead of Ht values, it is advised to use different red cell markers for the donor cells and those of the newborn. Then the pre- and post-transfusion red cell volumes (RCVs) can be calculated^{7–10}.

The haemoglobins of fetal (HbF) and adult red blood cells (HbA) differ and are easily separated analytically. The pre- and post-transfusion fetal RCVs can be calculated from the dilution of the HbF percentage after the transfusion and the Ht value of the donor blood. If the pre-transfusion RCV is calculated, then the pre-transfusion Ht value can be used for the calculation of the feto-placental blood volume. Furthermore, the post-transfusion RCV and Ht values can be used to calculate the post-transfusion FPV, and from that the volume of plasma fluid shifted to the interstitium.

In this study, the results of the calculation of the feto-placental blood volume based on pre- and post-transfusion Ht values were compared with those based on the changed HbF values. The fluid loss was estimated and therefore an adapted formula is proposed for the calculation of the volume of donor blood necessary for transfusion.

Methods

For this study, we selected 42 fetuses of which pre- and post-transfusion Ht and HbF values were available at their first transfusion. The fetuses received these transfusions between 19 and 36 weeks of gestation. At the start of the intrauterine treatment, eight fetuses were classified as mildly hydropic and seven as severely hydropic and 27 as non-hydropic^{11}. Fetal anaemia was caused by anti-Rhesus D antibodies (n= 33), and anti-Kell antibodies (n= 7). In two fetuses, other red cell antibodies caused the anaemia.

The intravascular transfusions were performed as described previously^{12}. Before the transfusion, a sample of blood was taken from the umbilical vein to determine the fetal Ht. The volume of donor blood (Ht 0.82 ± 0.01) necessary to increase the Ht to a desired value (0.35–0.45) is derived from a table, based on the FPVs for different gestational ages (GAs) (Nicolaides et al.^{2}). Donor blood is administered at a rate of 3–5 mL/minute. After the transfusion procedure, the needle was flushed with 0.9% saline, and at least 1 minute later, a sample was taken for determination of the post-transfusion Ht.

During the transfusion procedure, fetal Ht was derived from red blood cell counts and MCV values measured with the Sysmex K800 (Charles Goffin, IJsselstein, The Netherlands). The Ht values (or packed cell volumes) of the donor blood were determined by capillary high spin centrifugation. The percentages HbF and HbA were estimated using a Variant haemoglobins testing system (Biorad, Veenendaal, The Netherlands).

The Ht results were used to calculate retrospectively the individual FPVs, which have been described by two different methods^{1,2}.

The RCV_{initial} was then used to calculate the FPV_{HbF}, based on Ht_{initial} results

(4)

After transfusion, the post-transfusion Ht value should equal the total RCV (RCV_{final}) (=RCV_{initial}+V_{donor}* Ht_{donor.}) divided by the post-transfusion FPV (= FPV_{HbF}+F*V_{donor}).

(5)

F is the fraction of V_{donor} that is added to FPV_{HbF} to give the post-transfusion FPV.

(6)

and (1−F)*100 gives the percentage of the added volume that has disappeared from the fetal circulation during transfusion.

In these formulae, V_{donor} is the volume of transfused donor blood, Ht_{initial} is the fetal pre-transfusion haematocrit, Ht_{donor} is the donor haematocrit and Ht_{final} is the final haematocrit post-transfusion. HbF_{initial} and HbF_{final} are the pre- and post-transfusion HbF percentages.

The small sample volumes, taken before and after transfusion, and the effect of the percentage HbF in donor blood (<1%) are neglected in the formulae.

The statistical package SPSS-10 was used for data analysis. Student's t test for paired samples was used in comparing the calculated FPVs [means (SD)]. The relationship between FPV and GA was calculated by regression analysis. The outcomes of the regression analysis are given as estimates and standard errors and differences between the regression estimates for the slopes were analysed^{13}. P values < 0.05 were considered statistically significant.

Results

Figures 1A–C summarise the results of the calculation of the FPV of hydropic and non-hydropic fetuses, using the three different methods (Equations 1, 2 and 4) and plotted against GA. When matched for GA, the calculated FPV values for the hydropic and non-hydropic fetuses did not differ, as has also been shown by Nicolaides et al.^{2}. The results based on Equation 1 (Rodeck et al.^{1}) [Fig. 1A; FPV = 21.36 (±1.72) * GA − 390 (±48); r= 0.89] show a significantly steeper slope for the relationship between FPV and GA than those based on Equation 2 (Nicolaides et al.^{2}) and Equation 4 for HbF dilution (respectively P < 0.0001 and P < 0.01). The slope of the results based on Equation 2 [Fig. 1B: FPV = 9.90 (±1.02) * GA − 172 (±28); r= 0.84] differs also significantly (P < 0.01) from that of Equation 4 [Fig. 1C: (FPV = 15.10 (±1.47) * GA − 279 (±41); r= 0.85].

The comparison for paired results showed that the FPV, based on HbF results [mean volume 137 (82) mL] was significantly larger than that based on Ht values and Equation 2 [100 (54) mL; P < 0.0001]. On the other hand, the FPV–HbF values were always smaller than those calculated by Equation 1 [197 (110) mL; P < 0.0001].

We used the FPV values based on the HbF findings (FPV_{HbF}), the transfused volume, the end-transfusion RCV and the post-transfusion Ht values to calculate the post-transfusion FPV. The post-transfusion FPV is lower than the pre-transfusion FPV plus the transfused volume: only 69.9 ± 14.5% of the added volume is recovered. The absolute loss (L_{a}) of the plasma volume (mL) increased with the added volume [L_{a}=−10.5 (±4.8) + 0.56 (±0.81) *V_{donor}; r= 0.74; P < 0.0001] and the relative loss (L_{f}), expressed as a fraction of the transfused blood volume increased slightly with the transfused volume [L_{f}= 0.105 (±0.89) + 0.0039 (±0.001) *V_{donor}; r= 0.37; P < 0.05] However, stepwise multiple regression analysis with the variables GA, severity of hydrops, the transfused volume V_{donor} and FPV_{HbF} showed that the lost fraction (L_{f}) of the added volume increased only with the FPV before transfusion: [L_{f}= 0.080 (±0.076) + 0.0168 (±0.0005) * FPV_{HbF}; r= 0.49; P < 0.001].

Discussion

In this study, we compared three different methods to calculate the FPV for each individual fetus. In the first two methods, the final Ht values after transfusion were used. With our method, we determined first the initial total fetal RCV, based on the dilution of HbF by HbA during transfusion. It was followed by the calculation of the pre-transfusion FPV. Our method is independent of the final Ht value, contrary to the two other procedures.

The FPVs, determined from the RCV, based on percentages HbF pre- and post-transfusion, are between the volumes based on the equations using post-transfusion Ht values. The FPV values, estimated according to Rodeck et al.^{1}(Equation 1) differ significantly from those worked out with Equation 2 (Nicolaides et al.^{2}). Previously, Mandelbrot et al.^{4} have observed similar differences in FPV values resulting from these two types of calculations. The calculation will over-estimate the FPV if the formula does not take into consideration the difference between donor and post-transfusion Ht (Equation 1) and it will under-estimate it if the end-transfusion FPV has increased with a volume less than the added donor blood volume (Equation 2)^{4,5}.

Transfusion-induced fluctuations in plasma volume do not influence the calculation of FPV, based on dilution of specific red cell markers during transfusion. HbF is in a sense a perfect marker because 86–98% of the haemoglobins in the fetal cell is HbF, whereas in the donor cells it is mostly less than 1%^{14}. The haematocrit values Ht_{initial} and Ht_{donor} are stable values and the fetal pre- and post-RCV values can be determined.

Using sheep, Brace^{6} demonstrated that packed cell transfusions increase the fetal blood volume by smaller amounts than the transfused volume, because fluid is lost from the circulation. We showed that the same phenomenon occurs in the human fetus and calculated that approximately 30% of the transfused volume is lost, probably mainly into the fetal interstitial space, with minor amounts through the placenta^{15}. This fraction is the same as Brace^{6} found in fetal lambs at the end of the transfusion procedure and it increased to 50% at 90 minutes after the transfusion. Furthermore, we found, as Brace^{6} did, that the loss of plasma volume, as a fraction of the added donor blood, was positively related to the initial FPV. This effect was not influenced by the presence or severity of fetal hydrops.

Brace^{6} showed in the fetal lamb that the cardiovascular responses to transfusions are volume, and not GA dependent and the same may account for the human fetus during transfusion. Transfusion increases the fetal plasma endothelin levels. The increase of endothelin levels correlates positively with the volume of donor blood, expressed as the percentage of FPV and with the post-transfusion increase of the umbilical venous pressure^{16}. Then, the loss of fluid across the capillaries will result from an increase of capillary pressure due to the risen post-transfusion venous pressure. We found a negative correlation between FPV and the given volume of donor blood, expressed as the percentage of FPV (r=−0.45; P= 0.03). Thus, the fetuses with the smaller FPV (the younger ones) would get relatively larger volumes of donor blood than those with large FPVs. If the larger relative volume load results in a more increased post-transfusional venous pressure, then the loss of fluid from the vascular system should correlate positively with the volume of donor blood, expressed as the percentage of FPV. Although not statistically significant, we found almost the opposite correlation (r=−0.28; P= 0.07).

The larger relative volume loss in the bigger fetus may indicate that after transfusion, the capillary filtration coefficient becomes higher in the older fetuses than in the younger ones. In the younger fetus, with less capacity to adapt its capillary filtration, a volume load may therefore result in higher volume-induced pressure changes. Then, the younger fetus is at higher risk for volume overload than the older ones.

Because of the changes in the total blood volume during transfusion, the post-transfusion Ht is inaccurate. It also makes the calculation of FPV, based on this value, unreliable. If the FPV value is incorrect, then the required amount of donor blood is also estimated incorrectly. In analogy to Brace^{6}, we suggest therefore the following correction of the equation for the necessary amount of donor blood:

To keep this calculation simple, we have not corrected the ‘0.70’ value for the relationships between the loss of plasma volume and the added volume during transfusion or the FPV before transfusion.

The calculation of FPV, based on HbF measurements, is theoretically the best procedure. There are, however, some disadvantages on using red cell markers. Firstly, it is essential that substantial percentage changes of the red cell marker have occurred during the transfusion procedure. Thus, HbF as a marker for calculating FPV is most accurate at the first IUT. Other cell markers (e.g. RhD, Duffy, Kell, MN antigens^{8–10} or biotin labelling^{17}) can then be used for estimating FPV at subsequent transfusions but the time-consuming analyses will reduce the application for clinical use. Secondly, a sample is often taken during the transfusion to see whether more blood is needed. The Ht value of such an intermittent sample can be measured within a minute by cell counters, whereas the determination of HbF by cation exchange high performance liquid chromatography will take at least 20 minutes. The alkali denaturation^{7} method is faster but 5–10 minutes are still needed for this technique. However, recently developed equipment (Radiometer ABL 730/735) measures besides blood gas parameters, the sample Hb content and calculates the HbF percentage within 90 seconds. If this HbF value is accurate, then it makes the calculation of FPV based on HbF not only of scientific but also of clinical interest.

We conclude that using the difference between pre- and post-transfusion HbF is an appropriate method to calculate and reassess the feto-placental blood volume before and after transfusion. The corrected relationship between FPV and GA can be used for the calculation of the necessary donor red blood cell volumes in treating fetal haemolytic disease.

Acknowledgements

The authors would like to thank their colleagues, who performed the intrauterine transfusions (Professor H. Kanhai and the Drs F. Vandenbussche, F. Klumper and I. van Kamp) for their support. They would also like to thank M. Harvey PhD for the capillary high spin Ht values and Mrs J van Loon for supervising the HPLC measurements. Professor Kanhai and Dr S. Scherjon made helpful comments and Dr D. O. E. Gebhardt improved the English text.