Background
Infants with severe hemolytic disease of the fetus and newborn often require 1 or multiple intrauterine transfusions to treat fetal anemia. Intrauterine transfusions may have an inhibiting effect on fetal and neonatal erythropoiesis.
Objective
To quantify the effect of 1 or multiple intrauterine transfusions on the fetal erythropoiesis by assessing the fetal reticulocyte counts in a population with severe hemolytic disease of the fetus and newborn.
Study Design
This was an observational cohort study in infants admitted to the Leiden University Medical Center who received 1 or multiple intrauterine transfusions for hemolytic disease of the fetus and newborn caused by (Rh)D or Kell antibodies and were born between January 2005 and December 2018.
Results
A total of 235 patients were included, of whom 189 were patients with D-mediated hemolytic disease of the fetus and newborn and 46 with Kell-mediated hemolytic disease of the fetus and newborn. Absolute fetal reticulocyte count in D-mediated hemolytic disease of the fetus and newborn declined exponentially over the course of consecutive intrauterine transfusions, with a 62% decline after 1 intrauterine transfusion (95% confidence interval, 56–67). A similar exponential decline was observed in Kell-mediated hemolytic disease of the fetus and newborn, with 32% (95% confidence interval, 19–45) decline after 1 intrauterine transfusion. This decline was not associated with the varying gestational age at the time of the first intrauterine transfusion or the total number of intrauterine transfusions. The number of red blood cell transfusions for postnatal anemia was greater for infants with D and Kell-mediated hemolytic disease of the fetus and newborn with >2 intrauterine transfusions (median of 3 [interquartile range, 2–3] vs 2 [interquartile range, 1–3], P =.035, in D-mediated disease and median of 2 [interquartile range, 1–2] vs 1 [interquartile range, 1–1], P <.001, in Kell-mediated disease). Infants born after >2 intrauterine transfusions less often required exchange transfusion in D-mediated hemolytic disease of the fetus and newborn (19/89 [21%] vs 31/100 [31%], P =.039), compared with infants with 1–2 intrauterine transfusions.
Conclusion
Treatment with intrauterine transfusions causes an exponential decrease in fetal reticulocyte counts in both D- and Kell-mediated hemolytic disease of the fetus and newborn. Suppression of the compensatory erythropoiesis leads to prolonged postnatal anemia and an increased requirement of red blood cell transfusions after birth.
Hemolytic disease of the fetus and newborn (HDFN) is caused by an incompatibility between maternal and fetal erythrocyte antigens. HDFN is characterized by fetal and neonatal erythroid cell destruction due to maternal alloantibodies, which will induce compensatory erythropoiesis. In case of insufficient compensation, fetal or neonatal anemia may occur and intrauterine treatment with 1 or more red blood cell (RBC) transfusions may be indicated, as well as transfusions for persistent anemia after birth. ,
Why was this study conducted?
This study aimed to quantify the effect of intrauterine transfusions on fetal erythropoiesis in hemolytic disease of the fetus and newborn.
Key Findings
Treatment with intrauterine transfusions causes an exponential suppression of the compensatory erythropoiesis in hemolytic disease of the fetus and newborn and an increasing number of intrauterine transfusions is associated with a higher postnatal transfusion dependency.
What does this add to what is known?
In both D- and Kell-mediated hemolytic disease of the fetus and newborn, infants treated with multiple intrauterine transfusions are more likely to have a high postnatal transfusion dependency.
The exact effects of intrauterine transfusions (IUTs) with adult donor RBCs, which carry a different type of hemoglobin and have different oxygen binding and release characteristics compared with fetal red cells, is not known. In small populations, the effect of IUT(s) on various hematologic parameters such as fetal hematocrit, hemoglobin, leukocytes, and bilirubin has been studied. Reticulocyte counts at birth appeared to be lower in infants with HDFN who received 1 or multiple IUTs compared with infants who were not treated with IUTs, irrespective of the hemoglobin level at birth. , Treatment with IUTs is also associated with a greater number of transfusions after birth compared with infants with HDFN not treated with IUT(s). An inhibiting effect of donor blood on the fetal and neonatal erythropoiesis has been postulated before. The total number of IUTs per infant may be of clinical relevance to select infants at increased risk for a complicated postnatal course.
A wider understanding of the effects of IUTs on fetal and neonatal erythropoiesis is necessary to clarify the pathophysiologic mechanisms underlying both the intrauterine and postnatal course of HDFN. In this study, we specifically aimed to quantify the effect of 1 or multiple IUTs on fetal erythropoiesis by assessing the reticulocyte counts in a large population of fetuses and infants with severe HDFN.
Methods
Study population
All infants admitted to the Leiden University Medical Center (LUMC) who received treatment with 1 or multiple IUT(s) for the treatment of HDFN caused by (Rh)D or Kell antibodies and who were born between January 2005 and December 2018 were eligible for the study. In the Netherlands, all pregnant women are routinely screened for the presence of alloantibodies in pregnancy, and maternal blood samples with a positive screening result are sent to 1 of the 2 national referral laboratories (Sanquin Diagnostic Services or the Special Institute for Blood group Investigations). Thereafter, the clinical relevance of the antibody is evaluated by assessing the antibody specificity and by assessing whether the fetus is antigen-positive. If the fetus is positive, the risk on fetal hemolysis is assessed by serially determining the antibody titer and antibody-dependent cell-mediated cytotoxicity.
Referral to the LUMC as national specialized center is indicated if laboratory parameters are above determined cut-offs. These cut-offs are antibody titers tested in maternal serum ≥16 in D alloimmunization and ≥2 in Kell alloimmunization, or in case of an antibody-dependent cell-mediated cytotoxicity assay ≥50% in D alloimmunization and ≥30% in Kell alloimmunization. Subsequently, these high-risk pregnancies are monitored by serial Doppler measurements to assess the velocity of the blood flow in the middle cerebral artery. If MCA Doppler of the middle cerebral artery exceeds 1.5 multiples of the median or if signs of hydrops are present, treatment with a first or subsequent IUT is indicated. One or more IUTs can be administered until 34–35 weeks of gestation, after which induced delivery is preferred to IUT treatment. The IUT technique used in the Netherlands has been previously described. IUTs consist of irradiated, Parvovirus B19 and Cytomegalovirus seronegative packed erythrocytes, with an increased hematocrit of 0.80–0.85 L/L to minimize the risk of volume overload in the fetus. IUTs are preferably administered intravascularly, either into the placental cord insertion or into the intrahepatic part of the umbilical vein (often in combination with additional intraperitoneal transfusion), depending on the orientation of the placenta. To confirm the suspected fetal anemia, a fetal blood sample is taken before the procedure. Planned delivery at the LUMC and neonatal admission to the neonatal intensive care unit of the LUMC is recommended for all pregnancies after IUT.
Infant and fetal data were excluded from analyses in case of HDFN caused by other alloantibodies than D or Kell and major congenital malformations. Unsuccessful IUTs were defined as transfusion with a volume of less than 5 mL, as the pre-IUT blood sample is 5 mL, and were excluded.
In HDFN mediated by Kell antibodies, IUTs are generally needed earlier in gestation compared with D-mediated disease and erythroid suppression seems to be the predominant mechanism in producing fetal anemia rather than hemolysis. The results of infants with D and Kell-mediated HDFN were therefore reported separately.
Data collection
Data were extracted from the hospital’s patient database, including maternal and neonatal medical files and laboratory outcomes. Follow-up data on transfusions after discharge from the LUMC were collected from referral hospitals. Written consent was obtained from the parents or caregivers, and all personal data were coded before analysis. The following maternal and fetal data were recorded: number of previous births, antenatal intravenous immunoglobulin administration, maximum antibody titer, maternal age at first IUT, fetal gestational age at each IUT, total number of IUTs, volume dosage of IUT, IUT procedure access site (placental cord insertion or intrahepatic transfusion with or without intraperitoneal transfusion), fetal hemoglobin levels and leukocyte, platelet, and reticulocyte counts before every IUT. The following infant data were recorded: sex, gestational age at birth, birth weight, hemoglobin level at birth, reticulocyte count at birth, bilirubin level at birth, number of days of phototherapy, treatment with exchange transfusion (ET), treatment with postnatal RBC transfusion(s) the first 3 months after birth, and the number of postnatal transfusion(s) per infant (also known as “top-up” transfusions).
Referral hospitals receive the protocol for postnatal transfusions of the LUMC after discharge to their center to unify neonatal management. The current transfusion guideline of the department recommends a transfusion in term infants with HDFN when hemoglobin levels fall below 10.5 g/dL (6.5 mmol/L) for day 0–6, below 8.9 g/dL (5.5 mmol/L) for day 7–13, and below 7.2 g/dL (4.5 mmol/L) from day 14 onwards. A transfusion of 15 mL/kg irradiated packed erythrocytes less than 5 days old was advised throughout the study period, with a hematocrit of 0.50–0.65 L/L.
Exchange transfusion in the Netherlands is indicated within 24 hours after birth if the serum bilirubin level is above the cut-off values for exchange transfusion and proceeds to rise despite adequate intensive phototherapy (consisting of 4 phototherapy lamps), or if after 24 hours the bilirubin is above the cut-off values for exchange transfusion.
The study protocol and analysis plan were approved by the ethics committee of the LUMC (G19.041) and scientific committee of our department.
Primary and secondary outcomes
The primary outcome in this study was the suppression of fetal erythropoiesis, as defined by the decline (%) in absolute reticulocyte count (·10 9 /L) per consecutive IUT. The primary outcome was adjusted for the total number of IUTs per infant (as indicator of disease severity) and gestational age in weeks at time of the first IUT as measure for suppression of fetal erythropoiesis. The relative reticulocyte count, as expressed per thousand RBCs (‰), also was reported. Secondary outcomes were the change in leukocytes and platelets per IUT, the reticulocyte counts at birth (absolute and relative counts), ferritin levels at birth, the proportion of neonates requiring ET, and the proportion of neonates requiring RBC transfusion(s) after birth.
Statistical analysis
Data are presented as mean (± standard deviation) or median (interquartile range [IQR]) depending on the underlying distribution. The primary outcome is visualized in a boxplot. A linear mixed model was performed to account for the fact of repeated measurements and allow for covariate adjustment. Effect sizes are reported together with 95% confidence intervals (95% CI). The outcome variable of the model was the absolute reticulocyte count and the predicting variable the IUT number. The total number of IUTs per infant and gestational age in weeks at time of the first IUT were included as potential confounders. Data were log10-transformed to unskew the distribution. Changes on the linear scale for the log10-transformed values correspond to percentage changes on the raw scale. The change in leukocytes and platelets was assessed with a linear mixed model using a random intercept per individual. The reticulocyte counts at birth between the groups after a varying number of IUTs were tested with a Kruskal–Wallis test. The secondary outcomes of proportion of neonates requiring ET and RBC transfusion after birth were tested with a χ 2 test after categorizing the infants in 2 groups: with 1–2 IUTs and >2 IUTs. Statistical analyses were performed using IBM SPSS Statistics 25.0 (IBM Corp, Armonk, NY).
Results
During the studied period, 250 infants treated with IUT(s) for HDFN were born and admitted to the neonatal intensive care unit of the LUMC ( Figure 1 ). HDFN was caused by D alloimmunization (isolated or in combination with other antibodies) in 190 infants and by Kell alloimmunization in 46 infants. One infant with D alloimmunization was excluded due to a single unsuccessful IUT (<5 mL transfused, followed immediately by premature birth at 35 weeks of gestation). In total, 14 infants were excluded because of alloimmunization due to other alloantibodies.
Baseline characteristics
Baseline characteristics of the cohort are presented in Table 1 . The median gestational age at the first IUT was 29.0 weeks (IQR, 24.6–32.1) in D-mediated HDFN and 25.8 weeks (IQR, 22.9–28.6) in Kell-mediated HDFN. The median number of IUTs per fetus was 2 (IQR, 2–4) and 3 (IQR, 2–4).
| Variable | Study population, n=235 | D-mediated HDFN, n=189 | Kell-mediated HDFN, n=46 |
|---|---|---|---|
| Number of previous births, median (IQR) | 1 (1–2) | 2 (1–3) | 1 (1–1) |
| Antenatal IVIg administration, n (%) | 8 (3) | 6 (3) | 2 (4) |
| Maternal age at first IUT, y, mean ± SD | 32.0 ± 4.7 | 31.9 ± 4.7 | 32.3 ± 4.6 |
| Gestational age at first IUT, wk, median (IQR) | 28.0 (24.2–31.7) | 29.0 (24.6–32.1) | 25.8 (22.9–28.6) |
| Number of IUTs per fetus, median (IQR) | 3 (2–4) | 2 (2–4) | 3 (2–4) |
| Maximum antibody titer, median (IQR) | 256 (128–512) | 256 (128–512) | 128 (64–256) |
Hematologic parameters per IUT
Table 2 shows the pooled and unadjusted data of relevant hematologic parameters per consecutive IUT in D-mediated HDFN. IUTs were started at a median gestational age of 29 weeks (IQR, 24.6–32.1). The median transfusion volume was 56 mL (IQR, 37–80) at the first transfusion and increased with gestational age (ie, fetal weight).
| IUT number | 1 (n=189) | 2 (n=148) | 3 (n=89) | 4 (n=48) | 5 (n=16) | 6 (n=3) |
|---|---|---|---|---|---|---|
| Gestational age, wk, median (IQR) | 29.0 (24.6–32.1) | 30.0 (26.4–33.1) | 31.0 (28.7–33.3) | 33.0 (30.9–34.0) | 33.6 (31.1–34.3) | 34.1 a |
| Volume transfused blood, mL, median (IQR) | 56 (37–80) b | 63 (44–77) | 74 (57–89) | 77 (64–90) | 77 (59–85) | 94 a |
| IUT administration route c | ||||||
| Placental cord, n (%) | 127 (67) | 102 (69) | 61 (69) | 34 (71) | 11 (69) | 1 (33) |
| Intrahepatical, n (%) | 2 (1) | 1 (1) | 1 (1) | 0 (0) | 0 (0) | 0 (0) |
| Combination, n (%) | 55 (29) | 41 (28) | 27 (30) | 12 (25) | 5 (31) | 2 (67) |
| Hemoglobin level before IUT, g/dL, median (IQR) d | 6.9 (5.3–8.5) | 7.4 (6.1–8.9) | 7.4 (6.5–8.5) | 8.1 (6.8–8.9) | 8.5 (7.4–9.7) | 8.7 a |
| Reticulocyte count before IUT (‰), median (IQR) e | 177 (121–242) | 34 (2–72) | 3 (2–56) | 2 (2–4) | 2 (1–2) | 2 a |
| Reticulocyte count before IUT (·10 9 /L) – median (IQR) f | 297 (239–390) | 92 (5–176) | 7 (5–143) | 5 (4–12) | 5 (4–7) | 5 a |
b Two missing values (187/189), 2 missing values (146/148), 0 missing values, 0 missing values, 0 missing values, 0 missing values
c groups do not add up to 100% due to 9 missing values
d Two missing values (187/189), 1 missing value (147/148), 1 missing value (88/89), 0 missing values, 0 missing values, 0 missing values
e Twelve missing values (177/189), 10 missing values (138/148), 9 missing values (80/89), 4 missing values (44/48), 1 missing value (15/16), 0 missing values
f Six missing values (183/189), 2 missing values (146/148), 2 missing values (87/89), 0 missing values, 0 missing values.
The median hemoglobin level was low before every IUT but showed a gradual increase from the first IUT (hemoglobin 6.9 g/dL, IQR, 5.3–8.5) to consecutive IUTs, as is seen in fetal hemoglobin with increasing gestational age. The median absolute reticulocyte count before the first IUT was 297·10 9 /L (IQR, 239–390) and showed an exponential decline with 70% to 92·10 9 /L (IQR, 5–176) toward the second IUT and with an additional 28% decline to 7·10 9 /L (IQR, 5–143) toward the third IUT ( Figure 2 ). The relative reticulocyte counts showed a similar decline from 177‰ (IQR, 121–242) before the first IUT to 34‰ (IQR, 2–72) before the second IUT and further declined with consecutive IUTs.
