Hackney, David N. MD†; Knudtson, Eric J. MD*; Rossi, Karen Q. RN*; Krugh, Dave MT(ASCP)SBB‡; O’Shaughnessy, Richard W. MD*
The widespread use of Rh-D immune globulin has led to a relative increase in the importance of non–Rh-D isoimmunization as a cause of hemolytic disease of the newborn.1–3 The Rh antigen system is mainly composed of the antigens C, c, D, E, and e. The c antigen is a common cause of isoimmunization.4–14 Its relative ability to cause clinically significant hemolytic disease of the newborn, however, has been the focus of debate.13
Most practitioners, including those at our institution, manage anti-c isoimmunization in a manner similar to that for anti-D, without published support in the literature. The objective of our study was to review available cases of anti-c isoimmunization at our institution, identify cases of clinically significant disease, and determine whether management similar to that for anti-D is appropriate.
The Ohio State University isoimmunization program provides laboratory and consultative services for the detection and management of pregnancies complicated by isoimmunization. The Fetal Therapy Program at The Ohio State University has maintained a database of pregnancies affected by isoimmunization, with 1946 pregnancies followed at our institution since 1967.
MATERIALS AND METHODS
A review of our database was performed for the years 1967–2001 for pregnancies complicated by anti-c isoimmunization. Approval for this study was obtained from The Ohio State University institutional review board. Once identified, case records were assessed for completeness of their data. The following had to be present for a pregnancy to be included: 1) maternal serum antibody titer analyzed at The Ohio State University Medical Center Prenatal Reference Laboratory, 2) amniotic fluid ΔOD450 value and gestational age at the time of sampling, if amniocentesis was performed, 3) the nature and timing of any cordocenteses or intrauterine transfusions, including fetal hemoglobin results, 4) confirmation of an affected fetus through fetal or neonatal direct antiglobulin or antigen testing, 5) the initial neonatal hemoglobin level and newborn in-patient history of simple or exchange transfusions, and 6) the absence or presence of hydrops fetalis.
Only titers performed at our institution were included in the analysis, to eliminate any interlaboratory variation. When performing serologic investigation (ie, saline indirect antiglobulin test, antibody titer), The Ohio State University Medical Center Prenatal Reference Laboratory follows published guidelines endorsed by the American Association of Blood Banks15 or the equivalent standard of its time. Internal quality control is performed with each titer, and the procedure has not changed during the course of this period. The patient’s previous titer is repeated with each new titer, and a difference of two dilution changes is considered a significant difference. In addition, titers that were obtained after amniocentesis or cordocentesis were excluded from the analysis because they might have been elevated by these procedures. The critical titer is defined as the antibody titer that detects all cases of significant hemolytic disease of the newborn in patients with a first affected pregnancy. First affected pregnancy is defined as the first pregnancy after the identification of an antibody associated with hemolytic disease of the newborn.
Pregnancies affected by multiple antibodies were reviewed on a case-by-case basis. Patients with additional antibodies generally associated with hemolytic disease of the newborn (D, Kell, etc) whose titers of those antibodies were either significantly elevated or not analyzed separately were excluded from the analysis. Cases in which additional antibodies were either not generally associated with hemolytic disease of the newborn or remained of low titer were included, as were those in which the infant was negative for the additional antigen in question.
First affected pregnancies with anti-c isoimmunization were monitored according to the same criteria used for anti-D isoimmunization. Titers are performed at least every 4 weeks after confirmation of the antibody and titer at our institution. The paternal red blood cell antigenic phenotype is determined when paternity is certain, because approximately 20% of the random donor population will be c antigen negative. If the father is antigen negative, the fetus will be c antigen negative and unaffected by maternal anti-c antibody. If the father is homozygous, the fetus will be affected. The fetus of a heterozygous father has a 50% chance of being affected and is assumed to be affected. If paternal typing is not available or reliable, the fetus is assumed affected.
Amniocentesis for the measurement of fetal bilirubin in amniotic fluid is performed after 20 weeks’ gestation if the titer rises to a level of greater than or equal to 1:32, or at lower titer levels when there is a history of a prior affected child. Measurement and stratified display of the ΔOD450 in amniotic fluid after 28 weeks’ gestation was introduced by Liley.16 We use a modified Liley graph (Figure 1) developed by Colin McPherson and colleagues in the 1960s and based on patient care data obtained at The Ohio State University. Our modified Liley graph extends from 20 to 40 weeks’ gestation. The modified graph was subsequently validated.17,18 Only ΔOD450 values analyzed at our Prenatal Reference Laboratory are included here. Amniocentesis is repeated at intervals determined by the Liley graph, with more frequent procedures for higher zones.
In cases in which the paternal antigen was unknown or the father was heterozygous for the c antigen and invasive procedures were necessary, prenatal fetal antigenic phenotyping was performed on fetal blood cells obtained by cordocentesis, or (after 1995) by polymerase chain reaction testing of amniocytes from amniocentesis.
Cordocentesis is generally reserved for cases in which the ΔOD450 value has reached zone III, has plateaued or is rising in zone IIB, or there is hydrops fetalis and the fetus is less than 34 weeks. Hydrops fetalis is defined as fluid accumulation in at least two locations or a single effusion and anasarca. Criteria for fetal transfusion before 1986 included hydrops fetalis or amniotic fluid ΔOD450 in zone III. After 1986, fetal anemia was confirmed by cordocentesis and the fetus transfused if the hemoglobin level was less than 10 g/dL.
The management of fetal hemolytic disease remained largely unchanged over the duration of this study period. Serologic analysis supplemented by amniotic fluid analysis with fetal or neonatal transfusion as therapeutic modalities have been in place since the collection of this data began. Cordocentesis for direct analysis of fetal anemia was first used in 1986, and more recently middle cerebral artery peak velocity has been used as a noninvasive indicator of severe fetal anemia.19 Nonetheless, in all cases presented here, serologic and amniotic fluid analyses are the cornerstones of detection and management of fetal hemolytic disease.
Pregnancies in this cohort were considered affected or unaffected based on the neonatal or fetal direct antiglobulin test. A positive direct antiglobulin test indicates an affected fetus that is antigen positive; a negative direct antiglobulin test indicates an unaffected fetus. Affected pregnancies were then divided into two groups by clinical criteria of disease severity. Those in group A had less severely affected fetuses that did not require in utero therapy, had a newborn hemoglobin level of 11 g/dL or greater (the threshold used by Liley for severe disease in his studies16) and did not die from hemolytic disease of the newborn. Those in group B had fetuses with more severe hemolytic disease of the newborn, defined as the need for an intrauterine transfusion, the presence of hydrops fetalis, newborn hemoglobin level less than 11 g/dl, or perinatal death from hemolytic disease of the newborn.
The retrospective review identified anti-c antibodies affecting 102 pregnancies. For comparison, 966 anti-D sensitized pregnancies were seen during the same time period. The study population came from the clinics at our own institution and referrals from physicians throughout central and southeastern Ohio and neighboring regions. Of the original 102 cases of anti-c sensitization, 43 were excluded because we did not have access to direct antiglobulin test or newborn hemoglobin level data after deliveries outside our institution. Maximum titers from excluded patients ranged from 1:1 to 1:32, and the available data indicated that none of them had severe hemolytic disease of the newborn. Fifty-nine pregnancies in 56 women had complete records for interpretation.
Four patients were then excluded for the presence of potentially significant concurrent antibodies (one each with anti-D, Kell, E, and warm autoantibody). Thirteen patients had additional antibodies that were either low titers (titered separately) or not generally associated with hemolytic disease of the newborn; they were included in the analysis. This included seven affected patients with anti-E, two with anti-M, two with anti-Jka, and one each with anti-Fya and anti-Fyb. The final population used in this study thus comprised 55 pregnancies in 52 women.
Of these 55 pregnancies, 46 (84%) were affected and 9 (16%) were unaffected. The 46 affected pregnancies included 34 (74%) in group A and 12 (26%) in group B. Data regarding the individual patients is provided in Tables 1 and 2.
Titer of antibody in the first affected pregnancy or ultrasound evidence of hydrops fetalis identified all of the patients in group B, the more seriously affected pregnancies. Patient 1 from group B presented in her first affected pregnancy at 34 weeks with hydrops fetalis and a titer of 1:16. Amniotic ΔOD450 was in zone III. She was subsequently delivered, and the newborn hemoglobin level was 7.9 g/dL. No cause other than anti-c hemolytic disease of the newborn was identified for the anemia and hydrops. All of the other patients in group B had titers of 32 or greater in their first affected pregnancy. Fifty percent of our study group required only serologic titration of maternal antibody to monitor the isoimmunization.
Fifty-eight amniocenteses were performed in 23 pregnancies before any cordocenteses, based on the established criteria defined above. Thirty-seven were performed on 12 pregnancies in group A, and the remaining 21 were performed on 11 pregnancies in group B. Figure 1 plots the amniocenteses’ results on a modified Liley graph. Those patients in group A who underwent amniocentesis had ΔOD450 values lower than those in group B. All of the results from patients in group B reached zone III, whereas none of the patients in group A exceeded zone IIB. Middle cerebral artery peak velocity data is not available because the patients in this study preceded its routine use for assessment of fetal anemia at The Ohio State University Medical Center.
A total of 39 intravascular intrauterine transfusions were performed in six pregnancies. Fetal hemoglobins before the first transfusion ranged from 3.3 to 9.4 g/dL. In addition, intraperitoneal transfusions were performed twice each on two patients in the 1970s. Two patients underwent one cordocentesis each without an intrauterine transfusion, with fetal hemoglobin levels of 10.0 and 10.7 g/dL. They later received intrauterine transfusions when the fetal hemoglobin level fell to less than 10.0 g/dL. There were no deliveries for complications from any of these invasive procedures.
The infants in group A were delivered between 37 and 42 weeks’ gestation unless there were other obstetric indications for earlier delivery. The 12 infants in group B were delivered between 32 and 38 weeks’ gestation for isoimmunization. Three infants were delivered when fetal lung maturity was documented. One infant was delivered at 32 weeks after an unsuccessful intrauterine transfusion. Three pregnancies were delivered at 34, 37, and 35 weeks, respectively, for hydrops, dramatically higher serum titer associated with vaginal bleeding, and ΔOD450 in zone III because we do not institute intrauterine transfusions after 34 weeks’ gestation. The two patients treated with intraperitoneal transfusions in the 1970s were delivered at 33 and 34 weeks. One fetus was delivered at 36 weeks for a nonreassuring nonstress test, one was delivered at 38 weeks after six intrauterine transfusions, and the last was delivered at 33 weeks for unknown reasons at an outside hospital. Direct antiglobulin tests had a stronger reaction, on average, in group B infants. No infants in our study died as a direct result of anti-c hemolytic disease of the newborn.
Eight of the newborns underwent exchange transfusions for hemolytic disease of the newborn, including six in group B and two in group A (Tables 1 and 2). Several newborns in both groups received simple transfusions. Not enough pediatric information is available to discern which of these simple transfusions might have been for hemolytic disease of the newborn versus iatrogenic anemia associated with long hospital stays, prematurity, sepsis, or other complications.
We performed a literature search in MEDLINE (1966–2002) with the keywords “anti-c isoimmunization,” “isoimmunization,” “erythroblastosis fetalis,” and “hemolytic disease of the newborn.” Since the first reported case of anti-c isoimmunization in 1944,4 there have been only ten case series or individual reports of anti-c isoimmunization,5–14 with very few amniocentesis data and no cordocentesis data. The regional isoimmunization data banks in Norway,5 Sweden,6 and northern Wales7 have reported on a combined total of 68 anti-c patients. In the Norwegian series, nine of 32 infants underwent newborn transfusions. In the Swedish study, none of the 17 neonates affected by anti-c required phototherapy or exchange transfusions. A more recent analysis from northern Wales described three of 19 cases requiring exchange transfusions for hemoglobin levels of 7.4–9.6 g/dL.7 The largest analysis to date is the Oxford region study by Bowell et al13 in 1986, which reported on 177 anti-c isoimmunization patients: four had amniocentesis, none required cordocentesis, and three severe cases of hemolytic disease of the newborn were identified, based on neonatal data. Kozlowski et al14 evaluated 100 anti-c isoimmunized women who delivered c-antigen-positive infants. Fourteen percent had severe hemolytic disease, defined as the need for exchange transfusion.
From these national data banks, anti-c isoimmunization would seem to be a generally mild disease. However, a substantial number of patients in our study developed clinically significant anti-c isoimmunization. Other case reports of hydrops fetalis and death exist.10 A statement regarding the incidence of clinically significant hemolytic disease of the newborn among affected patients cannot be made from our data, however, because we cannot account for every patient in our geographic region.
Most practitioners manage anti-c isoimmunization in a manner similar to that for anti-D. Our data suggest that such management is appropriate. All of our seriously affected pregnancies were detected with a critical titer of 32 or greater or ultrasound evidence of hydrops fetalis in the first affected pregnancy or by amniotic fluid spectrophotometry in subsequent pregnancies. In our series, use of the critical titer of 1:32 without supplementation with ultrasound would have missed one seriously affected case, with a titer of 1:16. However, a titer of 1:16 is only one dilution lower than the critical titer of 1:32 used for anti-D and within the acceptable range of intralaboratory variation. Serologic evaluation, as noted before, is a cornerstone of management because only 50% of cases (23 of 46) with antigen-positive fetuses required more invasive testing than serologic titration of maternal antibody.
Liley’s original data16 on spectrophotometric evaluation of amniotic fluid was limited to 27–40 weeks’ gestation, and controversy exists regarding the evaluation and management of isoimmune fetal anemia with amniotic fluid ΔOD450 at lower gestational ages. Queenan et al20 have generated a graph for ΔOD450 from 14 weeks’ gestation through term. A variety of other graphs have been devised as well.21,22 At The Ohio State University, the modified graph developed by MacPherson has been validated and is successfully used for management of isoimmunized pregnancies after 20 weeks’ gestation.17,18 Amniotic ΔOD450 values were higher in group B patients, reaching zone III in every case, which indicates that amniotic fluid spectrophotometry reliably identifies severely affected fetuses.
The clinical criteria used to define group A and B in this study allow one to use the historic hemoglobin standards of the original Liley studies16 while also accounting for the changes that have occurred since his time. This is necessary when one uses a database that extends into the 1960s and includes cases that preceded cordocentesis. Two patients in our database underwent intraperitoneal transfusions in 1970 and 1978 without a determination of fetal hemoglobin beforehand. Although their ΔOD450 values in zone III indicated probable severe disease, hemoglobin data are not available.
Another change since the 1960s is the definition of the severity of fetal anemia. Recent gestational-age-specific hemoglobin values based on cordocentesis data have been proposed, which generally define “severe” anemia at levels much lower than those originally used by Liley.19,23 The hemoglobin levels of our patients are provided in Tables 1 and 2, and the readers can draw their own conclusions regarding severity. Ultrasound measurement of middle cerebral artery peak velocity is a noninvasive method that might be a useful addition to the care of anti-c isoimmunized pregnancies.19 However, only one fetus in our series had severe enough anemia according to Mari’s criteria to have been potentially detected by middle cerebral artery peak velocity.19
A limitation of this study is that the newborn data used to group these patients according to severity of disease were obtained retrospectively. However, antibody titers, the presence of hydrops fetalis, and amniotic fluid analyses on a modified Liley graph were predictive of fetal anemia less than 10 g/dL in this group of patients. None of the newborns in group A were born with hemoglobin levels below what was predicted by the modified Liley graph or the serum titers. Likewise, the fetuses in group B had ΔOD450 values or hydropic changes that were predictive of their resulting fetal or newborn hemoglobin levels. There were no preventable adverse fetal outcomes, and all patients were managed appropriately with the same modalities used to monitor anti-D isoimmunization.
It can be concluded from this study and a review of the literature that, although the relative incidence might not be high, anti-c isoimmunization is clearly capable of causing hemolytic disease of the newborn of clinical significance and warrants caution and close observation during pregnancy. The majority of patients, however, maintains low titers throughout the pregnancy and enjoys a benign neonatal outcome. The use of management similar to that for anti-D isoimmunization, including the use of a critical titer of 1:32 or greater, ultrasound monitoring for signs of hydrops fetalis, and the Liley graph, is recommended.
1. Chaves GF, Mulinare J, Edmonds LD. Epidemiology of Rh hemolytic disease of the newborn in the United States. JAMA 1991;265:3270–4.
2. Kornstad L. New cases of irregular blood group antibodies other than anti-D in pregnancy. Acta Obstet Gynecol Scand 1983;62:431–6.
3. Bowell PJ, Allen DL, Entwistle CC. Blood group antibody screening tests during pregnancy. Br J Obstet Gynecol 1986;93:1038–43.
4. McCall AJ, Race RR, Taylor GL. Rhesus antibody in Rh positive mother causing hemolytic disease of newborn. Lancet 1944;i:214.
5. Astrup J, Kornstad L. Presence of anti-c in the serum of 42 women giving birth to c-positive babies: Serologic and clinical findings. Acta Obstet Gynecol Scand 1977;56:185–8.
6. Filbey D, Hanson U, Wesstrom G. The prevalence of red cell antibodies in pregnancy correlated to the outcome of the newborn. Acta Obstet Gynecol Scand 1995;74:687–92.
7. Howard H, Martlew V, McFadyen I, Clarke C, Duguid J, Bromilow I, et al. Consequences for fetus and neonate of maternal red cell allo-immunization. Arch Dis Child Fetal Neonatal Ed 1998;78:F62–6.
8. Saenz RB. Hemolytic disease of the newborn due to anti-c. J Tenn Med Assoc 1993;1:9–11.
9. Hiehle WW, Altamirano MD, Kuhlman HH. Hemolytic disease of the newborn due to anti-c. Am J Obstet Gynecol 1960;79:600–3.
10. Appelman Z, Lurie S, Juster A, Borenstein R. Severe hemolytic disease of the newborn due to anti-c. Int J Gynaecol Obstet 1990;33:73–5.
11. Babinszki A, Berkowitz R. Haemolitic disease of the newborn caused by anti-c, anti-E and anti-Fya antibodies: Report of five cases. Prenat Diagn 1999;19:533–6.
12. Spong CY, Porter AE, Queenan J. Management of isoimmunization in the presence of multiple maternal antibodies. Am J Obstet Gynecol 2001;185:481–4.
13. Bowell PJ, Brown SE, Dike AE, Inskip MJ. The significance of anti-c alloimunization in pregnancy. Br J Obstet Gynecol 1986;93:1044–8.
14. Kozlowski CL, Lee D, Shwe KH, Love EM. Quantification of anti-c in haemolytic disease of the newborn. Trans Med 1995;5:37–42.
15. American Association of Blood Banks. Technical manual of the American Association of Blood Banks. 9th ed. Arlington, Virginia: American Association of Blood Banks, 1985.
16. Liley AW. Liquor amnii analysis in the management of the pregnancy complicated by rhesus sensitization. Am J Obstet Gynecol 1961;82:1359–70.
17. O’Shaughnessy R. Amniotic fluid spectraphotometry is useful after 20 weeks gestation in the care of pregnancies complicated by red blood cell isoimmunization. Am J Obstet Gynecol 1991;164:256.
18. O’Shaughnessy R, Kennedy M. Isoimmunization. In: Ling FW, Duff P, eds. Obstetrics and gynecology principles and practice. New York: McGraw Hill, 2001:308–26.
19. Mari G. Noninvasive diagnosis by Doppler ultrasonography of fetal anemia due to maternal red cell alloimmunization. N Engl J Med 2000;342:9–14.
20. Queenan JT, Tomai TP, Ural SH, King JC. Deviation in amniotic fluid optical density at a wavelength of 450 nm in Rh-immunized pregnancies from 14 to 40 weeks’ gestation: A proposal for clinical management. Am J Obstet Gynecol 1993;168:1370–6.
21. Sikkel E, Vandenbussche FP, Oepkes D, Meerman RH, Le Cessie S, Kanhai HH. Amniotic fluid delta OD 450 values accurately predict severe anemia in D-alloimmunization. Obstet Gynecol 2002;100:51–7.
22. Harman CR.Invasive techniques in the management of alloimmune anemia. In: Harman CR, ed. Invasive fetal testing and treatment. Boston: Blackwell Scientific, 1995:109–1, 121–2.
23. Nicolaides KH, Clewell WH, Mibashan RS, Soothill PW, Rodeck CH, Campell S. Fetal hemoglobin measurement in the assessment of red cell isoimmunization. Lancet 1988;1:1073–5.