The Kell blood group system consists of two proteins, Kell and XK. Kell is a highly polymorphic type II glycoprotein, which carries most of the more than 20 known Kell antigens. KEL1 is the strongest Kell immunogen, and antibodies against KEL1 (anti-K) are the major cause of Kell-related hemolytic disease of the fetus and newborn (HDFN).1 Kell alloimmunization in pregnancy, when untreated, is associated with severe fetal anemia, hydrops, and fetal death. It is the second major cause of HDFN after Rh-D alloimmunization, with a still increasing incidence of 3.2 in 1,000, affecting 1 in 10,000 neonates.2,3
In a large series of red cell alloimmunized pregnancies treated with intrauterine transfusions, the survival rate of fetuses with Rh-D–related anemia was 89%, compared with 58% in the Kell group.4 This difference was attributed mainly to late diagnosis and referral in the Kell group. An important and still unresolved question is which pregnancies should be referred and at what gestational age close monitoring in specialized centers should start. Traditionally, titers of Kell antibodies were considered predictive of the severity and time of onset of HDFN.5 However, even with relatively low antibody titers (less than 1:16) serious fetal morbidity has been reported.6,7 Another often advised method for risk estimation is evaluation of the obstetric history.8,9 Were previous children affected, and at what gestational age was the onset of the disease?
The aim of this study was to evaluate the role of the course and outcome of previous pregnancies and maternal serum Kell antibody titers as tools to estimate the optimal time for referral and close surveillance in Kell-alloimmunized pregnancies.
MATERIALS AND METHODS
The Leiden University Medical Center is the Dutch national center for referral of pregnancies complicated by red cell alloimmunization. Data on a consecutive cohort of patients, referred to the Leiden University Medical Center between January 1988 and April 2006, with a singleton pregnancy, a Kell antibody titer of at least 1:1 (with indirect antiglobulin test in saline), and a confirmed Kell-positive fetus treated with at least one intrauterine transfusion (IUT) were extracted from our database.
Since 1987, we perform ultrasound-guided intravascular blood transfusion for the treatment of severe fetal anemia. This technique has not changed in the last 18 years. In the 1980s and 1990s, the decision to perform fetal blood sampling with subsequent transfusion in case of anemia was based on either ultrasound detection of fetal hydrops or a combination of obstetric history, ultrasonographic measurements of fetal liver and spleen size, and venous Doppler parameters as described previously.4,10 Amniocentesis for amniotic fluid bilirubin levels was rarely used in our management of Kell alloimmunization. In 1986, a study by Nicolaides et al11 suggested that Liley’s chart could not be reliably used before 27 weeks of gestation. Almost all affected fetuses in our cohort required the first transfusion in the second trimester. Amniotic fluid analysis for bilirubin levels was shown to be unreliable in Kell alloimmunization in 1994 and has not been used in our practice since then.12 Since the late 1990s, we have increasingly relied on Doppler measurements of middle cerebral artery peak systolic flow velocities to predict fetal anemia, using the graph published by Mari in 2000.13 This diagnostic method was shown to work well in Rh-D– as well as in Kell-induced fetal anemia.14,15
Data recorded of the last known (index) pregnancy included last known maternal serum titer of Kell antibodies before the first IUT, number of IUTs performed, gestational age at first IUT, hemoglobin (Hb), and hematocrit levels at first IUT, presence or absence of hydrops, need for neonatal exchange transfusions, length of stay in the neonatal intensive care unit, and perinatal mortality related to Kell alloimmunization. Details on the previous pregnancies included the presence or absence of a Kell-positive child, Kell-related stillbirth, or neonatal death and number of IUTs if treatment occurred for a Kell-positive anemic child. Based on the obtained data, previous pregnancies were classified into two categories according to increasing severity: 1) Kell-negative child in previous pregnancy or no previous child, and 2) Kell-positive child in previous pregnancy. If a woman had more than one previous pregnancy, only the data of the most severely affected pregnancy was recorded. We examined the usefulness of subdividing category 2 into increasing levels of severity (eg, fetal or neonatal mortality and onset of fetal anemia before various cutoffs of gestational age). This resulted in groups of only 1 or 2 patients, numbers too small to draw meaningful conclusions. We chose, therefore, to base our analysis on the two broad categories described. The study was approved by the Leiden University Medical Center review board.
To test the significance of the obstetric history in predicting the severity of fetal disease, we compared both groups for each of the variables of the index pregnancies listed above. Statistical analysis of the differences was done using Mann-Whitney U test or Fisher exact test where appropriate. To test the significance of the correlation between the level of maternal antibody titers and onset of fetal anemia, Spearman’s rank sum test was used, after log transformation of the titers. A P value of.05 or less was considered statistically significant. All statistical analyses were performed using SPSS 11.0 (SPSS Inc., Chicago, IL).
A total of 44 women met our inclusion criteria. Three women were excluded due to the absence of fetal anemia at first IUT. Although in all cases the fetal Hb was more than 9 g/dL, the operator chose to transfuse some blood, making an analysis of the severity of fetal disease impossible. Characteristics of the remaining 41 women and their pregnancies are summarized in Table 1.
The median gestational age of pregnancies in category 1 (n=29) was 24 weeks 0 days at time of first IUT and 21 weeks 3 days for category 2 (n=12) (Fig. 1 and Table 1). This difference was statistically significant (P=.01). In the group of women without a previously affected child, two women received the first IUT before 20 weeks of gestation (18 weeks 4 days and 19 weeks 2 days); the earliest gestational age for the first IUT in women with a previous Kell-positive child was performed at 17 weeks 5 days.
The median maternal serum Kell antibody titer in both women with and without a previous Kell-positive child was 1:256 just before first IUT (Table 1). Figure 2 shows the gestational age of the fetus at which first IUT was performed plotted against maternal serum Kell antibody titer for the two categories of obstetric history. Increasing levels of antibody titers were associated with a decrease in gestational age at first IUT (Fig. 2). In neither the category of women with a previous affected child nor within the category of women without such a child was the observed inverse correlation significant (R2=−0.20, P=.62 and R2=−0.08, P=.72, respectively). Severe hydrops was present at the first transfusion in 16 of 41 fetuses. In all but one case the hydrops was the reason for referral and thus present at the first visit. This occurred more often in the group with an uneventful history, although the difference did not reach statistical significance (Table 1). Two cases were found with severe fetal anemia despite a low last-known antibody titer. In one case the titer of 1:4 had been found at 16 weeks of gestation, and again at 19 weeks, after which it had not been repeated since the patient was already monitored weekly in our center. At 30 weeks of gestation, middle cerebral artery peak systolic flow velocities indicated anemia, and the first of two transfusions was done for a fetal Hb of 7.2 g/dL. In the second case, a Kell antibody titer of 1:2 was found at 16 weeks of gestation. Amniocentesis revealed a Kell-positive fetus, this result was available at 17 weeks of gestation. At the next visit, at 19 weeks and 1 day, severe hydrops was already present. Transfusion the next day was done, showing an extremely low fetal Hb of 1.3 g/dL. After the second IUT, the hydrops disappeared. A total of 5 IUTs were needed. A healthy girl was born at 37 weeks of gestation, weighing 3,180 g. In Figure 3, a flow chart is provided for the suggested management of Kell alloimmunization in pregnancies based on our results.
In this study, we found that in women who had a previous Kell-positive child, severe fetal anemia occurred earlier on in the current pregnancy than in women without a previous Kell-positive child. However, even in Kell-immunized women without a previous Kell-positive child, severe fetal anemia may occur early in the second trimester. For clinical practice, we must therefore conclude that, irrespective of the obstetric history, early frequent fetal surveillance from around 17 weeks of gestation onward in these pregnancies is essential to allow timely intervention with blood transfusions. In our experience, IUTs are feasible even from 16.5 weeks of gestation onward, although certainly technically challenging.4
Median maternal serum Kell antibody titers were similar in both groups and were not predictive for the time of onset of severe fetal anemia. Maternal antibody titers were 1:32 or higher in all women at time of first IUT, except for two women with an antibody titer of 1:2 and 1:4. A limitation of our analysis of the correlation between antibody titer and severity of fetal anemia is the variable time between the last titer measurement and the first IUT. Possibly, titers may have increased during this interval. However, the data used in this study represent clinical reality, with management comparable to most other centers. No studies have yet shown a benefit of more frequent (eg, weekly) titer assessment. Importantly, the case described with severe early fetal anemia with a recent titer of 1:2 clearly shows that any higher cutoff level cannot be recommended. Again, in order not to miss any case of severe fetal anemia, the safest approach would be to closely monitor all Kell-alloimmunized pregnancies with a Kell titer of 1:2 or above.
This conclusion is in accordance with the observed lack of any correlation between maternal antibody titers and fetal outcome in a few previous small studies.6,16,17 Although retrospective by design, the strength of our study is that a relatively large consecutive cohort of patients was analyzed, all were treated using the same protocol, and all maternal serum testing was performed by the same laboratory. Possibly, in our cohort, women with a previous Kell-positive child might have been more vigorously screened for signs of fetal anemia than women without such a previous child. It is unlikely however that this has caused us to perform IUTs earlier on in pregnancy, since fetal Hb levels at first IUT within both categories of obstetric history were comparable.
The significant correlation between severity of the obstetric history and gestational age at time of first IUT is not easily explained, taking into account our other results: women with a previous Kell-positive child had the same level of Kell antibody titers as women without such a previous child, suggesting that the drop in fetal Hb is not dependent on the antibody titers alone. Blood transfusions with incompatible Kell-positive blood occurred more frequently in women with a previous Kell-positive child. However, since matching of both Kell and c antigens before blood transfusion in women before or within the fertile age category was introduced in the Netherlands in 1994, the risk for developing Kell antibodies through an unmatched transfusion has diminished.18
For clinical practice, we recommend that all pregnancies with Kell-alloimmunization should be regarded as high risk, deserving early referral and close monitoring by specialists in this field. As a guideline, we suggest referral of all women with a titer of 1:2 or above, at a gestational age of 16–17 weeks, irrespective of their obstetric history and levels of maternal serum antibody titers.
1. Lee S, Russo D, Redman CM. Functional and structural aspects of the Kell blood group system. Transf Med Rev 2000;14:93–103.
2. Tovey LA. Haemolytic disease of the newborn: the changing scene. Br J Obstet Gynaecol 1986;93:960–6.
3. Geifman-Holtzman O, Wojtowycz M, Kosmas E, Artal R. Female alloimunization with antibodies known to cause hemolytic disease. Obstet Gynecol 1997;89:272–5.
4. Van Kamp IL, Klumper FJ, Meerman RH, Oepkes D, Scherjon SA, Kanhai HH. Treatment of fetal anemia due to red-cell alloimmunization with intrauterine transfusions in the Netherlands, 1988–1999. Acta Obstet Gynecol Scand 2004;83:731–7.
5. Moise KJ Jr. Hemolytic disease of the fetus and newborn. In: Creasy RK, Resnik R, Iams JD, editors. Maternal-fetal medicine: principles and practice. 5th ed. Philadelphia (PA): Saunders; 2004. p 537–61.
6. Van Dijk BA, Dooren MC, Overbeeke MAM. Red cell antibodies in pregnancy: there is no “critical titre”. Transfus Med 1995;4:199–202.
7. Bowell PJ, Brown SE, Dike AE, Inskip MJ. The significance of anti-c alloimmunization in pregnancy. Br J Obstet Gynaecol 1986;93:1044–8.
8. Rodeck CH, Deans A. Red cell alloimmunisation. In: Rodeck CH, Whittle MJ, editors. Fetal medicine: basic science and clinical practice. London (UK): Churchill Livingstone; 1999. p. 785–804.
9. Nicolaides KH, Sebire NJ. Management of red cell isoimmunized pregnancies. In: Kurjak A, Chervernak FA, editors. Textbook of perinatal medicine. 2nd ed. Informa (UK): Oxon; 2006. p 1135–42.
10. Oepkes D, Brand R, Vandenbussche FP, Meerman RH, Kanhai HH. The use of ultrasonography and Doppler in the prediction of fetal haemolytic anaemia: a multivariate analysis. Br J Obstet Gynaecol 1994;101:680–4.
11. Nicolaides KH, Rodeck CH, Mibashan RS, Kemp JR. Have Liley charts outlived their usefulness? Am J Obstet Gynecol 1986;155:90–4.
12. Vaughan JI, Warwick R, Letsky E, Nicolini U, Rodeck CH, Fisk NM. Erythropoietic suppression in fetal anemia because of Kell alloimmunization. Am J Obstet Gynecol 1994;171:247–52.
13. Mari G. Noninvasive diagnosis by Doppler ultrasonography of fetal anemia due to maternal red-cell alloimmunization. Collaborative Group for Doppler Assessment of the Blood Velocity in Anemic Fetuses. N Engl J Med 2000;342:9–14.
14. Van Dongen H, Klumper FJCM, Sikkel E, Vandenbussche FP, Oepkes D. Ultrasound and Doppler in the prediction of fetal anemia in Kell alloimmunized pregnancies. Ultrasound Obstet Gynecol 2005;25:341–5.
15. Oepkes D, Seaward PG, Vandenbussche FP, Windrim R, Kingdom J, Beyene J, et al. Doppler ultrasonography versus amniocentesis to predict fetal anemia. N Engl J Med 2006;355:156–64.
16. Gottvall T, Hilden JO, Selbing A. Evaluation of standard parameters to predict exchange transfusions in the erythroblastotic newborn. Acta Obstet Gynecol Scand 1994;73:300–6.
17. Bowman J. Maternal Kell blood group alloimmunization. Obstet Gynecol 1992;79:39–44.
© 2007 by The American College of Obstetricians and Gynecologists. Published by Wolters Kluwer Health, Inc. All rights reserved.
18. Castel A, Van Dijk BA, Van den Boom FMLG Brand A, Engelfriet CP, Overbeeke MAM, et al. Prevention of immunization by c, E and K: backgrounds and gradual implementation. Ned Tijdschr Klin Chem 1996;21:3–7.