Background
KD is a systemic vasculitis in which the major complication is the development of coronary artery aneurysms (CAA). Treatment with intravenous immunoglobulin (IVIG) significantly lowers the incidence of CAA. Standard therapy for the treatment of KD is high dose IVIG (2 g/kg) and aspirin. Retreatment with IVIG is administered for persistence of fever ≥ 36 h after the first infusion. Persistence of fever after initial IVIG therapy is estimated to occur in approximately 10% - 20% of cases [
1].
IVIG is used in high doses, most frequently at 2 g/kg, as an immunomodulatory agent [
2]. It is a pooled blood product acquired from thousands of blood donors and it contains measurable levels of anti-A and anti-B (IgG subclass) as well as non-ABO erythrocyte antibodies (e.g. anti-D) [
3]. IVIG is considered to be a safe product that is generally well tolerated. Hemolysis is a rarely reported side effect of IVIG. It occurs more commonly in those patients who receive high-dose IVIG [
2,
4] as is used in the treatment of KD. In the literature, there are 6 reported cases of children with hemolytic anemia following IVIG treatment for KD [
4‐
7]. In this report, we describe 4 patients, all from a single centre, who developed hemolytic anemia following IVIG treatment for KD. To our knowledge, this is one of the largest case series describing this complication in this patient population.
Discussion
Hemolytic anemia is a potentially serious and possibly under recognized side effect of IVIG therapy. The presence of transient, passively acquired antibodies (positive direct antiglobulin test) with associated decreased haptoglobin levels and mild reticulocytosis following treatment with IVIG was noted many years ago [
9]. However, this rarely results in clinically significant anemia. There have only been 4 previous reports with a total of 6 patients described to have hemolysis following IVIG therapy for KD [
4‐
7]. In all 4 reports, the presumed mechanism was of a direct antibody mediated attack, but they only demonstrated specific blood group antibodies in 4 of 6 of these patients [
4‐
7] We found antibodies in the eluate of the 3 patients that were tested; in addition, all 4 patients had a positive direct antiglobulin test (see Table
1). To our knowledge this is the largest case series that demonstrated specific blood group antibodies following IVIG for Kawasaki disease, which further supports the hypothesis that direct antibody mediated attack is one of the potential mechanisms for hemolysis in this population.
The etiology of clinically significant hemolysis in patients with KD treated with IVIG is multifactorial. Isohemagglutinins (anti-A/B antibodies) present in IVIG can cause direct antibody attack on RBCs. Even though the titers of these antibodies, in commercially available products, must meet agreed-upon specifications by manufacturers through a Biologics Licence Application [
3], there are many documented cases of hemolytic anemia despite these limits [
10]. This phenomenon may be dose dependent given that in 6 of the 6 patients previously reported, and in all 4 of our patients (total 10/10 patients), acute hemolytic anemia occurred following a second dose of IVIG for persistent fever. All patients received a total of 4 g/kg prior to the clinically significant hemolysis. However, it is known that hemolytic anemia can occur even with doses as low as 1 g/kg [
10]. Factors that may increase the risk for hemolysis include the dose of IVIG, the titer of the isohemagglutinin antibody in the IVIG preparation, the strength of the patient's antigen expression [
5,
7], the affinity of the antibody for the antigen [
7], or a combination of these factors.
All our KD patients received Gammunex. Hemolytic anemia in KD has also been reported with Gammaguard [
7]. Isosmolar, liquid IVIG products, such as Gammaguard and Gammunex, seem to have higher isohemagglutinin titers than lyophilized products [
10]. However, lyophilized products are more osmolar and have other potential side effects including acute renal failure and thrombosis [
10].
There is a greater risk for hemolysis in patients with non-O blood groups [
10]. It is hypothesized that patients with non-O blood groups, with low concentrations (congenital or acquired) of soluble A and/or B substance in their plasma, are at increased risk for hemolysis due to their inability to neutralize the anti-A and/or anti-B isohemagglutinins present in the plasma after IVIG infusion [
11]. What is interesting about this hypothesis is that in all KD patients who developed hemolysis following IVIG (including our patients), the blood type in 6 was A+, 2 AB+, 1 B + and 1 O+. The patient who was O + also had cold agglutinins.
It is also plausible that patients with a baseline anemia, as is often seen in KD, may be more susceptible to hemolysis because of an absolute decrease in binding sites to disperse the antibodies. Other potential causes for hemolysis in KD include cold agglutinating anti-RBC antibodies [
4]. KD in itself may also be a predisposing factor for hemolysis given that there is immune dysregulation in this disease. This is supported by case reports describing hemolysis in patients with KD who did not receive IVIG [
12,
13].
Another possible mechanism for a decrease in hemoglobin with IVIG, independent of isoantibodies, is enhanced erythrocyte sequestration [
14]. Since IVIG contains high molecular weight IgG complexes, these can mimic immune complexes by activating complement [
14]. These complexes bind to complement receptor 1 on RBCs which leads to erythrophagocytosis and hence a reduction in hemoglobin [
14]. Predictors for this phenomenon include age and RBC ability to bind the IVIG immune complex-like moieties [
14].
In Canada, the prevalence of hemolysis following IVIG is unknown given the underdeveloped surveillance system of IVIG administration. However, in recent years, the Canadian Blood Services and Hema-Quebec have noted an increase in reporting of this complication [
15].
We witnessed an unusually high rate (16%) of clinically significant hemolysis over a 14-month period at our institution and this complication had not been noted previously. We are unaware of any important changes that occurred in the IVIG product during this time period to account for these findings. At the same time, we experienced a high retreatment rate and therefore we cannot exclude subtle changes in the IVIG product that could explain both these findings. Specific blood group antibodies were demonstrated in addition to the positive direct antiglobulin test, which suggests that the primary mechanism for hemolysis in our cases was through direct antibody mediated attack.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
RB - reviewed medical records, gathered patient data, literature review, manuscript preparation; BW - interpretation of hematology tests, manuscript review; RS - reviewed medical records, literature review, manuscript preparation. All authors read and approved the final manuscript.