IgP-HS is a frequent problem in PID patients. Whilst an underlying allergic mechanism, notably IgE- and/or IgG-mediated, has been proposed to account for these hypersensitivity reactions, the evidence supporting this hypothesis is fragile. Our case series and literature review validate the occurrence of IgP-HS even in patients with detectable IgA levels and/or no anti-IgA IgG. Hence, we decided to explore another mechanism, namely complement activation due to massive IgG infusion, considering that IgP-HS symptoms have numerous similarities with CARPA syndrome.

Our Work First Shows that Anti-IgA IgG Are not Rare in PID Patients, Especially in sIgA, and that Their Occurrence Is not Restricted to IgA-Deficient Individuals

In our centre, which uses a standardised commercial anti-IgA IgG assay, the prevalence of anti-IgA IgG was 6.2% in CVID patients, and 33% in sIgA. Anti-IgA IgG was also found in 2% of healthy controls. In the literature, the percentages varied widely from one study to another. We can formulate several hypotheses to explain these discrepancies between studies, mostly for technical reasons. First, most of the studies are old and used various, non-standardised methods of dosage. The positivity threshold also varied between studies. Furthermore, it should be noted that most of the studies involving healthy controls were performed on blood donors, in whom the presence of IgA was not verified (but sIgA is rare in the general population and is reported to be around 1:600 in Caucasians) [23]. In both our study and that of Petty et al. [24], anti-IgA IgG was found even in patients with detectable serum IgA. Thus, it seems that immunisation against IgA is possible in healthy individuals, and in PID patients, even if they have no IgA deficiency. As we found low anti-IgA IgG levels in some IgP products, we cannot exclude that the anti-IgA IgG contained in IgP contribute to those detected in the patients’blood in the days or months after an IgP infusion. However, anti-IgA IgG levels in these patients were higher than in healthy controls, and amongst the 6 anti-IgA IgG-positive subjects we described in our cohort, 3 of them never received IgP, so it is clear that their anti-IgA IgG are endogenous.

Our Work Also Shows that the Clinical Significance of Anti-IgA IgG Is Highly Controversial and Depends on the Assay Technique, as Highlighted by Others[6, 25]

The presence of anti-IgA antibodies seems to be a poor predictor of IgP-HS, for diverse reasons:

(i)

The presence of anti-IgA antibodies in a patient was neither sufficient nor essential to cause adverse reactions: in the literature, we identified several patients with anti-IgA Abs without IgP-HS, and patients with IgP-HS without anti-IgA Ab.

(ii)

Moreover, we did not identify specific patterns of patients with anti-IgA Abs who could be at risk of IgP-HS. Indeed, no association was found between the presence of IgP-HS and the anti-IgA Ab level, nor with the content in IgA in the IgP.

Furthermore, the results of a passive transfer experiment were not in favour of a pathogenicity of anti-IgA Abs. In that study, Robitaille et al. transferred anti-IgA-containing blood products to IgA-positive patients, and found no evidence for an increased risk of HS compared to anti-IgA-free blood products [26]. However, as we found anti-IgA IgG at low levels in some IgPs, we cannot exclude that they could play a role in some cases of IgP-HS.

In our case series, all three patients who tested positive for anti-IgA IgG and received IgRT, experienced IgP-HS (Table 2). This could argue in favour of a greater specificity of our anti-IgA IgG assay, which uses a standardised fluorescent enzyme immunoassay with a well-defined positivity threshold, unlike the various “in-house” techniques used in the literature. However, two of these three patients tolerated subsequent infusions of other IgP, with similar IgA content. Moreover, we did not test our patients for anti-IgA IgE, and cannot rule out a participation of their possible presence in the patients’ symptoms, even if the course of IgP-HS is not in line with an allergic mechanism.

Thus, even if it remains controversial, we do not think that a systematic assay of anti-IgA Abs is of great interest in PID patients, as its absence is not predictive of the absence of HS risk, and its presence is not a contraindication to the use of IgPs (regardless of whether they contain low or high levels of IgA).

We Also Found that IgP Infusion Can Activate the Complement Cascade, Suggesting a Possible Role in IgP-HS

We assessed the possibility that some of these HS reactions could be due to a recently described HS mechanism. In this CARPA mechanism, the classical or the alternative complement cascade is directly activated by the drug (respectively, by the generation of immune complexes, or by their homology with pathogenic viruses, such as with liposomal drugs). Then, the anaphylatoxins (C3a, C5a) that are generated by the complement cascade activate the allergy-mediator secretory cells (mast cells, basophils and tissue macrophages), resulting in the secretion of numerous highly effective vasoactive inflammatory mediators. These mediators then trigger the effector cells (endothelial cells, smooth muscle cells) inducing various responses (capillary leakage, bronchoconstriction, vasodilation, etc.) [14].

The hypothesis of complement activation by IgPs has long been raised in the literature. Yet, the data are inconsistent. Several studies found a complement activation (according to CH50, total C3 or C3a levels) and an elevation of circulating immune complexes after an IgP infusion [13, 27, 28], but these results were not confirmed by others [29]. Here, we provide data which suggest that IgPs were able to activate complement, both in vivo or in vitro. The elevation of the complement split product Bb could indicate an activation of the alternative pathway; it would be interesting to complete our results by an AH50 assay to confirm this point.

This raises the question of the clinical consequence of this complement activation. Indeed, some studies have shown an increased complement activation in patients with IgP-HS [13], although others found complement activation even in patients who well tolerated IgPs [9, 27]. In our in vitro experiments, the magnitude of the complement activation was not predictive of IgP-HS, as sC5b9 fold change was not different whether the serum came from a patient who tolerated IgP well or from a patient who tolerated it poorly. This is in line with studies about CARPA with other drugs, in which the in vitro assays were useful to predict the complement activation potential of a drug, but not to predict the individual risk of a patient having an infusion reaction [30, 31]. In our in vivo experiments, all the patients presented signs of complement activation after IgP infusion, regardless of whether they presented IgP-HS. We could only assess three patients with IgP-HS. The patient with the most severe reaction presented the highest levels of complement activation split products.

It could well be that the risk of IgP-HS is not strictly proportional to the complement activation level, but rather to the speed of anaphylatoxin clearance, and to the susceptibility of mast cells to release vasoactive mediators [32]. Indeed, Chanan-Khan et al. demonstrated that with the first IV infusion of pegylated liposomal doxorubicin, a drug frequently responsible for CARPA, the higher the dose rate, the higher the sC5b9 levels, and that HS reactions were more frequent with the higher rates of infusion [32]. It should be noted that our samples for sC5b9 assessment after infusion were collected in the hour following the infusion: the peak of plasma sC5b9 may occur very early during the infusion, and it is likely that we did not obtain the maximum level of sC5b9 in our patients with a single sample [32]. Moreover, it is important to acknowledge that we could only include a small number of patients in our in vitro and in vivo experiment, and that all the PID patients had been receiving IgP for months or years: a prospective study would be warranted in a larger cohort of patients to assess (i) if complement activation is greater at the start of IgRT, (ii) if a greater complement activation is associated with more severe and/or more frequent HS symptoms, and (iii) if complement activation decreases with time, as IgP-HS decreases with time in PID patients.

The link between anti-IgA Abs and complement activation also needs to be clarified. Indeed, in theory, the presence of anti-IgA Abs could promote complement activation by the formation of immune complex with IgA-containing IgP. In the study by Cunningham-Rundles et al., all the patients showed signs of complement activation with an increase in the C3a levels during the Ig infusion, whether they had anti-IgA Abs or not [28]. We could not assess the in vivo complement activation in anti-IgA IgG-positive patients in our study, but we found no difference in in vitro complement activation between patients with and those without anti-IgA IgG.

Finally, it is important to underline that some IgPs (such as Clairyg®, Gammagard®, Gammanorm®, Hizentra® and Cutaquig®) contain polysorbate 80 [33]. This compound, a derivative of polyethylene glycol, is known to be implicated in complement activation [34] and in CARPA reactions [35], as well as in IgE-mediated HS [36]. One can thus assume that this vehicle could also be implicated in some IgP-HS, and that complement activation tests and skin tests should be performed in these patients.

Overall, our results suggest that IgPs can activate the complement cascade, but further studies will be needed to clarify the clinical consequences of this complement activation.

It is important to underline that the infusion rate and the route could also contribute to the tolerance of IgP, and that some methods have been proposed to improve it. It is usually recommended to reduce the infusion speed [18, 37], to use simple analgesics if appropriate, to ensure adequate hydration, and to consider SC rather than the IV route if symptoms persist [10, 38]. Premedication with anti-histamine drugs and steroids can also be considered. An intentionally repeated exposure to IgP at short time intervals has also been proposed as a desensitisation way to improve IgP tolerance [18, 20,21,22].

In conclusion, our data and the findings of the literature review do not support a predictive value of anti-IgA Abs in IgP-HS. Breaking the myth of “IgA allergy” has major implications for PID patient care. Indeed, despite well-conducted IgRT, some patients with CVID or X-linked agammaglobulinaemia patients continue to experience recurrent or chronic bacterial respiratory tract infections, which may worsen bronchiectasis [37, 39]. One hypothesis to explain this could be the lack of IgA and IgM in conventional IgPs. In support of this, lower levels of IgM and IgA were associated with bronchiectasis in CVID patients [40], and low levels of IgA were associated with a higher risk of pulmonary and sinus infections [39]. These findings can support the use of IgA/IgM-enriched IgPs in selected severe PID patients [37, 41]. Our study and the high similarities with “CARPA” also suggest a role of complement activation, which will need to be confirmed in larger prospective studies in PID patients starting IgRT.