To our knowledge, this is the first study that explored the role of PEGylation in triggering hypersensitivity reactions using the Italian SRS database. In line with a recent Italian pharmacovigilance study, our analysis showed that drug-induced hypersensitivity reactions were mostly reported in female patients aged more than 40 years [18].

This study showed a disproportionate reporting of hypersensitivity reactions for most PEG+ medicinal products, i.e., peginterferon alpha-2b, peginterferon beta-1a, certolizumab pegol, pegvisomant, pegaspargase, and PEGylated liposomal doxorubicin, versus their respective PEG− comparators. However, when using all other drugs as comparators, no disproportionate reporting of hypersensitivity reactions for PEG+/PEG− medicinal product was observed. This may be due to a masking effect, owing to the presence of other medicinal products (e.g., antibiotics, other antineoplastic agents, and anti-inflammatory drugs) in the same database, which increase the threshold required for highlighting a signal [19].

A number of studies and clinical reports showed that PEG administration may be associated with moderate-to-severe hypersensitivity reactions, mainly due to complement system activation [4, 20] (complement activation-related pseudo allergy) or to the production of anti-PEG antibodies that may lead to an acceleration of the blood clearance (ABC phenomenon) of the PEGylated drug, resulting in efficacy loss and hypersensitivity reactions [21,22,23].

Many of the reported hypersensitivity reactions, mainly for PEG− medicinal products, had very long median times to onset. Indeed, the SMQ “hypersensitivity” does not only contain PTs pointing towards immediate hypersensitivity, but also PTs pertaining to delayed hypersensitivity (e.g., cutaneous adverse reactions). As a confirmation of this, most of the reports of hypersensitivity reactions for the medicinal products under investigation, especially those judged as serious, were cutaneous reactions.

Cutaneous hypersensitivity reactions have been frequently reported for interferons, and for peginterferon alpha-2b especially, and they have been extensively described in the literature [24,25,26,27,28,29]. A large body of evidence suggests that cutaneous hypersensitivity reactions are more frequent for peginterferon alpha as compared with its non-PEGylated form [28,29,30] and that switching from peginterferon alpha to conventional interferon alpha decreases the risk of severe allergic cutaneous reactions [30,31,32], suggesting that the likely culprit for the skin reaction may be the PEG component of peginterferon [30,31,32].

Additionally, pegaspargase was significantly associated with an increased reporting of hypersensitivity reactions. Although a large body of evidence proved that it causes less hypersensitivity as compared with asparaginase, its non-PEGylated form [33,34,35,36], such a difference could not be found in this study because asparaginase was not widely used in Italy during the study period (overall, only 632 packages were dispensed from 2018 to 2021) and no hypersensitivity reaction with this drug was identified in the Italian SRS.

Hypersensitivity reactions reported for certolizumab pegol mainly concerned cutaneous reactions such as hives and skin rashes. Cutaneous reactions, and psoriasiform skin eruptions in particular, are among the most common adverse reactions of tumor necrosis factor-α inhibitors and a considerable number of case reports describing adverse cutaneous reactions in patients treated with certolizumab have been published [37,38,39,40,41]. It has been proposed that increased levels of interferon-α due to tumor necrosis factor-α suppression and the activation of interleukin-23/T-helper-17 axis may play a crucial role in the development of such paradoxical reactions [42]. Therefore, cutaneous adverse reactions due to a tumor necrosis factor-α blockade seems to be a class effect.

It was postulated that the interaction between PEG and antibodies is stabilized and then enhanced when PEG is exposed on the surface of nanoparticles, suggesting that PEGylated nanocarriers might be even more immunogenic when compared with PEGylated drugs. This hypothesis seems to be confirmed looking at the data collected for liposomal doxorubicin, which is associated with important adverse events likely caused by complement activation, such as the hand-foot syndrome and infusion-related reactions [43]. As compared with the non-liposomal formulation, PEGylated liposomal doxorubicin is associated also with a wider range of cutaneous adverse reactions, reflecting its pharmacokinetics and tissue distribution [44].

Nevertheless, although a correlation between the presence of antibodies against methoxy-PEG and complement activation upon incubation with liposomal doxorubicin has not been found, the immunogenicity of this formulation is well known, and data drawn from liposomal doxorubicin experience cannot be transferred to other liposomal-based medicinal products. This is further demonstrated by the lack of a severe adverse effect registered in the case of PEGylated liposomes carrying irinotecan. In fact, PEG-induced complement activation-related pseudo allergy is dependent on several factors including liposome composition and physicochemical properties, along with the density of the PEG coating, the dose, and the frequency of administration. As an example, the immunogenicity of liposomal doxorubicin is attributed to both the negative charge and the oval shape of liposomes resulting from the crystallization of doxorubicin in the aqueous core. Equivalent formulations having a spherical shape did not lead to an increased complement activation [45]. In general, negatively or positively charged liposomes, with a cholesterol content higher than 45%, given at a high lipid dose through a slow infusion might be associated with a higher risk of the complement activation-related pseudo allergy phenomenon regardless of the presence of PEG moieties on the surface. As further proof of this, negatively charged non-PEGylated liposomes containing amphotericin B, used for the treatment of fungal infection, were found to increase levels of complement activation to a higher extent with respect to analogous non-charged liposomes [45].

In addition to these considerations, the results of our study are in line with those of a recently published pharmacovigilance study that explored the adverse events potentially associated with PEGylation by comparing ADR reports of PEG+ and PEG− drugs from the US Food and Drug Administration Adverse Event Reporting System database [46]. This study found that three immunogenicity-related PTs (i.e., rash, pruritus, and erythema) and the SMQs of hypersensitivity and anaphylactic reactions were slightly higher for PEGylated medicinal products, with a trend toward statistical significance, thus highlighting the importance of screening for anti-PEG antibodies to identify patients who may require a PEGylated drug with a reduced dosing strategy or the use of non-PEGylated drugs [46].

Furthermore, conflicting evidence exists on the role of PEG size in triggering hypersensitivity reactions [47]. Findings from this study suggest that PEG molecular weight is not statistically associated with an increased frequency of hypersensitivity reactions reporting; however, such statistical evidence does not necessarily translate into biological evidence, thus suggesting the need to conduct other studies investigating the role of PEG size in triggering hypersensitivity reactions.

The SRS is the cornerstone of pharmacovigilance for post-marketing surveillance of drugs and for the detection of new potential safety signals. In our study, we provided additional evidence concerning PEG-related hypersensitivity. To the best of our knowledge, this is the first pharmacovigilance study calculating the reporting rate of hypersensitivity reactions and to investigate the relationship between the frequency of hypersensitivity reaction reporting and PEG size for each of the medicinal products included in the analyses.

However, some limitations warrant caution. Many of the reports of hypersensitivity reactions for the medicinal products under investigation were long-term cutaneous reactions that were not necessarily due to hypersensitivity and/or allergic mechanisms, consequently accounting for a possible overestimation of the risk. The SRS has several known limitations that should be acknowledged, including under-reporting of suspected ADRs [48], selective over-reporting, missing demographic and clinical data, and a lack of a denominator (i.e., the total number of drug users), which precludes measuring the incidence of hypersensitivity reactions for both PEG+ and PEG− drugs. Furthermore, ROR estimates may also be affected by several biases ascribed to reporting trends in the SRS database. Finally, the analysis of SRS is mainly aimed at generating hypotheses, and causative relationships between the drugs and the studied ADRs can only be surmised. Indeed, drug-induced hypersensitivity reactions may also be due to specific excipients that are known to be involved in such adverse reactions, and delayed hypersensitivity specifically, rather than specific active principles [49]. Moreover, as it was not possible to stratify the analyses by route of administration, we compared drugs with more than one route of administration to drugs with only one route of administration (e.g., peginterferon beta-1a vs interferon beta-1a and pegaspargase vs asparaginase). However, it should be noted that there are no significant differences concerning the frequency of hypersensitivity reactions between the various parenteral routes of administration, but mostly between oral and parenteral routes of administration [50]. As we compared only drugs that are administered through parenteral routes of administration, we think this was unlikely to affect our analysis.

Finally, to really understand the role of PEG in the allergy phenomenon, more PEG+/PEG− liposomes comparators are needed, as lipid nanoparticles can activate the immune system by themselves and, as such, non-liposomal drugs cannot be used as controls in these studies. Therefore, findings from the analysis of the SRS need to be evaluated further through pharmacoepidemiological studies.