Adverse Reactions Associated with Intravenous Immunoglobulin Administration in the Treatment of Neurological Disorders: A Systematic Review

Intravenous immunoglobulin (IVIg) is a therapeutic preparation used to treat multiple neurological disorders including Guillain-Barre syndrome (GBS), chronic inflammatory demyelinating polyneuropathy (CIDP), dermatomyositis (DM), autoimmune encephalitis, and myasthenia gravis (MG) [1, 2]. While a large number of clinical trials have found IVIg therapy to be safe and generally well-tolerated, adverse reactions (ARs) have been reported. A recent retrospective observational study of IVIg-associated AR identified a higher rate of ARs occurring in IVIg treatment for neuromuscular diseases in particular [3]. Previous narrative literature reviews have been published, describing the ARs of IVIg in a broad scope of clinical applications [4]. To our knowledge, however, there have not been any systematic reviews that describe the current evidence regarding ARs of IVIg therapy in the treatment of neurological disorders specifically.

This review aims to summarise studies that report ARs of IVIg therapy used in more than one neurological disorder. Through the examination of studies that have included patients with multiple neurological conditions, insights may be gained into the characteristics of individual neurological diseases associated with IVIg-ARs. Only studies which report ARs of IVIg use across multiple neurological indications were included, to allow for comparison of the ARs associated with different neurological diseases. Patients with neurological or rheumatological disorders are often treated with high-dose IVIg (2 mg/kg) for immunomodulatory purposes, as opposed to the standard 0.4 mg/kg/month dosing used for IVIg replacement therapy in immunodeficiency. An attempt will be made to determine if the rate differs between patients receiving IVIg for neurological indications versus non-neurological indications. Additionally, we seek to determine if there are any specific neurological indications for IVIg use associated with a greater rate of AR.

In addition, we aim to determine patient-related risk factors, treatment-related risk factors, and indications of IVIg use that may be associated with a greater risk of developing an AR. We also seek to determine evidence for monitoring, premedication, and treatment strategies for IVIg-ARs. This review has significant clinical implications as it may assist clinicians in navigating the complex interplay of patient-related and treatment-related risk factors involved in the risk assessment of IVIg administration across various neurological conditions. Ultimately, this research seeks to facilitate the evidence-based prevention and treatment of IVIg-associated ARs that may arise in this subset of patients.

This review followed the methods of the PRISMA 2020 guidelines and was registered with PROSPERO, ID 321009 [5]. An electronic database search of PubMed, Embase, and Cochrane Library was performed in March 2022, to search for articles published from inception until March 26, 2022. The search was conducted using the following search terms “(IVIg OR intravenous immunoglobulin OR immunoglobulin G) AND (neurolog*) AND (adverse reaction OR adverse effect OR side effect OR allerg*).”

Articles were included if they fulfilled the following criteria: (1) primary clinical publication; (2) reports ARs of IVIg therapy use in more than one neurological disorder; and (3) available in full-text. The exclusion criteria were (1) secondary clinical publications, editorials, and conference abstracts; (2) articles written in a language other than English; and (3) articles unable to be retrieved in full-text. Case reports which only reported a singular case were not included as they did not meet the second inclusion criteria.

The titles and abstracts of the included publications were examined by two authors (M.J. and J.S.K.), with conflicts resolved through discussion with a third author (SB). Articles that were likely to meet the eligibility criteria were reviewed in full-text by two authors (M.J. and J.S.K.), and consensus achieved in cases of uncertainty through discussion with the other co-authors (S.B., J.S.K., A.K.G., J.G.K., and B.S.) before being excluded or included. Library requests were sent for any full texts that could not be directly accessed through the online databases.

Data were extracted using a standardised form and included the design of the study, the indications for IVIg therapy, number of patients with an AR, the type of AR, the rates of allergic and non-allergic reactions, treatments used for the AR, premedication utilised, and other preventative measures employed (such as prehydration and dose fractionation). Treatment details including the formulation, dose, rate, schedule, and route of infusion were also collected. For the included studies, the Cochrane risk of bias tool 2.0 [6] was used to critically analyse individual studies for the risk of selection bias, reporting bias, attrition bias, and other bias.

The search strategy results are summarised in Figure 1. The initial search strategy identified 2,196 publications. Following title and abstract review, 170 articles were viewed in full-text, resulting in 65 studies fulfilling the inclusion criteria: 48 cohort studies, 12 case series, three clinical trials, one case-control study, and one case-crossover study. The study characteristics and key findings of the included articles are summarised in online supplementary Table 1 (for all online suppl. material, see www.karger.com/doi/10.1159/000529110). The rates of individual ARs, suspected pathophysiology, risk factors, and possible preventative and management options suggested by the included literature are summarised in online supplementary Table 2.

In many studies, the definition of specific ARs was unclear or not specified. Reporting of AR data often lacked detail, particularly in articles reporting a large range of ARs. A large proportion of included studies were retrospective cohort studies or case series and therefore prone to selection bias. For a significant proportion of studies, it was unclear how the AR data were collected and what methods, if any, were used for systematically detecting ARs. For the studies which were randomised, the applicability of quality assessment tools was limited. For instance, while the efficacy of IVIg may have been studied in a placebo-controlled, well-blinded, randomised trial using standard laboratory measurements, the ARs of IVIg, on the other hand, may be collected via a self-assessment questionnaire. AR incidence rates varied greatly as AR data were rigorously collected by systematic survey and frequent laboratory studies in some studies, while other studies relied on spontaneous reporting of symptomatic ARs only. Incomplete reporting of results was also a common source of bias, as multiple studies reported AR categories (e.g., “dermatological AR”) without characterising the type (e.g., eczematous reaction) and severity of the AR. Furthermore, some studies, particularly those investigating a specific AR, neglected to report other non-severe ARs which may have been considered less significant, for example, injection site discomfort and hypertension.

Headache was one of the most frequently reported ARs in a number of studies [7, 32]. Headache was reported by 48% of the neurological patients during 68% of the treatments in one study, higher than that experienced by patients receiving IVIg for a non-neurological cause (41% and 22%, respectively) [33]. In a cohort of 50 patients receiving rapid-infusion IVIg for neuromuscular disorders, mild or moderate headache was reported in 22 patients (44%), and severe headache requiring hospitalisation was recorded in 1 patient (2%) [34]. A preexisting history of migraine was reported to be a risk factor for experiencing headache with IVIg infusion in several studies [8]. One study identified two statistically significant factors associated with post-IVIg headache: higher baseline systolic blood pressure and younger age [35]. Given findings of a higher pre-infusion systolic blood pressure in patients who experienced headache, this study suggested that dehydration is unlikely to be a major factor [35]. The findings of this study corroborate the fact that the evidence supporting rehydration before IVIg administration is poor (evidence level IV, recommendation D) compared with switching to a subcutaneous immunoglobulin (SCIg) (evidence level I, recommendation A) [35].

Possible management options are listed in online supplementary Table 2. The recommendation of switching to SCIg for headache is supported by one study where improvement in headache and nausea was reported in 50% of patients after switching from IVIg to SCIg [36]. Another study reported clinically relevant headache in 51% of patients after IVIg, and none following SCIg therapy [19]. A retrospective medical record review found that paracetamol and diphenhydramine were typically effective for mild to moderate headaches, and for severe headaches, prophylactic steroids (100–200 mg hydrocortisone intravenously or 100 mg prednisolone orally) were a successful premedication strategy [8]. Similarly, another study found that mild to moderate headaches were usually short-lived and respond to paracetamol, nonsteroid anti-inflammatory agents, or the slowing of infusion rates [37]. The use of paracetamol in the treatment of mild IVIg-induced headache was supported by two further studies [12, 26]. Vasomotor symptoms and headache may be controlled by reducing the flow rate according to one study [8]. On the contrary, another study reported that rate, dosage, and duration of infusion did not have a statistically significant influence on post-IVIg headache [35].

Aseptic meningitis was reported in eight of the included studies, with most of these reporting a rate between 0% and 11% [18, 20, 23, 26, 33, 37]. The development of aseptic meningitis did not correlate with age or type of underlying neuromuscular disease [37]. One study hypothesised that the entry of immunoglobulin or other components in the preparation into the subarachnoid space may trigger a hypersensitivity reaction, resulting in aseptic meningitis [7]. Aseptic meningitis was a clinical diagnosis made in 2.0% of paediatric patients in one study (n = 196), defined by the presence of at least three of the following: fever, headache, altered mental status, stiff neck, photophobia; specific CSF findings were not required for diagnosis [20]. On the other hand, a smaller study rigorously screened for cases of aseptic meningitis through neurological examinations before and after each infusion, and cerebrospinal fluid (CSF) analysis was indicated, reporting a much higher rate of aseptic meningitis (11%) in patients receiving high-dose IVIg for neuromuscular disorders [37]. CSF analysis in these 6 patients demonstrated pleocytosis in 4 patients (leukocyte count as high as 1,169 × 10⁶/L in 1 patient), eosinophilia in 3 patients, and IgG elevation in all patients; CSF cultures were negative in all patients [37]. One study reported 5 patients with headache associated with fever, nausea, and vomiting, although meningitis was documented with CSF studies in only 1 patient (1.6%) [7]. Aseptic meningitis developed in only two of 48 (4%) patients without a history of migraine, and in 4 of 8 (50%) patients with a history of migraine (p = 0.003) [37].

Most studies did not state whether patients who had developed aseptic meningitis were able to continue further infusions without recurrence; one study, however, reported that 2 patients received additional infusions without recurrence of aseptic meningitis, after having changed the IVIg lot or product [37]. In terms of preventative measures, one study reported the use of 1 L of normal saline as prehydration, as well as a graduated protocol starting with an initial rate of infusion of 3 g/h or the slowest rate practicable, with a gradual increase in rate of infusion by 1 g/h with subsequent infusions as tolerated [23]. This study reported that, from a cohort of 38 patients, only 1 case of presumed aseptic meningitis occurred, in the only patient for whom the prescribed graduated protocol was not followed [23]. Similarly, another study also supports lowering the infusion rate, using a low-concentration Ig product (3%), and adequate pre- and post-infusion hydration to reduce symptom severity [24]. On the other hand, one study did not find any benefit in slowing the rate of infusion and administration of corticosteroids for reducing the risk of aseptic meningitis [37]. For premedication, a large study reported a lower rate of aseptic meningitis (0.1%) in patients receiving diphenhydramine (n = 954) or paracetamol (n = 773) as premedication prior to receiving a home-based IVIg infusion; methylprednisolone premedication, however, did not appear beneficial (2.2% rate of aseptic meningitis) [25]. In terms of management options (see online suppl. Table 2), while mild to moderate headaches often respond well to paracetamol and nonsteroidal anti-inflammatory agents, patients with aseptic meningitis may not respond to simple measures and may require additional hospitalisation, close monitoring, and narcotic analgesics [37].

One study systematically measured blood pressure immediately before and after each infusion, on every day of the course, and found that individual courses of IVIg did not raise blood pressure [38]. Despite this, hypertension or “increased blood pressure” was reported in 11 studies with a rate between 0.5% [13] and 28% [18], although most studies did not explain how this diagnosis was made, and failed to comment on patients’ baseline blood pressure and pre-treatment hypertension status [10, 13, 15, 16, 18, 20, 21, 23, 25, 34, 39].

Hypotension was reported in 10 of the included articles and appeared to be reported frequently in paediatric cohorts [8, 10, 11, 13, 15, 16, 20, 21, 23, 39]. In a paediatric inpatient cohort of 115 patients, hypotension was the most frequent AR in 28.7% (all indications of IVIg) [16]. Hypotension is often reported in association with other symptoms; one study reported hypotension as a feature of one allergic reaction, accompanied by constrictive retrosternal chest pain, malaise, and somnolence occurring with at-home IVIg [15]. One case of hypotension was reported accompanying aseptic meningitis in a paediatric patient [20]. Another study reported a case of a severe hypotensive episode requiring vasopressor support in a patient with a recent myocardial infarction; further courses of IVIg were well tolerated [8].

Increased plasma viscosity is a well-recognised side effect of IVIg therapy, which may contribute to two of the more significant IVIg-ARs: headache and thromboembolic complications (TECs) [17, 38, 40, 41]. Cases of TEC of IVIg therapy reported include venous thromboembolism (VTE), myocardial infarction, acute ischaemic stroke, and arterial thrombosis [42, 43]. Two studies suggest that the hyperviscosity effect is likely secondary to hyperproteinemia induced by IVIg administration [41, 44]. Another study measured serum viscosity serially in 13 patients being treated with IVIg for either amyotrophic lateral sclerosis or IgM paraproteinemic polyneuropathy and found serum viscosity increased after IVIg in all patients by 0.1–1.0 centipoise (cp) (mean, 0.55 cp), which can impair blood flow and thereby trigger a TEC [40]. This is supported by a recent prospective study that investigated 20 patients receiving a 1–5-day course of IVIg therapy for various neurological diseases and found there to be an acute and cumulative rise in plasma hyperviscosity across a treatment course, with no concordant blood pressure rise [38]. One study suggests obtaining a baseline serum viscosity and to consider ongoing monitoring of these levels in elderly patients and other individuals with risk factors such as cryoglobulinaemia, monoclonal gammopathies, high lipoproteins, or preexisting vascular disease [40].

The risk of TEC of IVIg is important to recognise and has been consistently demonstrated in a large number of studies [8, 18, 24, 26, 31, 34, 39, 41, 45, 50]. A large case-crossover study involving 1.9 million admissions for TEC identified an odds ratio of 3.33 (1.34–8.30, p = 0.0097) for VTE within a 30-day window after exposure to IVIg for neurologic disease; no association was found with myocardial infarction or ischaemic stroke [42]. All index admissions for VTE were reported up to 120 days postexposure to IVIg for neurological conditions including CIDP, AIDP, MG, MS, and dermatomyositis [42]. Another study identified the risk of symptomatic TEC [51] to be significantly greater during courses given to IVIg-naïve neuropathy patients, reaching 12% in their analysis [49]. Therefore, they suggest the administration of all first-ever courses of IVIg to be undertaken in an exclusively hospital environment with adequate clinical monitoring and intensive care facilities [49]. In one study (n = 244), deep vein thrombosis (DVT) was reported in 6 patients (2.4%) [18]. In this study, DVT occurred after the first IVIg infusion in 4 of the patients [18]. Additionally, they found that age ≥60 years and age ≥75 years was not associated with an increased risk of DVT [18]. In a retrospective chart review (n = 61), the risk of developing TEC was suggested to be low in patients with immune-mediated neurological disease, at a rate of 4.91% per patient and 0.82% per infusion [46]. Another study (n = 61) identified a similar frequency of TEC at 4.4% despite 16 patients receiving prophylactic anticoagulation [48].

Possible risk factors for IVIg-associated TEC are listed in online supplementary Table 2. According to one study, patients being treated with IVIg for paraproteinemic polyneuropathy may have a baseline serum viscosity in, or even above, the high range of normal, and for these patients, even the slight serum viscosity rise with IVIg treatment may result in crossing a symptomatic “viscosity threshold” [40]. Further patient-related factors that may increase thromboembolic risk with IVIg may include: advancing age, advanced coronary atherosclerosis, or cerebrovascular disease as well as patients with borderline high serum viscosity due to cryoglobulins, high chylomicrons, triglycerides, or monoclonal gammopathies; for these patients, they recommend measuring serum viscosity at baseline before starting IVIg [40].

Two studies found that possessing ≥4 cardiovascular risk factors were significantly associated with TEC of IVIg, with an odds ratio of 10.5 (95% CI: 1.91–57.58) in the first study [45] and p value of 0.006 in the second study [49]. In the first study, 8 of 19 cases (42%) were treated for neurological indications, and a similar proportion of controls were treated under neurological conditions (45%) [52]. No single independent risk factor for TEC was found in this study [45]. The second study only included patients with neuropathy [49]. In this study, immobility and previous coronary disease were the only 2 individual patient risk factors found to be statistically associated with TECs of IVIg in the second study; age, gender, atrial fibrillation, diabetes, hypertension, and dyslipidaemia alone were not included [49]. Patients whose mobility is severely limited by the manifestations of their neurological condition may be predisposed toward the development of DVT; additional prophylactic measures may be considered in these patients [8]. There was one case of DVT reported in a patient with CIDP who was bedbound in one study [8] and another case reported in a patient whose mobility was limited by GBS [39]. Additionally, in another study, DVT was only reported in 2 patients: a patient with multiple sclerosis who was wheelchair-bound and another patient with inclusion body myositis with severe paraparesis [28]. As expected, in a cohort of 59 ICU patients receiving IVIg for various neurological and non-neurological indications, a higher rate of arterial or VTE (5.1%) was also reported [47]. One study reported the successful treatment of three out of four IVIg-induced arterial thromboses through administration of tissue plasminogen activator (tPA) [43].

Stroke was a reported AR in six of the included papers [7, 14, 25, 46, 51, 52]. One large case-crossover study reported no association between IVIg and ischaemic stroke [53]. Overall, stroke is an uncommon complication of IVIg treatment, with an approximate rate of 0.6% of all inpatient IVIg treatments based on a pharmacy record review of 498 patients in one study [52]. This study included a series of 16 cases of stroke associated with administration of IVIg; 10 of these patients were being treated for neuromuscular indications [52]. Of the 16 patients reported, 15 patients had one or more stroke risk factors [52]. A strong temporal relationship between stroke and IVIg administration was established in this series; of the 10 patients with a neurological indication for IVIg, 9 (90%) patients experienced stroke during or within 24 h of an IVIg infusion, and many developed symptoms while the infusion was being administered [52]. All cases of stroke involved large or medium-sized cerebral artery occlusion, and none were haemorrhagic [52]. These patients were managed with tissue plasminogen activator, anticoagulants, or antiplatelet agents [52]. One patient continued to receive identical IVIg treatment for their MG without further incident [52].

Several studies reported cases of myocardial infarction (MI) [18, 34, 47, 49, 51]. One case (0.4%) of MI was reported in a study of patients >60 years of age being treated for a dysimmune neuromuscular disease (n = 244) [18]. One small case-control study with approximately half of the cases and controls being treated for neurological indications found five cases of MI across all indications [45]. One case of MI was reported by another study in a cohort of 50 patients (2%) receiving rapid-infusion IVIg for neuromuscular indications. Chest pain without the diagnosis of MI was reported in five studies [7, 15, 16, 29, 34].

The occurrence of an isolated case of heart failure was reported in three among the analysed studies [7, 8, 34]. Both patients with reported cardiac ARs of IVIg therapy in one study had preexisting cardiac disease; it was suggested that the volume overload from IVIg likely resulted in the exacerbation of the patients’ heart failure [8]. Patients receiving IVIg for a neurological indication such as polymyositis which may be associated with secondary cardiomyopathy or other cardiovascular sequelae, may be at higher risk of developing heart failure with IVIg therapy [8]. The highest incidence of IVIg-associated heart failure (10%) was reported in a study of 50 patients with neuromuscular disorders [34].

Fatigue, lethargy, malaise, general weakness, or asthenia were reported in 16 of the included studies [7, 9, 12, 15, 20, 27, 31, 34, 36]. Myalgia, back, neck, and limb pain were also commonly reported [7, 8, 10, 11, 14, 15, 22, 24, 29, 34, 39]. Fever was reported in 20 studies [7, 11, 13, 15, 17, 20, 22, 24, 29, 32, 34]. Flu-like symptoms, including fever, almost always occur as immediate reactions, usually within the first hour of infusion [4, 13]. In the paediatric neurological patient cohort, one study reported a rate of 16.3% (10) and another study reported a lower rate of 4.6% (20).

Nausea and/or vomiting are common side effects reported in several studies, particularly in paediatric populations [7, 11, 13, 15, 20, 22, 25, 26, 29, 32, 34, 36, 39]. Nausea was reported in 10 (24%) of patients in one study [7]. Vomiting was reported in 29% of children treated with IVIg in one study, 25.6% of children in another study, and only 5.6% of children in another study [9, 17, 20]. Abdominal pain and diarrhoea are less frequently reported IVIg-ARs [14, 17, 20, 21, 25].

These were broadly defined as any combination of chills, chest tightness, skin flushing, and/or sweating. Vasomotor symptoms were reported in several studies, including 26 patients (2.95%) in 1 study of patients receiving IVIg for neurological disorders [8, 10, 11, 15, 16, 20, 21, 24, 26, 28, 34, 39].

Local injection site discomfort was reported in seven of the included studies [7, 13, 20, 22, 24, 25, 36]. Intravenous site discomfort was reported in 14 (33%) of neurological patients in one study [7] and in 4.6% of paediatric neuroimmunological patients in another study [20]. SCIg may be associated with a higher rate of a local injection site reaction compared with IVIg, at a rate between 50% [14] and 66.7% [54], as reported in two included studies.

Anaphylactic and anaphylactoid reactions are rare with IVIg therapy and have been reported in six included studies [8, 15, 18, 23, 30, 55]. Possible preventative options are listed in online supplementary Table 2. A recent large retrospective cohort study reported 37 anaphylactic reactions in 65,058 administrations for neurological indications, totalling to an incidence rate of 5.7 per 10,000 administrations (95% CI: 4.0–7.9) [55]. Age under 18 years and first IVIg administration were predictors of anaphylaxis across all IVIg indications [55]. One study (n = 46) reported one anaphylactic reaction with an at-home infusion of Tegeline^®[15]. One study suggests patients with a total IgA deficiency may experience an anaphylactic reaction due to IgE or IgG antibodies against IgA, which may react with the IgA in the IVIg preparation [56]. Most preparations nowadays, however, have addressed this with very low IgA levels in currently available products [56, 57]. Although the total absence of IgA is more common in patients receiving IVIg for combined variable immunodeficiency, it is rarely encountered in the neurological patient cohort; as such routine determination of IgA levels is not generally recommended in the latter subgroup. There is no significant increase in the risk of anaphylaxis with relative IgA deficiency in neurological patients [58] – one study (n = 90) reported relative IgA deficiency in two individuals (2.2%) who received 12 and 26 IVIg infusions each for inflammatory neuropathy without developing anaphylaxis [57].

Rashes and skin reactions were commonly reported ARs; however, the morphology of these was not described in most articles. Rash was reported in 18 of the included studies [7, 10, 13, 15, 17, 18, 20, 24, 26, 28, 36, 39, 59, 60]. Most of the evidence available on dermatological IVIg reactions was from case series. Types of skin reactions commonly described include generalised eczematous or erythematous eruptions [26, 59, 60], maculopapular [8, 17, 61], urticaria [8, 10, 14, 15, 28, 32], pompholyx [59, 61, 62], and desquamation [59, 62]. The eczematous skin reaction associated with IVIg often has a characteristic initial localisation to the palms and/or soles, before extending to the rest of the body [59, 61]. Localisation of skin reactions to the palms or soles have also been reported [8, 12, 28, 59, 60]. One study reported 4 cases of severe eczematous skin reactions occurring approximately 10 days after IVIg infusion [61]. A possible male predilection was observed with this AR; 8 of 9 cases of eczematous eruptions occurred in males in one retrospective case series [59], and 11 of 15 patients with IVIg-associated dermatological reactions were reported in males in another case series [60]. The time to onset of the skin eruption may vary: one case series of eczematous eruptions reported a median delay of 8.4 days after the start of IVIg infusion (range: 3–21) [59].

Complete resolution of eczematous skin reactions may occur either spontaneously or with systemic or topical steroid treatment [59, 61]. One case series reported resolution of three cases of pompholyx with 3 weeks of topical corticosteroids and zinc peroxide at 1% and discontinuation of IVIg [62]. These three cases of pompholyx occurred despite a very slow rate of infusion (12 h, 0.4 g/kg/day) used [62]. As seen in another retrospective case series, reintroducing the same type of IVIg may result in relapse in some, but not all patients, thus, switching IVIg preparation may be advised [59]. This study suggests that treatment withdrawal, however, is usually not required [59]. After developing a reaction, premedication with antihistamines and/or oral prednisolone was administered with subsequent IVIg doses in one case series; however, only 4 patients responded favourably to pre-treatment and continued with the IVIg [60]. Pruritis was reported in several studies, sometimes, but not always in association with a rash [8, 14, 25, 26, 59, 61]. One case series reported unintentional rapid and significant hair growth on the scalp and arms in 2 patients with androgenic alopecia after IVIg [63].

Isolated incidences of transfusion-related acute lung injury (TRALI) were reported in three studies [34, 47, 64]. One study reported an immediate-type reaction where the patient developed acute respiratory distress during the infusion, after receiving two thirds of the prescribed IVIg [64]. One study reported that the presumed TRALI occurred 1 day after the completion of IVIg; however, the patient had also received multiple other blood products which may have contributed to significant volume overload [47]. A retrospective case series reported the occurrence of TRALI in 2 patients receiving 0.5 g/kg Tegeline^® IVIg for acute polyradiculoneuritis and neurolupus, within 0.1 h and 3 h from time of infusion, respectively [64]. A retrospective cohort study reported 1 patient with TRALI (2%) receiving rapid-infusion IVIg for neuromuscular disease [34].

Some degree of renal impairment was reported in six studies; however, this AR was not clearly defined in most studies [8, 13, 18, 29, 47, 65]. In one study (n = 46), the average time to onset of acute renal failure was 3.7 days (range: 2–5 days), with the time to maximum serum creatinine occurring at 5.17 days after initiation of IVIg (range: 3–8 days) [65]. A retrospective cohort study of inflammatory neuropathy patients (n = 90) estimated the risk of Ig-related acute kidney injury (AKI) to be 0–0.1%, an order of magnitude lower than that reported in historical literature: 0.77%; they explained this result may be due to the decline in the use of sucrose-containing Ig products [57]. Several studies suggested that obtaining a baseline creatinine level before commencing IVIg was part of their standard practice; this measurement may assist in monitoring for renal impairment with IVIg infusions, and the identification of patients who are at a greater risk of IVIg renal impairment [29, 57]. Age ≥60 years and age ≥75 years were not associated with an increased risk of AKI in one retrospective cohort study (n = 244) [18]. A retrospective chart review (n = 64) specifically studied IVIg-associated AKI (≥0.5 mg/dL rise in serum creatinine within 10 days from the initiation of IVIg therapy), with cases reported in 6 patients (13%), all after the first cycle of IVIg with the sucrose-containing Carimune^® formula [65]. In this study, a baseline serum creatinine was obtained 2 days before IVIg for all patients, then measured 2 days after IVIg, and the maximum value was recorded within 10 days of the IVIg therapy [65]. No statistically significant predictors of this AR were found in this study, nor was association with the underlying primary disease [65]. A retrospective record review reported 1 case of acute oliguric renal failure in a patient with preexisting mild diabetic nephropathy after receiving 2 consecutive days of IVIg [8]. The role of prehydration in mitigating the risk of IVIg-induced renal impairment was not formally investigated in these studies.

It has been well recognised that IVIg therapy is associated with a decrease in haemoglobin, haematocrit, platelet, and white cell counts, as well as sodium levels; however, these are rarely clinically significant [8, 26, 28, 29, 32, 39, 44, 47, 50, 57, 66, 68]. Several studies focussing on laboratory abnormalities following IVIg found an association with anaemia [27, 39, 67, 68], a transient neutropenia [8, 28, 29, 32], and leukopenia [8, 26, 29, 32, 68]; however, in most cases reported, these abnormalities were reversible and of no clinical relevance. A small prospective cohort study (n = 15) reported a statistically significant, but not clinically significant hyponatremia in all patients studied [41]. Statistical analysis demonstrated that hyponatremia following IVIg may in fact be pseudohyponatremia due to an IVIg-induced hyperproteinemia [41]. Following IVIg therapy, a significant increase in the erythrocyte sedimentation rate [28, 44], total protein [26, 41, 44], and IgG levels [26, 44, 68] may be observed for up to a month [44]. A recent retrospective cohort study (n = 104) found that high serum total protein levels in paediatric neurological patients were significantly associated with the occurrence of IVIg-ARs (OR, 14.8; 95% confidence interval, 2.4–90.5; p < 0.01) [17]. A reversible, asymptomatic rise in liver enzymes, particularly transaminase levels is also common with IVIg administration [20, 23, 26, 39, 44, 69].

Haemolytic anaemia in association with IVIg infusions may be due to haemolysis as a result of passive immunisation from ABO antibodies [68]. One cohort study reported a high rate of haemolytic anaemia (31.6%; n = 11), three required blood transfusions, and 1 patient required discontinuation of IVIg [50]. Interestingly, this study noted that although most of their patients had received IVIg for haematological indications at the standard dose of 0.4 g/kg monthly, most of the cases of haemolysis (8 of 12) occurred in patients receiving IVIg at a greater dose for neurological disorders (0.4 g/kg per day for 5 days) [50]. Another study reported one case of a child with a haemolytic reaction with fever and lethargy associated with IVIg use [20]. One cohort study reported a case of severe haemolytic anaemia requiring discontinuation of IVIg in a patient with a paraneoplastic neurological syndrome [27]. In terms of management, findings from one study suggest that if a patient develops severe haemolysis from the first lot of IVIg, then therapy may be resumed using a different lot once their haemoglobin stabilised; no patients developed haemolysis from the second lot in their study [50]. It may be safe to continue the same lot of IVIg in cases of mild asymptomatic haemolysis, with close surveillance of haemoglobin levels [50].

In 1995, one case series reported 10 of 23 (44%) patients who received Gammagard^® between March 1993 and February 1994 and were screened were subsequently found to be anti-hepatitis C virus (HCV) positive [70]. In 1996, one retrospective record review reported the transmission of polymerase chain reaction (PCR)-proven HCV in 4 patients receiving IVIg who were all immunosuppressed with corticosteroids; 3 patients subsequently developed chronic hepatitis [7]. Revised manufacturing processes aiming to eliminate viral infection have significantly reduced this risk since the time of these studies. Our search revealed only one recent publication reporting possible transmission of viral hepatitis in the last 20 years. A prospective study from Iran conducted in 2013 reported 2 female cases with CIDP that showed positive results for hepatitis B virus surface antibody (0.8%) and hepatitis B surface antigen (0.8%) with ELISA after IVIg, with only 1 case was confirmed with PCR [71]. However, the authors stated that it was unclear whether the infection was transmitted through IVIg therapy or via another route [71].

The highest rate of infection was reported in a multicentre clinical trial of 17 patients receiving IVIg for paraneoplastic syndromes, 1 patient had a catheter infection, 2 developed sepsis, and 1 subsequently died from sepsis [72]. However, sepsis is a recognised cause of death in cancer patients, so it is difficult to attribute IVIg treatment as the sole cause of this complication [72]. Line-related infection was reported in 4 patients (6.78%) in a cohort of intensive care unit patients [47], and another case of catheter sepsis was reported in a retrospective cohort study (n = 244) [18].

Most studies did not specify the rate of infusion used. The association between IVIg-AR occurrence and infusion rate is still controversial; one retrospective cohort study (n = 104) found no statistically significant association between the infusion rate and incidence of ARs [17]. In a cohort of 50 patients receiving rapid-infusion IVIg, ARs were reported in 76% of patients including severe ARs in 22%; these figures are markedly higher than reported in other neuromuscular centres where standard infusion rates were used [34]. These ARs included five cases of congestive heart failure – the authors suggest that it would be prudent to avoid rapid infusion in patients with cardiac disease, although a statistically significant association was not established [34]. A retrospective record review suggested no relationship between headache and the rate of infusion; however, they did find this AR to be dose-dependent [7]. In one cohort study, serum viscosity increased with high-dose IVIg infusion (2 g/kg) in all patients regardless of rate of infusion [40].

Several studies have reported a dose-dependent risk of haemolytic anaemia, and thus, a higher rate was found in patients being treated with high-dose IVIg for neurological indications compared to non-neurological indications [50, 67, 68]. With respect to other ARs, a retrospective chart review (n = 424) did not find a statistically significant difference in the incidence of ARs in neurological patients compared to patients being treated for an immunodeficiency syndrome (2.28% compared with 1.94%, p = 0.6), despite the higher average dose of Ig used and older age in the former group [24]. A retrospective cohort study (n = 104) found that a statistically significant association of IVIg-ARs with duration, but not with dose of IVIg [17]. In this study, the median duration of IVIg therapy was 3 days (interquartile range of 3–5 days) [17]. One series involving 16 cases of IVIg-associated stroke did not establish an association between total dose of IVIg infused and the occurrence of stroke [52]. Similarly, a small cohort study of 6 patients receiving “ultra-high dose IVIg” at >2 g/kg/6 weeks at high frequency, and in some cases for a prolonged time period, found no ARs encountered; as such, the authors suggest that IVIg-related ARs are likely influenced by individual risks as well as the dose and rate of the infusion itself [73].

In a retrospective cohort study (n = 1176) of home-based IVIg use for both neurological and non-neurological indications, the mean number of ARs per patient increased from 1.4 in the low-dose patients to 3.6 in the high-dose patients, and then decreased in the very high-dose group [25]. In this study, the indications for IVIg infusions were neuroimmunological for the majority of the patients, including neuropathy (n = 651, 54.5%), MS (n = 104, 8.8%), neuromuscular junction disorders (n = 101, 8.6%), and myopathy (n = 14, 1.2%) [25]. Other indications included immunodeficiency (n = 165, 14.0%), connective tissue disorders, (n = 52, 4.4%), and other disorders (n = 99, 8.4%) [25]. Daily dose ≥35 g was found to be an independent predictor of AR occurrence in a retrospective cohort study [18] and was also significantly correlated with occurrence of TEC within 14 days after IVIg infusion in another study [49]. The IVIg dose found in one study was found to be correlated with the intensity of nausea (n = 54; r² = 0.12; p = 0.01) but not with the intensity of headache (n = 58, r² = 0.05; p = 0.1) [19].

Two retrospective cohort studies found no statistically significant association between IVIg preparation and incidence of ARs [17, 24]. Sucrose is used as a stabiliser in some IVIg preparations. There have been suggestions in the literature regarding a potential association between the use of a high sucrose formulation of IVIg (such as Sandoglobulin^®) with the occurrence of sucrose nephropathy and AKI; however, the evidence for this effect is equivocal [64]. One study reported that after 12 months of therapy, 15% of patients on Gamunex^®, 15.63% of patients on Privigen^®, and 13.89% of patients on Bivigam^® developed a ≥20% decline in GFR; however, only 5.2% of patients on the sucrose-containing Carimune^® formulation developed this AR [66]. Additionally, four of the five cases of AKI reported in a retrospective cohort study (n = 244) occurred in patients from the subgroup (n = 76) infused with sucrose-free IVIg, including three of four without previous renal dysfunction [18]. Conversely, 1 study reported 6 patients (13%) developing AKI after Carimune^® infusion [65]. One patient receiving SCIg for CIDP reported an urticarial rash with most infusions with the Hizentra^® brand, but not with Gammanorm^® (24) 1 case series reported 15 patients who had developed a generalised erythematous rash or desquamation after receiving IVIg; Intragram P^® was the most common brand implicated [60].

One retrospective cohort study reported that ARs were significantly more common in patients treated with sucrose-free IVIg [18]. Sugar-stabilised IVIg formula may be associated with eczematous skin reactions – this AR was the main reason for discontinuation (58%) of this product in the study [18]. A prospective case series suggested that maltose-containing stabilising agents in IVIg preparations may cause a self-limited, asymptomatic transaminitis [69]. One prospective cohort study also found the occurrence of transaminitis to be associated with sucrose and PEG/sorbitol-containing IVIg preparations as opposed to glycine-containing preparations; the underlying mechanism was unclear [26]. To reduce the risk of an anaphylactic reaction, an IVIg preparation containing lower amounts of IgA for patients with low serum IgA levels should be used [56, 57].

Several studies support the use of a routine full blood count (FBC), electrolytes-urea-creatinine (EUC), and liver function test (LFT) before commencing treatment [26, 69]. Obtaining a baseline FBC (focussing on haemoglobin, white cell count, platelet count, and absolute neutrophil count) and EUC (particularly creatinine) is advised to assist in identifying patients at greater risk of complications [56, 57]. Findings of one retrospective cohort study suggest that routine pre-infusion monitoring of FBC and EUC is not indicated; obtaining a baseline followed by annual screening and clinically indicated testing is safe and more appropriate with long-term IVIg use [57]. There was no significant evidence from the included studies to support the utility of baseline and ongoing screening for hepatitis B, C, and other blood-borne viruses in IVIg recipients.

Most studies did not report the use of a premedication regimen, and there were no randomised studies that formally investigated the efficacy of this strategy. The most common premedication given among the included studies was antihistamine [9, 24, 46]. A retrospective chart review (n = 420) reported a significantly lower incidence of ARs in the group of patients (65.7%) who received either a single agent or combined premedication regimen consisting of paracetamol (1 g), diphenhydramine (60 mg), and/or dexamethasone (12 mg) compared to those who did not (18.4% compared with 27.1%, p = 0.04) [24]. This study included 334 patients with neuroimmunological disorders and 86 patients with immunodeficiency syndromes as a control group [24]. Interestingly, the authors report that the effect of premedication was more significant in the neuroimmunological group: the AR rate was significantly lower in the premedicated group (18.2%) compared to the non-premedicated group (29.3%) (p = 0.02) [24]. On the other hand, in the immunodeficiency group, there was no significant difference in the incidence of ARs with (19.5%) or without (22.2%) premedication [24]. While premedication is used fairly commonly in routine practice, the use of corticosteroid premedication is rarely reported. Use of premedication strategies was reported most often in studies with paediatric cohorts [9, 13, 59]. One retrospective cohort study (n = 186) recently reported the use of antipyretics and antihistamines in all patients to reduce ARs; ARs were reported in 15 (3.8%) of a total of 394 IVIg doses [13].

Management of ARs is complex and varies depending on the nature of the reaction. Several studies suggest that discontinuation of IVIg is usually not required in the case of mild-moderate ARs if there is major clinical benefit of the IVIg for the patient [47, 59]. In a multicentre prospective cohort study (n = 544) across both neurological and non-neurological indications for IVIg, discontinuation, or modification of the IVIg regimen due to an AR was only required in 16 of the 119 patients (13.4%) reporting an AR [22]. A retrospective cohort study reported success with the use of scheduled nonsteroidal anti-inflammatory agents, antihistamines, and a litre of normal saline post-infusion for mild symptoms; however, the efficacy of these were not formally investigated in most studies [23].

Switching from an intravenous to subcutaneous mode of administration may be a useful strategy that may avoid the need to cease Ig therapy entirely. Our search revealed data from several small prospective observational studies to support the safety of SCIg as an alternative method of administration of Ig in neurologic disorders [14, 31, 36]. A recent prospective, observational, and longitudinal multicentre study investigated patients recently switched over from IVIg to high-dose, self-administered, and home-based SCIg in the treatment of neurological conditions [14]. Three of these patients had been switched over to SCIg due to poor tolerance with IVIg [14]. The occurrence of a local injection site reaction, including pain and pruritis, appeared to be the only AR reported at a markedly higher rate with SCIg than with IVIg in the literature [14]. Findings of several included studies strongly support switching to SCIg as an effective management strategy for patients experiencing headache and nausea with IVIg [14, 19, 36]. ARs of SCIg were generally reported to be mild and infrequent [36].

The findings of our review suggest that the safety of IVIg administration in patients with neurological disorders is likely similar to that in other indications [24]; however, some patient groups may be at an increased risk for ARs. Among the included studies, the reported incidence rate of ARs with IVIg for neurological indications, ranged between 0% [57, 73] and 100% [31]; the significant discrepancy observed likely relates to wide variability in sample size, follow-up duration, and more rigorous AR detection and reporting methods in some studies. True estimates of IVIg-AR incidence in neurological disease are likely to be closer to the 25.5% [20] and 34% [18] per patient rates reported by the two larger cohort studies published. The precise incidence of IVIg-ARs could not be established with meta-analysis due to heterogenous populations studied and discrepant findings. Despite one observational study [3] suggesting neuromuscular disease to be a significant risk factor for AR with IVIg, our review did not offer substantial evidence to support a marked difference in the AR rate between patients receiving IVIg for neurological indications versus non-neurological indications. However, there is a small amount of data to suggest a potentially lower risk of myocardial infarction and acute ischaemic stroke [42] as well as a possibly higher risk of haemolytic anaemia [50, 67, 68] in patients receiving IVIg for neurological indications, although this distinction requires further exploration.

Overall, the general safety profile of IVIg is favourable, particularly when compared to the AR rates of the alternative treatment options for neuroimmunological disease, such as corticosteroids and other immunosuppressive drugs. Nonetheless, the utility of IVIg is restricted by its limited availability; as such, immunosuppressive drug therapy may still be preferable in the long-term for some patients, despite higher AR rates.

The mechanisms underlying different IVIg-ARs are variable and are yet to be fully elucidated [17]. Baseline laboratory studies provide valuable information that can assist with patient risk assessment and AR monitoring. Various mechanisms contributing to ARs may include increased serum viscosity, haemolysis, allergic and immunological mechanisms, or effects relating to the physicochemical properties of the Ig or their excipients (as in IVIg-induced nephrotoxicity). Suspected pathophysiology mechanisms underlying specific ARs are described in online supplementary Table 2.

Mild ARs of IVIg administration for neurological disorders are common (see online suppl. Table 1) [10, 26]. Headache was the most frequently reported AR. The risk of severe ARs, such as TEC, anaphylaxis, renal impairment, and cardiovascular complications, is low, but cannot be dismissed, particularly in patients with multiple cardiovascular or thromboembolic risk factors [42, 45, 49]. As such, careful consideration of patient risk factors is required to guide an increase in monitoring practices, optimisation of treatment specifications (dose, rate, and preparation) and the use of a premedication regimen as appropriate for the individual. With respect to monitoring, clinicians should be aware that abnormalities of specific laboratory parameters, particularly neutropenia, anaemia, and transaminitis, are commonly reported with IVIg use; however, such results should be interpreted with caution in asymptomatic patients as they are typically reversible and of no clinical relevance [32, 44, 68].

In terms of patient-related risk factors, gender and age did not appear to be consistent predictors for IVIg-ARs in most studies [18], except in eczematous skin reactions where a male predominance was observed [59, 60]. Particular IVIg-ARs including fever, hypotension, and anaphylaxis appeared to be reported more frequently in paediatric cohorts. The standard risk factors expected to predispose complications in the general population, still apply in the neurological patient cohort receiving IVIg: coronary disease, other cardiovascular risk factors, and immobility are still important risk factors of TEC to consider. Our review did not aim to systematically compare the difference in AR rates between neurological conditions; however, we note that three cohort studies found no association between the underlying disease and the incidence of ARs [26, 34, 37]. A retrospective cohort study investigating IVIg use in a range of neurological disorders found epilepsy and peripheral nervous system disorders to be associated with a higher rate of ARs [17]. Neurological conditions, which may result in significant deterioration in mobility and functional status [42], increase patient’s serum viscosity (e.g., paraproteinemic polyneuropathy) or have significant cardiovascular and renal sequelae, which may increase patients’ risk of ARs. Overall, there is a lack of high-quality comparative data to definitively determine if any specific neurological indications are associated with a higher risk of ARs; this association was not formally investigated in most studies given the number of patients per indication was typically too small to allow for a statistically significant comparison between indications.

For most non-severe ARs, ceasing IVIg may be unnecessary if the patient is experiencing a significant degree of improvement with their neurological disease; switching IVIg preparation (manufacturer or brand) could be an alternative option. This may be a useful strategy particularly in cases of IVIg-associated transaminitis, renal impairment, skin reactions, and anaphylactic reactions, where there have been previous reports of an association of the AR to the preparation and its excipients, rather than the immunoglobulin itself. Other preventative strategies including premedication may play an important role in improving tolerability of IVIg. Patients’ risk of developing anaphylaxis, renal impairment, and TEC may be increased to some extent with the first infusion; thus, increased monitoring and vigilance during and after patients’ first IVIg infusion would be prudent. We found the evidence to support the association between IVIg infusion rate and the occurrence of ARs to be limited, although a daily dose of IVIg ≥35 g may increase TEC risk. There is small-scale observational evidence to support the use of antihistamines and paracetamol in both premedication and treatment regimens for mild to moderate ARs; however, randomised controlled trials are required to provide concrete evidence to guide the prevention and treatment of IVIg-ARs.

Severe ARs such as significant TEC, severe cutaneous reactions (including Stephens-Johnson syndrome and toxic epidermal necrolysis), and aseptic meningitis constitute absolute contraindications to the future use of IVIg for most patients. Transfusion-associated circulatory overload, TRALI, significant hypotension, renal impairment, moderate cutaneous reactions, and cases of recurrent mild-moderate AR constitute relative contraindications to IVIg. There may be exceptions in cases where there are no other viable alternatives, the association to IVIg is unclear, the risk-to-benefit ratio of persevering with IVIg is favourable, and there is an effective treatment available for the AR should it recur. In such cases, clinicians must have a clear plan for re-trialling IVIg that utilises preventative strategies, premedication, and increased monitoring, ideally with expert guidance from an immunology specialist.

There is evidence from small, nonrandomised studies to suggest that in patients who experience intolerable ARs (particularly, headache and nausea) but still derive benefit from IVIg for their neurological disease, switching to SCIg may be a viable option. Severe ARs with SCIg including cerebral infarction and myocarditis have still been reported, although SCIg may be associated with a lower incidence of nausea and headache. With SCIg therapy, however, injection site reactions may be more common than with IVIg. With training, the SC route allows for home-based treatment, offering another advantage of independence from hospital-based treatment. Currently, the widespread use of SCIg in autoimmune neurological disease is limited by the high dose requirement for immunomodulation and lack of proven efficacy data. Our review only revealed observational data from small studies supporting improved tolerability and clinical equivalence with SCIg, without direct comparison between IV and SC administration routes. Randomised studies with a longer observational period are required to confirm these results.

Limitations of this systematic review include the small sample size for some outcomes, making it difficult to generalise these results to the general population. There was significant heterogeneity among studies due to varied follow-up times, use of different treatment specifications, varying sample sizes, follow-up durations, and populations studied. In some studies, the AR data for patients receiving IVIg for both neurological and non-neurological indications were reported together, limiting an effective comparison of AR risk between the groups. The lack of robust evidence from randomised studies limits the ability of this review to provide recommendations regarding premedication regimen, and other prevention and treatment strategies. Some studies did not report the overall rate or incidence of ARs. A large number of studies reported the incidence rate of AR per patient or per treatment course, rather than per infusion; the former statistic can be potentially skewed by the number of infusions received. This review did not assess evidence from primary literature that was not available in English. However, strength of this review is that it included studies from 17 countries.

While the incidence of headache and other mild ARs are common, IVIg is generally well tolerated in neurological patients. The risk of severe ARs with IVIg therapy for neurological indications is low. However, there are subgroups of patients that may be at an increased risk of IVIg complications. These subgroups include patients with limited mobility, paraproteinaemia, and cardiomyopathy. Paediatric patients may be at an increased risk of AR, including anaphylaxis. Overall, this review identified a paucity of literature systematically studying the efficacy of approaches to prevent and treat IVIg-ARs. Described strategies include alternative rates of infusion, doses of IVIg, and premedication. Future research in this area is warranted to improve the prediction, prevention, and treatment of IVIg-AR in patients with neurological conditions.

The authors have no conflicts of interest to declare.

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Dr. Melinda Jiang made substantial contributions to the interpretation of data for the work, drafting the work, revising the work critically for important intellectual content, and for the final approval of the version to be published Dr. James Kimber made substantial contributions to the acquisition, analysis, and interpretation of data for the work. Dr. Aashray Gupta, Dr. Joshua Kovoor, and Dr. Brandon Stretton made substantial contributions to the interpretation of data for the work and revising the work critically for important intellectual content. Dr. Janakan Ravindran, Associate Professor Pravin Hissaria, and Dr. William Smith made substantial contributions to the interpretation of data for the work and revising the work critically for important intellectual content and for the final approval of the version to be published. Stephen Bacchi made substantial contributions to the conception, design of the work, acquisition, analysis, interpretation of data, drafting the work, revising the work critically, and for the final approval of the version to be published.

All data generated or analyzed during this study are included in this article and its online supplementary material. Further enquiries can be directed to the corresponding author.

Edited by: Hans-Uwe Simon, Bern.