Thromboembolic risk in subcutaneous versus intravenous immunoglobulin therapy: a systematic review and narrative meta-synthesis

Abstract

Intravenous immunoglobulin (IVIg) therapy is associated with thromboembolic complications, including deep vein thrombosis (DVT), pulmonary embolism (PE), ischaemic stroke, and myocardial infarction (MI). Subcutaneous immunoglobulin (SCIg) achieves lower peak serum IgG concentrations and may confer a reduced thromboembolic risk. We systematically reviewed the comparative thromboembolic safety of SCIg versus IVIg across neurological and immunological indications. We searched PubMed, Embase, Cochrane Library, Web of Science, Scopus, and ClinicalTrials.gov from inception to December 2024. Studies reporting thromboembolic events in patients receiving IVIg or SCIg for any indication were eligible. Two reviewers independently screened, extracted data, and assessed risk of bias. Where feasible, pooled incidence rates with 95% confidence intervals (CIs) were calculated using Freeman-Tukey double arcsine transformation. Thirty-four studies (approximately 12,800 patient-years of immunoglobulin exposure) were included. IVIg-treated patients (approximately 10,600 patient-years across 27 studies) experienced thromboembolic events at a pooled incidence of 1.8 per 100 patient-years (95% CI: 0.9–3.1). SCIg-treated patients (approximately 2,200 patient-years across 7 studies) had a pooled incidence of 0.14 per 100 patient-years (95% CI: 0.01–0.52). Only 3 thromboembolic events were documented in SCIg cohorts, all in patients with pre-existing prothrombotic risk factors. No head-to-head randomised trial comparing thromboembolic outcomes between routes exists. GRADE certainty was very low for the comparative estimate. Indirect evidence suggests that SCIg may be associated with a lower thromboembolic risk than IVIg, consistent with pharmacokinetic predictions. However, substantial differences in patient populations, follow-up duration, and event ascertainment between SCIg and IVIg cohorts preclude definitive conclusions. Prospective comparative trials with pre-specified thromboembolic endpoints are needed.

Introduction

Immunoglobulin therapy is a cornerstone of treatment for autoimmune neurological disorders, including chronic inflammatory demyelinating polyneuropathy (CIDP), myasthenia gravis (MG), Guillain-Barré syndrome (GBS), and stiff-person spectrum disorder (SPS), as well as primary immunodeficiency and selected haematological conditions [1–3]. Intravenous immunoglobulin (IVIg) has been the predominant route of administration for decades, with well-established efficacy in randomised controlled trials (RCTs) [4–6]. However, IVIg carries a recognised risk of thromboembolic complications—including deep vein thrombosis (DVT), pulmonary embolism (PE), ischaemic stroke, and myocardial infarction (MI)—attributed primarily to transient plasma hyperviscosity following rapid intravenous infusion [7–10].

The European Academy of Neurology/Peripheral Nerve Society (EAN/PNS) guideline on CIDP management recommends immunoglobulin therapy as first-line treatment, with both IVIg and SCIg recognised as valid administration routes [11]. For GBS, IVIg remains the standard acute treatment, while for SPS, IVIg has demonstrated efficacy in controlled trials [12]. In MG, IVIg is used for acute exacerbations and as maintenance therapy, with emerging evidence supporting SCIg as an alternative maintenance route [13, 14].

Subcutaneous immunoglobulin (SCIg) delivers equivalent total doses via slow subcutaneous absorption, producing lower peak serum IgG concentrations and more stable steady-state levels compared with IVIg [15–17]. This pharmacokinetic profile theoretically reduces plasma viscosity peaks and the associated thromboembolic risk [9]. The PATH trial demonstrated SCIg efficacy for CIDP maintenance [2], and the ADVANCE-CIDP 1 trial confirmed non-inferiority of SCIg versus IVIg [14]. However, no trial has been specifically powered to compare thromboembolic outcomes between routes.

We conducted a systematic review and narrative meta-synthesis to comprehensively evaluate the thromboembolic risk associated with SCIg versus IVIg, synthesise the available comparative evidence, identify patient-level risk factors, and delineate the evidence gaps requiring prospective investigation.

Methods

Protocol and registration

This systematic review was conducted in accordance with PRISMA 2020 reporting standards [18]. The review methodology was pre-specified and internally documented prior to data extraction. Given the nature of this study as a secondary analysis of published data, formal ethical approval was not required.

Search strategy

We searched PubMed, Embase, Cochrane Library, Web of Science, Scopus, and ClinicalTrials.gov from inception to December 2024. Search terms combined immunoglobulin-related terms (“intravenous immunoglobulin”, “IVIg”, “subcutaneous immunoglobulin”, “SCIg”) with thromboembolic outcome terms (“thromboembolism”, “deep vein thrombosis”, “pulmonary embolism”, “stroke”, “myocardial infarction”, “thrombosis”). Reference lists of included studies and relevant reviews were hand-searched. The FDA Adverse Event Reporting System (FAERS) database was searched for spontaneous thromboembolic reports attributed to immunoglobulin products.

Study selection and data extraction

Two independent reviewers (SF and SC) screened all records by title and abstract, followed by full-text review. Disagreements were resolved by discussion; in cases of persistent disagreement, a consensus meeting with structured criteria adjudication was used to reach final decisions. Data were extracted using a pre-specified form capturing: study design, geographic setting, indication, participant numbers, demographics, immunoglobulin dose and regimen, follow-up duration, thromboembolic event types and counts, and risk factor data.

Risk of bias and evidence certainty

RCTs were assessed using the Cochrane Risk of Bias 2 (RoB 2) tool [19]. Observational studies were assessed using the Newcastle-Ottawa Scale (NOS) [20]. Case series were appraised using the Joanna Briggs Institute (JBI) checklist. Evidence certainty was graded using GRADE methodology, considering risk of bias, inconsistency, indirectness, imprecision, and publication bias.

Synthesis approach

Given substantial heterogeneity across study designs, populations, and outcome definitions, formal meta-analysis was feasible only for a subset of studies. Where at least three studies reported comparable outcomes with extractable event counts, pooled incidence rates with 95% confidence intervals (CIs) were calculated using the Freeman-Tukey double arcsine transformation. Between-study heterogeneity was quantified using the I² statistic [21]. Funnel plot asymmetry was assessed using Egger’s test where 10 or more studies were available [22]. Analyses were performed in R (version 4.3.2) using the ‘meta’ and ‘metafor’ packages.

Eligibility criteria

We included RCTs, prospective and retrospective cohort studies, and large case series (n ≥ 10 patients) reporting thromboembolic events in adult patients (≥ 18 years) receiving IVIg or SCIg for any indication. Paediatric-only studies were excluded from the primary analysis, as the thromboembolic risk profile in children differs substantially from adults due to lower baseline cardiovascular risk. Single case reports were excluded due to high reporting bias.

Results

Study selection

The initial search retrieved 4,312 records, with 87 additional records from hand-searching and ClinicalTrials.gov. After deduplication, 3,891 records were screened, 312 underwent full-text review, and 34 studies met eligibility criteria (Fig. 1). Exclusion reasons at full-text review included: no extractable thromboembolic data (n = 118), single case report design (n = 72), paediatric-only populations (n = 41), and duplicate cohorts (n = 29).

Fig. 1.

PRISMA 2020 flow diagram illustrating the study selection process

Characteristics of included studies

Included studies comprised 11 RCTs, 14 prospective cohort studies, 7 retrospective cohort studies, and 2 large case series from 18 countries (Table 1). Publication dates ranged from 1992 to 2024. Total immunoglobulin exposure was approximately 12,800 patient-years (IVIg: approximately 10,600; SCIg: approximately 2,200).

Table 1.

Summary of Included Study Characteristics by Administration Route

Characteristic	IVIg Studies (n = 27)	SCIg Studies (n = 7)
Total patients (approx.)	~ 11,400	~ 820
Total patient-years (approx.)	~ 10,600	~ 2,200
Median follow-up	12–36 months	6–24 months
Study designs	8 RCTs, 12 prospective, 5 retrospective, 2 case series	3 RCTs, 3 prospective, 1 retrospective
Mean age range (years)	38.4–69.7	42.1–61.3
Indications	CIDP (44%), MG (26%), GBS (11%), PID (11%), Other (8%)	CIDP (57%), PID (29%), MG (14%)
Dose range	0.4–2.0 g/kg/cycle (q3–8 weeks)	0.1–0.4 g/kg/week
Geographic distribution	Europe (52%), N. America (33%), Asia-Pacific (11%), Mixed (4%)	Europe (57%), N. America (29%), Asia-Pacific (14%)

Thromboembolic events

Among IVIg-treated patients, DVT and PE were the most commonly reported thromboembolic complications, occurring in 0.3–5.0% of patients per treatment course. The highest rates (3.0–5.0%) were observed in elderly patients with multiple cardiovascular comorbidities. Temporal analysis consistently identified peak risk within 24–96 h of infusion completion, corresponding to post-infusion peaks in serum IgG concentration and plasma viscosity [9, 23]. Ischaemic stroke was reported in 0.1–1.2% of patients, with multifactorial mechanisms including hyperviscosity-mediated reduction in cerebrovascular flow [23, 24]. MI was less frequently reported, appearing predominantly in pharmacovigilance data and patients with established coronary disease [25] (Table 2).

Table 2.

Thromboembolic Event Incidence by Route and Outcome Type

Outcome	IVIg: Events/Studies	IVIg: Pooled Incidence (per 100 pt-yrs, 95% CI)	SCIg: Events/Studies
DVT/PE	89 events / 26 studies	1.2 (0.6–2.1)	2 events / 7 studies
Ischaemic stroke	28 events / 18 studies	0.4 (0.2–0.8)	1 event / 7 studies
MI	11 events / 12 studies	0.2 (0.05–0.4)	0 events / 7 studies
Composite	128 events / 27 studies	1.8 (0.9–3.1)	3 events / 7 studies
SCIg composite	—	—	0.14 (0.01–0.52)

In addition to thromboembolic events, IVIg is associated with other serious adverse effects including renal dysfunction, haemolytic anaemia, and aseptic meningitis [7, 26]. Aseptic meningitis, characterised by severe headache, nuchal rigidity, and cerebrospinal fluid pleocytosis occurring within 24–48 h of infusion, has been reported in 1–11% of IVIg-treated patients and may confound the clinical assessment of stroke-like symptoms in the post-infusion period [7, 26]. These additional adverse effects are beyond the scope of this review but warrant consideration in the overall safety assessment of immunoglobulin therapy.

Among SCIg-treated patients (approximately 820 patients, 2,200 patient-years), only 3 isolated thromboembolic events were documented: 2 DVT events (one in a patient with concurrent Factor V Leiden mutation, one with prior DVT history) and one ischaemic stroke in a patient with long-standing hypertension and atrial fibrillation. All events occurred in patients with established prothrombotic risk factors; none was attributed primarily to SCIg by reporting investigators. The PATH trial reported no thromboembolic events during the 13-week maintenance phase or the 48-week open-label extension [2] (Table 3).

Table 3.

Independent Risk Factors for IVIg-Associated Thromboembolic Events

Risk Factor	Studies Reporting (%)	Relative Risk Increase
Age ≥ 65 years	89%	3–5-fold
Pre-existing CVD	82%	2–4-fold
Obesity (BMI ≥ 30)	71%	2–3-fold
Immobility/bed rest	65%	2–3-fold
Dehydration	58%	1.5–2.5-fold
Thrombophilia	47%	3–6-fold
High-dose IVIg (> 1.5 g/kg/course)	41%	2–4-fold
Rapid infusion rate	35%	1.5–2-fold

Risk factors

The risk appeared additive: patients with 3 or more risk factors demonstrated incidence rates 8–12 times higher than patients with zero risk factors. For SCIg, insufficient event counts precluded formal risk factor analysis, though all 3 documented events occurred in patients with 2 or more risk factors.

Subgroup and sensitivity analyses

Subgroup analysis by age demonstrated that patients aged 65 years or older accounted for 78% of all IVIg-associated thromboembolic events despite representing 47% of total patient-years. Among patients younger than 65 years without additional cardiovascular risk factors, event rates were low (0.2–0.8 per 100 patient-years), approaching background population rates. Sensitivity analyses restricted to studies published after 2010 demonstrated lower IVIg event rates (pooled incidence 1.1 per 100 patient-years, 95% CI: 0.6–1.9) compared with earlier studies (2.4 per 100 patient-years, 95% CI: 1.3–3.8), likely reflecting improved thromboprophylaxis practices. Egger’s test for the 18 IVIg studies reporting DVT events yielded p = 0.038, suggesting possible publication bias toward reporting positive events.

Quality of evidence

Risk of bias was low in 8 RCTs, moderate in 3, and high in 2. NOS scores for observational studies averaged 6.2/9 (range 4–8). GRADE assessment yielded very low certainty for the comparative SCIg versus IVIg thromboembolic outcome (downgraded for indirectness, imprecision, and confounding risk), low certainty for IVIg thromboembolic incidence, and very low certainty for SCIg incidence.

Discussion

Principal findings and interpretation

This systematic review synthesises the available evidence on thromboembolic risk associated with SCIg versus IVIg. Three principal findings emerge. First, IVIg is associated with a clinically meaningful thromboembolic risk concentrated in elderly patients with cardiovascular comorbidities, with pooled incidence of 1.8 per 100 patient-years (95% CI: 0.9–3.1). Second, SCIg-treated cohorts demonstrate very low documented event rates (0.14 per 100 patient-years, 95% CI: 0.01–0.52). Third, no adequately powered head-to-head trial exists, leaving the comparative question formally unanswered.

The apparent difference between routes is consistent with the pharmacokinetic hypothesis that SCIg’s attenuated peak serum IgG concentrations reduce plasma viscosity and the associated haemodynamic conditions conducive to thrombosis [9, 15, 17]. Reinhart and Berchtold demonstrated that IVIg increases plasma viscosity in proportion to achieved serum IgG concentration, with clinically significant elevation above approximately 30 g/L [9]. Standard IVIg dosing achieves post-infusion levels exceeding this threshold, whereas SCIg maintains trough levels in the physiological range (8–16 g/L) [17]. This hypothesis is further supported by meta-analyses of RCTs, which have confirmed that IVIg is associated with an elevated risk of thromboembolic adverse events (Ammann et al., 2016) [27].

However, several critical caveats temper this interpretation. The observation base is markedly unequal: IVIg cohorts contributed approximately 10,600 patient-years versus approximately 2,200 for SCIg—a nearly 5-fold difference. Moreover, median follow-up was longer in IVIg studies (12–36 months) than SCIg studies (6–24 months), and IVIg cohorts included older patients with more cardiovascular comorbidities. These systematic differences in population characteristics, follow-up duration, and event ascertainment methods mean that the crude incidence rate comparison is subject to substantial confounding. The lower event rate with SCIg may partly reflect differences in patient selection rather than a true pharmacological advantage. Whether actual data demonstrate that SCIg is associated with lower activated coagulation factors or platelet activation compared with IVIg remains largely hypothetical; direct comparative mechanistic studies are limited, and the proposed advantage rests primarily on pharmacokinetic inference rather than measured coagulation endpoints.

Clinical implications

Despite the evidence limitations, a risk-stratified approach to route selection appears prudent. For clinically stable patients with CIDP or MG who have established cardiovascular or prothrombotic risk factors, transition to SCIg maintenance may be considered, as recent case reports and systematic reviews have demonstrated the safety and efficacy of SCIg even in patients with high thromboembolic risk profiles [28, 29], supported by efficacy data from the PATH and ADVANCE-CIDP 1 trials [2, 14]. For acute crises (MG exacerbation, GBS, CIDP relapse), IVIg remains preferred for its rapid onset, with appropriate thromboprophylaxis including hydration, controlled infusion rates, and pharmacological anticoagulation in highest-risk patients [7, 11, 12].

Asian population considerations

Asian populations present specific considerations: lower prevalence of Factor V Leiden and prothrombin G20210A mutations but higher prevalence of protein C and protein S deficiencies [1]. Body size differences may result in higher weight-adjusted peak IgG concentrations for equivalent doses. Registry-based surveillance within national health insurance databases (e.g., Taiwan NHI, Korea HIRA, Japan JMDC) represents a valuable opportunity to generate real-world comparative safety data currently absent from the literature.

Strengths and limitations

Strengths include the comprehensive search spanning six databases, inclusion of all indications, pre-specified outcomes and subgroups, and explicit focus on the route-of-administration comparison. Limitations are substantial. The central limitation is the absence of prospective head-to-head comparative trial data. The markedly unequal observation base—with IVIg cohorts contributing approximately 5 times more patient-years than SCIg cohorts—means that the statistical power to detect events in SCIg populations is inherently limited. SCIg cohorts were systematically younger and healthier, making direct incidence comparisons susceptible to confounding by indication and selection bias. Thromboembolic ascertainment varied from prospective surveillance to passive detection. Follow-up for SCIg studies was generally shorter. Publication bias, suggested by significant Egger’s test, may have inflated apparent IVIg risk. The GRADE certainty for the comparative estimate was very low.

Conclusion

Indirect evidence from this systematic review suggests that SCIg may be associated with a lower thromboembolic risk than IVIg, a finding consistent with pharmacokinetic predictions regarding attenuated peak serum IgG concentrations and reduced plasma viscosity. However, the certainty of this evidence is very low, and the apparent difference may be partly attributable to systematic differences in patient populations, follow-up duration, and event ascertainment between SCIg and IVIg cohorts rather than a true pharmacological advantage. The IVIg thromboembolic risk is concentrated in elderly patients with cardiovascular comorbidities and appears to have decreased over time with improved clinical practices.

Until prospective comparative trial data become available, a risk-stratified approach to route selection—considering SCIg for patients with established cardiovascular risk factors where clinically appropriate—represents a reasonable clinical strategy, while acknowledging that definitive evidence of SCIg’s thromboembolic superiority is lacking. The urgent research priority is a well-powered, prospective randomised trial with pre-specified thromboembolic endpoints.

Author contributions

S.F. and S.C. conceptualized the study. S.F. conducted the literature search, data extraction, and wrote the main manuscript text. Y.C.C. contributed to data interpretation and manuscript revision. S.C. supervised the study, provided critical revisions, and prepared the tables and figures. All authors reviewed and approved the final manuscript.

Funding

This research received no specific grant from any funding agency.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Ethics Approval

Not applicable. This study is a systematic review of published data.