Cardiac Magnetic Resonance Imaging Findings in Patients With Antineutrophil Cytoplasmic Antibody–Associated Vasculitides: A Systematic Review

Abstract

Objective

Our objective was to review the available literature on cardiac magnetic resonance imaging (cMRI) findings in patients with antineutrophil cytoplasmic antibody–associated vasculitides (AAV), evaluate its diagnostic utility, and assess its potential as a screening tool.

Methods

We systematically searched PubMed, Embase, Scopus, and Web of Science from inception to March 29, 2023, following Preferred Reporting Items for Systematic reviews and Meta‐Analyses (PRISMA) 2020 guidelines. English‐language studies involving adult patients diagnosed with AAV—eosinophilic granulomatosis with polyangiitis (EGPA), granulomatosis with polyangiitis (GPA), or microscopic polyangiitis (MPA)—using recognized classification criteria were included. Studies had to report specific cMRI parameters in at least three patients. Three independent reviewers conducted study selection, data extraction, and quality assessment.

Results

Of 2,251 studies, 30 met the inclusion criteria, encompassing 1,149 patients with AAV (87% with EGPA, 13% with GPA, and 0.3% with MPA). The mean patient age was 52 ± 5 years, with 50.4% being female. The mean left ventricular ejection fraction (LVEF) was 55.6% ± 11.3%, and 29% of patients had an LVEF less than 50%. Myocardial fibrosis, indicated by late gadolinium enhancement (LGE), was present in 49% of patients, with predominantly subendocardial or endocardial (23%), intramyocardial (14%), and subepicardial (10%) patterns. Patients in remission (26%), when compared to those not in remission (74%), exhibited higher proportions of LGE (55% vs 47%) and glucocorticoid use (77% vs 68%), despite similar rates of abnormal electrocardiograms (44% vs 42%).

Conclusion

This systematic review reveals a high prevalence of myocardial fibrosis detected by cMRI in patients with AAV, even during remission. Significant subclinical cardiac involvement may be missed by conventional diagnostic methods, underscoring the utility of cMRI during routine evaluation.

INTRODUCTION

Antineutrophil cytoplasmic antibody (ANCA)–associated vasculitides (AAV) represent a group of systemic autoimmune diseases characterized by inflammation of small‐ to medium‐sized blood vessels. The primary subtypes of AAV include granulomatosis with polyangiitis (GPA), microscopic polyangiitis (MPA), and eosinophilic granulomatosis with polyangiitis (EGPA). ¹ Recent advancements in classification criteria have improved the ability to accurately classify subtypes of AAV, guiding more tailored clinical and research approaches. For instance, the 2022 American College of Rheumatology (ACR)/EULAR classification criteria for EGPA, GPA, and MPA have demonstrated improved sensitivity and specificity compared to the outdated 1990 ACR criteria. ² , ³ , ⁴ Additionally, the 2012 Revised International Chapel Hill Consensus Conference provided a more refined nomenclature and classification system, distinguishing among large, medium, and small vessel vasculitis, thereby facilitating more precise clinical and research applications. ¹

SIGNIFICANCE & INNOVATIONS.

• This study is the first to systematically review cardiac magnetic resonance imaging (cMRI) findings in patients with antineutrophil cytoplasmic antibody–associated vasculitides (AAV).

• cMRI revealed myocardial fibrosis in 49% of patients, even those in remission, highlighting subclinical cardiac involvement.

• Incorporating cMRI into routine evaluations could improve the detection and management of cardiac involvement in patients with AAV.

Although EGPA, GPA, and MPA share overlapping immune‐mediated processes characterized by ANCA‐driven vascular inflammation, ⁵ EGPA differs by featuring prominent eosinophilic involvement. The management of AAV has been guided by comprehensive, evidence‐based recommendations, such as those developed by the ACR and the Vasculitis Foundation in 2021. These guidelines emphasize the use of rituximab over cyclophosphamide for remission induction in severe GPA and MPA due to its comparable efficacy and lower toxicity and recommend mepolizumab for remission induction in nonsevere EGPA cases. ⁶ Remission maintenance strategies also prioritize rituximab for severe GPA and MPA, while suggesting alternative immunosuppressives for EGPA, underscoring the importance of tailored therapeutic approaches based on disease severity and patient‐specific factors. ⁶

Cardiac involvement in AAV, although less common than renal or pulmonary involvement, poses significant diagnostic challenges. Conventional diagnostic tools such as electrocardiography (EKG) and echocardiography (TTE) may often fail to detect subtle myocardial abnormalities, especially in patients presenting with preserved left ventricular ejection fraction (LVEF). ⁷ , ⁸ This silent cardiac involvement can lead to underdiagnosis and delayed treatment, potentially resulting in adverse cardiovascular outcomes. ⁹ Cardiac magnetic resonance imaging (cMRI) emerges as a superior modality for identifying myocardial fibrosis, inflammation, and other subtle cardiac changes that are not readily apparent with traditional imaging techniques. ⁹ cMRI's ability to detect late gadolinium enhancement (LGE) allows for the visualization of myocardial scarring and fibrosis, which are critical markers of subclinical cardiac involvement. ¹⁰ The prognostic significance of LGE on cMRI has been well‐documented in other nonischemic cardiomyopathies, in which its presence correlates with increased risks of adverse cardiovascular events, including all‐cause death, heart failure hospitalization, and sudden cardiac death. ¹¹  Despite these advancements, the specific role and utility of cMRI in the context of AAV remain underexplored. Given the potential for cMRI to uncover subclinical cardiac involvement in patients with AAV, a systematic evaluation of existing literature is essential to elucidate its diagnostic and prognostic value in this patient population. This systematic review aims to synthesize current evidence on cMRI findings in patients with AAV, evaluate the diagnostic utility of cMRI, and assess its potential as a routine screening tool to detect silent cardiac involvement that may be missed by conventional diagnostic methods.

MATERIALS AND METHODS

Search strategy and information sources

This systematic review was conducted following the Preferred Reporting Items for Systematic reviews and Meta‐Analyses (PRISMA) 2020 guidelines ¹² to ensure a comprehensive and transparent synthesis of the available evidence on cMRI findings in patients diagnosed with AAV, including EGPA, GPA, and MPA. A systematic literature search was performed across four major databases: PubMed, Embase, Scopus, and Web of Science. The search strategy for each database was predetermined and developed with the assistance of a medical librarian to ensure a comprehensive retrieval of relevant studies. Search terms were tailored to capture studies focusing on cMRI findings in adult patients diagnosed with AAV. The final search was executed on March 29, 2023. Detailed search strategies for each database are available in the Supplementary Material.

Eligibility criteria and study selection process

Studies were included if they involved adult (>18 years old) patients diagnosed with AAV (EGPA, GPA, or MPA) according to recognized classification criteria published by the ACR, EULAR, Chapel Hill Consensus Conference, and Lanham criteria. Different versions of these criteria were deemed acceptable depending on the study publication date. Some studies did not specify the exact criteria used but were included if they involved patients diagnosed with AAV based on recognized standards. The intervention (or exposure) involved the reporting of specific cMRI parameters. Outcomes of interest included detailed cMRI findings such as LVEF, LGE, myocardial swelling, and pericardial effusion. All study designs were considered, including cohort studies, cross‐sectional studies, and case series, provided they were published in English and reported cMRI parameters (at least LGE) in three or more patients. Exclusion criteria encompassed studies that reported incorrect outcomes, provided insufficient data, used inappropriate study designs, were non‐English publications, or involved unsuitable patient populations. Two independent reviewers (IK and UB) initially screened the titles and abstracts of all identified studies for relevance and eligibility. During this screening phase, studies were evaluated based on the predetermined inclusion and exclusion criteria. Any disagreements between the primary reviewers were resolved by consulting a third reviewer. Following the initial screening, the full texts of potentially eligible studies were retrieved and assessed by two independent reviewers (IK and UB) to determine their suitability for inclusion in the review. The final decision to include or exclude studies was based on a thorough evaluation of each full‐text article against the established criteria. Any disagreements between the primary reviewers were resolved by discussion between the two reviewers.

Data extraction and risk of bias assessment

Three independent reviewers (IK, HA, and BC) conducted data extraction and bias assessments using predetermined forms to minimize subjective bias and ensure consistency. IK assessed all studies, while HA and BC each assessed half of the studies. Discrepancies between reviewers were resolved through discussion until a consensus was reached. The extraction encompassed demographic information (age, sex, and other relevant details), relevant medical history, laboratory values, medications administered, measures of vasculitis disease severity, specific cMRI findings (such as LVEF, LGE, myocardial edema, and pericardial effusion), patient outcomes, and remission data based on Birmingham Vasculitis Activity Score (BVAS) scoring or explicit statements by study authors.

The methodologic quality and potential risk of bias of the included studies were systematically evaluated using appropriate critical appraisal tools tailored to each study design, following recommendations from the Joanna Briggs Institute (JBI) and the Newcastle‐Ottawa Scale (NOS). ¹³ , ¹⁴ , ¹⁵ For cohort studies with control groups, the NOS was used, whereas cohort studies without control groups, cross‐sectional studies, and case series were appraised using the respective JBI critical appraisal checklists specific to each design. Scoring and categorization were performed accordingly. For the NOS, a maximum of 9 points was assigned; studies scoring 7 to 9 points were categorized as low risk of bias, 4 to 6 points as moderate risk, and 0 to 3 points as high risk. For the JBI checklists, items were scored as “yes,” “no,” or “unclear,” with studies achieving 75% or more of the total possible points classified as low risk of bias, 50% to 74% as moderate risk, and less than 50% as high risk. Bias assessments were performed only on full articles (21 of 30 included studies), as conference abstracts provided limited information, precluding adequate evaluation of methodologic quality. Detailed bias assessment tables for each selected study are included in the Supplementary Tables. Any modifications to the appraisal process during the review were documented and justified following best practices for systematic reviews. ¹²

Data synthesis and additional analyses

Because of the heterogeneity in study designs, populations, and reported outcomes, a meta‐analysis was not feasible. Instead, data were synthesized through comprehensive summary tables that organized key demographic, clinical, treatment, and cMRI findings. We performed a subgroup analysis between patients in remission—defined as a BVAS of less than or equal to 1, or by explicit statement of remission by the authors—and patients not in remission. The review did not include subgroup analyses based on different ANCA subtypes or perform sensitivity analyses.

Registration and deviations from the protocol

The review protocol was registered on PROSPERO (CRD42023464957) to ensure transparency and adherence to predefined methodologic standards. ¹² Although our systematic review closely followed the methodology outlined in our registered PROSPERO protocol, certain deviations were necessary due to practical considerations encountered during the study. We were unable to perform subgroup analyses based on vasculitis subtypes, disease stage, specific cMRI parameters, and demographics as initially planned, owing to insufficient or inconsistent data reporting across the included studies. For the risk of bias assessment, we adapted our approach by using the JBI critical appraisal checklists alongside the NOS to appropriately evaluate the diverse study designs included (cohort studies without control groups, cross‐sectional studies, and case series). Although we intended to extract specific cMRI parameters such as left ventricular end diastolic volume, left ventricular end systolic volume, left ventricular mass, right ventricular ejection fraction, right ventricular end diastolic volume, right ventricular end systolic volume, native T1, extracellular volume, T2 ratio, and early gadolinium enhancement, these were not consistently reported across studies; therefore, our synthesis focused on the most commonly reported cMRI findings such as LVEF, LGE, myocardial edema, and pericardial effusion. Lastly, we did not assess publication bias as planned because of the heterogeneity of the studies and lack of sufficient comparable data.

Availability of data and materials

All data generated or analyzed during this study are included in this published article and the Supplementary Materials, including the data extraction forms, synthesized data tables, and bias assessment forms and tables. Further information may be provided from the corresponding author upon reasonable request.

RESULTS

Study selection

From the initial 3,589 studies, after removing duplicates (1,318 through automatic deduplication in EndNote and an additional 20 through manual deduplication), 2,251 unique studies remained. During the abstract screening process, 2,181 studies were excluded based on relevance and the predetermined eligibility criteria. The full texts of the remaining 70 studies were then assessed for eligibility. Exclusions at this stage were due to incorrect outcomes, ¹⁶ , ¹⁷ insufficient data, ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ , ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ , ³¹ , ³² , ³³ , ³⁴ inappropriate study design, ³⁵ , ³⁶ , ³⁷ , ³⁸ , ³⁹ , ⁴⁰ , ⁴¹ , ⁴² , ⁴³ , ⁴⁴ , ⁴⁵ , ⁴⁶ , ⁴⁷ non‐English language, ⁴⁸ , ⁴⁹ , ⁵⁰ , ⁵¹ , ⁵² , ⁵³ and unsuitable patient population. ⁵⁴ , ⁵⁵ Ultimately, 30 studies met all inclusion criteria and were included in the review, ⁷ , ⁸ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ , ⁷² , ⁷³ , ⁷⁴ , ⁷⁵ , ⁷⁶ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸¹ , ⁸² , ⁸³ as shown in the PRISMA flow diagram in Figure 1.

Figure 1.

PRISMA flow diagram illustrating the study selection process.

Study characteristics

A total of 30 studies published between 2005 and 2022 were included in the systematic review, encompassing 1,149 patients diagnosed with AAV, as shown in Table 1. The majority of the studies were cohort designs (n = 21), followed by cross‐sectional studies (n = 7) and case series (n = 2). Sample sizes varied across the studies, ranging from 7 to 176 patients. Most studies focused on evaluating cardiac involvement in patients with AAV using cMRI. The primary objectives included detecting myocardial damage, assessing the incidence and prevalence of cardiac involvement, evaluating the diagnostic utility of cMRI, and exploring its potential in early detection and risk stratification. Common findings across the studies indicated a high prevalence of myocardial fibrosis and cardiac abnormalities detected by cMRI in patients with AAV. Myocardial fibrosis was frequently observed even in patients without overt cardiac symptoms or with preserved LVEF.

Table 1.

Summary of included studies^*

Study ID	Ref no.	Year	Study design	AAV, n	Primary outcome
Goebel et al (2005) a	3	2005	Case series	10	cMRI identifies reversible and irreversible myocardial damage in EGPA
Cereda et al (2017)	357	2017	Cohort	11	Cardiac disease detection in EGPA
Dennert et al (2010)	506	2010	Cross‐sectional	32	Incidence of cardiac involvement in EGPA
Dunogué et al (2015)	548	2015	Cohort	42	Prognosis prediction via cMRI in EGPA
Fijolek et al (2016)	624	2016	Cohort	33	cMRI detects cardiac involvement, monitors treatment efficacy in EGPA
Garcia‐Vives et al (2021)	677	2021	Cohort	131	Universal cardiac screening improves early detection in EGPA
Giollo et al (2021)	707	2021	Cross‐sectional	26	cMRI identifies myocardial fibrosis, aiding risk stratification in GPA
Greulich et al (2016)	740	2016	Cross‐sectional	63	Prevalence and patterns of cardiac involvement in rheumatic disorders
Greulich et al (2017) a	742	2017	Cohort	37	Evaluation of myocardial involvement in patients with AAV using T1 and T2 mapping techniques
Hansch et al (2009)	785	2009	Cross‐sectional	7	Myocardial FPP abnormalities in patients with EGPA with cardiac involvement
Hazebroek et al (2015)	814	2015	Cohort	91	Prevalence and prognostic significance of cardiac involvement for death in patients with EGPA and GPA
Hinojar et al (2014) a	837	2014	Cohort	7	Elevated T1 and T2 values, indicating diffuse myocardial injury in systemic inflammatory disease
Hua et al (2022) a	866	2022	Cohort	63	Cardiac involvement was detected in 37% of patients with EGPA by cMRI
Lagan et al (2021)	1,128	2021	Cohort	13	Myocardial fibrosis was observed in patients with stable EGPA without pulmonary fibrosis or inflammation
Marmursztejn et al (2010)	1,258	2010	Case series	8	cMRI detected cardiac lesions in patients with EGPA, and immunosuppressive therapy reduced abnormalities in most cases
Marmursztejn et al (2013)	1,260	2013	Cohort	20	cMRI detected subclinical myocardial lesions in patients with EGPA, with FDG‐PET distinguishing between fibrosis and inflammation
Marmursztejn et al (2009)	1,261	2009	Cross‐sectional	20	cMRI with delayed enhancement detected myocardial involvement in patients with EGPA, regardless of symptoms
Marmursztejn et al (2006) a	1,263	2006	Cohort	20	cMRI identified myocardial inflammation, perfusion defects, and fibrosis in patients with EGPA, highlighting subclinical cardiac involvement
Mavrogeni et al (2013)	1,297	2013	Cohort	28	cMRI detected higher cardiac fibrosis in ANCA‐negative EGPA
Miszalski‐Jamka et al (2013) a	1,363	2013	Cohort	50	Cardiac involvement detected in 82% of patients with EGPA, often with subendocardial fibrosis
Miszalski‐Jamka et al (2013)	1,366	2013	Cohort	21	Subclinical myocardial involvement on cMRI is prevalent in patients with EGPA and GPA, despite normal EKG and TTE
Miszalski‐Jamka et al (2014) a	1,367	2014	Cohort	51	Nonsteroid immunosuppression reduces myocardial damage on cMRI and dysfunction in patients with EGPA
Miszalski‐Jamka et al (2011) a	1,369	2011	Cohort	11	cMRI showed involvement in 82% of patients with GPA, often with fibrosis and ongoing inflammation
Neumann et al (2009)	1,471	2009	Cohort	49	Endomyocarditis in EGPA correlates with ANCA negativity, worsening cardiac outcomes
Pugnet et al (2017)	1,637	2017	Cross‐sectional	31	cMRI revealed abnormalities in 61% of patients with GPA, primarily LGE
Sartorelli et al (2022)	1,769	2022	Cohort	176	cMRI revealed abnormalities in 40% of patients with EGPA, primarily LGE
Szczeklik et al (2011)	1,976	2011	Cross‐sectional	20	Cardiac involvement detected in 90% of patients with EGPA, often with fibrosis
Szczeklik et al (2015) a	1,977	2015	Cohort	51	Nonsteroid immunosuppressive therapy may limit cardiac damage in EGPA
Wassmuth et al (2008)	2,143	2008	Case series	11	cMRI effectively detects myocardial injury in EGPA, even with preserved LVEF
Yune et al (2016)	2,230	2016	Cohort	16	cMRI can detect myocardial GLE in patients with active EGPA, even in the absence of cardiac symptoms

^*AAV, antineutrophil cytoplasmic antibody–associated vasculitides; ANCA, antineutrophil cytoplasmic antibody; cMRI, cardiac magnetic resonance imaging; EGPA, eosinophilic granulomatosis with polyangiitis; EKG, electrocardiography; FDG, fluorodeoxyglucose; FPP, first pass perfusion; GPA, granulomatosis with polyangiitis; ID, identifier; LGE, late gadolinium enhancement; LVEF, left ventricular ejection fraction; PET, positron emission tomography; Ref, reference; TTE, echocardiography.

^a These studies have been excluded from bias assessment due to inadequate information (conference abstracts).

Risk of bias in studies

Of the 30 included studies, 21 full‐text studies subjected to bias assessment, ⁸ , ⁵⁶ , ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁷ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷² , ⁷⁵ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸² , ⁸³ as seen in Table 2. Nine conference abstracts provided limited information, precluding adequate evaluation of methodologic quality. ⁷ , ⁶¹ , ⁶⁵ , ⁶⁶ , ⁷¹ , ⁷³ , ⁷⁴ , ⁷⁶ , ⁸¹ None of the evaluated studies were categorized as having a high risk of bias. Three studies (14%) were rated as low risk of bias, reflecting high methodologic quality with adequate selection processes, comparability, and outcome assessments. ⁸ , ⁵⁶ , ⁶² The remaining 18 studies (86%) were deemed to have a moderate risk of bias. ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶³ , ⁶⁴ , ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷² , ⁷⁵ , ⁷⁷ , ⁷⁸ , ⁷⁹ , ⁸⁰ , ⁸² , ⁸³ Common methodologic limitations identified across studies included inadequate consideration of confounding factors, incomplete follow‐up reporting, and insufficient strategies to address confounding variables. Several studies did not explicitly identify or adjust for potential confounders, which could affect the internal validity of their findings. ⁵⁷ , ⁵⁸ , ⁵⁹ , ⁶⁰ , ⁶³ , ⁶⁴ , ⁶⁸ , ⁷⁰ , ⁷⁷ , ⁷⁹ , ⁸³ Incomplete or unclear reporting of follow‐up and loss to follow‐up was observed in some studies. ⁶⁷ , ⁷² , ⁷⁵ , ⁷⁷ , ⁷⁹ , ⁸³ Additionally, some studies lacked clarity in participant inclusion criteria and did not fully describe the study settings, which may affect the generalizability of the results. ⁶⁴ , ⁶⁹ , ⁸² Despite these limitations, most studies demonstrated strengths in other domains. Exposure and outcome measurements were conducted in valid and reliable ways, with standardized criteria and appropriate use of cMRI to detect cardiac involvement, in all studies but one. ⁸² All but two studies employed appropriate statistical analyses consistent with their study designs and objectives. ⁵⁹ , ⁸³ Detailed bias assessment for each study can be accessed in the Supplementary Tables.

Table 2.

Bias assessment overview^*

Study ID	Ref no.	Design	Assessment tool	Total score	Risk of bias	Comments
Cereda et al (2017)	357	Cohort (control group)	NOS (9 points)	7/9	Low	Good selection and comparability
Dennert et al (2010)	506	Cross‐sectional	JBI (8 items)	5/8	Moderate	Confounding factors not identified
Dunogué et al (2015)	548	Cohort (no control group)	JBI (11 items)	9/11	Moderate	Confounding factors not addressed
Fijolek et al (2016)	624	Cohort (no control group)	JBI (11 items)	7/11	Moderate	Inadequate statistical analysis
Garcia‐Vives et al (2021)	677	Cohort (no control group)	JBI (11 items)	9/11	Moderate	Incomplete follow‐up
Giollo et al (2021)	707	Cross‐sectional	JBI (8 items)	8/8	Low	High methodologic quality
Greulich et al (2016)	740	Cross‐sectional	JBI (8 items)	6/8	Moderate	Confounding factors not addressed
Hansch et al (2009)	785	Cross‐sectional	JBI (8 items)	5/8	Moderate	Study setting not described
Hazebroek et al (2015)	814	Cohort (control group)	NOS (9 points)	8/9	Low	Comprehensive follow‐up
Lagan et al (2021)	1,128	Cohort (control group)	NOS (9 points)	6/9	Moderate	Incomplete follow‐up reporting
Marmursztejn et al (2009)	1,261	Cross‐sectional	JBI (8 items)	6/8	Moderate	Confounding factors not addressed
Marmursztejn et al (2010)	1,258	Case series	JBI (10 items)	8/10	Moderate	Incomplete participant inclusion
Marmursztejn et al (2013)	1,260	Cohort (no control group)	JBI (11 items)	9/11	Moderate	Confounding factors not addressed
Mavrogeni et al (2013)	1,297	Cohort (control group)	NOS (9 points)	5/9	Moderate	Loss to follow‐up not described
Miszalski‐Jamka et al (2013)	1,366	Cohort (control group)	NOS (9 points)	5/9	Moderate	Nonexposed cohort derivation not described
Neumann et al (2009)	1,471	Cohort (no control group)	JBI (11 items)	7/11	Moderate	Participants not free of outcome at baseline
Pugnet et al (2017)	1,637	Cross‐sectional	JBI (8 items)	6/8	Moderate	Confounding factors not addressed
Sartorelli et al (2022)	1,769	Cohort (no control group)	JBI (11 items)	8/11	Moderate	Confounding factors partially addressed
Szczeklik et al (2011)	1,976	Cross‐sectional	JBI (8 items)	6/8	Moderate	Confounding factors not addressed
Wassmuth et al (2008)	2,143	Case series	JBI (10 items)	6/10	Moderate	Inclusion criteria not clearly defined
Yune et al (2016)	2,230	Cohort (no control group)	JBI (11 items)	7/11	Moderate	Follow‐up completeness unclear

^*Nine of 30 studies without full articles (conference abstracts) have been excluded from bias assessment due to inadequate information. ID, identifier; JBI, Joanna Briggs Institute; NOS, Newcastle‐Ottawa Scale; Ref, reference.

Results of individual studies and results of syntheses

A total of 30 studies encompassing 1,149 patients with AAV were included in this systematic review. Among these patients, 87% of patients (n = 963) were diagnosed with EGPA, 13% of patients (n = 146) were diagnosed with GPA, and three patients were diagnosed with MPA, as shown in Table 3. The mean ± SD age was 52 ± 5 years, and the sex distribution was nearly equal, with 50.4% of patients (n = 532) being female. The comprehensive data extraction table detailing the results of each study was not included in the main text due to space limitations. However, this detailed table can be accessed in the Supplementary Material provided.

Table 3.

Demographic, clinical, and imaging characteristics by remission status. The asterisk applies to all values except Age.

Patient characteristic	All patients	Remission	No remission
Total patients with AAV, n	1,149	296	853
Age, mean (SD), y	52 (5)	51.2 (14.2)	48.4 (11.5)
Sex, n (%)
Female	532 (50.4)	151 (51)	381 (50)
Male	522 (49.5)	145 (49)	377 (50)
Disease type, n (%)
EGPA	963 (87)	245 (83)	718 (84)
GPA	146 (13)	51 (17)	95 (11)
ANCA status, n (%)
ANCA+	308 (37)	90 (49)	218 (34)
ANCA−	512 (61)	94 (51)	418 (66)
Clinical parameters
Hypertension, n (%)	195 (25)	78 (40)	117 (20)
Hyperlipidemia, n (%)	104 (18)	44 (27)	60 (15)
Diabetes, n (%)	60 (11)	21 (13)	39 (10)
Abnormal EKG, n (%)	336 (43)	68 (44)	268 (42)
LVEF, mean % (SD)	55.6 (11.3)	54.7 (14.9)	56.1 (8.3)
Glucocorticoids, n (%)	558 (70)	180 (77)	378 (68)
Other immunosuppresion, n (%)	484 (58)	118 (50)	366 (61)
Cardiac MRI parameters
Cardiac MRIs, n	1,006	270	736
Late GE, n (%)	496 (49)	149 (55)	347 (47)
Subendocardial or endocardial, n (%)	178 (23)	54 (25)	124 (22)
Intramyocardial, n (%)	100 (14)	39 (24)	61 (11)
Subepicardial, n (%)	55 (10)	20 (28)	35 (8)
No. of segments, mean (SD)	7.2 (2.5)	4.2 (7.6)	6.2 (4.8)
Early GE, n (%)	63 (20)	1 (3)	62 (22)
Pericardial effusion, n (%)	132 (18)	15 (9)	117 (22)
Myocardial edema, n (%)	65 (19)	4 (4)	61 (26)
Intraventricular thrombus, n (%)	13 (5)	3 (27)	10 (4)

^*Sex: denominators for female and male are based on the total number of patients. Disease type: denominators for EGPA and GPA are based on the total number of patients. ANCA status: denominators for ANCA+ and ANCA− are based on the total number of ANCA‐tested patients. Clinical parameters: denominators for all parameters are based on the total number of patients with available data. Cardiac MRI parameters: denominators for all parameters are based on the total number of patients with available MRI data. AAV, antineutrophil cytoplasmic antibody–associated vasculitides; ANCA, antineutrophil cytoplasmic antibody; EGPA, eosinophilic granulomatosis with polyangiitis; EKG, electrocardiography; GE, gadolinium enhancement; GPA, granulomatosis with polyangiitis; LVEF, left ventricular ejection fraction; MRI, magnetic resonance imaging.

Clinical characteristics and cMRI findings

The mean ± SD LVEF across all patients was 55.6% ± 11.3%, with 29% of patients (n = 133) exhibiting an LVEF below 50%, indicating varying degrees of ventricular function impairment. Hypertension was present in 25% of patients (n = 195), hyperlipidemia was present in 18% of patients (n = 104), and diabetes mellitus was present in 11% of patients (n = 60). Abnormal EKG findings were reported in 43% of patients (n = 336).

cMRI was performed on 1,006 patients. LGE was present in 49% of the patients (n = 496). The patterns of LGE involvement were predominantly subendocardial or endocardial (23%, n = 178), followed by intramyocardial (14%, n = 100) and subepicardial regions (10%, n = 55). The mean ± SD number of myocardial segments exhibiting LGE was 7.2 ± 2.5 based on the 17‐segment model, ⁸⁴ suggesting extensive myocardial involvement in some patients. Early gadolinium enhancement, indicative of active inflammation, was observed in 20% of patients (n = 63). Pericardial effusion was detected in 18% of patients (n = 132), myocardial edema was detected in 19% of patients (n = 65), and intraventricular thrombus was detected in 5% of patients (n = 13). As shown in Table 4, most of the available cMRI data pertain to EGPA (870 scans in 963 patients), whereas fewer MRI scans were reported for GPA (133 scans in 146 patients) and MPA (3 scans in 3 patients). LGE was seen in approximately half of patients with EGPA (50.9%, n = 443) and 39.0% of patients with GPA (n = 52), with a notably higher proportion of intramyocardial or subepicardial involvement in the latter. Because of the small number of patients with MPA (n = 3), no meaningful cMRI subgroup analysis was feasible for that subtype.

Table 4.

Comparison of key cMRI findings in EGPA, GPA, and MPA subgroups 1. The asterisk applies to all values.

Parameter	EGPA (n = 963)	GPA (n = 146)	MPA (n = 3)
cMRI, n	870	133	3
Late GE, n (%)	443 (50.9)	52 (39.0)	–
Subendocardial or endocardial	171 (22.6)	7 (30.4)	–
Intramyocardial	88 (13.2)	12 (38.7)	–
Subepicardial	41 (10.0)	14 (45.1)	–
Early GE, n (%)	52 (21.3)	11 (15.7)	–
Pericardial effusion, n (%)	121 (18.7)	11 (15.9)	–

^*Denominators for all parameters are based on the total number of patients with available MRI data. Cells containing en dashes were data that could not be extracted. cMRI, cardiac magnetic resonance imaging; EGPA, eosinophilic granulomatosis with polyangiitis; GE, gadolinium enhancement; GPA, granulomatosis with polyangiitis; MPA, microscopic polyangiitis; MRI, magnetic resonance imaging.

Several studies specifically examined the discrepancy between normal EKG or TTE findings and abnormal cMRI results in patients with AAV. Overall, eight studies ⁸ , ⁵⁸ , ⁵⁹ , ⁶⁴ , ⁶⁹ , ⁷⁰ , ⁷⁵ , ⁷⁸ reported that cMRI could detect subclinical myocardial involvement even when EKG and TTE appeared normal. Fijolek et al ⁵⁹ observed that 39.4% of patients had a normal EKG and 36% of patients had a normal TTE, yet all patients (100%) had abnormal cMRI findings. Marmursztejn et al ⁷⁰ (2013) found that 10 of 14 patients with abnormal cMRI still had a normal EKG, and 7 of 14 patients had a normal TTE. Similarly, Miszalski‐Jamka et al ⁷⁵ reported that 81% of patients with otherwise normal EKG or TTE exhibited LGE on cMRI. Some studies also provided paired follow‐up cMRI data, suggesting that cMRI abnormalities may regress partially or completely following immunosuppressive therapy, although only a small number of patients underwent repeat imaging. ⁵⁸ , ⁶⁹ , ⁷⁸

Comparison by remission status

When comparing patients in remission (26%, n = 296) to those not in remission (74%, n = 853), several notable differences emerged. A higher proportion of patients in remission exhibited LGE on cMRI compared to those not in remission (55% vs 47%, respectively). The rates of abnormal EKGs were similar between the two groups (44% in remission vs 42% not in remission). Patients in remission had a lower incidence of pericardial effusion (9% vs 22%) and myocardial edema (4% vs 26%) compared to those not in remission. The mean ± SD LVEF was slightly lower in patients in remission (54.7% ± 14.9%) compared to those not in remission (56.1% ± 8.3%). Regarding treatment, a higher percentage of patients in remission were receiving glucocorticoids (77%, n = 180) compared to those not in remission (68%, n = 378). Conversely, other immunosuppressive therapies were more commonly started in patients not in remission (61% vs 50%). Among the patients, 37% (n = 308) were ANCA‐positive, with a higher proportion of ANCA positivity observed in patients in remission (49% vs 34%). Hypertension and hyperlipidemia were more prevalent in patients in remission compared to those not in remission (hypertension: 40% vs 20%; hyperlipidemia: 27% vs 15%).

Synthesis of findings, reporting biases, certainty of evidence

The included studies consistently demonstrate a significant burden of subclinical cardiac involvement in AAV, particularly EGPA, as indicated by cMRI‐detected myocardial fibrosis even in patients who appear to be in clinical remission. This highlights the limitations of conventional diagnostic tools such as EKG and TTE, as well as the potential for irreversible cardiac damage if fibrosis persists. Although publication bias was not formally evaluated—largely due to heterogeneity in study populations, designs, and reported outcomes—most studies were deemed to have a moderate risk of bias, often stemming from inadequate control of confounders and variable cMRI protocols. Furthermore, the lack of randomized trials, reliance on observational data, and significant variability in patient characteristics limit the overall strength of the conclusions. Nevertheless, the repeated finding of LGE across multiple studies supports a reliable association between AAV and myocardial involvement, yielding a moderate level of certainty when weighing the consistency of results against methodologic constraints. Notably, we did not formally apply Grading of Recommendations Assessment, Development and Evaluation because of the narrative synthesis required by the heterogeneity of the included studies.

DISCUSSION

Our findings indicate a high prevalence of subclinical cardiac involvement in AAV patients, as evidenced by cMRI. Nearly half of the patients (49%) exhibited LGE, even among those in clinical remission. This underscores the superior sensitivity of cMRI in uncovering cardiac abnormalities that conventional diagnostic methods, such as EKG and TTE, often fail to detect. Our results align with previous studies focusing on patients with EGPA. Cereda et al ⁵⁶ reported that patients with EGPA in clinical remission exhibited significant myocardial fibrosis and reduced LVEF on cMRI, despite normal electrocardiogram findings. Similarly, Dunogué et al ⁵⁸ found that 59.5% of patients with EGPA had myocardial anomalies on cMRI, with LGE being particularly prevalent in those with cardiomyopathy. Fijolek et al ⁵⁹ demonstrated that all patients with EGPA in their cohort exhibited myocardial injury on cMRI, which was not detected by TTE or EKG.

Studies focusing on patients with GPA revealed similar patterns. Giollo et al ⁶² found that 32% of patients with GPA without known cardiovascular disease exhibited LGE indicative of myocardial fibrosis, which was absent in healthy controls. Greulich et al ⁷ , ⁶³ reported that 43% of patients with AAV exhibited LGE and elevated native T1 and T2 mapping values were common, independent of LGE presence. Hazebroek et al ⁸  found that 61% of patients with GPA had cardiac abnormalities detectable by cMRI, even without evident cardiac symptoms. The pattern of LGE predominantly involved subendocardial or endocardial regions, consistent with findings from Goebel et al ⁶¹ and Marmursztejn et al, ⁶⁸ , ⁶⁹ , ⁷⁰ , ⁷¹ indicating myocardial damage due to granulomatous inflammation and vasculitis.

Our review also highlights the significant impact of immunosuppressive therapy on cardiac involvement in patients with AAV. Fijolek et al ⁵⁹ observed that after treatment with glucocorticoids and cyclophosphamide, 81% of patients with EGPA showed improvements on cMRI, with some achieving complete remission, whereas Marmursztejn et al ⁶⁹ reported significant regression or normalization of cMRI abnormalities in patients receiving immunosuppressive therapy, demonstrating that early and aggressive immunosuppressive treatment can reduce myocardial inflammation, limit the progression of fibrosis, and improve cardiac function. Similarly, our review found that patients in remission had a lower incidence of active myocardial inflammation markers, such as myocardial edema and pericardial effusion, compared to those not in remission, suggesting that immunosuppressive therapy effectively reduces ongoing myocardial inflammation. However, the persistence of LGE in a significant proportion of patients in remission (55%) indicates that myocardial fibrosis may continue or remain despite clinical improvement, highlighting the importance of early and aggressive treatment to prevent irreversible cardiac damage. The higher usage of glucocorticoids among patients in remission in our review (77% vs 68%) further supports the role of immunosuppressive therapy in achieving remission and mitigating cardiac involvement.

Furthermore, cardiac involvement detected by cMRI is associated with adverse outcomes and higher mortality rates. Hazebroek et al ⁸ linked cardiac involvement to significantly higher all‐cause and cardiovascular death, whereas Pugnet et al ⁷⁸ noted that cardiac involvement was more prevalent in patients with longer disease duration and those experiencing disease relapse. Consistent with these findings, our review underscores the prognostic significance of cardiac involvement detected by cMRI. The high prevalence of myocardial fibrosis and its persistence during remission suggest that patients with cardiac involvement may be at increased risk of adverse outcomes, reinforcing the need for routine cardiac screening and early therapeutic interventions to improve long‐term prognosis. However, a potential indication bias must be acknowledged when interpreting higher LGE rates in patients classified as being in remission, as cMRI might be performed more often in individuals with previous or suspected cardiac complications. In addition, the decision to order advanced imaging may vary among treating physicians, further introducing selection bias. Future prospective and longitudinal studies with systematic cMRI screening—both in active disease and in remission—would help clarify the true prevalence and progression of cardiac involvement in AAV.

As previously discussed, the majority of the included studies were assessed as having a moderate risk of bias due to limitations such as inadequate consideration of confounding factors and incomplete follow‐up reporting. These methodologic shortcomings may influence the reliability of the synthesized findings. Specifically, the prevalence of myocardial fibrosis and other cardiac abnormalities could be affected by selection bias, measurement bias, and unaddressed confounders. The heterogeneity of study designs, patient populations, and cMRI protocols across studies also contributes to variability in the reported outcomes.

The studies included in this review have several limitations that may affect the interpretation of the findings. Many studies had small sample sizes, limiting the generalizability of the results. For instance, the study by Hansch et al ⁶⁴ included 7 patients, the study by Yune et al ⁸³ had 16 patients, and the study by Giollo et al ⁶² had 26 patients. The majority of studies were observational and cross‐sectional, making it difficult to establish causality or assess the progression of cardiac involvement over time. Methodologic limitations, such as heterogeneity in cMRI protocols and diagnostic criteria, were evident. Differences in imaging techniques, LGE assessment methods, and definitions of cardiac involvement can lead to variability in reported prevalence rates and hinder direct comparisons between studies. Some studies did not adequately control for confounding factors, such as comorbid cardiovascular risk factors and differences in disease severity or duration, which could influence cMRI findings. Moreover, cMRI findings in GPA should be interpreted in light of potential verification bias, as cMRI is often ordered only when cardiac involvement is suspected. Consequently, the prevalence of cardiac abnormalities in this subgroup may be artificially heightened compared to a more uniformly screened population. The risk of bias assessments indicated that most studies were of moderate quality, with potential biases related to selection, measurement, and confounding. The observational nature of these studies introduces the possibility of selection bias, and the lack of blinding may result in measurement bias.

Our systematic review has several limitations. We included only studies published in English, which may introduce language bias and exclude relevant research in other languages. Several included studies were conference abstracts with limited information, precluding a thorough assessment of methodologic quality. The heterogeneity among studies in terms of patient populations, disease activity, cMRI protocols, and reported outcomes limited our ability to perform a meta‐analysis, necessitating a narrative synthesis of the data. Additionally, the inclusion of studies with varying definitions of remission and inconsistent use of standardized disease activity scores may affect the comparability of findings related to disease activity and cardiac involvement. The reliance on published data limits the ability to explore individual patient data or unpublished outcomes that could provide a more comprehensive understanding of the prognostic significance of cardiac involvement in AAV.

To our knowledge, this is the first systematic review of cMRI findings in patients with AAV. The high prevalence of subclinical myocardial fibrosis detected by cMRI in patients with AAV underscores the need for routine cardiac assessment using advanced imaging techniques in this population. Clinicians should be aware that conventional diagnostic methods may not be sufficient to detect early cardiac involvement. Incorporating cMRI into the standard evaluation of patients with AAV, even those in clinical remission or without cardiac symptoms, may facilitate early detection and intervention, potentially improving long‐term cardiovascular outcomes. Policy‐wise, guidelines for the management of AAV should consider recommending cMRI as a valuable tool for cardiac assessment. Insurance coverage and resource allocation for advanced cardiac imaging in this patient population may need to be addressed to ensure accessibility. Future research should focus on longitudinal studies to assess the progression of myocardial fibrosis over time and its impact on clinical outcomes. Randomized controlled trials investigating the effectiveness of therapeutic interventions in reducing myocardial fibrosis detected by cMRI would provide valuable insights. Standardization of cMRI protocols and diagnostic criteria should be established to enhance comparability across studies.

AUTHOR CONTRIBUTIONS

All authors contributed to at least one of the following manuscript preparation roles: conceptualization AND/OR methodology, software, investigation, formal analysis, data curation, visualization, and validation AND drafting or reviewing/editing the final draft. As corresponding author, Dr Karageorgiou confirms that all authors have provided the final approval of the version to be published and takes responsibility for the affirmations regarding article submission (eg, not under consideration by another journal), the integrity of the data presented, and the statements regarding compliance with institutional review board/Declaration of Helsinki requirements.

Supporting information

Appendix S1: Supplementary Bias_Assessment_Tables_detailed

Appendix S2: Supplementary Master_data_extraction

Appendix S3: Supplementary Study protocol‐2

AAVcMRI search strategy.

Disclosure form.