Introduction

Idiopathic inflammatory myopathies (IIM) are a collection of systemic autoimmune connective tissue diseases, with an incidence of 9–14 cases/million people and mainly including dermatomyositis (DM), polymyositis (PM), immune-mediated necrotizing myopathy (IMNM), sporadic inclusion body myositis (IBM), and overlap myositis [12, 25]. Data from recent studies have shown a significant rate of cardiac involvement in individuals with IIM [14], where the incidence of subclinical cardiac involvement in PM/DM patients ranges from 13 to 72%, much higher than other symptoms [21].

Anti-mitochondrial antibody (AMA) is a common autoantibody which is closely related to primary biliary cirrhosis (PBC), but is also noted in other autoimmune diseases, such as scleroderma, autoimmune thyroid disease, and Sjogren’s syndrome [5]. Previously, we have observed 7 patients with AMA-related myositis from North America and found that 71% had cardiac involvement [1]. Due to the fact that most current reports are small sample data or case reports, there is no clear consensus on the association between AMA and heart damage in patients with IIM.

Therefore, in this study, we conducted a systematic review of the literature aiming to analyze the association between AMA and cardiac involvement in IIM patients and evaluate the diagnostic value of AMA for cardiac involvement in patients with IIMs.

Methods

Search strategy

In this meta-analysis, we performed a search independently in PubMed, Web of Science, Embase, and the Cochrane Library to identify studies published in English prior to November 19, 2021. A combination of the following keywords was used to retrieve studies: “idiopathic inflammatory myopathy,” “inflammatory muscle diseases,” “myositis,” “myopathy,” “polymyositis,” “immune-mediated necrotizing myopathy,” “inclusion body myositis,” “dermatomyositis,” “juvenile dermatomyositis” AND “heart,” “pericardial,” “pericardium,” “ventricular,” “valvular,” “echocardiography,” “cardiac involvement,” “myocarditis,” “cardiovascular,” “arrhythmia,” “coronary heart disease,” “atherosclerosis” AND “anti-mitochondrial antibody,” “anti-AMA antibody.” The reference lists of the retrieved articles were also manually screened independently to identify potentially relevant articles. The inclusion criteria were well specified and no discrepancies in search results were found.

Study selection

Studies were selected according to the following inclusion criteria: (1) patients with IIMs fulfilled the Bohan and Peter criteria [10, 16], the 2017 European League Against Rheumatism American College of Rheumatology classification criteria [7], Griggs [7], or Sontheimer [27] criteria; (2) availability of data pertaining to the AMA autoantibody status of patients with IIMs; (3) availability of adequate data to calculate true-positive (tp), false-positive (fp), true-negative (tn), false-negative (fn) rates; studies were excluded if they (1) were reviews, case reports, or letters; or (2) lacked a control group of AMA-negative IIM patients. For similar studies conducted by one research group, we gave priority to articles with a larger sample size.

Main outcome variables

Two reviewers independently browsed the potentially eligible articles and extracted the relevant data from each study. The following information was recorded: the first author’s name, publication year, country, disease type, detection method, the total number of cases and controls, frequency of AMA in cases and controls, and detailed information about cardiac involvement, if available. Disagreements were resolved by discussion.

Risk of bias assessment

Two reviewers independently assessed the quality of the studies included using the Quality Assessment of Diagnostic Accuracy Studies 2 (QUADAS-2) tool [29]. They performed the following assessment procedure: patient selection, index test, reference standard, and flow and timing. Each key domain includes two sections: risk of bias and applicability. If answers to all signaling questions for a domain are “yes,” then we could judge the risk of bias as low. If any question is answered with “no,” potential bias exists. Concerns about applicability are judged as “low,” “high,” or “unclear.” We define “yes” as one score.

Analysis

We used Stata 12.0 software to statistically analyze sensitivity (Sen), specificity (Spe), diagnostic ratio (DOR), positive likelihood ratio (PLR), and negative likelihood ratio (NLR), finally calculating the overall odds ratios (ORs) with 95% confidence intervals (CIs) and the summary receiver operating characteristic (SROC) curve. We used the P-value and I2 statistic to test for heterogeneity in every result. The selected studies are considered to be homogeneous if the P-value is greater than 0.1. When the P-value was > 0.1 or I2 was < 50%, a fixed-effects model was adopted. If P was < 0.1 or I2 was > 50%, we used a random-effects model. In addition, we performed a Deek’s funnel plot for the assessment of publication bias.

Results

Literature search and study characteristics

A total of 343 related articles were obtained following a search of the four databases, with 69 replicated studies. After the titles and abstracts were screened, 250 studies were excluded. Seven studies were included after the final evaluation, with a total of 2308 IIM patients (including 171 AMA-positive and 2137 AMA-negative patients). The flow diagram of the search is shown in Fig. 1. The detailed characteristics of the eligible studies are depicted in Table 1. The studies included were conducted in three different countries, with four studies [10, 16, 32, 33] conducted in China, two [18, 29] in Japan, and one [22] in the USA. Moreover, three studies [18, 22, 32] described the demographics, clinical symptoms, and laboratory tests of AMA-positive patients.

Fig. 1
figure 1

Flow diagram of studies selection process

Full size image
Table 1 Characteristics of the studies included in the meta-analysis
Full size table

Quality assessment

The quality score of each study is presented in Supplementary Table S1. Each study was assigned a score from 0 to 11 points, and higher scores indicated higher quality. The score of each study was more than 9 points. All the included studies received moderately high scores from the QUADAS‑2 quality assessments [30].

Findings of the selected studies

In the meta-analysis, we use the random-effects model to pool the sensitivity and specificity because the I2 value was over 50%. The overall pooled sensitivity and specificity were 0.29 (95% CI: 0.19–0.43) and 0.92, respectively (95% CI: 0.88–0.96; Fig. 2). The pooled PLR was 3.9 (95% CI: 2.82–5.38), NLR was 0.76 (95% CI: 0.66–0.88; Fig. 3), and DOR was 5 (95% CI: 3–7). The overall SROC curve is shown in Fig. 4, and the area under the SROC curve was 0.76 (95% CI: 0.72–0.79). The diagnostic accuracy of cardiac involvement in idiopathic inflammatory myopathy patients was not very high.

Fig. 2
figure 2

Forest plots of estimated sensitivity and specificity of anti-mitochondrial autoantibody testing for cardiac involvement in idiopathic inflammatory myopathies

Full size image
Fig. 3
figure 3

Forest plots of estimated PLR and NLR of anti-mitochondrial autoantibody testing for cardiac involvement in idiopathic inflammatory myopathies

Full size image
Fig. 4
figure 4

Summary receiver operating characteristic (SROC) curve of anti-mitochondrial autoantibody testing. Numbered circles show sensitivity (SENS) and specificity (SPEC) estimates for individual studies. Circle 1 represents the study by Maeda et al. [18]; circle 2, the study by Uenaka et al. [29]; circle 3, the study by Liu, Y et al. [16]; circle 4, the study by Huang, Y et al. [10]; circle 5, the study by Sabbagh. et al. [22]; circle 6, the study by Zhang, L et al. [32]; circle 7, the study by Lixi Zhang et al. [33]. The diamond represents the summary operating point (summary values for sensitivity and specificity) and the broken line delimits the 95% confidence region for the summary operating point. The numbers in brackets are the 95% confidence intervals. AUC area under the curve

Full size image

Publication bias

The Deek’s funnel plot shows t = 0.77, P = 0.48 (Fig. 5), so we consider there was no publication bias.

Fig. 5
figure 5

Deek’s funnel plot to evaluate the publication bias. Circle 1 represents the study by Maeda et al. [18]; circle 2, the study by Uenaka et al. [29]; circle 3, the study by Liu, Y et al. [16]; circle 4, the study by Huang, Y et al. [10]; circle 5, the study by Sabbagh. et al. [22]; circle 6, the study by Zhang, L et al. [32]; circle 7, the study by Lixi Zhang et al. [33]

Full size image

Discussion

In this meta-analysis, the pooled OR showed that AMA are associated with an increased risk of cardiac involvement in patients with IIM (OR = 5.07, 95% = 3.43–7.50). We also found that AMA autoantibody achieved the overall pooled sensitivity of 0.29 (95% CI: 0.19–0.43) and specificity 0.92 (95% CI: 0.88–0.96), and AUC was 0.76 (95% CI: 0.72–0.79). The results show that the diagnostic accuracy is not high, with low sensitivity and high specificity. The DOR represents the value that combines sensitivity and specificity, with a higher value meaning better diagnostic ability. The DOR in our study was 5 (95% CI: 3–7), suggesting the overall pooled accuracy was not high.

Although cardiac involvement in patients with IIM mainly presents subclinical symptoms, it has attracted increasing attention because it occurs frequently and even significantly worsens the prognosis [6, 9, 28]. Current research has found there is a correlation between cardiac involvement and the type of autoantibodies. For example, Behan et al. analyzed 55 patients with PM and found that 69% of PM patients complicated with cardiac involvement were anti-Ro antibody positive, and the main manifestation was complete heart block [2]. Gupta et al. revealed that 25% of patients had arrhythmias in 12 patients with anti-SRP-positive PM [8]. Albayda et al. showed that 71% of the 7 patients with AMA-positive myositis had myocarditis, arrhythmia, and cardiomyopathy [1]. Therefore, further research in this field may help us identify IIM patients who are prone to cardiovascular damage and perform timely interventions.

AMA are autoantibodies against a variety of mitochondrial antigens, which can be detected in the IIM patient’s serum by western blot and enzyme-linked immunosorbent assay. Although the positive rate of AMA in IIM is very low, it plays a role in cardiac involvement, including arrhythmia, myocarditis, heart failure, and cardiomyopathy [11, 13, 15, 19, 31]. At present, the role of AMA in the pathogenesis of cardiac involvement in IIM patients remains to be further elucidated; however, several theories have cast light upon possible mechanisms. Previous findings have shown that in dilated cardiomyopathy, the most enriched protein in the mitochondrial inner membrane is an organ-specific autoantigen [23], and the functionally active antibody–antigen complex of this structure reduces the rate of ADP/ATP exchange in cardiac mitochondria by blocking the substrate-binding site of the carrier protein [24], leading to an imbalance between energy supply and demand. This may be regarded as one of the mechanisms of AMA in cardiomyocyte dysfunction. Furthermore, AMA-related cardiac involvement may be related to mitochondrial dysfunction caused by cellular autoimmunity and humoral autoimmunity maintained by autoreactive T cells [19, 26].

We strictly followed the PRISM guidelines [20] to conduct the meta-analysis. However, there are still several limitations in the meta-analysis. First, despite making every effort to search for relevant studies, we may have ignored some studies not published online. Second, the present studies are mainly from the Asian population, which may lead to some population selection bias. Third, different ELISA kits were used, and the differences between the tests needs to be further confirmed. Finally, the small sample size of overall studies and the low positive rate for AMA could over- or underestimate the predictive ability of AMA for cardiac involvement of IIM patients. Therefore, further studies are required to obtain more definitive evidence.

Conclusion

Our findings indicate that the AMA autoantibody has a low sensitivity and high specificity for the diagnosis of cardiac involvement of IIM patients. Thus, the overall diagnostic value may not be very high. However, AMA still has a warning effect regarding cardiovascular damage in patients with IIM.