Immunosuppressive Therapy for Active Lymphocytic Myocarditis

The role of immunosuppression in the treatment of myocarditis is still debated because of the controversial results obtained in both children^1,2 and adults^3,4 presenting with either cardiac arrhythmias⁵ or heart failure.⁶ These discrepancies may be due to the heterogeneity of pathogenic components, including host immune response, type of infectious agent, and mechanism of cell damage (directly cytopathic or immunomediated). Nevertheless, it is recognized that despite a spontaneous resolution occurring in up to 40% of patients with acute myocarditis,⁷ some patients with chronic heart failure benefit from immunosuppression. Thus, it would be useful to identify biological markers of potential candidates for immunosuppression among the various forms of inflammatory myocardial disease.

In a retrospective analysis of patients with active lymphocytic myocarditis and chronic heart failure who failed to benefit from conventional supportive treatment, we tried to define the virological and immunologic profile of responders and nonresponders to immunosuppressive therapy.

Methods

Patient Population

From January 1997 to July 2000 at our Institutions, 652 patients underwent an endomyocardial biopsy, 478 because of idiopathic left ventricular dysfunction (ejection fraction <40%), 136 for severe ventricular arrhythmias with preserved contractility, and 38 because of unexplained chest pain. A diagnosis of active myocarditis was made in 112 patients, 80 with left ventricular dysfunction (16% out of 478 patients), 28 with ventricular arrhythmias, and 4 with unexplained chest pain. Among the 80 patients with active myocarditis and left ventricular dysfunction, 41 (29 men and 12 women with a mean age of 42.9±13.5 years) were characterized by lymphocytic myocarditis and chronic (lasting >6 months) heart failure (NYHA class III/IV) unresponsive to conventional therapy. These patients were treated for 6 months with prednisone and azathioprine, which are among the safest immunosuppressive strategies adopted in previous similar studies.^4,8 The remaining 39 patients were not treated with immunosuppression, 21 because of heart failure of recent onset and thus who still potentially susceptible to spontaneous recovery, 5 because of the presence of systemic disorders, 5 due to insulin-dependent diabetes, 3 due to renal failure, and 2 because of recent steroid therapy. Three patients were excluded from the study because of histological evidence of eosinophilic (1 patient), giant cell (1 patient), or granulomatous (1 patient) myocarditis.

Therapeutic Protocol and Definition of Improvement

All patients were on full conventional therapy including digitalis (0.25 mg daily), diuretics (furosemide 25 to 50 mg daily), ACE inhibitors (enalapril 20 mg bid daily), and carvedilol (25 to 50 mg daily) for at least 3 months. In addition, they received immunosuppressive therapy including prednisone 1 mg · kg⁻¹ · d⁻¹ for 4 weeks followed by 0.33 mg · kg⁻¹ · d⁻¹ for 5 months and azathioprine 2 mg · kg⁻¹ · d⁻¹ for 6 months.

The efficacy of therapy was evaluated at the end of the treatment. The patients were classified as responders if they had a decrease of at least one NYHA class and an improvement in ejection fraction ≥10% compared with baseline measurements. The patients were classified as nonresponders if NYHA class and ejection fraction failed to improve or deteriorated or if they had major events such as cardiogenic shock, heart transplantation, or cardiac death.

Clinical Studies and Follow-Up

Clinical assessment, resting ECG, Holter monitoring, and 2D echocardiography were performed at baseline, weekly during the first month, every 4 weeks for the remaining 5 months, and every 2 months until 1 year of follow-up. Cardiac catheterization, angiography, and biventricular endomyocardial biopsy were performed at baseline and at 1 and 6 months. Coronary angiography was performed only at baseline. Ejection fraction was calculated by echocardiography in the apical 4- and 2-chamber views using the modified Simpson’s method.

Endomyocardial biopsies (3 or 4 from each ventricular chamber) were performed in the septal-apical region of both ventricles. Two to 3 samples were immediately frozen in OCT compound with isopentane cooled in liquid nitrogen for molecular studies. The remaining tissue specimens were fixed in 10% buffered formalin and embedded in paraffin wax. Blood samples were collected at the time of cardiac catheterization and stored at −80°C.

All invasive cardiac procedures were performed after informed patient consent and approval by the Ethics Committees of our institutions.

Immunohistological Study

Four to six endomyocardial samples obtained from each patient were processed for histological and immunohistochemical studies. Multiple five-micron-thick sections were stained with hematoxylin-eosin, Miller’s elastic Van Gieson, Masson’s trichrome and examined by light microscopy. Specific stains like Ziehl-Nielsen and Giemsa were additionally used in the case of intracellular inclusions. The Dallas criteria⁹ were adopted for histological diagnosis of myocarditis. In all samples, immunohistochemistry for the characterization of inflammatory infiltrates was carried out as previously described.⁵

Molecular and Immunologic Retrospective Studies

At the end of the therapeutic study, a retrospective virological and immunologic characterization of the patients enrolled was performed by a pathologist who was blind to the clinical outcome of the 41 patients.

Two frozen myocardial specimens from each patient were used for polymerase chain reaction (PCR) and reverse transcriptase PCR analysis, as previously described.⁵ Ventricular myocardial samples obtained during cardiac surgery from 5 age-matched patients with stable angina and no systemic infection or histological evidence of myocarditis served as controls for the PCR analysis. Ten primer pairs were used to detect cardiotropic viruses DNA and RNA.^5,10 A nested PCR for the highly conserved 5′ noncoding region of hepatitis C virus (HCV) was performed to detect this virus. The positive and negative strands of viral RNA (HCV and enterovirus) were also detected to identify replicative infective forms. Each assay included negative controls (ie, the above described control patients and reaction mixture without template) and known positive controls (infected viral cells). The purified PCR products were sequenced directly on an automated ABI Model 310 A sequencer, as previously described.¹⁰ For nucleotide sequence analysis and comparisons, the programs Navigator, Fasta, and MAP of the Wisconsin Genetics Computer Group (version 9.1) sequence analysis software package were used. Viral types were identified when nucleotide comparisons revealed an identity of >95% with the known type. Blood samples of patients whose myocardium was infected by a viral agent were analyzed by PCR for the presence of the same virus. In the presence of positivity for HCV genome in the myocardium, the viral load in the serum was also evaluated at baseline and during follow-up. HCV antigen was assessed by the monoclonal antibody TORDJI-22 (Biogenex 1:60) specific for the HCV c100 protein, using 3Amino9ethylcarbazole as chromogen.

All patients at the time of the first endomyocardial biopsy underwent serological tests for the most common cardiotropic viruses and immunologic studies.⁵ The presence of circulating cardiac autoantibodies was evaluated by indirect immunofluorescence, as previously described.¹¹

Statistical Analysis

All values are expressed as mean±SD. Echocardiographic and hemodynamic data of the 2 groups were compared using Student’s t test for continuous variables and Fisher’s exact test for categorical data.

Results

Patient Population

No patient had a history of familial cardiomyopathy, autoimmune disease, recent pregnancy, or alcohol abuse. There were no significant differences in clinical, echocardiographic, or hemodynamic baseline data between responders and nonresponders to immunosuppressive therapy. The requirement for inotropic agents was similar in the 2 groups (Table 1). Clinical, echocardiographic, and hemodynamic data are summarized in Tables 1 through 3.

Immunohistological Studies

Histological analysis showed an active myocarditis with diffuse inflammatory infiltrates associated with focal necrosis of adjacent myocytes (meeting the Dallas criteria) associated with interstitial and focal replacement fibrosis in all left ventricle and right ventricle specimens. Endocardial thickening with prominent smooth muscle cells suggested long-lasting ventricular dilatation. The infiltrates included mainly activated T lymphocytes (CD45RO+) with a moderate amount of cytotoxic lymphocytes (CD8+).

Response to Therapy and Follow-Up

Out of 41 patients, 21 responded to immunosuppression, with improvement in NYHA class, reduction in cardiac volumes, and improved function (Figure 1) at 6-month and 1-year follow-up (Tables 2 and 4). Conversely, 20 patients failed to show improvement: 12 remained stationary, 3 underwent cardiac transplantation, and 5 died within the 6 months after immunosuppression (Table 3).

• Download figure
• Download PowerPoint

Control histology showed a healed myocarditis in responders (Figure 2) and myocyte degeneration with disappearance of inflammatory infiltrates in nonresponders (Figure 3).

• Download figure
• Download PowerPoint

• Download figure
• Download PowerPoint

In responders, a decline of heart rate, disappearance of gallop rhythm, increase in QRS voltages, and improvement in ECG repolarization abnormalities occurred within 1 week of treatment. However, in the 3 HCV-positive responders, cardiac function deteriorated a few months after discontinuation of immunosuppression. This was associated with histological evidence of relapsing myocarditis, which improved after the restoration of immunosuppressive therapy and persisted on a maintenance dose (0.33 mg/kg prednisone daily). During treatment, 6 patients had an increase in body weight >5 kg and 7 developed mild hypertension.

Virological and Immunologic Studies

The presence of sufficient target nucleic acid for PCR analysis was confirmed by amplification of β-globin for DNA and 3GPDH for RNA in all cases but one (case 21 among responders). Viral genomes were found in 17 nonresponders (85%) and in 3 responders (14%; P<0.001). Both positive and negative strands of enterovirus genome were present in positive patients. Transplanted and deceased patients had enterovirus infection (4 patients), adenovirus (3 patients), and adenovirus-enterovirus (1 patient) infection (Table 3).

Three of the responders were positive for HCV (15%), with positive and negative strands of HCV RNA in the myocardium and serum (Table 2). In these patients, the viral load in blood samples increased during immunosuppressive treatment and returned to the previous levels after its discontinuation. Blood samples from the other virus-positive cases analyzed by PCR were negative for the presence of viral genomes. Focal intracytoplasmic positivity for TORDJI-22 was observed in the myocytes of all 3 patients with HCV infection. None of the 5 controls was positive for any virus.

Sequencing analysis of enterovirus and adenovirus PCR amplimers showed a high homology with coxsackievirus B3 (accession number U57056: cases 10, 13, 15, 17, and 18), rhinovirus 14 (accession number KO2121: case 12), adenovirus 2 (accession number J01966: case 11), and adenovirus 5 (accession number S74067: cases 8, 12, 14, 20). Comparison of influenza virus A and Epstein-Barr virus showed high homology with human viral sequences (accession numbers V01104 and V015555 for case 3 and cases 2, 5, 6, 16, and 19, respectively). HCV-positive cases showed high homology for genotype 1. Apart from viruses, no specific agents like bacteria, rickettsiae, fungi, or protozoans were identified at histological examination.

Serological tests for cardiotropic viruses were positive in 6 patients (14.5%), 5 responders and 1 nonresponder. Among responders, the 3 with serological evidence of anti-HCV antibodies also had detectable HCV viral genome in the myocardium. The other 2 responders with positive serology had no evidence of myocardial viral infection. In nonresponders with serological evidence of influenza A infection, the genome of that virus was present in the myocardium. In the other nonresponders with myocardial viral infection, there was no positive serology for the viral agents involved. Positivity for antinuclear antibodies (titer ≥1/80) was found in 6 patients, 3 responders (all HCV positive) and 3 nonresponders.

Cardiac autoantibodies were present in 19 responders (90%), 15 patients (79%) with the organ-specific pattern of staining and 4 patients with the cross-reactive 1 pattern, and in none of the nonresponders (P<0.001; Tables 2 and 3).

Discussion

In our retrospective study, patients with active lymphocytic myocarditis and chronic heart failure had an opposite response to immunosuppressive therapy, despite similar duration and severity of symptoms and histological findings. We found that 85% of nonresponders had viral genomes in the myocardium and no detectable cardiac autoantibodies in the serum. Conversely, only 14% of responders had viral particles in the myocardium, but 90% had cardiac autoantibodies.

After the acute phase of idiopathic lymphocytic myocarditis, which undergoes spontaneous resolution in up to 40% of the cases,⁷ at least some patients with persisting myocarditis and chronic heart failure are likely to benefit from immunosuppression. The controversial results obtained in children^1,2 and in adults^3,4 presenting with either cardiac arrhythmias⁵ or heart failure⁶ are probably related to the heterogeneity of the patients included in such studies. Thus, at present, immunosuppression is recommended essentially for the treatment of eosinophilic,¹² granulomatous, giant-cell myocarditis and lymphocytic myocarditis associated with connective tissue diseases or with transplanted heart rejection.

Recently an upregulation of human leukocyte antigens in the myocardium was indicated as a marker of the beneficial effects of immunosuppression in lymphocytic myocarditis.⁸ In a randomized placebo-controlled study, Wojnicz et al⁸ identified for the first time an autoimmune mechanism in a homogeneous subgroup of patients with inflammatory dilated cardiomyopathy. However, the use of immunohistochemical semiquantitative methods and the absence of information on the type and persistence of the causal agent may explain the findings that only 70% of patients with myocarditis and HLA upregulation benefited from immunosuppression versus 30% of the placebo group.

In our study, responders showed an early cardiac improvement, with decline of heart rate, disappearance of gallop rhythm, and amelioration of ECG abnormalities within 1 week of treatment. This observation can be a useful indication when an ex iuvantibus treatment of myocarditis with immunosuppression is attempted in the absence of objective evidence. The high prevalence of a prompt response to immunosuppression, compared with previous animal¹³ and human⁴ studies, suggests the possibility that a favorable response is more common in myocarditis elicited by an autoimmune disease, either related to exposure of host cardiac antigens or to a molecular mimicry between epitopes shared by viral and host antigens.^14,15 The beneficial effect of nonspecific immunosuppressive agents used in our study could be explained by the combined suppression of humoral and cell-mediated cytotoxic and cardiodepressant effects induced by myocardial inflammation.

Among the various agents identified in nonresponders, enterovirus and adenovirus or their combination were associated with the worst clinical outcome (cardiac transplantation or death), in agreement with previous experimental studies¹⁶ and with the adverse predictive value of the presence of viral genome, particularly adenovirus and enterovirus in the myocardium of transplant recipients.^10,17 Preliminary studies suggest that such patients may benefit from the administration of β-interferon, which can provide complete myocardial clearance of viral genome and long-term improvement of cardiac function.¹⁸ In 3 responders, the type 1b of HCV was uniformly observed in association with cardiac autoantibodies and in the absence of hepatic or other systemic manifestations. HCV infection is known to be associated with several hepatic and extrahepatic autoimmune diseases, including essential mixed cryoglobulinemia, lymphocytic sialoadenitis, membranoproliferative glomerulonephritis, and panarteritis nodosa.^19–21 In such cases, corticosteroids are the first-choice therapy, although increased viral replication may occur, exposing viral antigens on the cell surface that are cleared by T lymphocytes with inflammatory reactivation when immunosuppressive drugs are withdrawn. This mechanism, already documented in HCV hepatitis,²² may explain the recurrence of myocarditis after immunosuppression withdrawal and the need in such patients for contemporary use of antiviral agents such as anti-sense oligonucleotides or ribozyme.

Serology for cardiotropic viruses failed to predict the presence of viral genome in the myocardium because it was positive in both patients with positive and negative PCR results. In particular, positive serology for influenza A virus was observed in 2 responders in the absence of the viral genome in the myocardium, casting doubts on its causal role. Therefore, our findings suggest that serology cannot be used as an alternative to endomyocardial tissue PCR for the diagnosis of viral myocarditis, in agreement with previous reports.²³

In conclusion, in patients with active lymphocytic myocarditis, the presence of detectable cardiac autoantibodies and the absence of viral genome in myocardial biopsy specimens represent a strong indication for immunosuppressive strategies. Patients with immunomediated myocardial damage related to HCV infection may also obtain partial benefit, although adjunct therapy with antiviral agent may be required.

Footnotes