Diagnosis, Treatment, and Outcome of Giant-Cell Myocarditis in the Era of Combined Immunosuppression

From the year 1991 through 2011, 32 patients with histologically verified GCM were seen at the Division of Cardiology, Helsinki University Central Hospital. The majority of diagnoses (29/32) were made after year 2000, that is, during the latter half of the study period. The medical records, laboratory test, imaging studies, and available biopsy material of all patients were retrospectively reviewed and analyzed for the present study. Of the 32 cases, 26 were diagnosed by endomyocardial biopsies (EMBs [n=23]) or surgical biopsies (n=3), 4 at autopsy, and 2 from explanted hearts post-transplantation. The median follow-up time calculated from symptom onset was 15.0 months (range, 0.3–90.3 months). Nine patients with ventricular tachycardia (VT) have been reported previously⁶ but we provide new clinical and follow-up data of them in the present report.

From 1991 through 2005, we were routinely using clinical examination, 12-lead ECG, laboratory tests, and echocardiography to explore the pathogenesis of an unknown myocardial disease. Selective coronary angiography and left-ventricular cineangiography were done whenever coronary artery disease was to be excluded. Angiography was not performed if coronary disease was considered unlikely, eg, in women aged <50 years free of risk factors for atherosclerosis. EMBs were taken if a chronic infiltrative or inflammatory myocardial disease or an aggressive myocarditis, like GCM, was considered possible. Biopsies were usually omitted, however, if chronic dilated cardiomyopathy was the likely diagnosis on clinical grounds. Right-ventricular septum was the target area for myocardial sampling. In 2005, we revised our diagnostic strategy and have since been actively using gadolinium-enhanced cardiac MRI and ¹⁸F-fluorodeoxyglucose positron emission tomography (¹⁸FDG-PET) combined with resting myocardial perfusion imaging with ^99mTc-tetrofosmin SPECT to identify and localize possible inflammatory myocardial processes. The details of these imaging methods in our hands were recently published.⁷ Concomitant with the active use of gadolinium-enhanced cardiac MRI and ¹⁸FDG-PET we changed our EMB policy from random right-ventricular septal sampling to targeted biopsies of myocardial areas showing signs of damage or inflammation at MRI or PET. Since then, left-ventricular biopsies were also acquired more often and we started to admit patients for repeat EMBs if the first samples showed nondiagnostic despite other findings suggestive of an inflammatory myocardial disease.⁷

In the early 1990s, patients with biopsy-verified GCM were treated with a combination of prednisone and azathioprine in addition to a general treatment for heart failure and arrhythmias. Since the first report of the Multicenter GCM Study Group,¹ our recommended treatment has been the triple combination of cyclosporine, prednisone, and azathioprine. Exceptionally, other immunosuppressive drugs like mycophenolate mofetil, methotrexate, or the T-cell antibody muromonab could be added or substituted for other agents. Our practice has been to initiate cyclosporine cautiously aiming at trough blood concentrations inside the lower therapeutic range for immunosuppression after cardiac transplantation (80–120 μg/L). Prednisone was started at 60 mg (or ≈0.75–1 mg/kg) per day decreasing the dose thereafter at 1- to 2-month intervals to 10 mg per day after 6 months. In very severe cases, steroids could be introduced with intravenous methylprednisolone, 500 to 1000 mg per day for 2 to 3 days. The duration of prednisone was not predefined, and continuation at a small dose (5–10 mg per day) was possible as long as was clinically considered necessary. Discontinuation of prednisone was considered if the patient had been stable for 6 to 12 months or if the side-effects were intolerable. The target dose of azathioprine was ≈1.5 to 2 mg/kg per day. The dose was adjusted according to blood count and liver enzyme surveillance. Once stabilized, the patients were seen 2 to 4 times a year in our outpatient cardiology service.

We initially identified from the medical records 36 patients with GCM, the diagnoses having been made by pathologists in the referring hospitals (n=18) or at our institution (n=18). The diagnosis of GCM was based on the presence of a widespread inflammatory infiltrate including lymphocytes, histiocytes, and multinucleated giant cells in association with myocyte necrosis and eosinophils. We adhered to the early criteria of the Multicenter GCM Registry¹ and, to avoid any mistaking of sarcoidosis for GCM, unequivocal granuloma formation excluded the diagnosis of GCM. Two experienced cardiac pathologists (K.S., A.R.-S.) reanalyzed the tissue samples. By the consensus of the reviewing pathologists, 4 of the initial diagnoses of GCM were converted to cardiac sarcoidosis. In 2 cases, the slides were unavailable for reanalysis. One diagnosis was based on autopsy and the other on EMB. The available histology reports described findings typical for GCM by experienced pathologists and both cases were retained in the study.

We used χ² or log-rank tests for comparison. Survival analyses were calculated, first, from the onset of symptoms and, second, from the date of diagnosis. The time point of symptom onset was considered the date of the first medical contact for symptoms compatible with GCM. The former analyses encompassed all patients, even the ones diagnosed at autopsy or after transplantation, whereas the latter ones only included the 26 patients who were identified by cardiac biopsies and underwent GCM-targeted treatment. Including in the analysis cases identified postmortem or post-transplantation assumes that there is no benign form of GCM and that the diagnosed cases are therefore representative of all those who have the disease. In this respect our methods are identical to the strategy used in the prior survival studies in GCM.^1,3 The primary end point for assessing outcome was heart transplantation or death. Transplant-free survival rates were compared using the log-rank test. In all tests, P<0.05 were considered statistically significant. All analyses were performed using SPSS-19 for Windows (SPSS Inc, IL).

The mean age (±SD) was 52.5±12.7 years (range, 35–69) in patients diagnosed at autopsy or transplantation and 49.5±11.1 years (range, 29–70) in patients diagnosed by lifetime biopsy. Of the 32 patients, 22 (69%) were female and 10 (31%) were male. The main presenting clinical manifestations were congestive heart failure in 10 of the 32 patients (31%), distal atrioventricular block in 10 patients (31%), sustained VT in 7 cases (22%), and a syndrome mimicking acute myocardial infarction with chest pain and ST-segment changes in 4 cases (13%). In 1 case, the first manifestation was an out-of-hospital cardiac arrest and death because of ventricular fibrillation. Six patients (19%) had associated autoimmune disorders: reactive arthritis, iritis, and thyreoiditis (n=1), vitiligo with orbital myositis (n=1), coeliac disease (n=1), psoriasis (n=1), rheumatoid arthritis (n=1), and hypothyroidism (n=1). Among the 32 patients, there were 2 siblings; both had coeliac disease. They had 1 healthy brother and their first-degree relatives did not have evidence of cardiomyopathy or autoimmune diseases.

Thirty-one patients underwent echocardiography during the initial diagnostic work-up. The mean ejection fraction was 38±13% with values <50% in 23 patients (74%). The left ventricle was dilated (end-diastolic diameter >55 mm in women or >60 mm in men) in 9 patients (28%). Twenty-one patients (68%) had locally thinned or thickened interventricular septum and 3 (10%) had aneurysms of the left ventricle (Figure 1). Twenty-two patients underwent coronary angiography; none had evidence of coronary artery disease.

Contrast-enhanced cardiac MRI was done in 9 patients and showed areas of late contrast enhancement in each. Twelve patients underwent ¹⁸FDG-PET combined with a myocardial perfusion study. Of them, 10 patients had focally enhanced ¹⁸FDG uptake that was superimposable on a ^99mTc-tetrofosmin perfusion defect in 9 patients and involved the septum in all. Two patients had a perfusion defect without a hot spot at PET.

A total of 28 (88%) of the 32 patients underwent 1 or more EMB sessions or surgical biopsies (Figure 2). The reported number of samples taken per session varied from 2 to 10 (median, 5.3). One of the procedures was complicated by pericardial tamponade requiring surgical drainage. GCM was diagnosed at the first EMB in 19/28 cases (sensitivity, 68%). Of the 9 patients with nondiagnostic histology initially, 7 underwent a second EMB session, which gave the diagnosis of GCM in 5 patients. Finally, 2 patients had a third EMB that exposed GCM in both. Thus, repeated procedures improved the yield of EMB from 68% to 93% (26/28 patients). In 3 patients, the repeat EMBs were taken from the left ventricle. All surgical biopsies were from the left ventricle. Typically, the first nondiagnostic biopsies showed nonspecific abnormalities such as myocyte degeneration, edema, and nuclear size variation. The surgical biopsies revealing the presence of GCM in 3 patients were done in association with (1) resection of a left-ventricular aneurysm, (2) placement of a left-ventricular assist device, and (3) treatment of cardiac tamponade (see above). The interval from the onset of symptoms to the biopsy diagnosis of GCM varied from 0.3 to 16 months (median, 3 months).

Four of the 32 cases were diagnosed at autopsy (Figures 2 and 3) and 2 from explanted hearts after transplantation. Five of these 6 patients had undergone diagnostic procedures including echocardiography (5/6), coronary and left-ventricular cineangiography (2/6), and EMB (2/6) but none had been studied with MRI or PET. Their diagnoses had been dilated cardiomyopathy (n=3), idiopathic atrioventricular block (n=1), and acute nonspecific myocarditis (n=1). The time from symptom onset to death or transplantation ranged from 1 to 66 months (median, 11 months).

The 26 biopsy-diagnosed patients were all treated with combined immunosuppression that included steroids (n=26), azathioprine (n=24), cyclosporine (n=20), mycophenolate mofetil (n=3), muromonab (n=1), gammaglobulin (n=1), and methotrexate (n=1). Table 1 specifies the combinations used and Table 2 summarizes the mean doses of prednisone, azathioprine, and cyclosporine as well as the blood cyclosporine concentrations during treatment. Seven patients received high-dose intravenous methylprednisolone. Twenty-five of the 26 patients were on β-blockers and 19 received antiarrhythmic drugs (mainly amiodarone). Intracardiac defibrillators (ICD) were implanted in 18 patients, and 9 patients received a permanent pacemaker solely for a high-grade atrioventricular block.

The most significant side-effects of treatment included cyclosporine-related elevations of serum creatinine in 10 patients (100–156 μmol/L); elevated liver enzymes, lymphocytopenia or pancreatic irritation attributed to azathioprine in 6 patients; and muscular weakness, marked weight gain, insomnia, or cataract from prednisone in 7 patients.

Figure 4 shows Kaplan-Maier curves for survival from the onset of symptoms for all 32 patients. Altogether 15 patients (47%) either died (n=5) or underwent transplantation (n=10), a median of 11.0 months (range, 0.3–66.0 months) from symptom onset. Among the 5 fatalities, GCM was diagnosed postmortem in 4 patients and the fifth patient died 2 weeks after the diagnostic EMB. All 5 deaths were arrhythmic. Of the 10 transplanted patients, 3 died of postoperative complications (large intracardiac thrombus, bleeding, multiple organ failure) within 38 days of surgery. The remaining 7 transplanted patients were alive at the end of our follow-up in December 2011. One patient had a recurrence of GCM in the graft 4.8 years post-transplantation. The Kaplan-Meier estimates of transplant-free survival from symptom onset (95% CI) were 69% (50%–83%) at 1 year, 58% at 2 years (39%–75%), and 52% at 5 years (34%–70%). Age, sex, or symptoms at presentation did not predict outcome.

Figure 5 shows the Kaplan-Meier curves for survival from onset of GCM-targeted therapy (ie, from diagnosis) in the 26 patients diagnosed by EMB or surgical biopsy. Their follow-up times ranged from 0.3 to 90.0 months (median, 14.5 months). One of the 26 patients died and 8 patients (31%) underwent transplantation while on immunosuppressive medication: 6 patients were transplanted for severe heart failure and 2 for life-threatening arrhythmias. All 8 listings for transplantation were made within 9 months of diagnosis, and the transplantations were accomplished within 1 year of diagnosis in 7 of the 8 cases. Three transplanted patients died early of postoperative complications (see above). Thus, at the end of follow-up 22 of 26 patients (85%) having received GCM-targeted treatment were alive. Seventeen (65%) were alive free of transplantation with a median of 35.0 months from diagnosis (range, 4.0–90.0 months). Table 3 compares the characteristics of patients with and without the primary outcome event (death or transplantation) during follow-up. The 1 patient who expired was diagnosed 2 weeks before death and thus received targeted therapy only for 2 weeks; therefore, data concerning her have been excluded from the treatment analysis in Table 3. Patients on cyclosporine had a trend toward a lower likelihood of cardiac transplant or death (4/20 versus 4/6, log rank P=0.086). Patients requiring a cardiac transplant had lower median left-ventricular ejection fraction during follow-up (26±7%) than patients surviving free of transplantation (45±11%, P=0.03). The treatment delay (time from symptom onset to diagnosis and treatment) did not correlate with poor prognosis (death/transplant) of GCM (P=0.71). The Kaplan-Meier estimate of transplant-free survival from diagnosis (95% CI) was 77% (56%–90%) at 1 year, 63% at 2 years (42%–81%), and 63% at 5 years (42%–81%).

During follow-up, 24 of the 26 patients receiving GCM-targeted treatment (92%) had 1 or more episodes of VT, 17 patients (65%) had sustained VTs, and 1 patient was resuscitated successfully from ventricular fibrillation. Of the 18 patients receiving an ICD, 4 experienced appropriate shocks for sustained VT (n=3) or ventricular fibrillation (n=1), and in another 4 patients VT was terminated with antitachycardia pacing. In the subgroup of transplant-free survivors, 10 of 17 (59%) had sustained VTs during follow-up and 3 received appropriate ICD shocks. During follow-up, 7 of the 26 patients receiving GCM-targeted therapy had The New York Heart Association class II to IV chronic heart failure with ejection fraction between 20% and 40%. Six of these patients required cardiac transplantation and 1 was stabilized with medical treatment only.