Author + information
- Received August 4, 2017
- Revision received September 1, 2017
- Accepted September 1, 2017
- Published online October 30, 2017.
- Michela Brambatti, MDa,
- Maria Vittoria Matassini, MDb,
- Eric D. Adler, MDa,
- Karin Klingel, MDc,
- Paolo G. Camici, MDd,e and
- Enrico Ammirati, MD, PhDe,f,∗ ()
- aDivision of Cardiology, Department of Medicine, University of California San Diego, San Diego, California
- bCardiology and Arrhythmology Clinic, Marche Polytechnic University, University Hospital, “Ospedali Riuniti,” Ancona, Italy
- cCardiopathology, Institute for Pathology, University Hospital Tübingen, Tübingen, Germany
- dSan Raffaele Hospital, Milan, Italy
- eSan Raffaele Vita-Salute University, Milan, Italy
- fTransplant Center and “De Gasperis” Cardio Center, Niguarda Hospital, Milan, Italy
- ↵∗Address for correspondence:
Dr. Enrico Ammirati, “De Gasperis” Cardio Center and Transplant Center, ASST Grande Ospedale Metropolitano Niguarda, Piazza Ospedale Maggiore 3, 20162 Milan, Italy.
Background Eosinophilic myocarditis (EM) is an acute life-threatening inflammatory disease of the heart. Neither large case series nor clinical trials on this specific myocarditis have been reported.
Objectives Based on a systematic revision of all published histologically proven cases, this study aimed to describe the clinical presentation, treatment, and outcome of EM.
Methods The study screened 443 manuscripts in MEDLINE and EMBASE on cases of EM published until June 2017. The authors identified 264 patients and included in the main analysis 179 patients admitted to hospital with histologically proven EM.
Results Median age was 41 years (interquartile range: 27 to 53 years) with similar prevalence in both sexes; pediatric cases (≤16 years of age) accounted for 10.1%. The main symptom at presentation was dyspnea (59.4%), with peripheral eosinophilia observed in 75.9%. Median left ventricular ejection fraction at presentation was 35% (interquartile range: 25% to 50%). The disorders most frequently associated with EM were hypersensitivity and eosinophilic granulomatosis with polyangiitis, which accounted for 34.1% and 12.8% of cases, respectively, whereas idiopathic or undefined forms accounted for 35.7% of cases. Steroids were administered in 77.7% of patients. A temporary mechanical circulatory support (n = 30) was instituted in 16.8% of patients. In-hospital death was 22.3% (n = 40), with the highest occurrence in the hypersensitivity form (36.1%; p = 0.026).
Conclusions EM has a poor prognosis during the acute phase, despite a publication bias that could have led to an overestimation of mortality. Associated conditions are identified in approximately 65% of cases. Specific trials and multicenter registries are needed to provide evidence-based treatments to improve in-hospital outcome.
- endomyocardial biopsy
- eosinophilic granulomatosis with polyangiitis
- eosinophilic myocarditis
- hypersensitivity myocarditis
Eosinophilic myocarditis (EM) is a rare form of myocardial inflammation, characterized by eosinophilic infiltration, and often accompanied by eosinophilia (1–3). The degree of eosinophilic infiltration of the myocardium is thought to depend on the underlying condition as well as the degree and duration of eosinophilic exposure. EM has been reported in association with hypersensitivity reactions (4,5); immune-mediated disorders, such as eosinophilic granulomatosis with polyangiitis (EGPA) (formerly Churg-Strauss syndrome) (6–8); undefined complex hypereosinophilic syndrome (HES) or its myeloproliferative variant (9–11); infections (12); and cancer (13), although the relative proportion of these associations remains incompletely understood (14–16). Furthermore, in a relatively large number of cases, the underlying cause of EM remains unknown. The clinical presentation can be variable, ranging from paucisymptomatic to acute fulminant myocarditis (also called acute necrotizing EM) or chronic restrictive cardiomyopathy (also called Loeffler cardiomyopathy or endomyocarditis) (15).
Only endomyocardial biopsies (EMBs) permit a definite diagnosis of EM (14,15), although clinicians often base the diagnosis on laboratory parameters such as troponin and eosinophilia and imaging, in particular cardiac magnetic resonance (CMR).
EM is often fatal with high in-hospital mortality, in particular when it presents as fulminant form (17), but no definite information on the actual mortality rate exists. EM is probably under-recognized and often discovered on postmortem examination (18). Currently, 18 original studies of EM (including between 2 and 8 patients each) have been published, and no systematic review on this subject is available (for the complete list of articles, see the Online Appendix). Hence, uncertainty remains regarding signs and symptoms at presentation, clinical course, and outcome. There are neither clear evidence-based guidelines nor consensus statements for the treatment of EM, although some evidences support the usefulness of steroids (16).
Our aim was therefore to conduct a systematic review on all published cases of EM. Our specific objectives were: 1) to determine the prevalence of systemic conditions associated with histologically proven EM; and 2) to describe clinical presentation, diagnostic findings, treatment strategies, and outcomes in specific forms of EM based on the underlying condition (i.e., idiopathic, hypersensitivity, EGPA, or HES-related EM).
The systematic review of the literature was undertaken according to the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analysis) guidelines for its design, implementation, analysis, and reporting.
The database was created using an EM specific search strategy in PubMed (MEDLINE) and EMBASE with the filter of languages (English, Italian, or French), and included all case reports and case-series published until June 30, 2017. Three independent authors (M.B., M.V.M., E.A.) performed the search using the term eosinophilic myocarditis. Reference lists of all studies previously identified as having met the inclusion criteria were manually reviewed for additional relevant publications. We excluded case reports and case series regarding tropical endomyocardial fibrosis, a specific cardiomyopathy associated with myocardial eosinophilic infiltrates, as we believe this is a separate entity. Case series were considered only if sufficient data were clearly available for each case.
Data extraction and management
Two independent authors (M.B., M.V.M.) performed analysis of the full texts and abstracts. All data were extracted from article texts, tables, and figures and included into our database (Online Appendix). Discrepancies were resolved by consensus or by a third author (E.A.), if necessary. We further classified the total population in cases with histologically proven EM and cases without histological evidence of EM and we focused on the former. Among these, we finally excluded biopsy-confirmed EM cases of patients with pre-existing advanced heart failure or those who had an out-of-hospital death (Figure 1).
Histologically proven EM was based on EMB data in 138 (77.1%) cases, on surgical biopsies (i.e., post–ventricular assist device [VAD] implantation) in 6 (3.3%) cases, on postmortem examination in 34 (19.0%) cases, and on EMB and autopsy in 1 (0.6%) case. Median time from admission to EMB was 1 day and first to third quartile interquartile range (IQR) was 1 to 6 days (n = 60). EMB was obtained from the right ventricle (RV) in 47 (83.9%) cases, from the left ventricle (LV) in 6 (10.7%) cases, and from both ventricles in 3 (5.4%) cases.
Coronary angiography was performed in 106 (59.2%) patients whereas coronary inspection was reported in 22 (64.7%) of 34 cases that underwent postmortem examination after in-hospital death. Occurrence of coronary vasospasm during angiography has been reported in 2 cases (n = 2 of 106, 1.9%). This is compatible with the diagnosis of Kounis syndrome, which is a hypersensitivity coronary disorder induced by exposure to drugs or other triggers, characterized by arterial (histamine-induced) vasospasm, thrombus formation, and vessel wall infiltration by eosinophils or mast cells (19). Postmortem examination revealed normal coronary arteries in 17 (77.3%) cases, signs of vasculitis in 3 (13.6%) cases, an atherosclerotic lesion (stenosis of 50%) in 1 (4.5%) case, and in 1 (4.5%) case thrombotic occlusion of the coronary.
For identification of associated systemic disorders and statistical analysis, see the Online Appendix.
Identification, screening, and inclusion of EM cases
Figure 1 summarizes the identification, screening, and inclusion process. A total of 443 publications were identified. A total of 227 manuscripts, including 18 case series (the largest series accounted for 8 patients), for a total of 264 clinical cases, met the criteria. First, we considered separately 65 cases lacking histological evidence of EM (see the Online Appendix for a list of nonhistologically proven EM): in 6 cases EMB was performed, but histology did not fully support the diagnosis whereas in 59 cases EMB was not performed. Thus, we identified 199 patients with histologically proven EM. Of these, 20 cases were excluded from the analysis due to pre-existing advanced heart failure such as patients on the heart transplantation (HTx) list, on left ventricular assist device (LVAD) support (n = 8), or with out-of-hospital death (death before admission [n = 12], mainly due to hypersensitivity reaction [n = 9]) (see the Online Appendix for a list of histologically proven EM excluded from the final analysis). Thus, 179 patients with histologically proven EM were included in the main analysis (1,4,7,9,12,13,18) (see the Online Appendix for a list of histologically proven EM): 3 cases were reported before 1985, 25 cases were reported between 1985 and 2000, and 151 cases were published after 2000. Ethnicity of patients was reported in a minority of cases of histologically proven EM (21.2%) and in most cases it was Caucasian (47.4%).
The characteristics of patients with histologically proven EM are reported in Table 1. Patients with histologically proven EM compared with the nonhistologically proven EM group (n = 65) were younger (median age 41 years vs. 46 years; p = 0.02) and less likely to have history of asthma (22.0% vs. 42.2%; p = 0.003). Of note, in histologically proven EM, peripheral blood eosinophilia was significantly less frequently observed (75.9% vs. 98.5%; p = 0.0001). At electrocardiogram, histologically proven EM had a higher prevalence of ST-segment elevation (38.6% vs. 16.0%; p = 0.004) and a lower left ventricular ejection fraction (LVEF) at first echocardiogram (35% vs. 43%; p = 0.03). Presence of an endoventricular thrombus was more frequently reported (28.3% vs. 13.7%; p = 0.02), and CMR has been more frequently performed (40.0% vs. 20.7%; p = 0.005) in nonhistologically proven EM. Finally, an increased number of cardiac arrests has been reported in histologically proven cases (27.2% vs. 4.6%; p = 0.0001).
Histologically proven EM were mostly reported by hospitals in North America (36.9%) Europe (24.6%), and Japan (21.8%). The remaining cases were reported from hospitals in South America (2.2%), New Zealand and Australia (5.6%), and the rest of Asia (8.9%) (Figure 2).
Among the 179 histologically proven cases, an associated systemic disorder was identified in 64.3% (Central Illustration, A and B). Hypersensitivity was the most frequently reported condition (n = 61, 34.1%), followed by EGPA (n = 23, 12.8%), HES (n = 15, 8.4%), and other known causes (n = 16, 8.9%). Other causes that accounted for <10 subjects per group were infections (n = 9, 5.0%), mainly attributable to Toxocara canis infection (also known as visceral larva migrans; n = 6, 3.4%), pregnancy-related EM (n = 3, 1.7%), malignancies (n = 2, 1.1%), overdose of clozapine for attempted suicide (n = 1, 0.6%), and other immune disorder–related EM (Omenn syndrome; n = 1, 0.6%). In 64 (35.7%) patients, EM was classified as idiopathic or undefined.
Overall, among all EM cases during the initial hospitalization, 40 (22.3%) patients died, 5 (2.8%) patients received a long-term VAD (n = 4; in 1 case a biventricular VAD) or a total artificial heart (TAH; n = 1). No patients underwent HTx. Hypersensitivity EM has been observed associated with the highest rate of in-hospital deaths (36.1%) and with the poorest mid-term outcome (53.7% freedom of death, HTx, TAH, or VAD at 120 days from the initial admission (Central Illustration, C). Median hospital stay was 14 days (IQR: 6 to 30 days) and the median time from admission to death or long-term VAD or TAH was 3 days (IQR: 1 to 9 days).
Idiopathic or undefined EM
In 62 patients EM was classified as idiopathic or undefined. Median age was 41 years (IQR: 24 to 58 years) (Table 2). At presentation, patients complained of dyspnea (68.3%) and chest pain (41.3%). Fever was present in 26.7% of cases. At admission, they presented with a normal electrocardiogram in 2.1% of cases. All patients with available data had increased troponin at admission while peripheral eosinophilia was detected in 72.9%. Median LVEF at admission was of 30% (IQR: 25% to 45%), and pericardial effusion was observed in 37.7% of cases, with cardiac tamponade in 4.9% of patients (Table 3). Eleven (18.0%) patients had endoventricular thrombosis (1 thrombus in the RV and 10 thrombi in the LV). The most prevalent late gadolinium enhancement (LGE) pattern at CMR was subendocardial (57.1%). Coronary angiography was performed in 47 patients, with normal findings in 45 patients. In 1 case coronary artery disease was detected at coronary angiography, and postmortem examination found coronary artery thrombosis. Moreover, in 1 case acute vasospasm was reported during angiography. Coronary inspection at autopsy revealed signs of vasculitis in another case, not observed at coronary angiography. Steroid therapy was administered in 82.8% of cases and a second immunosuppressive agent in 11.1% cases. There were 8 in-hospital deaths (12.5%); 1 patient (1.6%) received a TAH as bridge to transplant whereas 2 (3.1%) patients underwent to a long-term LVAD and biventricular VAD implantation, respectively, both subsequently explanted due to functional recovery (Table 4). Median LVEF at hospital discharge was 57% (IQR: 35% to 60%) based on the 30 patients with available data. After hospital discharge with a median follow-up of 150 days (IQR: 90 to 180 days), 1 HTx was performed.
EM associated with hypersensitivity
Hypersensitivity EM was diagnosed in 61 patients, representing the most common identified condition associated with histologically proven EM (61 of 115, 53.0%). The median age was 35 years (IQR: 25 to 47 years) (Table 2). Although not statistically significant, a higher proportion of pediatric patients (16.4%; overall p = 0.07) and a higher prevalence of male subjects (63.5%; overall p = 0.11) was observed in this group compared with EM associated with other conditions. At presentation, patients complained of dyspnea (47.5%), and chest pain (35.5%) followed by nonspecific symptoms (27.9%). Fever at presentation was more frequently reported in this group (54.2%; overall p = 0.004; vs. EGPA-related EM and idiopathic or undefined EM p < 0.05). Of note, hypersensitivity EM was less frequently associated with a history of asthma (8.2%; overall p < 0.001 vs. EGPA-related EM p < 0.01). Increased peripheral eosinophilia was detected only in 63.5% of patients, representing the group with the lowest prevalence of peripheral eosinophilia (overall p = 0.02 vs. all others p < 0.05). Median LVEF at admission was 35% (IQR: 27% to 49%); and pericardial effusion was observed in 30.8% with cardiac tamponade in 5.8% of patients (Table 3). Only 1 patient presented an LV thrombus. Eight patients underwent CMR without evidence of a prevalent LGE pattern. Coronary angiography was performed in 30 patients and showed coronary artery vasculitis in 1 (3.3%) case, and was normal in all the others (Table 3).
A temporary MCS was instituted in 12 (19.7%) cases (Table 4). Hypersensitivity EM was the form in which more frequently a cardiac arrest was observed (up to 44.6%; overall p = 0.009; p < 0.05 vs. idiopathic or undefined EM). Steroid therapy was administered in 42 (68.8%) patients and a second immunosuppressive agent in 8 (15.1%) cases. In-hospital mortality was higher (p = 0.026; p < 0.05 vs. idiopathic or undefined EM) with 22 deaths (36.1%) and 2 long-term LVAD implantations (3.3%) that were subsequently explanted due to functional recovery (Table 4). Median LVEF at hospital discharge was 60% (IQR: 45% to 60%) based on 31 patients with available data. A cardiac death was reported 2 months after hospital discharge after diagnosis of drug rash with eosinophilia and systemic symptoms (DRESS) syndrome.
Drugs more frequently associated with the development of hypersensitivity EM were antibiotics (36.5%, mainly mynocicline and β-lactam antibiotics), central nervous system agents (21.1%; primarily clozapine followed by carbamazepine), vaccines (7.7%), antitubercular agents (1.9%), and other agents in 32.8%. Data about drugs potentially involved were not available in 14.7%.
EM associated with EGPA
In 23 patients, EM was associated with EGPA. The median age at presentation was 50 years (IQR: 26 to 65 years) (Table 2). At presentation, dyspnea and chest pain were the most common symptoms (68.2% and 45.4%, respectively). When compared with other EM forms, this group showed the highest prevalence of history of asthma (up to 68.2% of cases, overall p < 0.001; p < 0.01 vs. idiopathic or undefined and vs. hypersensitivity) and the lowest incidence of fever (15.0% of patients; overall p = 0.004). At the time of admission, peripheral eosinophilia was observed in up to 90.9% of cases. Median LVEF at admission was of 33% (IQR: 27% to 54%), and pericardial effusion was observed in 27.3% with cardiac tamponade in 4.5% of patients (Table 3). An endoventricular thrombus was reported in 19.0% of cases (1 thrombus in the RV and 3 thrombi in the LV). In 8 patients who underwent CMR, the most prevalent LGE pattern was subendocardial (62.5%). At coronary angiography only 1 (6.7%) patient presented signs of coronary artery vasculitis. Echocardiographic, EMB, and CMR findings of an exemplificative case of fulminant EM associated with EGPA are presented in Online Videos 1 and 2 and Figure 3.
A temporary MCS was instituted in 5 (21.7%) cases. Steroid therapy was administered in 20 (87.0%) patients and a second immunosuppressive agent in 7 (33.3%) cases. Five (21.7%) patients died during the hospitalization (Table 4). Median LVEF at hospital discharge was 40% (IQR: 31% to 49%) based on 11 patients with available data. No further events were reported after hospital discharge with a median follow-up of 150 days (IQR: 68 to 349 days).
EM associated with HES
EM was associated with HES in 15 patients with median age of 41 years (IQR: 32 to 53 years) and the highest, albeit not statistically significant, prevalence of women (66.7%; overall p = 0.11) (Table 2). At presentation, patients reported dyspnea (71.4%) and chest pain (50.0%). Fever was present in 23.1% of cases. At admission, peripheral eosinophilia was commonly observed (up to 92.9% of patients). Median LVEF at admission was of 32% (IQR: 21% to 50%), and a pericardial effusion was observed in 40.0%, with cardiac tamponade in 6.7% of patients. An LV thrombus was reported in 4 (28.6%) cases (overall p = 0.036), and it was the EM form with the highest prevalence of endoventricular thrombi. Five patients performed a CMR that revealed an LGE subendocardial pattern in 40%. Coronary angiography showed normal anatomy in 7 cases, and in 1 case coronary vasospasm (Table 3). Only 1 (6.7%) patient needed a temporary MCS and 2 patients died during hospitalization (13.3%) (Table 4). Steroid therapy was administered in 86.7% cases and a second immunosuppressive agent was added in 5 (41.7%) cases. Median LVEF at discharge was 49% (IQR: 23% to 58%) based on 8 patients with available data. A sudden cardiac death was reported after hospital discharge at 3 years (16 months after withdrawal of steroids).
This is the first report based on a large series of published cases that provides data on the clinical presentation, diagnostic findings, treatment, and outcome of patients with histologically proven EM, comparing the characteristics of subcategories of EM associated with specific systemic conditions. We confirm the anecdotal high mortality of EM; in fact, in-hospital death was 22.3%, with a significant increased occurrence in the hypersensitivity form (36.1%) (Central Illustration).
A possible overestimation of mortality can be taken into account due to publication bias. In fact, EM cases not critically ill or with nonspecific signs or symptoms have a lower chance of being recorded or published. In this regard, the real mortality remains difficult to estimate. Furthermore, considering all the case reports (n = 244), including also nonhistologically proven cases of EM, the in-hospital mortality rate remains at 17.2% and hypersensitivity EM remains the form with the highest mortality (31.4%). These data are in line with a previous review on 22 cases of DRESS-associated myocarditis reporting a 55% mortality rate (5).
The fact that peripheral eosinophilia is absent in up to 25% of patients with EM probably contributes to the underdiagnosis of EM without EMB. This is indirectly demonstrated by the comparison with the cohort of 65 patients without histologically proven EM, where peripheral eosinophilia was observed in almost all cases (98.5%). In a study including 8 patients with EM, the changes of peripheral eosinophil count was monitored showing that in 3 of 4 patients with initial eosinophil count of <500/mm3, an increase to ≥500/mm3 occurred 7 to 12 days after onset, stressing the fact that absence of eosinophilia at admission does not exclude the diagnosis of EM (20). In our study, among patients without eosinophilia at admission only 5 of 39 (12.8%) patients developed peripheral eosinophilia, between the second and sixth days of hospitalization. Thus, white blood cell including eosinophil count should be repeated in the first days after admission to reduce the probability of missing the diagnosis.
In particular, the high-risk group of patients with hypersensitivity EM with the highest occurrence of cardiac arrest and in-hospital death can frequently lack peripheral eosinophilia (up to 35% to 40% of cases). In the previously mentioned study, 1 case (33%) out of 3 with hypersensitivity EM had no eosinophilia, in line with our findings. Thus, in patients with allergic reactions, in particular if temporally related to the assumption of minocycline and β-lactam antibiotics, or clozapine and carbamazepine, electrocardiogram and measurement of troponin levels could help exclude cardiac involvement. In fact, the absence of eosinophilia in this group may contribute to the high rate of diagnosis only after sudden death. Postmortem series have documented that in allergic EM the diagnosis was not suspected clinically antemortem due to the fact that clinical signs of cardiac involvement were often nonspecific (15).
The identification of EM-associated condition is relevant for the specific treatments (i.e., use of cyclophosphamide in EGPA-related EM, imatinib in myeloproliferative variant of PDGFRA-associated HES or albendazole in EM associated with Toxocara canis infection) of these patients although up to 35% of cases had an idiopathic or undefined EM. Asthma, which is 1 of the major diagnostic criteria of EGPA, was reported in approximately 21% of idiopathic or undefined EM, leading to the hypothesis that the final diagnosis of EGPA was missed at least in a proportion of these patients due to limited awareness of cardiologists about this condition and its diagnostic criteria.
Evidence of endocavitary thrombi was reported in 12.3% of histologically proven EM, mainly in disorders with persistent eosinophilia such as HES (28.6%) and EGPA (19.0%). This finding could support the role of anticoagulation during the acute phase of EM in particular in these subgroups of patients also for prevention of endocavitary thrombi, even if no trials have tested this hypothesis. It is supposed that the persistent eosinophilic damage of the endomyocardium can foster thrombotic and fibrotic transition from acute EM to Loeffler endomyocarditis. In the nonhistologically proven EM, the incidence of endoventricular thrombus was even higher (28.3% of cases). The presence of an LV thrombus can increase the risk of stroke during LV biopsy. Based on data from a large-volume center, the rate of stroke as complication of LV EMB is approximately 1:300, even if this refers to a mix series where myocarditis accounted for 44% of the diagnosis (21). Thus, echocardiography should be performed before EMB to exclude presence of endoventricular thrombi. In patients that are hemodynamically stable, CMR can be considered as first diagnostic tool providing alternative information to EMB. CMR can clearly identify thrombi or diffuse subendocardial fibrosis (subendocardial LGE pattern), furthermore in the context of a known myeloproliferative variant of HES or EGPA, CMR in association with cardiac biomarkers, electrocardiogram, and clinical signs could be reasonable enough to reach the diagnosis of EM.
It must be noted that based on an autoptic series of 69 cases of hypersensitivity EM, it has been reported that one-half of the cases have infiltrates with focal lesions that could be missed by EMB (22). They also observed that the RV was involved in 96% of cases, suggesting that cardiac damage is patchy, but diffuse. Finally, they observed that cardiac arrhythmias occurred in 42% (in our series 44.6% of cardiac arrest took place in hypersensitivity EM during hospitalization) and did not correlate with the degree of myocardial infiltration or presence of necrosis. In our report we were not able to estimate the accuracy of EMB (only 1 patient underwent both to EMB and postmortem examination, and were both diagnostic for EM). Nevertheless most of the diagnoses were based on RV EMB (83.9%) performed after a median of 3 days from admission. It is possible that in case of negative histological findings in highly suspected acute myocarditis, an image-guided (even if this approach must still validated in the clinical practice) or electrogram-guided EMB could help to increase the diagnostic accuracy of this approach.
Corticosteroids were frequently used (77.7%) in particular in systemic conditions where steroids have a primary indication such as EGPA (87.0%) and HES (86.7%). The hypersensitivity EM group is generally less treated with steroids (68.8%). This finding is quite surprising, but it could be potentially explained by an underestimation of the potential cardiac risk in patients with allergic reactions. This could reinforce the concept that immediate cessation of the offending drug may not be enough to treat hypersensitivity EM. Pathologic findings show that in hypersensitivity EM, eosinophilic myocardial infiltrates range from mild to severe, but correlated with neither the incidence of sudden cardiac death nor the levels of peripheral eosinophilia based on our findings. Thus, patients with allergic reactions and signs of cardiac involvement should be monitored and treated with high doses of intravenous corticosteroids, such as suggested in DRESS (5). It must be emphasized that no clinical trials tested the efficacy of steroids in patients with EM, even if the indication to this class of drug is often related to the associated condition.
We have observed that excluding HES-related and EGPA-related forms where corticosteroids have been widely used, there is a lower incidence of in-hospital death among those patients treated with steroids (n = 10 of 101; 9.9%) compared with those did not receive steroids (n = 23 of 35; 65.7%; p < 0.0001). Nevertheless, this analysis must be considered with caution, due to potential bias (e.g., patients that died too early to receive steroids).
Our study demonstrates that EM has often a fulminant presentation with abrupt impairment of LVEF and high risk of malignant arrhythmias. Thus, some form of circulatory support (inotropes or MCS) may be required in a significant proportion of cases. Despite a high rate of in-hospital mortality among patients with fulminant presentation of EM, recovery after MCS was common. Among 30 patients receiving a temporary MCS associated with steroids in 27 of 30 (90%), 26 (86.7%) survived, and 22 (73.3%) had a complete recovery of the cardiac function before discharge. Furthermore 4 patients who were implanted with long-term VAD had recovered and were explanted.
Our analysis cannot provide data regarding the risks of EM recurrence, future ventricular arrhythmias, or development of post-inflammatory dilated cardiomyopathy. Specific collaborative registries are needed to find an answer to these questions.
The main limitation of the present report is that all data are derived from published cases and this did not allow a check for accuracy and completeness of data. Furthermore, even if cases are reported from all over the world except for Africa (Figure 2), the results mainly reflect the characteristics of patients admitted to hospitals in North America, Europe, and Japan (83.6%), where EMB and CMR imaging are more widely available. Finally, our analysis focused on new onset of symptomatic EM, thus excluding patients with known advanced heart failure, in whom myocardial infiltrate of eosinophils have been observed as in those with prolonged exposure to dobutamine. It must be acknowledged that myocardial infiltrates of eosinophils in these patients generally occur in the absence of symptoms and is usually an accidental finding.
We describe the clinical presentation, diagnostic findings, associated systemic conditions, treatment, and outcome of EM based on a relatively large population derived by current literature. The present study tries to shed lights on a nosological entity with a high early mortality rate that is often neglected by physicians and cardiologists. Clinical trials and international registries are needed to increase awareness regarding this condition and to improve survival.
COMPETENCY IN MEDICAL KNOWLEDGE: EM affects mainly middle-aged individuals, causing dyspnea and ventricular dysfunction, but clinical features vary with specific forms of the disease. While about one-third of cases are idiopathic, another third are associated with hypersensitivity reactions to antimicrobial or neurological drugs. Approximately 25% of acute cases lack eosinophilia on admission and require endomyocardial biopsy for definitive diagnosis. With or without corticosteroid therapy, in-hospital mortality remains high.
TRANSLATIONAL OUTLOOK: Both multicenter registries and randomized trials are needed to develop more effective management strategies for patients with this disease.
Dr. Camici has served as a consultant for Servier. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. Drs. Brambatti and Matassini contributed equally to this work.
- Abbreviations and Acronyms
- cardiac magnetic resonance
- drug rash with eosinophilia and systemic symptoms
- eosinophilic granulomatosis with polyangiitis
- eosinophilic myocarditis
- endomyocardial biopsy
- hypereosinophilic syndrome
- heart transplantation
- interquartile range
- late gadolinium enhancement
- left ventricle/ventricular
- left ventricular assist device
- left ventricular ejection fraction
- right ventricle/ventricular
- total artificial heart
- ventricular assist device
- Received August 4, 2017.
- Revision received September 1, 2017.
- Accepted September 1, 2017.
- 2017 American College of Cardiology Foundation
- Fozing T.,
- Zouri N.,
- Tost A.,
- et al.
- Kindermann I.,
- Barth C.,
- Mahfoud F.,
- et al.
- Ammirati E.,
- Cipriani M.,
- Musca F.,
- et al.
- Ammirati E.,
- Stucchi M.,
- Brambatti M.,
- et al.
- Caforio A.L.,
- Pankuweit S.,
- Arbustini E.,
- et al.
- Ammirati E.,
- Cipriani M.,
- Lilliu M.,
- et al.
- Abdelghany M.,
- Subedi R.,
- Shah S.,
- Kozman H.
- Yilmaz A.,
- Kindermann I.,
- Kindermann M.,
- et al.