Cardiac arrhythmias are a complication of myocarditis. There are no large studies of in-hospital arrhythmia development and outcomes in pediatric patients with acute myocarditis. This was a retrospective 2-center review of patients ≤21 years hospitalized with acute myocarditis from 1996 to 2012. Fulminant myocarditis was defined as the need for inotropic support within 24 hours of presentation. Acute arrhythmias occurred at presentation and subacute after admission. Eighty-five patients (59% men) presented at a median age of 10 years (1 day to 18 years). Arrhythmias occurred in 38 patients (45%): 16 acute, 12 subacute, and 9 acute and subacute (1 onset unknown). Arrhythmias were associated with low voltages on the electrocardiogram (14 of 34, 41% vs 6 of 47, 13%; odds ratio [OR] 4.78, 95% confidence interval [CI] 1.60 to 14.31) and worse outcome (mechanical support, orthotopic heart transplant, or death; OR 7.59, 95% CI 2.61 to 22.07) but were not statistically significantly associated with a fulminant course, ST changes, initial myocardial function, lactate, creatinine level, C-reactive protein and/or erythrocyte sedimentation rate, or troponin I level, after adjusting for multiple comparisons. Subacute arrhythmias were associated with preceding ST changes (10 of 15, 67% vs 15 of 59, 25%, OR 5.87, 95% CI 1.73 to 19.93). All patients surviving to discharge had arrhythmia resolution or control before discharge (10 on antiarrhythmic), with 1 exception (patient with complete heart block requiring a pacemaker). At 1-year follow-up, there were 3 recurrences of ventricular arrhythmias, but no arrhythmia-related mortality. In conclusion, arrhythmias are common in pediatric patients with myocarditis, occurring in nearly 1/2 of all hospitalized children and are associated with a worse outcome. Early identification of subacute arrhythmias using electrocardiographic changes may help management. A majority of patients do not require continued postdischarge arrhythmia treatment.
Cardiac arrhythmias are a common complication of myocarditis, with an incidence ranging from 29% to 100%. The most common arrhythmias are ventricular tachycardia (VT), ventricular fibrillation (VF), and atrioventricular (AV) block. Published literature on arrhythmia development and outcomes in pediatric myocarditis has been limited to case reports and small patient populations studying subsets of cardiac arrhythmias. Friedman et al reviewed 12 pediatric patients with ventricular arrhythmias and occult myocarditis, and Chien et al and Wang et al both evaluated complete heart block (CHB) in 9 patients. The most substantial pediatric study to date by Lee et al described clinical outcomes of acute myocarditis in 36 children. The purpose of this study was to evaluate the incidence and outcomes of in-hospital arrhythmias in a larger cohort of pediatric patients with acute myocarditis and determine predictors of in-hospital arrhythmia development.
Methods
This was a 2-center retrospective review of all patients aged ≤21 years hospitalized with the diagnosis of myocarditis from 1996 to 2012. Medical records were reviewed with permission from the Internal Review Board at the Boston Children’s Hospital and Lucile Packard Children’s Hospital. For the purposes of this study, all patients had (1) a clinical diagnosis of myocarditis by a pediatric cardiologist, (2) prodromal viral or infectious symptoms ≤2 weeks before presentation, and (3) evidence of ventricular dysfunction (ejection fraction <55% by echocardiography within 24 hours of admission). A diagnosis of acute myocarditis also required at least 1 of the following: (1) active or borderline biopsy by Dallas criteria, (2) positive infectious origin of ventricular dysfunction, or (3) delayed enhancement on cardiac magnetic resonance imaging consistent with myocarditis. In the absence of one of these findings, we defined additional patients if (4) left ventricular end-diastolic volume z score was ≤3 and recovery of systolic function occurred within 1 year of diagnosis. Exclusion criteria included a history of congenital heart disease, cardiomyopathy, or arrhythmias.
Fulminant myocarditis was defined as the need for intravenous inotropic support to maintain cardiac output within 24 hours of presentation. Arrhythmias were classified as acute or subacute, depending on the timing of arrhythmia development. Acute arrhythmias were defined as those present before or during initial evaluation by the medical team at the admitting hospital (e.g., primary care physician, emergency room physician, or admitting intensive care physician). Subacute arrhythmias occurred after a patient had completed the initial evaluation. Emphasis was placed on rhythm disorders that were judged a priori to be clinically significant and included (1) sustained supraventricular tachycardia (SVT), (2) high-grade ventricular ectopy (frequent isolated monomorphic or polymorphic ventricular beats or couplets deemed high grade by a pediatric electrophysiologist), (3) sustained or nonsustained VT or VF, and (4) high-grade (Mobitz type II) or CHB. Reported low-grade arrhythmias (e.g., frequent atrial or infrequent single ventricular ectopic beats) were recorded, although these arrhythmias were not considered significant for the purpose of this study.
Left ventricular ejection fraction was defined as: normal ≥55%, mild dysfunction ≥45% and <55%, moderate dysfunction ≥30% and <45%, and severe dysfunction <30%. The Dallas criteria were used for endomyocardial biopsy determination of myocarditis. A clinical worse outcome was defined as the need for any of the following: mechanical support, orthotopic heart transplant (OHT), or death.
All electrocardiograms (ECGs) obtained during the hospital course were evaluated by 2 pediatric electrophysiologists. Timing of development and resolution of electrocardiographic changes including ST elevation or depression, low voltages, and left bundle branch block (LBBB) was recorded and compared with arrhythmia onset. Low voltage was defined as ≤0.5 mV in all limb leads or ≤1.0 mV in all precordial leads. Abnormal ST segments were defined as ≥0.2 mV ST elevation or depression in at least 2 leads. ST segments were not evaluated if the patient had CHB or an intraventricular conduction delay.
Statistical analysis was performed using SAS, version 9.3 (SAS Institute Inc., Cary, North Carolina). Continuous variables are presented as medians with ranges. Categorical variables are presented as counts with percentages. Univariate binary logistic regression models were built to predict associations between arrhythmias, worse outcomes, and subacute arrhythmias. Variables for the primary analysis were chosen before data collection. Because of the small numbers of patients and to account for multiple comparisons in the planned primary analysis of arrhythmia predictors, the Bonferroni correction was applied and statistically significant data were defined as a p value of <0.0056. Subsequent to the primary analysis, a secondary analysis of electrocardiographic findings and associations with worse outcomes and subacute arrhythmias was performed. Because these variables were not chosen a priori, this analysis was termed exploratory, and a standard p value of <0.05 was used.
Results
A total of 85 patients (59% men) presented at a median age of 10 years (range, 1 day to 18 years). Demographic data and presenting symptoms are listed in Table 1 . Arrhythmias occurred in 38 (45%) of 85 patients, acutely in 16, subacutely in 12, and both acute and subacute in 9. In 1 patient, the timing of arrhythmia onset was unknown. Among the 85 patients, a fulminant course was seen in 41 (50%). A majority (80%) received intravenous immunoglobulin, and 21% received steroids in addition to intravenous immunoglobulin. Sixteen patients did not receive either intravenous immunoglobulin or steroids.
| Variable | n | Arrhythmia (%) | No Arrhythmia (%) |
|---|---|---|---|
| Men | 85 | 20/38 (53) | 30/47 (64) |
| Median age | 85 | 8 yrs (1 day–17 yrs) | 14 yrs (8 days–18 yrs) |
| Presenting symptoms | |||
| Constitutional | 85 | 24/38 (63) | 34/47 (72) |
| Gastrointestinal | 85 | 17/38 (45) | 27/47 (57) |
| Upper respiratory symptoms | 85 | 12/38 (32) | 26/47 (55) |
| Chest pain | 85 | 10/38 (26) | 26/47 (55) |
| Dyspnea | 85 | 16/38 (42) | 19/47 (40) |
| Syncope | 85 | 6/38 (16) | 2/47 (4) |
| Palpitations | 85 | 3/38 (8) | 3/47 (6) |
| Laboratory values at presentation | |||
| Elevated troponin | 64 | 22/25 (88) | 37/39 (95) |
| Elevated C-reactive protein | 46 | 15/18 (83) | 23/28 (82) |
| Elevated erythrocyte sedimentation rate | 37 | 7/13 (54) | 18/24 (75) |
| Elevated creatinine | 81 | 9/36 (25) | 4/45 (9) |
| Elevated lactate | 51 | 10/25 (40) | 2/26 (8) |
| Elevated aspartate transaminase | 74 | 28/34 (82) | 26/40 (65) |
| Elevated brain natriuretic peptide | 22 | 9/12 (75) | 6/10 (60) |
| ID pathogen found | 64 | 16/32 (50) | 14/32 (44) |
| Initial echocardiogram | |||
| Mild | 85 | 11/38 (29) | 16/47 (34) |
| Moderate | 85 | 9/38 (24) | 14/47 (30) |
| Severe | 85 | 18/38 (47) | 17/47 (36) |
| Fulminant course | 82 | 24/37 (65) | 17/45 (38) |
| Mechanical support | 85 | 17/38 (45) | 6/47 (13) |
| OHT | 85 | 5/38 (13) | 0/47 (0) |
| Death | 85 | 6/38 (16) | 2/47 (4) |
| Median length of hospital stay (days) | 85 | 19 (5–244) | 8 (1–71) |
Median length of hospital stay for all patients surviving to discharge was 11 days (1 to 244 days). Patients with clinically significant arrhythmias had longer hospital stays than those without arrhythmias ( Table 1 ). Mechanical support was required in 23 patients (27%), 17 of whom had clinically significant arrhythmias. OHT was performed in 5 patients (6%) at a median time of 4 months (1 to 7 months). Death occurred in 8 patients (9%) at a median time of 1 month (5 hours to 9 months), 7 during initial hospitalization. There was no arrhythmia-related mortality.
On initial echocardiogram, left ventricular function was mildly depressed in 27 (32%), moderately depressed in 23 (27%), and severely depressed in 35 (41%). A pericardial effusion was seen in 25 patients (29%).
An endomyocardial biopsy was performed in 44 patients (52%), and of these, 30 (68%) had active lymphocytic myocarditis, 6 (14%) had borderline myocarditis, and 8 (18%) had no evidence of myocarditis. A cardiac magnetic resonance imaging was performed in 23 patients (27%), 15 (65%) of whom demonstrated delayed enhancement consistent with myocarditis. Infectious disease workup was completed in 64 patients (75%), with a pathogen identified in 30 (47%).
Initial laboratory values at presentation are listed in Table 1 . Of 81 patients with ECGs available for review, 20 (25%) had low voltages, 26 (33%) had ST changes (excluding 2 patients with unresolved CHB), and 7 (9%) had an LBBB pattern. A total of 38 patients (45%) had 54 clinically significant arrhythmias occur either at presentation (acute) or during their hospitalization (subacute). Sixteen patients had acute arrhythmias only, 12 subacute only, 9 acute and subacute, and 1 had unknown onset ( Figure 1 ). Clinically significant arrhythmias included 9 SVT (2 atrial flutter, 2 ectopic atrial tachycardia, and 5 not specified), 2 high-grade ventricular ectopy, 6 nonsustained VT, 21 sustained VT, 3 VF, 1 Mobitz type II AV block, and 12 CHB.
Compared with those patients without arrhythmias, patients with clinically significant arrhythmias were statistically significantly more likely to have low voltages on ECG (14 of 34, 41% vs 6 of 47, 13%; odds ratio [OR] 4.8, 95% confidence interval [CI] 1.6 to 14.3) and have a worse outcome (20 of 38, 53% vs 6 of 47, 13%; OR 7.6, 95% CI 2.6 to 22.1). After adjusting for multiple comparisons, clinically significant arrhythmias were not associated with a fulminant course, initial myocardial function, initial troponin I level, elevated initial C-reactive protein and/or erythrocyte sedimentation rate, elevated initial creatinine level, elevated initial lactate level, or ST changes on ECG (see Table 2 ). Nevertheless, we did observe greatly increased odds (OR >2.9) of arrhythmias associated with a fulminant course, ST changes, and elevated initial lactate and creatinine levels.
| Primary Analysis | n | Arrhythmia (%) | No Arrhythmia (%) | p | OR | 95% CI |
|---|---|---|---|---|---|---|
| Arrhythmia | ||||||
| Fulminant | 82 | 24/37 (65) | 17/45 (38) | 0.0160 | 3.0 | 1.2–7.5 |
| Any ST change | 79 | 15/32 (47) | 11/47 (23) | 0.0319 | 2.9 | 1.1–7.6 |
| Any low voltage | 81 | 14/34 (41) | 6/47 (13) | 0.0051 ∗ | 4.8 | 1.6–14.3 |
| Elevated lactate | 51 | 10/25 (40) | 2/26 (8) | 0.0135 | 8.0 | 1.5–41.6 |
| Elevated creatinine | 81 | 9/36 (25) | 4/45 (9) | 0.0102 | 4.5 | 1.4–14.3 |
| Elevated CRP/ESR | 53 | 17/21 (81) | 27/32 (84) | 0.7458 | 0.8 | 0.2–3.4 |
| Troponin I | 49 | 4.4 (0.1–50.0) | 2.0 (0.1–46.4) | 0.7219 | 1.0 | 0.96–1.1 |
| Myocardial function | 0.5783 | |||||
| Mild vs moderate | — | — | 0.9 | 0.3–2.9 | ||
| Mild vs severe | — | — | 1.5 | 0.6–4.2 | ||
| Worse outcome | 85 | 20/38 (53) | 6/47 (13) | 0.0002 ∗ | 7.6 | 2.6–22.1 |
Ventricular arrhythmias were the most common rhythm disturbance, occurring in 30 (79%) of 38 patients with arrhythmias and 35% of the entire patient cohort. Acute ventricular arrhythmias included high-grade ventricular ectopy (1) and sustained VT (15). Subacute ventricular arrhythmias included high-grade ventricular ectopy (1), nonsustained VT (6), sustained VT (6), and VF (3). Subacute ventricular arrhythmias developed at a median time of 2 days (range, 1 to 19 days) after admission. Median time to complete resolution of all ventricular arrhythmias, including those for which antiarrhythmic therapy was used, was 6 days (1 to 27). No patient received an implantable cardioverter-defibrillator.
CHB was seen at presentation in 11 patients and developed during hospitalization in 1 patient on hospital day 6. Among the 11 patients who presented with CHB, 9 resolved at a median time of 5 days (range, 0 to 23) and 2 died without resolution of CHB. The 1 patient who developed subacute CHB never resolved and was discharged home with a pacemaker. Temporary transvenous pacing wires were placed in 9 patients for a median time of 6 days (range, 3 to 19). Permanent pacemakers were placed in 2 patients during hospitalization. One patient presented in CHB and a pacemaker was placed after 17 days. Normal conduction returned 6 days later, and the patient was discharged in sinus rhythm and then lost to follow-up. The second patient developed CHB on hospital day 4. A pacemaker was placed 31 days later and the patient remains in heart block after 11 months of follow-up.
Electrocardiographic predictors for the development of subacute arrhythmias were explored by evaluating the presence of ST changes and/or low voltages before new arrhythmia development. Compared with patients who did not develop new arrhythmias, patients with subacute arrhythmias were more likely to demonstrate ST elevation or depression before subacute arrhythmia onset (10 of 15, 67% vs 15 of 59, 25%; OR 5.9, 95% CI 1.7 to 19.9) but not low voltages ( Table 3 ).
