Results:
Out of a total of 42,570,716 weighted pediatric admissions, 4165 were
among patients ≤20 years of age with ICDs at the time of admission.
Pediatric ICD admissions increased significantly from 2000 to 2016
(2.9/100,000 in 2000 to 16/100,000 in 2016, p = 0.002), with a steeper
increase from 2000-2009 (Figure 1). Admissions by underlying diagnoses
varied over time (Figure 2). Overall, cardiomyopathy was the most common
underlying diagnosis throughout all time periods. ICD admission with an
underlying diagnosis of primary arrhythmia decreased from 25% in 2006
to 16% of all admissions in 2016 (p = 0.06) (Figure 2). ICD admissions
with heart failure diagnosis increased significantly over time
accounting for 16% in 2000 and 33% in 2016 (p = 0.029, Figure 3).
Among patients with ICDs, there were a total of 54 in-hospital deaths
(1.3%) at a median age of 18 years (IQR 15-20 years) while 4111
(98.7%) survived to discharge (median age 17 yrs., IQR 14-19 yrs). The
total percentage of deaths among ICD hospitalizations increased,
starting with 0.7% in 2000. With the exception of a small rise in 2009
(2.2%), overall rates of deaths remained approximately 1.0-1.5% of ICD
admissions (Supplementary Figure 1).
The baseline characteristics of the pediatric ICD admissions overall and
by mortality status are shown in Table 1. The majority of the admissions
were between the ages of 13-20 years admitted to urban teaching
hospitals. Hospitalization rates were relatively similar by region and
by season. A majority (80%) were not elective admissions. An ICD
complication was present in 15% of the admissions. Cardiac procedures
or surgeries were fairly common; 30% of patients that died had some
form of cardiac procedure or surgery compared to 20% of those that were
discharged.
On univariate analysis, in-hospital death among admissions with ICDs was
significantly associated with male sex, non-Hispanic Black race,
underlying diagnosis of cardiomyopathy or congenital heart disease and
presence of heart failure (Table 2). Age group and insurance type were
not associated with in-hospital death. On initial multivariable analysis
(excluding heart failure from our model), an underlying diagnosis of
cardiomyopathy (OR 3.5, 95% CI 1.1 – 11, p = 0.04) and CHD (OR 4.8,
95% CI 1.5 – 15.6, p = 0.01) were significantly associated with
mortality, with a trend seen in non-Hispanic Black race (OR 1.9, 95% CI
0.99 – 3.8, p = 0.05) (Table 3). Notably, non-Hispanic Black patients
with ICDs more frequently had a diagnosis of heart failure or
cardiomyopathy (Table 4). In addition, while overall mortality was
1.3%, mortality among non-Hispanic Black race was 2.3%. Among all ICD
admissions and a concomitant diagnosis of heart failure, mortality was
3.7%. To assess if the relationship of race and heart failure, an
interaction term was included in the analysis and did not show an
interaction between race and heart failure diagnosis in this cohort,
suggesting that the relationship between non-Hispanic Black race and
death was not mediated by the presence of heart failure. Nevertheless,
because heart failure 1) has been shown in prior studies to be higher
among Black race and 2) is likely to be higher among cardiomyopathy
patients, yet 3) may be on the causal pathway, an exploratory analysis
was performed by including heart failure diagnosis in the multivariable
model. On exploratory multivariable analysis including heart failure,
only the presence of heart failure (OR 8.6, 95% CI 3.7-19.6,
p<0.0001) remained significantly associated with in-hospital
mortality (Table 5).