Emergency Department Management of Pediatric Patients With Cyanotic

Heart Disease and Fever

Abstract and Introduction

Abstract
Background: Children with cyanotic congenital heart disease (CCHD) are living
longer and presenting to the Emergency Department (ED) in larger numbers. A
greater understanding of their diagnoses and appropriate management strategies
can improve outcomes.
Objective: Our objective was to describe the ED diagnoses, management, and
dispositions of pediatric CCHD patients who present with fever.
Methods: We retrospectively analyzed pediatric ED patients age 18 years or
younger with a previous diagnosis of CCHD who presented with a fever from
January 2000 to December 2005.
Results: Of 809 total ED encounters, 248 (30.6%) were eligible for inclusion. Of
those meeting inclusion criteria, 59 (23.8%) required supplemental oxygen and
67 (27%) received intravenous fluid. ED diagnoses were febrile illness in 120
(48.4%), pneumonia in 35 (14.1%), upper respiratory infection in 19 (7.7%), viral
syndrome in 17 (6.9%), gastroenteritis in 17 (6.9%), otitis media in 10 (4.0%),
bronchiolitis in 5 (2.0%), pharyngitis in 3 (1.2%), croup in 3 (1.2%), bronchitis in
3 (1.2%), urinary tract infection in 3 (1.2%), mononucleosis in 2 (0.8%),
pericarditis in 2 (0.8%), influenza in 1 (0.4%), cellulitis in 1 (0.4%), bacteremia
in 1 (0.4%), and potential endocarditis in 1 (0.4%). In terms of patient
disposition, 53.2% were discharged, 44.4% were floor admissions, and 2.4%
were intensive care unit admissions.
Conclusions: A cardiac cause of fever in CCHD patients is rare. Because of
limited cardiopulmonary reserve, they might require supplemental oxygen,
intravenous fluids, and hospital admission.
Introduction
The mean incidence of cyanotic congenital heart disease (CCHD) is
approximately 1.4 cases per 1000 live births, which has changed little in the last
50 years.[1,2] Since that time, advances in medical and surgical treatment have
reduced the morbidity and mortality associated with CCHD significantly and, as a
result, approximately 85% of children with cyanotic lesions now live into
adulthood (age 21 years or older).[3] As recently as the 1960s, only 6.5% of
children born with cyanotic lesions experienced similar longevity. [2]
Consequently, there are now more children with CCHD presenting to the
Emergency Department (ED) than ever before, and these numbers will continue
to grow.
CCHD is the primary risk factor for infective endocarditis (IE) in children, and is
responsible for approximately 74% to 85% of pediatric cases.[4,5] Unfortunately,
the diagnosis in children can be especially challenging because they tend to
present with vague, nonspecific symptoms (e.g., fatigue, weakness, weight loss,
anorexia), and rarely exhibit the classic physical examination findings seen in
adults (i.e., Roth spots, Janeway lesions, Osler nodes, splinter hemorrhages, and
petechiae). Clinicians should be particularly suspicious when these children
present with a persistent low-grade fever of unknown etiology. [6–8] Although IE
should always be considered in the differential diagnosis of children with CCHD
who present to the ED with a fever, anecdotal evidence suggests that the most
common diagnoses encountered in these children mirror the general pediatric
population. More aggressive use of fluids and supplemental oxygen might be
required, however, due to their more tenuous cardiopulmonary physiology. Early
and aggressive treatment, even for routine childhood infections, is important in
these patients and should be directed toward optimizing volume status, gas
exchange, and the underlying cause of the fever.
Our primary study objective is to describe the most common diagnoses,
management strategies, and dispositions of pediatric patients with known CCHD
who present to the ED with a fever.

Methods
We performed a retrospective data analysis of ED medical records for patients
who presented to an urban, tertiary care pediatric hospital ED (approximately
50,000 annual visits) between January 2000 and December 2005. It was approved
by the Institutional Review Board. Patients were eligible for inclusion in the
study if they were aged 18 years or younger, had a pre-existing diagnosis of
CCHD (e.g., pre-repair, at an intermediate operative stage, or post-repair), and
either had a temperature at triage ≥38°C (100.4°F) or a recent history of a
temperature ≥38°C (100.4°F) per patient report. Patients were excluded if they
were electively admitted to the hospital or if they had undergone a surgical
procedure within the previous 30 days.
CCHD cases were identified by searching the hospital's medical record system
using the following keywords: AV canal, HLHS, TAPVR, TGA, Tetralogy of
Fallot, Fontan, Norwood, and Blalock. Two trained abstractors recorded patient
demographics and data relevant to the following outcomes measures: final ED
diagnosis (International Classification of Diseases, 9th edn. diagnostic code), at
least one desaturation episode (≥5% below baseline) requiring supplemental
oxygen, intravenous fluid requirement, chest radiography performed and their
results, whether blood cultures were ordered, and disposition status (discharged
home, admitted to floor, or admitted to the intensive care unit).
Results
From our initial search criteria, 809 total ED encounters by CCHD patients were
identified. After additional screening, 248 (30.6%) met our inclusion criteria (see
Figure 1). Median age was 1.6 years (interquartile range: 0.7–4.0 years). Fifty-
nine (23.8%) had at least one desaturation episode (defined as ≥5% below
baseline) requiring supplemental oxygen, and 67 (27%) received intravenous
fluid treatment. Chest radiography was performed in 171 (68.9%) patients and an
infiltrate was identified on 38 (22.2%) of them and pneumonia diagnosed in 35
(14.1%). Blood cultures were ordered for 117 (47.1%). The final ED diagnoses
and dispositions are shown in Table 1 and Table 2 gives a description of the cases
of endocarditis.
The types of CCHD present in our cohort included the following: atrioventricular
(AV) canal defect, Tetralogy of Fallot, transposition of the great arteries, total
anomalous pulmonary venous return (TAPVR), tricuspid atresia (TA),
hypoplastic left heart syndrome (HLHS), single ventricle, pulmonary atresia,
combined TA/Ebstein's anomaly, combined HLHS/double outlet right ventricle
(DORV), combined AV canal defect/single ventricle, combined AV canal
defect/TAPVR/DORV, and combined AV canal defect/Tetralogy of Fallot

Discussion
The current literature addressing the topic of fevers in children with CCHD
focuses almost exclusively on endocarditis. The majority of endocarditis cases in
children are associated with congenital heart anomalies, with 2.6% to 18% [9,10] of
this population affected by IE in longitudinal studies. To our knowledge, no
previous study has ever described the most common diagnoses in a population of
children with known CCHD who present to the ED with a fever. Our study
supports anecdotal reports that children with CCHD who present to the ED with a
fever most likely have diagnoses similar to the general pediatric ED population
(Table 3).[11] The incidence of endocarditis (0.4%) in our study population is
consistent with our hypothesis that endocarditis is a rare diagnosis, even in this
high-risk group. Our findings are similar to estimates extrapolated from a
previous retrospective analysis of cardiology consultations performed at the
Children's Hospital Boston ED. Of 503 consultations involving ED patients with
a known cardiac history, only 12 (2.4%) consults were due to a concern for
endocarditis, and only 3 (0.6%) of these patients were diagnosed with
endocarditis[12] (Table 2).
Children with CCHD (whether pre-repair, at an intermediate operative stage, or
post-repair) are at an increased risk of suffering from respiratory distress and
hypovolemic complications from routine childhood infections. Congenital heart
disease can actually be a risk factor for recurrent pediatric pneumonias (at least
two episodes per year or at least three during a lifetime). [13] Specifically, patients
with lesions that increase pulmonary blood flow (PBF), such as transposition of
the great arteries, truncus arteriosus, TAPVR, DORV with subpulmonary
ventricular septal defect (Taussig-Bing anomaly), single ventricle (in the 50% of
cases without pulmonary stenosis), and AV canal defects are thought to be more
susceptible to pneumonia because the elevated PBF increases alveolar secretions,
which are believed to act as a nidus for infection. [14–16] In our study population,
14.1% of patients were found to have pneumonia, compared with 1.8% in the
general pediatric ED population,[17] which supports the theory CHD can be a
predisposing factor for pneumonia.[15]
Current evidence also indicates that young children with CHD have a prolonged
and more complicated hospital course from lower respiratory tract infections.
According to the PICNIC (Paediatric Investigators Collaborative Network on
Infections in Canada) study, of all children (younger than 2 years of age)
admitted from the ED with respiratory syncytial virus bronchiolitis, children with
CHD had a hospital stay that was 1.4 days (95% confidence interval 1.2–1.6; p <
0.05) longer than children without comorbidities.[18,19]
A study by Savitsky et al. selected a high-acuity subset of pediatric patients
(mean age 20 months) with congenital heart disease (74% had complex lesions)
and described their most common ED presentations and diagnoses.[20] Respiratory
tract infections (36% of patient visits) and gastroenteritis (4.5% of patient visits)
were the most common infectious diagnoses. There were no documented cases of
endocarditis and only one (2.3%) documented case of sepsis in their population.
The most common ED presentation patterns of their high-acuity subpopulation,
independent of diagnosis, were acute respiratory distress/cyanosis (22% of
patients) and dehydration/hypovolemia (20% of patients).[20] These findings
demonstrate the importance of closely monitoring respiratory and volume status
in children with CHD, and suggest that it can be beneficial to intervene early in
their course, before decompensation.
Hypoxia, which can result from respiratory tract infections, vasoconstricts the
pulmonary vasculature, increases pulmonary vascular resistance, and reduces
PBF. This can be particularly problematic in children with lesions that decrease
PBF (e.g., TOF, DORV [with pulmonary stenosis], single ventricle [in the 50%
of cases with pulmonary stenosis], Ebstein's anomaly, tricuspid atresia,
pulmonary atresia, and HLHS). The reduction of an already low PBF increases
the right to left shunt fraction, which increases the amount of deoxygenated blood
in the systemic circulation. This feedback loop progressively worsens the
patient's hypoxemia. The situation is exacerbated when body temperature is
elevated (such as during a fever) or in the setting of acidosis (metabolic or
respiratory can occur alone or in combination), which induces a rightward shift of
the hemoglobin–O2 dissociation curve, increasing the body's overall oxygen
consumption, and leading to clinical deterioration and a potentially fatal situation.
[21]
Supplemental oxygen is an important intervention because of its vasodilatory
action on the pulmonary vasculature, which increases PBF and markedly
improves respiratory function.
In order to provide the most appropriate management for patients with CCHD, it
is important to understand the anatomic and physiologic effects of current
operative interventions. Cyanotic lesions usually cause either elevations or
reductions in PBF and the initial (stage 1) palliative procedures are often aimed at
normalizing blood flow to the pulmonary vasculature before more permanent
palliation and repair. The surgical procedures for anomalies causing single-
ventricle physiology (e.g., tricuspid atresia, HLHS, single ventricle, complicated
DORV, Ebstein's anomaly, unbalanced AV canal defects, or pulmonary atresia
with intact ventricular septum) occur in three stages to allow sufficient time for
the pulmonary vascular resistance to decrease and provide optimal conditions for
successful outcomes. Patients with other types of CCHD can have an initial
palliative shunt placed to increase blood flow to the lungs but, subsequently, will
often have lesion-specific repairs.[14]
Limitations
The limitations include those inherent in a retrospective study design, such as the
inability to eliminate the effects of bias and confounding on our results. A
potential source of bias in this study was the subjective nature of a clinician's
history taking, physical examination, and diagnostic decision making.
Because only 171 of the 248 (68.9%) patients had chest radiography performed,
and only 117 of the 248 (47.1%) had blood cultures ordered, there might have
been additional cases of lower respiratory tract infections, bacteremia, or
potential IE that were missed. Consequently, our results might have
underestimated their actual incidence in a population of children with CCHD and
fever.
A more in-depth analysis of the attending notes for the cases diagnosed as
"febrile illness" revealed that many of these were likely viral illnesses or illnesses
of unknown etiology based on the available data at the time of disposition. There
is a chance that some of the febrile illness diagnoses were bacteremia or IE,
because they are less likely to be definitively diagnosed when a patient is still in
the ED. This might have contributed to an underestimation of the actual incidence
of bacteremia or IE in a population of CCHD patients presenting to the ED with a
fever.
Because only 35 cases of pneumonia were diagnosed, but 38 chest radiographs
showed an infiltrate, there might have been additional cases of pneumonia either
misdiagnosed or included in the febrile illness category.
By using specific search terms to identify potential patients with CCHD, we
might have inadvertently missed cases, especially charts that contained
incomplete or inaccurate documentation of a patient's medical history, resulting
in misclassification bias. The decision to include patients with a recent history of
a measured fever, who were afebrile at presentation, also represents a potential
source of human error and measurement bias.
A possible threat to the external validity of our study was the use of data from a
single tertiary care pediatric hospital. The facility under study has a high rate of
transfers from community hospitals and other tertiary care medical centers, which
would tend to increase the average acuity level and complexity of patients
presenting to the their ED. Therefore, it is possible that our population might
have been more likely to have diagnostic tests (such as chest radiographs or blood
cultures) ordered, to require treatment with intravenous fluid, to suffer from
decreased oxygen saturations, to be diagnosed with a serious illness, or to be
admitted to the hospital. However, because the majority of diagnoses in our
population still mirrored those in the general pediatric ED population, this might
actually provide stronger support for our hypothesis because we might have
inadvertently selected a population of children with more complex cyanotic
congenital heart disease than what is normally seen at the majority of EDs.
Conclusions
A source of fever is frequently identified in patients with CCHD, but a specific
cardiac cause is rare. In fact, the most common ED diagnoses of children with
CCHD who present with a fever are similar to the most common diagnoses in the
general pediatric ED population. However, because of limited cardiopulmonary
reserve, they are more vulnerable than healthy children to infections that alter gas
exchange or volume status. Consequently, these patients are often, but not
always, admitted to the hospital and can require supplemental oxygen and
intravenous fluid.