Demographic characteristics
During the study period, 32 episodes of IE were identified (demographic and clinical characteristics shown in Table 1); considering the Duke modified criteria, 28 (87.5%) were definite IE and 4 (12.5%) possible IE.
Table 1. Demographic and clinical characteristics of the 32 episodes of infective endocarditis (IE).
|
IE episodes
(n=32)
|
Gender (male), n (%)
|
17 (53.1)
|
Age, median (IQR), months
|
9.1 (4.2-129.8)
|
Underlying condition
|
|
Congenital heart disease, n (%)
|
28 (87.5)
|
Prematurity, n (%)
|
8 (25.0)
|
Immunocompromised, n (%)
|
1 (3.1)
|
Other chronic conditions, n (%)
|
6 (18.8)
|
>1 underlying condition, n (%)
|
11 (34.4)
|
Central venous catheter, n (%)
|
20 (62.5)
|
Fever, n (%)
|
27 (84.4)
|
Shock, n (%)
|
16 (50.0)
|
Heart failure, n (%)
|
1 (3.1)
|
Vascular phenomena, n (%)
|
14 (43.8)
|
Septic embolisms, n (%)
|
11 (34.4)
|
Intracranial hemorrhage, n (%)
|
2 (6.3)
|
Janeway’s lesions, n (%)
|
1 (3.1)
|
Immunological phenomena, n (%)
|
1 (3.1)
|
Echocardiogram
|
|
Positive echocardiogram, n (%)
|
27 (84.4)
|
Left IE, n (%)
|
6 (18.8)
|
Right IE, n (%)
|
21 (65.6)
|
Seventeen (53.1%) episodes were in male patients. The median age at diagnosis was 9.1 (4.2-129.8) months; 17 (53.1%) episodes were diagnosed in infants younger than 1 year of age, 3 (9.4%) in patients between 1-5 years, 4 (12.5%) in patients between 6-10 years and 8 (25.0%) in patients between 11-15 years. There were two incidence peaks in patients younger than 1 year and those older than 10 years of age.
Regarding underlying conditions, 28 (87.5%) patients had CHD (Table 2), 8 (25.0%) were preterm infants (median gestational age at birth: 27.5 weeks, IQR: 25.0-32.8 weeks), 1 (3.1%) was immunocompromised (metastatic medulloblastoma) and 6 (18.8%) had other chronic conditions (3 had a polymalformative syndrome, 1 short bowel syndrome, 1 esophageal atresia, and 1 Down syndrome); in 11 (34.4%) episodes, more than one underlying condition was associated (6 episodes: CHD plus preterm birth, 5 episodes: CHD plus other chronic conditions).
Table 2. Congenital heart disease (CHD) diagnosis.
|
CHD (n=28)
|
Cyanotic heart defects, n (%)
|
14 (50.0)
|
Pulmonary atresia, n (%)
|
5 (17.9)
|
Hypoplastic left heart syndrome, n (%)
|
4 (14.3)
|
Double outlet right ventricle, n (%)
|
3 (10.7)
|
Tetralogy of Fallot, n (%)
|
1 (3.6)
|
Truncus arteriosus, n (%)
|
1 (3.6)
|
Acyanotic heart defects, n (%)
|
14 (50.0)
|
Patent ductus arteriosus, n (%)
|
3 (10.7)
|
Ventricular septal defect, n (%)
|
3 (10.7)
|
Atrial septal defect, n (%)
|
2 (7.1)
|
Aortic stenosis, n (%)
|
2 (7.1)
|
Bicuspid aortic valve, n (%)
|
2 (7.1)
|
Atrioventricular canal, n (%)
|
1 (3.6)
|
Persistent valve of coronary sinus, n (%)
|
1 (3.6)
|
Regarding the 28 episodes of IE diagnosed in patients with CHD, a corrective cardiac surgery or therapeutic cardiac catheterization was previously performed in 20 (71.4%). The median time of IE diagnosis after the cardiac procedure was 47.5 (18.3-348.0) days. At diagnosis, in 20 (62.5%) episodes the patients had an indwelling CVC, and the median time of IE diagnosis after CVC canalization was 8.5 (4.0-14.3) days. Eight out of 8 (100%) preterm infants, 1 out of 1 (100%) immunocompromised patient and 5 out of 6 (83.3%) patients with other chronic conditions had an indwelling CVC at diagnosis (i.e., 14 out of 15, 93.3%), compared to 6 out of 17 (35.3%) patients with CHD and no other associated condition (p<0.001).
Clinical manifestations
Presenting signs were fever in 27 (84.4%) episodes, shock in 16 (50.0%), heart failure in 1 (3.1%), vascular phenomena in 14 (43.8%) (11 (34.4%) septic embolisms, 2 (6.3%) intracranial hemorrhage, 1 (3.1%) Janeway’s lesions), immunological phenomena in 1 (3.1%) (glomerulonephritis and rheumatoid factor elevation in the same episode). The presenting signs in preterm infants compared to other patients were different: fever was only present in 4 out of 8 (50.0%) compared to 23 out of 24 (95.8%) of the non-preterm patients (p<0.001) and shock was present in 7 out of 8 (87.5%) compared to 9 out of 24 (37.5%) of the non-preterm patients (p<0.001).
Echocardiogram
An echocardiogram was performed in all episodes, transthoracic echocardiography was used in 29 (90.6%) episodes and both transthoracic and transesophageal echocardiography was used in 3 (9.4%). There were echocardiographic findings in 27 (84.4%), of which 21 (65.6%) were right IE and 6 (18.8%) left IE.
Findings were located at: native tricuspid valve in 10 (31.3%), prosthetic pulmonary valve in 3 (9.4%), native aortic valve in 3 (9.4%), entrance of the venae cavae in 3 (9.4%), prosthetic pulmonary conduit in 2 (6.3%), interventricular septum in 2 (6.3%; 1 right ventricle, 1 left ventricle), interatrial septum in 1 (3.1%, right atrium), ventricular septal defect in 1 (3.1%), origin of left carotid artery in 1 (3.1%), and right chambers in 1 (3.1%). 5 (15.6%) episodes were prosthetic valve IE.
Microbiology
Thirty-six microbiological isolates were obtained in the 32 episodes (Table 3).
Table 3. Isolated microorganisms of the 32 episodes of infective endocarditis.
|
Isolated microorganisms
(n=36a)
|
Gram-positive, n (%)
|
20 (55.6)
|
Coagulase-negative, n (%)
|
7 (19.4)
|
Staphylococcus aureus, n (%)
|
6 (16.7)
|
Enterococcus faecalis, n (%)
|
4 (11.1)
|
Viridans streptococci, n (%)
|
3 (8.3)
|
Gram-negative, n (%)
|
11 (30.6)
|
Non-fermenters, n (%)
|
5 (13.9)
|
Enterobacteriaceae, n (%)
|
5 (13.9)
|
HACEK, n (%)
|
1 (2.8)
|
Fungi, n (%)
|
4 (11.1)
|
Coxiella burnetii, n (%)
|
1 (2.8)
|
a 4 episodes with 2 isolates.
Blood cultures were positive in 30 (93.8%) episodes; in 4 (12.5%) episodes, more than one microorganism was isolated (2 microorganisms isolated per episode). In those 2 episodes without positive blood cultures, one Trichosporon inkin was isolated at surgical specimen and serology for Coxiella burnetii was positive (confirmed by PCR in surgical specimen and blood) in another case.
Of those 36 microbiological isolates, 20 (55.6%) were Gram-positive bacteria (GPB), 11 (30.6%) were GNB, 4 (11.1%) were fungi, and 1 (2.8%) was Coxiella burnetii.
Regarding the 20 GPB isolates: 7 (19.4%) were CoNS, 6 (16.7%) Staphylococcus aureus, 4 (11.1%) Enterococcus faecalis, and 3 (8.3%) viridans streptococci. All S. aureus isolates were methicillin-sensitive and all E. faecalis isolates were ampicillin-sensitive.
Regarding the 11 GNB isolates: 10 (27.8%) were non-HACEK GNB and 1 (2.8%) was a HACEK-group GNB (Aggregatibacter aphrophilus). Of those 10 non-HACEK GNB, 5 (13.9%) were non-fermenting GNB: 4 (11.1%) Pseudomonas aeruginosa and 1 (2.8%) Stenotrophomonas maltophilia, and 5 (13.9%) were Enterobacteriaceae: 2 (5.6%) Escherichia coli, 1 (2.8%) Klebsiella pneumoniae, 1 (2.8%) Enterobacter cloacae and 1 (2.8%) Serratia marcescens. Among those 10 non-HACEK GNB, 5 (13.9%) were MDRB: 1 MDR Pseudomonas aeruginosa, 1 extended spectrum beta-lactamase (ESBL) Escherichia coli, 1 carbapenemase-producing Klebsiella pneumoniae, and 2 AmpC beta-lactamase producing GNB (Enterobacter cloacae and Serratia marcescens).
Regarding the 4 fungi isolates, those were: 2 (5.6%) Candida albicans and 2 (5.6%) Trichosporon inkin.
In 10 (31.3%) episodes, patients were previously colonized by MDRB (clinical characteristics of those patients and comparison with non-colonized patients in Table 4).
Table 4. Clinical characteristics of multidrug resistant bacteria (MDRB) colonized patients and comparison with non-colonized patients.
|
MDRB colonized patients (n=10)
|
Non-colonized patients (n=22)
|
p
|
Gender (male), n (%)
|
3 (30.0)
|
14 (63.6)
|
0.077
|
Age, median (IQR), months
|
6.5 (5.0-28.1)
|
43.9 (1.7-157.2)
|
0.393
|
Underlying condition
|
|
|
|
Congenital heart disease, n (%)
|
9 (90.0)
|
19 (86.4)
|
0.773
|
Prematurity, n (%)
|
2 (25.0)
|
6 (27.3)
|
0.660
|
Immunocompromised, n (%)
|
1 (10.0)
|
0 (0)
|
0.132
|
Other chronic conditions, n (%)
|
2 (20.0)
|
4 (18.2)
|
0.903
|
>1 underlying condition, n (%)
|
4 (40.0)
|
7 (31.8)
|
0.652
|
Central venous catheter, n (%)
|
9 (90.0)
|
11 (50.0)
|
0.030*
|
Time after CVC, median (IQR), days
|
4.0 (1.5-15.0)
|
9.0 (6.0-15.0)
|
0.334
|
Intensive care unit admission, n (%)
|
9 (90.0)
|
17 (77.3)
|
0.393
|
Time of PICU/NICU stay, median (IQR), days
|
171.0 (57.0-233.0)
|
23.0 (4.5-58.5)
|
0.005*
|
MDRB IE, n (%)
|
3 (30.0)
|
0 (0)
|
0.007*
|
MDRB colonization was associated with MDRB EI (p=0.007), with the presence of indwelling CVC (p=0.030), and with longer time of PICU/NICU stay (p=0.005). Of those 10 patients, the colonizing bacteria were isolated in 3 (30.0%) of the patients’ blood cultures: 1 MDR Pseudomonas aeruginosa, 1 ESBL Escherichia coli, and 1 AmpC beta-lactamase producing Enterobacter cloacae.
Complications
PICU/NICU admission related to IE or its complications was required in 26 (81.3%) episodes, with a median stay of 26.0 (6.5-87.3) days. Complications secondary to IE were present in 25 (78.1%) episodes, with 17 (53.1%) of the episodes more than one complication. Registered complications were septic embolisms in 11 (34.4%) (8 (25.0%) pulmonary, 1 (3.1%) pulmonary and soft tissue, 1 (3.1%) cerebral, 1 (3.1%) cerebral and hepatic), shock in 8 (25.0%), heart failure in 7 (21.9%), valve regurgitation in 7 (21.9%), acute renal failure in 4 (12.5%), neurological complications in 3 (9.4%), local extension of infection in 3 (9.4%), and heart rhythm disturbances in 1 (3.1%). In-hospital mortality occurred in 2 (6.3%) episodes but was not attributed to IE or its complications, but to complications related to the patients’ underlying condition (pulmonary atresia and polymalformative syndrome (CHARGE syndrome) respectively).
Outcomes
There were 6 (18.8%) recurrences: 4 (12.5%) in-hospital relapses and 2 (6.3%) outpatient recurrences (1 relapse, 1 reinfection). At discharge, there were sequelae in 10 (31.3%) episodes, having 3 (9.4%) of the episodes more than one sequela: valvular regurgitation in 7 (21.9%) (2 (6.3%) of them associating heart failure), neurological sequelae in 3 (9.4%), heart rhythm disturbances in 1 (3.1%), valvular double lesion in 1 (3.1%), and chronic renal failure in 1 (3.1%). At follow-up (median time 3.4 years, IQR 0.7-7.2 years), 6 (18.8%) of those patients had persistent sequelae (1 (3.1%) with more than one sequela): valvular regurgitation in 4 (12.5%), heart failure in 2 (6.3%), neurological sequelae in 1 (3.1%), heart rhythm disturbances in 1 (3.1%), and chronic renal failure in 1 (3.1%).
Comparison of clinical characteristics and complications depending on causative microorganism of IE
Clinical characteristics and complications of IE episodes caused by non-HACEK GNB were compared to those episodes caused by GPB, and no statistically significant differences were found between them (Table 5 and Table 6).
Table 5. Comparison between clinical characteristics and isolated microorganism.
|
Bacterial IEa
(n=26)
|
Gram-positive bacteria (n=17)
|
Gram-negative bacteria (n=9)
|
p
|
Gender (male)
|
14 (53.8)
|
9 (52.9)
|
5 (55.6)
|
0.899
|
Age, median (IQR), months
|
6.5 (3.5-67.9)
|
18.0 (2.8-153.4)
|
5.9 (3.5-37.4)
|
0.467
|
Underlying condition
|
|
|
|
|
Congenital heart disease, n (%)
|
23 (88.5)
|
15 (88.2)
|
8 (88.9)
|
0.960
|
Prematurity, n (%)
|
7 (26.9)
|
4 (23.5)
|
3 (33.3)
|
0.592
|
Immunocompromised, n (%)
|
1 (3.8)
|
1 (5.9)
|
0 (0.0)
|
0.458
|
Other chronic conditions, n (%)
|
5 (19.2)
|
2 (11.8)
|
3 (33.3)
|
0.184
|
>1 underlying condition, n (%)
|
10 (38.5)
|
5 (29.4)
|
5 (55.6)
|
0.192
|
Central venous catheter, n (%)
|
18 (69.2)
|
10 (58.8)
|
8 (88.9)
|
0.114
|
Fever, n (%)
|
22 (84.6)
|
14 (82.4)
|
8 (88.9)
|
0.660
|
Shock, n (%)
|
15 (57.7)
|
8 (47.1)
|
7 (77.8)
|
0.131
|
Vascular phenomena, n (%)
|
9 (34.6)
|
6 (35.3)
|
3 (33.3)
|
0.920
|
Immunological phenomena, n (%)
|
1 (3.8)
|
1 (5.9)
|
0 (0.0)
|
0.458
|
Echocardiogram
|
|
|
|
|
Positive echocardiogram, n (%)
|
22 (84.6)
|
16 (94.1)
|
6 (66.7)
|
0.065
|
Left IE, n (%)
|
5 (19.2)
|
4 (23.5)
|
1 (11.1)
|
0.169
|
Right IE, n (%)
|
17 (65.4)
|
12 (70.6)
|
5 (55.6)
|
0.169
|
a Excluding Coxiella burnetii EI and Aggregatibacter aphrophilus IE.
Table 6. Comparison between complications and isolated microorganism.
|
Bacterial IEa
(n=26)
|
Gram-positive bacteria (n=17)
|
Gram-negative bacteria (n=9)
|
p
|
PICU admission, n (%)
|
21 (80.8)
|
12 (70.6)
|
9 (100)
|
0.070
|
Surgical management, n (%)
|
7 (26.9)
|
5 (29.4)
|
2 (22.2)
|
0.694
|
Septic embolisms, n (%)
|
7 (26.9)
|
4 (23.5)
|
3 (33.3)
|
0.592
|
Local extension of infection, n (%)
|
1 (3.8)
|
1 (5.9)
|
0 (0.0)
|
0.458
|
Heart failure, n (%)
|
5 (19.2)
|
2 (11.8)
|
3 (33.3)
|
0.184
|
Residual valve regurgitation, n (%)
|
5 (19.2)
|
5 (29.4)
|
0 (0.0)
|
0.070
|
Heart rhythm disturbances, n (%)
|
1 (3.8)
|
1 (5.9)
|
0 (0.0)
|
0.458
|
Neurological complications, n (%)
|
3 (11.5)
|
1 (5.9)
|
2 (22.2)
|
0.215
|
Acute renal failure, n (%)
|
3 (11.5)
|
3 (17.6)
|
0 (0.0)
|
0.180
|
Recurrences, n (%)
|
4 (15.4)
|
2 (11.8)
|
2 (22.2)
|
0.482
|
Mortality, n (%)
|
2 (7.7)
|
1 (5.9)
|
1 (11.1)
|
0.634
|
a Excluding Coxiella burnetii EI and Aggregatibacter aphrophilus IE.
Although children with non-HACEK GNB IE exhibited a higher proportion of PICU admission (100% vs. 70.6%), shock (77.8% vs. 47.1%), heart failure (33.3% vs. 11.8%), neurological complications (22.2% vs. 5.9%), more than one underlying condition (55.6% vs. 29.4%), and presence of indwelling CVC (88.9% vs. 58.8%), these distinctions did not reach statistical significance. On the other hand, children with GPB IE showed a trend towards higher rate of residual valve regurgitation (29.4% vs 0.0%).