CARDIOVASCULAR JOURNAL OF AFRICA • Volume 25, No 2, March/April 2014
AFRICA
51
using both clinical features and the level of CD
4
count. A
broad description of the echocardiographic features of IE was
documented for the study population as a whole. The clinical
features, laboratory results and echocardiographic findings were
compared in the HIV-infected and uninfected patients.
Patients with a clinical diagnosis of suspected IE had an
initial examination by transthoracic echocardiography (TTE)
to look for echocardiographic evidence of infection, document
haemodynamic status (valvular dysfunction, ventricular
dimensions and ventricular function) and to determine the
presence of predisposing conditions such as congenital, valvular
or degenerative heart disease. Echocardiographic evidence of
valve infection was accepted when any one of the following
findings was identified: vegetation, paravalvular extension
with abscess formation, or rupture and fistula formation,
and prosthetic valve dehiscence. Leaflet or cuspal thickening,
with/without areas of calcification found at TTE was taken as
evidence of underlying rheumatic heart disease.
Transoesophageal echocardiography (TEE) was performed
within 24 to 48 hours of admission if paravalvular extension was
suspected, or where TTE images were suboptimal, and in the
case of mechanical prosthetic valves. It was performed in most
patients except where patients were rushed to emergency surgery
on the basis of the transthoracic echo findings. Echocardiography
was performed on a Sequoia C256 (Acuson, Germany) cardiac
ultrasound machine, using a 5-MHz transthoracic transducer for
TTE, and a 7-MHz multiplane transoesophageal probe for TEE.
Statistical analysis
Baseline characteristics of all patients were evaluated to describe
the demographics and to identify any underlying risk factors.
Comparisons between HIV-infected and HIV-negative patients for
categorical outcomes (e.g. valvular assessment) were evaluated
by means of chi-square tests or Fischer’s exact tests. Where
the outcome was numerical (e.g. CD
4
count), Mann–Whitney
tests were used to compare mean ranks in the HIV-positive and
HIV-negative groups. The Student’s
t
-test was used to determine
differences between samples. The significance level was taken
at
p
<
0.05.
Results
Of the 91 patients screened for suspected endocarditis, 63
satisfied the diagnosis of definite IE by the Duke and 78 by the
modified Duke criteria. According to the modified Duke criteria,
77 patients were classified as definite IE (HIV infected,
n
=
17,
uninfected,
n
=
60), nine as possible IE, and five as rejected IE.
The analysis that follows was performed on the 77 patients with
definite IE (Table 1).
The mean participant age in the whole group was
approximately 30 years, with a slight male preponderance
(55%). Overall, there was a slight male predominance for the
occurrence of IE in both groups of patients: 55% (
n
=
43) were
male and 45.5% (
n
=
35) were female (Table 2). There were
no significant differences in age, admission weight (61 vs 59
kg), and temperature (36.5 vs 37°C) between the HIV-infected
and uninfected groups. Fever above 38°C was noted in four
HIV-infected patients. The source of infection was not apparent
in most patients, nor was there any information from the history
about a childhood history of rheumatic fever or valvular heart
disease.
Other clinical features of infective endocarditis did not
appear to be different in the two groups. Among the 17 HIV-
infected patients, 11 had clubbing and five had heart failure.
Hepatomegaly mirrored the findings of congestive heart failure
and was found in five (29.4%) HIV-infected patients, and in 28
(46.7%) HIV-negative patients (
p
=
0.024) (Table 2).
Thirty-six of the 78 patients with definite IE (46%) had positive
blood cultures; 29 (37%) of these were from the HIV-negative
patients, and seven (9%) were from the HIV-infected patients.
S
aureus
was the commonest infecting bacterium in both groups
of patients, and was found overall in 17 (47%) of those with
positive cultures. Of these, four were HIV-infected and 13 were
HIV negative (
p
=
ns). The second most common infecting
bacterium was
S viridans
(
n
=
7) (20%); of these one was in the
HIV-infected group. One HIV-infected patient had an unusual
organism; this was
Propionibacterium
.
Significantly, higher elevation in the sedimentation rate and
the C-reactive protein levels was noted in the HIV-infected group
compared to the HIV-negative group. In addition, serum albumin
level was markedly lower in the HIV-infected group (Table 3).
Echocardiographic findings
All but one patient had findings suggestive of IE on TTE. These
included the presence of vegetations, root abscesses or leaflet
aneurysms (Table 4). Apart from these changes, 60 patients
(78%) had leaflet thickening with/without reduced mobility or
calcium on the valve apparatus that was suggestive of rheumatic
heart disease; of these 51 were HIV negative, and nine were HIV
Table 1. Duke criteria versus modified Duke criteria in the
classifications of infective endocarditis
Duke criteria
Modified Duke criteria
HIV+
n
=
18 (%)
HIV–
n
=
73 (%)
Total
91
HIV+
n
=
18 (%)
HIV–
n
=
73 (%)
Total
91
Definite 16 (88.9) 47 (64.4) 63 17 (94.4) 60 (82.2) 77
Possible 2 (11.1) 21 (28.8) 23 1 (5.6)
8 (10.9)
9
Rejected 0 (0)
5 (6.8)
5 0 (0)
5 (6.8)
5
Table 2. Demographic data and clinical features in HIV-positive
and HIV-negative patients with infective endocarditis
Parameter
HIV+
n
=
17 (%)
HIV–
n
=
60 (%)
Total
n
=
78 (%)
p
-value
Age (years)
32 (22–50)* 31(12–64)* 63(80.8)
0.867
Gender: male
9 (53)
33 (55)
43(55.1)
1.000
female
8 (47)
27 (45)
35( 44.9)
Body weight (kg)
61 (41–82)* 59 (43–79)* 120 (153.8) 0.585
Fever
4 (23.5)
3 (5)
7 (9.1)
0.024
Clubbing
11 (64.7)
32 (53.3)
43 (55.8) 0.102
Splinter haemorrhages 2 (11.8)
3 (5)
5 (6.5)
0.304
Emboli/stroke
3 (17.6)
6 (10)
9 (11.7)
1.000
Splenomegaly
2 (11.8)
3 (5)
5 (6.5)
Heart failure/
hepatomegaly
5 (29.4)
28 (47)
33 (43)
0.204
Haematuria
3 (17.6)
19 (31.7)
22 (28)
*Mean values with the ranges bracketed.
