CARDIOVASCULAR JOURNAL OF AFRICA • Volume 25, No 2, March/April 2014
56
AFRICA
tertiary referral centre receiving patients who have already been
started on antibiotics. In febrile patients with elevated ESR,
clinicians at base hospitals (often far removed from laboratory
facilities) feel obliged to administer antibiotic therapy prior to
obtaining the results of initial investigations.
16
A deficiency in the modified Duke criteria becomes apparent
when diagnosing IE when blood cultures are negative; we have
shown that these patients often have elevated sedimentation rates
and C-reactive protein levels from repeated non-cardiac infection
and anaemia. Using the surgical findings on operated cases as a
gold standard, the positive predictive value of the modified Duke
criteria was only 72%.
The higher culture negativity in HIV-infected patients in our
study also raises the possibility of non-bacterial thrombotic
endocarditis (NBTE) in at least some of these cases.
17
Serial
negative blood cultures should alert the clinician to the possibility
of NBTE, which has been reported in HIV-infected patients.
16
These findings have significant implications for the prevalence
and diagnosis of IE in HIV-positive patients.
We found a similar rate of morbidity and mortality between
HIV-infected and HIV-negative patients, in keeping with the
data from Fowler
et al
., who found that overall morbidity and
mortality related to cardiac disease in AIDS was low.
18
There
were four deaths among our HIV-infected patients (23.6%), and
14 among the HIV-negative patients (23.3%). Three of the four
HIV-infected patients who died had CD
4
counts
<
100 /mm³. This
is in keeping with data from our centre showing that surgery in
HIV-infected patients with CD
4
counts of
>
400 /mm³ are likely
to have early surgical outcomes similar to that in HIV-uninfected
patients.
19
Study limitations
The small sample size of the study was a limiting factor. This
could have been due to the poor referral system from the base
hospital to our hospital, or in fact, that IE is not as common in
HIV-positive patients as we had presumed. Also not all patients
diagnosed with IE at TTE received a TEE. This was due to
various reasons, such as a patient’s inability to tolerate the TEE
probe and markedly elevated international normalised ratio
(INR) levels at the time of examination.
A further limitation in the study was the high rate of negative
blood cultures. This was most likely the result of administration
of antibiotics to the patient prior to referral to our institution, or
in the case of the HIV-positive patients, the possibility of NBTE.
Of the 91 patients initially screened, 77 were accepted as
having had a definite diagnosis of IE, according to the modified
Duke criteria. The remaining 14 were deemed not to have IE
and excluded from the analysis. Whether any patients in this
group had IE or not (true negative and false negative) could not
be determined with certainty since they were not subjected to
surgery.
We believe the modified Duke criteria were responsible
for the higher false-positive rates since it permits diagnosis of
IE based on the echocardiographic criteria in the absence of
positive blood cultures. While this reflected a potential flaw in
the study, since the diagnosis of IE was based on the finding of
vegetations in the absence of positive blood cultures, this study
highlights the difficulty in diagnosis when the blood cultures are
negative, placing more reliance on echocardiographic detection
of vegetations.
Unless supported by clinical features and bacteriological
evidence, vegetations alone are not diagnostic of IE because they
may represent healed infection. Furthermore, the diagnosis of IE
postoperatively is rendered more difficult by the now common
practice of leaving the chordal mechanisms intact. Five of the 33
patients clinically diagnosed as definite IE, and one possible IE,
had no evidence of infection at operation, supporting the need for
bacteriological confirmation of infection.
Not all patients however were referred for surgery. The low
CD
4
counts in the HIV-positive patients meant an even smaller
group of these patients were accepted for operation, as the
acceptable CD
4
level for surgery at our institution is CD
4
count
>
200 /mm³
In this study, attempts were made to more accurately define
valve pathology on echocardiography. According to Taams
et
al
., there are five distinct pathological features of IE that may
be seen clearly on TEE and these are (1) mitral stenosis with
vegetations; (2) myxomatous degeneration of leaflets with
vegetations; (3) chordal rupture with vegetations; (4) chordal
rupture without vegetations; and (5) mycotic aneurysms with
fistulous connections.
20
It is often difficult to decide on the
underlying pathology with this degree of accuracy with TTE.
In our study, harmonic imaging was employed to improve
the diagnostic value of TTE by improving the image quality,
as documented by Chirillo
et al
.
21
Harmonic imaging works
on the principle of limiting near-field artefacts, and because
the harmonic energy increases with the distance the ultrasound
wave propagates, most harmonics will result from the central
ultrasound beam rather than the weaker side lobe artefacts.
20
This modality is used primarily to enhance left ventricular
endocardial borders. Its use did not really increase the resolution
in visualising vegetations. Therefore, we usedTEE to differentiate
and define chordal rupture in association with vegetations, leaflet
prolapse and flail leaflets. Our surgical findings revealed that
even with TEE there were limitations, which were resolved at
surgery when the subtlety of the findings could not be dissected.
Conclusion
This study has shown no significant differences in the vegetation
characteristics between HIV-infected and uninfected patients.
Complications such as leaflet aneurysms and root abscesses
occurred in patients with CD
4
counts
<
250 /mm
3
in HIV-positive
subjects, and appeared to be of a larger size when compared to
the HIV-uninfected patients. The clinical outcome of medical
and surgical therapy was also similar in both groups.
References
1.
Moreillon P, Que YA. Infective endocarditis.
Lancet
2004;
363
:
139–149.
2.
The task force on the prevention, diagnosis and treatment of infective
endocarditis of the European Society of Cardiology (ESC) Guidelines
on the prevention, diagnosis and treatment of infective endocarditis
(new version 2009).
Eur Heart J
2009;
30
: 2369–2413.
3.
Durack DT. Evaluating and optimizing outcomes of surgery for endo-
carditis (editorial).
J Am Med Assoc
2003;
290
(24): 3250–3251.
4.
Abraham J, Veledar E, Lerakis S. Comparison of frequency of active
infective endocarditis by echocardiography in patients with bacteremia
