MAJOR ARTICLE

Quantitative Histological Examination

of Mechanical Heart Valves
Hubert Lepidi,1 Jean-Paul Casalta,1 Pierre-Edouard Fournier,1 Gilbert Habib,2 Frederic Collart,3 and Didier Raoult1
1
Unite des Rickettsies et des Pathogenes Emergents, Faculte de Medecine, Universite de la Mediterranee, and 2Service de Cardiologie
and 3Service de Chirurgie Cardiaque, Hopital de la Timone, Marseille, France

Background. Histological demonstration of microorganisms, vegetations, or active endocarditis in cardiac

valve tissue is included in the Duke criteria and is considered to be a criterion of confirmed infective endocarditis.
However, the histological features that characterize infective endocarditis are not accurately defined at the qualitative
and quantitative levels.
Methods. Pathologic analysis of tissue adjoining mechanical cardiac valves was undertaken retrospectively for
21 patients who underwent surgical removal of a mechanical valve because of suspected infective endocarditis and
69 patients who underwent surgical removal of a mechanical valve because of noninfectious dysfunction. To better
define the histological criteria for infective endocarditis, we used quantitative image analysis to compare these 2
groups of patients with respect to valvular fibrosis, calcifications, vegetations, patterns of inflammation, and
vascularization.
Results. Histologically, infective endocarditis in patients with mechanical valves was characterized by the
demonstration of microorganisms, vegetations, and significant neutrophil-rich inflammatory infiltrates with ex-
tensive neovascularization. In contrast, valve tissue specimens from patients with mechanical valves that were
removed because of noninfectious complications showed significant rates of extensive fibrosis and, when present,
inflammatory infiltrates that were mainly composed of macrophages and lymphocytes. A neutrophil surface area
with a cutoff value of 2% of the total valve tissue surface is highly predictive of (90%) and specific for (98%)
infective endocarditis.
Conclusions. When no microorganisms are detected and vegetations are lacking in tissue adjacent to a
mechanical valve, neutrophil-rich inflammation and extensive neovascularization might better histologically define
the term active endocarditis in the Duke criteria. This definition would allow differentiation between infective
endocarditis and inflammatory noninfectious valve processes in patients with mechanical cardiac valves.

Mechanical cardiac valves can be subjected to a diverse
spectrum of complications, including thromboembolic
complications, paravalvular leak or partial dehiscence,
durability problems caused by material degeneration,
and infective endocarditis [16]. Mechanical valve en-
docarditis is an infrequent complication (occurring in
1%6% of patients) but is a serious and possibly lethal
infection. Because bacteremia is a common feature
during infective endocarditis, the causative organism
is isolated in blood cultures in 190% of endocarditis
Received 23 July 2004; accepted 12 October 2004; electronically published 4
February 2005.
Reprints or correspondence: Dr. Didier Raoult, Unite des Rickettsies, Centre
National de la Recherche Scientifique, CNRS UMR 6020, WHO Collaborative
Center, IFR 48, Faculte de Medecine de la Timone, 27 Bd Jean Moulin, 13385
Marseille Cedex 5, France (Didier.Raoult@medecine.univ-mrs.fr).
Clinical Infectious Diseases 2005; 40:65561
 2005 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2005/4005-0002$15.00
cases. However, despite improved culture methods,
the results of blood cultures may be negative if the
patient has recently received antibiotics or if the or-
ganism is fastidious or requires special culture tech-
niques [710]. Alternative, nonculture laboratory di-
agnostic methods include serological tests, molecular
techniques, and pathologic examination [11]. Path-
ologic examination of resected valve tissues remains
the gold standard for the diagnosis of infective en-
docarditis, when these specimens are available [12].
Histological findings are included in the Duke criteria
and the von Reyn criteria and are considered to be
criteria of confirmed infective endocarditis [1315].
These criteria include the histological demonstration of
microorganisms in a vegetation or intracardiac abscess,
as well as pathologic lesions, such as vegetations and
active endocarditis at the histological level.
However, histological examination may fail to dem-
onstrate the presence of microorganisms in cardiac

Histological Examination of Mechanical Heart Valves CID 2005:40 (1 March) 655

valve tissue, and vegetations may be absent. In such cases, the
histological diagnosis of infective endocarditis can be made only
on the basis of the presence of inflammatory infiltrates in valve
tissues, which defines active endocarditis in the Duke criteria.
On the other hand, inflammation is often observed in the his-
tological examination of uninfected degenerative valves and can
cause confusion during diagnosis by raising the possibility of
infective endocarditis. Histological analysis requires subjective
interpretation, and, therefore, the analysis may not be consis-
tent between histopathologists. In our experience, when no
microorganisms are detected and vegetations are not evident
in valve tissues, there is no reproducibility in the final diagnosis
of infective endocarditis between different histopathologists or
with blind retesting of valve tissues by the same investigator
(H.L., unpublished observations). Futhermore, the concept of
active endocarditis in the Duke criteria has never been ac-
curately defined, either at the qualitative or quantitative level
[9], and no previous study has systematically examined the
sensitivity, specificity, and predictive values of histological ex-
amination findings for the diagnosis of infective endocarditis.
To better define the histological features of infective endo-
carditis associated with mechanical valves, we used quantitative
image analysis to retrospectively analyze the valvular histolog-
ical parameters, such as fibrosis, calcifications, vegetations, in-
flammatory infiltrates, and neovascularization, in tissue ad-
joining excised mechanical valves from 21 patients with and
69 patients without suspected mechanical valve endocarditis.
We analyzed the contribution of quantitative pathologic ex-
amination to the diagnosis of infective endocarditis associated
with mechanical valves.
PATIENTS, MATERIALS, AND METHODS
Case definition and patients. From April 1994 through Sep-
tember 2003, a total of 116 patients underwent surgical removal
of a mechanical valve at La Timone Hospital (Marseille,
France), of whom 90 patients had data analyzed for the pa-
thology of their infection. The patients were divided into 2
groups. In the endocarditis group, all 21 patients had either
definite or possible infective endocarditis, according to the pre-
operative application of the Duke criteriathat is, without the
pathologic criteria. In the control group, all 69 patients un-
derwent valvular removal for dysfunction that was presumably
noninfectious and were negative for infective endocarditis, ac-
cording to the preoperative application of the Duke criteria.
Patients were considered to have confirmed infective endocar-
ditis if microorganisms were detected by standard blood cul-
tures or cultures of valvular material. The details of the patients
clinical status, echocardiographic data, and preoperative diag-
nosis according to the Duke criteria, as well as valve tissue
samples, were obtained by physicians and surgeons at La Ti-
mone Hospital.
656 CID 2005:40 (1 March) Lepidi et al.
Histological analysis. All tissue samples that were excised
during the course of the surgical removal of mechanical valves
were collected in a sterile container without fixative or culture
medium. The entire sample was carried without delay to the
diagnostic microbiology laboratory for optimal recovery and
identification of microorganisms. The mechanical valves were
first examined grossly. The sewing ring was systematically sam-
pled, because infections usually begin around the sewing ring
and localize to the annulus. Other degenerative processes, such
as thrombi or other deposits, were also systematically sampled
[12]. After selecting valve tissue samples for bacteriological pro-
cedures, the remaining tissue samples were fixed in neutral
buffered formalin, were decalcified, if necessary, and were em-
bedded in paraffin.
Tissue specimens were then cut to 3 mm in thickness and
stained with hematoxylin-eosin-saffron by use of routine meth-
ods. Serial sections of each tissue specimen were also obtained
for special staining or immunohistochemical investigations.
Special stains, including periodic acid-Schiff, Giemsa, Gram,
Grocott-Gomori methenamine silver, and Warthin-Starry
stains, were used for the detection of bacteria and fungi.
Mechanical valve tissue specimens were divided into 3
groupsA, B, and Con the basis of the histological findings
and without knowledge of the findings of the preoperative ap-
plication of the Duke criteria and the bacteriological results for
each patient. Group A valve tissue specimens were those that
showed histological features of infective endocarditis when veg-
etations and/or polymorphonuclear leukocyte-rich valvular in-
flammation and/or microorganisms were present. Group B
valve tissue specimens were those that showed valvular inflam-
mation composed mainly of inflammatory mononuclear cells,
macrophages and lymphocytes, without vegetations or micro-
organisms. Group C valve tissue specimens were those that were
devoid of inflammation, vegetations, and microorganisms.
Quantitative image analysis. Fibrosis, calcifications, and
vegetations were analyzed in tissue samples by quantitative im-
age analysis, as described elsewhere [16, 17]. In brief, histo-
logical images were digitized and transferred to a computer
system. Using the image analyzer Samba 2005 (Samba Tech-
nologies), which is a specific interactive program that provides
visual control of analysis, we were able to analyze tissue sections
with respect to histological parameters. For each set of obser-
vations, the surfaces of 10 randomly chosen areas were studied
at a magnification of 100, and the surface areas of fibrotic
tissue, calcifications, and vegetations were measured. The av-
erage areas were calculated by comparison with the area of the
whole tissue sample. Mitral, aortic, and tricuspid valve tissue
samples were evaluated using the same procedure.
Immunohistochemical analysis. To quantitatively evaluate
the burden of polymorphonuclear leukocytes, macrophages,
and T lymphocytes, as well the relative proportion of neovas-
cularization in valve tissue specimens, paraffin sections were
stained with the antibodies anti-CD15 (Immunotech), anti-
CD68 (Dako), anti-CD3 (Dako), and factor VIIIrelated an-
tigen (Dako), respectively, using a peroxidase-based method,
as described elsewhere [17]. The antibodies anti-CD15 and anti-
CD3 were ready to use, whereas the antibodies anti-CD68 and
antifactor VIIIrelated antigen were used at a working dilution
of 1:1000. Another specific program was used to analyze the
CD15-positive, CD68-positive, CD3-positive, and factor VIII
positive surfaces to determine the percentages of the total sur-
face area covered by neutrophils, macrophages, T lymphocytes,
and endothelial cells, respectively.
Determination of cutoff values. The sensitivity, specificity,
and positive and negative predictive values of the expression
levels for leukocyte markers were determined to establish the
best cutoff value for the leukocyte marker(s) that were signif-
icantly associated with mechanical valve endocarditis. This was
analyzed using the cutoff values of 0.5%, 0.9%, and 2%
of the surface area of the examined cells that expressed the
markers among the population of 90 patients. The predictive
values of the cutoff values were evaluated in relationship to
the expected prevalence of infective endocarditis in the general
population. For that evaluation, we used Bayes theorem:
positive predictive value p SE PR / (SE PR) + (1SPE)
(1PR), and negative predictive value p SPE (1PR) / SPE
(1PR) + (1SE) PR, where SE is the sensitivity, PR is the
prevalence of the disease in the population, and SPE is the
specificity.
Statistical analysis and graphs. The Mann-Whitney U test
was used for the statistical comparisons of values that were
obtained for each histological parameter for valve tissue samples
from patients in the endocarditis group patients and the control
group. P ! .05 was considered to be statistically significant.
Bayesian curves and slopes of linear regression curves of the
expression levels for leukocyte markers were obtained using MS
Excel, version 7.0 (Microsoft).
RESULTS
Twelve patients, including 4 from the endocarditis group and
8 from the control group, were excluded from the study because
of the lack of valve tissue specimens available for histological
examination. In these cases, the valve tissue specimens that were
given to the pathologist contained only suture threads or a
noninflammatory thrombus. The mean ages  SDs of patients
in the endocarditis group and patients in the control group
were 56.52  17.38 and 58.27  13.83, respectively. The ratio
of men to women was 1.83 (11 men and 6 women) in the
endocarditis group and 0.79 (27 men and 34 women) in the
control group. In the endocarditis group, the mitral and aortic
valves were involved in 10 (58.82%) and 7 (41.17%) patients,
respectively; however, in the control group, mitral, aortic, and
Histological Examination of Mechanical Heart Valves CID 2005:40 (1 March) 657
tricuspid valves were involved in 34 (55.73%), 26 (42.62%),
and 1 (1.63%) patient(s), respectively. Both mitral and aortic
valves were removed from 2 patients in the endocarditis group
and from 5 patients in the control group. The causative or-
ganisms of mechanical valve endocarditis are summarized in
table 1. The results of microbiological testing were in accor-
dance with those reported in the literature [1]. Through the
use of special stains, microorganisms were histologically visu-
alized in 6 mechanical valve samples from the endocarditis
group: 3 samples from patients with staphylococcal endocar-
ditis, 2 from patients with streptococcal endocarditis, and 1
from a patient with fungal endocarditis (which was determined
to be due to Acremonium species).
Pathologic findings. The findings of the histological exam-
ination of valve tissue specimens are summarized in table 2,
stratified by diagnostic group. The valve tissue specimens from
most of the patients in the endocarditis group exhibited veg-
etations (12 patients [70.58%]) and moderate fibrosis (13 pa-
tients [76.47%]), and some exhibited calcifications (2 patients
[11.76%]). The valve tissue specimens from the patients in the
endocarditis group exhibited numerous inflammatory infil-
trates that were composed predominantly of polymorphonu-
clear leukocytes and abundant neovascularization (12 patients
[70.58%], figure 1). In contrast, valve tissue specimens from the
patients in the control group were devoid of vegetations and
demonstrated extensive fibrosis (42 patients [68.85%]), some-
times with calcifications (22 patients [36.06%]). In this group,
inflammatory infiltrates were rare and were focal in 23 patients
(table 2) and consisted mainly of macrophages and lymphocytes
Table 1. Etiologic organisms in 18 confirmed cases
of mechanical valve endocarditis.
No. (%)
Organism, by class of isolates
Staphylococci
Staphylococcus aureus 7 (38.88)
Staphylococcus epidermidis 2 (11.11)
Streptococci
Streptococcus oralis 2 (11.11)
Streptococcus mitis 1 (5.55)
Other bacteria
Acinetobacter species 2 (11.11)
Enterococcus cloacae 1 (5.55)
Actinobacillus actinomycetemcomitansa 1 (5.55)
Neisseria sicca 1 (5.55)
Fungi
Acremonium species 1 (5.55)
a
In this case, the mechanical valve was removed from a pa-
tient in the control group because of hemodynamic complica-
tions. However, the valve tissue samples exhibited histological
features of infective endocarditis (see definition for Group A in
Patients, Materials, and Methods), and A. actinomycetemcomi-
tans was isolated from the valve tissue culture (see Results).

Figure 1. Infective endocarditis in tissue adjacent to a mitral me-
chanical valve. Note the extensive destruction of valve tissue with a
vegetation and a dense inflammatory infiltrate that is mainly composed
of polymorphonuclear leukocytes (hematoxylin-eosin-saffron stain; original
magnification, 100).
with discrete neovascularization (figure 2). In 37 patients in
the control group (table 2), the valve tissue samples were non-
inflammatory (figure 3).
Some discrepancies existed. In the control group (table 2),
the valve tissue samples of 1 patient showed histological features
of infective endocarditis (group A) without vegetations. This
was considered to be a case of confirmed infective endocarditis
because Actinobacillus actinomycetemcomitans was isolated from
a valve culture (table 1) and blood cultures were sterile. For 2
patients in the endocarditis group (table 2), a diagnosis of
infective endocarditis was considered to be confirmed on the
basis of the preoperative application of the Duke criteria
despite valve tissue samples that were devoid of vegetations,
inflammatory infiltrates, and microorganisms (group C). Blood
and valve cultures yielded Neisseria sicca in one case and Staph-
ylococcus aureus in the other case. Finally, in 3 patients in the
endocarditis group, valve tissues samples showed inflammatory
infiltrates that were composed mainly of lymphocytes and mac-
rophages and did not exhibit vegetations and microorganisms
(group B, table 2). These 3 patients were considered to have
confirmed infective endocarditis on the basis of the preoper-
ative application of the Duke criteria. Blood and valve cultures
yielded S. aureus, Streptococcus oralis, and Acinetobacter species,
for 1 patient each.
Quantitative and statistical analyses of histological param-
eters and determination of cutoff values. Quantitative anal-
ysis of the histological findings for the valve tissue samples from
the endocarditis group and the control group is shown in fig-
658 CID 2005:40 (1 March) Lepidi et al.
ures 4 and 5. The valve tissue samples of the endocarditis group
were less fibrotic (P p .006) than were those of the control
group, but CD15 expression, which quantifies the surface area
covered by polymorphonuclear leukocytes, and neovasculari-
zation were more extensive in the endocarditis group (P p
.04 and P p .01, respectively). Moreover, all the valve tissue
specimens from the control group were devoid of vegetations.
In contrast, the quantitative analysis of the other valvular his-
tological parameterssurface area of calcifications and CD68
and CD3 expressiondid not show any statistically significant
differences between the endocarditis group and the control
group. Finally, the patients in the endocarditis group were seg-
regated according to the length of the interval whether 1
month, 2 months, or 3 monthsbetween the start of the an-
tibiotic treatment and surgical excision of the mechanical valves
(10, 3, and 2 patients, respectively). We observed that, among
all the histological parameters studied, only the expression of
CD15, CD68, and CD3 decreased in the valve tissue specimens
during that time (data not shown). This decrease was not sta-
tistically significant, probably because of the small number of
valve specimens studied.
Positive and negative predictive values of the 3 cutoff values
for the surface area of CD15 expression were calculated. The
2% cutoff value had a lower sensitivity than the 0.5% and
0.9% cutoff values (0.529 vs. 0.823 and 0.7, respectively) and
a slightly lower negative predictive value (0.882 vs. 0.946 and
0.916, respectively), but a higher specificity (0.983 vs. 0.868 and
0.901, respectively) and positive predictive value (0.9 vs. 0.636
and 0.667, respectively). Therefore, the 2% surface area cutoff
value was considered to be the most accurate.
DISCUSSION
Mechanical valve endocarditis continues to be a major com-
plication. Its case-fatality rate remains high when infections
occur with microorganisms such as S. aureus [1, 1820]. In
terms of the Duke criteria, regardless of whether the involved
valve is native or prosthetic, a definite diagnosis of infective
endocarditis can be made by histological demonstration of mi-
Table 2. Histological findings for tissue from mechanical tissue
valve specimens from valves removed because of suspected in-
fective endocarditis (endocarditis group) and valves excised be-
cause of dysfunction that was presumably noninfectious (control
group).
No. (%) of patients with specimen
in specified diagnostic group
Diagnostic Diagnostic Diagnostic
Study group group A group B group C
Endocarditis group (n p 17) 12 (70.58) 3 (17.64) 2 (11.76)
Control group (n p 61) 1 (1.63) 23 (37.70) 37 (60.65)
NOTE. Diagnostic groups are defined in Patients, Materials, and Methods.

Figure 2. Inflammatory degenerative valvular lesions in tissue adja-
cent to an aortic mechanical valve. Note the numerous inflammatory
infiltrates, which are mainly composed of mononuclear leukocytes, and
the fibrous degenerative changes (hematoxylin-eosin-saffron stain; original
magnification, 200).
croorganisms or by histological findings of vegetations and ac-
tive endocarditis. However, active endocarditis is not defined,
and the precise histological features that characterize infective
endocarditis are not standardized. Some investigators distin-
guish acute inflammation, which is defined as the presence of
polymorphonuclear leukocytes in valvular inflammatory infil-
trates, from chronic inflammation, which is defined as the pres-
ence of inflammatory infiltrates composed of mononuclear cells
(macrophages and lymphocytes) in the absence of neutrophils
[9]. However, as we observed, neutrophils, macrophages, and
lymphocytes are often mixed in valvular inflammatory infil-
trates. Quantitative studies of cardiac valves should help to
avoid subjective impressions that may lead to erroneous re-
ports. Moreover, quantitative studies can provide accurate re-
sults for the histological differentiation of infective endocarditis
from other inflammatory, but noninfective, valvular processes.
To achieve this aim, this histological diagnostic method was
evaluated in the present study of a larger number of mechanical
valves that were removed because of suspected infective en-
docarditis or that were removed because of noninfectious de-
generative changes and which, in the latter case, were used as
control specimens. The purpose of the present study was to
provide precise histological criteria that define the features of
infective endocarditis in patients from whom a mechanical
valve is removed because of dysfunction.
The histological diagnosis of infective endocarditis associated
with mechanical valves requires a different approach than the
approach for native valves. In patients with mechanical valves,
the infection is mainly situated at the host-prosthesis sewing
ring interface and localizes to the annulus, resulting in annular
abscesses without destruction of the material and small or ab-
sent vegetations [21, 22]. If vegetations are lacking, infective
endocarditis associated with mechanical valves may be under-
diagnosed on the basis of the findings of both echocardiography
and gross pathological examination. In our series, valve tissue
specimens from 6 patients with infective endocarditis were de-
void of vegetations (5 patients from the endocarditis group and
1 patient with infective endocarditis from the control group).
The presence of vegetations, extensive neutrophil-rich in-
flammation, and the visualization of microorganisms in valve
tissue are 3 well-known criteria for the histological diagnosis
of infective endocarditis [9, 13, 22]. However, the last criterion
does not apply in all cases in the diagnosis of infective endo-
carditis, because microorganisms may be destroyed by preop-
erative antibiotic treatment, and, consequently, may not be de-
tectable by histological methods. In our series, microorganisms
were visualized in valve tissues from only 6 patients in the
endocarditis group. On the other hand, the 2 other criteria
have not been previously quantified in the literature concerning
mechanical valve endocarditis. In the present study, we have
shown by use of quantitative analysis that vegetations, when
present, represent an important proportion of the valve tissue
area (an average of 28.23%, in the cases in the present study).
In contrast to T lymphocytes and macrophages, polymorpho-
nuclear leukocytes are the only inflammatory cells whose pres-
ence allows the differentiation of endocarditis from other in-
flammatory, but noninfective, processes. CD15 expression in
Figure 3. Noninflammatory degenerative valvular lesions in tissue
adjacent to a mitral mechanical valve. Note the extensive fibrosis of the
connective valve tissue with focal calcifications (hematoxylin-eosin-saf-
fron stain; original magnification, 200).

Histological Examination of Mechanical Heart Valves CID 2005:40 (1 March) 659


Figure 4. Quantification of histopathological changes in tissue from
excised mechanical valves from patients with suspected infective en-
docarditis (endocarditis group, n p 17 ) or with noninfectious dysfunction
(control group, n p 61). Fibrosis, calcifications, and vegetations were
quantified in valve tissue sections stained with hematoxylin-eosin-saffron.
Quantification of each parameter was evaluated by computer-assisted
analysis of digitized microscopic images. Results were normalized and
are expressed as a percentage of the total valve surface area. Columns
represent mean values  SE.
valve tissues represents the histological burden of neutrophils
[17]. In order to use CD15 expression as a diagnostic histo-
logical marker for infective endocarditis, we considered it to
be more useful to use the area of CD15 expression that had
the greatest positive predictive value (i.e., a surface area of
2%). As is true for the tissue of native cardiac valves, tissue
adjacent to mechanical valves is normally avascular [23]. How-
ever, when an inflammatory reaction appears in valve tissue, it
always creates neovascularization of variable abundance that
allows the entry of leukocytes into valve tissue and the devel-
opment, in turn, of an inflammatory reaction. Thus, neovas-
cularization may be a new histological criterion that could aid
pathologists in the recognition of infective processes in valve
tissue, because the degree of vessel formation is statistically
greater in infective endocarditis than in noninfective valvular
damage. In contrast, although noninfective valve processes can
include significant inflammatory infiltrates, we showed quan-
titatively that these infiltrates are mainly composed of mono-
nuclear leukocytes, lymphocytes, and macrophages, and are as-
sociated with poor neovascularization. Moreover, degenerative
changes, such as fibrosis, are more pronounced in noninfective
valve damage than in infective endocarditis.
We observed 2 cases of infective endocarditis without vege-
tations, inflammatory infiltrates, neovascularization, or micro-
660 CID 2005:40 (1 March) Lepidi et al.
organisms and 3 cases of infective endocarditis without vegeta-
tions or microorganism but with mononuclear cell inflammation.
These 5 false-negative cases were probably the result of errors in
the valve tissue sampling procedure. At La Timone Hospital, the
excised valve tissue samples are first processed by a bacteriologist,
who divides the sample to select some for bacteriological pro-
cedures, such as various culture systems and molecular tests [11].
As a result, the resected valve tissue samples that are submitted
for histological examination are often small. Because the infective
process may be confined to a small area of the valve tissue, the
part provided for histological examination may be devoid of the
infective process or may show a nonspecific inflammation. In
contrast, histological examination or culture of valve material,
when performed systematically, allowed us to diagnose a case of
infective endocarditis that had not been identified by the pre-
operative application of the Duke criteria.
In conclusion, we present the results of a quantitative anal-
ysis of several histological parameters for tissue from excised
mechanical valves to better define the histological features of
infective endocarditis. When vegetations are lacking and mi-
croorganisms fail to be detected in valve tissues by histological
examination, the pattern of inflammation may be considered
to be a key to formulating the diagnosis. We showed that
infective processes are characterized by inflammatory infil-
trates that are composed mainly of polymorphonuclear leu-
Figure 5. Quantification of histopathological changes in tissue from
excised mechanical valves from patients with suspected infective en-
docarditis (endocarditis group, n p 17 ) or with noninfectious dysfunc-
tion (control group, n p 61). Surface areas expressing CD15, CD68,
CD3, and factor VIII were quantified after immunostaining. Quantifi-
cation of each parameter was evaluated by computer-assisted analysis
of digitized microscopic images. Results were normalized and are ex-
pressed as a percentage of the total valve surface area. Columns
represent mean values  SE.
kocyteswith CD15 expression on 2% of the valvular sur-
face considered to be the cutoff valueand by extensive
neovascularization. These 2 histological features may better
define the term active endocarditis in the histopathological
Duke criteria for infective endocarditis and allow the differ-
entiation of infective endocarditis from inflammatory non-
infective valvular processes.
Acknowledgments
We express many thanks to J. Stephen Dumler and Esther Platt for their
thoughtful review of the manuscript.
Potential conflicts of interest. All authors: no conflicts.
References
1. Vongpatanasin W, Hillis LD, Lange RA. Prosthetic heart valves. N Engl
J Med 1996; 335:40716.
2. Schoen FJ. Surgical pathology of removed natural and prosthetic heart
valves. Hum Pathol 1987; 18:55867.
3. Alvarez L, Escudero C, Figuera D, Castillo-Olivares JL. The Bjork-Shiley
valve prosthesis: analysis of long-term evolution. J Thorac Cardiovasc
Surg 1992; 104:124958.
4. Lepidi H. Pathology of cardiac valve prostheses. Ann Pathol 1999; 19:
18794.
5. Masiello P, Cassano V, Di Benedetto G. Fibrous tissue ring: an un-
common cause of severe prosthetic valve stenosis. J Thorac Cardiovasc
Surg 1995; 109:1253.
6. Silver MD, Butany J. Mechanical heart valves: methods of examination,
complications, and modes of failure. Hum Pathol 1987; 18:57785.
7. Hoen B, Selton-Suty C, Lacassin F, et al. Infective endocarditis in
patients with negative blood cultures: analysis of 88 cases from a one-
year nationwide survey in France. Clin Infect Dis 1995; 20:5016.
8. Berbari EF, Cockrill FR, Steckelberg JM. Infective endocarditis due to
unusual or fastidious microorganisms. Mayo Clin Proc 1997; 72:
53242.
Histological Examination of Mechanical Heart Valves CID 2005:40 (1 March) 661
9. Morris AJ, Drinkovic D, Pottumarthy S, et al. Gram stain, culture, and
histopathological examination findings for heart valves removed be-
cause of infective endocarditis. Clin Infect Dis 2003; 36:697704.
10. Moreillon P, Que YA. Infective endocarditis. Lancet 2004; 363:13949.
11. Fournier PE, Raoult D. Nonculture laboratory methods for the diag-
nosis of infectious endocarditis. Curr Infect Dis Rep 1999; 1:13641.
12. Lepidi H, Durack DT, Raoult D. Diagnostic methods: current best
practices and guidelines for histologic evaluation in infective endo-
carditis. Infect Dis Clin N Am 2002; 16:33961.
13. Durack DT, Lukes AS, Bright DK, Duke Endocarditis Service. New
criteria for diagnosis of infective endocarditis: utilization of specific
echocardiographic findings. Am J Med 1994; 96:2009.
14. von Reyn CF, Levy BS, Arbeit RD, Friedland G, Crumpacker CS. In-
fective endocarditis: an analysis based on strict case definitions. Ann
Intern Med 1981; 94:50517.
15. Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the Duke
criteria for the diagnosis of infective endocarditis. Clin Infect Dis 2000;
30:3638.
16. Lepidi H, Fournier PE, Raoult D. Quantitative analysis of valvular
lesions during Bartonella endocarditis. Am J Clin Pathol 2000; 114:
8809.
17. Lepidi H, Fenollar F, Dumler JS, et al. Cardiac valves in patients with
Whipple endocarditis: microbiological, molecular, quantitative histo-
logic, and immunohistochemical studies of 5 patients. J Infect Dis 2004;
190:93545.
18. Zeien LB, Klatt EC. Cardiac valve prostheses at autopsy. Arch Pathol
Lab Med 1990; 114:9337.
19. Rose AG. Prosthetic valve endocarditis: a clinicopathological study of
31 cases. S Afr Med J 1986; 69:4415.
20. John MDV, Hibberd PL, Karchmer AW, Sleeper LA, Calderwood SB.
Staphylococcus aureus prosthetic valve endocarditis: optimal manage-
ment and risk factors for death. Clin Infect Dis 1998; 26:13029.
21. Arnett EN, Roberts WC. Valve ring abscess in inactive infective en-
docarditis: frequency, location and clues to clinical diagnosis from the
study of 95 necropsy patients. Circulation 1976; 54:1406.
22. Atkinson JB, Virmani R. Infective endocarditis: changing trends and
general approach for examination. Hum Pathol 1987; 18:6038.
23. Schoen FJ, Sutton MSJ. Contemporary pathologic considerations in
valvular heart disease. In: Virmani R, Atkinson JB, Fenoglio JJ, eds.
Cardiovascular pathology. WB Saunders, 1991:33453.