MAJOR ARTICLE
Gram Stain, Culture, and Histopathological
Examination Findings for Heart Valves
Removed because of Infective Endocarditis
Arthur J. Morris,1 Dragana Drinkovic,1 Sudha Pottumarthy,1 Marianne G. Strickett,1 Donald MacCulloch,1
Neil Lambie,2 and Alan R. Kerr3
Departments of 1Microbiology, 2Pathology, and 3Cardiothoracic Surgery, Green Lane Hospital, Auckland, New Zealand
Retrospective chart review was undertaken for 480 patients who underwent a total of 506 valve replacements
or repair procedures for infective endocarditis. The influence of preoperative antimicrobial treatment on
culture, Gram stain, and histopathological examination findings for resected valve specimens was examined.
When valves were removed before the end of treatment, organisms were seen on the Gram stain of ground
valve material performed in the microbiology laboratory and on Gram-stained histopathological sections in
231 (81%) of 285 and 140 (67%) of 208 specimens, respectively (P p .0007 ). Gram-positive cocci were either
cultured from or observed in excised valve tissue in 42 (67%) of 63 episodes involving negative preoperative
blood cultures. Positive Gram stain results for microbiological specimens should be reintroduced into the
definite pathological criteria for infective endocarditis. When deciding on how long to continue antimicrobial
therapy after valve replacement for endocarditis, valve culture results should be the only laboratory finding
taken into account, because it takes months for dead bacteria to be removed from sterile vegetations.
Most cases of bacterial endocarditis can be cured by
antimicrobial treatment alone. When infection causes
valve dysfunction, prompt valve replacement is required
for optimal outcome [1–3]. Valve replacement is also
indicated when infection has not been controlled by
antimicrobial treatment or when recurrent embolization occurs [1–4]. In these situations, surgery improves
outcome, and delays in surgery (e.g., in an attempt to
sterilize the valve) are associated with increased morbidity and mortality [1–3, 5–10].
After valve replacement is performed for active bacterial endocarditis, there is no consensus on how long
Received 10 September 2002; accepted 21 November 2002; electronically
published 4 March 2003.
Financial support: The Auckland Infectious Disease Education and Research
Trust.
Reprints or correspondence: Dr. Arthur J. Morris, Microbiology Laboratory,
Auckland City Hospital, Auckland 1003, New Zealand ([email protected]).
Clinical Infectious Diseases 2003; 36:697–704
2003 by the Infectious Diseases Society of America. All rights reserved.
1058-4838/2003/3606-0004$15.00
to continue antimicrobial treatment [1]. Although
some researchers have given consideration to surgical
findings and culture, Gram stain, or histopathological
examination findings [3, 4, 7, 10–13], others administer
treatment for a standard period after valve replacement
[5, 14–21]. No previous study has systematically examined the influence of preoperative antimicrobial
treatment on the culture, Gram stain, and histopathological findings for resected valves.
Previous reports have used different pathological criteria to define endocarditis with respect to the Gram
staining of resected tissue performed in the microbiology laboratory [22–25]. We have undertaken a retrospective review of all patients who underwent valve
replacement for infective endocarditis and have recorded the results of valve cultures, Gram stains of valve
specimens performed in the microbiology laboratory,
and histopathological examinations (Gram stains and
determination of the presence of acute inflammation).
Our first objective was to describe the evolution of these
findings while patients received antimicrobial treatment
Gram Stain, Culture, and Histopathology in IE • CID 2003:36 (15 March) • 697
for endocarditis. Our second objective was to compare the
findings for the Gram-stained specimens examined in the microbiology laboratory (hereafter referred to as “microbiology
Gram stains”) with the findings for Gram-stained histopathological sections (hereafter referred to as “histopathology Gram
stains”) in this series of patients.
PATIENTS AND METHODS
Patient population. Green Lane Hospital (Auckland, New
Zealand) has a large cardiothoracic surgical unit serving ∼1
million people, and local hospitals refer 17%–25% of their
patients with endocarditis to Green Lane Hospital for surgical
treatment [26, 27]. All patients who undergo valve replacement
surgery have a known or possible cause for valve dysfunction
entered into a computer database. All patients who underwent
valve replacement surgery from September 1963 through December 1999 and who had a diagnosis of endocarditis recorded
in the database had their charts reviewed. All relevant information was recorded for each patient, including the following
data: number of blood culture sets performed, number of positive blood culture sets, organism or organisms isolated, duration of antibiotic treatment before surgery, operation findings, and the results of microbiological (culture and Gram stain
findings) and histopathological (Gram stain findings and characteristics of the inflammatory cell infiltrate) examinations for
the resected valve.
Valve replacement was considered to have occurred during
active endocarditis if the patient had not completed a standard
antimicrobial therapy (SAT) regimen. For viridans streptococci,
enterococci, and staphylococci, the recommendations of Wilson
et al. [28] were used to define SAT. When these recommendations specified 4–6 weeks as the duration of treatment, 4
weeks was regarded as being SAT. For prosthetic valve endocarditis, 6 weeks was regarded as SAT, except for allografts, for
which 4 weeks was regarded as SAT. For other organisms on
native valves, 4 weeks was regarded as being SAT. Timing of
surgery was analyzed according to the proportion of SAT that
had been completed at the time of the operation. This approach
was taken because a given duration of treatment in days does
not convey sufficient meaning. For example, 14 days of penicillin treatment for an episode of penicillin-susceptible viridans
streptococcal endocarditis would represent completed treatment for a native valve if penicillin was combined with an
appropriate aminoglycoside, but as only one-half of a treatment
course if it was given as monotherapy, and as only one-third
of the recommended course if endocarditis involved a prosthetic valve. By analyzing treatment as the proportion of SAT
completed, the organism, valve type, and treatment regimen
could all be taken into consideration [28].
Days of receipt of parenteral antimicrobial treatment with
698 • CID 2003:36 (15 March) • Morris et al.
⭓1 active antibiotic were counted. Oral agents taken alone were
not counted in the duration of treatment. The duration was
determined from the start of appropriate treatment, not from
when diagnosis of endocarditis was made.
Definitions. The modified Duke criteria [24] were used
to define cases of endocarditis, with the additional criterion of
a positive microbiology Gram stain being included in the definite pathological criteria for endocarditis. Patients who had
negative preoperation blood culture results but who underwent
valve replacement surgery while receiving antibiotic treatment
were considered to have blood culture–negative endocarditis,
regardless of the findings of Gram stain, culture, or histopathological examination of samples of the valves removed or resected at the time of surgery. Patients for whom endocarditis
was not suspected before the operation (i.e., the patient did
not have fever and no blood samples were obtained for culture)
but who were discovered at the time of surgery to have infected
valves were deemed to have incubating endocarditis.
Acute inflammation in histopathology sections was defined
as the presence of polymorphonuclear leukocytes (PMLs) in
the inflammatory cell infiltrate. Chronic inflammation was defined as an inflammatory cell infiltrate (e.g., histiocytes and
lymphocytes) without PMLs being present.
Laboratory procedures.
Resected valve specimens were
sent for Gram staining and culturing in sterile containers without additives. After macroscopic examination, vegetations and
material that appeared to be infected were removed and ground
in 0.5 mL of trypticase-soy broth with a 15-mL tapered tissue
grinder. Gram stains of the ground material were made. Blood
and chocolate agar media were inoculated, examined regularly,
and discarded after 7 days’ incubation. Four liquid media (20
mL each) were inoculated: Sabouraud dextrose broth; biphasic
brain-heart infusion broth with a nutrient agar scope, with
4%–10% added CO2; hypertonic blood culture medium containing 10% sucrose; and anaerobic thioglycollate broth. Mechanical valves were placed in anaerobic thioglycollate broth.
Inoculated broths were incubated at 35C and inspected daily.
Routine subcultures were performed after 7 days onto Columbia agar that contained 5% defibrinated sheep RBCs. Plates
were aerobically and anaerobically incubated for 48 h before
discarding.
The Gram stain finding for ground valve material was reported without waiting for the findings of culture or histopathological examination. Isolates recovered from surgical specimens were compared with available preoperative blood culture
isolates. Obvious plate or broth culture contaminants were ignored and were not taken to indicate a positive culture (e.g.,
1 of 4 broths with a coagulase-negative Staphylococcus species
that had a susceptibility profile different from that of the blood
culture isolate).
For histopathological examination, specimens were placed in
formalin and sent to the histopathology laboratory for routine
processing and hematoxylin-eosin and Gram staining. Histopathological findings were reported by the pathologist assigned
to the case. No special reporting protocol was in place to provide a standardized report, but when endocarditis is mentioned
on the laboratory request form, it is standard practice for the
histopathology report to specifically state the presence or absence of organisms and to describe the composition of the
inflammatory cell infiltrate. In a limited number of cases in
which the histopathology report lacked sufficient detail, one of
us would reread the slides. Histopathological findings were recorded as organisms present irrespective of the inflammatory
cell infiltrate, acute inflammation without organisms, chronic
inflammation without organisms, or other findings (e.g., fibrosis or valve thickening).
Blood culture methods. Cultures of blood samples obtained from patients suspected of having endocarditis were incubated for 14 days until mid-1994, when the duration was
reduced to our current practice of 10 days’ incubation. Manual
blood culture methods were in use until mid-1986, when the
BACTEC system (Becton Dickinson) was introduced. This was
replaced by the BacT/Alert system (Organon Tecknika) in early
1995.
Statistical analysis. Analysis was performed with use of
the x2 test for 2 ⫻ 2 tables. P ! .05 was considered to be statistically significant.
Table 1. Etiologies of 506 episodes of infective endocarditis in
patients who underwent valve surgery.
Finding
Organism
Staphylococci
All
Staphylococcus aureus
Staphylococcus epidermidis
181 (36)
114 (23)
56 (11)
Other staphylococcia
11 (2)
177 (35)
Streptococci
Viridans streptococci
Streptococcus sanguis
Streptococcus mitis
56 (11)
43 (8)
17 (3)
Streptococcus bovis
Streptococcus anginosus
Otherb
12 (2)
10 (2)
Enterococcus speciesc
Abiotrophia species
d
Corynebacterium species
e
Other gram-positive organisms
HACEK group organismsf
Other gram-negative organismsg
Fungih
Culture-negative cases
Gram-positive cocci present
Acute inflammation in histopathological specimen
Definite endocarditisi
Possible endocarditisi
RESULTS
Patients. There were 480 patients who underwent 506 operations for treatment of infective endocarditis. Twenty patients
underwent 2 valve replacements and 3 patients underwent 3
valve replacements for treatment of endocarditis. Of the 480
patients, men predominated (71%). Only 5 patients were
known injection drug users. Aortic valves were the most commonly infected valves, either alone (277 [58%]) or in combination with other valves (27 [8%]). One hundred ninety-three
patients (40%) had undergone previous valve surgery, and 55
(11%) had experienced a previous episode of endocarditis.
Inclusion criteria for endocarditis. A total of 424 (84%)
of 506 episodes met our pathological criteria for endocarditis:
153 (30%) had positive valve culture results, 230 (45%) had
organisms present in either the microbiology or histopathology
Gram stain, and 41 (8%) did not have organisms seen or grown
but had acute inflammation in sections obtained for histopathological examination [24]. Of the 82 remaining episodes,
20 and 62 met the Duke clinical criteria for definite and possible
endocarditis, respectively [24].
Etiology of endocarditis. The etiologies of the 506 episodes
of endocarditis are shown in table 1. Episodes of blood culture–negative endocarditis that had positive valve culture results
No. (%) of
episodes
39 (8)
31 (6)
11 (2)
12 (2)
7
16 (3)
15 (3)
7
51 (10)
32 (6)
10 (2)
1
8
NOTE. Two patients had polymicrobial endocarditis with 2 isolates present
in multiple preoperative blood cultures: a viridans Streptococcus species and an
Enterococcus species; and S. epidermidis and a b-hemolytic Streptococcus species. One patient with incubating endocarditis with gram-positive cocci present
in a Gram-stained histopathological specimen is included in the culture-negative
group. Thirteen of the blood culture–negative cases (table 3) had positive valve
cultures and are included with the respective organisms, as are 9 of 10 cases
of incubating endocarditis that had positive valve cultures. HACEK, Haemophilus,
Actinobacillus, Cardiobacterium, Eikenella, and Kingella species.
a
S. capitis (2 episodes), S. warneri (2), S. cohnii (1), and nonspeciated
coagulase-negative staphylococci (6).
b
S. mutans (8 episodes), Group G Streptococcus species (6), S. pneumoniae (4), S. agalactiae (4), S. pyogenes (5), S. oralis (3), S. salivarius (2), S. mitior
(1), and nonspeciated streptococci (6).
c
E. faecalis (24 episodes), E. faecium (1), E. durans (1), and nonspeciated
enterococci (5).
d
C. jeikeium (6 episodes), C. pseudodiphtheriticum (3), C. diphtheriae (1),
and nonspeciated corynebacteria (2).
e
Dermobacter hominis (1 episode), Gemella morbillorum (1), Lactobacillus
casei (1), Listeria monocytogenes (1), nonspeciated lactobacilli (2), and a nonspeciated aerobic sporing bacillus (1).
f
Haemophilus species: H. parainfluenzae (3 episodes), H. influenzae (2), H.
paraphrophilus (2), and H. aphrophilus (1); Actinobacillus actinomycetemcomitans (1), Cardiobacterium hominis (2), Eikenella corrodens (1), Kingella kingae
(2), and Kingella denitrificians (2).
g
Serratia marcescens (2 episodes), Pseudomonas aeruginosa (2), Pseudomonas alcaligenes (1), Brucella abortus (1), Citrobacter diversus (1), Escherichia coli (1), Flavobacterium species (1), Klebsiella pneumoniae (1), Neisseria
mucosa (1), Neisseria pharyngis (1), Salmonella species (1), Stenotrophomonas
maltophilia (1), and a nonspeciated, nonlactose fermenting gram-negative bacillus (1).
h
Aspergillus fumigatus (2 episodes); Candida species: C. albicans (3), C.
glabrata (1), C. parapsilosis (1).
i
For clinical criteria, see [24].
Gram Stain, Culture, and Histopathology in IE • CID 2003:36 (15 March) • 699
(n p 13) are recorded alongside the respective organisms, as
are the 9 episodes of incubating endocarditis that had positive
valve culture results. Patients with blood culture–negative endocarditis who had negative valve culture results (n p 50) are
recorded as having cases of culture-negative endocarditis and
are grouped according to the histopathological and clinical criteria for endocarditis. One patient with incubating endocarditis
who did not have valve culture performed is included with the
culture-negative cases because gram-positive cocci were present
in histopathology Gram stains.
Findings of culture, Gram stain, and histopathological examinations. The results of culture, microbiology Gram stain,
histopathology Gram stain, and examination for the presence
of acute inflammation in patients with negative histopathology
Gram stain results are listed in table 2. Valves were seldom
culture positive after receipt of ⭓50% of SAT, but microbiology
Gram stain findings were positive for 160% of patients who
were still receiving antibiotic treatment. Nonviable bacteria persist for weeks to months in sterilized vegetations, and acute
inflammation may persist for weeks to months after microbiological cure (table 2 and figure 1). Chronic inflammation (i.e.,
inflammation without PMLs) made up a small but increasing
proportion of the histopathological findings late into and ⭐6
months after completion of SAT (figure 1).
Five (9%) of 53 valves from patients who were still receiving
treatment but who had completed SAT were culture positive
(table 3). One additional valve was also culture positive (patient
6; table 3). This patient had completed 4 weeks of combination
therapy with ampicillin and netilmicin for Enterococcus faecalis
native aortic valve endocarditis. Three weeks later, he received
ampicillin prophylaxis for a colonoscopy. At the time of surgery,
12 days after undergoing colonoscopy, an obviously infected
(as determined macroscopically), culture-positive aortic valve
was removed.
Blood culture–negative episodes. There were 63 episodes
of suspected endocarditis that involved negative blood cultures.
For the 44 episodes in which the number of negative blood
culture results was recorded, the median number of blood cultures was 4, and, for 20 episodes, there were ⭓5 negative blood
culture results. Twenty-three patients (37%) were receiving or
had recently received antibiotic treatment at the time that blood
samples were obtained for culture, and 4 (17%) of the 23 had
positive valve culture results.
Thirteen (21%) of 63 episodes that involved negative blood
culture results had positive valve culture results. The isolates
were all gram positive: Staphylococcus aureus (n p 1), coagulase-negative staphylococci (n p 6 ), Corynebacterium jeikeium
(n p 2), viridans streptococci (n p 3 ), and Abiotrophia defectiva (n p 1). Only the latter can be regarded as a fastidious
organism. A further 31 episodes (49%) had gram-positive cocci
Table 2. Gram stain, culture, and histopathological findings for 506 episodes of infective endocarditis that required removal or
resection of heart valves.
Microbiological findings
Variable
No. of
episodes
No. of
positive Gram
stain results/
total no.
of samples
stained (%)
No. of
positive
culture results/
total no. of
cultures (%)
Histopathological findings
No. of
samples with
organisms
present/no.
examined (%)
No. of
samples with
no organisms
found but acute
inflammation
presenta/no.
examined (%)
Proportion of standard duration antibiotic treatment
completed at time of operation
⭐25%
106
88/100 (88)
76/106 (72)
51/63 (81)
9/63 (14)
26%–50%
113
85/101 (84)
40/108 (37)
50/70 (71)
17/70 (24)
51%–75%
57
37/50 (74)
7/54 (13)
21/39 (54)
12/39 (31)
76%–100%
41
21/34 (62)
2/38 (5)
18/36 (50)
11/36 (31)
1100% but still receiving treatment
61
28/43 (65)
5/53 (9)
29/53 (55)
11/53 (21)
⭐1 month before operation
22
7/15 (47)
0/19 (0)
9/20 (45)
4/20 (20)
11 month but !6 months before operation
33
4/18 (22)
1/22 (5)
6/25 (24)
9/25 (36)
Negative blood cultureb
63
31/52 (60)
13/60 (22)
26/48 (54)
11/48 (23)
Incubating endocarditis presentc
10
7/8 (88)
9/9 (100)
4/6 (67)
2/6 (33)
Patient stopped treatment
a
Presence of polymorphonuclear leukocytes in the inflammatory cell infiltrate.
One or more negative preoperative blood cultures.
Endocarditis was not suspected before the patient underwent surgery (i.e., the patient did not have fever and blood samples were not obtained preoperatively
for culture). See table 4.
b
c
700 • CID 2003:36 (15 March) • Morris et al.
Figure 1.
Histopathological findings for heart valves removed because of infective endocarditis
seen in either the microbiology or histopathology Gram stains.
Overall, 44 episodes (70%) were due to gram-positive organisms, and 42 (67%) were due to gram-positive cocci. The clinical presentation of these cases favored streptococcal rather then
staphylococcal etiology.
The proportion of episodes that involved negative blood culture results did not change during the study period (14 [13%]
of 107 episodes before 1980, 24 [12%] of 200 episodes in the
1980s, and 25 [13%] of 199 episodes in the 1990s), nor did
the proportion change with the changes in blood culture systems (data not shown).
Incubating endocarditis.
Ten episodes of endocarditis
were discovered at the time of surgery (table 4). All had either
obvious or suspicious surgical findings of endocarditis.
Patients with definite endocarditis on the basis of the microbiology Gram stain findings. There were 59 episodes that
involved negative valve culture results that had neither organisms
seen on the histopathology Gram stain nor acute inflammation
noted on histopathological examination, or for which no histopathological examination was performed, but that had organisms seen on the microbiology Gram stain. If the microbiology
Gram stain result was ignored, these episodes would have been
defined as definite and possible endocarditis in 24 and 26 patients,
respectively, on the basis of the Duke clinical criteria [24]. The
remaining 9 episodes would have been rejected on the basis of
the Duke criteria, but, at the time of surgery, 8 of the patients
with episodes, who had been treated for 2–14 days before the
operation, were reported by the surgeon to have obviously infected valves (e.g., vegetations, cusp rupture, or perforation). The
single patient whose allograft aortic valve did not look obviously
infected had received 38 days of treatment for Streptococcus mutans endocarditis before surgery.
Microbiology and histopathology Gram stain findings. A
total of 293 episodes had both microbiology and histopathology
Gram stains performed. In 153 episodes (52%), organisms were
seen on both Gram stains; in 62 (21%), the microbiology Gram
stain was positive but the histopathology stain was negative; in
29 (10%), the histopathology Gram stain was positive but the
microbiology Gram stain was negative; and in 48 (16%), neither
Gram stain was positive. The microbiology Gram stain was
more likely to be positive (74% vs. 63%; P ! .0001).
Microbiology Gram stains were performed for the majority
of valves of each valve type (native valves, 82%; allografts, 73%;
bioprosthetic valves, 84%; and mechanical valves, 94%). Histopathological examination was performed for most native
valves and allografts, but it was seldom performed for episodes
of prosthetic valve endocarditis (native valves, 91%; allografts,
84%; bioprosthetic valves, 39%; and mechanical valves, 15%).
Gram Stain, Culture, and Histopathology in IE • CID 2003:36 (15 March) • 701
Table 3.
Patient
Characteristics of culture-positive patients who completed standard-duration antibiotic treatment.
Infected valve(s)
Treatment
(duration, days)
Microbiological findings
Surgical
Gram stain
Culture findings
Findings of
histopathological
examination
1
Aortic mitral, native
Combination penicillin
and gentamicin (20)
Para-aortic abscess
Gram-positive cocci
Streptococcus
a
sanguis
Both valves with PML
infiltration and grampositive cocci
2
Aortic, bioprosthetic (Hancock)
b-Lactam agents (43;
⭓19 in combination
with gentamicin)
Large vegetation extending to orifice of
right coronary artery
Gram-positive cocci
Staphylococcus
b
epidermidis
Not done
3
Aortic, native
Doxycycline and
rifampicin (43)
Nodular thickened valve,
granulations extending into annulus
Not done
Brucella abortus
Fibrosis, calcification, and
lymphocytic infiltration;
no PMLs
4
Aortic, allograft
Amphotericin (32)
Vegetations
Not done
Candida albicans
Vegetations with PMLs
and pseudohyphae
5
Aortic mitral,
mechanical
Amphotericin (52) and
itraconazole (39;
total therapy, 56)
Vegetations extensive
pannus
Septate, branching
hyphae
Aspergillus
fumigatus
Septate branching
hyphae
6
Aortic, native
Ampicillin and netilc
micin (28)
Vegetations
Gram-positive cocci
Enterococcus
faecalis
Vegetations with PMLs;
no organisms seen
NOTE.
a
b
c
PML, polymorphonuclear leukocyte.
MIC of penicillin, 0.03 mg/L. The MIC was determined for the blood culture isolate; it was not determined for the valve culture isolate.
Endocarditis due to penicillin-susceptible S. epidermidis 33 months after native aortic valve replacement with Hancock bioprosthetic valve.
Five-week gap between completion of treatment and surgery.
DISCUSSION
The first objective in this study was to report the evolution of
culture and Gram stain findings during treatment for infective
endocarditis. In this series, organisms were seen in most episodes that involved operation during antimicrobial treatment
but recovered in an ever-decreasing proportion of patients as
treatment progressed. There are few previous data on Gram
stain and culture findings for valves resected for endocarditis
[20, 29–34]. Furthermore, when authors do comment on the
detection of organisms, it is unclear whether they are referring
to the microbiology Gram stain, the histopathology Gram stain,
or both.
Past reviewers have differed in their interpretation of the
disparity between Gram stain and culture results. The opinion
of Jung et al. [35] was that this “proves that conversion of
blood cultures from positive to negative results does not necessarily mean that the organism was eradicated from the involved valve” (p. 512). Manhas et al. [21] concluded, “This
suggests that even when a full course of antibiotic therapy
has not been given, the valve was rendered sterile in a number
of patients” (p. 745). We agree with Manhas et al. [21] and
believe that seeing organisms on Gram stains of ground valve
material or by histopathological examination does not mean
that viable organisms are present. Culture results should be
the index of whether the surgery has been performed in an
infected field, because it may take months for dead bacteria
in a vegetation to be removed by phagocytosis and/or bacterial
cell lysis (table 2).
The second objective of this study was to report the microbiology Gram stain findings for resected valves. We are unsure
702 • CID 2003:36 (15 March) • Morris et al.
why positive microbiology Gram stain results are omitted in
the Duke criteria as definitive pathological evidence of endocarditis [23, 24]. It may have been because microbiology Gram
stain was thought to be insufficiently sensitive or specific or
because the Duke endocarditis database did not have enough
microbiology Gram stain findings stored to allow adequate
analysis. If the latter were the case, this study removes any data
deficit. If it were the former, we have shown that the microbiology Gram stain performs as well, if not better, than the
histopathology Gram stain. As for any concern over the specificity of microbiology Gram stain, we believe this to be a theoretical consideration only. Microbiology technologists are
skilled microscopists and regularly examine Gram-stained
smears of tissue specimens. In our experience, there is such an
abundance of organisms in cases of streptococcal endocarditis
that it only takes examination of a few microscopic fields for
organisms to be unambiguously identified. Although staphylococci can have variable size and staining characteristics, especially later in a course of antimicrobial therapy, a confident
assessment can be made in most instances. So, although the
original Duke article stated that “our … pathologic criteria are
essentially similar to those used in the definite category of von
Reyn, except for some minor improvements in wording” [23,
p. 202], we suggest that the rewording is not an improvement
because it dropped a positive microbiology Gram stain result
from being considered sufficient for the pathological definition
of endocarditis. We agree with the Beth Israel and St. Thomas
pathological definitions for endocarditis, which include microbiology Gram stain results [22, 25]. We suggest that future
modifications of the Duke criteria take our results into account.
Table 4.
Causes of incubating endocarditis discovered during valve replacement surgery for valve failure.
Microbiological findings
Patient
Infected valve
Surgical
Gram stain
Culture findings
Histopathological findings
1
Aortic, native
Cusp destruction with Not done
ragged edges
Enterococcus faecalis
Fibrinous nodules with PML infiltration; no
organisms seen
2
Aortic, allograft
Vegetations
NOS
Streptococcus mitis
Vegetations with PMLs and chronic
inflammatory cell infiltration; no
organisms seen
3
Aortic, native
Thickened soft valve
Gram-positive cocci
Streptococcus anginosus
Vegetations with PMLs and grampositive cocci
4
Mitral, mechanical Clot on suture line
Gram-positive cocci
Staphylococcus epidermidis Not done
5
Mitral, mechanical Partial dehiscence,
fibrinous deposits
Gram-positive cocci
S. epidermidis
6
Aortic, allograft
Vegetations
Gram-positive bacilli Corynebacterium species
7
Aortic, allograft
Vegetations
Gram-positive bacilli Corynebacterium species
Not done
8
Aortic, allograft
Vegetations
Not done
Vegetation with PML infiltration and gram-positive
cocci
9
Aortic, mechanical Partial dehiscence of
sewing ring
Gram-positive bacilli Corynebacterium species
Chronic inflammatory cells and gram-positive bacilli
10
Aortic, allograft
Gram-positive cocci
Vegetations with PML infiltration and gram-positive
cocci
Vegetations
Not done
Streptococcus oralis
Not done
No report in notes; slides and blocks not found
NOTE. Endocarditis was not suspected before the patient underwent surgery (i.e., the patient did not have fever and no blood samples were obtained for
culture). PML, polymorphonuclear leukocyte.
This is of particular relevance for prosthetic valve endocarditis,
for which material for histopathological examination is infrequently available.
We also suggest a change to the wording of the histopathological definition of endocarditis. The currently worded Duke
criteria for pathological lesions include “vegetation or intracardiac abscess confirmed by histologic examination showing
active endocarditis” [24, p. 636]. We believe this wording is
ambiguous because “active endocarditis” has been used previously to describe when surgery has been performed before
the end of antibiotic treatment or for when the valve culture
result is positive [1, 4, 34]. By “active endocarditis,” we believe
that the authors meant the presence of acute inflammatory cells
(i.e., PMLs) in the inflammatory cell infiltrate of the valve,
vegetation, or abscess wall tissue. We suggest that the word
“active” be removed and a phrase indicating the presence of
acute inflammatory cells in the vegetation or tissue be used to
define histopathological findings establishing endocarditis in
lesions when bacteria are not seen.
There were 5 instances in which valves were culture positive
when the patient was still receiving antibiotic treatment but
had completed SAT (table 3). It was rare, however, for commonly encountered organisms (i.e., staphylococci and streptococci) to be recovered from valve cultures after SAT had been
completed, a finding that provides support for commonly followed treatment guidelines [28].
Our results for patients with negative blood cultures are similar to those of Pesanti and Smith [36] in their series of 52
cases of endocarditis with negative blood culture results. In
their series, 25 patients had valvular tissue specimens obtained
for culture and Gram stain for organisms; 9 (36%) had organisms both seen and grown, and a further 6 (24%) had
organisms seen but not cultured [36]. Overall, 10 patients
(40%) had culture or Gram stain evidence of gram-positive
cocci. We saw or grew gram-positive cocci for 42 (67%) of 63
patients who had blood culture–negative endocarditis. In a
more recent study that described 32 patients with ⭓3 negative
blood culture results, the valve culture result was positive for
6 (19%), and, for 5 (83%) of these patients, fastidious organisms (i.e., nutritionally variant streptococci and a Haemophilus
species) were recovered [37]. Only one of our isolates, A. defectiva, can be regarded as fastidious. Our results support the
notion that initial treatment for blood culture–negative cases
can be based on the assumption that the majority of the organisms responsible are the same as those found in cases with
positive blood cultures.
To our knowledge, this study is the first to have reported a
series of cases of incubating endocarditis. We use this term
because we believe that these cases would have eventually become clinically evident. It seems highly improbable that they
would have resolved spontaneously. Isolated episodes of incubating endocarditis have been reported previously [4, 22, 38].
In conclusion, this report provides insight into the evolving
culture findings, Gram stain findings, and inflammation status
of heart valves before, during, and shortly after completing
treatment for infective endocarditis. We suggest that the microbiology Gram stain should be reinstated into the pathological criteria for definite endocarditis. Finally, we do not believe
positive Gram stain results for microbiological or histopathological specimens of resected valves necessarily indicate an in-
Gram Stain, Culture, and Histopathology in IE • CID 2003:36 (15 March) • 703
fected surgical field because of the considerable time delay between vegetation sterilization and disappearance of organisms.
20.
References
21.
1. Dinubile M. Surgery in active endocarditis. Ann Intern Med 1982; 96:
650–9.
2. Vlessis AA, Khaki A, Grunkemeier GL, Li H-H, Starr A. Risk, diagnosis
and management of prosthetic valve endocarditis: a review. J Heart
Valve Dis 1997; 6:443–65.
3. Moon MR, Stinson EB, Miller DC. Surgical treatment of endocarditis.
Prog Cardiovasc Dis 1997; 40:239–64.
4. Stinson E. Surgical treatment of infective endocarditis. Prog Cardiovasc
Dis 1979; 22:145–68.
5. Wilson W, Danielson G, Guiliani E, Washington J II, Jaumin P, Geraci
J. Cardiac valve replacement in congestive heart failure due to infective
endocarditis. Mayo Clin Proc 1979; 54:223–6.
6. Boyd A, Spencer F, Isom W, et al. Infective endocarditis: an analysis
of 54 surgically treated patients. J Thorac Cardiovasc Surg 1977; 73:
23–30.
7. Richardson J, Karp R, Kirklin J, Dismukes W. Treatment of infective
endocarditis: a 10-year comparative analysis. Circulation 1978; 58:
589–97.
8. Young J, Welton D, Raizner A, et al. Surgery in active infective endocarditis. Cardiovasc Surg 1979; 60:77–81.
9. Mills J, Utley J, Abbott J. Heart failure in infective endocarditis: predisposing factors, course and treatment. Chest 1974; 66:151–7.
10. Croft C, Woodward W, Elliott A, Commerford P, Bernard C, Beck W.
Analysis of surgical versus medical therapy in active complicated native
valve infective endocarditis. Am J Cardiol 1983; 51:1650–5.
11. Sethia B, Reece I, Forrester A, Davidson K. The surgical management
of extravalvular aortic root infection. Ann Thorac Surg 1984; 37:484–7.
12. David T, Bos J, Christakis G, Brofman P, Wong D, Feindel C. Heart
valve operations in patients with active infective endocarditis. Ann
Thorac Surg 1990; 49:701–5.
13. Kay P, Oldershaw P, Dawkins K, Lennox S, Paneth M. The results of
surgery for active endocarditis of the native aortic valve. J Cardiovasc
Surg (Torino) 1984; 25:321–7.
14. Pringle T, Webb S, Khan M, O’Kane H, Cleland J, Adgey J. Clinical,
echocardiographic and operative findings in active infective endocarditis. Br Heart J 1982; 48:529–37.
15. Anker E, Thaulow E, Forfang K, Rostad H. Surgical treatment of bacterial endocarditis: a review and follow-up of 36 patients. Acta Med
Scand 1981; 209:285–8.
16. Symbas P, Vlasis S, Zacharopoulos L, Lutz J. Acute endocarditis: surgical
treatment of aortic regurgitation and aortico-left ventricular discontinuity. J Thorac Cardiovasc Surg 1982; 84:291–6.
17. de Costa Lins R, Soares D, Van Berg L, et al. Surgical treatment of
active valvular infective endocarditis. Scand J Thorac Cardiovasc Surg
1988; 22:43–5.
18. Sweeney M, Reul G Jr, Cooley D, et al. Comparison of bioprosthetic
and mechanical valve replacement for active endocarditis. J Thorac
Cardiovasc Surg 1985; 90:676–80.
19. Ergin M, Raissi S, Follis F, Lansman S, Griepp R. Annular destruction
704 • CID 2003:36 (15 March) • Morris et al.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
in acute bacterial endocarditis: surgical techniques to meet the challenge. J Thorac Cardiovasc Surg 1989; 97:755–63.
Soyer R, Redonnet M, Bessou J, Mutel P, Hubscher C, Letac B. Valve
replacement in acute valve endocarditis. Thorac Cardiovasc Surg
1986; 34:149–52.
Manhas D, Mohri H, Hessel E, Merendina K. Experience with surgical
management of primary infective endocarditis: a collected review. Am
Heart J 1972; 84:738–47.
von Reyn CF, Levy BS, Arbeit RD, Friedland G, Crumpacker CS. Infective endocarditis: an analysis based on strict case definitions. Ann
Intern Med 1981; 94:505–18.
Durack DT, Lukes AS, Bright DK. New criteria for diagnosis of infective
endocarditis: utilization of specific echocardiographic findings. Duke
Endocarditis Service. Am J Med 1994; 96:200–9.
Li JS, Sexton DJ, Mick N, et al. Proposed modifications to the Duke
criteria for the diagnosis of infective endocarditis. Clin Infect Dis
2000; 30:633–8.
Lamas CC, Eykyn SJ. Suggested modifications to the Duke criteria for
the clinical diagnosis of native valve and prosthetic valve endocarditis:
analysis of 118 pathologically proven cases. Clin Infect Dis 1997; 25:
713–9.
Wells AU, Fowler CC, Ellis-Pegler RB, Luke R, Hannan S, Sharpe DN.
Endocarditis in a general hospital in Auckland, New Zealand. Q J Med
1990; 76:753–62.
Peat EB, Lang SDR. Infective endocarditis in a racially mixed community: a 10 year review of 78 cases. N Z Med J 1989; 102:33–6.
Wilson WR, Karchmer AW, Dajani AS, et al. Antibiotic treatment of
adults with infective endocarditis due to streptococci, enterococci,
staphylococci, and HACEK microorganisms. JAMA 1995; 274:1706–11.
Utley J, Mills J, Hutchinson J, Edmunds L, Sanderson R, Roe B. Valve
replacement for bacterial and fungal endocarditis: a report of six cases.
Am Surgeon 1964; 30:766–9.
Wilson W, Danielson G, Guiliani E, Washington J II, Jaumin P, Geraci
J. Valve replacement in patients with active infective endocarditis. Circulation 1978; 58:585–8.
Tuna I, Orszulak T, Schaff H, Danielson G. Results of homograft aortic
valve replacement for active endocarditis. Ann Thorac Surg 1990; 49:
619–24.
Becker R, Frishman W, Frater R. Surgery for mitral valve endocarditis.
Chest 1979; 75:314–9.
Kinsley R, Colsen P, Bakst A. Emergency valve replacement for primary
infective endocarditis. S Afr Med J 1978; 53:86–8.
Blumberg EA, Robbins N, Adimora A, Lowy FD. Persistent fever in
association with infective endocarditis. Clin Infect Dis 1992; 15:983–90.
Jung J, Saab S, Almond C. The case for early surgical treatment of leftsided primary infective endocarditis: a collective review. J Thorac Cardiovasc Surg 1975; 70:509–18.
Pesanti EL, Smith IM. Infective endocarditis with negative blood cultures an analysis of 52 cases. Am J Med 1979; 66:43–50.
Kupferwasser LI, Darius H, Müller AM, et al. Diagnosis of culture
negative endocarditis: the role of the Duke criteria and the impact of
transesophageal echocardiology. Am Heart J 2001; 142:146–52.
Chuard C, Antley CM, Reller LB. Clinical utility of cardiac valve Gram
stain and culture in patients undergoing native valve replacement. Arch
Pathol Lab Med 1998; 122:412–5.