In Vitro Resistance to Thrombin-Induced Platelet Microbicidal Protein in Isolates of Staphylococcus aureus from Endocarditis Patients Correlates with an Intravascular Device Source
In Vitro Resistance to Thrombin-Induced Platelet Microbicidal Protein in Isolates of Staphylococcus aureus from Endocarditis Patients Correlates with an Intravascular Device Source
Vance G. Fowler 0 1
Jr. 0 1
Lauren M. McIntyre 0 1 2
Michael R. Yeaman 0 1 3
Gail E. Peterson 0 1
L. Barth Reller 0 1 4
G. Ralph Corey 0 1
Dannah Wray 0 1
Arnold S. Bayer 0 1 3
0 Received 3 March 2000; revised 3 July 2000; electronically published 8 September 2000. Research guidelines of Duke University and UCLA were followed in the conduct of the clinical research in this study. Financial support: National Institutes of Health: AI-01647 (to V.G.F.), AI-39108 (to A.S.B.), and AI-39001 (to M.R.Y.). of Infectious Diseases, Duke University Medical Center , Durham, NC 27710
1 Department of Medicine, Divisions of
2 Department of Agronomy, Purdue University , West Lafayette, Indiana
3 St. John's Cardiovascular Research Center, Research and Education Institute, and Division of Infectious Diseases, Harbor-UCLA Medical Center, Torrance, and UCLA School of Medicine , Los Angeles, California
4 Clinical Microbiology Laboratory, Duke University Medical Center , Durham, North Carolina
Platelet microbicidal proteins (PMPs) are small antimicrobial peptides secreted by mammalian platelets. In vitro resistance of Staphylococcus aureus strains to PMPs correlates with more extensive disease in experimental infective endocarditis (IE). To determine whether this same relationship exists in human S. aureus IE, we evaluated the in vitro PMP susceptibility phenotype of isolates from 58 prospectively-identified patients with definite S. aureus IE. On multivariate analyses, patients with S. aureus IE complicating an infected intravascular device were significantly more likely to have IE caused by a PMP-resistant strain (P p .0193). No correlations were detected between in vitro PMP resistance among S. aureus strains and the severity of human IE. This work supports the concept that in vitro PMP resistance in clinical S. aureus strains is associated with important clinical characteristics of S. aureus endovascular infections in vivo.
A host defense role for platelets against endovascular
infections has been linked to platelet microbicidal proteins (PMPs)
. PMPs are released from platelets at sites of endovascular
damage or microbial colonization and kill common
bloodstream pathogens, including Staphylococcus aureus, in vitro.
Because the isolation of S. aureus from the bloodstream is
frequent, but S. aureus endovascular infections (e.g., infective
endocarditis [IE]) are uncommon, PMPs may play a role in
preventing serious endovascular infections .
Because organisms exhibiting PMP-resistance in vitro have
a distinct survival advantage for induction and/or propagation
of endovascular infections , we reasoned that in vitro PMP
resistance in clinical S. aureus strains might also be related to
a more severe clinical course in patients with IE caused by such
strains. Recent in vivo data support this hypothesis. Animal
models infected with a PMP-resistant strain of S. aureus had
S. aureus strains. Blood culture isolates were identified as S.
aureus and stored for future use, as described elsewhere .
Patient selection. Patients with S. aureus bacteremia were
prospectively identified between September 1994 and May 1998,
evaluated for IE, and followed up for 12 weeks after the date of the
initial positive blood culture, as described elsewhere [10, 11]. For
the current study, only isolates from patients with clinically or
pathologically definite IE were further analyzed. IE was defined
according to the Duke criteria . Embolic events were defined as
any clinical or radiographic evidence of thromboembolic episodes
(e.g., a cerebrovascular embolic event or a cutaneous or deep-tissue
infarct). A cerebrovascular embolic event was defined as the
presence of both clinical (i.e., focally abnormal neurologic examination
results) and radiologic evidence of an acute thromboembolic event.
Renal involvement complicating IE was defined as the presence of
hematuria plus a positive urine culture for S. aureus. Septic shock
was defined according to Consensus guidelines .
Echocardiography. Specific echocardiographic findings were
analyzed as individual indicators of IE, as described elsewhere ;
57 of the 58 study patients underwent some form of
echocardiography. Nine of the 58 study patients with definite IE underwent
transthoracic echocardiography (TTE) only, 4 underwent
transesophageal echocardiography (TEE) only, and 44 received both
TTE and TEE. All echocardiograms were read at the time of the
procedure by an experienced echocardiographer. To reduce
potential bias, all echocardiograms were later blindly reinterpreted by
one of the authors (G.E.P.). Vegetation size was calculated in square
millimeters, as described elsewhere . The median value was 61.5
mm2; vegetations 161.5 mm2 were defined as large.
In vitro PMP assays. Because all our prior data on the
relationships between in vitro PMP susceptibility phenotypes in
S. aureus and in vivo infection progression involved the use of
thrombin-induced PMP-1 (tPMP-1), we used this peptide in the
current study. The preparation of tPMP-1, bioactivity assays of
tPMP-1 preparations, and in vitro assays for tPMP-1 susceptibility
phenotypes were performed as detailed elsewhere . In vitro PMP
assays were performed previously in 18 of the 58 isolates as part
of another report .
Our previous in vitro analyses of common bloodstream
pathogens for tPMP-1 susceptibility profiles revealed little overlap of
strains from patients with well-defined IE versus strains from
patients without any evidence of IE, at a breakpoint of >70% survival
in the 2 h in vitro assay [2, 6]. For this reason, we categorized all
strains exhibiting >70% survival following a 2 h exposure to
tPMP1 as resistant for the current study. All assays and data analyses
of tPMP-1 in vitro studies were performed by 2 investigators
(A.S.B. and M.R.Y.), who were blinded as to the clinical details
of the patients.
Statistical analyses. Bivariate analyses were conducted
between in vitro tPMP-1 susceptibility profiles on individual patient
isolates (i.e., resistant vs. susceptible) and clinical outcome variables
in these patients with IE, using logistic regression. Odds ratios
(ORs) were computed, and the Wald x2 test was used to determine
significance; inferences were confirmed using exact methods .
Relationships were considered significant when the 2-sided P value
was !.05 (table 1). The OR for relapse compared with all other
outcome variables was not estimable, because none of the patients
with IE caused by a tPMP-1resistant S. aureus isolate experienced
Platelet Microbicidal Protein in S. aureus IE
a clinical or microbiologic relapse of their infection. Therefore,
relapse was not considered in further statistical evaluation models.
To identify any independent associations between in vitro
tPMP1 resistance and specific clinical parameters of the IE study
population, multivariate logistic regression analysis was performed.
Clinical variables found to be significant in the bivariate analyses
were considered as candidates for inclusion in the multivariate
analysis model. Thus, we modeled hemodialysis dependence, injection
drug use, and intravascular device source in a multivariate logistic
regression. Hemodialysis dependence did not independently
contribute significantly in the multivariate model, owing to a
correlation between hemodialysis dependence and an intravascular
device source (Kendalls t, 0.64). Therefore, our final multivariate
model contained injection drug use and intravascular device source
as potential independent correlates of patient outcomes.
Patient selection, categorization, and clinical findings.
Sixtyfive patients had definite S. aureus IE . Of these 65 patients,
blood culture isolates from 58 patients were available for
tPMP1 susceptibility profiling. Among these 58 patients with definite
IE, 54 were clinically confirmed and 4 were histopathologically
confirmed. Data from these 58 patient-isolate pairs formed the
basis for the further analyses in this investigation. Twenty (34%)
of the 58 patients were infected with methicillin-resistant S.
aureus. Twenty-eight patients (48%) had community-acquired
IE, whereas 30 patients (52%) had hospital-acquired or nursing
homeacquired IE. Injection drug use was the presumed route
of infection in 8 patients, all of whom developed
communityacquired S. aureus IE. Four of these 8 patients had right-sided
IE, 3 had left-sided IE, and 1 patient had both left- and
Embolic events (n p 23) were confirmed by computed
tomography in 11 patients (cerebrovascular and/or deep-tissue
embolic events), by chest radiograph in 3 patients (septic
pulmonary emboli), and by clinical examination in 9 patients with
Relationship of clinical and echocardiographic features of IE
with tPMP-1 susceptibility profiles. Isolates were susceptible
to tPMP-1 in 19 patients (33%) and resistant to tPMP-1 in 39
isolates (67%; table 1). Patients with tPMP-1resistant strains
were significantly more likely to have developed IE as a
consequence of an infected intravascular device (OR, 9.524; 95%
confidence interval [CI], 2.62846.306; P p .002). This
association between tPMP-1 resistance and the presence of an
intravascular device persisted when multivariate logistic
regression analysis was employed (OR, 5.77; 95% CI, 1.44529.543;
P p .0193). These patients with tPMP-1resistant strains were
also more likely to be hemodialysis dependent than were
patients with IE due to tPMP-1-susceptible strains.
Several clinical and echocardiographic parameters had been
predicted a priori to be associated with tPMP-1 resistance in
vitro (table 1). No statistically significant association was
detected between tPMP-1 resistance in vitro and any of these
clinical or echocardiographic parameters.
The relationships of various study patient characteristics
were evaluated. The presence of an intravascular device was
strongly correlated with dependence on hemodialysis (Kendalls
t, 10.63623; P p .0001) but was negatively correlated with
injection drug use (Kendalls t, 20.38644; P p .0035). Thus,
the correlation noted above between tPMP-1 resistance in vitro
and hemodialysis dependence was based in the intravascular
device infection source, rather than in any parameters specific
for hemodialysis. Injection drug use was strongly correlated
with community acquisition (Kendalls t, 1 0.41; P p .0018).
Patients who were injection drug users were more likely to
be infected with tPMP-1susceptible strains (7 of 8 patients).
This association was present in both bivariate (OR, 0.045; 95%
CI, 0.0020.288; P p .0056) and multivariate (OR, 0.099; 95%
CI, 0.0050.692; P p .045) logistic regression analyses.
The present study demonstrates that patients with S. aureus
IE complicating an infected intravascular device were 10-fold
more likely to have tPMP-1resistant blood culture isolates
than were patients with IE arising from either an inapparent
primary focus or a noncatheter source. This finding is consistent
with previous results . The association between intravascular
device infection and tPMP-1resistant staphylococci is also
consistent with the clinical observation that an indwelling
catheter is, overall, an important risk factor for invasive
staphylococcal infections [11, 15].
Hemodialysis dependence was significantly associated with
IE caused by tPMP-1resistant strains in bivariate analyses.
This finding is consistent with previous observations that
hemodialysis-dependent patients are at increased risk for
staphylococcal infection [10, 15] and may be due to the strong
correlation between hemodialysis dependence and the presence of
an intravascular device, rather than any intrinsic characteristics
of the hemodialysis population. However, other features of this
patient subset might also contribute to the propensity of their
tPMP-1resistant staphylococcal strains to induce IE (e.g.,
platelet dysfunction related to renal failure ).
When S. aureus colonization occurs at vascular catheter sites,
platelets accrue at such sites and likely release PMPs. Thus, the
survival of colonizing S. aureus strains at these sites would
depend, at least in part, on their intrinsic tPMP-1 susceptibility
profile . It follows, then, that intravascular infections caused
by tPMP-1resistant strains of S. aureus may be more likely to
disseminate and induce IE than are tPMP-1susceptible
counterpart strains. More than two-thirds of the bacteremic S. aureus
isolates from this patient cohort with IE were resistant to
tPMP1 in vitro, paralleling our prior observations that
tPMP-1resistant strains are associated with a more progressive form of
experimental IE than that seen with isogenic tPMP-1susceptible
strains [4, 5]. These differences in clinical severity between
tPMP1resistant and tPMP-1susceptible strains in experimental IE
have been ascribed to an enhanced capacity of the resistant
bacteria to survive at, and disseminate from, the infected vegetation
site, in the presence of locally secreted tPMP-1 .
Clinical and echocardiographic features associated with
tPMP-1 resistance in animal IE models  were not found to
be significantly associated with tPMP-1 resistance in the clinical
IE strains evaluated in this study. There are several possible
explanations for these latter disparities. First, the duration of
clinical IE varied widely, providing a potentially confounding
variable. For example, the reported duration of symptoms
among the study patients ranged from 1 to 14 days. This wide
range in disease duration in patients with clinical IE provided
an inconstant interval for the development of clinical
complications and/or larger valvular vegetations, regardless of the
tPMP-1 susceptibility profile of their infecting strains. By
contrast, the duration of experimental IE in the animal studies was
equivalent in animals infected by either the tPMP-1susceptible
or tPMP-1resistant strains. Secondly, experimental IE is
induced by a large intravenous bacterial inoculum given as a
bolus challenge (106107 colony-forming units) , thus
magnifying the chances of developing complications of IE at the
site of catheter-induced valvular trauma. Last, our relatively
small sample size may have hindered detection of clinical or
echocardiographic parameters that are associated with in vitro
Of the 8 injection drug users in our study population, 7 (88%)
were infected with tPMP-1susceptible isolates. This
association may be due in part to colonization of these injection
drugusing patients by a clonally related staphylococcal strain
. Outbreaks of infection due to a single bacterial strain
among drug users sharing injection paraphernalia commonly
occur . Similarly, it is possible that tPMP-1 resistance may
be associated with clonally related S. aureus strains among the
patients in this current report. Although we cannot exclude this
possibility, pulsed-field gel electrophoresis (PFGE) data
obtained from a subset of study isolates suggest that multiple
strains were present. None of the 5 isolates in the present study
that were previously evaluated by PFGE  shared identical
PFGE banding patterns.
In summary, the present study further supports the concept
that, among bacteremic S. aureus strains, resistance to
tPMP1 is associated with specific clinical characteristics of S. aureus
endovascular infections in vivo. The presence of an
intravascular device in a patient with S. aureus IE is strongly associated
with resistance to tPMP-1 in the infecting strain. Moreover,
these data underscore the concept that PMPs play an important
role in limiting IE in the setting of vascular device infections
caused by susceptible strains; in contrast, it appears that
resistant strains can circumvent PMP activities at sites of vascular
catheter infections and have a survival advantage in terms of
inducing IE. Future studies will be necessary to identify whether
tPMP-1 resistance in a pathogen infecting an intravascular
device also increases the likelihood of local or metastatic
complications in addition to IE. Moreover, the mechanism by which
tPMP-1 resistance develops in such strains remains to be
1. Azizi N , Li C , Shen AJ , Bayer AS , Yeaman MR . Staphylococcus aureus elicits release of platelet microbicidal proteins in vitro [abstract no. G54] . In: Programs and abstracts of the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy . Washington, DC: American Society for Microbiology, 1996 : 153 .
2. Bayer AS , Cheng D , Yeaman MR , et al. In vitro resistance to thrombininduced platelet microbicidal protein among clinical bacteremic isolates of Staphylococcus aureus correlates with an endovascular infectious source . Antimicrob Agent Chemother 1998 ; 42 : 3169 - 72 .
3. Dhawan VK , Yeaman MR , Cheung AL , Kim E , Sullam PM , Bayer AS . Phenotypic resistance to thrombin-induced platelet microbicidal protein in vitro is correlated with enhanced virulence in experimental endocarditis due to Staphylococcus aureus . Infect Immun 1997 ; 65 : 3293 - 9 .
4. Dhawan VK , Bayer AS , Yeaman MR . In vitro resistance to thrombin-induced platelet microbicidal protein is associated with enhanced progression and hematogenous dissemination in experimental Staphylococcus aureus infective endocarditis . Infect Immun 1998 ; 66 : 3476 - 9 .
5. Kupferwasser LI , Yeaman MR , Shapiro SM , Nast CC , Bayer AS . Beneficial effects of thrombin-induced platelet microbicidal peptide in experimental Staphylococcus aureus endocarditis . Circulation 1999 ; 100 (Suppl): 1163A .
6. Wu T , Yeaman MR , Bayer AS . In vitro resistance to platelet microbicidal protein correlates with endocarditis source among bacteremic staphylococcal and streptococcal isolates . Antimicrob Agent Chemother 1994 ; 38 : 729 - 32 .
7. Yeaman MR , Puentes SM , Norman DC , Bayer AS . Partial purification and staphylocidal activity of thrombin-induced platelet microbicidal protein . Infect Immun 1992 ; 60 : 1202 - 9 .
8. Yeaman MR , Sullam PM , Dazin PF , Bayer AS . Platelet microbicidal protein alone and in combination with antibiotics reduces Staphylococcus aureus adherence to platelets in vitro . Infect Immun 1994 ; 62 : 3416 - 23 .
9. 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 .
10. Fowler VGJ , Kong LK , Corey GR , et al. Recurrent Staphylococcus aureus bacteremia: pulsed-field gel electrophoresis findings in 29 patients . J Infect Dis 1999 ; 179 : 1157 - 61 .
11. Fowler VGJ , Sanders LL , Sexton DJ , et al. Infective endocarditis due to Staphylococcus aureus: 59 prospectively identified cases with follow-up . Clin Infect Dis 1999 ; 28 : 106 - 14 .
12. American College of Chest Physicians/Society of Critical Care Medicine Consensus Conference . Definitions for sepsis and organ failure and guidelines for the use of innovative therapies in sepsis . Crit Care Med 1992 ; 20 : 864 - 74 .
13. Sanfilippo AJ , Picard MH , Newell JB , et al. Echocardiographic assessment of patients with infectious endocarditis: prediction of risk for complications . J Am Coll Cardiol 1991 ; 18 : 1191 - 9 .
14. Good P. Permutation tests: a practical guide to resampling methods for testing hypotheses . New York : Springer, 1994 .
15. Hoen B , Paul-Dauphin A , Hestin D , Kessler M. EPIBACDIAL: a multicenter prospective study of risk factors for bacteremia in chronic hemodialysis patients . J Am Soc Nephrol 1998 ; 9 : 869 - 76 .
16. Andrassy K , Ritz E , Bommer J. Effects of hemodialysis on platelets . Contributions Nephrol 1987 ; 59 : 26 - 34 .
17. Tuazon CU , Sheagren JN . Increased rate of carriage of Staphylococcus aureus among narcotic addicts . J Infect Dis 1974 ; 129 : 725 - 7 .
18. Craven DE , Rixinger AI , Goularte TA , McCabe WR. Methicillin-resistant Staphylococcus aureus bacteremia linked to intravenous drug abusers using a shooting gallery . Am J Med 1986 ; 80 : 770 - 6 .