Abstract
BACKGROUND/AIMS. The aetiology of infective endocarditis (IE) can be determined directly from surgically removed valve tissue using broad‐range bacterial rDNA polymerase chain reaction (PCR) followed by sequencing. We sought to assess the value of this methodology in a routine clinical setting.
METHODS. Broad‐range PCR with primers targeting conserved bacterial rDNA sequences was applied to directly analyse valve samples from 56 patients operated on for diagnosed or suspected IE. Identification of the aetiological agent was performed by partial DNA sequencing of the 16S and 23S rDNA genes.
RESULTS. The final diagnosis was definite IE in 36 patients and possible IE in 2 patients, while the diagnosis of IE was rejected in 18 patients. PCR analysis from removed valve tissue was positive in 25 patients with IE. Molecular identification was consistent with the blood culture finding in 20 of these patients. The PCR approach was the only method to yield the aetiological diagnosis in additional 4 patients (2 Staphylococcus species, 1 Streptococcus bovis, 1 Bartonella quintana), all of whom had received antimicrobials before blood cultures were taken. The mean duration of preoperative antimicrobial treatment for the patients with PCR‐positive valves was 19.6 (range 1–58) days.
CONCLUSIONS. Bacterial DNA may persist during treatment in infected valves for long periods. The PCR method is especially useful when the causative agent of IE is fastidious or when the specimen is taken during antimicrobial treatment.
| Abbreviations | ||
| CRP | = | C‐reactive protein |
| IE | = | infective endocarditis |
| PCR | = | polymerase chain reaction |
Introduction
Since the beginning of the 1990s, we have used broad‐range bacterial polymerase chain reaction (PCR) targeting rRNA genes to identify bacterial pathogens directly from diseased tissue Citation1–4. The use of this method in patients with infective endocarditis (IE) to identify rare and/or difficult‐to‐culture causative agents has previously been described in several case reports Citation5–9. Goldenberger et al. Citation10 were the first to publish their results on the systematic use of the PCR approach in the aetiological diagnostics of IE in a clinical setting. In the 21st century, a number of papers from several groups have addressed the utility of the molecular methods in different patient populations to identify the aetiology of endocarditis from resected valve tissue Citation11–17. The PCR approach has been found useful especially in blood culture‐negative IE and in IE caused by fastidious or non‐culturable microbes, the foremost being the Bartonella species, Coxiella burnetii, and Tropheryma whippleiCitation16,Citation18,19.
In 1994, broad‐range bacterial rDNA PCR combined with species identification by DNA sequencing was adopted as a routine diagnostic tool to examine valve tissue removed during thoracic surgery from patients treated for diagnosed or suspected IE in the Infectious Diseases Unit, Turku University Hospital, Turku, Finland. We have evaluated the usefulness of this method in the aetiological diagnosis of our patients undergoing surgery for IE, focusing special attention on the persistence of bacterial DNA in valve tissue during antimicrobial treatment of endocarditis. In addition, the association between the persistence of bacterial DNA in infected heart valves and the clinical condition of the patient at the time of cardiac surgery was evaluated.
Key messages
Broad‐range bacterial rDNA polymerase chain reaction (PCR) followed by sequencing from surgically removed valve tissue is valuable in the aetiological diagnostics of blood culture‐negative infective endocarditis (IE).
In a clinical setting, it is not necessary to employ the PCR approach to examine valve tissue from patients, in whom preoperative blood cultures have yielded the aetiological diagnosis of IE.
The PCR assay cannot be used to monitor the efficacy of antimicrobial therapy in IE.
Patients and methods
Patients
From the beginning of 1994 to the end of 2004, 56 valve samples from patients undergoing surgery for diagnosed or suspected IE were analysed by the broad‐range bacterial PCR. Additionally, conventional microbiological methods were used Citation4. The PCR assays were requested by the attending clinicians and performed at the Department of Medical Microbiology, University of Turku. The patients included in this study form a prospective part of a larger follow‐up study on patients treated for IE in our hospital from 1980 onwards. For these patients, details regarding the diagnostic classification, the presence of neurological manifestations, and the utility of the serum C‐reactive protein (CRP) value in assessing the outcome of IE have been described earlier Citation20–22. Turku University Hospital is a 1000‐bed teaching facility with a cardiothoracic department, serving as a tertiary referral center for the southwestern part of Finland.
For each patient, data were collected on age, sex, blood culture findings, involved valves, preoperative CRP value, the duration of preoperative antimicrobial therapy and the indication for valve surgery, as well as the microbiological and histological findings of the tissue specimens taken during cardiac surgery. In addition, each case of IE was classified both preoperatively and postoperatively using the Duke diagnostic criteria Citation23. The recommendations given by the American Heart Association Citation24 and the European Society of Cardiology Citation25 for the treatment of IE were used to assess whether the antimicrobial treatment given to each patient before surgery could be deemed adequate, i.e. effective and long enough towards the respective causative agent to constitute a complete course of IE. The clinical response to therapy was evaluated based on resolution of fever and normalization of the parameters of inflammation. For the purposes of the present study, a patient was defined as having had adequate antimicrobial treatment before surgery if the duration of the preoperative antimicrobial treatment effective for the causative agent was ⩾28 days, the clinical symptoms of infection had resolved, and the preoperative CRP value was ⩽15 mg/L.
DNA purification
DNA was extracted from the fresh tissue samples after proteinase K (0.1 mg/mL) digestion (56°C, 2–17 hours) with two phenol‐chloroform‐isoamyl alcohol extractions followed by one ether wash, as described earlier Citation6.
PCR and sequencing
The primers, reagents, and conditions used in the broad‐range bacterial 23S and 16S rDNA PCR assays have been described previously Citation2. All samples were initially screened for the presence of bacterial DNA by amplification of the 23S rRNA gene with oligonucleotide primers MS 37 and MS 38 Citation2. Special care was taken to avoid carryover contamination of samples with amplicons. Strict measures were employed to separate the pre‐PCR facilities from the post‐PCR areas. Gamma irradiation was used to eliminate possible remnants of environmental DNA from water, and UV light treatment of the PCR tubes containing the final PCR mix was used before addition of template Citation1,Citation26. Inhibition of the PCR was ruled out by successful amplification of beta globin gene sequences. Bacterial identification was performed by sequencing of either or both of the partially amplified 23S or 16S rRNA genes. The 16S rDNA PCR product was preferred for sequencing because of the more abundant sequence data available in public databanks, whereas the 23S rDNA assay provided better sensitivity. Sequence analysis was performed as described previously Citation2–4.
Statistical analyses
Fisher's exact test and two‐tailed t test were used in statistical analyses.
Results
The valve samples of a total of 56 patients operated on for diagnosed or suspected IE were examined. The study collection included the valves of 81.2% (56 patients) of all 69 patients who had undergone cardiac surgery for diagnosed or suspected IE during the study period. Results of the microbiological findings and preoperative serum CRP value for each patient are given in , as are the data regarding the final clinical diagnosis, indication for valve surgery and the duration of antimicrobial treatment before surgery.
Table I. Microbiological findings and clinical characteristics of 56 patients operated on for diagnosed or suspected infective endocarditis (IE).
Preoperative classification of IE
Of the 56 patients, 28 were preoperatively defined as having definite IE (patients 1–14, 16–20, 24–30, 33 and 34), and 10 possible IE (patients 15, 21–23, 31, 32 and 35–38). For 18 patients (patients 39–56), the clinical suspicion of IE was not high, but the operating surgeon considered this disease as one possible explanation for the development of rapidly increasing valvular regurgitation, prosthetic valve dehiscence, or acute heart failure, and decided to send the valve samples also for microbiological examination. These included three patients who were operated on for increasing valvular regurgitation after the completion of antimicrobial treatment for IE 4 months, 27 months, and 11 years earlier, respectively.
Microbiological and histological findings
PCR‐positive valves (Table )
Valve tissue material of 25 patients contained bacterial rDNA by PCR analysis. Preoperative blood culture analysis as well as bacterial culture and Gram staining of resected material were positive in five of these patients, these findings being consistent with the direct molecular identification. In all five, the causative bacterial species was identified by sequencing at least to the genus level: two cases were caused by the Streptococcus species, two cases by the Staphylococcus species, and one case by Enterococcus faecalis. Histological examination revealed an acute inflammatory cell infiltrate as a sign of acute IE in all of these valves.
The remaining 20 broad‐range rDNA PCR‐positive valve specimens were negative by bacterial culture. Bacterial identification by sequencing was successful in 19 of these specimens, the result being consistent with the preoperative blood culture finding in 15 patients. The causative agent of IE was Staphylococcus species in seven patients, Streptococcus species in four patients, Abiotrophia species in two patients, and Streptococcus pneumoniae in two patients. In four patients with PCR‐positive but culture‐negative valves, blood cultures were negative and the aetiology of IE was identified only by the PCR method. These causative agents included Staphylococcus aureus, Staphylococcus species,Streptococcus bovis, and Bartonella quintana, one case of each. Of these 19 PCR‐positive valve specimens, histological examination revealed signs of acute IE in 18 and no signs of acute IE in one (patient 19). Sequencing was not successful from one remaining PCR‐positive valve specimen (patient 25), which showed no signs of IE in histological examination. Of these 20 PCR‐positive but culture‐negative valve specimens, Gram staining was positive in 3, negative in 16, and not performed in 1.
PCR‐negative valves (Table )
In 31 patients, the valves were negative by both broad‐range bacterial rDNA PCR analysis and bacterial culture. Included were nine patients with a preoperative diagnosis of definite or possible IE, in whom blood cultures yielded the aetiological diagnosis. Among these nine patients, histological examination revealed signs of acute IE in the valve specimens of four (patients 26, 27, 30, 33) and no signs of IE in the specimens of five (patients 28, 29, 31, 32, 34). In patient 30, whose blood cultures grew Streptococcus pneumoniae, Gram staining revealed few Gram‐positive cocci on the valve tissue.
In four patients with prosthetic valves (patients 35–38), all microbiological findings remained negative, but the histological findings revealed acute IE. In the remaining 18 patients, microbiological findings were negative and histological examination of the valve specimens revealed no signs of IE.
Final clinical diagnoses
The diagnosis remained unchanged at surgery for the 28 patients with a preoperative diagnosis of definite IE. In eight patients, the diagnosis changed from possible to definite IE based on findings consistent with IE in the histological examination of the removed valve tissue. Included were four patients with native valve IE (patients 15, 21–23) and four patients with prosthetic valve endocarditis (patients 35–38). In two patients (patients 31 and 32), the final diagnosis was possible IE. In 18 patients (patients 39–56), the diagnosis of IE was rejected. The clinical characteristics of the patients with a final diagnosis of IE are presented in . Surgery revealed intracardiac abscesses in seven patients (patients 3–5, 11, 15, 16, 21).
Table II. Characteristics of 38 episodes of infective endocarditis.
Adequacy of preoperative antimicrobial treatment of IE
Preoperative antimicrobial treatment was deemed inadequate in all five patients with PCR‐positive and culture‐positive valves (patients 1–5) on the basis of too short a duration. Of the 19 patients with culture‐negative IE from which bacterial DNA sequences were successfully identified, antimicrobial treatment was deemed adequate in 2 (patients 7 and 10). Antimicrobial treatment was deemed adequate also for one additional patient with a PCR‐positive valve specimen, from which sequencing did not reveal specific identification from the amplified rDNA product of the expected length (patient 25). For the remaining 17 patients, either the duration of the treatment was too short and/or the response to antimicrobial agents poor on the basis of clinical signs and/or elevated preoperative CRP values.
Of the 13 patients with IE who had PCR‐negative and culture‐negative valves, preoperative antimicrobial therapy was deemed adequate in 5 (patients 26, 28, 29, 31, 32) and inadequate in 8 (patients 27, 30, 33–38).
Association between microbiological findings and clinical characteristics
The duration of preoperative antimicrobial treatment in PCR‐positive versus PCR‐negative patient groups was compared. The mean duration (SD) of preoperative antimicrobial treatment was 19.6 days (18 days) (range 1–58 days) for the 25 patients with IE who had PCR‐positive valves, and 22.6 days (23 days) (range 0–60 days) for the 13 patients with IE who had PCR‐negative valves (P = 0.65, two‐tailed t test). The difference in the proportion of patients having received inadequate preoperative antimicrobial treatment in PCR‐positive versus PCR‐negative groups was also tested. Antimicrobial treatment was deemed inadequate in 22 (88%) of the 25 patients with PCR‐positive valves and in 8 (62%) of the 13 patients with PCR‐negative valves (P = 0.06; Fisher's exact test).
The mean duration of preoperative antimicrobial treatment was 4 days (2.3 days) (range 2–8 days) for the 5 patients with IE whose valve specimens were positive by both PCR and culture, and 23.5 days (15.5 days) (range 1–58 days) for the 20 patients with IE whose valve specimens were positive by PCR but negative by culture (P<0.001, two‐tailed t test for differing variances).
In‐hospital mortality of IE was 26.3%. Mortality of IE was 32% (8/25) for the patients with PCR‐positive valves and 15.4% (2/13) for those with PCR‐negative valves (P = 0.18, Fisher's exact test).
Accuracy of microbiological methods
When the final diagnosis of definite IE or possible IE were regarded as the golden standard, the sensitivity, specificity, and positive and negative predictive values of the PCR method were 65.8%, 100%, 100%, and 58%, respectively. In these analyses, the positive PCR finding based on an amplicon of the expected length as visualized by agarose gel electrophoresis from patient 25 was defined as true positive, although sequencing was not successful. The sensitivity, specificity, and positive and negative predictive values of the bacterial culture of valve tissue were 13.2%, 100%, 100%, and 35%, respectively, and of the Gram staining, 25%, 100%, 100%, and 37%, respectively.
Discussion
In this prospective clinical investigation, we used broad‐range bacterial PCR combined with sequencing to study valve specimens from patients undergoing surgery for diagnosed or suspected IE. Most of the aetiological agents identified from affected valve tissue were common pathogens causing IE, which had been identified already preoperatively based on blood culture findings. In four patients, the PCR approach was the only method to provide the aetiological diagnosis. One case involved the first and, so far, only patient with diagnosed Bartonella endocarditis in Finland Citation6. As described previously, the diagnosis was confirmed by demonstration of a strongly positive titre of serum antibodies to Bartonella quintanaCitation6. For more specific Bartonella quintana serologic analysis, Western immunoblotting has also been successfully used Citation27, but this assay was not available when our patient was examined. The other causative agents that were identified based on presence of specific rDNA sequences from blood culture‐negative IE patients were Staphylococcus aureus, Staphylococcus species, and Streptococcus bovis, one case of each in three patients who had all received antimicrobial therapy before blood cultures were taken. Thus, our results corroborate those of previous studies in that the broad‐range bacterial rDNA PCR approach is useful especially when the causative agent is fastidious or when the patient has preoperatively received antimicrobial treatment leading to negative blood cultures Citation3,4,Citation10,Citation15.
In the present study, two cases of Abiotrophia spp. endocarditis were identified by the PCR approach from the removed valve tissue. Members of the genus Abiotrophia, formerly classified as nutritionally variant streptococci, are rare but important causative agents of endocarditis Citation28. These organisms do not usually grow on conventional blood agar plates; thus, Abiotrophia endocarditis is sometimes falsely grouped into the category of blood culture‐negative endocarditis. In both of our patients, however, the organism was preoperatively isolated from blood cultures.
The mean duration of the preoperative treatment in the patients with broad‐range bacterial rDNA PCR‐positive valves was 19.6±18 days, indicating that bacterial DNA may persist in infected valves for several weeks during an apparently effective antimicrobial therapy. In our patients, the causative agent of IE was recognizable by PCR from resected valve tissue for up to 58 days after commencement of parenteral antimicrobial treatment, whereas bacterial cultures of valve tissue became negative within only a few days. This renders the PCR approach especially useful in the aetiological diagnostics of patients with blood culture‐negative IE, who have been treated with antimicrobials for long periods before cardiac surgery becomes necessary due to heart failure or rapidly progressing valvular regurgitation. In a clinical setting, the PCR approach is most beneficial in this type of patient with IE. It is not necessary to employ the molecular methods to examine valve tissue from patients, in whom preoperative blood cultures have already yielded the aetiological diagnosis of IE.
In the beginning of this study, we regarded the finding of long‐term persistence of bacterial DNA on heart valves as perplexing, even confusing. However, the same finding has been made by other groups. For example, in the series of IE patients reported by Podglajen et al. Citation14, the mean duration of the preoperative antimicrobial treatment was 24.6 days (range, 1–75 days) for the patients with PCR‐positive valves. Our initial hypothesis was that a positive PCR test from the valve tissue might indicate an inadequate response to the preoperatively given antimicrobial treatment, in which case the PCR result could be used as an aid to monitor the efficacy of antimicrobial treatment. To test this assumption, we analysed the association between the positive PCR test and clinical and histopathological findings of the patient at the time of cardiac surgery. Among the 25 patients with PCR‐positive valves, the preoperative antimicrobial treatment was deemed inadequate on clinical basis in 22 (88%) patients. Moreover, of all 24 patients with PCR‐positive valves from which sequencing was successful, all but one (96%) showed signs of acute IE in histological examination. Bearing in mind that histopathological examination of affected heart valves is regarded as the gold standard in the diagnosis of IE Citation17, the association between the positive PCR test and an inadequate response to antimicrobial treatment cannot be excluded based only on the findings of the present study.
The association between the molecular and histological findings from resected valves has been evaluated also in other studies. In the patients examined by Gauduchon et al. Citation13, the presence of bacterial DNA in heart valve tissue correlated with the presence of histological findings of IE, while PCR amplification was always negative when the patient had no histological evidence of IE. None of the three patients in the present study who underwent surgery after successful treatment of IE exhibited bacterial DNA in their heart valves. Yet, a few other studies have shown that bacterial DNA is detectable in the valve tissue even after the completion of antimicrobial therapy for IE. Of particular note are the results of Rovery et al. Citation17 showing that the PCR result is sometimes positive for years after an apparently successful treatment of IE. They found specific bacterial DNA in valves removed from 11 of 30 patients who had completed the treatment for IE; of note, 6 patients with PCR‐positive valves up to 850 days after the diagnosis had no histological evidence of IE. Based on these findings, the authors warn against using the PCR method as a tool to monitor the treatment of IE Citation17. Similar findings have been made by Lang et al. Citation29 who found bacterial DNA in the valves of three patients in whom IE had been treated by antibiotic therapy 5, 12 and 18 months previously. They conclude that demonstration of bacterial DNA within resected heart valves suggests either recurrence of infection, treatment failure or the persistence of bacterial debris within the cardiac vegetation. At the present time, factors associated with the persistence of bacterial DNA in heart valves remain undefined. Further studies are warranted to seek a deeper understanding of this phenomenon. It remains possible that the capacity of many causative agents of IE to produce biofilms Citation30 may contribute to the bacterial nucleic acid persistence in the valve tissue.
In contrast, it is quite evident that a negative PCR test at the time of surgery does not mean that the patient has responded favourably to the given antimicrobial therapy. Of all 13 patients with definite IE who had PCR‐negative valves in this study, the antimicrobial treatment was deemed inadequate in as many as 8 (62%) patients.
Recently, it has been proposed that the molecular diagnostics should be included in the Duke diagnostic classification as a major criterion Citation11,12. The proposal is based on previous findings showing that a number of cases of possible IE can be classified as definite, if the results of the molecular tests are taken into account. In the present study, inclusion of the PCR results would not have affected the reclassification. In four patients with PCR‐positive valves, the diagnosis was postoperatively changed from possible IE to definite IE, but this was done on the basis of histological findings. On the whole, the concept of including molecular diagnostics into the classification schema appears sensible and would certainly facilitate the aetiological diagnosis of blood culture‐negative IE. It is to be hoped that some unsolved questions regarding the persistence of bacterial nucleic acids in heart valves after IE will be answered first.
Among the microbiological tests on the removed valve tissue, PCR was associated with the highest sensitivity (65.8%). In the patients with IE, PCR was positive from all 5 valves from which bacteria were cultured and, in addition, from 20 valves which were culture‐negative. On the other hand, all 18 patients with no IE had PCR‐negative valves, indicating an excellent specificity (100%) of this test.
In the present study, the results from direct molecular analysis and bacterial culture were identical for ten specimens and to some extent different for three specimens, while for seven specimens the causative agent was identified by sequencing only to the genus level. It is of note that this prospective 11‐year study was based on a broad‐range bacterial rDNA PCR methodology developed for routine clinical analysis. Accuracy of molecular identification within certain bacterial genera may be improved by amplification of a more variable target gene than rDNA, e.g. use of rpoB encoding the RNA polymerase beta subunit has allowed a better resolution among the Staphylococcus species Citation31. However, use of rDNA as primary target remains an excellent routine approach because of the extensive databases. Thus far, multibacterial infections are problematic in broad‐range PCR‐based routine applications. Extensive molecular cloning and sequencing experiments of clone libraries as described in other settings Citation32,33 are not feasible due to high material cost and time limitations. In cases of endocarditis, dual or polymicrobial infections are admittedly rare. However, novel technologies, e.g. based on diagnostic arrays of DNA probes, may in the future provide useful tools for more accurate microbial identification also of mixed bacterial infections directly from clinical samples Citation34.
Although for patient 25, a PCR product of the expected length was observed by agarose gel electrophoresis, direct sequencing did not provide a readable signal. This was probably due to the low quality and quantity of the original sample and amplicon, respectively. It is of note that cloning experiments were not performed in this study. Also the negative PCR finding on the valve tissue of patient 30, which showed only a few Gram‐positive cocci on Gram staining, deserves attention. The patient's blood cultures grew Streptococcus pneumoniae, and the duration of the preoperative antimicrobial treatment was only 3 days. Thus, the negative PCR result may be explained e.g. by a difference in sampling of the study specimens. Inhibition of the PCR was ruled out by successful amplification of beta globin gene sequences from the clinical samples.
In conclusion, the PCR approach was especially useful, when the causative agent of IE was fastidious or when the specimen was taken during antimicrobial treatment. In a clinical setting, it is reasonable to limit the use of the PCR approach from resected valve tissue to patients with blood culture‐negative IE. Bacterial DNA may persist during treatment in infected valve tissue for long periods, up to 58 days in the present study. The PCR assay cannot be used to monitor the efficacy of antimicrobial therapy in IE. The implications of nucleic acid persistence remain undefined at the present time. Further studies are warranted to seek a deeper understanding of this phenomenon.
Conflicts of interest
None of the authors have any conflicts of interest in connection with this paper and none have received financial support for the study.
Acknowledgements
We thank Mrs Tiina Haarala for excellent technical work.
References
- Nikkari S., Merilahti‐Palo R., Saario R., Söderström K. O., Granfors K., Skurnik M., et al. Yersinia‐triggered reactive arthritis. Use of polymerase chain reaction and immunocytochemical staining in the detection of bacterial components from synovial specimens. Arthritis Rheum 1992; 35: 682–7
- Kotilainen P., Jalava J., Meurman O., Lehtonen O. P., Rintala E., Seppälä O. P., et al. Diagnosis of meningococcal meningitis by broad‐range bacterial PCR with cerebrospinal fluid. J Clin Microbiol 1998; 36: 2205–9
- Rantakokko‐Jalava K., Nikkari S., Jalava J., Eerola E., Skurnik M., Meurman O., et al. Direct amplification of rRNA genes in diagnosis of bacterial infections. J Clin Microbiol 2000; 38: 32–9
- Kupila L., Rantakokko‐Jalava K., Jalava J., Nikkari S., Peltonen R., Meurman O., et al. Aetiological diagnosis of brain abscesses and spinal infections: application of broad‐range bacterial polymerase chain reaction analysis. J Neurol Neurosurg Psychiatry 2003; 74: 728–33
- Drancourt M., Mainardi J. L., Brouqui P., Vandenesch F., Carta A., Lehnert F., et al. Bartonella (Rochalimaea) quintana endocarditis in three homeless men. N Engl J Med 1995; 332: 419–23
- Jalava J., Kotilainen P., Nikkari S., Skurnik M., Vänttinen E., Lehtonen O. P., et al. Use of the polymerase chain reaction and DNA sequencing for detection of Bartonella quintana in the aortic valve of a patient with culture‐negative infective endocarditis. Clin Infect Dis 1995; 21: 891–6
- Célard M., de Gevigney G., Mosnier S., Buttard P., Benito Y., Etienne J., et al. Polymerase chain reaction analysis for diagnosis of Tropheryma whippelii infective endocarditis in two patients with no previous evidence of Whipple's disease. Clin Infect Dis 1999; 29: 1348–9
- Roux V., Eykyn S., Wyllie S., Raoult D. Bartonella vinsonii subsp. berkhoffii as an agent of afebrile blood culture‐negative endocarditis in a human. J Clin Microbiol 2000; 38: 1698–700
- Nikkari S., Gotoff R., Bourbeau P. P., Brown R. E., Kamal N. R., Relman D. A. Identification of Cardiobacterium hominis by broad‐range bacterial polymerase chain reaction analysis in a case of culture‐negative endocarditis. Arch Intern Med 2002; 162: 477–9
- Goldenberger D., Ku˝nzli A., Vogt P., Zbinden R., Altwegg M. Molecular diagnosis of bacterial endocarditis by broad‐range PCR amplification and direct sequencing. J Clin Microbiol 1997; 35: 2733–9
- Millar B., Moore J., Mallon P., Xu J., Crowe M., Mcclurg R., et al. Molecular diagnosis of infective endocarditis—a new Duke's criterion. Scand J Infect Dis 2001; 33: 673–80
- Bosshard P. P., Kronenberg A., Zbinden R., Ruef C., Böttger E. C., Altwegg M. Etiologic diagnosis of infective endocarditis by broad‐range polymerase chain reaction: a 3‐year experience. Clin Infect Dis 2003; 37: 167–72
- Gauduchon V., Chalabreysse L., Etienne J., Célard M., Benito Y., Lepidi H., et al. Molecular diagnosis of infective endocarditis by PCR amplification and direct sequencing of DNA from valve tissue. J Clin Microbiol 2003; 41: 763–6
- Podglajen I., Bellery F., Poyart C., Coudol P., Buu‐Hoi A., Bruneval P., et al. Comparative molecular and microbiologic diagnosis of bacterial endocarditis. Emerg Infect Dis 2003; 9: 1543–7
- Breitkopf C., Hammel D., Scheld H., Peters G., Becker K. Impact of a molecular approach to improve the microbiological diagnosis of infective heart valve endocarditis. Circulation 2005; 111: 1415–21
- Greub G., Lepidi H., Rovery C., Casalta J‐P., Habib G., Collart F., et al. Diagnosis of infectious endocarditis in patients undergoing valve surgery. Am J Med 2005; 118: 230–8
- Rovery C., Greub G., Lepidi H., Casalta J‐P., Habib G., Collart F., et al. PCR detection of bacteria on cardiac valves of patients with treated bacterial endocarditis. J Clin Microbiol 2005; 43: 163–7
- Houpikian P., Raoult D. Blood culture‐negative endocarditis in a reference center. Etiologic diagnosis of 348 cases. Medicine 2005; 84: 162–73
- Lisby G., Gutschik E., Durack D. T. Molecular methods for diagnosis of infective endocarditis. Infect Dis Clin N Am 2002; 16: 393–412
- Heiro M., Nikoskelainen J., Hartiala J. J., Saraste M. K., Kotilainen P. M. Diagnosis of infective endocarditis: sensitivity of the Duke vs von Reyn criteria. Arch Intern Med 1998; 158: 18–24
- Heiro M., Nikoskelainen K., Engblom E., Kotilainen E., Marttila R. J., Kotilainen P. Neurologic manifestations of infective endocarditis. A 17‐year experience in a teaching hospital in Finland. Arch Intern Med 2000; 160: 2781–7
- Heiro M., Helenius H., Sundell J., Koskinen P., Engblom E., Nikoskelainen J., et al. Utility of serum C‐reactive protein in assessing the outcome of infective endocarditis. Eur Heart J 2005; 26: 1873–81
- Durack D. T., Lukes A. S., Bright D. K., and the Duke Endocarditis Service. New criteria for diagnosis of infective endocarditis: utilization of specific echocardiographic findings. Am J Med 1994; 96: 200–9
- Wilson W. R., Karchmer A. W., Dajani A. S., Taubert K. A., Bayer A., Kaye D., et al. Antibiotic treatment of adults with infective endocarditis due to streptococci, enterococci, staphylococci, and HACEK microorganisms. American Heart Association. JAMA 1995; 274: 1706–13
- Horstkotte D., Follath F., Gutschik E., Lengyel M., Oto A., Pavie A., et al. Guidelines on prevention, diagnosis and treatment of infective endocarditis executive summary. The task force on infective endocarditis of the European Society of Cardiology. Eur Heart J 2004; 25: 267–76
- Kwok S., Higuchi R. Avoiding false positives with PCR. Nature 1989; 339: 237–8
- Benslimani A., Fenollar F., Lepidi H., Raoult D. Bacterial zoonoses and infective endocarditis, Algeria. Emerg Infect Dis 2005; 11: 216–24
- Roggenkamp A., Leitritz L., Baus K., Falsen E., Heesemann J. PCR for detection and identification of Abiotrophia spp. J Clin Microbiol 1998; 36: 2844–6
- Lang S., Watkin R. W., Lambert P. A., Littler W. A., Elliott T. S. Detection of bacterial DNA in cardiac vegetations by PCR after the completion of antimicrobial treatment for endocarditis. Clin Microbiol Infect 2004; 10: 579–81
- Presterl E., Grisold A. J., Reichmann S., Hirschl A. M., Georgopoulos A., Graninger W. Viridans streptococci in endocarditis and neutropenic sepsis: biofilm formation and effects of antibiotics. J Antimicrob Chemother 2005; 55: 45–50
- Drancourt M., Raoult D. rpoB gene sequence‐based identification of Staphylococcus species. J Clin Microbiol 2002; 40: 1333–8
- Nikkari S., Lopez F. A., Lepp P. W., Cieslak P. R., Ladd‐Wilson S., Passaro D., et al. Broad‐range bacterial detection and the analysis of unexplained death and critical illness. Emerg Infect Dis 2002; 8: 188–94
- Eckburg P. B., Bik E. M., Bernstein C. N., Purdom E., Dethlefsen L., Sargent M., et al. Diversity of the human intestinal microbial flora. Science 2005; 308: 1635–8
- Roth S. B., Jalava J., Ruuskanen O., Ruohola A., Nikkari S. Use of an oligonucleotide array for laboratory diagnosis of bacteria responsible for acute upper respiratory infections. J Clin Microbiol 2004; 42: 4268–74