Skip to main page content

• HOME
• Current Issue
• Archive
• Alerts
• About ASM
• Contact us
• Tech Support
• Journals.ASM.org

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Hermsen, E. D. Articles by Rupp, M. E. Search for Related Content PubMed PubMed Citation Articles by Hermsen, E. D. Articles by Rupp, M. E.

 Previous Article  |  Next Article 

Journal of Clinical Microbiology, September 2008, p. 2924-2929, Vol. 46, No. 9
0095-1137/08/$08.00+0     doi:10.1128/JCM.00623-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

Pharmacoeconomic Analysis of Microbiologic Techniques for Differentiating Staphylococci Directly from Blood Culture Bottles

Elizabeth D. Hermsen,^1,2,3* Sara S. Shull,¹ Donald G. Klepser,² Peter C. Iwen,⁴ Amy Armbrust,⁴ Jodi Garrett,⁴ Alison G. Freifeld,³ and Mark E. Rupp³

Department of Pharmaceutical and Nutrition Care, The Nebraska Medical Center, Omaha, Nebraska 68198,¹ Department of Pharmacy Practice, College of Pharmacy, University of Nebraska Medical Center,² Section of Infectious Diseases, Department of Internal Medicine, College of Medicine, University of Nebraska Medical Center,³ Department of Pathology and Microbiology, College of Medicine, University of Nebraska Medical Center, Omaha, Nebraska 68198⁴

Received 1 April 2008/ Accepted 1 July 2008

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Differentiating staphylococci in blood cultures is a critical issue, particularly when only one of two cultures is positive by Gram staining for staphylococci. New tests for the identification of Staphylococcus aureus allow faster results and definitive treatment compared to the tube coagulase test interpreted at 24 h (TCT24). These newer tests, peptide nucleic acid fluorescence in situ hybridization (PNA-FISH) and real-time PCR (RT-PCR), offer improved sensitivity at higher cost. Data suggest that the tube coagulase test may be interpreted at 4 h (TCT4) with little loss of sensitivity. The impact of variability in turnaround time, sensitivity, specificity, and cost on comparative cost-effectiveness is unknown. Our aim was to establish the cost-effectiveness of TCT24, PNA-FISH, RT-PCR, and TCT4 for direct identification of staphylococci in blood cultures. Decision analysis comparing these strategies was done from the institutional perspective. Besides test variables, other variables included patient risk factors, empirical treatment, and follow-up cultures. Probability and cost estimates came from the literature and institutional data. Base case estimates were derived from institutional rates of 73% contamination when coagulase-negative staphylococci were identified, 67.6% prevalence of risk factors, and 12.4% prevalence of S. aureus when one of two cultures yielded staphylococci. Sensitivity analysis was done across a range of probabilities and costs. In the base case, TCT4 and TCT24 were more cost-effective than RT-PCR and PNA-FISH ($78 versus $120 versus $165 per patient, respectively). The advantage of TCT4 and TCT24 remained robust upon sensitivity analysis. TCT4 should be further evaluated as a rapid, cost-effective means for identification of S. aureus in blood cultures.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Staphylococci are the most commonly identified organisms from blood cultures. Two immediate questions confront clinicians when they initially encounter blood cultures that are positive by Gram staining for staphylococci, particularly when only one of two cultures is positive (one blood culture consists of blood collected from a single venipuncture and typically inoculated into an aerobic and an anaerobic bottle). First, a determination of whether the organism is coagulase negative or coagulase positive must be made. The latter implies that Staphylococcus aureus is present, a highly virulent bacterium associated with significant morbidity and mortality that requires prompt treatment with appropriate antibiotics (3, 4, 11, 18). Second, if coagulase-negative staphylococci (CoNS) are identified, the clinician must decide whether the positive culture indicates contamination or true infection. Although CoNS are normal skin flora and common blood culture contaminants, they are increasingly implicated as pathogens, making the distinction between contamination and infection difficult (13, 14, 17). The proportion of positive blood cultures as a function of the total number of blood cultures collected is one of the most frequently used tools to help delineate contamination from true infection (15), but patient-specific factors must also be considered.

The ability to rapidly distinguish between the presence of S. aureus and CoNS in a blood culture is essential for optimal patient management. Both patient outcomes and the costs associated with treatment are affected by the method used for identification. Unnecessary or inappropriate empirical therapy may be eliminated or replaced with targeted therapy. Follow-up cultures, used to discriminate contamination from true infection, may be eliminated, decreasing laboratory costs, or collected sooner, expediting diagnostic accuracy and appropriate treatment. The length of the hospital stay and/or resource utilization may be decreased by not treating CoNS contamination or by diagnosing an S. aureus infection more quickly (1, 5). Therefore, careful evaluation of the available options to differentiate staphylococci is essential in order to select the alternative that generates accurate results in a timely manner at a reasonable cost.

Currently, most clinical microbiology laboratories use either the tube coagulase test or the latex agglutination test to differentiate S. aureus from CoNS. These tests are performed on the overnight culture obtained from the positive blood culture and thus are not interpreted until 24 h after the blood culture becomes positive. During this time, clinicians frequently make empirical therapy decisions because of the possibility that the organism is S. aureus. New tests are available that allow more rapid identification of staphylococci directly from blood. They include the peptide nucleic acid fluorescence in situ hybridization (PNA-FISH; AdvanDx, Woburn, MA) assay, which can be interpreted in as little as 3 hours, and real-time PCR (RT-PCR; Cepheid, Sunnyvale, CA), which can be interpreted in under 1 hour. Both of these tests have improved sensitivity, but at a higher cost, compared to the tube coagulase test interpreted at 24 h (TCT24). Recent data suggest that the tube coagulase test may be performed directly from blood and may also be interpreted at 4 h (TCT4) with little loss of sensitivity or specificity (2).

The impact of variability in the turnaround time, sensitivity, specificity, and acquisition cost on the cost-effectiveness of these tests is currently unknown. Estimating these relationships through decision modeling is an efficient way to quantify the benefit and risk of using any of these tests when data from well-designed prospective trials are not available. The objective of this study was to compare the cost-effectiveness of TCT24, PNA-FISH, RT-PCR, and TCT4 for differentiation of staphylococci directly from blood culture bottles when one of two blood cultures yields gram-positive cocci in clusters.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Model design. This study used a clinical decision model to compare the cost-effectiveness of four organism identification strategies: TCT24, PNA-FISH, RT-PCR, and TCT4. Our model was based on patients with one of two blood cultures yielding staphylococci by Gram staining. The decision analysis was conducted from the perspective of the health care institution. TreeAge Pro 2007 (TreeAge Software, Williamstown, MA) was used to construct the decision analysis model.

Diagnostic strategies. The diagnostic strategies included various methods used for detection of S. aureus directly from blood culture bottles. Data suggest that the tube coagulase test can be performed directly from blood culture bottles and interpreted earlier than the standard 24-hour turnaround time (2). The methodologies for both TCT24 and TCT4 were the same, with the exception of the turnaround time. More specifically, several drops of the positive blood culture broth were added to a tube containing 0.5 ml of rabbit plasma with EDTA (BBL, Cockeysville, MD). The tubes were incubated at 35°C for either 4 or 24 h for TCT4 or TCT24, respectively. Any clotting was interpreted as indicative of S. aureus. The time required for the laboratory technician to perform the test and read and report the results was approximately 5 minutes per specimen.

The PNA-FISH assay involved a number of steps (AdvanDx, Woburn, MA). First, a smear was prepared on a microscope slide using a drop of the positive blood culture broth and a fixation solution. Hybridization was then performed by adding S. aureus PNA to the microscope slide and incubating the slide at 55°C for 90 min. The slide was subsequently washed with a preheated, prepared wash solution and incubated at 55°C for an additional 30 min. Finally, one drop of mounting medium was added to the smear prior to the slide being observed using a fluorescence microscope. S. aureus was identified as multiple bright-green fluorescent clusters of cocci. We estimated that the time required by the laboratory technician for preparation of the test and to read and report the results for a single specimen was approximately 50 min, with 5 min added for each additional specimen within the batch.

The RT-PCR test was performed by inoculating the provided cartridge with the positive blood culture broth and inserting the cartridge into the required instrument (Cepheid, Sunnyvale, CA). Four reactions occurred simultaneously within a single cartridge. The reactions targeted S. aureus protein A (Spa), the mecA gene, a proprietary sequence of the SCC mec cassette, and the S. aureus-specific orfX gene and included an internal control for determination of the presence of S. aureus and methicillin-resistant S. aureus. We estimated that the time required by the laboratory technician for preparation of the test and reading and reporting the results was approximately 5 min per specimen.

Probability estimates. Probability and cost estimates for the base case and sensitivity analysis came from the literature and institutional data. Our institution is a 689-bed academic medical center with three adult intensive-care units, a pediatric intensive-care unit, a neonatal intensive-care unit, and large hematopoetic stem cell and solid-organ transplant populations. Using sensitivity analysis, the probabilities and costs were varied over a range of plausible values to assess the impact of each variable on the final conclusion. Sensitivity analysis allows one to see if changing the probability or cost of an event changes the conclusion drawn in the base case. If the conclusion does not change, the model is considered robust, generating enhanced confidence in the model's usefulness for decision making. If the conclusion changes, then the model is sensitive to changes in the estimates, suggesting that the model may be less useful in different environments as a decision-making tool.

Baseline probabilities. The decision model and variables used for the various methods of testing are shown in Fig. 1A to D. The first chance node for modeled patients is to evaluate the presence of risk factors for harboring true infection when one of two blood cultures yields CoNS. The presence of any of the following risk factors would cause increased suspicion of true infection due to CoNS rather than contamination: (i) a white blood cell count of <2,000 cells/mm³ or >12,000 cells/mm³ plus >10% bands and a body temperature of <36°C or 38.5°C or systolic blood pressure of <90 mm Hg, (ii) a neonate, or (iii) the presence of a central line or other indwelling intravascular device (e.g., a pacemaker or vascular graft). The probability of the presence of these risk factors was determined from a review of a convenience sample of 182 patients with one of two blood cultures yielding either CoNS or S. aureus at our institution.

View larger version (70K): [in this window] [in a new window]

FIG. 1. Decision model and variables for TCT24, TCT4, PNA-FISH, and RT-PCR. p, probability; c, cost; Coag +, coagulase-positive staphylococci (S. aureus); Coag –, CoNS; RiskPos, presence of risk factors for true infection plus coagulase-positive staphylococci; NoRiskPos, lack of risk factors for true infection plus coagulase-positive staphylococci; TruePos, true coagulase positive; TrueNeg, true coagulase negative; FalsePosCont, false coagulase positive plus contamination; TrueNegCont, true coagulase negative plus contamination; FalseNegCont, false coagulase negative plus contamination; Vanco, vancomycin; BloodCult, blood culture. (A) Decision model and variables for TCT24. (B) Decision model and variables for TCT4. (C) Decision model and variables for PNA-FISH. (D) Decision model and variables for RT-PCR.

The second chance node determines the probability of the presence of coagulase-positive staphylococci (S. aureus) when one of two blood cultures is initially characterized by Gram staining as yielding staphylococci.

The third chance node reflects the probability that a test result will be either a true positive or a true negative (Table 1), which was calculated from the sensitivity and specificity of the diagnostic tests and the underlying institutional rate of S. aureus presence when one of two blood cultures is positive by Gram staining for staphylococci. The sensitivities and specificities of TCT24 (90% and 100%, respectively) and TCT4 (85% and 100%, respectively), directly from blood culture bottles, were derived by our microbiology laboratory by testing consecutive blood cultures initially characterized by Gram staining as yielding staphylococci (personal communication). The sensitivity and specificity of the PNA-FISH assay (98.5% and 98.5%, respectively) were taken from the manufacturer's data (S. aureus PNA FISH Staphylococcus aureus culture identification kit package insert; AdvanDx, Woburn, MA). The sensitivity and specificity of the RT-PCR test (99.7% and 99.1%, respectively) were taken from the manufacturer's data for a commercially available kit (BD GeneOhm StaphSR assay package insert; BD Diagnostics, Quebec, Canada), although the other modeled RT-PCR variables are associated with an investigational kit (Cepheid, Sunnyvale, CA).

View this table: [in this window] [in a new window]

TABLE 1. Model variable values and sensitivity ranges

The final chance node addresses the rate of blood culture contamination with CoNS and is determined primarily through the use of follow-up blood cultures drawn when the initial culture becomes positive by Gram staining for staphylococci. A contamination rate of 73% was used in this analysis and was derived from a previously conducted study at our institution (Table 1) (E. D. Hermsen and S. S. Shull, presented at the 43rd Annual Meeting of the Infectious Diseases Society of America, San Francisco, CA, 2005).

Baseline costs. Cost data are presented from the institutional perspective and are expressed in U.S. dollars. The costs of vancomycin ($15/g), blood cultures ($29.50/culture and identification), TCT24 ($3.02/test), and TCT4 ($3.02/test) were based on the costs to our institution, including supplies and labor. The cost for PNA-FISH ($66.06/test) is based on batching the test runs three times daily with eight tests per batch, including supplies and labor but excluding the $4,500 start-up cost for instrumentation. The cost for RT-PCR ($45/test) is according to the manufacturer and includes supplies and labor but excludes $135,000 in start-up costs for instrumentation. All costs were varied through a range of values using sensitivity analysis (Table 2).

View this table: [in this window] [in a new window]

TABLE 2. Costs of tests

Assumptions. The model is based on several assumptions consistent with the standard of practice at our institution. (i) All blood cultures characterized by Gram staining as yielding staphylococci are subcultured to an agar plate, which is interpreted at 24 h via a latex agglutination test, yielding definitive organism identification results. (ii) Two follow-up blood cultures are ordered for all patients with at least one risk factor for true infection due to CoNS. (iii) The time horizon of the model is 48 h after the initial Gram stain result. This period is required for definitive identification of the organism at 24 h and determination of contamination or true infection with CoNS via follow-up cultures, which are available within 48 h. Thus, the relevancy of the test attributes declines beyond the 48-hour time horizon. (iv) Due to workflow efficiencies in the microbiology laboratory, the PNA-FISH test is batched and run three times daily. Thus, the time to interpretation for PNA-FISH is 8 h, despite a 3-hour processing time. (v) Vancomycin is the agent utilized for empirical treatment to allow coverage of methicillin-resistant S. aureus and methicillin-resistant Staphylococcus epidermidis. The dosing regimen is 1 g intravenously every 12 h. Patients without risk factors for true infection receive only a single dose of vancomycin until the coagulase status is identified and true infection versus contamination with CoNS is determined via follow-up cultures. (vi) Clinicians will wait up to 1 hour for differentiation of staphylococci before beginning empirical therapy and collecting follow-up blood cultures.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
Baseline probabilities. The base case probabilities and the range of probabilities used in the sensitivity analysis are shown in Table 1. Based on a review of patients with one of two blood cultures yielding either CoNS or S. aureus, 67.6% of the patients had one or more of the risk factors for true infection present. The probability of 12.4% for the presence of S. aureus when one of two blood cultures is initially characterized by Gram staining as yielding staphylococci was derived from institutional data from our microbiology laboratory.

Model results. In the base case, TCT4 and TCT24 ($78 per patient) were equally cost-effective and more cost-effective than RT-PCR and PNA-FISH ($120 versus $165 per patient, respectively) for the identification of staphylococci directly from blood culture bottles.

Sensitivity analyses. One-way sensitivity analyses were conducted on all variables. Three variables accounted for 95.5% of the model's variability: cost of blood culture, cost of vancomycin, and probability of contamination when CoNS was identified in a blood culture. A series of two-way sensitivity analyses were conducted using these variables. Regardless of the values used for these variables, the cost-effectiveness relationship between the tests was unchanged. Thus, the advantages of TCT4 and TCT24 were robust.

A series of threshold analyses were also conducted to determine the level at which either RT-PCR or PNA-FISH would become a more cost-effective alternative to TCT4 or TCT24. Holding all other variables at base case levels, the RT-PCR test would have to cost less than $9.81/test in order to be the most cost-effective alternative. Similarly, the cost per test for PNA-FISH would have to be less than $2.61.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 
To our knowledge, this is the first published study to model the cost-effectiveness of these four strategies for differentiating staphylococci directly from blood culture bottles. PNA-FISH offers more rapid bacterial identification and enhanced sensitivity in identifying the coagulase status at an acquisition cost 22 times greater than that of TCT24 (Tables 1 and 2). RT-PCR offers more rapid bacterial identification and enhanced sensitivity and determination of methicillin susceptibility at an acquisition cost 15 times greater than that of TCT24 (Tables 1 and 2). TCT4 offers more rapid bacterial identification with slightly lower sensitivity (85% versus 90%) at a cost equal to that of TCT24. In a single year at our institution, the excess acquisition cost differential between PNA-FISH or RT-PCR and TCT4 or TCT24 is $77,850 or $51,846, respectively. Moreover, the newer test methods require substantial capital start-up costs. These increased costs would be justified if the utilization of the test methods resulted in improved infection management and better outcomes for patients while helping hospitals cope with the problem of blood culture contamination. Therefore, these comparisons are important as laboratories make decisions about how to distribute scarce economic resources while optimizing patient care.

While sensitivity and specificity are generally important comparative attributes of diagnostic tests, the variation in sensitivity and specificity among TCT24, PNA-FISH, RT-PCR, and TCT4 had a minimal effect on our model. The practice of subculturing Gram stain-positive blood cultures to agar plates strongly affects the impact of test sensitivity and specificity on the model. Access to these plates allows definitive identification of the organism (via latex agglutination) 24 h after the culture becomes positive, preventing organism misidentification beyond 24 h and significantly decreasing the probability of a missed infection. Plate subculturing, therefore, increases the acceptability of the lower sensitivity achieved with TCT24 and TCT4 and may obviate the need for the higher sensitivity and associated expense of using PNA-FISH and RT-PCR. Furthermore, because the presence of CoNS can be defined as infection or contamination within 48 h based on follow-up blood cultures, the relevancy of the sensitivity and specificity of diagnostic tests declines beyond this time point.

With an increased demand for more rapid reporting of results in the laboratory, a shorter turnaround time for PNA-FISH and RT-PCR appears attractive (6, 7, 12, 16). The PNA-FISH assay can be interpreted at 3 hours, a significant difference from TCT24. However, this time saving may be more theoretical than practical. For instance, the testing of individual samples at the time the blood culture bottle is determined to be Gram stain positive for staphylococci, in order to derive the greatest benefit from the attenuated interpretation time, may not be feasible for most institutions due to workflow issues. Rather, in most laboratories, samples would be batched to increase efficiencies necessitated by manpower and labor and supply cost considerations. Batches may be run as little as one to three times a day. The value of a short time to interpretation, when its benefit cannot be fully realized, renders the PNA-FISH assay less cost-effective.

We modeled an investigational RT-PCR kit (Cepheid, Sunnyvale, CA) because of the rapid generation of results (within 1 hour) and ease of use in contrast to the commercially available RT-PCR kit (BD GeneOhm StaphSR assay; BD Diagnostics, Quebec, Canada), which would be batched similarly to PNA-FISH due to labor intensity. Thus, unlike PNA-FISH, RT-PCR may feasibly be performed on individual samples, allowing rapid species identification and determination of the methicillin susceptibility of the isolate. However, similar to PNA-FISH, this time-saving benefit may be more theoretical than actual. Obtaining bacterial identification sooner may result in the administration of fewer unnecessary doses of vancomycin, assuming that physicians will wait 2 hours for the results of testing before starting empirical therapy. However, in the model, this scenario occurs only when patients who have no risk factors for true infection have a blood culture contaminated with CoNS. Furthermore, rapid bacterial identification is problematic when it is erroneous. The outcome associated with a misidentification of S. aureus, when bacterial growth is actually due to CoNS contamination in patients with risk factors for true infection, is the administration of unnecessary antimicrobial therapy. Conversely, a potential benefit of RT-PCR is earlier decision making related to patient isolation, which would be made possible by more rapid access to information about methicillin resistance. However, little is understood about how the timing of isolation affects patient outcomes and costs. Another potential benefit of RT-PCR is an earlier change from vancomycin to an antistaphylococcal penicillin when methicillin susceptibility is confirmed because an antistaphylococcal penicillin is considered optimal therapy for methicillin-susceptible S. aureus. However, factors affecting the timing of isolation and a switch to optimal therapy due to the determination of methicillin susceptibility are beyond the scope of this model.

TCT4 provides a more rapid result at a cost equivalent to that of TCT24. Importantly, the choice between TCT4 and TCT24 results in differential administration of vancomycin. Specifically, in contrast to TCT24, TCT4 prevents a missed dose of vancomycin in patients with no risk factors for true CoNS infection but who are actually harboring S. aureus. The impact on patient outcomes resulting from the elimination of a single dose of vancomycin in the first 48 h of detection of possible bacteremia is unknown and is not considered in this model. However, it is reasonable to assume that missing a dose would result in inadequate therapy, which is not optimal for patient outcomes (8-10) Thus, the prevention of a missed dose is a potential advantage of TCT4 over TCT24.

In conclusion, our model highlights two unexpected findings. The variation in sensitivity and specificity and time to interpretation of TCT24, PNA-FISH, RT-PCR, and TCT4 have limited impact on the cost-effectiveness of the four strategies. Thus, the increased sensitivity and rapid turnaround time of the newer tests do not appear to be worth the increased price. Of note, our model does not address the methicillin susceptibility of S. aureus. Future research on the impact of this factor is needed. Based on our current model results, we believe that TCT4 should be investigated further as a cost-effective means of rapidly differentiating staphylococci directly from blood culture bottles.

 
We are grateful to Melissa Maiefski and Monica Hanson, both at the University of Nebraska Medical Center College of Pharmacy, and to the Clinical Microbiology Laboratory personnel for their assistance with data collection. E.D.H., S.S.S., D.G.K., P.C.I., A.A., J.G., and A.G.F. have no conflict of interest. M.E.R. has received 3M and Becton Dickinson research grants.

No financial support was received for this study.

 
* Corresponding author. Mailing address: Department of Pharmaceutical and Nutrition Care, The Nebraska Medical Center, 988485 Nebraska Medical Center, Omaha, NE 68198-8485. Phone: (402) 559-4287. Fax: (402) 559-3350. E-mail: ehermsen{at}nebraskamed.com

Published ahead of print on 9 July 2008.

Top ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

 

1
1. Bates, D. W., L. Goldman, and T. H. Lee. 1991. Contaminant blood cultures and resource utilization. The true consequences of false-positive results. JAMA 265:365-369.[Abstract/Free Full Text]
2
2. Chapin, K., and M. Musgnug. 2003. Evaluation of three rapid methods for the direct identification of Staphylococcus aureus from positive blood cultures. J. Clin. Microbiol. 41:4324-4327.[Abstract/Free Full Text]
3
3. Cosgrove, S. E., Y. Qi, K. S. Kaye, S. Harbarth, A. W. Karchmer, and Y. Carmeli. 2005. The impact of methicillin resistance in Staphylococcus aureus bacteremia on patient outcomes: mortality, length of stay, and hospital charges. Infect. Control Hosp. Epidemiol. 26:166-174.[CrossRef][Medline]
4
4. Cosgrove, S. E., G. Sakoulas, E. N. Perencevich, M. J. Schwaber, A. W. Karchmer, and Y. Carmeli. 2003. Comparison of mortality associated with methicillin-resistant and methicillin-susceptible Staphylococcus aureus bacteremia: a meta-analysis. Clin. Infect. Dis. 36:53-59.[CrossRef][Medline]
5
5. Dunagan, W. C., R. S. Woodward, G. Medoff, J. L. Gray III, E. Casabar, M. D. Smith, C. A. Lawrenz, and E. Spitznagel. 1989. Antimicrobial misuse in patients with positive blood cultures. Am. J. Med. 87:253-259.[Medline]
6
6. Forrest, G. N., S. Mehta, E. Weekes, D. P. Lincalis, J. K. Johnson, and R. A. Venezia. 2006. Impact of rapid in situ hybridization testing on coagulase-negative staphylococci positive blood cultures. J. Antimicrob. Chemother. 58:154-158.[Abstract/Free Full Text]
7
7. Grobner, S., and V. A. Kempf. 2007. Rapid detection of methicillin-resistant staphylococci by real-time PCR directly from positive blood culture bottles. Eur. J. Clin. Microbiol. Infect. Dis. 26:751-754.[CrossRef][Medline]
8
8. Ibrahim, E. H., G. Sherman, S. Ward, V. J. Fraser, and M. H. Kollef. 2000. The influence of inadequate antimicrobial treatment of bloodstream infections on patient outcomes in the ICU setting. Chest 118:146-155.[CrossRef][Medline]
9
9. Kollef, M. H. 2000. Inadequate antimicrobial treatment: an important determinant of outcome for hospitalized patients. Clin. Infect. Dis. 31(Suppl. 4):S131-S138.[CrossRef][Medline]
10
10. Kollef, M. H., G. Sherman, S. Ward, and V. J. Fraser. 1999. Inadequate antimicrobial treatment of infections: a risk factor for hospital mortality among critically ill patients. Chest 115:462-474.[CrossRef][Medline]
11
11. Mekontso-Dessap, A., M. Kirsch, C. Brun-Buisson, and D. Loisance. 2001. Poststernotomy mediastinitis due to Staphylococcus aureus: comparison of methicillin-resistant and methicillin-susceptible cases. Clin. Infect. Dis. 32:877-883.[CrossRef][Medline]
12
12. Oliveira, K., G. W. Procop, D. Wilson, J. Coull, and H. Stender. 2002. Rapid identification of Staphylococcus aureus directly from blood cultures by fluorescence in situ hybridization with peptide nucleic acid probes. J. Clin. Microbiol. 40:247-251.[Abstract/Free Full Text]
13
13. Rupp, M. E. 1997. Coagulase-negative staphylococcal infection: an update regarding recognition and management. Curr. Clin. Top. Infect. Dis. 17:51-87.[Medline]
14
14. Rupp, M. E., and G. L. Archer. 1994. Coagulase-negative staphylococci: pathogens associated with medical progress. Clin. Infect. Dis. 19:231-243.[Medline]
15
15. Schifman, R. B., C. L. Strand, F. A. Meier, and P. J. Howanitz. 1998. Blood culture contamination: a College of American Pathologists Q-Probes study involving 640 institutions and 497,134 specimens from adult patients. Arch. Pathol. Lab. Med. 122:216-221.[Medline]
16
16. Stratidis, J., F. J. Bia, and S. C. Edberg. 2007. Use of real-time polymerase chain reaction for identification of methicillin-resistant Staphylococcus aureus directly from positive blood culture bottles. Diagn. Microbiol. Infect. Dis. 58:199-202.[CrossRef][Medline]
17
17. Weinstein, M. P. 2003. Blood culture contamination: persisting problems and partial progress. J. Clin. Microbiol. 41:2275-2278.[Free Full Text]
18
18. Whitby, M., M. L. McLaws, and G. Berry. 2001. Risk of death from methicillin-resistant Staphylococcus aureus bacteraemia: a meta-analysis. Med. J. Aust. 175:264-267.[Medline]

---

Journal of Clinical Microbiology, September 2008, p. 2924-2929, Vol. 46, No. 9
0095-1137/08/$08.00+0     doi:10.1128/JCM.00623-08
Copyright © 2008, American Society for Microbiology. All Rights Reserved.

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Reprints and Permissions Copyright Information Books from ASM Press MicrobeWorld Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Hermsen, E. D. Articles by Rupp, M. E. Search for Related Content PubMed PubMed Citation Articles by Hermsen, E. D. Articles by Rupp, M. E.