Skip Navigation

Difference in Time to Detection: A Simple Method to Differentiate Catheter-Related from Non—Catheter-Related Bloodstream Infection in Immunocompromised Pediatric Patients

  1. Aditya H. Gaur1,3,
  2. Patricia M. Flynn1,3,a,
  3. Mary Anne Giannini2,
  4. Jerry L. Shenep1,3, and
  5. Randall T. Hayden2,a
  1. 1Departments of Infectious Diseases, St. Jude Children's Research Hospital, Memphis, Tennessee
  2. 2Departments of Pathology, St. Jude Children's Research Hospital, Memphis, Tennessee
  3. 3Department of Pediatrics, University of Tennessee Health Science Center, Memphis, Tennessee
  1. Reprints or correspondence: Dr. Randall T. Hayden, Dept. of Pathology, St. Jude Children's Research Hospital, 332 N. Lauderdale St., Memphis, TN 38105-2794 (randall.hayden{at}stjude.org).
 
Next Section

Abstract

Current methods for diagnosis of catheter-related infection (CRI) are cumbersome and may require removal of the central venous catheter (CVC). A prospective study was conducted to validate the difference in time to detection (DTD) of cultures of blood samples obtained simultaneously from a peripheral vein (PV) and from the CVC for differentiation of CRI and non-CRI. During a 15-month period, 9 episodes were categorized as CRI and 24 as non-CRI. The median DTD for patients with CRI was significantly higher than that for patients with non-CRI (457 vs. -4 min; P < .001). The optimum cutoff point for diagnosis of CRI was a DTD of ⩾120 min (sensitivity, 88.9%; specificity, 100%). With pretest probability of CRI ranging from 28% to 54%, the positive predictive value of a DTD of ⩾120 min for the diagnosis of CRI was 100%; the negative predictive value was 89%–96%. On the basis of findings from this study, which is the largest, to date, to involve pediatric patients with tunneled CVCs and the first to use paired quantitative blood cultures as a “criterion standard,” DTD was found to be a simple, reliable tool for diagnosis of CRI in hospitals that use continuously read blood culture systems.

In patients with central venous catheters (CVCs), catheter-related infections (CRIs) are a prominent cause of morbidity, excess hospital costs, and, in some cases, mortality. Early efforts to remove all putatively infected catheters were problematic, because many of these catheters have subsequently been found to not be infected. In patients with surgically placed CVCs, such as children with cancer, removal of the CVC puts the patients at risk of having to undergo additional surgery and can also delay chemotherapy or other necessary treatment. Current guidelines advocate specific interventions, such as administration of antibiotic lock therapy or removal of the catheter, in certain patients with CRI [1]. During the past 2 decades, numerous diagnostic methods have been proposed to differentiate CRIs from non-CRIs [2]. Despite these proposals, there is no single method for diagnosis of CRI that is accurate and reliable, that allows a diagnosis with the catheter in situ, and that is easy to perform. Although the quantitative catheter-tip culture technique is sensitive and specific, it is of no value in salvaging a CVC [2]. Paired quantitative blood cultures are useful for in situ diagnosis of CRI [3–6], but they have had limited application because of the expense and the labor-intensive process involved.

In 1998, with use of an automated, continuously monitored blood culture system, Blot et al. [7] noted that a difference between the peripheral and CVC blood cultures in time to detection (DTD) of >120 min was highly sensitive and specific for the diagnosis of CRI [7]. This time differential was theoretically based on the difference in the bacterial load between CVC and peripheral vein (PV)–obtained blood cultures for patients with CRI [4] and the linear relationship between the inoculum size at incubation and the time to positivity of the blood culture [7–9]. Since then, 3 studies have confirmed the utility of DTD for the diagnosis of CRI in adults with cancer [10–12], and 1 study of adults in a medical-surgical ICU did not confirm the utility [13]. Although 4 of these studies [7, 10, 11, 13] used the catheter-tip culture/clinical criteria as the criterion standard, 1 study [12] used paired quantitative blood cultures. The results were somewhat disparate and lacked both data specific to pediatric patients and an evaluation of whether changes in culture media type would influence results.

The present study assessed the validity of DTD for the diagnosis of CRI in pediatric, immunocompromised patients with tunnelled CVC in place using paired quantitative blood cultures [4] as the criterion standard to define CRI. In addition, we evaluated the differences in cutoff points for DTD for 2 different media types, the BACTEC Plus Aerobic/F (AF; Becton Dickinson) culture vial and the BACTEC MYCO/F Lytic (MFL; Becton Dickinson) culture vial.

Previous SectionNext Section

Materials and Methods

Patient population. This prospective study was performed at St. Jude Children's Research Hospital (SJCRH; Memphis, TN) from June 2001 through September 2002. SJCRH is a primarily pediatric oncology research and treatment facility with ∼1800 hospital admissions per year.

Collection of blood samples and culture processing. The BACTEC 9240 blood culture system (BAC; Becton Dickinson) began to be used at SJCRH in June 2001 with the intention of replacing the existing 1.5 mL ISOLATOR Lysis Centrifugation System (ISO; Wampole Laboratories). The volume of blood inoculated into the BAC was based on the patients' total blood volume. Patients were stratified into 5 weight classes, with individual blood sampling graduated as shown in table 1. Sample volume was not to exceed 1% of the patient's calculated blood volume (determined on the basis of normal population data [14]). The volume of blood inoculated into the ISO remained unchanged from past practice (1.5 mL per culture for all weight classes except for the lowest weight class).

Table 1
View larger version:
Table 1

Patient weight and volume of blood obtained for culture.

On the first day of a suspected bloodstream infection, samples were obtained for culture both percutaneously and from indwelling CVCs. Patients in weight classes 3–5 had inoculation of both ISO and BAC systems. The latter included AF cultures (targeting detection of aerobic organisms) and MFL cultures (targeting detection of mycobacteria and yeast).

For ISO, quantitation of organisms was determined by total colony counts on 1 sheep-blood agar plate and 2 chocolate agar plates, with extrapolation of the total to colony-forming units (cfu) per milliliter of blood. Counts were reported on a scale of 1–400 cfu/mL or as >400 cfu/mL. BAC bottles were incubated in an automated, continuously read incubation and detection system.

Definitions. An infectious episode was defined as documented case of (noncontaminant) bacteremia or fungemia. An organism was considered to be a contaminant if it was typical skin flora (i.e., diphtheroids, except for Corynebacterium jeikeium; Bacillus species, except for Bacillus cereus; Propionibacterium species; coagulase-negative staphylococci; and Micrococcus species) and if it was isolated from only 1 culture receptacle. The AF culture vial, the MFL culture vial, and the ISO were considered 3 culture receptacles. The paired quantitative culture criteria for the ISO formed the criterion standard to differentiate CRI from non-CRI [1, 4]. A CRI occurred when the colony count in a blood sample obtained from a CVC was ⩾5-fold higher than the count for a blood sample obtained from a PV; a <5-fold difference was considered to indicate a non-CRI. Cultures of samples obtained from CVCs and PVs within 3 h of each other were considered to be paired cultures. Time to detection was defined as the time from the start of culture incubation in BAC to the initial indication of a positive culture result (i.e., change in fluorescence) in either AF or MFL culture vials. DTD was calculated as the time to detection for the PV BAC culture minus the time to detection for the CVC BAC culture (similar BAC media type [i.e., AF or MFL]).

Inclusion criteria. The study included only those episodes for which blood samples for paired ISO and BAC cultures were obtained from both PVs and CVCs and for which the results for ⩾1 ISO and both BAC bottles (i.e., both AF and/or both MFL cultures) were positive. Cases for which the diagnosis of CRI could not be determined because both the CVC and PV quantitative cultures yielded >400 cfu/mL were excluded. When assessing the significance of a single positive BAC culture result, infectious episodes in which a diagnosis of CRI or non-CRI could be established on the basis of ISO cultures but only a single BAC bottle of a pair were analyzed. We also examined application of the DTD cut point established in this study for infectious episodes in which paired BAC culture results were positive but paired quantitative cultures were not.

Data collection and analysis. After SJCRH Institutional Review Board approval was obtained, data were abstracted from the microbiology laboratory and patient medical records. Statistical analyses were performed using SAS software, version 8.1 (SAS Institute), and StatXact-5, version 5.0.3 (Cytel Software). The DTD for CRI versus non-CRI was compared using the nonparametric, median, 2-sample test. All tests were 2-sided, with a significance level of .05. To determine the optimal threshold, a receiver operating characteristic curve was constructed using the DTD for episodes of CRI and non-CRI for both AF and MFL cultures. Interpretation of DTD relative to the cutoff point was described by likelihood ratio and posttest probability. The positive likelihood ratio is the increase in the odds of having CRI when the DTD was greater than or equal to the cutoff point for that particular media type, and it is defined as (sensitivity)/(1-specificity). When determining the predictive value, prevalence of CRI was calculated as the number of infectious episodes considered to be CRI divided by the total number of infectious episodes during that period. The prevalence of CRI was determined from data collected at SJCRH and from previously published pediatric oncology literature [15–17]. The exact 95% CIs of all the parameters were determined using StatXact-5 software.

Previous SectionNext Section

Results

Validity Analysis

During the 15-month study period, there were 269 infectious episodes, 33 (12%) of which met the inclusion criteria for analysis. In all 33 episodes included in the validity analysis, the results of paired AF cultures were positive, and the results of paired MFL cultures were positive in 30 episodes. The median age of patients included in the analysis was 11.5 years (range, 1.6–19.1 years), with 64% of patients having an underlying hematological malignancy and 70% having neutropenia. A majority (82%) of the children had a Hickman catheter in place.

Of the 33 episodes, 24 (73%) were non-CRI and 9 (27%) were CRI, as determined by the criterion standard. The spectrum of microorganisms isolated, the colony counts, and the DTDs (AF culture data only) are shown in table 2. Although a majority of the organisms isolated in episodes deemed to be CRI were gram positive and were likely to have cutaneous origin, a significant proportion of gram-negative organisms were isolated from patients with non-CRI. Unlike most non-CRIs, 8 of 9 CRIs were associated with high-grade bacteremia or fungemia (colony count in catheter-obtained blood, >400 cfu/mL).

Table 2
View larger version:
Table 2

Findings for episodes classified as catheter-related infection or non—catheter-related infection.

The median DTD for CRI (456.6 min; range, 40.8–6205.2 min) was significantly greater than that for non-CRI (-4.2 min; range, -646.8 to 90.6 min; P < .001). In 30 of 33 episodes, paired MFL culture results were also positive (data not shown). In these cases, the median DTD for the 7 CRIs (431.4 min; range, 60.6–1104.6 min) was significantly greater than that for the 23 non-CRIs (0 min; range, -3147.6 to 2824.2; P = .003).

The optimum threshold for a test using DTD to differentiate CRI from non-CRI on the basis of AF and MFL culture data can be seen from the receiver operating characteristic curve (figure 1). In case of the AF media, the optimum cutoff point for diagnosis of CRI was ⩾120 min (sensitivity, 89% [95% CI, 51.8%–99.7%]; specificity, 100% [95% CI, 85.8%–100%]). With 100% specificity, the estimate of the positive likelihood ratio of the test is infinity. The 95% lower bound of the positive likelihood ratio is 3.64 (i.e., at minimum, with a 5% probability of error, a DTD of ⩾120 min is 3.6 times more likely to be found for a patient with CRI than for one without CRI). Although the sensitivity and specificity for MFL media using the same cutoff point are 86% (95% CI, 42%–99.6%) and 91%, respectively, the specificity improves to 96% (95% CI, 78.1%–99.9%) if a cutoff point of ⩾150 min is used (figure 1). With the latter cutoff point, the positive likelihood ratio is 21.5 (95% lower bound, 1.92), meaning a DTD of ⩾150 min is 21 times more likely to be found for a patient with CRI than for a patient with non-CRI.

Figure 1
View larger version:
Figure 1

Receiver operating characteristic curve plotting the sensitivity and 1-specificity for various differences in time to detection (DTD) cut points. DTD thresholds of 30–210 min were evaluated, in 30 min increments. Selected thresholds are shown for each media type (shortest to longest duration, depicted from right to left, respectively), with optimal cut points indicated in filled symbols. AF, BACTEC Plus Aerobic/F culture vial (Becton Dickinson); MFL, BACTEC MYCO/F Lytic culture vial (Becton Dickinson).

In this study, the positive predictive value or posttest probability of a DTD of ⩾120 min indicating a CRI with paired positive AF culture results is 100% (95% CI, 63%–100%), and the negative predictive value is 96% (95% CI, 79.7%–99.9%). On the basis of both the prevalence of CRI in pediatric oncology patients at SJCRH (40%) and the prevalence reported from other institutions (28%–54%) [15–17], the positive predictive value would be 100% and the negative predictive value would be 88.6%–95.9%. A positive predictive value for a DTD of ⩾150 min being indicative of a CRI with paired MFL culture results in this study is 85.7% (95% CI, 42.1%–99.6%), and the negative predictive value is 95.6% (95% CI, 78.1%–99.9%). Using the prevalence estimates mentioned earlier, the positive predictive value is 89.3%–96.2% and the negative predictive value is 85.4%–94.6%.

There was some difference in the volume of blood inoculated into each bottle of the pair in 33% of the episodes categorized as CRI and in 62% of those deemed to be non-CRI (data not shown). Of the 33 episodes, in only 1 (non-CRI episode 6 in table 2) was there a >3-fold difference in the volume of blood obtained from the PV and CVC inoculated into an AF culture each.

Subset Analysis

Patients with positive results of AF and/or MFL cultures of either CVC- or PV-obtained blood samples, as well as positive results of paired quantitative cultures. Of the 10 instances in which the results of 1 of the pair of AF cultures were positive, 6 were CRI and 4 were non-CRI, on the basis of our criterion standard. On the basis of these 10 cases, it appears that, although the sensitivity for a positive CVC culture result and a negative PV culture result for diagnosis of CRI is fairly high (83%), the results are not very specific (50%). A negative CVC culture result, however, in the presence of a positive PV culture result could be helpful in ruling out a CRI (specificity, 83%). Although suggestive, these numbers are too small to perform inferential statistical analysis.

Patients with paired positive CVC/PV AF and/or MFL cultures but paired negative quantitative cultures. During the first 6 months of the study, there were 14 episodes in which the results of paired BAC cultures were positive but the results of both ISO cultures were negative. In 10 of these episodes, paired AF culture results were positive, and 10 episodes had paired positive MFL culture results. If we apply the cutoff points of 120 and 150 min for AF and MFL culture pairs, respectively, the DTD for all 10 AF culture pairs and for 8 of 10 MFL culture pairs indicated non-CRI.

Previous SectionNext Section

Discussion

This study demonstrates that, in immunocompromised pediatric patients with tunneled CVCs in place, DTD can reliably distinguish CRI from non-CRI. The sensitivity, specificity, and positive and negative predictive values for a DTD of ⩾120 min (for paired positive AF culture results) for the diagnosis of CRI are 88.9%, 100%, 100%, and 88.6%–95.9%, respectively. This cutoff is similar to those reported in adult oncology patients [7, 10–12]. The consistency of results between the present study and previous studies comes despite differences in age group, catheter type, blood culture system, culture media, and criterion standard. One previous study [13] found DTD to be inadequate for the diagnosis of CRI; this may be because of such factors as small sample size and differences in clinical setting, including the number of antibiotic-pretreated patients. None of the patients included for analysis in the present study were receiving antimicrobials at the time that positive blood culture results were obtained, compared with 78% of the patients studied by Rijnders et al. [13].

Whether variations in the blood culture media and/or the metabolic indicator used to detect microorganism growth influences the cutoff DTD value used for CRI determination has not been previously reported. Unlike previously published studies [7, 10–13], we examined the utility of DTD to categorize a CRI when using MFL culture, a media originally developed and marketed to detect fungal and mycobacterial infection. Like other investigators [18, 19], we found that this media type is useful for isolating other bacterial organisms [20]. The optimal threshold for DTD when using MFL culture vials is 150 min, with a positive likelihood ratio of 21.5 and sensitivity and specificity of 86% and 96%, respectively. Therefore, media type should be considered when validating and interpreting DTD values. Although determination of CRI is optimal using AF culture—related data, when necessary, MFL culture—related data can be used instead.

There are some important limitations of this study. Although the sample size is small, the compelling results of this study and their consistency with previously published data lend credence to the conclusions. Second, the question of external validity of the study must be addressed. On the basis of age, sex, catheter type, weight class, presence of neutropenia, and diagnosis of underlying disease, there was no difference between the 33 patients included for analysis and the 64 patients who had bacteremia but could not be categorized as having CRI or non-CRI during the first 6 months of the study (data not shown). In addition, according to DTD findings, when the 10 patients who had positive paired AF culture results but not positive ISO culture results were examined, all 10 were categorized as having non-CRI. This is expected, because, in our experience [20], the ISO system is a less sensitive detection system than the BAC system, which uses a weight-based volume of blood inoculation; therefore, the ISO system would be more likely to detect the high-grade bacteremia associated with CRI than the low-grade bacteremia typical of non-CRI (table 2). Another limitation arises from inclusion of culture data from paired blood culture bottles (with blood samples obtained via PV and CVC) with differing volumes of blood inoculum (data not shown). Although such a difference might alter absolute times to detection, it did not seem to influence the results of this study. This could partially be ascribed to the relatively large difference in time to detection seen in cases of CRI, such that comparatively minor differences resulting from differences in the volume of blood inoculated may not have affected the test accuracy. Finally, no study can prove the value of a method beyond the sensitivity and specificity of the stated criterion standard, put at 79% and 94%, respectively, for paired quantitative cultures by one recent meta-analysis [2]. Despite this limitation, DTD appears at least as accurate for diagnosis of CRI as paired quantitative blood cultures. In addition, although time to detection can be calculated for any positive culture, this is not always true of colony counts, which are difficult to accurately assess beyond a certain upper limit (400 cfu/mL for our laboratory). DTD should therefore prove useful in determining the relatedness of an infection to a CVC even in the setting of high colony counts (i.e., >400 cfu/mL) in cultures of samples obtained from both CVCs and PVs, when paired quantitative blood cultures are not interpretable.

To summarize, it is important to consider the following when applying the results of this study.

When calculating DTD, paired CVC- and PV-obtained blood cultures should each be inoculated to identical media, preferably AF (in the case of the BAC system).

In patients with a double-lumen catheter, only 1 lumen may be infected, and there could be a major difference in the bacterial load in blood samples obtained from the 2 lumens. Therefore, samples obtained from both lumens should be cultured, and the culture with shortest time to detection should be used for calculating the DTD.

Major discrepancies in volume of blood inoculated from a CVC and PV should be viewed with caution in the determination of CRI by DTD.

The performance of this test in partially treated patients remains to be assessed and may be difficult to study.

This test is best used when the results of CVC- and PV-obtained blood cultures are both positive; the significance of only 1 positive culture result remains to be determined.

DTD of paired CVC-obtained and PV-obtained blood cultures is an easy and rapid way to distinguish CRI from non-CRI in immunocompromised pediatric patients. Larger studies must be performed to confirm the promising results seen in this patient population and to answer questions concerning the influence of volume discrepancies, as well as the potential impact of differences in culture media and growth detection systems. Because many hospitals in the United States use similar blood culture systems, this relatively simple test could have widespread application.

Previous SectionNext Section

Acknowledgment

We thank Joan Hu, Department of Biostatistics, SJCRH, for assisting in the analysis in this study.

Previous SectionNext Section

Footnotes

  • ↵a Senior coauthors.

  • Financial support: National Cancer Institute Cancer Center Support (CORE grant P30 CA 21765) and American Lebanese Syrian Associated Charities.

  • Received November 12, 2002.
  • Accepted March 3, 2003.
Previous Section
 

References

  1. ↵
    1. Mermel LA,
    2. Farr BM,
    3. Sherertz RJ,
    4. et al
    . Guidelines for the management of intravascular catheter-related infection. Clin Infect Dis 2001;32:1249-72.
    FREE Full Text
  2. ↵
    1. Siegman-Igra Y,
    2. Anglim AM,
    3. Shapiro DE,
    4. et al
    . Diagnosis of vascular catheter-related bloodstream infection: a meta-analysi. J Clin Microbiol 1997;35:928-36.
    Abstract/FREE Full Text
  3. ↵
    1. Wing EJ,
    2. Norden CW,
    3. Shadduck RK,
    4. et al
    . Use of quantitative bacteriologic techniques to diagnose catheter-related sepsi. Arch Intern Med 1979;139:482-3.
    Abstract/FREE Full Text
  4. ↵
    1. Flynn PM,
    2. Shenep JL,
    3. Stokes DC,
    4. et al
    . In situ management of confirmed central venous catheter-related bacteremi. Pediatr Infect Dis J 1987;6:729-34.
    MedlineWeb of Science
    1. Flynn PM,
    2. Shenep JL,
    3. Barrett FF
    . Differential quantitation with a commercial blood culture tube for diagnosis of catheter-related infectio. J Clin Microbiol 1988;26:1045-6.
    Abstract/FREE Full Text
  6. ↵
    1. Capdevila JA,
    2. Planes AM,
    3. Palomar M,
    4. et al
    . Value of differential quantitative blood cultures in the diagnosis of catheter-related sepsi. Eur J Clin Microbiol Infect Dis 1992;11:403-7.
    CrossRefMedlineWeb of Science
  7. ↵
    1. Blot F,
    2. Schmidt E,
    3. Nitenberg G,
    4. et al
    . Earlier positivity of central-venous-versus peripheral-blood cultures is highly predictive of catheter-related sepsi. J Clin Microbiol 1998;36:105-9.
    Abstract/FREE Full Text
    1. Rogers MS,
    2. Oppenheim BA
    . The use of continuous monitoring blood culture systems in the diagnosis of catheter related sepsi. J Clin Pathol 1998;51:635-7.
    Abstract
  9. ↵
    1. Haimi-Cohen Y,
    2. Vellozzi EM,
    3. Rubin LG
    . Initial concentration of Staphylococcus epidermidis in simulated pediatric blood cultures correlates with time to positive results with the automated, continuously monitored BACTEC blood culture syste. J Clin Microbiol 2002;40:898-901.
    Abstract/FREE Full Text
  10. ↵
    1. Blot F,
    2. Nitenberg G,
    3. Chachaty E,
    4. et al
    . Diagnosis of catheter-related bacteraemia: a prospective comparison of the time to positivity of hub-blood versus peripheral-blood culture. Lancet 1999;354:1071-7.
    CrossRefMedlineWeb of Science
  11. ↵
    1. Malgrange VB,
    2. Escande MC,
    3. Theobald S
    . Validity of earlier positivity of central venous blood cultures in comparison with peripheral blood cultures for diagnosing catheter-related bacteremia in cancer patient. J Clin Microbiol 2001;39:274-8.
    Abstract/FREE Full Text
  12. ↵
    1. Seifert H,
    2. Cornely O,
    3. Seggewiss K,
    4. et al
    . Bloodstream infection in neutropenic cancer patients related to short-term nontunnelled catheters determined by quantitative blood cultures, differential time to positivity, and molecular epidemiological typing with pulsed-field gel electrophoresi. J Clin Microbiol 2003;41:118-23.
    Abstract/FREE Full Text
  13. ↵
    1. Rijnders BJ,
    2. Verwaest C,
    3. Peetermans WE,
    4. et al
    . Difference in time to positivity of hub-blood versus nonhub-blood cultures is not useful for the diagnosis of catheter-related bloodstream infection in critically ill patient. Crit Care Med 2001;29:1399-403.
    CrossRefMedlineWeb of Science
  14. ↵
    1. Kellogg JA,
    2. Ferrentino FL,
    3. Liss J,
    4. et al
    . Justification and implementation of a policy requiring two blood cultures when one is ordere. Lab Med 1994;25:323-30.
  15. ↵
    1. Viscoli C,
    2. Garaventa A,
    3. Boni L,
    4. et al
    . Role of broviac catheters in infections in children with cance. Pediatr Infect Dis J 1988;7:556-60.
    MedlineWeb of Science
    1. Castagnola E,
    2. Garaventa A,
    3. Viscoli C,
    4. et al
    . Changing pattern of pathogens causing broviac catheter-related bacteraemias in children with cance. J Hosp Infect 1995;29:129-33.
    CrossRefMedlineWeb of Science
  17. ↵
    1. Das I,
    2. Philpott C,
    3. George RH
    . Central venous catheter-related septicaemia in paediatric cancer patient. J Hosp Infect 1997;36:67-76.
    CrossRefMedlineWeb of Science
  18. ↵
    1. Fuller DD,
    2. Davis TE Jr,
    3. Denys GA,
    4. et al
    . Evaluation of BACTEC MYCO/F lytic medium for recovery of mycobacteria, fungi, and bacteria from bloo. J Clin Microbiol 2001;39:2933-6.
    Abstract/FREE Full Text
  19. ↵
    1. Vetter E,
    2. Torgerson C,
    3. Feuker A,
    4. et al
    . Comparison of the BACTEC MYCO/F Lytic bottle to the isolator tube, BACTEC Plus Aerobic F/bottle, and BACTEC Anaerobic Lytic/10 bottle and comparison of the BACTEC Plus Aerobic F/bottle to the Isolator tube for recovery of bacteria, mycobacteria, and fungi from bloo. J Clin Microbiol 2001;39:4380-6.
    Abstract/FREE Full Text
  20. ↵
    1. Gaur AH,
    2. Giannini MA,
    3. Flynn PM,
    4. et al
    . Optimizing blood culture practices in pediatric immunocompromised patients: evaluation of media types and blood culture volum. Pediatr Infect Dis 2003;22:545-52.
    CrossRefWeb of Science
| Table of Contents

This Article

  1. Clin Infect Dis. (2003) 37 (4): 469-475. doi: 10.1086/376904
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most