Skip Navigation

Time to Positivity in Staphylococcus aureus Bacteremia: Possible Correlation with the Source and Outcome of Infection

  1. Riad Khatib1,2,
  2. Kathleen Riederer1,
  3. Sajjad Saeed1,
  4. Leonard B. Johnson1,2,
  5. Mohamad G. Fakih1,2,
  6. Mamta Sharma1,
  7. M. Shamse Tabriz1, and
  8. Amir Khosrovaneh1
  1. 1Section of Infectious Diseases, Department of Medicine, St. John Hospital and Medical Center, Detroit, Michigan
  2. 2Wayne State University School of Medicine, Detroit, Michigan
  1. Reprints or correspondence: Dr. Riad Khatib, Dept. of Medical Education, St. John Hospital and Medical Center, 22101 Moross Rd., Detroit, MI 48236 (riad.Khatib{at}stjohn.org).
 
Next Section

Abstract

Background. Staphylococcus aureus bacteremia often persists and causes metastatic infections. It is unknown whether the time between blood culture incubation and growth detection (i.e., the time to positivity) in a continuously monitored system—a probable surrogate marker of bacteremia severity—correlates with outcome.

Methods. We performed a prospective, observational study involving adult inpatients who had S. aureus bacteremia between 1 January 2002 and 30 June 2003 at a 600-bed teaching hospital. Measurements included time to positivity in initial blood culture series, duration of bacteremia, rate of metastatic infection, and outcome.

Results. A total of 376 S. aureus bacteremias (⩾1 positive blood culture result) were reported for 357 patients aged 18–103 years (median age, 59 years); 64 bacteremias were excluded because blood was drawn after antibiotic therapy was started (n = 59) or through an intravascular catheter (n = 5). The source of infection was identified in 244 series (78.2%). Metastatic infection was detected in 25 bacteremias (8.0%). The mortality rate was 25.6%. The duration of bacteremia (determined in 251 series) was 1–59 days (median duration, 1 day; 70th percentile, 3 days). The time to positivity ranged from 4.2 to 98.2 h (median time to positivity, 15.5 h) and was significantly shorter for patients with an endovascular source of infection (14.9 ± 5.4 vs. 19.5 ± 10.6 h; P < .0005), extended duration (i.e., ⩾3 days) of bacteremia (14.1 ± 4.2 vs. 18.6 ± 9.2 h; P < .0005), and metastatic infection (12.9 ± 5.9 vs. 18.0 ± 9.3 h; P = .007). Analysis of a range of cutoff values demonstrated that a time to positivity of ⩽14 h yielded the best sensitivity and specificity for predicting the source and outcome of infection. Logistic regression analyses revealed that a time to positivity of ⩽14 h was an independent predictor of an endovascular source of infection (P < .0005), extended bacteremia (P < .0005), metastatic infection (P < .0005), and attributable mortality (P = .017).

Conclusions. Time to positivity in S. aureus bacteremia may provide useful diagnostic and prognostic information. Growth of S. aureus within 14 h after the initiation of incubation may identify patients with a high likelihood of endovascular infection sources, delayed clearance, and complications.

Staphylococcus aureus bacteremia is known for its propensity to cause metastatic infection [1–4]. Early recognition of patients who have endocarditis or are at risk for related complications often represents a challenge [2, 3]. There are clinical criteria suggestive of metastatic infection, but they remain insensitive [1, 3]. The characteristics of bacteremia may provide diagnostic information. For instance, in persons with endocarditis, bacteremia is continuous and associated with higher loads (up to 200 cfu/mL) [1], whereas in persons with gingival disease, bacteremia usually has a short duration and is associated with lower loads (up to 5 cfu/mL) [5]. Because quantitative blood cultures are not routinely performed, the time between incubation onset and growth detection (defined as the time to positivity), revealed by results of a continuously monitored blood culture system, can be used as a surrogate marker for bacteremia severity [6]. This assumption was corroborated in simulated pediatric blood culture systems spiked with Staphylococcus epidermidis [7]. Furthermore, the differential time to detection between culture of peripheral blood and culture of blood obtained from an intravenous catheter (reflecting a higher inoculum in blood drawn from the catheter) had been advocated for in vivo diagnosis of intravenous catheter—associated bacteremia [8–12]. We studied time to positivity in S. aureus bacteremia and correlated the results with the source and outcome of infection.

Previous SectionNext Section

Methods

This study was conducted at St. John Hospital and Medical Center, a 600-bed teaching hospital in Detroit, Michigan. Blood culture results obtained from 1 January 2002 through 30 June 2003 were monitored. All consecutive adult inpatients (age, ⩾18 years) with ⩾1 blood culture positive for S. aureus were identified and monitored prospectively. The following information was recorded for each patient: demographic characteristics, clinical characteristics, implicated source of infection, duration of bacteremia, metastatic infections, and outcome. Blood culture was repeated after 2–3 days initially and then every 1–5 days (median interval, every 2 days) thereafter until clearance (for most patients). Follow-up observation up to 100 days after initial hospitalization was done in the outpatient setting and/or through review of records associated with subsequent hospitalization. The conduct of research followed human experimentation guidelines of the US Department of Health and Human Services and St. John Hospital and Medical Center.

Blood culture and time to positivity. Approximately 10 mL of blood was inoculated into Bactec Plus aerobic and anaerobic bottles, which were then transported to the laboratory and incubated in an automated continuous monitoring system. The Bactec 9240 system (BD Diagnostic Systems) was used during the first 6 months of the study, and the BacT/Alert system (BioMerieux) was used thereafter. Both systems monitor CO2 production every 10 min by means of a fluorescent signal (Bactec) or a colorimetric signal (BacT/Alert). Bottles with positive results were examined by Gram staining, and their contents were subcultured. Species identification and susceptibility tests were performed using Staphaurex latex agglutination (Remel) or the Vitek identification and susceptibility cards (bioMùrieux Vitek). Time to positivity, defined as the time between the start of incubation and the start of the alert signal (as documented by the monitoring system), was recorded for each bottle from the initial blood culture series.

The study was approved by the institutional review board and the human investigation committee at St. John Hospital. Requirement of patient consent was exempted because of the observational nature of the investigation.

Definitions. The initial blood culture series comprised ⩾1 blood culture performed during the first assessment of each patient. Bacteremia was defined as ⩾1 positive blood culture result with systemic manifestations of infection. The duration of bacteremia was calculated as the number of days between the first and the final positive blood culture result. Extended bacteremia was defined as bacteremia with a duration of ⩾3 days [13]. Source of bacteremia was determined according to Centers for Disease Control and Prevention definitions of nosocomial infections [14] and the presence of clinical signs with isolation of S. aureus from the presumed source. Sources included intravenous catheter, heart valve (endocarditis), other endovascular sources (aneurysm, vascular graft, or other endovascular devices), respiratory tract, soft tissue, bone, urinary tract, miscellaneous, and unknown (systemic signs of infection without an obvious source or in the presence of multiple potential sources). Endocarditis was diagnosed by means of the modified Duke criteria [15]. Metastatic infections were defined as distant foci that were anatomically unrelated to the source. Outcomes included cure, death, and relapse. Relapse was defined as recurrence of bacteremia within 100 days after initial hospitalization. Attributable mortality was defined as death with persistent signs of sepsis or positive blood culture results.

Statistical methods. Data are presented as incidences or rates for categorical variables and as ranges, medians, and 70th percentiles for continuous variables. The statistical significance of observed differences was assessed by means of the χ2 test or Fisher's exact test for categorical variables; Student's t test, analysis of variance, or the Mann-Whitney U test for continuous variables; nonparametric correlation (Spearman's ρ) for association; the log-rank test for Kaplan-Meier time-to-positivity analyses; and logistic regression for multivariate analyses. We also constructed a receiver operating characteristic (ROC) curve by plotting the rate of true-positive results (i.e., sensitivity) against the rate of false-positive results (i.e., 1 - specificity) over a range of cutoff values of time to positivity. All tests were performed using SPSS, version 10 (SPSS). P < .05 was considered to be statistically significant.

Previous SectionNext Section

Results

Study population. A total of 376 S. aureus bacteremias were reported for 357 patients aged 18–103 years (median age, 60 years) during the study period; 64 bacteremias (17.0%) were excluded because blood for initial culture was obtained after starting antibiotic therapy (59 bacteremias) or through a central intravascular catheter (5 bacteremias). Of the remaining 312 bacteremias, S. aureus was detected with other pathogens in 17 (5.4%) and with common contaminants in 20 (6.4%); they were included in the study. Patient characteristics are summarized in table 1.

Figure 1
View larger version:
Figure 1

Time to positivity stratified according to infection source. See Methods for the definition of time to positivity. IE, infective endocarditis; IVC, intravascular catheter; O-endovasc: other endovascular source; UK, unknown. Top of bars, mean values; +, individual isolates.

Figure 2
View larger version:
Figure 2

Comparisons of time to positivity (up to 30 h) in initial blood culture series according to source of infection (A; log-rank test statistic, 27.64), duration of bacteremia (B; log-rank test statistic, 33.33), and presence of metastatic infection (C; log-rank test statistic, 15.42). See Methods for the definition of time to positivity.

Table 1
View larger version:
Table 1

Demographic and clinical characteristics of 376 patients with Staphylococcus aureus bacteremia.

Oxacillin-resistant S. aureus isolates were found in 163 blood culture series (52.2%). Metastatic infection was detected in 25 bacteremias (8.0%), relapse of bacteremia was noted in 13 instances (4.2%; i.e., 5.6% of patients who completed 100 days of follow-up), the all-cause mortality rate was 25.6% (91 patients), and the attributable mortality rate was 11.9% (42 patients). The duration of bacteremia, which was determined for 251 (80.4%) of 312 bacteremias that included follow-up blood cultures, was 1–59 days (median duration, 1 day; 70th percentile, 3 days). Extended bacteremia (duration, ⩾3 days) was noted in 111 series (32.1%). Transthoracic and/or transesophageal echocardiography was performed for 158 patients (44.2%; i.e., 71% of 223 patients with extended bacteremia).

Time to positivity. Time to positivity was recorded for 931 individual aerobic and/or anaerobic bottles from the initial positive blood culture series (range, 1–4 bottles with positive results per series). Time to positivity ranged from 4.2 to 98.2 h; there was a 3.7-h median difference (range, 0–75.4 h) in the time to positivity between bottles with positive results in individual series. The shortest time to positivity in any bottle and the mean time to positivity for all bottles correlated strongly (ρ = 0.916; P < .0005). For our analysis, we used the shortest time to positivity (median time to positivity, 15.5 h). Growth was faster in patients with oxacillin-susceptible isolates, compared with patients with oxacillin-resistant isolates (mean time to positivity [±SD], 16.2 ± 7.8 vs. 18.8 ± 10.0 h; P = .01).

Correlation of time to positivity with the source of infection. We stratified time to positivity according to the source of infection (figure 1). The mean time to positivity (±SD) was shortest for endocarditis (12.1 ± 4.0 h; P = .002) and intravenous catheters (15.1 ± 5.3 h; P = .004) and longest for bone (20.6 ± 9.5 h; P = .04) and the respiratory tract (24.8 ± 15.3 h; P = .001). Kaplan-Meier analysis of time to positivity in series with endovascular sources (intravenous catheter, heart valve, and other endovascular sources) and nonendovascular sources (all other sources, excluding series with an unknown source) demonstrated statistically significant differences (figure 2A). ROC analysis revealed that a time to positivity of ⩽14 h was associated with the best sensitivity and specificity (74% and 57%, respectively). Stepwise forward logistic regression analysis (with endovascular source as the dependent variable and age, comorbid conditions, type of onset, cardiovascular or orthopedic prosthesis, and oxacillin susceptibility as covariates) revealed that a time to positivity of ⩽14 h was an independent predictor of isolation of S. aureus from an endovascular source (table 2).

Table 2
View larger version:
Table 2

Stepwise multiple logistic regression analyses of a time to positivity of ⩽14 h for predicting an endovascular source and the outcome of Staphylococcus aureus infection.

Correlation of time to positivity with bacteremia duration. Correlation with bacteremia duration was determined for 261 (83.7%) of 312 bacteremias that included follow-up blood cultures. Bacteremias without follow-up cultures were more likely to be from females (32 [62.7%] of 51 vs. 108 [41.4%] of 261; P = .005), to have a single positive blood culture result (19 [48.8%] of 51 vs. 49 [18.8%] of 261; P = .02), and to have an unknown source of bacteremia (20 [37.3%] of 41 vs. 60 [23.5%] of 255; P = .003) than were patients with follow-up cultures. Time to positivity correlated negatively with bacteremia duration (ρ = -0.336; P = .01). Stratification of time to positivity according to the duration of bacteremia (1–2 days [50th percentile], 3–6 days [70th percentile], and ⩾7 days [90th percentile]) revealed significant differences between bacteremia with a duration of 1–2 days and bacteremia that lasted for 3–6 days (P < .0005) or ⩾7 days (P = .009), whereas the difference in the time to positivity between 3–6-day bacteremia and ⩾7-day bacteremia was not significant.

Kaplan-Meier analysis of time to positivity in patients who were and patients who were not associated with extended bacteremia revealed statistically significant differences (figure 2B). A time to positivity of ⩽14 h was associated with higher risk for extended bacteremia (relative risk [RR], 2.02; 95% CI, 1.48–2.74; P < .0005). It was also the best independent predictor in regression analysis after entering the previously mentioned covariates, as well as the time to initiation of antibiotic therapy (table 2).

To avoid any bias associated with excluding patients for whom bacteremia cleared but blood culture was not repeated, we estimated the duration of bacteremia for all patients on the basis of culture results plus clinical characteristics, assuming that the final positive culture result represented the last day of bacteremia in patients whose fever had resolved. Patients who died within 4 days after blood was initially obtained for culture were assumed to have bacteremia until death and were censored thereafter. All analyses were performed again using the estimated duration, and results similar to those in the initial analysis were obtained (data not shown).

Correlation of time to positivity with the outcome. Kaplan-Meier analysis of time to positivity revealed statistically significant differences between bacteremias that were and bacteremias that were not associated with metastatic infections (figure 2C). Additionally, a time to positivity of ⩽14 h was associated with significant risk for developing metastatic infection (RR, 5.07; 95% CI, 2.08–12.33; P < .0005); it was also an independent predictor of the development of metastatic infection and of attributable mortality (table 2). The difference in the mean time to positivity (±SD) between series that were and series that were not associated with relapse did not reach statistical significance (13.2 ± 5.4 vs. 17.8 ± 9.2 h; P = .08).

Previous SectionNext Section

Discussion

The time to positivity in blood cultures can be influenced by multiple factors, including the inoculum, the volume of blood drawn, the incubation conditions, and the presence of growth inhibitors, such as antibiotics [16]. A potential limitation in our study is that the volume of blood drawn for culture was not tracked. We assumed that variations occurred randomly and that they were corrected for by the large number of cases studied. Incubation conditions were probably stable, because of the automation involved in most routine microbiology procedures. The presence of antibiotics should be a minor confounding factor in initial blood cultures. Furthermore, we excluded all patients who were receiving antibiotics at the time of initial blood cultures. Thus, the difference in time to positivity is most likely related to bacteremia load. Our patient population was similar to those reported elsewhere with respect to patient characteristics, source of bacteremia, and the prevalence of metastatic infection [2, 3, 13], but the relapse rate was somewhat lower, although a similar rate has been reported [19]. We hypothesized that higher-load S. aureus bacteremias originate from endovascular sources and are associated with more-frequent complications. Our investigation supports this hypothesis and suggests that time to positivity may provide useful diagnostic and prognostic information. Faster growth in blood cultures was associated with a higher probability of an endovascular source, delayed clearance, and complications. ROC curves revealed that a time to positivity of ⩽14 h was associated with satisfactory sensitivity and specificity. The presence of other organisms in blood cultures did not influence the results (data not shown).

The patient's immune status, the presence of neutropenia, and the duration of illness before blood is drawn might also be important determinants of bacteremia severity. The small number of patients with these conditions and our failure to track the duration of illness precluded a meaningful assessment of these potential variables. Additional studies are necessary to clarify the effect of immune status and duration of illness on time to positivity.

Bacteremias originating from endovascular foci are known to have high bacteria loads [1] and are thus likely to be associated with shorter times to positivity. Additionally, more-severe bacteremias would be expected to cause a higher incidence of metastatic infection and to be more difficult to clear.

The difference in time to positivity between patients with oxacillin-susceptible isolates and patients with oxacillin-resistant isolates was interesting. Although oxacillin-resistant isolates are known to have a longer doubling period [17, 18], time to positivity from clinical samples with oxacillin-resistant or oxacillin-susceptible isolates has not been reported. Additional studies are needed to verify this observation, to assess its relevance, and to determine whether cutoff values should be stratified according to oxacillin-susceptibility.

In conclusion, although the sensitivity and specificity of time to positivity are suboptimal, this marker may provide useful diagnostic and prognostic information for patients with S. aureus bacteremia. Time to positivity may aid in identifying the source of infection in patients with multiple potential sources, optimize the utility of diagnostic studies, and influence considerations about whether to remove an existing intravenous catheter.

Previous SectionNext Section

Acknowledgments

We thank Dr. Ruth Moore and Alice Mar for their assistance with statistical analyses and careful copyediting of the manuscript.

Financial support. Medical Education Funds, St. John Hospital.

Potential conflicts of interest. All authors: no conflicts.

  • Received December 7, 2004.
  • Revision received April 12, 2005.
Previous Section
 

References

  1. ↵
    1. Mandell GL,
    2. Bennett JE,
    3. Dolin R
    1. Bayer AS,
    2. Scheld WM
    . Endocarditis and intravascular infections. In: Mandell GL, Bennett JE, Dolin R, editors. Mandell, Douglas and Bennett's principles and practice of infectious diseases 5th ed. Vol. 1. Philadelphia: Churchill Livingstone; 2000. p. 857-902.
  2. ↵
    1. Fowler VG,
    2. Olsen MK,
    3. Corey GR,
    4. et al
    . Clinical identifiers of complicated Staphylococcus aureus bacteremia. Arch Intern Med 2003;163:2066-72.
    Abstract/FREE Full Text
  3. ↵
    1. Johnson LB,
    2. Almoujahed MO,
    3. Ilg K,
    4. Maolood L,
    5. Khatib R
    . Staphylococcus aureus bacteremia: compliance with standard treatment, long-term outcome and predictors of relapse. Scand J Infect Dis 2003;35:782-9.
    CrossRefMedlineWeb of Science
  4. ↵
    1. Archer GL,
    2. Climo MW
    . Staphylococcus aureus bacteremia—consider the source. N Engl J Med 2001;344:55-6.
    CrossRefMedlineWeb of Science
  5. ↵
    1. Everett ED,
    2. Hirschmann JV
    . Transient bacteremia and endocarditis prophylaxis: a review. Medicine 1977;56:61-77.
    Medline
  6. ↵
    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 sepsis. J Clin Microbiol 1998;36:105-9.
    Abstract/FREE Full Text
  7. ↵
    1. Haim-Cohen Y,
    2. Vellozi 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 system. J Clin Microbiol 2002;40:898-901.
    Abstract/FREE Full Text
  8. ↵
    1. Blot F,
    2. Nitenberg G,
    3. Chachaty E,
    4. et al
    . Diagnosis of catheter-related bacteremia; a prospective comparison of the time to positivity of hub-blood versus peripheral-blood culture. Lancet 1999;354:1071-7.
    CrossRefMedlineWeb of Science
    1. Gaur AH,
    2. Flynn PM,
    3. Giannini MA,
    4. Shenep JL,
    5. Hayden RT
    . Difference in time to detection: a simple method to differentiate catheter-related from non-catheter-related bloodstream infection in immunocompromised pediatric patients. Clin Infect Dis 2003;37:469-75.
    Abstract/FREE Full Text
    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 patients. J Clin Microbiol 2001;39:274-8.
    Abstract/FREE Full Text
    1. Seifert H,
    2. Cornely O,
    3. Seggwiss K,
    4. et al
    . Bloodstream infection in neutropenic cancer patients related to short-term nontunneled catheters determined by quantitative blood cultures, differential time to positivity, and molecular epidemiological typing with pulse-field gel electrophoresis. J Clin Microbiol 2003;41:118-23.
    Abstract/FREE Full Text
  12. ↵
    1. Raad I,
    2. Hann HA,
    3. Alakech B,
    4. Chatzinikolaou C,
    5. Johnson MM,
    6. Tarrand J
    . Differential time to positivity: a useful method for diagnosing catheter-related bloodstream infections. Ann Intern Med 2004;140:18-25.
    Abstract/FREE Full Text
  13. ↵
    1. Fowler VG Jr,
    2. Sanders LL,
    3. Sexton DJ,
    4. et al
    . Outcome of Staphylococcus aureus bacteremia according to compliance with recommendations of infectious diseases specialists: experience with 244 patients. Clin Infect Dis 1998;27:478-86.
    Abstract/FREE Full Text
  14. ↵
    1. Garner GS,
    2. Jarvis WR,
    3. Emory TG,
    4. Horan TC,
    5. Hughes JM
    . CDC definitions for nosocomial infections. Am J Infect Control 1988;16:128-40.
    CrossRefMedlineWeb of Science
  15. ↵
    1. Li JS,
    2. Sexton DJ,
    3. Mick N,
    4. et al
    . Proposed modifications to the Duke criteria for the diagnosis of infective endocarditis. Clin Infect Dis 2000;30:633-8.
    Abstract/FREE Full Text
  16. ↵
    1. Plorde JJ,
    2. Tenover FC,
    3. Carlson LG
    . Specimen volume versus yield in the BACTEC blood culture system. J Clin Microbiol 1985;22:292-5.
    Abstract/FREE Full Text
  17. ↵
    1. Okuma K,
    2. Iwakawa K,
    3. Tunidge JD,
    4. et al
    . Dissemination of new methicillin-resistant Staphylococcus aureus clones in the community. J Clin Microbiol 2002;40:4289-94.
    Abstract/FREE Full Text
  18. ↵
    1. Laurent F,
    2. Lelievre H,
    3. Cornu M,
    4. et al
    . Fitness and competitive growth advantage of new gentamicin-susceptible MRSA clones spreading in French hospitals. J Antimicrob Chemother 2001;47:277-83.
    Abstract/FREE Full Text
  19. ↵
    1. Malanoski GJ,
    2. Samore MH,
    3. Pefanis A,
    4. Karchmer AW
    . Staphylococcus aureus catheter-associated bacteremia. Arch Intern Med 1995;155:1161-6.
    Abstract/FREE Full Text
| Table of Contents

This Article

  1. Clin Infect Dis. (2005) 41 (5): 594-598. doi: 10.1086/432472
  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