Chest ACCP Career Connection
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Guest Access | Sign In via User Name/Password
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Shafazand, S.
Right arrow Articles by Weinacker, A. B.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Shafazand, S.
Right arrow Articles by Weinacker, A. B.
(Chest. 2002;122:1727-1736.)
© 2002 American College of Chest Physicians

Blood Cultures in the Critical Care Unit*

Improving Utilization and Yield

Shirin Shafazand, MD and Ann B. Weinacker, MD, FCCP

* From the Division of Pulmonary and Critical Care, Department of Medicine, Stanford University, Stanford, CA.

Correspondence to: Ann Weinacker, MD, FCCP, Assistant Professor of Medicine, Stanford University Medical Center, Division of Pulmonary and Critical Care, 300 Pasteur Dr, Room H3142, Stanford, CA 94305-5236; e-mail: annw{at}stanford.edu


   Abstract
TOP
 Abstract
Introduction
 Utilization and Yield of...
 Methods to Improve Blood...
 Contaminant Blood Cultures:...
 Predictors of Bacteremia
 Conclusion
 References
 
Sepsis is a common cause of morbidity and death in critically ill patients, and blood culture samples are often drawn in an effort to identify a responsible pathogen. Blood culture results are usually negative, however, and even when positive are sometimes difficult to interpret. Distinguishing between true bacteremia and a false-positive blood culture result is important, but complicated by a variety of factors in the ICU. False-positive culture results are costly because they often prompt more diagnostic testing and more antibiotic prescriptions, and increase hospital length of stay. A number of factors influence the yield of blood cultures in critically ill patients, including the use of antibiotics, the volume of blood drawn, the frequency with which culture samples are drawn, and the site from which the culture samples are taken. Skin preparation techniques, handling of the cultures in the microbiology laboratory, and the type of blood culture system employed also influence blood culture yield. Attempts to identify predictors of true bacteremia in critically ill patients have been disappointing. In this review, we discuss factors that influence blood culture yield in critically ill patients, suggest ways to improve yield, and discuss true bacteremia vs false-positive blood culture results. We also discuss the costs and consequences of false-positive blood culture results, and list noninfectious causes of fever in the ICU.

Key Words: bacteremia • blood cultures • critically ill • ICU • sepsis


   Introduction
TOP
 Abstract
 Introduction
Utilization and Yield of...
 Methods to Improve Blood...
 Contaminant Blood Cultures:...
 Predictors of Bacteremia
 Conclusion
 References
 
Sepsis, a systemic inflammatory response to infection, is an important cause of morbidity and mortality in critically ill patients, resulting in frequent diagnostic testing, greater prescription of antibiotics, and increased length of hospitalization. In a prospective multicenter study1 of critically ill patients, clinically suspected sepsis occurred in approximately 9% of adults admitted to ICUs and had a mortality rate of up to 60%. However, microbiological infection was documented in only 71% of patients with suspected severe sepsis, and bacteremia was documented in 53%. Bacteremia (including fungemia) in critically ill patients has an average attributable mortality of 26%,2 with Candida spp3 and enterococci4 responsible for 38% and 31% of mortality, respectively.5 However, the microbiologic diagnosis of bacteremia acquired after admission is often difficult in the ICU, complicated by the fact that many critically ill patients receive empiric antibiotic therapy. In fact, antibiotic consumption is 10 times greater in ICUs than on general hospital wards.6 The significant health and potential economic impact of bacteremia underscore the importance of early detection and treatment of nosocomial infections in critically ill patients.

Blood cultures are the standard for diagnosing bacteremic episodes. Critically ill patients are at increased risk of acquiring nosocomial bacteremia because of their underlying disease processes, and because of the presence of invasive devices such as endotracheal tubes and central venous catheters. Of 111 consecutive episodes of nosocomial bacteremia in a medical-surgical ICU during a 36-month period, IV catheter infections were the most common sources of bacteremia. Coagulase-negative staphylococcus and Staphylococcus aureus were the most commonly isolated organisms. The overall mortality rate of patients with nosocomial bacteremia has been reported to be 31.5%.7 Because of the increased risk and associated high mortality of bacteremia, there is generally a low threshold for drawing blood culture samples in the ICU. Unfortunately, the overall yield of blood cultures is low.8 To increase the efficiency with which blood culture samples are obtained, a number of investigators have tried to identify predictors of bacteremia.9 10 11 12 13 14 15 16 17 However, there is currently no consensus on acceptable predictors of bacteremia in critically ill patients. In addition, up to 50% of positive blood culture results represent contamination of the culture, rather than true bacteremia,8 18 19 20 and distinguishing between true-positive culture results and contaminant cultures is often difficult.

In this article, we review the utilization and diagnostic yield of blood cultures in the ICU, and report the incidence of contaminated blood cultures and their economic impact. We explore potential predictors of bacteremia and highlight means of improving blood culture yield.


   Utilization and Yield of Blood Cultures in the ICU
TOP
 Abstract
 Introduction
 Utilization and Yield of...
Methods to Improve Blood...
 Contaminant Blood Cultures:...
 Predictors of Bacteremia
 Conclusion
 References
 
The utilization of blood cultures varies according to physician experience and an estimation of the probability of bacteremia in a given patient population. While there is no good estimate of the number of blood culture samples drawn in ICU settings on a monthly basis, there is some suggestion that this number is excessive. This is especially true in teaching hospitals, perhaps due to the inexperience of house-officers.21 In a study in a medical ICU, Gross et al22 implemented a protocol to limit the number of blood culture samples drawn during a 3-week period. Solitary blood culture samples were eliminated. Four blood culture sets were recommended for clinical suspicion of endocarditis, drawn 30 min to 1 h apart. For all other suspected cases of bacteremia, up to three blood culture sets were permitted. In patients with no change in their clinical status and with initially negative blood culture results, no further samples were drawn despite persistent fever. Study results were compared to a representative 3-week period prior to initiation of the protocol. During the study, the number of blood cultures ordered for suspected septic episodes decreased from 3.0 to 2.2 sets. The number of blood culture samples drawn was reduced from 1.2 cultures per discharge to 0.3 cultures per discharge with no untoward effects on patient care.22 A follow-up survey was conducted 2 months later, at a time when new residents rotating through the ICU were not familiar with the study protocol. Once again, an increase in the number of blood cultures per suspected septic episode was observed.

Despite a tendency to obtain repeated blood cultures in ICUs across the country, the overall yield is often disappointing. In a study of 645 admissions to a trauma and surgical ICU over a 1-year period, 32% of patients had at least one set of blood culture samples drawn while in the ICU. Of these culture results, only 7.5% were positive, and only 4% represented true bacteremia. All fungi, Gram-positive cocci, and Gram-negative rods were considered pathogens if they grew from two or more blood cultures or if they were isolated from another infectious focus (sputum, urine, wound) in addition to blood. Coagulase-negative staphylococci were considered pathogenic if there was evidence of urinary tract, CNS, or vascular catheter infection within 48 h of the blood culture. Bacillus, Micrococcus, and Proprionibacterium spp were regarded as contaminants.21 The most frequent causes of bacteremia in this study were catheter-related infections, and coagulase-negative staphylococci were the most frequent isolates. The overall incidence of true bacteremia was 3.6 per 100 ICU admissions. Similarly, in another study of septic patients in a surgical ICU, the incidence of positive blood culture results was 3.2 per 100 ICU admissions.23 In this study, sepsis was defined as a systemic inflammatory response syndrome associated with positive blood culture results. The authors, however, made no distinction between positive blood culture results and true bacteremia.

Surgical ICU patients are reported to be two to six times more likely than medical ICU patients to have bacteremia,5 21 24 despite the lower incidence of chronic illness in surgical patients. The greater number of procedures and invasive devices used in surgical ICU patients may account for the increased rates of bacteremia. Given the surgical population of patients studied by Darby et al,21 the incidence of bacteremia in their patients was unexpectedly low. Most of the culture samples were drawn while the patients were receiving antibiotics, and up to one third of the samples were solitary sets. These two factors at least partially account for the low incidence of bacteremia reported in this study, a rate lower than a comparable European study of mostly surgical ICUs with a reported incidence of bacteremia of 5.4 per 100 ICU admissions.25

In a review of nosocomial bacteremia in an adult ICU consisting predominately of medical patients, the incidence of bacteremia, during a 5-year observation period (from 1991 to 1995), was 3.7 per 100 ICU admissions.5 Pseudomonas spp, Enterococcus spp, and coagulase-negative staphylococci were the most frequently isolated organisms, and up to 75% of the bacteremic episodes were attributed to colonized intravascular catheters.

Methodologic differences aside, these findings are surprisingly similar to the previously discussed results by Darby et al.21 Despite these similarities, the yield of blood cultures in medical ICU patients has not been well studied, and results from the surgical critical care unit cannot be readily extrapolated to other ICU populations.


   Methods to Improve Blood Culture Yield
TOP
 Abstract
 Introduction
 Utilization and Yield of...
 Methods to Improve Blood...
Contaminant Blood Cultures:...
 Predictors of Bacteremia
 Conclusion
 References
 
The overall yield of blood cultures can be affected by several factors (Table 1 ). Solitary blood cultures, intermittent bacteremia, a low number of colony forming units detectable in blood, antimicrobial properties of blood components, and concurrent antibiotic treatment can all lower the yield of blood cultures. The specific blood collection system used (BACTEC; Becton Dickinson Diagnostic Instruments Systems; Sparks, MD, and DuPont Isolator; Wampole Laboratories; Cranbury, NJ) can play a role in increasing yields in particular clinical settings.26


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

  		Table 1. Factors Influencing Blood Culture Yield

 
Several authors have addressed the issue of the optimal number of blood cultures needed to detect bacteremia. In their analysis of 500 episodes of septicemia, Weinstein et al20 noted that 91.5% of all bacteremic episodes were detected by the first blood culture. An additional 7.8% were identified by a second blood culture.20 For patients with continuous bacteremia (eg, endocarditis) in whom one blood culture result was positive, the probability that subsequent culture results would be positive was > 95%. In cases of intermittent bacteremia, approximately 85% of second culture results were positive after an initial positive culture result. Of note, when the initial culture was judged to be contaminated (based on clinical criteria), the probability of a second positive blood culture result was < 5%, and with further testing this decreased to < 1%. When the second culture result was positive, it almost invariably grew a different (contaminating) organism. Based on this and similar studies, for the vast majority of patients there is no evidence that obtaining more than two or three sets of blood cultures (defined as blood cultures of samples drawn from two or three different sites) within a 24-h period improves diagnostic yield. Drawing more than one culture set is useful, however, since obtaining two or three sets can help distinguish true bacteremia from contaminated cultures.18 19 26 27 28 29

Several technical factors, including skin preparation methods, blood incubation time, type of culture media, and the type of blood collecting system used, can improve the yield of cultures (Table 2 ). Venipuncture is the preferred method for obtaining blood culture samples. Blood culture samples drawn from intravascular catheters are not optimal, as they can become contaminated with organisms colonizing the hub or the walls of the catheter. If it becomes necessary to obtain culture samples from these catheters, strict aseptic technique should be followed, while all efforts are made to draw a second set of culture samples from a peripheral venipuncture.


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

  		Table 2. Guidelines to Improve Blood Culture Yield

 
Aseptic blood collection techniques and careful processing of cultures in the microbiology laboratory have been shown to decrease the rate of occurrence of skin contaminants.29 30 One issue of interest is the choice of material used for skin disinfection. None of the commercially available antiseptic products work instantly; a minimum exposure time is required for each product to be effective. While this time varies by agent, it is generally recommended that sufficient time be allowed for the solution to dry prior to venipuncture.27 31 Materials often used include tincture of iodine, povidone-iodine alone or in combination with isopropyl alcohol, and recently, chlorhexidine. Published studies and recommendations vary in their choices of material for skin disinfection.

In one randomized study,32 blood culture contamination rates were observed after disinfection of the venipuncture site with either 10% povidone-iodine or 2% iodine tincture. A 70% isopropyl alcohol applicator was used for a 1-min scrub prior to the application of either disinfectant, and a 2-min drying time before phlebotomy was recommended for both povidone-iodine and iodine tincture.32 A total of 3,851 blood cultures were studied; 376 results were positive, of which 120 cultures were contaminated. Compared to povidone-iodine, iodine tincture was responsible for a statistically significant reduction in skin flora contamination (2.4% compared to 3.8% after disinfection with povidone-iodine; odds ratio, 1.6; p = 0.01). The difference noted may in part be due to the faster onset of action of tincture of iodine compared to povidone-iodine. The recommended drying time for povidone-iodine to have maximal antiseptic effect is 2 to 3 min. In a busy hospital setting such as the ICU, the nurses or house-officers may not wait for the required skin contact time prior to drawing a blood culture sample, rendering povidone-iodine less effective.

Another randomized controlled trial33 demonstrated that 0.5% alcoholic chlorhexidine applied to the skin 15 to 30 s prior to venipuncture reduced the incidence of blood culture contamination compared to povidone-iodine applied in a similar manner. Alcoholic chlorhexidine requires less time than povidone-iodine to achieve skin antisepsis (seconds compared to minutes), and this may lead to its greater usefulness in clinical practice. Current evidence suggests that iodine tincture, alone or in combination with isopropyl alcohol, or alcoholic chlorhexidine are both more effective than povidone-iodine in reducing the rates of contamination with skin flora. To date, no direct comparisons have been made between chlorhexidine and tincture of iodine. Nonetheless, the weight of the evidence in the literature advocates the use of some type of a disinfecting agent, and care must be taken to allow adequate time for effective antisepsis prior to venipuncture.

Few studies have evaluated the optimal timing of blood cultures. In clinical conditions with continuous bacteremia, such as endocarditis or septic thrombophlebitis, this issue is probably not very relevant. In cases of intermittent bacteremia, a report29 from the 1950s that has become the basis for traditional teaching suggests that bacteremia precedes the onset of fever and chills by 1 to 2 h. According to this report, blood culture results are often negative at the time fever occurs. For obvious reasons, drawing blood samples during the 1-h time window prior to fever can be practically impossible. Nonetheless, drawing multiple2 3 culture sets in a 24-h period has been shown to be sufficient to detect intermittent bacteremia.20 The ideal interval between blood cultures is not well known, but likely has less impact on the yield than was once thought.34 In a study of the optimal time interval between blood cultures, Li et al34 demonstrated that similar yields were obtained when samples were collected simultaneously, within 2 h, or within 24 h of the initial blood culture. In critically ill patients who are hemodynamically unstable, two blood culture sets should be drawn promptly prior to initiation of empiric antibiotic treatment. In less urgent cases, blood can be drawn over a 24-h period. In patients receiving antibiotics, it has been suggested that culture samples drawn close to the time that antibiotic concentrations have reached trough levels may improve yield.29 35 However, this has not been studied objectively and may not be clinically practical.

The volume of blood drawn is perhaps the most important factor in increasing the sensitivity of blood cultures. Adult patients with clinically significant bacteremia often have a low number of colony-forming units per milliliter of blood (1 to 10 cfu/mL). There is a direct relationship between the volume of blood obtained and the yield of the culture. Various studies reveal that at least 10 mL, but optimally 20 to 30 mL, of blood provide the highest yield and lowest number of false-negative blood culture results. Drawing > 30 mL of blood per culture does little to improve the sensitivity of the blood culture, and contributes to iatrogenic causes of anemia.26 34 36 37 38 39 In a study of blood culture practices at the University of Wisconsin Hospital and Clinics, investigators noted that 15% of all adult blood culture samples were collected in 3.5-mL pediatric blood culture tubes. Five percent of samples collected in the appropriate adult 10-mL tubes contained < 5 mL of blood. The sensitivity for detecting Gram-positive bacteremia with culture samples containing 10 mL of blood was 93%, compared to only 74% when smaller volumes of blood were sampled.38 A survey of 158 US clinical microbiology laboratory directors revealed comparable results; only 20% of the 71 laboratories responding to the survey reported routinely recording the volume of blood submitted, and up to 88% of these laboratories accepted samples that contained <= 5 mL of blood.38 Given the seemingly widespread practice of obtaining low volume blood samples, educating house-staff, phlebotomists and nurses about the adequate amount of blood needed for culture is a simple measure to increase the yield of blood cultures.

Blood culture yield is also decreased by blood components such as phagocytes, antibodies, and complement factors that are known to have bactericidal activity, impairing the isolation of microorganisms from blood cultures.20 Use of the anticoagulant sodium polyanetholsulfonate (SPS) 0.025% in blood culture media has been shown to help counteract these bactericidal effects by inhibiting phagocytosis and lysozymal activity.26 29 40 In addition to having antiphagocytic properties, SPS inactivates aminoglycosides. Although it has the potential drawback of inhibiting the growth of fastidious organisms such as Neisseria gonorrhoeae, Gardnerella vaginalis, and Neisseria meningitidis, increasing concentrations of SPS increase the recovery of Gram-negative bacteria.29 41 42 Virtually all commercially available blood culture bottles and media now use SPS.41

Many critically ill patients are on empiric antibiotic treatment at the time blood culture samples are drawn. In one study, up to 65% of blood culture samples were obtained from patients receiving antibiotics.21 Antibiotic therapy may significantly decrease the yield of blood cultures.29 43 Recently, blood culture yield in febrile hospitalized patients during 72 h of antibiotic therapy was retrospectively compared to the yield of blood culture samples obtained in the same patients prior to antibiotic administration.44 Fifty-six of 139 eligible patients (40%) had positive blood culture results prior to antibiotics. Of these, only 26 patients (less than half) continued to have positive culture results within 72 h of antibiotic use. Patients who had S aureus, aerobic Gram-negative bacilli, and streptococci isolated from preantibiotic cultures were more likely to have persistently positive blood culture results within 72 h of antibiotic use. Only one patient had a blood culture result during the 72-h study period that was positive for an organism that had not been isolated from her preantibiotic blood cultures. Isolation of S aureus from preantibiotic blood cultures and a diagnosis of endocarditis were the only independent predictors of positive blood culture results after antibiotic use. The data from this study suggest that successive blood cultures during the initial 72 h of antibiotic administration rarely yield new and clinically useful information.

The impact of antibiotic use on blood culture yield can be minimized by obtaining an adequate volume of blood for culture. Culturing 10 mL of blood per 100 mL of culture broth dilutes the concentrations of antibiotics and of neutralizing serum bactericidal activity in the cultures.26 Several manufacturers have marketed blood culture media containing antibiotic-adsorbent resins. Such media include BacT/Alert FAN media (Organon Teknika; Marcy L’etoile, France), a variety of resin-containing BACTEC culture media, and the antimicrobial removal device (ARD; Becton Dickinson Microbiology Systems; Cockeysville, MD) used with manual blood culture systems (eg, Septi-Chek; Becton Dickinson Microbiology Systems).39 Several reports have been published on the use of the antimicrobial removal device, but the results of these studies have been conflicting.45 46 47 Routine use of the antimicrobial removal device is generally not recommended; it may be useful in the setting of septic patients who have been receiving antibiotics and have had persistently negative blood culture results after 48 h of incubation.46

Most microbiology laboratories use automated blood culture systems, and the nonradiometric BACTEC systems are the most widely used automated systems.46 No one system currently available can isolate all the organisms that may be present in bacteremic patients. Blood cultures in patients with fungal and mycobacterial infections deserve special mention. Fungi are important nosocomial pathogens, especially in the critical care setting. Risk factors for the development of fungemia include the prolonged use of broad-spectrum antibiotics, the presence of intravascular catheters, the use of parenteral nutrition, and various immunocompromised states. Recovery of fungi in blood cultures is best achieved by lysis-centrifugation systems (DuPont Isolator) rather than conventional blood culture systems. The lysis of RBCs, centrifugation, and inoculation on antibiotic-free solid media increase the yield of microorganisms such as yeast, mycobacteria, filamentous fungi, and Legionella. The DuPont Isolator system also allows for faster recovery of these organisms and quantitation of organisms in blood cultures.

The BACTEC system (and other conventional systems such as BacT/Alert) is superior to the DuPont Isolator system for the isolation of Streptococcus pneumoniae, Pseudomonas aeruginosa, anaerobic bacteria, and other Streptococcus species.39 The use of a combination of BACTEC and DuPont Isolator systems allows for the detection of a broad range of bacterial and fungal infections. However, the regular use of both systems for all patients is neither indicated nor economically feasible. For routine clinical situations, the conventional culture systems are adequate. Lysis-centrifugation systems (eg, DuPont Isolator) should be used when there is suspicion of a fungal or mycobacterial infection, and in conjunction with expertise from the infectious disease and microbiology departments. To optimize blood culture yield, each individual microbiology lab should be consulted for information about which of the commercially available conventional systems are readily available and appropriate for the suspected pathogens. A more detailed discussion of blood culture media and systems is beyond the scope of this article, and the reader is referred to excellent reviews of this topic.27 37 39 48 49 50 51


   Contaminant Blood Cultures: Costs and Consequences
TOP
 Abstract
 Introduction
 Utilization and Yield of...
 Methods to Improve Blood...
 Contaminant Blood Cultures:...
Predictors of Bacteremia
 Conclusion
 References
 
The low yield of blood cultures and the large proportion of false-positive results may lead to extensive diagnostic testing, excessive antibiotic use, and prolonged hospitalization. Studies in unselected hospitalized patients have shown that only 7.5% to 12.4% of all results of blood culture sets are positive; of these, up to 4.9% are considered contaminants.20 21 52

Bates et al8 prospectively studied blood culture results obtained in a single academic center during a 6-month period. The primary end point was resource utilization starting 2 days after blood culture samples were drawn. Lengths of stay, as well as pharmacy, IV antibiotic, microbiology, and total charges, were evaluated. A total of 1,516 blood culture episodes, defined as the 48-h period following initial blood culture, were identified. Of the 219 positive culture results, 104 cultures were considered contaminants, based on clinical judgement. Blood cultures from samples drawn through intravascular catheters were contaminated 6% of the time, compared to 3% drawn from peripheral venipunctures. A single blood culture set was drawn in 22% of episodes; in 55% of culture samples drawn from intravascular catheters, paired peripheral samples to confirm the presence of true bacteremia were not obtained. Contaminated blood cultures were also more likely to occur in uncooperative patients with altered mental status, in whom it was often difficult to achieve and maintain sterile technique during venipuncture.

In the same study, after controlling for severity of illness and diagnosis, contaminated blood cultures increased unnecessary IV antibiotic charges by 39%. By escalating the number of repeat blood cultures, contaminated cultures also led to 80% higher microbiology laboratory charges. There was a trend toward increased length of stay, but this was not statistically significant. Although this study has several limitations and results may not be readily generalized, the authors highlight the fact that contaminated blood cultures comprise a significant proportion of all positive blood culture results, often generating costs that far exceed the costs of the test itself.8

There is a paucity of data on the incidence and costs of contaminants in the critical care setting. Schwenzer and colleagues17 studied blood cultures from a surgical ICU over a 2-year period; of 1,411 blood cultures, there were 122 positive culture results, and 51 of these were considered contaminants. Coagulase-negative staphylococci were considered true pathogens only if isolated from both a central line and peripheral venipuncture, from two peripheral blood culture sets, or from patients who were clinically septic and had a single positive culture result from a central venous catheter. Most other skin organisms were considered contaminants. No cost data were provided in this study, and no good estimates exist in current literature. However, it is likely that the costs and consequences of contaminant blood cultures parallel or exceed estimates available for unselected hospital admissions. Differentiating contaminants from true pathogens is often difficult. In one of the earlier studies on this topic, MacGregor and Beaty19 determined that pneumococci, group A Streptococci, Enterobacteraceae, Bacteriodes, Candida, and Haemophilus spp were true pathogens in their series of patients. Diphtheroids, nonhemolytic streptococci, and Bacillus spp were usually contaminants. Polymicrobial infections occurred in injection drug users or in patients who had leukemia or were receiving chemotherapy. In this study, clinical criteria alone were used to determine whether a positive blood culture result represented true bacteremia or contamination. When two or more blood culture samples were drawn, 69% of patients with a clinical course suggestive of bacteremia had repeatedly positive culture results. This number may be deceptively low, since paired blood culture samples were not drawn, and subsequent culture samples were often drawn while patients were receiving antibiotics. In contrast, only 11% of patients classified as having a contaminated blood culture had positive blood culture results on repeated testing, and often different organisms were isolated. This study highlights the importance of drawing multiple (ie, two or more) blood culture samples as an aid to differentiating between true bacteremia and contaminated cultures.

While there are no standardized criteria, several general observations can help guide the clinician in determining the clinical significance of positive blood culture results (Table 3 ). False-positive culture results tend to require prolonged incubation periods before isolation of an organism.18 19 Contaminating organisms often do not grow from multiple consecutive blood culture sets. Contamination is usually due to skin flora such as diphtheroids and Bacillus species, whereas Enterobacteriaceae spp, Pseudomonas spp, S pneumoniae, S aureus, and Candida albicans are rarely contaminants. Coagulase-negative staphylococci present a unique problem; although the majority of isolates are contaminants, 6 to 15% are true pathogens. In addition, coagulase-negative staphylococci are the most common cause of catheter-related infections.53 The significance of coagulase-negative staphylococcus isolates should be judged on the basis of a patient’s clinical condition and laboratory data.18 19 39 54 The recovery of multiple organisms also suggests contamination rather than true bacteremia, except in immunosuppressed or neutropenic patients, and in critically ill patients with intravascular catheters or clinical findings suggesting intra-abdominal infection.55


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

  		Table 3. True Bacteremia vs Contaminated Blood Cultures

 
Central venous catheters present a particular challenge to the physician trying to distinguish between true bacteremia and contaminated blood cultures. On a yearly basis, an estimated 850,000 catheters become colonized, resulting in 50,000 cases of catheter-related bacteremia.56 The body of literature on the diagnosis and management of central catheter infections is rapidly growing. The multiplicity of methods used in diagnosing such infections and the variability of results are at times confusing. Identifying IV catheters as a source of bacteremia is often a diagnosis of exclusion, as usually there are no focal findings at the catheter insertion site. Positive blood culture results obtained from intravascular catheters may reflect colonization, contamination with skin flora, or true catheter-related bacteremia. In an attempt to avoid unnecessary removal of catheters, quantitative blood cultures have been used to diagnose catheter-related bacteremia. Paired samples of blood are obtained from the catheter hub and from a peripheral venipuncture site for quantitative culture. A fivefold to 10-fold increase in the colony count obtained from the catheter sample compared to the peripheral blood specimen is considered indicative of catheter-related bacteremia.53 57 Quantitative blood cultures are not readily available in all hospitals, however, and their use is not without controversy.58

The continuous-monitoring blood culture systems currently available (eg, BacT/Alert) have thus been used instead to determine the differential time to positivity of blood culture samples obtained from central venous lines compared to peripheral vein samples. Blot et al59 noted that earlier positive culture results from central catheters were predictive of catheter-related infections. In their population of oncology patients with long-term (> 30 days) and often tunneled catheters, this method was highly sensitive (96.4%) and specific (100%) when the differential was > 120 min.59 60 Rijnders and colleagues61 prospectively evaluated this technique in 100 critically ill patients with short-term intravascular catheters. These investigators, however, were unable to demonstrate a significant difference in differential time to positivity between patients with catheter-related bacteremia and patients with other causes of bacteremia.61 Due to the small number of central catheter infections in this study, strong conclusions cannot be made. The validity of this method for noncancer patients and in the setting of multiple short-term central catheters requires further study prior to recommendations for its routine use as an alternative to quantitative blood cultures.

In general, to distinguish between contaminated cultures and true episodes of bacteremia, clinicians must consider any positive blood culture result in its proper clinical context. A positive blood culture result is more likely a true-positive result when the pretest probability of bacteremia is high, as in a patient with predisposing risk factors and a clinical course suggestive of sepsis.


   Predictors of Bacteremia
TOP
 Abstract
 Introduction
 Utilization and Yield of...
 Methods to Improve Blood...
 Contaminant Blood Cultures:...
 Predictors of Bacteremia
Conclusion
 References
 
Theoretically, it may be possible to decrease the number of false-positive culture results by diminishing the number of blood culture samples drawn in patients who are at low risk for bacteremia. Prediction rules to determine the pretest probability of bacteremia in patients admitted to the hospital have been developed.11 12 14 15 17 62 63 64 65 These rules have been limited in value, however. They are often retrospective in nature, analyze a subset of patients, lack external validity, and are not readily generalized to the critically ill population.

In an effort to determine the probability of bacteremia in hospitalized patients, Bates et al11 performed a prospective observational study of 1,007 blood culture episodes (defined as blood culture samples drawn in a 48-h time period) in a tertiary care setting. Factors correlating with true bacteremia included major comorbidity, body temperature >= 38.3°C, shaking chills, a clinical examination significant for an acute abdomen, or IV drug use. True bacteremia was also associated with the presence of an underlying disease considered to be "rapidly" fatal in 1 month (eg, acute untreated leukemia) or "ultimately" fatal within 5 years (eg, metastatic colon cancer). Coma, multiple trauma or burns, cardiopulmonary arrest within 24 h of the study period, ARDS, liver failure, bowel perforation, and severe pancreatitis were considered major comorbidities. Points were assigned for each identified risk factor. Patients with three or more risk factors or with 10 points were determined to be at high risk for bacteremia and had a 14% probability of true bacteremia. This is in contrast to the low-risk group with no identifiable risk factors, who had a 2% likelihood of true bacteremia on blood culture results.11 Unfortunately, this prediction model cannot be generalized to the critically ill population of patients, most of whom already fall into the high-risk category defined by this study.

Traditionally, the presence of fever in the ICU patient has been the impetus for extensive imaging and for drawing frequent blood culture samples to rule out bacteremia. Although infectious causes of fever lead to considerable morbidity and mortality and warrant further investigation, noninfectious etiologies of fever should also be considered (Table 4 ). When considered in the context of other clinical and laboratory clues, the presence of fever may increase the suspicion of bacteremia. However, fever is not a reliable predictor of bacteremia in the ICU population. Ironically, a proportion of bacteremic patients are afebrile, making the task of predicting sepsis and bacteremia in the ICU a difficult one. In one of the few studies looking at predictors of bacteremia in the critically ill, Schwenzer et al17 retrospectively compared 24 surgical ICU patients who had one or more positive blood culture results to 48 control patients with negative culture results, matching for admitting diagnosis, procedures, severity of illness on admission, and age. The decision to obtain blood culture samples was left to the discretion of the attending physician, but often was determined clinically in the presence of body temperature of >= 38.5°C, hypothermia, hypotension, leukocytosis, and changes in mental status. Clinical data were reviewed 5 days before and after initial blood culture samples were drawn. There was a 5% true-positive rate. Patients with bacteremia were three times more likely to die than matched control subjects. This study, however, was unable to identify any clinical predictors of bacteremia in this population of surgical ICU patients. In particular, there was no correlation between the presence of fever, hypotension, and leukocytosis (factors that often raise the suspicion of nosocomial infection in the ICU) and bacteremia.


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

  		Table 4. Noninfectious Etiology of Fever in the ICU

 
Even less is known about clinical predictors of bacteremia in medical ICU patients, many of whom have several comorbidities and are immunocompromised. Studies in this area are made difficult by the heterogeneous nature of ICU patients, the lack of agreement on diagnosis of certain conditions (eg, ventilator-associated pneumonias), and the presence of clinical states that mimic sepsis. Regardless, the considerable impact of nosocomial infections and bacteremia necessitates the early diagnosis and treatment of these infections.


   Conclusion
TOP
 Abstract
 Introduction
 Utilization and Yield of...
 Methods to Improve Blood...
 Contaminant Blood Cultures:...
 Predictors of Bacteremia
 Conclusion
References
 
Sepsis is an important cause of morbidity and mortality in the critical care unit. Bacteremia in critically ill patients has an average attributable mortality of 26%,2 but clinical parameters are often not reliable predictors of bacteremia. As the standard for diagnosing bacteremia, blood cultures have low yield. The volume of blood drawn in adult patients is the single most important factor improving blood culture sensitivity, and at least 10 mL, but optimally 20 mL, of blood should be drawn. Two sets of culture samples drawn from separate venipuncture sites will help the clinician distinguish contaminants from true-positive results. Blood culture samples drawn from intravascular devices should be avoided, or if necessary paired with a peripheral venipuncture. Although fever is not a specific predictor of bacteremia, its presence often triggers an elaborate diagnostic workup. Blood cultures remain a valuable diagnostic tool. Every effort should be made to improve the yield of this diagnostic modality, and results obtained should be interpreted in light of clinical and other laboratory data. Once bacteremia is identified, repeated cultures with each temperature elevation, especially in patients who are clinically unchanged, are unnecessary. The judicious use of cultures, while paying attention to factors that improve blood culture yield and decrease contamination rates, will improve the utility of blood cultures as diagnostic tools in critically ill patients.


   Footnotes
 
Abbreviation: SPS = sodium polyanetholsulfonate

Received for publication September 6, 2001. Accepted for publication April 23, 2002.


   References
TOP
 Abstract
 Introduction
 Utilization and Yield of...
 Methods to Improve Blood...
 Contaminant Blood Cultures:...
 Predictors of Bacteremia
 Conclusion
 References
 

  1. Brun-Buisson, C, Doyon, F, Carlet, J, et al (1995) Incidence, risk factors, and outcome of severe sepsis and septic shock in adults: a multicenter prospective study in intensive care units; French ICU Group for Severe Sepsis. JAMA 274,968-974[Abstract]
  2. Pittet, D, Tarara, D, Wenzel, RP Nosocomial bloodstream infection in critically ill patients: excess length of stay, extra costs, and attributable mortality. JAMA 1994;271,1598-1601[Abstract]
  3. Wey, SB, Mori, M, Pfaller, MA, et al Hospital-acquired candidemia: the attributable mortality and excess length of stay. Arch Intern Med 1988;148,2642-2645[Abstract]
  4. Landry, SL, Kaiser, DL, Wenzel, RP Hospital stay and mortality attributed to nosocomial enterococcal bacteremia: a controlled study. Am J Infect Control 1989;17,323-329[CrossRef][ISI][Medline]
  5. Edgeworth, JD, Treacher, DF, Eykyn, SJ A 25-year study of nosocomial bacteremia in an adult intensive care unit. Crit Care Med 1999;27,1421-1428[CrossRef][ISI][Medline]
  6. Bergmans, DC, Bonten, MJ, Gaillard, CA, et al Indications for antibiotic use in ICU patients: a one-year prospective surveillance. J Antimicrob Chemother 1997;39,527-535[Abstract/Free Full Text]
  7. Rello, J, Ricart, M, Mirelis, B, et al Nosocomial bacteremia in a medical-surgical intensive care unit: epidemiologic characteristics and factors influencing mortality in 111 episodes. Intensive Care Med 1994;20,94-98[CrossRef][ISI][Medline]
  8. Bates, DW, Goldman, L, Lee, TH Contaminant blood cultures and resource utilization: the true consequences of false-positive results. JAMA 1991;265,365-369[Abstract]
  9. Bates, DW, Lee, TH Rapid classification of positive blood cultures: prospective validation of a multivariate algorithm. JAMA 1992;267,1962-1966[Abstract]
  10. Ram, S, Mylotte, JM, Pisano, M Rapid classification of positive blood cultures: validation and modification of a prediction model. J Gen Intern Med 1995;10,82-88[ISI][Medline]
  11. Bates, DW, Cook, EF, Goldman, L, et al Predicting bacteremia in hospitalized patients: a prospectively validated model. Ann Intern Med 1990;113,495-500[ISI][Medline]
  12. Bates, DW, Sands, K, Miller, E, et al Predicting bacteremia in patients with sepsis syndrome: Academic Medical Center Consortium Sepsis Project Working Group. J Infect Dis 1997;176,1538-1551[ISI][Medline]
  13. Bensousan, TA, Vincent, F, Damaj, G, et al Predicting infection in critically ill surgical patients: usefulness of bacteriuria. Intensive Care Med 1995;21,621-622[Medline]
  14. Bossink, AW, Groeneveld, AB, Thijs, LG Prediction of microbial infection and mortality in medical patients with fever: plasma procalcitonin, neutrophilic elastase-{alpha}1-antitrypsin, and lactoferrin compared with clinical variables. Clin Infect Dis 1999;29,398-407[ISI][Medline]
  15. Leibovici, L Predicting bacteremia [letter]. Ann Intern Med 1991;114,703[Medline]
  16. Leibovici, L, Greenshtain, S, Cohen, O, et al Bacteremia in febrile patients: a clinical model for diagnosis. Arch Intern Med 1991;151,1801-1806[Abstract]
  17. Schwenzer, KJ, Gist, A, Durbin, CG Can bacteremia be predicted in surgical intensive care unit patients? Intensive Care Med 1994;20,425-430[CrossRef][ISI][Medline]
  18. Aronson, MD, Bor, DH Blood cultures. Ann Intern Med 1987;106,246-253[ISI][Medline]
  19. MacGregor, RR, Beaty, HN Evaluation of positive blood cultures: guidelines for early differentiation of contaminated from valid positive cultures. Arch Intern Med 1972;130,84-87[CrossRef][ISI][Medline]
  20. Weinstein, MP, Reller, LB, Murphy, JR, et al The clinical significance of positive blood cultures: a comprehensive analysis of 500 episodes of bacteremia and fungemia in adults; I. Laboratory and epidemiologic observations. Rev Infect Dis 1983;5,35-53[ISI][Medline]
  21. Darby, JM, Linden, P, Pasculle, W, et al Utilization and diagnostic yield of blood cultures in a surgical intensive care unit. Crit Care Med 1997;25,989-994[CrossRef][ISI][Medline]
  22. Gross, PA, Van Antwerpen, CL, Hess, WA, et al Use and abuse of blood cultures: program to limit use. Am J Infect Control 1988;16,114-117[CrossRef][ISI][Medline]
  23. Pittet, D, Thievent, B, Wenzel, RP, et al Bedside prediction of mortality from bacteremic sepsis: a dynamic analysis of ICU patients. Am J Respir Crit Care Med 1996;153,684-693[Abstract]
  24. Craven, DE, Kunches, LM, Lichtenberg, DA, et al Nosocomial infection and fatality in medical and surgical intensive care unit patients. Arch Intern Med 1988;148,1161-1168[Abstract]
  25. Daschner, FD, Frey, P, Wolff, G, et al Nosocomial infections in intensive care wards: a multicenter prospective study. Intensive Care Med 1982;8,5-9[CrossRef][ISI][Medline]
  26. Washington, JA, II, Ilstrup, DM Blood cultures: issues and controversies. Rev Infect Dis 1986;8,792-802[ISI][Medline]
  27. Washington, JA Collection, transport, and processing of blood cultures. Clin Lab Med 1994;14,59-68[ISI][Medline]
  28. Washington, JA Blood cultures: an overview. Eur J Clin Microbiol Infect Dis 1989;8,803-806[CrossRef][ISI][Medline]
  29. Chandrasekar, PH, Brown, WJ Clinical issues of blood cultures. Arch Intern Med 1994;154,841-849[CrossRef][ISI][Medline]
  30. Schifman, RB, Pindur, A The effect of skin disinfection materials on reducing blood culture contamination. Am J Clin Pathol 1993;99,536-538[ISI][Medline]
  31. Stratton, CW Blood cultures and immunocompromised patients. Clin Lab Med 1994;14,31-49[ISI][Medline]
  32. Little, JR, Murray, PR, Traynor, PS, et al A randomized trial of povidone-iodine compared with iodine tincture for venipuncture site disinfection: effects on rates of blood culture contamination. Am J Med 1999;107,119-125[CrossRef][ISI][Medline]
  33. Mimoz, O, Karim, A, Mercat, A, et al Chlorhexidine compared with povidone-iodine as skin preparation before blood culture: a randomized, controlled trial. Ann Intern Med 1999;131,834-837[Abstract/Free Full Text]
  34. Li, J, Plorde, JJ, Carlson, LG Effects of volume and periodicity on blood cultures. J Clin Microbiol 1994;32,2829-2831[Abstract/Free Full Text]
  35. Mylotte, JM, Tayara, A Blood cultures: clinical aspects and controversies. Eur J Clin Microbiol Infect Dis 2000;19,157-163[CrossRef][ISI][Medline]
  36. Ilstrup, DM, Washington, JA, II The importance of volume of blood cultured in the detection of bacteremia and fungemia. Diagn Microbiol Infect Dis 1983;1,107-110[CrossRef][ISI][Medline]
  37. Henry, NK, McLimans, CA, Wright, AJ, et al Microbiological and clinical evaluation of the isolator lysis-centrifugation blood culture tube. J Clin Microbiol 1983;17,864-869[Abstract/Free Full Text]
  38. Mermel, LA, Maki, DG Detection of bacteremia in adults: consequences of culturing an inadequate volume of blood. Ann Intern Med 1993;119,270-272[Abstract/Free Full Text]
  39. Weinstein, MP Current blood culture methods and systems: clinical concepts, technology, and interpretation of results. Clin Infect Dis 1996;23,40-46[ISI][Medline]
  40. Traub, WH, Kleber, I Inactivation of classical and alternative pathway-activated bactericidal activity of human serum by sodium polyanetholsulfonate. J Clin Microbiol 1977;5,278-284[Abstract/Free Full Text]
  41. Wilson, ML, Weinstein, MP General principles in the laboratory detection of bacteremia and fungemia. Clin Lab Med 1994;14,69-82[ISI][Medline]
  42. Eng, J Effect of sodium polyanethol sulfonate in blood cultures. J Clin Microbiol 1975;1,119-123[Abstract/Free Full Text]
  43. McKenzie, R, Reimer, LG Effect of antimicrobials on blood cultures in endocarditis. Diagn Microbiol Infect Dis 1987;8,165-172[CrossRef][ISI][Medline]
  44. Grace, CJ, Lieberman, J, Pierce, K, et al Usefulness of blood culture for hospitalized patients who are receiving antibiotic therapy. Clin Infect Dis 2001;32,1651-1655[CrossRef][ISI][Medline]
  45. Peterson, LR, Shanholtzer, CJ, Mohn, ML, et al Improved recovery of microorganisms from patients receiving antibiotics with the antimicrobial removal device. Am J Clin Pathol 1983;80,692-696[ISI][Medline]
  46. Doern, GV Manual blood culture systems and the antimicrobial removal device. Clin Lab Med 1994;14,133-147[ISI][Medline]
  47. Appleman, MD, Swinney, RS, Heseltine, PN Evaluation of the Antibiotic Removal Device. J Clin Microbiol 1982;15,278-281[Abstract/Free Full Text]
  48. Oud, L, Krimerman, S, Salam, N, et al Role of blood culture systems in the evaluation of epidemiological features of coagulase-negative staphylococcal bloodstream infection in critically ill patients. Eur J Clin Microbiol Infect Dis 1999;18,899-901[CrossRef][ISI][Medline]
  49. Pohlman, JK, Kirkley, BA, Easley, KA, et al Controlled clinical evaluation of BACTEC Plus Aerobic/F and BacT/Alert Aerobic FAN bottles for detection of bloodstream infections. J Clin Microbiol 1995;33,2856-2858[Abstract]
  50. Pohlman, JK, Kirkley, BA, Easley, KA, et al Controlled clinical comparison of Isolator and BACTEC 9240 Aerobic/F resin bottle for detection of bloodstream infections. J Clin Microbiol 1995;33,2525-2529[Abstract]
  51. Wilson, ML General principles of specimen collection and transport. Clin Infect Dis 1996;22,766-777[ISI][Medline]
  52. Roberts, FJ, Geere, IW, Coldman, A A three-year study of positive blood cultures, with emphasis on prognosis. Rev Infect Dis 1991;13,34-46[ISI][Medline]
  53. Mermel, LA, Farr, BM, Sherertz, RJ, et al Guidelines for the management of intravascular catheter-related infections. Clin Infect Dis 2001;32,1249-1272[CrossRef][Medline]
  54. Souvenir, D, Anderson, DE, Jr, Palpant, S, et al Blood cultures positive for coagulase-negative staphylococci: antisepsis, pseudobacteremia, and therapy of patients. J Clin Microbiol 1998;36,1923-1926[Abstract/Free Full Text]
  55. Rello, J, Quintana, E, Mirelis, B, et al Polymicrobial bacteremia in critically ill patients. Intensive Care Med 1993;19,22-25[CrossRef][ISI][Medline]
  56. Collin, GR Decreasing catheter colonization through the use of an antiseptic-impregnated catheter: a continuous quality improvement project. Chest 1999;115,1632-1640[Medline]
  57. Reimer, LG Catheter-related infections and blood cultures. Clin Lab Med 1994;14,51-58[ISI][Medline]
  58. Reimer, LG, Wilson, ML, Weinstein, MP Update on detection of bacteremia and fungemia. Clin Microbiol Rev 1997;10,444-465[Abstract]
  59. Blot, F, Nitenberg, G, Chachaty, E, et al Diagnosis of catheter-related bacteraemia: a prospective comparison of the time to positivity of hub-blood vs peripheral-blood cultures. Lancet 1999;354,1071-1077[CrossRef][ISI][Medline]
  60. Blot, F, Schmidt, E, Nitenberg, G, et al Earlier positivity of central-venous- vs peripheral-blood cultures is highly predictive of catheter-related sepsis. J Clin Microbiol 1998;36,105-109[Abstract/Free Full Text]
  61. Rijnders, BJ, Verwaest, C, Peetermans, WE, et al Difference in time to positivity of hub-blood vs nonhub-blood cultures is not useful for the diagnosis of catheter-related bloodstream infection in critically ill patients. Crit Care Med 2001;29,1399-1403[CrossRef][ISI][Medline]
  62. Peduzzi, P, Shatney, C, Sheagren, J, et al Predictors of bacteremia and gram-negative bacteremia in patients with sepsis: The Veterans Affairs Systemic Sepsis Cooperative Study Group. Arch Intern Med 1992;152,529-535[Abstract]
  63. Niederman, MS, Fein, AM Predicting bacteremia in critically ill patients: a clinically relevant effort? Intensive Care Med 1994;20,405-406[CrossRef][ISI][Medline]
  64. Mozes, B, Milatiner, D, Block, C, et al Inconsistency of a model aimed at predicting bacteremia in hospitalized patients. J Clin Epidemiol 1993;46,1035-1040[CrossRef][ISI][Medline]
  65. Marcus, EL, Clarfield, AM Predicting mortality in patients with bacteremia. J Am Geriatr Soc 1998;46,1327-1328[Medline]



This article has been cited by other articles:


Home page
 ChestHome page
R. Berger, A. Weinacker, and S. Shafazand
Modification of Blood Culture Recommendations?
Chest, June 1, 2003; 123(6): 2157 - 2158.
[Full Text] [PDF]

Home page
 ChestHome page
R. P. Byrd Jr, T. M. Roy, A. Weinacker, and S. Shafazand
Anaerobic Blood Cultures: Useful in the ICU?
Chest, June 1, 2003; 123(6): 2158 - 2159.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF) Free
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Article Archive
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (10)
Right arrow