Bacteremia often misdirects, but can point to infection source
Click Here to Manage Email Alerts
Each day, blood and cerebrospinal fluid cultures are considered the most significant tests that are requested from the clinical microbiology laboratory by essentially all hospitals. Objectively speaking, the term “bacteremia” simply refers to the detection or presence of a bacterial organism in the bloodstream. No more or no less. Analogous to a chemical assay that identifies elevated potassium (hyperkalemia) or diminished sodium (hyponatremia) levels, the detection of bacteremia in and of itself does not necessarily nor specifically point to any one precise infectious process any more than the other so-called “emias” define the explicit entity that is responsible for an aberrant amount of these ions, which are vital for the functioning of all living cells. In effect, the recognition of their presence merely calls attention to any one of several potentially dangerous pathophysiologic disorders disrupting normal electrolyte homeostasis or, in the case of bacteremia, to the fact that an endogenous or exogenous microbiologic organism has, for an array of possible reasons, invaded the routinely sterile blood. In fact, bacteremia is one of only two microbiology test results (positive cerebrospinal fluid culture being the other) that is included on the list of laboratory-critical values put forth by both the College of American Pathologists and the American Society for Clinical Pathology and adopted and employed by almost all certified laboratories.
The moment bacteremia is suspected to be present, hospital policy mandates that laboratory personnel immediately inform the patient’s clinician and document this communication in the permanent medical record, even before the organism is identified.
Blood cultures
Rightfully or wrongfully so, the discovery and reporting of bacteria in the blood tends to provoke a sense of emergency, often eliciting urgent reactions and intervention orders from health care providers. Furthermore, nearly always and without explanation, the knowledge of bacteremia induces a frightened and somewhat visceral response from patients and their family members. Yet, even though bacteremia is commonly and most times incorrectly used synonymously with the somewhat colloquial term “septicemia,” this laboratory test is neither required nor included in the sequential (sepsis-related) organ failure assessment (SOFA) clinical criteria score outlined in the newly revised 2016 Third International Consensus Definitions for Sepsis and Septic Shock. In fact, as pointed out by the members of the Society of Critical Care Medicine, culture-positive “sepsis” from any site is historically observed in only 30% to 40% of cases. Likewise, in a large series of nonselected patients hospitalized with community-acquired pneumonia (CAP), which is the leading cause of sepsis, blood cultures yielded positive results for a probable pathogen in only 5% to 14% of patients who were tested, and positive cultures did not result in better clinical outcomes or improvements in antibiotic selection. The relatively futile and at times misleading results associated with incorporating blood cultures in the diagnostic workup in hospitalized patients with CAP led to the guideline recommendations that the tests should be reserved for individuals with severe CAP or underlying immune system defects known to diminish the clearance of bacteria in the blood (eg, complement deficiencies, asplenia).
That being said, the prevalence of physician blood culture requests for hospitalized patients persists unabated. Requests such as “blood culture x 2 for temperature >38.4°C” are often entered as standing orders. In one 234-bed, presumably representative urban acute-care hospital, the average number of monthly blood culture sets (comprising an aerobic and anaerobic bottle) is estimated to be about 1,000, of which approximately 60 (6%) yield an organism, and approximately 20 to 30 sets (2% to 3%) are contaminated (about one-third to one-half of all isolates). Unfortunately, on many occasions, clinicians inappropriately treat blood culture contaminants with unwarranted antibiotics as if they had clinical relevance. This misguided practice results in prolonged hospital stays, escalated expenditures and unintended complications, not to mention a misdirected diagnosis. On the contrary, the determination that a blood culture isolate is a contaminant falls under the purview and is essentially dictated by laboratory policy (ie, common skin flora bacterial species in one culture bottle), precluding further identification and antimicrobial susceptibility testing. Although the culture report includes instructions to the clinician to notify the laboratory if further workup is desired, clinical information is rarely provided, increasing the risk that certain low-grade indolent infections (particularly those involving foreign body implants) may go undiagnosed and untreated.
Other than those instances when outpatients are being evaluated for possible infective endocarditis or perhaps fever of undetermined origin, most blood cultures are drawn in hospitalized patients. Theoretically, these cultures are obtained when a patient demonstrates signs and/or symptoms or has documented laboratory and/or imaging data that raise the clinical suspicion of infection. However, in reality, we know that the documentation of fever alone serves as the stimulus for ordering cultures in an overwhelmingly large percentage of cases. Interestingly, the requirement that at least two of four criteria be met — one being a temperature greater than 38°C (100.4°F) — to qualify as manifesting the systemic inflammatory response syndrome (SIRS) and therefore possible sepsis has been appropriately questioned. In fact, it has been eliminated from the new consensus definition of sepsis owing to its relative insensitivity in reliably predicting an infection, let alone sepsis.
Regrettably, to avoid running afoul of CMS mandates, almost all hospital sepsis protocols continue to abide by the SIRS criteria when formulating their sepsis response, which when enacted, demands a set of urgent actions, ignoring the fact that infection is just one of a dozen or more conditions (most unrelated to infection) that would satisfy the SIRS criteria. Accordingly, based on a checklist that typically includes blood cultures, nonphysician personnel may initiate a “code sepsis.”
Undoubtedly, during “curbside consults” by physician colleagues we all have been frequently asked for our input on “how to treat” Escherichia coli or perhaps enterococcus, for example, reported to be growing in their patient’s blood culture. Frankly, the correct response should be to question the details of the clinical scenario, aside from the presence of fever, that served as the impetus for collecting these cultures in the first place, all the while explaining that we do not treat “bugs” but rather disease processes primarily or secondarily involved with bacteria. In other words, “What are you treating?”
In the United States, about 250,000 of the millions of patients from whom blood cultures are annually obtained are determined to have bacteremia or fungemia, which are associated with mortality rates of 16% to 40%. The vast majority of bacteremia cases occur in hospitalized patients, thereby posing the understandably difficult task of determining whether the individual happened to die with bacteremia or because of it in many cases. In other words, did the laboratory finding of bacteremia have any greater influence on the patient’s demise than other potentially severe “emia” manifestations of the primary disease process, such as acidemia, azotemia and lactatemia?
BSIs
Ostensibly, the term bloodstream infection (BSI) in and of itself is a misnomer, as the blood per se is rarely infected but rather serves as a virtual conduit for bacterial organisms that primarily or secondarily have been introduced into this normally sterile body fluid and in most instances remain present there for only a brief time. The CDC’s National Healthcare Safety Network (NHSN) categorizes all bacteremias as either primary or secondary. Simply stated, a BSI is considered to be primary when the laboratory confirms the presence of a bacterial organism that is not included on the NHSN’s list of common commensals or related to an infection at another body site. In one early study conducted to specifically examine the etiologies of bacteremia, nearly one-half (44.7%) of episodes were considered to be primary and almost equally divided between intravascular catheter or central line-associated bloodstream infections (CLABSIs) and those having no obvious portal of entry. For several reasons, it is important to make an effort to determine whether a bacteremic episode is primary or has occurred secondarily to a localized focus of infection.
CLABSIs
Knowing that CLABSIs account for a large proportion of primary BSIs with potential complications (eg, nosocomial endocarditis, vertebral osteomyelitis, etc.) and result in prolonged hospital stays and increased health care expenditures, these inadvertent infections now occupy a front seat in CMS’ Hospital Inpatient Quality Reporting Program of hospital-acquired infections (HAIs) and are considered a “zero-tolerance” event. Along with the punitive financial ramifications for hospitals that report CLABSIs, this information is now openly available for public scrutiny, opening the door to warranted, and at times unwarranted, criticism and concern. Even though there is a clear lack of specificity in the NHSN’s surveillance definition of a CLABSI, each individual hospital or health care system is charged with regularly reporting their rates of CLABSIs, lending the determination of what is a CLABSI to the influences of subjectivity and distinct variability. Nonetheless, these rates are of noteworthy importance in helping to assess the effectiveness of infection prevention policies and procedures, as well as in epidemiologic investigations of infection outbreaks. As pointed out in the Infectious Diseases Society of America/American Society for Microbiology’s recently published Guide to Utilization of the Microbiology Laboratory for Diagnosis of Infectious Diseases, no microbiologic gold standard for diagnosis of a CLABSI exists, therefore making it one of exclusion. Although definitive data regarding the value of various diagnostic methods for detecting CLABSIs are lacking, the three that appear to be the most widely accepted and recommended include:
l the determination of the differential time to culture growth from blood obtained via a central venous catheter vs. a culture from a peripheral site (>2 hours);
l paired quantitative blood cultures from both sites (fivefold more organisms); and
l catheter tip or segment cultures employing a semi-quantitative method, assuming this is accompanied by a positive blood culture.
However, the authors of the guidelines admittedly acknowledged that none of these diagnostic methods are routinely performed in most clinical laboratories.
In recognition of the fact that the specificity and reliability of the NHSN BSI surveillance definition somewhat suffered from objectivity and needed to be better aligned with clinical judgment, the CDC developed a modification of this definition in 2013 to include the term “mucosal barrier injury laboratory-confirmed bloodstream infection,” intent on identifying a subset of BSIs that are reported as CLABSIs but are more likely related to the integrity of the mucosal barrier and not the central line.
To a certain degree, once a CLABSI diagnosis is suspected or confirmed, the management is straightforward. In most cases, this involves the removal of the central venous catheter and the administration of appropriate and adequate antimicrobial therapy which, according to IDSA guideline recommendations, varies dependent upon the type of catheter and the pathogen involved.
Clues from bacteremia
Granted that most infections are not accompanied by bacteremia, the BSIs that are deemed to be secondary presumptively arise from either a local or deep tissue focus. In these cases, it is the degree of tissue injury of normally colonized tissues or organs (eg, gastrointestinal tract, skin) or those that are normally sterile but for a variety of reasons now harbor potential bacterial pathogens (eg, lung, urinary tract, cardiac valves) that serve as the determining pathophysiologic factor for whether or not bacteria will be introduced into the bloodstream via either direct translocation or lymphatic spread.
Again, irrespective of bacteremia being classified as transient (eg, related to a dental procedure), intermittent (eg, undrained closed-space infection) or persistent (eg, endocarditis, mycotic aneurysm, suppurative thrombophlebitis), the bacteremia is not the infectious process but rather a manifestation of a significant infectious process someplace else in the body — no different than hyperkalemia is a sign of tumor lysis syndrome. It would seem more plausible to assign the development of organ dysfunction(s) as a result of a dysregulated inflammatory host response to the primary infectious process site (eg, necrotizing fasciitis), rather than to the accompanying secondary bacteremia.
In many ways, the identification of one or more bacterial organisms in the blood greatly helps to determine the likely primary infectious process (eg, E. coli suggesting either a gastrointestinal, urinary or biliary focus), or serves to forestall additional invasive diagnostic interventions, even though the primary focus is clinically and radiographically evident (eg, Staphylococcus aureus and vertebral osteomyelitis). In addition, finding certain specific bacterial pathogens in the blood may raise a clinical suspicion of a process besides the obvious infection, such as the relationship between Streptococcus bovis Biotype I (Streptococcus gallolyticus) and colon cancer. On the other hand, the identification of a bacteremic isolate may lead to the administration of a narrow-spectrum antibiotic specifically chosen to “cover” the single organism, all the while ignoring the fact that the disease process, having led to the secondary bacteremia, is always polymicrobial (eg, Bacteroides fragilis and perforated diverticulitis).
Other issues in bacteremia
Other aspects of bacteremia that deserve clarification or revision include, but are not limited to, the following: number, timing and volume of cultures; selection of culture medium and duration of incubation; and duration of antimicrobial therapy and follow-up cultures.
Number, timing and volume of cultures
The success of detecting bacteremia is dependent upon and influenced by many factors. Without much clinical thought, orders for blood cultures are routinely entered in sets of two to be collected from two separate venous puncture sites 15 minutes apart, even though the optimal number and timing of sets should be based on the pretest probability of infection and each individual clinical scenario. In fact, many hospital computer physician order entry systems offer this option only. Although the yield of pathogen detection increases from 90% with two blood cultures to more than 99% with four blood cultures taken over a 24-hour period, the importance of immediate antibiotic therapy should take precedence. Although it is rarely monitored closely, the recommended volume of blood for optimally recovering an organism is 20 mL per set.
Selection of culture medium and duration of incubation
Manufacturers of modern automated blood culture instruments used in almost all clinical microbiology laboratories lay claim to a near-flawless sensitivity in capturing a pathogen if indeed it is present in the blood sample. Consequently, cultures are deemed negative and discarded after 3 to 5 days of incubation. However, most clinicians are unaware that, for fastidious and other infrequently encountered bacterial species, as well as mycobacteria and certain fungi, ensuring a diagnosis may require special transport and culture media in addition to prolonged periods of incubation. Unfortunately, the laboratory often is not kept abreast of the necessary information.
Duration of antimicrobial therapy and follow-up cultures
Because of the limited quantity of published studies on the optimal duration of antimicrobial therapy for infections accompanied by bacteremia, the concept that this subset of infections requires both longer courses and intravenously delivered antibiotics is pervasive in the general medical community. However, in everyday clinical practice, the bulk of patients with infections are treated on an outpatient basis, many of whom may have undocumented associated bacteremia but receive the same durations of treatment. The results of recent comparative studies involving gram-negative bacilli (predominantly Enterobacteriaceae) show that shorter courses of antibiotic therapy are as effective as longer ones. Moreover, the commonly adopted maxim that in all cases of bacteremia, it is imperative to use repeat blood cultures to document bacterial eradication has at least in one contemporary randomized study roven to be unjustified for gram-negative bacteremia.
As we have previously stressed, when it comes to bacteremia, as well as almost all things “infectious,” each and every stakeholder would be best served and surely benefit from the involvement of the infectious diseases specialist. Again, to borrow a quote, “If not us, who? If not now, when?”
- References:
- Bearman GM, et al. Arch Med Res. 2005;doi:10.1016/arcmed.2005.02.005.
- Canzoneri CN, et al. Clin Infect Dis. 2017;doi:10.1093/cid/cix648.
- Chotiprasitsakul D, et al. Clin Infect Dis. 2018;doi:10.1093/cid/cix767.
- Epstein L, et al. Infect Control Hosp Epidemiol. 2016;doi:10.1017/ice.2015.245.
- Freeman JT, et al. Infect Control Hosp Epidemiol. 2016;doi:10.1086/668431.
- Genzen JR, et al. Am J Clin Path. 2011;doi:10.1309/AJCP9IZT7BMBCJRS.
- Lee A, et al. J Clin Microbiol. 2007;doi:10.1128/JCM.01555-07.
- Mandell LA et al. Clin Infect Dis. 2007;doi:10.1086/511159.
- Marra R, et al. Am J Infect Control. 2010;doi:10.1016/j.ajic.2009.11.012.
- Mermel LA, et al. Clin Infect Dis. 2009;doi:10.1086/599376.
- Miller JM, et al. Clin Infect Dis. 2018;doi:10.1093/cid/ciy381.
- Singer M, et al. JAMA. 2016;doi:10.1001/jama.2016.0287.
- Weinstein MP, et al. Clin Infect Dis. 1997;24:584-602.
- For more information:
- Larry M. Bush, MD, FACP, is an affiliated professor of biomedical sciences at the Charles E. Schmidt College of Medicine, Florida Atlantic University, and affiliated associate professor of medicine at the University of Miami Miller School of Medicine, JFK Medical Center, Palm Beach County, Florida.
- Donald Kaye, MD, MACP, is a professor of medicine at Drexel University College of Medicine, associate editor of the International Society for Infectious Diseases’ ProMED-mail, section editor of news for Clinical Infectious Diseases and an Infectious Disease News Editorial Board member.
Disclosures: Bush and Kaye report no relevant financial disclosures.