Clinical Syndromes of Device-Associated Infections

Epidemiology and Pathogenesis

With focused prevention efforts, rates of central line associated bloodstream infections (CLABSIs) in the US have decreased dramatically. The Centers for Disease Control and Prevention (CDC) National Healthcare Safety Network (NHSN) reporting system noted an approximately 50% decline in CLABSIs between 2008 and 2016. Still, device-associated infections comprise approximately 25% of all healthcare-associated infection (HAI) reported to NHSN, and CLABSIs represent the most common device-associated infection experienced by hospitalized children. ^, Critically ill children who develop CLABSIs have longer duration of intensive care unit (ICU) and hospital stay, and the crude attributable mortality has been estimated to be 19%. ^, In addition, using a propensity score matched case-control study, the attributable cost and length of stay between matched CLABSI cases and non-CLABSI controls was found to be $55,646 (2011 US dollars) and 19 days, respectively. Children with parenteral nutrition needs are also at high risk for CLABSI and, when diagnosed and admitted to an inpatient unit, have a prolonged length of stay (8 days), with $28,375 in mean cost (2015 US dollars). Children requiring extracorporeal membrane oxygenation (ECMO) are at especially high risk for CLABSI, with rates as high as 24.9 per 1000 ECMO days with risk factors including duration of bypass and open chests postoperatively if a cardiothoracic procedure has taken place.

The terminology describing catheter-related infections can be confusing. Catheter-related bloodstream infection (CRBSI) identifies the catheter as the source of infection. However, CLABSI is a term used by NHSN to facilitate surveillance (and therefore has been widely adopted as a surrogate for CRBSI in the literature describing its epidemiology) and may overestimate the true incidence of CRBSI. In addition, several different types of central venous catheters (CVCs) are used in children. A CVC generally is classified by its intended duration of use (short term vs. long term); its site and mode of insertion (peripheral, femoral, subclavian; tunneled vs. nontunneled); or other specific characteristics (antibiotic coated, multilumen). Lastly, multiple types of catheter-related infections can be identified. Infection can occur at the exit site, in the subcutaneous tunnel or pocket, or in the bloodstream. Pathogens can be introduced during insertion, manipulation, calibration, and flushing of monitoring devices (all modes of direct contamination influenced by hand hygiene and antiseptic practices), or through contaminated parenteral fluids, flush solutions, or skin antiseptics. Translocation of organisms from colonizing sites with resultant transient bacteremia also can lead to CLABSI ( Box 100.3 ).

Gram-positive organisms, including coagulase-negative staphylococci (CoNS), Staphylococcus aureus, and enterococci, are most commonly isolated in intravascular catheter-associated infections. ^, In fact, although incidence density rates of CLABSI have decreased overall in recent years, the incidence density rate of S. aureus CLABSIs among pediatric ICUs in the US has remained the same. Identification of gram-negative bacilli, including Enterobacter spp., P. aeruginosa, Klebsiella spp. , and Escherichia coli, has increased to approximately one-third of reported infections ( Box 100.4 ). Candida spp. are increasingly prevalent and make up 25% of nosocomial bloodstream infection (BSI) in the US. Among 29 hospitals participating in the Children’s Hospital Association Quality Transformation Network, the majority of CLABSIs in pediatric ICUs were caused by a single organism, most commonly gram-positive cocci. Of S. aureus isolates, methicillin-resistant S. aureus (MRSA) represented nearly 50%. Gram-negative bacilli were identified in one-third of cases and Candida spp. in 15%. Other relatively common organisms include Corynebacterium, Cutibacterium, and Bacillus spp., Serratia marcescens, Acinetobacter spp., Citrobacter spp., and nontuberculous mycobacteria. The frequency and distribution of organisms may be influenced by the child’s exposure to (or level of illness associated with) the ICU, particular medications, such as intravenous (IV) prostanoid agents, or underlying cancer or chemotherapy. ^,

Catheter-Related Factors

Risk for infection varies by catheter type. Rates are higher for CVCs than for peripheral venous or arterial catheters. Although some studies showed similar infection risks for tunneled and nontunneled catheters, ^, later randomized studies showed that tunneling of femoral catheters and internal jugular catheters is associated with lower infection risk ( Box 100.5 ). Peripherally inserted central catheters (PICCs) now are used more commonly because of their perceived lower risk for infection; however, two outpatient studies demonstrated a 10% PICC infection rate, ^, and analyses of critically ill patients have revealed that the risk for catheter infection is similar among patients with a PICC and those with a centrally inserted percutaneous catheter. A correlation between increasing number of lumens and incidence of infection has been noted in some studies but not others. ^, In addition, the presence of multiple venous catheters was found to be an independent risk factor associated with nosocomial BSIs.

The duration of catheter placement is directly related to the risk for infection, a concept that has been substantiated in multiple pediatric studies. ^, Biofilm formation, which increases with dwell time, clearly increases the opportunity for catheter colonization. In addition, each manipulation of a catheter, stopcock, or needleless device provides an opportunity for the introduction of organisms. ^{, ,} Outbreaks of bacteremia have resulted from contamination of pressure transducers. ^, Replacing the catheter over a guidewire does not decrease the incidence of infection and may even increase it. Finally, the effect of transparent versus gauze dressing use on infection risk remains an unresolved issue. A meta-analysis found an association between CLABSI and transparent dressing use; however, the source studies for the meta-analysis were of low quality.

Certain infusates are more likely than others to be associated with BSI. Use of parenteral nutrition fluids increases the risk for CLABSI in multiple pediatric studies. ^{, , ,} In addition, receipt of blood transfusion has been identified as an independent risk factor for CLABSI in pediatric ICU patients. ^, Intrinsically or extrinsically contaminated infusates also can cause CLABSI. Although intrinsic contamination of IV fluids is rarely reported in the US, episodes of extrinsic contamination of IV fluids and medications continue to occur.

Host Factors

Host factors influence the risk of CLABSI. Skin flora varies according to a person’s age, hygiene, underlying medical conditions, and medications. Topical antimicrobial agents may alter the nature and density of local flora, and chlorhexidine gluconate (CHG) is now the preferred agent for preinsertion decontamination and daily maintenance (see “Prevention” section later). Systemic antimicrobial agents alter stool and skin flora and increase the risk for infection by Candida spp. ^, Factors that lead to a break in the integrity of skin or mucous membranes can increase the risk for catheter infections. Thus patients with immature skin (premature infants), impaired integrity of the gut epithelium (patients with short-gut syndrome), ^, or burn patients may be at risk for bacteremia and subsequent colonization of an indwelling catheter, although the true contribution of these putative mechanisms (as opposed to confounding risks such as illness severity and increased need for intravascular access) remains unclear. Similarly, patients with a distant site of infection (e.g., an abscess) or who develop a spontaneous bacteremia are also at risk for developing a secondary CRBSI.

Children undergoing cancer chemotherapy are at risk for infection because of neutropenia; infection rates in these patients are increased 2- to 4-fold by the use of indwelling CVCs. In addition, mucositis may raise the risk for bacterial gastrointestinal wall translocation and bacteremia and may account for differences in the microbiology of BSIs that occur in neutropenic, compared with nonneutropenic, patients. In a prospective study of more than 24,000 episodes of nosocomial BSIs, neutropenic patients were at greater risk for infection due to Candida spp., enterococci, and viridans group streptococci. However, it remains unclear whether the increased infection rate and variable microbiology are attributable to mucositis or to the associated illness severity, immune suppression, and care environment of such patients. The CDC incorporated a category for CLABSI surveillance to capture this known as mucosal barrier injury−laboratory confirmed bloodstream infection (MBI-LCBI); these infections include primary BSI due to select organisms in patients with neutropenia or in patients with allogeneic hematopoietic stem cell transplantation with gastrointestinal graft-versus-host disease or diarrhea. ^,

Patients with human immunodeficiency virus (HIV) infection are at greater risk for a variety of infections; use of CVCs for nutritional support and for medications increases the risk for bacteremia in HIV-infected and other patients. ^, Patients with uremia are at higher risk for CLABSI, particularly those with hemodialysis catheters. In patients undergoing hemodialysis, the type of bloodstream access used influences the risk for infection; the risk is lowest with native arteriovenous fistulas, intermediate with artificial arteriovenous grafts, and highest with CVCs. ^,

Clinical factors related to the severity of illness also have been found to influence the risk for infection. In a cohort of critically ill pediatric patients, the need for transport out of the care unit and for a procedure performed within an ICU were both independently associated with a 3- to 4-fold increase in the risk for developing a nosocomial BSI. Need for ICU exposure, ICU placement of the central line, mechanical ventilation, medical cardiac disease, malignancy, genetic syndrome, metabolic disease, presence of a gastrostomy tube, and history of prior CVC also have been associated with CLABSI in children. ^{, , ,}

Clinical Manifestations and Laboratory Diagnosis

Fever usually is present in patients with catheter-related infections. Patients with localized infection of the exit site, pocket, or tunnel frequently have infection in the absence of fever. When a catheter-related infection is suspected, the exit site should be examined for the presence of local suppuration or cellulitis; however, normal appearance of the exit site does not exclude a local or systemic infection involving the catheter. Any catheter malfunction (including decrease in flow or unidirectional flow) should prompt an evaluation for infection. Findings suggestive of disseminated infection include the presence of emboli in the retina, skin, bone, and viscera (lungs, kidneys, liver, and spleen) and organ dysfunction due to immune complex deposition (e.g., nephritis).

Infections attributable to a CVC include exit, tunnel, and pocket infections as well as BSI ( Table 100.1 ). ⁹²¹⁰³ Establishing that the catheter is the source of infection is not always straightforward. For example, a BSI in a patient with an indwelling catheter can originate from an undocumented source of infection (e.g., a postoperative wound infection or a urinary tract infection [UTI]) rather than from the catheter. However, pathogens commonly associated with CRBSI such as staphylococci should keep it high on the differential diagnosis if no alternative infection source is identified. ^{, ,} Of note, adjunctive laboratory evaluation such as C-reactive protein, procalcitonin, and white blood cell count may be part of an infectious work-up but are not specific in predicting CRBSIs. ^,

Several methods have been used to identify the catheter as the source of infection, although no microbiologic gold standard for diagnosis of catheter-related BSIs exists. Proposed methods include quantitative cultures of blood obtained through the catheter and a peripheral vein simultaneously; quantitative and semiquantitative cultures of a catheter segment; and differential time to blood culture positivity ( Table 100.1 ). ^, Until recently, the most commonly accepted methods of diagnosing a CRBSI have involved either quantitative or semiquantitative cultures of the catheter tip, which is falling out of favor due to poor predictive value.

Differential time to positivity of paired blood cultures is the simplest of the three methods and does not require specialized laboratory culture methods (other than a continuous-monitoring blood culture system) or catheter removal but requires that peripheral and central cultures are drawn simultaneously and are of equal volume. If the same organism is isolated from both blood cultures and the time to positivity of the catheter-obtained specimen is > 2 hours shorter than that of the peripherally obtained culture, catheter colonization and CRBSI are likely. ^{, ,} Compared with quantitative and semiquantitative methods, differential time to positivity of > 2 hours had a sensitivity of 93% and a specificity of 75% for catheters in place for > 30 days and a sensitivity of 81% and a specificity of 92% for catheters in place for <30 days. Several smaller studies have demonstrated similar results. In an analysis of semiquantitative cultures obtained from different lumens of multilumen catheters in pediatric oncology patients with double-lumen catheters in place and no peripheral blood cultures available, PPV of more than 5-fold difference in CFU/mL of isolates from samples from the two lumens was 92% predictive of a CRBSI.

Few studies have evaluated the optimal timing of blood cultures, although timing should be guided by patient acuity. Compared with obtaining multiple specimens at the same time, obtaining multiple samples over a 24-hour period appears to increase the ability to detect intermittent bacteremia. ^, In patients already receiving antibiotics, samples obtained close to the time that antibiotic concentrations have reached trough levels (i.e., just before next dose) theoretically could improve recovery of organisms in blood cultures ; however, this issue has not been studied and may not be practical clinically.

The volume of blood that is obtained is the most important factor in successfully recovering bacteria and fungi. ^, Studies have demonstrated that the likelihood of growth is lower and the time to detection is delayed when small volumes ( < 0.5 mL) of blood are used to inoculate blood culture bottles ; however, advancements in blood culture detection system technology continue to improve efficiency. Although multiple cultures enhance the recovery of pathogens, the volume of blood cultured may be more important than the total number of blood cultures obtained. One study found that the pathogen recovery rate at 24 hours was 72% for a large-volume (6 mL) single culture compared with a 47% combined yield of two smaller (2 mL) samples inoculated into separate culture bottles. Similar results were found in studies of adult patients in which standard “adult-volume” cultures (mean, 8.7 mL) had a higher detection rate (92%) than “low-volume” cultures (mean, 2.7 mL), in which the detection rate was only 69%. Investigators estimated that the yield of adult blood cultures increased approximately 3% per mL of blood cultured. For pediatric patients, however, the appropriate blood volume is typically dictated by the child’s weight and age.

Other technical issues can affect the sensitivity of blood cultures. Diluting the blood into the blood culture broth enhances recovery of pathogens, perhaps by diluting antimicrobial agents (if applicable) and blood components such as phagocytes, antibodies, and complement factors that are known to have bactericidal activity. The ideal blood-to-broth ratio depends on the blood culture system used, but a ratio between 1:5 and 1:10 generally is considered ideal ; it is essential to identify the optimal blood volume suitable for the culture system in one’s clinical microbiology laboratory. For recovery of certain bacteria, fungi, and mycobacteria, special handling of blood cultures is required. For example, filamentous fungi require solid media for growth, Malassezia furfur requires lipid supplementation, and certain gram-negative species and some mycobacteria need prolonged incubation (see Chapter 286 ). Failure of culture to identify a pathogen in the patient with continued clinical evidence of infection should prompt consideration of less common, more fastidious organisms and discussion with microbiology laboratory personnel about additional collection, culture, and detection techniques.

Distinguishing between cultures that represent “contamination” and those that represent “true” BSI can be challenging, especially when a skin organism (e.g., CoNS, Bacillus spp., micrococci) is isolated. Obtaining multiple cultures can clarify most situations. A study using a mathematical model of blood cultures positive for CoNS in patients with a CVC found that the PPV of a single positive culture (if only one culture was obtained) was 55%; however, if only 1 of 2 cultures obtained was positive, the PPV was 20%, and if only 1 of 3 cultures was positive, the PPV was only 5%. Investigators developed a similar model for blood culture positivity depending on sampling site. If 2 of 2 cultures were positive, the PPV was 98% if both samples were obtained from a peripheral vein, 96% if 1 sample was obtained through a catheter and the other was obtained through the vein, and only 50% if both samples were obtained through a catheter. Furthermore, the distinction between pathogen and contaminant is influenced by age and underlying conditions. Although CoNS often is considered a true pathogen in neonates, this issue remains controversial. Because the isolation of CoNS from a blood culture often results in a clinical intervention and administration of vancomycin, it is particularly important to obtain multiple blood cultures. Overuse of vancomycin has implications for adverse events, and the continued increase of vancomycin resistance among gram-positive organisms.

Repeatedly positive blood culture results despite appropriate antimicrobial therapy suggest the presence of an intravascular focus of infection such as the catheter itself, complicating suppurative thrombophlebitis or endocarditis (especially if the catheter has been removed). ^, Among patients with intravascular catheters, the risk for endocarditis is highest with Swan-Ganz catheters and lowest with peripheral catheters. In an autopsy series, 53% of patients who had undergone pulmonary artery catheterization had right-sided endocardial lesions; 7% had infective endocarditis. Centrally placed catheters often traverse and damage the tricuspid valve, increasing the risk for right-sided endocarditis. Thrombi within the heart can become infected or can obstruct outflow.

Ultrasonography is useful in detecting thrombosis of vessels or formation of vegetations. The value of a negative transthoracic ultrasonography (TTE) result is debated; in adults, the use of transesophageal echocardiography (TEE) increases the likelihood of detecting intracardiac vegetations. ^, Gadolinium-enhanced magnetic resonance venography can be useful in detecting central venous thromboses.

Management and Outcome

Exit site infections can be treated by removal of the catheter and local care. If the catheter remains in place, local topical therapy can be successful in controlling this type of infection; however, systemic antimicrobial therapy and catheter removal usually are required if an exit site infection extends to involve the soft tissues that surround the catheter (i.e., the subcutaneous tunnel or pocket) or is associated with a CLABSI. Limited data guide the management of CLABSI in children. Even in adults, there are insufficient randomized or controlled studies to address optimal management of CLABSI. ^, To help address this uncertainly, the Infectious Diseases Society of America has published guidelines for the diagnosis and management of CLABSI. Although focused on adult patients, this guideline contains recommendations for the management of children with CLABSI and is endorsed by the Pediatric Infectious Diseases Society.

Blood cultures are obtained ideally prior to initiating antimicrobial therapy, but laboratory testing should not delay empiric treatment in a critically ill child. Empiric therapy in children with suspected CLABSI should include an antimicrobial agent with activity against gram-positive bacteria, such as nafcillin, oxacillin, or vancomycin, and an agent effective against gram-negative bacteria, including Pseudomonas spp., such as ceftazidime or cefepime. The empiric use of both an antipseudomonal β-lactam and an aminoglycoside may be appropriate in severely ill patients or when infection with a resistant gram-negative organism is suspected. In institutions in which MRSA is prevalent, the use of vancomycin is appropriate. Aztreonam or fluoroquinolones can be used when β-lactam allergy is present and broad-spectrum gram-negative coverage is required.

Limited data guide clinical decisions regarding the need for catheter removal ( Table 100.2 ). In adults with CLABSI, it is recommended that most nontunneled CVCs be removed. In children, removal of a catheter may not always be feasible because of the potential for complications associated with reinsertion and limited vascular access sites. Children treated without catheter removal should be monitored closely and additional blood cultures drawn; the catheter should be removed with clinical deterioration or persistent or recurrent CLABSI. In patients with CLABSI associated with a tunneled catheter or implantable device such as a port catheter, the decision to remove the catheter is more complicated; however, good evidence favors the removal of the CVC in patients with evidence of a tunnel infection or infection in the subcutaneous pocket of an implanted device. When culture data are available, treatment decisions can be tailored to the specific organism isolated. Several studies have reported successful treatment of CLABSI without catheter removal, depending on the pathogen identified.

The aforementioned management guideline also includes the use of antimicrobial lock therapy in management of CLABSI when catheter salvage is the goal. The concept underpinning lock therapy is that bacteria embedded in biofilms on intraluminal surfaces of catheters may be killed more effectively by antibiotic concentrations 100–1000 times higher than standard concentrations of IV antibiotics. Lock solutions typically are mixed with heparin in sufficient volumes to fill the lumen, allowed to dwell, then removed. This approach is currently recommended in conjunction with systemic therapy for select organisms.

Few data exist regarding the duration of antibiotic therapy for CLABSI. The duration of therapy depends in part on the pathogen; whether the catheter is removed; and whether infection is complicated. Common complications associated with CLABSI include septic thrombosis, venous thrombosis, endocarditis, osteomyelitis, or other metastatic foci of bacteria. For complicated infections, the duration of therapy is based on the length of therapy needed to treat the suppurative complication. There are no data to determine the optimal duration of IV versus oral antibiotics for the treatment of CLABSI when the catheter is removed. Certain antibiotics with excellent oral bioavailability may be considered an alternative to parenteral therapy after the catheter has been removed (or lock therapy begun), the BSI has cleared, and the patient has shown clinical improvement. Some pathogen-specific recommendations are provided later.