A Clinical guide to pediatric infectious disease
Laboratory Diagnosis of Pediatric Infectious Disease
A variety of laboratory tests are used in pediatric infectious disease. Useful techniques include blood cultures, Gram stain, serology, and determination of minimal inhibitory concentration (MIC). This chapter discusses these methods and addresses the most common questions regarding the interpretation of these tests.
Interpretation of a Positive Culture
Clinicians are often required to determine whether a bacterial culture is “real,” that is, represents a true invasive bacterial infection. The accurate interpretation of a positive bacterial culture requires an understanding of colonization, contamination, and invasive infection.
Distinguishing Colonization, Contamination, and Invasive Infections
Colonization refers to the isolation of bacteria or fungi from an area of the body that is not expected to be sterile, including skin surfaces and tracheostomy tubes. These surfaces are in contact with the outside environment, and the presence of bacteria is not indicative of invasive disease. Within hours after birth, neonates become colonized with a variety of bacteria and even fungal pathogens. Cultures of their nose, mouth, or skin would be positive for numerous bacteria and yeast. Usually, organisms residing in these areas do so without ill effects to the host.
Contamination refers to the presence of an organism isolated from a site expected to be sterile. These organisms are not causing disease; rather, their isolation is the result of a break in sterile technique during sampling. Contaminated body fluids can include blood, peritoneal fluid, and even cerebrospinal fluid.
Invasive infection refers to the presence of bacteria and fungi that are causing clinical disease.
Colonization versus Invasive Infection
In evaluating a positive culture from a nonsterile area, the first issue is whether the culture represents colonization or true infection. This is typically an assessment required with a skin or endotracheal tube culture. One must rely on the overall clinical picture of the patient. A skin culture obtained in the context of cellulitis or obvious wound infection would likely represent a true infection. A positive skin culture obtained on intact or normal-looking skin usually represents colonization.
In evaluating an endotracheal tube culture, it is also important to consider the entire clinical picture. Many studies have evaluated the optimal method for determining whether an endotracheal tube or tracheostomy tube culture represents a ventilator-associated pneumonia (invasive disease) or routine colonization. Generally, a clinical decision is made involving an evaluation of the patient's vital signs, respiratory status, oxygenation, and chest x-ray. An endotracheal tube culture obtained in a stable patient without respiratory deterioration is often deemed colonization. A positive endotracheal tube culture taken in the setting of decreasing oxygenation, an increase in ventilatory status, fever, leukocytosis, or new pulmonary infiltrates would likely represent a true invasive pathogen.
Contaminant versus Invasive Infection
The second exercise frequently facing the clinician is determining whether an isolate from a sterile body site (such as blood or cerebrospinal fluid) is a contaminant or a true infection. The following case provides an example of the importance of making an accurate determination.
A term newborn with a history of intrauterine drug exposure was evaluated for a low-grade fever. A blood culture was obtained and grewStaphylococcus aureus. Repeat blood cultures were negative, and this isolate was thought to be a contaminant.
Over the next 2 weeks, the child became increasing irritable; this was attributed to drug withdrawal. Phenobarbital was given with little effect. After 2 weeks, the infant's knee joint became swollen and ultimately ruptured, discharging purulent fluid. Culture of this fluid grew S. aureus.
Group A streptococcus (Streptococcus pyogenes)
Coagulase-negative staphylococcus (Staphylococcus epidermidis)
This story stresses the importance of an accurate determination of whether a positive blood culture constitutes a true infection. Although the bacteremia itself may be transient, the bacteria may come to rest at secondary sites such as the brain or bone. Therefore, it is wise to treat aggressively all pathogens that are thought to be “real,” even if repeat blood cultures are negative. Infection at secondary sites, left untreated, can cause major morbidity and even mortality.
A good system to use is the “always/sometimes” system. Certain pathogens, when isolated from a sterile site (such as blood), are always considered true infections and require treatment. These pathogens should not readily be considered contaminants.
A good system to use is the “always/sometimes” system. Certain pathogens, when isolated from a sterile site (such as blood), are always considered true infections and require treatment. These pathogens should not readily be considered contaminants.
There are some bacteria that, when isolated in blood or spinal fluid, can be either invasive pathogens or contaminants. Often, these organisms reside on skin surfaces, facilitating the contamination in cultures. These can be referred to as the “sometimes” pathogens. Despite these organisms often being contaminants, care must be taken in making this designation. Generally speaking, the more immunocompromised a patient (particularly in the setting of indwelling catheters), the greater the chance of skin flora becoming a true invasive pathogen.
As in the evaluation of a skin or endotracheal tube culture, a clinical context is needed. A well patient with a single positive blood culture for coagulase-negative staphylococcus usually warrants no therapy. A febrile oncology patient with an indwelling catheter and a blood culture for Corynebacterium or Bacillus species should be strongly considered for treatment because these organisms represent pathogens in this patient population.
Quantitation of bacterial colonies is an additional technique considered helpful in interpretation of blood cultures. Peripheral cultures that grow more than 50 colony forming units (CFUs) likely represent infection, whereas peripheral blood cultures with less than 5 CFUs usually represents contamination. Unfortunately, the ability to perform quantitative culture is often not available.
The Gram stain is often the first laboratory test returned to a clinician caring for an ill child, usually reported as an organism growing in blood or cerebrospinal fluid. Correct interpretation of a positive Gram stain in the blood or cerebrospinal fluid can
lead to optimal antibiotic management, which in turn can have a significant effect on patient outcome. One way to approach the correct interpretation of a Gram stain is the “Gram stain game”. This can be done by dividing the isolate into major categories based on Gram stain morphology. The possible isolates with a given Gram stain morphology can then be evaluated in regard to the particular patient.
Gram Stain Game
The Gram stain divides bacteria into two groups. Gram-positive bacteria retain crystal violet stain, whereas gram-negative bacteria do not and appear red. Gram stains are also designated in terms of morphology and configuration; bacteria are typically designated as cocci, rods, clusters, and chains (Table 1.1).
TABLE 1.1. Gram Stain Game
Gram-positive Cocci in Clusters
Clusters of gram-positive cocci usually represent Staphylococcus species infection. Staphylococci are divided traditionally into two groups based on the coagulase test. S. aureus is the most common coagulase-positive isolate. Coagulase-negative staphylococci include more than 30 species, the most common of which is Staphylococcus epidermidis. S. epidermidis is a common infection of surgical catheters and implanted foreign bodies and is a common contaminant found in blood cultures. Methicillin-resistant staphylococci should also be considered in the initial evaluation of any Gram stain showing gram-positive cocci in clusters. Once thought to occur primarily in hospitalized patients, a large percentage of community-acquired S. aureus infections are now methicillin resistant. When this Gram stain is isolated from an ill patient and empiric treatment is warranted, vancomycin provides empiric coverage.
Gram-positive Cocci in Chains
In children, chains of gram-positive cocci are usually seen in five organisms:
- Group A streptococcus
- Group B streptococcus
- Streptococcus pneumoniae
- Viridans streptococci
These are classified in numerous ways. One of the most frequently used systems classifies the ability of the organism to lyse red blood cells on a blood agar plate. β-Hemolytic streptococci produce complete hemolysis on sheep blood agar. α-Hemolytic organisms produce partial hemolysis, resulting in a green or a grayish zone surrounding colonies. Nonhemolytic organisms impart no hemolysis on blood auger plate.
Group A streptococcus is a β-hemolytic organism that is implicated in pharyngitis as well as toxic shock syndrome.
Group B streptococcus is also β-hemolytic and is a common colonizer of the female genital tract that is a primary cause of neonatal sepsis and meningitis.
Streptococcus pneumoniae is the major cause of otitis media and bacterial pneumonia in toddlers. It is typically α-hemolytic.
Viridans streptococcus is not a single organism but rather a large number of different streptococci. These α-hemolytic streptococci reside in the gastrointestinal (GI) tract and are implicated in transient bacteremia, endocarditis, and GI-related septicemia.
Enterococcus species are nonhemolytic organisms residing in the GI tract that is often implicated in endocarditis and septicemia secondary to abdominal trauma.
Clinical context can help the clinician make the best guess as to the offending pathogen. A neonate who is clinically septic and has a blood culture isolate of gram-positive cocci in chains is likely to have group B streptococcus; less likely are viridans streptococci and enterococci. Empiric treatment with ampicillin and gentamicin covers these pathogens best. A toddler with a diagnosis of bacterial meningitis and gram-positive cocci in chains seen in the cerebrospinal fluid is likely to have Streptococcus pneumoniae. In this case, the best empiric treatment is vancomycin plus a third-generation cephalosporin. A patient who has intraabdominal sepsis with gram-positive cocci in chains isolated from nonblood or peritoneal culture is likely to have viridans streptococci or enterococci because these organisms reside in the GI tract. In this case, ampicillin with gentamicin or vancomycin provides good coverage.
The gram-positive rods are an unusual group of pathogens that can cause disease in specialized conditions. In evaluating this group, a helpful first step is to determine whether the organism is acid-fast positive.
Gram-positive rods that are acid-fast positive include the following:
Gram-positive rods that are not acid-fast positive include the following:
- Listeria monocytogenes
Rhodococcus (formally Corynebacterium) equi is a gram-positive rod that may be partially acid fast. Primarily an animal pathogen, it has become recognized as a human pathogen that causes respiratory disease and sepsis in immunocompromised hosts. The organism may cause a necrotizing pneumonia associated with bacteremia that can resemble Mycobacterium tuberculosis or Nocardia species infection. These bacteria are resistant to β-lactam antibiotics and often require prolonged treatment with vancomycin.
Nocardia species are weakly acid fast. In immunocompromised patients, including those with chronic granulomatous disease and those receiving long-term corticosteroid treatment, chronic pneumonia may be seen. Pulmonary disease with dissemination to skin and brain is frequently encountered. Nocardia pneumonia often appears as a consolidation process resembling mycobacterial disease. Long-term treatment with trimethoprim-sulfamethoxazole is usually needed.
Listeria monocytogenes affects neonates and immunocompromised patients. Neonates have early-onset and late-onset syndromes similar to group B streptococcal infection. Listeria is an aerobic, non-spore-forming, motile gram-positive rod that is usually easily identified by the laboratory once the initial morphology of gram-positive rod is confirmed. Antibiotic therapy includes the combination treatment of ampicillin and gentamicin.
Bacillus species are non-acid-fast gram-positive rods. The major member of this species in terms of past and present historical importance remains Bacillus anthracis. Anthrax is caused when spores are inhaled, digested, or come in close contact with body surfaces. Skin infections begin as painless lesions around which vesicles develop. The vesicles then progress to an eschar and are often accompanied by marked edema. Inhalation anthrax begins with flulike illness which progresses to respiratory failure, sepsis, and meningitis. Although anthrax remains rare, it should be considered in the proper epidemiologic and clinical contexts.
The other major Bacillus species is Bacillus cereus, an endotoxin-producing spore. This organism can cause endophthalmitis and neonatal sepsis. A gram-positive rod isolated from a septic neonate, even if the lab indicates it is not Listeria species, should not automatically be considered a contaminant. B. cereus is resistant to β-lactam antibiotics and requires vancomycin for treatment.
Corynebacterium species are commonly referred to as diphtheroids and are nonmotile, non-acid-fast organisms that are the most common gram-positive rod species isolated in blood cultures. Corynebacteria are common in the environment and are part of the normal skin flora. Although frequently the cause of contamination in blood cultures, their isolation can also signify the presence of true invasive disease.Corynebacterium diphtheriae, the cause of diphtheria, is rare in developed countries. Corynebacterium jeikeium, previously known asCorynebacterium group JK, is an opportunistic pathogen increasingly seen in hematology and oncology patients. Corynebacterium striatum is reported as a cause of serious infection in the immunocompromised host. Patients with leukopenia and indwelling catheters are also particularly susceptible. C. jeikeium is resistant to most antibiotics and requires vancomycin for treatment. Specific identification of an organism labeled diphtheroid or corynebacteria should be strongly considered if the patient is immunosuppressed or seriously ill.
Clostridium species are anaerobic bacilli that reside in the GI tract and the environment. These bacteria usually require anaerobic media to grow. These organisms often secrete toxins that are involved in the pathogenesis of the disease process.
These include Clostridium perfringens (gas gangrene), Clostridium botulinum (associated with paralysis), Clostridium difficile (antibiotic-associated diarrhea), and Clostridium tetani (tetanus).
It is helpful to divide the gram-negative rods into two basic groups: those that are community acquired and those causing nosocomial infections.
The community-acquired gram-negative rods include the following:
- Escherichia coli
- Haemophilus influenzae
The incidence of haemophilus type B disease has greatly diminished since the introduction of the conjugate vaccine. After the newborn period (when E. coli predominates), a toddler with fever, gastroenteritis, and gram-negative rods in the blood is likely to have a Salmonellaspecies infection. A third-generation cephalosporin in these instances is usually adequate.
The nosocomial gram-negative rods are a different group of organisms that includes the following:
These organisms are noteworthy for their rapid induction of β-lactamase production. Even if they are initially sensitive to semisynthetic penicillins or cephalosporins, within a few days, these organisms can become resistant.
β-Lactam Sensitivity Cannot Be Assumed
Never rely on a third-generation cephalosporin alone when faced with a gram-negative sepsis in a severely ill patient in a hospital or intensive care setting. Empiric treatment with more broad-spectrum coverage (such as amikacin, imipenem, or a fluoroquinolone) should be considered until final identification and sensitivities are known. Numerous studies point to the significance of the early correct treatment of nosocomial gram-negative infection on overall morbidity and mortality.
Extended-Spectrum β-Lactamase (ESBL)
In addition to empiric treatment for infection with presumed β-lactamase-producing organisms, understanding the correct laboratory diagnosis of these pathogens is critical.
Various organisms, typically gram-negative bacteria, can produce extended-spectrum β-lactamase (ESBL) by a variety of means. These β-lactamases are often secondary to encoded plasmids and cause structural changes in the β-lactamase protein that can increase its activity to include the third- and fourth-generation cephalosporins and most semisynthetic penicillins. Patients most likely to develop these ESBL-producing organisms are those with prolonged hospitalization or multiple courses of broad-spectrum antibiotics.
The initial clinical suspicion for ESBL begins with evaluation of the antibiotic resistance panel of an isolated organism. ESBL-producing organisms typically have reduced susceptibility to all cephalosporins and semisynthetic penicillins. The National Committee for Clinical Laboratory Standards (NCCLS) has proposed screening criteria for ESBL production; an MIC greater than 4 µg/mL for ceftdzidime is thought to be particularly worrisome. Confirmatory testing involves “double-disk” Kirby-Bauer methodology. Zones of inhibition around an isolate are measured against disks of third-generation cephalosporins alone and then in combination with disks containing the β-lactamase inhibitor clavulanic acid. A more than 5-mm increase in the zone of inhibition when the disks are combined usually indicates ESBL production. If the laboratory does confirm an ESBL-producing strain, β-lactam antibiotics such as semisynthetic penicillins or cephalosporins should not be used, regardless of in vitro antibiotic reporting. Most authorities suggest carbapenems (imipenem) as front-line therapy for ESBL-producing organisms.
Two major organisms are in the group of gram-negative diplococci, the most important being Neisseria meningitidis. N. meningitidis remains a major cause of overwhelming sepsis and meningitis. Moraxella catarrhalis is a common cause of lower respiratory infection and otitis media in children; it occasionally can cause invasive bacteremic disease. Neisseria species remain susceptible to penicillin, whereasMoraxella species have a higher incidence of β-lactam production; these organisms require second- or third-generation cephalosporins.
Serologic Diagnosis in Infectious Disease
Many agents in pediatric infectious disease are difficult to culture. These include atypical bacteria, such as Rickettsia, Legionella, andChlamydia species, and viral pathogens. The measurements of serum antibodies provide the clinician with a valuable tool in diagnosing infection.
Two major points regarding serologic diagnosis must be remembered when using this diagnostic technique. First is that most serology measures immunoglobulin G (IgG) antibodies that readily cross the placenta. These maternal antibodies can persist in an infant's circulation for many months and sometimes for up to a year, complicating the interpretation of these tests in young children.
The second potential pitfall is that there is a very high incidence of “background” IgG serology. Most individuals have been exposed to numerous pathogens and are positive for a variety of IgG antibodies. A patient with acute hepatitis and an IgG antibody for hepatitis A does not have proven acute hepatitis A infection. The diagnosis of acute infection often must be made by more sophisticated means. Specific IgM serology often is used because IgM represents the earliest antibody response and is often elevated in the setting of acute infection. The diagnosis of acute infection can also be made by a single IgG titer; a sufficiently high IgG titer is thought not to represent background serology and can only be present in the context of an acute infection. An additional method of diagnosing an acute or recent infection is to use paired serology to show a rise in IgG antibodies; a four-fold rise in IgG antibody titers measured several weeks apart is also taken as evidence of recent infection.
Direct Immunofluorescent Antibody
In addition to serology, other methods can be used to document the presence of difficult-to-culture organisms. Direct fluorescent antibody (DFA) staining employs the use of specific antibodies linked to a light-emitting material. The antibody binds to the infecting pathogen and can then be visualized by a specially equipped microscope to detect the emitted fluorescence. The advantage of this technique is that it can be done rapidly, often in less than 24 hours.
Minimal Inhibitory Concentration
The MIC refers to the lowest concentration of an antibiotic that visibly inhibits the in vitro growth of an organism. This information can help to determine whether a given antibiotic can be used for treatment of specific bacteria. There are several ways of determining the MIC.
Dilution testing is the major methodology in determining MIC (Fig. 1.1). An early method for dilution testing actually used rows of standard test tubes (macrodilution method). Typically, eight or more concentrations of an antibiotic were used
with twofold dilutions (16, 8, 4, 2, and 1 µg/mL). A standard amount of bacteria, usually 5 × 105 CFUs/mL, is inoculated in each test tube. In 24 hours, the tubes are examined, and the first concentration in which there is no visible growth is determined to be the MIC.
FIG. 1.1. Minimal inhibitory concentration (MIC).
Tube dilution has been modified by miniaturization techniques. Microdilution trays that contain 96 wells of 100µL each allow up to 12 antibiotics to be tested at one time. Visible growth in a specific tube dilution can be determined visually or by automated photometers.
Another method for calculating MIC is the disk diffusion method (Kirby-Bauer method), which involves a bacterial inoculation of 1 to 2 ×108CFUs/mL to an agar plate, on which is placed an antibiotic disk. After 24 hours of incubation, diameters of the zone of inhibition around the antibiotic disks are measured to the nearest millimeter (Fig. 1.2). This zone-of-inhibition diameter is related to the susceptibility of the isolate (i.e., linear regression analysis of zone diameter plotted against log 2 MIC values). This test is attractive in that it is simple and can be performed in community hospital laboratories. However, the disadvantage of this test is that it provides only a qualitative result rather than a quantitative MIC.
Nonetheless, this diffusion method remains in place in many laboratories and can be helpful in many basic infections.
The NCCLS provides guidelines for methodology in determining MICs. It also determines the “breakpoints” (i.e., the MIC results for a given drug that are interpreted as susceptible, intermediate, or resistant). The breakpoints are made through an analysis of several factors, including data on achievable drug concentrations, the results of clinical trials, and appraisal of toxicities.
Although the MIC and breakpoints are often used in selecting a particular antibiotic, additional variables need to be considered. Drug concentrations within the patient may change over time or may not represent the actual free drug concentration at the site of infection. Drugs that are excreted renally may actually have higher concentrations in the urine; drugs that do not cross the blood-brain barrier may actually have reduced levels within the cerebrospinal fluid. Multiple additional factors, such as toxin production, individual host response, and concentration of antibiotic at the site of infection, are important and can determine response. Breakpoints often need to be reevaluated as new data emerge regarding the development of antibiotic resistance. The “90-60 rule” has been proposed; this rule states that a susceptible isolate will respond 90% of the time, whereas a resistant isolate will respond only about 60% of the time.
FIG. 1.2. Disc diffusion method (Kirby-Bauer).
Pharmacodynamics in Infectious Disease
In addition to MIC calculations, the advancing field of pharmacodynamics has led to additional refinement in the way clinicians use the MIC and administer antibiotics.
For antibiotics in this group, such as aminoglycosides and the fluoroquinolones, the rate of killing of an organism is dependent on the concentration of the antibiotic above the MIC. High ratios of peak drug concentration to MIC (10:1) are correlated with lower development of resistance and improved clinical outcome. The use of once-daily aminoglycoside dosing is based on this principle; increasing the milligram per kilogram dose given once a day increases the peak concentration and therefore increases the ratio of peak drug concentration to MIC. Additionally, the risk for nephrotoxicity with aminoglycosides is thought to be related to increased trough concentration for an extended period of time. It is possible that this could be reduced with once-daily dosing because there is an increased period of time with lower undetectable drug concentrations.
Although there has been increasing acceptance of once-daily aminoglycoside dosing in adults, it has yet to become standard practice in children. Most published data regarding the use of once-daily dosing in adults are based on a short duration of therapy. Children may have different volumes of distribution and excretion patterns than adults, leading to different outcomes. There is concern that a higher dose given once daily, often 6 mg/kg, may be associated with higher ototoxicity if used for an extended periods. Although once-daily dosing remains an attractive theory, it has not been definitively proved in pediatric studies. It is possible that once-daily dosing will gain increasing acceptance in pediatrics as clinical experience with this therapy increases.
This pharmacodynamic pattern is seen with β-lactam antibiotics such as penicillin, cephalosporins, and carbapenems. Maximal killing is related not to the concentration ratio above the MIC, but rather to the time above the MIC. For penicillins and cephalosporins, the plasma concentration needs to exceed the MIC for 50% to 70% of the dosing interval for maximal effectiveness. In the case of carbapenems, the plasma concentration should exceed the MIC for at least 40% of the dosing interval. It is for this reason that these drugs are often considered for continuous infusions, so that the plasma concentrations will continually be above the MIC. Like once-daily therapy for aminoglycoside administration, there is minimal clinical experience in pediatrics, and it is not used extensively in children's hospitals at this time.
The art of optimal antibiotic administration is an evolving one. MIC determinations, clinical experience, the site of infections, and pharmacodynamics will all play a role in future studies regarding antibiotic use.
American Academy of Pediatrics. Report of the Committee on Infectious Disease 2003, 26th ed.
Feigin R, Cherry J. Textbook of pediatric infectious diseases, 5th ed. Philadelphia: WB Saunders, 2003.
Long SS, Pickering LK, Proter CG, eds. Principles and practice of infectious diseases, 2nd ed. Philadelphia: Churchill Livingstone