Misdiagnosis

The bacteriologic identification of K. kingae relies on the following: typical Gram stain results, showing pairs or short chains of plump, gram-negative bacilli with tapered ends; β-hemolysis; pitting of the agar surface; failure to grow on MacConkey medium; weak oxidase activity and a negative catalase reaction; production of acid from glucose; and, with rare exceptions, production of acid maltose (8,9). However, K. kingae tends to retain crystal violet dye and, therefore, it may appear to be gram-positive, and laboratories unfamiliar with its cultural and staining features may misidentify the bacterium altogether or dismiss invasive isolates as culture contaminants. Identification of K. kingae, however, is not difficult, and many commercial instruments and technologies such as VITEK 2 (bioMérieux, Marcy-l’Etoile, France), matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF), or 16S rDNA gene sequencing, correctly identify the organism.

Characteristics and Manifestations

Recent studies have demonstrated the presence of a polysaccharide capsule on the surface of K. kingae that exhibits between-strains chemical variation and, most probably, antigenic variation (17,18). These characteristics may assist K. kingae in evading the immune response by facilitating the successive colonization of the host by strains harboring different capsular types. The polysaccharide capsule probably enables mucosal colonization, the survival of the organism in the bloodstream, and invasion of deep body sites (17,18). In addition, all K. kingae strains produce and secrete a potent repeats-in-toxin (RTX) that exhibits a wide range of cytotoxic activity and is particularly deleterious to macrophage-like cells, leukocytes, and synovial cells and, to a lesser degree, to respiratory epithelial cells that improve the organism’s chances of surviving in the host and of invading skeletal tissues (12,19).

Similar to meningococci, K. kingae is carried on the oropharyngeal epithelium (20,21), and the colonized mucosa is the portal of entry of the organism to the bloodstream from which it may disseminate to 3 areas for which the bacterium shows particular tropism: joints, bones, or the endocardium (22,23). Damage to the upper respiratory surfaces by previous or concurrent viral infections or stomatitis appears to facilitate bloodstream invasion by K. kingae (16).

K. kingae isolates show remarkable genomic diversity and, to date, 37 multilocus sequence typing (MLST) and 74 pulsed field gel electrophoresis (PFGE) clones have been identified (24–26). Carried K. kingae organisms differ in their invasive capabilities (27), and simultaneous carriage of ≥1 genotype is unusual (28). Whereas some strains, which are frequently isolated from healthy carriers, are seldom if ever detected in patients with clinical disease, others are rarely carried asymptomatically, but are responsible for a high proportion of invasive disease (24,27,28). However, a few strains appear to possess an optimal balance between transmissibility and invasiveness and are common among healthy carriers and among infected patients (24,27,28). A recent study has found that certain virulent K. kingae clones, characterized by a distinct combination of PFGE and MLST profiles, are substantially associated with bacteremia with no focal infection, skeletal system invasion, or endocarditis, which suggests biological specialization for invading specific host tissues (25).

Most young children in whom an invasive K. kingae disease developed have been otherwise healthy. In contrast, children >4 years of age and adults who become infected frequently have underlying conditions such as congenital heart diseases, chronic renal failure, or a variety of primary immunodeficiencies (16).

The prevalence rate in healthy children during the second year of life ranges between 10% and 12% (7,28), which coincides with the peak attack rate of invasive infections (16). The colonization rate drops substantially in older children and adults (7,28,29). Pharyngeal carriage of K. kingae and occurrence of disease before a child is 6 months of age are exceptions, indicating that maternal immunity and limited social contact provide protection (7,16,30).

Diagnosis of K. kingae Infections

The clinical features of invasive K. kingae infections (other than endocarditis) are usually mild, and diagnosis requires a high level of suspicion by clinicians. Many patients with K. kingae–associated joint or bone disease are afebrile and blood leukocyte counts, C-reactive protein levels, and erythrocyte sedimentation rates are frequently normal (16,31,32). Although bacteremia with no focus is the second most frequent manifestation of invasive K. kingae infections, the condition is probably unsuspected and the diagnosis is likely missed in a large number of cases. The current guidelines for managing illness in young, febrile children with no apparent source of infection, which use body temperature and leukocyte count as criteria for obtaining blood cultures (33), are not sensitive enough for detecting occult K. kingae bacteremia because many infected children have a low-grade fever and may or may not have leukocytosis.

Detection of K. kingae infections by culture is highly dependent on the use of adequate laboratory techniques. The recovery of K. kingae from synovial fluid and bone exudates seeded onto routine solid culture media is suboptimal (8). The yield of cultures can be substantially improved by inoculating clinical specimens into aerobic blood culture vials (BCVs) from a variety of commercial systems (8). Attempts to isolate the organism from synovial fluid or bone exudates on routine solid media succeeded in 2 of 25 patients, whereas inoculation of these specimens into aerobic BACTEC (Becton Dickinson, Cockeysville, MD, USA) BCVs yielded the organism in all cases after a median incubation of 4 days (8). When specimens from BCVs determined to be positive by the automated blood culture instrument were subcultured onto a blood–agar plate of trypticase soy agar with 5% sheep blood hemoglobin or chocolate agar, K. kingae grew readily, indicating that routine solid media are able to support the nutritional requirements of the organism. This observation suggests that skeletal system exudates exert a detrimental effect on this fastidious bacterium, and dilution of purulent material in a large volume of broth decreases the concentration of inhibitory factors, improving its recovery (8). In studies conducted in Israel and France in which BCVs were routinely inoculated with synovial fluid aspirates from young children who had arthritis, K. kingae was isolated in 48% of the patients with culture-proven disease (8,9). Conversely, when BCVs are not used, many K. kingae infections will be overlooked and labeled as culture-negative septic arthritis (34).

In recent years, development of NAAAs has further improved the diagnosis of K. kingae from skeletal system exudates. Use of this novel technique facilitates detection of difficult-to-culture organisms, enables bacteriologic diagnosis in patients already treated with antibacterial drugs, reduces time to detection, and facilitates precise identification of unusual species (10–14). The procedure consists of extraction of bacterial DNA from the synovial fluid sample, a PCR amplification step in which primers target the 16S rDNA, the 23S rDNA, or the rpo genes that are ubiquitous to all bacteria, then sequencing of the amplicon, and comparison of results with those kept in a broad database curator (such as GenBank) to enable precise species identification. Alternately, the specimen may be subjected to amplification by PCR by using species-specific primers that recognize the most plausible pathogens. Use of NAAAs has confirmed that in countries where this topic has been studied, K. kingae is the most common etiologic agent of septic arthritis in children <3 years of age. NAAA use detected the presence of K. kingae–specific DNA sequences, including in cases in which seeding of synovial fluid specimens into BCVs failed to recover the organism and shortened the time required to detect and identify the bacterium from 3–4 days to <24 hours (10–14). It was, then, natural that this sensitive approach was consequently adopted to study respiratory colonization by K. kingae and its connection to invasive disease.

In an outbreak investigated by Bidet et al., the cause of the skeletal infections was determined by sequential use of real-time PCR targeting the K. kingae–specific toxin rtxA and the cpn60 genes (35). The same method was used to investigate the prevalence of K. kingae among attendees of the index daycare facility. The organism was recovered by culture in 6 asymptomatic carriers, whereas the NAAs detected 5 additional carriers (35). MLST and sequencing of the rtxA amplicon were performed directly on the joint fluid sample from the child who had arthritis and on the recovered pharyngeal isolates. All carriage isolates and the arthritis strain belonged to MLST 25 and shared rtxA allele 1, which is among the most common genotypes involved in joint and bone infections in France (26,35). It should be pointed out that, despite the increased sensitivity of NAAAs, when evaluating the efficacy of prophylactic antibacterial drug administration for eradicating K. kingae from colonized children, cultures have the obvious advantage of detecting living bacteria, whereas the viability of K. kingae organisms in PCR-positive/culture-negative specimens is questionable.

In addition detecting invasive K. kingae disease in patients, searching for asymptomatic but colonized children is crucial in the investigation of an outbreak in a daycare facility to assess the full extent of the contagion, identify attendees at risk for clinical disease, and evaluate the effects of prophylactic antibacterial drugs. Because of the high density of the resident bacterial flora and the relatively slow growth of K. kingae, detecting the organism in pharyngeal cultures is difficult. A differential and selective medium consisting of blood agar with 2 mg/mL of vancomycin (BAV medium) added has been developed to improve recovery of K. kingae from respiratory cultures (36). This formulation facilitates recognition of β-hemolytic K. kingae colonies by inhibiting growth of competitive gram-positive bacteria. In a blinded evaluation, the BAV medium detected 43 (97.7%) of 44 pharyngeal cultures positive for the organism; 10 (22.7%) positive cultures were identified on routine blood-agar plates (p<0.001) (36). If the original BAV formulation (36) or a similar medium (23) had not been used, and a chocolate-based agar substituted (in which the faint ring of hemolysis surrounding K. kingae colonies could not be recognized), carriers of the organism might not have been detected among 27 asymptomatic attendees of a daycare facility where a cluster of invasive disease occurred (37).