OBJECTIVE. Kingella kingae, a Gram-negative coccobacillus, is being increasingly recognized as an invasive pathogen in children, causing mainly bacteremia and arthritis; however, there have been only a few studies on K kingae infections to date, mostly small-scale series. The aim of this study was to report our experience with invasive K kingae infections in children who were hospitalized at a major tertiary medical center in Israel.
METHODS. The medical charts of 62 children with proven invasive K kingae infections were reviewed: 42 with positive blood culture results and 20 with positive synovial fluid culture results.
RESULTS. Most infections occurred among previously healthy children aged 5 to 22 months. Eighty percent had a mild concurrent illness of the respiratory or gastrointestinal tract. A chronic underlying disease was documented in 19% of the 1- to 15-year-old children with bacteremia. Three patients had persistent bacteremia, identified by 2 positive blood cultures drawn 1 to 4 days apart. Four (10%) patients from the bacteremia group had endocarditis, and 2 required emergency cardiac surgery. Only a mild-to-moderate elevation of serum inflammatory markers was noted except for patients with endocarditis or a prolonged course of arthritis. Patients with bacteremia received a diagnosis significantly later than those with arthritis, with no other between-group differences in age, month of disease onset, and inflammatory marker levels. All K kingae isolates were resistant to vancomycin and clindamycin.
CONCLUSIONS. Our large series indicates that invasive K kingae infections occur in previously healthy children, mostly during the first 2 years of life; affected older children usually have an underlying medical condition. The infection generally elicits only a mild inflammatory response unless accompanied by endocarditis. Despite its low virulence, K kingae might cause a life-threatening heart disease that requires emergent, aggressive treatment.
Kingella kingae is a slow-growing Gram-negative coccobacillus that is being increasingly recognized as an invasive pathogen in children.1,2K kingae infection causes mainly osteoarticular disease and bacteremia and is 1 of the HACEK (Haemophilus species, Actinobacillus actinomycetemcomitans, Cardiobacterium hominis, Eikenella corrodens, K kingae) group of fastidious microorganisms that are known to cause endocarditis on rare occasions. In addition, few cases of meningitis, endophthalmitis, and dactylitis have been reported.1,2 A study of healthy children attending child care centers in southern Israel yielded a pharyngeal carriage rate of >70%.3 In 2 studies of invasive K kingae outbreaks in child care centers in Israel and the United States, up to 45% of pharyngeal cultures grew K kingae.4,5 The high carrier rates combined with the low incidence of identified K kingae infections is possibly explained by a low virulence of this bacterium.
Only a few series of pediatric invasive K kingae infections have been published to date. The largest was published by Yaguspky et al6–8 from southern Israel and included 82 children; they described the clinical characteristics of K kingae infections and found a higher prevalence of infections in the second half of the year. Small clusters of patients from the United States and Australia were also reported.4,9,10 Given the recent emergence of invasive K kingae infections, more data from larger studies may further our understanding of the spectrum of diseases that is caused by this microorganism. The aim of this study was to investigate the epidemiologic, clinical, and laboratory features of a large group of children with culture-proven invasive K kingae infection.
The study was conducted at Schneider Children's Medical Center of Israel, a tertiary pediatric hospital located in central Israel. According to hospital protocol, blood cultures are routinely drawn from patients with high fever or suspected localized bacterial infections, such as arthritis, osteomyelitis, pneumonia, or pyelonephritis, and placed in BACTEC Peds Plus (Becton Dickinson, Franklin Lakes, NJ) bottles. In cases of arthritis, joint fluid is aspirated and sent for cell count and direct microscopic examination and cultured in BACTEC Peds Plus bottles and on solid media. Positive culture samples undergo direct Gram staining and are spread on culture plates that contain 5% sheep hemoglobin and chocolate agar. Findings of Gram-negative rods, often arranged in pairs or short chains, suggest K kingae; the diagnosis is confirmed by identification of the known biochemical characteristics of the microorganism.11
For this study, we reviewed the medical charts of all children with a positive blood or synovial fluid culture result for K kingae, tested at Schneider Children's Medical Center of Israel (SCMI) from January 1996 to December 2006. Epidemiologic, clinical, and laboratory data were gathered, and findings in patients with bacteremia or arthritis were compared. Season/month of disease onset was analyzed against a hypothetical equal monthly distribution. Antibiogram data were analyzed as well. The study was approved by the institutional review board of SCMI.
Continuous variables were compared by Mann-Whitney U test. Because of its normal distribution, body temperature was compared by t test. Proportions were compared by χ2 or Fisher's exact test, depending on group size. Linear regression was used to determine the trend of annual infection rate throughout the study period. A 2-tailed P value of <.05 was considered statistically significant. In a few cases of missing data, the number of patients for whom information was available is given in parentheses.
Sixty-two children with culture-proven invasive K kingae infections were identified. Forty two (68%) had positive blood culture results, and 20 (32%) had positive synovial fluid culture results. Eighty-seven percent of the children were previously healthy; all were older than 5 months, and only 1 was older than 2 years (Fig 1). Eight children had preexisting medical conditions, and their age range was 10 months to 15 years (median: 5 years; Fig 1). Eighty percent of children had a concurrent acute illness of the respiratory or gastrointestinal tract. The infections occurred throughout the year, with no statistically significant deviation from a hypothetical equal monthly distribution for all patients (P = .3), for bacteremia group only (P = .7), or for arthritis group only (P = .2). Infections were not more frequent in the second half of the year (1-tailed P-value for all patients P = .13, for bacteremia group P = .12, and for arthritis group P = .7). The median rate of infections per study year was 12 cases per 100000 visits to the emergency department, with no statistically significant change throughout the study period.
K kingae Bacteremia
The age range of the 42 children with bacteremia was 5 months to 15 years, but only 2 of the of the previously healthy children were older than 18 months. Clinical and laboratory data of the patients are summarized in Table 1. Three children had persistent bacteremia, with 2 positive culture results from blood samples drawn 1 to 4 days apart, before antibiotic treatment.
Eight children had a relevant medical history: chronic renal failure requiring hemodialysis (2), peritoneal dialysis (1), cancer (1), congenital heart disease (2), or failure to thrive (2). All patients who were older than 28 months had a preexisting disease (Fig 1). An acute concurrent disease was documented in 79% of children, including aphthous stomatitis, upper respiratory tract infections, and acute gastroenteritis.
Serum levels of inflammatory markers are presented in Table 1. Only mild-to-moderate elevations of C-reactive protein level and erythrocyte sedimentation rate were found in most patients. Only 5 patients had an absolute neutrophil count >10000 cells per mm3. Of the 16 patients who underwent echocardiographic studies, 4 were found to have endocarditis, including 2 who were previously healthy, 1 with transposition of the great arteries that was repaired in infancy, and 1 with a prosthetic mitral valve. In all 4, the levels of inflammatory markers were significantly higher than in the rest of the bacteremia group: mean white blood cell count was 22085 ± 17697 cells per mm3 (P = .03), C-reactive protein level was 7.6 ± 4.9 mg/dL (P = .005), and erythrocyte sedimentation rate was 84.3 ± 21.4 mm/hour (P = .03). The clinical course of the 2 children with endocarditis and a preexisting cardiac disease was uneventful; however, the 2 previously healthy patients required urgent mitral valve replacement because of destruction of the valve and acute heart failure.
Ten (24%) of the children with positive blood culture results had osteoarticular complaints/signs involving the shoulder (2), ulna (1), left palm (1), hip (2), knee (3), and ankle (1). For 8 children, a bone scan was performed; results were negative in 6 cases and demonstrated mild synovitis in 2 cases (both involved the knees). These joints were not aspirated because clinical findings were absent and ultrasound testing was normal. One patient had overt knee arthritis and underwent a joint aspiration that yielded 17680 white blood cells per mm3, but Gram-stain and synovial fluid culture results were negative. Another child had complaints compatible with hip joint involvement, yet ultrasound examination was normal. Because of a rapid resolution of his symptoms, additional testing was not performed. Thus, we did not identify any case of Kingella osteomyelitis in our cohort.
Initial antibiotic treatment varied and included ceftriaxone, amikacin/gentamicin, ampicillin or penicillin when bacteremia was suspected on presentation; cefazolin or cefuroxime when an osteoarticular infection was suspected; or oral amoxicillin, amoxicillin/clavulanic acid, and cephalexin when occult bacteremia or pneumonia was suspected. The children with chronic health conditions were treated initially with empirical broad-spectrum antibiotic combinations. No child was started on vancomycin or clindamycin alone. There were no treatment failures, because all patients were hospitalized and improved clinically, and repeat blood culture results remained negative. Even the patients who were started on oral β-lactams improved rapidly, with fever subsiding to normal within 1 to 2 days. Children were treated for 7 to 10 days intravenously, except for those with endocarditis, who received a much longer course of 4 to 6 weeks.
K kingae Arthritis
Twenty children had positive K kingae joint fluid culture results, and none had a concomitant positive blood culture result. All were younger than 22 months and were previously healthy. Their clinical and laboratory data are presented in Table 1.
Sixteen (80%) of the K kingae infections in this group were accompanied by another acute disease, including upper respiratory infections, aphthous stomatitis, and acute gastroenteritis. Only large joints were involved: knee (7), ankle (7), hip (5), and shoulder (1). The levels of inflammatory markers were only mildly elevated (Table 1). The joint symptoms resolved within a median of 6 days (range: 2–30). A prolonged course (>3 weeks) was documented for 2 patients with very high C-reactive protein levels at presentation (11.9 and 17.0 mg/dL). No patient had an absolute neutrophil count >10000 cells per mm3.
The median white cell count, performed in 13 joint fluid specimens, measured 130000 cells per mm3 (range: 5200–299000 cells per mm3), with 90% to 96% neutrophils. Although a low synovial fluid cell count is usually not associated with bacterial joint infection, low levels were observed for 3 patients (5200, 15800, and 20600 cells per mm3). Direct Gram staining was negative in all joint fluids. All positive culture results originated from BACTEC bottles and not from culture plates. Patients were treated with intravenous cefuroxime (150 mg/kg per day) or cefazolin (100 mg/kg per day). Children were treated for 14 to 21 days intravenously, with another 2 to 3 weeks of oral antibiotics.
Two patients with hip infections underwent surgical drainage. Recurrent aspirations were performed on 1 patient with a hip infection and 1 with a knee infection.
K kingae Bacteremia Versus Arthritis
Comparison of the 2 patient groups showed that patients with bacteremia were characterized by a significantly longer time from symptom onset to diagnosis and a slightly higher mean temperature at presentation than patients with arthritis (Table 1). There were no between-group differences in age, month of appearance of the infection, or inflammatory marker levels. In addition, no differences were found between groups when we combined the 30 patients with osteoarticular complaints (10 with concurrent bacteremia and 20 with culture-positive arthritis) and compared the clinical and laboratory data with the 32 patients with bacteremia and without osteoarticular symptoms.
Presence of Concurrent Illness
There were no differences in clinical or laboratory parameters between patients with and without a concurrent illness.
Antibiotic Susceptibility of Isolates
Thirty-nine antibiograms were available for analysis. All K kingae isolates were sensitive to β-lactam antimicrobial drugs (except for 1 that was resistant to methicillin), aminoglycosides (except for 1 that was resistant to gentamicin and amikacin), tetracycline, minocycline, colistin, rifampin, erythromycin, quinolones, and trimethoprim-sulfamethoxazole (except 1). All K kingae isolates were resistant to vancomycin, fusidic acid, and clindamycin.
This study investigated the epidemiologic, clinical, and laboratory characteristics of K kingae infection in a large series of children from a major medical center in central Israel. The young age of our patients and the slight male predominance are in accordance with previous data from southern Israel1,7,8; however, most (68%) of the invasive infections were bacteremias, in contrast to the previous studies, in which >60% of the infections were osteoarticular. The lower prevalence of arthritis in our study may be attributable in part to differences in hospital and laboratory protocols regarding drawing and handling blood and synovial fluid cultures. Also in contrast to previous studies from Israel,7 we found no significant disparity in the monthly distribution of the infections and no increase in incidence in the second half of the year. We found a constant rate of infections throughout the 11-year study period.
The shorter time to diagnosis in the arthritis group is probably explained by the higher likelihood that a child with an inflamed joint will be referred to a hospital, compared with a patient with fever and signs of a viral respiratory illness. The higher mean temperature in the bacteremia group is of little clinical significance and may result from the longer time to hospital referral in this group.
Eighty percent of the patients had a concurrent illness that breached the natural mucosal barriers (upper respiratory tract infections, stomatitis, or diarrhea), suggesting that this may be a predisposing factor for invasion by K kingae, which is a common habitant of the normal oropharyngeal flora of children: K kingae was found in the pharynx of up to 45% to 70% of children who attended child care centers.3–5 Nevertheless, despite the high carrier rate, few cases of invasive K kingae infections are identified each year. Presumably, the reason lies in both host and bacterial factors, including low bacterial virulence, slow bacterial growth rate, presence of inhibitory factors in the synovial fluid,12 low concentration of microorganisms in the synovial fluid exudate,6 and the sensitivity of K kingae to common antimicrobial drugs that are administered for other reasons. In addition, the diagnosis of K kingae infection might occasionally be missed because of technical problems with the culture. The low ability of solid culture media to grow K kingae compared with BACTEC bottles is widely known.12 Polymerase chain reaction may aid in identifying microorganisms that infect joints,13,14 and a recent study that used this technique demonstrated that K kingae now seems to be the most common bacterial cause of arthritis in children.15
The resolution of symptoms in the 3 patients with persistent bacteremia (untreated up to 4 days) without overt clinical complications, supports a low virulence of K. kingae. In addition, osteomyelitis and arthritis may resolve spontaneously without antibiotic treatment, as was demonstrated in 4 children with a delayed diagnosis of K. kingae infection.16
Eight (19%) patients with K kingae bacteremia had a significant preexisting medical condition and were considerably older than children who were previously healthy. Thus, the presence of a chronic medical condition, such as renal failure or a prosthetic heart valve, seems to increase the risk for K kingae bacteremia or endocarditis in older children.
The serum levels of inflammatory markers in our patients were relatively low. Previous studies also reported a mild inflammatory response to K kingae infection, with normal or almost normal white blood cell counts, C-reactive protein levels, or erythrocyte sedimentation rates.10,16,17 It is interesting that we found no differences in inflammatory marker levels between the bacteremia and arthritis groups. In addition, there was no apparent effect of the concurrent acute respiratory or gastrointestinal illnesses, which could increase acute-phase reactant levels per se. The low white cell count in 3 joint aspirations suggests that K kingae arthritis can sometimes trigger only a weak local inflammatory response that can be erroneously diagnosed as reactive or viral arthritis, especially in the presence of low systemic inflammatory markers.
K kingae antibiotic susceptibility and resistance in our study were similar to previous reports of resistance to vancomycin and occasionally to clindamycin, fusidic acid, and trimethoprim-sulfamethoxazole.1,18 The K kingae isolate that was resistant to gentamicin is a novel finding.
This study of a large group of 62 children with invasive K kingae infections substantiates previous findings of young age, slight male predominance, and association with mild concomitant illness involving mucosal surfaces. An additional important finding was the mild-to-moderate inflammatory response in most patients that was unaffected by infection site or concurrent illness. Important exceptions were the patients with K kingae endocarditis, 2 of whom had a devastating disease and required urgent heart surgery, and the patients with a prolonged course of arthritis. The possibility of endocarditis with rapid deterioration and native-valve destruction warrants close monitoring of patients with K kingae bacteremia and immediate treatment.
- Accepted March 3, 2008.
- Address correspondence to Gal Dubnov-Raz, MD, MSc, Department of Pediatrics, Mt Scopus, Hebrew University-Hadassah Medical Center, Jerusalem 91120, Israel. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Kingella kingae is being increasingly recognized as an invasive pathogen in children. Only a few, mostly small-scale studies have been published to date. Given the increasing rate of invasive K kingae infections identified, more data from larger studies are needed.
What This Study Adds
K kingae infects older children only in the presence of an underlying medical condition and elicits a mild inflammatory response. K kingae might cause a life-threatening heart disease (10% of bacteremia cases) that requires emergent treatment.
- ↵Kuhls TL. Kingella. In: Feigin RD, Cherry JD, eds. Textbook of Pediatric Infectious Diseases. 4th ed. Philadelphia, PA: WB Saunders; 1998:1495–1499
- ↵Kiang KM, Ogunmodede F, Juni BA, et al. Outbreak of osteomyelitis/septic arthritis caused by Kingella kingae among child care center attendees. Pediatrics.2005;116 (2). Available at: www.pediatrics.org/cgi/content/full/116/2/e206
- ↵Yagupsky P, Dagan R, Howard CB, Einhorn M, Kassis I, Simu A. Clinical features and epidemiology of invasive Kingella kingae infections in southern Israel. Pediatrics.1993;92 (6):800– 804
- ↵Yagupsky P, Peled N, Katz O. Epidemiological features of invasive Kingella kingae infections and respiratory carriage of the microorganism. J Clin Microbiol.2002;40 (11):4180– 4184
- ↵Slonim A, Steiner M, Yagupsky P. Immune response to invasive Kingella kingae infections, age-related incidence of disease, and levels of antibody to outer-membrane proteins. Clin Infect Dis.2003;37 (4):521– 527
- ↵Von Graevenitz A, Zbinden R, Mutters R. Actinobacillus, Capnocytophaga, Eikenella, Kingella, Pasteurella, and other fastidious or rarely encountered gram-negative rods. In: Murray PR, Baron EJ, Jorgensen JH, Pfaller MA, Yolken RH, eds. Manual of Clinical Microbiology. 8th ed. Washington, DC: American Society for Microbiology; 2003:614–615
- ↵Yagupsky P, Dagan R, Howard CB, Einhorn M, Kassis I, Simu A. High prevalence of Kingella kingae in joint fluid from children with septic arthritis revealed by the BACTEC blood culture system. J Clin Microbiol.1992;30 (5):1278– 1281
- ↵Verdier I, Gayet-Ageron A, Ploton C, et al. Contribution of a broad range polymerase chain reacting to the diagnosis of osteoarticular infections caused by Kingella kingae: description of twenty-four recent pediatric diagnoses. Pediatr Infect Dis J.2005;24 (8):692– 696
- ↵Chometon S, Benito Y, Chaker M, Boisset S, Ploton C, Berard J, Vandenesch F, Freydiere AM. Specific real-time polymerase chain reaction places Kingella kingae as the most common cause of osteoarticular infections in young children. Pediatr Infect Dis J.2007;26 (5):377– 381
- ↵Yagupsky P, Katz O, Peled N. Antibiotic susceptibility of Kingella kingae isolates from respiratory carriers and patients with invasive infections. J Antimicrob Chemother.2001;47 (2):191– 193
- Copyright © 2008 by the American Academy of Pediatrics