Abstract
As the result of improved bacteriological techniques, Kingella kingae is a slow-growing Gram-negative coccobacillus that is emerging as an important cause of spondylodiscitis in children younger than 3 years of age. The high pharyngeal carrier rates of this slow-growing Gram-negative coccobacillus combined with the low incidence of identified K. kingae infections is possibly explained by a low virulence of this bacterium. The use of specific real-time polymerase chain reaction (PCR) on blood samples and throat swabs opens new prospects in the bacteriological investigations of young children suspected to have spondylodiscitis, an approach that could prevent, in the future, unnecessary invasive interventions.
Introduction
Childhood spondylodiscitis remains an uncommon and often missed ailment in the non-talkative toddler from 1 to 3 years of age. It should be considered as a diagnosis in children who present with a refusal to walk, gait disturbances, back pain or abdominal pain. Laboratory findings, such white blood cell count (WBC), C-reactive protein (CRP) and erythrocyte sedimentation rate (ESR), often provide non-specific information [1]. Blood cultures are often the only means available to select the antimicrobial therapy, but, surprisingly, show a high percentage of negative results (88–100%) [1–3]. The indication for more invasive procedures such as biopsy or needle aspiration is currently not established. The literature shows success rates for needle aspiration and open biopsy in identifying the causative organism that ranges from 0 to 63% [3–6]. However, due to the surgical and anaesthetic risks, these interventions are still not standard diagnostic procedures for most authors. Since the 1980s, the reported number of cases of Kingella kingae osteoarticular infections [7–13], including spondylodiscitis [5, 14, 15], has markedly increased, mainly due to improvements in culture techniques (inoculation of blood culture vials with bone/joint aspirates) [13, 16] and to the use of molecular methods [17–21]. Currently, we believe that most cases of spondylodiscitis in children less than 3 years of age are caused by K. kingae. Thus, the purpose of this communication is to report the first encouraging results of a new approach to the bacteriological diagnosis of spondylodiscitis in those children, enabling us to avoid unnecessary invasive procedures.
Materials and methods
After approval from the Children's Hospital Review Board (09-029R, Mat-Ped 09-008R), we prospectively started recruiting all children under 3 years of age admitted to our institution for spondylodiscitis since January 2009. The diagnosis of spondylodiscitis was confirmed by magnetic resonance imaging (MRI). Biological evaluation included peripheral WBC, left-band shift percentage, platelets counts, CRP levels and ESRs. Bacteriological investigations included blood cultures and K. kingae-specific polymerase chain reaction (PCR) assays from peripheral blood and throat swabs. The first two cases are reported here as representative examples of this new technique.
Microbiological methods
Since 2007, we have been using in our hospital a novel real-time PCR assay that targets the RTX toxin gene [22]. The assays were designed to detect two independent gene targets from the K. kingae RTX toxin locus, genes rtxA and rtxB. Primers and probes were created using Primer Express (version 2.0; Applied Biosystems) on the basis of the K. kingae strain 269-492 RTX gene locus (accession number EF067866) [22]. Basic Local Alignment Search Tool (BLAST) searches across a public database failed to reveal any significant cross-homology. TaqMan Universal PCR Master Mix with AmpErase UNG (Applied Biosystems) was used with 0.5 μl of input DNA and nuclease-free water (Promega). Each PCR analysis was done twice. A PCR-positive sample collected from a culture-proven K. kingae sample was used as a positive control. To determine the specificity of the qPCR assays, we tested the DNA extracted from related species and organisms that are frequently encountered in osteoarticular infections; qPCR assays showed no cross-reactivity with common osteoarticular pathogens [22]. To evaluate the analytical sensitivity of the rtxA and rtxB qPCR assays, we used a range of serially diluted K. kingae genomic DNA (3 × 106–3 × 10−2 cfu). The assay exhibited a sensitivity of 30 cfu, which is ten times more sensitive than a previously published semi-nested broad-ranged 16S rRNA gene PCR assay [22].
Case 1
A 17-month-old toddler arrived in the emergency department with a 4-day history of fever up to 38.5°C, inability to walk and sit up, and pain during diaper change. On examination, she had no fever, but presented lower lumbosacral pain on palpation and on lumbar spine flexion. The neurological examination was otherwise normal. Laboratory tests revealed a serum WBC of 7,300 leucocytes/mm3 (normal range 6,000–17,000 leucocytes/mm3), a CRP level less than 10 mg/l (normal <10 mg/l) and an ESR of 16 mm/h (normal range 0–10 mm/h). On conventional X-rays, there was only a slightly decreased height of the disk space. An MRI was performed, which demonstrated loss of normal T2 hyperintensity in the L4–L5 disc with adjacent vertebral body T2 hyperintensity, indicative of bone marrow oedema. In addition, abscess of the L5 vertebral body and a non-compressive spinal subdural abscess were noted. These findings were suggestive of discitis and osteomyelitis. Classical blood culture was negative, but K. kingae-specific qPCR assays were positive in peripheral blood and throat swab. The patient responded well to intravenous cefuroxime and body plaster cast, and improved durably after the end of antibiotic treatment.
Case 2
A 13-month-old girl was admitted to our hospital with a 20-day history of back pain and inability to ambulate, sit up and crawl. On examination, she had no fever, but presented lower lumbosacral pain on lumbar spine flexion; the neurological examination was normal. Laboratory tests revealed a serum WBC of 14,000 leucocytes/mm3 (normal range 6,000–17,000 leucocytes/mm3), a CRP level of 26 mg/l (normal <10 mg/l) and an ESR of 60 mm/h (normal range 0–10 mm/h). The conventional X-rays demonstrated a decreased height of the disk space with erosions of adjacent vertebral endplates. The MRI investigation was very similar to those of the previous child, with loss of normal T2 hyperintensity in the L4–L5 disc, adjacent vertebral body T2 hyperintensity and bone abscesses. Classical blood culture was negative but K. kingae-specific qPCR assays were positive in peripheral blood and throat swab. The patient responded well to intravenous cefuroxime and body plaster cast, and improved durably after the end of antibiotic treatment.
Discussion
Childhood spondylodiscitis may affect the entire spine, but it is more common in the lumbar region in children less than 3 years of age. Currently, the aetiology appears to be a bacterial infection, usually caused by low-virulent and atypical organisms like K. kingae. The pathogenesis of K. kingae was recently attributed to the production of a potent cytotoxin (RTX) that may play a crucial role in the colonisation of the respiratory tract, in the invasion of the bloodstream and in damaging bones and joints [23]. The disease seems to be self-limiting in most cases [6] and, therefore, the indication for more invasive procedures, such as biopsy or needle aspiration, is currently not established. In spondylodiscitis, the inoculation usually occurs by haematogenous spread to the disc and then to the vertebral body [2]. Studies of vertebral body and disc vascular anatomy have demonstrated that there is, in the young child, an important network of vessels that go through the cartilaginous vertebral endplates and enters the annulus [24]. The venous drainage of the vertebral column is ensured by rich venous vertebral plexuses and by the basivertebral veins [25]. For this reason, it is surprising that blood cultures are not more often positive in children with spondylodiscitis. One can, therefore, legitimately ask the question of whether these spondylodiscitis are not due to atypical organisms that are difficult to isolate on conventional agar media. In these two reported children, specific qPCR assays were positive using whole blood and from a throat swab, thus, suggesting the diagnosis of lumbar spondylodiscitis caused by K. kingae. The presence of the RTX cytotoxin of K. kingae in the bloodstream was interpreted in our cases as the cause of the osteoarticular damage, even if we could not prove formally that the spondylodiscitis was actually caused by K. kingae, as we did not have specific culture material from the spine. These first results encourage us to pursue our study on specific K. kingae PCR assays in the blood and throat, in order to prove the responsibility of this organism for spondylodiscitis in young children. However, further studies will have to correlate the simultaneous existence of positive qPCR for K. kingae in the blood, in the throat and in the disk needle aspiration. At the same time, we are currently studying the throat carrier rate of K. kingae RTX+ among healthy children less than 3 years of age and those with upper airway infection. We hope that a widespread use of specific qPCR in the blood and on throat swabs may, in the future, decrease the indication of disk needle aspiration.