The Epidemiology of Pediatric Bone and Joint Infections in Cambodia, 2007–11

There are limited data on osteoarticular infections from resource-limited settings in Asia. A retrospective study of patients presenting to the Angkor Hospital for Children, Cambodia, January 2007–July 2011, identified 81 cases (28% monoarticular septic arthritis, 51% single-limb osteomyelitis and 15% multisite infections). The incidence was 13.8/100 000 hospital attendances. The median age was 7.3 years, with a male/female ratio of 1.9:1; 35% presented within 5 days of symptom onset (median 7 days). Staphylococcus aureus was cultured in 29 (36%) cases (52% of culture-positive cases); one isolate was methicillin-resistant (MRSA). Median duration of antimicrobial treatment was 29 days (interquartile range 21–43); rates of surgical intervention were 96%, and 46% of children had sequelae, with one fatality. In this setting osteoarticular infections are relatively common with high rates of surgical intervention and sequelae. Staphylococcus aureus is the commonest culturable cause, but methicillin-resistant S. aureus is not a major problem, unlike in other Asian centers.


Material and Methods
Cases from January 2007-July 2011 were identified through hospital and laboratory records. Data collected included age, gender, residence, admission details, comorbidities, clinical presentation, laboratory investigations, surgical interventions, antibiotic treatment and outcomes. Weight-for-age Z-scores were calculated for children aged <5 years [24].
Disease episodes were defined as single treatment episodes, and relapses as recurrent disease episodes following improvement. Monoarticular septic arthritis described single joint infections; single-limb osteomyelitis included adjacent infected long bones within a limb and bone-plus-adjacent-joint involvement, and multisite infections included non-adjacent sites of infection, and/or non-musculoskeletal sites. Mandibular/foot infections were considered separately.
Data were analyzed using Stata 11.1 (StataCorp, TX, USA). Fisher's exact and Kruskal-Wallis tests were used for comparisons between groups for categorical and continuous variables, respectively. Multivariable logistic regression was used to assess risk factors for binary outcomes.
Ethical approval was granted by the AHC Institutional Review Board and the Oxford Tropical Research Ethics Committee, UK.

Results
Of 81 patients, 60 (74%) had a single episode of osteoarticular infection, and 21 had primary episodes followed by relapse(s). The median age (range) was 7.3 years (0-14); boys were almost twice as commonly afflicted. Trauma was seen in 56% of cases, 41% of which was penetrating. Where vaccination status was known, 52 (74%) children had received age-appropriate vaccinations. Five of 23 (22%) under-5 s were moderately to severely undernourished.
For primary episodes, details across clinical categories and symptoms are represented in Tables 1  and 2. There were no significant differences in white cell count (WCC), ESR or CRP between clinical groups. Crude incidence for non-relapses was 13.8/100 000 attendances; temporal changes by clinical group are depicted in Fig. 1. Median follow-up time (interquartile range [IQR]; range) was 28.5 days (0-140; 0-1339).
Many patients sought treatment before attendance at our institution: 48% had been reviewed in a private clinic and/or by traditional healers, and 24% had taken antibiotics.

Monoarticular septic arthritis
Forty-eight percent of cases presented within 5 days of symptom onset; 68% had fever, joint pain and decreased movement. The hip was involved in 43%, and the knee in 39% of cases (Fig. 2). Microbiological results are presented in Table 3.
Aspiration/drainage was performed in 21 of 22 cases, with 55% having multiple aspirations; five cases required subsequent arthrotomies. Six had arthrotomies without preceding aspiration. The median (IQR) duration of antimicrobial treatment was 27 days (10-29), with 10 days of intravenous treatment. Neither hip joint involvement nor arthrotomy was associated with sequelae (p ¼ 0.66, 1.0).
A higher mean admission WCC was associated with acute presentations (21.1 vs. 11.7 Â 10 9 /l; p < 0.001). Mean values for WCC, ESR and CRP were non-significantly higher in those with sequelae.  Table 4), one with Haemophilus influenzae and two with mixed infections (one S. aureus/Escherichia coli; one S. aureus/beta-hemolytic streptococcus). No methicillin-resistant S. aureus (MRSA) was isolated. One case was positive for acid-fast bacilli on microscopy.
All but two patients had surgery. The median duration of antimicrobial therapy (IQR) was 30 days (25-43) with 12 days of intravenous therapy.
Both mandibular cases were associated with poor dentition; one had health care-associated MRSA osteomyelitis. The three cases with foot osteomyelitis all occurred post-puncture wounds.

Multisite infections
Multisite infections were invariably associated with S. aureus. They included most cases (7/8) admitted to intensive care and the only fatality.

Discussion
This study is one of the larger, recent, single-center series of pediatric osteoarticular infection, and the only study, to our knowledge, from Cambodia.
Staphylococcus aureus was the most commonly cultured pathogen, but MRSA was not a major problem, unlike in other regional studies [13,14]. Six percent of cases cultured H. influenzae, demonstrating the importance of empirical cover for this in the absence of adequate vaccination.  Demographic and microbiological details for sample-positive monoarticular septic arthritis cases (n ¼

13)
Patient characteristics More than 95% of patients required surgical intervention, and rates of infectious sequelae were high, although we did not find any associations with published risk factors for poor outcomes [15,26]. Delays in presentation and inappropriate preliminary treatment are likely to be contributing factors. A further concern is the use of low-dose/low-frequency oral antibiotic regimens in treatment to improve compliance. Studies of suitable regimens in our setting are needed.
Study limitations include incomplete follow-up; this and the sample size have made it unfeasible to model risk factors for relapse. Diverse treatment approaches were used, making it difficult to identify optimal regimens. We did not look for Kingella kingae, which is an important pathogen elsewhere [27,28]. Nevertheless, we demonstrate relevant features of these infections in Cambodian children, which can be used as a baseline for modifications to treatment approaches and the monitoring of epidemiological trends.  Cases of mandibular (n ¼ 2), calcaneal (n ¼ 2) and metatarsal osteomyelitis (n ¼ 1) have been excluded from this analysis. a Numbers in square brackets denote status at last visit; however, further follow-up was planned for these two patients beyond the study end period.