Bone and Joint Infections in Children and Adolescents in Luanda, Angola

: We reviewed the characteristics of children hospitalized for bone and joint infections in Luanda, Angola. In a retrospective chart review of 45 patients with childhood osteomyelitis or septic arthritis, 51% of the patients had sickle cell disease, and these patients presented with lower hemoglobin and needed blood transfusion more frequently ( p < 0.05). Out of all patients, 64% underwent surgical procedures; a pathological fracture occurred in 31% of the patients.


Introduction
Unless diagnosed and treated promptly, bone and joint infections in children are devastating and even fatal diseases that especially in resource-poor settings cause serious and long-lasting sequelae [1]. In children, osteomyelitis (OM) and septic arthritis (SA) are most often hematogenous [2]. Common causative agents in tropical regions include S.aureus, respiratory pathogens, and Salmonella s.p.p. [2,3]. The prognosis is generally favorable in acute cases if the antibiotic treatment is started promptly [1]. Potential sequelae include chronic osteomyelitis, pathological fracture, sequestra, and growth disturbance [1,2].
The global disease burden of sickle cell disease (SCD) is highest in sub-Saharan Africa. The birth prevalence of SCD is 1125 per 100,000 compared with 43 per 100,000 in Europe [4,5]. Children with SCD are prone to osteomyelitis [1,2]. A special problem is to distinguish osteomyelitis from a vaso-occlusive crisis [6,7]. Bone and joint infections are the most common musculoskeletal complication in children under 10 years old with SCD, surpassed in adolescence only by malleolar ulcers [8]. SCD patients with hematogenous bone or joint infection are traditionally deemed to require surgery more frequently and are believed to experience complications such as epiphysiodesis or joint destruction more often than patients without SCD. Few trials have studied the efficacy and safety of antibiotic treatment among patients with SCD [9]. Besides bone and joint infections, patients with SCD suffer from osteonecrosis and bone infarction [10]. Bone infarction among SCD patients is significantly more common that osteomyelitis, which only covers 1-2% of musculoskeletal complaints of SCD patients [11]. Unfortunately, the differential diagnosis between osteomyelitis and bone infarction in SCD is challenging [10,11].
Distinguishing osteomyelitis from bone infarction in patients with SCD is a challenge even for a multidisciplinary team of pediatric infectious disease specialists, hematologists, and orthopedic surgeons. Body temperature (>38 • /100.4 • F for infection and <38 • /100.4 • F for infarction), leukocytosis, and C-reactive protein levels may be used in the differential diagnosis. Fever is rare in infarction, but mildly elevated inflammatory markers may cause confusion [10]. X-rays are often normal. Bone scan and MRI are specific [10,12], but poor availability limits their use in resource-poor settings. Pain at multiple sites has been thought to be more characteristic of bone infarction than osteomyelitis.
The treatment of bone infarction in patients with SCD is supportive with hydration and analgesics whereas prompt antibiotic management and possibly invasive procedures are required in osteomyelitis. Every effort should be undertaken in the differential diagnosis. Knowledge of the pattern of disease of osteomyelitis is useful in this respect, but still very few publications have reported the characteristics of osteomyelitis in resourcepoor settings.
The aim of this study is to characterize the features of osteomyelitis in Luanda, Angola, with special emphasis on the special challenges that the clinicians face in a sub-Saharan African setting, such as severe cases of osteomyelitis in patients without SCD, osteomyelitis in SCD patients, and difficulties in distinguishing osteomyelitis from vaso-occlusive crisis or osteonecrosis in patients with SCD.

Materials and Methods
In this retrospective chart review, the case records of all patients at age 0 to 15 years admitted to the Department of Surgery in David Bernandino Children s Hospital, Luanda, Angola were systematically searched from January 1st to December 31st 2014 to identify patients diagnosed with bone and joint infections (the main disease entity was osteomyelitis or septic arthritis). We collected demographic, laboratory, and radiographic data regarding the treatment and outcome. No universal screening for SCD could be applied during this period. Only if sickle-cell anemia was suspected by the attending clinician was hemoglobin electrophoresis used to identify cases of sickle-cell anemia. A total of 45 bone and joint infections (the main disease entity was osteomyelitis in 40 cases and septic arthritis in 5 cases) were identified. Of the patients, 26 (58%) were males and 19 (42%) females. The mean and median age was 6 years and the age of patients included in the study varied from 1 month to 13 years. Median age of patients with and without sickle-cell disease were 6 years (range 1 month to 13 years) and 3 years (range 1 year to 13 years) Malnutrition was defined as weight for age z-score < 2. Children with and without SCD were compared to identify potential difference in clinical findings, blood analysis, duration of fever, number of surgical procedures, and prognosis. Skin traction was used in small children, and skeletal traction was used in older children and adolescents. Statview ® (version 5.0.1., Abacus Corporation, Baltimore, MD, USA) was used in the data analysis. An unpaired T-test was used to calculate p-values and p < 0.05 was considered significant. We were not able to perform preset follow-up for the patients, so the data analysis consists of inpatient data.

Results
Some cases were multifocal; in all, 61 bones and joints were involved. Mean and median duration of illness (identified in 31/45 patients) were 39 and 14 days, respectively (range 7 days to 1 year). No patient went to the hospital in under 7 days from the onset of symptoms and signs. Of them, 42% (19/45) presented within 2 weeks. A total of 13% (6/45) of the patients were malnourished.
Of them, 7% (3/45) of the patients arrived to hospital with an altered level of consciousness, and 18% (8/45) were dyspneic. None of the patients had convulsions on admission. The rough assessment of the attending clinician regarding the general condition was that it was poor in 13% (6/45). Three patients (6%) had edema at the ward. Median (interquartile range, IQR) blood hemoglobin, erythrocyte sedimentation rate (ESR), and white blood cell count (WBC) were 5.7 g/dL (5.0), 115 mm/h (61), and 15,400/mm 3 (10,600), respectively. A total of 87% (39/45) of the patients had changes in X-ray. Potential discharge from the focus was cultured in only five cases, the yield being 2 cases of Staphylococcus aureus and isolated cases of Klebsiella, Enterobacter, and Citrobacter spp. A total of 25 patients were tested for HIV, but all had negative serology.
SCD was screened in 60% (27/45) of patients and the disease was found in 85% of the screened patients (23 patients). A comparison of the SCD patients versus those without SCD found is given in Table 1. There was no significant difference between groups, except that the patients with SCD presented with significantly lower hemoglobin (p < 0.05), and significantly more likely required blood transfusion (p < 0.05). Furthermore, 74% (14/19) of the children who attended with a history of less than 2 weeks had SCD.  SCD was screened in 60% (27/45) of patients and the disease was found in 85% of the screened patients (23 patients). A comparison of the SCD patients versus those without SCD found is given in Table 1. There was no significant difference between groups, except that the patients with SCD presented with significantly lower hemoglobin (p < 0.05), and

Discussion
This report highlights the severity of bone and joint infections in children in a sub-Saharan setting. Recent studies have shown that acute osteomyelitis in previously healthy non-neonatal children in a high-income setting can be treated with a short antibiotic treatment (a total course of 3 weeks in osteomyelitis, 2 weeks in septic arthritis) [13,14]. The recovery is fastest if the patient presents within 5 days of the onset of symptoms, this window obviously being the "golden period" of the treatment of osteomyelitis [14].
Unfortunately, none of the patients in this series presented within 7 days from the disease onset, and we assume that at least in part because of this reason, the recovery was rather slow, as over 80% of patients were still febrile after one week of treatment.
Empirical antibiotic treatment is commonly given to children with SCD suffering from bone pain of unknown cause [15]. Understanding the pattern of symptoms in osteomyelitis may help to better target the treatment. The clinical features and findings used to differentiate osteomyelitis from a vaso-occlusive crisis in SCD patients are fever (> 38.0 • C) and pain in a single area in the diaphysis of a long bone [16]. In our series, fever was present in 8-9 out of 10 patients. Multiple bone involvement was however not that uncommon a finding, as seen in Figure 1.
Patients with SCD probably presented with low hemoglobin and thus needed blood transfusion more often. In patients with major sickling hemoglobinopathies, Salmonella is the most common pathogen, causing osteoarticular infections in the USA and Europe, whereas S.aureus is the leading pathogen in sub-Saharan Africa and the Middle East [17]. However, the prevalence of different pathogens and their sensitivity pattern to antibiotics are not well known in sub-Saharan Africa. Our study did not much illuminate this problem, as only 5 cases were appropriately analyzed bacteriologically. Moreover, without bacterial cultures an aseptic osteonecrosis may be mistaken for an active infection, and this was a clear limitation when interpreting the results. Because of this we have started using PCR technique for the bone and joint samples. We also have used this method to identify the agents causing bacterial meningitis from cerebrospinal fluid impregnated in the filter paper strips which were mailed to Finland for bacteriology [18].
If acute osteomyelitis is treated early, fractures are extremely rare, because they usually associate with delayed diagnosis or resistant pathogens, such as methicillin-resistant S. aureus [1,18]. Approximately one third of our patients had a fracture, and in other settings in Africa pathological fractures are also known to be commonly found in chronic osteomyelitis [19].
Our study has limitations. PCR was not used to identify the causative agent. The analysis was retrospective, and the diagnostic criteria for osteomyelitis and septic arthritis could not be pinpointed; we had to rely on the clinical symptoms, laboratory analysis, and X-rays. Our study compared children from a diverse age range that differ significantly in terms of bone structure and stage of cognitive development. The pathogenesis of osteomyelitis changes as the prevalence of hematogenous etiology declines after growth plate closure [1]. Thus, we chose to include all under 15-year-olds, realizing the developmental diversity of this group.
The patients already had a long duration of illness on admission. This reflects the reality in a sub-Saharan Africa with a significant delay in patient presentation. Moreover, the duration of symptoms was recorded as reported by the parents, which may lead to underestimation. All patients were not systematically evaluated for sickle cell disease, but targeted diagnostic assessment such as electrophoresis was available if hemoglobinopathy was suspected. Systematic screening for SCD might have increased detection, so the prevalence of SCD reported in this series should be considered as the lower margin. Unfortunately, we were not able to arrange a prescheduled follow-up. Thus, we are unable to formulate guidelines regarding the optimal diagnostic method or treatment.
In conclusion, we noted that in a significant minority of osteomyelitis, there is absence of fever or multifocal symptoms, making the differential diagnosis of osteomyelitis and bone infarction in patients with SCD difficult. It is unsurprising that a large proportion of patients suffering from SCD require blood transfusion during hospitalization, as patients with SCD present with low hemoglobin. Treatment guidelines cannot be formulated due to the lack of routine long-term follow-up. More research regarding the diagnosis and treatment of osteomyelitis in resource-scarce settings is urgently needed.