Chronic non-bacterial osteomyelitis: a comparative study between children and adults

Patients diagnosed with CRMO/CNO or SAPHO syndrome between March 2004 and September 2016 with follow-up of at least 6 months were studied. All patients were diagnosed and treated at the Departments of Pediatrics and Adolescent Medicine, Rheumatology and Immunology, or Orthopedic surgery at our institution. Data, collected anonymously by reviewing electronic medical charts, included

We studied 34 patients (24 pediatric and 10 adult) with a median follow-up of 4.8 years in pediatric and 4.3 years in adult patients (Table 1).

Median age at onset of symptoms was 11.8 years (range 6.3–17.9) in the pediatric group and 28.3 years (18.1–54.2) in the adult group. Median age at diagnosis was 12.3 years (range 7.9–18.9) in the pediatric group and 32.5 years (22–56) in the adult. Median time to diagnosis was shorter in the pediatric group than in the adult (0.3 vs. 1.0 years respectively). In both groups females (58.5% vs. 60%) predominated. The number of physicians consulted before diagnosis was 2.1 (range 1–3) in children and 2.5 [1‐4] in adults. 37.5% of pediatric patients had a family history of chronic immune-mediated diseases in first- or second-degree relatives, including CNO (in a monozygotic co-twin sister), compared with 20% of adult patients.

95.9% of the pediatric and 100% of the adult patients reported pain as a first symptom, accompanied by local swelling in 79.2% of children and 50.0% of adults. Local signs of inflammation (heat, redness) were reported in 50% of children and 10% of adults. None of the patients presented with fever.

Associated diseases were found in 41.5% of pediatric and 50% of adult patients, with skin lesions predominating in both groups (8 children and 4 adults, 33.3% vs. 40% respectively). Severe acne was most frequent in children (7, 29.2%), whereas palmoplantar pustulosis (PPP) was most frequent in adults (2, 20%), (Table 1). Skin disease preceded the diagnosis of CNO in 16% of pediatric patients and 30% of adult patients and appeared after diagnosis of CNO in respectively 16 and 10% of pediatric and adult patients within 2 and 7 years.

Ocular involvement was found in 12.5% of children and 10% of adult patients. One child had coeliac disease.

X-ray images were performed in all patients and demonstrated abnormalities in 69.6% of pediatric patients and 75% of adult patients. MRI of clinically overt lesions was conducted in all pediatric and adult patients. CT imaging was conducted more frequently in adult patients than in pediatric patients (80% vs.12.5%). Imaging techniques used to detect additional lesions differed between children and adults. Whereas whole-body MRI was performed solely in children (21%), technetium bone scans were deployed more frequently in adults (80% vs. 58.5%). Osteolytic lesions were seen in 30.4% of pediatric and 50% of adult patients and hyperostotic lesions in 30.4% of pediatric and 10% of adult patients.

Mean numbers of bone lesions were comparable between the groups, with 3.1 in children (range 1–7) and 3.0 in adults [1‐9]. Unifocal involvement did not differ between pediatric and adult patients (41.7% vs. 40%). Sites of lesions differed, however, with involvement of clavicle (pediatric 41.7% vs. adult 10%) and long-bone metaphyses (pediatric 33.3% vs. adult 10%) more frequent in children. In contrast, involvement of the sternum was more frequent in adults (pediatric 8.3% vs. adult 50%), as was involvement of the vertebrae (pediatric 21% vs. adult 50%). Figure 1 presents the distribution of bone lesions.

Associated arthritis was found in 50% of pediatric and 30% of adult patients; the joints involved mainly adjoined bone lesions (Table 1).

Overlapping features with SpA were observed in 6 pediatric and one adult patient (25% vs. 10% respectively) of whom 5 were males. All patients demonstrated involvement of the sacroiliac joint. Sacroiliitis was present at time of diagnosis of CNO in all 7 patients. Only one of 6 tested patients was HLA-B27 positive. Table 2 shows the characteristics of this subgroup of CNO patients.

Bone biopsies were performed at least once in 19 children (79.4%) and 9 adults (90%) to exclude infectious osteomyelitis and malignancy. Histopathologic results at initial presentation did not differ between the two groups concerning inflammation and fibrosis of the bonemarrow (Table 1).

Results of bacterial culture were available in 17 children and 7 adults. Amplification of 16S ribosomal (r) RNA was performed in 6 children and 2 adults. Mycobacterial PCR was performed in 2 children and one adult. Bacterial pathogens were detected in 3 patients: One biopsy specimen grew Staphylococcus epidermidis and Propionibacterium acnes on enrichment-culture taken from sternal lesion in an adult patient with severe acne. PCR revealed Staphylococcus epidermidis, Neisseria sp., and Streptococcus oralis in one child’s tibial lesion, and Enterobacter klebsiella at low bacterial load in the second rib of an adult. Bacterial culture yielded no organisms in either the tibial or the second-rib lesion. All detected organisms were interpreted as contaminants.

Laboratory tests: 20.8% of pediatric but no adult patient exhibited leukocytosis. 66.7% of pediatric and 50% of adult patients presented with elevated C-reactive protein values > 5 mg/l. 63.7% of pediatric and 57% of adult patients had ESR values > 15 mm/hour (Fig. 2). The HLA-B27 phenotype was found in 3 of 15 tested children (20%) and in none of 5 tested adults; it occurs in 8% of inhabitants of our geographical region. [10]. ANA were demonstrated in 3 of 21 tested children (14.3%), with titers ranging between 1:320–1:5120, but were absent in 3 adults tested.

Starting with a working diagnosis of bacterial infection, 62.5% of pediatric and 55.6% of adult patients received intravenous and/or oral antibiotics with a mean duration of 4 weeks (range 0–25) in children and 5.1 weeks (range 0–26) in adults. Further treatment strategies in children included immobilization with plaster casts in 5, prostaglandin-E infusions in 3, hyperbaric oxygenation in 2, and surgical resection in one. Four adults underwent therapy with prostaglandin-E, gabapentin, surgical resection, or high-frequency radio ablation (one patient each).

Non-steroidal anti-inflammatory drugs (NSAIDs) were the most common first-line treatment, used in 83.3% of pediatric patients and all adults. 75% of children, but only 20% of adults, received NSAIDs continuously. Remission was observed in 3 of 20 children (15%) and in one adult (10%). Systemic steroids were used as a second-line treatment in 29.2% of children and 20% of adults, with transient response in all but one child. Topical steroids were used in one patient from each group (4.2% vs. 10%). Further treatment included disease-modifying anti-rheumatic drugs (DMARDS), which were used in 2 (8.3%) children and 3 (30%) adults. Children were treated with either methotrexate (MTX) or sulfasalazine (SSZ). One adult patient was treated with MTX, a second was treated with SSZ first-line, followed by MTX, and a third was treated with MTX first-line followed by leflunomide and than by cyclosporine A. Remission was observed in only one adult patient, who received MTX. The use of bisphosphonates differed between groups. These drugs were administered in 13 children (54.1%) but only one adult (10%). Bisphosphonate treatment led to remission in 9 of 13 children (69.4%) and in the adult patient. Another difference was seen in biological treatment. Children received tumor necrosis factor (TNF)-α-antagonists twice as often as adults (20.8% vs. 10%) with remissions seen in 3 of 5 children (59.9%), but not in the adult. Table 3.

The course of disease was chronic and remittent in the majority of children (57%), whereas in adults it was predominantly chronic (83%).

At last follow-up visit 62.5% of pediatric patients and 30% of adult patients were in remission. Complications were observed in 33.3% of pediatric patients and 60% of adults. Complications in children included fracture of the vertebra in 2, bone deformity in 1, and chronic pain in 2. Complications in adults were fracture of a vertebra and chronic pain (3 patients each, no overlap). Fracture rates were 8.3% in children and 30% in adults. (Table 3).