Video

Efforts toward producing CNO/CRMO classification criteria show first results


 

REPORTING FROM EULAR 2019 CONGRESS

Surveys and consensus techniques have been instrumental in identifying much needed candidate criteria toward the classification of chronic nonbacterial osteomyelitis (CNO), according to recent findings from international surveys of pediatric rheumatologists that were presented at the European Congress of Rheumatology.

Melissa Oliver, MD, a pediatric rheumatologist at Riley Hospital for Children, Indianapolis, and colleagues recently undertook the multiphase study as part of an international collaborative effort led by the Childhood Arthritis and Rheumatology Research Alliance to establish consensus-based diagnostic and classification criteria for CNO, an autoinflammatory bone disease of unknown cause that primarily affects children and adolescents. CNO is also known as chronic recurrent multifocal osteomyelitis (CRMO). If this disease is not diagnosed and treated appropriately in a timely fashion, damage and long-term disability is possible. In the absence of widely accepted, consensus-driven criteria, treatment is based largely on expert opinion, Dr. Oliver explained in an interview.

“There is an urgent need for a new and more robust set of classification criteria for CRMO, based on large expert consensus and the analysis of a large sample of patients and controls,” she said.

There are two proposed diagnostic criteria, the 2007 classification of nonbacterial osteitis and the 2016 Bristol diagnostic criteria for CRMO, but both are derived from single-center cohort studies and have not been validated, Dr. Oliver explained.

The list of candidate items that have come out of the study is moving clinicians a step closer toward the design of a practical patient data collection form that appropriately weighs each item included in the classification criteria.

The study employed anonymous survey and nominal group techniques with the goal of developing a set of classification criteria sensitive and specific enough to identify CRMO/CNO patients. In phase 1, a Delphi survey was administered among international rheumatologists to generate candidate criteria items. Phase 2 sought to reduce candidate criteria items through consensus processes via input from physicians managing CNO and patients or caregivers of children with CNO.

Altogether, 259 of 865 pediatric rheumatologists (30%) completed an online questionnaire addressing features key to the classification of CNO, including 77 who practice in Europe (30%), 132 in North America (51%), and 50 on other continents (19%). Of these, 138 (53%) had greater than 10 years of clinical practice experience, and 108 (42%) had managed more than 10 CNO patients.

Initially, Dr. Oliver and colleagues identified 33 candidate criteria items that fell into six domains: clinical presentation, physical exam, laboratory findings, imaging findings, bone biopsy, and treatment response. The top eight weighted items that increased the likelihood of CNO/CRMO were exclusion of malignancy by bone biopsy; multifocal bone lesions; presence of bone pain, swelling, and/or warmth; signs of fibrosis and/or inflammation on bone biopsy; typical location of CNO/CRMO lesion, such as the clavicle, metaphysis of long bones, the mandible, and vertebrae; presence of CNO/CRMO–related comorbidities; normal C-reactive protein (CRP) or erythrocyte sedimentation rate (ESR); and typical MRI findings of CNO/CRMO.

By phase 2, candidate items, which were presented to 39 rheumatologists and 7 parents, were refined or eliminated using item-reduction techniques. A second survey was issued to 77 of 82 members of a work group so that the remaining items could be ranked by their power of distinguishing CNO from conditions that merely mimicked the disease. The greatest mean discriminatory scores were identified with multifocal lesions (ruling out malignancy and infection) and typical location on imaging. Normal C-reactive protein and/or an erythrocyte sedimentation rate more than three times the upper limit of normal had the greatest negative mean discriminatory scores.

The next steps will be to form an expert panel who will use 1000minds software to determine the final criteria and identify a threshold for disease. The investigators hope to build a large multinational case repository of at least 500 patients with CNO/CRMO and 500 patients with mimicking conditions from which to derive a development cohort and an external validation cohort. So far, 10 sites, including 4 in Europe, have obtained approval from an institutional review board. The group has also submitted a proposal for classification criteria to the American College of Rheumatology and the European League Against Rheumatism, Dr. Oliver said.

Dr. Oliver had no disclosures to report, but several coauthors reported financial ties to industry.

SOURCE: Oliver M et al. Ann Rheum Dis. Jun 2019;78(Suppl 2):254-5, Abstract OP0342. doi: 10.1136/annrheumdis-2019-eular.1539.

Next Article: