Abstract
Objective Chronic nonbacterial osteomyelitis (CNO) is a rare autoinflammatory bone disease that is gaining recognition from clinicians and researchers. We aim to publish data from our cohort of patients with CNO living in the northwestern United States to increase the awareness of specific demographics, characteristics, and presentation of this rare disease.
Methods A retrospective chart review was performed of our electronic medical records. Patients with complete chart records who met criteria for a diagnosis of CNO from 2005 to 2019 were included. Extracted data including patient demographics, bone biopsy results, and lesion locations on advanced imaging were analyzed. King County census data were used to calculate the annual new case rate within our center.
Results A total of 215 CNO cases were diagnosed at our large tertiary pediatric hospital. The majority of cases were of White race residing in Washington’s most populous county, King County. Most cases were diagnosed in 2016 to 2019, showing a significant increase in the annual case rate from 8 to 23 per million children in King County, though there did not appear to be a seasonal predilection. Biopsy rate decreased from 75% to 52%. One hundred fifty-two (71%) children had family history of autoimmunity. With increasing use of whole-body magnetic resonance imaging (WB-MRI), results showed 68% had multiple lesions.
Conclusion CNO has been diagnosed at an increased rate in recent years. WB-MRI may assist in identifying other lesions that may be asymptomatic on presentation. Bone biopsy is still required in some children at the time of diagnosis.
Chronic nonbacterial osteomyelitis (CNO), also known as chronic recurrent multifocal osteomyelitis (CRMO), is an autoinflammatory bone disease that mainly affects children, with an average age of onset around 9 years.1,2,3,4,5The ratio of females to males is about 2:1.1,2,3,4,5 Due to the low awareness among the medical community and waxing and waning pain patterns, the diagnosis of CNO is often delayed or missed. The incidence rate of CNO has been reported to be as rare as 4 in every 1,000,000 children between 2006 and 2008 in Germany.6 However, previous reports showed that CNO had similar incidence rates as bacterial osteomyelitis in children.7,8 The incidence rate of CNO in the US remains unknown. Thus, we conducted a retrospective study within a single large pediatric referral center with unique catchment to determine the annual case rate and demographics of CNO cases within our region. In addition, the initial distribution of bone lesions within the skeletal system was explored.
METHODS
Inclusion and exclusion criteria. Institutional review board approval (#2375) was obtained from the authors’ tertiary, multidisciplinary pediatric hospital prior to the study. Consent was waived due to the study’s retrospective nature. Children with CNO were identified by evaluating the electronic medical database (launched before 2005) using standard International Classification of Diseases, 9th revision, Clinical Modification (ICD-9-CM; codes 730.1 or 730.2 for chronic and unspecified osteomyelitis); ICD-10-CM codes (M86.30 for chronic multifocal osteomyelitis, unspecified site); chief complaints of bone pain; CNO; or CRMO for clinic visits from January 1, 2005, to December 31, 2019. Inclusion criteria were (1) age < 18 years at the diagnosis of CNO; and (2) residence within the Washington (WA), Alaska (AK), Montana (MT), Idaho (ID; WAMI) region. Exclusion criteria were (1) patients who moved to WAMI region with a preexisting diagnosis made outside of Seattle Children’s Hospital; and (2) incomplete medical record for a provider to confirm the diagnosis.
Other regional providers. Within WA, Seattle Children’s Hospital was the only pediatric rheumatology group until 2015. Regional clinics were available in Seattle (WA), Bellevue (WA), Tri-Cities (WA), Missoula (MT), Billings (MT), and Anchorage (AK). Madigan Army Medical Center (Pierce County, WA) hired a pediatric rheumatologist in 2015 and the service population was limited to children from military base. Mary Bridge Hospital (Pierce County, WA) hired a pediatric rheumatologist in 2016. A pediatric rheumatologist was hired at St. Luke in Boise (ID) in 2018. One of the coauthors (YZ) joined Seattle Children’s Hospital in July 2013 and made efforts to raise awareness of CNO in the community.
Case confirmation and chart extraction. A board-certified pediatric rheumatologist (YZ) reviewed the diagnoses of CNO using clinical, laboratory, imaging, and pathological data based on the following criteria: presence of multifocal bone lesions on imaging with typical features (sclerosis and/or hyperostosis on radiograph, or hyperintensity within bone marrow on magnetic resonance imaging [MRI]) at typical sites (metaphysis of long bones, pelvic bones, vertebrae, clavicle, and mandible) or absence of malignancy and infection on bone biopsy if only 1 lesion was present or atypical features/sites were seen in cases with multifocal lesions.2,9,10 Patients with severe cytopenia defined as white blood cell count < 3 × 103/mm3, hemoglobin < 10 g/dL, or platelet < 100 × 103/mm3; low alkaline phosphatase; or vitamin C level were not included. Demographic data including zip code, age at diagnosis, age at onset, month of onset, month of diagnosis, whether a bone biopsy was performed, and affected bone sites on MRI were extracted.
Statistics. Baseline characteristics were summarized using descriptive statistics. Annual case rate was calculated by averaging data between 2016 and 2019 and adjusting to the census data of Washington state in 2010 and 2018. The mean number of cases per year within King County between 2 intervals (2005-2015 vs 2016-2019) was compared using nonparametric tests through IBM SPSS Statistics for Windows (IBM Corp). The cut-off of 2016 was determined by the year when whole-body imaging was made accessible for all patients with CNO. The seasonal distributions and months of disease onset, geographic distribution of patients, as well as the anatomical locations of initially affected bone sites were analyzed using Tableau (Tableau Software Inc). Chi-square analysis was used to compare the distribution of cases across seasons. A P value < 0.05 was considered statistically significant.
RESULTS
Epidemiology. The total number of cases between 2005 and 2019 identified through various sources was 244. Twenty-nine were excluded due to 1 case of incorrect diagnosis, 1 case of incomplete record, 2 cases diagnosed prior to 2005, and 25 cases diagnosed in children who resided outside of WAMI area. There were 215 CNO cases within the WAMI area diagnosed at Seattle Children’s Hospital. There were approximately 9 CNO cases newly diagnosed per year between 2005 and 2015, and 30 new cases per year between 2016 and 2019 within WAMI region (P < 0.001). The annual number of new cases from King County and WAMI area are shown in Figure 1. Within King County where Seattle Children’s Hospital resides, 3 CNO cases per year between 2005 and 2015 and 10 cases per year between 2016 and 2019 were newly diagnosed. Based on the census data from the US Census Bureau in 2010 and 2018, children aged < 18 years in King County made up 20% of the total population of 1,931,249 or 2,233,163, respectively (www.census.gov/quick-facts/kingcountywashington). The population of King County increased by 11% (roughly 1.7 to 1.9 million) between 2000 and 2010 and 13% (roughly 1.9 to 2.2 million) between 2010 and 2018. Therefore, the estimated average annual new case rate increased from 8 (years 2005–2015) to 23 (years 2016–2019) per million children.
Increase of annual number of new cases of chronic nonbacterial osteomyelitis within Pacific Northwest region (dashed line) and King County (solid line) between 2005 and 2019.
Overall demographics, family history, and associated conditions. Among 215 patients, 188 (87%) were from Washington state and 73 (34%) from King County. Within the WAMI area from 2005 to 2015, the median ages of onset and diagnosis were 9.9 (IQR 7.6–11.9) years and 11.0 (IQR 8.4–12.7) years, respectively (Table 1). This has remained essentially unchanged during the years of 2016 to 2019, with median ages at onset and diagnosis being 9.4 (IQR 7.0–11.6) years and 10.5 (IQR 8.0-12.6) years, respectively. Out of the 215 subjects in the study, 111 children (52%) were girls. In terms of racial identification, 166 (77%) subjects identified as White, 8 (4%) as Black, 7 (3%) as Asian or Pacific Islander, 2 (1%) as Native Americans, 12 (6%) as other, and 20 (9%) did not wish to indicate. Time to diagnosis increased from a median of 4.9 (IQR 1.8–14.9) months in 2005–2015 to 7.7 (IQR 3.0–19.9) months in 2016–2019, but this was not significantly different (P = 0.10).
Demographic and clinical characteristics of CNO population.
A total of 152 (71%) patients had a family history of rheumatological or autoimmune disease in first- or second-degree relatives, with only 33 (15%) patients having a first-degree relative affected. Among the 152 patients, 46 had a family history of rheumatoid arthritis, 33 with psoriasis (PsO), 22 with inflammatory bowel disease (IBD), 8 with ankylosing spondylitis, 5 with psoriatic arthritis, 3 with unspecified inflammatory arthritis, and 4 with CNO. Associated conditions in patients were found in 91 (42%) patients including 28 (13%) with PsO, 21 (10%) with arthritis, 13 (6%) with IBD, and 3 (1%) with severe cystic acne (Table 1).
Rheumatologists and orthopedic surgeons diagnosed 189/215 (88%) of the CNO cases, with diagnoses made by rheumatologists and orthopedic surgeons in 133 (62%) and 56 (26%) cases, respectively. In terms of referrals to the rheumatology department, orthopedics referred more cases than any other specialty, with a total of 77 cases, followed by primary care physicians with 46 cases, urgent care physicians with 14 cases, infectious disease physicians with 10 cases, oral surgeons with 9 cases, gastroenterologists with 5 cases, and oncologists with 5 cases. A dermatologist, occupational therapist, endocrinologist, otolarygologist, and physical medicine and rehabilitation physician referred 1 case each; there were 44 cases of unidentifiable referring sources.
Bone biopsy was performed in 74 of 99 patients (75%) between 2005 and 2015 and in 60 of 116 patients (52%). Open biopsy was the main biopsy approach throughout, with 71 (96%) between 2005 and 2015 and 55 (92%) between 2016 and 2019.
Geographic and seasonal distribution. Of all 215 subjects, 188 resided in Washington state, with the 2 leading zip codes (5 and 7 cases) located in Snohomish County, the third most populous county. The county with the most cases of 73, King County, is the most populous county in Washington. Pierce County is the second most populous county; however, more cases were reported in Snohomish than in Pierce County with 45 and 10 cases, respectively (Supplementary Data 1, available with the online version of this article). Most cases were in western Washington and the distribution of cases and population density of these areas were consistent. King and Snohomish Counties make up 63% of all the cases in Washington (39% and 24% respectively) but account for only 40% of the state’s population.
There was no significant difference in seasonal distribution of new cases within Washington state or the WAMI region based on the month of diagnosis or the month of disease onset (P = 0.46; Supplementary Data 2, available with the online version of this article).
Distribution of lesion sites within skeleton. From 2005 to 2015, among 99 patients, 70 had regional MRI, 2 whole-body MRI (WB-MRI), and 56 bone scintigraphy within 3 months of diagnosis, whereas 16 patients had no advanced imaging. In contrast, between 2016 and 2019, among 116 patients, 83 had regional MRI, 63 WB-MRI, 12 bone scintigraphy, and 9 had no advanced imaging. Lesions reported by radiologists are summarized in Figure 2 and Supplementary Data 3 (available with the online version of this article). Within those who had regional MRI, 45 (29%) had a unifocal lesion and 108 (71%) had multifocal lesions (eg, multiple vertebrae, distal femoral, proximal tibial lesions) within 3 months of diagnosis. Within those who had WB-MRI, 20 (31%) had a unifocal lesion and 45 (69%) had multifocal lesions within 3 months of diagnosis.
Heatmap of skeleton lesions in body detected by using (A) regional MRI (n = 153) and (B) whole-body MRI (n = 65). Dark gray indicates more commonly affected areas and light gray indicates less commonly affected areas. Long bones were divided into proximal, diaphysis, and distal segments. Pelvis was divided into ilium, periacetabulum, and ischium/pubis on each side. Skull, mandible, and sacrum were divided into left and right halves. Hand bones including carpal bones were counted as 1 unit. Foot bones were divided into hindfoot (talus and calcaneus) and fore-/midfoot. Each vertebra was counted as 1 individual bone. MRI: magnetic resonance imaging.
Skeletal lesions were detected using either regional MRI or WB-MRI (Figures 2A,B). Lesions were marked proximal, diaphysis, or distal based on the site of the lesion for long bones, and all were marked left or right to assess for bilateral and symmetrical involvement. Long bones included the humerus, radius, ulna, femur, tibia, and fibula.
The regional MRI recorded 6 patients with head and face lesions (4%), 13 (9%) with upper torso lesions, 12 (8%) with spine lesions, 44 (29%) with lower torso lesions, 11 (7%) with upper extremities lesions, and 97 (63%) with lower extremities lesions. Among those with regional MRI, 12 out of 19 patients who had spine MRIs had active spinal lesions. The WB-MRI recorded 8 (12%) patients with head and face lesions, 6 (9%) with upper torso lesions, 12 (19%) with spine lesions, 17 (26%) with lower torso lesions, 7 (11%) with upper extremity lesions, and 46 (71%) with lower extremity lesions. Common skeletal sites (sites with > 15 counts) detected by the regional MRI were lower extremities (tibia and femur) and lower torso (sacrum and periacetabulum). With the WB-MRI, the most common skeletal sites were lower extremities (tibia > femur > fibula).
Regional MRI demonstrated higher lesional counts in the right lower extremity, though with WB-MRI, there were similar numbers of lesions in bilateral lower extremities. Bones that had the highest bilateral involvement within the same patient were the sacrum (21 left and 16 right), periacetabulum (15 left and 16 right), proximal tibia (19 left and 23 right) and proximal femur (18 left and 17 right; Supplementary Data 3, available with the online version of this article).
DISCUSSION
We reported the increasing number of CNO cases from a tertiary pediatric center that has a large catchment in the US Northwest. Due to the lack of a national registry and single-payer system in the US, it is difficult to assess the incidence rate of a rare disease. Our center is the main regional referral center for pediatric rheumatologic diseases, complex oral surgery, and bone tumors within the WAMI region and evaluates suspected CNO cases. The rise of the new CNO cases since 2016 has increased nearly 3-fold, and is likely due to the increasing awareness of this condition within our center, as well as further collaboration between rheumatology and referring physicians from oral maxillofacial surgery, orthopedic surgery, and infectious disease. Similar increases have been reported in the UK and Germany.6,10 In addition, bone biopsies were performed more commonly between 2005 and 2015, and thus, the diagnosis could have been hastened. On the other hand, time to diagnosis had increased in recent years, perhaps owing to the difficulty of accessing rheumatology clinic or delay of obtaining whole-body imaging due to insurance denials. Further research is required to determine the cause for delayed diagnosis.
The population increase from our region was unlikely the cause of the rising case rate of CNO. The population of King County increased by 11% (roughly 1.7 to 1.9 million) between 2000 and 2010 and by 13% (roughly 1.9 to 2.2 million) between 2010 and 2018. The 3-fold increase of new patients with CNO significantly outpaced the population growth. Across counties, the small number of cases reported from Pierce County, the second most populous within Washington state, could be associated with the presence of a nearby children’s hospital to which CNO patients in that region might have been referred to the non-SCH rheumatologist hired in 2016 at a nearby hospital. Collaboration among pediatric rheumatology clinics to maximize access for all eligible patients will be important to diagnose and treat all suspected patients in a timely manner.
Geographic distribution of cases did not show a clear trend of increasing cases in any specific area. The demographic distribution of our cohort is comparable to that of previous reports1,3,4,6,10 It may represent the relative awareness of this condition among community physicians and can be used for future advocacy projects through further education. Seasonal distribution showed variation across the entire year but did not show specific predilection.
The increasing use of WB-MRI in our center has been accompanied with the complete discontinuation of bone scintigraphy, which uses ionizing radiation and is less sensitive. The identified lesional sites in this study are consistent with previously reported locations.1,2,4,5 Although regional MRI demonstrated increased propensity for right lower extremity lesions, WB-MRI noted comparable incidences in both lower extremities. With the majority of patients having multiple lesions and the possibility of missing bilateral and asymptomatic lesions, WB-MRI is a critical tool that allows early diagnosis.
The bone biopsy rate has decreased from 75% to 52%, likely due to the increasing use of WB-MRI that revealed multifocal and symmetric lesions more often during 2016 to 2019. Therefore, physicians were more likely confident with the diagnosis of CNO without needing to request a bone biopsy. In addition, increasing awareness of this condition among orthopedic surgeons also avoided unnecessary biopsy in typical cases, and collaboration with rheumatology colleagues led to increasing cases with a trial of medications prior to bone biopsy when red flags such as cytopenia, fever, or highly elevated inflammation markers were absent.
Our study has several limitations. First, the data from nearby hospitals since 2015 were not available so we cannot estimate the CNO incidence rate of the entire region. Second, the retrospective nature of the study resulted in missing data, such as incomplete family history. Last, the lack of a uniform diagnostic algorithm may have led to variation of practicing patterns during initial clinical and imaging investigations.
The annual rate of CNO cases in King County seen in our center has increased from 8 to 23 per million children. There was no apparent seasonal difference at onset or diagnosis. Access to WB-MRI is important for the diagnosis of CNO. Further research including systematic reviews may improve the ability of clinicians to expedite accurate diagnosis and treatment of this rare autoinflammatory bone disease.
ACKNOWLEDGMENT
We would like to thank the donors to Seattle Children’s CRMO Research Fund to make this work possible.
Footnotes
The authors declare no conflicts of interest relevant to this article.
YZ receives research funds from the Childhood Arthritis and Rheumatology Research Alliance, American College of Rheumatology/European Alliance of Associations for Rheumatology, BMS, and the Washington Research Foundation.
- Accepted for publication March 8, 2022.
- Copyright © 2022 by the Journal of Rheumatology
REFERENCES
ONLINE SUPPLEMENT
Supplementary material accompanies the online version of this article.
