Background
Chronic recurrent multifocal osteomyelitis (CRMO) is an auto-inflammatory osteopathy that predominantly affects children [
1‐
3]. It is characterized by bone pain, soft tissue swelling and osteolysis. SAPHO syndrome (synovitis, acne, pustulosis, hyperostosis and osteitis) includes similar bony lesions, and the dermatologic manifestations may be absent in childhood [
4]. A recent series suggests that unifocal and/or non-recurrent non-bacterial osteomyelitis may represent different levels of severity within the same clinical spectrum, and a new classification has been proposed to include all of these disorders under the umbrella term of chronic non-bacterial osteomyelitis (CNO) [
5]. Within CNO, the existing entity of CRMO represents the more severe end of the spectrum [
5]. In this manuscript, the term CRMO is retained to describe this subgroup of CNO patients.
Although majority of patients with CRMO have resolution of symptoms post-pubertally, the bone pain in active disease is severe [
6‐
9]. In addition, long-term studies reveal that up to a quarter of patients have persistent disease, with risk for permanent bony deformities, poorer quality of life, and difficulty in achieving vocational goals [
1,
2,
10,
11]. There is no cure, and the goal of management is effective pain control [
1,
2]. No single agent has proven to be consistently effective, although analgesics, nonsteroidal anti-inflammatory agents (NSAIDs) [
5,
12], sulfasalazine, methotrexate, corticosteroids and infliximab have all been tried [
3]. There is a need to develop new therapies for treatment resistant patients.
Histology [
6,
12,
13] reveals inflammatory changes, and increased osteoclasts and bone resorption characterize early lesions. Plain radiographs of early CRMO document osteolytic lesions with periosteal reaction [
9,
11‐
13]. Magnetic resonance imaging (MRI) [
14] is a non-invasive imaging modality that is highly sensitive to active and remitted inflammatory lesions in bone and soft tissues in CRMO [
14,
15]. Active lesions occur with increased signal intensity on short tau inversion recovery (STIR) or fat-saturated T2-weighted images and decreased signal intensity on T1-weighted images, consistent with inflammation [
15]. Conversely, resolved CRMO is reflected by no signs of inflammation by MRI [
5].
Because bisphosphonates inhibit bone resorption, have pain modifying effect and demonstrate anti-inflammatory action [
3], they seemed excellent candidates for use in CRMO. In year 2003, when our first patient was treated, there were several reports of beneficial effect of bisphosphonates in adult patients with inflammatory osteitis, such as symphysitis pubis and ankylosing spondylitis [
16,
17], and in non-inflammatory pediatric bone disorders, such as osteogenesis imperfecta [
18] and polyostotic fibrous dysplasia of the bone [
19]. In addition, there was one existing description in an abstract form of successful use of pamidronate for rapid pain relief in a pediatric patient with long-standing CRMO [
20]. These encouraging reports prompted us to pursue pamidronate for those patients with persistently active CRMO despite conventional treatment.
Urinary N-telopeptide/urine creatinine (uNTX/uCr) is a marker of collagen-1 breakdown. It is traditionally used to monitor treatment response to bisphosphonates in adult patients with bone diseases characterized by accelerated bone turnover, such as Paget's disease of bone [
21,
22]. It has not been previously studied in pediatric inflammatory osteopathies.
We present our clinical experience using pamidronate in nine children with CRMO, in whom we prospectively collected data on clinical features, pain, MRI and uNTX/uCr response.
Discussion
We have presented data on clinical and MRI response to pamidronate in nine pediatric patients with persistent CRMO. Dramatic pain relief, followed by an impressive improvement in bone inflammation was documented in all patients. Four children had MRI confirmed relapses 12 – 18 months following pamidronate treatment, with prompt response to re-treatment.
Although the individual course of CRMO can vary, long-term studies emphasize the potentially debilitating nature of this condition. A 2007 review of 39 pediatric CRMO patients by Jansson et al reported a median disease duration of 21 months, with 51% of patients having ongoing active disease in the preceding 18 month period [
3]. Other studies have revealed disease duration for up to 14–25 years [
1,
2,
15]. Huber et al reporting on 23 Canadian patients [
2] and Catalano-Pons et al reporting on 40 patients from a multi-center study in France[
25] confirm that for patients with persistent disease, the adverse effects can extend into future and include: 1) poorer quality of life[
2]; 2) risk for bone abnormalities (kyphosis, vertebral fractures, and limb length discrepancy); and 3) potentially worse vocational outcome [
1‐
4,
9,
25,
26]. Even those patients who eventually recover have significant pain during CRMO flares, with adverse effect on general well-being. Although anti-inflammatory medications and surgery are effective for some patients, there is a need to find effective treatment for refractory cases.
Pamidronate was chosen because of beneficial reports in one pediatric patient with CRMO [
20] and in several adult patients with inflammatory osteitis [
16,
17,
27]. During our study period, other reports in the literature emerged, confirming the efficacy of bisphosphonates for pain relief in pediatric and adult onset SAPHO syndrome [
4,
21], in pediatric patients with chronic inflammatory lesions of the mandible [
28], and in further cases of CRMO [
29,
30]. According to the new proposed classification of chronic non-bacterial osteomyelitis (CNO) as an umbrella term for CRMO, SAPHO syndrome, and unifocal/non-recurrent non-bacterial osteomyelitis [
5], the equally beneficial effects of bisphosphonates in these disorders seem to be supportive of a shared pathogenesis.
Pathogenesis of CRMO/CNO continues to be poorly understood. It is not known what triggers the initial episode, and why some patients have more persistent disease than others. However, both local and systemic increase of tumor necrosis factor alpha (TNF-α) has been documented in active CRMO/CNO [
3]. In keeping with this observation, beneficial effects of anti-TNF agents have been reported in a few patients each with CRMO [
31], SAPHO syndrome [
32], and inflammatory bowel disease related CRMO [
33]. Indeed, some of the anti-inflammatory effects of bisphosphonates are proposed to be secondary to their ability to suppress proinflammatory cytokines, such as TNF-α, interleukin (IL)-6, and IL-1 [
16].
The dosing and scheduling regimen for pamidronate in our study was adapted empirically from the schedule used in treating osteogenesis imperfecta [
18]. Although we had designed our study to allow for a maximum cumulative pamidronate dose of 11.5 mg/kg/year, no patient required this maximum dose.
The most striking clinical observation was rapid pain relief within 48 hours following the first pamidronate treatment, regardless of the location of CRMO lesions. The effective pain relief following pamidronate has now been observed in a variety of pediatric CNO disorders. The 1999 abstract by Seibel et al was the first to document dramatic pain relief following pamidronate in an adolescent boy with a 6-year history of intermittently active CRMO [
20]. This patient received 30 mg of IV pamidronate, with complete pain resolution within 4 days. His radiographs gradually improved over one year, although they did not completely normalize. The beneficial effect of pamidronate in childhood onset SAPHO syndrome is illustrated by Kerrison et al who reported dramatic pain relief in 7 pediatric patients following an initial dose of pamidronate [
4]. In keeping with the observation that some children with SAPHO syndrome may lack dermatological features, these authors included three patients with bone lesions only, and who therefore were similar to our CRMO patients. The skin lesions in the remaining patients consisted of psoriasis/pustular psoriasis. The dose of pamidronate varied between a maximum of 30–60 mg/day, given as 3-day cycles. The frequency of pamidronate cycles was determined by symptom recurrence. Five patients had clinical relapse of osteitis, which responded to repeat pamidronate in all patients. The authors did not comment on dermatological response following pamidronate. Finally, complete pain resolution and decrease in adjacent soft tissue swelling was documented by Compeyrot-Lacassagne et al in two pediatric patients with chronic inflammatory lesions of the mandible within one week of pamidronate [
28]. Similar to Kerrison's and our series, both of these patients had recurrence of inflammatory osteitis, 6 and 17 months after the initial treatment. CT scan together with clinical features and laboratory measurements was used to monitor disease activity.
Although the rapid pain relief after bisphosphonates in multiple bone disorders is well documented, its etiology remains incompletely understood. Ghilardi et al have studied osteolytic bone lesions in malignancies, and suggest that overactive osteoclasts induce acidic environment, which allows bone resorption by osteoclasts, contributing to the bone destruction and damage to primary afferent fibers [
34]. Bisphosphonates neutralize the nociceptive acidic environment rapidly by inactivating osteoclasts. This mechanism could explain the rapid pain relief in our patients. The prompt improvement in soft tissue swelling adjacent to bone lesions, which was observed in several of our CRMO patients, most likely reflects the anti-cytokine effects of pamidronate.
Compared to the above reports on effective analgesic effects of pamidronate in diverse pediatric CNO disorders, our study was unique in that all patients had correlating MR images to document bone inflammation and its resolution at pre-determined time points [
4,
20]. The advantage of obtaining MRI at baseline and during follow-up is multifold. MRI is extremely sensitive in detecting bone marrow edema and soft tissue edema that are present in inflammatory processes such as CRMO. It is more sensitive indicator of disease activity than the lytic changes seen on plain radiography. It is especially helpful in evaluating lesions that are difficult to visualize by plain radiography, such as spine and pelvis [
35]. At baseline, all patients demonstrated bone and/or soft tissue inflammation by MRI. MRI improved remarkably after the first pamidronate treatment, with further resolution with subsequent treatments, supporting the proposed anti-inflammatory role of bisphosphonates [
4,
27]. Because CRMO can result in pain from non-inflammatory bone changes such as vertebral body compression fractures from spinal CRMO [
3], we recommend a repeat MRI assessment in cases of suspected flare, particularly if repeat treatment with pamidronate is contemplated.
In this study we show for the first time data on bone remodeling markers in pediatric CRMO patients. In adult onset SAPHO syndrome, increased levels of serum osteocalcin have been observed in some patients, suggesting greater importance of bone resorption compared with bone formation in this disorder [
21]. Although histology reveals increased osteoclasts in early CRMO lesions [
8], and early radiographs reveal osteolysis, we were not able to show generalized increase in bone resorption markers or in markers of bone formation compared to age-specific norms. Four patients had baseline uNTX/uCr values above the 75
th % percentile for age, but we feel that this increase most likely reflects the on-set of puberty, rather than increase from CRMO related osteolysis. All patients had a decrease in uNTX/uCr following pamidronate, which was expected. However, while it is also expected that uNTX/uCr values gradually begin to increase after bisphosphonates are discontinued, it is interesting that no patient relapsed while his/her bone turnover remained suppressed. It is therefore tempting to speculate that bone and adjacent soft tissue inflammation may require functioning osteoclasts for clinical manifestations. However, further studies are required on the role of osteoclasts and on the potential use of uNTX/uCr in pediatric CRMO.
In long-term studies 20–30% of CRMO patients may develop acne, psoriasis, synovitis, inflammatory bowel disease and/or ankylosing spondylitis [
3,
5,
36]. The low rate of these associations in our patients most likely reflects the well documented lag in presentation of the extra-osseous manifestations, and their true incidence will only become apparent with longer follow-up.
Similar to other reviews of pamidronate in CNO, the only observed side-effect in our study was transient acute phase response with the first pamidronate dose [
4]. Osteonecrosis of the jaw (ONJ) has emerged as a potential concern for adult patients treated with bisphosphonates. ONJ is in most cases initiated by tooth removal, and most reported cases have had a history of malignancy [
37]. To date, ONJ has not been reported in pediatric patients [
23,
37]. A comprehensive review by Malmgren et al of all 64 pediatric osteogenesis imperfecta patients in Sweden who were treated with pamidronate from 1991–2005 revealed no cases of ONJ. A total of 38 dental surgical procedures were performed in 22 of these 64 patients at the age of 3.4–31.9 years after 0.03–7.9 years of treatment [
37]. However, until more pediatric data is collected, we suggest the following: 1) a thorough dental assessment pre-pamidronate; 2) wisdom tooth extraction pre-pamidronate in adolescent patients if recommended by a dentist; and 3) postponement of dental braces or elective dental extractions for at least 6 months following the final pamidonate treatment.
The "minimum effective dose" and ideal duration of pamidronate treatment for CRMO is not known. In this case series pamidronate was used only when conventional treatment was ineffective, and symptoms were of long standing. A much lower dose of pamidronate was required for MRI documented CRMO resolution following relapse, suggesting that early CRMO may be more amenable to bisphosphonate treatment than persistent disease. Pamidronate was discontinued once MRI abnormalities had resolved. At that time, uNTX/uCr remained suppressed, and ongoing pamidronate treatment would not have likely offered additional benefit. However, our experience as well as that of others [
4,
28] suggests that many pediatric CRMO/CNO patients who are treated with pamidronate eventually flare. We have restricted repeat treatment to those patients who have MRI confirmed relapse of bone inflammation, and whose flare is unresponsive to NSAIDs.
Finally, pamidronate has been the only intravenous bisphosphonate available for pediatric patients. However, other intravenous bisphosphonates are becoming available for this age group, including zoledronic acid, which, because of its higher relative potency [
38], may achieve the same effect as pamidronate with less number of infusions and possibly longer time interval until relapse.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
PMHM conceived of the study, and participated in its design and coordination, and drafted the manuscript. XW, DK and WF participated in the design for diagnostic imaging of the patients, and contributed equally to analysis of radiographs and magnetic resonance imaging. ANA contributed to study design and was responsible for data analysis, and JDK contributed to overall study design and helped draft the manuscript. All authors read and approved the final manuscript.