Rheumatology 2008;47:13971399

Advance Access publication 11 July 2008

doi:10.1093/rheumatology/ken249

Clinical outcome in children with chronic recurrent multifocal

osteomyelitis
C. Catalano-Pons1, A. Comte2, J. Wipff3, P. Quartier4, A. Faye5, D. Gendrel1, A. Duquesne6,
R. Cimaz6 and C. Job-Deslandre3

KEY

WORDS:

Chronic recurrent multifocal osteomyelitis, Children, Outcome.

<18 yrs at diagnosis; in 28 cases, a bone biopsy with microbial

investigation was performed and excluded an infectious origin.
A chronic non-bacterial osteitis was diagnosed when there was
only one bony lesion during the course of the disease. A patient
was defined as responder to treatment if there was absence of pain
and improvement in inflammatory markers.
Questionnaires were sent to assess disease activity and
educational and vocational achievement. Disease activity was
defined as the presence of pain or use of medical treatment. Pain
was evaluated using a visual analogue scale of 0100. Patients
were asked to indicate if they had received treatment, and to
specify the drug. Global impact of the disease was determined
with another visual analogue scale graduated from 0 to 100.
Physical function was assessed using the HAQ including eight
domains: dressing and grooming, arising, eating, walking,
hygiene, reach, grip and activity. The final score ranged from 0
(no disability) to 3 (severe disability). Educational achievement
was evaluated by the highest educational level reached and
vocational achievement by asking about employment. Oral
consent was obtained from the parents and, when possible, the
patients, as recommended in France. In accordance with the
French rules, no ethical committee agreement was needed before
sending the questionnaire to the patient or the parents.
The Pearson correlation was used in a first analysis to
determine as to which parameters were correlated with the
duration of active disease and with the response to NSAIDs.
Then a multivariate regression was used to construct a multivariate model for the duration of active disease including
biological parameters (ESR, CRP, platelets, white blood cells),
age at the onset, sex, initial and total number of bony lesions and
response to NSAID treatment. The logistic regression was used to
construct a model for the response to NSAID treatment including
the same parameters.

Introduction
Chronic recurrent multifocal osteomyelitis (CRMO) is a noninfectious inflammatory bone disease of unknown aetiology, first
reported by Giedion et al. in 1972 [1]. The disease causes
multifocal lytic bone lesions with swelling and pain. It could be
associated with the other symptoms of SAPHO syndrome
(synovitis, acne, pustulosis, hyperostosis, osteitis) [2], or with
sacroiliitis, psoriasis [3] and IBD [4]. The course of the disease is
characterized by periodic exacerbations and remissions but longterm outcome remains unclear. Most of the studies in the
literature are case reports or short series [3, 58]. Huber et al.
[9] determined the long-term clinical outcome of 23 patients with
CRMO. Most subjects had no disease activity or sequelae but
25% of patients had persistent disease.
The aim of our study was to determine clinical outcome of
children with CRMO.

Patients and methods

We retrospectively reviewed clinical, biological and radiological
data of children with CRMO diagnosed before 1 January 2007
and followed up at five French paediatrics centres (Hopitaux
Saint-Vincent-de-Paul, Cochin, Necker, Robert Debre, Paris and
Hopital Edouard Herriot, Lyon, France). The diagnosis of
CRMO was made based on multifocal bony lesions in the absence
of infectious origin, typical X-ray pattern (osteolysis with
sclerosis) and high uptake on technetium bone scan (which was
performed in 34 out of 40 cases), chronic evolution and age

1
General Pediatric Department, Saint-Vincent-de-Paul Hospital, Paris, 2General
Pediatric Department, Saint-Jacques Hospital, Besancon, 3Rheumatology
Department, Cochin Hospital, 4Pediatric Immunology and Haematology
Department, Necker Hospital, 5General Pediatric Department, Robert Debre
Hospital, Paris and 6Pediatric Department, Edouard Herriot Hospital, Lyon, France.

Results
Forty patients were evaluated (34 females and 6 males) with a
median age of 10 yrs (range 117) at first symptoms. There was
statistically no difference between the male and female patients
with regard to demographic and clinical data (age at disease onset,

Submitted 21 November 2007; revised version accepted 9 June 2008.

Correspondence to: C. Catalano-Pons, Service de Pediatrie Generale, Centre
hospitalier dAvignon, 305, rue Raoul Follereau, 84902 Avignon cedex 9, France.
E-mail: [email protected].

1397
The Author 2008. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected]

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 24, 2015

Objective. To determine the clinical outcome of children with chronic recurrent multifocal osteomyelitis (CRMO).
Methods. We retrospectively reviewed clinical, biological and radiological data of children with CRMO at five French paediatric centres.
Outcome data were obtained through review of hospital charts and questionnaires sent to all patients to assess disease activity and
educational and vocational achievement.
Results. Forty patients were assessed (34 females and 6 males) with a median age at diagnosis of 11.5 yrs (range 217). Median number of
initial bony lesions was 2 at onset, and 3.5 over disease course. Median time since diagnosis was 3.5 yrs (range 0.515) and median duration
of active disease 2.7 yrs (range 0.513.5). Nine (22.5%) patients had psychological or physical sequelae. Twenty-nine children (72.5%)
responded to the questionnaire. Twenty-six had no physical disability as judged by the HAQ 01, two had moderate disability (HAQ: 12) and
one had severe disability (HAQ: 23). Seventeen patients (58.6%) had active disease at follow-up (after 6 months to 15 yrs since diagnosis)
and continued to have pain (median value of visual analogue scale: 10/100). CRMO had interfered with patients education in two cases.
Conclusions. Clinical outcome of children with CRMO is generally good, but a sizeable proportion of patients have active disease at followup, and a minority of patients can have a severe and prolonged disease course despite intensive treatments. Further studies are required to
determine predictive factors for severe disease.

C. Catalano-Pons et al.

1398
TABLE 1. Bony sites affected
Site affected

TABLE 3. Pearson correlation for duration of active disease

Number (%) of subjects

Distal tibia
Pelvis
Proximal tibia
Metatarsals
Vertebrae
Proximal femur
Clavicle
Calcaneus
Distal femur
Distal fibula
Sternum
Distal radius
Scapula
Distal humerus
Patella
Ribs
Proximal humerus
Distal ulna
Mandible

23
13
11
11
10
9
8
8
5
5
5
4
4
3
3
3
2
1
1

(57.5)
(32.5)
(27.5)
(27.5)
(25)
(22.5)
(20)
(20)
(12.5)
(12.5)
(12.5)
(10)
(10)
(7.5)
(7.5)
(7.5)
(5)
(2.5)
(2.5)

Age at disease onset (yrs)

Sex
Initial number of localizations
Total number of localizations
NSAID response
WBC (per mm3)
Platelet count (per mm3)
ESR (mm/1st h)
CRP (mg/l)

0.385
0.04
0.246
0.262
0.53

0.004
0.146
0.337
0.166

0.01
0.8
0.02
0.01
0
0.9
0.4
0.04
0.34

Bold values indicate P < 0.05.

Means for
non-responders

Means for
responders

8.7
8F/1M
4
5
45
14
7600
345 428
7.5

9.7
25F/5M
2
4
28
5
8560
348 000
2.7

NS
NS
0.03
NS
NS
NS
NS
NS
0.05

Bold values indicate P < 0.05. F, female; M, male; NS, not significant.

initial and total number of lesions, biological parameters,

percentage of response to NSAIDs and duration of active disease).
Median age at diagnosis was 11.5 yrs (range 217). Initial
symptoms were bony pain in all patients, associated with localized
swelling and arthritis secondary to local bony inflammation in
four children. Three patients suffered from enthesitis, two at onset
of the disease and one during the disease course. Nine patients had
fever and one pustulosis plantaris. Biological evaluation at onset
of the disease usually showed normal white blood cell count,
haemoglobin and platelet count and slightly increased inflammatory markers with a median value of ESR of 26 mm/1st h and of
CRP of 5 mg/l (ref <5 mg/l).
Median number of initial bony lesions for each patient was 2
(range 17) at onset and 3.5 (range 111) during the disease
course. Fifteen patients (37.5%) had only one bone lesion at onset
of the disease and four patients (10%) had a unifocal course. The
sites affected are summarized in Table 1. Twenty-eight children
underwent bone biopsies that showed non-specific subacute or
chronic inflammation with mild fibrosis. Microbial investigations
remained negative.
All patients except one were initially treated with NSAIDs.
Eleven patients received antibiotics, nine corticosteroids, eight
SSZ, four MTX, three intravenous bisphosphonates, two etanercept and one AZA. The efficacy of NSAIDs seemed to be linked
to the initial number of bony sites (Table 2): patients who did not
respond to NSAIDs had statistically more lesions at onset
(P 0.03). There was no significant difference of age, sex and
biological parameters between the responders and non-responders
to the NSAID.
Median age at last follow-up was 14.5 yrs (range 6.527).
Median time since diagnosis was 3.5 yrs (range 0.515) and

median duration of active disease 2.7 yrs (range 0.513.5). The

duration of active disease was linked to the initial number of
localizations (P < 0.001) and the total number of localizations
(P < 0.001) and inversely correlated with the age at onset
(P < 0.001) in a multivariate regression analysis. ESR at onset
was correlated with the duration of active disease (Pearson
correlation, P 0.04) in a linear regression analysis, but not in a
multivariate regression model (Table 3). The other biological
parameters (CRP, white blood cell count, haemoglobin and
platelets) were not associated with a prolonged evolution of the
disease.
Eight patients (20%) had physical sequelae, including leg length
discrepancies in three patients, bony overgrowth of the clavicle in
one patient, generalized growth failure in three patients (ranging
from
1 to
2 S.D.) and vertebral compression fracture due to
corticosteroids in one child. One patient suffered from depression
linked to the disease. Associated features were psoriasis in two
patients, pustulosis in one and Takayasu arteritis in one. Twentynine children (72.5%) responded to the questionnaire. Twenty-six
had no physical disability (HAQ: 01), two had moderate
disability (HAQ: 12) and one had severe disability (HAQ: 23).
Seventeen patients (58.6%) had active disease at last follow-up
(after 6 months to 15 yrs since diagnosis) and still suffered from
pain (median value of visual analogue scale: 10/100), all requiring
analgesics. Fifteen (51.7%) patients considered that CRMO had
some global repercussion on their life (median value of visual
analogue scale: 10/100). CRMO had interfered with the patients
education in two cases, who failed to obtain a degree because of
hospitalization and pain. The first patient was a 9-yr-old boy at
the beginning of the disease, with a duration of the disease of
7.5 yrs. He received successively antibiotics, NSAID, corticosteroids, intravenous bisphosphonates and etanercept. He suffered
from overgrowth of the mandible and generalized growth failure
of
1.5 S.D. The second patient was a 7-yr-old girl at the beginning
of the disease, with a duration of the disease of 15 yrs. She was
treated successively with NSAIDs, SSZ, corticosteroids, MTX and
AZA. She suffered from generalized growth failure of
2 S.D. No
other patient thought that CRMO had affected their ability to
obtain a degree or a job.

Discussion
In our study as reported in the literature [2, 7], 85% of the cases of
CRMO occurred in females with a median age at onset of the
disease of 10 yrs [2, 3]. An association with SAPHO syndrome or
psoriasis has been reported in up to 23% of the cases [3], but was
rare in our study, affecting only three patients. Initial symptoms
were bone pain in all patients and fever in only nine patients.
Inflammatory markers were increased in 65% of the patients in
one series [10], and in 68% of our cases. Median number of total
bony lesions was 3.5, in agreement with what is described in the
literature [3, 9]. Lower limbs were mostly affected, usually with
lesions of the metaphysis as in previous reports [5, 6, 8, 9].

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 24, 2015

TABLE 2. Multivariate model for response to NSAID treatment

Age at disease onset (yrs)

Sex ratio
Initial number of lesions
Total number of lesions
ESR (mm/1st h)
CRP (mg/l)
WBC (per mm3)
Platelet count (per mm3)
Duration of active disease (yrs)

Pearson correlation
coefficient

Chronic recurrent multifocal osteomyelitis

Despite this fact, we must underline that we were able to gather

information on the majority (72.5%) of the patients, which is a
high rate of response for such type of study.
In conclusion, our study suggests that clinical outcome of
children with CRMO is generally good, but a sizeable proportion
of patients have active disease at follow-up, and a minority of
patients can have a severe and prolonged disease course despite
intensive treatments. Further studies are required to confirm the
risk factors for severe and persistent disease, and then to treat
those patients more aggressively in order to prevent a poor longterm outcome.

Rheumatology key messages

 Clinical outcome of children with CRMO is generally good.
 About one-fifth of the patients can have severe and prolonged
disease course.
 Younger age and high number of bony sites at onset seems to be
predictive of poorer outcome.

Disclosure statement: The authors have declared no conflicts of

interest.

References
1 Giedion A, Holthusen W, Masel LF, Vischer D. Subacute and chronic symmetrical
osteomyelitis. Ann Radiol 1972;15:32942.
2 Beretta-Piccoli BC, Sauvin MJ, Gal I et al. Synovitis, acne, pustulosis, hyperostosis,
osteitis (SAPHO) syndrome in childhood: a report of ten cases and review of the
literature. Eur J Pediatr 2000;159:594601.
3 Schultz C, Holterhus PM, Seidel A et al. Chronic recurrent osteomyelitis in children.
Pediatr Infect Dis J 1999;18:100813.
4 Carpenter E, Jackson MA, Friesen CA, Scarbrough M, Roberts CC. Crohnsassociated chronic recurrent multifocal osteomyelitis responsive to infliximab.
J Pediatr 2004;144:5414.
5 Quelquejay C, Job-Deslandre C, Hamidou A, Benosman A, Adamsbaum C. Chronic
recurrent multifocal osteomyelitis in children. J Radiol 1997;78:11521.
6 Coinde E, David L, Cottalorda J et al. Chronic recurrent multifocal osteomyelitis in
children: report of 17 cases. Arch Pediatr 2001;8:57783.
7 Job-Deslandre C, Krebs S, Kahan A. Chronic recurrent multifocal osteomyelitis:
five-year outcomes in 14 pediatric cases. Joint Bone Spine 2001;68:24551.
8 Girschick HJ, Raab P, Surbaum S et al. Chronic non-bacterial osteomyelitis in
children. Ann Rheum Dis 2005;64:27985.
9 Huber AM, Lam PY, Duffy CM et al. Chronic recurrent multifocal osteomyelitis: clinical outcomes after more than five years of follow-up. J Pediatr
2002;141:198203.
10 Brown T, Wilkinson RH. Chronic recurrent multifocal osteomyelitis. Radiology
1988;166:4936.
11 Jurik AG, Helmig O, Ternowitz T, Moller BN. Chronic recurrent multifocal
osteomyelitis: a follow-up study. J Pediatr Orthop 1988;8:4958.
12 Girschick HJ, Krauspe R, Tschammler A, Huppertz HI. Chronic recurrent
osteomyelitis with clavicular involvement in children: diagnostic value of different
imaging techniques and therapy with non-steroidal anti-inflammatory drugs. Eur
J Pediatr 1998;157:2833.
13 Ishikawa-Nakayama K, Sugiyama E, Sawazaki S et al. Chronic recurrent multifocal
osteomyelitis showing marked improvement with corticosteroid treatment.
J Rheumatol 2000;27:13189.
14 Holden W, David J. Chronic recurrent multifocal osteomyelitis: two cases of sacral
disease responsive to corticosteroids. Clin Infect Dis 2005;40:6169.
15 Gallagher KT, Roberts RL, MacFarlane JA, Stiehm ER. Treatment of chronic
recurrent multifocal osteomyelitis with interferon gamma. J Pediatr 1997;131:4702.
16 Andersson R. Effective treatment with interferon alpha in chronic recurrent multifocal
osteomyelitis. J Interferon Cytokine Res 1995;15:8378.
17 Seibel MJ, Farahmand I, Zeigler R. Successful treatment of chronic recurrent
multifocal osteomyelitis (CRMO) with intravenous bisphosphonates: a case report.
Exp Clin Endocrinol Diabetes 1999;107:115.
18 Guignard S, Job-Deslandre C, Sayag-Boukris V, Kahan A. Therapeutic use of
pamidronate in SAPHO syndrome. Joint Bone Spine 2002;69:3926.
19 Deutschmann A, Mache CJ, Bodo K, Zebedin D, Ring E. Successful treatment of
chronic recurrent multifocal osteomyelitis with tumor necrosis factor alpha-blockage.
Pediatrics 2005;116:12313.
20 Duffy CM, Lam PY, Ditchfield M, Allen R, Graham HK. Chronic recurrent multifocal
osteomyelitis: review of orthopaedic complications at maturity. J Pediatr Orthop
2002;22:5015.

Downloaded from http://rheumatology.oxfordjournals.org/ by guest on March 24, 2015

Involvement of the clavicle has classically been described [11], but

was found in only 20% of our patients, as reported in other
studies [7, 9]. Bone biopsy was only performed when diagnosis was
uncertain and showed non-specific chronic inflammation, as
described by Girschick et al. [8].
As in SAPHO syndrome, NSAIDs are usually effective with a
response in up to 80% of the cases [3, 12]. The fraction of good
responders has been probably overstated because inflammatory
markers were usually only slightly increased and the patients pain
evaluation was rather good despite the persistence of the disease.
In our study, they were used in almost all cases, with 73%
responders. Corticosteroids can be used for severe relapsing cases
with a very high response rate [13, 14]. In a small number of
patients, improvement has been observed after treatment with
SSZ, MTX, colchicine [9], IFN-
[15], IFN- [16], bisphosphonates [17, 18] and infliximab [4, 19].
Although the early literature suggested that CRMO is a selflimiting condition without sequelae, more recent studies have
found that symptoms can be prolonged. Some cases that remained
active for many years have been reported [11]. In this study,
median duration of active disease was 2.7 yrs with a median time
since diagnosis of 3.5 yrs. More than half of the patients had active
disease at follow-up, continued to have pain and considered that
CRMO had global consequencies on their life. In their review of
the literature, Schultz et al. [3] reported data similar to ours, with
similar median duration from diagnosis to remission (2 yrs) and
similar median time from onset of illness to last follow-up
(3.5 yrs). In Hubers study, median overall duration of active
disease was 5.7 yrs. More than 25% of the patients had persistent
CRMO activity at the time of evaluation, a median of 12.4 yrs
later [9]. In comparison with our study in which 20% of the
patients had physical sequelae, only 7% of patients in Schultz
et al.s review [3] had long-term sequelae such as growth
retardation, bone deformities, kyphosis and thoracic outlet
syndrome. This may be due to a recruitment bias because all of
our patients were referred to tertiary care hospitals and were
therefore likely to be more severely affected. Duffy et al. [20] also
found a high rate of orthopaedic complications at maturity in a
cohort of 12 adults, including deformities in seven patients and
leg-length inequality in five patients, requiring surgery in one.
Huber et al. [9] confirmed those results with nearly half of the
cohort suffering from bone deformities.
In spite of this significant morbidity associated with CRMO,
only 3 patients out of 29 had moderate or severe disability when
measured by the HAQ. CRMO had interfered with patients
education in only two cases, and no patient felt that CRMO had
affected their ability to obtain a job. In Huber et al.s [9] long-term
follow-up (median time since diagnosis of 13 yrs), 78% of subjects
had HAQ scores of 0, indicating no or minimal impairment of
physical function; in our study with a median time since diagnosis
of 3.5 yrs, an HAQ of 0 was present in 59% of the cases. In Huber
et al.s [9] study, impairments were seen in all domains of
measurement of a quality-of-life test, but appeared more frequently
in non-physical aspects of health. CRMO had affected education
for seven patients out of 23 and the ability to work in three patients.
Very few studies have analysed predictive factors for persistent
evolution of CRMO. Girschick et al. [12] showed that there was
no significant difference in the duration of NSAID treatment
between the unifocal and multifocal forms. In our study, we
showed that young patients at the onset of the disease with a high
number of bony sites seemed to be at risk for persistent disease.
Moreover, a high number of bony sites at the onset was also a risk
factor for a poor response to NSAIDs. Sex and biological
parameters did not seem to influence the evolution or the response
to the treatment. However, our group of patients is rather
small and these results have to be confirmed by larger series.

1399