1. health and fitness
2. disease

Primary subacute haematogenous osteomyelitis
in children
M. N. Rasool
From the University of Natal, Durban, South Africa
etween 1990 and 1998 we saw 21 children with
primary subacute haematogenous osteomyelitis.
Pain, swelling and a limp had been present for two to
12 weeks with little functional impairment.
Laboratory tests were non-contributory. The lesions
were classified radiologically into metaphyseal,
diaphyseal, epiphyseal and vertebral. There were 24
sites involved, with most (20) being in the tibia; 17
lesions were in the diaphysis, five in the metaphysis
and two in the epiphysis. The diagnosis was confirmed
histologically in all cases. Staphylococcus aureus was
cultured in six patients. Healing occurred in all
patients after treatment with antibiotics for six weeks
and radiological improvement was seen after three to
six months.
Subacute osteomyelitis develops as a result of
increased host resistance and decreased bacterial
virulence. The radiological features can mimic various
benign or malignant bone tumours and non-pyogenic
infections. Histological confirmation is necessary to
avoid a delay in diagnosis.
B
J Bone Joint Surg [Br] 2001;83-B:93-8.
Received 20 March 2000; Accepted after revision 17 July 2000
Primary subacute haematogenous osteomyelitis may be difficult to diagnose because the characteristic signs and symptoms of the acute form of the disease are absent. It has an
insidious onset, lacks a systemic reaction and may mimic
various benign and malignant conditions, resulting in delay in
1-7
diagnosis and treatment. The condition has assumed importance in recent years and is becoming more common in the
8-10
UK.
Although osteomyelitis is frequent in Northern Niger11
1
ia and East Africa it is only occasionally seen in the acute
form.
The primary form of subacute haematogenous osteomyelitis, which occurs mainly in children, must be
distinguished from subacute osteomyelitis which has been
modified by inadequate or partial treatment with antibiotics,
M. N. Rasool, FCS (SA), Consultant Orthopaedic Surgeon
University of Natal, Private Bag 7, Congella 4013, Durban, South
Africa.
©2001 British Editorial Society of Bone and Joint Surgery
0301-620X/01/111180 $2.00
VOL. 83-B, NO. 1, JANUARY 2001
and from other forms of the condition such as chronic
recurrent multifocal osteomyelitis and the SAPHO
syndrome (synovitis, acne, pustulosis, hyperostosis and
osteitis). In previous reports either the primary form has not
been distinguished from that modified by anti6,10,12,13
biotics
or both adults and children have been
1,2,5,6,9
included.
2
King and Mayo described the disease as following an
indolent course because of host resistance combined with low
3
virulence of the infecting organism. Gledhill described four
7
radiological types in long bones. Roberts et al expanded and
modified the classification into six forms to include the spine,
the lesions being classified as metaphyseal, diaphyseal, epiphyseal and vertebral. Their classification is based on anatomical location, morphology and the similarity of the lesions
to various neoplasms.
The diaphyseal forms of subacute osteomyelitis in children
have received less attention in the literature; six cases have
14
been described in South Africa. Most other reports include
1,5,6
mainly adult patients.
The aim of this paper is to describe the spectrum of
primary subacute haematogenous forms of osteomyelitis seen
in children, and the difficulties in making the diagnosis. The
high incidence of diaphyseal lesions and the importance of
obtaining a histological diagnosis are highlighted.
Patients and Methods
The hospital records and radiographs of 21 children with
primary subacute haematogenous osteomyelitis seen at the
King Edward VIII Hospital, Durban between 1990 and 1998,
were reviewed. These formed 7% of all cases of osteomyelitis
treated in this period. Patients with chronic illness or immune
deficiencies and those who had had previous antibiotic therapy or acute osteomyelitis were excluded. The criteria for the
1-3,7
subacute course as previously described.
included symptoms which had been present for more than two weeks, mild
to moderate pain of insidious onset with intermittent symptoms and little or no functional impairment, no systemic
manifestations, negative blood cultures and positive findings
on plain radiographs.
There were 15 boys and six girls with a mean age of 7.5
years (2 to 12). The length of symptoms ranged from two
weeks to three months. Pain was the main complaint,
93
94
M. N. RASOOL
Table I. Characteristics of primary subacute haematogenous osteomyelitis in 21 patients
Type
I
Metaphyseal
(no cortical erosion)
Classification
7
(Roberts et al )
Differential
diagnosis
Number
of cases
(a) Solitary localised punched-out zone
of radiolucency
(b) Localised radiolucent zone with
sclerotic margin (Brodie's abscess)
Eosinophilic granuloma
0
-
Osteoid osteoma
2
Distal femur
3
Proximal tibia
Proximal tibia
Distal tibia
3
Tibial shaft
14
Tibial shaft
Ulnar shaft
II
Metaphyseal
(cortical erosion)
Large radiolucent zone
Osteosarcoma
Eosinophilic granuloma
Tuberculosis
Fungal infection
III
Diaphyseal (cortical)
Localised cortical and periosteal reaction
Osteoid osteoma
IV
Diaphyseal (periosteal)
Periosteal reaction
Linear (single) or laminated (multiple)
Ewing's sarcoma
Round-cell tumours
Leukaemia
V
Epiphyseal
Radiolucent defect in epiphysis with
fine sclerotic margin
Chondroblastoma
Tuberculosis
2
VI
Vertebral
Destructive process involving vertebral body
Osteosarcoma
Tuberculosis
Fungal infection
0
usually a mild ache which did not interfere with sleep. Six
children remembered an episode of minor trauma. All the
patients were afebrile.
The lesions were classified using the system of Roberts
7
et al (Table I). Technetium bone scintigraphy, MRI and CT
were also carried out. All patients underwent surgical
exploration and biopsy. At operation, granulation tissue and
a little pus were obtained from the metaphyseal lesions and
small amounts of granulation tissue from those in the
epiphysis. The latter were approached extra-articularly
under fluoroscopy with a fine drill and a curette. Metaphyseal lesions were also curetted. Specimens of bone and
granulation tissue were sent for microscopic investigations
including those for tuberculosis and fungi and also for
histological examination.
All the affected limbs were immobilised for four to six
weeks after biopsy to allow healing and to avoid pathological fractures. Intravenous cloxacillin was given for four
or five days followed by oral treatment for six weeks. All
patients were assessed clinically and radiologically at intervals of about six weeks. The mean follow-up was for 2.4
years (six months to four years).
Results
Physical examination revealed minor local soft-tissue swelling in 14 patients with some tenderness. The range of
movement of the adjacent joint in epiphyseal and metaphyseal lesions was normal. Twelve patients had a slight
limp. Laboratory investigations revealed a haemoglobin level
of 10.5 to 13.1 g%, and a white cell count of 6.3 to 12.9 ±
9
10 /L with a normal differential count. The ESR ranged from
9 to 49 mm/hr (Westergren).
The symptoms resolved within three to four weeks after
treatment had commenced. Leg length and angular deformity
were assessed clinically and radiologically. Clinical examination showed no leg-length discrepancy and there was no
Sites
Proximal femur
Distal femur
Fig. 1
Radiograph showing a well-defined radiolucent lesion in the metaphysis of the distal
femur surrounded by dense sclerosis (Brodie’s abscess; type-1b lesion).
radiographic evidence of growth disturbance. There were 24
lesions in 21 patients. The sites were in the tibia (20), femur
(three) and ulna (one) (Table I). All patients had unifocal
lesions except one in whom both tibiae and the ulna were
involved (see Fig. 5). There were two type-1b lesions (Fig.
1), three type-II (Fig. 2), three type-III (Figs 3 and 4), 14
type-IV (Fig. 5) and two type-V (Figs 6 and 8). There were
no type-Ia or type-VI lesions in this group. Metaphyseal
lesions did not show radiological extension into the epiphysis
and those in the epiphysis did not penetrate the articular
surface or extend into the metaphysis. There were no pathological fractures.
THE JOURNAL OF BONE AND JOINT SURGERY
PRIMARY SUBACUTE HAEMATOGENOUS OSTEOMYELITIS IN CHILDREN
95
Fig. 2a
Fig. 3
Radiographs showing cortical thickening with a central radiolucent
nidus in the left tibial diaphysis (type-III lesion).
Fig. 2b
Radiographs showing a) a radiolucent cavity in the proximal
tibia with cortical erosion and marginal sclerosis (type-II
lesion) and b) six months later, with healing of the metaphysis
and mild sclerosis after curettage and administration of
cloxacillin.
Technetium bone scans were positive in all cases, and
helped to identify the two other sites in the patient with
multifocal lesions.
MRI showed replacement of normal fatty marrow with a
high signal intensity on T2-weighted images in two patients
with diaphyseal type-IV lesions (Fig. 7). CT helped to localise the extent of an epiphyseal lesion in the femoral head
(Fig. 8). In patients with epiphyseal lesions there was no
evidence of disturbance of the articular surface at follow-up
(Fig. 6b). Radiological healing was seen at three to six
months. Residual sclerosis was observed in metaphyseal lesions (Fig. 2b); all epiphyseal (Fig. 6b) and diaphyseal infections (Fig. 4b) healed completely.
No extension of metaphyseal lesions into the epiphysis
was found at operation. Epiphyseal lesions were also
VOL. 83-B, NO. 1, JANUARY 2001
Fig. 4a
Fig. 4b
Figure 4a – Radiograph of the right tibia showing cortical
thickening and a central radiolucent nidus (type-III lesion).
Figure 4b – At 11 months there is healing of the cortical
lesion.
96
M. N. RASOOL
Radiographs showing a) a single-layered
periosteal reaction of the shaft of the
right ulna with increased density of the
medullary canal (type-IV lesion) and b)
bilaterally symmetrical periosteal reaction of the shafts of the tibiae with slight
increased density of the medullary canal
(type-IV lesion) in the same patient.
Fig. 5a
Fig. 5b
Fig. 6a
Radiographs showing a) a radiolucent defect with slight marginal
sclerosis in the right distal femoral epiphysis (white arrow) (type-V
lesion) and b) healing of the epiphyseal lesion without articular
involvement or growth disturbance two years later.
Fig. 6b
confined to the epiphysis and did involve the articular
surface. In the cortical lesions granulation tissue was
found in a nidus surrounded by reactive thickened cortex.
The nidus was curetted. In the periosteal lesions a rectangular core of bone consisting of periosteum, cortex and
medullary contents was removed with an osteotome. The
cortex and periosteum were slightly thickened; no pus was
found but there was granulation tissue within the cancellous bone.
Culture of granulation tissue and of medullary curettings
was positive for Staphylococcus aureus in only six patients.
The histological features of chronic pyogenic infection
characterised by inflammatory cells, plasma cells and polymorphonuclear leukocytes were seen in all specimens.
Discussion
Primary subacute haematogenous osteomyelitis remains
uncommon, but the incidence is increasing in the UK compared with that of the acute form. Some authors have reported
a change in the clinical characteristics and epi1,8-10
demiology of osteomyelitis in recent years.
Subacute
osteomyelitis develops when there is an altered host-pathogen
relationship as a result of increased host resistance and
1-3,7
decreased bacterial virulence.
The acute process may also
be masked by antibiotics administered early in the clinical
course. True primary subacute haematogenous osteomyelitis
occurs mainly in children without a history of previous
antibiotic treatment.
THE JOURNAL OF BONE AND JOINT SURGERY
PRIMARY SUBACUTE HAEMATOGENOUS OSTEOMYELITIS IN CHILDREN
Fig. 7
MRI of a type-IV lesion showing a high signal intensity (T2-weighted image) in the marrow of the right
tibial shaft.
Fig. 8
CT shows a circumscribed radiolucent defect in the head of the left
femur (type-V lesion).
Differences in the virulence of strains of staphylococci
have been identified and variations in host susceptibility
15
demonstrated. The clinical course in subacute osteomyelitis
has an insidious onset with mild symptoms. An occasional
history of minor trauma has been noted in this and other
5,9,12,16
and may be regarded as a predisposing factor.
series
VOL. 83-B, NO. 1, JANUARY 2001
97
Laboratory data, apart from a slightly raised ESR, do not
support a diagnosis of infection and the radiological presentation may be suggestive of a benign or malignant neoplasm.
7
3
Roberts et al modified Gledhill’s classification into six
types (Table I) including four basic forms of the disease
occurring in the long bones and spine, defined as metaphyseal, diaphyseal, epiphyseal and vertebral. The vertebral
1,2,5,9,17
form is usually seen in adults.
Some authors have
reported the presence of subacute osteomyelitis in other sites,
2,13,18
6
such as the calcaneum,
pelvis, clavicle and metatarsal
9
bones.
Metaphyseal lesions are the most common and occur
1,2,4,6,7,13
mainly in the tibia.
Type 1a is a punched-out localised zone of lucency and can mimic forms of histiocytosis especially an eosinophilic granuloma. Type 1b resembles a typical Brodie’s abscess with dense sclerotic margins
and may be mistaken for an osteoid osteoma. Type-II cavities
erode the metaphyseal cortex and may resemble osteosarcoma; they may cross the growth plate, but no disturbance
4-6
of growth has been reported. Lesions similar to type II
19
have been reported in cystic tuberculosis and also in fungal
20
infection.
There have been several recent reports of the epiphyseal
form of the condition (type V) which, however, remains
2,6,13,16,21
rare.
The lesions appear as a central lucency, with a
faint sclerotic margin and may resemble a chondroblastoma.
They usually occur within the epiphysis in children less than
six years old and do not cross the epiphyseal plate. Similar
19
lesions have been reported in tuberculosis.
Subacute diaphyseal osteomyelitis has received less attention in the literature than the other forms. Most reports mainly
1,2,5,7
involve adults.
There are two types, cortical (type III)
and periosteal (type IV). In this study the commonest site was
the diaphysis, usually of the tibia. Similar findings have been
14
reported by Hoffman et al from South Africa. This is at
variance with previously reported studies on the distribution
of subacute osteomyelitis in which the lesions were mainly in
1,2,4,6,7,13
the metaphysis.
In this series the cortical form (type III) consisted radiologically of a localised cortical and periosteal reaction with a
small central lucency resembling an osteoid osteoma. Diaphyseal osteomyelitis, characterised by a linear periosteal reaction
(type IV), was the commonest subtype. This was single (eight
cases) or finely laminated (four cases) with only a slight
increase in the density of the medullary canal on plain
radiographs. There was bilateral symmetrical tibial involvement in one patient (Fig. 5b), which has not been previously
reported. Unlike other studies, the diaphyseal lesions showed
a less aggressive periosteal reaction. According to Harris and
1
Kirkaldy-Willis diaphyseal involvement indicated diffuse
22
local involvement, starting in the metaphysis. Cole described diaphyseal forms as aggressive, destructive lesions with
the formation of ‘onion-skin’ or ‘sunburst’ subperiosteal new
2
bone. King and Mayo also described lesions with a considerable amount of layered periosteal reaction. The radiological
picture in type IV is indistinguishable from that of round-cell
98
M. N. RASOOL
tumours of bone especially Ewing’s sarcoma, and also from
leukaemia. This differential diagnosis has been
6,14,23
emphasised.
Primary subacute osteomyelitis commonly involves a sin22
gle bone, and Cole emphasised that it is distinct from
chronic multifocal osteomyelitis of which the recurrent type
and the SAPHO syndrome are the forms which it may
resemble. These conditions are usually recurrent and may be
22,24-28
associated with skin infections.
Although multifocal
13,14
lesions were seen in this and other studies,
the lesions
resolved with treatment, and were not recurrent or associated
with skin changes.
The recommended treatment for subacute osteomyelitis
with a lucent lesion or nidus has been curettage, biopsy and
culture followed by immobilisation and antibiotics. In diaphyseal lesions with a periosteal reaction a core of bone should
be taken which includes periosteum, cortex and medullary
contents. The diagnosis is usually confirmed by histological
examination which shows an inflammatory component with
scattered lymphocytes, plasma cells and granulation tis2,4,6,7,14
Staphylococcus aureus is regarded as the causasue.
tive organism by most authors. It is, however, difficult to
identify and was positively isolated in only 29% of our
14,21,22,24
patients, which is similar to other reports.
Only
7
Roberts et al have reported a higher incidence of positive
culture (61%), and this may have been due to the larger
number of metaphyseal lesions in their series from which
more material for culture could be obtained. There are single
16
reports of other rare organisms such as Pneumococcus,
3
6
Klebsiella and Kingella kingae.
Cloxacillin is the antibiotic of choice in the treatment of
subacute osteomyelitis and is given orally for six weeks after
an initial intravenous course for up to five days.
12
13
18
Bogoch et al, Ross and Cole and Hamdy et al suggest
that antibiotics alone may be adequate and that surgery
should be reserved for “aggressive lesions” and those which
do not respond to antibiotics. It is generally agreed that
treatment should not be started until exploration and biopsy
2,3,5-7,14
have been performed.
Tuberculosis and fungal lesions
are common in South Africa and can mimic subacute
osteomyelitis.
In this series there were no late complications such as
chronic osteomyelitis, growth changes, pathological fractures
or recurrence. Although solitary metaphyseal lesions have
been shown to cross the growth plate in other studies in this
series no disturbances of growth or compromise of joint
2-6,12,13
Similarly, as observed previfunction were found.
2,7,21
ously,
epiphyseal lesions were confined to the epiphysis
with no connection with the metaphysis or the joint margin.
13
Only Ross and Cole have reported arrest of growth in a
child with metaphyseal and epiphyseal lesions in the proximal femur.
True primary subacute haematogenous osteomyelitis represents a favourable host-pathogen response and the diaphyseal
form appears to be more commonly seen in children in South
Africa. It is important to re-emphasise that it is a clinical
entity which is distinct from the acute form and from those
types in which the clinical presentation has been modified by
the administration of antibiotics. These lesions are frequently
confused with a variety of benign and malignant bone
tumours and non-pyogenic infections.
The author thanks B. Katia for the typing of this manuscript.
No benefits in any form have been received or will be received from a
commercial party related directly or indirectly to the subject of this article.
References
1. Harris NH, Kirkaldy-Willis WH. Primary subacute pyogenic osteomyelitis. J Bone Joint Surg [Br] 1965;47-B:526-32.
2. King DM, Mayo KM. Subacute haematogenous osteomyelitis. J Bone
Joint Surg [Br] 1969;51-B:458-63.
3. Gledhill RB. Subacute osteomyelitis in children. Clin Orthop 1973;96:
57-69.
4. Kandel SN, Mankin HJ. Pyogenic abscess of the long bones in
children. Clin Orthop 1973;96:108-17.
5. Cabanela ME, Sim FH, Beabout JW, Dahlin DC. Osteomyelitis appearing as neoplasms: a diagnostic problem. Arch Surg 1974;109:68-72.
6. Lindenbaum S, Alexander H. Infections simulating bone tumours: a
review of subacute osteomyelitis. Clin Orthop 1984;184:193-203.
7. Roberts JM, Drummond DS, Breed AL, Chesney J. Subacute
hematogeneous osteomyelitis in children: a retrospective study. J
Pediatr Orthop 1982;2:249-54.
8. Winters JL, Cahen I. Acute haematogenous osteomyelitis. J Bone
Joint Surg [Am] 1960;42-A:691-704.
9. Jones NS, Anderson DJ, Stiles PJ. Osteomyelitis in a general
hospital: a five-year study showing an increase in subacute osteomyelitis. J Bone and Joint Surg [Br] 1987;69-B:779-83.
10. Craigen MAC, Watters J, Hackett JS. The changing epidemiology of
osteomyelitis in children. J Bone Joint Surg [Br] 1992;74-B:541-5.
11. Bryson AF. The problems of orthopaedic surgery in Northern Nigeria.
Postgrad Med J 1962;38:99-104.
12. Bogoch E, Thompson G, Salter RB. Foci of chronic circumscribed
osteomyelitis (Brodie’s abscess) that traverse the epiphyseal plate. J
Pediatr Orthop 1984;4:162-9.
13. Ross ERS, Cole WG. Treatment of subacute osteomyelitis in childhood. J Bone Joint Surg [Br] 1985;67-B:443-8.
14. Hoffman EB, de Beer JD, Keys G, Anderson P. Diaphyseal primary
subacute osteomyelitis in children. J Pediatr Orthop 1990;10:250-4.
15. Adlam C, Pearce JH, Smith H. Virulence mechanisms of staphylococci grown in vivo and in vitro. J Med Microbiol 1970;3:147-55.
16. Andrew TA, Porter K. Primary subacute epiphyseal osteomyelitis: a
report of three cases. J Pediatr Orthop 1985;5:155-7.
17. Waldvogel FA, Vasey H. Osteomyelitis: the past decade. New Engl J
Med 1980;303:360-70.
18. Hamdy RC, Lawton L, Carey T, Wiley J, Marton D. Subacute
hematogenous osteomyelitis: are biopsy and surgery always indicated?
J Pediatr Orthop 1996;16:220-3.
19. Rasool MN, Govender S, Naidoo KS. Cystic tuberculosis of bone in
children. J Bone Joint Surg [Br] 1994;76-B:113-7.
20. Fialk MA, Marcove RC, Armstrong D. Cryptococcal bone disease: a
manifestation of disseminated cryptococcosis. Clin Orthop 1981;
158:219-23.
21. Green NE, Beauchamp RD, Griffin PP. Primary subacute epiphyseal
osteomyelitis. J Bone Joint Surg [Am] 1981;63-A:107-14.
22. Cole WG. The management of chronic osteomyelitis. Clin Orthop
1991;264:84-9.
23. Willis RB, Rozencwait R. Pediatric osteomyelitis masquerading as
skeletal neoplasia. Orthop Clin North Am 1996;27:625-34.
24. Gamble JG, Rinsky LA. Chronic recurrent multifocal osteomyelitis:
a distinct clinical entity. J Pediatr Orthop 1986;6:579-84.
25. Murray SD, Kehl DK. Chronic recurrent multifocal osteomyelitis: a
case report. J Bone Joint Surg [Am] 1984;66-A:1110-2.
26. Björkstén B, Gustavson KH, Eriksson B, Lindholm A, Nordström
S. Chronic recurrent multifocal osteomyelitis and pustulosis palmoplantaris. J Pediatr 1978;93:227-31.
27. Jurik AG, Helmig O, Ternowitz T, Moller BN. Chronic recurrent
multifocal osteomyelitis: a follow-up study. J Pediatr Orthop 1988;
8:49-58.
28. Letts M, Davidson D, Birdi N, Joseph M. The SAPHO syndrome in
children: a rare cause of hyperostosis and osteitis. J Pediatr Orthop
1999;19:297-300.
THE JOURNAL OF BONE AND JOINT SURGERY