The limping child: epidemiology, assessment

and outcome
S. U. Fischer, T. F. Beattie
From the Royal Hospital for Sick Children, Edinburgh, Scotland

e investigated the epidemiology, assessment and

outcome of acute atraumatic limp in 243
children under the age of 14 years presenting to a
paediatric accident and emergency department (AED)
over a period of six months. Data were collected at
presentation and medical notes were re-examined after
18 to 21 months.
The incidence of limp was 1.8 per thousand. The
male:female ratio was 1.7:1 and the median age 4.35
years. Limp was mainly right-sided (54%) and painful
(80%); 33.7% of the children had localised pain in the
hip. A preceding illness was found in 40%. The main
diagnosis was irritable hip/transient synovitis
(39.5%); Perthes disease accounted for 2%. Most
patients (77%) were managed entirely in the AED.
Acute atraumatic limp is a common problem in
children presenting to the AED. Most can be safely
managed there if guidelines are followed and will have
a benign outcome. Further studies are needed to
identify the role of preceding illness in the aetiology of
acute atraumatic limp.

J Bone Joint Surg [Br] 1999;81-B:1029-34.

Received 2 December 1998; Accepted after revision 12 May 1999

Limping is a common reason for children to present to

accident and emergency departments (AED). A proportion
will have a preceding history of injury, but often this is
absent. The main concerns are not to miss serious pathology and to begin appropriate management for the underlying
condition. Potentially serious diseases include bone or joint
sepsis, primary or metastatic tumours of bone, Legg-CalvPerthes disease (Perthes disease) and slipped femoral capi-

S. U. Fischer, MD, Senior House Officer, Accident and Emergency Care

T. F. Beattie, FRCS Ed, Consultant in Accident and Emergency Care
Accident and Emergency Department, Royal Hospital for Sick Children,
Sciennes Road, Edinburgh EH9 1LF, UK.
Correspondence should be sent to Mr T. F. Beattie.
1999 British Editorial Society of Bone and Joint Surgery
0301-620X/99/69607 $2.00
VOL. 81-B, NO. 6, NOVEMBER 1999

tal epiphysis (SFCE). The list of differential diagnoses

1
includes a wide spectrum of diseases (see Table VIII).
Previous studies have either been retrospective or have
2-13
evaluated diagnostic criteria or investigations.
Most have
concentrated on specific diseases such as irritable hip, which
includes transient synovitis/toxic synovitis or orthopaedic
2-12,14,15
Often, however, children do not present to
infections.
their general practitioner or the AED with a specific diagnosis
and doctors faced with the limping child must make an
appropriate assessment and ensure correct management. Tay12
16
13,17
lor and Clarke, Fink et al, and others,
have developed
an algorithm for reducing admissions from this condition with
particular reference to painful hips, most of which have a
benign course. There is, however, no prospective study which
has examined children with an atraumatic limp who present
to the AED and, as a result, a systematic approach for this
condition has not been established.
We have therefore examined the problem of atraumatic
limping with a view to establishing the pathology to be considered and to determine a safe and cost-effective method
of assessment.

Patients and Methods

The Royal Hospital for Sick Children in Edinburgh is a
teaching hospital providing paediatric services to the city and
its surroundings. The area has a population of 764 600 of
18
which 135 796 (17.8%) are below the age of 14 years. There
are 69 506 boys and 66 290 girls. This is the only paediatric emergency department in the city with an annual attendance in 1996 of 28 200 patients under 14 years of age.
We studied all children with a history of atraumatic limp
who presented to the AED between 1 January 1996 and 30
June 1996 either by self referral or from their general
practitioner. The criteria for inclusion and exclusion are
given in Table I. The protocol involved answering a questionnaire which was attached to the clinical notes. It included general details together with relevant clinical data.
Additionally, the attendance record book and the computerised attendance list of the AED were searched daily to
ensure that all limping children were registered for the
study. If details were missed on the initial presentation,
completion was achieved at the following attendance or
obtained by telephone.
1029

1030

S. U. FISCHER,

Table I. The criteria for inclusion and exclusion into the study
Criteria for inclusion
Age between one and 14 years (provided they had started walking)
Acute hip, groin, thigh, knee or lower-leg pain
Painful or painless limp
Abnormality in gait, refusal to walk or bear weight
Any acute limp regardless of a pre-existing orthopaedic
problem on the contralateral side
Criteria for exclusion
Major or minor trauma resulting in immediate symptoms
Any other limb trauma causing the above symptoms
Long-standing orthopaedic, neurological or other problems,
regardless if ipsilateral or contralateral

Protocols and guidelines for the management of the

limping child are available for all staff in the AED. The
approach currently applied was used for this study. A
history is taken and any limping child without preceding
trauma is included. The general condition of the child is
assessed and a full clinical examination made. The temperature is taken on arrival using an axillary thermometer;
a full blood count (FBC) and the ESR are taken unless the
child is symptomless when first seen. If the problem is
localised, further investigations are arranged. If abnormality in the hip is suspected, ultrasonography of the hips is
carried out by a consultant radiologist. Transient synovitis
is diagnosed if an effusion, synovial thickening or both are
present on this examination. Irritable hip is defined as any
abnormality of the hip found on history or clinical examination, but not on assessment by imaging. Plain radio-

Back
Hip
Knee
Lower limb*
Not localised
None
Total

Number

Percentage

4
82
47
44
16
50

1.6
33.7
19.3
18.1
6.6
20.6

243

graphs of the hip or affected anatomical site are obtained if

an orthopaedic problem is suspected or if the ultrasound is
negative. Hip or joint aspiration is only considered if there
is sufficient clinical evidence of septic arthritis, but is not
carried out in the AED.
After the initial clinical assessment and investigation,
each child is assigned a working diagnosis and treated
accordingly. Indications for admission of a child with limp
include severe pain and/or significant pathology, such as
12
bone or joint sepsis, neoplasia or SCFE. Conditions such
as Perthes disease, Khlers disease, transient synovitis or
irritable hip can be followed up in the outpatient department of the AED or by the relevant specialty service. All
others, including soft-tissue injury (STI) and upper respiratory tract infection (URTI), are followed up in the AED and
reviewed within five days by a senior member of the
medical staff.
Each child was entered into the study and counted as one
entry. If the symptoms settled and the child was discharged,
but presented again with a limp to the AED within the
study period, it was registered as a new entity regardless of
whether attendance was for a similar or a new and different
problem. All the medical notes were re-examined from
May until September 1997 to complete the details of
subsequent investigations, arrangements for follow-up, the
final diagnosis and the date of discharge.

Results
During the period of the study 244 affected children were
seen. The records of one were incomplete and this patient
was excluded from the revue. This left 243 children with
complete data available at least one year after their initial
presentation. Three attended twice, but all had problems
which had completely resolved between the two visits.
There were 152 boys (62.5%) and 91 girls (37.5%). The
median age was 4.35 years (25th centile 2.9; 75th centile
7.5). The incidence of atraumatic limp was 180/100 000
during the period studied. The limp was right-sided in 131
patients (54%) and left-sided in 102 (42%). Ten patients
(4%) presented with a limp on both sides.

Table II. The distribution of pain for the 243 children

presenting with a limp
Site of pain

T. F. BEATTIE

100

* refers to rest of leg excluding hip and knee

Table III. The site of pain in relation to the final diagnosis
Site
Final diagnosis
Irritable hip
Transient synovitis
No final diagnosis
Perthes' disease
SCFE
Osteomyelitis
STI/muscular strain
Fracture/toddler's fracture
Others
Total number (%)

Total
number
of patients
39
57
72
5
1
4
38
2
25
243

Back

Hip

Knee

Lower
limb*

Not
localised

1
1
2

26
39
3
4
1
3
6

3
5
15
2
12
10

1
15
2
18
2
6

4 (1.6)

82 (33.7)

47 (19.3)

44 (18.1)

16 (6.6)

2
4
7

2
1

None
7
8
31
1
3
50 (20.6)

* refers to rest of leg excluding hip and knee

THE JOURNAL OF BONE AND JOINT SURGERY

THE LIMPING CHILD: EPIDEMIOLOGY, ASSESSMENT AND OUTCOME

1031

Table IV. Final diagnosis in relation to the duration of symptoms

Final diagnosis

Range of duration
of symptoms (days)

Median duration of symptoms

at presentation (days)

Irritable hip
Transient synovitis
No final diagnosis
Perthes' disease
SCFE
Osteomyelitis
STI/muscular strain
Fracture/toddler's fracture
Others

0 to
0 to
0 to
2 to
30
0 to
0 to
2 to
0 to

1
1
1
30
30
0.5
1
3
2

60
47
152
123
7
90
4
62

Table V. Preceding illnesses recorded in 97 children before

presentation
Preceding illness

Number

Percentage of
total number

Upper respiratory tract infection

(non-specific)

61

25

Earache

13

5.3

Rash

2.5

Tonsillitis

2.1

Gastrointestinal problems

12

Total

97

40

Table VI. Diagnosis of previous limps, unrelated to presenting

limp
Percentage of
total number

Diagnoses of previous limps

Number

Irritable hip
No final diagnosis
Perthes' disease
Osteomyelitis
STI/muscular strain
Fracture/toddler's fracture
Others

19
17
3
1
5
6
11

7.8
7.0
1.2
0.4
2.1
2.5
4.5

Total

62

25.5

Table VII. Outcome in relation to the working diagnosis for the 243 children presenting with a limp

Working diagnosis
Irritable hip
Transient synovitis
Limp ? Cause
Perthes' disease
SCFE
Osteomyelitis
Septic arthritis
STI/muscular strain
Fracture/toddler's fracture
Others
Total (%)

Follow-up
in AED
33
47
30
1
18
6
3
138 (57.0)

Discharged
with no
follow-up
2
1
24
18
1
3

VOL. 81-B, NO. 6, NOVEMBER 1999

Specialty
follow-up

6
6
3

6
3
6
3

1
3
3
-

3
-

49 (20.0)

Pain was a presenting feature in 193 children (80%),

with the site being localised by all but 16 of them. The
remaining 50 patients (20%) had no pain (Table II). The
relationship between the site of pain and the final diagnosis
is illustrated in Table III.
The median time of onset to presentation was one day
(25th centile 0; 75th centile 4). Approximately one-third
presented on the day of onset and most (85%) within one
week. A small number showed a substantial delay in presentation, but were seen within six months (Table IV).
A preceding illness was recorded in 40% (Table V). A
further 39 (16%) had a convincing URTI at presentation
and 25% had had a previous limp, unrelated to that with
which they presented (Table VI).
The working diagnoses on leaving the AED after the first
presentation and assessment are shown in Table VII. Most
(77%) of the patients were discharged with no involvement

Admitted

28 (11.5)

7
28 (11.5)

Total
47
57
63
4
1
3
3
39
7
19
243 (100.0)

of other specialties. About 75% of these were followed up

in the AED and 25% were discharged with no follow-up
being arranged. The remaining 23% were referred to a
specialty department and equal numbers were followed up
as outpatients or were admitted.
Two children initially diagnosed as having an STI/
muscular strain and a possible calcaneal fracture, respectively, developed osteomyelitis. In both, the protocol had
not been followed and the initial investigations were
incomplete. The FBC and ESR had not been determined.
The one child with Langerhans cell histiocytosis was
diagnosed by a consultant radiologist after the patient had
left the department. He was followed up earlier than
planned. No other significant diagnoses were missed on
the first attendance.
On re-examination of the medical notes 18 to 21 months
after first presentation, most children (94%) had been dis-

1032

S. U. FISCHER,

T. F. BEATTIE

Table VIII. Differential diagnoses of acute atraumatic limp

Number
Inflammatory
Transient or toxic synovitis
Irritable hip
Juvenile rheumatoid arthritis
Ankylosing spondylitis
Others:
Reactive arthritis secondary to viral illness
or Henoch-Schnlein purpura
Monoarthropathy (knee)
Arthralagia
Tendinitis/tenosynovitis
Total
Infectious
Osteomyelitis
Septic arthritis
Discitis
Others:
Cellulitis (secondary to eczema)
Inguinal lymphadenitis
Paronychia with cellulitis
Total

57
39

23.5
16.0

4
2
1
1
105

1.6
0.8
0.4
0.4
42.7

1.6

3(2)
1
1
9

1.2
0.4
0.4
3.6

5
1

2.1
0.4

1
1
2
10

0.4
0.4
0.8
4.1

1
1

0.4
0.4

0.8

2
1

0.8
0.4

38
1
1
43

15.7
0.4
0.4
17.7

2
1
3

0.8
0.4
1.2

171

70.4

Developmental or acquired
Legg-Calv-Perthes disease
Slipped capital femoral epiphysis
Developmental hip dislocation
Torsional deformities
Acquired limb-length discrepancy
Others:
Sclerotic lesion of L5 pedicle
Popliteal cyst
Khlers disease
Total

Neoplasia
Malignant tumours
Langerhans' cell histiocytosis
Acute lymphoblastic leukaemia
Benign bone tumours
Total

Trauma or overuse
Stress fracture
Patellofemoral pain/chondromalacia patellae
Spondylolisthesis
Herniated nucleus pulposus
Others:
STI/muscular strain
Osgood-Schlatters disease
Haematoma of thigh
Total

Metabolic
Rickets
Hyperparathyroidism

Haematological
Sickle-cell disease

Others:
Hypermobility syndrome
? Intermittent torsion of testis/?epididymo-orchitis
Total
Total

charged; 6% remained under review. The list of differential

diagnoses of acute atraumatic limping children is presented
in Table VIII. Almost 40% were diagnosed as having
irritable hip/transient synovitis. In about 30% no definite
diagnosis was made.

Percentage

Discussion
In this prospective study every 58th child who was seen in
the AED had an acute atraumatic limp which was, therefore, a common and significant problem.
THE JOURNAL OF BONE AND JOINT SURGERY

THE LIMPING CHILD: EPIDEMIOLOGY, ASSESSMENT AND OUTCOME

The main diagnosis was irritable hip/transient synovitis

which has a benign outcome. In about 30% of patients no
final diagnosis was made but all recovered without intervention, and had no subsequent problems.
The age distribution was approximately the same for
both sexes with an increased incidence in the group aged
under five years. More boys than girls were affected. This
may be a reflection of the male predisposition to injury or
18,19
other illnesses.
Most experienced pain, but this was localised to the hip
in less than half. Approximately 10% of patients with
positive radiological findings in the hip had no hip pain,
indicating that the site of pain is not a reliable indicator of
the site of pathology.
Most children presented on the day when symptoms
started, but in a significant few there was a notable delay.
Some of these had visited their general practitioner. None
had previously been seen with the same problem in the
AED or other departments of the hospital. In this group
Perthes disease and SCFE were of significance since these
children tended to present late, as has been noted in pre3,9,13
vious studies.
Infections of the upper respiratory tract are thought to be
particularly important as a predisposing factor in the irrita3,5,7,8,10,15,20,21
ble hip.
This has never been proven by an
identifiable cause or agent, but our results confirm the need
6,22
for further studies.
Around 25% of patients had a history of a previous limp,
mainly due to an irritable hip. One of these was subsequently diagnosed as having Perthes disease. The incidence of the development of Perthes disease after irritable
hip/transient synovitis varies between 0% and 20% in
3,6,7,15,20,23,24
different studies.
In our study it was 2%.
There are some drawbacks to our definitions. Young
children find it difficult to remember injury and its temporal
relationship to the subsequent development of pain is often
unclear. No final diagnosis and STI/muscular strain were
often diagnosed. Undoubtedly, many of these were attributable to some form of injury, acute and occult. In our study,
we diagnosed two boys, aged 3 and 11 years, respectively,
with stress fractures of their second metatarsal bone and no
history of trauma. No other fracture or toddlers fracture
was diagnosed late, indicating good agreement with our
definition.
No patient presented with septic arthritis. This may be
due to the excellent uptake (>95%) of Haemophilus influ25
enzae type b vaccine in our area.
There were no difficulties in diagnosing Perthes disease
or SCFE using our protocol. The incidence of SCFE may
be low because of the age cut-off. In two children the
diagnosis of osteomyelitis was delayed and, in retrospect,
their management was not in accordance with our
guidelines.
The protocol proved to be effective in detecting two
other patients with serious conditions, acute lymphoblastic
leukaemia and Langerhans cell histiocytosis, both with
VOL. 81-B, NO. 6, NOVEMBER 1999

1033

significant findings on radiographs of the hip. The latter

was diagnosed by routine reporting. This supports the view
that radiographs taken in the AED should be reported on by
a senior radiologist.
Most patients (94%) were discharged. The remainder
were still being followed up by the appropriate specialty
service in September 1997. More than 75% of the children
presenting with an acute atraumatic limp were managed
entirely in the AED, without major invasive investigation
or expensive involvement of other specialties.
We believe that the protocol used as described under
methods is a cost-effective approach which, when followed
correctly, has proved to be safe and accurate.
We would like to thank Mr Malcolm Macnicol for his advice and also Dr
Richard K Ferguson, Debbie Hamilton, Louise Cowan, Medical Records
and the medical and nursing staff from the AED for their professional help
and continuing support.
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. Dabney KW, Lipton G. Evaluation of limp in children. Curr Opin
Pediatr 1995;7:88-94.
2. Briggs RD, Baird KS, Gibson PH. Transient synovitis of the hip
joint. J R Coll Surg Edinb 1990;35:48-50.
3. Bickerstaff DR, Neal LM, Brennan PO, Bell MJ. An investigation
into the etiology of irritable hip. Clin Pediatr Phila 1991;30:353-6.
4. Del Beccaro MA, Champoux AN, Bockers T, Mendelman PM.
Septic arthritis versus transient synovitis of the hip: the value of
screening laboratory tests. Ann Emerg Med 1992;21:1418-22.
5. Haueisen DC, Weiner DS, Weiner SD. The characterisation of transient synovitis of the hip in children. J Pediatr Orthop 1986;6:11-7.
6. Landin LA, Danielsson LG, Wattsgard C. Transient synovitis of the
hip. J Bone Joint Surg [Br] 1987;69-B:238-42.
7. Sharwood PF. The irritable hip syndrome in children: a long-term
follow-up. Acta Orthop Scand 1981;52:633-8.
8. Vidigal EC, da Silva OL. Observation hip. Acta Orthop Scand 1981;
52:191-5.
9. Kunnamo I, Kallio P, Pelkonen P, Hovi T. Clinical signs and
laboratory tests in the differential diagnosis of arthritis in children. Am
J Dis Child 1987;141:34-40.
10. Kunnamo I. Infections and related risk factors of arthritis in children:
a case control study. Scand J Rheumatol 1987;16:93-9.
11. McCarthy PL, Wasserman D, Spiesel SZ, Dolan TF, Jekel JF.
Evaluation of arthritis and arthralgia in the pediatric patient. Clin
Pediatr Phila 1980;19:183-90.
12. Taylor GR, Clarke NM. Management of irritable hip: a review of
hospital admission policy. Arch Dis Child 1994;71:59-63.
13. Choban S, Killian JT. Evaluation of acute gait abnormalities in
preschool children. J Pediatr Orthop 1990;10:74-8.
14. Gavalas M, Potts H, Galasko CS. Bone infection and the limping
child in the Accident & Emergency Department: a diagnosis to be
considered. Arch Emerg Med 1992;9:323-5.
15. Hollingworth P. Differential diagnosis and management of hip pain in
childhood. Br J Rheumatol 1995;34:78-82.
16. Fink AM, Berman L, Edwards D, Jacobson SK. The irritable hip:
immediate ultrasound guided aspiration and prevention of hospital
admission. Arch Dis Child 1995;72:110-4.
17. Abu-Arafeh I, Russell G. Recurrent limb pain in schoolchildren. Arch
Dis Child 1996;74:336-9.
18. Scottish Health Statistics 1996. Information and Statistics Division.
The Natinal Health Service in Scotland ISD Publication Edinburgh,
1996:2.
19. Thompson GH, Scoles PV. Bone and joint disorders Section I:
Orthopaedic problems. In: Behrman RE, Kliegman RM, Arvin AM,
eds. Nelson textbook of pediatrics. 15th ed. Philadelphia: WB Saunders, 1996:1915-42.

1034

S. U. FISCHER,

20. Parsch K. The painful hip joint in the child: differential diagnosis and
therapy of coxitis fugax, Perthes disease and septic coxitis. Padiatr
Padol 1992;27:55-61.
21. Bernd L, Niethard FU, Graf J, Kaps HP. Transient hip joint inflammation (coxitis fugax). Z Orthop Ihre Grenzgeb 1992;130:529-35.
22. Hardinge K. The etiology of transient synovitis of the hip in
childhood. J Bone Joint Surg [Br] 1970;52-B:100-7.

T. F. BEATTIE

23. Kallio P, Ryoppy S, Kunnamo I. Transient synovitis and Perthes

disease. J Bone Joint Surg [Br] 1986;68-B:808-11.
24. Erken EH, Katz K. Irritable hip and Perthes disease. J Pediatr
Orthop 1990;10:322-6.
25. Salisbury DM, Begg N. Haemophilus influenzae Type b (Hib). In:
Immunisation against infectious disease. London: HMSO, 1996:77.

THE JOURNAL OF BONE AND JOINT SURGERY