The Limping Child: an Algorithm to Outrule Musculoskeletal Sepsis

1
The Limping Child an Algorithm to Outrule
Musculoskeletal Sepsis

• R. A. Delaney
• Final Year Medicine, University College Cork.
• Department of Trauma Orthopaedics,
• Cork University Hospital.

2
Introduction
The Limping Child

• 2001-2003 19.3 of CUH orthopaedics admissions
  aged under 16 years were limping children
• Diagnostic challenge
• Kocher et al J Bone Joint Surg Am.
  2004861629-35
• Conditions with different management prognosis
  present a similar clinical picture

3
Background
The Limping Child

• Current policy at CUH is to admit all limping
  children for observation investigation - most
  efficient strategy?
• Predictive models developed elsewhere are region-
  and even institution-specific
• Luhmann et al. J Bone Joint Surg Am.
  200486-A956-62.

4
Aims
The Limping Child

• Identify which variables are predictive of
  infection in the child presenting with a limp
• Create a clinically useful algorithm of these
  variables to exclude musculoskeletal sepsis
  from the differential diagnosis

5
Methods
The Limping Child

• Retrospective review of medical records of all
  patients aged 0 15 years who presented to CUH
  with acute atraumatic limp, hip or knee pain over
  a 3 year period
• Source Hospital In-Patient Enquiry (HIPE)
• Statistical analysis using Minitab Release 14

6
Descriptive Data
7
Main Diagnoses
The Limping Child
Total no. of admissions 304 286 patients
8
Slipped Capital Femoral Epiphysis
The Limping Child
9
Slipped Capital Femoral Epiphysis
The Limping Child
10
Perthes Disease
The Limping Child
11
Osteomyelitis
12
Main Diagnoses
The Limping Child

• Total number of cases 304
• 231 Transient synovitis
• 30 Slipped capital femoral epiphysis
• 15 Infection 10 septic arthritis
• 5 osteomyelitis
• 9 Legg-Calve-Perthes disease
• 19 other includes toddlers , osteochondritis
  dissecans, eosinophilic granuloma of ischium
• 259 ?unneccessary admissions for a total of 779
  days

13
Gender
The Limping Child

• 213 boys, 91 girls
• Male Female 2.31
• Transient synovitis 31
• SCFE 11.3
• Perthes disease 81
• Infection 1.51, p 0.532

14
Age
The Limping Child

• Mean Age
• Infection
• 66.9 months
• No Infection
• 72.4 months
• p 0.623

15
Length of Hospital Stay
The Limping Child

• Average length of stay 2.9 nights
• Range 1 24 nights
• 125 (41) patients stayed 1 night
• Range Mode
• Transient synovitis 1-15 nights 1
• Septic arthritis 6-23 nights
  16
• Osteomyelitis 5-24 nights
  16

16
Statistical Analysis
17
Univariate Analysis - history
The Limping Child
18
Univariate Analysis - investigations
The Limping Child
19
Univariate Analysis - CRP
The Limping Child

• C-reactive protein assay only available 09.00 -
  17.00
• 68 of the 304 admissions had CRP data
• 11 (87) bone/joint infection
• 57 (24) no infection
• Mean CRP 54.5 g/dl infection
• 6.5 g/dl no infection
• P-value 0.040

20
The Limping Child
TEMPERATURE ON ADMISSION (oC)
Mean 37.9oC
Mean 36.6oC
p lt 0.001
No Infection
Bone/Joint Infection
GROUP
DESCRIPTION
21
The Limping Child
TEMPERATURE ON ADMISSION (oC)
No Infection
Bone/Joint Infection
GROUP DESCRIPTION
22
Multivariate Regression Analysis
The Limping Child

• Duration of Symptoms at Presentation
• greater than 1 but less than 5 days
• Temperature on Admission
• gt37.0oC
• ESR
• gt35 mm/h
• CRP not included in multivariate regression
  analysis as this would not reflect clinical
  practice at CUH numbers were not sufficient

23
The Limping Child
24
Conclusions
The Limping Child

• Musculoskeletal sepsis can be ruled out with 99
  certainty in a child with none of the 3
  predictors
• A policy to discharge all children with zero
  predictors during the study period would have
  resulted in
• No case of true infection missed
• 154 (51) patients saved admission
• 462 bed-days saved

25
Conclusions

• Savings on
• Stress disruption to children families
• Admissions
• Unnecessary investigations
• Money
• Prospective validation of this predictive model
  in a new patient population will determine its
  value in clinical practice

26
Acknowledgements

• Mr. A. McGuinness, FRCS (Tr Orth)
• Consultant Orthopaedic Surgeon,CUH/SMOH
• Mr. J. Street, PhD FRCS (Tr Orth)
• Mr. B. Lenehan, MCh
• Ms. L. OSullivan, MSc, Statistician
• Ms. C. Twomey, Medical Records, CUH

27
Thank You