Design and performance of a fracture monitoring transducer

M. Evans, J. Kenwright* and J.L. Cunningham

Oxford Orthopaedic Engineering Centre, *Nuffield Orthopaedic Centre, Headington, Oxford OX3 7LD, UK

Received February 1987, accepted June 1987

ABSTRACT

The answer to the question, ‘when is a fracture healed?’ is not simple, since the heal+ process is profressiue and it is not possible to specify a time when the fracture can be said to have healed. In the past the assessment of fracture healing has, in the main, been subjctive, relyin upon the skill of the interpreter. A more objective method would be an advantagefor many

reasons, and since the bone is intended to be load bearirg it is reasonable to assess healing by measuring the mechanical integrity of the bone. To do this a ‘clamp on ’ transducer has been developed which, when fitted to the support column of an external fixator, enables the stgfness of a fracture to be determined during the healing process. Over the past Gyears this system has been usedfor both clinical and research work. It has enabled various forms of treatment to be evaluated in terms of ‘rate of

healing’ and it also indicates the safe point at which the fixator can be removed.

Keywords: Musculoskeletal system, bone, fracture healing

INTRODUCTION

Clinical assessment of fracture healing is usually subjective, relying upon the detection of movement (‘feel’) by the surgeon, the patient’s response in terms of pain and confidence, and radiographic evidence of callus and primary bone union.

A more objective method of measuring the strength of a healing fracture would be useful in assessing many aspects, such as the effectiveness of different forms of treatment, the pattern and rate of healing and the stage at which treatment can be discontinued.

DETERMINING THE MECHANICAL STRENGTH OF FRACTURES

Comparing the deflection of a structure with the loads applied to it can indicate its strength, provided that characteristics of the materials and the geometry of the structure are known. This principle is used by clinicians when judging the strength of a fracture by ‘feel’.

In recent years several methods of measuring the deflection, or movement, in long-bone fractures under known loading conditions have been used’-*. In all of these situations direct coupling to the bone had to be achieved to provide the necessary accuracy. In the case of fractures reduced by exter- nal fixation, the bone screws provide the necessary contact with the bone for deflection measurement. With earlier methods of direct measuremen?, it has been necessary to remove the fixator support column in order to take readings, thereby risking re- fracture.

An alternative to measuring ‘fracture movement’ directly is to determine the proportions of the applied load being carried by the external fixator

Reprints from Mr M. Evans

0141-5425/88/010064-06 $03.00 64 J. Biomed. Eng. 1988, Vol. 10, January

and the fracture. With this method the fracture can be considered as a structural member of variable stiff- ness, in parallel with the external fixator (Fz”ure 1). As the bone heals, it will carry proportionally more of the load.

Figure 1 Load distribution through a fracture and an external fixator

Sensing element

Figure 2 The fracture monitoring transducer

One method of detecting the loads in a tixator is by bonding strain gauges directly onto the fixator frame’. However, the complication of instrument- ing each fixator makes this technique unsuitable for long-term patient monitoring. A better solution would be a device which could easily be fitted temporarily to the fixator without requiring any modifications to the fixator system.

We describe a ‘clamp on’ strain gauged transducer which has been used together with the Oxford external fixator’*~” over the past 6 years to monitor fracture healing. This approach provides for the measurement of fracture stiffness and calibration at regular intervals, it also enables comparisons to be made between the progress of individual patients.

FRACTURE MONITORING TRANSDUCER DESIGN CONSIDERATIONS

A number of problems needed careful consideration when designing this transducer. The clamping arrangement is critical since any movement between the transducer and the support column results in substantial errors. To minimize slip between the square section support column (used on the Oxford fixator) and the transducer clamps, small hardened steel balls are set into the clamps to provide a positive ‘bite’ onto the fixator column and a location for subsequent applications. The transducer must withstand handling without damage yet be sensitive to small deflections in the column.

To determine the principal bending moment, deflections in two planes were monitored. Two sens- ing elements were used for this purpose, arranged at 90’ to each other; torsion was also monitored using this arrangement (Figure 2). The design of the transducer sensing elements was a compromise to satisfy conflicting requirements. The material and physical dimensions of the sensing element were selected after considering loading conditions in the fixator and transducer. The deflection produced in the column for a given load produces a proportional deflection in the transducer sensing elements.

In bending, the load induced in the transducer will be proportional to the stiffness of the sensing element, i.e. the moment of inertia (Ze) and the

Fracture moniton’ng tranrducer~ M. Evans et al.

modulus of elasticity (E,). In torsion, the transducer loading is proportional to the polar moment of inertia (Je) and the modulus of rigidity (G,). Therefore to minimize the clamp loading (minimum clamp slip) low values of Z,, E,, Je and G, should be selected. However the signal from the transducer (surface strain in the sensing elements) is proportional to Z, and Je. Therefore, the best compromise is to select a material with a low modulus of elasticity and reasonable section propor- tions. This led to a semicircular sensing element 8.0 mm wide and 2.0 mm thick, with an inside radius of 7.25 mm; the material selected was titanium- 115. The low modulus of elasticity (approximately half that of stainless steel) produced a compliant transducer, whilst the high strength of this material gave the required handling protection and long-term stability.

STRAIN GAUGE SELECTION AND POSITIONING

Initially, the strain gauges used to monitor bending were bonded along the axis of the element on the line of maximum strain. This arrangement gave ‘bend- ing’ signals of approximately 20 microstrain per Nm, but a significant signal was produced when subjected to torsion, assumed to be due to the warp- ing of the element. But, by mounting the gauges in the conventional torsion mode, i.e. four gauges at

d

/

- I IO

Bending moment (Nml

Figure 3 Transducer calibration (bending). Cross talk from q , torsion channel; and H, opposite bending channel

J. Biomed. Eng. 1988, Vol. 10. January 65

Fracture monitorirg transducer: M. Evans et al.

45O the axis of the element, an appreciable degree of compensation was achieved at the expense of loss of signal. This was considered acceptable in order to reduce ‘cross-talk’. The gauges finally selected were MM: EA 05062TH 120 (Welwyn Strain Measure- ment Ltd), all being taken from the same batch to minimize thermal drift.

Six strain gauges were mounted symmetrically about the centre line of each element, four on the outside of the curve and two on the inside. A full bridge was used to monitor the bending loads and two gauges from each element formed the ‘torsion bridge’.

TRANSDUCER PERFORMANCE

The final design of the transducer produced a unit which gives a linear response to within 1% on all three channels, with a cross-talk of less than 5% between channels when sensing bending, and less than 15 % when sensing torsion. The sensitivity in bending is 14.5 microstrain per Nm and in torsion 18.6 microstrain per Nm applied torque to the lixator column. The results are presented in Figures 3 and 4. Thermal drift is less than 2 microstrain per Nm per OC between 0 and 40°C. Repeated applica- tion and removal of the transducer produces an error of no more than 5%.

APPLICATIONS AND RESULTS

Over the past 6 years the healing of over 100 tibia1 fractures has been monitored using the transducer, which has also been used in both experimental” and clinical’ trials to assess the effect of a small controlled micromovement on the rate of fracture

Torsion (N/m)

Figure 4 Transducer calibration (torsion); q , cross talk from bending channels

66 J. Biomed. Eng. 1988, Vol. 10, January

Figure 5 Patient testing

Load F

1

Figure 6 Fixator calibration

Transducer f

----t

Defl

F= TX : K = Colibrotion factor

Fixotor stiffness = - I S, AY

:. Ay 1 3 sx (1)

signal

Fracture monitoring transducer: M. Evans et al.

I F Lood

Transducer sign01 T

----c

System stiffness = S,

Fixotor stiffness = S,

Fracture stiffness = Sz

s, = s,+ s, = 5 (2)

s, = 5 - s,

( Fracture stiffness 1

From equotion I Ay = - Sx

Fs, :. s, = - - s,

sz&-l] (3)

Figure 7 Fracture stiffness calculation

5000

healing. The progress of knee arthrodeses and leg lengthening have also been monitored using this transducer. The method used to obtain measurements of fracture stiffness, and some selected results, are presented below.

The precalibrated transducer is fitted to the patient’s fixator and the three channel amplifier is zeroed with no load applied to the fractured limb. In the bending test (Fz&re 5) the patient’s leg is supported in a cradle standing on a Kistler forceplate and a downward load is applied manually just above the knee. In the axial loading test the patient’s foot rests on the forceplate with the patient in the normal sitting position, and a downward load is applied to the knee.

The signals from the transducer and the forceplate are digitized and fed to a Digital Equipment Corporation PDP 11/34 computer for subsequent analysis. By modelling the geometry of the patient’s fixator, a value of fixator axial stiffness (F&.oz 6) and bending stiffness can be found. Using the fixator stiffness values the fracture stiffness can be found directly from the transducer and forceplate signals as shown in Figure 7.

Initial experience suggested that there were threshold values of fracture stiffness (Figure 8) above which the ftxator could be safely removed and weight bearing permitted, with a very low risk of refracture. Below these threshold values the incidence of refracture increased dramatically. Hence, if the fracture stiffness values obtained from a patient exceeded these threshold values, the frac- ture was said to be ‘clinically consolidated’ and the fixator could be removed. However, care must be exercised in the interpretation of the fracture

4000

- 1000

(

-F 75

60

‘

I

c A,

Figure 8 Fracture stiffness results up to fixator removal. +, Safe for free weight bearing;), high incidence of refracture

J. Biomed. Eng. 1988, Vol. 10, January 67

Fracture monitoring transducer: M. Evans et al.

Time after injury (weeks)

Figure 9 Fracture stiffness results for patient with closed tibia1 fracture. a, Clinical consolidation; -, axial stiffness; .“, bending stiffness

Time after qury (weeks)

Figure 10 Fracture stiffness results for patient with comminuted tibia1 fracture with puncture wound. q , Clinical consolidation; -, axial stiffness; ..., bending stiffness

IOOC

8OC

1 60C

E

d 4OC

2oc

I I I . . . I I I

0 4 8 12 16 20 24 28 Time after injury (weeks)

Figure 11 Fracture stiffness results for patient with open comminuted tibia1 fracture. q , Clinical consolidation; -, axial stiffness; ..., bending stiffness

stiffness results if loose bone screws are suspected. Churches et ~1.” showed that if one bone screw was loose then the fixator stiffness decreases by 23%. This decrease in fixator stiffness allows more of the applied load to be taken by the fracture, and will produce a false indication of fracture healing.

Typical stiffness results for healing tibia) fractures with differing initial injury grades are given in Figures 9-11. Figure 9 shows the fracture stiffness results for a patient who sustained a closed mid-shaft tibia1 fracture with a butterfly fragment. The frac- ture stiffness values for clinical consolidation were reached at 16 weeks after fracture when the fixator was removed. Fracture stiffness results for a patient who sustained a cornminuted tibial fracture with a puncture wound of < 1 cm are given in Figure 10. Radiographs of this fracture taken shortly after being stabilized by external fixation and immediately before fixator removal, at 21 weeks after fracture, are shown in Figure 12. Finally, the stiffness results for a patient with a cornminuted fracture with a wound of > 1 cm are shown in Figure 11. Fracture stiffness values for clinical consolida- tion were achieved at 24.5 weeks after fracture, and the fixator was removed at 27 weeks.

Figure 12 Radiographs of fracture for patients shown in Figure 10. Left, at the time of application of fixator; right, at 20 weeks post operation

66 J. Biomed. Eng. 1988, Vol. 10, January

CONCLUSIONS Stand 1972; 43: 264-75.

The transducer described in this paper has been used for the monitoring of fracture healing in over 100 patients during the past 6 years at Oxford, it has proved to be reliable and has become an accepted measurement technique for assessing the mechanical integrity of a healing fracture. The abiiity to attach the transducer to any number of patients has enabled comparative data to be compiled for a wide range of injuries and can provide valuable informa- tion on different forms of fracture management using external fixation.

5.

6.

7.

8.

REFERENCES

9.

10.

11.

12.

13.

Jorgensen TE. Measurements of stability of crural fractures treated with Hoffman osteotaxis. 4: The complicated terminal phase of healing of crural fractures. Acta Orthop

Stand 1972; 43: 280-91. Burny F, Bourgois R, Donkerwolcke M, Moulart F. Utilisation clinique de janges de constianinte - situation actuelle et perspectives d’avenir. Attu Orthop B& 1978; 44:

895-920. Jernberger A. Measurement of stability of tibia1 fractures. Acta Orthop Scandinauica 1970; 135 (suppl.). Beaupre GS, Hayes WC, Jofe MH, White III AA. Monitoring fracture site properties with external fixation. J Biomech Eng 1983; 105: 120-6. Kenwright J, Richardson JB, Goodship AE et al. Effect of controlled axial micromovement on healing of tibia1 frac- tures. Lancet 1986; ii: 1185-7. Churches AE, Tanner KE, Harris JD. The Oxford exter- nal fixator: fixator stiffness and the effect of bone pin loosening. Erg Med 1985; 14: 3-l 1.

Goodship AE, Kenwright J. The influence of induced micromovement upon the healing of experimental tibia1 fractures. J Bone Jt Surg 1985; 67: 650-5. Evans M, Kenwright J, Tanner KE. Analysis of single sided external fracture fixation. Erg Med 1979; 8: 133-7. Kenwright J, Harris JD, Evans M. External fracture fixa- tion - the analysis and design of a new system. Erg Med 1979; 8: 138-42.

Lippert FG, Hirsch C. The three dimensional measure- ment of tibia1 fracture motion by photogrammetry. Clin Orthop 1974; 105: 30-43. Jorgensen TE. Measurements of stability of crural fractures treated with Hoffmann osteotaxis. 1: Method and measure- ment of deflection on autopsy crura. Acta Orthop Stand 1972; 43: 188-206. Jorgensen TE. Measurements of stability of crural fractures treated with Hoffman osteotaxis. 2: Measurement of crural fractures. Acta Orthop &and 1972; 43: 207-18. Jorgensen TE. Measurements of stability of crural fractures treated with Hoffman osteotaxis. 3: The uncomplicated terminal phase of healing of crural fractures. Acta Orthop

Fracture monitoring transducer: M. Evans et al.

@ British Crown copyright 1988

J. Biomed. Eng. 1988, Vol. 10, January 69