Europearl Journal ofRadiology, 14 (1992) 185-188 Ci 1992 Elsevier Science Publishers B.V. All rights reserved. 0720-048X/92/$05.00 185

EURRAD 00236

Leg length inequality measured by ultrasound and clinical methods

Stanislaw Junk, Terje Terjesen, Ivar Rossvoll and Martinus BrAten Department of Orthopaedic Surgery, Trondheim University Hospital. Trondheim. Norway

(Received 31 July 1991: accepted after revision 13 September 1991)

Key Words: Ultrasound, leg length inequality; Measurement, ultrasound

Abstract

Leg length inequality (LLI) was determined by ultrasound and two clinical methods in 100 subjects without previous disorders of the lower extremities. The mean LLI by ultrasound was 4.1 mm (range: O-16 mm). An LLI of 10 mm or more was found in 4% of the subjects. The mean inter-observer variation by ultrasound was 0.9 mm (range: + 5-8 mm), which was significantly less than that obtained by clinical methods. The 95”, confidence interval of inter-observer variation (k 2 SD) by ultrasound was f 5 mm and this interval is assumed to represent the accuracy ofthe method. Because LLI measurement by ultrasound is more reliable than clinical methods. ultrasound is recommended for routine use in clinical practice.

Introduction

Various radiographic techniques including CT scans are regarded as the most accurate methods for measure- ment of leg length inequality (LLI) [ 1- 31. However, for screening purposes the use of radiography seems to be hardly acceptable, even when low-dose methods are used.

Clinical methods have been reported to be rather inaccurate in most studies [ 1,4,5]. However, a better accuracy was found by Aspegren et al. [ 61 and this was also our general impression based on routine clinical work.

Because there is disagreement with regard to the accuracy of clinical methods and because of the radiation hazards involved when radiography is em- ployed, new methods for determination of LLI are de- sirable. A new method should be harmless for the patients and give reliable results. The ultrasound tech- nique reported by Terjesen et al. [7] seems to be promising in this respect, but further evaluation of the reliability of the method is required.

Address for reprints: Terje Terjesen, M.D., Department of Ortho- paedic Surgery, Trondheim University Hospital, N-7006 Trond- heim, Norway

The aims of the present study were: 1. to evaluate the reliability of the ultrasound technique and of clinical methods. 2. to determine the LLI by ultrasound in a normal group of subjects.

Material and Methods

Leg length inequality was determined in 100 adult persons (56 male and 44 female) without previous frac- tures or any other disorders of the lower extremities. The mean age of females was 34 years (range: 16-64 years) and of males 33 years (range: 18-66 years).

The subjects answered a questionnaire to assess whether there was any relationship between LLI and the following parameters: height of the subjects, their occupation, and symptoms from the back, hip and knee. They were also asked if they were aware of any length inequality of their lower limbs.

LLI was determined by ultrasound and by two clini- cal methods. In order to calculate the inter-observer variation, all the measurements were made indepen- dently by two observers.

186

Ultrasound technique This was described in detail by Terjesen et al. [ 71. A

real-time apparatus with a 5 MHz transducer was used (Sonoline SL-1, Siemens AG). The subject was in the erect position with straight knees (Fig. 1). The transdu- cer was fixed in a holding device on a specially con- structed rack and moved vertically up and down with a crank handle. The level of the transducer was measured by a millimeter scale on the rack. Ultrasound scanning from the antero-lateral aspect of the hip region was performed, and the leg length was measured as the level of the most proximal, but still clear, contour of the femoral head (Fig. 2). Each limb was measured twice by the same examiner and the average value was used.

Clinical methods

Iliac crest (IC). The subject stood with straight knees and equal weight on both lower limbs. By manual pal- pation of the iliac crests from behind, their levelness was assessed by visual inspection. If a difference was ob-

Fig. 1. The subjects were examined in the erect position with straight knees. Ultrasound measurements were done from the

antero-lateral aspect.

Fig. 2. Ultrasound image showing the anterolateral outline of the proximal part ofthe femoral head (arrows). This level represents the

leg length by ultrasound.

served, wooden boards of known thickness (5 mm increments) were placed under the shorter leg until the levels of the iliac crests were thought to be equal.

Iliac spine-medial malleolus (ISMM). With the patient

supine the leg length was measured as the distance between the anterior superior iliac spine and the distal margin of the medial malleolus. Special attention was paid to perform the measurements with the legs in neu- tral position.

Statistical analysis. Student’s t-test and ANOVA were used to test for dif- ferences between group means. The Chi-square test was applied for differences in the distribution of inter-obser- ver variation. Differences were considered significant at P values below 0.05.

Results

The distribution of LLI measured by ultrasound is shown in Fig 3. The mean LLI was 4.1 mm (range: O-16 mm, SD 3.0). There was no significant difference between females (mean LLI 3.9 mm) and males (mean LLI 4.3 mm). Seventy percent of the subjects had LLI of 5 mm or less and 96% had 10 mm or less. The remaining 4% had LLI from 11 mm to 16 mm.

The range of inter-observer variation by ultrasound and the two clinical methods is shown in Table 1. The mean interobserver variation (observer l-observer 2)

Ill 567 1234

I I

I

8 9 10 11 12 13 14 15 16 17 18 19 20 LLI, ultrasound (mm)

Fig. 3. Histogram showing the distribution of leg length inequality by the number of subjects according to the amount of LLI (mm).

was 0.9 mm (range: - 5-8 mm, SD 2.5) for ultrasound, - 1.0 mm (range: - 15-20 mm, SD 4.8) for the IC method, and - 1.0 mm (range: - 15-10 mm, SD 5.3) for the ISMM. The number of inter-observer differ- ences greater than 5 mm was significantly higher by clinical methods (ultrasound/IC, P = 0.04; ultra- sound/ISMM, P < 0.001) (Table 1). The difference between LLI measured by ultrasound and LLI meas- ured by clinical methods was less than 10 mm in approximately 95% of the subjects (Table 2).

Fifty-seven% of the subjects had had at least one episode of low-back pain and 19% had a record of working disability because of back pain. The subgroup with such complaints had a mean LLI of 5.1 mm while the rest of the subjects had 3.5 mm. The difference was not significant (P = 0.07). Hip complaints were re- ported by 15 y0 and knee complaints by 19 % . There was no significant relationship between these complaints and LLI.

The mean height of females was 168 cm (range:

TABLE 1

Inter-observer variation in leg length inequality. Measurements by ultrasound and two clinical methods, showing the number of sub- jects according to the degree of difference

Method Inter-observer variation (mm)

o-5 6-10 11-15 16-20

Ultrasound 91 3 0 0

Clinical Iliac crest 90 7 2 1 ISMM* II 20 3 0

* ISMM = anterior superior iliac spine to medial malleolus.

187

TABLE 2

The difference between LLI measured by ultrasound and LLI measured by each of the clinical methods, showing the number of subjects according to the difference

Method Observer Difference (mm)

o-5 6-10 11-16

Iliac crest I 61 28 5 2 21 26 3

ISMM* 1 70 27 3 2 72 23 5

* ISSM = anterior superior iliac spine to medial malleolus.

158-179cm) and of males 179 cm (range: 168-194 cm). No significant correlation between LLI and the height of the individuals occurred.

Nineteen individuals were aware of their LLI; the mean LLI in this subgroup was 6.4 mm. Only two of them had LLI of 11 mm or more.

Discussion

The upper normal limit of LLI should be given as the mean LLI + 2 SD, as this range includes 95% of the population. Using this definition, the upper limit of the present series measured by ultrasound was 10 mm, which is consistent with previous reports using radio- graphic methods [ 8,9].

The precision of the ultrasound method was good, as the standard deviation of inter-observer differences was 2.5 mm, which means that 95% of the differences between two observers was less than 5 mm. This is in accordance with the experience of Terjesen et al. [7], who evaluated the same ultrasound technique and used erect posture radiography for comparison. Thus, the accuracy of our ultrasound method is assumed to be + 5 mm. The ultrasound technique is rapid and easy to

perform; it takes approximately 5 minutes and can be done as a routine out-patient procedure. We have used this technique in adolescents and children down to approximately 10 years of age. Below this age, the re- sults would probably be less reliable because small children often lack sufficient concentration to stand with perfectly straight knees for the time required to perform the measurements.

There is no consensus with regard to the accuracy of clinical methods in LLI measurements. Most authors state that clinical methods like the iliac crest and the ISMM are notoriously inaccurate [4,5,10,11]. Differ- ences up to 25-30 mm between radiography and clini- cal assessment have been reported [4,5]. However,

188

Aspegren et al. [6] found agreement between clinical and radiographic measurements in 78% when 6 mm was used as the limit. This indicates a fairly good accu- racy, which is consistent with the results of the present study. Because 95% of the subjects had agreement within 10 mm between ultrasound and clinical methods and because 2 SD of inter-observer differences by clini- cal methods was 10 mm, the accuracy of clinical measurements is assumed to be f 10 mm. Thus, we agree with Gofton [ 121 who maintained that clinical methods are crude but sufficiently sensitive for ordinary use. This is also in keeping with Morscher and Figner [ 131 and Aspegren et al. [ 61. Although a level of accuracy of 10 mm is less than desired for screening purposes, the clinical methods will detect most cases with clinically significant LLI and should not be aban- doned.

A significant correlation between LLI of 10 mm or more and low-back pain has been reported [3,8,9]. In our material the number of subjects with LLI more than 10 mm was too small to be meaningful in this respect. However, there was a trend towards greater LLI in those with low-back symptoms compared to those without such complaints.

All patients with low-back pain should be examined with regard to LLI. We recommend ultrasound as the primary imaging method in subjects with a clinically detected LLI of more than 10 mm. If such an inequality is not confirmed by ultrasound, radiographic exami- nation should be omitted. However, if an LLI of more than 10 mm is confirmed by ultrasound, the evaluation should be supplemented by erect posture radiography to determine LLI by the most accurate method and to evaluate the effects of LLI on the lumbar spine.

Acknowledgments

This study was supported by a grant from the Norwegian Research Council for Science and the Humanities according to an exchange agreement between Norway and Poland.

References

1

2

3

4

5

6

I

8

9

10

11

12

13

Clarke GR. Unequal leg length: an accurate method of detection and some clinical results. Rheum Phys Med 1972; 11: 385-390. Aitken AGF, Flodmark 0, Newman DE, Kilcoyne RF, Shuman WP, Mack LA. Leg length determination by CT digital radiogra- phy. AJR 1985; 144: 613-615. Friberg 0. Clinical symptoms and biomechanics of lumbar spine and hip joint in leg length inequality. Spine 1983; 8: 643-655. Cleveland RH, Kushner DC, Ogden MC, Herman TE, Kermond W, Correia JA. Determination of leg length discrepancy: a com- parison of weight-bearing and supine imaging. Invest Radio1 1988; 23: 301-304. Friberg 0, Nurminen M, Korhonen K, Soininen E, Manttari T. Accuracy and precision of clinical estimation of leg length ine- quality and lumbar scoliosis: comparison of clinical and radio- logical measurements. Int Disabil Studies 1988; 10: 49-53. Aspegren DD, Cox JM, Trier KK. Short leg correction: a clinical trial of radiographic vs. non-radiographic procedures. J Man Phys Ther 1987; 10: 232-238. Terjesen T, Benum P, Rossvoll I, Svenningsen S, Isern AEF, Nordbo T. Leg-length discrepancy measured by ultrasonogra- phy. Acta Orthop Stand 1991; 62; 121-124. Rush WA, Steiner HA. A study of lower extremity length ine- quality. AJR 1946; 56: 616-623. Giles LGF, Taylor JR. Low-back pain associated with leg length inequality. Spine 1981; 6: 510-521. Nichols PJR, Bailey NTJ. The accuracy of measuring leg-length differences. An ‘observer error’ experiment. Br Med J 1955; 2: 1247-1248. Fisk JW, Baigent ML. Clinical and radiological assessment of leg length. N Z Med J 1975; 81: 477-480. Gofton JP. Persistent low back pain and leg length disparity. J Rheum 1985; 12: 747-750. Morscher E, Figner G. Measurement of leg lengthIn: Hunger- ford DS, ed. Progress in orthopaedic surgery: leg length discre- pancy; the injured knee. New York: Springer-Verlag, 1977; 21-27.