Reliability of carpal angle determinations

Reliability of carpal angle determinations The radioscaphoid, radiolunate, and radiocapitate angles of nine lateral projections of the wrist (three in flexion, three in extension, and three in neutral position) of three fresh cadaver specimens were measured. Seven orthopedic surgeons (six hand surgeons and one orthopedic surgeon) made the measurements with a standard goniometer using both the axial and tangential methods of angle determination. The overall standard dcvlatlon for all measurements was 5.2 degrees, and no significant difference in variability between axial and tangential methods was found. By comparing the same angles from different wrist positions, the amount of flexion-extension motion of the capitate, scaphoid, and lunate with respect to the radius was estimated. To assess the accuracy of such a method of carpal motion determination, a more accurate stereoradiographic . method of analysis of carpal kinematics was utilized. The overall estimated error of this standard goniometric method of carpal motion determination averaged 7.4 degrees. (J HAND SURG 1989jI4A:I017·21.)

Marc Garcia-Elias, MD, Kai-Nan An, PhD, Peter C. Amadio, MD, William P. Cooney, MD, and Ronald L. Linscheid, MD, Rochester, Minn.

Objective measurements of carpal' bone alignment are currently required more frequently to assess the functional status of the wrist.' :" Commonly, these measurements are made with a simple goniometer. Angles between the different bones are usually traced on the lateral radiographic projection of the wrist using either the so-called axial method,' : 2. 6-8 the tangential method ,":S or a combination of both methods': II. 12 (Fig. 1). The axial method of drawing carpal axes was first described by Linscheid et al, I . 2 for assessment of carpal mal alignment problems. The lunate axis was defined by a line connecting the midpoints of the convex proximal and the concave distal joint surfaces. The longitudinal axis of the scaphoid was traced by connecting the inidpoints of its distal and proximal poles. Sarrafian and associates' found that an axis line perpendicular to a line tangential to the poles of the lunate may be more reproducible than the one obtained using the axial method. Gilula and Weeks' found more reproducible results by tracing the scaphoid axis tangential to the From the Biomechanics Laboratory, Department of Orthopedics, Mayo Clinic/Mayo Foundation, Rochester, Minn.

Received for publication Sept , 12, 1988; accepted in revised form Jan, 24, 1989. 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 . Reprint requests: Kai-Nan An, PhD, Biomechanics Laboratory, Mayo Clinic, Rochester, MN 55905, . 3/1/11524

palmar outlines of the proximal and distal poles. In our study, a similar concept has been applied to the capitate by tracing an axis tangential to the dorsal profile of the proximal and distal thirds of the capitate. Whether or not these modifications to the initial description of carpal axis result in a better reproducibility remains unknown . These angular measurements have been used not only to evaluate carpal malalignrnents,': 2, 4 , S, 14. IS but also as a tool to quantify individual carpal bone motion."?: II. 13 Little is known, however, about the variability and reliability of these measurements. The goals of this study were as follows: (1) to evaluate and compare the interobservcr variations encountered in the use of these methods of carpal angle determination, and (2) to assess the accuracy of these methods in measuring relative motion of selected carpal bones.

Materials and methods Three fresh-frozen human cadaveric wrists were used in this experiment. Through a longitudinal mid-dorsal incision, the capitate, lunate, proximal pole of the scaphoid, and the distal radius were exposed. A dorsoradial incision was made to approach the distal scaphoid . Four radiodense tantalum pellets were inserted into predrilled holes of varying depth made in each of these carpal bones. Care was taken to avoid injury to any ligaments during insertion of the marker. Each specimen was mounted in neutral rotation TIlE JOURNAL OF llANO SURGERY

1017

The Journal of

1018

Garcia-Elias et al.

HAND SURGERY

Table I. Average results of measurements of different carpal angles using two methods of carpal angle determination (axial and tangential): (n = 7 observers)

AXIAL

Angle

(/)

(2)

Mean ± SD

Mean ± SD

I.N

RS RL RC RS RL RC RS RL RC RS RL RC RS RL RC RS RL RC RS RL RC RS RL RC RS RL RC

52.t ± 10.7 ± -7.8 ± -12.0 ± -41.6 ± -74.4 ± 94.3 ± 37.7 ± 55.7 ± 51.9 ± 10.0 ± -6.6 ± 2.4 ± -23.6 ± -62.7 ± 99.8 ± 46.0::t 73.4 ± 23.6 ± -26.8 ± -25.0 ± -13.6 ± -4t.1 ± -65.1 ± 93.4 ± 22.3 ± 66.8 ±

52.1 1t.1 -22.3 -12.3 -43.3 -87.7 89.1 37.0 43.7 54.6 11.7 -21.8 3.7 -25.8 -71.4 102.7 46.3 59.7 19.0 -31.7 -36.3 -13.8 -36.4 -73.3 95.3 22.t 54.8

I.F

2.N

2.E

2.F

3.N

3.E

within a transparent (Plexiglas) frame by two Steinmann's pins inserted transversely through both the radius and ulna. The specimens were balanced into the desired wrist positions by loading the carpal motor tendons (extensor carpi ulnaris [ECU], extensor carpi radialis brevis [ECRB], extensor carpi radialis longus [ECRL], flexor carpi radialis [FCR], flexor carpi ulnaris [FCU]. Biplanar radiographs were obtained with the wrist in neutral, flexion, and extension positions. From the digitized marker locations on the orthogonal radiographs, the three-dimensional coordinates of each marker could be reconstructed by use of a specially designed software program. The data on these reference locations were then used to calculate the relative rotation of the scaphoid, lunate, and capitate with respect to the radius during flexion-extension motion using another specially designed software program of rigid body motion determination. A more detailed account of this three-dimensional kinematic analysis is documented in previous publications.": 17 In those studies, this technique was found to be very accurate;" Because of the reliability of this technique, the ranges obtained for the lunate, scaphoid, and capitate were assumed to be the "true" ranges of motion for these bones in this study. 0

Tangential method

Film

I.E

Fig. 1. Scaphoid (S), lunate (L), capitate (C). and radius (R) axis as determined using either the axial orthe tangential method. Axial method uses the midpoints of the proximal and distal articular surfaces of the carpals to obtain longitudinal axis. The tangential method determines carpal axis using tangential lines to the palmar outline of the scaphoid, the dorsal of the capitate, and the distal of the lunate.

Axial method

3.F

4.3 5.6 4.8 4.9 7.t 8.5 3.1 4.9 6.7 1.9 4.8 4.2 6.7 4.4 2.7 7.5 5.6 1.7 3.1 7.4 3.8 3.9 9.6 2.9 5.5 3.7 5.2

0

± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

5.6 4.8 6.4 6.3 5.9 9.6 4.5 3.6 5.6 3.8 2.5 1.9 6.0 5.7 12.1 2.1 6.8 2.7 4.t 8.9 3.6 6.9 5.7 8.6 5.5 3.6 4.7

(I). Specimen number; position (n: neutral; f: flexion; e: extension); (2), RL: radioJunate; RS: radioscaphoid; RC: radiocapitate.

The lateral radiographs of the same three wrists in the three different positions (neutral, flexion, and extension) were examined by seven different observers. The observers were seven orthopedic surgeons, six hand surgeons with more than 3 years' experience in hand surgery and one general orthopedic surgeon. The observers measured the radioscaphoid (RS), radiolunate (RL), and radiocapitate (RC) angles on each of the nine radiographs using both the tangential method (TM) and the axial method (AM) (Fig. I). All the observers received instruction in how to measure these angles before their participation. A standard plastic goniometer (Protee AG, Bern, Switzerland), marked in two-degree intervals was used. Twenty-seven angular measurements for each method, axial and tangential, were done by each observer (three wrists, three positions, and three angles).

Vol. 14A, No.6 November 1989

Reliability of carpal angle determinations

1019

Table II. Average difference (mean ± SO) in degrees between the true and the calculated total range of flexion extension of the scaphoid, lunate, and capitate of three different specimens using the two methods (axial, tangential) of carpal angle determination (n = 7 observers) Axial method Specimen no. I

-3.7 ± 6.9 NS 5.3 ± t2.t NS -7.8 ± 12.2 NS

Scaphoid Lunate Capitate 'p

I

Specimen no. 2

-16.6 ± 14.1* -13.4 ± 6.6t -0.8 ± 4.1 NS

Tangential method

I

I

Specimen no. 2

Specimen no. 3

Specimen no. I

-5 .0 ± 4.1*

-8.6 ± 4.5t

-15.0 ± 6.2t

6.3 ± 8.51 NS -6.6 ± 9.9 NS

-10.8 ± 8.2*

2.4 ± 9.6 NS 1.0 ± 5.7 NS

-5.8 ± 11.6 NS

I Specimen no. 3 -2.8 ± 5.2 NS -2.4 ± 6.5 NS -2.8 ± 10.9 NS

< 0.05.

tp < 0.005 NS. Not significant. (I test: 110: Mu

= 0).

By comparing the angular values obtained in flexion and extension, the total amount of flexion-extension motion (FEM) of the lunate, scaphoid, and capitate for the three wrists could be calculated. Statistical analysis The variability of measurements made by the seven observers (n = 9) using eaeh of the two methods of angle determination (AM, TM) was analyzed by comparing the standard deviations of these measurements for each of the three specific angles (RS, RL, and RC) with a paired 1 test. The variability of measurements combining all angles, positions, and wrists (n = 27) also was analyzed by comparing the standard deviations of these measurements for each of the two methods of angle determination (AM, TM) using a paired 1 test. To determine the accuracy of the measurements of individual carpal bone rotations by the two standard goniometric techniques (AM and TM), the range of motion of each carpal bone as calculated with these two methods was matched with the range of motion obtained by use of the above explained stereoradiographic method. The accuracy was evaluated by testing the ability of these methods to predict the true range, with the null hypothesis that the difference between the range calculated and the true range was equal to zero. Differences were considered significant at a level of p

< 0.05.

Results Reproducibility of the methods of carpal angle determination When angles on the same set of films were tabulated for the seven observers, a great deal of variation was

observed. This is illustrated in Table I, which includes the mean and standard deviation (SO) obtained for each angle, each film, and each method of angle determination. Combining all the angles from all positions and wrists (n = 27), the average SO when the axial method was used was not significantly different from the average SO obtained when using the tangential method (AM: 4.98 ± 1.96; TM: 5.46 ± 2.37; P = 0.37). When comparisons were done for each of the three angles separately (n = 9), the differences between each of the two methods of carpal angle determination still were not significant (RS angle , average SO: AM: 4.54, TM: 4.97, p = 0.38; RC angle, average SO: AM: 4.50, TM: 6.20, p = 0.33; RL angle, average . SO: AM: 5.90, TM: 5.27, p = 0.61). The overall SO for all the angles and both methods considered together was 5.2 degrees. Accuracy of the axial and tangential methods in predicting range of motion of the carpal bones The average differences between the calculated range of motion for the scaphoid, lunate and capitate and the true range are tabulated in Table II for each specimen and each of the two methods (AM and TM) of carpal angle determination. Such methods showed ability to predict the true excursion of the capitate in all the three specimens whatever the method of carpal angle determination used. These methods were accurate in predicting lunate excursion in only two of the three specimens studied. Accuracy was even poorer for the scaphoid range determination: a significant level of accuracy was obtained in only one of the three specimens. The average error when calculation of all angles was based on data obtained by the axial method was 'of 7.7 ± 5.7 degrees. By use of the tangential method,

1020

The Journal of HAND SURGERY

Garcia-Elias e/ aI.

the error averaged 7.2 ± 5.2 degrees. The error for the two methods combined was 7.4 ± 5.5 degrees. No statistically significant difference was observed between both methods (p = 0.43). None of the observers exhibited a significantly higher accuracy over the others with either method. Discussion This study shows that different observers measuring the same angles from the same routine films are consistent to ± 5.2 degrees and that their measurements of carpal bone motion are an average of 7.4 degrees different than the true measurement of that carpal angle. Several factors can be responsible for this interobserver variability. Assessment of carpal bone axis is one important factor. Recognition of the overlapping profiles of the carpal bones very often becomes a difficult task . even for experienced surgeons. Especially difficult is the distinction between the proximal contour of the scaphoid and the lunate or the recognition of all the dorsal outline of the scapho id in lateral views of the carpus . When such anatomic landmarks are not optimally represented, tracing of longitudinal axes becomes a somewhat subjective procedure. That may explain in part our findings. Other possible factors affecting accuracy and reliability are the quality of the radiographs used, the mineraI content of the carpal bones , the positioning of the wrist when the radiographs are taken, and the accuracy of the goniometric devices used . To prevent these factors from influencing our results, we used more than one benchmark specimen. Only the best specimens available were used in this study; the best possible radiographs were obtained, and positioning was carefully assessed. Despite all these precautions, accuracy was not similar in all three specimens. The error committed in specimen No.2 was significantly greater than in the other specimens (Table II). In that case, the carpal bones showed a small decrease of bone density, which made any angle determination more difficult. Since osteopenia is a frequent finding in patients where these angles are measured, inclusion of the data from this specimen was accepted. For similar reasons, specimen No.3 with abnormal carpal angles was also included. Malrotation of the carpal bones may complicate determination of carpal bone axes. The fact that the error from true in specimen No. I, the most normal, was intermediate between that of the osteopenic specimen No .2 and the abnormal specimen No.3, suggests that the range of errors we observed reflects the range likely to be encountered in clinical practice. It might be argued that with more observers, or more

observations per observer, other values for interobserver reliability might be obtained. As other studies evaluating the reliability of clinical goniometry have shown similar interobserver variability, it is more likely that the limit of roughly ± 5 degrees is inherent in clinical goniometry.":" Given what is probably an inherent interobserver variability of ± 5 degrees in clinical goniometry, our variability of 5.2 degrees for carpal angle measurement is unlikely to be reduced greatly by extending the study with more observers or more specimens. The second goal of this study was to establish the reliability of these goniometric methods when used to study carpal kinematics. Several studies have addressed the kinematic behavior of the carpus in normal and pathologic conditions using standard goniometric measurements of carpal angles on lateral projections of the wrist . J. 7. 8. 9 . 11-1l If we consider the following : (I) that most of the movement of the carpal bones has not two but six degrees of freedom,": 17 (2) that any determination of intercarpal angle has a SD of 5.2 degrees between examiners, and (3) that any carpal motion determination has a potential error of 7.4 degrees from true, as found in this study, the utilization of carpal angles measured with a goniometer from lateral radiographs does not seem to be a very reliable method to quantify exact degrees of carpal motion . The use of this method as a research tool to quantify intracarpal kinematics cannot be recommended. With this study, however, we do not advise against goniometry for carpal angle determination in clinical practice . Such measurements are a necessary part of the clinical evaluation of carpal malalignment. These angles will provide initial data on which further investigations can be based (tomograms, cineradiography, etc.). Therapeutic decisions, however, should probably not be made exclusively on the basis of changes found in the amplitude of those angles if those changes are less than five degrees. Conclusions The use of goniometrically derived intracarpal angles to evaluate carpal malalignment problems has some limitations due mainly to difficulties in determining carpal axis. The interobserver variation in tracing these angles with a goniometer is ±5.2 degrees and , therefore, any calculation of carpal motion based in those angles must be regarded as potentially inaccurate to that extent. Although this variability limits the usefulness of goniometric measurement of carp al alignment to study carpal kinematics in the research laboratory, in the clinical selling, goniometry should still be encouraged as

Vol. 14A, No .6 November 1989

Reliability of carpal angle determinations

an easy way to obtain valuable information about the general pattern of carpal alignment.

REFERENCES I. Linscheid RL, Dobyns JH, Beabout JW, Bryan RS. Traumatic instability of the wrist. Diagnosis, classification and pathomechanics. J Bone Joint Surg 1972;54A:161232. 2. Dobyns JH, Linscheid RL, Chao EYS, Weber ER, Swanson GE. Traumatic instability of the wrist. Instructional Course Lecture, AAOS 1975;24:182-99 . 3. Sarrafian SK, Melamed JL, Goshgarian GM . Study of wrist motion in flexion and extension. Clin Orthop 1977;126 :153-9. 4. Talcisnik J. Wrist: anatomy, function, and injury. Instructional Course Lecture. AAOS 1978;27 :61-87. 5. Gilula LA, Weeks PM. Post-traumatic ligamentous instabilities of the wrist. Radiology 1978;129:641-51. 6. Palmer AK, Dobyns JH, Linscheid RL. Management of post-traumatic instability of the wrist secondary to ligament rupture. J HAND SURG 1978;3:507-32. 7. Chaise F, Roger B, Witvoet J, Laval-Jeantet M. Analyse de l'incidence radiologique de profil du carpe normal. Ann Radial 1985;28:381-6. 8. Schuhl JF. Leroy B, Comtet JJ. Radiological study of the mobility of the scapho-lunate joint. Ann Chir Main 1985;4: 143-8. 9. Schuh! JF, Comtet 11. Biodynamics of the wrist: radiologic approach 10 scapholunate instability. J HAND SURG 1985; lOA:1006-8. 10. Nakamura R, Hori M, Imamura T. Horii E, Miura T. Clinical significance of the angles between bones of the wrist. Presented at the Third Congress of International Federation of Societies for Surgery of the Hand. Tokyo 1986. II. Nakamura R, Miura T, Kino Y, Horibe Y. Clinical sig-

12.

13. 14.

15.

16. 17.

18.

19.

20.

21.

22.

1021

nificance of the angles between carpal bones in functional radiography of the wrist joint. Orthop Traum Surg (Japan) 1985;28:1481-93. Nishikawa T. Normal value of carpal bone angles and gap of the wrist in motion study. J Japanese Orthop Traum J 1985;30:1119-30. Imamura K. Cineradiographic study of the wrist motion . J Jpn Orthop Assoc 1987;61:499-5 10. Rasmussen P, Schwab JP, Johnson RP. Symptomatic scaphoid nonunions treated by Russe bone grafting. Orthop Review 1985;14:41-7. Garcia-Elias M, Vall A, Salo JM, Lluch AL. Carpal alignment after different surgical approaches to the scaphoid: A comparative study. J HAND SURG 1988; 13A:604-12. Chao EYS, Morrey BF. Three-dimensional rotation of the elbow. J Biomech 1978;11:57-73. Ruby LK, Cooney WP, An KN, Linscheid RL, Chao EYS. Relati ve motion of selected carpal bones: a kinematic analysis of the normal wrist. J HAND SURG 1988;13A :I-IO. Hamilton GF, Lachenbruch PA. Reliability of goniom eters in assessing finger joint angle. Phys Ther 1969; 49:465-9. Mitchell WS, Millar J, Sturrock RD. An evaluation of goniometry as an objective parameter for measuring joint motion. Scot Med J 1975;20:57-9. Boone DC, Azen SP, Lin CM, Spence C, Baron C, Lee L. Reliability of goniometric measurements. Phys Ther 1978;58: 1355-60. . Solgaard S, Carlsen A, Kramhoft M, Petersen VS. Reproducibility of goniometry of the wrist : Scand J Rehabil Med 1986;18:5-7. Low JL. The reliability of joint measurement. Physiotherapy 1976;62:227-9.