- Email: [email protected]
Geometry of the proximal humerus and implications for prosthetic design
Geometry of the proximal humerus and implications for prosthetic design Ralph Hertel, MD, Ulf Knothe, MD, and Franz T. Ballmer, MD, Berne, Switzerland
The purpose of this study was to add critical information to the data already available on anthropometry of the proximal humerus. Two hundred macerated humeri were examined. Measurements were taken either directly on the bones or on standardized radiographic projections. The methodology was validated and showed a mean interobserver correlation of 0.94 ⫾ 0.067. Results were expressed in mean values, first SD, and minimum and maximum values, as well as the 10th and 90th percentiles. The frontal radius of the head ranged between 21 and 26.5 mm (10th respectively 90th percentile). The frontal diameter of the base of the head ranged between 39.4 and 50 mm. The head height ranged between 14.4 and 18.8 mm. The frontal radius– head height ratio ranged between 0.64 and 0.77. The inclination of the head ranged between 132° and 142°. The medial offset ranged between 3.9 and 8.6 mm. The posterior offset ranged between ⫺0.4 and 3.2 mm. The greater tuberosity offset (distance between the axis of the proximal humerus and the most medial insertion point of the supraspinatus tendon) ranged between 2.5 and 9.2 mm. Retrotorsion ranged between 7° and 38.5°. The distance from the bicipital groove to the head equator ranged between 6 and 10.5 mm. The anatomy of the proximal humerus showed a wide range for variables such as the medial offset and the greater tuberosity offset but was surprisingly constant for the inclination and relative dimensions of the head. The implications for prosthetic design are as follows: stem design and insertion should respect the insertion facet of the supraspinatus, a constant head inclination is an adequate approximation, only one head height per radius is required, and the capability for adjustment of medial offset is mandatory. (J Shoulder Elbow Surg 2002;11:331– 8.) From the Department of Orthopaedic Surgery, Inselspital, University of Berne, Switzerland. Reprint requests: Ralph Hertel, MD, Department of Orthopaedic Surgery, Inselspital, University of Berne, 3010 Berne, Switzerland (E-mail: [email protected]). Copyright © 2002 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2002/$35.00 ⫹ 0 32/1/124429 doi:10.1067/mse.2002.124429
INTRODUCTION Dimensional data of the proximal humerus are important for optimal design of prosthetic humeral components. Previous studies have shown considerable variability in the size and position of the head in relation to the shaft.1,2,4,6-8,11-18 The purpose of this study was to add possibly critical information to the already available data. In addition to addressing methodological issues such as the determination of the equatorial plane, we introduced new, practically relevant measures. Interpretation of the results revealed a potential for misinterpretation and therefore deserve specific attention. MATERIALS AND METHODS The Institute of Anatomy of the Universities of Berne and Fribourg provided 278 macerated humeri. Seventy-eight bones with obvious deformities such as degenerative arthritis, undefined critical line due to rotator cuff disease, and preparation artifacts were excluded from the study. Of the 200 remaining bones, 106 were right humeri and 94 were left humeri. Information regarding age, sex, and handedness was not available. Total bone length, distance from sulcus to equator, and frontal and sagittal diameter of the base of the humeral head were measured directly on the specimens. The retrotorsion was measured as follows: the equator of the humeral head was plotted on the bone, as outlined in the section definitions. The humeri were placed on a flat surface with the anterior aspect facing down. By doing so, the rotation of the bone was entirely controlled by the orientation of the trochlea (anterior tangent to the trochlea). A photographic projection that used a 100-mm objective in line with the axis of the humerus was used to visualize and measure the amount of retrotorsion (Figure 1, C). All other measurements were made on the basis of 2 standardized radiologic projections in the frontal and sagittal plane (ie, parallel and orthogonal to the equatorial plane). The source-to-film distance was 120 cm. The mean magnification factor was 1.0178. The data were corrected for magnification. The technique of measurement was validated in a pilot study on 20 specimens in which the mean interobserver correlation coefficient was 0.94 (SD, 0.067; minimum, 0.74; maximum, 0.99). Parametric tests were used for statistical analysis. The normal distribution of the data was confirmed by graphical analysis. Correlation coefficients
331
332
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg July/August 2002
Figure 1 A-D, Measured distances and angles. ARC, Surface arc of the head; AX, anatomic axis of the proximal humerus; CCD, head inclination angle; CD, greater tuberosity offset (or critical distance); CL, critical line; CP, critical point; CYL, best-fitting cylinder; EQ, head equator; FBD, frontal diameter of the base of the head; FR, radius of curvature of the head in the frontal plane; HH, head height; MO, medial offset of the head; PO, posterior offset of the head; RT, retrotorsion of the humeral head; SBD, sagittal diameter of the base of the head; SE, distance between sulcus and equator.
and P values were calculated. The data are presented as mean values, first SD, range, and 10th and 90th percentiles.
Measured variables and definitions Axis of the proximal humerus was obtained through serial midpoint determination on both radiographic projec-
tions (outer cortex) in a proximal segment defined by the 10% and 40% total bone length marks. Best-fitting cylinder is defined as the largest cylinder fitting into the proximal medullary cavity between 10% and 40% of total bone length on both radiographic projections. Center of the head was determined geometrically and
J Shoulder Elbow Surg Volume 11, Number 4
double-checked with circular templates on both radiographic projections. Clearance best-fitting cylinder to greater tuberosity is defined as the shortest distance from the lateral wall of the best-fitting cylinder to the critical point measured on the anteroposterior radiographic projection. Critical line is the most medial insertion line of the supraspinatus tendon. In well-preserved macerated bone without rotator cuff disease, this line is clearly visible as a small step off at the edge of the subchondral bone plate. Critical point is the intersection between the critical line and the humeral head equator. Distance sulcus-equator is the distance between the midpoint of the bicipital groove and the critical point measured directly on the bones. Frontal diameter of the base of the head is the diameter of the head segment at the anatomic neck measured directly on the bones. Head equator is the equatorial line drawn directly on the bones and obtained as follows: the bone was mounted in a frame that allowed a 45° oblique medial-to-lateral projection of the humeral head. The humerus was rotated until the base of the head was perpendicular to the axis of the proximal humerus (Figure 2). The distal articular midpoint was defined as the mid distance between the anterior and posterior tangent to the head as viewed from the 45° oblique medial-to-lateral projection. A line was then drawn on the head starting from the distal articular midpoint running parallel to the axis of the proximal humerus and bisecting the head. This method was derived from Tillet et al.17 Head height determines the height or thickness of the humeral head and was measured in the anteroposterior radiographic projection. Head inclination angle is defined as the inclination of the anatomic neck in the frontal plane. The angle was measured on the anteroposterior radiographic projection and was formed by the intersection of the axis of the proximal humerus and the line perpendicular to the anatomic neck. Head surface arc represents the arc of articular surface available in the frontal plane. The angle was measured on the anteroposterior radiographic projection. Medial offset of the head is the shortest distance between the axis of the proximal humerus and the center of the head measured on the anteroposterior projection. Posterior offset of the head is the shortest distance between the axis of the proximal humerus and the center of the head measured on the lateral projection. Retrotorsion of the humeral head is the angle between the projection of the equatorial plane and the plane tangent to the anterior aspect of the trochlea measured on a photographic projection. Radius of the curvature of the head in the frontal plane is the radius of the head measured on the anteroposterior radiographic projection. Radius of the curvature of the head in the sagittal plane was calculated based on the sagittal diameter of the base of the head and the head height. Sagittal diameter of the base of the head is the maximal distance between the anterior and posterior head-collum junction in a plane perpendicular to the equatorial plane measured directly on the bones.
Hertel, Knothe, and Ballmer
333
Figure 2 Methodology. For the identification of the head equator, the bones were mounted in a special frame and manually rotated until the base of the head was perpendicular to the axis of the humerus (A and B). The midpoint a was determined. A line was then drawn on the head starting from the distal articular midpoint running parallel to the axis of the proximal humerus and bisecting the head (C).
Total bone length is the distance between the superior tangent to the humeral head and the inferior tangent to the trochlea measured directly on the bones. Greater tuberosity offset (or critical distance) is the short-
334
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg July/August 2002
Figure 3 Histograms of the most relevant variables.
est distance between the axis of the proximal humerus and the critical point measured directly on the bones (Figures 3-5).
RESULTS Some variables showed a large amount of scatter, whereas others, in particular angles and ratios, showed minimal variations. The results are summarized in Table I. Head dimensions
The mean radius of curvature of the head showed a 12% difference in the frontal and sagittal planes. The sagittal diameter of the base of the head was 6% smaller than the frontal diameter. The base of the resected head best fit an ellipse, with the long-axis in line with the head equator (retrotorsion axis). The mean head height measured 71% of the radius in the frontal plane.
Shaft dimensions
Total bone length and best-fitting cylinders were quite variable. The diameters of the best-fitting cylinders were not only related to the size and to the outer shape of the bone but also to the cortical thickness (ie, osteoporosis). The anatomic axis, on the contrary, was not altered by variation of cortical thickness. 11.5% of the specimens featured a marked bend (⬎5°) in the medullary cavity of the proximal humerus. The angulation, when present, was located at 13.5 ⫾ 0.72 cm (minimum, 13; maximum, 15), distal to the superior humeral head tangent. The shape of the proximal medullary cavity was roughly cylindrical in the remaining specimens. Head position in relation to shaft
The position of the head in relation to the shaft was quite scattered for variables such as the medial and posterior offset, the greater tuberosity offset, and the
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg Volume 11, Number 4
335
Figure 4 High-offset (1), standard (2), and low-offset (3) morphotypes. With decreasing critical distance, the introduction of a straight-stemmed prosthesis becomes increasingly difficult.
Figure 5 The introduction of a straight-stemmed canal fitting prosthesis may damage the insertion of the supraspinatus tendon.
retrotorsion. In contrast, the head inclination and the distance from the sulcus to the equator showed little variation (Figures 3 and 4). The distance between the lateral border of the bestfitting cylinder and the most medial insertion line of the supraspinatus tendon (clearance) was negative in 49.5% of the specimens. In these cases introduction of a canal-fitting, straight-stem prosthesis without fin would damage the insertion facet of the supraspinatus tendon (Figure 5). Adding a 4-mm flange to the canal-fitting shaft would damage the insertion facet of the supraspinatus in 94.5% of the measured specimens.
Correlations
Relevant correlations are shown in Table II. Of particular interest was the high correlation between the radius and the height of the head, as well as the high correlation between the frontal and sagittal radius. This suggests that the relative shape of the head is largely constant and that it can be accurately calculated based on the frontal radius value. On the other hand, there was no measurable correlation between the size of the head and its
336
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg July/August 2002
Table I Anthropometrical measurements
Dimensions of head Radius in frontal plane (mm) Radius in sagittal plane (mm) Ratio of frontal sagittal radius (mm) Diameter of head base in frontal plane (mm) Diameter of head base in sagittal plane (mm) Ratio of frontal-sagittal diameter Head height (mm) Ratio of frontal radius–head height Head surface arc (°) Dimensions of shaft Total length of humerus (cm) Best-fitting cylinder (mm) Clearance: Best-fitting cylinder to greater tuberosity (mm) Relative position of head on shaft Medial offset (mm) Posterior offset (mm) Greater tuberosity offset (mm) Head inclination angle (°) Retrotorsion (°) Distance of sulcus to equator (mm)
10th Percentile
90th Percentile
30 26 1.01 57 53.5 1.15 22 0.85 163
21 18.5 0.83 39.5 37 0.88 14.4 0.64 138
26.5 23 0.94 50 46 1.04 18.8 0.77 153
24.5 6 ⫺7.9
36.8 21 5.5
30 9 ⫺3.6
34.5 14 3.3
1.7 ⫺3 ⫺2.3 128 ⫺9 3.5
11.5 5.3 12 145.5 50 13.7
3.9 ⫺0.4 2.5 132 7 6
8.6 3.2 9.2 142 38.5 10.5
Mean
SD
Minimum
24 21 0.88 44.5 42 0.94 17 0.71 145
2.2 1.8 0.04 4 3.8 0.06 1.7 0.05 5.95
19 17 0.78 36 33.5 0.85 12.5 0.56 130
31.6 11.5 ⫺0.1
2.3 2.09 2.6
6 1.4 5.6 137 23.3 8.3
1.81 1.43 2.58 3.62 11.75 1.72
Maximum
Table II Correlations Variable a
Variable b
Frontal head radius Diameter of head base in frontal plane Head height Head inclination angle Medial offset Greater tuberosity offset Total length Total length Head height Total length Distance of sulcus to equator Diameter of head base in frontal plane Medial offset Posterior offset Total length Frontal head radius Retrotorsion
Sagittal head radius Diameter of head base in sagittal plane Frontal head radius Greater tuberosity offset Greater tuberosity offset Medial offset Frontal head radius Diameter of head base in frontal plane Total length Retrotorsion Total length Medial offset Head inclination angle Retrotorsion Medial offset Posterior offset Distance of sulcus to equator
relative position on the shaft. High- and low-offset morphotypes (Figure 4) did not correlate with the size of the bone nor with the head inclination angle. DISCUSSION The aim of this study was to add detail information to already available databases on the geometry of the proximal humerus. Although the results are largely in accordance with published data, the interpretation
Correlation factor
P value
0.84 0.75 0.7 0.69 0.64 0.64 0.62 0.61 0.57 0.27 0.27 0.26 0.23 0.22 0.14 0.11 0.04
⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 .1724 ⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 .0007 .0009 .0017 .0157 .1062 .5055
of several measurements was quite discordant. To optimize statistical power, measurements were taken from a large number of specimens—the largest thus far. In an effort to reduce methodological errors, the definition of landmarks was made as precisely and as reproducibly as possible and data acquisition was kept relatively simple. Critical issues included determination of the equatorial plane, determination of the anatomic axis of the proximal humerus, and determination of the center of the head. The described methodology provided a high degree of accuracy. Other
J Shoulder Elbow Surg Volume 11, Number 4
authors have determined the humeral axis by either manual antegrade15 or retrograde1 reaming. The potential methodological errors inherent to these techniques have not been assessed. Calculation of the axis based on computer-assisted mapping of the surface of the bone2,16,18 can be influenced by bone ridges. Comparing our data with surface-mapped data,2,16,18 it can be noted that the axis calculated on the base of surface mappings lies slightly more anterior (1-2 mm). The poorly defined, irregular, and S-shaped articular margin complicates determination of retrotorsion. Construction of the head equator according to the presented protocol, which is a modification of the method of Tillet et al,17 was reproducible. In order to highlight the range of variation of normal anatomy, results were also expressed in percentiles. The range between the 10th and 90th percentile comprises 80% of the values, which was judged to reflect the clinically relevant variation of normality better than the description of only the mean and first SD. As expected, the variation in the total length of the humeral specimens and in the size of the humeral head was large.2,15,18 However, the ratio of the frontal radius of the head versus head height remained remarkably constant (0.71 ⫾ 0.05). Pearl and Volk15 found a similar ratio of 0.73 ⫾ 0.04. The surface arc is yet another way to express the same relationship. It measured 145° ⫾ 5.9°. Pearl and Volk reported a surface arc of 150° ⫾ 5°, and Jobe and Iannotti8 found a slightly larger value, 159° ⫾ 8.5°. The implication for prosthetic design is that, for a given radius, only one anatomic head height is necessary for near-normal reconstruction of the outer contour of the proximal humerus. The radius of the head showed a 12% difference between the frontal and sagittal planes. Iannotti et al7 reported an 8% mismatch. Thus, the head is not exactly spherical. The impact of this finding on prosthetic design remains unclear. Choosing the right head in the frontal plane would result in an overlap of the head in the sagittal plane of about 3 mm. We did not observe a gradual decrease in the radius of the head in the frontal plane, as suggested by Kapandji.9 The question of whether the reason was methodological (ie, if the technique used might have been too coarse to detect small variations of the radius) remains unanswered. The relative position of the head on the shaft was quite variable. The mean medial offset was 6 ⫾ 1.8 mm. McPherson et al12 reported 7.6 ⫾ 3.2 mm, whereas Pearl and Volk15 reported slightly larger values, with a mean medial offset of 9.7 ⫾ 1.7 mm. Variations in the medial offset are of paramount importance in the design of humeral prosthetic components in order to respect the most medial insertion line
Hertel, Knothe, and Ballmer
337
of the rotator cuff. To snugly position the head against the most medial insertion line of the supraspinatus, a shaft with variable medial offset is necessary. The mean posterior offset was relatively small (mean, 1.4 ⫾ 1.44 mm). McPherson et al12 measured 1.9 ⫾ 1.7 mm, and Ballmer et al1 reported 2 mm. Using a surface-mapping method, Walch and Boileau2,18 found slightly larger values (mean 2.5 mm), and Roberts et al16 measured a mean value as high as 4.7 mm. These differences are probably due to the use of the frontal—as opposed to the equatorial reference plane. The offset of the greater tuberosity is a crucial variable. It indicates the distance between the insertion of the innermost fibers of the supraspinatus to the axis of the proximal humerus. On its basis, we classified the humeri into 3 morphotypes: a so-called standard humerus, a high-offset morphotype, and a low-offset morphotype (Figure 4). The implication is that straight-stemmed prostheses cannot be introduced in a low-offset humerus without relevant damage to the insertion facet of the supraspinatus tendon (Figure 5). Flanges and press-fit designs represent an additional hazard. Implantation of a straightstemmed, canal-fitting shaft would have caused damage to the insertion of the rotator cuff in 49.5% of the measured specimens. The inclination of the humeral head did not vary much (mean, 137° ⫾ 3.6°) and occurred in only 1 plane. Other investigators found similar mean values ranging from 130° to 141° and similar SDs ranging from 3.4° to 5°.2,3,12,14,18 Considering minimal and maximal values, published inclinations range from 120° to 145°.2,7,15 It is questionable whether the extreme values should be used to support arguments in favor of or against variable inclination designs because these values are single observations that may well be outliers and that have no specific representation in the Gauss curve. Therefore, use of the second SD or, better, the 10th and 90th percentiles appears to reflect more adequately the practical needs. Assuming a possible maximum mismatch of ⫹5° or ⫺5° (corresponding to the 10th and 90th percentiles), the effect on the outer shape of the proximal humerus would remain irrelevant. With 5° of valgus error, the mean humeral width2,14,18 would be increased by only 1.5 mm, or 3%, and the humeral height2,14,18 would be affected even less. Therefore, when the aim is to reconstruct normal or near-normal anatomy of the proximal humerus variable prosthetic head inclinations are not required. All authors of anatomic studies have found large variations in retrotorsion. The mean values ranged from 18° to 33°.1-3,5,6,14,15,18 In vivo radiologic studies10 seem to have less scattered results. The reason is unclear, as methodological errors would be expected to be larger in the in vivo measurements. When
338
Hertel, Knothe, and Ballmer
comparing data, it is essential to consider that measurements taken using the epicondylar axis differ by about ⫺10° from measurements taken using the anterior tangent to the trochlea.6 We prefer the anterior tangent because it can be determined with precision on dry bones and adequately assessed during surgery (perpendicular to the axis of the forearm). The relative position of the bicipital groove to the head equator showed little variation (mean distance, 8.3 mm) and is, as we and others17 believe, a useful intraoperative landmark for positioning the head in the right retrotorsion. Tillet et al17 found a mean distance of 9 mm. However, other authors1,4,11 found little or no correlation at all. The reason for these contrasting conclusions is most likely methodological. So far determination of the equatorial plane of the humeral head has not been addressed with sufficient clarity. Clear definitions are missing, and reproducibility has not been tested. We believe that adequate definition of the anatomic landmarks is the key issue to obtain useful results. In conclusion, the anatomy of the proximal humerus was largely scattered for variables such as the medial offset, the greater tuberosity offset, and the retrotorsion. Yet, for the inclination of the humeral head, the head radius– height ratio, and the distance between the bicipital groove to the equator, the practically relevant variance was minimal. These findings yield pertinent implications for prosthetic design and implantation. REFERENCES
1. Ballmer FT, Lippitt SB, Romeo AA, Matsen FAI. Humeral prosthetic arthroplasty: surgically relevant considerations. J Shoulder Elbow Surg 1993;2:296-304. 2. Boileau P, Walch G. The three-dimensional geometry of the proximal humerus. Implications for surgical technique and prosthetic design. J Bone Joint Surg Br 1997;79:857-65. 3. Clarke IC, Gruen TAW, Sew Hoy A, Hirschowitz D, Maki S,
J Shoulder Elbow Surg July/August 2002
4. 5. 6. 7.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
18.
Amstutz HC. Problems in gleno-humeral surface replacements— real or imagined? IMechE 1979;8:161-75. Doyle A, Burks R. Comparison of humeral head retroversion with the humeral axis/biceps groove relationship: a study in live subjects and cadavers. J Shoulder Elbow Surg 1998;7:453-7. Edelson G. Variations in the retroversion of the humeral head. J Shoulder Elbow Surg 1999;8:142-5. Hernigou P, Duparc F, Filali C. Humeral retroversion and shoulder prosthesis [in French]. Rev Chir Orthop Reparatrice Appar Mot 1995;81:419-27. Iannotti JP, Gabriel JP, Schneck SL, Evans BG, Misra S. The normal glenohumeral relationships. An anatomical study of one hundred and forty shoulders. J Bone Joint Surg Am 1992;74:491500. Jobe CM, Iannotti JP. Limits imposed on glenohumeral motion by joint geometry. J Shoulder Elbow Surg 1995;4:281-5. Kapandji IA. Physiologie articulaire, 5th ed. Malone S.A. Editor: Paris; 1980. Kronberg M, Brostrom LA, Soderlund V. Retroversion of the humeral head in the normal shoulder and its relationship to the normal range of motion. Clin Orthop 1990;253:113-7. Kummer F, Perkins R, Zuckerman J. The use of the bicipital groove for alignment of the humeral stem in shoulder arthroplasty. J Shoulder Elbow Surg 1998;7:144-6. McPherson EJ, Friedman RJ, An YH, Chokesi R, Dooley RL. Anthropometric study of normal glenohumeral relationships. J Shoulder Elbow Surg 1997;6:105-12. Pearl ML, Kurutz S. Geometric analysis of commonly used prosthetic systems for proximal humeral replacement. J Bone Joint Surg Am 1999;81:660-71. Pearl ML, Volk AG. Retroversion of the proximal humerus in relationship to prosthetic replacement arthroplasty. J Shoulder Elbow Surg 1995;4:286-9. Pearl ML, Volk AG. Coronal plane geometry of the proximal humerus relevant to prosthetic arthroplasty. J Shoulder Elbow Surg 1996;5:320-6. Roberts S, Foley A, Swallow H, Wallace W, Coughlan D. The geometry of the humeral head and the design of prostheses. J Bone Joint Surg Br 1991;73:647-50. Tillet E, Smith M, Fulcher M, Shanklin J. Anatomic determination of humeral head retroversion: the relationship of central axis of the humeral head to the bicipital groove. J Shoulder Elbow Surg 1993;2:225-56. Walch G, Boileau P. Morphological study of the humeral proximal epiphysis [abstract]. J Bone Joint Surg Br 1992;74:14.
The purpose of this study was to add critical information to the data already available on anthropometry of the proximal humerus. Two hundred macerated humeri were examined. Measurements were taken either directly on the bones or on standardized radiographic projections. The methodology was validated and showed a mean interobserver correlation of 0.94 ⫾ 0.067. Results were expressed in mean values, first SD, and minimum and maximum values, as well as the 10th and 90th percentiles. The frontal radius of the head ranged between 21 and 26.5 mm (10th respectively 90th percentile). The frontal diameter of the base of the head ranged between 39.4 and 50 mm. The head height ranged between 14.4 and 18.8 mm. The frontal radius– head height ratio ranged between 0.64 and 0.77. The inclination of the head ranged between 132° and 142°. The medial offset ranged between 3.9 and 8.6 mm. The posterior offset ranged between ⫺0.4 and 3.2 mm. The greater tuberosity offset (distance between the axis of the proximal humerus and the most medial insertion point of the supraspinatus tendon) ranged between 2.5 and 9.2 mm. Retrotorsion ranged between 7° and 38.5°. The distance from the bicipital groove to the head equator ranged between 6 and 10.5 mm. The anatomy of the proximal humerus showed a wide range for variables such as the medial offset and the greater tuberosity offset but was surprisingly constant for the inclination and relative dimensions of the head. The implications for prosthetic design are as follows: stem design and insertion should respect the insertion facet of the supraspinatus, a constant head inclination is an adequate approximation, only one head height per radius is required, and the capability for adjustment of medial offset is mandatory. (J Shoulder Elbow Surg 2002;11:331– 8.) From the Department of Orthopaedic Surgery, Inselspital, University of Berne, Switzerland. Reprint requests: Ralph Hertel, MD, Department of Orthopaedic Surgery, Inselspital, University of Berne, 3010 Berne, Switzerland (E-mail: [email protected]). Copyright © 2002 by Journal of Shoulder and Elbow Surgery Board of Trustees. 1058-2746/2002/$35.00 ⫹ 0 32/1/124429 doi:10.1067/mse.2002.124429
INTRODUCTION Dimensional data of the proximal humerus are important for optimal design of prosthetic humeral components. Previous studies have shown considerable variability in the size and position of the head in relation to the shaft.1,2,4,6-8,11-18 The purpose of this study was to add possibly critical information to the already available data. In addition to addressing methodological issues such as the determination of the equatorial plane, we introduced new, practically relevant measures. Interpretation of the results revealed a potential for misinterpretation and therefore deserve specific attention. MATERIALS AND METHODS The Institute of Anatomy of the Universities of Berne and Fribourg provided 278 macerated humeri. Seventy-eight bones with obvious deformities such as degenerative arthritis, undefined critical line due to rotator cuff disease, and preparation artifacts were excluded from the study. Of the 200 remaining bones, 106 were right humeri and 94 were left humeri. Information regarding age, sex, and handedness was not available. Total bone length, distance from sulcus to equator, and frontal and sagittal diameter of the base of the humeral head were measured directly on the specimens. The retrotorsion was measured as follows: the equator of the humeral head was plotted on the bone, as outlined in the section definitions. The humeri were placed on a flat surface with the anterior aspect facing down. By doing so, the rotation of the bone was entirely controlled by the orientation of the trochlea (anterior tangent to the trochlea). A photographic projection that used a 100-mm objective in line with the axis of the humerus was used to visualize and measure the amount of retrotorsion (Figure 1, C). All other measurements were made on the basis of 2 standardized radiologic projections in the frontal and sagittal plane (ie, parallel and orthogonal to the equatorial plane). The source-to-film distance was 120 cm. The mean magnification factor was 1.0178. The data were corrected for magnification. The technique of measurement was validated in a pilot study on 20 specimens in which the mean interobserver correlation coefficient was 0.94 (SD, 0.067; minimum, 0.74; maximum, 0.99). Parametric tests were used for statistical analysis. The normal distribution of the data was confirmed by graphical analysis. Correlation coefficients
331
332
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg July/August 2002
Figure 1 A-D, Measured distances and angles. ARC, Surface arc of the head; AX, anatomic axis of the proximal humerus; CCD, head inclination angle; CD, greater tuberosity offset (or critical distance); CL, critical line; CP, critical point; CYL, best-fitting cylinder; EQ, head equator; FBD, frontal diameter of the base of the head; FR, radius of curvature of the head in the frontal plane; HH, head height; MO, medial offset of the head; PO, posterior offset of the head; RT, retrotorsion of the humeral head; SBD, sagittal diameter of the base of the head; SE, distance between sulcus and equator.
and P values were calculated. The data are presented as mean values, first SD, range, and 10th and 90th percentiles.
Measured variables and definitions Axis of the proximal humerus was obtained through serial midpoint determination on both radiographic projec-
tions (outer cortex) in a proximal segment defined by the 10% and 40% total bone length marks. Best-fitting cylinder is defined as the largest cylinder fitting into the proximal medullary cavity between 10% and 40% of total bone length on both radiographic projections. Center of the head was determined geometrically and
J Shoulder Elbow Surg Volume 11, Number 4
double-checked with circular templates on both radiographic projections. Clearance best-fitting cylinder to greater tuberosity is defined as the shortest distance from the lateral wall of the best-fitting cylinder to the critical point measured on the anteroposterior radiographic projection. Critical line is the most medial insertion line of the supraspinatus tendon. In well-preserved macerated bone without rotator cuff disease, this line is clearly visible as a small step off at the edge of the subchondral bone plate. Critical point is the intersection between the critical line and the humeral head equator. Distance sulcus-equator is the distance between the midpoint of the bicipital groove and the critical point measured directly on the bones. Frontal diameter of the base of the head is the diameter of the head segment at the anatomic neck measured directly on the bones. Head equator is the equatorial line drawn directly on the bones and obtained as follows: the bone was mounted in a frame that allowed a 45° oblique medial-to-lateral projection of the humeral head. The humerus was rotated until the base of the head was perpendicular to the axis of the proximal humerus (Figure 2). The distal articular midpoint was defined as the mid distance between the anterior and posterior tangent to the head as viewed from the 45° oblique medial-to-lateral projection. A line was then drawn on the head starting from the distal articular midpoint running parallel to the axis of the proximal humerus and bisecting the head. This method was derived from Tillet et al.17 Head height determines the height or thickness of the humeral head and was measured in the anteroposterior radiographic projection. Head inclination angle is defined as the inclination of the anatomic neck in the frontal plane. The angle was measured on the anteroposterior radiographic projection and was formed by the intersection of the axis of the proximal humerus and the line perpendicular to the anatomic neck. Head surface arc represents the arc of articular surface available in the frontal plane. The angle was measured on the anteroposterior radiographic projection. Medial offset of the head is the shortest distance between the axis of the proximal humerus and the center of the head measured on the anteroposterior projection. Posterior offset of the head is the shortest distance between the axis of the proximal humerus and the center of the head measured on the lateral projection. Retrotorsion of the humeral head is the angle between the projection of the equatorial plane and the plane tangent to the anterior aspect of the trochlea measured on a photographic projection. Radius of the curvature of the head in the frontal plane is the radius of the head measured on the anteroposterior radiographic projection. Radius of the curvature of the head in the sagittal plane was calculated based on the sagittal diameter of the base of the head and the head height. Sagittal diameter of the base of the head is the maximal distance between the anterior and posterior head-collum junction in a plane perpendicular to the equatorial plane measured directly on the bones.
Hertel, Knothe, and Ballmer
333
Figure 2 Methodology. For the identification of the head equator, the bones were mounted in a special frame and manually rotated until the base of the head was perpendicular to the axis of the humerus (A and B). The midpoint a was determined. A line was then drawn on the head starting from the distal articular midpoint running parallel to the axis of the proximal humerus and bisecting the head (C).
Total bone length is the distance between the superior tangent to the humeral head and the inferior tangent to the trochlea measured directly on the bones. Greater tuberosity offset (or critical distance) is the short-
334
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg July/August 2002
Figure 3 Histograms of the most relevant variables.
est distance between the axis of the proximal humerus and the critical point measured directly on the bones (Figures 3-5).
RESULTS Some variables showed a large amount of scatter, whereas others, in particular angles and ratios, showed minimal variations. The results are summarized in Table I. Head dimensions
The mean radius of curvature of the head showed a 12% difference in the frontal and sagittal planes. The sagittal diameter of the base of the head was 6% smaller than the frontal diameter. The base of the resected head best fit an ellipse, with the long-axis in line with the head equator (retrotorsion axis). The mean head height measured 71% of the radius in the frontal plane.
Shaft dimensions
Total bone length and best-fitting cylinders were quite variable. The diameters of the best-fitting cylinders were not only related to the size and to the outer shape of the bone but also to the cortical thickness (ie, osteoporosis). The anatomic axis, on the contrary, was not altered by variation of cortical thickness. 11.5% of the specimens featured a marked bend (⬎5°) in the medullary cavity of the proximal humerus. The angulation, when present, was located at 13.5 ⫾ 0.72 cm (minimum, 13; maximum, 15), distal to the superior humeral head tangent. The shape of the proximal medullary cavity was roughly cylindrical in the remaining specimens. Head position in relation to shaft
The position of the head in relation to the shaft was quite scattered for variables such as the medial and posterior offset, the greater tuberosity offset, and the
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg Volume 11, Number 4
335
Figure 4 High-offset (1), standard (2), and low-offset (3) morphotypes. With decreasing critical distance, the introduction of a straight-stemmed prosthesis becomes increasingly difficult.
Figure 5 The introduction of a straight-stemmed canal fitting prosthesis may damage the insertion of the supraspinatus tendon.
retrotorsion. In contrast, the head inclination and the distance from the sulcus to the equator showed little variation (Figures 3 and 4). The distance between the lateral border of the bestfitting cylinder and the most medial insertion line of the supraspinatus tendon (clearance) was negative in 49.5% of the specimens. In these cases introduction of a canal-fitting, straight-stem prosthesis without fin would damage the insertion facet of the supraspinatus tendon (Figure 5). Adding a 4-mm flange to the canal-fitting shaft would damage the insertion facet of the supraspinatus in 94.5% of the measured specimens.
Correlations
Relevant correlations are shown in Table II. Of particular interest was the high correlation between the radius and the height of the head, as well as the high correlation between the frontal and sagittal radius. This suggests that the relative shape of the head is largely constant and that it can be accurately calculated based on the frontal radius value. On the other hand, there was no measurable correlation between the size of the head and its
336
Hertel, Knothe, and Ballmer
J Shoulder Elbow Surg July/August 2002
Table I Anthropometrical measurements
Dimensions of head Radius in frontal plane (mm) Radius in sagittal plane (mm) Ratio of frontal sagittal radius (mm) Diameter of head base in frontal plane (mm) Diameter of head base in sagittal plane (mm) Ratio of frontal-sagittal diameter Head height (mm) Ratio of frontal radius–head height Head surface arc (°) Dimensions of shaft Total length of humerus (cm) Best-fitting cylinder (mm) Clearance: Best-fitting cylinder to greater tuberosity (mm) Relative position of head on shaft Medial offset (mm) Posterior offset (mm) Greater tuberosity offset (mm) Head inclination angle (°) Retrotorsion (°) Distance of sulcus to equator (mm)
10th Percentile
90th Percentile
30 26 1.01 57 53.5 1.15 22 0.85 163
21 18.5 0.83 39.5 37 0.88 14.4 0.64 138
26.5 23 0.94 50 46 1.04 18.8 0.77 153
24.5 6 ⫺7.9
36.8 21 5.5
30 9 ⫺3.6
34.5 14 3.3
1.7 ⫺3 ⫺2.3 128 ⫺9 3.5
11.5 5.3 12 145.5 50 13.7
3.9 ⫺0.4 2.5 132 7 6
8.6 3.2 9.2 142 38.5 10.5
Mean
SD
Minimum
24 21 0.88 44.5 42 0.94 17 0.71 145
2.2 1.8 0.04 4 3.8 0.06 1.7 0.05 5.95
19 17 0.78 36 33.5 0.85 12.5 0.56 130
31.6 11.5 ⫺0.1
2.3 2.09 2.6
6 1.4 5.6 137 23.3 8.3
1.81 1.43 2.58 3.62 11.75 1.72
Maximum
Table II Correlations Variable a
Variable b
Frontal head radius Diameter of head base in frontal plane Head height Head inclination angle Medial offset Greater tuberosity offset Total length Total length Head height Total length Distance of sulcus to equator Diameter of head base in frontal plane Medial offset Posterior offset Total length Frontal head radius Retrotorsion
Sagittal head radius Diameter of head base in sagittal plane Frontal head radius Greater tuberosity offset Greater tuberosity offset Medial offset Frontal head radius Diameter of head base in frontal plane Total length Retrotorsion Total length Medial offset Head inclination angle Retrotorsion Medial offset Posterior offset Distance of sulcus to equator
relative position on the shaft. High- and low-offset morphotypes (Figure 4) did not correlate with the size of the bone nor with the head inclination angle. DISCUSSION The aim of this study was to add detail information to already available databases on the geometry of the proximal humerus. Although the results are largely in accordance with published data, the interpretation
Correlation factor
P value
0.84 0.75 0.7 0.69 0.64 0.64 0.62 0.61 0.57 0.27 0.27 0.26 0.23 0.22 0.14 0.11 0.04
⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 .1724 ⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 ⬍.0001 .0007 .0009 .0017 .0157 .1062 .5055
of several measurements was quite discordant. To optimize statistical power, measurements were taken from a large number of specimens—the largest thus far. In an effort to reduce methodological errors, the definition of landmarks was made as precisely and as reproducibly as possible and data acquisition was kept relatively simple. Critical issues included determination of the equatorial plane, determination of the anatomic axis of the proximal humerus, and determination of the center of the head. The described methodology provided a high degree of accuracy. Other
J Shoulder Elbow Surg Volume 11, Number 4
authors have determined the humeral axis by either manual antegrade15 or retrograde1 reaming. The potential methodological errors inherent to these techniques have not been assessed. Calculation of the axis based on computer-assisted mapping of the surface of the bone2,16,18 can be influenced by bone ridges. Comparing our data with surface-mapped data,2,16,18 it can be noted that the axis calculated on the base of surface mappings lies slightly more anterior (1-2 mm). The poorly defined, irregular, and S-shaped articular margin complicates determination of retrotorsion. Construction of the head equator according to the presented protocol, which is a modification of the method of Tillet et al,17 was reproducible. In order to highlight the range of variation of normal anatomy, results were also expressed in percentiles. The range between the 10th and 90th percentile comprises 80% of the values, which was judged to reflect the clinically relevant variation of normality better than the description of only the mean and first SD. As expected, the variation in the total length of the humeral specimens and in the size of the humeral head was large.2,15,18 However, the ratio of the frontal radius of the head versus head height remained remarkably constant (0.71 ⫾ 0.05). Pearl and Volk15 found a similar ratio of 0.73 ⫾ 0.04. The surface arc is yet another way to express the same relationship. It measured 145° ⫾ 5.9°. Pearl and Volk reported a surface arc of 150° ⫾ 5°, and Jobe and Iannotti8 found a slightly larger value, 159° ⫾ 8.5°. The implication for prosthetic design is that, for a given radius, only one anatomic head height is necessary for near-normal reconstruction of the outer contour of the proximal humerus. The radius of the head showed a 12% difference between the frontal and sagittal planes. Iannotti et al7 reported an 8% mismatch. Thus, the head is not exactly spherical. The impact of this finding on prosthetic design remains unclear. Choosing the right head in the frontal plane would result in an overlap of the head in the sagittal plane of about 3 mm. We did not observe a gradual decrease in the radius of the head in the frontal plane, as suggested by Kapandji.9 The question of whether the reason was methodological (ie, if the technique used might have been too coarse to detect small variations of the radius) remains unanswered. The relative position of the head on the shaft was quite variable. The mean medial offset was 6 ⫾ 1.8 mm. McPherson et al12 reported 7.6 ⫾ 3.2 mm, whereas Pearl and Volk15 reported slightly larger values, with a mean medial offset of 9.7 ⫾ 1.7 mm. Variations in the medial offset are of paramount importance in the design of humeral prosthetic components in order to respect the most medial insertion line
Hertel, Knothe, and Ballmer
337
of the rotator cuff. To snugly position the head against the most medial insertion line of the supraspinatus, a shaft with variable medial offset is necessary. The mean posterior offset was relatively small (mean, 1.4 ⫾ 1.44 mm). McPherson et al12 measured 1.9 ⫾ 1.7 mm, and Ballmer et al1 reported 2 mm. Using a surface-mapping method, Walch and Boileau2,18 found slightly larger values (mean 2.5 mm), and Roberts et al16 measured a mean value as high as 4.7 mm. These differences are probably due to the use of the frontal—as opposed to the equatorial reference plane. The offset of the greater tuberosity is a crucial variable. It indicates the distance between the insertion of the innermost fibers of the supraspinatus to the axis of the proximal humerus. On its basis, we classified the humeri into 3 morphotypes: a so-called standard humerus, a high-offset morphotype, and a low-offset morphotype (Figure 4). The implication is that straight-stemmed prostheses cannot be introduced in a low-offset humerus without relevant damage to the insertion facet of the supraspinatus tendon (Figure 5). Flanges and press-fit designs represent an additional hazard. Implantation of a straightstemmed, canal-fitting shaft would have caused damage to the insertion of the rotator cuff in 49.5% of the measured specimens. The inclination of the humeral head did not vary much (mean, 137° ⫾ 3.6°) and occurred in only 1 plane. Other investigators found similar mean values ranging from 130° to 141° and similar SDs ranging from 3.4° to 5°.2,3,12,14,18 Considering minimal and maximal values, published inclinations range from 120° to 145°.2,7,15 It is questionable whether the extreme values should be used to support arguments in favor of or against variable inclination designs because these values are single observations that may well be outliers and that have no specific representation in the Gauss curve. Therefore, use of the second SD or, better, the 10th and 90th percentiles appears to reflect more adequately the practical needs. Assuming a possible maximum mismatch of ⫹5° or ⫺5° (corresponding to the 10th and 90th percentiles), the effect on the outer shape of the proximal humerus would remain irrelevant. With 5° of valgus error, the mean humeral width2,14,18 would be increased by only 1.5 mm, or 3%, and the humeral height2,14,18 would be affected even less. Therefore, when the aim is to reconstruct normal or near-normal anatomy of the proximal humerus variable prosthetic head inclinations are not required. All authors of anatomic studies have found large variations in retrotorsion. The mean values ranged from 18° to 33°.1-3,5,6,14,15,18 In vivo radiologic studies10 seem to have less scattered results. The reason is unclear, as methodological errors would be expected to be larger in the in vivo measurements. When
338
Hertel, Knothe, and Ballmer
comparing data, it is essential to consider that measurements taken using the epicondylar axis differ by about ⫺10° from measurements taken using the anterior tangent to the trochlea.6 We prefer the anterior tangent because it can be determined with precision on dry bones and adequately assessed during surgery (perpendicular to the axis of the forearm). The relative position of the bicipital groove to the head equator showed little variation (mean distance, 8.3 mm) and is, as we and others17 believe, a useful intraoperative landmark for positioning the head in the right retrotorsion. Tillet et al17 found a mean distance of 9 mm. However, other authors1,4,11 found little or no correlation at all. The reason for these contrasting conclusions is most likely methodological. So far determination of the equatorial plane of the humeral head has not been addressed with sufficient clarity. Clear definitions are missing, and reproducibility has not been tested. We believe that adequate definition of the anatomic landmarks is the key issue to obtain useful results. In conclusion, the anatomy of the proximal humerus was largely scattered for variables such as the medial offset, the greater tuberosity offset, and the retrotorsion. Yet, for the inclination of the humeral head, the head radius– height ratio, and the distance between the bicipital groove to the equator, the practically relevant variance was minimal. These findings yield pertinent implications for prosthetic design and implantation. REFERENCES
1. Ballmer FT, Lippitt SB, Romeo AA, Matsen FAI. Humeral prosthetic arthroplasty: surgically relevant considerations. J Shoulder Elbow Surg 1993;2:296-304. 2. Boileau P, Walch G. The three-dimensional geometry of the proximal humerus. Implications for surgical technique and prosthetic design. J Bone Joint Surg Br 1997;79:857-65. 3. Clarke IC, Gruen TAW, Sew Hoy A, Hirschowitz D, Maki S,
J Shoulder Elbow Surg July/August 2002
4. 5. 6. 7.
8. 9. 10. 11. 12. 13. 14. 15. 16. 17.
18.
Amstutz HC. Problems in gleno-humeral surface replacements— real or imagined? IMechE 1979;8:161-75. Doyle A, Burks R. Comparison of humeral head retroversion with the humeral axis/biceps groove relationship: a study in live subjects and cadavers. J Shoulder Elbow Surg 1998;7:453-7. Edelson G. Variations in the retroversion of the humeral head. J Shoulder Elbow Surg 1999;8:142-5. Hernigou P, Duparc F, Filali C. Humeral retroversion and shoulder prosthesis [in French]. Rev Chir Orthop Reparatrice Appar Mot 1995;81:419-27. Iannotti JP, Gabriel JP, Schneck SL, Evans BG, Misra S. The normal glenohumeral relationships. An anatomical study of one hundred and forty shoulders. J Bone Joint Surg Am 1992;74:491500. Jobe CM, Iannotti JP. Limits imposed on glenohumeral motion by joint geometry. J Shoulder Elbow Surg 1995;4:281-5. Kapandji IA. Physiologie articulaire, 5th ed. Malone S.A. Editor: Paris; 1980. Kronberg M, Brostrom LA, Soderlund V. Retroversion of the humeral head in the normal shoulder and its relationship to the normal range of motion. Clin Orthop 1990;253:113-7. Kummer F, Perkins R, Zuckerman J. The use of the bicipital groove for alignment of the humeral stem in shoulder arthroplasty. J Shoulder Elbow Surg 1998;7:144-6. McPherson EJ, Friedman RJ, An YH, Chokesi R, Dooley RL. Anthropometric study of normal glenohumeral relationships. J Shoulder Elbow Surg 1997;6:105-12. Pearl ML, Kurutz S. Geometric analysis of commonly used prosthetic systems for proximal humeral replacement. J Bone Joint Surg Am 1999;81:660-71. Pearl ML, Volk AG. Retroversion of the proximal humerus in relationship to prosthetic replacement arthroplasty. J Shoulder Elbow Surg 1995;4:286-9. Pearl ML, Volk AG. Coronal plane geometry of the proximal humerus relevant to prosthetic arthroplasty. J Shoulder Elbow Surg 1996;5:320-6. Roberts S, Foley A, Swallow H, Wallace W, Coughlan D. The geometry of the humeral head and the design of prostheses. J Bone Joint Surg Br 1991;73:647-50. Tillet E, Smith M, Fulcher M, Shanklin J. Anatomic determination of humeral head retroversion: the relationship of central axis of the humeral head to the bicipital groove. J Shoulder Elbow Surg 1993;2:225-56. Walch G, Boileau P. Morphological study of the humeral proximal epiphysis [abstract]. J Bone Joint Surg Br 1992;74:14.