- Email: [email protected]
Association Between Lunate Morphology and Carpal Collapse Patterns in Scaphoid Nonunions
Association Between Lunate Morphology and Carpal Collapse Patterns in Scaphoid Nonunions Steven C. Haase, MD, Richard A. Berger, MD, PhD, Alexander Y. Shin, MD From the Department of Surgery, University of Michigan Health System, Ann Arbor, MI; and the Department of Orthopedic Surgery, Mayo Clinic, Rochester, MN.
Purpose: Type I lunates have a single distal facet for articulation with the capitate; type II lunates have an additional (medial) hamate facet on the distal articular surface. We retrospectively reviewed a series of patients with scaphoid nonunions to determine if there was an association between lunate morphology and the degree of carpal instability observed. Association between lunate morphology and the location of the scaphoid fracture (proximal or waist) was also investigated. Methods: Radiographs were evaluated for 45 patients with established scaphoid nonunions. Lunate morphology, scaphoid fracture location, and radiolunate angle were determined. Results: Type I lunates were present in 21 patients. Of these, 15 were found to have a dorsal intercalated segment instability pattern (radiolunate angle greater than 15°). By contrast, only 4 of the patients with type II lunates exhibited this pattern of instability. No significant association was found between lunate morphology and the scaphoid fracture location. Conclusions: Type II lunate morphology is associated with significantly decreased incidence of dorsal intercalated segment instability (DISI) deformity in cases of established scaphoid nonunion (p ⫽ .0002). Lunate morphology, however, was not significantly associated with the location of the scaphoid fracture in these cases (p ⫽ .19). (J Hand Surg 2007;32A: 1009 –1012. Copyright © 2007 by the American Society for Surgery of the Hand.) Type of study/level of evidence: Prognostic IV. Key words: Carpal instability, DISI, lunate, morphology, scaphoid nonunion.
he lunate has often been described as the cornerstone or keystone of the wrist.1,2 This title is well deserved, as this carpal bone is uniquely positioned at the middle of both the transverse and coronal arches of the carpus. Critical ligamentous attachments between the lunate, the scaphoid, and the triquetrum allow for stability within the proximal carpal row. Additionally, the lunate is the intercalated segment between the radius and distal carpal row. Variations in lunate morphology have been described in several different ways. In 1966, Antuna Zapico divided lunates into three types (I, II, and III) based on whether the proximal surface was curved or angulated.3 Watson’s three lunate types (D, V, and N) are based on the lateral appearance on radiography.4 In 1990, Viegas classified lunates by their
T
distal articular morphology5,6: type I lunates with a single capitate facet and type II lunates with an additional hamate facet on the medial portion of the distal articular surface, as shown in Figure 1. In the same year, Burgess also described this finding, along with matching variations of hamate morphology, referring to these as type I and type II midcarpal joints.7 The presence of a medial (hamate) facet has been linked to increased incidence of proximal hamate arthrosis.8 –10 Compared with type I lunates, type II lunates have considerably different kinematics during radial-ulnar deviation of the wrist.11 Anecdotally, we have noticed a tendency for wrists with type II lunate morphology to exhibit less dorsal intercalated segment instability (DISI) deformity, in cases where this deformity is commonplace. The Journal of Hand Surgery
1009
1010
The Journal of Hand Surgery / Vol. 32A No. 7 September 2007
Figure 1. Type I lunates (L) have a single distal facet that articulates with the capitate (C). Type II lunates have an additional facet for articulation with the hamate (H). (T ⫽ triquetrum, S ⫽ scaphoid) [Copyright Mayo Foundation].
DISI deformity12 describes the abnormally extended posture of the lunate bone relative to the longitudinal axis of the radius, as seen on lateral radiographs. This radiographic manifestation of carpal instability is often seen in cases of scapholunate dissociation or scaphoid nonunion. Both of these cases are forms of the carpal instability dissociative (CID) pattern, and both lead to abnormal extension of the lunate, as it is dissociated from the flexion moment of the distal scaphoid. The purpose of this study was to evaluate the association, if any, between lunate morphology and the presence of DISI deformity. We also wanted to determine if there was any association between lunate morphology and the site of the scaphoid fracture (proximal pole or waist). We believe these associations, if present, may lead to new understandings regarding instabilities of the wrist. Our null hypotheses were as follows: (1) Lunate morphology is NOT associated with the presence of DISI deformity. (2) Lunate morphology is NOT associated with the location (proximal or waist) of scaphoid fractures. To test these hypotheses, we set out to perform a retrospective review of a group of patients predisposed to DISI deformity (patients with scaphoid nonunions).
Materials and Methods Institutional review board approval was granted for this study. Surgical records for the years 1994 through 2003 were searched, and 52 consecutive patients who had undergone vascularized bone grafts for established scaphoid nonunion were identified. The records for these patients were reviewed, with specific regard to preoperative radiographic studies.
Seven patients were excluded due to lack of adequate preoperative radiographs. The remaining 45 patients were included in this analysis. Preoperative wrist x-ray films were reviewed for each patient, determining radiolunate angle, scaphoid fracture location, and lunate morphology (Fig. 2). The radiolunate angle was determined from lateral wrist radiographs using the tangential method.13 The patient was determined to have DISI deformity if the radiolunate angle was greater than 15°.14 Lunate morphology and scaphoid fracture location were determined by examination of standard posterioanterior wrist radiographs. If a medial lunate facet could be identified, the lunate was classified as type II.5 Statistical analysis was performed using the chisquare test to determine if there was a statistically significant relationship between the variables measured. Statistical significance was set at p ⬍ .05.
Results The study group was composed of 37 males and 8 females. The average age was 23 (range 13– 66). Type I lunates were identified in 21 (47%) cases. Table 1 shows the distribution of lunate morphology and DISI deformity. Chi-square analysis shows that this relationship is statistically significant (p ⫽ .0002). Nine patients with type I lunates sustained a proximal pole fracture. Of the patients with type II lunates, 15 had proximal pole fractures. On chi-square analysis, this distribution was not statistically significant. Further analysis of the data was conducted to see if there was an association between fracture location and the development of DISI deformity. Of the proximal pole fracture patients, 5 developed DISI deformity, whereas 14 patients with scaphoid waist fractures developed DISI deformity. This was found to be a significant association (p ⫽ .002).
Discussion Scaphoid nonunion is a condition well known to predispose to carpal instability, specifically DISI deformity. Upon close examination of a cohort of scaphoid nonunion patients, we found a significantly decreased incidence of DISI deformity among those
Table 1. Lunate Morphology and DISI Deformity Lunate Morphology
DISI Present
DISI Absent
Type I Type II
15 4
6 20
Haase, Berger, and Shin / Lunate Morphology and Carpal Instability
1011
Figure 2. Posterioanterior (A) and lateral radiographs (B) of a typical patient’s wrist with a type I lunate, showing evidence of DISI deformity. Posterioanterior (C) and lateral (D) radiographs of a type II lunate wrist, without evidence of DISI deformity.
patients with type II lunates. One possible interpretation of this finding is that type II lunate morphology is protective against DISI deformity in this clinical setting. We theorize that the lunate’s added articulation with the hamate lends some additional stability that resists abnormal extension. It could be that the triquetrum’s usual “extension moment,” caused by force transmitted through its screw-like articulation with the hamate, is halted by the lunatohamate articulation present in type II midcarpal joints. The reported incidence of type II lunate in the literature ranges from 46% to 73%.5,7,8,15,16 Our finding of 53% compares favorably with these reports, despite our relatively small sample size. Sagerman et al specifically examined whether or not lunate morphology could be accurately predicted
with routine radiographs.9 They found that only 74% of type I lunates and 66% of type II lunates could be correctly identified on plain x-rays. Relying on plain radiographs for our determinations may have caused us to overlook some type II lunates with very narrow medial (hamate) facets. It is possible that these very narrow medial facets are less clinically important, but additional studies would be needed to confirm this. X-ray findings of an anatomic relationship do not always have clinical implications. There is, however, ample evidence that the lunatohamate articulation, when present, is clinically relevant. Many authors have noted that wrists with type II lunates are predisposed to proximal hamate arthrosis at this location.6,9 In a review of 186 wrist MRIs, Malik et al found that 77% of lunates with type II morphology
1012
The Journal of Hand Surgery / Vol. 32A No. 7 September 2007
were physically apposed (articulated) with the hamate on static magnetic resonance images. Furthermore, hamate edema, which was found in 5% of studies, only occurred in wrists with type II lunates.15 Finally, a biomechanical study by Nakamura et al has shown that there are real differences in the kinematics of wrists with type I versus type II lunates.11 All of this evidence points to the conclusion that this lunatohamate articulation is clinically and biomechanically important. We believe it adds substantially to the stability of the proximal row in cases where dissociative instability is likely. Further biomechanical studies are indicated to detect the exact effect the type II lunate has on carpal instability. Regarding the other findings of this study, it is possible the rejection of our second hypothesis is invalid due to a beta error. Sample sizes, although satisfactory to make chi-square a valid test, were indeed small in this study. Analysis of the data also revealed an association between fracture location and DISI deformity. The increased incidence of DISI deformity in more distal fracture patterns correlates nicely with the findings of previous studies,17 although this was not one of our study’s a priori hypotheses. In conclusion, our primary null hypothesis is rejected. There is a statistically significant association between lunate morphology and the development of DISI deformity. There was, however, no significant association detected between lunate morphology and the location of the scaphoid fracture.
2. 3. 4.
5.
6. 7. 8.
9.
10.
11.
12.
13. 14.
Received for publication March 25, 2007; accepted in revised form June 11, 2007. 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. Corresponding author: Steven C. Haase, MD, University of Michigan Health System, 2130 Taubman Center, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0340; e-mail: [email protected]. Copyright © 2007 by the American Society for Surgery of the Hand 0363-5023/07/32A07-0010$32.00/0 doi:10.1016/j.jhsa.2007.06.005
References 1. Sennwald G, Segmuller G. Base anatomique d’un nouveau concept de stabilité du carpe. [Anatomic basis of a new
15.
16.
17.
concept of stability of the carpus]. Int Orthop 1986;10: 25–30. Amadio PC. Carpal kinematics and instability: a clinical and anatomic primer. Clin Anat 1991;4:1–12. Taleisnik J. The wrist. New York: Churchill Livingstone, 1985:171–172. Watson HK, Yasuda M, Guidera PM. Lateral lunate morphology: an x-ray study. J Hand Surg 1996;21A:759 – 763. Viegas SF, Wagner K, Patterson R, Peterson P. Medial (hamate) facet of the lunate. J Hand Surg 1990;15A:564 – 571. Viegas SF. The lunatohamate articulation of the midcarpal joint. Arthroscopy 1990;6:5–10. Burgess RC. Anatomic variations of the midcarpal joint. J Hand Surg 1990;15A:129 –131. Viegas SF, Patterson RM, Hokanson JA, Davis J. Wrist anatomy: incidence, distribution, and correlation of anatomic variations, tears, and arthrosis. J Hand Surg 1993;18A:463– 475. Sagerman SD, Hauck RM, Palmer AK. Lunate morphology: can it be predicted with routine x-ray films? J Hand Surg 1995;20A:38 – 41. Nakamura K, Patterson RM, Moritomo H, Viegas SF. Type I versus type II lunates: ligament anatomy and presence of arthrosis. J Hand Surg 2001;26A:428 – 436. Nakamura K, Beppu M, Patterson RM, Hanson CA, Hume PJ, Viegas SF. Motion analysis in two dimensions of radialulnar deviation of type I versus type II lunates. J Hand Surg 2000;25A:877– 888. Linscheid RL, Dobyns JH, Beabout JW, Bryan RS. Traumatic instability of the wrist. Diagnosis, classification, and pathomechanics. J Bone Joint Surg 1972;54A:1612– 1632. Gilula LA, Weeks PM. Post-traumatic ligamentous instabilities of the wrist. Radiology 1978;129:641– 651. Nakamura R, Hori M, Imamura T, Horii E, Miura T. Method for measurement and evaluation of carpal bone angles. J Hand Surg 1989;14A:412– 416. Malik AM, Schweitzer ME, Culp RW, Osterman LA, Manton G. MR imaging of the type II lunate bone: frequency, extent, and associated findings. AJR Am J Roentgenol 1999;173:335–338. Aufauvre B, Herzberg G, Garret J, Berthonneaud E, Dimnet J. A new radiographic method for evaluation of the position of the carpus in the coronal plane: results in normal subjects. Surg Radiol Anat 1999;21:383–385. Moritomo H, Viegas SF, Elder KW, Nakamura K, Dasilva MF, Boyd NL, et al. Scaphoid nonunions: a 3dimensional analysis of patterns of deformity. J Hand Surg 2000;25A:520 –528.
Purpose: Type I lunates have a single distal facet for articulation with the capitate; type II lunates have an additional (medial) hamate facet on the distal articular surface. We retrospectively reviewed a series of patients with scaphoid nonunions to determine if there was an association between lunate morphology and the degree of carpal instability observed. Association between lunate morphology and the location of the scaphoid fracture (proximal or waist) was also investigated. Methods: Radiographs were evaluated for 45 patients with established scaphoid nonunions. Lunate morphology, scaphoid fracture location, and radiolunate angle were determined. Results: Type I lunates were present in 21 patients. Of these, 15 were found to have a dorsal intercalated segment instability pattern (radiolunate angle greater than 15°). By contrast, only 4 of the patients with type II lunates exhibited this pattern of instability. No significant association was found between lunate morphology and the scaphoid fracture location. Conclusions: Type II lunate morphology is associated with significantly decreased incidence of dorsal intercalated segment instability (DISI) deformity in cases of established scaphoid nonunion (p ⫽ .0002). Lunate morphology, however, was not significantly associated with the location of the scaphoid fracture in these cases (p ⫽ .19). (J Hand Surg 2007;32A: 1009 –1012. Copyright © 2007 by the American Society for Surgery of the Hand.) Type of study/level of evidence: Prognostic IV. Key words: Carpal instability, DISI, lunate, morphology, scaphoid nonunion.
he lunate has often been described as the cornerstone or keystone of the wrist.1,2 This title is well deserved, as this carpal bone is uniquely positioned at the middle of both the transverse and coronal arches of the carpus. Critical ligamentous attachments between the lunate, the scaphoid, and the triquetrum allow for stability within the proximal carpal row. Additionally, the lunate is the intercalated segment between the radius and distal carpal row. Variations in lunate morphology have been described in several different ways. In 1966, Antuna Zapico divided lunates into three types (I, II, and III) based on whether the proximal surface was curved or angulated.3 Watson’s three lunate types (D, V, and N) are based on the lateral appearance on radiography.4 In 1990, Viegas classified lunates by their
T
distal articular morphology5,6: type I lunates with a single capitate facet and type II lunates with an additional hamate facet on the medial portion of the distal articular surface, as shown in Figure 1. In the same year, Burgess also described this finding, along with matching variations of hamate morphology, referring to these as type I and type II midcarpal joints.7 The presence of a medial (hamate) facet has been linked to increased incidence of proximal hamate arthrosis.8 –10 Compared with type I lunates, type II lunates have considerably different kinematics during radial-ulnar deviation of the wrist.11 Anecdotally, we have noticed a tendency for wrists with type II lunate morphology to exhibit less dorsal intercalated segment instability (DISI) deformity, in cases where this deformity is commonplace. The Journal of Hand Surgery
1009
1010
The Journal of Hand Surgery / Vol. 32A No. 7 September 2007
Figure 1. Type I lunates (L) have a single distal facet that articulates with the capitate (C). Type II lunates have an additional facet for articulation with the hamate (H). (T ⫽ triquetrum, S ⫽ scaphoid) [Copyright Mayo Foundation].
DISI deformity12 describes the abnormally extended posture of the lunate bone relative to the longitudinal axis of the radius, as seen on lateral radiographs. This radiographic manifestation of carpal instability is often seen in cases of scapholunate dissociation or scaphoid nonunion. Both of these cases are forms of the carpal instability dissociative (CID) pattern, and both lead to abnormal extension of the lunate, as it is dissociated from the flexion moment of the distal scaphoid. The purpose of this study was to evaluate the association, if any, between lunate morphology and the presence of DISI deformity. We also wanted to determine if there was any association between lunate morphology and the site of the scaphoid fracture (proximal pole or waist). We believe these associations, if present, may lead to new understandings regarding instabilities of the wrist. Our null hypotheses were as follows: (1) Lunate morphology is NOT associated with the presence of DISI deformity. (2) Lunate morphology is NOT associated with the location (proximal or waist) of scaphoid fractures. To test these hypotheses, we set out to perform a retrospective review of a group of patients predisposed to DISI deformity (patients with scaphoid nonunions).
Materials and Methods Institutional review board approval was granted for this study. Surgical records for the years 1994 through 2003 were searched, and 52 consecutive patients who had undergone vascularized bone grafts for established scaphoid nonunion were identified. The records for these patients were reviewed, with specific regard to preoperative radiographic studies.
Seven patients were excluded due to lack of adequate preoperative radiographs. The remaining 45 patients were included in this analysis. Preoperative wrist x-ray films were reviewed for each patient, determining radiolunate angle, scaphoid fracture location, and lunate morphology (Fig. 2). The radiolunate angle was determined from lateral wrist radiographs using the tangential method.13 The patient was determined to have DISI deformity if the radiolunate angle was greater than 15°.14 Lunate morphology and scaphoid fracture location were determined by examination of standard posterioanterior wrist radiographs. If a medial lunate facet could be identified, the lunate was classified as type II.5 Statistical analysis was performed using the chisquare test to determine if there was a statistically significant relationship between the variables measured. Statistical significance was set at p ⬍ .05.
Results The study group was composed of 37 males and 8 females. The average age was 23 (range 13– 66). Type I lunates were identified in 21 (47%) cases. Table 1 shows the distribution of lunate morphology and DISI deformity. Chi-square analysis shows that this relationship is statistically significant (p ⫽ .0002). Nine patients with type I lunates sustained a proximal pole fracture. Of the patients with type II lunates, 15 had proximal pole fractures. On chi-square analysis, this distribution was not statistically significant. Further analysis of the data was conducted to see if there was an association between fracture location and the development of DISI deformity. Of the proximal pole fracture patients, 5 developed DISI deformity, whereas 14 patients with scaphoid waist fractures developed DISI deformity. This was found to be a significant association (p ⫽ .002).
Discussion Scaphoid nonunion is a condition well known to predispose to carpal instability, specifically DISI deformity. Upon close examination of a cohort of scaphoid nonunion patients, we found a significantly decreased incidence of DISI deformity among those
Table 1. Lunate Morphology and DISI Deformity Lunate Morphology
DISI Present
DISI Absent
Type I Type II
15 4
6 20
Haase, Berger, and Shin / Lunate Morphology and Carpal Instability
1011
Figure 2. Posterioanterior (A) and lateral radiographs (B) of a typical patient’s wrist with a type I lunate, showing evidence of DISI deformity. Posterioanterior (C) and lateral (D) radiographs of a type II lunate wrist, without evidence of DISI deformity.
patients with type II lunates. One possible interpretation of this finding is that type II lunate morphology is protective against DISI deformity in this clinical setting. We theorize that the lunate’s added articulation with the hamate lends some additional stability that resists abnormal extension. It could be that the triquetrum’s usual “extension moment,” caused by force transmitted through its screw-like articulation with the hamate, is halted by the lunatohamate articulation present in type II midcarpal joints. The reported incidence of type II lunate in the literature ranges from 46% to 73%.5,7,8,15,16 Our finding of 53% compares favorably with these reports, despite our relatively small sample size. Sagerman et al specifically examined whether or not lunate morphology could be accurately predicted
with routine radiographs.9 They found that only 74% of type I lunates and 66% of type II lunates could be correctly identified on plain x-rays. Relying on plain radiographs for our determinations may have caused us to overlook some type II lunates with very narrow medial (hamate) facets. It is possible that these very narrow medial facets are less clinically important, but additional studies would be needed to confirm this. X-ray findings of an anatomic relationship do not always have clinical implications. There is, however, ample evidence that the lunatohamate articulation, when present, is clinically relevant. Many authors have noted that wrists with type II lunates are predisposed to proximal hamate arthrosis at this location.6,9 In a review of 186 wrist MRIs, Malik et al found that 77% of lunates with type II morphology
1012
The Journal of Hand Surgery / Vol. 32A No. 7 September 2007
were physically apposed (articulated) with the hamate on static magnetic resonance images. Furthermore, hamate edema, which was found in 5% of studies, only occurred in wrists with type II lunates.15 Finally, a biomechanical study by Nakamura et al has shown that there are real differences in the kinematics of wrists with type I versus type II lunates.11 All of this evidence points to the conclusion that this lunatohamate articulation is clinically and biomechanically important. We believe it adds substantially to the stability of the proximal row in cases where dissociative instability is likely. Further biomechanical studies are indicated to detect the exact effect the type II lunate has on carpal instability. Regarding the other findings of this study, it is possible the rejection of our second hypothesis is invalid due to a beta error. Sample sizes, although satisfactory to make chi-square a valid test, were indeed small in this study. Analysis of the data also revealed an association between fracture location and DISI deformity. The increased incidence of DISI deformity in more distal fracture patterns correlates nicely with the findings of previous studies,17 although this was not one of our study’s a priori hypotheses. In conclusion, our primary null hypothesis is rejected. There is a statistically significant association between lunate morphology and the development of DISI deformity. There was, however, no significant association detected between lunate morphology and the location of the scaphoid fracture.
2. 3. 4.
5.
6. 7. 8.
9.
10.
11.
12.
13. 14.
Received for publication March 25, 2007; accepted in revised form June 11, 2007. 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. Corresponding author: Steven C. Haase, MD, University of Michigan Health System, 2130 Taubman Center, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0340; e-mail: [email protected]. Copyright © 2007 by the American Society for Surgery of the Hand 0363-5023/07/32A07-0010$32.00/0 doi:10.1016/j.jhsa.2007.06.005
References 1. Sennwald G, Segmuller G. Base anatomique d’un nouveau concept de stabilité du carpe. [Anatomic basis of a new
15.
16.
17.
concept of stability of the carpus]. Int Orthop 1986;10: 25–30. Amadio PC. Carpal kinematics and instability: a clinical and anatomic primer. Clin Anat 1991;4:1–12. Taleisnik J. The wrist. New York: Churchill Livingstone, 1985:171–172. Watson HK, Yasuda M, Guidera PM. Lateral lunate morphology: an x-ray study. J Hand Surg 1996;21A:759 – 763. Viegas SF, Wagner K, Patterson R, Peterson P. Medial (hamate) facet of the lunate. J Hand Surg 1990;15A:564 – 571. Viegas SF. The lunatohamate articulation of the midcarpal joint. Arthroscopy 1990;6:5–10. Burgess RC. Anatomic variations of the midcarpal joint. J Hand Surg 1990;15A:129 –131. Viegas SF, Patterson RM, Hokanson JA, Davis J. Wrist anatomy: incidence, distribution, and correlation of anatomic variations, tears, and arthrosis. J Hand Surg 1993;18A:463– 475. Sagerman SD, Hauck RM, Palmer AK. Lunate morphology: can it be predicted with routine x-ray films? J Hand Surg 1995;20A:38 – 41. Nakamura K, Patterson RM, Moritomo H, Viegas SF. Type I versus type II lunates: ligament anatomy and presence of arthrosis. J Hand Surg 2001;26A:428 – 436. Nakamura K, Beppu M, Patterson RM, Hanson CA, Hume PJ, Viegas SF. Motion analysis in two dimensions of radialulnar deviation of type I versus type II lunates. J Hand Surg 2000;25A:877– 888. Linscheid RL, Dobyns JH, Beabout JW, Bryan RS. Traumatic instability of the wrist. Diagnosis, classification, and pathomechanics. J Bone Joint Surg 1972;54A:1612– 1632. Gilula LA, Weeks PM. Post-traumatic ligamentous instabilities of the wrist. Radiology 1978;129:641– 651. Nakamura R, Hori M, Imamura T, Horii E, Miura T. Method for measurement and evaluation of carpal bone angles. J Hand Surg 1989;14A:412– 416. Malik AM, Schweitzer ME, Culp RW, Osterman LA, Manton G. MR imaging of the type II lunate bone: frequency, extent, and associated findings. AJR Am J Roentgenol 1999;173:335–338. Aufauvre B, Herzberg G, Garret J, Berthonneaud E, Dimnet J. A new radiographic method for evaluation of the position of the carpus in the coronal plane: results in normal subjects. Surg Radiol Anat 1999;21:383–385. Moritomo H, Viegas SF, Elder KW, Nakamura K, Dasilva MF, Boyd NL, et al. Scaphoid nonunions: a 3dimensional analysis of patterns of deformity. J Hand Surg 2000;25A:520 –528.