- Email: [email protected]
Risk Factors of Lunate Collapse in Kienböck Disease
EDITOR’S CHOICE
Risk Factors of Lunate Collapse in Kienböck Disease Wouter F. van Leeuwen, MD,*† Matthew A. Tarabochia, MD,* Arnold H. Schuurman, MD, PhD,† Neal Chen, MD,* David Ring, MD, PhD‡
Purpose Not all patients with Kienböck disease progress to collapse of the lunate and carpal malalignment, but it is difficult to determine which patients are at risk. We aimed to identify demographic or anatomical factors associated with more advanced stages of Kienböck disease. Methods We included all 195 eligible patients with Kienböck disease and available preoperative posteroanterior and lateral radiographs. We compared the mean age, sex distribution, mean ulnar variance, radial height, radial (ulnarward) inclination, palmar tilt, anteroposterior distance, and lunate type among the different Lichtman stages of Kienböck disease and performed ordinal logistic regression analysis. Results We found that patients with more negative ulnar variance had more advanced stages of Kienböck disease (adjusted odds ratio, 1.4). An increase in age was also independently associated with a higher Lichtman stage of Kienböck disease (adjusted odds ratio, 1.02). Conclusions Our findings suggest that more negative ulnar variance may be related to a greater magnitude of lunate collapse in Kienböck disease. Additional long-term study is needed to confirm the longitudinal relationship of negative ulnar variance with progressive Kienböck disease. (J Hand Surg Am. 2017;42(11):883e888. Copyright Ó 2017 by the American Society for Surgery of the Hand. All rights reserved.) Type of study/level of evidence Prognostic II. Key words Kienböck disease, lunatomalacia, ulnar variance, lunate collapse, Lichtman stage.
K
osteonecrosis of the lunate. The cause is unknown but some theories implicate mechanical influences on the lunate. Several anatomical risk factors are proposed and include negative ulnar variance,1e10 variations in lunate shape11,12 and radial (ulnarward) inclination,7,13,14 and a greater palmar tilt of the distal radial articular surface.14 IENBÖCK DISEASE IS IDIOPATHIC
From the *Department of Orthopaedic Surgery, Hand and Upper Extremity Service, Massachusetts General Hospital, Harvard Medical School, Boston, MA; the †Department of Plastic, Reconstructive and Hand Surgery, University Medical Center Utrecht, Utrecht, The Netherlands; and the ‡Department of Surgery and Perioperative Care, Dell Medical School, The University of Texas at Austin, Austin, TX.
Kienböck disease may lead to collapse of the lunate and eventually collapse of the proximal carpal row with proximal migration of the capitate and fixed volar flexion of the scaphoid, ultimately with progression to pancarpal arthrosis.15,16 However, Kienböck disease does not seem to progress in a notable number of patients and is occasionally diagnosed in the absence of symptoms.17e19 This work was performed at Massachusetts General Hospital, Boston, MA. Corresponding author: David Ring, MD, PhD, 1400 Barbara Jordan Blvd., Suite 1.114AC MC: R1800, Austin, TX 78723; e-mail: [email protected]. 0363-5023/17/4211-0005$36.00/0 http://dx.doi.org/10.1016/j.jhsa.2017.06.107
Received for publication July 27, 2016; accepted in revised form June 30, 2017. No benefits in any form have been received or will be received related directly or indirectly to the subject of this article.
Ó 2017 ASSH
r
Published by Elsevier, Inc. All rights reserved.
r
883
884
KIENBÖCK COLLAPSE RISK FACTORS
Unfortunately, we are currently unable to predict which patients are likely to progress to lunate collapse or osteoarthritis. Previous reports suggest that lunate collapse may be associated with differences in lunate shape12 and ulnar variance.20,21 The purpose of this study is to identify anatomical or demographic risk factors for lunate collapse and carpal malalignment in a cross-sectional cohort of patients with Kienböck disease. We hypothesize that there are no demographic or anatomical factors independently associated with more advanced disease stage according to the Lichtman radiographic classification of Kienböck disease.15,16 PATIENTS AND METHODS The review board at Massachusetts General Hospital approved the protocol for this retrospective study and granted a waiver of informed consent. We screened the radiology databases of 2 tertiary academic medical centers and 2 affiliated institutions for “Kienböck,” “lunatomalacia,” “lunate,” and/or “avascular necrosis,” as well as common misspellings and synonyms. We then manually reviewed the radiographic images and electronic medical records and included all 195 eligible adult patients with a confirmed diagnosis of Kienböck disease and in whom preoperative posteroanterior and lateral wrist radiographs were available. The radiographs were obtained between 1999 and 2016. We included the first available set of radiographs demonstrating Kienböck disease. In patients who had their initial diagnosis of Kienböck disease made based on magnetic resonance imaging (MRI) scans, we included the preoperative radiograph that was obtained closest to the date of the MRI. In 7 patients, the MRI was obtained a median of 23 days (interquartile range, 10e57 days) after the initial radiographs, and in 1 patient, radiographs were obtained 69 days after an MRI showed changes consistent with Kienböck disease. Outcome measures and study variables Our primary outcome measure was Lichtman stage of Kienböck disease. One investigator (W.F.v.L.) reviewed and staged all radiographs.15 Stage IIIB was defined as lunate collapse in the presence of a radioscaphoid angle of more than 60 as measured on lateral wrist radiographs.22 The radioscaphoid angle was measured between the longitudinal axis of the radius and a line tangential to the proximal and distal volar poles of the scaphoid.23 Another investigator (M.A.T.) performed all radiographic measurements J Hand Surg Am.
r
and was blinded to the study purpose to minimize observer bias. The ulnar variance, radial height, and radial (ulnarward) inclination were measured on posteroanterior wrist radiographs. We determined the central reference point—the midpoint between the volar and dorsal corners of the ulnar margins of the distal radial articular surface—as described by Medoff.24 We used the method of perpendiculars to measure the ulnar variance, radial height, and radial (ulnarward) inclination: a line was drawn through the longitudinal axis of the radius and a perpendicular line was constructed through the central reference point.25 Ulnar variance was measured as the distance between the perpendicular and a parallel line tangential to the distal surface of the ulnar head. The radial height was defined as the difference in axial length between the line through the central reference point and the tip of the radial styloid measured along the longitudinal axis of the radius. Radial (ulnarward) inclination was determined as the angle between the perpendicular and a line through the central reference point tangential to the tip of the radial styloid (Fig. 1).24 In addition, the absence (type 1 lunate) or presence (type 2 lunate) of a medial facet on the lunate in articulation with the hamate was determined as proposed by Viegas et al.26 Palmar tilt and anteroposterior distance were measured on lateral radiographs. Palmar tilt was measured as the angle between a perpendicular to the longitudinal axis of the radial shaft and a line connecting the apex of the volar and dorsal rims of the distal radial articular surface, which at the same time represents the anteroposterior distance.24 In addition, we recorded the age at time of the first radiograph and sex as explanatory variables to control for potential confounding because previous research has demonstrated that there might be a correlation between age and sex and ulnar variance.3 Statistical analysis We used a 1-way analysis of variance to compare the mean age, ulnar variance, radial height, radial (ulnarward) inclination, palmar tilt, and anteroposterior distance among the different Lichtman stages of Kienböck disease, and we used the Bonferroni method for pair-wise comparisons. To compare the proportions of men and lunate types among the different stages of Kienböck disease, we used the Fisher exact test. We entered all explanatory variables into an ordinal logistic regression model to adjust for Vol. 42, November 2017
KIENBÖCK COLLAPSE RISK FACTORS
885
an effect size of 0.58 using a 1-way analysis of variance.
FIGURE 1: Measurement technique. a, Central reference point; b, ulnar variance; c, radial (ulnarward) inclination; d, radial height.
potential confounding by any of the included covariates and we calculated odds ratios (OR), 95% confidence intervals, and P values. In addition, we performed 2 multivariable logistic regression analyses to identify factors that were independently associated with lunate collapse (Lichtman stage 3A) and carpal collapse (Lichtman stage 3B), and calculated ORs, 95% confidence intervals, and P values. We inverted the ulnar variance for multivariable analysis so that negative ulnar variance is represented by positive values for ease of calculation and interpretation of the results. The ORs, therefore, represent the risk of more advanced disease per each millimeter more negative ulnar variance. A P value of less than 0.05 was considered statistically significant. Demographic data We included 195 patients with Kienböck disease with a mean age of 49 years (SD, 15; range, 18e83 years), of whom 87 (45%) were men. Power analysis A post hoc power analysis demonstrated that 195 patients provided us with 99% statistical power to detect the observed difference in ulnar variance among the Lichtman stages of Kienböck disease with J Hand Surg Am.
r
RESULTS In bivariate analysis, patients with Lichtman stage 3A (mean difference, e1.2 mm; Bonferroni adjusted P < .05) and Lichtman stage 3B (mean difference, e1.0 mm; Bonferroni adjusted P < .05) Kienböck disease had more negative ulnar variance than patients with Lichtman stage 2 disease. There was also a difference in mean age stratified by Lichtman stage (P < .05; Table 1). Using multivariable ordinal logistic regression analysis controlling for potential confounding by any of the included covariates, we found that patients with more negative ulnar variance had a higher risk of having more advanced Kienböck disease (adjusted OR, 1.4; P < .05; Table 2). In addition, we found that an increase in age was independently associated with a higher Lichtman stage of Kienböck disease (adjusted OR, 1.02; P < .05; Table 2). In multivariable logistic regression analyses, we found that more negative ulnar variance (adjusted OR, 1.5 [per mm]; P < .05) was independently associated with lunate collapse (Lichtman stage 3A), whereas both a more negative ulnar variance (adjusted OR, 1.3; P < .05; Appendix A; available on the Journal’s Web site at www.jhandsurg.org) and an increase in age (adjusted OR, 1.03; P < .05; Appendix A; available on the Journal’s Web site at www.jhandsurg.org) were independently associated with carpal malalignment (Lichtman stage 3B). DISCUSSION Not all patients with Kienböck disease progress to collapse of the lunate, but it is difficult to determine which patients are at risk. We aimed to identify demographic or anatomical factors associated with lunate collapse and subsequent carpal malalignment among 195 patients with Kienböck disease and found that older patients and patients with more negative ulnar variance are more likely to have more advanced disease. The older age is likely a reflection of the longer duration of the disease, but it is less clear whether the ulnar variance contributed to or resulted from the longstanding Kienböck disease, although we feel that the former is more likely. This study had a number of limitations. First, we performed all measurements on radiographs obtained for regular care, and no standardized protocol for obtaining the radiographs specifically for this study was used. We consider this a minor Vol. 42, November 2017
886
KIENBÖCK COLLAPSE RISK FACTORS
TABLE 1.
Bivariate Analysis (n [ 195) Lichtman Stage Stage I (n ¼ 9)
Stage II (n ¼ 94)
Stage IIIA (n ¼ 37)
Stage IIIB (n ¼ 43)
Stage IV (n ¼ 12)
38 (12)
50 (14)
45 (16)
51 (13)
54 (14)
< .05
2 (22)
41 (44)
15 (41)
23 (53)
6 (50)
.47
e1.1 (1.2)
e1.7 (1.7)
e2.9 (1.6)
e2.7 (2.0)
e2.6 (2.4)
< .05
P Value
Demographics Age, y, mean (SD) Male sex, n (%) Radiographic parameters Ulnar variance, mm (SD) Radial height, mm (SD) Radial inclination,
(SD)
Type 2 lunate, n (%) AP distance, mm (SD)
Palmar tilt,
(SD)
9.9 (1.6)
11 (2.5)
12 (4.9)
11 (2.2)
12 (2.1)
.25
22 (3.1)
23 (3.4)
22 (4.5)
23 (4.0)
24 (4.0)
.73
2 (22)
41 (44)
19 (51)
18 (42)
3 (25)
.40
15 (5.6)
17 (2.1)
17 (2.9)
17 (2.3)
17 (7.4)
.44
8.5 (6.8)
7.1 (5.6)
8.5 (5.2)
8.1 (5.4)
5.0 (6.8)
.31
AP, anteroposterior. Bold indicates statistical significance (P < .05).
TABLE 2.
Ordinal Logistic Regression Analysis (n [ 195) OR (95% CI)
SE
P Value
1.02 (1.0e1.04)
0.010
< .05
1.5 (0.80e2.8)
0.47
.20
1.4 (1.2e1.6)
0.11
< .05
Radial height
1.05 (0.96e1.1)
0.044
.27
Radial inclination
0.99 (0.92e1.1)
0.037
.73
Type 2 lunate
0.79 (0.44e1.4)
0.23
.41
Palmar tilt
1.01 (0.96e1.1)
0.025
.83
Anteroposterior distance
1.01 (0.89e1.1)
0.062
.93
Demographics Age Male sex Radiographic parameters Negative ulnar variance*
95% CI, confidence interval; SE, standard error. Bold indicates statistical significance (P < .05). *We inverted the values of ulnar variance so that negative ulnar variance is represented with positive values.
limitation, because—although ulnar variance changes slightly with pronation and supination27— we do not feel the forearm position is biased by the stage of Kienböck disease. Second, we deliberately included only the first radiograph demonstrating Kienböck disease for each patient, and we did not evaluate disease progression longitudinally. We also consider this a minor limitation, because it is counterbalanced by our relatively large crosssectional sample of patients who presented with different stages of Kienböck disease and our data show a clear trend of more negative ulnar variance among patients with more advanced stages of Kienböck disease. J Hand Surg Am.
r
Our observation that patients with Lichtman stage 3A and 3B Kienböck disease have more negative ulnar variance than patients with stage 2 agrees with the observation made by Goeminne et al,20 who also found a significant difference in ulnar variance between patients with and without a collapsed lunate. Conversely, Iwasaki et al28 report no differences in ulnar variance and radial inclination among 24 patients with stages 2, 3A, and 3B Kienböck disease, and Mirabello et al29 also found no correlation between ulnar variance and the severity of lunate collapse or the age of onset of Kienböck disease. Our large sample enabled us to perform multivariable (ordinal) logistic regression analyses to identify Vol. 42, November 2017
KIENBÖCK COLLAPSE RISK FACTORS
factors associated with lunate collapse, adjusting for potential confounding by age, sex, ulnar variance, and other anatomical factors. We found that, in addition to more negative ulnar variance, a higher age was independently associated with lunate collapse. Nakamura et al3 report a correlation between ulnar variance and age and sex; however, our regression found that these factors were independently related to more advanced changes of Kienböck disease. Finally, the absence of a medial facet on the lunate was, in contrast to the study of Rhee et al,12 not associated with more advanced disease stage in our study. Our findings should be interpreted in the light of the limitation that, in a few cases of more advanced disease, the lunate was deformed to such an extent that determination of the presence or absence of a medial hamate facet was based on the anatomical congruity between the triquetrum, capitate, and/or hamate. There are several potential interpretations of our findings. First, it is possible that lunate deformity and/or carpal malalignment affect the radiographic measurements. In an effort to minimize observer bias, measurements were performed by an independent investigator who was not familiar with the study purpose. Second, ulnar variance may actually change as a result of progressive Kienböck disease and carpal malalignment. Kristensen et al30 report pseudolengthening of the distal radius due to osteoarthritic changes of the wrist, resulting in a more negative ulnar variance. Third, there may be an actual correlation between a more negative ulnar variance and more advanced Kienböck disease through a doseresponse relationship. Our data cannot discriminate between these 3 possible explanations, but we feel the latter is most plausible. The association between negative ulnar variance and Kienböck disease is inconsistent in the literature. There are a comparable number of studies that do1,5,8e10 and do not support a correlation.2e4,6,7,18 It is difficult to use anatomical variation alone to explain etiology given that Kienböck disease occurs in ulnar positive wrists, as well as the fact that most people with negative ulnar variance never develop Kienböck disease.10,31 Assuming that—as our study suggests— more negative ulnar variance is associated with a greater magnitude of lunate collapse in Kienböck disease, it is possible that the association of ulnar variance with Kienböck disease in previous reports is an artifact of the fact that more advanced disease is more likely to come to attention and be diagnosed.1,5,8e10 If this is true, it would create a false perception that ulnar negativity causes Kienböck disease. J Hand Surg Am.
r
887
We were somewhat surprised at the independent association between both age and negative ulnar variance and lunate collapse in this cross-sectional sample of patients with Kienböck disease and consider the findings preliminary. One explanation is if there is a dose-response relationship of ulnar variance with lunate collapse, if patients generally develop Kienböck disease at an early age, this would be a reasonable finding. However, this finding is inconsistent with prior work and may be spurious. In addition, the magnitude of the ORs is relatively small, suggesting that they may not be key factors for disease progression. Additional long-term study is needed to confirm the longitudinal relationship of negative ulnar variance with progression of Kienböck disease. REFERENCES 1. Gelberman RH, Salamon PB, Jurist JM, Posch JL. Ulnar variance in Kienböck’s disease. J Bone Joint Surg Am. 1975;57(5):674e676. 2. Kristensen SS, Thomassen E, Christensen F. Ulnar variance in Kienböck’s disease. J Hand Surg Br. 1986;11(2):258e260. 3. Nakamura R, Tanaka Y, Imaeda T, Miura T. The influence of age and sex on ulnar variance. J Hand Surg Br. 1991;16(1):84e88. 4. D’Hoore K, De Smet L, Verellen K, Vral J, Fabry G. Negative ulnar variance is not a risk factor for Kienböck’s disease. J Hand Surg Am. 1994;19(2):229e231. 5. Bonzar M, Firrell JC, Hainer M, Mah ET, McCabe SJ. Kienböck disease and negative ulnar variance. J Bone Joint Surg Am. 1998;80(8):1154e1157. 6. Muramatsu K, Ihara K, Kawai S, Doi K. Ulnar variance and the role of joint levelling procedure for Kienböck’s disease. Int Orthop. 2003;27(4):240e243. 7. Thienpont E, Mulier T, Rega F, De Smet L. Radiographic analysis of anatomical risk factors for Kienböck’s disease. Acta Orthop Belg. 2004;70(5):406e409. 8. Afshar A, Aminzadeh-Gohari A, Yekta Z. The association of Kienböck’s disease and ulnar variance in the Iranian population. J Hand Surg Eur Vol. 2013;38(5):496e499. 9. Stahl S, Stahl AS, Meisner C, et al. Critical analysis of causality between negative ulnar variance and Kienböck disease. Plast Reconstr Surg. 2013;132(4):899e909. 10. van Leeuwen WF, Oflazoglu K, Menendez ME, Ring D. Negative ulnar variance and Kienböck disease. J Hand Surg Am. 2016;41(2): 214e218. 11. Lamas C, Carrera A, Proubasta I, Llusa M, Majo J, Mir X. The anatomy and vascularity of the lunate: considerations applied to Kienböck’s disease. Chir Main. 2007;26(1):13e20. 12. Rhee PC, Jones DB, Moran SL, Shin AY. The effect of lunate morphology in Kienböck disease. J Hand Surg Am. 2015;40(4): 738e744. 13. Tsuge S, Nakamura R. Anatomical risk factors for Kienböck’s disease. J Hand Surg Br. 1993;18(1):70e75. 14. Jafari D, Shariatzadeh H, Mazhar FN, Ghahremani MH, Jalili A. Radial inclination and palmar tilt as risk factors for Kienböck’s disease. Am J Orthop (Belle Mead NJ). 2012;41(11):E145eE146. 15. Allan CH, Joshi A, Lichtman DM. Kienböck’s disease: diagnosis and treatment. J Am Acad Orthop Surg. 2001;9(2):128e136. 16. Lichtman DM, Lesley NE, Simmons SP. The classification and treatment of Kienböck’s disease: the state of the art and a look at the future. J Hand Surg Eur Vol. 2010;35(7):549e554. 17. Taniguchi Y, Nakao S, Tamaki T. Incidentally diagnosed Kienböck’s disease. Clin Orthop Relat Res. 2002;395:121e127.
Vol. 42, November 2017
888
KIENBÖCK COLLAPSE RISK FACTORS
18. Tsujimoto R, Maeda J, Abe Y, et al. Epidemiology of Kienböck’s disease in middle-aged and elderly Japanese women. Orthopedics. 2015;38(1):e14ee18. 19. van Leeuwen WF, Janssen SJ, Ter Meulen DP, Ring D. What is the radiographic prevalence of incidental Kienböck disease? Clin Orthop Relat Res. 2016;474(3):808e813. 20. Goeminne S, Degreef I, De Smet L. Negative ulnar variance has prognostic value in progression of Kienböck’s disease. Acta Orthop Belg. 2010;76(1):38e41. 21. Ledoux P, Lamblin D, Wuilbaut A, Schuind F. A finite-element analysis of Kienböck’s disease. J Hand Surg Eur Vol. 2008;33(3): 286e291. 22. Goldfarb CA, Hsu J, Gelberman RH, Boyer MI. The Lichtman classification for Kienböck’s disease: an assessment of reliability. J Hand Surg Am. 2003;28(1):74e80. 23. Larsen CF, Mathiesen FK, Lindequist S. Measurements of carpal bone angles on lateral wrist radiographs. J Hand Surg Am. 1991;16(5):888e893. 24. Medoff RJ. Essential radiographic evaluation for distal radius fractures. Hand Clin. 2005;21(3):279e288.
J Hand Surg Am.
r
25. Steyers CM, Blair WF. Measuring ulnar variance: a comparison of techniques. J Hand Surg Am. 1989;14(4):607e612. 26. Viegas SF, Wagner K, Patterson R, Peterson P. Medial (hamate) facet of the lunate. J Hand Surg Am. 1990;15(4):564e571. 27. Yeh GL, Beredjiklian PK, Katz MA, Steinberg DR, Bozentka DJ. Effects of forearm rotation on the clinical evaluation of ulnar variance. J Hand Surg Am. 2001;26(6):1042e1046. 28. Iwasaki N, Genda E, Minami A, Kaneda K, Chao EY. Force transmission through the wrist joint in Kienböck’s disease: a two-dimensional theoretical study. J Hand Surg Am. 1998;23(3): 415e424. 29. Mirabello SC, Rosenthal DI, Smith RJ. Correlation of clinical and radiographic findings in Kienböck’s disease. J Hand Surg Am. 1987;12(6):1049e1054. 30. Kristensen SS, Soballe K. Kienböck’s disease—the influence of arthrosis on ulnar variance measurements. J Hand Surg Br. 1987;12(3):301e305. 31. Schuurman AH, Maas M, Dijkstra PF, Kauer JM. Assessment of ulnar variance: a radiological investigation in a Dutch population. Skeletal Radiol. 2001;30(11):633e638.
Vol. 42, November 2017
888.e1
KIENBÖCK COLLAPSE RISK FACTORS
APPENDIX A.
Logistic Regression Analysis (n [ 195) Lunate Collapse (> 3A) (n ¼ 92)
Carpal Collapse (> 3B) (n ¼ 55)
Odds Ratio (95% CI)
SE
P Value
Age
1.01 (0.99e1.03)
0.011
.38
Male sex
1.08 (0.54e2.2)
0.39
.82 < .05
Odds Ratio (95% CI)
SE
P Value
1.03 (1.0e1.1)
0.013
< .05
Demographics 1.9 (0.90e4.0)
0.73
.09 < .05
Radiographic parameters Negative ulnar variance*
1.5 (1.2e1.8)
0.14
Radial height
1.1 (0.97e1.3)
0.076
.13
Radial inclination
0.97 (0.89e1.1)
0.044
Type 2 lunate
0.88 (0.46e1.7)
0.29
Palmar tilt
1.03 (0.97e1.1)
0.029
1.0 (0.89e1.1)
0.055
Anteroposterior distance
1.3 (1.1e1.6)
0.12
1.0 (0.89e1.1)
0.062
.99
.56
1.0 (0.92e1.1)
0.047
.93
.71
0.61 (0.30e1.2)
0.22
.18
.32
1.0 (0.94e1.1)
0.031
.98
.95
1.01 (0.89e1.1)
0.063
.90
95% CI, confidence interval; SE, standard error. Bold indicates statistical significance (P < .05). *We inverted the values of ulnar variance so that negative ulnar variance is represented with positive values.
J Hand Surg Am.
r
Vol. 42, November 2017
Risk Factors of Lunate Collapse in Kienböck Disease Wouter F. van Leeuwen, MD,*† Matthew A. Tarabochia, MD,* Arnold H. Schuurman, MD, PhD,† Neal Chen, MD,* David Ring, MD, PhD‡
Purpose Not all patients with Kienböck disease progress to collapse of the lunate and carpal malalignment, but it is difficult to determine which patients are at risk. We aimed to identify demographic or anatomical factors associated with more advanced stages of Kienböck disease. Methods We included all 195 eligible patients with Kienböck disease and available preoperative posteroanterior and lateral radiographs. We compared the mean age, sex distribution, mean ulnar variance, radial height, radial (ulnarward) inclination, palmar tilt, anteroposterior distance, and lunate type among the different Lichtman stages of Kienböck disease and performed ordinal logistic regression analysis. Results We found that patients with more negative ulnar variance had more advanced stages of Kienböck disease (adjusted odds ratio, 1.4). An increase in age was also independently associated with a higher Lichtman stage of Kienböck disease (adjusted odds ratio, 1.02). Conclusions Our findings suggest that more negative ulnar variance may be related to a greater magnitude of lunate collapse in Kienböck disease. Additional long-term study is needed to confirm the longitudinal relationship of negative ulnar variance with progressive Kienböck disease. (J Hand Surg Am. 2017;42(11):883e888. Copyright Ó 2017 by the American Society for Surgery of the Hand. All rights reserved.) Type of study/level of evidence Prognostic II. Key words Kienböck disease, lunatomalacia, ulnar variance, lunate collapse, Lichtman stage.
K
osteonecrosis of the lunate. The cause is unknown but some theories implicate mechanical influences on the lunate. Several anatomical risk factors are proposed and include negative ulnar variance,1e10 variations in lunate shape11,12 and radial (ulnarward) inclination,7,13,14 and a greater palmar tilt of the distal radial articular surface.14 IENBÖCK DISEASE IS IDIOPATHIC
From the *Department of Orthopaedic Surgery, Hand and Upper Extremity Service, Massachusetts General Hospital, Harvard Medical School, Boston, MA; the †Department of Plastic, Reconstructive and Hand Surgery, University Medical Center Utrecht, Utrecht, The Netherlands; and the ‡Department of Surgery and Perioperative Care, Dell Medical School, The University of Texas at Austin, Austin, TX.
Kienböck disease may lead to collapse of the lunate and eventually collapse of the proximal carpal row with proximal migration of the capitate and fixed volar flexion of the scaphoid, ultimately with progression to pancarpal arthrosis.15,16 However, Kienböck disease does not seem to progress in a notable number of patients and is occasionally diagnosed in the absence of symptoms.17e19 This work was performed at Massachusetts General Hospital, Boston, MA. Corresponding author: David Ring, MD, PhD, 1400 Barbara Jordan Blvd., Suite 1.114AC MC: R1800, Austin, TX 78723; e-mail: [email protected]. 0363-5023/17/4211-0005$36.00/0 http://dx.doi.org/10.1016/j.jhsa.2017.06.107
Received for publication July 27, 2016; accepted in revised form June 30, 2017. No benefits in any form have been received or will be received related directly or indirectly to the subject of this article.
Ó 2017 ASSH
r
Published by Elsevier, Inc. All rights reserved.
r
883
884
KIENBÖCK COLLAPSE RISK FACTORS
Unfortunately, we are currently unable to predict which patients are likely to progress to lunate collapse or osteoarthritis. Previous reports suggest that lunate collapse may be associated with differences in lunate shape12 and ulnar variance.20,21 The purpose of this study is to identify anatomical or demographic risk factors for lunate collapse and carpal malalignment in a cross-sectional cohort of patients with Kienböck disease. We hypothesize that there are no demographic or anatomical factors independently associated with more advanced disease stage according to the Lichtman radiographic classification of Kienböck disease.15,16 PATIENTS AND METHODS The review board at Massachusetts General Hospital approved the protocol for this retrospective study and granted a waiver of informed consent. We screened the radiology databases of 2 tertiary academic medical centers and 2 affiliated institutions for “Kienböck,” “lunatomalacia,” “lunate,” and/or “avascular necrosis,” as well as common misspellings and synonyms. We then manually reviewed the radiographic images and electronic medical records and included all 195 eligible adult patients with a confirmed diagnosis of Kienböck disease and in whom preoperative posteroanterior and lateral wrist radiographs were available. The radiographs were obtained between 1999 and 2016. We included the first available set of radiographs demonstrating Kienböck disease. In patients who had their initial diagnosis of Kienböck disease made based on magnetic resonance imaging (MRI) scans, we included the preoperative radiograph that was obtained closest to the date of the MRI. In 7 patients, the MRI was obtained a median of 23 days (interquartile range, 10e57 days) after the initial radiographs, and in 1 patient, radiographs were obtained 69 days after an MRI showed changes consistent with Kienböck disease. Outcome measures and study variables Our primary outcome measure was Lichtman stage of Kienböck disease. One investigator (W.F.v.L.) reviewed and staged all radiographs.15 Stage IIIB was defined as lunate collapse in the presence of a radioscaphoid angle of more than 60 as measured on lateral wrist radiographs.22 The radioscaphoid angle was measured between the longitudinal axis of the radius and a line tangential to the proximal and distal volar poles of the scaphoid.23 Another investigator (M.A.T.) performed all radiographic measurements J Hand Surg Am.
r
and was blinded to the study purpose to minimize observer bias. The ulnar variance, radial height, and radial (ulnarward) inclination were measured on posteroanterior wrist radiographs. We determined the central reference point—the midpoint between the volar and dorsal corners of the ulnar margins of the distal radial articular surface—as described by Medoff.24 We used the method of perpendiculars to measure the ulnar variance, radial height, and radial (ulnarward) inclination: a line was drawn through the longitudinal axis of the radius and a perpendicular line was constructed through the central reference point.25 Ulnar variance was measured as the distance between the perpendicular and a parallel line tangential to the distal surface of the ulnar head. The radial height was defined as the difference in axial length between the line through the central reference point and the tip of the radial styloid measured along the longitudinal axis of the radius. Radial (ulnarward) inclination was determined as the angle between the perpendicular and a line through the central reference point tangential to the tip of the radial styloid (Fig. 1).24 In addition, the absence (type 1 lunate) or presence (type 2 lunate) of a medial facet on the lunate in articulation with the hamate was determined as proposed by Viegas et al.26 Palmar tilt and anteroposterior distance were measured on lateral radiographs. Palmar tilt was measured as the angle between a perpendicular to the longitudinal axis of the radial shaft and a line connecting the apex of the volar and dorsal rims of the distal radial articular surface, which at the same time represents the anteroposterior distance.24 In addition, we recorded the age at time of the first radiograph and sex as explanatory variables to control for potential confounding because previous research has demonstrated that there might be a correlation between age and sex and ulnar variance.3 Statistical analysis We used a 1-way analysis of variance to compare the mean age, ulnar variance, radial height, radial (ulnarward) inclination, palmar tilt, and anteroposterior distance among the different Lichtman stages of Kienböck disease, and we used the Bonferroni method for pair-wise comparisons. To compare the proportions of men and lunate types among the different stages of Kienböck disease, we used the Fisher exact test. We entered all explanatory variables into an ordinal logistic regression model to adjust for Vol. 42, November 2017
KIENBÖCK COLLAPSE RISK FACTORS
885
an effect size of 0.58 using a 1-way analysis of variance.
FIGURE 1: Measurement technique. a, Central reference point; b, ulnar variance; c, radial (ulnarward) inclination; d, radial height.
potential confounding by any of the included covariates and we calculated odds ratios (OR), 95% confidence intervals, and P values. In addition, we performed 2 multivariable logistic regression analyses to identify factors that were independently associated with lunate collapse (Lichtman stage 3A) and carpal collapse (Lichtman stage 3B), and calculated ORs, 95% confidence intervals, and P values. We inverted the ulnar variance for multivariable analysis so that negative ulnar variance is represented by positive values for ease of calculation and interpretation of the results. The ORs, therefore, represent the risk of more advanced disease per each millimeter more negative ulnar variance. A P value of less than 0.05 was considered statistically significant. Demographic data We included 195 patients with Kienböck disease with a mean age of 49 years (SD, 15; range, 18e83 years), of whom 87 (45%) were men. Power analysis A post hoc power analysis demonstrated that 195 patients provided us with 99% statistical power to detect the observed difference in ulnar variance among the Lichtman stages of Kienböck disease with J Hand Surg Am.
r
RESULTS In bivariate analysis, patients with Lichtman stage 3A (mean difference, e1.2 mm; Bonferroni adjusted P < .05) and Lichtman stage 3B (mean difference, e1.0 mm; Bonferroni adjusted P < .05) Kienböck disease had more negative ulnar variance than patients with Lichtman stage 2 disease. There was also a difference in mean age stratified by Lichtman stage (P < .05; Table 1). Using multivariable ordinal logistic regression analysis controlling for potential confounding by any of the included covariates, we found that patients with more negative ulnar variance had a higher risk of having more advanced Kienböck disease (adjusted OR, 1.4; P < .05; Table 2). In addition, we found that an increase in age was independently associated with a higher Lichtman stage of Kienböck disease (adjusted OR, 1.02; P < .05; Table 2). In multivariable logistic regression analyses, we found that more negative ulnar variance (adjusted OR, 1.5 [per mm]; P < .05) was independently associated with lunate collapse (Lichtman stage 3A), whereas both a more negative ulnar variance (adjusted OR, 1.3; P < .05; Appendix A; available on the Journal’s Web site at www.jhandsurg.org) and an increase in age (adjusted OR, 1.03; P < .05; Appendix A; available on the Journal’s Web site at www.jhandsurg.org) were independently associated with carpal malalignment (Lichtman stage 3B). DISCUSSION Not all patients with Kienböck disease progress to collapse of the lunate, but it is difficult to determine which patients are at risk. We aimed to identify demographic or anatomical factors associated with lunate collapse and subsequent carpal malalignment among 195 patients with Kienböck disease and found that older patients and patients with more negative ulnar variance are more likely to have more advanced disease. The older age is likely a reflection of the longer duration of the disease, but it is less clear whether the ulnar variance contributed to or resulted from the longstanding Kienböck disease, although we feel that the former is more likely. This study had a number of limitations. First, we performed all measurements on radiographs obtained for regular care, and no standardized protocol for obtaining the radiographs specifically for this study was used. We consider this a minor Vol. 42, November 2017
886
KIENBÖCK COLLAPSE RISK FACTORS
TABLE 1.
Bivariate Analysis (n [ 195) Lichtman Stage Stage I (n ¼ 9)
Stage II (n ¼ 94)
Stage IIIA (n ¼ 37)
Stage IIIB (n ¼ 43)
Stage IV (n ¼ 12)
38 (12)
50 (14)
45 (16)
51 (13)
54 (14)
< .05
2 (22)
41 (44)
15 (41)
23 (53)
6 (50)
.47
e1.1 (1.2)
e1.7 (1.7)
e2.9 (1.6)
e2.7 (2.0)
e2.6 (2.4)
< .05
P Value
Demographics Age, y, mean (SD) Male sex, n (%) Radiographic parameters Ulnar variance, mm (SD) Radial height, mm (SD) Radial inclination,
(SD)
Type 2 lunate, n (%) AP distance, mm (SD)
Palmar tilt,
(SD)
9.9 (1.6)
11 (2.5)
12 (4.9)
11 (2.2)
12 (2.1)
.25
22 (3.1)
23 (3.4)
22 (4.5)
23 (4.0)
24 (4.0)
.73
2 (22)
41 (44)
19 (51)
18 (42)
3 (25)
.40
15 (5.6)
17 (2.1)
17 (2.9)
17 (2.3)
17 (7.4)
.44
8.5 (6.8)
7.1 (5.6)
8.5 (5.2)
8.1 (5.4)
5.0 (6.8)
.31
AP, anteroposterior. Bold indicates statistical significance (P < .05).
TABLE 2.
Ordinal Logistic Regression Analysis (n [ 195) OR (95% CI)
SE
P Value
1.02 (1.0e1.04)
0.010
< .05
1.5 (0.80e2.8)
0.47
.20
1.4 (1.2e1.6)
0.11
< .05
Radial height
1.05 (0.96e1.1)
0.044
.27
Radial inclination
0.99 (0.92e1.1)
0.037
.73
Type 2 lunate
0.79 (0.44e1.4)
0.23
.41
Palmar tilt
1.01 (0.96e1.1)
0.025
.83
Anteroposterior distance
1.01 (0.89e1.1)
0.062
.93
Demographics Age Male sex Radiographic parameters Negative ulnar variance*
95% CI, confidence interval; SE, standard error. Bold indicates statistical significance (P < .05). *We inverted the values of ulnar variance so that negative ulnar variance is represented with positive values.
limitation, because—although ulnar variance changes slightly with pronation and supination27— we do not feel the forearm position is biased by the stage of Kienböck disease. Second, we deliberately included only the first radiograph demonstrating Kienböck disease for each patient, and we did not evaluate disease progression longitudinally. We also consider this a minor limitation, because it is counterbalanced by our relatively large crosssectional sample of patients who presented with different stages of Kienböck disease and our data show a clear trend of more negative ulnar variance among patients with more advanced stages of Kienböck disease. J Hand Surg Am.
r
Our observation that patients with Lichtman stage 3A and 3B Kienböck disease have more negative ulnar variance than patients with stage 2 agrees with the observation made by Goeminne et al,20 who also found a significant difference in ulnar variance between patients with and without a collapsed lunate. Conversely, Iwasaki et al28 report no differences in ulnar variance and radial inclination among 24 patients with stages 2, 3A, and 3B Kienböck disease, and Mirabello et al29 also found no correlation between ulnar variance and the severity of lunate collapse or the age of onset of Kienböck disease. Our large sample enabled us to perform multivariable (ordinal) logistic regression analyses to identify Vol. 42, November 2017
KIENBÖCK COLLAPSE RISK FACTORS
factors associated with lunate collapse, adjusting for potential confounding by age, sex, ulnar variance, and other anatomical factors. We found that, in addition to more negative ulnar variance, a higher age was independently associated with lunate collapse. Nakamura et al3 report a correlation between ulnar variance and age and sex; however, our regression found that these factors were independently related to more advanced changes of Kienböck disease. Finally, the absence of a medial facet on the lunate was, in contrast to the study of Rhee et al,12 not associated with more advanced disease stage in our study. Our findings should be interpreted in the light of the limitation that, in a few cases of more advanced disease, the lunate was deformed to such an extent that determination of the presence or absence of a medial hamate facet was based on the anatomical congruity between the triquetrum, capitate, and/or hamate. There are several potential interpretations of our findings. First, it is possible that lunate deformity and/or carpal malalignment affect the radiographic measurements. In an effort to minimize observer bias, measurements were performed by an independent investigator who was not familiar with the study purpose. Second, ulnar variance may actually change as a result of progressive Kienböck disease and carpal malalignment. Kristensen et al30 report pseudolengthening of the distal radius due to osteoarthritic changes of the wrist, resulting in a more negative ulnar variance. Third, there may be an actual correlation between a more negative ulnar variance and more advanced Kienböck disease through a doseresponse relationship. Our data cannot discriminate between these 3 possible explanations, but we feel the latter is most plausible. The association between negative ulnar variance and Kienböck disease is inconsistent in the literature. There are a comparable number of studies that do1,5,8e10 and do not support a correlation.2e4,6,7,18 It is difficult to use anatomical variation alone to explain etiology given that Kienböck disease occurs in ulnar positive wrists, as well as the fact that most people with negative ulnar variance never develop Kienböck disease.10,31 Assuming that—as our study suggests— more negative ulnar variance is associated with a greater magnitude of lunate collapse in Kienböck disease, it is possible that the association of ulnar variance with Kienböck disease in previous reports is an artifact of the fact that more advanced disease is more likely to come to attention and be diagnosed.1,5,8e10 If this is true, it would create a false perception that ulnar negativity causes Kienböck disease. J Hand Surg Am.
r
887
We were somewhat surprised at the independent association between both age and negative ulnar variance and lunate collapse in this cross-sectional sample of patients with Kienböck disease and consider the findings preliminary. One explanation is if there is a dose-response relationship of ulnar variance with lunate collapse, if patients generally develop Kienböck disease at an early age, this would be a reasonable finding. However, this finding is inconsistent with prior work and may be spurious. In addition, the magnitude of the ORs is relatively small, suggesting that they may not be key factors for disease progression. Additional long-term study is needed to confirm the longitudinal relationship of negative ulnar variance with progression of Kienböck disease. REFERENCES 1. Gelberman RH, Salamon PB, Jurist JM, Posch JL. Ulnar variance in Kienböck’s disease. J Bone Joint Surg Am. 1975;57(5):674e676. 2. Kristensen SS, Thomassen E, Christensen F. Ulnar variance in Kienböck’s disease. J Hand Surg Br. 1986;11(2):258e260. 3. Nakamura R, Tanaka Y, Imaeda T, Miura T. The influence of age and sex on ulnar variance. J Hand Surg Br. 1991;16(1):84e88. 4. D’Hoore K, De Smet L, Verellen K, Vral J, Fabry G. Negative ulnar variance is not a risk factor for Kienböck’s disease. J Hand Surg Am. 1994;19(2):229e231. 5. Bonzar M, Firrell JC, Hainer M, Mah ET, McCabe SJ. Kienböck disease and negative ulnar variance. J Bone Joint Surg Am. 1998;80(8):1154e1157. 6. Muramatsu K, Ihara K, Kawai S, Doi K. Ulnar variance and the role of joint levelling procedure for Kienböck’s disease. Int Orthop. 2003;27(4):240e243. 7. Thienpont E, Mulier T, Rega F, De Smet L. Radiographic analysis of anatomical risk factors for Kienböck’s disease. Acta Orthop Belg. 2004;70(5):406e409. 8. Afshar A, Aminzadeh-Gohari A, Yekta Z. The association of Kienböck’s disease and ulnar variance in the Iranian population. J Hand Surg Eur Vol. 2013;38(5):496e499. 9. Stahl S, Stahl AS, Meisner C, et al. Critical analysis of causality between negative ulnar variance and Kienböck disease. Plast Reconstr Surg. 2013;132(4):899e909. 10. van Leeuwen WF, Oflazoglu K, Menendez ME, Ring D. Negative ulnar variance and Kienböck disease. J Hand Surg Am. 2016;41(2): 214e218. 11. Lamas C, Carrera A, Proubasta I, Llusa M, Majo J, Mir X. The anatomy and vascularity of the lunate: considerations applied to Kienböck’s disease. Chir Main. 2007;26(1):13e20. 12. Rhee PC, Jones DB, Moran SL, Shin AY. The effect of lunate morphology in Kienböck disease. J Hand Surg Am. 2015;40(4): 738e744. 13. Tsuge S, Nakamura R. Anatomical risk factors for Kienböck’s disease. J Hand Surg Br. 1993;18(1):70e75. 14. Jafari D, Shariatzadeh H, Mazhar FN, Ghahremani MH, Jalili A. Radial inclination and palmar tilt as risk factors for Kienböck’s disease. Am J Orthop (Belle Mead NJ). 2012;41(11):E145eE146. 15. Allan CH, Joshi A, Lichtman DM. Kienböck’s disease: diagnosis and treatment. J Am Acad Orthop Surg. 2001;9(2):128e136. 16. Lichtman DM, Lesley NE, Simmons SP. The classification and treatment of Kienböck’s disease: the state of the art and a look at the future. J Hand Surg Eur Vol. 2010;35(7):549e554. 17. Taniguchi Y, Nakao S, Tamaki T. Incidentally diagnosed Kienböck’s disease. Clin Orthop Relat Res. 2002;395:121e127.
Vol. 42, November 2017
888
KIENBÖCK COLLAPSE RISK FACTORS
18. Tsujimoto R, Maeda J, Abe Y, et al. Epidemiology of Kienböck’s disease in middle-aged and elderly Japanese women. Orthopedics. 2015;38(1):e14ee18. 19. van Leeuwen WF, Janssen SJ, Ter Meulen DP, Ring D. What is the radiographic prevalence of incidental Kienböck disease? Clin Orthop Relat Res. 2016;474(3):808e813. 20. Goeminne S, Degreef I, De Smet L. Negative ulnar variance has prognostic value in progression of Kienböck’s disease. Acta Orthop Belg. 2010;76(1):38e41. 21. Ledoux P, Lamblin D, Wuilbaut A, Schuind F. A finite-element analysis of Kienböck’s disease. J Hand Surg Eur Vol. 2008;33(3): 286e291. 22. Goldfarb CA, Hsu J, Gelberman RH, Boyer MI. The Lichtman classification for Kienböck’s disease: an assessment of reliability. J Hand Surg Am. 2003;28(1):74e80. 23. Larsen CF, Mathiesen FK, Lindequist S. Measurements of carpal bone angles on lateral wrist radiographs. J Hand Surg Am. 1991;16(5):888e893. 24. Medoff RJ. Essential radiographic evaluation for distal radius fractures. Hand Clin. 2005;21(3):279e288.
J Hand Surg Am.
r
25. Steyers CM, Blair WF. Measuring ulnar variance: a comparison of techniques. J Hand Surg Am. 1989;14(4):607e612. 26. Viegas SF, Wagner K, Patterson R, Peterson P. Medial (hamate) facet of the lunate. J Hand Surg Am. 1990;15(4):564e571. 27. Yeh GL, Beredjiklian PK, Katz MA, Steinberg DR, Bozentka DJ. Effects of forearm rotation on the clinical evaluation of ulnar variance. J Hand Surg Am. 2001;26(6):1042e1046. 28. Iwasaki N, Genda E, Minami A, Kaneda K, Chao EY. Force transmission through the wrist joint in Kienböck’s disease: a two-dimensional theoretical study. J Hand Surg Am. 1998;23(3): 415e424. 29. Mirabello SC, Rosenthal DI, Smith RJ. Correlation of clinical and radiographic findings in Kienböck’s disease. J Hand Surg Am. 1987;12(6):1049e1054. 30. Kristensen SS, Soballe K. Kienböck’s disease—the influence of arthrosis on ulnar variance measurements. J Hand Surg Br. 1987;12(3):301e305. 31. Schuurman AH, Maas M, Dijkstra PF, Kauer JM. Assessment of ulnar variance: a radiological investigation in a Dutch population. Skeletal Radiol. 2001;30(11):633e638.
Vol. 42, November 2017
888.e1
KIENBÖCK COLLAPSE RISK FACTORS
APPENDIX A.
Logistic Regression Analysis (n [ 195) Lunate Collapse (> 3A) (n ¼ 92)
Carpal Collapse (> 3B) (n ¼ 55)
Odds Ratio (95% CI)
SE
P Value
Age
1.01 (0.99e1.03)
0.011
.38
Male sex
1.08 (0.54e2.2)
0.39
.82 < .05
Odds Ratio (95% CI)
SE
P Value
1.03 (1.0e1.1)
0.013
< .05
Demographics 1.9 (0.90e4.0)
0.73
.09 < .05
Radiographic parameters Negative ulnar variance*
1.5 (1.2e1.8)
0.14
Radial height
1.1 (0.97e1.3)
0.076
.13
Radial inclination
0.97 (0.89e1.1)
0.044
Type 2 lunate
0.88 (0.46e1.7)
0.29
Palmar tilt
1.03 (0.97e1.1)
0.029
1.0 (0.89e1.1)
0.055
Anteroposterior distance
1.3 (1.1e1.6)
0.12
1.0 (0.89e1.1)
0.062
.99
.56
1.0 (0.92e1.1)
0.047
.93
.71
0.61 (0.30e1.2)
0.22
.18
.32
1.0 (0.94e1.1)
0.031
.98
.95
1.01 (0.89e1.1)
0.063
.90
95% CI, confidence interval; SE, standard error. Bold indicates statistical significance (P < .05). *We inverted the values of ulnar variance so that negative ulnar variance is represented with positive values.
J Hand Surg Am.
r
Vol. 42, November 2017