Abstract This study examines the osteological changes in the hands and fingers of rock climbers that result from intense, long-term mechanical stress placed on these bones. Specifically, it examines whether rock climbing leads to metacarpal and phalange modelling in the form of increased cortical thickness as well as joint changes associated with osteoarthritis. This study also attempts to identify specific climbing-related factors that may influence these changes, including climbing intensity and frequency of different styles of climbing. Radiographs of both hands were taken for each participant and were scored for radiographic signs of osteoarthritis using an atlas method. Total width and medullary width were measured directly on radiographs using digital calipers and used to calculate cross-sectional area and second moment of area based on a ring model. We compared 27 recreational rock climbers and 35 non-climbers for four measures of bone strength and dimensions (cross-sectional area, second moment of area, total width and medullary width) and osteoarthritis. A chi-squared test for independence was used to compare climber and non-climber osteoarthritis scores. For each measure of bone strength climbers and non-climbers were compared using a manova test. Significant manova tests were followed by principal components analysis (PCA) and individual anova tests performed on principal components with eigenvalues greater than one. A second PCA was performed on the climber subsample and the first principal component was then used as the dependent variable in linear regression variable selection procedures to determine which climbing-related variables affect bone thickness. The results suggest that climbers are not at an increased risk of developing osteoarthritis compared with non-climbers. Climbers, however, do have greater cross-sectional area as well as second moment of area. Greater total width, but not meduallary width, indicates that additional bone is deposited subperiosteally. The strength of the finger and hand bones are correlated with styles of climbing that emphasize athletic difficulty. Significant predictors include the highest levels achieved in bouldering and sport climbing. Keywords: bone modelling, cortical bone, mechanical stress, osteoarthritis

Materials and methods Data collection Twenty-seven climbers and 35 non-climbers were recruited for participation in this study following a protocol approved by the Human Subjects Review Board at the University of Tennessee. Participants were asked to complete a questionnaire that included personal bio-relevant data (mass, height and age) as well as information concerning participation in rock climbing (years of participation, types of climbing engaged in, frequency and highest level of difficulty achieved in different types of climbing). Highest levels of climbing in sport and traditional climbing were converted to an interval scale. In addition to the questionnaire, a posterior–anterior radiograph of the right hand and a lateral radiograph of the left hand were taken for each participant with a Trex Hologic X-ray machine and standard film-to-tube distance of 40 inches. All postero–anterior radiographs were scored using a single-blind approach (P.A.K.) for the radiographic changes associated with OA using an atlas method (Altman et al. 1995). These radiographic signs include marginal osteophytosis, joint space narrowing, sclerosis and subchondral cysts. An unaffected site was scored as 0, and possible, definite or severe involvement was scored as 1, 2 or 3, respectively. Three joints were scored for each ray. The carpal–metacarpal, metacarpal–phalangeal and interphalangeal joints were scored for the thumb. For the fingers, the metacarpal–phalangeal and both interphalangeal joints were scored. A randomly selected subset of ten individuals was scored a second time and a kappa test for intraobserver reliability indicates high repeatability (Kappa coefficient = 0.92, Z-score = 12.23, P < 0.0001). Total bone and medullary width were measured for 12 bones in each radiograph using digital calipers by one of us (A.M.C.). The metacarpals and proximal and medial phalanges of the second to fifth rays were measured from the radiograph of the right hand radiograph. The proximal, medial and distal phalanges of the second to fifth rays were measured on the left hand radiograph. Metacarpal measurements were taken at midshaft (Roy et al. 1994), while proximal and medial phalanges were measured at two-thirds of the shaft length from the proximal end following Bollen & Wright (1994). Distal phalanges on lateral radiographs were measured just proximal to the apical tuft. Replicate total width measurements were taken for 20 participants (A.D.S.), and a manova test of interobserver reliability did not reveal a significant difference (Wilks’ lambda, F = 0.40, P = 0.98). Four bone dimensions were calculated as measures of bone strength and to determine the location (subperiosteal/endosteal) of bone modelling in the fingers. Cross-sectional area was calculated for each bone using the ring model described by Roy et al. (1994) and provides a measure of the compressive and tensile strength of the bone. Second moment of area, which is proportional to bending strength, was also calculated using the ring model. The second moment of area was also used as a measure of torsional strength because in a circular ring the polar moment of inertia, which is proportional to torsional strength, is simply twice the second moment of area (Roy et al. 1994). Total bone width and medullary width were used to determine the location of bone (subperiosteal/endosteal) modelling. All bone measures were scaled by body mass prior to analyses. The linear measurements were scaled by reported body mass0.33 and area was scaled by reported body mass0.67. Second moment of area was scaled by the product of body mass and bone length as recommended by Ruff (2000). Analyses Because the prevalence and severity of OA were low, the OA scores were dichotomized to either 0 (no radiographic signs) or 1 (any radiographic sign). Climbers and non-climbers were compared using a chi-squared test for independence, testing the null hypothesis that climbers and non-climbers had the same levels of OA development, and as an alternative considered the hypothesis that climbers and non-climbers were statistically different. To examine which finger joints contributed to group differences, we regressed group membership against the OA scores for each joint using logistic regression. Climbers and non-climbers were compared for the four measures of bone dimension (area, second moment of area, total width and medullary width) using individual multivariate analysis of variance (manova) tests. We tested the null hypothesis that no group differences exist, and considered the alternative that significant group differences are present. We followed each significant manova test with a principal component analysis (PCA) performed on the entire sample and retained only those components with eigenvalues greater than one for further analysis. We then compared the principal component scores for climber and non-climbers using multiple anova tests. A Bonferroni adjustment was made to the significance level (α = 0.0025) for all between-group comparisons to control for type-one errors associated with multiple tests. To understand how different styles and length of participation in climbing might influence bone strength and dimension, we followed significant manova tests with an additional PCA using only the climber subset. We then used the first principal component as the dependent variable in a linear regression analysis to determine which, if any, climbing variables were significant predictors. We examined nine single-variable models: years of participation in rock climbing, hours of sport climbing per week, hours of traditional climbing per week, hours of bouldering per week, highest difficulty level sport climbing, highest difficulty level traditional climbing, highest difficulty level bouldering, hours of grip training (hand and finger exercises) per week and hours climbing on indoor gyms per week. In the case that multiple variables were significant predictors we used partial correlation analysis to determine which of the independent variables have a stronger correlation with the first principal component of the measures of bone strength. In cases where participants did not provide hours of climbing style, the hours of participation was recorded as zero. In cases where participants did not provide information on highest level of climbing achieved, we treated this as missing data and the case was removed from the regression analysis.

Discussion Results from the OA analysis seem surprising in light of the known connection between joint stress and OA, as well as between age and OA. One explanation we considered is that the climbers who participated in this study may not climb at sufficient levels to incur the joint stress necessary to cause joint damage; however, the four climbers with the highest achieved levels of sport climbing and bouldering (5.13b−5.14b and V9–V12, which are considered elite levels of difficulty) had no indications of OA (scores of 0 for all joints). Another possibility is that the climbers in this study are simply too young to detect any changes consistent with OA; however, the climbers are on average older than the non-climbers. Another possibility is that individuals with weaker hands or with OA-related difficulties may self-select themselves out of climbing. If an individual tries rock climbing and does not excel at it, or worse, experiences pain from it, they may decide to quit (or never take up) the sport. Those remaining long-term and at elite levels may be those who had healthier hands before beginning. The higher incidence of OA among the non-climbers, however, is most likely a spurious result of the sample. Size accounts for the majority of sample variation: climbers have stronger fingers than non-climbers. The eigenvectors associated with the first principal component for second moment of area, cross-sectional area and total width all have positive coefficients. Climbers have higher principal component scores than non-climbers, indicating that climbers have greater second moment of area, cross sectional area and total width for bones of the fingers and hands. The second principal component, for all three measures/variables, describes shape variation within the hand and fingers. Climbers and non-climbers are not significantly different along the second axis, indicating no major shape differences between groups. Subsequent components describe additional shape changes, although patterns are difficult to discern and no significant difference between groups exists. Climbers have greater cross-sectional area than non-climbers, indicating that additional bone has been deposited to accommodate the mechanical stress associated with rock climbing. Analyses of total width and medullary width reveal that bone is being deposited on the subperiosteal surface, but not endosteally. These results conform to mechanical expectations. Increases in torsional and bending strength are made by increasing second moment of area and polar moment of inertia, and these measures are greater in climbers. Because both measures are dependent not only on the cross-sectional area but also on how far that material is distributed from the neutral axis (the centre line of the bone cross-section), greater gains in strength are made if the same amount of material is added subperiosteally rather than endosteally. One significantly complicating factor in comparing climbers and non-climbers, however, is the systemic response of bone to physical activity (Lieberman, 1996). Climbers may be a more active group than non-climbers and the increased cortical thickness may be part of a systemic response to that activity. The significant correlation of climbing ability with bone strength, however, suggests that the bone response is specific to climbing stress and not to overall activity level. There does not appear to be a relationship between earlier initiation of climbing and thicker cortical bone. Most climbers (21 out of 27) in this study had finger and hand epiphyses that had fully fused (age > 16.5 years) when they began climbing, although the same number had probably not reached full skeletal maturity (age < 25 years). The age at which participants started climbing is not correlated with the first principal component of the second moment of area data ( ). In fact, several climbers began climbing well after total skeletal maturity but their bones are among the strongest ( ). These results suggest that it is possible for adults to add bone subperiosteally in metacarpals and phalanges. Alternatively, climbers may be, as a group, generally more active over their lifetime, including prior to engaging in rock climbing. Thus, the fact that the climbers who began later in life have thicker hand and finger bones may reflect higher activity levels prior to skeletal maturity. Open in a separate window These results contradict the general findings that pre- and post-pubescent individuals add bone mainly on the endosteal surface (Bass et al. 2002). Optimization models suggest haversian remodelling to be dominant over modelling in distal segments to prevent additional energy expenditure resulting from accelerating additional mass in distal limb segments during motion (Lieberman et al. 2003). The additional mass in the hands and fingers is so small in relation to the whole limb that it possibly has little effect on energetic costs. These findings do support equilibrium models that predict bone changes that maintain stress strain ratios below a specific threshold (Frost, 1987). Bouldering and sport climbing, in which physical/athletic difficulty is the primary emphasis, are important determinants of bone strength. Climbing difficulty is, at least partially, inversely related to the size of hand-holds and directly related to the degree of overhang. More difficult sport climbs and boulder problems are often considered more difficult because hand-holds are smaller and the rock face steeper. Steeper rock requires a greater proportion of the body mass to be supported by the hands and arms, while smaller holds decrease the area of the fingers over which body mass can be distributed. Both of these aspects produce greater stresses within the hands and fingers, and thus the relationship between higher climbing difficulty and bone strength is expected. Highest level in bouldering and sport climbing are probably similar measures of physical ability and encountered stress and we do not consider them to be independent predictors (Pearson correlation coefficient = 0.74). Because bouldering involves short pieces of rock as compared with sport climbing, which generally involves cliffs of 40–100 feet, the average climbing movements on a bouldering route are more difficult than the average climbing movement on a sport climbing route of comparable difficulty. As a result, bouldering routes will have smaller hand-holds on steeper rock than sport climbing routes of comparable difficulty. Thus, bouldering is the style of climbing where participants are likely to generate peak levels of mechanical stress and is expected to have a greater effect. The significant correlation of highest difficulty level achieved in bouldering with measures of bone strength, after controlling for highest difficulty level achieved in sport climbing, supports this conclusion. The hours of bouldering per week is a measure of the frequency of the maximal stress associated with bouldering. It may be that more frequent bouldering also leads to thicker cortical bone, but hours and highest achieved level of bouldering are mildly correlated (Pearson correlation coefficient = 0.55). The more a participant engages in bouldering, the more proficient they become and the more difficult routes they can attempt. The hours of bouldering, however, do not have a significant correlation with measures of bone strength after controlling for the highest level achieved in bouldering. The highest level of difficulty achieved in bouldering, by contrast, maintains a significant correlation with measures of bone strength after controlling for hours of bouldering. This suggests that the intensity of stress encountered is more important than the frequency of stress.

Conclusion The results of this study suggest that the mechanical stress generated during rock climbing is sufficient to stimulate the bone deposition response. The relationship between measures of bone thickness and sport climbing and bouldering, and not traditional climbing or years of climbing, indicate that bone remodels to accommodate high-intensity mechanical stress and not to frequent low-intensity stresses, even if maintained over long periods of time. The results also suggest that it is possible for adults to deposit new bone subperiosteally, even if they have already reached skeletal maturity. The results do not support the findings from other studies that climbers have a higher incidence or earlier onset of OA.

Disclaimer The research presented in this paper was not conducted under the auspices of the Federal Bureau of Investigation. The opinions expressed herein are those of the authors and do not reflect the US Department of Justice or the Federal Bureau of Investigation.

Acknowledgments We would like to thank Cathy Graves for taking the radiographs, and Dr Gitta Lubke and Dr Richard Jantz for statistical assistance. We thank Dr Christopher Ruff for suggestions during the development of this work. We also thank the two anonymous reviewers for helpful comments and criticisms.