Fractures are a major cause of morbidity, death, and health-care expense in both humans and horses.1,2 Catastrophic fractures in horses, which result in euthanasia without treatment, also create a financial loss for horse owners.2 There is strong evidence that accumulated bone tissue modeling and remodeling in response to exercise can result in reduced bone quality in racehorses and contribute to catastrophic fractures.2–5 Bone quality describes the characteristics of a bone, other than bone mass, that contribute to its ability to resist fracture.6,7 Improved prevention strategies are needed to reduce the incidence and impact of fractures. To be successful, a fracture prevention strategy must be able to identify subjects that are at increased risk of fracture. To achieve this, assessment of an animal's bone quality or bone tissue resistance to fracture is needed. The assessment should be easily applied in a clinical setting in a straightforward and minimally invasive manner. In addition, practical, easily applied methods for monitoring the fracture resistance of bone tissue and response of bone tissue to an intervention would be a valuable tool for clinical investigators.
Currently, the most widely studied methods used to identify humans at risk of fracture are based on measurements of bone density.8 Although bone density is one of the determinants of fracture risk, other factors such as bone turnover rate, microdamage accumulation, bone matrix properties, and bone geometry also contribute to bone quality.6,7 Therefore, methods for in vivo assessment of bone quality beyond traditional bone density–based modalities are needed to identify individuals at risk of fracture.
A novel, clinically applicable in vivo bone indentation technique known as RPI has been used to assess bone tissue mechanical properties in people.9–11 This technique can be used to distinguish between a group of patients with osteoporosis-related fractures and an age-comparable control group.9 The RPI instrument used in those studies9–11 consisted of a reference probe (22-gauge hypodermic needle) positioned on the surface of a bone and a test probe (coaxial within the reference probe) that indented the bone surface. The instrument could perform a series of up to 20 indentation cycles and record measurements of total InDist and InDist increase from the 1st to 20th cycles. The distances were measured with respect to the reference probe. This testing protocol required that a patient remain still, and 5 or more separate measurements were obtained through the skin over a period of 8 to 10 minutes. Total InDist and InDist increase are related to the ability of bone to resist microfracture.9,12
To be adopted for use in horses, a clinically applicable method of assessing the fracture resistance of bone would ideally be performed on standing animals because anesthesia is costly and anesthetic recovery adds risks of morbidity and death. However, performing multiple 20-cycle indentation measurements with the previously described RPI instrument without limb movement in a standing horse would be difficult. Therefore, an RPI instrument that involved measurement of a single impact indentation cycle into a bone surface was developed.13,14 This instrument is portable and has a handpiece that contains a single, solid indentation probe that penetrates the skin and is subsequently impacted into a bone. The handpiece is connected to a laptop computer by a USB cable. This HRPII potentially may have wide applicability as a clinical test of fracture resistance of bone tissue in horses, other domestic animals, and humans.
The previously described RPI instrument was designed for use through the skin of people.9,10,12 Because of the coaxial design of the probes, it was unnecessary to measure the effect of the skin on in vivo measurements for human patients. In contrast, the HRPII uses a single indentation probe, and the effect of testing through the skin is unknown. Furthermore, the relationship between in vivo and ex vivo bone indentation measurements in the same horse by use of an RPI instrument (such as the HRPII) is unknown. This information is important for the interpretation of longitudinal in vivo investigations and relating those findings to postmortem evaluations, when applicable.15
The objectives of the study reported here were to describe the use of an HRPII in horses without bone fractures; determine BMSi values for 6 regions of third metacarpal bones; characterize the influence of age, sex, and body weight on BMSi measurements; determine the influence of skin on BMSi measurements; and compare in vivo and ex vivo BMSi measurements. Results of this study will help investigators assess the RPI technique in horses and the influence of various testing conditions on BMSi values.
Materials and Methods
Animals
Specimens were obtained from 12 horses that had been donated to the university for euthanasia because of a variety of chronic, unmanageable conditions. The horses (4 geldings and 8 females) ranged in age from 20 days to 28 years (median, 12.5 years); body weight ranged from 70 to 572 kg (median, 460 kg). Breeds represented included Quarter Horse (n = 4), Thoroughbred (2), Standardbred (1), Saddlebred (1), Trakehner (1), Arab (1), Morgan (1), and Pony of America (1). This study was approved by an institutional animal care and use committee.
Experimental design
In vivo testing was performed on 6 horses immediately before they were euthanized (3 were anesthetized for testing, and 3 were sedated and restrained in a standing position for testing). Anesthetized horses were initially sedated by IV administration of xylazine hydrochloride (1.1 mg/kg), butorphanol tartrate (0.02 mg/kg), and 5% guaifenesin solution (given to effect). Anesthesia was induced by IV administration of ketamine hydrochloride (2.2 mg/kg) and diazepam (0.06 mg/kg) and maintained with isoflurane in oxygen administered via a semiclosed circle rebreathing system. For the standing horses, detomidine hydrochloride (0.01 mg/kg, IV) and butorphanol tartrate (0.01 mg/kg, IV) were administered to induce sedation, which was followed by high palmar and palmar metacarpal nerve blocks with 2% mepivicaine hydrochloride solution. Ex vivo testing was performed on specimens obtained from these 6 horses after euthanasia and on specimens obtained postmortem from an additional 6 horses to evaluate the influence of skin on test results. All horses were euthanized by IV administration of an overdose of pentobarbital sodium (39 g) and phenytoin sodium (5 g).
Paired third metacarpal bones were tested. Results of the aforementioned initial in vivo testing for 6 horses were used to compare testing conditions; each of those 12 bones was delineated into 6 regions, which provided a total of 72 paired testing sites for comparison. After initial data analysis, it was determined that an additional 6 pairs of third metacarpal bones would be necessary to elucidate the effect of skin on test results. The use of data obtained from live horses for the second group of 6 bone pairs was considered unnecessary because of the small differences detected between in vivo and ex vivo test results.
The 6 testing regions were on the dorsal aspect of each third metacarpal bone, a location where soft tissue cover is thin and the bone surface is easily accessible. The dorsal surface was divided into 6 testing regions/bone, which comprised 3 testing areas along the length of the bone (proximal [metaphyseal], middle [diaphyseal], and distal [metaphyseal]) and medial and lateral areas, respectively (Figure 1). Third metacarpal bones were tested immediately before the initial 6 horses (3 anesthetized and 3 sedated but standing) were euthanized, within 2 hours after horses were euthanized and specimens collected, or after bones were stored frozen at −20°C for 3 or 5 weeks.
Photograph of the left forelimb obtained from a representative horse. Notice the 6 testing regions (medial and lateral for proximal [P], middle [M], and distal [D], respectively) indicated on the dorsal surface of the third metacarpal bone for use in InDist measurements. The demarcation between regions P and M and regions M and D is the junction of the diaphysis and the proximal and distal metaphysis of the bone, respectively. The length of each region is approximately 5 cm.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Photograph of the left forelimb obtained from a representative horse. Notice the 6 testing regions (medial and lateral for proximal [P], middle [M], and distal [D], respectively) indicated on the dorsal surface of the third metacarpal bone for use in InDist measurements. The demarcation between regions P and M and regions M and D is the junction of the diaphysis and the proximal and distal metaphysis of the bone, respectively. The length of each region is approximately 5 cm.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Photograph of the left forelimb obtained from a representative horse. Notice the 6 testing regions (medial and lateral for proximal [P], middle [M], and distal [D], respectively) indicated on the dorsal surface of the third metacarpal bone for use in InDist measurements. The demarcation between regions P and M and regions M and D is the junction of the diaphysis and the proximal and distal metaphysis of the bone, respectively. The length of each region is approximately 5 cm.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
HRPII
The HRPIIa was a handheld instrument consisting of an indentation probe, guide housing with probe attachment hub, and handle that moved over the guide housing to activate an internal trigger. There was a laptop computer containing software for the indentation data analysis and a data transmission cable to connect the probe to a laptop computer. The instrument used an internal reference point for InDist measurement rather than a hypodermic needle reference probe that has been used in other RPI instruments. The HRPII made a measurement of InDist from this internal reference point during a single impact indentation cycle onto a bone surface. The indentation probe was a solid stainless steel rod that could be removed and sterilized. The tip of the probe was a 90° cone shape with a radius of 5 μm at the point. The probe was positioned perpendicular to a bone surface for InDist measurement. The trigger mechanism within the handpiece was activated once a predetermined force was reached as the handle was pushed over the guide housing. On trigger activation, the probe was automatically driven into the bone. Peak force was approximately 40 N for approximately 1 millisecond. The distance that the probe moved into the bone relative to its position prior to trigger release was the InDist. These measurements were recorded automatically during each test and processed by the analysis software.
RPI testing technique
Indentation testing was performed by use of the same measurement technique in each horse in accordance with guidelines used for other RPI instruments.9,12 Hair was clipped from the skin over the length of each third metacarpal bone. Skin was prepared for testing by washing for approximately 5 minutes with a disinfectant scrub of 4% chlorhexidine alternated with rinses of 70% alcohol. After washing was completed, the 6 testing regions were marked on the skin to ensure consistency of location between testing conditions.
The HRPII was used during indentation testing (Figure 2). For all measurements, operators ensured that the skin and periosteum were removed prior to testing and the probe was perpendicular to the bone. The instrument was designed to be initially placed in contact with the bone surface to establish the probe in a testing position. This was performed by pressing the probe through the skin so that the handle and trigger were not activated. The handle of the instrument was held securely with the operator's dominant hand. When necessary, the operator's other hand was used to make contact with the limb while also making contact with the guide housing at the neck of the instrument for additional stabilization and guidance without applying force on the instrument during measurement. Once the probe was positioned on the bone surface, its position was adjusted to ensure that the probe tip was not in the same location where the probe tip contacted the bone during insertion through the skin and periosteum. The operator's nondominant hand was relaxed or removed, and the handle was steadily depressed over approximately 1 second until the trigger point was reached. This steady force established a reference InDist. The measured InDist was the additional distance (beyond the reference InDist) that the probe penetrated the bone as a result of the triggered impact.
Schematic illustrations for operation of the HRPII. A—The operator's nondominant hand is used to push the instrument through the soft tissue by pressing down on the guide housing at a location just above the probe. B—Once the operator is ready to obtain the measurement, the operator's nondominant hand is removed from the probe to prevent interference with the applied forces. The dominant hand gradually compresses the instrument for 1 second, which causes the probe to anchor into the bone. C—Once the HRPII is fully compressed, the impacting mechanism triggers, which generates an impact and pushes the probe into the sample. The distance the probe moves into the sample, compared with its position just before impact, is the measured InDist.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Schematic illustrations for operation of the HRPII. A—The operator's nondominant hand is used to push the instrument through the soft tissue by pressing down on the guide housing at a location just above the probe. B—Once the operator is ready to obtain the measurement, the operator's nondominant hand is removed from the probe to prevent interference with the applied forces. The dominant hand gradually compresses the instrument for 1 second, which causes the probe to anchor into the bone. C—Once the HRPII is fully compressed, the impacting mechanism triggers, which generates an impact and pushes the probe into the sample. The distance the probe moves into the sample, compared with its position just before impact, is the measured InDist.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Schematic illustrations for operation of the HRPII. A—The operator's nondominant hand is used to push the instrument through the soft tissue by pressing down on the guide housing at a location just above the probe. B—Once the operator is ready to obtain the measurement, the operator's nondominant hand is removed from the probe to prevent interference with the applied forces. The dominant hand gradually compresses the instrument for 1 second, which causes the probe to anchor into the bone. C—Once the HRPII is fully compressed, the impacting mechanism triggers, which generates an impact and pushes the probe into the sample. The distance the probe moves into the sample, compared with its position just before impact, is the measured InDist.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
After each indentation measurement was obtained, the probe tip was elevated slightly from the bone surface by the operator's nondominant hand and shifted 2 to 3 mm to a new location within the same region for additional measurements beneath the skin and periosteum. A total of 10 reliable indentation measurements were obtained for each bone region for each testing condition for each horse, except for the first horse, in which only 5 reliable measurements were obtained. Within each testing region, measurements deemed to have been obtained by use of poor technique (as judged by the operator at the time of the measurement) were discarded, and an additional measurement was obtained. Poor technique included noticeable slippage of the probe during impact, visible deviation of the probe from perpendicular alignment during the measurement, reaching the impact trigger point in a rapid, poorly controlled manner (< 1 second), visible movement of the bone or limb as a result of probe impact, or movement of a horse during in vivo measurement. Immediately after the series of indentation measurements were obtained for each bone region, 5 measurements were obtained for a PMMA block (2.5 × 2.5 × 1.1 cm) by the same operator and same technique. Testing of the PMMA block was used as a measure of probe calibration and for estimating relative bone strength.13
Indentation testing conditions
The same HRPII and PMMA block material were used for all measurements. Testing conditions evaluated were through the skin, on a visibly exposed bone surface, in live horses, and postmortem evaluation. Testing was performed sequentially on each region of a bone until all measurements for that testing condition were obtained. Testing always was performed first through the skin by use of the described technique. Tests on visibly exposed bone were performed by making a 5-cm incision through the skin, subcutaneous tissues, and periosteum in the center of each bone region. A scalpel handle was used to gently remove the periosteum from the bone surface and establish a testing area that was visibly free of soft tissues. Hank's balanced salt solution was used to keep the exposed bone moist during testing.16
Bones of 3 horses were used for testing immediately after the horses were euthanized. Bones of the remaining 3 horses were stored frozen before testing. Those bones were harvested with the skin and soft tissues intact, wrapped in gauze, soaked with Hank's balanced salt solution, placed in a plastic bag, and frozen at −20°C for 3 (2 horses) or 5 (1 horse) weeks. Bones were allowed to thaw at room temperature (21°C) for 24 hours before they were tested. Not all testing conditions could be applied for all horses because of the skin incision and bone exposure required for testing the exposed bone surface. As a result, ex vivo testing through the skin could not be performed after bones were exposed during in vivo testing.
Data analysis
Data were maintained in a spreadsheet with information for each horse that included age, sex, body weight, limb, test region, test condition (live vs postmortem and through the skin vs exposed bone), bone InDist, PMMA InDist, and calculated BMSi. The BMSi was determined for each testing procedure by use of the mean of the bone InDist and mean of the corresponding measurements for the PMMA block by use of the following equation:
BMSi = (mean PMMA InDist/mean bone InDist) × 100
Thus, a larger value for BMSi resulted from a lower bone InDist, which reflects greater resistance to microfracture.9 The CV for all bone and PMMA InDist measurements were calculated for each successive measurement between 2 and 10 for bone and between 2 and 5 for PMMA to examine the progression of measurement error associated with an increasing number of measurements in the same region.
Because there was a lack of variability data for the HRPII in horses, we used both total InDist measurements reported for humans by use of another RPI instrument9,b as well as preliminary data for a bone from an equine cadaver to determine a sample size of 12 horses. Sample size was based on detecting a difference of > 1 SD between testing conditions with type 1 and type 2 error of 0.05 and 0.2, respectively. The study was conducted in stages, and data analysis was performed to compare testing conditions for an initial group of 6 horses. As a result of that initial analysis, it was determined that further in vivo testing was unnecessary to achieve the study aims; thus, the second group of 6 pairs of third metacarpal bones were only tested ex vivo.
To assess agreement between BMSi values obtained for the various testing conditions (live vs postmortem and through the skin vs exposed bone), only data for each condition, paired on the basis of testing location (matched for horse, limb, and bone region), were used. Difference plots were constructed with 95% limits of agreement calculated by use of the combined CV for the 2 test conditions.17,18 The mean percentage difference between the paired values and associated 95% confidence limits were also plotted to indicate any bias between the testing conditions. These plots were used to visually assess and compare results for the 2 measurement conditions of interest. Although difference plots are often used to compare a novel test with a criterion-referenced standard test, there was no criterion-referenced standard established for BMSi measured by use of RPI; therefore, difference plots were used to evaluate agreement between the testing conditions.18,19
Least squares linear regression was used to examine the relationship between BMSi and age and body weight of horses. A mixed-model ANOVA with repeated measures and compound symmetry covariance matrix was used to examine the effect of age, sex, limb, bone region, and testing condition on BMSi values. Horse was treated as a random variable in the analysis. Age was categorized as foal (< 1 year old) and adult (≥ 1 year old) for the ANOVA. The BMSi data were assessed for normality by examination of a normal probability plot and frequency histogram. Multiple comparisons were made by use of the Tukey method on the difference of least squares means. Values of P < 0.05 were considered significant. Statistical softwarec was used to perform initial sample size calculations, regression analysis, and the ANOVA. Difference plots and 95% confidence limits were constructed in a separate spreadsheet software program.d
Results
Indentation testing procedures
Use of the HRPII was straightforward, and it easily penetrated through the skin of live horses (both anesthetized and sedated standing horses). The amount of time required to obtain a single set of measurements in all 6 regions of a single third metacarpal bone was 8 to 10 minutes. This included a total of 90 indentations (60 on bone [10 in each of the 6 regions] and 30 on the PMMA block). The CV for InDist measurements obtained on the PMMA block was low, ranging from 1.6% to 2.6% for the 5 indentations performed after each bone measurement. This represented random error attributable to the HRPII, the PMMA material, and the operator. A total of 3,902 InDist measurements were obtained for equine bone. The CV for 6 indentations obtained in bone was 8.8%. This value did not change significantly with a greater number of indentations obtained in the same region. Sedation and local anesthesia allowed measurements to be obtained with minimal movement of standing horses. The impact of the indentation probe onto the bone surface did not elicit a response from standing horses, and limb movements were related to typical positional adjustments or body sway as a result of sedation. There were 140 (3.6%) unacceptable measurements as a result of poor technique or movement. The majority (127/140 [91%]) of these unacceptable measurements were in 3 of the 4 horses with the lowest body weight. One source of unacceptable measurements, which was controlled more easily in standing horses than in recumbent anesthetized horses or horses during postmortem testing, was the stability of the bone during the procedure. Weight of a standing horse provided excellent stability of the third metacarpal bone for testing. Maintaining bone stability during the procedure was more cumbersome for the operator during in vivo testing performed on recumbent anesthetized horses or ex vivo testing on limbs than during in vivo testing on standing sedated horses.
BMSi
The BMSi values differed among regions of the third metacarpal bone in adult horses in a consistent pattern (Figure 3). In adult horses, the distal region had a significantly (P = 0.01) lower BMSi than did the middle region. The distal region had a lower BMSi, but not significantly different (P = 0.096; power = 64%), compared with that of the proximal region. There was no difference (P = 0.33; power = 88%) between the BMSi for the middle region and the proximal region. The same pattern of BMSi values among the 3 regions was not detected for the foals (n = 2). For all horses, there was no significant difference detected for BMSi between the lateral and medial aspects of the limbs (P = 0.19; power = 32%) or between the left and right limbs (P = 0.82; power = 82%).
Plot of mean and 95% confidence limits for BMSi of the third metacarpal bone by bone region (distal, middle, and proximal), aspect (lateral and medial), and limb (right and left) of 10 adult horses. The BMSi is defined as (mean impact InDist into a calibration material [ie, PMMA block] divided by the mean InDist into bone) × 100. The BMSi differs slightly but consistently across regions of the third metacarpal bone. The distal region has the lowest BMSi, the middle region has the highest BMSi, and the proximal region has an intermediate BMSi. Data for mean BMSi represents data collected for each adult horse (n = 10) for the various testing conditions.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Plot of mean and 95% confidence limits for BMSi of the third metacarpal bone by bone region (distal, middle, and proximal), aspect (lateral and medial), and limb (right and left) of 10 adult horses. The BMSi is defined as (mean impact InDist into a calibration material [ie, PMMA block] divided by the mean InDist into bone) × 100. The BMSi differs slightly but consistently across regions of the third metacarpal bone. The distal region has the lowest BMSi, the middle region has the highest BMSi, and the proximal region has an intermediate BMSi. Data for mean BMSi represents data collected for each adult horse (n = 10) for the various testing conditions.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Plot of mean and 95% confidence limits for BMSi of the third metacarpal bone by bone region (distal, middle, and proximal), aspect (lateral and medial), and limb (right and left) of 10 adult horses. The BMSi is defined as (mean impact InDist into a calibration material [ie, PMMA block] divided by the mean InDist into bone) × 100. The BMSi differs slightly but consistently across regions of the third metacarpal bone. The distal region has the lowest BMSi, the middle region has the highest BMSi, and the proximal region has an intermediate BMSi. Data for mean BMSi represents data collected for each adult horse (n = 10) for the various testing conditions.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Effect of age, sex, and body weight on BMSi assessed by use of linear regression analysis
A linear relationship (r2 = 0.71; P = 0.01) was found between body weight and BMSi (Figure 4). Linear regression analysis revealed a weak relationship between age and BMSi that, when combined with body weight, did not add to the strength of the body weight relationship with BMSi. The regression equation for body weight against BMSi was BMSi = (0.03 × body weight) + 76. Measurements used for this analysis were obtained by use of fresh cadaveric bones with the surface exposed for testing. Sex did not have an effect on BMSi when included in either regression models or the ANOVA.
Plot of mean BMSi of the third metacarpal bone versus body weight of 8 horses. Error bars represent the 95% confidence limits of the mean for each horse. Results for females (circles) and males (crosses) are plotted separately; however, the regression line (BMSi = [0.03 × body weight] + 76) is based on combined data for all 8 horses. Mean BMSi for each horse was calculated from data collected on exposed bones during postmortem examination at 12 testing sites involving both third metacarpal bones.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Plot of mean BMSi of the third metacarpal bone versus body weight of 8 horses. Error bars represent the 95% confidence limits of the mean for each horse. Results for females (circles) and males (crosses) are plotted separately; however, the regression line (BMSi = [0.03 × body weight] + 76) is based on combined data for all 8 horses. Mean BMSi for each horse was calculated from data collected on exposed bones during postmortem examination at 12 testing sites involving both third metacarpal bones.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Plot of mean BMSi of the third metacarpal bone versus body weight of 8 horses. Error bars represent the 95% confidence limits of the mean for each horse. Results for females (circles) and males (crosses) are plotted separately; however, the regression line (BMSi = [0.03 × body weight] + 76) is based on combined data for all 8 horses. Mean BMSi for each horse was calculated from data collected on exposed bones during postmortem examination at 12 testing sites involving both third metacarpal bones.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Comparison of testing conditions
Live versus postmortem testing—There were 69 pairs of BMSi data available for comparison of results for live horses versus specimens within 2 hours after horses were euthanized. These data included results for only adult horses (n = 10). Examination of the difference plot of live versus postmortem testing revealed that the mean ± SD percentage bias (–1.4 ± 2.7%) between live and postmortem BMSi was small and the 95% confidence limits of the mean included 0 (Figure 5). There was a symmetric distribution of data points about the x-axis in the difference plot. These findings suggested a lack of constant bias between BMSi values obtained for live horses and for specimens during the early postmortem period.
Difference plot between live and postmortem BMSi as a percentage of the combined mean plotted against the combined mean BMSi for the 2 methods. Each circle represents a pair of results, 1 live and 1 postmortem, obtained at a single testing location (matched for horse, limb, and region). Mean percentage bias between live and postmortem values (solid black line) and 95% confidence limits of the mean (dotted-and-dashed lines) are indicated. Notice that the 95% confidence limits of the mean percentage bias include 0 (horizontal gray line). The upper and lower 95% limits of agreement (dotted lines) are constructed from the combined CV of the paired values for BMSi and predict the range for 95% of future measurements.18
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Difference plot between live and postmortem BMSi as a percentage of the combined mean plotted against the combined mean BMSi for the 2 methods. Each circle represents a pair of results, 1 live and 1 postmortem, obtained at a single testing location (matched for horse, limb, and region). Mean percentage bias between live and postmortem values (solid black line) and 95% confidence limits of the mean (dotted-and-dashed lines) are indicated. Notice that the 95% confidence limits of the mean percentage bias include 0 (horizontal gray line). The upper and lower 95% limits of agreement (dotted lines) are constructed from the combined CV of the paired values for BMSi and predict the range for 95% of future measurements.18
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Difference plot between live and postmortem BMSi as a percentage of the combined mean plotted against the combined mean BMSi for the 2 methods. Each circle represents a pair of results, 1 live and 1 postmortem, obtained at a single testing location (matched for horse, limb, and region). Mean percentage bias between live and postmortem values (solid black line) and 95% confidence limits of the mean (dotted-and-dashed lines) are indicated. Notice that the 95% confidence limits of the mean percentage bias include 0 (horizontal gray line). The upper and lower 95% limits of agreement (dotted lines) are constructed from the combined CV of the paired values for BMSi and predict the range for 95% of future measurements.18
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Exposed bone versus through skin—There were 126 pairs of data available for comparison of BMSi obtained with exposed bone versus through the skin. Results for both foals and adult horses were included. Examination of the difference plot revealed that the mean ± SD percentage bias (–2.4 ± 1.8%) between exposed bone and through the skin BMSi was small and the 95% confidence limits of the mean did not include 0 (Figure 6). There was a symmetric distribution of data points about the x-axis. Removing 18 pairs of data for the 2 foals from this analysis resulted in a mean percentage bias (–3.9 ± 1.8%) for adult horses that was farther from 0 than the value when data for foals were included in the analysis. These findings suggested a small constant bias when BMSi values obtained through the skin were compared with values obtained on exposed bone, particularly in adult horses. Values of BMSi obtained through the skin were marginally higher than values obtained on exposed bone.
Difference plot between exposed bone and through the skin BMSi as a percentage of the combined mean plotted against the combined mean of the paired testing conditions. Notice that the 95% confidence limits of the mean percentage bias do not include 0 (horizontal gray line). See Figure 5 for remainder of key.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Difference plot between exposed bone and through the skin BMSi as a percentage of the combined mean plotted against the combined mean of the paired testing conditions. Notice that the 95% confidence limits of the mean percentage bias do not include 0 (horizontal gray line). See Figure 5 for remainder of key.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Difference plot between exposed bone and through the skin BMSi as a percentage of the combined mean plotted against the combined mean of the paired testing conditions. Notice that the 95% confidence limits of the mean percentage bias do not include 0 (horizontal gray line). See Figure 5 for remainder of key.
Citation: American Journal of Veterinary Research 77, 1; 10.2460/ajvr.77.1.39
Effect of age, sex, limb, bone region, and testing conditions assessed by use of a mixed-model ANOVA
There was no effect on BMSi for sex, limb, aspect of bone (medial or lateral), or testing on live or postmortem specimens. When age was included in the statistical model as a categorical variable, it had a significant effect on BMSi. Foal BMSi values were significantly (P < 0.001) lower than BMSi values for adult horses, with a least squares means ± SE estimate of 69.8 ± 3.1 for foals and 93.7 ± 1.4 for adults. There was also a significant interaction between age and testing condition (exposed bone or through the skin). For foals, no significant (P = 0.2) difference was detected between BMSi obtained on exposed bone and through the skin, whereas for adults, the least squares mean BMSi was significantly (P = 0.01) higher when obtained through the skin (95.4 ± 1.4) than when obtained on exposed bone (92.0 ± 1.4). There was also a significant interaction between age and bone region. There was a significant (P = 0.01) difference in BMSi between the distal and middle regions of adult horses, whereas no differences in BMSi values were detected among bone regions of foals. When data for foals were excluded from the analysis, similar findings were detected, with a significant effect of bone region and exposure type on BMSi of adult horses.
Discussion
For the study reported here, the HRPII was a clinically applicable RPI instrument that involved the use of bone microindentation testing to measure resistance of bone tissues to microfracture. Bone strength can be broadly defined as the ability to resist fracture or, alternatively, as the integration of bone mass and bone quality.7,9,20–22 Within this broad definition, a distinction is made between strength of bone as a material, which is an intrinsic property of the tissue, and strength of bone as a structure, which is determined by the architecture and geometry of the bone in addition to its intrinsic material properties.20 The intrinsic material properties and ability of bone to resist fracture are strongly influenced by its hierarchic composition.23 It is these intrinsic material properties that the HRPII was designed to quantify.
The HRPII was easy to use in both standing and recumbent horses and provided investigators with a method to assess BMSi in vivo in a variety of clinical applications. Measurements were successfully obtained from multiple horses under several testing conditions. Keeping the probe in a stable position perpendicular to the bone surface during testing was critical to obtaining acceptable measurements. It has been reported13 that keeping the angle of the probe within 10° of a perpendicular orientation to the bone surface is optimal for obtaining consistent measurements. Although we did not measure the angle of the probe relative to the bone surface during the present study, efforts were made while obtaining measurements to adhere to this guideline. In addition, applying a controlled force to the handle over at least 1 second results in more consistent BMSi measurements.13 During the process of obtaining InDist measurements, the operator subjectively determined whether the technique for each indentation measurement was acceptable. Because of the ease of obtaining additional measurements, there was a tendency to obtain additional measurements if there was any question about the technique. The majority (127/140 [91%]) of the repeated measurements were in horses with less body weight. We believe this was attributable to the fact that the smaller radius of curvature of the dorsal third metacarpal bone region made it more challenging to determine perpendicular probe alignment with the bone surface. The low CV for InDist measurements obtained for the PMMA block reflected the precision of the instrument (ie, small measurement error) for standardized conditions. The higher CV reported for bone InDist measurements was a reflection of the variability in bone tissue as a result of its anisotropic properties and hierarchic structure.24,25 The CVs reported for other measurement techniques of the mechanical properties of bone tissue are similar to the CV reported here.26
The finding in the present study that there was no overall bias between measurements obtained for live horses and for specimens during the early postmortem period is important, although not surprising. Mechanical testing of bone has primarily been performed on cadaver specimens. Postmortem testing of bones of horses with a catastrophic fracture can yield insights into predisposing conditions or risk factors associated with the fracture.15 However, it is important for investigators to modify risk factors and monitor the response to an intervention. The HRPII may allow longitudinal studies to be performed with repeated measurements of BMSi in specific horses, thereby enhancing the understanding of factors that may contribute to fracture risk. The finding that BMSi values were similar between live horses and postmortem specimens will allow direct interpretation of the results of such longitudinal studies and how those findings may relate to measurements obtained from horses with a catastrophic fracture. This finding also has implications for other types of testing of bone (eg, drilling and cutting), whereby studies on cadaver specimens are often performed to determine the bone response characteristic of surgery in live horses.
Interestingly, the finding that there was a small constant bias for BMSi values between exposed bone and through the skin is in contrast to findings with another RPI instrument, whereby soft tissue over a tibia specimen from a human cadaver did not significantly affect measurement results.12 There are 2 notable differences between the results reported here and those of that previous study.12 First is the thickness and type of skin through which testing was performed. Mean ± SD skin thickness over the dorsal surface of the equine third metacarpal bone (2.75 ± 0.37 mm) is more than 2 times as great as the skin thickness of the distal aspect of the pelvic limb in humans (1.11 ± 0.05 mm).27,e Equine skin also contains a higher density of hair than does human skin.28 Second is the difference in the specific design of the instruments tested. The previously tested RPI instrument used in humans has a reference probe to penetrate the skin, whereas the test probe is located within the reference probe and does not directly contact the skin during testing. The HRPII of the present study had a single solid probe that penetrated the skin and was positioned on the bone surface prior to impact. Although the solid probe was sharp and penetrated equine skin easily, an increase in resistance as a result of skin contact on the probe during impact onto a bone surface may result in a marginally lower InDist measurement. This would, in effect, result in a marginally higher BMSi. Although it is important to know that this small constant bias exists between testing through the skin and on exposed bone, this small effect does not prevent the use of the HRPII to assess bone and its resistance to microfracture in standing horses.
Results for the ANOVA were in close agreement with the finding of a small constant bias resulting in marginally higher BMSi values in measurements obtained through the skin. The interaction between age and testing condition (exposed bone or through the skin) indicated that there was a difference detected in adult horses but not in foals. The slightly larger bias found when foal data were excluded from the calculation of the mean percentage bias for the difference plot also supported this finding. Thinner skin of foals may have less impact on the BMSi than does the thicker skin of adult horses. Furthermore, only 2 foals were included in this analysis. Although a small constant bias of 2% to 3% was detected, adjustments could easily be made to compare testing conditions of exposed bone and through the skin, if necessary. In addition, comparing BMSi within the same testing condition would not require adjustment. These results should be interpreted in light of their potential clinical relevance because the level of bias was small relative to the measured BMSi and the potential differences that may be evident for various disease conditions. Further studies should be performed to elucidate the magnitude of any potential effect of skin thickness on HRPII measurements in other species, particularly if measurements are to be made with and without skin overlying the bone.
The linear relationship between body weight and mean BMSi of the third metacarpal bone was not unexpected considering the known relationships between bone loading and other bone material properties.29 Other investigators found that adjustment for body weight and age improves bone mineral density measurements on the third metacarpal bone of horses obtained by use of dual-energy x-ray absorptiometry.30 However, in the present study, inclusion of age did not improve the regression model over use of body weight alone. The weak relationship between age and BMSi may reflect the dynamic nature of the measurements being performed with the HRPII. Investigators of another study31 performed on equine femurs found that age did not affect the fracture resistance of cortical bone when testing was performed under dynamic conditions. Similar to the horses of that study,31 the horses for the present study ranged from 20 days to 28 years of age. This wide age range is advantageous when assessing a linear relationship; however, it should be mentioned that further refinement of the relationship should be possible with data from a larger number of horses.
In adult horses, BMSi had a consistent pattern from the proximal to distal regions of the third metacarpal bone. Specifically, the middle region (diaphysis) had higher BMSi values than did the distal region (distal metaphysis), and the proximal region (proximal metaphysis) had intermediate values. The difference between the distal and proximal regions was not significant (P = 0.096); however, there was insufficient power in the statistical analysis to conclude that no differences existed between the proximal and distal regions. There was also insufficient power to conclude that no difference in BMSi existed between the medial and lateral aspects of the bone. This study was primarily designed to examine the effect of various testing conditions on BMSi of horses, and inclusion of results on regional differences was of secondary interest. In a study32 of material property distributions in equine third metacarpal bones, significant regional (levels in that study) differences were reported for elastic modulus, yield and ultimate strain, yield and ultimate stress, and ultimate strain energy density between locations corresponding to the middle and distal regions of the study reported here. It was beyond the scope of the present study to evaluate relationships that may have existed between BMSi and these other reported bone properties that had regional variation; however, such regional differences will be an important area of future study.
The results reported here represented an assessment of the performance of the HRPII when used in horses under various testing conditions. However, there were several limitations that warrant discussion. The broad range of age, breed, body weight, and health status were used to obtain a robust set of data with the HRPII for a variety of testing conditions. Although ideal for that purpose, limits were placed on our ability to make inferences from the data, particularly considering there were only 2 foals and some of the effects detected were related to age. Another limitation of the study was the fact that not all horses were used for all testing conditions. It was not possible to perform both live and postmortem comparisons and exposed bone to through the skin comparisons in the same animal. Once a skin incision was made and soft tissues were removed from the bone surface in a testing region, further measurements through the skin were not possible in the same manner as if the skin were intact. Another reason that not all testing conditions were represented in all horses was the use of horses that were being euthanized for reasons unrelated to the present study. Several horses were unavailable for testing before euthanasia early during the study; hence, the initial data analysis was conducted and study staging was performed after the first 6 in vivo tests.
It is important to state that the HRPII was a relatively new instrument and there currently was no criterion-referenced standard clinical test available for comparison of BMSi results. Despite this, bone indentation testing has been evaluated as a laboratory method to assess hardness and the mechanical properties of bone tissue. Indentation testing has been performed at the macroscale, microscale, and nanoscale levels.33 Testing variables such as the size, shape, and material composition of the probe; indentation velocity and hold time; and angle of application of the probe can all affect results. Bone variables such as viscoelastic properties, hydration status, and orientation of the bone sample can also affect results obtained during laboratory indentation testing.24,33–35 Consistent laboratory testing protocols with controlled conditions on cadaver bone are necessary to yield reliable results, and direct application of these techniques in vivo is not possible. Hardness testing of bone requires the use of microindenter probes designed initially for hardness testing of metals and other materials.34 These probes are positioned on the bone surface by use of a predetermined loading force and are maintained there for a dwell time of 10 to 15 seconds. The size of the indent in the bone surface is measured optically to ultimately determine the hardness value obtained through knowledge of the probe geometry and calculated depth of the resulting indentation. Although similar to hardness testing, it is important to point out that testing performed with the HRPII is not hardness testing that previously has been applied to bone. Hardness testing is not performed under conditions of material fracture. In contrast, the HRPII created a microfracture on the bone surface, and direct measurement of the depth of travel of the probe (InDist) was performed. This was used as a measure of the local bone tissue resistance to microfracture. The HRPII probe impacted the bone during a period of < 1 millisecond; thus, the measurement closely reflected the dynamic process of bone fracture. Finally, because of the inherent variability for in vivo testing and the specific design of the HRPII, the result of testing (BMSi) was determined as the ratio of the InDist measured for the PMMA calibration material to the InDist measured for bone, which was expressed as a percentage.
The basic premise that the HRPII can be used to measure the resistance of bone to microfracture has been established.9,12 This may allow insights into factors that are considered determinants of fracture risk, such as properties of the bone matrix, microdamage accumulation, and bone turnover rate in a single, clinically feasible test. Future studies in horses should address the issue of the safety of indentation testing in clinical settings as well as investigate the relationship between BMSi, fracture risk, and other bone property measurements. The present study did not address the issue of clinical safety because horses tested in vivo were euthanized immediately following testing. The size of the indentations on the bone surface measured less than 0.5 mm in diameter and less than 0.3 mm in depth.14 It seems unlikely that these surface indentations in bone, being comparable in size to other bone porosities such as blood vessels and resorption spaces, would result in a substantial increase in fracture risk. We are unaware of any major adverse effects of microindentation testing in humans in multiple clinical trials.9–11 However, further examination of this aspect of clinical testing is warranted in horses given the elevated fracture risk and potential for catastrophic injury in racehorse populations.2,3 Clinical studies9,10 in humans conducted with bone microindentation testing have revealed higher InDist (lower BMSi) values in patients with osteoporotic and atypical femoral fractures, compared with results for control patients. In a recent study,11 the HRPII was used to perform bone microindentation testing in a group of women with type 2 diabetes and an age-matched control group. Investigators of that study11 found significantly lower BMSi values for the diabetic group and suggested that the testing may provide insights into reasons why patients with chronic type 2 diabetes are at increased risk for fragility fractures despite physiologically normal bone mineral density. Results of bone microindentation testing with the HRPII for the present study should provide a basis for additional studies to establish this technique for clinical applications in horses as well as other domestic animals and humans.
Acknowledgments
Supported in part by National Science Foundation grants CMMI-1031056 (Lescun and Chandresekar) and CMMI-1031244 (Hansma and Yang). Development and construction of the handheld reference point indentation instrument used for this work was supported by the National Institutes of Health (grant No. RO1 GM 065354).
Dr. Hansma is an employee and member of the board of directors of Active Life Scientific Inc, which produces reference point indentation instruments.
All data collection for this study was performed at Purdue University.
ABBREVIATIONS
| BMSi | Bone material strength index |
| CV | Coefficient of variation |
| HRPII | Handheld reference point indentation instrument |
| InDist | Indentation distance |
| PMMA | Polymethylmethacrylate |
| RPI | Reference point indentation |
Footnotes
Osteoprobe RPI instrument, Active Life Scientific Inc, Santa Barbara, Calif.
BioDent RPI instrument, Active Life Scientific Inc, Santa Barbara, Calif.
SAS, version 9.2, SAS Institute Inc, Cary, NC.
Microsoft Excel, version 14.0.7128.5000, Microsoft Corp, Redmond, Wash.
Volkering ME. Variation of skin thickness over the equine body and the correlation between skin fold measurement and actual skin thickness. Doctoral thesis, Faculty of Veterinary Medicine, University of Utrecht, Utrecht, The Netherlands, 2009.
References
1. Budhia S, Mikyas Y, Tang M, et al. Osteoporotic fractures: a systematic review of US healthcare costs and resource utilization. Pharmacoeconomics 2012; 30: 147–170.
2. Clegg PD. Musculoskeletal disease and injury, now and in the future. Part 1: fractures and fatalities. Equine Vet J 2011; 43: 643–649.
3. Riggs CM. Fractures—a preventable hazard of racing Thoroughbreds? Vet J 2002; 163: 19–29.
4. Entwistle RC, Sammons SC, Bigley RF, et al. Material properties are related to stress fracture callus and porosity of cortical bone tissue at affected and unaffected sites. J Orthop Res 2009; 27: 1272–1279.
5. Whitton RC, Trope GD, Ghasem-Zadeh A, et al. Third metacarpal condylar fatigue fractures in equine athletes occur within previously modelled subchondral bone. Bone 2010; 47: 826–831.
6. Burr DB. Bone quality: understanding what matters. J Musculoskelet Neuronal Interact 2004; 4: 184–186.
7. Cole JH, van der Meulen MCH. Whole bone mechanics and bone quality. Clin Orthop Relat Res 2011; 469: 2139–2149.
8. Stagi S, Cavalli L, Iurato C, et al. Bone health in children and adolescents: the available imaging techniques. Clin Cases Min Bone Metab 2013; 10: 166–171.
9. Diez-Perez A, Güerri R, Nogues X, et al. Microindentation for in vivo measurement of bone tissue mechanical properties in humans. J Bone Miner Res 2010; 25: 1877–1885.
10. Güerri-Fernández RC, Nogués X, Quesada Gómez JM, et al. Microindentation for in vivo measurement of bone tissue material properties in atypical femoral fracture patients and controls. J Bone Miner Res 2013; 28: 162–168.
11. Farr JN, Drake MT, Amin S, et al. In vivo assessment of bone quality in postmenopausal women with type 2 diabetes. J Bone Miner Res 2014; 29: 787–795.
12. Hansma P, Turner P, Drake B, et al. The bone diagnostic instrument II: indentation distance increase. Rev Sci Instrum 2008; 79: 064303.
13. Bridges D, Randall C, Hansma PK. A new device for performing reference point indentation without a reference probe. Rev Sci Instrum 2012; 83: 044301.
14. Randall C, Bridges D, Guerri R, et al. Applications of a new handheld reference point indentation instrument measuring bone material strength. J Med Device 2013; 7: 410051–410056.
15. Stover SM, Murray A. The California Postmortem Program: leading the way. Vet Clin North Am Equine Pract 2008; 24: 21–36.
16. Habelitz S, Marshall GW, Balooch M, et al. Nanoindentation and storage of teeth. J Biomech 2002; 35: 995–998.
17. Jensen AL, Kjelgaard-Hansen M. Method comparison in the clinical laboratory. Vet Clin Pathol 2006; 35: 276–286.
18. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995; 346: 1085–1087.
19. Dewitte K, Fierens C, Stöckl D, et al. Application of the Bland-Altman plot for interpretation of method-comparison studies: a critical investigation of its practice. Clin Chem 2002; 48: 799–801.
20. Turner CH, Burr DB. Basic biomechanical measurements of bone: a tutorial. Bone 1993; 14: 595–608.
21. Link TM. Osteoporosis imaging: state of the art and advanced imaging. Radiology 2012; 263: 3–17.
22. Seeman E, Delmas PD. Bone quality—the material and structural basis of bone strength and fragility. N Engl J Med 2006; 354: 2250–2261.
23. Akkus O, Yeni YN, Wasserman N. Fracture mechanics of cortical bone tissue: a hierarchical perspective. Crit Rev Biomed Eng 2004; 32: 379–426.
24. Riches PE, Everitt NM, Heggie AR, et al. Microhardness anisotropy of lamellar bone. J Biomech 1997; 30: 1059–1061.
25. Rho JY, Currey JD, Zioupos P, et al. The anisotropic Young's modulus of equine secondary osteons and interstitial bone determined by nanoindentation. J Exp Biol 2001; 204: 1775–1781.
26. Currey JD, Pitchford JW, Baxter PD. Variability of the mechanical properties of bone, and its evolutionary consequences. J R Soc Interface 2007; 4: 127–135.
27. Tan CY, Statham B, Marks R, et al. Skin thickness measurement by pulsed ultrasound: its reproducibility, validation and variability. Br J Dermatol 1982; 106: 657–667.
28. Tregear RT. Hair density, wind speed, and heat loss in mammals. J Appl Physiol 1965; 20: 796–801.
29. Evans GP, Behiri JC, Vaughan LC, et al. The response of equine cortical bone to loading at strain rates experienced in vivo by the galloping horse. Equine Vet J 1992; 24: 125–128.
30. McClure SR, Glickman LT, Glickman NW, et al. Evaluation of dual energy x-ray absorptiometry for in situ measurement of bone mineral density of equine metacarpi. Am J Vet Res 2001; 62: 752–756.
31. Kulin RM, Jiang F, Vecchio KS. Effects of age and loading rate on equine cortical bone failure. J Mech Behav Biomed Mater 2011; 4: 57–75.
32. Les CM, Stover SM, Keyak JH, et al. The distribution of material properties in the equine third metacarpal bone serves to enhance sagittal bending. J Biomech 1997; 30: 355–361.
33. McKoy B, Kang Q, An YH. Chapter 14: Indentation testing of bone. In: An Y, Draughn RA, eds. Mechanical testing of bone and the bone-implant interface. Boca Raton, Fla: CRC Press, 2000; 233–239.
34. Huja SS, Katona TR, Roberts WE. Chapter 16: Microhardness testing of bone. In: An Y, Draughn RA, eds. Mechanical testing of bone and the bone-implant interface. Boca Raton, Fla: CRC Press, 2000; 247–256.
35. Ziv V, Wagner HD, Weiner S. Microstructure-microhardness relations in parallel-fibered and lamellar bone. Bone 1996; 18: 417–428.
