Medical Engineering & Physics 36 (2014) 364–370
Contents lists available at ScienceDirect
Medical Engineering & Physics journal homepage: www.elsevier.com/locate/medengphy
Editor’s comment: This paper addresses an important issue faced by orthopedic surgeons when sawing through bone: how to minimize thermal damage to adjacent healthy tissue? The authors have conducted a systematic investigation into the effects of sawing parameters on the depth of osteonecrosis at the cut surface – a factor implicated in the aseptic loosening of orthopedic implants. These findings will be of interest to surgeons who use such tools in their clinical practice. Richard Black, Editor in Chief
Effect of applied force and blade speed on histopathology of bone during resection by sagittal saw Thomas P. James a,∗ , Gerard Chang b , Steven Micucci b , Amrit Sagar a , Eric L. Smith c , Charles Cassidy d a
Department of Mechanical Engineering, Tufts University, 200 College Avenue, Medford, MA 02155, USA Tufts University School of Medicine, 145 Harrison Avenue, Boston, MA 02111, USA c Tufts Medical Center, 800 Washington Street, Boston, MA 02111, USA d Tufts University School of Medicine, 800 Washington Street, Boston, MA 02111, USA b
a r t i c l e
i n f o
Article history: Received 25 June 2013 Received in revised form 13 November 2013 Accepted 1 December 2013 Keywords: Orthopedics Bone sawing Thermal necrosis Histopathology
a b s t r a c t A sagittal saw is commonly used for resection of bone during joint replacement surgery. During sawing, heat is generated that can lead to an increase in temperature at the resected surface. The aim of this study was to determine the effect of applied thrust force and blade speed on generating heat. The effect of these factors and their interactions on cutting temperature and bone health were investigated with a full factorial Design of Experiments approach for two levels of thrust force, 15 N and 30 N, and for two levels of blade oscillation rate, 12,000 and 18,000 cycles per minute (cpm). In addition, a preliminary study was conducted to eliminate blade wear as a confounding factor. A custom sawing fixture was used to crosscut samples of fresh bovine cortical bone while temperature in the bone was measured by thermocouple (n = 40), followed by measurements of the depth of thermal necrosis by histopathological analysis (n = 200). An analysis of variance was used to determine the significance of the factor effects on necrotic depth as evidenced by empty lacunae. Both thrust force and blade speed demonstrated a statistically significant effect on the depth of osteonecrosis (p < 0.05), while the interaction of thrust force with blade speed was not significant (p = 0.22). The minimum necrotic depth observed was 0.50 mm, corresponding to a higher level of force and blade speed (30 N, 18,000 cpm). Under these conditions, a maximum temperature of 93 ◦ C was measured at 0.3 mm from the kerf. With a decrease in both thrust force and blade speed (15 N, 12,000 cpm), the temperature in the bone increased to 109 ◦ C, corresponding to a nearly 50% increase in depth of the necrotic zone to 0.74 mm. A predictive equation for necrotic depth in terms of thrust force and blade speed was determined through regression analysis and validated by experiment. The histology results imply that an increase in applied thrust force is more effective in reducing the depth of thermal damage to surrounding bone than an increase in blade speed. © 2013 IPEM. Published by Elsevier Ltd. All rights reserved.
1. Introduction The use of power saws in orthopedics has existed since at least 1890 as evident from US Patent 436,804 awarded to M.J. Roberts for an “Electro Ostetome” that relied upon a unique reciprocating mechanism to rapidly translate the blade with an electric motor. In contemporary joint replacement surgery, reciprocating saws have
∗ Corresponding author. Tel.: +1 978 771 8945. E-mail address: [email protected] (T.P. James).
largely been replaced by battery powered tools that oscillate the blade at high frequency, 10,000–20,000 oscillations per minute (cpm), through a relatively small angle, typically 4–5◦ . Oscillating (sagittal) saws are used, for example, in total knee arthroplasty to prepare flat surfaces on the condylar ends of the femur and tibia in order to facilitate attachment of an artificial knee implant. Unfortunately, heat is generated during the bone sawing process, which can lead to higher temperatures that over time can cause cellular damage at the resected surface [1]. During sawing, a surgeon has control over applied thrust force and blade speed, which are two parameters that can affect temperature. Blade speed is
1350-4533/$ – see front matter © 2013 IPEM. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.medengphy.2013.12.002
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
proportional to the product of blade length and oscillation rate of the saw. The primary motivation behind the implementation of temperature abatement strategies during bone resection and preparation by sagittal saw has been the identification of a link between osteonecrosis and aseptic implant loosening. During bone cutting, heat energy derived from shear deformation of bone and friction between the blade and the bone can cause osteonecrosis at or near the resection plane [1,2]. The local effect of heat on bone is further enhanced by cortical bone’s relatively poor thermal conductivity at 0.58 ± 0.018 W/mK in the longitudinal direction [3] and low temperature threshold for irreversible osteocyte cell death at 47–50 ◦ C [4,5]. Eriksson and Albrektsson found that heating rabbit cortical bone to 50 ◦ C for 1 min led to implant loosening and significantly reduced osteointegration when compared to a control [6]. In addition, mechanical deformation from sawing can cause micro-cracks that can lead to osteocyte apoptosis [7]. As a result, the absence of osteocytes triggers targeted osteoclastic bone resorption of the damaged area [8]. Sawing experiments have been conducted to determine the effect of saw blade design and kerf (saw cut) irrigation on cutting temperature. Toksvig-Larsen et al. found that different sagittal saw blade designs did not significantly reduce temperatures below the critical limit for heat induced bone necrosis [9]. The effect of irrigation on the necrotic threshold temperature is less clear, with Toksvig-Larsen and Ryd [10] reporting no difference in temperature due to irrigation of the kerf, while Wächter and Stoll [11] demonstrated that intermittent sawing combined with irrigation was able to adequately reduce bone temperatures. However, it appears that in practice surgeons are reluctant to forcefully irrigate the saw blade as it is difficult to inject fluids into the kerf. In addition, the kerf in total knee arthroplasty is typically obstructed by the cutting guide. Considering limited success in controlling temperature with saw blade design and practical difficulties with kerf irrigation, parameters affecting operation of the sagittal saw deserve further scrutiny. Unfortunately, there is a paucity of data on the relationship between sawing parameters and the resultant bone temperature. There is, however, significant literature on operational parameters affecting temperature during bone drilling. Acknowledging that sawing and drilling of bone may differ in the mechanics of chip creation and removal, the literature on bone drilling provides important insight and guidance into the methodologies used to investigate the problem of heat generation during bone removal. A comprehensive review of the literature on the relationship between drilling parameters and thermal osteonecrosis was recently provided by Augustin et al. [12]. From the parameters identified, the most relevant to sawing are feed rate and drill speed, which are analogous to thrust force and blade speed, respectively. The investigative method traditionally used in research on bone drilling is to measure the maximum temperature rise with a thermocouple placed in the bone. A multifactorial study on drilling parameters revealed that an increase in drill speed was associated with a higher temperature rise in the bone as measured by thermocouple, whereas an increase in feed force generally resulted in a lower temperature rise [13]. In addition to temperature measurement by thermocouple, the effect of higher temperatures on health of the boney bed can be deduced by histopathological analysis. Although this method has not been applied to bone sawing research, a recent histology study on process parameters related to bone drilling revealed that high temperatures resulted in osteocyte death as evident from an increased depth of empty lacunae from the surface of the drilled hole [14]. The objective of the current study is to determine the effect of applied thrust force and blade speed on health of the boney bed by
365
using thermocouples to measure temperature in the bone near the resected surface and by studying the histopathology of the bone to determine a necrotic depth. Drawing upon the research on bone drilling, it was hypothesized that an increase in thrust force and blade speed would correspond to a decrease in the necrotic depth. To test the hypothesis, a custom bone sawing fixture was developed and a full factorial Design of Experiments (DOE) approach was employed. Prior to conducting the DOE, initial studies were pursued to determine proper sample size for statistical significance and to determine if wear was a confounding factor. 2. Materials and methods 2.1. Oscillating saw and fixture A custom apparatus was designed for sagittal sawing experiments (Fig. 1). The fixture was fitted with an AC electric oscillating tool (Fein Multimaster, Fein Power Tool, Inc., Pittsburgh, PA, USA). This tool is primarily used in wood and metal working applications, but has the same 4–5◦ sweep angle and rate of oscillation as contemporary sagittal saws. However, unlike cordless surgical saws, where blade speed can fluctuate as a condition of the battery’s state of charge, the corded oscillating saw maintains a constant blade speed while under load as confirmed in screening studies by stroboscope (Nova Strobe, Monarch Instrument, Amherst, NH, USA). This is more suitable for cutting experiments where the effect of blade speed is a parameter being investigated. The oscillating saw was attached to vertical guide rails with low friction linear bearings to facilitate orthogonal cross-cutting of the bone workpiece, Fig. 1. Applied thrust force, normal to the workpiece surface, was generated by the weight of the tool and adjusted with a counterweight pulley system. A force gage (MG20, Mark 10 Company, Copiague, NY) with an accuracy of ±0.4 N was placed beneath the center of the saw blade to confirm the range of thrust forces being investigated. Oscillation rate was controlled by the tool’s variable speed dial and prior to each cut was confirmed by stroboscope.
Fig. 1. Sawing apparatus used to cross cut bone under various conditions of blade speed and thrust force.
366
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
2.2. Saw blade Stainless steel sagittal saw blades, 12 teeth per inch (tpi), 0.80 mm thick, 64 mm depth of cut (Stryker 2108-107-009, Stryker Surgical, Kalamazoo, MI, USA) were fitted to the oscillating saw with a custom fabricated stainless steel blade adapter. The adapter added 20 mm to the saw blade depth of cut, which proportionally affects tip speed. A combination of the blade length and adapter length where chosen specifically to result in a blade with an effective 84 mm depth of cut, which falls within the 80–90 mm range commonly used in knee and hip replacement surgery. 2.3. Bone samples Two fresh bovine femurs, taken from the hind legs of one 18–24 month adult cow, were acquired from an approved abattoir (Blood Farm, Groton, MA, USA). A bandsaw was used to remove 100 mm long sections from the mid-diaphysis regions of both femurs. Three uniform rectangular planks were then prepared by bandsaw from the thicker cortical sections and finish machined with an upright mill. The sections were milled to an equal thickness and width, resulting in a cross section of 25.9 mm by 5.8 mm. The primary osteon direction was aligned to the sample length. Finally, a series of holes, 1.6 mm in diameter, 3.0 mm deep, and equally spaced at 4.0 mm center-to-center, were drilled perpendicular to the surface of the bone to provide a light press fit with the thermocouple. During workpiece preparation, the bone was kept moist through misting with a physiologic solution (Hank’s Balanced Salt Solution, Life Technologies, Grand Island, NY, USA). The feed and speed during milling was sufficient to prevent heating of the bone. This was confirmed by histology prior to conducting the sawing experiments. Densities of the three workpieces (ave. = 2.04 g/cm3 , SD = 0.018 g/cm3 ) were consistent with those of healthy human femoral cortical bone [15]. By visual inspection, the bone appeared of uniform density with no signs of porosity or other defects. Prior to and during the sawing experiments, the unused bone workpieces were soaked in physiologic solution and kept refrigerated. 2.4. Experimental design An initial study was conducted to eliminate the possibility of blade wear as a confounding factor. Cutting times were measured over 40 consecutive cross cuts with a single blade, for an applied thrust force of 30 N and a blade speed of 12,000 cpm. Volumetric cutting rates were determined from sample cross sectional area and thickness of the kerf over the cutting time. An optical microscope was used to examine teeth for chips and to measure the cutting edge radius of several teeth before and after the 40 cuts. Concurrent with the blade wear study, a histopathology study was performed to: (1) gather data for a power analysis, (2) determine the feasibility of assessing osteocyte viability on histopathological analysis, and (3) determine the amount of thermal damage caused by the bone preparation process, which included the duration of sample preparation and microtome cutting. Following these preliminary studies, a two factor, two level, full factorial Design of Experiments (DOE) was implemented. The DOE approach is statistically efficient when determining significance of one or more variables on an outcome, especially when potential interactions of variables may exist. The outcome measure for the DOE was depth of thermal necrosis in micrometers (m) as measured from the cut surface. The two factors considered were applied thrust force, normal to the workpiece surface, and oscillation rate (blade speed). Two levels were sufficient to represent the factors based on a preliminary study of cutting rate that indicated a linear response to changes in both factors. The upper (+)
and lower (−) values for thrust force were 30 N (6.74 lbs) and 15 N (3.37 lbs), respectively. Thrust force was chosen to represent a reasonable range that a surgeon would employ during sawing [16]. The upper (+) and lower (−) values for blade speed were 18,000 cpm and 12,000 cpm, which covered the operating speeds for typical sagittal saws [15]. Each of the four unique trials for thrust force and blade speed (++, −−, +−, −+) was replicated five times for a total of 20 experiments based on the power analysis from preliminary data (target power = 0.90, ˛ = 0.05, significant effect = 80 m, SD = 50 m). The level of significance is approximately 10% of the anticipated depth of thermal necrosis. While implementing the experiment, randomization was employed wherever possible to eliminate setup bias. The order of the cuts was randomized, as were the use of two different saw blades and three different bone workpieces. 2.5. Temperature measurement J-type thermocouple probes (1.6 mm diameter, grounded, 304 SS Sheath, #JQSS-116G-6, Omega Engineering Inc., Stamford, CT, USA) were attached to a 16-channel data acquisition unit (#9213, National Instruments Corp., Austin, TX, USA). Calibration of the thermocouples was performed with an ice bath and boiling water. Temperature data was acquired with a custom interface (LabView© 2009, National Instruments Corp.) at a rate of 20 Hz. 2.6. Sawing operations Upon setting the counterweight for the proper thrust force, the sagittal saw was turned on, lowered by hand to a resting position above the surface of the bone and then released, allowing the saw blade teeth to engage the bone. Cuts were made 0.30 mm from the thermocouple hole, resulting in 4.0 mm between cuts. Repeatability of the precise location of a saw cut from the edge of the thermocouple hole was accomplished by placing 0.30 mm shim stock between the blade teeth and the thermocouple wire as the bone was positioned with a translatable (screw drive) machinist’s vise. Calibrated thermocouples were used to measure temperature for a total of 60 s, allowing a 5 s lead time prior to cutting for measurement of a baseline workpiece temperature. After resection of the bone, the saw was turned off to minimize vibration while the temperature measurement continued. A stopwatch was used, alongside thermocouple output recordings, to confirm cut times for each trial. Sufficient time was afforded between runs to allow the bone sample and the blade to return to a baseline ambient temperature (22 ◦ C). Following sawing, the backside of each specimen (opposite the cut surface) was removed by milling to reduce the volume of bone requiring decalcification. After milling, one face on the bone specimen was marked with ink to preserve proper orientation during handling and the sample was placed in 10% formaldehyde to preserve it for histopathological analysis. 2.7. Histopathology Standard histological preparations were performed for the control and for each of the 20 experimental specimens. Immediately following experiments, the specimens were fixed in 10% formaldehyde solution for 16 h and then decalcified in a solution of 40 ml 65 vol% nitric acid, 20 ml 10 vol% formaldehyde and 340 ml distilled water for 48 h. Following decalcification, the bone samples were embedded in paraffin wax and 5 m thick slices were taken from the middle of each specimen, perpendicular to the cut surface using a rotary microtome, Fig. 2. The bone slices were mounted to slides, stained with hematoxylin–eosin, and evaluated with an inverted microscope and measurement software (Omnimet Enterprise, Buehler, Lake Bluff, IL, USA). The distance from the cut surface to the first viable
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
Fig. 2. Preparation of cortical bone samples for histopathology. The figure is not to scale. The surface cut by the saw, ABCD, is perpendicular to the thin slice of bone, EFGH, that was prepared by microtome. Length AB 25.9 mm, thickness BC 5.8 mm, and depth FG 2.0 mm. Measurements were taken from line EF in the direction of FG to determine the depth of osteonecrosis.
367
Fig. 3. Volumetric cutting rate for one blade over 40 consecutive cross-cuts in bovine cortical bone (thrust force = 30 N, blade speed = 12,000 oscillations per minute).
osteocytes was measured from 10 high-powered fields for each specimen. Two technicians, operating independently and blind to the experimental conditions, each measured the depth of the necrotic zones (n = 400; 4 trials in the DOE × 5 replications per DOE trial × 10 depth measurements per sample × 2 operators independently taking measurements). 2.8. Statistical analysis Averages of the depths of empty lacunae for each experimental replicate were expressed as confidence intervals (95% probability). An analysis of variance was performed to determine if the measured changes in depth of osteocyte cell death, related to various levels of blade speed and thrust force, were statistically significant (p < 0.05). Through regression analysis of the full factorial DOE, an equation was developed to predict the depth of osteocyte cell death. Decision limits were calculated to assess the statistical significance of the factor effects (force and speed) and their interaction (t-test, ˛ = 0.05). The statistical analysis was performed using the Design of Experiments functionality within Minitab (ver. 16, Minitab Inc., State College, PA, USA). 3. Results Volumetric cutting rates for 40 consecutive cuts of cortical bone with one blade revealed that wear was not a significant factor when considering the DOE plan of 10 cuts per blade, Fig. 3. Through optical inspection, there was no observable chipping of teeth or changes in the cutting edge radii. As such, two blades were used to make 10 cuts each during the five replications of the full factorial experiment. Histological results for the control specimen consistently showed osteocytes within the lacunae (dark spots within the lighter circles as indicated by the arrows), Fig. 4. This demonstrated that the bone preparation process resulted in negligible thermal damage as osteocytes were visible within a few micrometers of the milled edges of the sample. In addition, there was no apparent degradation of osteocyte viability from the time of bone procurement to the time that histology studies were completed. Following preliminary experiments on blade wear and sample preparation technique, the DOE was conducted to determine the effect of applied thrust force and blade speed on the health of the boney bed. As shown in Fig. 5, the thermal effects of sawing can immediately be seen by the discoloration at the cut surface. Heat decreases the affinity of extracellular protein collagen for the pink eosin stain and promotes affinity for the blue hematoxylin stain.
Fig. 4. Histopathology of the control specimen. Viable osteocytes are shown within the lacunae (dark spots indicated by arrows) near the milled surfaces of the prepared workpieces.
Fig. 5. Representative microphotograph used to identify the depth of thermal damage to cortical bone during sagittal sawing (applied force = 30 N, blade speed = 18,000 cycles per minute). The depth of osteonecrosis (solid horizontal line) is approximately 420 m from the cut surface as defined by the dispersion of empty lacunae. The cut surface is marked by line segment EF (previously defined in Fig. 2).
368
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
Table 1 Regression coefficients from full factorial design of experiments. Factor
Coefficienta
p value
Force Speed Force × speed
−69.36 −49.41 23.11
0.002 0.017 0.220
a
Relative to +1 and −1 representing the high and low factor levels.
Factors and Interactions
2.179 (Decision Limit)
A
B
AB
A B 0
1
2 3 Standardized Effect
Fig. 7. Mean maximum temperature measured at 0.30 mm from the resection plane as a function of applied thrust force and blade speed. Error bars indicate ±1 SD.
Force Speed 4
Fig. 6. Pareto chart of standardized effects showing the relative effect of force, speed, and the interaction of force with speed, on the depth of osteonecrosis. Effect values below the decision limit of 2.179 are not considered statistically significant (t-test, ˛ = 0.05).
Thermal damage at the cut surface can be seen to extend deep within the bone as indicated by the absence of osteocytes from their lacunae. The dispersion of empty lacunae formed a well-demarcated line against the viable osteocytes, indicating the probable depth of osteonecrosis, Fig. 5. Both experimental factors, thrust force and blade speed, individually showed a statistically significant effect on the depth of osteonecrosis (p < 0.05), with the effect of force greater than the effect of speed as apparent from the relative magnitudes of the coefficients, Table 1. The negative values of the coefficients indicate that as each factor is increased from their low level (−) to their high level (+), the depth of the necrotic zone decreases. The interaction of force with speed had a measurable effect on the depth of osteonecrosis, but the factor interaction was not statistically significant (p = 0.22). The t-test decision limit was determined to be 2.179 (˛ = 0.05) according to the Pareto analysis of standardized effects, Fig. 6. The combination of low blade speed (12,000 cpm) and low force (15 N) generated the highest recorded temperature of 109.1 ◦ C (SD = 1.0 ◦ C), Fig. 7, corresponding the deepest necrotic zone at an average depth of 737 m (SD = 110 m), Fig. 8. Meanwhile, high blade speed (18,000 cpm) and high force (30 N) resulted in the lowest temperature of 93.0 ◦ C (SD = 4.5 ◦ C) and caused the least amount of thermal damage corresponding to an average necrotic depth of 500 m (SD = 52 m). The time–temperature profile demonstrated that a temperature above the necrotic threshold of approximately 50 ◦ C was sustained for 21–35 s, with the longer time duration corresponding to the higher temperature of 109.1 ◦ C. A regression analysis for the full factorial design resulted in an equation to predict the depth of osteonecrosis from the cut surface, D (m), as a function of applied thrust force, F (N), and blade speed, S (oscillation rate of the saw in cycles per minute): D = 1397 − 24.65 (F) − 3.956 × 10−2 (S) + 1.027 × 10−3 (S)(F) (1)
Fig. 8. Mean depth of osteonecrosis as a function of applied thrust force and blade speed. Error bars indicate ±1 SD.
While the interaction of force and speed was not determined to be statistically significant, it must be included in Eq. (1) (hierarchical rule) for the purpose of predicting the depth of osteonecrosis [17]. Following development of the regression equation, an independent set of experiments was conducted for verification of the predicted values. Following the same procedures as described for the DOE, sawing was done at an intermediate speed of 15,000 cpm for an applied thrust force of 15 N and 30 N. Upon analysis of the histopathology, the predicted depth of osteonecrosis from Eq. (1) fell within 10% of the measured depth (n = 60). 4. Discussion The depth of osteonecrosis during resection of bone by sagittal saw was characterized in terms of applied thrust force and blade speed, both of which can be controlled intraoperatively. Aggressive sawing accomplished by pushing firmly on the saw, while running at high blade speed, resulted in less thermal damage to bone than cutting with less applied force at lower speed. The difference was statistically significant and clinically relevant with an increase of nearly 50% in average necrotic depth from 0.50 mm to 0.74 mm when both force and speed were decreased. An increase in temperature during cutting is primarily due to frictional rubbing between the blade and the bone as well as from the energy required to create chips through deformation and
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
fracture [18]. When a surgeon applies a greater thrust force during sawing, the cutting rate increases as larger chips are created due to an increase in the depth of cut per tooth [16]. While deformation and fracture energy to create larger chips may generate more heat, the mean maximum temperature decreased with greater applied force. Considering the results at a fixed speed of 12,000 cpm, a force of 15 N resulted in a temperature of 109.1 ◦ C, but as the force was increased to 30 N the temperature dropped to 98.2 ◦ C (p < 0.05). While greater deformation and fracture may have generated more heat at the tooth–chip interface under greater thrust force, the likely reason for the drop in temperature was less time for heat conduction into the surrounding bone as the cutting rate increased. While holding thrust force constant, the effect of blade speed on temperature rise was less clear. With a change in blade speed, there are two competing factors that impact temperature rise: (1) heat generation due to blade friction, which is greater at higher blade speed, and (2) heat conduction, where similar to higher thrust force, higher blade speed results in greater cutting rates and therefore less time for heat transfer. While holding thrust force constant at 15 N, an increase in blade speed from 12,000 cpm to 18,000 cpm resulted in a significant (p = 0.035) mean maximum temperature reduction from 109.1 ◦ C to 105.0 ◦ C as shown in Fig. 7. However, at a higher force of 30 N, an increase in blade speed did not conclusively produce a decrease in temperature. While the mean maximum temperature declined from 98.1 ◦ C (SD = 1.3 ◦ C) to 93.0 ◦ C (SD = 4.5 ◦ C), the higher standard deviation associated with a mean of 93.0 ◦ C resulted in a lack of statistical significance (p = 0.115). The temperatures measured here are similar to those determined by Toksvig-Larsen et al. while sawing bovine cortical bone with an air powered sagittal saw operating at 16,000 oscillations per minute [1]. An average maximum temperature of 109 ◦ C was measured, but the thrust force was not specified as cutting was done by hand and thermocouples were placed in the blade rather than in the bone. However, research by Wächter and Stoll demonstrated that there was no significant difference in temperature when comparing the results of thermocouples placed in the blade or in bovine mandible when cutting with a reciprocating saw [11]. Temperature when cutting mandible bone was observed to be a minimum of 110 ◦ C and a maximum of 150 ◦ C for sacrificial thermocouples placed directly in the path of the blade. While these temperatures are higher than measured here with the thermocouple placed at 0.3 mm from the resected surface, the distance from the blade to the thermocouple is significant as the temperature from transient conduction decays exponentially [19]. It is also worth noting that the standard deviation reported by Toksvig-Larsen was 23 ◦ C during hand sawing, as compared to 1–5 ◦ C in the present research where a weight balance was used to prescribe a constant load. The standard deviations from thermocouple measurement reported here are more in line with those reported by Augustin et al. during a study on bone drilling where a fixture was used to control feed force and a plate was used to accurately control the distance between the hole and thermocouple. For example, the standard deviation in temperature measurement during drilling was 0.7–3.1 ◦ C when considering 24 combinations of drill bit diameter and drill speed, with the majority falling below 1.0 ◦ C [13]. Based on the author’s experience in bone sawing research, repeatable temperature measurements require the use of a gage block or precision translation stage to precisely control the gap between the thermocouple and the blade due to a rapid decay in temperature as the distance increases. The amount of heat generated and the time for heat transfer must also be standardized by using bone samples of equal cross sectional area and a fixture to control applied thrust force. In addition, it is important to harvest fresh bone samples from a single source and to maintain the samples in a moist condition. Finally, blade speed must be chosen
369
carefully to avoid regions that precipitate out-of-plane deflection (observed by strobe) during high speed oscillation resonance. Blade deflection normal to the oscillation plane causes the blade to initially bounce on the workpiece surface before eventually creating a kerf. This undesirable effect, i.e. the blade hunting to settle into a kerf, adds error to the precise distance that must be maintained between the edge of the blade and the thermocouple wire. Given the challenges of consistent temperature measurement, histopathology provided another means to directly determine the effect of heat on bone health. While previous histopathology studies related to bone sawing do not exist, a comparison to bone drilling provides similar results. Karaca et al. determined that an increase in applied force while drilling bovine cortical bone resulted in less thermal damage as evidenced by the depth of empty lacunae [14]. In this same study, it was confirmed that lower temperatures resulted when the drill speed was increased from 400 to 800 revolutions per minute, which increased the cutting rate. At the higher speed, the average necrotic depth from the edge of the drill bit was approximately 200 m. The histology method employed here has some limitations in that the depth of thermal necrosis was observed following immediate fixation and decalcification of the bone sample, but the effect of temperatures above the necrotic threshold may damage cells deeper within the bone over a longer time horizon [4]. In addition, the immediate death of osteocytes based on histological preparation and observation of empty lacunae may require a temperature that is well above the necrotic threshold temperature of approximately 50 ◦ C. In the results presented here, there is correlation between the mean maximum temperature rise measured by thermocouple and the greater depth of empty lacunae, but the results could be strengthened in future work by implanting a series of thermocouples to capture a temperature profile at progressively greater distances from the cut surface. This would provide insight into the actual temperature at the observed boundary of osteocyte viability. In addition, these experiments were conducted in vitro without the effect of cooling that is present in vivo due to an active vasculature. The ranges of thrust force and blade speed investigated herein were chosen to reflect reasonable values observed in clinical practice. The regression equation to predict the depth of osteonecrosis, Eq. (1), is only applicable within the ranges investigated. While extending the range of speed and force would provide a greater degree of applicability, there are some physical limitations to what is possible. For example, at speeds lower than 12,000 cpm and forces higher than 30 N, a phenomenon was observed whereby the teeth could not sweep freely back and forth over the complete oscillation angle, but rather would burrow straight into the bone. At thrust forces less than 10 N, the blade cut intermittently as the teeth failed to penetrate the bone. These and other sawing conditions are also affected by blade thickness, which could be investigated in future work. A 0.8 mm blade was used for this study, but thicker blades may be specified by implant suppliers to coincide with the slot width in their custom cutting guides. It takes more energy to create a thicker kerf [16], which may result in more heat being generated and transferred during sawing. Knowledge of these findings would be of value to current orthopedic surgeons, especially surgeons in training that might tend to timidly saw with gentle pressure, as well as choose lower speeds for more control during the beginning stages of their practice. However, there are limitations to these findings based on clinical technique where the application of a thrust force is constantly changing as the surgeon guides the saw through the bone. In practice, surgeons repeatedly advance the saw and drawback in an ebb and flow manner, thereby sawing over a wide range of applied forces, rather than the constant force investigated in this study. In addition, high-speed forceful cuts can reduce the amount of control a surgeon has over the saw. Therefore, a balance must be struck
370
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
for every surgeon to make as aggressive a cut as possible without sacrificing precision and control. 5. Conclusions It is feasible to assess the severity of thermal damage from sagittal sawing by measuring the depth of empty lacunae from the resected surface. Experiments on applied thrust force (15 N and 30 N) and blade speed (12,000 cpm and 18,000 cpm) indicated that greater force and higher speed resulted in less thermal damage to the bone. An increase in applied thrust force had a greater impact on reducing the depth of thermal necrosis than an increase in blade speed. Competing interests None declared. Funding No external funding was received. Internal funding was received from the School of Engineering at Tufts University, Medford, MA, USA. Ethical approval Not required. Acknowledgements The authors would like to acknowledge Tobi Quinto, Tufts Medical Center, Boston, MA, for assistance with histopathology. The research was supported with internal funding from Tufts Medical Center and the School of Engineering at Tufts University, Medford, MA, USA.
References [1] Toksvig-Larsen S, Ryd L, Lindstrand A. On the problem of heat generation in bone cutting. J Bone Joint Sur 1991;73-B:13–5. [2] Ark TW, Neal JG, Thacker JG, Edlich RF. Influence of irrigation solutions on oscillating bone saw blade performance. J Biomed Mater Res 1998;43(2):108–12. [3] Davidson SRH, James DF. Measurement of thermal conductivity of bovine cortical bone. Med Eng Phys 2000;22:741–7. [4] Lundskog J. Heat and bone tissue. An experimental investigation of the thermal properties of bone and threshold levels for thermal injury. Scand J Plast Reconstr Surg 1972;9:11–3. [5] Eriksson AR, Albrektsson T. Temperature threshold levels for heat-induced bone tissue injury: a vital-microscopic study in the rabbit. J Prosthet Dent 1983;50(1):101–7. [6] Eriksson AR, Albrektsson T. The effect of heat on bone regeneration: an experimental study in the rabbit using the bone growth chamber. J Oral Maxillofac Surg 1984;42:705–11. [7] Noble B. Bone microdamage and cell apoptosis. Eur Cell Mater 2003;6:46–55. [8] Pizzoferrato A, Ciapetti G, Stea S, Toni A. Cellular events in the mechanisms of prosthesis loosening. Clin Mater 1991;7(1):51–81. [9] Toksvig-Larsen S, Ryd L, Lindstrand A. Temperature influence in different orthopaedic saw blades. J Arthroplasty 1992;7(1):21–4. [10] Toksvig-Larsen S, Ryd L. Temperature elevation during knee arthroplasty. Acta Orthop Scand 1989;60(4):439–42. [11] Wächter R, Stoll P. Increase of temperature during osteotomy. In vitro and in vivo investigations. Int J Oral Maxillofac Surg 1991;20(4):245–9. [12] Augustin G, Zigman T, Davila S, Udilljak T, Staroveski T, Brezak D, et al. Cortical bone drilling and thermal osteonecrosis. Clin Biomech 2012;27:313–25. [13] Augustin G, Davila S, Mihoci K, Udiljak T, Vedrina DS, Antabak A. Thermal osteonecrosis and bone drilling parameters revisited. Arch Orthop Trauma Surg 2008;128:71–7. [14] Karaca F, Aksakal B, Kom M. Influence of orthopaedic drilling parameters on temperature and histopathology of bovine tibia: an in vitro study. Med Eng Phys 2011;33:1221–7. [15] Wall JC, Chatterji SK, Jeffery JW. Age-related changes in the density and tensile strength of human femoral cortical bone. Calcif Tissue Int 1979;27(2): 105–8. [16] James TP, Pearlman JJ, Saigal A. Rounded cutting edge model for the prediction of bone sawing forces. J Biomech Eng 2012;134:071001. [17] Antony J. Design of experiments for engineers and scientists. 1st ed. London: Butterworth–Heinemann; 2003. [18] Krause WR, Bradbury DW, Kelly JE, Lunceford EM. Temperature elevations in orthopaedic cutting operations. J Biomech 1982;15(4):267–75. [19] Kreith F, Manglik RM, Bohn MS. Principles of heat transfer. 7th ed. Stamford: Cengage Learning; 2011.
Contents lists available at ScienceDirect
Medical Engineering & Physics journal homepage: www.elsevier.com/locate/medengphy
Editor’s comment: This paper addresses an important issue faced by orthopedic surgeons when sawing through bone: how to minimize thermal damage to adjacent healthy tissue? The authors have conducted a systematic investigation into the effects of sawing parameters on the depth of osteonecrosis at the cut surface – a factor implicated in the aseptic loosening of orthopedic implants. These findings will be of interest to surgeons who use such tools in their clinical practice. Richard Black, Editor in Chief
Effect of applied force and blade speed on histopathology of bone during resection by sagittal saw Thomas P. James a,∗ , Gerard Chang b , Steven Micucci b , Amrit Sagar a , Eric L. Smith c , Charles Cassidy d a
Department of Mechanical Engineering, Tufts University, 200 College Avenue, Medford, MA 02155, USA Tufts University School of Medicine, 145 Harrison Avenue, Boston, MA 02111, USA c Tufts Medical Center, 800 Washington Street, Boston, MA 02111, USA d Tufts University School of Medicine, 800 Washington Street, Boston, MA 02111, USA b
a r t i c l e
i n f o
Article history: Received 25 June 2013 Received in revised form 13 November 2013 Accepted 1 December 2013 Keywords: Orthopedics Bone sawing Thermal necrosis Histopathology
a b s t r a c t A sagittal saw is commonly used for resection of bone during joint replacement surgery. During sawing, heat is generated that can lead to an increase in temperature at the resected surface. The aim of this study was to determine the effect of applied thrust force and blade speed on generating heat. The effect of these factors and their interactions on cutting temperature and bone health were investigated with a full factorial Design of Experiments approach for two levels of thrust force, 15 N and 30 N, and for two levels of blade oscillation rate, 12,000 and 18,000 cycles per minute (cpm). In addition, a preliminary study was conducted to eliminate blade wear as a confounding factor. A custom sawing fixture was used to crosscut samples of fresh bovine cortical bone while temperature in the bone was measured by thermocouple (n = 40), followed by measurements of the depth of thermal necrosis by histopathological analysis (n = 200). An analysis of variance was used to determine the significance of the factor effects on necrotic depth as evidenced by empty lacunae. Both thrust force and blade speed demonstrated a statistically significant effect on the depth of osteonecrosis (p < 0.05), while the interaction of thrust force with blade speed was not significant (p = 0.22). The minimum necrotic depth observed was 0.50 mm, corresponding to a higher level of force and blade speed (30 N, 18,000 cpm). Under these conditions, a maximum temperature of 93 ◦ C was measured at 0.3 mm from the kerf. With a decrease in both thrust force and blade speed (15 N, 12,000 cpm), the temperature in the bone increased to 109 ◦ C, corresponding to a nearly 50% increase in depth of the necrotic zone to 0.74 mm. A predictive equation for necrotic depth in terms of thrust force and blade speed was determined through regression analysis and validated by experiment. The histology results imply that an increase in applied thrust force is more effective in reducing the depth of thermal damage to surrounding bone than an increase in blade speed. © 2013 IPEM. Published by Elsevier Ltd. All rights reserved.
1. Introduction The use of power saws in orthopedics has existed since at least 1890 as evident from US Patent 436,804 awarded to M.J. Roberts for an “Electro Ostetome” that relied upon a unique reciprocating mechanism to rapidly translate the blade with an electric motor. In contemporary joint replacement surgery, reciprocating saws have
∗ Corresponding author. Tel.: +1 978 771 8945. E-mail address: [email protected] (T.P. James).
largely been replaced by battery powered tools that oscillate the blade at high frequency, 10,000–20,000 oscillations per minute (cpm), through a relatively small angle, typically 4–5◦ . Oscillating (sagittal) saws are used, for example, in total knee arthroplasty to prepare flat surfaces on the condylar ends of the femur and tibia in order to facilitate attachment of an artificial knee implant. Unfortunately, heat is generated during the bone sawing process, which can lead to higher temperatures that over time can cause cellular damage at the resected surface [1]. During sawing, a surgeon has control over applied thrust force and blade speed, which are two parameters that can affect temperature. Blade speed is
1350-4533/$ – see front matter © 2013 IPEM. Published by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.medengphy.2013.12.002
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
proportional to the product of blade length and oscillation rate of the saw. The primary motivation behind the implementation of temperature abatement strategies during bone resection and preparation by sagittal saw has been the identification of a link between osteonecrosis and aseptic implant loosening. During bone cutting, heat energy derived from shear deformation of bone and friction between the blade and the bone can cause osteonecrosis at or near the resection plane [1,2]. The local effect of heat on bone is further enhanced by cortical bone’s relatively poor thermal conductivity at 0.58 ± 0.018 W/mK in the longitudinal direction [3] and low temperature threshold for irreversible osteocyte cell death at 47–50 ◦ C [4,5]. Eriksson and Albrektsson found that heating rabbit cortical bone to 50 ◦ C for 1 min led to implant loosening and significantly reduced osteointegration when compared to a control [6]. In addition, mechanical deformation from sawing can cause micro-cracks that can lead to osteocyte apoptosis [7]. As a result, the absence of osteocytes triggers targeted osteoclastic bone resorption of the damaged area [8]. Sawing experiments have been conducted to determine the effect of saw blade design and kerf (saw cut) irrigation on cutting temperature. Toksvig-Larsen et al. found that different sagittal saw blade designs did not significantly reduce temperatures below the critical limit for heat induced bone necrosis [9]. The effect of irrigation on the necrotic threshold temperature is less clear, with Toksvig-Larsen and Ryd [10] reporting no difference in temperature due to irrigation of the kerf, while Wächter and Stoll [11] demonstrated that intermittent sawing combined with irrigation was able to adequately reduce bone temperatures. However, it appears that in practice surgeons are reluctant to forcefully irrigate the saw blade as it is difficult to inject fluids into the kerf. In addition, the kerf in total knee arthroplasty is typically obstructed by the cutting guide. Considering limited success in controlling temperature with saw blade design and practical difficulties with kerf irrigation, parameters affecting operation of the sagittal saw deserve further scrutiny. Unfortunately, there is a paucity of data on the relationship between sawing parameters and the resultant bone temperature. There is, however, significant literature on operational parameters affecting temperature during bone drilling. Acknowledging that sawing and drilling of bone may differ in the mechanics of chip creation and removal, the literature on bone drilling provides important insight and guidance into the methodologies used to investigate the problem of heat generation during bone removal. A comprehensive review of the literature on the relationship between drilling parameters and thermal osteonecrosis was recently provided by Augustin et al. [12]. From the parameters identified, the most relevant to sawing are feed rate and drill speed, which are analogous to thrust force and blade speed, respectively. The investigative method traditionally used in research on bone drilling is to measure the maximum temperature rise with a thermocouple placed in the bone. A multifactorial study on drilling parameters revealed that an increase in drill speed was associated with a higher temperature rise in the bone as measured by thermocouple, whereas an increase in feed force generally resulted in a lower temperature rise [13]. In addition to temperature measurement by thermocouple, the effect of higher temperatures on health of the boney bed can be deduced by histopathological analysis. Although this method has not been applied to bone sawing research, a recent histology study on process parameters related to bone drilling revealed that high temperatures resulted in osteocyte death as evident from an increased depth of empty lacunae from the surface of the drilled hole [14]. The objective of the current study is to determine the effect of applied thrust force and blade speed on health of the boney bed by
365
using thermocouples to measure temperature in the bone near the resected surface and by studying the histopathology of the bone to determine a necrotic depth. Drawing upon the research on bone drilling, it was hypothesized that an increase in thrust force and blade speed would correspond to a decrease in the necrotic depth. To test the hypothesis, a custom bone sawing fixture was developed and a full factorial Design of Experiments (DOE) approach was employed. Prior to conducting the DOE, initial studies were pursued to determine proper sample size for statistical significance and to determine if wear was a confounding factor. 2. Materials and methods 2.1. Oscillating saw and fixture A custom apparatus was designed for sagittal sawing experiments (Fig. 1). The fixture was fitted with an AC electric oscillating tool (Fein Multimaster, Fein Power Tool, Inc., Pittsburgh, PA, USA). This tool is primarily used in wood and metal working applications, but has the same 4–5◦ sweep angle and rate of oscillation as contemporary sagittal saws. However, unlike cordless surgical saws, where blade speed can fluctuate as a condition of the battery’s state of charge, the corded oscillating saw maintains a constant blade speed while under load as confirmed in screening studies by stroboscope (Nova Strobe, Monarch Instrument, Amherst, NH, USA). This is more suitable for cutting experiments where the effect of blade speed is a parameter being investigated. The oscillating saw was attached to vertical guide rails with low friction linear bearings to facilitate orthogonal cross-cutting of the bone workpiece, Fig. 1. Applied thrust force, normal to the workpiece surface, was generated by the weight of the tool and adjusted with a counterweight pulley system. A force gage (MG20, Mark 10 Company, Copiague, NY) with an accuracy of ±0.4 N was placed beneath the center of the saw blade to confirm the range of thrust forces being investigated. Oscillation rate was controlled by the tool’s variable speed dial and prior to each cut was confirmed by stroboscope.
Fig. 1. Sawing apparatus used to cross cut bone under various conditions of blade speed and thrust force.
366
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
2.2. Saw blade Stainless steel sagittal saw blades, 12 teeth per inch (tpi), 0.80 mm thick, 64 mm depth of cut (Stryker 2108-107-009, Stryker Surgical, Kalamazoo, MI, USA) were fitted to the oscillating saw with a custom fabricated stainless steel blade adapter. The adapter added 20 mm to the saw blade depth of cut, which proportionally affects tip speed. A combination of the blade length and adapter length where chosen specifically to result in a blade with an effective 84 mm depth of cut, which falls within the 80–90 mm range commonly used in knee and hip replacement surgery. 2.3. Bone samples Two fresh bovine femurs, taken from the hind legs of one 18–24 month adult cow, were acquired from an approved abattoir (Blood Farm, Groton, MA, USA). A bandsaw was used to remove 100 mm long sections from the mid-diaphysis regions of both femurs. Three uniform rectangular planks were then prepared by bandsaw from the thicker cortical sections and finish machined with an upright mill. The sections were milled to an equal thickness and width, resulting in a cross section of 25.9 mm by 5.8 mm. The primary osteon direction was aligned to the sample length. Finally, a series of holes, 1.6 mm in diameter, 3.0 mm deep, and equally spaced at 4.0 mm center-to-center, were drilled perpendicular to the surface of the bone to provide a light press fit with the thermocouple. During workpiece preparation, the bone was kept moist through misting with a physiologic solution (Hank’s Balanced Salt Solution, Life Technologies, Grand Island, NY, USA). The feed and speed during milling was sufficient to prevent heating of the bone. This was confirmed by histology prior to conducting the sawing experiments. Densities of the three workpieces (ave. = 2.04 g/cm3 , SD = 0.018 g/cm3 ) were consistent with those of healthy human femoral cortical bone [15]. By visual inspection, the bone appeared of uniform density with no signs of porosity or other defects. Prior to and during the sawing experiments, the unused bone workpieces were soaked in physiologic solution and kept refrigerated. 2.4. Experimental design An initial study was conducted to eliminate the possibility of blade wear as a confounding factor. Cutting times were measured over 40 consecutive cross cuts with a single blade, for an applied thrust force of 30 N and a blade speed of 12,000 cpm. Volumetric cutting rates were determined from sample cross sectional area and thickness of the kerf over the cutting time. An optical microscope was used to examine teeth for chips and to measure the cutting edge radius of several teeth before and after the 40 cuts. Concurrent with the blade wear study, a histopathology study was performed to: (1) gather data for a power analysis, (2) determine the feasibility of assessing osteocyte viability on histopathological analysis, and (3) determine the amount of thermal damage caused by the bone preparation process, which included the duration of sample preparation and microtome cutting. Following these preliminary studies, a two factor, two level, full factorial Design of Experiments (DOE) was implemented. The DOE approach is statistically efficient when determining significance of one or more variables on an outcome, especially when potential interactions of variables may exist. The outcome measure for the DOE was depth of thermal necrosis in micrometers (m) as measured from the cut surface. The two factors considered were applied thrust force, normal to the workpiece surface, and oscillation rate (blade speed). Two levels were sufficient to represent the factors based on a preliminary study of cutting rate that indicated a linear response to changes in both factors. The upper (+)
and lower (−) values for thrust force were 30 N (6.74 lbs) and 15 N (3.37 lbs), respectively. Thrust force was chosen to represent a reasonable range that a surgeon would employ during sawing [16]. The upper (+) and lower (−) values for blade speed were 18,000 cpm and 12,000 cpm, which covered the operating speeds for typical sagittal saws [15]. Each of the four unique trials for thrust force and blade speed (++, −−, +−, −+) was replicated five times for a total of 20 experiments based on the power analysis from preliminary data (target power = 0.90, ˛ = 0.05, significant effect = 80 m, SD = 50 m). The level of significance is approximately 10% of the anticipated depth of thermal necrosis. While implementing the experiment, randomization was employed wherever possible to eliminate setup bias. The order of the cuts was randomized, as were the use of two different saw blades and three different bone workpieces. 2.5. Temperature measurement J-type thermocouple probes (1.6 mm diameter, grounded, 304 SS Sheath, #JQSS-116G-6, Omega Engineering Inc., Stamford, CT, USA) were attached to a 16-channel data acquisition unit (#9213, National Instruments Corp., Austin, TX, USA). Calibration of the thermocouples was performed with an ice bath and boiling water. Temperature data was acquired with a custom interface (LabView© 2009, National Instruments Corp.) at a rate of 20 Hz. 2.6. Sawing operations Upon setting the counterweight for the proper thrust force, the sagittal saw was turned on, lowered by hand to a resting position above the surface of the bone and then released, allowing the saw blade teeth to engage the bone. Cuts were made 0.30 mm from the thermocouple hole, resulting in 4.0 mm between cuts. Repeatability of the precise location of a saw cut from the edge of the thermocouple hole was accomplished by placing 0.30 mm shim stock between the blade teeth and the thermocouple wire as the bone was positioned with a translatable (screw drive) machinist’s vise. Calibrated thermocouples were used to measure temperature for a total of 60 s, allowing a 5 s lead time prior to cutting for measurement of a baseline workpiece temperature. After resection of the bone, the saw was turned off to minimize vibration while the temperature measurement continued. A stopwatch was used, alongside thermocouple output recordings, to confirm cut times for each trial. Sufficient time was afforded between runs to allow the bone sample and the blade to return to a baseline ambient temperature (22 ◦ C). Following sawing, the backside of each specimen (opposite the cut surface) was removed by milling to reduce the volume of bone requiring decalcification. After milling, one face on the bone specimen was marked with ink to preserve proper orientation during handling and the sample was placed in 10% formaldehyde to preserve it for histopathological analysis. 2.7. Histopathology Standard histological preparations were performed for the control and for each of the 20 experimental specimens. Immediately following experiments, the specimens were fixed in 10% formaldehyde solution for 16 h and then decalcified in a solution of 40 ml 65 vol% nitric acid, 20 ml 10 vol% formaldehyde and 340 ml distilled water for 48 h. Following decalcification, the bone samples were embedded in paraffin wax and 5 m thick slices were taken from the middle of each specimen, perpendicular to the cut surface using a rotary microtome, Fig. 2. The bone slices were mounted to slides, stained with hematoxylin–eosin, and evaluated with an inverted microscope and measurement software (Omnimet Enterprise, Buehler, Lake Bluff, IL, USA). The distance from the cut surface to the first viable
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
Fig. 2. Preparation of cortical bone samples for histopathology. The figure is not to scale. The surface cut by the saw, ABCD, is perpendicular to the thin slice of bone, EFGH, that was prepared by microtome. Length AB 25.9 mm, thickness BC 5.8 mm, and depth FG 2.0 mm. Measurements were taken from line EF in the direction of FG to determine the depth of osteonecrosis.
367
Fig. 3. Volumetric cutting rate for one blade over 40 consecutive cross-cuts in bovine cortical bone (thrust force = 30 N, blade speed = 12,000 oscillations per minute).
osteocytes was measured from 10 high-powered fields for each specimen. Two technicians, operating independently and blind to the experimental conditions, each measured the depth of the necrotic zones (n = 400; 4 trials in the DOE × 5 replications per DOE trial × 10 depth measurements per sample × 2 operators independently taking measurements). 2.8. Statistical analysis Averages of the depths of empty lacunae for each experimental replicate were expressed as confidence intervals (95% probability). An analysis of variance was performed to determine if the measured changes in depth of osteocyte cell death, related to various levels of blade speed and thrust force, were statistically significant (p < 0.05). Through regression analysis of the full factorial DOE, an equation was developed to predict the depth of osteocyte cell death. Decision limits were calculated to assess the statistical significance of the factor effects (force and speed) and their interaction (t-test, ˛ = 0.05). The statistical analysis was performed using the Design of Experiments functionality within Minitab (ver. 16, Minitab Inc., State College, PA, USA). 3. Results Volumetric cutting rates for 40 consecutive cuts of cortical bone with one blade revealed that wear was not a significant factor when considering the DOE plan of 10 cuts per blade, Fig. 3. Through optical inspection, there was no observable chipping of teeth or changes in the cutting edge radii. As such, two blades were used to make 10 cuts each during the five replications of the full factorial experiment. Histological results for the control specimen consistently showed osteocytes within the lacunae (dark spots within the lighter circles as indicated by the arrows), Fig. 4. This demonstrated that the bone preparation process resulted in negligible thermal damage as osteocytes were visible within a few micrometers of the milled edges of the sample. In addition, there was no apparent degradation of osteocyte viability from the time of bone procurement to the time that histology studies were completed. Following preliminary experiments on blade wear and sample preparation technique, the DOE was conducted to determine the effect of applied thrust force and blade speed on the health of the boney bed. As shown in Fig. 5, the thermal effects of sawing can immediately be seen by the discoloration at the cut surface. Heat decreases the affinity of extracellular protein collagen for the pink eosin stain and promotes affinity for the blue hematoxylin stain.
Fig. 4. Histopathology of the control specimen. Viable osteocytes are shown within the lacunae (dark spots indicated by arrows) near the milled surfaces of the prepared workpieces.
Fig. 5. Representative microphotograph used to identify the depth of thermal damage to cortical bone during sagittal sawing (applied force = 30 N, blade speed = 18,000 cycles per minute). The depth of osteonecrosis (solid horizontal line) is approximately 420 m from the cut surface as defined by the dispersion of empty lacunae. The cut surface is marked by line segment EF (previously defined in Fig. 2).
368
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
Table 1 Regression coefficients from full factorial design of experiments. Factor
Coefficienta
p value
Force Speed Force × speed
−69.36 −49.41 23.11
0.002 0.017 0.220
a
Relative to +1 and −1 representing the high and low factor levels.
Factors and Interactions
2.179 (Decision Limit)
A
B
AB
A B 0
1
2 3 Standardized Effect
Fig. 7. Mean maximum temperature measured at 0.30 mm from the resection plane as a function of applied thrust force and blade speed. Error bars indicate ±1 SD.
Force Speed 4
Fig. 6. Pareto chart of standardized effects showing the relative effect of force, speed, and the interaction of force with speed, on the depth of osteonecrosis. Effect values below the decision limit of 2.179 are not considered statistically significant (t-test, ˛ = 0.05).
Thermal damage at the cut surface can be seen to extend deep within the bone as indicated by the absence of osteocytes from their lacunae. The dispersion of empty lacunae formed a well-demarcated line against the viable osteocytes, indicating the probable depth of osteonecrosis, Fig. 5. Both experimental factors, thrust force and blade speed, individually showed a statistically significant effect on the depth of osteonecrosis (p < 0.05), with the effect of force greater than the effect of speed as apparent from the relative magnitudes of the coefficients, Table 1. The negative values of the coefficients indicate that as each factor is increased from their low level (−) to their high level (+), the depth of the necrotic zone decreases. The interaction of force with speed had a measurable effect on the depth of osteonecrosis, but the factor interaction was not statistically significant (p = 0.22). The t-test decision limit was determined to be 2.179 (˛ = 0.05) according to the Pareto analysis of standardized effects, Fig. 6. The combination of low blade speed (12,000 cpm) and low force (15 N) generated the highest recorded temperature of 109.1 ◦ C (SD = 1.0 ◦ C), Fig. 7, corresponding the deepest necrotic zone at an average depth of 737 m (SD = 110 m), Fig. 8. Meanwhile, high blade speed (18,000 cpm) and high force (30 N) resulted in the lowest temperature of 93.0 ◦ C (SD = 4.5 ◦ C) and caused the least amount of thermal damage corresponding to an average necrotic depth of 500 m (SD = 52 m). The time–temperature profile demonstrated that a temperature above the necrotic threshold of approximately 50 ◦ C was sustained for 21–35 s, with the longer time duration corresponding to the higher temperature of 109.1 ◦ C. A regression analysis for the full factorial design resulted in an equation to predict the depth of osteonecrosis from the cut surface, D (m), as a function of applied thrust force, F (N), and blade speed, S (oscillation rate of the saw in cycles per minute): D = 1397 − 24.65 (F) − 3.956 × 10−2 (S) + 1.027 × 10−3 (S)(F) (1)
Fig. 8. Mean depth of osteonecrosis as a function of applied thrust force and blade speed. Error bars indicate ±1 SD.
While the interaction of force and speed was not determined to be statistically significant, it must be included in Eq. (1) (hierarchical rule) for the purpose of predicting the depth of osteonecrosis [17]. Following development of the regression equation, an independent set of experiments was conducted for verification of the predicted values. Following the same procedures as described for the DOE, sawing was done at an intermediate speed of 15,000 cpm for an applied thrust force of 15 N and 30 N. Upon analysis of the histopathology, the predicted depth of osteonecrosis from Eq. (1) fell within 10% of the measured depth (n = 60). 4. Discussion The depth of osteonecrosis during resection of bone by sagittal saw was characterized in terms of applied thrust force and blade speed, both of which can be controlled intraoperatively. Aggressive sawing accomplished by pushing firmly on the saw, while running at high blade speed, resulted in less thermal damage to bone than cutting with less applied force at lower speed. The difference was statistically significant and clinically relevant with an increase of nearly 50% in average necrotic depth from 0.50 mm to 0.74 mm when both force and speed were decreased. An increase in temperature during cutting is primarily due to frictional rubbing between the blade and the bone as well as from the energy required to create chips through deformation and
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
fracture [18]. When a surgeon applies a greater thrust force during sawing, the cutting rate increases as larger chips are created due to an increase in the depth of cut per tooth [16]. While deformation and fracture energy to create larger chips may generate more heat, the mean maximum temperature decreased with greater applied force. Considering the results at a fixed speed of 12,000 cpm, a force of 15 N resulted in a temperature of 109.1 ◦ C, but as the force was increased to 30 N the temperature dropped to 98.2 ◦ C (p < 0.05). While greater deformation and fracture may have generated more heat at the tooth–chip interface under greater thrust force, the likely reason for the drop in temperature was less time for heat conduction into the surrounding bone as the cutting rate increased. While holding thrust force constant, the effect of blade speed on temperature rise was less clear. With a change in blade speed, there are two competing factors that impact temperature rise: (1) heat generation due to blade friction, which is greater at higher blade speed, and (2) heat conduction, where similar to higher thrust force, higher blade speed results in greater cutting rates and therefore less time for heat transfer. While holding thrust force constant at 15 N, an increase in blade speed from 12,000 cpm to 18,000 cpm resulted in a significant (p = 0.035) mean maximum temperature reduction from 109.1 ◦ C to 105.0 ◦ C as shown in Fig. 7. However, at a higher force of 30 N, an increase in blade speed did not conclusively produce a decrease in temperature. While the mean maximum temperature declined from 98.1 ◦ C (SD = 1.3 ◦ C) to 93.0 ◦ C (SD = 4.5 ◦ C), the higher standard deviation associated with a mean of 93.0 ◦ C resulted in a lack of statistical significance (p = 0.115). The temperatures measured here are similar to those determined by Toksvig-Larsen et al. while sawing bovine cortical bone with an air powered sagittal saw operating at 16,000 oscillations per minute [1]. An average maximum temperature of 109 ◦ C was measured, but the thrust force was not specified as cutting was done by hand and thermocouples were placed in the blade rather than in the bone. However, research by Wächter and Stoll demonstrated that there was no significant difference in temperature when comparing the results of thermocouples placed in the blade or in bovine mandible when cutting with a reciprocating saw [11]. Temperature when cutting mandible bone was observed to be a minimum of 110 ◦ C and a maximum of 150 ◦ C for sacrificial thermocouples placed directly in the path of the blade. While these temperatures are higher than measured here with the thermocouple placed at 0.3 mm from the resected surface, the distance from the blade to the thermocouple is significant as the temperature from transient conduction decays exponentially [19]. It is also worth noting that the standard deviation reported by Toksvig-Larsen was 23 ◦ C during hand sawing, as compared to 1–5 ◦ C in the present research where a weight balance was used to prescribe a constant load. The standard deviations from thermocouple measurement reported here are more in line with those reported by Augustin et al. during a study on bone drilling where a fixture was used to control feed force and a plate was used to accurately control the distance between the hole and thermocouple. For example, the standard deviation in temperature measurement during drilling was 0.7–3.1 ◦ C when considering 24 combinations of drill bit diameter and drill speed, with the majority falling below 1.0 ◦ C [13]. Based on the author’s experience in bone sawing research, repeatable temperature measurements require the use of a gage block or precision translation stage to precisely control the gap between the thermocouple and the blade due to a rapid decay in temperature as the distance increases. The amount of heat generated and the time for heat transfer must also be standardized by using bone samples of equal cross sectional area and a fixture to control applied thrust force. In addition, it is important to harvest fresh bone samples from a single source and to maintain the samples in a moist condition. Finally, blade speed must be chosen
369
carefully to avoid regions that precipitate out-of-plane deflection (observed by strobe) during high speed oscillation resonance. Blade deflection normal to the oscillation plane causes the blade to initially bounce on the workpiece surface before eventually creating a kerf. This undesirable effect, i.e. the blade hunting to settle into a kerf, adds error to the precise distance that must be maintained between the edge of the blade and the thermocouple wire. Given the challenges of consistent temperature measurement, histopathology provided another means to directly determine the effect of heat on bone health. While previous histopathology studies related to bone sawing do not exist, a comparison to bone drilling provides similar results. Karaca et al. determined that an increase in applied force while drilling bovine cortical bone resulted in less thermal damage as evidenced by the depth of empty lacunae [14]. In this same study, it was confirmed that lower temperatures resulted when the drill speed was increased from 400 to 800 revolutions per minute, which increased the cutting rate. At the higher speed, the average necrotic depth from the edge of the drill bit was approximately 200 m. The histology method employed here has some limitations in that the depth of thermal necrosis was observed following immediate fixation and decalcification of the bone sample, but the effect of temperatures above the necrotic threshold may damage cells deeper within the bone over a longer time horizon [4]. In addition, the immediate death of osteocytes based on histological preparation and observation of empty lacunae may require a temperature that is well above the necrotic threshold temperature of approximately 50 ◦ C. In the results presented here, there is correlation between the mean maximum temperature rise measured by thermocouple and the greater depth of empty lacunae, but the results could be strengthened in future work by implanting a series of thermocouples to capture a temperature profile at progressively greater distances from the cut surface. This would provide insight into the actual temperature at the observed boundary of osteocyte viability. In addition, these experiments were conducted in vitro without the effect of cooling that is present in vivo due to an active vasculature. The ranges of thrust force and blade speed investigated herein were chosen to reflect reasonable values observed in clinical practice. The regression equation to predict the depth of osteonecrosis, Eq. (1), is only applicable within the ranges investigated. While extending the range of speed and force would provide a greater degree of applicability, there are some physical limitations to what is possible. For example, at speeds lower than 12,000 cpm and forces higher than 30 N, a phenomenon was observed whereby the teeth could not sweep freely back and forth over the complete oscillation angle, but rather would burrow straight into the bone. At thrust forces less than 10 N, the blade cut intermittently as the teeth failed to penetrate the bone. These and other sawing conditions are also affected by blade thickness, which could be investigated in future work. A 0.8 mm blade was used for this study, but thicker blades may be specified by implant suppliers to coincide with the slot width in their custom cutting guides. It takes more energy to create a thicker kerf [16], which may result in more heat being generated and transferred during sawing. Knowledge of these findings would be of value to current orthopedic surgeons, especially surgeons in training that might tend to timidly saw with gentle pressure, as well as choose lower speeds for more control during the beginning stages of their practice. However, there are limitations to these findings based on clinical technique where the application of a thrust force is constantly changing as the surgeon guides the saw through the bone. In practice, surgeons repeatedly advance the saw and drawback in an ebb and flow manner, thereby sawing over a wide range of applied forces, rather than the constant force investigated in this study. In addition, high-speed forceful cuts can reduce the amount of control a surgeon has over the saw. Therefore, a balance must be struck
370
T.P. James et al. / Medical Engineering & Physics 36 (2014) 364–370
for every surgeon to make as aggressive a cut as possible without sacrificing precision and control. 5. Conclusions It is feasible to assess the severity of thermal damage from sagittal sawing by measuring the depth of empty lacunae from the resected surface. Experiments on applied thrust force (15 N and 30 N) and blade speed (12,000 cpm and 18,000 cpm) indicated that greater force and higher speed resulted in less thermal damage to the bone. An increase in applied thrust force had a greater impact on reducing the depth of thermal necrosis than an increase in blade speed. Competing interests None declared. Funding No external funding was received. Internal funding was received from the School of Engineering at Tufts University, Medford, MA, USA. Ethical approval Not required. Acknowledgements The authors would like to acknowledge Tobi Quinto, Tufts Medical Center, Boston, MA, for assistance with histopathology. The research was supported with internal funding from Tufts Medical Center and the School of Engineering at Tufts University, Medford, MA, USA.
References [1] Toksvig-Larsen S, Ryd L, Lindstrand A. On the problem of heat generation in bone cutting. J Bone Joint Sur 1991;73-B:13–5. [2] Ark TW, Neal JG, Thacker JG, Edlich RF. Influence of irrigation solutions on oscillating bone saw blade performance. J Biomed Mater Res 1998;43(2):108–12. [3] Davidson SRH, James DF. Measurement of thermal conductivity of bovine cortical bone. Med Eng Phys 2000;22:741–7. [4] Lundskog J. Heat and bone tissue. An experimental investigation of the thermal properties of bone and threshold levels for thermal injury. Scand J Plast Reconstr Surg 1972;9:11–3. [5] Eriksson AR, Albrektsson T. Temperature threshold levels for heat-induced bone tissue injury: a vital-microscopic study in the rabbit. J Prosthet Dent 1983;50(1):101–7. [6] Eriksson AR, Albrektsson T. The effect of heat on bone regeneration: an experimental study in the rabbit using the bone growth chamber. J Oral Maxillofac Surg 1984;42:705–11. [7] Noble B. Bone microdamage and cell apoptosis. Eur Cell Mater 2003;6:46–55. [8] Pizzoferrato A, Ciapetti G, Stea S, Toni A. Cellular events in the mechanisms of prosthesis loosening. Clin Mater 1991;7(1):51–81. [9] Toksvig-Larsen S, Ryd L, Lindstrand A. Temperature influence in different orthopaedic saw blades. J Arthroplasty 1992;7(1):21–4. [10] Toksvig-Larsen S, Ryd L. Temperature elevation during knee arthroplasty. Acta Orthop Scand 1989;60(4):439–42. [11] Wächter R, Stoll P. Increase of temperature during osteotomy. In vitro and in vivo investigations. Int J Oral Maxillofac Surg 1991;20(4):245–9. [12] Augustin G, Zigman T, Davila S, Udilljak T, Staroveski T, Brezak D, et al. Cortical bone drilling and thermal osteonecrosis. Clin Biomech 2012;27:313–25. [13] Augustin G, Davila S, Mihoci K, Udiljak T, Vedrina DS, Antabak A. Thermal osteonecrosis and bone drilling parameters revisited. Arch Orthop Trauma Surg 2008;128:71–7. [14] Karaca F, Aksakal B, Kom M. Influence of orthopaedic drilling parameters on temperature and histopathology of bovine tibia: an in vitro study. Med Eng Phys 2011;33:1221–7. [15] Wall JC, Chatterji SK, Jeffery JW. Age-related changes in the density and tensile strength of human femoral cortical bone. Calcif Tissue Int 1979;27(2): 105–8. [16] James TP, Pearlman JJ, Saigal A. Rounded cutting edge model for the prediction of bone sawing forces. J Biomech Eng 2012;134:071001. [17] Antony J. Design of experiments for engineers and scientists. 1st ed. London: Butterworth–Heinemann; 2003. [18] Krause WR, Bradbury DW, Kelly JE, Lunceford EM. Temperature elevations in orthopaedic cutting operations. J Biomech 1982;15(4):267–75. [19] Kreith F, Manglik RM, Bohn MS. Principles of heat transfer. 7th ed. Stamford: Cengage Learning; 2011.
