Effect of Drilling Speed on Dental Implant Insertion Torque

The long-term success of dental implants depends on the quality of the implant-bone interface. Branemark et al proposed that bone healing with concomitant osseointegration is related to primary dental implant stability or resistance to micromotion. This stability is in turn connected to several factors including: implant design, surgical technique, bone quality, and drilling speed.^1–3  Although the combined effect of all these factors on primary dental implant stability is commonly measured by insertion torque, this paper investigates how rotational velocity of the drill effects implant insertion torque.

Numerous in vivo and in vitro studies evaluate primary implant stability by measuring maximum insertion torque during implant placement.^4–6  Insertion torque could vary due to a variation in the size, shape, and direction of the osteotomy site with free hand drilling. Furthermore, the pressure on the drill and drilling speed also affect the geometry of the osteotomy even with the same sequence of drills.

Previous in vitro experiments on polyurethane simulated bone blocks were used to reveal the relative contribution of trabecular and cortical bone to insertion torque.⁷  That previous work introduced a new form of “mechanical spectroscopy” using plots of insertion torque as a function of implant axial displacement. To ensure that the center of rotation during drilling of the pilot hole was concentric with the center of rotation during implant insertion, a modified commercial Mark-10 torque measurement unit (Figure 1; Mark-10 Corporation, Copaigue, NY) was used for drilling and implant insertion procedures. This system provided a consistent pressure on the drills. The maximum rotational speed of the device was 15 rpm. This slow rotation created a concern that the cutting surfaces of the drill bits used to make the pilot holes were chipping rather than cutting the polyurethane.⁸  This, in turn, raised concerns about the quality of the simulated bone surface presented to the implant during insertion, which could ultimately affect the torque required for implant placement.

The current paper investigates whether slow drilling speeds produce pilot hole surfaces that result in different insertion torques than pilot holes made with higher speed drilling. Our hypothesis is that the final insertion torque values do not significantly differ between the low-speed drilling protocol, 15 rpm, and a relatively high-speed drilling protocol, 1500 rpm. Testing this hypothesis required that we develop a new method for securing test pieces into a benchtop drill press and the MARK 10 implant placement and torque sensing machine in a such a way that the center of rotation of the pilot hole remained the center of rotation when placing the implant. After discussing the results of low- vs high-speed drilling of pilot holes, we then discuss whether the conclusions from previous work⁷  are expected to remain valid in light of the new results presented in this paper.

Simulated bone blocks made of polyurethane rigid foam (Sawbones, Pacific Research Laboratories Inc, Vashon, Wash) are used: 0.08 g/mL block to mimic trabecular bone, and a 3-mm thick 0.64 g/mL sheet overlayer to mimic cortical bone. Two types of titanium implant are used (Nobel Active, Nobel Biocare AB, Göteborg, Sweden) for insertion torque testing: narrow platform (NP) 3.5-mm diameter and regular platform (RP) 4.3-mm diameter. Both implants are nominally 13 mm in length. Actual measured lengths show the implants to be 12.5-mm long (Figure 2).

In our previous study, samples were clamped into the Mark 10 system using a vice integrated with the sample platform. In the current study, we needed a method for transferring the sample from a separate drilling machine onto the Mark 10 while maintaining the same center of rotation. To accomplish this, the polyurethane blocks were imbedded into green wax (Rigidax WI-Green 24-12, M. Argüeso & Co, Inc, Muskegon, Mich) in an aluminum boat after melting the wax by partially submerging the boat in a 90°C water bath for 5 minutes. During placement in the melted wax, samples were lined up with evenly spaced detents on the side of the boat (Figure 3). A thermocouple probe (USB-TC01, National Instruments, Austin, Tex) was used to ensure the temperature within the block stayed below the critical temperature for bone necrosis, 47°C, during the imbedding process.^9–12  The boat was then placed into a freezer for 5 minutes to cool the wax.

For the high-speed (1500 rpm) holes, the aluminum boat was inserted into a custom jig on the table of a benchtop drill press (Jet-8, Jet Tools North America, La Vergne, Tenn). A series of equally spaced holes were then drilled along the long axis of the polyurethane block. Drilling consisted of using a sequence of 3 drill bits with final drills of 2.8/3.2 mm diameter for NP implant and 3.2/3.6 mm for RP implants. For clarity, we will refer to the smaller bit as a 2.8-mm bit and the larger as a 3.6-mm bit. The aluminum boat was then removed from the drill press and transferred over to a commercial torque measurement unit (Mark-10; Figure 1) for implant insertion. A second custom jig, identical to the one used on the drill press, was used on the Mark-10 stage. This ensured that the center of rotation used to drill the pilot holes coincided with the center of rotation used to insert the implants. This was necessary because eccentric motion of the implant during insertion easily caused enlargement and change of cross-sectional shape of the pilot hole. For the low speed (15 rpm) holes, the aluminum boat was inserted directly into the Mark-10 system. The pilot holes were created and then the implant was placed. Details regarding how the Mark-10 system was used to insert the implants were previously described.⁷ 

Morphological characterization of the holes created by low- and high-speed drilling was carried out at both macro and microscopic length scales. At the macro scale, the shape and diameter were measured. At the microscale, roughness of the drilled surface was measured. These roughness measures were destructive since the samples had to be cut in longitudinal section along the long axis of the pilot holes, leaving them unusable for further testing. On a separate set of samples, prepared in an identical manner, implants were inserted and torque vs implant axial displacement curves were measured as described previously.⁷ 

To characterize the pilot holes at the macro scale, 5 pilot holes were drilled (3 holes with a final diameter of 3.2 mm and 2 holes with a final diameter of 3.6 mm) using each drilling protocol. Each hole was imaged using a digital SLR camera (Evolt E-520, Olympus, Tokyo, Japan) with an Olympus Zuiko 35-mm macro lens and analyzed in ImageJ (an open source Java image processing program developed at the National Institutes of Health). The grayscale images were converted to binary using the threshold function. The measure function in ImageJ was used to obtain the major axis, circularity, and roundness of each hole. The given information was then used to calculate the area and the perimeter of each corresponding hole.

To measure surface roughness of the pilot holes, a diamond saw was used to cut the polyurethane blocks along the axis of symmetry for each pilot hole. Using a stylus profilometer (Form Talysurf, Model PGI-1200, Taylor Hobson, Inc, Leicester, UK), a 2-mm length of the dense overlayer was analyzed in the direction of the long axis of the hole. Analysis was restricted to this region because our previous study⁷  showed that most of the insertion torque arises from interaction of the implant with this dense overlayer. Profilometer data were analyzed using a matrix based programming platform (MATLAB, MathWorks Inc, Natick, Mass). Although root mean square and average roughness are more conventional topographical measures, they were not the most appropriate for our analysis. The chipping phenomenon hypothesized to occur during low-speed drilling is likely to have removed large volumes of material. If the characteristic lengths of these removed volumes approached the thickness of an implant thread along the screw's axial direction, then we expect to find a decrease in insertion torque. Thread thickness for the implants used was approximately 0.5 mm. Thus, the profilometer data was analyzed in 0.5-mm increments to determine if characteristic lengths of the removed volumes exceeded the width of an implant thread. We slid the 0.5-mm analysis region down the 2-mm measured section of the hole and report the average and maximum peak-to-peak differences found within each 0.5-mm section.

For each implant type, insertion torque as a function of implant displacement was measured for 24 pilot holes (6 holes drilled with a final diameter of 3.2 mm at 15 rpm, 6 holes drilled with a final diameter of 3.6 mm at 15 rpm, 6 holes drilled with a final diameter of 3.2 mm at 1500 rpm, and 6 holes drilled with a final diameter of 3.6 mm at 1500 rpm). Torque and displacement data were collected as a function of time for both sets of holes. Following data collection, 6 plots were generated representing torque vs displacement data from implants inserted using low- and high-speed drilling protocols.

We did consider use of an implant stability quotient (ISQ) system to evaluate implant primary stability but rejected this idea. We did an extensive pilot study on ISQ measurements by resonance frequency analysis (RFA) and found it has a good repeatability but poor sensitivity. For example, 2 implants having 77 and 75 ISQ, respectively, showed accurate repeated ISQ measurements. However, their insertion torque values ranged from 80 NTcm²  for the 77 ISQ sample to 62 NTcm²  for the ISQ 75 sample. That is, small differences in ISQ were associated with large differences in insertion torque. Therefore, we abandoned the ISQ system and created our own system to measure insertion torque values. This is consistent with another primary implant stability study using clinical results to show that there was no correlation between ISQ and insertion torque.¹³ 

For the macroscopic analysis, each of the 5 holes drilled with a low-speed drill and analyzed in ImageJ, were paired with a similar sized hole drilled with a high-speed drill. After pairing the pilot holes, a paired sample bootstrap analysis was done to compare the 5 pairs of perimeters and areas of the 2 types of pilot holes. The microscopic analysis showed a lack of normality and symmetry in the profilometer distributions. Thus, we elected to run a Wilcoxon rank sum test to compare the largest peak-to-peak distances within a 0.5-mm analysis region for each 2.0-mm line analyzed in the low-speed drilled holes and each line analyzed in the high-speed drilled holes. Finally, we used a Wilcoxon signed-rank test to compare medians of the maximum and final insertion torques. In this test, each maximum insertion torque value was paired with the final insertion torque value of the same curve. The difference in medians of the average final torques between the 2 drilling protocols was analyzed using a Wilcoxon rank sum test. All analyses were performed using functions such as smean.cl.boot and wilcox.test in

R 3.5.1 (Vienna, Austria) and P values <.05 were considered statistically significant. Statistical significance was set at P < .05 since this was the condition used for significance in our previous work on implant insertion torques.⁷ 

Methodology, statistical analysis, results, and conclusions were reviewed by Dr Yuping Wu, professor of Statistics at Cleveland State University.

Typical binary images of pilot holes drilled to accommodate RP (Figure 4) and NP (Figure 5) implants are shown pairing low speed drilling (Figures 4a and 5a) next to high-speed drilling (Figures 4b and 5b). Observing the holes drilled using a 3.6-mm diameter drill bit (Figure 4) reveals a smoother edge for the high-speed drilled hole (Figure 4b) compared to the low-speed drilled hole (Figure 4a). Although the high-speed drilled hole is circular, the low-speed drilled hole is more elliptically shaped. When comparing the low- and high-speed holes drilled using a 2.8-mm diameter drill bit (Figure 5), both holes are circular; however, the low-speed drilled hole has a rougher edge (Figure 5a) while the high-speed drilled hole has a smoother edge (Figure 5b). The average perimeter of the 3.6-mm diameter holes drilled using a high- and a low-speed drilling protocol are 13.4 mm ± 0.07 mm and 14.1 mm ± 0.14 mm, respectively. The average perimeter of the 2.8-mm diameter holes drilled using a high- and low-speed drilling protocol are 10.5 mm ± 0.15 mm and 11.3 mm ± 0.15 mm, respectively. At a 0.05 significance level, we observe a difference between the perimeter of low-speed drilled holes and the high-speed drilled holes. Furthermore, on average, a 0.70-mm increase in perimeter is observed when drilling with a lower speed drill. With the same level of confidence, we observe a difference in the area of the holes. On average a 0.54 mm² decrease in area is seen when drilling with the lower speed drill. Data collected through ImageJ for perimeter and area data are displayed in Table 1. Typical profilometer plots are shown for a pilot hole drilled using a 3.6-mm drill bit (Figure 6). The displacement in the Z direction ranges from −70 μm to 65 μm for the holes drilled using the low speed and from −40 μm to 40 μm for the holes drilled using the high-speed drilling protocol. Although the features with amplitudes <10 μm are similar in both low- and high-speed drilled cases, the larger features with amplitudes >60 μm are more prevalent in the lower speed drilled holes. The average difference between the maximum and minimum peak (in a 0.5-mm analysis region) for all types of holes are shown in Table 2. Both types of holes drilled, with 2.8-mm and 3.6-mm diameter drill bits, show a mean peak-to-peak distance of ∼100 μm when using the low-speed drilling protocol. The high-speed drilling protocol resulted in mean peak-to-peak distance of <61 um.

Typical torque (τ) vs displacement (D) curves are shown for RP (Figure 7) and NP (Figure 8) implants. Ten implants were inserted into pilot holes drilled using low- (Figures 7a and 8a) and high- (Figures 7b and 8b) speed drilling. All data is displayed using the same axis ranges: torque ranging from 0 to 120 Ncm on the y-axis and displacements ranging from 0 to 13 mm on the x-axis.

RP implants show that both low- and high-speed drilled holes produce the same general curve shape: a monotonically rising torque with a plateau in the 4–6 mm range, a peak around 11 mm, and then a drop in torque toward the end of the insertion. The low-speed drilled holes show a drop off from maximum torque to final torque of 16% ± 4.6% (Figure 7a). The high-speed drilled holes show a smaller dropoff of 11% ± 3.5% (Figure 7b). A Wilcoxon signed-rank test showed a significant difference between the max and final insertion torque values for the low- and the high-speed drilling protocol.

NP implants show a monotonic increase to maximum torque (Figure 8). As evidenced by the range of torque values displayed for a given insertion distance, there was more curve-to-curve variability when implants were inserted into low speed drilled holes (Figure 8a) compared to the high-speed drilled holes (Figure 8b). This difference was less pronounced with the RP implants (Figure 7). The NP implant high-speed drill average curve shows a maximum torque of 71 Ncm ± 0.9 Ncm with a drop in maximum torque of ∼1% to 70 Ncm ± 0.8 Ncm. The average low-speed drill curve shows a maximum torque of 73 Ncm ± 2.1 N drop in maximum torque of ∼1% to 72 Ncm ± 2.0 Ncm.

To visualize differences between low- and high-speed drill cases, averages for each set of curves are shown in Figure 9. When looking at a specific displacement for the RP implant (Figure 9a), the high-speed drilled holes (dashed lines) produced slightly higher torque values than the low-speed drilled holes (solid lines) until the peak is reached at about 11 mm. When comparing the NP curves (Figure 9b), the 2 curves have torque values that differ from one another by less than the error bars. The RP implant high-speed drill average curve shows a maximum torque of 93 Ncm ± 3.2 Ncm (N = 4) with a drop in maximum torque of ∼16% to 78 ± 3.6 Ncm. The average low-speed drill curve shows a maximum torque of 93 ± 1.6 Ncm (N = 4) with a drop in maximum torque of ∼11% to 83 Ncm ± 3.6 Ncm.

This study quantitatively contrasts insertion torque vs axial displacement between a low- and a high-speed drilling protocol. An example of torque curves as a function of implant displacement is seen in Figures 7 and 8. These curves show a RP implant having similar torque vs displacement curves for both low- and high-speed drilling protocols. As observed in Figures 7a and b, the insertion torque values increase to a max torque value as the displacement value increases, and then experience a dropoff due to the reverse bevel design near the abutment of the implant.⁷  Both Figures 7a and b reveal a significant difference between the maximum insertion torque and the final insertion torque (P value = .13 and a 90% CI of [12.0–15.4] for the high-speed drilled holes and a CI of [7.46–11.3] for the low-speed drilled holes). Note that our hypothesis was that there would not be a difference in insertion torques between the low- and high-speed drilled holes at a P value of .05 or lower. Thus, P values larger than .05 support the hypothesis. It is evident in Figure 9a that, between the high- and low-speed drilling protocols, the high-speed drilling protocol yields a slightly higher average torque for a given displacement. However, there is no significant difference observed in average final torque values between the 2 different drilling protocols' very wide 95% CI of [−14.7–1.36] which includes 0 (P value = .114).

Furthermore, torque values as a function of NP implant displacement are seen in Figures 8a and b. These curves show NP implants having similar profiles for both a low- and high-speed drilling protocol. As observed in Figure 8, the torque curves monotonically increase to maximum torque. Although each of the high-speed curves (Figure 8b) contains similar torque values for a given displacement, there is more variability in the torque values for the low-speed drilled holes (Figure 8a). Again, this may be because high-speed drilling may cut through the material generating a consistent smooth material-implant interface, while the low-speed drill could be chipping away the material causing a rough material-implant interaction and generating slight variability in torque measurements. With all of this considered, Figure 9b reveals no significant difference in average final torque values between the two drilling protocols (P value = .06 with a 95% CI of −3.62–0.90). Although this P value is much closer to our target level of significance, it is still greater than the .05 level we used to determine significance in our previous study.⁷  These results support the null hypothesis: no significant difference exists in final insertion torque between the 2 drilling protocols.

Although the mechanical function of the holes drilled with low- and high-speed protocols appear similar with respect to insertion torque when placing dental implants, the fine structure of the holes is different. By looking at the pilot hole images, the edge of holes drilled using a low-speed drill are visibly different from holes drilled using the high-speed drill. This qualitative finding is supported by our quantitative analyses. Holes drilled using a lower speed drilling protocol showed a smaller area compared to high-speed drilled holes (Table 1). In addition, profilometer data showed the mean peak-to-peak distance (in a 0.5-mm analysis region) is about 45 μm higher for the low-speed drilled holes across both types of implants (Table 2). On average, we also observed a higher average perimeter for holes drilled using a lower speed drilling protocol (Table 1). Such difference in the perimeter could be due to a chipping mechanism caused by the low-speed drill ultimately, leading to multiple edges instead of one smooth edge. Though there were measureable differences in hole shape, size, and roughness between the 2 drilling protocols, these differences were not sufficient to cause a significant change in the insertion torque during implant placement.

In light of the new results presented here, it is useful to look back at the conclusions we previously stated regarding data collected using a low-speed drill.⁷  Our previous data supported 3 conclusions. First, for Nobel Biocare RP and NP implants, the system showed that retrograde slopes near the top of the implant reduce the final insertion torque when placing the implant into thin dense cortical layers supported over lower density under layers. Second, final torque values depend primarily on the thickness of cortical bone when this layer is 1–3 mm thick. Third, there is a synergistic effect when inserting implants into thin dense layers over lower density blocks.⁷  The results presented here show no statistically significant differences in the torque vs displacement curves between low- and high-speed drilled pilot holes when placing Nobel Biocare RP and NP implants into Sawbones blocks similar in construction to the blocks used in our previous work. Thus, the conclusions stated in that work would be expected to remain in force had that data been collected using high-speed drilled pilot holes.

Although beyond the scope of the data presented here, it is worth asking how one might expect the results to vary if we were to investigate implants with a different thread design. For example, the self-tapping thread used here requires drilling an undersized pilot hole. Upon insertion, these self-tapping implants apply a large compressive force to the walls of the pilot hole. Thus, small imperfections in circularity and/or surface roughness may well become irrelevant to the insertion torque. A non-self-tapping design that used a larger pilot hole might well be more sensitive to the small changes in pilot hole geometry we measured. This, in turn, would result in a measurable difference in implant insertion torque between the low- and high-speed drilled holes.

Previous in vivo^14,15  and in vitro^16,17  studies have suggested that high insertion torque values prevent adverse micro motion under implant loading, which is desirable for improved implant integration particularly immediately after implant placement. Insertion torque value has been positively correlated with primary implant stability, which is influenced by the interplay between implant macrothread design, surface roughness, and the surrounding bone if the same surgical techniques are used.

Implant macro geometry and associated drilling dimensions provide optimized implant placement, stress distribution, and lower degrees of micromotion, thus improving the conditions for bone formation under immediate and early loading conditions. This interplay of implant design and surgical technique has been regarded as a crucial step in the rehabilitation process. Nobel Active implants have standard V-shaped thread design. Other implant systems on the market have square, reverse buttress, or knife-edge aggressive thread design. To eliminate other variables, we standardized our study using the V thread design only. Within the limitations of this study, we are unable to make statements regarding how other non-V thread designs would affect the outcome measurements in terms of insertion torque and surface roughness values. Our experimental method can certainly be used in the future to test and compare other systems to determine how their thread designs may impact primary implant stability in vitro.

A new drilling apparatus and protocol is presented that enables us to compare insertion torque of dental implants after low- and high-speed drilling of pilot holes. A difference in perimeter and area of the 2 types of holes was observed (alpha = 0.05). Moreover, profilometer data revealed a significant difference in peak-to-peak distances between the low- and high-speed drilled holes (P = .03). However, the torque vs displacement data show that there is no significant difference in final insertion torque between the 2 drilling protocols (P = .06 for NP and P = .25 for RP). Therefore, we conclude that the roughened surface, increased perimeter, and decreased area caused by the low-speed drilling protocol does not significantly impact insertion torque. This allows retention of the conclusions published in our previous work even under conditions of more clinically relevant high-speed drilling of pilot holes.⁷ 

Abbreviations

ISQ:

implant stability quotient

NP:

narrow platform

RFA:

resonance frequency analysis

RP:

regular platform

We thank Christian Nguyen, DDS, and Jeehee Kim for useful discussions and assistance with data collection. We also thank Yuping Wu, professor of Statistics at Cleveland State University, for her careful and independent review of the manuscript. This work was funded in part by a grant from the American Academy of Implant Dentistry (AAID), and some of the work was performed in facilities funded by NIH grant number 1 C06 RR12463-01. The authors have no conflicts of interest to report relevant to this work