Pullout Strength After Multiple Reinsertions in Radial Bone Fixation

Abstract

Background: Due to bone cutting loss from self-tapping screws (STS), progressive destruction of bone can occur with each reinsertion during surgery. When considering the use of jigs that utilize multiple insertions such as those seen in ulnar and radial shortening osteotomy systems, or scenarios where a screw needs to be removed and reinserted due to some technical issue, this can be concerning, as multiple studies examining the effects of multiple reinsertions and the relationship between insertional torque and pullout strength have had mixed results. Methods: Insertional torque and pullout strength were experimentally measured following multiple reinsertions of STS for up to 5 total insertions for various densities and locations along radial sawbone shafts. Results: Torque and pullout strength were significantly greater in middle segments of the radial shaft. Our trials corroborate previous literature regarding a significant reduction in fixation between 1 and 2 insertions; beyond this, there was no significant difference between pullout strength across all segment locations as well as bone densities for 3 to 5 insertions. There was a moderate to high correlation of insertional torque to pullout strength noted across all bone densities and segments (Pearson r = 0.663, P < .001). Conclusion: While reinsertion of STS between 1 and 2 insertions has been shown to significantly differ in pullout strength, beyond this, there does not appear to be a significant difference in up to 5 insertions at any specific region of radial bone across a range of sawbone densities. Further insertions may be considered with caution.

Introduction

Surgical screws are essential in the fixation of long bone injuries and have been extensively studied to determine their holding power in bone. Both self-tapping and non-self-tapping screws have been shown to be effective in fixating fractures. However, self-tapping screws (STS) have several advantages including ease and speed of application due to elimination of a step.¹⁴ Eliminating this tapping step saves time and minimizes the opportunity for a reduction to be lost without an increase in operative risk. Shortening surgery time and the number of instruments used would theoretically additionally reduce the risk, while minimal, of infections and other complications associated with increased operating time and duration of anesthesia.^{5,14,18,19,20,23,26,27,29,30,34,35}

Due to the bone cutting loss from STS, progressive destruction of bone within the pilot hole can occur with each reinsertion. Screw removal and reinsertion, therefore, has been considered as a factor that can affect pullout strength. This concept is important to evaluate for a variety of reasons. When considering implants which are placed with a jig such as the distal radial ulnar joint Scheker prosthesis (Aptis Medical, Louisville, Kentucky) or the Rayhack ulnar and radial shortening systems (Wright Medical, Memphis, Tennessee), both of which employ non-self-tapping screws due to the need for multiple insertions, scenarios where more screws are required may cause concern as many orthopedic fragment systems only include STS. A study evaluating multiple reinsertions of STS would be of practical benefit as the operating surgeon may be able to use STS with these jigs should the situation warrant consideration.

Studies have established that factors involved in obtaining a maximum holding force include insertional torque, cortical depth, pilot hole size, thread diameter, and screw position relative to bone center.^{1,3,4,7,13,17,24,25,32} Yet as many factors are involved in optimizing screw holding force, there are thus a variety of errors in screw placement that can occur which reduce the screw holding force.⁸ Scenarios arise where a screw has to be removed and reinserted. Some of these scenarios are seen when initial pilot hole depth measurements are found to be too long or short, when screws require repositioning, when a plate requires repositioning, or when a rereduction of fractured fragments is necessary.

There are several studies that have evaluated the effect of multiple reinsertions on pullout strength, with contrasting results.^{9,11,21,22,31} A study by Foley et al using a porcine rib model analyzed the effect with both self-tapping and non-self-tapping screws, and reported no significant reduction in pullout strength in up to 3 reinsertions.¹¹ A separate study by Schatzker et al involving a canine femoral model with STS documented the same pushout resistance with up to 12 reinsertions into the same hole when inserted with 80% of their torque out value.³¹ Results from these studies, however, are limited by model choice as these models do not clinically or mechanically compare with human bone models. Thus, Matityahu et al studied and compared the reinsertion of STS in polyurethane synthetic bone models to human cadaveric tibias, and they demonstrated a significant decrease in pullout strength.²² Specifically, the study attempted 5 reinsertions of STS in both models and discovered that cadaveric bone exhibited a significant decrease in pullout strength after the second and third reinsertions in metaphyseal and diaphyseal bone, respectively.²² In addition, in a 2016 study, Marmor et al used different types of bone surrogates to represent normal cortical and cancellous bone and osteoporotic cortical and cancellous bone.²¹ While previous studies have found that screw pullout strength does not correlate with screw insertion torque, Marmor et al found that similar to a reduction in pullout strength, reinsertion led to a significant reduction in insertion torque.^7,21,28 They further postulate that during a reinsertion a surgeon may inadvertently increase the drill-hole size by inserting the screws in a slightly different angle than the perpendicular angle to the drill hole, suggesting that the surgeon’s experience as well as screwing technique (manual vs. power drive) may influence a reduction in insertion torque with reinsertions.²¹

As there are notable differences in the results of these studies, the goal of this study is to provide further evidence on how many times STS can be reinserted before the holding power in bone becomes clinically irrelevant, using a clinically and mechanically relevant model. Furthermore, reinsertion of STS will be completed in 3 synthetic bone models with different bone densities (normal/healthy, intermediate, and low) to evaluate the impact of reinsertions of STS in various bone densities. Our hypothesis is that at 4 insertions, the decrease in pullout strength will not be significant in the healthy bone model; furthermore, we hypothesize that it will take a fewer number of insertions for the pullout strength to become irrelevant in the synthetic bone models with lower bone densities.

Materials and Methods

Twenty-seven synthetic left radii were acquired from Pacific Research Laboratories (Vashon, Washington); the catalog number of this fourth-generation composite bone model was 3407. According to the manufacturer, the cortical regions were made of a mixture of glass fiber and epoxy resin while the cancellous regions and center canal regions were made of solid rigid polyurethane foam. The dimensions of all models can be seen in Figure 1. The material properties of these models (Table 1) have been shown in biomechanical studies to simulate the behavior of human bone.^{10,12,15,16,36} The cortical composition of these models are based on ASTM D-638 and D-695 and the specifications of material properties for the cancellous compositions were based on ASTM D-1621 and ASTM F-1839-08. The stainless steel STS were obtained from Synthes (Paoli, Pennsylvania). The STS were 50.0 mm length with a 3.5 mm diameter.

Figure 1.

Left radius.

Note. The dimensions of all models used in this study were 250.0 mm in length, 16.0 mm width at the center of the diaphysis, 35.0 mm at the base, and 27.0 mm at the head of the radius. The canal at the center of the diaphysis was 5.0 mm in diameter. All models were acquired from sawbones. This image was reproduced with permission from Pacific Research Laboratories.

Table 1.

Material Properties of Fourth-Generation Composite Bone Model No. 3407.

Composition/Model	Density (g/cc)	Longitudinal tensile strength (MPa)	Longitudinal tensile modulus (GPa)	Compressive strength (MPa)	Compressive modulus (GPa)	Transverse tensile strength (MPa)	Transverse tensile modulus (GPa)
Cortical/all models	1.64	106.0	16.0	157.0	16.7	93.0	10.0
Cancellous/Normal Bone Density Model (No. 3407)	0.27			6.0	0.155
Cancellous/Intermediate Bone Density Model (No. 3407-6)	0.22			4.3	0.108
Cancellous/Low Bone Density Model (No. 3407-5)	0.16			2.2	0.058

The radial bone midshaft was divided into 3 segments of 50.0 mm in length and a hole was drilled in the middle of each segment with a drill bit 2.5 mm in diameter in the anterior to posterior direction at approximately 90° to the surface of the bone. Screws were inserted and removed, and torque was measured with a TT03 digital torque gauge (Mark-10, Copiague, New York) during insertions and removals (Figure 2). The need to extend the flute past the cortex is of importance to maximize the bone screw interface, but there is a risk of damaging structures which lie immediate to the trans cortex. Original studies have shown that 2.0 mm past the trans cortex was necessary for maximum pullout strength.⁶ There are newer studies which show no significant benefit to inserting the screw greater than 1.0 mm past the trans cortex in healthy bone, with some literature addressing the need for 2.0 mm only for osteoporotic bone.^2,4,32,33 To limit this as a potential variable, all screws were inserted beyond 2.0 mm past the trans cortex by an orthopedic surgeon to ensure maximum fixation. Pullout strength was determined for each configuration by using the Instron 5960 materials testing system (Canton, Massachusetts) following the ASTM F543-07. For all trials, the rate of the pullout was 5.0 mm/min and a preload of 10.0 N was established to improve the repeatability of each trial. Pullout forces were recorded at a frequency of 100.0 Hz for each different bone density model and for each segment.

Figure 2.

Mechanical setup.

Note. A 50-mm portion of bone (arrow A) is rigidly fixed via a clamp (B). The bone specimen is fixed in place so that (C) the surgeon can insert or remove the screw at 90° to the bone with a vertically oriented torque sensor to accurately read the torque of the revolutions (curved arrow). Once the screw is in place and fixed to the Instron 5960 system for testing, pullout strength testing via an applied force (upward arrow) is initiated.

In this study, 26 sawbones were split into 3 groups (Table 2). The 27th sawbone, taken from the normal density group, was used to collect experimental data after 1 insertion and 1 reinsertion. During these trials, any damage or disruption to the synthetic bone model was documented. These trials were completed on each set of left radii with different bone densities, yielding a total of 81 diaphyseal trials, where a single screw was used and reused for each specific pilot hole. Statistical comparisons of torque and pullout strength across different bone densities, locations on the shaft, and number of reinsertions were made using an analysis of variance (ANOVA) with an alpha level of 0.05. Pearson correlation was used to evaluate the relationship between insertional torque and pullout strength.

Table 2.

Study Groups.

Group No.	No. of total sawbones (n = 27)	No. of normal-density sawbones (n = 9)	No. of intermediate-density sawbones (n = 9)	No. of low-density sawbones (n = 9)	No. of total insertions
0	1	1	0	0	2
1	9	3	3	3	3
2	9	3	3	3	4
3	8	2	3	3	5

Results

Torque measured in the first insertion for the middle portion had an average value of 61.3 ± 10.5 N·cm, which was significantly greater than that of both the head and base, which had average values of 51.5 ± 9.4 N·cm and 46.3 ± 7.5 N·cm, respectively (P < .001). Variation in trabecular bone densities had no influence on the insertional torque in the base, the middle, and in the head segments across all numbers of reinsertions (P > .05), yet the insertional torque remained significantly higher for the middle portion across each bone density for each number of insertions (P < .05 for 3, 4, and 5 insertions). Insertional torque for 2 reinsertions (3 total insertions) or more was not statistically significant in both the base and middle samples of the bone (P = .117 and P = .071, respectively), yet the head portion showed a significantly smaller insertional torque at 5 insertions (21.1 ± 5.6 N·cm) compared with 4 and 3 insertions (29.7 ± 6.4 N·cm and 27.3 ± 6.3 N·cm), respectively (P = .026).

The average pullout forces after a single insertion were 1623.0 N, 2318.0 N, and 1633.0 N for the base, middle, and head segments, respectively. After a single reinsertion trial, these values were reduced to 1422.0 N, 1991.0 N, and 1470.0 N. Thereafter, pullout loads following 1 reinsertion (more than 2 insertions) were not statistically significant with regard to number of insertions in all 3 segments with P values of .656, .577, and .123 for the base, middle, and head segments, respectively (Figure 3). As was the case with insertional torque, pullout strength was significantly greater in the middle segment compared with the head and base segments for each number of reinsertions (P < .05). Last, variation in trabecular bone density had no influence on the pullout strength in the base, middle, and the head segments across all numbers of reinsertions (P > .05).

Figure 3.

Pullout strength following 3, 4, and 5 insertions.

Note. Pullout loads following 2 reinsertions or more were not statistically significant with regard to number of insertions in all 3 segments (p > .05). Pullout strength was significantly greater in the middle segment compared with the head and base segments for each number of reinsertions (P < .05). Variation in trabecular bone density had no influence on the pullout strength in the base, middle, and the head segments across all numbers of reinsertions (p > .05).

There was a significant moderate to high correlation of insertional torque to pullout strength that was noted across all bone densities (low density: Pearson r = 0.718, P < .001; intermediate density: Pearson r = 0.611, P = .001; normal density: Pearson r = 0.735, P < .001) and both the head and the base segments (head: Pearson r = 0.620, P < .001; base: Pearson r = 0.543, P = .005). The middle segment showed a weak correlation, although this was not significant (Pearson r = 0.135, P = 0.531). The combined data of all segments and all bone densities across all insertions showed a high correlation between insertional torque and pullout strength (Pearson r = 0.663, P < .001).

Discussion

Screw removal and reinsertion is an event that is necessary for procedures involving a template or trialing before final implant placement, and the applicability of STS should be considered with these procedures; in addition, scenarios can arise where plate or screw repositioning, poor surgical technique, or suboptimal length will invariably lead to screw removal and reinsertion. While STS have practical applicability, there is a reasonable concern as well as experimental evidence that multiple screw reinsertions can affect pullout strength due to its path cutting design and stripping. While our pretesting data and previous studies demonstrate a statistically significant decline in pullout strength after a second insertion, this study aimed to explore the effects of multiple insertions thereafter. Our study suggests that reinsertion of STS beyond 1 time yields no significant difference in pullout strength of STS in long bone fixation for 3, 4, and 5 total insertions. The results also show differences in pullout strength between different segments of bone with the middle segment being significantly stronger than both the head and base segments. Of further clinical significance, this study also suggests no difference in insertional torque and pullout strength for 3, 4, and 5 total insertions for all 3 bone densities, revealing that reinsertions of STS screws can be safely performed without regard to bone density. Furthermore, screws inserted in different locations in the body have differences in pullout strength due to variations in density and bone architecture. Through testing of our synthetic radius model, we found that the center of the diaphyseal portion of the synthetic radius had almost a 50% greater pullout strength relative to diaphyseal bone approaching either the head of the radius or the distal radius across all insertions. Last, while there is conflicting literature on whether or not pullout strength is correlated with insertional torque, the results of our study provide further evidence of the positive correlation between these 2 variables.

Strengths of the current study include the reliability of the reported results which stem from the uniform applicability of the human synthetic bone model. When compared with human cadaver long bones, the fourth-generation composite synthetic model chosen for this study has proven to be equivalent regarding pullout strength of cortical screws, as well as overall screw purchase and shear stress.^{10,12,15,16,36} As discussed above, the synthetic radius model varies in material properties along the length of the bone, just as is the case in human bone, as indicated by the increased average pullout strength seen in the denser middle segment as a likely result of varying lengths of purchase between cortices. The uniformity of the specimens also allowed reproducible positioning for screw placement in each segment, as pilot hole location was based on plotted points from identical anatomic landmarks and manufacturing designs. Human cadaveric bone models have significant variability in shape and bone density properties, making them less reliable for comparing pullout strength. Furthermore, while some previous studies have used a custom jig to accurately reinsert a screw, our study simulated a clinical situation as each screw was reinserted by an attending orthopedic surgeon.

Limitations of the study include the use of a synthetic bone model; while the synthetic model has been shown to have similar material properties to human bone and its uniformity allows for reproducible results, the results of this study still cannot be absolutely generalized for human bone. Likewise, the results of this study were specific to our chosen screws and may not extrapolate directly to other implants that may have different designs. Furthermore, other limitations of the current study include the lack of controlling for insertional torque and screw position across the diameter of the specimens. Both of these aspects have been shown to have an effect on screw pullout forces, and this study could have benefitted from an increased sample size when considering this limitation and the multiple variables involved (bone density, number of insertions, and location on bone).^1,7 An increased sample size may have helped elucidate more findings as well. Specifically, we noticed an increase in the average difference in pullout strength between the middle portion compared with the head and base portions for each additional reinsertion (P = .03 for 3 insertions, P = .00002 for 4 insertions, and P = .000004 for 5 insertions). A similar, but not as precise trend was noted with insertional torque. Future studies could determine whether there is a relative rate of change in the biomechanical properties of more dense regions of bone compared with less dense regions with multiple reinsertions.

In conclusion, while reinsertion of STS between 1 and 2 insertions has been shown to significantly differ in regard to pullout strength, our study demonstrates that beyond this, there does not appear to be a significant difference in pullout strength in up to 5 insertions at any specific region of the radial bone and across a wide range of sawbone densities, indicating that further insertions may be considered with caution.