Pull out Strength of Dual Outer Diameter Pedicle Screws Compared to Uncemented and Cemented Standard Pedicle Screws: A Biomechanical in vitro Study

Abstract

Objective

To analyze the potential of the dual outer diameter screw and systematically evaluate the pull‐out force of the dual outer diameter screw compared to the uncemented and cemented standard pedicle screws with special regard to the pedicle diameter and the vertebra level.

Methods

Sixty vertebrae of five human spines (T ₆–L ₅) were sorted into three study groups for pairwise comparison of the uncemented dual outer diameter screw, the uncemented standard screw, and the cemented standard screw, and randomized with respect to bone mineral density (BMD) and vertebra level. The vertebrae were instrumented, insertion torque was determined, and pull‐out testing was performed using a material testing machine. Failure load was evaluated in pairwise comparison within each study group. The screw‐to‐pedicle diameter ratio was determined and the uncemented dual outer diameter and standard screws were compared for different ratios as well as vertebra levels.

Results

Significantly increased pull‐out forces were measured for the cemented standard screw compared to the uncemented standard screw (+689 N, P < 0.001) and the dual outer diameter screw (+403 N, P < 0.001). Comparing the dual outer diameter screw to the uncemented standard screw in the total study group, a distinct but not significant increase was measured (+149 N, P = 0.114). Further analysis of these two screws, however, revealed a significant increase of pull‐out force for the dual outer diameter screw in the lumbar region (+247 N, P = 0.040), as well as for a screw‐to‐pedicle diameter ratio between 0.6 and 1 (+ 488 N, P = 0.028).

Conclusions

For clinical application, cement augmentation remains the gold standard for increasing screw stability. According to our results, the use of a dual outer diameter screw is an interesting option to increase screw stability in the lumbar region without cement augmentation. For the thoracic region, however, the screw‐to‐pedicle diameter should be checked and attention should be paid to screw cut out, if the dual outer diameter screw is considered.

Introduction

Failure of pedicle screw fixation due to loosening or cutting out is a major reason for revision in spine surgery, especially in the older population with osteoporotic bone. One possibility to increase the mechanical strength of the screw–bone interface is cement augmentation of the pedicle screw. Numerous biomechanical studies have demonstrated increased pull‐out strength of pedicle screws after augmentation with bone cement1, 2, 3, 4, 5, 6, 7. Furthermore, clinical long‐term studies have also proven the good functional outcome and low revision rates using cement‐augmented pedicle screws8, 9, 10, 11, 12, 13. However, cement augmentation also entails several problems, such as the risk of cement leakage and subsequent embolism or complications during the removal of the screws in case of revision. Especially with larger cement volumes used for osteoporotic bone condition, cement may leak into the spinal canal intraoperatively. In clinical studies, cement leakages rates of 26%, 39%, or even 73% have been reported, which mostly remain asymptomatic. In rare cases, however, pulmonary cement embolism or persistent thigh pain occur10, 14, 15, 16.

Due to these drawbacks of cement augmentation, many different techniques and screw designs have been developed to increase failure load of pedicle screws without the use of bone cement. A promising method is the use of expandable screws. Two different versions of these screws were presented: one screw design with an expansive screw tip and another design with an expandable anchor in the middle of the screws7, 17, 18, 19, 20, 21. Both designs have been shown to increase pull‐out force compared to the standard pedicle screw design. Compared to the cemented screws, however, controversial results were achieved7, 17, 18, 19, 20, 21. Conical screws were developed to better fit to the anatomical situation in the pedicle, which had been shown to converge in the anteroposterior direction22, 23. While a slight increase in pull‐out force for the conical screws was reported in a study on 12 porcine vertebrae, this tendency could not be confirmed in other studies on larger samples, which did not find an improvement22, 24, 25, 26. In addition, dual‐lead pedicle screws with either thin or thick crest did not show a significant effect on the pull‐out force27.

Another new method to potentially increase screws stability is the use of screws with dual outer diameter. These screws have a larger outer thread diameter and, therefore, a higher thread depth in the proximal part of the screw (Fig. 1). This potentially increases the cancellous as well as cortical purchase without the need for cement augmentation28. In a biomechanical study on artificial bones, a dual outer diameter screw with a slightly different design only achieved smaller pull‐out forces compared to many other designs with varying inner and outer diameter, including conical screws25. Another pull‐out study on human vertebrae (T₁₀–L₅) compared dual outer diameter screws with standard screws and fenestrated cemented screws, resulting in a distinct and significant increase of the cemented screws compared to all other screw types, but only a much smaller and not significant increase of the dual outer diameter screws compared to the standard screws29. This study, however, was limited by a low number of tested screws (10 per group), which also made it impossible to further analyze the results for the vertebra level. Furthermore, paired comparison of the dual outer diameter screws with the uncemented standard screws as well as the cemented screws would have been interesting.

Figure 1.

Tested pedicle screws: (A ) standard pedicle screw with a thread depth of 1 mm and (B ) dual outer diameter pedicle screw with a thread depth of 1 mm in the distal part and 1.75 mm in the proximal part.

Due to the lack of more detailed studies, the purpose of our biomechanical in vitro study was to systematically evaluate the pull‐out force of dual outer diameter screws compared to standard pedicle screws and cemented pedicle screws by means of paired comparison in individual vertebrae in randomized study groups. Furthermore, as the dual outer diameter screw potentially increases cancellous and cortical purchase, it is supposable that the relation of the outer screw diameter and pedicle diameter plays an important role. Therefore, further aims of our study were to analyze the influence of the pedicle diameter and the vertebral level on the potential increase of pull‐out force of the dual outer diameter screws compared to the standard screws. Thus, the main hypotheses were: (i) that the mean failure loads of standard pedicle screws, dual outer diameter screws, and cement augmented standard screws were equal; and (ii) that there was no effect of the pedicle diameter and the vertebral level on the pull‐out force of the dual outer diameter screws. We further analyzed the correlations between pull‐out force, bone mineral density (BMD), and insertion torque for the different screw types.

Methods

Specimens and Study Design

With the permission of the Institute of Anatomy of the University of Tübingen, 10 human spines that were fixed in ethanol and glycerol were extracted30. The study was approved by the ethical committee of the University Hospital Tübingen.

Initial CT scans were performed to evaluate the quality of the vertebrae and to exclude vertebrae with severe fracture, osteolysis, or deformation. Of the 10 spines, 5 were chosen, from which all vertebrae of the levels T₆ to L₅ could be included in the study, except for 2 thoracic vertebrae, which were fused bony. Those were replaced by 2 vertebrae of the level T₅, resulting in 60 vertebrae altogether (35 thoracic and 25 lumbar). As 1 of the 2 fused vertebrae turned out to be fine during preparation, it was included as an additional, 61st vertebra. The bone mineral density (BMD) of all vertebrae was determined using dual‐energy X‐ray absorptiometry (DEXA) (Lunar Prodigy, GE Healthcare Medical Systems, München, Germany) with two repeated observations (mean BMD, 0.86 g/cm²).

To minimize bias from variation in anatomic and implant factors, a paired study design was used, comparing two conditions at a time within the same vertebrae. The vertebrae were randomized into three study groups (two groups of 20 and one group of 21 vertebrae) according to their level and their BMD: (i) uncemented standard pedicle (USP) screw versus cemented standard pedicle (CSP) screw; (ii) uncemented standard pedicle screw versus uncemented dual outer diameter (DOD) screw; and (iii) uncemented dual outer diameter screw versus cemented standard pedicle screw.

Specimen Preparation and Instrumentation

All the vertebrae were prepared separately, dissected from soft tissues and instrumented by the same experienced surgeon. All screws were taken from the same screw system (Revere, Globus Medical, Audubon, PA, USA; Fig. 1). The screw length was adjusted according to the individual vertebra (35, 40, 45 or 50 mm), while the screw diameter was only adapted to the spinal region (thoracic or lumbar). Standard pedicle screws with an outer diameter of either 5.5 mm in the thoracic or 6.5 mm in the lumbar vertebrae were used. The dual outer diameter screws had the same screw diameter in their lower end, while their upper end was 1.5 mm larger, resulting in outer diameters of 7 mm in the thoracic and 8 mm in the lumbar regions. During screw insertion, the torque was recorded for each screw using an analog torque sensor (Torsiometer 760/30, Stahlwille, Wille GmbH, Wuppertal, Germany).

For the cementation, spinal bone cement (three spines: KyphX HV‐R, Elmdown, London, England; two spines: Verebroplastic, DePuy International, Blackpool, England) was used. In this case, the screws were removed again after initial placement and reimplanted after cement injection. According to our previous findings, 1 mL of cement was used in the thoracic and 3 mL in the lumbar vertebrae6. Prior to mechanical testing, the cement was allowed to harden at room temperature for at least 12 h. CT scans were performed to document proper screw position.

Pull‐out Testing

Mechanical pull‐out testing was performed using a material testing machine (MTS 858 Mini Bionix, Eden Prairie, MN, USA) equipped with a 25‐kN load cell6. For the fixation in the testing machine, each vertebra was embedded in polymethylmethacrylate (PMMA) resin (Technovit Universal Fluid and Powder 2060, Heraeus Kulzer GmbH, Wehrheim, Germany). As pedicle fracture had been observed during pull‐out of cemented screws, care was taken that the embedding line was below the pedicle to avoid this interfering behavior6. During the mounting of the specimens, attention was paid to ensure proper alignment of the longitudinal screw axis with the vertical pulling direction of the testing machine6. Pull‐out testing was performed in displacement controlled mode with a velocity of 10 mm/min, while data was recorded with a sample rate of 10 Hz.

Data Analysis

From the recorded force‐displacement curves, the maximum peak load (failure load) of each screw was determined. Statistical analysis was performed using the software SPSS (IBM SPSS Statistics, Version 23.0 for Windows, IBM Corporation, Armonk, New York, USA). The data were tested for normal distribution using the Shapiro–Wilk test. For comparison of failure load and insertion torque within each of the three study groups, the Wilcoxon signed‐rank test was used (paired testing). To further analyze the effect of the dual outer diameter, the smallest pedicle diameters were measured from the initial CT scans and the relation of the outer screw diameter and the pedicle diameter was calculated. Plotting the enhancement of the failure load by the dual outer diameter screw compared to the standard pedicle screw (study group B) over the sorted screw‐to‐pedicle ratio, an area of this ratio with better performance of the dual outer diameter screw could be determined. Accordingly, the vertebrae were divided into three subgroups: below, within and above this area. For each subgroup, paired comparison with the Wilcoxon signed‐rank test was performed again. The influence of the increased outer screw diameter was further tested for two separate groups of the vertebra level (lumbar and thoracic) for the whole study group using the Mann–Whitney U‐test (unpaired testing). To evaluate correlations between failure load, insertion torque and BMD, Spearman's rank correlation coefficient (ρ) was determined.

Results

Failure Load

Considering failure load, the Shapiro–Wilk test revealed that data were not normally distributed in all tested groups, yielding to the application of non‐parametric methods. The pull‐out forces found in uncemented and cemented standard pedicle screws were significantly different (USP: 713 N; CSP: 1402 N; P < 0.001; Table 1). Paired comparison of uncemented standard and uncemented dual outer‐diameter screws in group B, however, provided a distinct but not significant difference (USP: 680 N; DOD: 829 N; P = 0.114). In contrast, a significant difference in pull‐out force was found for the uncemented dual outer diameter screw compared to the cemented standard screw (DOD: 842 N; CSP: 1245 N; P < 0.001) in group C.

Table 1.

Effect of screw type and cementation on the insertion torque and the failure load during pull‐out testing of the pedicles screws from human vertebrae in the three study groups

	Group A			Group B			Group C
	Uncemented standard	Cemented standard	P value	Uncemented standard	Uncemented DOD*	P value	Uncemented DOD*	Cemented standard	P value
Failure load (N)	713 ± 356	1402 ± 767	<0.001†	680 ± 371	829 ± 383	0.114	842 ± 383	1245 ± 578	<0.001†
Insertion torque (Ncm)	119 ± 37	124 ± 39	0.706	119 ± 47	172 ± 89	0.001*	157 ± 78	122 ± 61	0.011†
Bone mineral density (g/cm2)	0.86 ± 0.19			0.86 ± 0.19			0.83 ± 020
Number	20			21			19‡

Mean ± standard deviation are reported.;

^* Dual outer diameter screw.;

^† Statistically significant (P < 0.05).;

^‡ One vertebra of this group had to be excluded due to a broken pedicle prior to testing.

Influence of Screw‐to‐pedicle Diameter and Vertebra Level

For the comparison of the dual outer diameter and the uncemented standard screw (group B), two further influence parameters, which are screw‐to‐pedicle diameter ratio and vertebra level, were analyzed. When the vertebrae of this group were sorted according to the ratio of outer screw diameter (of the DOD screw) and pedicle diameter (Fig. 2), a range of six consecutive vertebrae was detected, where with the DOD screw better results were achieved than with the standard screw. This range belonged to a screw‐to‐pedicle diameter ratio between 0.6 and 1. Therefore, data was divided into three groups according to this diameter ratio: <0.6, between 0.6 and 1.0 and >1, and paired comparison was repeated for each of these areas separately. A significantly increased failure load (P = 0.028) was found for specimens with a diameter ratio between 0.6 and 1.0, while no significant effect was determined for specimens with a diameter ratio above this range (Table 2). Regarding the failure load of the dual outer diameter screw compared to the uncemented standard screw in the thoracic and lumbar area separately, a significant increase (P = 0.04) was achieved in the lumbar area, while no significant enhancement was measured in the thoracic area (P = 0.18, Table 3).

Figure 2.

Difference of the failure load in the dual outer diameter (DOD) and the uncemented standard pedicle (USP) screws of group B plotted against the outer screw‐to‐pedicle diameter ratio. For a diameter ratio between 0.6 and 1, all DOD screws achieved higher failure loads than the respective standard screws.

Table 2.

Failure load during pull‐out testing for the uncemented standard and dual outer diameter screw (group B) for different ranges of the outer screw‐to‐pedicle diameter ratio (D)

	Failure load (N)
	D* < 0.6	0.6 < D* < 1	D* > 1
Uncemented standard	1136 ± 244	756 ± 146	575 ± 407
Uncemented DOD†	1115 ± 318	1244 ± 437	594 ± 243
P value	‡	0.028§	0.972
Number	2	6	13

Mean values and standard deviations are reported.;

^* Diameter ratio.;

^† Dual outer diameter screw.;

^‡ P‐value was not calculated because only two specimens were available.;

^§ Statistically significant (P < 0.05).

Table 3.

Failure load for the uncemented standard and dual outer diameter screws in the thoracic and lumbar region (all study groups)

	Thoracic region		Lumbar region
	Failure load (N)	Number	Failure load (N)	Number
Uncemented standard	569 ± 229	24	876 ± 435	17
Uncemented DOD*	663 ± 268	25	1123 ± 432	15
P value	0.180		0.040†

Mean value and standard deviation are reported.;

^* Dual outer diameter screw.;

^† Statistically significant (P < 0.05).

Bone Mineral Density and Insertion Torque

Bone mineral density in each group was normally distributed with similar mean values (A: 0.86 g/cm²; B: 0.86 g/cm²; C: 0.83 g/cm²) and standard deviations (Table 1). Considering the insertion torque, almost no difference was found between the cemented and uncemented standard pedicle screw (around 120 Ncm for all subgroups; Table 1). It was interesting to see, however, that a significant increase of the mean insertion torque up to 172 Ncm (group B) was found for the DOD screw compared to the cemented (P < 0.001) and the uncemented (P < 0.011) standard screws.

Correlations

In addition, the correlations between failure load, insertion torque, and BMD were evaluated (Table 4). For both failure load and insertion torque, moderate to high correlations (0.5–0.68) to BMD were achieved for all screw types. For the relation of failure load and insertion torque, moderate correlation (P = 0.66) for the cemented standard screws, high correlation (P = 0.79) for the uncemented standard screw, and particularly high correlation (P = 0.93) for the uncemented DOD screws were found.

Table 4.

Correlation between the failure load during pull‐out testing, insertion torque, and bone mineral density (BMD) for the different pedicle screws (Spearman rank correlation coefficient ρ, all values significant, P < 0.05)

	Uncemented standard	Uncemented DOD	Cemented standard
Failure load/insertion torque	0.79	0.93	0.66
Failure load/BMD	0.51	0.65	0.64
Insertion torque/BMD	0.67	0.68	0.50

Discussion

With the purpose of increasing the stability of pedicle screws in osteoporotic vertebrae without using bone cement, many different screw designs have been developed (e.g. conical screws, expandable screws, or dual lead screws). Another recently developed screw type is the dual outer diameter screw, which is evaluated in the current study. These screws have a larger outer diameter in the proximal part which shall optimize purchase for both cancellous and cortical bone28. In this study, pull‐out force and insertion torque were measured in three different study groups, randomized for BMD and vertebra level, allowing paired comparison of the new screw to uncemented and cemented standard screws.

One purpose of our study was to evaluate whether the increase in contact area due to the increased outer diameter could achieve similar effects as cement augmentation. Regarding our results for study group C, where the uncemented dual outer diameter screw was compared to the cemented standard screw, we can clearly state that this was not the case. As for the comparison of the cemented and uncemented standard screws in group A, a distinct and significant increase in pull‐out force was measured for the cemented screws, rejecting our first hypothesis. These results correspond well with many other studies comparing uncemented pedicle screws of various designs to cemented pedicle screws1, 2, 3, 4, 5, 6, 7, 29. Only for the expansive screw designs, similar pull‐out forces as for the cemented standard screw have been reported in a few studies18, 19.

The most interesting results were achieved in group B, where the dual outer diameter screw and the uncemented standard screw were compared. Regarding the entire group, a distinct but not significant difference was found. This confirms the findings of Christodoulou et al., who report similar results using exactly the same screw designs29. In contrast, a study of Amaritsakul et al., which compared screws with several variations of the inner and outer diameter, report contradictory findings25. In that study, the dual outer diameter screw achieved relatively low pull‐out forces compared to other screws. Closer examination of the screw designs tested in that study, however, reveals that compared to the other screws, those dual outer diameter screws did not have a larger outer diameter in the proximal part, but a smaller outer diameter in the distal part. Other screw designs, where a large difference of outer and inner diameter (i.e. a large thread depth) was combined with a large overall diameter, showed good results. This might indicate that the contact area of the bone and screw is essential. Similar findings have been demonstrated at various times in connection with bone cement. In cemented implants, failure almost always occurs at the cement–bone interface, and it has been shown that failure strength increases with increasing contact area of the bone and cement31, 32, 33.

As the contact area was found to be an important parameter for screw strength, this would suggest an increased pull‐out force for the dual outer diameter screw compared to the standard screw. Our results (study group B), however, did not confirm this hypothesis. For this purpose, the influence of further parameters, vertebra level and screw‐to‐pedicle diameter ratio, was analyzed. Separate evaluation of the thoracic and lumbar vertebrae revealed that for the thoracic area, the results were similar to the entire group, but in the lumbar area, increased performance of the dual outer diameter screw was achieved. This finding demonstrates that, after all, there is a potential area of application for this new screw type. As one main difference between the lumbar and thoracic vertebrae is their dimension, we also evaluated the ratio of the screw and pedicle diameters. It was interesting to see that for a ratio of 0.6–1, which means that the screw diameter was equal to 60%–100% of the pedicle diameter, the pull‐out force could be significantly increased with the dual outer diameter screw compared to the standard screw. In contrast, almost equal pull‐out forces were measured for screws with a larger outer diameter than the pedicle (screw‐to‐pedicle diameter ratio >1). In this area, for 6 of the 13 vertebrae, the pull‐out force was even reduced by the screws with the larger outer diameter. This confirms our observations during experimentation that some of the screws already cut out the pedicles before testing due to their too large diameter. These large screw‐to‐pedicle diameter ratios as well as cut outs occurred because all vertebrae of one region (thoracic, lumbar) were instrumented with screws of the same diameter and this diameter was chosen to match the standard screws. The diameter ratio, however, was calculated for the outer screw diameter of the dual outer diameter screws, which was 1.5 mm larger than that of the standard screw.

Considering this aspect, an interesting study was carried out by Misenhimer et al. 23. They inserted screws with different diameters into human vertebrae and measured the resulting pedicle diameter in CT scans. If the pedicle diameter was increased after screw insertion, they concluded that the pedicle was already damaged and the screw was too large. They came to the conclusions that the screw diameter should only have a maximum of 80% of the pedicle diameter and that the main stability of the screw comes from cancellous and not cortical purchase. These findings basically confirm our results, as we also found that there is an upper boarder for the screw diameter. The fact that in our study this boarder was located at 100% of the pedicle diameter instead of 80% can be explained by the use of dual outer diameter screws. If this upper boarder is recalculated using the outer diameter of the standard screw, this results in 79% for the thoracic and 81% for the lumbar vertebrae, which is surprisingly close to the 80% of Misenhimer et al. Our results further indicated that partial penetration of the cortical bone by the screw does not matter, as long as it is not damaged too much, and that cancellous and cortical purchases are both important for screw stability. Unfortunately, our study cannot make a statement regarding very small screw‐to‐pedicle diameter ratios, as only two vertebrae were measured with a ratio below 0.6, with similar results for the two considered screw types. Further results in this group would be interesting, however, to test the hypothesis of Misenhimer et al. that the cancellous purchase is more important than the cortical purchase.

In summary, we could identify two potential application areas for the dual outer diameter screws: vertebrae in the lumbar region as well as vertebrae with a screw‐to‐pedicle diameter ratio between 0.6 and 1. To evaluate possible intersections of these two application areas, the number of lumbar and thoracic vertebrae was evaluated for the different diameter ratios. For a screw‐to‐pedicle diameter ratio between 0.6 and 1, 5 out of 6 vertebrae were lumbar, while for a ratio >1, 10 out of 12 vertebrae were thoracic. Therefore, we could recommend the dual outer diameter screws for clinical application to increase screw stability in the lumbar region, if cement augmentation is not an option. Considering the thoracic region, the screw‐to‐pedicle diameter ratio should be checked and the dual outer diameter screw is only recommendable for a ratio <1. In general, due to the large and sharp‐edged thread of the dual outer diameter screw, attention should be paid to avoid cutting out during the application of these screws.

According to a previous study of our own and many other studies, high correlations are found between BMD, insertion torque and pull‐out force6, 34, 35, 36. This is a strong indication for the clinical use of preoperative BMD as well as intraoperative insertion torques as predictors for screw stability. The relation between insertion torque and pull‐our force, however, has to be treated carefully. As shown in our study, it is only meaningful when calculated for each screw type separately. If, for example, screws are cemented, or if conical screws are used, there are other factors that influence insertion torque and screw stability separately37. For the dual outer diameter screws, in particular, very high correlation of 0.93 between insertion torque and pull‐out force was found, indicating that the insertion torque might be a good intraoperative predictor for screw stability for these screws.

A limitation of the present study was the use of embalmed vertebra specimens, which could have changed the mechanical properties of bone and, therefore, the absolute pull‐out force. However, this should not have affected the relative comparison of the pull‐out forces for different screw types and implantation techniques during paired comparison in the individual vertebrae. Furthermore, the applied fixation technique with ethanol and glycerol does not influence the morphology and macroscopic anatomy of the bone, and it was also used in a previous study6. To evaluate the difference between the dual outer diameter and the standard pedicle screw more extensively, more vertebrae with screws of these two groups should be tested. In particular, the small number of screws with a screw‐to‐pedicle diameter ratio <0.6 can be seen as a limitation of this study, as this effect could not be evaluated. Further analysis of this group would lead to extended knowledge concerning cancellous and cortical purchase.

Conclusions

Our biomechanical in vitro study confirmed the results of previous studies concerning the predictive ability of preoperatively determined BMD and intraoperatively measured insertion torque concerning pull‐out strength, particularly in the uncemented screws.

We demonstrated that dual outer diameter screws could not achieve comparable results during pull‐out testing as cemented standard pedicle screws, which still can be considered as the gold standard for increasing screw stability. Comparison to the uncemented standard pedicle screws, however, revealed significantly increased pull‐out forces for the dual outer diameter screws in the lumbar region as well as in case of a screw‐to‐pedicle ratio between 0.6 and 1. For clinical application, the use of dual outer diameter screws is an interesting option to increase screw stability in the lumbar region without cement augmentation. For the thoracic region, however, the screw‐to‐pedicle diameter should be checked and attention should be paid to screw cut out, if the dual outer diameter screw is considered.