A Rapid and Safe Minimally Invasive Procedure for Percutaneous Pedicle Screw Removal: A Case-Control Study and Technical Description

Introduction

The incidence of thoracolumbar and lumbar fractures resulting from traumatic events is high; thus, they are a leading cause of hospitalization and disability.¹ Conservative treatment, especially bed rest, can cause deep venous thrombosis, respiratory system disease, pressure ulcers, and psychological distress.^2,3 Consequently, surgery is the most frequently used treatment, especially among relatively young patients who require adequate stability and restoration of normal spinal alignment.⁴ Non-fusion surgery with short posterior segmental fixation has reportedly shown good therapeutic effects in recent decades.^4–6 Percutaneous pedicle screw fixation can alleviate the disadvantages of the open approach, such as stripping-induced muscle denervation, increased intramuscular pressure, ischemia, back pain, and physical dysfunction.⁷ Meanwhile, the clinical efficacy and radiological results are similar to those of the conventional open technique.^8–11 Theoretically, non-fusion surgery preserves the regional mobility of the affected spine. However, all implants used in non-fusion surgeries should be removed because of the potential risk of implant breakage caused by long-term fretting wear of the non-fusion segments.¹ The advantages of implant removal after bone consolidation include early mobilization, minimal soft tissue irritation, and regional mobility preservation.⁸

Previous studies have focused on implant removal to reduce potential hazards and improve clinical outcomes.^1,12–14 Nevertheless, to our knowledge, only one cohort study with 10 cases¹⁵ and one case report¹⁶ have documented techniques for removing percutaneous pedicle screws, and no case-control research on surgical technology is available. Both studies required specialized surgical instruments, limiting their applicability. The aim of this case-control study was to introduce a rapid and safe minimally invasive method for removing percutaneous pedicle screws without the need for special surgical instruments.

Materials and Methods

Setting

This retrospective study included 92 patients who underwent percutaneous pedicle screw removal after successful posterior non-fusion fixation of traumatic thoracolumbar or lumbar fractures at our hospital between May 2016 and February 2023. The first 57 patients underwent the conventional method from May 2016 to June 2021. Subsequently, we developed a modified method through communication and discussion in our department. Thus, the last 35 patients underwent the modified method between July 2021 and February 2023.

Participants

The inclusion criteria were: (1) implant removal after non-fusion posterior percutaneous pedicle screw fixation; (2) initial fractures at the level from T10 to L5; and (3) bony consolidation of the affected vertebrae confirmed by radiological images before removal. The exclusion criteria were: (1) fractures above the level of T10; (2) initial fractures treated using open posterior fixation or other fixation techniques; (3) initial surgical complications, such as infection and neurological deficits; and (4) unavailability of complete clinical and radiological data.

This study was conducted in accordance with the principles of the Declaration of Helsinki and approved by the ethics committee of Yantaishan Hospital (approval no. 2023105). Written informed consent was obtained from each patient, or their authorized signatory, whose images or clinical data were used.

Variables and Data Measurement

All patients who had undergone non-fusion percutaneous posterior fixation for traumatic thoracolumbar or lumbar fractures required follow-up assessments using X-rays every 3 months. Computed tomography scanning was recommended if the healing status of the fractured vertebrae was uncertain. When vertebral healing was observed, implant removal was suggested.¹⁷ Patient demographics (age, sex, body mass index, alcohol use, and current smoking status), clinical (hypertension and diabetes mellitus), surgical (affected levels, number of screws, time of surgery, and blood loss), and treatment cost characteristics were extracted from medical records and analyzed. The number of screws was varied between four and six (not fixed); thus, we compared the average time (surgical time/number of screws) and average blood loss volume (blood loss volume/number of screws) instead of the absolute surgical time and blood loss volume to account for individual differences.

Surgical Technique

All operations were performed by experienced and qualified surgeons. While screw systems varied, the methods for hardware removal remained consistent.¹⁵ All procedures were performed under general anesthesia with the patient in the prone position. For conventional removal surgery, we cut through the soft tissue, explored the hardware, and removed the plugs, rods, and screws individually. In the modified removal surgery, the previous skin incision was re-opened, cutting through all layers until the scalpel reached the end of the screw and plug. The assistant then widened the incision using a vessel clamp, allowing the operator to insert a system-specific set screw-retaining compressor into the plug hole and drive it out. The same process was repeated to remove other plugs on the ipsilateral side. After removing all the plugs on one side, the rod was removed using the push-and-pull method. The rod was pushed into the cephalic or caudal soft tissue out of the screw saddle of the last screw and then pulled out from the closest incision. Next, we used the blunt end of a Kirschner wire (diameter: 1.5 mm) to slide at the bottom of the polyaxial screw saddle. The aim was to blindly probe the hole at the bottom. When the Kirschner wire contacted with the edge of the hole, it would smoothly slide into the axis channel of the screw. Then, the system-specific screwdriver could easily be inserted into the screw saddle using the guide of the Kirschner wire, and both the screw and the Kirschner wire could be removed together. Afterwards, plugs, rods, and polyaxial screws on the contralateral side were removed using the same methods. Finally, the incisions were closed after removing all the screws. A typical case is illustrated in Figure 1.

Figure 1 Typical case. Male patient aged 25 years with a history of L3 fracture caused by a car accident 1 year before. (a) Anteroposterior X-ray film before hardware removal. (b) Lateral X-ray film before hardware removal showing bone healing. (c) Incision distraction by the vessel clamp and system-specific set screw retaining compressor insertion. (d) Kirschner wire insertion in a blind manner with the blunt end ahead. (e) Lateral X-ray after Kirschner wire insertion into the axial channel of a percutaneous pedicle screw. (f) Status of the screw with Kirschner wire after removal. A sharp end can be observed on the opposite side of the screw head, indicating that the blunt end was in the screw.

Statistical Methods

Continuous variables are expressed as mean ± standard deviation, and categorical variables are expressed as frequencies or percentages. Independent-sample t-tests and chi-square tests were performed to analyze differences in continuous and categorical variables, respectively, between groups. SPSS (version 25.0, IBM SPSS, Armonk, NY, USA) was used for statistical analyses. A P-value of <0.05 was considered statistically significant.

Results

The conventional group comprised 35 men and 22 women aged 19–69 years. The modified group included 25 men and 10 women aged 25–67 years.

Table 1 presents the results of univariate analysis of demographic, clinical, surgical, and total cost characteristics. No significant differences were observed in the demographic or clinical characteristics, or affected levels. However, the average surgical time (P=0.000) was significantly shorter, while the average blood loss volume (P=0.002) and total cost (P=0.000) were significantly lower in the modified group.

Table 1 Comparison of Demographic, Clinical, and Surgical Characteristics and the Total Treatment Cost Between the Two Groups

Discussion

Our study revealed that, compared with the conventional method, our modified method can shorten surgical time, reduce blood loss volume, and reduce total treatment cost, and it is a rapid and safe minimally invasive procedure for percutaneous pedicle screw removal.

The decision to remove implants after non-fusion fixation remains controversial.¹ Considering the drawbacks of preserving implants, such as local pain, the potential risk of implant breakage, dysfunction, damage to the spine caused by the implants,^18–20 and rare complications such as hypersensitivity reactions,²¹ most clinicians consider implant removal after non-fusion surgery beneficial.¹ Cappuccio et al² considered removal mandatory when fixation involved L2 or lower segments. Finite element analysis also recommended the removal of fixation to mitigate adjacent segment stress.²² Jeon et al²³ followed a cohort of patients over 2 years and reported that back pain and physical dysfunction were significantly lower in the implant removal group than in the control group. Only a few authors have argued against the removal of implants, even if screw breakage may occur.¹² Secondary kyphosis,^24–27 recurrent instability,²⁸ and fixed vertebral body compression fractures^2,29 were major concerns causing some authors to oppose implant removal. However, complications are rare. In our hospital, >50% of patients who received non-fusion percutaneous pedicle screw fixation complained of back pain after surgery, regardless of whether the pain was in the operative area or the region below. The exact mechanism of implant-related pain remains unclear but may involve local microenvironments, metallic properties, surgical site infections, or allergic reactions.^30,31 Complete remission of back pain is infrequent;³² however, back pain caused by non-fusion fixation should be relieved after implant removal, consequently resulting in improved quality of life.¹ Thus, all patients who received non-fusion percutaneous pedicle screw fixation were recommended for hardware removal after bone healing at our hospital.

Implant removal surgery has been widely used for non-fusion in recent years; however, most studies have reported that the range of segmental motion can be regained after implant removal.^1,6,18,23 Ko et al³³ reported that the regained range of motion was 9.12° based on a mean removal time of 12.2 months. Jeon et al²³ discovered that the mean regain in the segmental motion angle was 4.2° based on a mean removal time of 18.3 months. Wu et al¹ concluded that patients could gain a better segmental range of motion within 2 years; however, the range of motion in those who underwent hardware removal after 2 years was significantly decreased. This decrease might be caused by spontaneous stiffness of the facet joints due to long-term fixed segment immobilization. Oh et al⁸ also reported that early implant removal within 1 year after percutaneous pedicle screw fixation resulted in better range of motion than did implant removal after 1 year. Axelsson et al¹⁴ reported that segmental mobility can be restored even after a minimum of 1.5 years after hardware removal. In our practice, when patients’ radiographic images reveal fractured vertebra consolidation, we recommend removing the hardware as early as possible. However, due to variations in patients’ schedules and intentions, they often choose a timing that suits their convenience. In our research, we did not include the hardware removal time and fracture type at primary trauma. These variables might be more closely related to the fracture prognosis rather than our study aim, which centered on the technique.

To our knowledge, no case-control study has reported minimally invasive techniques for removing percutaneous pedicle screws. In 2004, Salerni¹⁵ reported on a cohort of 10 patients who had undergone microremoval of percutaneous pedicle screws using a tubular retractor, loupe magnification, and headlight. A case report by Grasso et al¹⁶ in 2020 detailed a method involving the use of a lumbar Caspar self-retractor and a surgical microscopic view. In contrast, our method does not require specialized instruments. To our knowledge, the Caspar retractor, tubular retractor, loupe magnification, headlight, and microscope are unavailable in most spinal departments in China and, possibly, other developing countries, limiting their use. Rather than Caspar and tubular retractors, our approach involves incision distraction using a vessel clamp, effectively creating space similar to that achieved with retractors. Even if the plug accidentally falls from the driver, it can be easily retrieved from the retracted space. However, the vessel clamp should not be loosened until the plug has been removed from the incision. To remove the polyaxial screw, Salerni¹⁵ inserted a spinal needle directly into the axis of the screw to guide the direction of the screwdriver. The author stated that because the head of the polyaxial screw was locked to the rod, it was typically inconsistent with the screw axis. Consequently, a C-arm fluoroscope might be necessary, and an additional skin incision might be chosen as the entry point for the needle between the two extended lines of the two axes. Then, the removal of the locking devices could be performed with greater precision.¹⁵ Our method involves the use of a 1.5-mm-diameter Kirschner wire instead of a spinal needle, offering certain advantages. Firstly, it is readily available in various diameters as it is commonly used in all subspecialties of orthopedics. Secondly, the diameter of the hollow shaft of the main percutaneous pedicle screw is 1.6 mm, and the diameter of the guide pin used in the primary operation is usually 1.5 mm. Thus, the 1.5 mm diameter Kirschner wire can effectively guide the screwdriver, similar to the guide pin, at its maximum. In cases whereby inserting a 1.5 mm Kirschner wire is challenging, the operator can use one with a 1.2 mm diameter instead. Furthermore, the Kirschner wire has one sharp and another blunt end. The operator can use the blunt end to probe the hole, which is safer than using the sharp spinal needle, which might damage the soft tissue or even a large vessel,³⁴ if inserted outside the vertebrae, or if the sharp end breaches the anterior wall of the vertebrae during insertion or screw-driving. The C-arm was not needed in our practice because, regardless of the direction of the screw saddle, it could easily be inserted into the end of the screw using the Kirschner wire guide.

Our results revealed that our modified method can shorten the surgical time, reduce blood loss, and reduce the total cost compared with the conventional method. The conventional method explores the hardware extensively, which wastes time, especially in cases where the fractured vertebrae are located at L3 and below. This region is the deepest due to lumbar lordosis, or when patients have excess fat or muscular tissue. The biggest contributing factor to prolonged surgical time is the difficulty in easily placing the screwdriver at the end of the screw without guidance. This issue arises because the polyaxial screw head is locked to the rod and typically inconsistent with the screw axis; this also increases incision bleeding volume.¹⁵ However, in the modified method, all the procedures may be performed by probing and feeling. Exploring the hardware is almost unnecessary, which can save time and consequently reduce blood loss volume. One additional Kirschner wire is required to perform the modified procedure, incurring an additional cost of 20–30 Chinese Yuan in our hospital. However, this cost is easily offset by the time saved during surgery. Shorter anesthesia duration and reduced anesthesia drug usage further contribute to overall cost savings.

Our study had several limitations. Firstly, it was a retrospective case-control study based on a review of medical records. This introduces a potential risk of bias due to misinterpretation of the descriptions, and the overall evidence level was relatively low. Secondly, the study included a small number of patients and was conducted at a single spine center, which may have reduced the confidence level of the study. A large-scale multicenter study should be conducted to further verify our conclusions. In addition, the grouping of patients was based on the date of the hardware removal surgery, as we did not previously propose a modified technique. The growth and experience of the operators may have impacted the results, leading to potential bias.

Conclusion

The modified technique can shorten the surgical time and reduce blood loss volume and the total treatment cost compared to the conventional method. We developed the technique based on our clinical experience; the learning curve for this procedure appears to be relatively flat, and it does not require additional surgical instruments. This technique is suitable for utilization in primary hospitals. Further large-scale multicenter studies are needed to validate these findings.