Pedicle Screw Loosening After Lumbar Spinal Stenosis Surgery and an Analytical Review of Recurrent Loosening: Based on Revision Surgery Experience

Introduction

Pedicle screw fixation is widely used for lumbar spinal fusion [1]. However, because these systems restrict physiological spinal motion, complications such as breakage of screws and rods and pedicle screw loosening can occur. Screw loosening, in particular, has been reported at varying rates across studies and remains a challenging complication when encountered [2]. Reported failure rates range from less than 1% to 15% in patients without osteoporosis treated with rigid systems and can reach up to 60% in patients with osteoporosis [3]

Microfractures caused by excessive loading and bone quality, as assessed by T-score values measured by dual-energy X-ray absorptiometry (DXA), are key prognostic factors for screw loosening [4,5]. Although screw loosening is more frequently reported in osteoporotic vertebrae than in non-osteoporotic vertebrae (T-score >−2.5), it remains a clinically significant problem even in patients without osteoporosis [4].

Previous studies have primarily focused on the revision of screw loosening in osteoporotic spines [6]. Various strategies have been proposed to restore fixation after loosening, including the use of expandable or cannulated screws augmented with hydroxyapatite or calcium phosphate cement [7]. In patients with osteoporosis, cement augmentation applied during either the index surgery or revision procedures has been shown to reduce the rate of screw loosening [8]. However, no standardized algorithm has been established for revision surgery in patients without osteoporosis.

Recent revision strategies for screw loosening, in addition to expandable and cannulated screws, have incorporated bone grafting as an alternative to hydroxyapatite or calcium phosphate cement. Lea et al described this technique in a series of 7 patients involving 10 screws [7].

In the present study, we adopted a multimodal revision approach in patients without osteoporosis. The screw tract was reinforced with bone grafting and cement augmentation, and the loosened screws were replaced with longer and larger-diameter screws to enhance biomechanical stability. This strategy aimed to improve implant fixation and reduce the risk of recurrent loosening. Although no patients had osteoporosis, bone mineral density was measured, comorbidities were assessed, and patients were stratified to evaluate factors associated with recurrent screw loosening after revision.

In the literature, revision strategies for pedicle screw loosening have predominantly been evaluated in osteoporotic spines, in which cement augmentation, graft-based techniques, and the use of longer and larger-diameter screws are commonly used [6]. However, data regarding the application of these methods in patients without osteoporosis remain limited [7].

To address this gap, the present study implemented a combined revision approach – distinct from previously described techniques in osteoporotic spine surgery – and applied it to cases of screw loosening in patients without osteoporosis with lumbar spinal stenosis.

In this study, we hypothesize that a low T-score and multilevel instrumentation in screw-rod constructs used for non-osteoporotic lumbar spinal stenosis surgery may be important risk factors for recurrent screw loosening.

Material and Methods

INCLUSION AND EXCLUSION CRITERIA:

Patients older than 18 years who underwent instrumentation for lumbar spinal stenosis and were subsequently reoperated due to screw loosening were included in the study. Surgical indications for lumbar spinal stenosis were determined based on the patients’ clinical presentation and radiological imaging findings. Neurogenic claudication with a walking distance of less than 100 m, together with radiological evidence on axial MRI of a central canal diameter less than 10 mm and an interpedicular distance less than 15 mm, was used as the surgical indication. When identifying patients with screw loosening, we excluded those who had undergone multiple prior spine surgeries or who had a postoperative infection, cement-augmented pedicle screws used in the index procedure, osteoporosis, or known metabolic bone disease (eg, Paget’s disease), and those who underwent instrumentation for indications other than lumbar spinal stenosis (eg, spondylolisthesis or thoracolumbar fracture).

RADIOLOGICAL ASSESSMENT:

After lumbar spinal stenosis surgery, postoperative 3-month CT scans were evaluated to radiologically assess screw loosening following the index procedure. Following lumbar spinal stenosis surgery, the presence of a radiolucent zone wider than 1 mm was considered the criterion for defining screw loosening [9,10]. CT scanning was used to identify this radiolucent zone [9]. In addition, the presence of a double-halo sign was considered a diagnostic criterion for screw loosening (Figure 1). During the assessment of screw loosening, the imaging studies were reviewed independently by 2 neurosurgeons and 1 radiologist. At the time of evaluation, the assessors did not have access to the patients’ clinical information. In cases of disagreement, a consensus decision was reached based on the conspicuity of the radiolucent zone, the presence of the double-halo sign, and the integrity of the screw-bone interface. For slices in which the criteria were deemed insufficient or equivocal, the reviewers agreed to classify the finding as no screw loosening. In addition, CT assessments were performed using bone and soft-tissue windows in accordance with standard institutional protocols. This approach was intended to enhance the consistency and reliability of the radiological evaluation. The primary endpoint of this study was defined as a short-term radiological assessment based on CT images obtained at postoperative month 3 in all patients. The presence or absence of screw loosening was evaluated on the basis of this 3-month CT examination. Data on longer-term radiological and clinical outcomes were not available in our study.

BONE MINERAL DENSITY MEASUREMENT:

Bone mineral density measurements were performed using a HOLOGIC ASY-00409 DXA scanner at our institution. Measurements were obtained from the lumbar spine (L1–L5), according to the standard protocol. The device undergoes regular daily quality control and calibration using phantom measurements in accordance with the procedures recommended by the manufacturer. The classification of bone mineral density results was based on the World Health Organization criteria, with T-score threshold values applied as follows: (1) normal is a T-score of −1.0 or higher; (2) osteopenia is a T-score between −1.0 and −2.5; and (3) osteoporosis is a T-score of −2.5 or lower.

This methodological framework enables assessment of patients’ bone quality based on standardized measurement techniques [11]. In our study, the sample was classified as patients without osteoporosis, with those having a T-score of −1 or higher considered to have high bone quality, and those with a T-score between −2.5 and −1 considered to have low bone quality.

SURGICAL TECHNIQUE:

For lumbar instrumentation revision, the screws identified as loosened and requiring revision were determined preoperatively using CT, and the patients were subsequently operated on under general anesthesia. Using the previous incision, the paraspinal muscles were bilaterally dissected (all patients had undergone Ponte osteotomy during the initial surgery). After removal of the loosened screws, the walls of the screw tract were examined with a probe (1-mm-diameter titanium alloy). Subsequently, bone grafts (allografts) were placed into the extracted screw tracts. Thereafter, semi-cured bone cement (calcium phosphate) was applied into each tract at a volume of 3 cc [12]. The extracted screws (titanium alloy, grit-blasted with aluminum oxide) were replaced with screws 5 mm longer and 1 mm wider [13]. The screws were advanced until the final full turn, at which point curing of the bone cement was awaited. Just before complete cement hardening, the screws were given a final full turn and secured in place. A polyaxial screw-rod system from the same manufacturer was used in all patients, and screws produced by the same manufacturer were also preferred for revision procedures. Screws with suspected loosening on preoperative imaging were directly inspected intraoperatively, and a lack of fusion was observed.

This surgical approach does not allow for the independent evaluation of each revision technique’s contribution. Therefore, the effect of all techniques used has been considered as a whole and applied as a bundled revision strategy.

GROUP ALLOCATION:

Patients who underwent lumbar instrumentation and subsequently required revision due to screw loosening identified at the 3-month follow-up were divided into two groups based on their outcomes. At the 3-month postoperative follow-up, patients with radiological evidence of screw stability were classified as group A, whereas those showing radiological evidence of screw loosening were classified as group B (Figure 2).

STATISTICAL ANALYSIS:

All analyses were performed using IBM SPSS version 25. Descriptive data were presented as mean, median, standard deviation, minimum, and maximum values. Chi-square analysis was used for the evaluation of categorical variables. The Shapiro-Wilk test was applied to assess the normality of data distribution. Since the data did not show a normal distribution, the Mann-Whitney U test was used for comparisons of quantitative variables between independent groups. A P value of ≤0.05 was considered statistically significant.

Results

Between January 2021 and January 2025, a total of 1063 lumbar spinal stenosis cases were operated on in our clinic, of which 542 (51%) underwent single-level instrumentation. Screw loosening occurred in 17 patients with multi-segment instrumentation. Of the 17 patients included in the study, 7 (41.2%) were male, with a mean age of 60.94±11.11 years. The mean number of screws implanted per patient was 7.41±2.09, while the mean number of loosened screws during follow-up was 2.76±1.25. In patients who underwent revision surgery, the mean number of screws that subsequently loosened was 1.65±1.90. In these 17 patients, the proportion of loosened screws after the initial surgery was 37.2% of all implanted screws, whereas the frequency of screw loosening after revision surgery was calculated as 59.8%. Among these 17 patients, no recurrent loosening was observed in 47.1% (n=8), whereas 52.9% (n=9) experienced recurrent screw loosening (Figure 3). Selected clinical characteristics of the patients are presented in Table 1. These rates are presented as descriptive frequencies derived from different patient populations and different denominators and are not intended for direct numerical comparison or for drawing conclusions regarding the superiority of one surgical approach over another. Accordingly, they should not be interpreted as comparative measures of effectiveness.

In the comparison between patients without recurrent screw loosening (group A) and those with recurrent loosening (group B), no significant differences were observed with respect to age group (≤60 vs >60 years; P=0.111), body mass index (BMI; P=0.419), bone mineral density (P=0.198), presence of diabetes (P=0.218), smoking (P=0.627), or alcohol consumption (P=0.735). The distribution of groups A and B according to selected clinical variables is presented in Table 2.

In the comparison between group A and group B, no significant differences were observed in mean age (P=0.111) or BMI (P=0.419). However, the mean T-score was significantly higher in group A (−0.70±0.93) than in group B (−1.18±0.77; P=0.021). A comparison of selected clinical variable means between the groups is presented in Table 3. Multivariate analysis was not performed in our study. The primary reason for this was the insufficient sample size, which did not allow for the achievement of adequate statistical power. This limitation represents one of the methodological constraints of our study.

Discussion

LIMITATIONS:

As a single-center study, our data may not reflect variability in surgical techniques, patient characteristics, or follow-up protocols across institutions, potentially limiting external validity. Although the incidence of screw loosening in our cohort is consistent with rates reported in the literature, the limited number of events restricts comprehensive evaluation of this complication and further constrains generalizability. Moreover, the inferences drawn from the selected endpoints and observations are specific to this cohort and should not be extrapolated. Additionally, multivariate analysis was not performed; therefore, the reported associations are unadjusted and may be influenced by residual confounding. Accordingly, these findings should be interpreted as observational associations rather than causal inferences.

The relatively short follow-up period represents an important limitation, given the potentially progressive nature of screw loosening. Loosening and the need for revision surgery often become more evident over longer follow-up periods. Extended follow-up would improve the accuracy of loosening rate estimates and allow more reliable assessment of the long-term effectiveness of surgical techniques. Accordingly, this study should be regarded as a preliminary investigation that may inform future research incorporating clinical outcome measures and extended CT follow-up. Larger studies with longer follow-up are needed to more accurately determine the true incidence of screw loosening and to enable more robust evaluation of risk factors and revision strategies.

In this study, revision for pedicle screw loosening was performed using a combined approach that included bone grafting, cement augmentation, and the use of longer and larger-diameter screws. Consequently, the independent effectiveness of each revision technique could not be assessed. The findings therefore reflect the overall short-term radiological stability achieved with the combined strategy rather than the effect of any single intervention. This represents an important methodological limitation of the study.

Conclusions

The results of this study indicate that, even in individuals without osteoporosis, bone quality and the number of instrumented segments are important determinants of pedicle screw stability. Accordingly, in patients undergoing revision for screw loosening, evaluation should consider not only the presence of osteoporosis but also T-score values and the extent of instrumentation.

However, these findings should be interpreted as exploratory observations. Given the current sample size and short-term radiological follow-up, the data do not support strong predictive conclusions. Larger studies with extended follow-up are required to further clarify these associations and enable more robust inferences.