Predictors of Persistent Postoperative Numbness Following Lumbar Fusion in Patients Older Than 75 Years: A Minimum 2-Year Follow-up
Article information
Abstract
Objective
To evaluate the preoperative and perioperative predictors of persistent leg numbness following lumbar fusion in patients aged ≥ 75 years.
Methods
This single-center retrospective study examined 304 patients aged ≥ 75 years who underwent lumbar fusion for lumbar degenerative disease (102 men, 202 women; mean age, 79.2 [75–90] years). The visual analogue scale (VAS) score for leg numbness was examined preoperatively and at 2 years postoperatively. The persistent leg numbness group included patients with a 2-year postoperative VAS score for leg numbness ≥ 5 points. The demographic data were also reviewed. A multivariate stepwise logistic regression analysis was performed for variables with univariate analysis values of p < 0.2 on univariate analysis.
Results
In total, 71 patients (23.4%) experienced persistent postoperative leg numbness. Multivariate logistic regression analysis revealed that a history of lumbar decompression, longer symptom duration, and a preoperative VAS score for leg numbness ≥ 5 points were associated with greater postoperative persistent leg numbness following lumbar fusion. In contrast, other factors, such as sex, body mass index, vertebral fracture, diabetes mellitus, depression, symptom duration, dural injury, operative time, and estimated blood loss, were not.
Conclusion
A history of preoperative lumbar decompression, longer symptom duration, and greater preoperative VAS scores for leg numbness were preoperative predictors of persistent postoperative leg numbness following lumbar fusion in older patients. Although lumbar fusion is expected to improve leg numbness, surgeons should consider the surgical history, duration, and preoperative numbness intensity and explain the potential postoperative persistent leg numbness in advance.
INTRODUCTION
Most countries face serious healthcare issues, such as increased life and health expectancy and population aging. This demographic shift will continue and by 2050 approximately half of the world’s population will live in aging countries [1]. In particular, Japan has become a super-aged society, with 28.4% of the population aged ≥ 65 years, and the mean life expectancy for Japanese women in 2019 was 87.5 years [2]. Conversely, the Healthy Life Years of Japanese women was 75.4 years in 2019 [2]. Thus, a 12-year time lag exists between Healthy Life Years and the mean life expectancy. To close this time lag and extend Healthy Life Years, it is important to focus on the health of patients aged ≥ 75 years. Concurrently, in the future, surgical treatments, such as lumbar fusion, will become the treatment of choice for older patients who require treatment for lumbar degenerative diseases to improve their quality of life [3-5].
With an increasing focus on patient health and satisfaction, spine surgeons are increasingly expected to reduce postoperative leg numbness [6]. Moreover, 3 previous reports have described the predictors of increased postoperative persistent leg numbness following lumbar surgery. These reports have demonstrated that increased postoperative persistent leg numbness following lumbar decompression surgery may be related to preoperative leg numbness intensity, diabetes mellitus, dural injury, preoperative symptom duration, and preoperative severity of spinal canal stenosis [7-9]. However, these previous studies examined patients of all ages. To the best of our knowledge, no studies have focused on postoperative persistent leg numbness following lumbar fusion in older adults despite its importance in an aging society. Therefore, in this study, we aimed to evaluate the specific demographic, preoperative, and perioperative predictors associated with greater postoperative persistent leg numbness following lumbar fusion in patients aged ≥ 75 years.
MATERIALS AND METHODS
1. Patients
The study was approved by the Institutional Review Board of Hakodate Central General Hospital (No. 2023-9). We conducted this study according to the 1964 Helsinki Declaration and its later amendments. Informed consent was obtained from all individual participants included in the study.
We retrospectively reviewed the medical records of patients aged ≥ 75 years who underwent posterior lumbar interbody fusion (PLIF) or transforaminal lumbar interbody fusion (TLIF) for lumbar degenerative diseases at our hospital between March 2013 and October 2019. The exclusion criteria were as follows: (1) follow-up period < 2 years; (2) additional lumbar fusion for adjacent segment degeneration; (3) surgery for spinal injury, infection, or tumor; and (4) postoperative infection. Neurologists were consulted in advance if a neurologic disease such as polyneuropathy was suspected. In our hospital, preoperative and postoperative epidural nerve block were not performed. In total, 304 patients met the inclusion criteria. The 2-year follow-up rate was 80.9% (304 of 376 patients). Of the 72 dropouts, 47 did not return to the doctor, 14 required revision surgery, 7 had postoperative infections, and 4 died of systemic diseases.
2. Operative Procedures
The operative procedures were performed in an open setting. We performed PLIF/TLIF in patients with spondylolisthesis or neuroforaminal stenosis, including far-lateral disc herniation. Patients with spondylolisthesis requiring bilateral decompression underwent PLIF, whereas those with neuroforaminal stenosis or spondylolisthesis requiring unilateral decompression underwent TLIF. Pedicle screws were positioned correctly. The upper articular processes, lower articular processes, and lamina were removed to expand the central spinal canal. After removing the nucleus pulposus, the endplate cartilage was scraped using a curette and a suitable polyetheretherketone cage was inserted into the intervertebral space with an autogenous bone graft. Two box-type cages were used for PLIF and a banana-shaped cage was used for TLIF.
3. Data Collection
Patient demographics and operative factors were reviewed. These variables included age, sex, body mass index (BMI), type of surgery, pathogenesis, alcohol, smoking, presence of preexisting vertebral fractures, central spinal canal stenosis, history of lumbar spine surgery, diabetes mellitus, depression, dural injury, symptom duration, serum calcium level, opioids use, anticonvulsants (pregabalin or mirogabalin) use, number of fusion segments, operation time, and estimated blood loss (EBL). To determine the appropriate cutoff value, we have performed receiver operating characteristic curve analysis in the univariate analysis (Supplementary Fig. 1). Slimness and obesity were defined as BMI < 20 and ≥ 30 kg/m2, respectively. Central spinal canal stenosis was defined as Schizae grade C or D [10]. History of previous lumbar spine surgery was defined as previous decompression, including discectomy, in the fusion range. Depression was confirmed from the medical records. We evaluated the visual analogue scale (VAS) scores for low back pain (LBP), leg pain, and leg numbness preoperatively, and at 1 and 2 years postoperatively. Improvement was evaluated using minimum clinically important differences (MCIDs). We set MCIDs as a 2-point improvement of the VASs for LBP, leg pain, and numbness based on a previous study [11]. Since the MCID score is 2, patients with preoperative VAS less than 2 were excluded from analysis. Spinopelvic sagittal parameters including sagittal vertical axis (SVA), lumbar lordosis (LL), pelvic tilt (PT), and pelvic incidence (PI) were measured using standing radiographs of the whole spine preoperatively and 2 years postoperatively.
4. Statistical Analysis
As a scale for leg numbness intensity evaluation, the presence of postoperative persistent leg numbness was defined as a VAS score for leg numbness ≥ 5 points at the 2-year postoperative follow-up evaluation [9]. Patients with a VAS score for leg numbness ≥ 5 and < 5 points were categorized into the persistent numbness (PN) and non-PN (N-PN) groups, respectively. All statistical analyses were performed using JMP Pro version 16.0 statistical software (SAS Institute, Cary, NC, USA). Fisher exact probability test was performed for the univariate analysis to compare categorical variables between the PN and N-PN groups. Multivariate stepwise logistic regression analysis was performed to investigate the predictors of postoperative persistent leg numbness after lumbar fusion using variables with p < 0.2 in the univariate analysis. Statistical significance was set at p < 0.05.
RESULTS
In total, 304 patients who underwent primary PLIF/TLIF were included in the analysis. Table 1 shows the patients’ demographic characteristics and clinical outcomes. This study included 102 male and 202 female patients with a mean age of 79.2 ± 3.5 (75–90) years at the time of surgery. The mean BMI was 24.7 ± 11.8 kg/m2, and the mean follow-up period was 46.0 ± 18.9 months. Pathologies included lumbar spondylolisthesis in 176 patients (57.9%), lumbar spinal stenosis (LSS) with foraminal stenosis in 122 patients (40.1%), and lumbar disc herniation in 6 patients (2.0%). Surgery included PLIF in 149 patients (49.0%) and TLIF in 155 patients (51.0%). The number of fusion segments was one in 194 cases (63.8%), 2 in 89 cases (29.3%), 3 in 18 cases (5.9%), and 4 in three cases (1.0%). Additionally, we have investigated a subgroup analysis based on the number of fusion segments and the result was shown in Supplementary Table 1. The mean symptom duration was 24.6 ± 44.6 months. The mean operation time and EBL were 202.3 ± 64.2 minutes and 246.0 ± 213.9 mL, respectively. The mean preoperative VAS score for LBP, leg pain, and leg numbness were 5.1 ± 3.2, 6.5 ± 3.3, and 5.4 ± 3.5 points, respectively. The mean operation times of PLIF and TLIF were 205.0 ± 67.1 minutes and 199.7 ± 61.3 minutes, respectively. There was no significant difference between the operation time of 2 techniques (p = 0.473). A total of 71 patients (23.4%) were categorized into the PN group. In the N-PN group, 3 patients (1.3%) had worsening numbness from the preoperative period; in the PN group, 24 patients (33.8%) had worsening numbness (p < 0.001). There was no significant difference in pathogenesis between PN and N-PN groups (Supplementary Table 2) (p = 0.244).
Characteristics of patient demographics
The changes in the VAS scores for LBP, leg pain, and leg numbness are presented in Fig. 1. The preoperative VAS scores for LBP and leg numbness were significantly higher in the PN than in the N-PN group, whereas there was no significant difference in the VAS scores for leg pain between the 2 groups. The postoperative VAS scores for LBP, leg pain, and leg numbness were significantly higher in the PN than in the N-PN group at 1 and 2 years postoperatively. Reaching MCIDs between PN and N-PN groups were shown in Supplementary Table 3. Rate of reaching MCIDs of LBP, leg pain, and leg numbness in PN group were 74.2%, 70.5%, and 36.5%, respectively, and those in N-PN group were 80.6%, 89.3%, and 93.5%, respectively. Although there was no significant difference in rate of reaching MCID of LBP (p = 0.289), those of leg pain and leg numbness were significantly lower in PN group than N-PN group (p < 0.001 each).
Comparison of the change of visual analogue scales (VAS) between the persistent numbness (PN) and non-PN (N-PN) groups. (A) VAS for low back pain. (B) VAS for leg pain. (C) VAS for leg numbness. Preop, preoperative; PO, postoperative. *p < 0.05, **p < 0.001.
Preoperative opioid use was 19 patients in PN group (24.4%) and 52 patients in N-PN group (23.0%), and preoperative anticonvulsants use was 37 patients in PN group (52.1%) and 118 patients in N-PN group (50.6%). There were no significant differences in preoperative use of opioid and anticonvulsants between the 2 groups (opioid, p = 0.808; anticonvulsants, p = 0.828). Postoperative opioid use was 9 patients in PN group (12.7%) and 39 patients in N-PN group (16.7%), and postoperative anticonvulsants use was 23 patients in PN group (32.4%) and 18 patients in N-PN group (7.7%). Although there was no significant difference in postoperative opioid use between the 2 groups (p = 0.461), the rate of postoperative anticonvulsants use was significantly higher in PN group than N-PN group (p < 0.001).
The majority of patients in our study sample were aged ≥ 78 years (56.6%), women (66.4%), had no slimness (87.8%), not obese (94.7%), no alcohol (89.5%), no smoking (83.2%), and had a symptom duration of < 16 months (64.5%). The majority of patients underwent TLIF (51.0%). Regarding baseline comorbidities, most patients had central spinal canal stenosis (75.0%), no foraminal stenosis (51.3%), no existing vertebral fractures (69.7%), no history of lumbar decompression (84.2%), no diabetes mellitus (79.6%), no depression (94.1%), serum calcium level ≥ 9.0 mg/dL (80.3%), and no dural injury (94.4%). Regarding intraoperative factors, most patients had < 3 fusion segments (92.8%), operative time < 254 minutes (78.9%), and EBL < 220 mL (58.2%). Most patients had a preoperative VAS score for LBP ≥ 5 points (58.2%), leg pain (74.0%), and leg numbness ≥ 5 (62.2%). The patient demographics are presented in Table 2.
Patient population
Univariate analyses were performed on the PN and N-PN groups to verify the effects of each preoperative and intraoperative factor on persistent leg numbness following lumbar fusion in older patients. Table 3 presents the results of the analysis. A history of previous lumbar decompression (p < 0.001), serum calcium level < 9.0 mg/dL (p = 0.175), symptom duration ≥ 16 months (p = 0.014), EBL ≥ 220 mL (p = 0.040), preoperative VAS score for LBP ≥ 5 points (p = 0.073), and preoperative VAS score for leg numbness ≥ 5 points (p < 0.001) were detected as variables with p-values < 0.2. However, other factors, such as age, sex, slimness, obesity, type of surgery, alcohol, smoking, central spinal canal stenosis, vertebral fracture, diabetes mellitus, depression, dural injury, symptom duration, number of fusion segments, operation time, and preoperative VAS score for leg pain ≥ 5 points, were not associated with postoperative persistent leg numbness following lumbar fusion (p ≥ 0.2 each).
Univariate analysis between PN and N-PN groups
The results of comparison of pre- and postoperative spinopelvic sagittal parameters between PN and N-PN groups were shown in Supplementary Table 4. There were no significant differences in spinopelvic sagittal parameters including pre- and postoperative SVA, LL, PI, PT, and PI-LL between the PN and N-PN groups.
Predictors of an increased risk of postoperative persistent leg numbness in older patients were determined using multivariate stepwise logistic regression analysis. Table 4 presents the results of this analysis. Two of the 4 factors detected as variables (p < 0.2) in the univariate analysis were significant in the multivariate analysis. A history of previous lumbar decompression (odds ratio [OR], 3.89; p < 0.001), symptom duration ≥ 16 months (OR, 2.11; p = 0.013), and a preoperative VAS score for leg numbness ≥ 5 points (OR, 3.63; p < 0.001) were associated with greater postoperative persistent leg numbness in older patients.
Independent risk factors for postoperative persistent leg numbness in the elderly patients
DISCUSSION
Managing leg numbness after lumbar fusion is important for postoperative patient care. A previous study showed that leg numbness had a greater impact than that of leg/back pain on patient satisfaction in patients who underwent lumbar decompression [6,12]. Indeed, in clinical settings, we frequently encounter patients who complain that leg pain has disappeared, but numbness is still persistent. Given that lower extremity and trunk muscle strengths in older adults decrease with age [13-16], management of postoperative leg symptoms becomes more important for maintaining activities of daily living. Therefore, the predictors of persistent leg numbness following lumbar surgery in older patients are of great interest. Several studies have investigated the predictors of postoperative persistent leg numbness [7,9]. However, previous studies have investigated the predictors of postoperative persistent leg numbness in all generations. Furthermore, all previous studies have investigated decompression surgery, and no study has focused on persistent leg numbness after lumbar fusion. Thus, the predictors of postoperative persistent leg numbness after lumbar fusion in older patients remain unclear. Here, we first investigated the predictors of postoperative persistent leg numbness following lumbar fusion in patients aged ≥ 75 years and found that a history of previous lumbar decompression, symptom duration ≥ 16 months, and preoperative VAS score for leg numbness ≥ 5 points were predictors of persistent leg numbness.
Prolonged recovery of leg numbness following lumbar spine surgery has been known [8,17], and alleviation of leg numbness by surgery for LSS has been reported to be more difficult than other neurological symptoms, such as muscle weakness or pain [18,19]. Huang and Sengupta [17] showed that although leg pain dramatically improved within 6 weeks following lumbar decompression, leg numbness improved at a much slower speed than pain and gradually improved from 3 months to 1 year of follow-up. Furthermore, Oba et al. [8] reported that postoperative leg numbness following lumbar decompression improved until 2 weeks postoperatively, after which it reached a plateau and showed no improvement. Additionally, at 1 year postoperatively, leg numbness remained significantly higher than leg pain. These previous studies have shown that leg numbness following lumbar spine surgery is associated with poorer recovery than leg pain.
We defined PN in cases of VAS score ≥5 points, which is obviously of significant intensity, according to a previous study, because mild residual numbness may not significantly impact postoperative patient satisfaction [9]. As a result, in the current study, the incidence of postoperative persistent leg numbness was 23.4%. Previous studies have reported that the incidence of postoperative PN following lumbar surgery was 15.5%–74.1% [7-9]. This result suggested that the incidence of postoperative leg numbness following lumbar fusion in older adults is equal to that following lumbar decompression. The results of this study are consistent with those of previous studies. This study suggested that a history of lumbar decompression is a predictor of persistent postoperative leg numbness after lumbar fusion in older patients. The impact of previous lumbar decompression on a range of persistent postoperative symptoms, including back pain, leg pain, and leg numbness, is known [20-22]. Postoperative paravertebral tissue fibrosis and disc degeneration, including recurrent disc herniation, retained disc fragment, and internal disc disruption, are caused by lumbar decompression [23-25]. Therefore, persistent postoperative symptoms, including leg numbness, may be caused by multiple lumbar surgeries that affect the progression of lumbar degeneration. Older patients often have a history of surgery; thus, surgeons planning lumbar fusion should consider the impact of previous lumbar decompression on persistent postoperative leg numbness.
The use of the VAS or Numerical Rating Scale (NRS) score for leg numbness is standard in previous studies analyzing leg numbness after lumbar spine surgery [6,8,9]. Oba et al. evaluated recovery from leg numbness following decompression surgery for LSS with VAS for leg numbness [8]. Furthermore, Ogura et al. [6,9] used NRS for leg numbness to assess patients’ satisfaction and risk factors following decompression surgery for LSS. Therefore, we believe that the research method of the current study, which is using VAS for leg numbness, is appropriate.
In the present study, the patients with persistent postoperative numbness had higher rates of postoperative LBP and leg pain. Furthermore, reaching MCIDs of leg pain, and leg numbness were associated with postoperative PN. Persistent back pain, leg pain, and numbness after lumbar spine surgery are known as failed back surgery syndrome [20-22,26,27]. Moreover, leg pain alone can be related to the sacroiliac joint, which is not an uncommon presentation after spinal fusion [28]. Thus, in this study, patients in the PN group may have failed back surgery syndrome or sacroiliitis after spinal surgery.
Several previous studies have also explored whether the symptom duration of preoperative leg numbness is related to postoperative PN [8,29,30]. Their study investigated the relationship between postoperative persistent leg numbness following lumbar decompression, spinal endoscopic surgery, or postoperative lumbar epidural hematoma, resulting in showing that postoperative PN was associated with longer symptom duration. Here, the current study has demonstrated a longer preoperative symptom duration is a risk factor of persistent leg numbness following lumbar fusion in the elderly patients. A longer duration of preoperative leg numbness is associated with a longer compression of the nerve root, which may lead to irreversible damage to the nerve root and the accompanying PN [31].
The current study suggested that a preoperative VAS score for leg numbness ≥ 5 points can predict persistent postoperative leg numbness. Similarly, previous studies reported that greater preoperative leg numbness was a risk factor for postoperative persistent leg numbness following lumbar surgery [7,9]. Hara et al. [7] reported that 65% of patients with severe preoperative numbness still showed residual leg numbness at 2 years postoperatively. Furthermore, Ogura et al. [9] reported that patients with postoperative persistent leg numbness following lumbar decompression had greater preoperative leg numbness than patients with no PN. Altogether, the findings of previous studies were consistent with those of our work in that greater preoperative leg numbness was a risk factor for postoperative PN following lumbar surgery.
The current study had 2 strengths. First, we investigated postoperative persistent leg numbness in patients who underwent lumbar fusion 2 years postoperatively with a sufficient sample size, even in this older adult cohort. Second, to our knowledge, this is the first study to focus on predictors of PN following lumbar fusion in older patients. When a surgeon plans to perform lumbar fusion in an older patient, the results of the current study may help explain the risk of persistent postoperative leg numbness and contribute to the choice of treatment.
However, the study had several limitations. First, it had a retrospective design. Therefore, if some older patients who were pathologically eligible for lumbar fusion could not undergo surgery because of systemic problems, a selection bias could exist. Moreover, because the symptoms could not be determined from medical records alone, which often lacked detailed patient complaints, we could not classify the patient’s leg numbness as cauda equina syndrome or radiculopathy. Second, we used the VAS score, a patient-reported outcome measure, to evaluate persistent leg numbness after lumbar fusion. Generally, as patients may use “numbness” to express various symptoms, such as paresthesia, dysesthesia, or loss of sensation, patient-reported leg numbness might include a variety of sensory disturbances [8]. Furthermore, elderly patients may have bias in description of their symptoms, numbness in particular. Distinguishing the difference between a type of sensory symptoms was difficult in the current study. Third, this was a single-center study. Additionally, in our hospital, we do not routinely perform electromyography, cervical spine magnetic resonance imaging (MRI), or thoracic spine MRI for patients with lumbar disease. Therefore, we were not able to evaluate polyneuropathy or spinal cord compression in this study. Despite these limitations, the current study identified 2 independent predictors of persistent postoperative leg numbness following lumbar fusion in older patients. Additional information on our cohort from further prospective long-term follow-up studies at multiple facilities is required to make this study more meaningful.
CONCLUSION
This work showed that a history of preoperative lumbar decompression, longer symptom duration, and a greater preoperative VAS score for leg numbness were independent predictors of persistent postoperative leg numbness following lumbar fusion in older patients. When surgeons plan to perform lumbar fusion in older patients, they should pay attention to the patient’s surgical history, symptom duration, and preoperative numbness intensity, and explain in advance the potential for postoperative persistent leg numbness in patients who expect recovery of leg numbness following lumbar fusion.
Supplementary Materials
Supplementary Tables 1-4 and Fig. 1 can be found via https://doi.org/10.14245/ns.2347312.656.
A subgroup analysis based on the number of fusion segments
A subgroup analysis based on pathogenesis
Reaching MCIDs between PN and N-PN groups
Comparison of pre- and postoperative spinopelvic sagittal parameters between PN and N-PN groups
Receiver operating characteristic curves for age (A), symptom duration (B), estimated blood loss (C), operation time (D), and serum calcium level (E). AUC, area under the curve.
Notes
Conflict of Interest
The authors have nothing to disclose.
Funding/Support
This study received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Author Contribution
Conceptualization: TT; Data curation: TT, SF; Formal analysis: TT; Methodology: TT; Project administration: TT, MK; Visualization: TT; Writing – original draft: TT; Writing – review & editing: MK, SF, FO, YS, YH, TH, KK, TY, YT, KM, NI.