Cervical spondylotic myelopathy (CSM) is a spinal cord compression disorder caused by cervical degeneration and is a leading cause of spinal cord dysfunction in middle-aged and elderly individuals [5]. The pathophysiology of CSM is complex, involving intervertebral disc degeneration, ligamentum flavum hypertrophy, posterior longitudinal ligament ossification, and osteophyte formation, all of which contribute to spinal cord compression and neural dysfunction [8,13].

The prognosis of CSM is affected by multiple factors, including age, symptom duration, preoperative neurological status, degree of spinal cord compression, and surgical technique [8,13]. Age is a significant prognostic factor, with younger patients showing better postoperative recovery than older ones, likely due to enhanced nerve regeneration and fewer degenerative changes [7,29]. Patients with longer symptom duration tend to have a worse prognosis due to chronic nerve damage from prolonged spinal cord compression [7]. Preoperative neurological status are also crucial for prognosis, with higher Japanese Orthopaedic Association (JOA) scores linked to better postoperative functional outcomes [10].

Chronic inflammation is a key factor in the progression of CSM. Karadimas et al. [13] demonstrated that chronic spinal cord compression triggers a prolonged inflammatory response, leading to neuronal and glial cell damage, and exacerbates spinal cord injury by releasing cytokines, including tumor necrosis factor-α and interleukin-1β. Chronic inflammation may worsen spinal cord injury and hinder postoperative neurological recovery by disrupting the blood-spinal barrier and promoting inflammatory cell infiltration.

Imaging evaluation, particularly magnetic resonance imaging (MRI), is highly valuable for diagnosing and predicting the prognosis of CSM. T2-weighted MRI signal intensity changes are strongly correlated with the severity of spinal cord injury and postoperative outcomes [17]. MRI assesses the degree of spinal cord compression and identifies concomitant pathological changes, such as intramedullary high signal intensity, which is often associated with poor clinical outcomes [4,24].

Anterior cervical decompression and fusion (ACDF) and laminoplasty (LAMP) are the primary surgical treatments for CSM, each with distinct advantages and limitations. ACDF allows for the direct removal of anterior compressive lesions and preserves cervical stability through fusion, thereby maintaining the cervical lordotic angle postoperatively and enhancing neurological recovery [30,31]. However, ACDF is associated with greater surgical trauma, longer operative times, and an increased risk of perioperative complications, such as dysphagia and graft loosening [11]. In contrast, LAMP relieves spinal cord compression indirectly by expanding the spinal canal, making it a simpler procedure that is particularly suitable for patients with multisegmental lesions [10,23]. Nevertheless, long-term outcomes with LAMP may be compromised by residual anterior compression syndrome or the development of postoperative kyphotic deformity [10,11]. Overall, studies indicate that ACDF provides superior postoperative neurological recovery and cervical stability, whereas LAMP offers the benefits of reduced surgical trauma and fewer complications [10,11,23,30,31].

The prognosis of CSM is influenced by multiple factors, and understanding their interactions and underlying mechanisms is crucial for optimizing treatment strategies and improving patient outcomes. This study aimed to analyze the clinical data of 158 CSM patients to identify key prognostic factors, enhance prognostic accuracy, and offer stronger guidance for clinical practice.

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the First Affiliated Hospital of Nanjing Medical University (Ethics Committee Number : 2023-SR-801). Given the retrospective nature of this investigation, the Ethics Committee waived the requirement for informed consent. This study is not a clinical trial; therefore, a clinical trial number is not applicable.

A total of 158 patients were included in the study. According to previous reports, we divided them into good and poor prognosis groups based on whether their improvement rate was ≥50% [11,15,21,28]. The good prognosis group consisted of 53 men (65.43%) and 28 women (34.57%) with a mean age of 52.15±9.27 years. The body mass index (BMI) in this group was 24.77±3.00. The group had a mean preoperative JOA score of 13.72±1.41 and a regulatory T-cell count of 5.91%±1.87%. The spinal cord compression ratio averaged 0.62±0.14, and the lesion length was 8.59±4.22 mm. In this group, the number of surgical levels was 1.99±0.81. Among them, 7.41% (6/81) had diabetes, 29.63% (24/81) had hypertension, and 8.64% (7/81) had a history of smoking. The poor prognosis group included 52 men (67.53%) and 25 women (32.47%), with a mean age of 60.71±9.63 years. This group had a preoperative JOA score of 12.71±1.47 and a regulatory T-cell count of 4.54%±2.14%. The spinal cord compression ratio was 0.54±0.14, and the lesion length was 13.85±4.56 mm. The BMI in this group was 24.67±3.48, and the number of surgical levels was 2.43±0.88. In this group, 20.77% (16/77) had diabetes, 36.36% (28/77) had hypertension, and 15.58% (12/77) were smokers. Significant differences between the good and poor prognosis groups were observed in age, symptom duration, preoperative JOA score, regulatory T-cell count, spinal cord compression ratio, and history of diabetes (p<0.05). No significant differences were found in sex, BMI, or history of hypertension or smoking (p>0.05). Table 1 presents the demographic and baseline clinical characteristics of the two groups.

A logistic regression analysis was conducted to assess the effects of age, sex, BMI, symptom duration, preoperative JOA score, regulatory T-cell count, spinal cord compression ratio, lesion length, the number of surgical levels, and history of hypertension, diabetes, and smoking history on CSM prognosis (Supplementary Table 1). The findings showed that age had a significant negative effect on the improvement rate (p<0.05), with older patients experiencing lower postoperative improvement rates, as reflected by the regression coefficient of -0.0059 (95% CI, -0.0080 to -0.0037). Symptom duration was also a significant negative predictor with a regression coefficient of -0.0062 (95% CI, -0.0099 to -0.0026), indicating that longer symptom duration correlates with a lower improvement rate. Preoperative JOA score (p<0.05) and regulatory T-cell count (p<0.05) were positively correlated with the improvement rate, with regression coefficients of 0.0258 (95% CI, 0.0110 to 0.0405) and 0.0176 (95% CI, 0.0071 to 0.0281), respectively. A regression coefficient of 0.1899 (95% CI, 0.0202 to 0.3596) indicated that the spinal cord compression ratio was significantly positively correlated with the improvement rate (p<0.05), suggesting that a higher spinal cord compression ratio leads to greater postoperative improvement. Shorter lesion length was associated with a better prognosis, as demonstrated by a negative correlation with improvement rate (p<0.05), with a regression coefficient of -0.0109 (95% CI, -0.0161 to -0.0056). More surgical levels are associated with a worse prognosis, with a regression coefficient of -0.0333 (95% CI, -0.0587 to -0.0080). Diabetes was associated with a significantly lower improvement rate (p<0.05), with a regression coefficient of -0.0708 (95% CI, -0.1319 to -0.0097), indicating that diabetes negatively impacts postoperative improvement in CSM patients. No statistically significant associations were found between the other variables and the improvement rate. A forest plot (Fig. 3) illustrates the results of the multiple regression analysis.

A stratified analysis was conducted to better understand the impact of each variable on different patient subgroups (Tables 2-4). Patients were categorized into subgroups based on age, symptom duration, and preoperative neurological function, as these factors significantly affect prognosis. To evaluate whether these factors altered the effect of other variables on the improvement rate, stratification was performed using cutoffs of 50 years for age, 12 months for symptom duration, and 12 points for preoperative JOA score. Cut-off values for stratified analysis were determined based on previously published literature [6,9,12,18,25]. Age, duration of symptoms, and preoperative JOA scores exhibit stable independent predictive value in most stratifications, highlighting their central role in prognostic evaluation due to their strong association with improvement rates. Stratification by age reveals that the recovery process of elderly patients is significantly driven by Treg cell counts and the spinal compression ratio (Table 2). Simultaneously, an increase in the number of surgical segments and a history of diabetes significantly weaken the prognosis. Conversely, the improvement rate in younger patients is primarily influenced by the negative impact of prolonged symptom duration. Although Treg cells play a significant role their contribution is more limited compared to the elderly group. In the stratification by symptom duration (Table 3), patients with long-term symptoms show a positive correlation between improvement rates and spinal compression ratios as well as Treg cells. However, lesion length and multi-segment surgery significantly inhibit recovery. Patients with short-term symptoms rely more on the improvement of preoperative JOA scores. Notably, a history of diabetes exhibits a stronger negative effect in the long-term symptom group but does not reach a significant level in the short-term group, indicating that metabolic disorders have a more prominent prognostic impact on patients with chronic disease courses. Stratification based on preoperative JOA scores reveals that the improvement rate in the high-score group is closely related to Treg cells and anatomical factors (Table 4). In contrast, the prognosis of the low-score group is primarily limited by both age and duration of symptoms, further confirming that preoperative neurological status can independently predict outcomes from other variables.

To further examine the influence of surgical extent, patients were stratified into three groups based on the number of operated levels (surgical levels ranged from 1 to 5 in both groups) : single-level, double-level, and multi-level (Table 5). Each subgroup included more than 30 patients, ensuring adequate statistical power for multivariate analysis. This stratification was clinically justified, as surgical complexity, tissue trauma, and the risk of postoperative complications increase with the number of operated segments. In patients undergoing single-level surgery, preoperative neurological status was positively associated with postoperative improvement, while lesion length was a negative predictor. As surgical complexity increased, additional prognostic variables became significant. In double-level procedures, advanced age, longer symptom duration, diabetes history, and greater lesion length negatively impacted recovery, whereas higher regulatory T cell counts were associated with improved outcomes. In multi-level surgeries (≥3 segments), age and symptom duration remained strong negative predictors, and preoperative neurological status continued to correlate positively with recovery. Regulatory T cell counts approached statistical significance in this group, suggesting potential clinical relevance. Collectively, these findings indicate that while age and symptom duration consistently predict poorer outcomes with increasing surgical complexity, the prognostic value of neurological function and immune status is modulated by the extent of surgery. Therefore, stratified prognostic assessment and individualized perioperative management strategies are essential for optimizing outcomes in clinical practice.

These findings demonstrate significant heterogeneity in prognostic mechanisms among different patient subpopulations. Treg cell counts are significantly positively correlated with improvement rates in all stratification analyses, suggesting their stable role in postoperative recovery. Additionally, the influence of the number of surgical segments varies based on individual characteristics, with a more pronounced negative effect in older or longer-duration patients. These results emphasize the need for precise decompression and optimization of the immune microenvironment in elderly or long-term symptom patients, minimizing surgical trauma. Conversely, younger or short-term symptom patients should have a shorter window from diagnosis to surgery to reduce subsequent neurological damage.

Additionally, we developed a predictive model for the improvement rate using multiple regression analysis. Significant predictors included age, symptom duration, preoperative JOA score, Treg cell count, spinal cord compression ratio, lesion length, and diabetes history. Using these predictors, the improvement rate can be calculated as follows : improvement rate = 0.5662 − 0.0059 × age − 0.0062 × duration of symptoms + 0.0258 × preoperative JOA score + 0.0176 × count of regulatory T cells + 0.1899 × spinal cord compression ratio − 0.0109 × lesion length − 0.0333 × number of surgical levels − 0.0708 × diabetes. The model’s performance metrics indicate a strong explanatory capability, with an R² of 0.687, explaining 68.7% of the variability in the improvement rate. This suggests that the model effectively captures key factors influencing improvement. The adjusted R² of 0.661, which accounts for the number of predictors, further supports the model’s robustness in explaining data variability. The F-statistic was 26.51, with a p-value near zero (8.82e-31), confirming the overall statistical significance of the model. Additionally, the log-likelihood was 107.37, AIC was -188.75, and Bayesian information criterion was -148.93, all indicating a well-fitted model. The Durbin-Watson statistic was 1.312, suggesting no severe autocorrelation in the residuals. The overall model demonstrated strong predictive stability and a solid explanatory power, effectively capturing key determinants of improvement rate.

CSM is a severe form of cervical spondylosis that often requires surgical intervention. With the rise in global life expectancy and the growing elderly population, the prevalence of CSM is expected to increase, making pre- and postoperative risk factor analysis critical. This study assessed the impact of various factors on the postoperative prognosis of CSM patients. The findings revealed that key factors influencing prognosis included age, symptom duration, JOA score, imaging features, regulatory T-cell count, the number of surgical levels and history of diabetes. These factors can be directly assessed, which may enhance diagnostic and treatment efficiency and improve patient outcomes.

Age was identified as a significant predictor of outcomes in CSM patients. Younger patients generally show better postoperative recovery, consistent with findings from multiple studies [29]. Their improved prognosis is likely due to greater neurological resilience and fewer comorbidities [26]. In the stratified analysis by symptom duration and preoperative JOA score, age was significantly negatively correlated with the improvement rate (p<0.05), confirming that age is a consistent negative prognostic factor, independent of other variables. These findings emphasize the need for more personalized postoperative management and support for older patients.

Preoperative JOA scores were identified as a critical factor in predicting CSM outcomes. Consistent with previous studies, patients with higher preoperative JOA scores demonstrated better postoperative recovery [7]. Therefore, a higher preoperative JOA score is a strong predictor of favorable postoperative outcomes, suggesting that preoperative neurological function is key to prognosis.

Neurological status should be carefully evaluated when making surgical decisions. For patients with lower JOA scores, intensive rehabilitation or other strategies should be employed to improve neurological function and prognosis. MRI T2-weighted images often show a correlation between high signal intensity and the severity of spinal cord injury and recovery, reflecting conditions such as spinal cord edema, ischemia, or gliosis. The extent of injury, quantified by spinal cord compression ratio and lesion length on MRI, was significantly correlated with the improvement rate. Notably, in patients with longer symptom duration and lower JOA scores, a higher spinal cord compression ratio (indicating less compression) was positively associated with recovery across various subgroups. Therefore, timely decompression surgery is critical once CSM is diagnosed. Lesion length consistently emerged as a negative predictor across all stratifications, influencing the improvement rate and requiring consideration in preoperative planning. Comprehensive rehabilitation plans should be developed for patients with larger lesions. The study also showed that longer symptom duration resulted in poorer postoperative prognosis due to irreversible spinal cord damage, underscoring the importance of early intervention to prevent further deterioration.

It has been reported that an increase in the number of surgical segments significantly impacts surgical complexity and postoperative complications in the ACDF group [11]. A greater number of surgical segments is associated with longer operative times, increased blood loss, and a higher risk of postoperative adjacent segment degeneration. Our study also confirms that patients undergoing surgery on multiple segments tend to have a poorer prognosis. Further stratified analyses based on additional factors revealed that the prognostic impact was especially prominent in older patients, who often present with longer symptom durations and poorer preoperative neurological status. Consistently, stratification by the number of surgical levels showed that in patients undergoing multi-level procedures, age, symptom duration, and immune status were significant predictors of outcome—likely reflecting the increased surgical trauma and complexity associated with extensive operations. These findings suggest that older individuals and those with prolonged symptom history face physiological constraints on neurological recovery, underscoring the critical importance of early intervention in these high-risk populations.

In the context of diabetes and CSM, patients without diabetes show better postoperative recovery compared to those with diabetes [12,14]. Machino et al. [16] reported that hemoglobin A1c levels and diabetes duration are important predictors of the JOA recovery rate. Our study also found that patients with a history of diabetes had a lower postoperative improvement rate, potentially due to complications such as macrovascular and microvascular disease, demyelination, and diabetes-related peripheral neuropathy.

The impact of immunological factors on prognosis was also assessed. Karadimas et al. [13] proposed that prolonged spinal cord compression leads to ischemic changes and glial cell activation, which in turn contributes to a chronic pathological environment that may exacerbate spinal cord injury. Fan et al. [7] reported that chronic inflammation disrupts the local spinal cord environment by damaging the blood-spinal cord barrier and promoting ongoing inflammatory cell infiltration, which hinders nerve regeneration and repair. Treg, key immune regulators, are known to improve prognosis in various diseases [1,22]. Previous studies have shown that Treg cells play a critical role in the repair of acute spinal cord injury, and a reduction in their number can impede tissue remodeling [27]. Treg cells promote spinal cord recovery by modulating macrophage polarization and secreting amphiregulin [3]. Raposo et al. [20] demonstrated that Treg cells migrate from peripheral tissues to the injury site in response to chemokine signaling. Therefore, a reduction in the overall proportion of Treg cells in the body may decrease the number of cells available to infiltrate the injured area [20]. Our study found a positive correlation between peripheral Treg cell count and postoperative improvement rate in CSM patients, suggesting that enhancing immune function could improve outcomes. In the context of chronic spinal cord injuries like CSM, Treg cells may limit immune system attacks by inhibiting excessive immune responses and controlling the release of inflammatory mediators, thereby improving prognosis. In age-stratified analysis, Treg cells had a significant impact on the improvement rate in patients over 50 years, consistent with Birmingham et al. [2]’s findings that older patients have poorer asthma control due to decreased Treg function. This link between Treg count and improved postoperative outcomes in CSM highlights the potential of immunomodulation as a therapeutic approach. Enhancing Treg function, whether through drugs, cell therapy, or bioactive interventions, may offer new possibilities for improving surgical outcomes in CSM patients.

To our knowledge, this is the first study to specifically investigate the association between peripheral regulatory T cell counts and postoperative prognosis in CSM. Previous studies have primarily focused on the role of Treg cells in acute spinal cord injury and autoimmune-related inflammation, highlighting their involvement in neuroprotection and tissue repair [3,20,27]. In contrast, our study focuses on CSM, and demonstrates that peripheral regulatory T cell counts are positively correlated with neurological recovery following surgical decompression. This expands the current understanding of immune modulation in chronic degenerative spinal disorders and offers insights distinct from prior immunological studies centered around acute traumatic or autoimmune spinal conditions.

This study has several limitations. First, as a single-center retrospective study, it may be subject to selection and information bias. Second, the relatively small sample size may limit the generalizability of the results. Third, important factors like postoperative rehabilitation were not comprehensively analyzed. Finally, the short follow-up period restricted the assessment of long-term prognosis, due to the retrospective study design and ethical and technical limitations, long-term radiographic follow-up of patients is not possible, so the incidence of pseudarthrosis and its potential impact on prognosis cannot be assessed. Future research should aim to conduct large-sample, multi-center prospective studies to verify these findings and improve generalizability, incorporate regular postoperative imaging follow-ups to more precisely assess the impact of surgical outcomes on prognosis. Additionally, the role of Treg cell count in CSM prognosis warrants further exploration, including an investigation into its mechanisms. Studies on the impact of different surgical methods and postoperative rehabilitation on CSM prognosis will help optimize treatment strategies. Moreover, longer follow-up periods are necessary to thoroughly assess long-term outcomes and refine the prognosis evaluation system for CSM patients.

Several risk factors influencing the prognosis of CSM surgery were identified, including age, preoperative JOA score, imaging characteristics, symptom duration, Treg cell count, the number of surgical levels and history of diabetes. These findings provide valuable insights to guide clinicians in surgical decision-making. Future research should focus on exploring the interactions between these factors and other potential prognostic variables to optimize treatment strategies for CSM patients.