Lumbar degenerative disease is the most common cause of radicular pain in the able-bodied population [1]. The persistence of radicular symptoms or neurological deficits following conservative treatment significantly worsens the quality of life and greatly impacts work, especially in physically and mentally active individuals [2]. Railway workers are often subjected to monotonous activities that can result in microtrauma to the musculoskeletal system and accelerate degenerative changes in the spine [3].

The treatment of lumbar degenerative diseases remains controversial and has not been standardized. This is partly due to the diverse clinical presentations, pathological conditions, and the absence of a direct correlation between neurological symptoms and neuroimaging data in some cases [4,5]. In addition, the timing of conservative versus surgical treatment, as well as their long-term clinical effectiveness, remain controversial [6–8].

To our knowledge, there are no studies on the long-term postoperative outcomes and risk factors for unsatisfactory long-term clinical outcomes after conservative and surgical treatment of lumbar degenerative disease in railway workers. Railway workers tend to be a homogeneous population and so have advantages as a study sample [3,8]. In addition, their workload poses risk factors for spinal degeneration, making them a unique population in which to study the effectiveness of various treatments.

In this study, we evaluated the return-to-work rate and long-term clinical effectiveness of delayed surgical treatment following prolonged (>6–12 weeks) conservative versus early surgical treatment following routine (6–12 weeks) non-operative treatment for persistent back pain and sciatica in railway workers.

The preoperative patient data are presented in Table 1. No intergroup differences were observed in any of the analyzed parameters (p>0.05).

Long-term postoperative outcomes were obtained in 82.5% (571/692) of the Delayed group and 80.8% (1,364/1,687) of the Early group (Table 2). The analysis revealed significantly better clinical results, fewer complications and reoperation rates, and a higher rate of return to work in the Early group than in the Delayed group (p<0.05). Moreover, in the lumbar interbody fusion subgroup, there were no intergroup differences in fusion rates.

In the Delayed group, 25.7% (147/571) had unsatisfactory long-term clinical outcomes compared to the Early group (15.3% [209/1,364]) (p<0.001). During the surgery, numerous difficulties were encountered in the Delayed group, including enlarged epidural veins, adhesive scar tissue, increased intraoperative bleeding, the need for prolonged use of electrocautery, hemostatic agents and drains, and unintentional durotomies. The above-mentioned technical difficulties were much rarer in the Early group during the initial surgical intervention.

Using the binary logistic regression model, the factors associated with unsatisfactory long-term clinical outcomes common to patients in both study groups were as follows: age >50 years, body mass index >30 kg/m², smoking status, presence of motor neurological deficits, and mental component score (MCS) SF-36 score <15 points (Table 3). Factors specific to unsatisfactory long-term clinical outcomes in the Delayed group included male sex, presence of type II diabetes mellitus, intensity of pain in the lower extremities >40 mm, ODI >48 points, physical component score (PCS) SF-36 score <18 points, use of narcotic analgesics before surgical treatment, and light-to-medium and heavy-to-very heavy preoperative workload (Table 3).

The goals of treatment for lumbar degenerative disease include the elimination of pain and restoration of pre-morbid physical activity and quality of life [11–14]. The outcome of such treatment is directly related to the selection of the most appropriate treatment based on the patient’s underlying disease and clinical presentation [15,16]. The timeliness of surgical treatment can be critical for preventing the progression of neurological deficits such as acute cauda equina syndrome [17]. However, for degenerative lumbar cases without emergency surgical indications, there are currently no well-accepted criteria for prioritizing surgical treatment over further non-operative treatment after failing the standard 6–12 weeks of non-operative care [7]. Although it is well established that surgical treatment is indicated when conservative treatment is ineffective, a clear strategy for appropriate therapy, the timing of surgery, and options for recurrent clinical symptoms are not well-established [18]. Possibly, some contradictions in the success of surgical versus non-surgical treatment in the literature are partly due to the heterogeneity of the subjects as well as their preoperative prognostic indicators. Therefore, the risk factors for unsatisfactory long-term clinical outcomes of conservative versus surgical treatment remain largely unexplored. In this study, we followed up patients who had failed an initial 6–12 weeks of conservative treatment and were then either treated with early or delayed surgery after failing further non-operative treatment. The Delayed group was offered surgery but refused and only consented after further non-operative treatment failed. To our knowledge, this is the first study to examine long-term clinical outcomes and return-to-work rates, as well as the risk factors for unsatisfactory long-term clinical outcomes following early versus delayed surgical treatment for lumbar degenerative disease in a relatively homogeneous population of physical laborers.

Our study included patients with only one surgical level, which was because most of the patients treated in our Neurosurgery Center had one clinically significant level. We also aimed to maximize the homogeneity of our cohorts and minimize the risk of bias; an increase in the number of operated levels tends to increase clinical and radiological variables compared to a single-level operation.

As is true for most operations, identifying and utilizing preoperative prognostic factors is essential for planning neurosurgical care, improving long-term clinical outcomes, and restoring work intensity levels in railway workers. We identified the following risk factors for unfavorable long-term outcomes following conservative and surgical treatment in railway workers. The common negative prognostic factors for both groups were age >50 years, body mass index >30 kg/m², smoking status, presence of motor deficits, and MCS <15 points (Table 3). Specific negative prognostic indicators for the Delayed group were male sex, presence of type II diabetes mellitus, lower extremity pain >40 mm, ODI >48%, PCS <18 points, use of narcotic analgesics before surgery, and light-to-medium and heavy-to-very heavy preoperative workloads.

There is a paucity of literature on early versus late surgical treatment for chronic degenerative lumbar disease. Most studies only compared operative versus non-operative treatment rather than the timing of the operative treatment. According to Davison et al. [19] over a 10-year follow-up period in a cohort of 531,980 patients identified in an insurance database with lumbar stenosis or spondylolisthesis, successful conservative therapy was noted in 98.3% (523,031) cases, whereas failure occurred in only 1.7% (8,949) cases, which served as a predictor for choosing surgical treatment. Ineffective conservative treatment resulted in a twofold increase in financial costs. The risk factors for ineffective conservative treatment were smoking status, obesity, and long-term opioid use. As is common in large database studies, the study has a paucity of data regarding patient demographics, the nature of both conservative and operative treatment, and outcome measures. In a prospective study by Atlas et al. [20], surgically treated patients had greater satisfaction and better functional outcomes than those in the non-operative group at 10 years. Similar results were obtained in a randomized clinical trial by Malmivaara et al. [21] at a 2-year follow-up. In contrast, a large prospective, randomized, multicenter study conducted by Weinstein et al. [22] did not demonstrate the superiority of surgical over non-operative treatment for pain and functional state according to ODI at 2 years. However, the study used an initial “intention-to-treat” methodology for the analysis. Therefore, patients who were initially randomized to receive non-operative treatment and then successfully treated by crossing over to surgical treatment were still categorized as having received successful non-surgical treatment. Similar results were obtained in a prospective cohort study by Gugliotta et al. [23] at 104 weeks in both conservative and surgical groups. Several systematic reviews of randomized clinical trials confirm the greater effectiveness of surgical treatment compared to non-operative treatment for long-term improvement of pain and functional status [23–26].

Our study adds to the evidence found in these studies that supports surgical treatment over continued conservative treatment. However, none of these studies was specifically designed to examine early versus late surgical treatment. Our data strongly suggest that after an initial failure of 6–12 weeks of non-operative treatment, early surgery results in better outcomes than further delaying surgery with more non-operative treatment.

This study has several limitations. First, this was a single-center retrospective study with the limitations associated with such studies. However, one can also posit that because this was a single-center study, variations in the quality and type of non-surgical and surgical care were significantly reduced, such that the strengths outweigh the limitations. Second, although we recorded long-term outcomes, we lacked early postoperative outcomes. However, most patients and surgeons are more interested in the final long-term outcomes, as opposed to the early postoperative outcomes. Third, we applied very strict inclusion criteria, enrolling only those who had failed non-operative treatment for 6–12 weeks and either agreed to or refused surgery. While the demographics of the group that refused surgery were similar to those of patients who opted for surgery, there may be some unmeasured differences between the two groups. For example, patients who refuse recommended surgery may have a higher level of distrust in the surgeons, a greater fear of surgery, less faith in medical treatment, or some other psychological or attitudinal differences. Such distrust has the potential to decrease the placebo effect of any treatment. As we did not evaluate the psychological attitudes of the patients, we cannot rule out such an effect. Therefore, we cannot exclude the possibility that the poorer outcomes in the Delayed group could have been, in part, due to some psychological factors that were not assessed, as this was a spine study conducted by neurosurgeons and not psychologists. We also did not assess the possibility of psychosomatic or secondary gain issues contributing to the differences in outcomes. In contrast, the Delayed group may have refused surgery initially because their symptoms did not bother them enough to opt for surgery. Despite similar preoperative outcome scores, it is possible that they had a more sanguine attitude toward their disability. This could have canceled out any negative psychological factor. In addition, due to the lack of psychologists on staff at our Neurosurgery Center and the retrospective design of the study, we were unable to assess the potential influence of psychological factors on the results of surgical intervention. Fourth, we did not perform further analyses based on the exact type and level of the procedure other than to divide them into microdiscectomy, laminectomy, and fusion groups. There was no difference between the Early and Delayed groups in terms of the number of procedures performed. Fifth, we performed a two-group analysis without multifactorial regression analysis because no variable was statistically significant. In addition, we considered the possibility of identifying various risk factors that, if taken into account in preoperative planning, could improve long-term clinical outcomes. Sixth, in the retrospective analysis of medical records, there was no information on the extent of motor neurological deficits in most cases, which precludes us from demonstrating the degree of motor impairment. Finally, because this was a relatively homogeneous population, the results may not apply to workers with different types of physically demanding jobs. However, to test the hypothesis that the timing of surgery can impact surgical results, we believe it is more important to reduce as many variables as possible.

Despite these weaknesses, the strengths of the study are the large number of patients who self-selected early versus delayed surgery in a relatively homogeneous population of railway workers treated by the same group of caregivers.

In patients with lumbar degenerative disease, structured conservative treatment is an effective method for achieving clinical improvement in >50% of cases. After such treatment for 6–12 weeks failed, early surgical intervention appeared to result in better clinical outcomes, reduced postoperative complications, higher return to preoperative work levels, and minimal reoperation rates. Univariate regression analysis identified age >50 years, body mass index >30 kg/m², smoking status, presence of motor deficits, and MCS <15 points as negative predictors in both the Early and Delayed groups. Specific negative prognostic indicators for the Delayed group were male sex, type II diabetes mellitus, lower extremity pain >40 mm, ODI >48%, PCS <18 points, use of narcotic analgesics before surgery, and light-to-medium and heavy-to-very heavy preoperative workloads.

Our limitations (single-center retrospective study, lack of analysis of early postoperative outcomes, psychological and psychosomatic factors, comparisons of the type of surgery performed, and multivariate regression analysis) call for verification with further prospective multicenter studies involving a larger number of respondents.