INTRODUCTION
Low back pain (LBP) is the leading cause of disability worldwide and is estimated to increase in prevalence due to aging and an expanding population [1]. There are several etiologies that play into the pathophysiology of LBP, with disc herniation accounting for 5–20 cases per 1,000 [2]. Recurrent lumbar disc herniation (rLDH) refers to radiologic reappearance of disc material previously operated on found after an initial symptom-free period [3]. When symptomatic, many of these patients undergo surgical discectomy after conservative management has failed or when associated with profound neurological deficit [4]. Surgical options may include open, microscopic, and endoscopic discectomy [5]. The recurrence of disc herniation after discectomy remains a significant clinical concern, occurring in up to 25% of cases and associated with risk factors such as smoking, obesity, diabetes, and preoperative radiographic changes [6].
Until now, no review has thoroughly examined this comprehensive list of factors for recurrent disc herniation. Previous works have fallen short only investigating radiographic and clinical risk factors (such as body mass index [BMI], smoking, and diabetes) risk factors independently [7,8]. Abdel-Fattah et al. [9] performed a comprehensive review investigating both radiographic and clinical risk factors, but was deficient in numerous radiographic factors such as disc height index (DHI), sagittal range of motion (SROM) and disc laterality, all of which were incorporated in the present study. Additionally a previous systematic review erroneously included studies from the same study sample, which could have skewed data [10]. Lastly, this study is an up-to-date systematic review that includes recent published data as opposed to reviews published more than 5 years ago [6-8].
With an ever increasing volume of lumbar discectomies performed within the United States [11], it is essential for a holistic review of risk factors to understand the real risks of recurrence and better preoperative patient counseling. This study aimed to report up-to-date findings of all risk factors both clinical and radiologic.
MATERIALS AND METHODS
1. Search Strategy
In an attempt to identify all studies that focused on the investigation of patients with rLDH after surgery, the authors queried the MEDLINE/PubMed and Embase databases without time restrictions or filters placed. The systematic search encompassed all eligible studies available up to October 2025. RLDH was defined as the reappearance of herniated disc material at the same level, accompanied with symptomatology following a free interval following surgery irrespective for the need for revision surgery. The following search terms were utilized to identify all relevant publications: (“risk factor” OR “prognostic factor” OR “epidemiologic factors” OR “multivariate analysis”) AND (“prolapsed disc” OR “herniated disc” OR “disc displacement” OR “disc herniation” OR “disc prolapse” OR “prolapsed disk” OR “disk herniation” OR “disk displacement”) AND (“recurrent” OR “relapse” OR “reoperation” OR “rehospitalization” OR “readmission”). Adherence to Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) guidelines was strictly followed to ensure accurate reporting of the findings [12].
2. Screening Process
All publications were first screened by title and abstract by 2 independent authors, and any disagreements were resolved by a third author. Duplicates were identified and removed at the conclusion of this process. Once all studies were screened for relevance then the articles’ full texts were downloaded and reviewed for final inclusion before the data collection process. Studies were excluded if they were not originally published in English, did not focus on lumbar spine surgery, lacked a recurrent disc herniation cohort, included patients who failed to improve after surgery, lacked a control group, or included overlapping patient populations with another included study. Studies were also excluded if they were case reports, reviews (meta-analysis, systematic review), propensity score-matched analysis (controlled for clinically relevant variables) or were not a full patient-based study (abstract, commentary, guidelines, letter to the editor, animal study, surgical technique).
3. Data Collection
All studies without exclusion criteria proceeded to data extraction, during which the following study characteristics were collected when available: first authors’ last name, year of publication, type of study (retrospective/prospective), duration of follow-up, and surgical procedure. The following variables were collected for the recurrent and nonrecurrent cohorts: total number of patients, age, BMI, number of men, smokers, diabetic patients, type of disc herniation (protrusion, extrusion, sequestration), side of procedure (left/right), levels of surgery (L5–S1 versus all other levels), DHI, patients with Modic changes, and distribution of disc degeneration. Disc degeneration was classified using the modified Pfirrman system, a published extension of the original 5-grade Pfirrman scale that expands it to 8 categories to allow more granular assessment of degeneration severity [13]. The modified system defines 8 levels of degeneration; however only grades I–VII were observed in our dataset, as grade VIII (complete collapse) was not present. Accordingly, studies reported degeneration as modified Pfirrman grades I–III and VI versus grades IV, V, and VII which were grouped accordingly for analysis. SROM was defined as the difference in angular measurements between the extension and flexion angles in the sagittal plane. Lastly, DHI was calculated by dividing the height of the intervertebral disc by the height of the superior vertebral body.
4. Quality Assessment
All included studies in this meta-analysis also underwent an analysis of bias with the Newcastle-Ottawa Scale, where a score of 0–2 indicated poor quality, a score of 3–5 was fair quality, and >6 was considered high quality [14]. Potential publication bias was assessed through visual inspection of a funnel plot (Figs. 1–3). The near-symmetric distribution of studies around the pooled effect estimate suggested a minimal likelihood of publication bias.
5. Statistical Analysis
All statistical analysis was completed utilizing the statistical software program R (https://www.r-project.org/). Continuous variables were presented as mean±one standard deviation or as mean (range) when standard deviation was not available. Pooled weighted odds ratios (ORs) were utilized to evaluate categorical variables as potential risk factors for rLDH, while pooled weighted standardized mean differences (SMDs) were used for continuous variables. If a variable had 3 or fewer studies, they were not included in the meta-analysis. All data were presented using forest plots that demonstrate the estimated effects (OR/SMD), the 95% confidence interval (CI), and the calculated weights. Across all variables investigated, heterogeneity was evaluated with the I2 statistic, which shows the percentage of variation between studies excluding the random chance effect. I2>50% was defined as significant heterogeneity across the studies. Findings were reported according to the random effects model (DerSimonian and Laird approach), if there was significant heterogeneity and fixed effect model if there was insignificant heterogeneity.
RESULTS
1. Study Identification and Demographics
The search and review process for the identification and selection of articles for meta-analysis, according to the inclusion and exclusion criteria of the study, is illustrated in the PRISMA flow chart (Fig. 4). A total of 511 articles from MEDLINE/PubMed and 755 from Embase were identified. After removing duplicates, including relevant references from article bibliographies, screening abstracts, and reviewing full texts, 51 studies involving 52,479 patients met the inclusion criteria (Table 1) [15-65]. Among these patients, 6,416 (12.2%) experienced recurrent disc herniation. The included studies consisted of 5 prospective (9.8%) and 46 retrospective studies (90.2%).
2. Risk Factors
1) Overview of evaluated risk factors
A comprehensive analysis of risk factors for rLDH was conducted across multiple domains. Clinical factors assessed included age at the time of index discectomy [18-24,26,28,32,34-39,41,42,46,49-61,64,65], sex [15-30,32-41,44-52,54-59,62-65], both of which were not significantly associated with recurrence. In contrast, a BMI>30 kg/m2 [18-20,23,26,28,32,34-36,38-41,43,46,50-52,54-56,59,60,64,65], and diabetes mellitus [17,19,20,22-25,32-36,38,39,41,46,48,49,51,54-56,58-65] were all significantly associated with increased odds of recurrence. Radiographic risk factors were also explored. The presence of Modic changes on preoperative imaging [15,18,23-25,27,30,32,33,38,39,47-50,54,55,59,64,65] and increased SROM [18,23,32,34,35,38,49,50,52,56,59] significantly elevated the risk of recurrence. Other risk factors, including smoking, herniation type (protrusion, extrusion, sequestration), disc level (L5–S1), DHI, and Modified Pfirrmann grade, did not show a significant association with recurrence.
2) Baseline demographics and clinical risk factors
Among the 33 studies that reported age at the time of index discectomy, there was no statistically significant difference in age between patients with and without recurrent disc herniation (SMD, 0.20; 95% CI, -0.33 to 0.74) [18-24,26,28,32,34-39,41,42,46,49-61,64,65]. Similarly, the 41 studies reporting gender found no significant difference between the groups (OR, 1.16; 95% CI, 1.00–1.33) [15-30,32-41,44-52,54-60,62-65]. Conversely, patients with recurrent disc herniation had significantly higher rates of diabetes, and higher BMI compared to nonrecurrent disc herniation group. Analysis of 27 studies assessing the risk of recurrent disc herniation in diabetic patients revealed significantly higher odds of recurrence among diabetics (OR, 1.48; 95% CI, 1.23–1.77) [17,19,20,22-25,32-36,38,39,41,46,48,49,51,54-56,58,59]. The meta-analysis of the 26 studies that reported BMI showed significantly higher BMI in the recurrent group (SMD, 0.48; 95% CI, 0.26–0.70) [18-20,23,26,28,32,34-36,38-41,43,46,50-52,54-56,59].
However, the meta-analysis of 43 studies reporting smoking status, showed no significant association between smoking and rLDH (OR, 1.16; 95% CI, 1.00–1.33) [15-25,29,33-38,40,41,44,46-52,54-60,62-65]. Pooled analysis demonstrated recurrence rates of 12.7% for endoscopic discectomy, 11.4% for microscopic discectomy, and 15.3% for open discectomy. Microscopic discectomy was associated with a significantly lower recurrence rate compared with endoscopic discectomy (p=0.007), whereas both microscopic and endoscopic approaches demonstrated significantly lower recurrence rates than open discectomy (p<0.001). Demographics and clinical risk factors’ forest plots are shown in Fig. 5.
3) Radiographic risk factors
Seventeen studies have examined the impact of Modic changes on the risk of recurrent disc herniation [15,18,23-25,27,30,32,33,38,39,47-50,54,55,59]. Meta-analysis revealed the presence of Modic changes on preoperative images significantly increased the odds of recurrence (OR, 2.97; 95% CI, 2.20–4.01). An additional significant radiographic risk factor was an increased SROM, as shown by the meta-analysis of 11 studies (SMD, 2.15; 95% CI, 0.35–3.94) [18,23,32,34,35,38,49,50,52,56,59,64]. Other radiographic characteristics, such as protrusion (OR, 0.14; 95% CI, 0.77–1.68), extrusion (OR, 0.90; 95% CI, 0.66–1.22), sequestration (OR, 1.26; 95% CI, 0.85–1.87), L5–S1 disc herniation compared to other levels (OR, 0.96; 95% CI, 0.83–1.11), DHI (SMD, 2.49; 95% CI, -1.48 to 6.47), and Pfirrmann grades 4 and 5 compared to grades 1–3 (OR, 1.34; 95% CI, 0.85–2.10), did not significantly increase the risk of recurrence of disc herniation. Radiographic risk factors’ forest plots are shown in Figs. 6 and 7.
DISCUSSION
RLDH globally represents the most frequent complication after lumbar discectomy [66]. Shriver et al. [67] performed a systematic review of postoperative outcomes following lumbar discectomy and found that wound infection, hematoma, and neurological deficits all fell short in comparison to rLDH. Furthermore complications of reherniated disc included continued LBP, lasting nerve damage, paralysis, disability, and death [68]. This notable high frequency and complication burden in disc recurrence warrants attention and further research. The purpose of this study is to provide practitioners with a comprehensive overview of risk factors to guide preoperative optimization and risk assessment. Previous meta-analyses have lacked extensivity [6,8,9], and adequate research protocol [10].
The key findings of this analysis are that elevated BMI, diabetes, presence of preoperative Modic changes, and increased SROM were significantly associated with recurrent disc herniation.
1. Obesity
Thirty-seven percent of adults within the United States are classified as obese [69], and this high prevalence of disease impacts spine surgery significantly [70]. Elevated BMI imposes greater mechanical loading on the intervertebral disc potentially leading to a higher incidence of recurrent disc herniation. Consistent with this, Juković-Bihorac et al. [71] investigated the histopathologic effects of obesity within the intervertebral disc and demonstrated significant differences in chondrocyte proliferation, tears, clefts, granular changes, and mucous degeneration in patients with BMI>25 kg/m2. These weight-bearing effects destabilize the integrity of the annulus fibrosis and predispose obese patients to disc extrusions and protrusions [71]. Altered biomechanics and histology due to weight increases the likelihood of disc herniation along with recurrence, and within our systematic review, a BMI>30 kg/m2 was significantly higher in the recurrent herniation group. Previous systematic reviews investigating obesity as a risk factor have been performed and demonstrate significant increase in postoperative recurrence [6,7,10]. We suggest that clinicians counsel patients regarding the increased risk of recurrence associated with elevated BMI and when feasible, encourage weight reduction prior to surgery [43].
2. Diabetes
A relationship has been established between diabetes and degenerative lumbar disease [72]. Park et al. [73] demonstrated that type 2 diabetes was significantly associated with lumbar spine disorders and frequent spinal procedures. Additional studies within the literature reflect this [74-77]. Jiang et al. [78] conducted animal studies investigating the histopathological effects of diabetes and found that uncontrolled glucose levels induce autophagy within the intervertebral discs in diabetic rats. The effects of hyperglycemia on the lumbar spine have also been studied radiologically, demonstrating increased fatty infiltration within the paraspinal muscles, and increased endplate cartilage [79]. These changes alter the biomechanical forces of the spine and provide less support surrounding the intervertebral disc [79]. After a comprehensive analysis of risk factors within the recurrent group in the current study, it was found that a significant proportion of patients were diabetic. Abdel-Fattah et al. [9] investigated risk factors within their meta-analysis and found diabetes demonstrated higher odds of rLDH. These findings highlight the need for optimized diabetes management and patient counseling regarding the increased risk of recurrent disc herniation [80].
3. Intervertebral SROM
The investigation of SROM proved radiological significance, demonstrating an increased odds within the recurrent group. Conventional radiography is a common initial imaging modality in the primary evaluation of disc herniation [30]. Although x-rays are widely available, current literature investigating lumbar SROM and disc herniation remains limited. Fujiwara et al. [81] performed a cadaveric study investigating the segmental flexibility of the spine where he hypothesized spinal instability is related to biomechanical stress on the affected disc. This hypothesis is consistent with our findings of increased SROM within the recurrent group. We believe that increased SROM propagates increased axial load on the disc due to movement because of weakness in surrounding degenerative structures. In a cadaveric study, Tanaka et al. [82] proposed that degenerative changes represent the primary cause of increased SROM. Previous meta-analysis performed by Brooks et al. [8] reported consistent weighted findings along with support of the instability and degenerative theories. We suggest that patients with severe SROM should be counseled regarding their elevated risk of recurrence and the potential need for lumbar fusion in the future.
4. Modic Changes
Modic changes were first described in 1988 and are defined as vertebral bone marrow signal intensity changes associated with degenerative disc disease [83]. Modic changes capture a pathophysiological snapshot of endplate degeneration [84]. Consistent with our study, Modic changes surrounding the intervertebral disc demonstrated increased odds for recurrent disc herniation amongst numerous studies [8,85,86]. The determinants of Modic changes reflect degeneration and an interplay with many of the risk factors discussed above such as smoking, obesity, and occupational load [83]. In summary, structural damage triggers an inflammatory cascade which results in immune mediated reactions or microbial invasion [87]. These cascades of events activate intracellular pathways leading to adipogenesis, osteoclastogenesis, and weakening of structural integrity [87]. This consequently leads to disc fragility and propulsion. This histopathology is explanatory and consistent with the findings of our study and recent meta-analysis in the literature [8-10]. As a significant risk factor, Modic changes can be evaluated preoperatively for risk assessment in the recurrence of disc herniation.
5. Other Risk Factors
Other risk factors for rLDH found to be nonsignificant within our study were age, sex, smoking, laterality of initial disc herniation, herniation type, herniation level, DHI, and high Pfirrmann grades. Although these risk factors were nonsignificant in our analysis, certain risk factors were significant in previous systematic reviews. Differing systematic reviews concluded that herniation type, advanced age, and smoking were significant factors within the recurrent group [9,10]. Dissimilarities between systematic reviews can be accounted due to the absence in exclusion of studies with overlapping cohorts [10]. Duplication of patient cases and their characteristics can negatively affect weighted results and external validity [88]. These discrepancies may also be attributable to earlier studies published 3–10 years ago, as many systematic reviews and recent analyses predate or omit several newer publications on this topic [6-8]. Although minor variations were found between the reviews, core risk factors such as BMI, diabetes, and Modic changes remained consistent [9,10].
6. Future Directions
Future directions include patient education and counseling through education programs. Augmented diabetic instructive programs have been found to be significantly effective in reduction of hemoglobin A1c in comparison to those without [89]. Added diabetic education should be incorporated in postoperative care. Weight loss is an additional recommendation. Within the literature weight loss as management has been associated with the reversal of disc herniation symptoms [90,91]. Reduced mechanical stress has been shown to promote matrix remodeling and neovascularization, which are the proposed radiologic mechanisms underlying symptom reversal [92]. Utilization of radiographic parameters such as SROM and Modic changes can be assessed preoperatively for better patient counseling that they are at high risk of recurrence and might need lumbar fusion in the future.
7. Limitations
This study had limitations inherent to most meta-analysis, including the potential for missed studies and incomplete data largely due to retrospective design of many included studies. Secondly our meta-analysis was limited by an imbalance, as fewer studies focused on radiographic risk factors. This low power could prevent generalization and external validity. Furthermore, definitions of the symptom-free interval used to identify recurrent disc herniation varied among the included studies, which may have introduced inconsistencies in case classification. Additionally, most of the included studies were retrospective, the selection of experimental and control groups was related to the outcome of disc herniation, which could introduce selection bias to some extent. Although subgroup analyses comparing surgical techniques demonstrated statistically significant differences in recurrence rates, these results may be influenced by beta error due to large cumulative sample size, potentially overstating the clinical magnitude of these associations. Lastly, statistical heterogeneity was found within significant factors such as smoking and SROM. Meta-regression was considered to explain some of the high heterogeneity, however, with the limited number of included studies this was not possible in the current study.
CONCLUSION
Our study found the following factors to be associated with recurrent disc herniation: BMI above 30 kg/m2, diabetes, Modic changes, and SROM. Knowledge of these risk factors also provides insight into preoperative target goals to mitigate complications. We encourage further research to investigate additional risk factors that may exist and to validate those factors.
PDF Links
PubReader
ePub Link
XML Download
Full text via DOI
Download Citation
Print
