Incidence of subsequent surgical decompression following minimally invasive approaches to treat lumbar spinal stenosis: A retrospective review

Surgical decompression is the definitive treatment for managing symptomatic lumbar spinal stenosis; however, select patients are poor surgical candidates. Consequently, minimally invasive procedures have gained popularity, but there exists the potential for failure of therapy necessitating eventual surgical decompression.


I N T RODUC T ION
Symptomatic lumbar spinal canal stenosis is a debilitating condition and increasingly affects the aging population, with a mean prevalence upwards of 39%. 1 Surgical decompression, which is performed in approximately 30% of cases, 2 is considered the gold-standard treatment for patients with symptomatic lumbar spinal stenosis who are refractory to conservative management. 35][6][7][8] Furthermore, the selection of appropriate surgical candidates can be challenging in a growing elderly population with several coexisting comorbidities.As a result, minimally invasive techniques have been developed to help decompress the lumbar spine, including Percutaneous Image-guided Lumbar Decompression (PILD) and placement of lumbar Interspinous Spacer Devices (ISD).While these minimally invasive techniques may provide another therapeutic option, evidence regarding their long-term benefits and the eventual need for subsequent surgical decompression is lacking. 9hus, this large-scale claims-based retrospective analysis aims to better understand the incidence and characteristics of patients who require surgical decompression after receiving one of the minimally invasive techniques to treat symptomatic lumbar spinal stenosis.Our primary outcome was to assess the rate of conversion to surgical decompression in patients who undergo minimally invasive techniques for the management of symptomatic lumbar spinal stenosis.The secondary outcome was to understand the characteristics of patients who require subsequent surgical interventions and the factors associated with their risk of receiving additional surgical decompression.

Data source
All data were collected using claims contained in the PearlDiver patient record database (PearlDiver Technologies, IN, USA) from January 1, 2010 to October 31, 2021.The PearlDiver database contains blinded demographic and claims information for over 92 million patients from all payer types.The provided data were deidentified and compliant with the Health Insurance Portability and Accountability Act.The database is organized based on the diagnostic codes within the International Classification of Diseases (ICD)-9 and ICD-10 classifications.Claims for procedures are classified under the Current Procedural Terminology (CPT) codes.Finally, prescription claims for prescription and brand-name drugs are classified according to the Uniform System of Classification (USC) and the U.S. Food and Drug Administration (FDA) National Drug Code (NDC) Directory.The data are compiled and adjudicated by PearlDiver using these specific ICD, CPT, USC, and NDC codes billed to the insurance company by the physician or provider in question.We selected this database for analysis to ensure that a large population of patients was analyzed, thereby reducing the risk of Type II error.

Eligibility criteria
We considered all patients who were managed with minimally invasive spinal techniques, including PILD or Conclusions: Minimally invasive procedures may provide an additional option to treat symptomatic lumbar spinal stenosis in patients who are poor surgical candidates or who do not desire open decompression; however, there still exists a subset of patients who will require subsequent surgical decompression.Factors such as gender and prior surgical decompression increase the likelihood of subsequent surgery, while older age, obesity, and a higher Charlson Comorbidity Index score reduce it.These findings aid in selecting suitable surgical candidates for better outcomes in the elderly population with lumbar spinal stenosis.

Study outcomes
Our primary outcome was to evaluate the incidence and time to surgical decompression within 5 years following selected minimally invasive techniques.Secondary outcomes included (i) an evaluation of the time period in which the majority of patients required additional surgical decompression, and (ii) odds of demographic/ comorbid factors in patients who underwent additional decompression surgery.

Evaluated comorbid conditions
1][12][13] Major demographic variables assessed included age ≥65 years, gender, and a diagnosis of obesity.Older age was selected as a surrogate to represent patients who may be poor surgical candidates due to comorbid conditions with an associated increase in postoperative morbidity. 14Gender was selected to help capture differences in spinal anatomy that may occur between males and females, with women possibly having a higher propensity for more advanced spine disease. 15,16Obesity was chosen as a marker for patients who may be considered poor surgical candidates due to an increased risk of postoperative complications. 17valuated comorbid conditions included tobacco use and alcohol abuse at any time prior to the minimally invasive decompression, and persistent opioid use for 1 year prior to the minimally invasive decompression.][20] Persistent opioid use was considered to identify patients who may have altered pain perception and function, thus impacting the perceived success of a minimally invasive approach and increasing the likelihood of progression to surgical management. 21,22Certain spinal pathologies were also evaluated that may predispose patients for the failure of minimally invasive techniques or are associated with more severe stenosis, including lumbar spondylosis with myelopathy, lumbar spondylolisthesis, lumbar radiculopathy, and a history of prior surgical intervention (i.e., surgical decompression, hemilaminectomy, laminotomy, or laminectomy).
Finally, we also assessed the odds of having a Charlson Comorbidity Index (CCI) of two or greater in patients who underwent additional surgical decompression.The CCI assesses the degree of patient morbidity based on 17 pre-defined comorbid conditions that are weighted from one to six, with six indicating severe morbidity. 23igher CCI scores are also predictive of increased mortality risk. 23The CCI cut-off of ≥2 was selected to identify patients with a high risk of mortality after surgery and possibly being poor-surgical candidates.

Measurement of outcome data
Demographic data were presented for each group of patients who underwent minimally invasive decompression and those who subsequently required surgical decompression.Gender and age were dichotomized to male or female and ≥65 years or <65.All comorbid conditions and spinal pathologies were dichotomized as having the presence or absence of a diagnosis claim.CCI scores were dichotomized as having a total score of ≥2, or <2.Time to subsequent surgical decompression was reported in median days with an interquartile range (IQR).

Statistical analyses
Differences in the characteristics of patients who underwent minimally invasive techniques with additional surgical decompression versus those that did not require additional surgery were analyzed using a chi-squared test.Associated factors for requiring additional surgical decompression were assessed using univariable logistic regression models inclusive of demographic, comorbid, and spinal pathology variables.Univariable model results were reported as odds ratios (OR) and 95% confidence intervals (CIs).All covariates from this analysis with a p-value ≤0.10 were included in a preliminary multivariable logistic regression model.Covariates were then removed from the multivariable model in a stepwise manner if they were not significantly associated with requiring subsequent surgical decompression at a p-value of 0.05, after adjusting for other variables in the model.Any confounding effect of the removed covariates was further assessed by examining the OR for the remaining covariates in the model and ensuring that they did not change by more than 15%.The final parsimonious model consisted of only those variables with a p-value <0.05, after determining that none of the removed variables were confounders.The fit of the final model was assessed using the Hosmer-Lemeshow goodness-of-fit test.Hypothesis testing was conducted at a 5% type I error rate (alpha = 0.05).

Subgroup analysis
For the primary outcome (incidence of subsequent surgical decompression), data were further dichotomized into the type of minimally invasive technique used, namely, PILD or ISD placement.Further secondary regression analysis was not performed on each individual subgroup due to an anticipated low number of patients and reduced power.

Data management
All data were analyzed in aggregate form using R Statistical Software version 4.1.0in the PearlDiver Software.

Incidence of subsequent spinal decompression
Across all minimally invasive interventions, the incidence of additional spinal decompression surgery was 6.56% (346 of 5278 patients) within 5 years of receiving the primary intervention.Among these, 56.06% (194 patients) of claims for spinal decompression occurred within the first year after minimally invasive management.The median time [IQR] of undergoing subsequent spinal decompression was 326 days [140, 776].Further subgroup analysis revealed that 82% of patients (284 of 346 patients) who received subsequent surgical decompression had a claim for initial PILD.

Characteristics of patients undergoing subsequent spinal decompression
Across patients who required additional spinal decompression surgery within the 5 years of a minimally invasive intervention, a significantly higher proportion were male (p = 0.001), and a significantly lower proportion of patients were aged 65 years and older (p < 0.001) or with a diagnosis of obesity (p < 0.001).Further, a significantly lower proportion of patients had a CCI score of three or greater (p < 0.001).In contrast, for spinal pathologies, a higher proportion of patients had a claim for a prior surgical decompression before receiving a minimally invasive intervention followed by surgical decompression (p < 0.001).None of the other evaluated factors, comorbidities, or spinal pathologies were found to differ between both cohorts.A full list of evaluated covariates can be viewed in Table 1.

Factors associated with claims for additional spinal decompression following minimally invasive intervention
Univariable regression modeling revealed that certain factors were associated with lower or higher odds of requiring additional surgical decompression.Specifically, factors associated with reduced OR [95% CI] of progressing to additional surgical decompression included (i) age 65 years and older by 38% (OR: 0.62 [0.49, 0.79], p < 0.001); (ii) obesity by 41% (OR: 0.59 [0.46, 0.77], p < 0.001); or (iii) a CCI score of three or greater by 36% (OR: 0.64 [0.50, 0.81], p < 0.001).In contrast, covariates associated with a greater OR [95% CI] of progressing to surgical decompression included (i) male gender by 42% (OR: 1.42 [1.14, 1.78], p = 0.001) or (ii) having a claim for T A B L E 1 Characteristics of patients who received a minimally invasive procedure versus those that progressed to subsequent surgical decompression.

DI SC US SION
Overall, our results suggest that minimally invasive procedures to treat symptomatic lumbar spinal stenosis may be a viable option for patients who may be poor surgical candidates for decompression.However, a small subset of patients (6.56%) still require subsequent surgical decompression, the majority of which occurs within the first year of the minimally invasive approach.Factors that appear to be associated with the need for additional surgical decompression include having a claim for prior surgical decompression and male gender.In contrast, a reduction in odds appears to be seen with older age, obesity, and higher CCI scores.Taken together, our 5-year analysis supports the durability of minimally invasive procedures to treat symptomatic lumbar spinal stenosis on a case-by-case basis.The results of this database analysis are concordant with prior single-center studies highlighting the low percentage of patients requiring surgical decompression after initially undergoing PILD [24][25][26] and/or ISD placement. 26,27It is important to note that this low percentage is not necessarily indicative of the success of minimally invasive approaches, but may rather signify the unsuitability of these groups of patients for surgical decompression.Furthermore, these groups of patients may have also opted for a trial of minimally invasive methods prior to consideration of surgical decompression.While our observations do suggest that physicians should consider offering such minimally invasive procedures for symptomatic lumbar spinal stenosis, select patients with a prior history of surgical decompression should be counseled on the possibility of additional repeat surgery.This is likely due to the natural history of progression of lumbar spinal stenosis, which may warrant repeat surgical decompression at the same operating levels or previously unaffected adjacent levels. 28,29everal associated factors identified through regression modeling warrant discussion.First, reasons supporting gender as an associated factor remain unclear.Studies have highlighted a higher propensity for more advanced spine disease in women. 15,16However, a prior study assessing risk factors of failure with minimally invasive surgical lumbar decompression did not identify an association with gender. 30While the literature remains sparse on gender-related outcomes after spine surgery, recent studies have supported that the degree of improvement after spine surgery is equivalent among men and women. 15,31Second, our findings should also be interpreted in light of the poor initial surgical candidacy of select patients.Older age and CCI scores of three or greater were associated with reduced odds of progressing to surgical decompression.3][34] Due to this, both surgeons and geriatric patients may be reluctant to pursue surgical decompression and instead may prefer to continue conservative therapy even in situations where this strategy has proven ineffective. 35While age as an independent risk factor has been studied, it is often confounded by the high prevalence of medical comorbidities in this age group, yet the association remained significant in our multivariable analysis. 36he association between CCI scores of three or greater and reduced odds for surgical decompression is also consistent with this finding since overall morbidity burden is predictive of mortality risk. 23Similar to the geriatric population, patients with higher CCI scores may elect to proceed with conservative or minimally invasive approaches after discussion with the surgical team.Finally, obesity also poses a unique challenge for open surgical decompression, which is likely why these patients were more commonly found to receive minimally invasive approaches.The risk obesity poses in the surgical patient for peri-operative complications has well been established, 37,38 and may limit candidacy for open surgical decompression.
Interestingly, our subgroup analysis revealed that a large majority of patients who received subsequent surgical decompression had a claim for PILD as opposed to an ISD.PILD solely decompresses the posterior ligamentum flavum.In contrast, an ISD not only increases the central canal area but also increases space within the lateral recess and intervertebral foramen, thereby treating both central canal and neuroforaminal stenosis.Since patients with lumbar spinal stenosis can have varying combinations of these, it may be expected that a lower incidence of subsequent surgical decompression was observed in patients receiving an ISD.

Limitations
Our analysis comes with limitations.First, although we identified patients who had lumbar spinal stenosis with neurogenic claudication, we were unable to identify details on the severity or type of stenosis (e.g., central canal, lateral recess, neuroforaminal stenosis).Second, ISDs have the same CPT code as interlaminar devices, and this could impact our overall findings.Third, an inherent limitation of database analysis is that complete certainty cannot be placed into lumbar spinal stenosis being the primary indication for surgical lumbar decompression.Although the database was queried for surgical CPT codes that are commonly used to treat lumbar stenosis, such as decompression, hemilaminectomy, laminotomy, and laminectomy, these same surgeries are also frequently offered for spine pathology unrelated to lumbar spinal stenosis.Fourth, patients who received additional surgical decompression may have received it at a different level(s) than where the minimally invasive intervention was performed.Fifth, a small subset of patients underwent both PILD and ISD placement, and these patients were not stratified in this analysis due to concerns of statistical power.It is plausible that the factors associated with surgical decompression differ between these approaches.Sixth, certain covariates evaluated in our regression model had OR that exceeded or were below 1.0, yet these were still non-significant in our overall model due to wide confidence intervals.This highlights that many more patients are likely required to further understand the characteristics of patients undergoing PILD or ISD placement.Finally, although the data from PearlDiver is adjudicated, claims can still be incorrectly recorded.Further, the database is limited to a subset of all-payer claims, which while believed to be an accurate representation of the entire population, still may have a low-capture rate for the investigated minimally invasive procedures and may impact our results.

CONC LUSIONS
Minimally invasive options (PILD or ISD) may provide an adequate option to treat symptomatic lumbar spinal stenosis in patients who may be considered poor surgical candidates or who do not desire initial operative decompression; however, there still exists a subset of patients who will require subsequent surgical decompression.Factors such as male gender and a history of prior surgical decompression are associated with an increased likelihood of subsequent surgery, while age 65 years and older, a diagnosis of obesity, and higher Charlson Comorbidity Index scores are associated with a reduced likelihood.These results can help inform the selection of appropriate surgical candidates and improve patient outcomes in the growing elderly population with lumbar spinal stenosis.Further research is needed to better understand the long-term benefits and outcomes of minimally invasive techniques in the management of this condition.
Significant factors for subsequent surgical decompression, as found using multivariate logistic regression modeling.(OR: 0.71 [0.56, 0.91], p = 0.007), respectively, after adjusting for other covariates in the model.The remaining covariates were not found to be associated factors in the final parsimonious model.The Hosmer-Lemeshow goodness-of-fit test was non-significant (p = 0.064), indicating adequate model fit.