Introduction

Spinal fusion introduces an immobile construct subject to greater bending moments and shear stresses at adjacent spinal segments, resulting in accelerated degeneration and symptoms of back and/or leg pain [5, 19, 1, 14]. Symptomatic manifestation of the degenerated adjacent segments, termed adjacent segment disease (ASD), may require decompression of the adjacent segment and extension of the fusion construct [1].

The minimally invasive transforaminal lumbar interbody fusion (MIS-TLIF) and lateral lumbar interbody fusion (LLIF) are two well-established surgical techniques utilized to treat adjacent segment disease [1, 22, 11, 9]. The MIS-TLIF involves a posterior approach for direct decompression of the neural elements, while LLIF involves a transpsoas approach for indirect decompression and direct access to the intervertebral disc [12]. Direct access to the disc space allows for larger interbody cage placement and, hence, greater correction of sagittal imbalance, coronal imbalance, and restoration of disc height [22, 11, 12].

Although there are studies examining direct versus indirect decompression in treating adjacent segment disease, these studies fail to directly compare the postoperative clinical outcomes in patients undergoing MIS-TLIF versus LLIF. Investigation into the clinical outcomes between surgical techniques for ASD may aid in surgical decision-making and management of postoperative expectations for patients. We aim to investigate this gap by querying patients undergoing MIS-TLIF or LLIF for adjacent segment disease in a single-surgeon database.

Methods

Patient population

Patient consent and Institutional Review Board approval (ORA #14051301) were acquired preceding the start of the current study. Patient records were queried from July 2009 to September 2022. Patients were included if they underwent MIS-TLIF or LLIF for adjacent segment disease. Patients were excluded if they were diagnosed with neoplasm, infection, or acute fracture. Patients were favored to undergo MIS-TLIF if they required direct decompression, had abnormal vascular anatomy, or had obstructive iliac crests. LLIF was favored if patients required greater correction of sagittal and/or coronal imbalance. All patients undergoing LLIF underwent additional posterior instrumentation.

Data collection

Patient demographics, perioperative characteristics, and inpatient complications were recorded. Age, gender, ethnicity, insurance type, body mass index (BMI), smoking status, hypertensive status, diabetic status, American Society of Anesthesiologists (ASA) classification, and Charlson Comorbidity Index (CCI) score were the demographic information collected. The number of contiguous fused levels, fusion level, operative time, estimated blood loss, length of stay, postoperative day (POD) 0 and 1 VAS pain, POD 0 and 1 narcotic consumption, and day of discharge were the perioperative characteristics recorded. Acute renal failure, altered mental status, postoperative anemia, arrhythmia, aspiration or reintubation, atelectasis, fever of unknown origin, ileus, urinary incontinence, nausea/vomiting, pleural effusion, pneumonia, pulmonary embolism, urinary retention, urinary tract infection, and venous thromboembolism were inpatient complications recorded.

The primary outcomes collected in this study were PROMs evaluating physical function, pain, and disability. The SF-12 PCS was utilized to evaluate physical function. The VAS back and VAS leg evaluated back and leg pain, respectively. The ODI was utilized to evaluate disability. These PROMs were collected at preoperative and postoperative 6-week, 12-week, 6-month, and 1-year time points.

Statistical analysis

Patients undergoing surgery for adjacent segment disease were stratified into two cohorts based on whether they underwent MIS-TLIF or LLIF. MCID achievement was determined through comparison of PROM improvement to established thresholds of 2.5 for SF-12 PCS, 2.1 for VAS back, 2.8 for VAS leg, and 14.9 for ODI [16, 15]. Categorical and continuous variables were compared between cohorts using the Fisher exact and Mann–Whitney U tests, respectively. After comparing the demographic characteristics between cohorts, propensity score-matched for age, BMI, smoking status, hypertensive status, and CCI scores. Postoperative improvement of PROMs was calculated using Wilcoxon ranked sign tests. All data were analyzed using Stata 17.0 (StataCorp LP, College Station, TX, USA). An alpha value of 0.05 was utilized to determine statistical significance.

Results

Prior to propensity score matching, a total of 69 patients were identified. There were 54 total patients, with 22 and 32 patients undergoing MIS-TLIF and LLIF, respectively, after matching for demographic characteristics. No significant differences between cohorts were calculated after propensity score matching. Demographic characteristics were skewed towards white (73.6%) patients. The proportion of patients receiving private insurance (44.4%) and workers’ compensation (44.4%) were equal. For self-identified gender, 37.0% were female. Patients typically had a moderate comorbidity burden, with 20.4%, 37.0%, and 16.7% with positive smoking, hypertensive, and diabetic statuses, respectively. The mean age, BMI, and CCI scores were 56.0 years, 32.0 kg/m2, and 2.2, respectively. Unmatched demographics are in Table 1, while matched demographics are in Table 2.

Table 1 Unmatched patient demographics
Full size table
Table 2 Matched patient demographics
Full size table

For perioperative characteristics, the fusion level and POD 0 narcotic consumption were significantly different between MIS-TLIF and LLIF cohorts (p ≤ 0.009, both). For fusion level, LLIF patients tended to undergo fusion at L3–L4 (37.5%), L2–L3 (28.1%), or L4–L5 (15.6%). MIS-TLIF patients tended to undergo fusion at L4–L5 (54.6%), L5–S1 (27.3%), or L3–L4 (18.2%). Most patients in either surgical technique underwent single-level fusion (90.7%). One patient underwent 3-level LLIF for degenerative scoliosis at L2–S1 and had a previous ALIF at L5–S1. For POD 0 narcotic consumption, MIS-TLIF patients utilized significantly greater narcotic consumption in OME of 116.7 compared to 59.8 in LLIF (p = 0.009). No other significant differences were noted in perioperative characteristics. The mean operative time, estimated blood loss, and length of stay were 170.4 min, 91.8 mL, and 46.9 h, respectively. Perioperative characteristics are in Table 3.

Table 3 Perioperative characteristics
Full size table

No significant differences were reported between cohorts for inpatient postoperative complications. Overall, most patients’ hospital courses were complicated by fever of unknown origin (20.4%), urinary retention (13.0%), or nausea/vomiting (7.4%). One patient undergoing MIS-TLIF suffered from altered mental status, while a separate patient undergoing MIS-TLIF suffered from atelectasis. Postoperative inpatient complications are in Table 4.

Table 4 Inpatient complications
Full size table

Patients undergoing MIS-TLIF reported significant postoperative improvement in 6-week to 1-year VAS back, 6-week to 1-year VAS leg, and 6-month to 1-year ODI (p ≤ 0.035, all). Patients undergoing LLIF reported significant postoperative improvement in 6-month SF-12 PCS, 6-week through 1-year VAS back, 12-week through 6-month VAS leg, and 6-month to 1-year ODI (p ≤ 0.035, all). No significant preoperative or postoperative differences in PROMs were calculated between cohorts. Greater than 50% of patients achieved MCID in MIS-TLIF for all domains. Greater than 50% of patients achieved MCID in LLIF for SF-12 PCS, VAS back, and VAS leg. No significant differences in MCID achievement were calculated between cohorts. Assessment of raw PROM scores is in Table 5, and MCID achievement rates are in Table 6.

Table 5 Mean patient reported outcome measures
Full size table
Table 6 MCID achievement
Full size table

Discussion

Independent of surgical technique, patients undergoing MIS-TLIF or LLIF for adjacent segment disease reported significant postoperative improvement in pain and physical function. Patients undergoing LLIF also demonstrated significant postoperative improvement in physical function. No significant differences were calculated between cohorts for postoperative PROM scores or MCID achievement rates. Either MIS-TLIF or LLIF are effective surgical interventions for adjacent segment disease.

Biomechanical finite element analysis (FEA) models have demonstrated considerably increased shear stress at the adjacent segments following a lumbar fusion [4, 8, 10]. One study calculated the shear stress at the upper adjacent segment to be as high as 43% in the fused segment model [4]. A separate article utilized FEA to determine significantly increased stress at the nucleus pulposus and annulus fibrosus after posterior lumbar interbody fusion (PLIF) and as disc degeneration progressed [8]. One additional study formed a FEA model with LLIF and bilateral pedicle screw placement for treatment of adjacent segment disease [10]. The authors calculated that placement of the interbody cage with bilateral pedicle screws in the adjacent segment provides sufficient structural stability and protective effects on the interbody cage placement compared to standalone cage placement [10]. Overall, these FEA studies demonstrate the increased stress applied to the adjacent segments due to an immobile construct and the efficacy in extending the immobile construct for adjacent segment disease.

In addition to these biomechanical studies, previous articles have examined the clinical and radiographic outcomes of utilizing direct and indirect decompression for adjacent segment disease. One study examining standalone LLIF versus laminectomy and posterior lateral fusion demonstrated that equivalent postoperative clinical outcomes in VAS back, VAS leg, and ODI scores, with significantly higher segmental lordosis (SL), lumbar lordosis (LL), and disc height in patients undergoing LLIF [11]. A separate article demonstrated that patients undergoing oblique lateral interbody fusion (OLIF) versus TLIF for adjacent segment disease had similar postoperative clinical outcomes in VAS back, VAS leg, and ODI, with the OLIF cohort experiencing greater improvement in sagittal balance [9]. A separate study examining postoperative clinical outcomes in OLIF versus PLIF reported clinical postoperative improvement in VAS, ODI, and Japanese Orthopedic Association (JOA) scores for the treatment of adjacent segment disease [21]. However, patients undergoing OLIF demonstrated superior VAS outcomes [21]. One case series demonstrated that patients undergoing LLIF for adjacent segment disease had significant improvements to disability and physical function outcomes, with improvements in SL, LL, and disc height [22]. Although none of these studies directly evaluated MIS-TLIF versus LLIF, these overall findings indicate that both indirect and decompression methods are effective in the treatment of adjacent segment disease in appropriately selected patients.

In this present study, patients undergoing MIS-TLIF for adjacent segment disease underwent surgical intervention at L4–5 and had higher POD 0 narcotic consumption compared to LLIF. Though the L4–5 intervertebral disc space is accessible through LLIF, the lumbar plexus courses more ventrally compared to the more cranial levels and may obscure up to 50% of the safety corridor [18, 13, 3]. Furthermore, abnormal vascular anatomy and high-riding iliac crests may preclude selection of LLIF at this level [7]. For postoperative pain management, patients undergoing MIS-TLIF required higher POD-0 narcotic consumption. One explanation may be due to the extensive retraction and dissection of the paraspinal muscles in TLIF [17, 20, 6]. As such, patients undergoing MIS-TLIF may require greater immediate postoperative pain management.

Patients undergoing either MIS-TLIF or LLIF for adjacent segment disease reported significant postoperative improvement for pain and disability outcomes in the current study. Patients demonstrated similar postoperative PROM scores and MCID achievement in pain and disability outcomes. Our results correspond to the majority of the previously cited literature [22, 11, 9, 21]. That is, patients undergoing direct versus indirect decompression demonstrated similar postoperative clinical outcomes in pain and disability outcomes [11, 9]. Overall, our findings suggest that either MIS-TLIF or LLIF are effective surgeries in treating adjacent segment disease.

For physical function, only patients undergoing LLIF reported significant postoperative improvement in physical function. Although patients undergoing MIS-TLIF did not report significant postoperative improvement in physical function, the mean SF-12 PCS score improved during the postoperative period. As such, the lack of significant improvement in the MIS-TLIF cohort may be the result of insufficient power. Still, the lack of significant differences between cohorts indicates the similar efficacy between MIS-TLIF and LLIF for adjacent segment disease.

This present study has several limitations. The lower patient numbers in each cohort may introduce insufficient power in our longitudinal analysis due to loss of follow-up. Furthermore, loss to follow-up introduces selection bias. One article studied this phenomenon and reported superior postoperative outcomes in patients who were initially lost to follow-up in clinic compared to patients who continued to follow in clinic [2]. Additionally, PROMs are inherently subjective and therefore are susceptible to response bias. As the most important aspect of patient selection was individual clinical evaluation, surgical technique was selected based on the anatomical characteristics highlighted in the Patient population. As such, the study design precludes the ability to entirely negate biases in patient selection. However, as the findings of this study indicate non-inferiority of either surgical technique in appropriately indicated patients, such biases may not have negatively influenced these results. Usage of a single-surgeon database limits the generalizability of these findings, as our cohort was typically White males. Despite this demographic predominance, the patients in our study typically reported a higher comorbidity burden. Future investigations may incorporate a multicenter analysis to aid in greater patient collection and generalizability.

Conclusion

Patients undergoing either MIS-TLIF or LLIF for adjacent segment disease demonstrated significant postoperative improvement in pain and disability outcomes. Patients undergoing LLIF additionally demonstrated postoperative improvement in physical function. Patients undergoing either MIS-TLIF or LLIF demonstrated equivalent postoperative improvement and MCID achievement rates in physical function, pain, and disability outcomes. MIS-TLIF and LLIF are similarly effective in treating adjacent segment disease.