Elective Lumbar Decompression With Fusion Is Associated With Increased Mortality Compared With Decompression Alone: A National Propensity-Matched Analysis of Aging Cohorts

Abstract

Study Design

Propensity–matched cohort study.

Objectives

To evaluate short/mid-term safety and mortality of elective lumbar decompression alone (LD) versus lumbar decompression with fusion (LDF) for degenerative lumbar disease in octogenarians and the aging patient.

Methods

The TriNetX research network was queried for adults with degenerative lumbar disease who underwent LD or LDF (2012–2025). Patients with deformity, instability, spondylolisthesis, or congenital malformation were excluded. Patients were stratified by decade of life (60–69, 70–79, 80–89 years) and matched 1:1 for demographics, steroid use, lipoprotein disorders, and modified Frailty Index-5 comorbidities. Outcomes were compared at 30 days, 90 days, 1 year, and 2 years. The primary outcome was mortality; secondary outcomes included medical and surgical complications.

Results

Across all age decades, LDF was associated with >5-fold higher 30-day mortality (OR 5.16–6.34, all p<0.0001) and >3.5-fold higher 90-day mortality (OR 3.83–6.35, all p<0.0001) compared to LD. In octogenarians, 30-day mortality was significantly higher with LDF versus LD (4.19% vs. 0.68%, p<0.0001). At 90 days, LDF was associated with higher rates of wound infection, sepsis, respiratory complications, DVT/PE, stroke, revision surgery, ED visits, and readmissions (all p<0.03). By 2 years, LDF was associated with higher mortality, revision surgery, and chronic pain (all p<0.009).

Conclusions

In this large, frailty-matched national analysis, LDF was consistently associated with higher mortality and morbidity compared to LD for degenerative lumbar disease. Octogenarians demonstrated six-fold higher 30-day mortality with LDF, underscoring the importance of careful patient selection for elective lumbar fusion.

Level of Evidence

III.

Introduction

The percentage of the population over the age of 65 in the United States has continued to rise over the last few decades, increasing from 12% in 2000 to a projected 20% by 2030. ¹ The prevalence of degenerative lumbar spine disease has risen in parallel with this increased expected life expectancy. Recent estimates from Medicare data found the prevalence of spinal degenerative disease was 27.3%, increasing with age. ² Rates of lumbar fusion have increased substantially, with elective lumbar fusions for degenerative spine disease rising by 276% in the United States from 2002 to 2014. The number of octogenarians receiving fusions has also increased in recent years.^3-5 With advancing age, however, patients often experience greater physiologic fragility and reduced physiologic reserve, leading to increased susceptibility to perioperative risk. ⁶ Despite these considerations, modern spine surgery increasingly treats older patients with limited physiologic reserve.

The decision between decompression alone (LD) and decompression with fusion (LDF) remains one of the most common decisions in lumbar surgery. Fusion is most clearly indicated in cases of deformity, instability, high-grade spondylolisthesis, and iatrogenic destabilization.^7,8 However, among patients with degenerative disease without instability, substantial practice variation in surgical management persists, and fusion may represent an elective augmentation rather than a strict necessity.^9-11 A large body of literature compares LD and LDF and consistently demonstrates that fusion procedures are associated with longer operative times, greater blood loss, higher infection risk, increased cardiopulmonary stress, and prolonged hospital stay.^12-16 Prior studies also report higher complication rates with fusion compared to decompression alone, particularly in older adults.^17-19 Systematic reviews report that fusion procedures are associated with a relative risk of complications approximately 1.77 times that of decompression alone, with similar findings for major complications across multiple studies.^20,21 However, most comparative studies focus on relative risk estimates rather than absolute mortality differences, and the absolute mortality risk attributable to fusion in patients who may be candidates for decompression alone remains poorly defined.

Surgeons typically counsel patients using absolute rather than relative risk, and even small increases in mortality may meaningfully influence surgical decision-making in elderly patients. ²² Although frailty has been associated with worse perioperative outcomes and mortality in spinal surgery cohorts, few studies report age-stratified mortality in frailty-matched patients with degenerative lumbar disease in whom fusion is not a mandatory indication.^23,24 Therefore, the purpose of the present study was to compare morbidity and absolute mortality between LD and LDF in patients with degenerative lumbar disease without instability. We hypothesized that fusion would be associated with higher complication rates and increased mortality, with the absolute mortality difference between procedures increasing with age.

Methods

Study Design

A retrospective cohort study was conducted using the TriNetX national health research network database (2012-2025), which aggregates de-identified electronic health record data from over 150 million patients across multiple healthcare organizations. Because all data within TriNetX are de-identified, Institutional Review Board was not required. All data was extracted on February 16, 2026. The study aimed to compare matched cohorts of patients in their 7^th, 8^th, and 9^th decades of life who underwent either a lumbar decompression (LD) or a lumbar decompression and fusion (LDF).

Cohort Definitions

Both the lumbar decompression and the lumbar decompression and fusion cohorts were defined in the TriNetX system using Current Procedural Terminology (CPT) and International Classification of Diseases, Revision 10 (ICD-10). The lumbar decompression and fusion cohort was defined by patients having a primary degenerative indication (ICD-10-CM-M47, M48.05/06/07, M51, or M54.15/16/17) which occurred at least one day before the patient’s first ever instance of a lumbar fusion (CPT-2253(3/4), 225(58/85), 2261(2/4), or 2263(0/2/3/4)) without any instance of a diagnosis for deformity, instability, or malformation that would require fusion (ICD-10-CM-M40, M41, M43.(0/1/5/8/9), Q76) (Appendix 1).

The lumbar decompression cohort was defined by patients having a primary degenerative indication (ICD-10-CM-M47, M48.05/06/07, M51, or M54.15/16/17) without a previous lumbar fusion (CPT-2253(3/4), 225(58/85), 2261(2/4), or 2263(0/2/3/4)) or instrumentation (CPT-1004126, 1004130, 2284(0/1)) which occurred at least one day before the patient’s first ever instance of a lumbar decompression (CPT-63030, 6304(7/8)) without any instance of a diagnosis for deformity, instability, or malformation that would require fusion (ICD-10-CM-M40, M41, M43.(0/1/5/8/9), Q76) or a prior lumbar fusion (CPT-2253(3/4), 225(58/85), 2261(2/4), or 2263(0/2/3/4)) or instrumentation (CPT-1004126, 1004130, 2284(0/1)) (Appendix 1).

To stratify by age a cohort was created for lumbar decompression such that only patients who received their initial surgery between ages 80-89 were included. A cohort of patients receiving initial surgery between ages 70-79 and 60-69 were also created. This process was repeated for the lumbar decompression and fusion cohort.

Propensity Score Matching

Baseline characteristics of the unmodified cohorts were statistically different. As a result, the cohorts were balanced using the TriNetX propensity score matching (PSM) system. The internal PSM system employs a 1:1 nearest-neighbor match process. Patients were matched for sex, race, BMI, lipoprotein metabolism disorders, long-term steroid use, and frailty characteristics derived from the modified Frailty Index-5 questionnaire (heart failure, hypertension, dependency, diabetes, and chronic obstructive pulmonary disease).

The TriNetX interface only permits pairwise comparisons. The two cohorts, LD and LDF, were compared at each age decade such that patients in their 60s who received an LD were compared to those in their 60s who received an LDF. Similar comparisons were performed for patients in their 70s and 80s.

Outcome Definitions

The outcomes of interest in this study at 30 days and 90 days postoperatively included: all-cause mortality, postoperative wound infection, sepsis, respiratory complications, cardiac events, deep vein thrombosis (DVT)/pulmonary embolism (PE), stroke, revision surgery, emergency service use, and hospital readmission.

The outcomes of interest in this study at 1 year and 2 years postoperatively included: all-cause mortality, revision surgery, opioid use disorder, and chronic pain. Some outcomes were not able to be compared between groups but provide insight into the success of a decompression surgery or a decompression and fusion surgery. Non-comparative descriptive statistics for these outcomes were included. For decompression these outcomes included: post-laminectomy syndrome. For decompression and fusion these outcomes included: pseudoarthrosis, mechanical complications of a prosthesis, and infectious/inflammatory complications of a prosthesis.

Outcomes were defined by CPT and ICD-10 codes, and the definitions for each outcome are listed in Appendix 2 and Appendix 3.

Statistical Analysis

TriNetX provides an internal Compare Outcomes analytic option, which was used for the purposes of this study. This analysis provides the number of patients in each cohort and the number with the outcome of interest. For each outcome the analysis provides the risk, odds ratio, 95% confidence interval (CI), and p-value drawn from a two-sided z-test of proportions. A p-value < 0.05 was considered statistically significant.

Results

Demographics

Prior to the TriNetX Propensity Score Match (PSM) a total of 11,849 LDF patients and 23,011 LD patients were identified in their 60s. 7,675 LDF patients and 16,714 LD patients were identified in their 70s, and 1,514 LDF patients and 3,300 LD patients were identified in their 80s. Following the TriNetX PSM, the count of patients receiving each operation was the same, such that 11,433 patients were included in the 60-69 year old cohort, 7,464 in the 70-79 year old cohort, and 1,487 in the 80-89 year old cohort. Cohorts were similar with respect to matched covariates except for rates of COPD (7.2% LDF versus 7.8% LD, p = 0.0533) and BMI < 20 (1.4% LDF versus 0.8% LD, p < 0.0001) in 60-69 year olds, and rates of long-term steroid use (5.2% LDF versus 4.2% LD, p = 0.0067) and BMI 20-29.9 (8.4% LDF versus 9.4% LD, p = 0.0309) in 70-79 year olds (Table 1).

Table 1.

Demographics of Patients Receiving Lumbar Decompression vs Lumbar Decompression and Fusion by Age

​	Demographics	Patients age 60-69 (%), N=11,433	P-value	Patients age 70-79 (%), N=7,464	P-value	Patients age 80-89 (%), N=1,487	P-value
LDF	White	8905 (77.9%)	0.8109	6185 (82.9%)	0.794	1241 (83.5%)	0.5498
LD	​	8920 (78.0%)	​	6197 (83.0%)	​	1253 (84.3%)	​
LDF	Female	4366 (38.2%)	0.4384	2833 (38.0%)	0.6613	551 (37.1%)	0.5199
LD	​	4423 (38.7%)	​	2859 (38.3%)	​	568 (38.2%)	​
LDF	Black or African American	1427 (12.5%)	0.9362	600 (8.0%)	0.9281	108 (7.3%)	0.7267
LD	​	1431 (12.5%)	​	603 (8.1%)	​	113 (7.6%)	​
LDF	Essential (primary) hypertension	7006 (61.3%)	0.2477	5118 (68.6%)	0.4161	1140 (76.7%)	0.2921
LD	I10	7091 (62.0%)	​	5164 (69.2%)	​	1164 (78.3%)	​
LDF	Disorders of lipoprotein metabolism and other lipidemias	5715 (50.0%)	0.1217	4479 (60.0%)	0.4317	1015 (68.3%)	0.476
LD	E78	5832 (51.0%)	​	4526 (60.6%)	​	1033 (69.5%)	​
LDF	Diabetes mellitus	3319 (29.0%)	0.4944	2550 (34.2%)	0.863	540 (36.3%)	0.7319
LD	E08-E13	3366 (29.4%)	​	2560 (34.3%)	​	549 (36.9%)	​
LDF	Body mass index [BMI]	2854 (25.0%)	0.8544	1962 (26.3%)	0.6957	441 (29.7%)	0.8412
LD	Z68	2842 (24.9%)	​	1941 (26.0%)	​	446 (30.0%)	​
LDF	Heart failure	1746 (15.3%)	0.3901	1223 (16.4%)	0.8253	315 (21.2%)	0.6847
LD	I50	1793 (15.7%)	​	1233 (16.5%)	​	306 (20.6%)	​
LDF	Other chronic obstructive pulmonary disease	820 (7.2%)	0.0533	851 (11.4%)	0.8774	200 (13.5%)	0.7486
LD	J44	897 (7.8%)	​	857 (11.5%)	​	206 (13.9%)	​
LDF	Problems related to care provider dependency	663 (5.8%)	0.2166	394 (5.3%)	0.716	102 (6.9%)	0.4577
LD	Z74	620 (5.4%)	​	404 (5.4%)	​	92 (6.2%)	​
LDF	Long term (current) use of steroids	544 (4.8%)	0.8518	385 (5.2%)	0.0067	102 (6.9%)	0.4361
LD	Z79.5	538 (4.7%)	​	315 (4.2%)	​	113 (7.6%)	​
LDF	Body mass index [BMI] 19.9 or less, adult	158 (1.4%)	< 0.0001	88 (1.2%)	0.3461	32 (2.2%)	0.3492
LD	Z68.1	95 (0.8%)	​	76 (1.0%)	​	25 (1.7%)	​
LDF	Body mass index [BMI] 20-29, adult	779 (6.8%)	0.6731	625 (8.4%)	0.0309	182 (12.2%)	0.0939
LD	Z68.2	763 (6.7%)	​	700 (9.4%)	​	213 (14.3%)	​
LDF	Body mass index [BMI] 30-39, adult	1302 (11.4%)	0.4563	938 (12.6%)	0.7864	251 (16.9%)	0.6272
LD	Z68.3	1338 (11.7%)	​	949 (12.7%)	​	261 (17.6%)	​
LDF	Body mass index [BMI] 40 or greater, adult	254 (2.2%)	0.1039	230 (3.1%)	0.4704	62 (4.2%)	0.075
LD	Z68.4	219 (1.9%)	​	215 (2.9%)	​	44 (3.0%)	​

[bold] indicating significance.

30-and 90-Day Outcomes

Across all decades, LDF patients demonstrated higher 30-day mortality (60s OR 5.16, 70s OR 5.31, 80s OR 6.34, all p < 0.0001) compared to LD patients. The absolute mortality rate in 80-89 year olds receiving a lumbar decompression and fusion was 4.19%, compared to 0.68% in lumbar decompression patients (p < 0.0001). Absolute mortality rate was higher in LDF patients compared to LD patients at 70-79 years old (3.64% versus 0.59%, p < 0.0001) and 60-69 years old (2.16% versus 0.43%, p < 0.0001). Similarly, patients of increasing age consistently demonstrated increasing absolute outcome event rates (Table 2).

Table 2.

Comparative 30-Day Outcomes in Patients Receiving Lumbar Decompression vs Lumbar Decompression and Fusion by Age

​	80s LDF patients with outcome (%)	80s LD patients with outcome (%)	P value	OR (95% CI)	70s LDF patients with outcome (%)	70s LD patients with outcome (%)	P value	OR (95% CI)	60s LDF patients with outcome (%)	60s LD patients with outcome (%)	P value	OR (95% CI)
Mortality	61 (4.19%)	≤ 10* (0.68%)	< 0.0001	6.34 (3.236, 12.424)	131 (1.79%)	25 (0.34%)	< 0.0001	5.306 (3.455, 8.148)	108 (0.96%)	21 (0.19%)	< 0.0001	5.161 (3.231, 8.242)
Wound infection	27 (1.86%)	17 (1.16%)	0.1236	1.609 (0.873, 2.966)	126 (1.72%)	83 (1.12%)	0.0023	1.537 (1.163, 2.032)	179 (1.60%)	134 (1.19%)	0.0081	1.354 (1.081, 1.697)
Sepsis	31 (2.27%)	≤ 10* (0.71%)	0.0007	3.244 (1.584, 6.644)	109 (1.55%)	56 (0.78%)	< 0.0001	2.006 (1.451, 2.773)	114 (1.04%)	49 (0.44%)	< 0.0001	2.387 (1.706, 3.34)
Respiratory Complications	26 (2.23%)	≤ 10* (0.78%)	0.0029	2.907 (1.396, 6.055)	147 (2.34%)	46 (0.69%)	< 0.0001	3.46 (2.48, 4.827)	160 (1.61%)	55 (0.53%)	< 0.0001	3.103 (2.281, 4.22)
Cardiac Events	14 (1.30%)	13 (1.18%)	0.8036	1.101 (0.515, 2.354)	92 (1.48%)	41 (0.66%)	< 0.0001	2.277 (1.573, 3.295)	88 (0.87%)	24 (0.23%)	< 0.0001	3.736 (2.377, 5.872)
DVT/PE	18 (1.34%)	14 (1.01%)	0.4267	1.328 (0.658, 2.682)	135 (1.96%)	40 (0.57%)	< 0.0001	3.503 (2.458, 4.993)	139 (1.30%)	53 (0.49%)	< 0.0001	2.71 (1.973, 3.723)
Stroke	​	​	​	​	31 (0.44%)	13 (0.18%)	0.0061	2.407 (1.258, 4.603)	34 (0.31%)	12 (0.11%)	0.001	2.877 (1.489, 5.558)
Revision Surgery	​	​	​	​	127 (1.70%)	79 (1.06%)	0.0008	1.618 (1.22, 2.147)	216 (1.89%)	131 (1.15%)	< 0.0001	1.661 (1.335, 2.067)
ED visits	146 (9.82%)	121 (8.14%)	0.1088	1.229 (0.955, 1.582)	629 (8.43%)	467 (6.26%)	< 0.0001	1.379 (1.218, 1.561)	840 (7.35%)	579 (5.06%)	< 0.0001	1.487 (1.333, 1.658)
Readmissions	762 (51.24%)	283 (19.03%)	< 0.0001	4.472 (3.793, 5.272)	2920 (39.12%)	979 (13.12%)	< 0.0001	4.257 (3.923, 4.619)	4074 (35.63%)	1092 (9.55%)	< 0.0001	5.243 (4.873, 5.641)

*indicating value ≤10. [bold] indicating significance.

At 30 days, LDF was associated with higher rates of sepsis (OR 2.01-3.24, all p ≤ 0.0007), respiratory complications (OR 2.91-3.46, all p ≤ 0.0029), and readmissions (OR 4.26-5.24, all p < 0.0001) compared to LD patients. In 60-69 year olds and 70-79 year olds, LDF had a higher association with wound infection (OR 1.35-1.54, all p ≤ 0.0081), cardiac events (OR 2.28-3.74, all p < 0.0001), DVT/PE (OR 2.71-3.50, all p < 0.0001), stroke (OR 2.41-2.88, all p ≤ 0.0061), revision surgery (OR 1.62-1.66, all p ≤ 0.0008), and ED visits (OR 1.38-1.49, all p < 0.0001). Among 80-89 year old patients, LDF was associated with higher rates of sepsis (2.27% vs. 0.71%, p = 0.0007), respiratory complications (2.23% vs. 0.78%, p = 0.0029), and readmission (51.24% vs. 19.03%, p < 0.0001) (Table 2).

At 90 days, mortality rates remained higher in patients receiving LDF compared to LD (OR 3.83-6.35, all p < 0.0001). Absolute mortality rate was higher in LDF patients compared to LD patients at 80-89 years old (7.02% versus 1.93%, p < 0.0001), 70-79 years old (3.64% versus 0.59%, p < 0.0001), and 60-69 years old (2.16% versus 0.43%, p < 0.0001). For all age groups, LDF patients exhibited higher rates of all perioperative complications, except cardiac events, compared to LD patients (all p < 0.05) (Table 3).

Table 3.

Comparative 90-Day Outcomes in Patients Receiving Lumbar Decompression vs Lumbar Decompression and Fusion by Age

​	80s LDF patients with outcome (%)	80s LD patients with outcome (%)	P value	OR (95% CI)	70s LDF patients with outcome (%)	70s LD patients with outcome (%)	P value	OR (95% CI)	60s LDF patients with outcome (%)	60s LD patients with outcome (%)	P value	OR (95% CI)
Mortality	105 (7.02%)	29 (1.93%)	< 0.0001	3.832 (2.523, 5.818)	241 (3.64%)	39 (0.59%)	< 0.0001	6.347 (4.517, 8.917)	242 (2.16%)	48 (0.43%)	< 0.0001	5.109 (3.744, 6.971)
Wound infection	53 (3.55%)	33 (2.19%)	0.0258	1.643 (1.057, 2.552)	188 (2.82%)	112 (1.67%)	< 0.0001	1.706 (1.346, 2.16)	308 (2.75%)	192 (1.70%)	< 0.0001	1.637 (1.365, 1.964)
Sepsis	63 (4.49%)	19 (1.32%)	< 0.0001	3.524 (2.098, 5.917)	159 (2.49%)	67 (1.02%)	< 0.0001	2.464 (1.848, 3.285)	199 (1.82%)	85 (0.76%)	< 0.0001	2.414 (1.87, 3.116)
Respiratory Complications	46 (3.84%)	23 (1.74%)	0.0013	2.25 (1.355, 3.735)	203 (3.56%)	77 (1.27%)	< 0.0001	2.866 (2.199, 3.735)	216 (2.17%)	92 (0.88%)	< 0.0001	2.51 (1.963, 3.208)
Cardiac Events	29 (2.61%)	27 (2.40%)	0.753	1.089 (0.64, 1.852)	125 (2.22%)	61 (1.07%)	< 0.0001	2.089 (1.534, 2.843)	133 (1.31%)	68 (0.66%)	< 0.0001	1.993 (1.486, 2.673)
DVT/PE	39 (2.85%)	23 (1.62%)	0.0279	1.78 (1.057, 2.996)	186 (2.97%)	63 (0.99%)	< 0.0001	3.069 (2.302, 4.093)	206 (1.93%)	85 (0.78%)	< 0.0001	2.513 (1.949, 3.24)
Stroke	​	​	​	​	45 (0.70%)	21 (0.33%)	0.0029	2.16 (1.286, 3.631)	57 (0.52%)	21 (0.19%)	< 0.0001	2.759 (1.672, 4.554)
Revision Surgery	37 (2.42%)	≤ 10* (0.66%)	< 0.0001	3.767 (1.867, 7.603)	153 (2.26%)	94 (1.39%)	0.0002	1.642 (1.267, 2.128)	293 (2.56%)	173 (1.51%)	< 0.0001	1.712 (1.416, 2.07)
ED visits	270 (17.68%)	205 (13.43%)	0.0012	1.385 (1.137, 1.687)	956 (14.10%)	659 (9.72%)	< 0.0001	1.525 (1.372, 1.694)	1333 (11.66%)	897 (7.85%)	< 0.0001	1.55 (1.418, 1.694)
Readmissions	820 (53.70%)	333 (21.81%)	< 0.0001	4.159 (3.552, 4.869)	2705 (39.90%)	965 (14.23%)	< 0.0001	4 (3.679, 4.349)	4206 (36.79%)	1220 (10.67%)	< 0.0001	4.872 (4.54, 5.228)

*indicating value ≤10. [bold] indicating significance.

1-and 2- Year Outcomes

Mortality differences persisted at 1- and 2- years postoperatively. By 2 years postoperatively, absolute mortality rate was higher in LDF patients compared to LD patients at 80-89 years old (13.24% versus 4.61%, OR 3.16, p < 0.0001), 70-79 years old (8.04% versus 3.24%, OR 2.62, p < 0.0001), and 60-69 years old (5.98% versus 2.03%, OR 3.08, p < 0.0001). The highest mortality OR between LDF and LD patients was observed in the 80-89 year old population (Tables 4 and 5).

Table 4.

Comparative 1-Year Outcomes in Patients Receiving Lumbar Decompression vs Lumbar Decompression and Fusion by Age

​	80s LDF patients with outcome (%)	80s LD patients with outcome (%)	P value	OR (95% CI)	70s LDF patients with outcome (%)	70s LD patients with outcome (%)	P value	OR (95% CI)	60s LDF patients with outcome (%)	60s LD patients with outcome (%)	P value	OR (95% CI)
Mortality	159 (10.63%)	51 (3.40%)	< 0.0001	3.381 (2.445,4.675)	461 (6.31%)	131 (1.80%)	< 0.0001	3.68 (3.023,4.481)	487 (4.35%)	126 (1.13%)	< 0.0001	3.978 (3.264, 4.848)
Revision surgery	61 (4.00%)	29 (1.90%)	0.0006	2.149 (1.373,3.364)	309 (4.14%)	183 (2.45%)	< 0.0001	1.718 (1.427,2.069)	529 (4.63%)	279 (2.44%)	< 0.0001	1.94 (1.674, 2.248)
Opioid use disorder	​	​	​	​	20 (0.27%)	≤ 10* (0.14%)	0.0654	2.014 (0.942,4.305)	59 (0.53%)	34 (0.30%)	0.0082	1.757 (1.151, 2.681)
Chronic pain	88 (9.06%)	92 (9.98%)	0.4973	0.899 (0.661,1.222)	430 (8.99%)	369 (7.46%)	0.0059	1.226 (1.06,1.417)	628 (8.68%)	544 (7.23%)	0.0011	1.22 (1.083, 1.375)
Post Laminectomy syndrome	​	48 (3.26%)	​	​	​	220 (3.04%)	​	​	​	385 (3.46%)	​	​
Pseudoarthrosis	​	​	​	​	355 (5.07%)	​	​	​	794 (7.54%)	​	​	​
Mechanical complications of prosthesis	​	​	​	​	121 (1.68%)	​	​	​	198 (1.81%)	​	​	​
Infectious and inflammatory complications of prosthesis	​	​	​	​	96 (1.30%)	​	​	​	133 (1.18%)	​	​	​

*indicating value ≤10. [bold] indicating significance.

Table 5.

Comparative 2-Year Outcomes in Patients Receiving Lumbar Decompression vs Lumbar Decompression and Fusion by Age

​	80s LDF patients with outcome (%)	80s LD patients with outcome (%)	P value	OR (95% CI)	70s LDF patients with outcome (%)	70s LD patients with outcome (%)	P value	OR (95% CI)	60s LDF patients with outcome (%)	60s LD patients with outcome (%)	P value	OR (95% CI)
Mortality	191 (13.24%)	67 (4.61%)	< 0.0001	3.16 (2.368,4.218)	588 (8.04%)	236 (3.24%)	< 0.0001	2.616 (2.241,3.054)	638 (5.98%)	215 (2.03%)	< 0.0001	3.078 (2.631, 3.601)
Revision surgery	69 (4.68%)	33 (2.24%)	0.0003	2.145 (1.407,3.268)	415 (5.56%)	240 (3.22%)	< 0.0001	1.772 (1.507,2.084)	660 (6.06%)	319 (2.93%)	< 0.0001	2.138 (1.865, 2.45)
Opioid use disorder	​	​	​	​	29 (0.39%)	16 (0.22%)	0.05	1.826 (0.991,3.365)	81 (0.76%)	46 (0.43%)	0.0015	1.786 (1.243, 2.567)
Chronic pain	113 (12.06%)	122 (13.85%)	0.2561	0.853 (0.649,1.122)	578 (12.08%)	556 (11.24%)	0.1927	1.086 (0.959,1.229)	845 (12.44%)	787 (11.18%)	0.0216	1.129 (1.018, 1.252)
Post Laminectomy syndrome	​	61 (4.28%)	​	​	​	325 (4.49%)	​	​	​	543 (5.12%)	​	​
Pseudoarthrosis	​	​	​	​	383 (5.47%)	​	​	​	862 (8.62%)	​	​	​
Mechanical complications of prosthesis	​	​	​	​	165 (2.29%)	​	​	​	248 (2.38%)	​	​	​
Infectious and inflammatory complications of prosthesis	​	​	​	​	112 (1.52%)	​	​	​	152 (1.41%)	​	​	​

*indicating value ≤10. [bold] indicating significance.

Rates of revision surgery were consistently higher in LDF patients compared to LD patients across all age cohorts (OR 1.72-2.15, all p ≤ 0.0006). Rates of chronic pain diagnosis were not statistically different between 80-89 year old LDF and LD patients (p > 0.05), though LDF patients experienced higher rates of chronic pain in both 70-79 (8.99% versus 7.46%, p = 0.0059) and 60-69 (8.68% versus 7.23%, p = 0.0011) year old patients. There were not enough data to comment on the prevalence of opioid use disorder (OUD) in 80-89 year olds, but OUD prevalence was similar in 70-79 (0.27% versus 0.14%, p = 0.0654), but different in 60-69 (0.53% LDF versus 0.30% LD, p = 0.0082) year old patients (Tables 4 and 5).

Absolute rates of post-laminectomy syndrome in LD patients were observed between 4.28% and 5.12% at 2-years postoperatively. Absolute rates of pseudoarthrosis were observed at 5.47% in 70-79 year olds and 8.62% in 60-69 year olds. Absolute rates of mechanical prosthesis complications were observed at 2.29% in 70-79 year olds and 2.38% in 60-69 year olds. Absolute rates of infectious/inflammatory prosthesis complications were observed at 1.52% in 70-79 year olds and 1.41% in 60-69 year olds (Table 5).

Discussion

The number of individuals with lumbar degenerative disease has increased over time, and the number of older patients undergoing surgical intervention has risen accordingly. In patients without instability, there is variability in surgical decision-making. The present study utilized a large national database to compare morbidity/mortality between lumbar decompression alone (LD) and lumbar decompression with fusion (LDF) in patients with degenerative lumbar disease with no instability. Given the lack of literature evaluating the absolute mortality rates in LD and LDF patients, our study examined the absolute mortality in LD and LDF patients, as well as the absolute mortality differences among age groups.

We found that fusion was associated with higher mortality across all age groups. Importantly, absolute mortality differences increased with advancing age and were largest in octogenarians. These differences persisted through two years of follow-up. In addition, fusion was associated with higher rates of complications, readmissions, emergency department visits, and revision surgery compared to LD. In patients aged 80–89 years, the absolute mortality rate in LDF patients was 7.02% compared to 1.93% in LD patients (p<0.0001). Notably, the absolute mortality differences persisted at two years postoperatively, with mortality of 13.24% in LDF patients compared to 4.61% in LD patients (p<0.0001). Collectively, these findings demonstrate that fusion is associated with a clinically meaningful absolute mortality increase in elderly patients, particularly those aged 80–89 years.

The mortality differences observed in our study are large enough to influence patient-level decision-making. In octogenarians, who are frequently afflicted with multiple medical comorbidities, the baseline surgical risk is already elevated compared to younger cohorts, making further risk amplification especially consequential.^25,26 The mortality risk observed in our study may reframe operative planning in elderly populations. Surgical planning in older adults should emphasize frailty status and comorbidities that affect surgical stress response and recovery, often requiring multidisciplinary optimization. ²⁷ Our findings at two years postoperatively do not suggest that fusion should be categorically avoided in elderly patients, but rather, that when fusion is considered in discretionary settings, particularly in the absence of instability, it should be undertaken with careful justification, clear indication, and transparent risk discussion.^27,28

The differences observed between LDF and LD patients in this study likely reflect the greater physiologic burden imposed by fusion procedures. Consistent with prior literature, fusion is associated with longer operative times, increased blood loss, and greater surgical invasiveness, all of which contribute to heightened perioperative stress.^14,16,29 In our cohort, this translated into significantly higher 90-day rates of wound infection, sepsis, respiratory complications, DVT/PE, stroke, revision surgery, emergency department visits, and readmissions (all p<0.03). While prior systematic reviews and database studies have demonstrated higher overall morbidity with fusion, our findings extend this work by delineating a broad range of specific postoperative complications within a single comparative cohort.^16,20,30,31 This more granular characterization suggests that the impact of fusion may not be strictly limited to general complication rates, but spans infectious, cardiopulmonary, thromboembolic, and healthcare utilization outcomes. These findings reinforce the concept that fusion is associated with amplified perioperative physiologic stress beyond that of decompression alone.^20,32 In populations with limited physiologic reserve, this added burden may contribute to the observed increase in adverse events and has important implications for surgical decision-making and risk stratification.^33,34

Mortality differences increased across all age cohorts studied, with the greatest divergence observed in octogenarians at 30-, 90-, and 2-year postoperative time points (4.19% vs 0.68%, p<0.0001 at 30 days; 7.02% vs 1.93%, p<0.0001 at 90 days; 13.24% vs 4.61%, p<0.0001 at 2 years). These findings suggest that fusion may function as a physiologic risk multiplier in advanced age. This effect is likely driven by diminished physiologic reserve, defined as an age-related reduction in the capacity to tolerate acute physiologic stress. ³⁵ In this context, the increased operative burden of fusion may exceed adaptive capacity in older patients, contributing to the disproportionate rise in mortality observed with advancing age. These findings highlight the importance of incorporating physiologic reserve into preoperative risk stratification when considering more invasive procedures such as fusion. ³⁵

Our findings have direct implications for surgical practice in elderly populations. The absolute mortality differences observed in this study, especially among octogenarians, highlight the importance of incorporating absolute risk into preoperative counseling.^36-38 Presenting these data in absolute terms may provide more clinically meaningful context than relative comparisons alone, particularly when modest percentage differences translate into substantial mortality gaps. In this setting, the increased mortality and complication rates associated with fusion may shift the balance toward less invasive intervention in frail or physiologically vulnerable patients, while also emphasizing the need for careful perioperative planning. Targeted strategies such as early mobilization, vigilant cardiopulmonary monitoring, and prompt recognition of complications, may help mitigate the physiologic stress and adverse outcomes identified in older adults.^39-45 Integrating these findings into shared decision-making and tailored perioperative management supports informed, patient-centered care that aligns operative choice with individual risk tolerance and goals.

Previous studies consistently report higher perioperative complication rates with LDF compared to LD, including longer operative times, increased blood loss, and greater medical complications.^17,30,46 However, these studies rarely isolate mortality as a distinct outcome, and many include patients with spondylolisthesis or radiographic instability where fusion has traditionally been considered indicated, even though recent high-quality RCTs demonstrate decompression alone is noninferior to fusion in these populations.^47,48 In contrast, our study focused specifically on patients without spinal instability, revealing that fusion was associated with a substantially increased risk for mortality and postoperative complications. Across all age groups, LDF was associated with a more than fivefold higher 30-day mortality risk and a greater than threefold higher 90-day mortality risk compared to LD alone (all p<0.0001). Octogenarians experienced particularly pronounced risk, with 30-day mortality of 4.19% versus 0.68% in LD patients, notably higher than the 0.6% versus 0.3% mortality rates reported in Medicare claims data that included patients with traditional fusion indications. ⁴⁹ Furthermore, LDF was associated with higher rates of wound infection, sepsis, cardiopulmonary complications, thromboembolism, revision surgery, and readmissions within 90 days (all p<0.03). Critically, elevated mortality persisted at two years (all p<0.009), suggesting the fusion’s potential impact extends well beyond the immediate perioperative period. By isolating fusion in the absence of instability, our study provides a clearer estimate of the inherent procedural risk, quantifying the incremental mortality and morbidity associated with fusion itself and highlighting the need for careful patient selection and perioperative planning.

Mortality differences persisted at two years, with the largest gap observed in octogenarians (13.24% vs 4.61%, p<0.0001), and revision surgery rates remained consistently higher in fusion cohorts across age groups. In contrast, differences in chronic pain were modest, with no significant difference in octogenarians and only slightly higher rates in patients aged 60–79. Taken together, these findings suggest that the potential symptomatic benefits of fusion may be limited relative to its substantial survival and complication risks, including the increased likelihood of requiring revision surgery. Clinically, this underscores the importance of carefully weighing the incremental risks of fusion against its modest gains in pain or function, supporting individualized decision-making that prioritizes patient age, physiologic reserve, and overall risk tolerance.

Like all retrospective database studies, this investigation has inherent strengths and limitations. TriNetX-based analyses are observational and subject to coding inaccuracies and misclassification, particularly when identifying spinal pathologies. The absence of radiographic data limits assessment of disease severity, which may have influenced surgical selection and outcomes. For instance, patients selected for fusion may have had more severe stenosis, sagittal imbalance, mechanical back pain and occult instability not effectively captured within the ICD coding of TriNetX. By excluding patients with coded instability, deformity, and spondylolisthesis, this study intentionally targeted the population with clinical equipoise in whom fusion represents a discretionary augmentation rather than a biomechanical necessity. However, this approach may limit generalizability to the broader fusion population and because degenerative instability is inconsistently captured by ICD coding, some patients with uncoded biomechanical indications for fusion may remain in the cohort, contributing to residual confounding. Although propensity score matching was performed, residual and unmeasured confounding cannot be excluded, and unmeasured confounders represent an inherent limitation of claims-based analyses and should be considered when interpreting the comparative results.

Likewise, mortality outcomes should be interpreted cautiously as TriNetX reports all-cause mortality and does not discriminate for healthcare related mortality. Furthermore, mortality ascertainment may differ from the capture of other clinical outcomes, potentially resulting in overestimation of mortality rates relative to institution-specific datasets.

Inherent to data retrieved from TriNetX, several operative variables like the number of levels fused, surgical techniques such as open versus minimally invasive, duration of operation, overall blood loss, surgical approach, and whether surgery was revision or index represent variables that are unavailable for more granular review. The identification of “first ever” surgical events was limited to records within participating organizations and does not rule out prior spine surgery that was performed outside the database. This may potentially introduce left-censoring bias, particularly among older patients who may have longer histories of spine care and should be considered when interpreting the results. These variables are known to influence perioperative complication and mortality risk and may have contributed to unmeasured confounding between cohorts.

Differences between participating and non-participating institutions may limit generalizability. Finally, functional and patient-reported outcomes were not evaluated, which are important measures of postoperative recovery and overall treatment effectiveness in spine surgery.

Despite these limitations, this study possesses several important strengths. The TriNetX Research Network aggregates de-identified electronic health record data from geographically diverse healthcare organizations, enabling analysis within a large, nationally representative cohort. ⁵⁰ Frailty matching strengthened internal validity by controlling for comorbidity-related confounding. Age stratification allowed evaluation across decades of life, enhancing clinical applications. Finally, the emphasis on absolute risk represents a key contribution, aligning findings with real-world counseling and shared decision-making practices.

Conclusion

In this large national propensity-matched analysis of patients with degenerative lumbar disease, lumbar decompression with fusion was consistently associated with higher mortality and complication rates compared with decompression alone across all age groups. The absolute mortality difference increased substantially with advancing age, with octogenarians demonstrating the greatest survival disparity. These findings suggest that lumbar fusion may represent a distinct physiologic intervention whose risk profile extends beyond that of decompression alone, particularly in older adults. While fusion remains essential when instability or deformity is present, its use in discretionary settings should be accompanied by careful patient selection, comprehensive preoperative optimization, and vigilant postoperative monitoring. Incorporating absolute mortality risk into shared decision-making may help surgeons and patients more accurately balance the expected benefits of fusion against its measurable survival implications.

ORCID iDs

Alec M. Giakas https://orcid.org/0000-0002-0082-1648

Mitchell K. Ng https://orcid.org/0000-0002-5831-055X

Molly E. Milano https://orcid.org/0009-0000-8343-1292

Ali Farooqi https://orcid.org/0000-0001-7763-6152

Alan Hilibrand https://orcid.org/0000-0001-8811-9687

Ethical Considerations

Not required for public data analysis of the TriNetX dataset.