Abstract
Background The US population is aging, with the number of individuals aged 85 or older expected to triple by 2060. As such, spine surgeons are operating on an increasing number of elderly patients. This propensity score–matched retrospective analysis aimed to determine whether 30-day postoperative outcomes from spinal fusion surgery differ between super-elderly patients (aged ≥85 years) and younger patients (≤65 years).
Methods Patients in the 2015 to 2023 National Surgical Quality Improvement Program datasets were included if they were treated by an orthopedic or neurological surgeon with a primary Current Procedural Terminology code corresponding to spinal fusion at any level. Patients were separated into a younger cohort (≤65 years) and a super-elderly cohort (≥85 years). Propensity score matching was performed to balance demographics, comorbidities, surgical indications, and specific surgical procedures. Univariate statistical analysis followed by false-discovery rate correction was performed using Python 3.10.16.
Results Following matching, there were 1298 patients in each cohort. There was a 2.6% higher rate of pneumonia (P < 0.01), a 1.8% greater risk of reintubation (P < 0.01), a 1.2% increase in failure to wean from the ventilator after 48 hours (P < 0.01), a 1.9% higher rate of urinary tract infection (P < 0.01), a 1.2% higher risk of myocardial infarction (P < 0.01), and a 1.2% increase in mortality rate (P < 0.01) in the super-elderly compared with younger patients. Older patients were also more likely to be discharged to a skilled care facility (P < 0.01) and had a shorter operative duration (P < 0.01) and longer hospital stay (P < 0.01) relative to younger patients.
Conclusions Although we identified statistically significant increases in the risk of several adverse outcomes in the super-elderly population, the magnitude of these differences ranged from 1.2% to 2.6%, implying minimal clinical significance. Considering the increased risk of medical complications independent of surgical intervention in the super-elderly population, the minor increase in the risk of complications observed in this study can be considered acceptable, suggesting that advanced age alone is not a contraindication to surgery.
Clinical Relevance The results of this study provide insights into the safety of spinal arthrodesis in the growing super-elderly population.
Level of Evidence 4.
Introduction
Modern medicine has greatly improved human life expectancy, and as such, many developed countries are experiencing a gradual increase in elderly populations. In the United States alone, the elderly population, defined as individuals aged 65 years or older, is expected to nearly double by 2060, while the number aged 85 or older are projected to triple.1,2 As a result, spine surgeons are operating on an increasing number of elderly patients. In the United States, the mean age at lumbar fusion increased from 54.6 to 59.9 years between 2004 and 2015.3 A similar trend was observed in Korea, where the proportion of patients aged 70 or older undergoing lumbar fusion increased from 24% to 38% between 2010 and 2018, while the proportion younger than 60 decreased from 40% to 27%.3 This trend is expected to continue, with the projected number of lumbar fusions nearly doubling in patients aged 75 to 84 by 2040.4
Concerns about comorbidities and reduced bone quality have historically led older age to be viewed as a relative contraindication to spinal surgery. Due to the increasing age of the US population, there has been a growing interest in evaluating the safety of surgical intervention in older patients within the literature. Some studies have shown that older patients have higher incidences of adverse surgical outcomes, including pneumonia, sepsis, in-hospital mortality, and postoperative ileus following spinal fusion, compared with younger patients.5–9 In particular, interest has emerged regarding surgical outcomes in the super-elderly. There are multiple definitions of the super-elderly patient, with a significant number of surgical studies focusing on patients aged 85 or older, a population that is expected to triple within the next 40 years.1,10–12
The existing literature presents conflicting findings, and many surgeons remain hesitant to operate on super-elderly patients. Furthermore, most studies have focused on specific procedures or patient populations, and few have directly compared outcomes across a broad range of spinal fusion surgeries in a large, diverse cohort. To address this gap, we performed a propensity score–matched analysis using a national surgical database to compare 30-day outcomes between super-elderly (≥85 years) and younger (≤65 years) patients undergoing spinal fusion at any level. We hypothesized that increased age at the time of surgery would not independently increase the risk of adverse outcomes.
Methods
Study Population
The American College of Surgeons National Surgical Quality Improvement Program (NSQIP) database includes demographic and clinical variables, along with 30-day postoperative data acquired directly from the medical records of patients from more than 609 participating hospitals.12,13 All data are deidentified, and thus, this study is exempt from institutional review board oversight.
Patient Selection
We accessed data from the 2015 to 2023 NSQIP participant use files. Patients were eligible for inclusion if they were treated by either an orthopedic or neurological surgeon and had a primary Current Procedural Terminology (CPT) code corresponding to spinal fusion at any level.
Patients were separated based on age at the time of surgery into a younger (≤65 years) and super-elderly (≥85 years) cohort. To reduce heterogeneity, patients in the younger group were excluded if their primary CPT code was not present in the super-elderly population. Additionally, patients in the younger group were excluded if their postoperative diagnostic (ICD-10) code was not present in the super-elderly population (Supplemental Table 1).
Surgical Indications
We established 5 categories of surgical indication for spinal fusion, which included trauma or fracture, degenerative disease or deformity, infection, complication of a prior procedure, and other. Patients were placed into each of these categories based on their postoperative ICD-10 code. The codes defining each indication are listed in Supplemental Table 1.
Propensity Score Matching
Propensity score matching was used to balance baseline characteristics between the younger and super-elderly cohorts. The propensity score was estimated using logistic regression.
The propensity model included demographic variables (sex, race/ethnicity, and body mass index), comorbidities (smoking status, diabetes, congestive heart failure, chronic obstructive pulmonary disease, functional status, ascites, ventilator dependence, hypertension, renal failure, and steroid use), principal procedure, other procedures, and diagnosis category. Number of levels fused was identified from the number of add-on procedural codes present for each patient. Type of decompression was determined based on the most complex decompression code present for each patient. Missing covariate data were handled using the missing indicator method: missing values were imputed with the median (continuous variables) or mode (binary/categorical variables), and a binary indicator for missingness was included in the propensity model for each variable with incomplete data.
We performed 1:1 greedy nearest neighbor matching without replacement using a caliper of 0.2 standard deviations of the logit propensity score.14 To ensure comparability on critical variables, we required exact matching on primary CPT code and additional procedural CPT codes. Covariate balance was assessed using the absolute standardized mean difference, with a standardized mean difference threshold of <0.1 indicating adequate balance.
The Modified 5-Item Frailty Index (mFI-5) was calculated from the proportion of patients with diabetes, history of congestive heart failure, chronic obstructive pulmonary disease, hypertension requiring medication, and functional dependence before and after matching.
Outcomes
We compared the matched sets of cohorts for several standard and custom outcome measures included in the NSQIP database. Standard outcome measures included wound dehiscence, reintubation, pulmonary embolism, failure to wean from the ventilator after 48 hours, stroke, cardiac arrest, myocardial infarction (MI), deep vein thrombosis, return to operating room for any reason, still in hospital at 30 days after surgery, and units of blood transfused postoperatively.
For superficial and deep surgical site infection (SSI), pneumonia, urinary tract infection, and sepsis, we sought to capture only new, postoperative instances. Thus, we established new variables based on those present in the database, excluding any patients who already had the complication at the time of surgery.
We also developed 3 unique outcome measures based on the CPT and ICD-10 codes associated with a patient’s return to the operating room. These measures included infection requiring surgical intervention, hematoma or seroma requiring surgical intervention, and durotomy. The codes used to define these outcomes are listed in Supplemental Table 1.
Additionally, patients were compared for differences in discharge destinations. We grouped several patient responses together to establish 4 potential destinations: death, acute care, home, and skilled care. The specific variables used to define each of these destinations are listed in Supplemental Table 1.
Finally, patients were compared for differences in operative duration and total length of hospital stay.
Statistical Analysis
Following propensity score matching, the matched cohorts were compared for each outcome measure. Categorical outcomes were compared using Pearson’s χ² test or Fisher’s exact test when expected cell frequencies were <5. Continuous outcomes were compared using the Mann-Whitney U test. Risk differences with 95% confidence intervals were calculated for binary outcomes. P values were adjusted for multiple comparisons using the Benjamini-Hochberg false discovery rate method, with statistical significance set at adjusted P < 0.05. All analyses were performed using Python version 3.10.16.
Results
Patient Characteristics
Prior to propensity score-matching, there were 75,020 younger patients and 1359 super-elderly patients. There was a much larger percentage of the super-elderly cohort treated for vertebral fractures, accounting for 264 (19.4%) patients, compared with 1650 (2.2%) patients in the younger cohort. Accordingly, there were only 961 (70.7%) super-elderly patients treated for degenerative conditions, 14 (1.0%) for infection, 25 (1.8%) for complications of previous procedures, and 41 (3.0%) for other reasons compared with 65,278 (87.0%), 421 (0.6%), 2374 (3.2%), and 1770 (2.4%), respectively, in the younger cohort (Table 1).
Patient characteristics before and after matching.
Primary Analysis
Following propensity score matching, there were 1298 patients in each of the cohorts (super-elderly and younger). All demographics, comorbidities, and surgical indications were successfully balanced. The mFI-5, calculated from patient characteristics following matching, showed no significant difference between the 2 cohorts (Table 1). Balance was also achieved for the individual spinal fusion CPT codes, number of levels fused, type of decompression performed, and revision procedure status (Table 2; Supplemental Table 2).
Procedural details before and after matching.
The results showed that there was no statistically significant difference in the risk of superficial SSI (adjusted P = 0.81), deep SSI (adjusted P = 1.00), wound dehiscence (adjusted P = 0.69), pulmonary embolism (adjusted P = 0.10), stroke (adjusted P = 0.44), cardiac arrest (adjusted P = 0.06), deep vein thrombosis (adjusted P = 0.06), sepsis (adjusted P = 1.00), or hospital stay longer than 30 days postoperatively (adjusted P = 0.06). Additionally, there was no significant difference in the number of patients requiring a return to the operating room for any reason (adjusted P = 0.81). There was no significant difference between the super-elderly and younger patients in terms of specific reasons for returning to the operating room, including infection (adjusted P = 0.67), hematoma or seroma (adjusted P > 0.99), or durotomy (adjusted P > 0.99). Finally, the number of units of blood transfused postoperatively was not significantly different between the 2 groups (adjusted P = 0.74).
In contrast, the super-elderly cohort was associated with a 2.6% higher rate of pneumonia (48 super-elderly vs 14 younger; adjusted P < 0.01), a 1.7% higher risk of reintubation (29 super-elderly vs 6 younger; adjusted P < 0.01), a 1.2% greater risk of failure to wean from the ventilator after 48 hours (23 super-elderly vs 7 younger; adjusted P < 0.01), a 1.9% increase in risk of urinary tract infection (UTI) (43 super-elderly vs 18 younger; adjusted P < 0.01), and a 1.2% greater risk of MI (21 super-elderly vs 5 younger; adjusted P < 0.01). There were 21 patients in the super-elderly cohort who died within 30 days of surgery, compared with 5 patients in the younger cohort (adjusted P < 0.01). Similarly, the super-elderly patients were more likely to be discharged to a skilled care facility (420 super-elderly vs 132 younger; adjusted P < 0.01). Accordingly, the younger patients were more likely to be discharged home (631 super-elderly vs 1099 younger; adjusted P < 0.01).
We also found that the super-elderly patients had a mean operative time of 174.7 ± 86.3 minutes, which was significantly less than the 198.9 ± 104.1 operative duration in the younger cohort (adjusted P < 0.01). The super-elderly also had an average length of hospital stay of 5.6 ± 4.9 days, which was significantly longer than the 4.5 ± 5.2 day average stay in the younger cohort (adjusted P < 0.01; Table 3).
Thirty-day surgical outcomes.
Discussion
In this large-scale, national database study, we assessed whether advanced chronological age independently increases adverse outcomes following spinal fusion. After matching for demographics, comorbidities, surgical indication, and procedural details, our sample included 95.5% of all super-elderly patients who met inclusion criteria and demonstrated no significant difference in systemic frailty or major surgical complications compared with the younger cohort. Among the included complications were SSI, wound dehiscence, reoperation, or durotomy. Importantly, we observed higher rates of postoperative medical complications among super-elderly patients. However, these differences were small in absolute magnitude, suggesting that while advanced age may be associated with modest increases in certain medical events, chronological age alone does not represent a clinically meaningful increase in 30-day surgical complications after spinal fusion.
Outcomes
Several medical complications reached statistical significance when comparing outcomes by chronological age. However, these findings require careful interpretation with respect to absolute magnitude, as statistical significance alone is insufficient for clinical inference. In our analysis, pneumonia, reintubation, failure to wean from mechanical ventilation, urinary tract infection, and MI were all more frequent in the 85+ cohort, with absolute risk differences ranging from approximately 1.1% to 2.6% across individual outcomes. Conversely, superficial and deep SSI, wound dehiscence, durotomy, and reoperation outcomes had no differences in outcomes between cohorts. While large, matched cohorts are capable of detecting small absolute differences, absolute risk magnitude is more relevant to surgical decision-making than P values alone. Accordingly, effect size and plausible clinical mechanisms must be considered when interpreting statistically significant findings. One plausible explanation for these results is that medical complications are more common in elderly patients independent of surgical intervention, underscoring the importance of comparison to baseline geriatric risk.
More specifically, pneumonia demonstrated the largest absolute difference between cohorts, with elderly patients experiencing a 2.6% higher postoperative pneumonia rate (3.7% vs 1.0%). Population-based epidemiologic data indicate that adults aged 85 years or older have a community-acquired pneumonia incidence of approximately 52.3 cases per 1000 person-years independent of surgical intervention, compared with 18.2 cases per 1000 person-years among adults aged 65 to 69 years, reflecting a substantial age-related increase in baseline pneumonia risk.13 In this context, the observed postoperative difference in pneumonia rates in our cohort, while statistically significant, is smaller than the expected baseline risk attributable to advanced age and thus can be attributed to age-associated susceptibility rather than a true surgical-specific risk. Moreover, pneumonia is a leading cause of hospitalization among elderly individuals, further supporting the notion that this population carries an inherently elevated risk for pulmonary infection regardless of surgical status.14
We also observed a 1.9% increase in UTI incidence (3.3% vs 1.4%). Like pneumonia, UTIs are among the most frequently diagnosed infections in older adults and are the second most common cause of hospitalizations in patients older than 65 years.15 The literature regarding pneumonia and UTI risk is supported by what is known about immune function as one ages. Namely, age-related changes in immune function, including immunosenescence, are characterized by reduced naive T cell populations, impaired cytokine signaling, and diminished adaptive immune response.16 In combination with increased comorbidity burden and health care exposure, these physiological changes predispose elderly patients to infection independent of surgical intervention, suggesting that the observed differences reflect baseline age-related vulnerability rather than a direct effect of spinal fusion.
Last, we also observed a small increase in cardiopulmonary events including MI (1.6% vs 0.4%), reintubation (2.2% vs 0.5%), and failure to wean from mechanical ventilation at 48 hours (1.8% vs 0.5%). Although statistically significant, these differences were small in absolute magnitude and are consistent with well-established epidemiological data demonstrating that the incidence of cardiovascular and respiratory complications rises substantially with advancing age, independent of surgical intervention. For example, data presented by the US Centers for Disease Control and Prevention in 2019 found that the prevalence of heart disease increased with age, reported by 1% of adults aged 18 to 44, 3.6% of adults aged 45 to 54, 9.0% of adults aged 55 to 64, 14.3% of adults aged 65 to 74, and 24.2% of adults aged 75 years or older.17 In parallel, aging is associated with progressive declines in respiratory physiology, including reduced ventilatory responsiveness to hypoxia and hypercapnia, diminished pulmonary reserve, and impaired gas exchange, which increase vulnerability to respiratory failure.18 These findings are consistent with prior literature demonstrating that advancing age independently increases baseline cardiovascular and pulmonary risk through cumulative atherosclerotic burden and reduced physiological reserve, irrespective of operative status.19,20 Overall, this finding suggests that the modest increase in cardiopulmonary events in the present study likely reflects age-related susceptibility rather than a clinically meaningful increase in postoperative risk due to spinal fusion surgery.
Super-elderly patients were more frequently discharged to skilled care and experienced longer hospital stays compared with younger patients, likely due to discharge planning logistics and age-related expectations for functional recovery. Additionally, they had shorter operative times, and considering the lack of differences in surgical complexity as shown in Table 2, likely reflects surgical tailoring to minimize operative duration in older adults. In the absence of differences in major surgical outcomes or reoperation rates, these utilization patterns should be interpreted as expected differences rather than evidence of increased surgical risk attributable to chronological age.
Taken in totality, we feel that these findings support the safety of spinal fusion in carefully selected super-elderly patients and suggest that advanced chronological age should not be considered a contraindication for surgery. However, as the interpretation of these results is subjective, we suggest that decisions be made based on the individual surgeons’ comfort with the risks reported by this study. Furthermore, surgical exclusion criteria should incorporate physiologic reserve, comorbidities, and functional status rather than age alone. As demonstrated by our cohort, patients with advanced age may achieve acceptable short-term outcomes comparable to younger cohorts after spinal fusion.
Limitations
This study has several limitations. First, as a retrospective analysis of the NSQIP database, our findings are subject to selection bias, as elderly patients offered surgery are likely healthier and more carefully selected than the general elderly population. Second, NSQIP captures outcomes limited to the first 30 postoperative days and does not include long-term complications, functional outcomes, or patient-reported measures. Third, granular measures of frailty, functional status, and social support are not fully captured, which may influence postoperative recovery and discharge disposition. To address this limitation, we controlled for the major components of systemic frailty, resulting in a balance of the mFI-5 between the cohorts. Finally, the relatively smaller sample size in the 90+ subgroup may limit power to detect very small differences in rare outcomes.
Conclusion
In this large, propensity score–matched analysis of a national surgical database, advanced chronological age was not associated with clinically meaningful increases in short-term adverse surgical outcomes following spinal fusion. Although patients aged 85 years or older demonstrated statistically significant increases in several postoperative medical complications, the absolute magnitude of these differences was small and consistent with well-established age-related baseline risks rather than procedure-specific effects. Importantly, no differences were observed in major surgical complications, reoperation, or infection-related outcomes. Differences in discharge disposition, operative duration, and length of hospital stay likely reflect age-related functional considerations and perioperative decision-making rather than increased surgical risk. Taken together, we feel that these findings suggest that chronological age alone should not be considered a contraindication to spinal fusion, and surgical decision-making should prioritize physiological reserve, comorbidity burden, and functional status when evaluating elderly patients for operative intervention.
Footnotes
Funding The authors received no financial support for the research, authorship, and/or publication of this article.
Declaration of Conflicting Interests Harry M. Mushlin, MD, is a consultant for Depuy Synthes, VB Spine, and Stryker.
IRB Exemption All information within the American College of Surgeons National Surgical Quality Improvement Program database is deidentified per the deidentification standard defined in section §164.514(a) of the Health Information Portability and Accountability Act Privacy Rule. Thus, this study was exempt from informed consent, and institutional review board approval was not required.
- This manuscript is generously published free of charge by ISASS, the International Society for the Advancement of Spine Surgery. Copyright © 2026 ISASS. To see more or order reprints or permissions, see http://ijssurgery.com.
References
- 1.↵Shreya D , Fish PN , Du D . Navigating the future of elderly healthcare: a comprehensive analysis of aging populations and mortality trends using national inpatient sample (NIS) data (2010-2024). Cureus. 2025;17(3):e80442. 10.7759/cureus.80442
- 2.↵Orimo H , Ito H , Suzuki T , Araki A , Hosoi T , Sawabe M . Reviewing the definition of “elderly”. Geriatrics Gerontology Int. 2006;6(3):149–158. 10.1111/j.1447-0594.2006.00341.x
- 3.↵Kwon WK , Theologis AA , Kim JH , Moon HJ . Lumbar fusion surgery in the era of an aging society: analysis of a nationwide population cohort with minimum 8-year follow-up. Spine J. 2024;24(8):1378–1387. 10.1016/j.spinee.2024.03.003
- 4.↵Neifert SN , Martini ML , Hanss K , et al . Large rises in thoracolumbar fusions by 2040: a cause for concern with an increasingly elderly surgical population. World Neurosurg. 2020;144:e25–e33. 10.1016/j.wneu.2020.06.241
- 5.↵Kobayashi K , Imagama S , Ando K , et al . Complications associated with spine surgery in patients aged 80 years or older: Japan Association of Spine Surgeons with Ambition (JASA) multicenter study. Global Spine J. 2017;7(7):636–641. 10.1177/2192568217716144
- 6.↵Badhiwala JH , Karmur BS , Hachem LD , et al . The effect of older age on the perioperative outcomes of spinal fusion surgery in patients with lumbar degenerative disc disease with spondylolisthesis: a propensity score-matched analysis. Neurosurgery. 2020;87(4):672–678. 10.1093/neuros/nyz444
- 7.↵Narayanan R , Tarawneh OH , Dalton J , et al . Clinical outcomes of revision lumbar fusion in the aging population. Spine (Phila Pa 1976). 2025;50(8):E151–E157. 10.1097/BRS.0000000000005216
- 8.↵Kato S , Taniguchi Y , Kawamura N , et al . Outcomes of adult spinal deformity surgery in octogenarians - how old is too old. Eur Spine J. 2026;35(2):388–394. 10.1007/S00586-025-09716-W/FIGURES/1
- 9.↵Watanabe T , Kanayama M , Takahata M , et al . Perioperative complications of spine surgery in patients 80 years of age or older: a multicenter prospective cohort study. J Neurosurg Spine. 2020;32(4):622–630. 10.3171/2019.9.SPINE19754
- 10.↵Fahim DK , Woznicki AR , Fahim DJ , Black J . Safety and outcomes of elective spine surgery in the super-elderly (over 85) patient population. Brain and Spine. 2025;5:105520. 10.1016/j.bas.2025.105520
- 11.↵Mui T , Shigematsu H , Ikejiri M , Kawasaki S , Tanaka Y . Perioperative complications in patients aged ≥85 years undergoing spinal surgery: a retrospective comparative study of pre-old and old patients in Japan. Asian Spine J. 2024;18(6):856–866. 10.31616/asj.2024.0215
- 12.↵Miura Y , Nishio K , Kitamura Y , Goto T , Yano M , Matsui S . Surgical risk assessment for super-elderly patients. Geriatr Gerontol Int. 2022;22(4):271–277. 10.1111/ggi.14340
- 13.↵Stupka JE , Mortensen EM , Anzueto A , Restrepo MI . Community-acquired pneumonia in elderly patients. Aging health. 2009;5(6):763–774. 10.2217/ahe.09.74
- 14.↵Shi T , Denouel A , Tietjen AK , et al . Global and regional burden of hospital admissions for pneumonia in older adults: a systematic review and meta-analysis. J Infect Dis. 2020;222(Suppl 7):S570–S576. 10.1093/infdis/jiz053
- 15.↵Rowe TA , Juthani-Mehta M . Urinary tract infection in older adults. Aging Health. 2013;9(5). 10.2217/ahe.13.38
- 16.↵Castle SC , Uyemura K , Fulop T , Makinodan T . Host resistance and immune responses in advanced age. Clin Geriatr Med. 2007;23(3):463–479, . 10.1016/j.cger.2007.03.005
- 17.↵Heart Disease Prevalence - Health, United States. https://cdc.gov/nchs/hus/topics/heart-disease-prevalence.htm. 3 March 2026.
- 18.↵Sharma G , Goodwin J . Effect of aging on respiratory system physiology and immunology. Clin Interv Aging. 2006;1(3):253–260. 10.2147/ciia.2006.1.3.253
- 19.↵North BJ , Sinclair DA . The intersection between aging and cardiovascular disease. Circ Res. 2012;110(8):1097–1108. 10.1161/CIRCRESAHA.111.246876
- 20.
