The Impact of Isolated Baseline Cannabis Use on Outcomes Following Thoracolumbar Spinal Fusion: A Propensity Score-Matched Analysis

Abstract

Background

There is limited literature evaluating the impact of isolated cannabis use on outcomes for patients following spinal surgery. This study sought to compare 90-day complication, 90-day readmission, as well as 2-year revision rates between baseline cannabis users and non-users following thoracolumbar spinal fusion (TLF) for adult spinal deformity (ASD).

Methods

The New York Statewide Planning and Research Cooperative System (SPARCS) database was queried between January 2009 and September 2013 to identify all patients who underwent TLF for ASD. Inclusion criteria were age ≥18 years and either minimum 90-day (for complications and readmissions) or 2-year (for revisions) follow-up surveillance. Cohorts were created and propensity score-matched based on presence or absence of isolated baseline cannabis use. Baseline demographics, hospital-related parameters, 90-day complications and readmissions, and two-year revisions were retrieved. Multivariate binary stepwise logistic regression identified independent outcome predictors.

Results

704 patients were identified (n=352 each), with comparable age, sex, race, primary insurance, Charlson/Deyo scores, surgical approach, and levels fused between cohorts (all, p>0.05). Cannabis users (versus non-users) incurred lower 90-day overall and medical complication rates (2.4% vs. 4.8%, p=0.013; 2.0% vs. 4.1%, p=0.018). Cohorts had otherwise comparable complication, revision, and readmission rates (p>0.05). Baseline cannabis use was associated with a lower risk of 90-day medical complications (OR=0.47, p=0.005). Isolated baseline cannabis use was not associated with 90-day surgical complications and readmissions, or two-year revisions.

Conclusion

Isolated baseline cannabis use, in the absence of any other diagnosed substance abuse disorders, was not associated with increased odds of 90-day surgical complications or readmissions or two-year revisions, though its use was associated with reduced odds of 90-day medical complications when compared to non-users undergoing TLF for ASD. Further investigations are warranted to identify the physiologic mechanisms underlying these findings.

Level of Evidence: III

Introduction

Cannabis is the third most commonly used recreational substance in the United States (US), following alcohol and tobacco.¹ As of May 2021, 36 states and four territories have medical marijuana programs, with 18 states having legalized it for nonmedical use. This wider acceptance of cannabis at the statewide level may reflect its use as a recreational or therapeutic drug. In 2018, there were 11.8 million estimated young adults who had used cannabis in the past year.² The increasing acceptance of cannabis as a medical therapy has led to an expanding list of prescribed indications, including sleep cycle dysrhythmias, chemotherapy-induced nausea and vomiting, chronic pain, and multiple sclerosis.³ While its efficacy has been demonstrated for these applications, chronic adverse events have also been described, including increased risks of stroke and cardiovascular accidents.^4-7

Adult spinal deformity (ASD) is classified as a spectrum of disorders that present in late adolescence or adulthood, and include adult spinal scoliosis, iatrogenic spinal deformity, and primary degenerative sagittal imbalance.⁸ With increasing life expectancies, the prevalence of ASD is rising, and more surgical corrections are expected to occur.⁹ Although risk factors such as smoking and hypertension are known to increase rates of infection, neurological, and cardiopulmonary complications, very few studies describe potential perioperative ramifications of cannabis in this orthopaedic patient population.¹⁰ Recently, one study of patients undergoing common orthopaedic procedures identified an association between cannabis and decreased risk of postoperative mortality. However, it also showed an increased risk of certain morbidities (e.g. cardiac disease, stroke, and heart failure).¹¹ Similarly, other studies found that cannabis use was associated with increased perioperative risk of myocardial infarction, thromboembolism, neurologic complications, respiratory complications, and sepsis.^12,13 Best et al.¹⁴ observed a high prevalence of cannabis consumption among total joint arthroplasty patients and described associated postoperative complications of polysubstance misuse. Others report associated cannabis use with lower bone mineral density and increased fracture risk.^14,15 However, some patients with musculoskeletal injuries have endorsed cannabis in conjunction with their postoperative care as a potentially effective analgesic.¹⁶ There is much still unknown and debated about the post-surgical effects of cannabis in patients undergoing orthopaedic procedures.

With shifting public sentiment, expanding decriminalization, and a lack of objective data on the potential consequences of cannabis use, it is imperative to identify how baseline cannabis use impacts postoperative outcomes of patients with adult spinal deformity (ASD) undergoing thoracolumbar fusion (TLF). This study sought to compare demographics, hospital-related parameters, 90-day complications, 90-day readmissions, and revision surgery between isolated baseline cannabis users and non-users in such circumstances.

Methods

Study Design

This was a retrospective cohort study using propensity score-matched patients from a large, all-payer, statewide database. Due to the deidentified nature of the data, this study was deemed to be exempt by the IRB and no informed consent was needed given the deidentified nature of the data queried.

Data Source

Data was extracted from the New York State Statewide Planning and Research Cooperative System (SPARCS) database. SPARCS is a comprehensive, statewide database reporting system that tracks and collects de-identified patient characteristics, diagnoses, treatments, services, and charges for all inpatient and outpatient visits at Article 28 licensed hospitals in New York State. All patients are assigned a distinct tracking identification number, and patient data can be readily extracted and analyzed across multiple admissions. Diagnostic and procedural data are classified according to the International Classification of Diseases, 9th revision, Clinical Modifications (ICD-9-CM).

Patient Population

All patients aged >18 years with ASD who underwent TLF surgery of two or more vertebral levels from January 2009 to September 2013 were identified by ICD-9-CM codes (81.00, 81.04, 81.05, 81.06, 81.07, 81.08, 81.30, 81.34, 81.35, 81.36, 81.37, 81.38, 81.62, 81.63, 81.64). ASD was defined using an established set of ICD-9-CM codes (idiopathic scoliosis [737.30, 737.32] or degenerative disc disease [737.10, 737.20, 756.19, 754.2, 737.19, and 737.29]).^17,18 Subjects were only included for analysis if they had 90-day surveillance for complications and all-cause readmissions as well as minimum two-year follow-up surveillance for revision surgery. Within this cohort, any patient with an ICD-9-CM code of nondependent or dependent cannabis use prior to and/or including the day of their primary TLF surgery was categorized as an isolated baseline cannabis user (304.30, 304.31, 304.32, 305.20, 305.21, and 305.22). Patients ≥18 years of age undergoing ≥2-level TLF with absent history of cannabis use were also identified.

Exclusion Criteria

Patients were excluded for any of the following conditions or pathologies: rheumatoid arthritis, osteomyelitis, traumatic fracture, pathologic fracture, any type of cancer or metastasis, Charcot-Marie-Tooth disease, mucopolysaccharidosis, osteitis deformans, fibrosa cystica, osteoporosis, poliomyelitis, rickets, Pott’s disease, tuberculosis, recognized scoliotic conditions of spina bifida/myelomeningocele, neurofibromatosis, cerebral palsy, Prader-Willi syndrome, Marfan syndrome, or any congenital musculoskeletal disease. They were also excluded if they had a co-diagnosis of another substance or poly-substance dependence/abuse (alcohol, cocaine, barbiturates, opioids, amphetamines, hallucinogens, antidepressants, and mixed/other substances) and if they did not meet follow-up surveillance criteria.

Data Collection

Extracted data included patient demographics and baseline factors (age, sex, race, Charlson/Deyo score, and primary insurance), hospital-related parameters (length of stay, hospital charges), surgical approach and number of levels fused, rates of 90-day complications (individual and overall medical and surgical complications, and overall complications), 90-day all-cause readmissions, and subsequent revisions with at least two-year follow-up surveillance.

Statistical Analysis

Cannabis users were 1:1 propensity score-matched to individuals undergoing the same index TLF procedure without a baseline diagnosis of cannabis use by age, sex, race, primary insurance, Charlson/Deyo score, number of fused spinal levels, fusion approach, and history of tobacco use. Multivariate binary stepwise logistic regression models were then performed to identify independent predictors of postoperative outcomes (covariates: cannabis use, age, sex, race, primary insurance, Charlson/ Deyo score, number of levels fused, fusion approach, and tobacco use history). All analysis was performed using SPSS® version 24.0 (IBM Corp., Armonk, NY, USA). A p-value < 0.05 was considered statistically significant.

Results

A total of 704 propensity score-matched patients were included in this study, with 352 patients in each of the Cannabis and Non-Cannabis groups. No significant differences were identified between the two patient groups according to their age, sex, race, primary method of payment, Charlson/Deyo scores, surgical approaches, and number of levels fused (p>0.160) (Table 1).

Table 1.

A Summary of The Differences in Patient Demographics, Operative Parameters, and Hospital-Related Parameters Representation Between a 1:1 Propensity Score-Matched Cohort of Non-Cannabis and Cannabis Users Who Underwent Thoracolumbar Fusion For Adult Spinal Deformity

		Non-Cannabis	Cannabis	p-value
n		352	352	–
Age (years)		±42.6 12.6	42.7 ±12.1	0.895
Charlson/Deyo Score		0.1 ± 0.4	0.1 ± 0.4	0.686
Approach	Anterior	6.8%	7.7%	0.663
	Posterior	91.2%	92.3%
Number of Levels Fused	2 or 3	85.2%	85.8%	0.583
	4 to 8	11.9%	11.6%
	≥9	2.8%	2.6%
Sex	Male	68.2%	67.9%	0.936
	Female	31.8%	32.1%
Race	White	61.8%	61.9%	0.785
	Black	20.2%	19.9%
	Hispanic	9.4%	10.2%
	Asian or Pacific Islander	0.6%	0.6%
	Native American	0.0%	0.6%
	Other	8.0%	6.8%
Primary Insurance	Medicare	27.8%	27.6%	0.160
	Medicaid	7.5%	11.3%
	Private Insurance Primary	47.4%	40.6%
	Insurance Self-Pay	1.1%	2.3%
	No Charge	0.0%	0.0%
	Other	16.2%	18.2%
History of Smoking		13.9%	15.3%	0.594
Total Charges (US dollars)		$82,135	$78,492	0.454
Length of Stay (days)		4.2 ± 4.0	4.3 ± 4.8	0.730

Cannabis users exhibited comparable individual 90-day medical complications when compared to non-cannabis users, including cardiac complications, infection, acute respiratory distress syndrome, and pneumonia (Table 2). However, Cannabis users experienced lower rates of overall 90-day medical complications than Non-Cannabis users (2.0% vs. 4.1%, p=0.018). The Cannabis and Non-Cannabis cohorts had comparable individual and overall 90-day surgical complications (p^≥0.317). Overall complication rates were higher among Non-Cannabis compared to cannabis users (4.8% vs. 2.4%, p=0.013). 90-day readmissions and two-year revision rates were similar between cohorts (p^≥0.157).

Table 2.

A Comparison of The Individual Medical and Surgical Complications, Overall Complications, Readmissions, and Revisions Between a 1:1 Propensity Score-Matched Cohort of Non-Cannabis and Cannabis Users Who Underwent Thoracolumbar Fusion For Adult Spinal Deformity

Outcomes		Cannabis Non-	Cannabis	value p-
90-Day Medical Complications	Cardiac Complications	0.1%	0.3%	0.563
	Urinary Infection Tract	0%	0.1%	0.317
	Hematoma	0.1%	0.1%	1.000
	Infection	0.1%	0%	0.317
	Anemia	2.7%	3.4%	0.431
	Acute Respiratory Distress Syndrome	0.3%	0.1%	0.563
	Pulmonary Embolism	0.1%	0%	0.317
	Deep Thrombosis Vein	0.1%	0.1%	1.000
	Acute Infarction Myocardial	0.1%	0.6%	0.178
	Pneumonia	0.6%	0.4%	0.704
	Acute Renal Failure	0.6%	0.6%	1.000
	Overall Medical Complications	4.1%	2.0%	0.018
90-Day Surgical Complications	Wound Infection	0.1%	0%	0.317
	Implant-Related	0.4%	0.3%	0.654
	Dural Tear	1.1%	0.7%	0.401
		0.9%	0.6%	0.524
90-Day Overall Complications		4.8%	2.4%	0.013
90-Day Readmission		0%	0.6%	0.157
Two-Year Revision		1.1%	0.6%	0.412

Regression analysis revealed that baseline cannabis use was not associated with an increase in odds of incurring 90-day surgical complications, 90-day readmissions, or two-year revisions. Baseline cannabis use was observed to be negatively associated with 90-day medical complications (OR=0.5, 95% CI: 0.3–0.8, p=0.005).

Discussion

As the favorable perception of cannabis use continues to grow and more states in the US pass decriminalizing legislation, the prevalence of cannabis usage is expected to increase. With an expanding pool of individuals with ASD and a persistent debate about the effect of cannabis in neuropathic pain management, it is essential to investigate the cannabis use and its potential effects on surgical outcomes.¹⁹ This study focused on the impact of isolated baseline cannabis use on patients who underwent thoracolumbar spinal fusion. This study found that cannabis use was negatively associated with 90-day medical complications. This is consistent with a previous study, by Moon et al.¹¹ that noted an associated decrease in mortality among cannabis users undergoing common orthopaedic procedures. Another study, by Nguyen et al.²⁰ involving traumatic brain injury patients found that cannabis users had reduced odds of death than nonusers. These findings are also corroborated by Singer et al.²¹ who reported on an intensive care unit (ICU) analysis of trauma patients, demonstrating lower mortality rates among cannabis users compared to non-users. However, this is not consistent with a study by Jakoi et al.²² that showed no significant differences in complications between cannabis users and non-users undergoing lumbar spinal fusion, however this study was performed with only 102 patients from a single surgeon.

The present study found that isolated baseline cannabis users undergoing TLF had lower 90-day medical and total complication rates than non-cannabis users. This differs from the results of the previous literature, which observed an association between cannabis use and cerebrovascular as well as cardiovascular diseases.^4,5,11,23-25 In contrast to these studies, this investigation excluded patients with multi- or poly-substance use from the analysis, potentially explaining the disparity in results observed by the removal of possibly confounding substance use outside of cannabis. This approach is particularly important, as highlighted by a study by Best et al.,¹⁴ in which, poly-substance-consuming patients undergoing TKA and THA had higher rates of surgery-related complications, including acute infection and anemia compared to non-users. The current study did not reveal difference in surgical complications between the two cohorts.

The cohorts in the present study demonstrated lower readmission rates when compared to the spine literature. A recent meta-analysis looking at 30-day readmission rates of spine surgeries found that single institutions reported the highest 30-day readmission rate at 6.6%, while multicenter studies reported the lowest at 4.7%.²⁶ Overall rates ranged between 4.2% and 7.4%.26 Baaj et al.²⁷ reported a 24.8% 90-day readmission rate for patients who underwent lumbar spinal fusion using the same data source (SPARCS) between 2005 and 2014. The discrepancies between these findings may be attributed to the stringent exclusion criteria for the patient cohort, strictly excluding patients with poly-substance use disorder patients and those with a variety of systemic comorbidities.

This study was not without limitations. As a retrospective review of a single statewide database, there is information associated with cannabis use that is not present in the database. For example, the purpose of cannabis use (medical versus recreational), cumulative exposure, and administration routes (inhalation, vaporization, ingestion, etc.) all could contribute to outcomes.⁷ Ultimately, such parameters would have allowed a more comprehensive assessment of isolated baseline cannabis use and further characterization of respective TLF patients. With the onset of the ICD-10-CM coding system giving providers broader diagnoses of cannabis use disorders, patients may be better stratified in future studies. Given the legal status of cannabis at the time this data was recorded, cannabis use was likely underreported and the percentage of users was likely higher than reported in this study. As such, the full spectrum of users may not entirely be represented by the study sample herein. The sample size (n=352 per cohort) stemmed in part from the exclusion of patients with any other documented drug use. Tobacco users were not part of the exclusion criteria, as this would have eliminated a large proportion of patients, and was instead controlled for in the propensity score-matching.

Conclusion

Compared to patients with ASD who underwent TLF without baseline cannabis use, patients with isolated baseline cannabis use were found to have no increase in odds of incurring 90-day surgical complications or readmissions or revisions two years postoperatively, though reduced odds of experiencing 90-day medical complications were observed. Future prospective, randomized-controlled studies could help further characterize the impact of isolated cannabis use on the postoperative course of surgical patients undergoing complex procedures such as thoracolumbar fusion for adult spinal deformity.