Should venous thromboembolism chemoprophylaxis be utilized following spinal fusions?—a systematic review and literature-based decision analysis

Introduction

Venous thromboembolism (VTE), which encompasses deep vein thrombosis (DVT) and pulmonary embolism (PE), is a frequent post-operative complication following spinal surgery, with elective procedures reporting a pooled incidence of 3.3% (1). VTE increases both patient morbidity and healthcare costs, making effective prevention strategies essential (2,3). Preventive strategies for VTE include both mechanical and pharmacologic methods. Mechanical methods, such as compression stockings and devices, help reduce blood stasis. Pharmacologic methods, also termed ‘VTE Chemoprophylaxis’, include unfractionated heparin (UH) or low-molecular-weight heparin (LMWH), and work by reducing blood clot formation.

While VTE chemoprophylaxis (CPX) has proven effective in many surgical fields, its risk-to-benefit ratio in spinal surgery is less clear. This uncertainty is largely due to inconsistency and heterogeneity across studies. This includes differing VTE and hematoma diagnostic criteria, variable surgical protocols, and insufficient patient risk stratification. Recent reviews and meta-analyses have perpetuated this uncertainty, indicating with moderate-quality evidence that mechanical prophylaxis yields similar rates of VTE events compared to CPX following elective spine surgery (1,4). These findings suggest that routine use of VTE CPX after elective spine surgery may not be clearly supported (4).

Despite this recent research, there is still no strong consensus in guidelines that address VTE CPX use in spinal surgery. This is likely due to the lack of robust high-quality evidence to perform clinical decision analysis (CDA) to adequately address the risks and benefits of VTE CPX. The American College of Chest Physicians (ACCP) 2012 guidelines recommend mechanical prophylaxis for non-orthopedic spinal surgical patients, but still recommend adding CPX for those at high VTE risk, once bleeding risk is deemed low (5). These guidelines were emphasized in the ACCP’s 2020 policy statement (6). On the other hand, the North American Spine Society (NASS) acknowledged in their updated 2025 guidelines that evidence supporting the use of VTE prophylaxis is insufficient and conflicting (7). As a result, current VTE prophylaxis practices vary significantly by institution and provider (8).

The lack of clear guidelines is further complicated by case-specific factors that influence VTE rates after spinal surgery, such as surgery type and patient risk. CDA may offer an answer when direct comparative evidence is limited because it integrates the best available data to quantify competing benefits and harms of rare, high-stakes events that would require prohibitively large or lengthy prospective studies to evaluate (9-11). This study aims to systematically review the available literature and construct an exploratory clinical decision model of perioperative VTE CPX versus no CPX in patients undergoing spinal fusion surgery. Given the limitations and heterogeneity of the underlying evidence, the analysis is intended to illustrate potential trade-offs and generate hypotheses for future prospective investigation, not to support definitive clinical recommendations. We present this article in accordance with the PRISMA reporting checklist (available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0125/rc).

Methods

Identification of studies

A comprehensive literature review was performed across several electronic databases, including PubMed, the Cochrane Library, and Embase, covering the period from inception to September 2025. The dates of the search took place between June 2025 and September 2025. The details of our search strategy can be found in Appendix 1.

Assessment of eligibility

The following criteria were followed for study inclusion: (I) adult patients greater than 18 years of age who underwent spinal fusion for any reason; (II) perioperative use of any form of VTE prophylaxis (CPX or mechanical prophylaxis); (III) reported specific outcomes of interest (occurrence of post-operative VTE, DVT, PE, and epidural hematoma); and (IV) randomized controlled trial (RCT) or retrospective cohort study design. We excluded studies performed in pediatric patients and non-human subjects. Five reviewers independently screened titles and abstracts using a standardized electronic form, excluding studies that did not meet relevance criteria. Two authors then independently evaluated the full texts of remaining articles for eligibility with the senior author resolving any disagreements.

Data extraction

Data was extracted by two authors independently, including study objectives, demographics, baseline characteristics, surgical indications, and surgical outcomes. The primary extracted outcome measures included VTE event incidence (noted as DVT, PE, or DVT in the studies) and incidence of epidural hematoma. If probabilities of individual studies showed conflicting outcomes, they were not excluded. Utility values from the current literature were extracted for outcomes of VTE events and epidural hematomas (12). A senior author reviewed utility values to ensure alignment with real-world practice.

Study definitions

For this study, the CPX cohort involved patients from every included study who received LMWH, UH, enoxaparin, dalteparin, and aspirin in the perioperative period. The control cohort consisted of patients from all included studies who received placebo or mechanical interventions such as compression devices without accompanying CPX. These groups were separated for data extraction and subsequent analysis. For the purpose of this study, DVTs included deep and superficial VTE, PE, or DVT as supplied by the authors of the included studies. Expected utility is derived by multiplying each terminal outcome’s utility by its weighted-average probability and summing these values for each treatment strategy.

Statistical analysis

The TreeAge Pro software 2022 (13) and Excel 2024 were used for the statistical analysis of this study. A utility values decision tree model was used to analyze the collected data. We developed a decision analytic model to depict the potential outcomes of spinal fusion procedures with and without VTE prophylaxis. For each outcome category (epidural hematoma and VTE), model probabilities were calculated as sample-size weighted averages of those reported in randomized trials and observational studies included in the systematic review. Calculated weighted base-case event probabilities were as follows: ‘VTE CPX and developed epidural hematoma’ 0.017, ‘VTE CPX and experienced VTE event’ 0.023, ‘No CPX and developed epidural hematoma’ 0.025, ‘No CPX and experienced VTE event’ 0.031. A clinical decision tree was developed by applying the weighted averages of outcome probabilities and the utility values derived from included studies. Weighted probabilities and utility values were combined to conduct a case-based analysis to determine the expected utility for each branch point outcome. Subsequent sensitivity analysis was completed to determine the strength of the final expected utilities.

Results

Search results

Our search identified 621 articles using our search terms across all databases (102 from PubMed, 11 from Cochrane, 508 from Embase). Of these, 26 full-text articles were assessed for inclusion eligibility, and of these, six studies were ultimately deemed eligible for final inclusion.

Study characteristics, patient demographics, and extracted outcomes

A total of 6 studies met final inclusion criteria for this systematic review and CDA (see Figure 1) (14-19). Baseline characteristics and patient demographics are presented in Tables 1,2. Table 3 displays the extracted data from included studies stratified by treatment group. Reported outcomes included rates of VTE, DVT, PE, and epidural hematoma.

Figure 1 PRISMA diagram of article selection process. RCT, randomized controlled trial.

Clinical decision tree model

A decision analytic model with rollback analysis was modeled to evaluate the outcomes of VTE CPX and No VTE CPX in patients undergoing spinal fusion to determine the preferred condition. The model was constructed using TreeAge Pro 2022 (13), as seen in Figure 2.

Figure 2 Clinical decision tree with “VTE chemoprophylaxis” or “No VTE chemoprophylaxis” treatment options made using TreeAge (12). After rollback analysis, no VTE chemoprophylaxis is the preferred treatment option, showing an expected utility of 0.88 in comparison to 0.79 for the VTE chemoprophylaxis treatment option. Numbers beneath branches represent conditional probabilities. At terminal nodes, the first value represents the outcome utility, and P represents the joint probability of reaching that outcome. Values displayed at chance or decision nodes represent expected utilities calculated by rollback analysis. chxEH = weighted average probability of developing an epidural hematoma if received chemoprophylaxis, chxVTE = weighted average probability of experiencing a VTE event if received chemoprophylaxis; nochxEH = weighted average probability of developing an epidural hematoma if did not receive chemoprophylaxis; nochxVTE = weighted average probability of experiencing a VTE event if did not receive chemoprophylaxis. EH, epidural hematoma; VTE, venous thromboembolism.

The decision tree begins with either condition (VTE CPX or no VTE CPX) and from each strategy branches into two possible clinical trajectories: development or non-development of epidural hematoma. Subsequently, those who did not develop an epidural hematoma were divided based on the occurrence or non-occurrence of VTE events.

Transition probabilities for each branch point were derived from the included studies in our analysis. Utility values were also derived from the literature and the clinical expertise of the senior author for discrepancies. All probabilities and utility values were entered manually and assigned to their respective branches. For example, the probability of developing an epidural hematoma following CPX was set at 0.017, indicated by “chxEH” in Figure 2. Each end node represents the final outcome of that event and the events prior to it. The first value of each node represents the utility value of that event, while the second value represents the probability of that event occurring, taking into account the previous events if there were branch points. Rollback analysis, utilizing derived utility values and weighted averages of each event, demonstrated that VTE CPX (n=1,049) yielded a final expected utility of 0.79 compared to 0.88 for No VTE CPX (n=974), highlighted by the label above the bottom branching. A higher expected utility is favorable because it reflects the choice or pathway that would maximize the likelihood of positive end results while also balancing associated risks. We acknowledge that this model is oversimplified as it does not take into account factors such as age, obesity, prior VTE, specific surgery types, and operative time, which all would significantly impact the CDA.

Sensitivity analysis

One-way sensitivity analyses demonstrated that the model conclusions remained stable across plausible variations in key parameters, including utility values and probabilities for VTE and epidural hematomas. Probabilities were tested by decreasing and increasing by 50% of the extracted value to try and encompass a broad range of possible outcomes seen in the literature. For similar reasons, utility values were also tested by increasing and decreasing up to 50% of the extracted value. Across the tested probability ranges, expected utility remained higher for ‘No VTE chemoprophylaxis’ (0.85–0.94) than for ‘VTE chemoprophylaxis’ (0.78–0.79) under the model assumptions. Across the tested utility ranges, expected utility remained higher for ‘No VTE chemoprophylaxis’ (0.87–0.89) than for ‘VTE chemoprophylaxis’ (0.78–0.79) when looking at branch point outcomes of ‘Epidural hematoma’ and ‘No epidural hematoma’ with a VTE event. On the other hand, when utility sensitivities were tested for the ‘No epidural hematoma and no VTE event,’ there was significant overlap between ‘No VTE chemoprophylaxis’ (0.45–0.97) and ‘VTE chemoprophylaxis’ (0.40–0.97). This shift in decision was seen when utility values were changed to 0.8 and 0.9 utility values, respectively. This overlap suggests that the preferred strategy may depend on the utility assigned to this outcome and should therefore be interpreted in the context of the model assumptions. Internal validation was done with the senior clinician auditing the tree structure, formulas, probabilities, and utility assignments to ensure correct input and output from calculations. External clinical validation was not done.

Discussion

Based on heterogeneous outcome and health-utility estimates from the available literature, our exploratory decision-tree analysis found limited evidence of incremental benefit from VTE CPX after spinal fusion with the set model assumptions and should not be interpreted as definitive guidance for clinical practice. Prospective, adequately powered studies using standardized outcome definitions are needed to generate higher-quality comparative evidence and externally validate the model’s findings. This model therefore cannot resolve the existing clinical controversy due to the heterogeneous and conflicted underlying evidence as outlined below.

We found that the lower expected utility of VTE CPX was not driven by an increased risk of epidural hematoma, as one might expect given the pathophysiologic bleeding risk of CPX. Rather, the ‘No VTE chemoprophylaxis’ group was associated with a higher rate of epidural hematoma than the ‘VTE chemoprophylaxis’ group, although this difference was not statistically significant. This paradoxical difference in outcomes may be in part because patients who had a higher baseline risk of epidural hematomas were less likely to be given VTE CPX, although it is said with caution due to the heterogeneity of the evidence used.

Nevertheless, these pooled epidural hematoma rates conflict with reports of increased symptomatic hematomas with CPX in other elective spine cohorts. Specifically, Thota et al., was a retrospective propensity-matched cohort study that showed increased risk of symptomatic hematomas requiring reoperation among patients receiving VTE CPX compared to those who did not (20). Differences in study design, population, and outcome definitions contribute to this discrepancy. Despite differences in the epidural hematoma directionality, our analysis is consistent with Thota et al. in that VTE rates were not significantly different between CPX and non-CPX groups.

Similarly to our results, Anaspure et al. evaluated CPX prevention of VTE incidence in spinal trauma surgeries and found no significant difference in rates of VTE or other bleeding complications (21). Likewise, Rahmani and Muralidharan’s systematic reviews and meta-analyses showed no statistically significant differences in VTE and bleeding complications when comparing VTE CPX groups to non-CPX groups of elective spine surgery patients (1,4). Collectively, these findings call into question the universal application of VTE CPX in elective spine surgery and support a more risk-stratified approach including multiple confounding factors, which our exploratory decision analysis lacks.

As mentioned elsewhere, a universal strategy is difficult to justify due to the substantial variability between studies in baseline event rates, which suggests that the net benefit of CPX is likely context- and patient-dependent. Across our pooled studies, the rate of VTE events ranged from 0% to 7.4% (14-19), and epidural hematomas ranged from 0% to 8.7% (14-19). Although CDA offers a framework for integrating disparate event probabilities and outcome utilities, the heterogeneity of the available evidence underscores the need for a more robust, externally validated model before reliable conclusions regarding net benefit can be drawn.

Given this variability, it is also important to consider factors beyond prophylaxis strategy that may influence postoperative VTE and epidural hematoma risk. Patient-specific factors likely do play a role in altered VTE rates. McClendon Jr et al. demonstrated that sex, VTE history, IVC filter type, and hospital stay had a statistically significant effect on post-operative VTE events for spinal fusion patients (17). These factors were not accounted for in the CDA model.

Comparison to guidelines

Our findings are broadly consistent with the risk-based approach suggested by the ACCP and reflected in the recently updated 2025 NASS guidelines. The ACCP guidelines advocate for mechanical prophylaxis in spinal surgery patients and reserve VTE CPX for patients with high risk for VTE after bleeding concerns are mitigated (5). Similarly, the updated NASS guidelines report that VTE CPX leads to “no increase in hemorrhagic complications” and that its addition to mechanical prophylaxis “does not appear to provide additional benefit in the prevention of thromboembolic events” for patients undergoing elective spinal surgeries (6). These statements were substantiated by our pooled spinal fusion data, though must be taken with caution due to heterogeneity between pooled studies.

However, NASS also concluded that there is “insufficient and conflicting evidence for or against superior effectiveness of combined CPX and mechanical prophylaxis vs. either alone”. Because our results are merely exploratory, we make a similar conclusion as the evidence we used was both heterogeneous and used model-based assumptions.

Clinical implications

The most relevant clinical impact of this CDA is its ability to model as a reproducible and transparent tool for risk assessment, which may potentially guide institution-level or individual provider protocols until further evidence emerges. This is particularly relevant given the current heterogeneity in provider practices (7).

In the context of spinal fusion surgery, providers should continue to exercise caution when considering pharmacologic VTE prophylaxis and our results should not overshadow stratified risk approaches as they are merely exploratory. Patients who are at higher risk for thromboembolic events or with a prior history of DVT or PE still require tailored chemoprophylactic strategies, pending multidisciplinary evaluations that may not be captured by a CDA regardless of its modeling parameters.

Limitations

Our rigorous inclusion criteria prioritized randomized trials and higher-quality evidence, which resulted in a limited pool of six studies, potentially raising concerns about statistical power and generalizability of our derived probabilities. This limited pool of studies may reflect a broader limitation within the literature itself; that being an absence of RCTs investigating combined VTE CPX and mechanical prophylaxis vs. VTE mechanical prophylaxis alone for spine surgeries. Additional limitations to our study involve heterogeneity in CPX regimens (including timing and medications), fusion techniques, and patient populations across the examined studies. This heterogeneity precluded subgroup analyses by spinal region, chemoprophylactic regimen, or operative indication, limiting the granularity of our model. Furthermore, postoperative inpatient vs. outpatient management was inconsistently reported and could not be modeled in our CDA, despite being a confounding variable that could potentially influence VTE and bleeding risk.

Lastly, limitations inherent to CDA itself must be acknowledged. For instance, the model did not factor in clinical covariates that shape VTE and bleeding risk, including age, obesity, prior VTE, multilevel fusion, operative time, revision surgery, and mechanical prophylaxis use. Even more, the utility values for specific outcomes, though anchored in literature and expert input, may be considered ambiguous as they may not fully capture patient preferences or long-term outcomes, which could vary by demographic or psychosocial factors. As noted elsewhere, the model’s results were highly sensitive to the utilities assigned to certain outcomes. Therefore, our conclusions are conditional on the specified model parameters and assumptions. Other limitations include rare-event modeling bias and absent real-world validation.

Conclusions

The clinical decision exploratory model cannot resolve the existing clinical controversy, but suggests that in patients undergoing spinal fusion surgery, VTE CPX may not provide a clear benefit when the weighted probabilities of adverse outcomes and associated health utility values are considered. Despite limitations in the underlying evidence, including potential confounding factors and heterogeneity, this exploratory clinical decision model provides an adaptable framework that can be refined as higher-quality evidence becomes available to better assess VTE CPX in spinal fusion surgery.

Future high-quality prospective multicenter cohorts, risk-stratified modeling, and pragmatic clinical trials are required to establish robust guidelines defining specific patient and surgery parameters, such as location and indication for spinal fusion. Additionally, while the 2009 and 2025 NASS guidelines justifiably questioned the ethics of controlled trials concerning VTE prophylaxis vs. no-prophylaxis, they have since called for more RCTs examining combined mechanical-pharmacological vs. mechanical-only VTE prophylaxis (1). Such a change would expand the pool of eligible studies for our CDA, allowing for subgroup analysis while enhancing the specificity and granularity of the model.

Acknowledgments

None.

Footnote

Reporting Checklist: The authors have completed the PRISMA reporting checklist. Available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0125/rc

Peer Review File: Available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0125/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jss.amegroups.com/article/view/10.21037/jss-2026-0125/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Rahmani R, Eaddy S, Stegelmann SD, et al. Chemical prophylaxis and venous thromboembolism following elective spinal surgery: A systematic review and meta-analysis. N Am Spine Soc J 2024;17:100295. [Crossref] [PubMed]
2. Kim J, Kim TH. Extended Postoperative Risk of Venous Thromboembolism After Degenerative Spine Surgery: A Nationwide Cohort Study. Spine (Phila Pa 1976) 2026; [Epub ahead of print]. [Crossref] [PubMed]
3. Goz V, McCarthy I, Weinreb JH, et al. Venous Thromboembolic Events After Spinal Fusion: Which Patients Are at High Risk?. J Bone Joint Surg Am 2014;96:936-42. [Crossref] [PubMed]
4. Muralidharan A, Gong DC, Baumann AN, et al. Venous thromboembolism chemoprophylaxis is not supported following elective spine surgery: a systematic review and meta-analysis of randomized controlled trials. J Spine Surg 2025;11:242-55. [Crossref] [PubMed]
5. Gould MK, Garcia DA, Wren SM, et al. Prevention of VTE in Nonorthopedic Surgical Patients. Chest 2012;141:E227S-77S.
6. Henke PK, Kahn SR, Pannucci CJ, et al. Call to Action to Prevent Venous Thromboembolism in Hospitalized Patients: A Policy Statement From the American Heart Association. Circulation 2020;141:e914-31. [Crossref] [PubMed]
7. North American Spine Society. (2025). Evidence-based clinical guideline for antithrombotic therapies for adults undergoing spine surgery (2nd Edition). Spine.org. Available online: https://www.spine.org/Research/Clinical-Guidelines
8. Adeeb N, Hattab T, Savardekar A, et al. Venous Thromboembolism Prophylaxis in Elective Neurosurgery: A Survey of Board-Certified Neurosurgeons in the United States and Updated Literature Review. World Neurosurg 2021;150:e631-8. [Crossref] [PubMed]
9. Martin BI, Mirza SK, Spina N, et al. Trends in Lumbar Fusion Procedure Rates and Associated Hospital Costs for Degenerative Spinal Diseases in the United States, 2004 to 2015. Spine (Phila Pa 1976) 2019;44:369-76. [Crossref] [PubMed]
10. Sheikh SR, Thompson NR, Benzel E, et al. Can We Justify It? Trends in the Utilization of Spinal Fusions and Associated Reimbursement. Neurosurgery 2020;86:E193-202. [Crossref] [PubMed]
11. Tosteson AN, Lurie JD, Tosteson TD, et al. Surgical treatment of spinal stenosis with and without degenerative spondylolisthesis: cost-effectiveness after 2 years. Ann Intern Med 2008;149:845-53.
12. Dijkland SA, Voormolen DC, Venema E, et al. Utility-Weighted Modified Rankin Scale as Primary Outcome in Stroke Trials: A Simulation Study. Stroke 2018;49:965-71. [Crossref] [PubMed]
13. TreeAge Software, Inc. TreeAge Pro 2024 [Software]. Williamstown, MA; 2024. Available online: https://www.treeage.com/
14. Lunardini D, Ayyash OM, Nwasike CO, et al. An Evaluation of the Efficacy and Safety of Pharmacologic VTE Prophylaxis After Spine Surgery. The Spine Journal 2013;13:S80-1.
15. Shapiro JA, Stillwagon MR, Padovano AG, et al. An Evidence-Based Algorithm for Determining Venous Thromboembolism Prophylaxis After Degenerative Spinal Surgery. Int J Spine Surg 2020;14:599-606. [Crossref] [PubMed]
16. Kiguchi MM, Schobel H, TenEyck E, et al. The risks and benefits of early venous thromboembolism prophylaxis after elective spinal surgery: A single-centre experience. J Perioper Pract 2022;32:286-94. [Crossref] [PubMed]
17. McClendon J Jr, Smith TR, O'Shaughnessy BA, et al. Time to Event Analysis for the Development of Venous Thromboembolism After Spinal Fusion ≥ 5 Levels. World Neurosurg 2015;84:826-33. [Crossref] [PubMed]
18. Dornbush C, Maly C, Bartschat N, et al. Chemoprophylaxis Timing Is Not Associated With Postoperative Bleeding After Spinal Trauma Surgery. Clin Neurol Neurosurg 2023;225:107590. [Crossref] [PubMed]
19. Godil SS, Dewan MC, Parker SL, et al. Mechanical Versus Chemical Prophylaxis for Deep Venous Thrombosis in Patients Undergoing Lumbar Spinal Fusion: Comparative Effectiveness and Cost-Benefit. The Spine Journal 2013;13:S118-9.
20. Thota DR, Bagley CA, Tamimi MA, et al. Anticoagulation in Elective Spine Cases: Rates of Hematomas Versus Thromboembolic Disease. Spine (Phila Pa 1976) 2021;46:901-6. [Crossref] [PubMed]
21. Anaspure OS, Baumann AN, Fiorentino A, et al. The Effectiveness and Safety of Chemoprophylaxis in the Surgical Management of Spinal Trauma: A Systematic Review and Meta-Analysis. World Neurosurg 2025;194:123554. [Crossref] [PubMed]