Cannabis use increases surgical, medical, and psychosocial complications after lower extremity fracture fixation and shows compounded risk with concurrent nicotine use
1https://ror.org/046rm7j60grid.19006.3e0000 0001 2167 8097Department of Orthopaedic Surgery, University of California, Los Angeles, CA USA
2https://ror.org/046rm7j60grid.19006.3e0000 0000 9632 6718David Geffen School of Medicine at UCLA, 1225 15th Street – Suite 3144B, Santa Monica, Los Angeles, CA 90404 USA
3https://ror.org/051fd9666grid.67105.350000 0001 2164 3847Case Western Reserve University School of Medicine, Cleveland, OH USA
4https://ror.org/03taz7m60grid.42505.360000 0001 2156 6853University of Southern California, Los Angeles, CA USA
5https://ror.org/0377srw41grid.430779.e0000 0000 8614 884XPopulation and Quantitative Health Sciences, Case Western Reserve University and the Center for Clinical Informatics Research and Education, The MetroHealth System, Cleveland, OH USA
Abstract
Cannabis use is rapidly increasing in the United States, yet its perioperative implications remain poorly understood, particularly in trauma populations. Despite expanding legalization, the perioperative effects of cannabis on surgical outcomes remain uncertain. This retrospective cohort study used electronic health record data to evaluate the association between cannabis use, with and without concurrent nicotine exposure, and postoperative outcomes following orthopedic lower extremity trauma fixation. We conducted a retrospective cohort study using the TriNetX Research Network, a global federated electronic health record platform that aggregates de-identified clinical data from more than 90 health systems. Adult patients who underwent surgical fixation of lower extremity fractures between 2015 and 2023 were stratified into four cohorts: cannabis only users (n = 2421), nicotine only users (n = 36,966), concurrent users (n = 2338), and non-users. Primary outcomes included surgical and medical postoperative complications; secondary outcomes included psychosocial outcomes and coagulation parameters (PT, aPTT). Binary outcomes were evaluated using absolute risk differences, risk ratios (RR), and 95% confidence intervals (CIs). Continuous variables were compared using independent samples t-tests assuming unequal variances. Cannabis-only users demonstrated higher rates of deep implant infection, nonunion or malunion, reoperation, transfusion, readmission, anxiety, and depression compared with matched non-users. Nicotine-only users showed higher rates of wound complications, infection, nonunion or malunion, reoperation, transfusion, pneumonia, myocardial infarction, death, opioid use, chronic pain, readmission, anxiety, and depression compared with matched non-users. Concurrent cannabis-and-nicotine users had higher rates of superficial wound infection, deep implant infection, reoperation, amputation, transfusion, pneumonia, respiratory failure, opioid use, chronic pain, readmission, and anxiety compared with matched cannabis-only users. Coagulation measures were not consistently different across exposure groups. Cannabis use was associated with surgical and psychosocial complications in this large retrospective cohort. Nicotine use was associated with broader adverse effects, and concurrent cannabis-and-nicotine exposure was associated with compounded perioperative risk. These findings support further prospective evaluation of perioperative substance use screening and counseling strategies in surgical populations. Prospective studies quantifying cannabis exposure are needed to define dose–response effects and mechanisms.
Supplementary Information
The online version contains supplementary material available at 10.1038/s41598-026-55842-w.
Introduction
Cannabis use has increased rapidly across the United States, with per capita consumption rising more than 15-fold since 19921. Since 2022, the number of daily cannabis users has surpassed that of daily alcohol users, with nearly half of cannabis consumers reporting daily or near-daily use1. As of 2021, approximately 19% of U.S. adults reported cannabis use, with the highest prevalence among young adults, a demographic frequently represented in orthopedic trauma populations2–4. As of 2025, recreational cannabis is legal in 24 states and Washington, D.C., with several others considering active legislation5. With expanding legalization and normalization of use, national exposure rates are expected to continue rising in the coming years.
Cannabis is now widely accessible, and its growing acceptance has outpaced our understanding of its perioperative and systemic safety. Evaluating cannabis-related health outcomes in large clinical populations is increasingly important as its use becomes common among surgical patients.
Cannabis contains bioactive cannabinoids, primarily delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD), which act on CB1 and CB2 receptors expressed throughout the cardiovascular, neurologic, musculoskeletal, and immune systems6,7. These receptor-mediated effects influence vascular tone, inflammation, wound repair, and immune regulation, which may alter postoperative recovery and complication risk.
Despite widespread legalization, the impact of cannabis exposure on perioperative and surgical outcomes remains uncertain. Prior studies have produced conflicting results, with some linking cannabis use to increased infection and delayed healing, and others reporting no measurable effect. Some reports associate cannabis use with surgical site infection, delayed healing, and altered coagulation profiles, potentially increasing thromboembolic and cardiovascular risk8–10. Others have found no such associations, highlighting the need for larger, more representative analyses11. Concurrent nicotine use has also been implicated as a compounding factor, though the extent of its interaction with cannabis remains poorly defined12. The lack of high-quality, population-level evidence on these interactions has limited clinicians’ ability to counsel patients on perioperative risk and recovery expectations.
This knowledge gap limits the ability of orthopedic surgeons to counsel patients and optimize perioperative care. This study represents one of the largest retrospective cohort analyses examining the independent and combined associations of cannabis and nicotine use with postoperative and psychosocial outcomes following surgical fixation for lower extremity fractures.
Methods
Data source and study design
We conducted a retrospective cohort study using the TriNetX Research Network (Cambridge, MA; www.trinetx.com), a federated electronic health record platform that aggregates de-identified clinical data from more than ninety health systems and one hundred thirty million patients worldwide. The platform provides real time access to structured clinical data, including demographics, diagnoses, procedures, medications, and laboratory test results. TriNetX applies internal data quality controls that include schema validation, duplicate suppression, and temporal consistency checks to ensure analytic reproducibility across institutions. Data access was supported by the National Center for Advancing Translational Sciences grant UM1TR004528. The Research Database was accessed on 08/02/2025 and final statistical analyses were completed on 08/12/2025.
All TriNetX data are de-identified in accordance with the Health Insurance Portability and Accountability Act Privacy Rule (45 CFR §164.514[b][1]) using expert-determined methods such as data obfuscation and minimum cell sizes of ten or more individuals per value. Therefore, the dataset does not contain protected health information and is exempt from both HIPAA and institutional review board oversight.
Cohort selection
Adults aged eighteen years or older who underwent surgical fixation of a lower extremity fracture between January 2015 and December 2023 were eligible for inclusion. This study interval was selected to ensure adequate ICD-10-CM classification and incorporate at least a full year of follow-up. The index event was defined as the date of the first qualifying lower extremity surgery within the study window. Exposure cohorts were defined using preoperative ICD-10-CM documentation of cannabis-related disorders or nicotine dependence. Cannabis exposure was identified using F12 codes, and nicotine exposure was identified using F17 codes; specific subcodes are listed in Supplemental Table 1. Patients were categorized as cannabis-only users, nicotine-only users, concurrent users, or non-users without documented cannabis or nicotine use. These diagnosis codes reflect documented clinical diagnoses rather than direct measures of cannabis or nicotine use behavior. Each cohort was used to generate matched pairs for independent comparisons to assess the association of cannabis and nicotine exposure with postoperative outcomes. Cannabis-only and nicotine-only cohorts were each compared with matched non-users to estimate the associations of each exposure relative to patients without documented cannabis or nicotine use. Concurrent users were compared with matched cannabis-only users to estimate the incremental association of nicotine exposure among patients with documented cannabis use. Therefore, the concurrent-user analysis should be interpreted as the additional risk associated with concurrent nicotine use among cannabis users, rather than as a comparison with substance non-users. Figure 1 demonstrates the stepwise cohort creation process for subsequent analyses.
Index event and follow-up
The index event for each patient was the first recorded date of lower extremity fracture surgery during the study period. Follow-up began one day after the index event and continued for 365 days. Patients were censored at their last available encounter if follow-up ended before one year. For each outcome assessed, patients with preexisting documentation of that outcome prior to the index event were excluded from the respective analysis to ensure evaluation of incident postoperative events. Laboratory and diagnostic outcomes were assigned based on the earliest qualifying code or result after surgery.
Outcomes
Primary and secondary outcomes included both medical and surgical complications commonly reported following lower extremity trauma surgery. Outcomes were defined using CPT and ICD-10 codes mapped through TriNetX’s Unified Medical Language System (UMLS) ontology. Events of interest included wound dehiscence, superficial wound infection, deep implant infection, nonunion or malunion, nerve palsy, reoperation, amputation, pneumonia, nosocomial infection, acute respiratory distress syndrome (ARDS), respiratory failure, deep vein thrombosis (DVT), pulmonary embolism (PE), transfusion, myocardial infarction (MI), acute kidney injury, stroke, mortality, all-cause 1 year readmission, anxiety, depression, opioid use, and chronic pain. Laboratory endpoints included prothrombin time (PT) and activated partial thromboplastin time (aPTT).
Statistical analysis
Propensity score matching (PSM) was applied to reduce baseline confounding between exposure groups. A 1:1 greedy nearest-neighbor algorithm with a caliper of 0.10 pooled standard deviations was employed for each pairwise comparison. Matching variables included age at index, sex, race, BMI, HbA1c, osteoporosis (with and without pathological fracture), atherosclerotic cardiovascular disease, chronic kidney disease, liver disease, COPD, peripheral arterial disease, polysubstance abuse (opioid related disorders, alcoholic related disorders, and other psychoactive substance related disorders), polytraumatized patients, preoperative antibiotics, fracture location, open versus closed fracture types, as well as type of fixation procedure. Additionally, we included all procedure codes used in initial cohort creation as covariates to minimize confounding based on surgical approach. Covariate balance was assessed using standardized mean differences (SMD < 0.10). Specific codes used for both cohort creation and for propensity score matching can be found in Supplemental Table 1.Pre-match: cannabis-only users Pre-match: nicotine-only users Pre-match: concurrent users Cannabis only users (N = 2433) Non-users (N = 196,622) SMD Nicotine only users (N = 37,649 ) Non-users (N = 196,622) SMD Concurrent Users (N = 4749 ) Cannabis only users (N = 2433) SMD N, Mean %, SD N, Mean %, SD N, Mean %, SD N, Mean %, SD N, Mean %, SD N, Mean %, SD Age, mean (SD) 36.289 16.307 53.022 24.711 0.799 49.653 17.47 53.022 24.711 0.157 39.158 14.295 36.289 16.307 0.187 Male, % 1569 64.49% 81,592 41.50% 0.473 25,582 67.95% 141,023 71.72% 0.235 3134 65.99% 1569 64.49% 0.032 Race (White) 1311 53.88% 141,024 71.72% 0.376 20,005 53.14% 81,591 41.50% 0.082 2488 52.39% 1311 53.88% 0.03 BMI, mean (SD) 27.761 6.721 28.486 7.249 0.104 27.962 7.061 28.486 7.249 0.073 26.932 6.709 27.761 6.721 0.124 18–25 kg/m2 740 30.42% 51,279 26.08% 0.096 12,022 31.93% 51,279 26.08% 0.129 1930 40.64% 740 30.42% 0.215 25–30 kg/m2 718 29.51% 56,127 28.55% 0.021 12,289 32.64% 56,127 28.55% 0.089 1651 34.77% 718 29.51% 0.113 30–35 kg/m2 450 18.50% 39,011 19.84% 0.034 8257 21.93% 39,011 19.84% 0.052 994 20.93% 450 18.50% 0.061 35–40 kg/m2 237 9.74% 21,176 10.77% 0.034 4224 11.22% 21,176 10.77% 0.014 499 10.51% 237 9.74% 0.025 40–60 kg/m2 144 5.92% 13,146 6.69% 0.032 2486 6.60% 13,146 6.69% 0.003 283 5.96% 144 5.92% 0.002 HbA1c, mean (SD) 6.031 1.841 6.341 1.781 0.172 6.411 2.055 6.341 1.781 0.036 6.166 2.162 6.031 1.841 0.068 4–5.7% 235 9.66% 21,512 10.94% 0.042 4501 11.96% 21,512 10.94% 0.032 672 14.15% 235 9.66% 0.139 5.7–6.5% 123 5.06% 19,576 9.96% 0.187 3560 9.46% 19,576 9.96% 0.017 318 6.70% 123 5.06% 0.07 6.5–8% 61 2.51% 12,236 6.22% 0.183 2054 5.46% 12,236 6.22% 0.033 154 3.24% 61 2.51% 0.044 8–10% 65 2.67% 9149 4.65% 0.106 1962 5.21% 9149 4.65% 0.026 180 3.79% 65 2.67% 0.063 Comorbidities Hyperlipidemia 256 10.52% 49,089 24.97% 0.385 8766 23.28% 49,089 24.97% 0.039 630 13.27% 256 10.52% 0.085 Cerebrovascular diseases 141 5.80% 19,303 9.82% 0.150 3894 10.34% 19,303 9.82% 0.018 393 8.28% 141 5.80% 0.097 Diabetes mellitus 203 8.34% 30,640 15.58% 0.225 5654 15.02% 30,640 15.58% 0.016 466 9.81% 203 8.34% 0.051 CAD 117 4.81% 21,225 10.80% 0.225 4410 11.71% 21,225 10.80% 0.029 338 7.12% 117 4.81% 0.098 Other diseases of liver 100 4.11% 8909 4.53% 0.021 2496 6.63% 8909 4.53% 0.092 352 7.41% 100 4.11% 0.142 CKD 105 4.32% 20,234 10.29% 0.231 2871 7.63% 20,234 10.29% 0.093 234 4.93% 105 4.32% 0.029 Osteoporosis 87 3.58% 21,684 11.03% 0.289 2765 7.34% 21,684 11.03% 0.128 138 2.91% 87 3.58% 0.038 COPD 80 3.29% 10,568 5.38% 0.103 6692 17.78% 10,568 5.38% 0.395 521 10.97% 80 3.29% 0.302 PAD 47 1.93% 7988 4.06% 0.125 2343 6.22% 7988 4.06% 0.098 140 2.95% 47 1.93% 0.066 Polytrauma 159 6.54% 7304 3.72% 0.128 2193 5.83% 7304 3.72% 0.099 500 10.53% 159 6.54% 0.143
For binary outcomes, absolute risk differences, risk ratios (RR) and 95% confidence intervals (CIs) were calculated. Kaplan–Meier survival analysis with log-rank testing was used for time-to-event data, and hazard ratios (HR) were derived from Cox proportional hazards models. Continuous outcomes, including PT and aPTT, were compared using independent samples t-tests assuming unequal variances. Outcomes were excluded from analysis if present prior to the index date, unless specifically included per protocol. We considered a P value of less than 0.05 to be statistically significant. All statistical analyses were performed within the TriNetX Analytics environment using validated, platform-embedded algorithms to maintain data security and reproducibility.
Ethics and data use
This study was conducted using de-identified data from the TriNetX Research Network and was performed in accordance with all relevant guidelines and regulations. The TriNetX Research Network complies with the HIPAA Privacy Rule under 45 CFR §164.514(b)(1) through expert-determined de-identification standards, and all data were de-identified before analysis. The study did not involve human subjects research as defined under 45 CFR §46.102 because the study team accessed only preexisting de-identified data and did not access protected health information, interact with individual subjects, or obtain identifiable private information. Therefore, institutional review board approval was not required. Because the study involved only secondary analysis of de-identified data, informed consent from subjects or their legal guardians was not required.
In accordance with TriNetX privacy rules, event counts of fewer than 10 patients are displayed as “ ≤ 10” to protect patient confidentiality. Statistical estimates including risk ratios and confidence intervals for these outcomes are computed by the TriNetX platform using the actual underlying counts prior to obfuscation and are reported as calculated.
Results
Cohort sizes and matching
Each of the four exposure cohorts was successfully balanced using one-to-one propensity score matching to create comparable cohorts for analysis. The analytic sample included 2421 cannabis-only users, 36,966 nicotine-only users, and 2338 concurrent users, each matched to non-user or cannabis-only comparators. Baseline demographic and clinical characteristics were well balanced across all matched cohorts, with standardized mean differences below 0.10 for every covariate. These balanced cohorts formed the basis for all subsequent outcome comparisons. Results of pre and post matching can be found in Table 1. Across comparisons of cannabis-only users and combined cannabis and nicotine users, baseline characteristics including age, sex, race, fracture type, and comorbidities were balanced with standardized mean differences below 0.10, confirming successful covariate matching. However, covariate balance was not fully achieved for HbA1c in the cannabis-only cohort (SMD = 0.146) and concurrent user cohort (SMD = 0.102), despite each sub-category being well balanced (Supplemental Fig. 1). All trauma-specific covariates, including fracture location, open versus closed fracture status, type of fixation procedure, polytrauma, and preoperative antibiotic administration, were included in matching and achieved standardized mean differences below 0.10 in all cohorts.
Cannabis-only users
Cannabis-only users experienced higher rates of several postoperative complications compared with matched non-users in this cohort. Specific values and results of these outcome analyses can be found in Table 2. Surgical complications were increased in cannabis-only users compared with non-users, including deep implant infection (RD: 1.467%; RR: 2.260; 95% CI: 1.453–3.515), nonunion or malunion (RD: 2.138%; RR: 1.462; 95% CI: 1.153–1.853), and reoperation (RD: 3.387%; RR: 1.299; 95% CI: 1.122–1.505). Nerve palsy was also more frequent among cannabis-only users (RD: 0.644%; RR: 2.171; 95% CI: 1.127–4.181), although this finding was considered exploratory given the absence of a clear mechanistic basis. No significant differences were observed for wound dehiscence (RD: 0.416%; RR: 1.774; 95% CI: 0.901–3.495), superficial wound infection (RD: 0.413%; RR: 1.144; 95% CI: 0.850–1.539), nosocomial infection (RD: 0%; RR: 1.000; 95% CI: 0.417–2.415), irrigation and debridement (RD: 0.337%; RR: 1.805; 95% CI: 0.835–3.901), amputation (RD: 0%; RR: 1.000; 95% CI: 0.417–2.398), and DVT (RD: 0%; RR: 1.000; 95% CI: 0.417–2.384).Metric Cannabis-only users vs non users Nicotine-only users vs non-users Concurrent users vs cannabis-only users Cannabis only users with outcome Non-users with outcome RD (%) 95% CI (%) P-value RR 95% CI Nicotine only users with outcome Non-users with outcome RD (%) 95% CI (%) P-value RR 95% CI Concurrent users with outcome Cannabis only users with outcome RD (%) 95% CI (%) P-value RR 95% CI Wound Dehiscence 23 13 0.416 (− 0.069, 0.902) 0.0926 1.774 (0.901, 3.495) 459 262 0.536 (0.394, 0.679) < 0.0001 1.754 (1.508, 2.04) 28 21 0.303 (− 0.284, 0.89) 0.3118 1.336 (0.761, 2.345) Superficial Wound Infection 89 79 0.413 (− 0.619, 1.639) 0.3759 1.144 (0.85, 1.539) 1844 1319 1.745 (1.413, 2.076) < 0.0001 1.435 (1.34, 1.538) 137 85 2.823 (1.438, 4.209) < 0.0001 1.703 (1.309, 2.217) Deep Implant Infection 63 28 1.467 (0.696, 2.238) 0.0002 2.26 (1.453, 3.515) 1172 769 1.114 (0.881, 1.347) < 0.0001 1.531 (1.40, 1.675) 97 56 1.795 (0.763, 2.827) 0.0007 1.742 (1.26, 2.408) Nonunion or Malunion 158 109 2.138 (0.812, 3.464) 0.0016 1.462 (1.153, 1.853) 1296 916 1.063 (0.812, 1.313) < 0.0001 1.421 (1.308, 1.545) 107 82 1.133 (− 0.021, 2.288) 0.0542 1.317 (0.994, 1.746) Nerve Palsy 28 13 0.644 (0.113, 1.175) 0.0174 2.171 (1.127, 4.181) 362 281 0.241 (0.101, 0.382) 0.0008 1.304 (1.116, 1.523) 23 27 − 0.162 (− 0.775, 0.451) 0.6043 0.864 (0.497, 1.502) Irrigation and Debridement 18 ≤ 10* 0.337 (− 0.097, 0.772) 0.1278 1.805 (0.835, 3.901) 250 190 0.172 (0.058, 0.285) 0.003 1.328 (1.10, 1.603) 25 19 0.282 (− 0.288, 0.851) 0.3318 1.34 (0.74, 2.427) Amputation ≤ 10* ≤ 10* 0 (− 0.364, 0.363) 1 1 (0.417, 2.398) 17 ≤ 10* 0.019 (− 0.009, 0.047) 0.1772 1.701 (0.779, 3.715) ≤ 10* 0 0.428 (0.163, 0.693) 0.0015 – (− , −) Reoperation 356 274 3.387 (1.494, 5.28) 0.0005 1.299 (1.122, 1.505) 4824 3903 2.491 (2.027, 2.956) < 0.0001 1.236 (1.188, 1.286) 432 335 4.149 (2.029, 6.268) 0.0001 1.29 (1.132, 1.47) DVT ≤ 10* ≤ 10* 0 (− 0.364, 0.363) 0.999 1 (0.417, 2.384) 119 100 0.052 (− 0.027, 0.13) 0.1998 1.189 (0.912, 1.551) ≤ 10* ≤ 10* 0 (− 0.377, 0.376) 0.9985 0.999 (0.417, 2.396) PE 40 33 0.289 (− 0.397, 0.976) 0.4091 1.212 (0.767, 1.915) 699 647 0.141 (− 0.052, 0.333) 0.1526 1.08 (0.972, 1.201) 39 39 0 (− 0.734, 0.734) 1 1 (0.644, 1.553) Transfusions 77 46 1.386 (0.46, 2.313) 0.0033 1.705 (1.189, 2.446) 1264 1009 0.751 (0.49, 1.013) < 0.0001 1.273 (1.164, 1.370) 94 69 1.15 (0.043, 2.258) 0.0417 1.371 (1.01, 1.86) Pneumonia 22 26 − 0.175 (− 0.759, 0.408) 0.556 0.844 (0.48, 1.485) 897 727 0.508 (0.277, 0.74) < 0.0001 1.236 (1.122, 1.361) 36 20 0.753 (0.092, 1.414) 0.0252 1.842 (1.07, 3.172) Nosocomial Infection ≤ 10* ≤ 10* 0 (− 0.361, 0.364) 0.9941 1 (0.417, 2.415) 123 79 0.12 (0.044, 0.196) 0.0019 1.559 (1.176, 2.067) ≤ 10* ≤ 10* 0 (− 0.375, 0.376) 0.999 1 (0.417, 2.399) Respiratory Failure 22 24 − 0.071 (− 0.661, 0.52) 0.8138 0.933 (0.525, 1.659) 897 855 0.171 (− 0.074, 0.416) 0.1706 1.067 (0.973, 1.17) 38 20 0.876 (0.177, 1.576) 0.0138 1.942 (1.136, 3.378) ARDS ≤ 10* ≤ 10* 0 (− 0.361, 0.361) 1 1 (0.417, 2.398) 179 147 0.087 (− 0.009, 0.182) 0.0757 1.218 (0.979, 1.514) ≤ 10* ≤ 10* 0 (− 0.374, 0.374) 1 1 (0.417, 2.398) Myocardial Infarction 13 ≤ 10* 0.13 (− 0.263, 0.524) 0.5162 1.312 (0.576, 2.986) 392 339 0.154 (0.005, 0.302) 0.0424 1.161 (1.005, 1.342) ≤ 10* 12 − 0.086 (− 0.487, 0.315) 0.6758 0.837 (0.362, 1.933) Acute Kidney Injury 30 28 0.124 (− 0.537, 0.785) 0.7126 1.101 (0.66, 1.836) 818 846 − 0.056 (− 0.297, 0.185) 0.648 0.978 (0.89, 1.075) 42 28 0.716 (− 0.057, 1.49) 0.0688 1.548 (0.963, 2.487) Stroke 13 ≤ 10* 0.129 (− 0.265, 0.522) 0.5221 1.307 (0.573, 2.991) 262 274 − 0.031 (− 0.158, 0.096) 0.6337 0.96 (0.811, 1.136) ≤ 10* 12 − 0.087 (− 0.487, 0.313) 0.6706 0.834 (0.361, 1.927) Death 44 46 − 0.082 (− 0.844, 0.68) 0.833 0.957 (0.635, 1.441) 1828 1538 0.797 (0.493, 1.10) < 0.0001 1.19 (1.113, 1.271) 41 42 − 0.043 (− 0.801, 0.715) 0.9118 0.976 (0.637, 1.495) Anxiety 88 49 2.621 (1.437, 3.805) < 0.0001 2.145 (1.521, 3.024) 1351 1055 1.472 (1.143, 1.802) < 0.0001 1.424 (1.316, 1.541) 109 84 1.885 (0.298, 3.473) 0.0194 1.388 (1.053, 1.83) Depression 66 35 1.522 (0.648, 2.397) 0.0006 2.009 (1.339, 3.014) 894 627 0.901 (0.674, 1.127) < 0.0001 1.492 (1.348, 1.65) 72 63 0.54 (− 0.546, 1.626) 0.3288 1.18 (0.846, 1.646) Opioid use 18 ≤ 10* 0.345 (− 0.089, 0.778) 0.1183 1.831 (0.847, 3.959) 306 91 0.602 (0.494, 0.709) < 0.0001 3.431 (2.716, 4.333) 37 17 0.972 (0.321, 1.622) 0.0031 2.31 (1.304, 4.089) Chronic Pain 103 95 0.63 (− 0.606, 1.867) 0.3169 1.149 (0.875, 1.509) 2086 1451 2.385 (2.021, 2.749) < 0.0001 1.53 (1.434, 1.633) 149 100 2.678 (1.184, 4.171) 0.0004 1.549 (1.211, 1.98) Readmission 183 141 1.735 (0.328, 3.142) 0.0157 1.298 (1.05, 1.605) 2260 1502 2.051 (1.734, 2.367) < 0.0001 1.505 (1.412, 1.604) 281 172 4.662 (2.972, 6.352) < 0.0001 1.634 (1.363, 1.958)
Among medical complications, cannabis-only users had higher transfusion rates (RD: 1.386%; RR: 1.705; 95% CI: 1.189–2.446). No significant differences were seen for PE (RD: 0.289%; RR: 1.212; 95% CI: 0.767–1.915), pneumonia (RD: –0.175%; RR: 0.844; 95% CI: 0.480–1.485), respiratory failure (RD: –0.071%; RR: 0.933; 95% CI: 0.525–1.659), ARDS (RD: 0%; RR: 1.000; 95% CI: 0.417–2.398), myocardial infarction (RD: 0.130%; RR: 1.312; 95% CI: 0.576–2.986), acute kidney injury (RD: 0.124%; RR: 1.101; 95% CI: 0.660–1.836), stroke (RD: 0.129%; RR: 1.307; 95% CI: 0.573–2.991), and death (RD: –0.082%; RR: 0.957; 95% CI: 0.635–1.441). Coagulation parameters were not significantly different (PTT: 35.478 vs 34.809 s, P = 0.6097; PT: 13.585 vs 14.213 s, P = 0.0976).
Psychosocial complications were also more common, including anxiety (RD: 2.621%; RR: 2.145; 95% CI: 1.521–3.024), depression (RD: 1.522%; RR: 2.009; 95% CI: 1.339–3.014), and readmission (RD: 1.735%; RR: 1.298; 95% CI: 1.050–1.605). No significant differences were observed for opioid use (RD: 0.345%; RR: 1.831; 95% CI: 0.847–3.959) or chronic pain (RD: 0.630%; RR: 1.149; 95% CI: 0.875–1.509). Comparative risk ratios for outcomes can be found in Fig. 2. Survival curves for reoperation for cannabis only users can be found in Fig. 3.
Nicotine only users
Nicotine-only users demonstrated elevated rates of surgical, medical, and psychosocial complications relative to matched non-users. Specific values and results of these outcome analyses can be found in Table 2. Surgical risks were consistently higher in nicotine-only users, including wound dehiscence (RD: 0.536%; RR: 1.754; 95% CI: 1.508–2.040), superficial wound infection (RD: 1.745%; RR: 1.435; 95% CI: 1.340–1.538), deep implant infection (RD: 1.114%; RR: 1.531; 95% CI: 1.400–1.675), nosocomial infection (RD: 0.120%; RR: 1.559; 95% CI: 1.176–2.067), nonunion or malunion (RD: 1.063%; RR: 1.421; 95% CI: 1.308–1.545), nerve palsy (RD: 0.241%; RR: 1.304; 95% CI: 1.116–1.523), irrigation and debridement (RD: 0.172%; RR: 1.328; 95% CI: 1.100–1.603), and reoperation (RD: 2.491%; RR: 1.236; 95% CI: 1.188–1.286). No significant differences were observed for amputation (RD: 0.019%; RR: 1.701; 95% CI: 0.779–3.715), DVT (RD: 0.052%; RR: 1.189; 95% CI: 0.912–1.551), or PE (RD: 0.141%; RR: 1.080; 95% CI: 0.972–1.201).
Medical complications with significantly higher rates among nicotine only users included transfusion (RD: 0.751%; RR: 1.273; 95% CI: 1.164–1.370), pneumonia (RD: 0.508%; RR: 1.236; 95% CI: 1.122–1.361), myocardial infarction (RD: 0.154%; RR: 1.161; 95% CI: 1.005–1.342), and death (RD: 0.797%; RR: 1.190; 95% CI: 1.113–1.271). No significant differences were found for respiratory failure (RD: 0.171%; RR: 1.067; 95% CI: 0.973–1.170), ARDS (RD: 0.087%; RR: 1.218; 95% CI: 0.979–1.514), acute kidney injury (RD: –0.056%; RR: 0.978; 95% CI: 0.890–1.075), or stroke (RD: –0.031%; RR: 0.960; 95% CI: 0.811–1.136). Coagulation parameters showed a significantly shorter PT among nicotine only users (14.026 vs 14.750 s; P < 0.0001) and no significant difference in PTT (35.202 vs 35.362 s; P = 0.6112) (Table 3).Outcome Comparison Mean (Group 1) SD (Group 1) Mean (Group 2) SD (Group 2) t P-value Partial thromboplastin time (PTT) Cannabis-only users vs non-users 35.478 16.5 34.809 12.883 0.511 0.6097 Nicotine-only users vs non-users 35.202 15.58 35.362 15.471 − 0.508 0.6112 Concurrent users vs cannabis only users 34.484 15.091 35.612 17.867 − 0.837 0.4029 Prothrombin time (PT) Cannabis-only users vs non-users 13.585 3.663 14.213 6.477 − 1.659 0.0976 Nicotine-only users vs non-users 14.026 5.579 14.75 6.845 − 6.723 < 0.0001 Concurrent users vs cannabis-only users 13.758 5.085 13.604 3.519 0.512 0.6091
Psychosocial morbidity was also greater among nicotine-only users, including anxiety (RD: 1.472%; RR: 1.424; 95% CI: 1.316–1.541), depression (RD: 0.901%; RR: 1.492; 95% CI: 1.348–1.650), opioid use (RD: 0.602%; RR: 3.431; 95% CI: 2.716–4.333), chronic pain (RD: 2.385%; RR: 1.530; 95% CI: 1.434–1.633), and readmission (RD: 2.051%; RR: 1.505; 95% CI: 1.412–1.604). Comparative risk ratios for outcomes can be found in Fig. 4. Survival curves for reoperation for cannabis only users can be found in Fig. 5.
Concurrent users
Concurrent cannabis-and-nicotine users had higher postoperative risk compared with cannabis-only users. Specific values and results of these outcome analyses can be found in Table 2. Surgical complications were notably higher, including superficial wound infection (RD: 2.823%; RR: 1.703; 95% CI: 1.309–2.217), deep implant infection (RD: 1.795%; RR: 1.742; 95% CI: 1.260–2.408), amputation (RD: 0.428%; RR not estimable), and reoperation (RD: 4.149%; RR: 1.290; 95% CI: 1.132–1.470). No significant differences were observed for wound dehiscence (RD: 0.303%; RR: 1.336; 95% CI: 0.761–2.345), nosocomial infection (RD: 0%; RR: 1.000; 95% CI: 0.417–2.399), nonunion or malunion (RD: 1.133%; RR: 1.317; 95% CI: 0.994–1.746), nerve palsy (RD: –0.162%; RR: 0.864; 95% CI: 0.497–1.502), irrigation and debridement (RD: 0.282%; RR: 1.340; 95% CI: 0.740–2.427), DVT (RD: 0%; RR: 0.999; 95% CI: 0.417–2.396), and PE (RD: 0%; RR: 1.000; 95% CI: 0.644–1.553).
Among medical complications, concurrent users had higher rates of transfusion (RD: 1.150%; RR: 1.371; 95% CI: 1.010–1.860), pneumonia (RD: 0.753%; RR: 1.842; 95% CI: 1.070–3.172), and respiratory failure (RD: 0.876%; RR: 1.942; 95% CI: 1.136–3.378). No significant differences were observed for ARDS (RD: 0%; RR: 1.000; 95% CI: 0.417–2.398), myocardial infarction (RD: –0.086%; RR: 0.837; 95% CI: 0.362–1.933), acute kidney injury (RD: 0.716%; RR: 1.548; 95% CI: 0.963–2.487), stroke (RD: –0.087%; RR: 0.834; 95% CI: 0.361–1.927), or death (RD: –0.043%; RR: 0.976; 95% CI: 0.637–1.495).
Psychosocial complications were also elevated, including anxiety (RD: 1.885%; RR: 1.388; 95% CI: 1.053–1.830), opioid use (RD: 0.972%; RR: 2.310; 95% CI: 1.304–4.089), chronic pain (RD: 2.678%; RR: 1.549; 95% CI: 1.211–1.980), and readmission (RD: 4.662%; RR: 1.634; 95% CI: 1.363–1.958). No significant differences were observed for depression (RD: 0.540%; RR: 1.180; 95% CI: 0.846–1.646). Coagulation parameters were not significantly different (PTT: 34.484 vs 35.612 s, P = 0.4029; PT: 13.758 vs 13.604 s, P = 0.6091). Comparative risk ratios for outcomes can be found in Fig. 6. Survival curves for reoperation for cannabis only users can be found in Fig. 7.
Sensitivity analysis for cannabis exposure ascertainment
A sensitivity analysis was performed to assess ascertainment of cannabis exposure in the TriNetX database. According to the National Survey on Drug Use and Health (NSDUH), approximately 6.8% of Americans age 12 or older report having a cannabis use disorder. Applying this prevalence to more than 150 million patients in the TriNetX Research database network would predict roughly 10.2 million cannabis users. However, only 1,666,438 patients (1.11%) in the database carried an ICD-10-CM diagnosis code for cannabis-related disorders. This discrepancy suggests substantial underascertainment of cannabis exposure in structured electronic health record diagnosis fields.
Sensitivity analysis for unmeasured confounding
E-values were calculated for all statistically significant associations to assess vulnerability to unmeasured confounding (Supplemental Table 2). Among cannabis-only users, E-values ranged from 1.92 (reoperation, readmission) to 3.95 (deep implant infection). Among nicotine-only users, E-values ranged from 1.59 (myocardial infarction) to 6.32 (opioid use). Among concurrent users, E-values ranged from 1.90 (reoperation) to 4.05 (opioid use). The majority of significant associations had E-values exceeding 2.0, indicating that a substantial degree of unmeasured confounding would be required to fully explain the observed effects. Full E-value measurements can be found in Supplemental Table 2.
Discussion
In this large retrospective cohort, cannabis use was associated with higher rates of surgical and psychosocial complications after orthopedic trauma surgery. Nicotine use was linked to a broader range of adverse outcomes, and concurrent exposure was associated with compounded risk. These findings indicate that documented cannabis and nicotine use may identify patients at increased risk for postoperative morbidity in surgical populations.
These findings are consistent with emerging evidence that cannabis exposure may adversely influence perioperative outcomes through immune, vascular, and metabolic mechanisms. Experimental data show that cannabinoids modulate immune and inflammatory responses by suppressing cytokine signaling, promoting regulatory immune phenotypes, and reducing monocyte-derived interleukin-1β production13,14. Such effects may weaken host defenses, increasing susceptibility to infection and delayed tissue repair.
The increased rates of nonunion and malunion among cannabis users may reflect dose-dependent effects on bone remodeling; however, this mechanistic interpretation remains speculative because TriNetX does not capture cannabis dose, frequency, route, chronicity, or recency of exposure. Prior experimental studies suggest that low cannabinoid concentrations may promote osteoblast differentiation, whereas higher concentrations may inhibit bone formation and enhance osteoclast activity, resulting in a biphasic response15,16. Because exposure intensity could not be quantified in this cohort, these biological mechanisms should be interpreted as plausibility arguments rather than direct evidence of causation. In addition to dose-related effects, age may influence the relationship between cannabis use and outcomes, with some literature indicating more frequent medical complications in older populations undergoing procedures such as arthroplasty and fragility fracture fixation, compared with procedures performed for high energy trauma in younger patients8–12.
Although we observed an association between cannabis use and increased rates of nerve palsy, this finding should be considered exploratory. There is currently no high-quality evidence directly linking cannabis exposure to peripheral nerve injury, and consensus guidelines do not recognize neuropathy or nerve damage as established complications of cannabis use17. This association may reflect coding artifact, differences in injury pattern, surgical positioning, or residual confounding rather than a causal effect of cannabis exposure. Further studies with procedure-level detail, laterality, neurologic examination findings, and adjudicated nerve injury outcomes are needed to determine whether this association is reproducible.
Nicotine’s known vasoconstrictive and anti-osteogenic effects likely contribute to its broader and more pronounced association with adverse surgical and medical outcomes. The consistency of these findings with prior literature also supports the validity of the analytic approach and strengthens the observed cannabis related signals18,19. Its inclusion in this study also serves as an internal control, reinforcing the validity of our methodology. The higher complication rates observed among concurrent cannabis-and-nicotine users suggest that combined exposure may compound perioperative risk. Prior work has identified similarly elevated complication rates among patients with combined cannabis and tobacco exposure after total joint arthroplasty and ankle fracture fixation20,21. While both substances may affect immune cell recruitment and resolution of inflammation, their combined influence may plausibly worsen wound healing through overlapping mechanisms.
No increase in venous thromboembolism was observed, although prior studies have reported conflicting results. Cannabis exposure has been linked to platelet activation and endothelial dysfunction, which may promote a prothrombotic state even when standard coagulation assays appear normal22,23. Consistent with this possibility, LoPolito et al. found prolonged thromboelastography (TEG) reaction times among trauma patients with cannabis exposure, suggesting delayed clot initiation24. These findings indicate that advanced coagulation assays may better detect subtle cannabis-related changes and help refine perioperative risk stratification.
Cannabis, nicotine, and concurrent use were each associated with higher rates of anxiety, depression, chronic pain, and readmission. These findings underscore the psychosocial dimension of substance use in surgical recovery and the importance of multidisciplinary perioperative care that includes mental-health and substance-use assessment. Although some literature suggests cannabis may reduce opioid requirements, we observed no protective effect, and concurrent use was associated with greater opioid utilization. Comprehensive perioperative screening and counseling integrating pain management, mental health, and substance use evaluation are warranted. This approach is consistent with ASRA Pain Medicine consensus guidelines, which recommend universal preoperative screening for cannabis use, including assessment of product type, route of administration, timing of last use, amount, and frequency17.
Limitations
This study has several limitations inherent to retrospective analyses of electronic health record data. Cannabis and nicotine exposure classification relied on ICD-10-CM diagnosis codes rather than direct measures of substance use behavior. These codes may undercapture true exposure because of inconsistent screening, patient nondisclosure, stigma, and institutional coding variation, which could bias results toward the null. Therefore, the cannabis cohort likely represents patients with documented cannabis-related diagnoses rather than all patients who use cannabis. The database does not capture cannabis dose, frequency, route, or recency of use, and detailed socioeconomic or behavioral variables were unavailable. Although residual confounding is possible, rigorous propensity score matching and adjustment for major comorbidities, fracture characteristics, and polysubstance use strengthen the validity of our findings. Because this study evaluated multiple outcomes across three cohort comparisons, the use of a nominal P < 0.05 threshold without formal multiplicity adjustment increases the possibility of type I error. Statistically significant findings, particularly borderline associations, should therefore be interpreted cautiously. Although 1:1 propensity score matching generated matched cohorts, TriNetX performs post-matching analyses using independent-sample methods rather than paired or conditional approaches. Accordingly, paired t tests, conditional logistic regression, stratified Cox models, and IPTW could not be performed, and results should be interpreted as matched-cohort rather than conditional matched-pair analyses. Injury Severity Scores and time to antibiotic administration could not be directly assessed, but surrogate variables for polytrauma and preoperative antibiotic exposure were incorporated to mitigate these effects. Despite these limitations, the study’s large sample size and multi-institutional design provide an estimate of cannabis- and nicotine-associated perioperative risk.
Conclusion
Cannabis use was associated with adverse perioperative outcomes after lower extremity fracture fixation. Using a large, multi-institutional electronic health record dataset, this study identified associations between cannabis and nicotine exposure and adverse surgical and psychosocial outcomes. These results support further prospective evaluation of standardized substance use screening and counseling strategies and highlight the need for prospective, dose response, and mechanistic investigations to guide clinical practice and public health policy.
Supplementary Information
Below is the link to the electronic supplementary material.
Acknowledgement
Dr. Hamad is supported by the University of California, Los Angeles (UCLA) Regenerative Musculoskeletal Medicine Training Program, funded by the National Institute of Arthritis and Musculoskeletal and Skin Diseases (NIAMS) of the National Institutes of Health (NIH) under National Research Service Award 2T32AR059033. The data were provided by the Clinical and Translational Science Collaborative of Northern Ohio, funded by the National Center for Advancing Translational Sciences (NCATS) of the NIH (UM1TR004528). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.
Data availability
The data that support the findings of this study are derived from the TriNetX Research Network, a federated electronic health record platform that aggregates de-identified electronic health records from participating healthcare organizations. The data are accessible only to institutions with active TriNetX subscriptions and cannot be shared publicly due to licensing restrictions and patient privacy regulations. Researchers affiliated with member institutions can access the underlying data directly via the TriNetX Analytics platform to replicate or extend the analyses described in this study.
Declarations
Competing interests
The authors declare that they have no financial or personal relationships that could have influenced the work reported in this manuscript.