A large matched cohort study of over 13,000 patients found that a pre-existing diagnosis of cannabis use disorder was associated with roughly triple the rate of medical complications within 90 days of wrist fracture surgery. However, the observational design and reliance on insurance claims data mean the study cannot determine whether cannabis itself, or the many unmeasured health and social factors that travel with this diagnosis, drives the difference.

Cannabis use is increasing rapidly across the United States, and the prevalence of cannabis use disorder among surgical patients appears to be doubling each decade. Hand and wrist surgeons, who perform thousands of open reduction and internal fixation procedures for distal radius fractures annually, need reliable data to inform preoperative risk conversations. This study represents one of the first large-scale investigations specifically examining whether a CUD diagnosis predicts worse postoperative outcomes in this common orthopedic procedure, arriving at a moment when clinicians are navigating conflicting signals about cannabis as both a therapeutic tool and a potential surgical risk factor.

Distal radius fractures are among the most common upper extremity injuries treated surgically, and open reduction and internal fixation remains the standard of care for unstable fractures in working-age adults. As cannabis legalization expands across the United States, clinicians have observed rising rates of cannabis use disorder among their surgical patients, yet the orthopedic literature has offered limited data on how this diagnosis relates to postoperative outcomes in hand surgery specifically. This study draws on the PearlDiver/Mariner insurance claims database to assemble a matched cohort of over 13,000 patients, comparing 90-day outcomes between those with and without an ICD-coded CUD diagnosis, while controlling for tobacco use, alcohol abuse, opioid dependence, and multiple medical comorbidities through a 1:5 matching strategy and subsequent multivariable logistic regression.

The results showed CUD patients had significantly higher rates of virtually every tracked complication. The composite 90-day complication rate was 15.24% in the CUD group versus 5.76% in controls, yielding an adjusted odds ratio of 3.10 (95% CI: 2.69 to 3.57). Individual complications with the largest relative increases included myocardial infarction (OR 9.17), acute kidney injury (OR 3.86), pulmonary embolism (OR 3.50), and cerebrovascular accident (OR 3.27). Emergency department visits were more than doubled (OR 2.25) and readmissions were modestly elevated (OR 1.36). Notably, the prevalence of CUD in this surgical population doubled from 4% to 8% over the study decade. The authors acknowledge that the administrative data design cannot establish causality and that unmeasured confounders, including cannabis use frequency, route, and perioperative timing, remain significant limitations. They call for prospective research with more granular exposure characterization.

CUD as a Surgical Risk Marker: Large Matched Cohort Reveals 90-Day Outcome Gaps in Distal Radius ORIF

A patient with a documented history of cannabis use disorder walks into your hand surgery clinic for wrist fracture fixation. Should that diagnosis change your preoperative conversation? A new matched cohort study of more than 13,000 patients suggests the answer is yes, but for reasons that are more complex than the headlines might imply. The paper appears to claim that cannabis use disorder independently raises the risk of postoperative complications after wrist surgery. What it actually tested is whether the presence of an ICD billing code for cannabis abuse or dependence, at some point prior to surgery, correlates with worse 90-day outcomes in an insurance claims database. Those are meaningfully different questions. To the authors’ considerable credit, the study is well-powered, controls for an impressively thorough set of confounders including tobacco, alcohol, and opioid dependence, applies Bonferroni correction across multiple comparisons, and offers the first large-scale analysis of CUD-specific outcomes in hand trauma surgery. The epidemiological trend data showing a doubling of CUD prevalence in this surgical population over a single decade is genuinely valuable for clinical planning purposes.

The central methodological problem, however, is one that no amount of matching can fully resolve in administrative data: CUD is not just a cannabis exposure variable. It is a behavioral diagnosis that travels with an entire constellation of unmeasured factors, from housing instability and nutritional deficits to polysubstance use patterns and differential healthcare engagement. Identifying a patient as having CUD in an insurance database is like noting a car has been towed multiple times. It tells you something about the driver’s situation but not specifically what road conditions caused each incident, or whether fixing the car’s brakes would prevent the next tow. The 85% tobacco co-use rate in the CUD group is particularly instructive. Even after matching on tobacco status, residual differences in smoking intensity, duration, and perioperative cessation compliance could account for a substantial share of the observed complication gap. No sensitivity analysis, such as an E-value calculation, was performed to quantify how strong an unmeasured confounder would need to be to explain away the associations. This matters because the most clinically actionable question, whether advising patients to stop cannabis before surgery would actually reduce their complication rates, is precisely the question this study cannot answer.

To a patient, I would say that a CUD history belongs in the preoperative conversation as part of a broader risk picture, not as a reason for alarm in isolation. To a colleague, I would emphasize that this is useful signal-generating data, not a mandate for cannabis-specific perioperative protocols. To a policymaker, I would argue that the doubling of CUD prevalence demands investment in prospective research with validated exposure measures, not restrictive surgical access policies built on observational associations. The study also leaves entirely unexamined any potential analgesic benefits of cannabis, or whether CUD patients might use fewer opioids postoperatively, a tradeoff that matters clinically even if it lies outside the scope of this analysis. When a high-comorbidity behavioral diagnosis is associated with worse outcomes in administrative data, the critical question is always: how much of the outcome difference belongs to the exposure itself versus the constellation of factors that travel with it? Without that decomposition, which requires richer data than claims databases can provide, the clinically actionable signal remains uncertain even when the statistical signal is strong.

This study sits in the early, signal-generating phase of research on cannabis use disorder and orthopedic surgical outcomes. Prior matched cohort studies using similar administrative databases have identified comparable complication patterns in CUD patients undergoing joint arthroplasty and lumbar spine surgery, suggesting the association is not procedure-specific but rather reflects a broader perioperative risk profile associated with this diagnostic category. The hand surgery literature has until now lacked a large-scale dataset specifically addressing this question, making this paper a welcome addition even as it shares the inherent limitations of the database study genre.

From a pharmacological standpoint, clinicians should be aware that cannabis interacts with cytochrome P450 enzymes, potentially altering the metabolism of perioperative medications including certain anesthetics, anticoagulants, and analgesics. The potential for increased airway reactivity in heavy cannabis smokers is also relevant during intubation and recovery. However, the safety implications specific to distal radius ORIF, a procedure typically performed under regional anesthesia, may differ from those in major arthroplasty or spine surgery. Clinicians reviewing this data should incorporate CUD history into their preoperative assessment as one element among many, much as they would inquire about tobacco use, and should engage patients in a frank discussion about overall perioperative optimization rather than singling out cannabis cessation as a standalone evidence-based recommendation.

This is a retrospective matched cohort study using administrative insurance claims data, placing it in the lower-to-middle tier of the evidence hierarchy for clinical questions about therapeutic risk. While the large sample size and rigorous matching strategy lend statistical credibility, the reliance on billing codes rather than clinical records means the study cannot characterize actual cannabis exposure, verify diagnosis timing, or account for unmeasured confounders. The single most important inference constraint is that no causal relationship between cannabis use and surgical complications can be established from this design.

The findings are directionally consistent with prior administrative database studies that found CUD associated with increased complications following lower extremity joint arthroplasty and lumbar spine surgery, suggesting a pattern that extends across orthopedic subspecialties. The tobacco literature on distal radius fracture outcomes, including work by Galivanche and colleagues using NSQIP data and by Hall and colleagues in a multicenter series, has similarly identified substance use as a risk marker in this fracture population, lending biological plausibility to the broader association between behavioral health diagnoses and surgical outcomes. What this study adds is the first CUD-specific dataset in the hand surgery context and a valuable temporal trend analysis showing that this is a rapidly growing clinical population. What it cannot do is distinguish whether the association reflects a cannabis-specific biological mechanism, a marker of overall health complexity, or an artifact of differential healthcare engagement and documentation practices.

The most consequential analytic choice was the absence of a sensitivity analysis for unmeasured confounding. Calculating an E-value, a metric that quantifies the minimum strength an unmeasured confounder would need to explain away the observed association, would have provided readers with a concrete measure of robustness. Given the extensive comorbidity burden in the CUD group, including 85% tobacco co-use, it is plausible that even moderate unmeasured confounding could substantially attenuate the reported effect sizes. Additionally, a propensity score analysis with explicit balance verification through standardized mean differences, rather than p-values alone, would have strengthened confidence that the matched groups were truly comparable on observable characteristics. Without these steps, the reported odds ratios should be interpreted as upper-bound estimates of association rather than precise effect estimates.

The most likely overinterpretation is that cannabis use directly causes surgical complications after wrist fracture repair. The study identifies an association between a formal ICD-coded cannabis use disorder diagnosis and worse outcomes, but this diagnosis captures a specific, relatively severe diagnostic category, not all cannabis users. Recreational or occasional cannabis consumers are not represented in this dataset, and extrapolating these findings to the broader cannabis-using population would substantially exceed what the evidence supports. Similarly, the dramatic myocardial infarction odds ratio of 9.17 may appear to implicate cannabis as powerfully cardiotoxic in the surgical setting, but this figure rests on only 17 events among 2,297 CUD patients, making it highly sensitive to even small amounts of residual confounding or coding error. The study provides no data on whether cannabis cessation before surgery would reduce complications, and interpreting it as evidence for such a recommendation is premature.

This study contributes the first large-scale matched analysis of cannabis use disorder and postoperative outcomes in distal radius fracture surgery, revealing a statistically robust association between CUD diagnosis and substantially elevated 90-day complication rates. It does not establish that cannabis causes these complications, does not demonstrate that perioperative cessation would improve outcomes, and cannot distinguish between cannabis-specific effects and the broader comorbidity burden carried by patients with this diagnosis. For now, CUD history belongs in the preoperative risk conversation as a flag for comprehensive perioperative planning, not as an independently actionable risk factor.

Does this study prove that cannabis causes complications after wrist surgery?

No. The study shows an association between a cannabis use disorder diagnosis and higher complication rates, but it cannot prove causation. Patients with CUD carry many other health conditions and social factors that independently predict worse surgical outcomes. The study design does not allow researchers to separate the effect of cannabis itself from these other factors.

Should I stop using cannabis before having wrist surgery?

This study does not provide direct evidence that stopping cannabis before surgery reduces complication risk. However, it is always wise to discuss all substance use, including cannabis, with your surgical team so they can make the best perioperative plan for your individual situation. Your surgeon may have additional reasons based on your overall health profile to recommend reducing or stopping use before a procedure.

If I use cannabis occasionally, does this study apply to me?

The study specifically looked at patients with a formal cannabis use disorder diagnosis, which represents a more severe pattern of use that meets specific clinical criteria. Occasional or low-frequency cannabis users were not identified or studied in this dataset. The findings should not be directly applied to people who use cannabis recreationally or occasionally.

Why were complication rates so much higher in the cannabis use disorder group?

The honest answer is that we do not know for certain. While the researchers tried to account for common overlapping conditions like tobacco use, alcohol abuse, and depression, patients with a CUD diagnosis often face additional challenges that insurance claims data cannot capture, including housing instability, nutritional deficits, and more complex patterns of substance use. Some or all of the elevated complication rates may be explained by these unmeasured factors rather than by cannabis itself.

Have thoughts on this? Share it: