Abstract
We aimed to examine the prevalence of documented cannabis use and its association with survival among patients with HF using the Optum Clinoformatics dataset, large, national informatics dataset. We demonstrated low period prevalence of cannabis use among patients with heart failure and there was no statistically significant difference (p = 0.76) in all-cause mortality between individuals who use cannabis versus non-users.
Keywords: Cannabis, Heart failure, Prevalence, Survival
1. Introduction
Cannabis use is increasing across the U.S., including among older adults aged >65 years [1]. Simultaneously, the prevalence of heart failure (HF) continues to rise, and heart failure is more prevalent in the elderly. Cannabis is becoming more available for both medicinal and recreational use, but the prevalence of cannabis use in patients with HF is not well studied. Prior population-based data estimate that cannabis use disorder affects approximately 0.3–1.3% of individuals [2], with increasing trends over the past decade, but these figures may underestimate use in patients with HF. In prior work, we evaluated urine drug screening in a large health system. We found that approximately 1 in 3 patients with HF tested positive for cannabis, suggesting that use may be more common in this population than previously recognized [3]. Clinician opinions and practices vary widely regarding cannabis and HF [4]. Patients with HF may use cannabis to manage symptoms of HF, such as discomfort related to chronic abdominal and leg swelling, pain, and anxiety [5]. Cannabis use may be beneficial in patients who have early satiety, anorexia, and cardiac cachexia also markers of end stage HF [6]. This should also be taken into context with some emerging data suggesting cardiovascular harm with cannabis use. While data for the cardiovascular effects of cannabis on atrial fibrillation and coronary artery disease are emerging, no data exist on the impact of cannabis use on survival and all-cause mortality among patients with heart failure.
Categorizing cannabis use in patients with HF and their association with outcomes has important implications for clinical care, including transplant candidacy, counseling, and long-term management. We aimed to examine the prevalence of documented cannabis use and its association with survival among patients with HF using a large, national informatics dataset.
2. Methods
2.1. Study population
Using the Optum Clinformatics® dataset (Optum, Inc., Eden Prairie, MN, USA), we performed a retrospective, longitudinal analysis to identify patients hospitalized with a diagnosis of HF over a 10-year period from January 1, 2010, to May 31, 2021. Optum includes patients with commercial or Medicare insurance, including member information, medical and pharmacy claims, laboratory results, and inpatient stays, and has been validated in prior studies. To be included, individuals needed to have at least 6 months of continuous enrollment in the database before their first identified diagnostic code for HF (index date) and they had to be above 18 years old. International Classification of Diseases, Tenth Revision (ICD-10-CM) diagnosis codes were used to identify patients with primary discharge diagnosis of HF (I50) and cannabis use (F12). Patients identified with HF were categorized into those who used cannabis and non-users. We defined users as follows: I) Individuals who use cannabis (ICD-10-CM code F12) and had a principal inpatient diagnosis of heart failure (ICD-10-CM code I50), and II) non-users as patients identified with ICD codes for HF but without an accompanying ICD-10-CM code for cannabis use. We excluded patients with less than 6 months of continuous enrollment. We evaluated all-cause mortality between individuals who use cannabis and non-users. The Indiana University Institutional Review Board waived informed consent because the data were deidentified. We followed the Strengthening the Reporting of Observational Studies in Epidemiology reporting guideline for study design, execution, analysis, interpretation, and reporting. We addressed missing data as follows to retain all individuals in the analysis. For demographic variables (sex, race, and insurance type), we created an “Unknown” category. For comorbidities and risk factors (e.g., smoking, substance use), the absence of relevant ICD codes in a patient's record was interpreted as the absence of the condition and coded as 0. This approach is standard in claims-based analyses and allowed us to minimize the impact of missing data while maintaining the full study population.
2.2. Statistical analysis
The primary outcome was all-cause mortality among patients with HF. We also evaluated the prevalence of cannabis use in this cohort. Continuous variables were summarized by mean and standard deviation, and categorical variables by counts and percentages. We estimated propensity scores using logistic regression. Individuals who use cannabis were then matched 1:5 to non-users using the nearest-neighbor algorithm based on propensity scores. The propensity score model for the HF population includes the following covariates: age at first HF diagnosis, sex, race/ethnicity, insurance type, diabetes mellitus, chronic liver disease, hypertension, atrial fibrillation/atrial flutter, acute myocardial infarction, chronic kidney disease, and alcohol, cocaine, or tobacco use. Cox proportional hazards regression models were used to compare all-cause mortality between both groups. Matched sets were included as a frailty term to account for the matching structure.
3. Results
We identified 389,280 inpatients with HF between January 1, 2011, and May 31, 2021. After excluding 94,496 with <6 months of enrollment, 294,784 remained in the final cohort. Cohort included cannabis users (5541) and non-users (289,243). We then performed propensity matching at a ratio of 1:5, and matching was based on demographics, payer status, and clinical comorbidities for the entire matched set and by cannabis use status. After propensity matching, the sample size of the matched dataset was 33,246 (5,541 individuals who use cannabis, and 27,705 cannabis non-users) (Table 1). The average age of patients in the matched dataset at first HF diagnosis was 61.5 years (±12.1). There were 31% (10,365) women, and 35% (11,561) had Medicare insurance. The cohort had 6.9% (2287) Hispanic, 30% (9986) Black, and 56% (18,671) White participants. The commonest comorbidities were chronic kidney disease 78% (25,797), diabetes mellitus 61% (20,331), atrial fibrillation/flutter 54% (17,946), and hypertension 41% (13,591. The period prevalence of cannabis use in patients with HF was 1.88% [5541/294,784 (95% CI: 1.85, 1.95)] (Table 2). A higher prevalence of cannabis use was observed among males (2.7%), Black patients (3.2%), and Medicare low-income subsidy (3.5%) or Medicare Dual beneficiaries (3.3%). Over the 10-year study period, there were 2059 all-cause deaths among 5541 individuals who use cannabis, with a median survival of 4.40 years (95% CI: 4.18, 4.62). In the non-user group, 11,350 all-cause deaths were observed among 27,705 non-users, with a median survival of 4.47 years (95% CI: 4.39, 4.59). There was no statistically significant difference (p = 0.76) in all-cause mortality between the groups (Fig. 1).
Table 1.
Demographics, payer status, and clinical features of Patients with Heart Failure (inpatients) are summarized for the entire matched set and by cannabis use status, after propensity score matching at the ratio 1:5.
| HF cohort (N = 33,246) | Cannabis user (N = 5541) |
Cannabis non-users (N = 27,705) | Absolute Standardized mean difference | ||
|---|---|---|---|---|---|
| Age at 1st Heart Failure diagnosis | 61.5 (12.1) | 61.6 (11.1) | 61.5 (12.0) | 0.010 | |
| Gender | Female | 10,365 (31%) | 1738 (31%) | 8627 (31%) | 0.005 |
| Male | 22,867 (69%) | 3801 (69%) | 19,066 (69%) | 0.005 | |
| Unknown | 14 (<0.1%) | 2 (<0.1%) | 12 (<0.1%) | 0.004 | |
| Race | Asian | 286 (0.9%) | 61 (1.1%) | 225 (0.8%) | 0.028 |
| Black | 9986 (30%) | 1641 (30%) | 8345 (30%) | 0.011 | |
| Hispanic | 2287 (6.9%) | 422 (7.6%) | 1865 (6.7%) | 0.033 | |
| White | 18,671 (56%) | 3061 (55%) | 15,610 (56%) | 0.022 | |
| Unknown | 2016 (6.1%) | 356 (6.4%) | 1660 (6.0%) | 0.018 | |
| Insurance | Commercial | 5179 (16%) | 828 (15%) | 4351 (16%) | 0.021 |
| Medicare | 11,561 (35%) | 1946 (35%) | 9615 (35%) | 0.009 | |
| Medicare Dual / LIS | 13,191 (40%) | 2196 (40%) | 10,995 (40%) | 0.001 | |
| Missing/ Inconsistent | 3315 (10.0%) | 571 (10%) | 2744 (9.9%) | 0.013 | |
| Diabetes mellitus | 20,331 (61%) | 3324 (60%) | 17,007 (61%) | 0.029 | |
| Hypertension | 13,591 (41%) | 2299 (41%) | 11,292 (41%) | 0.015 | |
| Chronic kidney disease | 25,797 (78%) | 4296 (78%) | 21,501 (78%) | 0.002 | |
| Chronic liver disease | 4852 (15%) | 861 (16%) | 3991 (14%) | 0.032 | |
| Acute myocardial infarction | 3240 (9.7%) | 583 (11%) | 2657 (9.6%) | 0.031 | |
| Atrial fibrillation/flutter | 17,946 (54%) | 2994 (54%) | 14,952 (54%) | 0.001 | |
| Alcohol use | 1754 (5.3%) | 594 (11%) | 1160 (4.2%) | 0.251 | |
| Cocaine use | 377 (1.1%) | 220 (4.0%) | 157 (0.6%) | 0.230 | |
| Tobacco use | 20,158 (61%) | 4165 (75%) | 15,993 (58%) | 0.376 | |
| Any substance use | 20,552 (62%) | 4265 (77%) | 16,287 (59%) | 0.379 |
Table 2.
Period Prevalence of Cannabis Use in the Heart Failure Cohort.
| Characteristic | Incidence of Cannabis use (N = 5541) |
|
|---|---|---|
| Gender | Female | 1738/151641 (1.1%) |
| Male | 3801/143092 (2.7%) | |
| Unknown | 2/51 (3.9%) | |
| Race | Asian | 61/5506 (1.1%) |
| Black | 1641/51149 (3.2%) | |
| Hispanic | 422/26934 (1.6%) | |
| White | 3061/194264 (1.6%) | |
| Unknown | 356/16931 (2.1%) | |
| Payer status | Commercial | 828/33096 (2.5%) |
| Medicare | 1946/173053 (1.1%) | |
| Medicare Dual | 1102/33705 (3.3%) | |
| Medicare Low-Income Subsidy | 1094/30966 (3.5%) | |
| Missing + Inconsistent | 571/23964 (2.4%) | |
| Comorbidities | Diabetes mellitus | 3324/156345 (2.1%) |
| Hypertension | 2299/175647 (1.3%) | |
| Chronic kidney disease | 4296/201650 (2.1%) | |
| Chronic liver disease | 861/19376 (4.4%) | |
| Acute myocardial infarction | 583/17295 (3.4%) | |
| Atrial fibrillation/flutter | 2994/194613 (1.5%) | |
| Other substance use | Alcohol use | 594/8372 (7.1%) |
| Tobacco use | 4165/151249 (2.8%) | |
| Cocaine use | 220/712 (30.9%) | |
| Any substance use | 4265/153696 (2.8%) |
Fig. 1.
Kaplan Meier curve of the primary outcome of all-cause mortality among patients with heart failure in the propensity-matched set (N = 294,784).
4. Discussion
This study offers a contemporary evaluation of medical record-documented cannabis use among patients with HF and its association with survival in a nationally representative dataset. Findings suggest that few patients (1.88% of patients hospitalized with HF) had a concomitant diagnostic code for cannabis use, with higher prevalence observed among males, Black patients, and Medicare beneficiaries. All-cause mortality was similar between those who use cannabis and non-users.
These findings suggest there may not be a need for screening for cannabis use since cannabis use may not be a risk factor for poorer outcomes, given that both cohorts have similar survival after propensity matching. Furthermore, even if a diagnosis of cannabis use/abuse or dependence is made, treatment (i.e. motivational enhancement therapy with cognitive-behavioral therapy) can be difficult to access. Data regarding the cannabis use among patients with HF is limited. In another study of populations admitted for acute decompensated HF, patients with diagnosis codes for cannabis use disorders showed lower in-hospital mortality and shorter hospital stays [7]. A prior study showed that medically authorized patients had an increased risk for HF-related ED visits or hospitalizations following cannabis authorization [8]. A cross-sectional study of a United States (US) inpatient sample showed an increased risk of heart failure associated with cannabis use [9]. Among adults, following noncardiac surgery, one study reported that cannabis use disorders were associated with a 26% increase in the odds of major cardiovascular events, which include acute heart failure [10]. Overall, the existing evidence regarding the association between cannabis and outcomes in HF remains inconclusive.
No consensus exists for heart transplant listing for patients who use cannabis. Health care providers generally agreed that a validated cannabis use disorder screening questionnaire would be helpful and that 6 months of abstinence from cannabis is sufficient prior to HT listing [4]. Significant heterogeneity exists regarding cannabis use as it relates to heart transplantation, and no guidelines exist for cannabis screening apart from during advanced HF therapies evaluation. Significantly more Black patients were deemed ineligible for heart transplants due to substance use in a recent study, and this may have been due to inequities in testing [11]. The lack of mortality difference between those who use cannabis and non-users differs from prior reports, which had suggested cardiovascular harm, highlighting the need for further investigation [12]. The strengths of this study include a large, diverse sample size and a more contemporary timeline to evaluate HF outcomes and trends, given the increasing use of cannabis due to more widespread legalization in the US.
4.1. Limitations
Misclassification bias is common in claims-based analysis. Claims data may not be sensitive to identifying patients who use cannabis because physicians do not consistently document medical cannabis use in medical records [13]. Despite our attempts to propensity match the two patient characteristics, the likelihood of residual confounding remains, especially regarding social determinants of health, which may not be adequately captured in a claims dataset. Our analysis required that individuals have a minimum of 6 months' continuous enrollment in a health insurance plan and may overrepresent those with social and financial stability able to maintain enrollment for this duration and underestimate the representation of patients with lower income. Another limitation of our study was that we lumped a range of severity of cannabis use (mild, moderate or severe cannabis use) together. Our sample size from a claims data set limits the generalizability of our findings. In addition, we did not stratify patients by heart failure subtype (reduced vs. preserved ejection fraction).
5. Conclusions
This paper suggests that there may be no difference in survival among patients hospitalized with HF regardless of cannabis use or non-use, and a low prevalence of cannabis use among hospitalized patients with HF is ∼2%. Further research should focus on i) establishing the cardiovascular impact of cannabis use in patients with HF among different populations/datasets, ii) assessing the impact of medicinal cannabis use on symptoms of HF, such as anorexia/cardiac cachexia, anxiety, and shortness of breath, and iii) assessing the cardiovascular impact of cannabis use on heart transplant recipients.
CRediT authorship contribution statement
Onyedika J. Ilonze: Writing – review & editing, Writing – original draft, Visualization, Validation, Methodology, Formal analysis, Data curation, Conceptualization. Xiaochun Li: Writing – review & editing, Project administration, Methodology, Investigation, Formal analysis, Data curation. Fangqian Ouyang: Project administration, Methodology, Investigation, Formal analysis. Larry A. Allen: Writing – review & editing, Project administration, Methodology, Investigation. Robert L. Page: Writing – review & editing, Project administration, Methodology, Investigation, Conceptualization. Sara K. Quinney: Writing – review & editing, Writing – original draft, Investigation. Denise Christina Vidot: Writing – review & editing, Investigation, Formal analysis. Sarah Chuzi: Writing – review & editing, Writing – original draft, Investigation. Khadijah Breathett: Writing – review & editing, Writing – original draft, Project administration, Investigation.
Disclosures
Dr. Onyedika Ilonze is supported by a Career Development Award from the American Heart Association (26CDA1591379) and the Winn Excellence in Clinical Trials Career Development Award. Dr. Sarah Chuzi is supported by a Career Development Award from the American Heart Association(25CDA1453935). These awards are unrelated to the current work.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
None.
References
- 1.Pravosud V., Lum E., Vali M., et al. Cannabis use among older adults. JAMA Netw. Open. 2025;8(5) doi: 10.1001/jamanetworkopen.2025.10173. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Chouairi F., Mullan C.W., Ravindra N., et al. Brief report: Cannabis and opioid use disorder among heart failure admissions, 2008-2018. PLoS One. 2021;16(9) doi: 10.1371/journal.pone.0255514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Ilonze O.J., Helmkamp L.J., CK McIlvennan, et al. Trends in urine drug screening and positivity among patients with heart failure: insights from a large health system. Am. Heart J. 2025;290:226–229. doi: 10.1016/j.ahj.2025.07.008. [DOI] [PubMed] [Google Scholar]
- 4.Ilonze O.J., Knapp S.M., Chernyak Y., et al. Cannabis use and heart transplant listing: a survey of clinician practices. PLoS One. 2024;19(12) doi: 10.1371/journal.pone.0310778. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Ilonze O.J., Vidot D.C., Breathett K., et al. Cannabis use and heart transplantation: disparities and opportunities to improve outcomes. Circ. Heart Fail. 2022;15(12) doi: 10.1161/CIRCHEARTFAILURE.122.009488. Dec. e009488. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Ilonze O.J., Parsly Read-Button L., Cogswell R., et al. Controversies and conundrums in cardiac Cachexia: key questions about wasting in patients with HFrEF. JACC Heart failure. 2024;12(10):1645–1660. doi: 10.1016/j.jchf.2024.03.003. [DOI] [PubMed] [Google Scholar]
- 7.Ajibawo T., Ajibawo-Aganbi U., Jean-Louis F., Patel R.S. Congestive heart failure hospitalizations and Cannabis use disorder (2010-2014): National Trends and outcomes. Cureus. 2020;12(7) doi: 10.7759/cureus.8958. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dubois C., Eurich D.T., Dyck J.R.B., Hyshka E., Hanlon J.G., Zongo A. Risk of heart failure-related events in patients exposed to medical Cannabis: a longitudinal cohort study. Am J Med Open. 2026;15 doi: 10.1016/j.ajmo.2025.100120. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Kalla A., Krishnamoorthy P.M., Gopalakrishnan A., Figueredo V.M. Cannabis use predicts risks of heart failure and cerebrovascular accidents: results from the National Inpatient Sample. J. Cardiovasc. Med. (Hagerstown) 2018;19(9):480–484. doi: 10.2459/JCM.0000000000000681. [DOI] [PubMed] [Google Scholar]
- 10.Ashrafian S., Ahrens E., Wachtendorf L.J., et al. Association of cannabis use with major cardiovascular and cerebrovascular events after surgery or interventional procedures. Am. J. Addict. 2025;34(5):517–527. doi: 10.1111/ajad.70029. [DOI] [PubMed] [Google Scholar]
- 11.Chuzi S., Cabrera E., Ilonze O., et al. Race, substance use, and evaluation for heart transplantation: insights from a large urban medical center. The Journal of heart and lung transplantation: the official publication of the International Society for Heart Transplantation. 2024;43(9):1521–1524. doi: 10.1016/j.healun.2024.05.011. [DOI] [PubMed] [Google Scholar]
- 12.Jeffers A.M., Glantz S., Byers A.L., Keyhani S. Association of cannabis use with cardiovascular outcomes among US adults. J. Am. Heart Assoc. 2024;13(5) doi: 10.1161/JAHA.123.030178. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Carrell D.S., Cronkite D.J., Shea M., et al. Clinical documentation of patient-reported medical cannabis use in primary care: toward scalable extraction using natural language processing methods. Subst. Abus. 2022;43(1):917–924. doi: 10.1080/08897077.2021.1986767. [DOI] [PMC free article] [PubMed] [Google Scholar]