A large, diverse cohort of young adults followed for 35 years showed no significant association between cumulative lifetime cannabis use and the development of high blood pressure. However, most participants used cannabis infrequently, so the findings apply mainly to light-to-moderate use and cannot confirm safety for heavy or daily users of today’s high-potency products.

As cannabis legalization accelerates across the United States and millions of adults integrate regular use into their lives, the question of whether long-term cannabis exposure raises blood pressure has become clinically urgent. Hypertension remains the single largest modifiable contributor to cardiovascular mortality globally, and clinicians need reliable longitudinal data to counsel patients. This study represents the longest and most analytically sophisticated attempt to answer that question, making its findings directly relevant to primary care conversations about cardiovascular risk in cannabis users.

Cannabis has well-documented acute cardiovascular effects, including transient increases in heart rate and short-term blood pressure changes mediated through endocannabinoid system activation and sympathetic nervous system modulation. Whether these acute effects translate into chronic hypertension risk with long-term cumulative exposure has remained an open and clinically important question. The CARDIA study, a well-established multicenter prospective cohort originally designed to study cardiovascular risk factor development in young adults, provided a unique platform to address this gap. Investigators constructed a novel cumulative exposure variable, cannabis-years, from self-reported use data collected at 10 in-person examinations over 35 years, and applied marginal structural models with inverse probability weighting to account for the complex interplay of time-varying confounders.

The primary result was a null association: the adjusted hazard ratio for incident hypertension per cannabis-year was 0.99 (95% CI, 0.97 to 1.00; P=0.18). This finding held across restricted cubic spline analyses, stratified analyses by sex, race, alcohol use, and tobacco smoking, and an alternate exposure measure (days of use in the past month). Importantly, 2,478 incident hypertension cases accrued over 88,292 person-years, providing substantial statistical power. However, the median cannabis-years in the cohort was only 0.2 at year 35, meaning the study primarily characterizes light-to-occasional users. The authors acknowledge that the study cannot address risk among heavy, daily, or high-potency users, and they call for future studies with biomarker-confirmed exposures, potency data, and enrollment of more diverse and heavily exposed populations.

35 Years of Cannabis Use and Blood Pressure: What the CARDIA Study Actually Shows

If you wanted to design the best possible observational study to ask whether a lifetime of cannabis use raises blood pressure, you would want 35 years of follow-up, repeated exposure measurements, and sophisticated causal methods. The CARDIA study delivers all three and finds no significant association. But the most important number in the paper may be the median cannabis exposure: 0.2 cannabis-years, the equivalent of about 73 days of use spread across 35 years. That context shapes everything. What this study genuinely contributes is a methodological advance. It is the first to combine a quantitative cumulative exposure metric with marginal structural modeling over decades of follow-up, and it does so in a well-characterized cardiovascular cohort with standardized blood pressure measurements. The convergence of every sensitivity analysis on the same null finding is not trivial. For the exposure range this cohort actually experienced, the evidence against a hypertension signal is now quite strong. I find that reassuring for the many patients I see whose cannabis use is occasional or modest, and who deserve to hear that the best available longitudinal data does not suggest their blood pressure is being quietly pushed upward by their use pattern.

The central problem, however, is one of dose and generalizability. Studying whether a glass of wine per week causes liver disease tells you very little about whether a bottle per day does. This study mostly observed the equivalent of occasional sippers. A median of 0.2 cannabis-years at year 35 means the population with heavy, sustained exposure was sparse, and the statistical power to detect effects in that tail was limited. Compounding this, one cannabis-year in 1986 is not pharmacologically equivalent to one cannabis-year in 2021. THC concentrations have risen dramatically, routes of administration have diversified, and product formulations bear little resemblance to what participants were using in the early decades of this cohort. Measuring cannabis-years without adjusting for potency is like measuring alcohol exposure in drink-years without distinguishing between light beer and grain alcohol. Marginal structural models are a genuinely important methodological tool for handling time-varying confounding, and the authors deserve credit for applying them here. But these models are like a sophisticated weather forecast: they account for more variables than a simple forecast and are genuinely better, yet they cannot predict what they haven’t been told to measure. Unmeasured confounders, including dietary patterns, stress exposure, and genetic susceptibility, remain unaddressed.

For my colleagues in primary care, I view this study as supporting measured reassurance for patients whose cannabis use is light to moderate. It does not, however, license blanket statements about cardiovascular safety for daily users of high-potency concentrates or edibles. When speaking with policymakers, I would emphasize that this evidence argues against population-level alarm about typical cannabis use and hypertension, but it cannot justify abandoning cardiovascular monitoring as legal markets evolve toward more intense products and patterns. For patients in my practice, I would say this: the best long-term study we have found no link between your level of use and high blood pressure, and that is genuinely good news, but if your use is heavier than what most people in this study reported, we should still keep an eye on your cardiovascular health. Null findings in well-designed longitudinal studies are important and should be communicated with appropriate confidence, but the boundaries of “well-designed” always include exposure range, population, and time period, and a null result at the 20th percentile of use does not answer the question at the 90th percentile.

This study represents the most advanced stage of observational longitudinal research on cannabis and hypertension currently available. It moves the field beyond cross-sectional snapshots and short-duration follow-ups that have produced conflicting signals, and it introduces a cumulative exposure metric that more faithfully captures the chronic nature of the clinical question. For clinicians assessing cardiovascular risk in cannabis users, the null finding at typical exposure levels shifts the conversation away from presumptive harm and toward nuanced, individualized assessment.

From a pharmacological standpoint, the acute hemodynamic effects of cannabis, including tachycardia and biphasic blood pressure responses, remain well documented and clinically relevant in patients with pre-existing cardiovascular conditions, regardless of this study’s long-term null finding. Drug interactions with antihypertensive medications, particularly those metabolized through CYP3A4 and CYP2C9 pathways, also warrant ongoing attention. Clinicians should not interpret this null result as eliminating the need for blood pressure monitoring in regular cannabis users, but rather as evidence that supports continued standard cardiovascular screening without adding cannabis-specific alarm in patients with light-to-moderate use patterns.

This is an original prospective cohort study using data from the CARDIA multicenter cardiovascular cohort, with repeated exposure assessments and marginal structural Cox regression. In the evidence hierarchy, prospective cohort studies rank below randomized controlled trials but above cross-sectional and case-control designs. The single most important inference constraint is that, despite the sophisticated analytic approach, this remains an observational study and cannot establish causality; the possibility of residual unmeasured confounding can never be fully excluded.

The null finding aligns with prior analyses from the same CARDIA cohort that found no association between cannabis-year quantiles and systolic or diastolic blood pressure, and with Corroon and colleagues’ earlier work in NHANES and MESA cohorts showing no link between duration of regular cannabis use and blood pressure levels. It stands in contrast to cross-sectional studies that have reported associations between recent cannabis use and elevated systolic blood pressure, though those studies are methodologically weaker and unable to distinguish acute from chronic effects. This study extends the literature by introducing a continuous cumulative exposure metric and applying time-varying confounding methods not previously used in this domain, strengthening confidence in the null finding within the observed exposure range.

The most consequential analytic choice was the use of cannabis-years as a continuous variable with a very low median. If the investigators had restricted the analysis to participants with cannabis-years above the median, or had categorized exposure into clinically meaningful tiers such as light, moderate, and heavy cumulative use, a dose-response relationship at higher exposure levels might have become detectable, or power limitations in those strata might have become more apparent. Additionally, applying a potency-adjusted cannabis-year metric, had such data been available, could have produced a different signal given the substantial increase in THC content over the study period. The authors’ decision to also examine restricted cubic splines was prudent and helps rule out nonlinear relationships within the observed range, but the sparseness of data at high cumulative exposure still limits what any modeling approach can reveal.

The most likely overinterpretation is to cite this study as proof that cannabis does not cause high blood pressure and is therefore cardiovascularly safe. This exceeds what the evidence supports on multiple levels. First, the cohort’s median cannabis exposure was extremely low, so the null finding primarily applies to light-to-occasional users. Second, the study addresses only incident hypertension and does not speak to other cardiovascular endpoints such as arrhythmia, acute coronary events, or vascular stiffness. Third, it cannot account for the pharmacological differences between historical cannabis products and today’s high-potency concentrates. A null finding within a specific exposure distribution is valuable evidence, but it is not a universal safety guarantee.

This study provides the most methodologically rigorous longitudinal evidence to date that cumulative cannabis use at typical population levels does not significantly predict incident hypertension over 35 years. It does not establish cardiovascular safety for heavy, daily, or high-potency cannabis users, nor does it address other cardiovascular outcomes. For clinical practice today, it supports measured reassurance for light-to-moderate users while reinforcing the need for individualized cardiovascular risk assessment as cannabis use patterns and product potency continue to evolve.

Does this study prove that cannabis is safe for my heart?

No. The study found no link between typical cannabis use and developing high blood pressure over 35 years, which is reassuring for light-to-moderate users. However, it only looked at hypertension and did not examine other heart-related risks. It also included very few heavy or daily users, so it cannot confirm safety at higher levels of use or with modern high-potency products.

Should I stop worrying about blood pressure if I use cannabis occasionally?

This study suggests that occasional cannabis use is unlikely to meaningfully increase your risk of developing hypertension over time. That said, routine blood pressure monitoring is recommended for all adults regardless of cannabis use, and you should discuss your individual risk factors with your physician. Cannabis still causes temporary heart rate increases that matter if you have existing heart conditions.

Does this study apply to people who use cannabis daily or use concentrates?

Not directly. The median lifetime cannabis exposure in this cohort was very low, equivalent to about 73 days of use spread across 35 years. Most participants were light or occasional users. Heavy daily users and those who rely on high-potency concentrates or edibles were underrepresented, so the null finding cannot reliably be applied to those use patterns.

Why does this study matter if it found nothing?

Null findings in well-designed studies are scientifically important because they help clinicians avoid over-warning patients about risks that may not exist at typical exposure levels. This study used the most rigorous methods available for this type of question and provides 35 years of data. Knowing where risk does not appear is just as clinically valuable as knowing where it does.

Have thoughts on this? Share it: