Is cannabis exposure associated with childhood leukemia rates across US states? | Research

The takeaway

A geospatial analysis found that US states with higher cannabis use had higher pediatric acute lymphoid leukemia rates, with cannabis-legal states showing elevated rates compared to non-legal states, though the ecological design limits causal conclusions.

Researchers studying potential carcinogenic effects of cannabis and anyone interested in the relationship between cannabis legalization and population health outcomes.

Legal states: higher ALL rates (p<0.001)

What the researchers found

Cannabis use was independently associated with pediatric ALL rates in multiple regression models adjusted for other substances, income, and ethnicity. A dose-response relationship was observed across cannabis use quintiles (Chi.Sq.=613.79, p=3.04x10^-70). Cannabis-legal states had higher ALL rates (2.395 vs 2.127/100,000, p=5.05x10^-10). Spatiotemporal models showed associations with specific cannabinoids including THC and cannabigerol.

Why it matters

Pediatric ALL rates in the US have risen 93.5% between 1975 and 2016 for unclear reasons. If cannabis exposure contributes to this rise, it would be a significant public health concern as cannabis use and legalization expand.

The numbers in context

ALL rates rose 93.5% from 1975-2016; cannabis association β=3.33 (bivariate); legal states: 2.395 vs 2.127/100,000 (p=5.05x10^-10); quintile effect Chi.Sq.=613.79; 33/35 e-Values supportive of causality

How the study worked

Ecological geospatial analysis combining state-level ALL rates from CDC/NCI SEER database with drug use data from the National Survey of Drug Use and Health (74.1% response rate), cannabinoid concentration data from the DEA, and census data. Analyzed using robust regression, spatiotemporal models, inverse probability weighting, and causal inference methods including e-Values.

What this study cannot tell us

Ecological design (state-level correlations cannot prove individual-level causation). Many potential confounders at the state level. Cannabis use data is self-reported. The authors make strong causal claims that exceed what ecological data can support. Multiple comparisons increase risk of spurious findings.

How to read the evidence

Ecological geospatial analysis with sophisticated statistical methods but fundamental design limitations. State-level correlations cannot establish individual-level causation.

When this study was published

Published in 2021; individual-level studies are needed to test these ecological associations.

The bigger picture

This study makes a provocative causal claim based on ecological data. While the statistical associations are striking, ecological studies correlating state-level exposures with state-level disease rates cannot establish that individual cannabis users' children develop leukemia. The findings warrant further investigation with individual-level data.

Questions still open

Do individual-level studies of prenatal cannabis exposure show elevated leukemia risk? Could the ecological association reflect other factors that differ between cannabis-legal and non-legal states? What biological mechanism could link cannabinoid exposure to leukemogenesis?

Common questions

Does cannabis cause childhood leukemia?

This ecological study found higher childhood leukemia rates in states with more cannabis use and legal cannabis. However, the study compares state-level averages and cannot prove that individual cannabis exposure causes leukemia. Individual-level studies are needed.

How strong is the evidence?

The statistical associations are striking, but the ecological study design is a fundamental limitation. States differ in many ways beyond cannabis policy, and these unmeasured factors could explain the associations.

Read the original research

Cannabinoid exposure as a major driver of pediatric acute lymphoid Leukaemia rates across the USA: combined geospatial, multiple imputation and causal inference study.

BMC cancer, 21(1), 984

View on PubMed ↗Publisher’s paper ↗

Citation

Reece, Albert Stuart; Hulse, Gary Kenneth. (2021). Cannabinoid exposure as a major driver of pediatric acute lymphoid Leukaemia rates across the USA: combined geospatial, multiple imputation and causal inference study.. BMC cancer, 21(1), 984. https://doi.org/10.1186/s12885-021-08598-7