Kabat · Cancer causes & control : CCC 2021 · Meta-analysis / sensitivity analysis · n=6 studies

On recent meta-analyses of exposure to glyphosate and risk of non-Hodgkin's lymphoma in humans.

Level 4 - case-series / case-control

Meta-analysis and sensitivity analysis of observational studies (5 case-control, 1 cohort)

PubMed 33447891 · doi:10.1007/s10552-020-01387-w · record verified 2026-08-26

What was done

The authors conducted sensitivity analyses using data from 6 observational studies (5 case-control studies and 1 cohort study) included in a prior meta-analysis on glyphosate exposure and non-Hodgkin lymphoma (NHL) risk. They assessed how summary estimates change based on exposure definitions (highest reported exposure level vs. ever-exposure) and latency assumptions (lag periods). They also calculated a summary risk estimate for ever-exposure using updated case-control study data.

What was found

Summary risk estimates varied substantially depending on latency and exposure criteria: - Highest reported exposure with a 20-year cohort lag: RR = 1.41 (95% CI: 1.13 to 1.76). - Highest reported exposure with a 15-year cohort lag: RR = 1.25 (95% CI: 1.01 to 1.25). - Ever-exposure with no lag period: RR = 1.05 (95% CI: 0.87 to 1.28).

Why it matters

This study shows that reported associations between glyphosate and NHL in meta-analyses are sensitive to analytical choices regarding exposure thresholds and assumed latency periods, yielding different conclusions depending on the model chosen.

Limits

The underlying evidence base is small (6 observational studies, predominantly case-control designs susceptible to recall and selection bias). The abstract provides no information on individual study quality, residual confounding, or direct exposure quantification.

Cited by