Relative risk reduction is not enough: why cardiorenal trials should routinely report absolute risk reduction and number needed to treat
Abstract
Cardiorenal outcome trials have transformed the management of cardiovascular and kidney disease, particularly through glucagon-like peptide-1 receptor agonists and sodium-glucose cotransporter-2 inhibitors. Treatment effects are commonly communicated using relative measures, including hazard ratios and relative risk reductions. Although essential for estimating treatment effects, these measures do not directly communicate the absolute magnitude of benefit experienced by patients. Absolute risk reduction (ARR) and number needed to treat (NNT) provide complementary information by incorporating underlying event risk and duration of follow-up. The distinction is clinically important. Similar relative treatment effects can produce substantially different absolute benefits depending on baseline risk, outcome definitions, follow-up duration, and study population characteristics. Conversely, a NNT from one trial should not be used to rank therapies across studies because differences in populations and endpoints strongly influence absolute event rates. For time-to-event outcomes, ARR should be estimated from cumulative event probabilities at a specified time point, preferably using appropriate survival methods. Hazard ratios reflect a lower estimated hazard over follow-up and should not be interpreted as equivalent to reductions in cumulative risk, and differences in annualized incidence rates should not be directly converted into conventional NNT estimates. We argue that relative and absolute measures should be presented together. Routine reporting of relative treatment effects, baseline event risks, ARR, and NNT, with a specified time horizon and method of estimation, would improve transparency and clinically meaningful interpretation. This approach would help clinicians, patients, guideline committees, and policymakers understand not only whether a therapy works, but also the magnitude of benefit that may reasonably be expected.
Related
- Characterizing resonant solitary states as relative equilibria via normal form reduction
- Haar-Bayesian Pure-State Prediction under Relative-Entropy Loss: Arbitrary-Effect Reduction and Global Optimality
- Absolute and relative invariants for low-dimensional topological insulators
- Scalar curvature and an infinite-dimensional hyperkähler reduction
- Dimension reduction for characterizing sexual dimorphism in biomechanics of the temporomandibular joint
BibTeX
@article{zaitoon2026relative,
title = {{Relative risk reduction is not enough: why cardiorenal trials should routinely report absolute risk reduction and number needed to treat}},
author = {Hussein Zaitoon and Muhammad Abdul-Ghani and Ralph A. DeFronzo},
journal = {eClinicalMedicine},
year = {2026},
month = oct,
doi = {10.1016/j.eclinm.2026.104258},
url = {https://linkinghub.elsevier.com/retrieve/pii/S2589537026005110},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Relative risk reduction is not enough: why cardiorenal trials should routinely report absolute risk reduction and number needed to treat
AU - Hussein Zaitoon
AU - Muhammad Abdul-Ghani
AU - Ralph A. DeFronzo
PY - 2026
DA - 2026/10/12
JO - eClinicalMedicine
DO - 10.1016/j.eclinm.2026.104258
UR - https://linkinghub.elsevier.com/retrieve/pii/S2589537026005110
AB - Cardiorenal outcome trials have transformed the management of cardiovascular and kidney disease, particularly through glucagon-like peptide-1 receptor agonists and sodium-glucose cotransporter-2 inhibitors. Treatment effects are commonly communicated using relative measures, including hazard ratios and relative risk reductions. Although essential for estimating treatment effects, these measures do not directly communicate the absolute magnitude of benefit experienced by patients. Absolute risk reduction (ARR) and number needed to treat (NNT) provide complementary information by incorporating underlying event risk and duration of follow-up. The distinction is clinically important. Similar relative treatment effects can produce substantially different absolute benefits depending on baseline risk, outcome definitions, follow-up duration, and study population characteristics. Conversely, a NNT from one trial should not be used to rank therapies across studies because differences in populations and endpoints strongly influence absolute event rates. For time-to-event outcomes, ARR should be estimated from cumulative event probabilities at a specified time point, preferably using appropriate survival methods. Hazard ratios reflect a lower estimated hazard over follow-up and should not be interpreted as equivalent to reductions in cumulative risk, and differences in annualized incidence rates should not be directly converted into conventional NNT estimates. We argue that relative and absolute measures should be presented together. Routine reporting of relative treatment effects, baseline event risks, ARR, and NNT, with a specified time horizon and method of estimation, would improve transparency and clinically meaningful interpretation. This approach would help clinicians, patients, guideline committees, and policymakers understand not only whether a therapy works, but also the magnitude of benefit that may reasonably be expected.
SN - 2589-5370
ER -