Caffeine and Experimental Pain Sensitivity in Humans—Systematic Review and Meta-Analysis

Introduction: Acute caffeine may modify laboratory-evoked pain, but its relevance to clinical analgesia is uncertain. We systematically reviewed controlled experiments of caffeine administered alone in healthy humans, distinguishing pain intensity, threshold, tolerance, and induced hypoalgesia. Methods: Six databases were searched through 1 September 2026 using PRISMA 2020 methods. Nine experiments (216 analyzed participants) were included; five (130 participants) supplied pain-intensity contrasts for an exploratory REML random-effects meta-analysis with Hartung–Knapp confidence intervals. Results: The pooled effect was uncertain (Hedges’ g = 0.61, 95% CI −0.33 to 1.54; p = 0.146; I2 = 89.0%). Exclusion of the influential graph-derived cold-immersion estimate reduced g to 0.16 (95% CI −0.22 to 0.53; I2 = 24.7%). Pain threshold, tolerance, and induced hypoalgesia were not combined with pain intensity. Protocol differences could be described but not tested as reliable moderators. Adverse-event ascertainment varied substantially. Conclusions: The available evidence does not establish a consistent generalized hypoalgesic effect of caffeine alone. It can inform standardization of experimental pain research, but cannot justify caffeine treatment for clinical pain.