Cannabinoid efficacy for pain: dose, chronicity and route

Cannabis-based therapies are widely used for chronic pain, yet their mechanisms and therapeutic windows remain incompletely defined. This review synthesizes preclinical and clinical evidence on how dose, route of administration, treatment duration, and chemical composition shape analgesic efficacy and adverse-effect liability for Δ⁹-tetrahydrocannabinol (THC), cannabidiol (CBD), and select cannabis-derived terpenes. Across rodent pain models, acute THC reliably produces antinociception, but its therapeutic window is narrow because analgesic doses overlap with CB₁ receptor-mediated side effects such as sedation, hypothermia, hyperphagia, and motor impairment. Repeated THC exposure leads to tolerance and dependence. In contrast, CBD shows limited acute efficacy in naïve and inflammatory models but demonstrates more consistent benefit with repeated dosing in neuropathic and chemotherapy-induced pain, often without cannabimimetic adverse effects. Terpenes such as linalool, β-caryophyllene, myrcene, limonene, α-terpineol, and α-bisabolol exhibit independent antinociceptive and anti-inflammatory properties and are thought to pharmacologically interact with cannabinoids in a dose-, ratio-, and route-dependent "entourage" effect that either enhance or constrain therapeutic benefit. This review also focuses on integrating machine learning-based behavioral phenotyping of rodents to refine cannabinoid analgesia preclinical research. Computer vision pose-estimation and unsupervised clustering approaches enable high-resolution quantification of spontaneous and evoked natural behaviors, allowing the analytical dissociation of true analgesia from sedation, ataxia, or reduced exploration. By coupling these behavioral pipelines with pharmacokinetic and circuit-level analyses, emerging frameworks will define therapeutic windows with greater precision and improve the translational relevance of cannabinoid-based pain therapeutics.