Caffeine has been an additive in analgesics for many years. However, the analgesic adjuvant effects of caffeine have not been seriously investigated since a pooled analysis conducted in 1984 showed that caffeine reduces the amount of paracetamol (acetaminophen) necessary for the same effect by approximately 40%. In vitro and in vivo pharmacological research has provided some evidence that caffeine can have anti-nociceptive actions through blockade of adenosine receptors, inhibition of cyclo-oxygenase-2 enzyme synthesis, or by changes in emotion state. Nevertheless, these actions are only considered in some cases. It is suggested that the actual doses of analgesics and caffeine used can influence the analgesic adjuvant effects of caffeine, and doses that are either too low or too high lead to no analgesic enhancement. Clinical trials suggest that caffeine in doses of more than 65 mg may be useful for enhancement of analgesia. However, except for in headache pain, the benefits are equivocal. While adding caffeine to analgesics increases the number of patients who become free from headache [rate ratio = 1.36, 95% confidence interval (CI) 1.17 to 1.58], it also leads to more patients with nervousness and dizziness (relative risk = 1.60, 95% CI 1.26 to 2.03). It is suggested that long-term use or overuse of analgesic medications is associated with rebound headache. However, there is no robust evidence that headache after use or withdrawal of caffeine-containing analgesics is more frequent than after other analgesics. Case-control studies have shown that caffeine-containing analgesics are associated with analgesic nephropathy (odds ratio = 4.9, 95% CI 2.3 to 10.3). However, no specific contribution of caffeine to analgesic nephropathy can be identified from these studies. Whether caffeine produces nephrotoxicity on its own, or increases nephrotoxicity due to analgesics, is yet to be established.