Thirty clinical studies involving more than 10,000 patients conducted during the last 20 years have been analyzed to assess the value of caffeine as an analgesic adjuvant. Although most studies included patients with postpartum uterine cramping or episiotomy pain, some involved patients with pain from oral surgery or headache. In 21 of 26 studies, the relative potency estimates of an analgesic with caffeine to an analgesic without caffeine is greater than one. The pooled relative potency estimates in each of several major categories of combination analgesics are significantly greater than one. The overall pooled relative potency estimate is 1.41, with 95% confidence limits of 1.23 to 1.63; that is, to obtain the same amount of response from an analgesic without caffeine requires a dose that is approximately 40% greater than one with caffeine.