Objective: Un®ltered coffee raises serum LDL cholesterol in humans, owing to the presence of the diterpenes cafestol and kahweol. Norwegians with a chronic high intake of un®ltered coffee also had elevated serum levels of lipoprotein(a), an LDL-like particle which is insensitive toward dietary interventions. We now experimentally studied the in¯uence of coffee diterpenes on lipoprotein(a) levels. Design: Four randomised controlled trials. Subjects: Healthy, normolipidemic volunteers. Interventions: Coffee, coffee oil, and pure diterpenes for 4±24 weeks. Main outcome measures: The circulating level of lipoprotein(a). Results: In 22 subjects drinking ®ve to six strong cups of cafetiere coffee per day, the median fall in lipoprotein(a) was 1.5 mg/dL after two months (P 0.03), and 0.5 mg/dL after half a year (P b 0.05), relative to 24 ®lter coffee drinkers. Coffee oil doses equivalent to 10±20 cups of un®ltered coffee reduced lipoprotein(a) levels by up to 5.5 mg/dL (P`0.05) in two separate trials (n 12±16 per group). A puri®ed mixture of cafestol and kahweol, as well as cafestol alone, were also effective in reducing Lp(a) levels (n 10). Averaged over the four trials, each 10 mg/d of cafestol (plus kahweol)Ðthe amount present in two to three cups of cafetiere coffeeÐdecreased Lp(a) levels by 0.5 mg/dL or 4% from baseline values after four weeks (n 63). Conclusions: Coffee diterpenes are among the few dietary exceptions shown to in¯uence serum lipoprotein(a) levels. However, the Lp(a)-reducing potency of coffee diterpenes may subside in the long run, and their adverse side effects preclude their use as lipoprotein(a)-reducing agents.