The high prevalence of liver diseases such as chronic hepatitis and cirrhosis underscores the need for efficient and cost-effective treatments. The potential benefit of silymarin (extracted from the seeds of Silybum marianum or milk thistle) in the treatment of liver diseases remains a controversial issue. Therefore, the objective of this review is to assess the clinical efficacy and safety of silymarin by application of systematic approach. 525 references were found in the databases, of which 84 papers were retained for closer examination and 36 were deemed suitable for detailed analysis. Silymarin has metabolic and cell-regulating effects at concentrations found in clinical conditions, namely carrier-mediated regulation of cell membrane permeability, inhibition of the 5-lipoxygenase pathway, scavenging of reactive oxygen species (ROS) of the R-OH type and action on DNA-expression, for example, via suppression of nuclear factor (NF)-kappaB. Pooled data from case record studies involving 452 patients with Amanita phalloides poisoning show a highly significant difference in mortality in favour of silibinin [the main isomer contained in silymarin] (mortality 9.8% vs 18.3% with standard treatment; p < 0.01). The available trials in patients with toxic (e.g. solvents) or iatrogenic (e.g. antispychotic or tacrine) liver diseases, which are mostly outdated and underpowered, do not enable any valid conclusions to be drawn on the value of silymarin. The exception is an improved clinical tolerance of tacrine. In spite of some positive results in patients with acute viral hepatitis, no formally valid conclusion can be drawn regarding the value of silymarin in the treatment of these infections. Although there were no clinical end-points in the four trials considered in patients with alcoholic liver disease, histological findings were reported as improved in two out of two trials, improvement of prothrombin time was significant (two trials pooled) and liver transaminase levels were consistently lower in the silymarin-treated groups. Therefore, silymarin may be of use as an adjuvant in the therapy of alcoholic liver disease. Analysis was performed on five trials with a total of 602 patients with liver cirrhosis. The evidence shows that, compared with placebo, silymarin produces a nonsignificant reduction of total mortality by -4.2% [odds ratio (OR) 0.75 (0.5 - 1.1)]; but that, on the other hand, the use of silymarin leads to a significant reduction in liver-related mortality of-7% [OR: 0.54 (0.3 - 0.9); p < 0.01]. An individual trial reported a reduction in the number of patients with encephalopathy of -8.7% (p = 0.06). In one study of patients with cirrhosis-related diabetes mellitus, the insulin requirement was reduced by -25% (p < 0.01). We conclude that available evidence suggests that silymarin may play a role in the therapy of (alcoholic) liver cirrhosis. Silymarin is has a good safety record and only rare case reports of gastrointestinal disturbances and allergic skin rashes have been published. This review does not aim to replace future prospective trials aiming to provide the 'final' evidence of the efficacy of silymarin.