Seventeen clinical trials (91 patients) and 41 case reports (79 patients) were included.
Methodological flaws in the primary studies included a failure to adequately define the botanical agents used; and case reports that did not clearly distinguish between adverse events due to toxicity and those due to herb-drug interactions.
Garlic (1 before-and-after study and 3 case reports). Concomitant use of garlic and paracetamol may be associated with changes in pharmacokinetic variables of paracetamol after 1 to 3 months' treatment (1 before-and-after study with 16 patients). Concomitant use of garlic and warfarin was associated with an increase in the international normalised ratio (INR) in 2 case reports. One case of hypoglycaemia was reported in a diabetic taking garlic and chlorpropamide.
Ginkgo (1 RCT, 1 before-and-after study and 4 case reports).
Ginkgo interactions include spontaneous bleeding when combined with warfarin (2 case reports), raised blood-pressure when combined with a thiazide diuretic (1 case report), and coma when combined with trazodone (1 case report). Ginkgo did not modify the half-life of antipyrine (1 RCT before-and-after study with 25 patients), and did not modify hormonal plasma levels after stimulation tests with luteinising hormone-releasing hormone or thyrotropin-releasing hormone (1 before-and-after study with 7 patients)
St. John's wort (2 RCTs, 2 non-randomised controlled trials, 6 before-and-after comparisons, 4 case series and 26 case reports).
St. John's wort was reported to lower the blood concentration of cyclosporin (3 case series with 45 patients and 5 case reports), amitriptyline (1 non-blind study with 12 patients), digoxin (1 placebo controlled trial with 25 patients), indinvar (1 before-and-after study with 8 patients), warfarin (1 placebo-controlled crossover RCT with 10 healthy patients, and 7 case reports), phenprocoumon (1 case report), and theophylline (1 case report). It also caused intermenstrual bleeding when used concomitantly with oral contraceptives (1 case series with 8 patients and 4 case reports), and delirium or mild serotonin syndrome when used concomitantly with loperimide (1 case report) or selective serotonin re-uptake inhibitors (7 case reports).
St. John's wort in combination with alcohol was found to result in no change in cognitive function (1 placebo-controlled, double-blind crossover RCT with 32 patients), or in vigilance with either alcohol alone or in combination with valarian (1 non-blind study with 12 patients). Four out of five before-and-after studies found an increase or a trend to increase the metabolic capacity of cytochrome P450 enzymes.
Ginseng (1 placebo-controlled RCT with 227 patients, 1 before-and-after with 14 patients, and 3 case reports).
Ginseng was reported to lower the blood concentration of alcohol (before-and-after with 14 patients), and not change any safety parameters when used in combination with influenza vaccine (1 RCT with 227 patients). However, it was found to induce mania (1 case report) or insomnia and headaches (1 case report) when used with phenelzine.
Kava (1 placebo-controlled RCT and 2 case reports).
Kava was reported to not change safety parameters when taken with alcohol (1 RCT with 20 patients). However, it increased 'off' periods in Parkinson patients taking levadopa (1 case report), and caused a semi-comatosed state when used with alprazolam, cimetidine and terazosin (1 case report). No interactions were found for echinacea and saw palmetto.