OBJECTIVES: There is a worldwide increase of use of herbal products but the safety data is often lacking. The use of TU-HUO-CHI-SHENG-TANG (THCST), which contains trace amounts of aristolochic acid (AA), raises concern. This study aimed to actively monitor the safety of THCST used for treating knee osteoarthritis. METHODS: A multi-center, prospective, and observational study was designed to apply Good Clinical Practice to monitor and quantify any adverse event (AE) and adverse drug reaction (ADR) occurred to the patients who took THCST five grams a day for four weeks. Complete blood counts, liver and kidney function tests, and RBP (retinol binding protein) and NAG (N-acetyl-ß-D glucosaminidase) were also performed before and after use. Every reported/detected AE were independently assessed by site investigators and study nurses in terms of its severity and causal relationship with the study drug. Subsequently, a severe or suspected adverse drug reaction (ADR) would be submitted to the panel for final decision. RESULTS: A total of 71 eligible subjects were included and 291 AEs reported. There was no significant detectable change of liver and kidney function. All AEs were either mild or moderate. The most often reported AEs were rash, cough, abdominal fullness, somnolence, muscle cramps and diarrhea, with the incidence rates of 14.98, 12.91, 12.91, 11.88, 10.33 and 10.33 per 103 person-days, accordingly and of 7.98, 6.88, 6.88, 6.33, 5.51 and 5.51 per 103 person-sachets, accordingly. Four AEs were judged to be ADRs, which included one event each of skin discoloration, flashes, diarrhea and tachycardia. CONCLUSIONS: The safety profile of THCST for human use was established to some degree, which might provide THCST more evidence to become an acceptable alternative in the treatment of knee osteoarthritis.