BACKGROUND: As more than 80% of Korean hospital laboratories don't use glycerol blank for the triglyceride measurement, we investigated free glycerol interference on the measurement of triglyceride concentration. METHODS: We collected 237 specimens which had more than 300mg/dL of triglyceride measured by Abbott (without glycerol blank, USA) reagent from the 217 patients visiting Catholic University Hospital of Taegu Hyosung from September, 1995, to May, 1996. We retested them with Youngdong (without glycerol blank, Korea) reagent, and IRC reagent (with glycerol blank, Japan), and also measured free glycerol concentration with Youngdong reagent. Then we examined the clinical records of the patient showing increased glycerol concentration. RESULTS: Average triglyceride concentration measured by Abbott reagent was 448.0+/- 165.9 mg/dL and average glycerol interference to triglyceride concentration was 3.4+/- 6.7%, There were 8 patients (3.4%) who revealed more than 10 % of glycerol interference. Except these patients the average glycerol interference was 2.4+/- 1.4%. Among these 8 patients, 3 patients were treated with Frucenil and other 3 were with Intralipose(R), and the other 2 were diagnosed as diabetes and preeclampsia respectively. In vitro test, Frucenil mixed with serum in 1% (v/v) increased triglyceride concentration as high as 640.6% with Abbott reagent, but it did only 48.4% with IRC reagent. CONCLUSIONS: Triglyceride measurement by non glycerol blank method exceeded the recommended limit of National Cholesterol Education Program's precision, and this method extremely overestimated the triglyceride concentration for the patient with the glycerol containing fluid therapy. So all laboratories would be encouraged to use the glycerol blank method for triglyceride measurements.