Article Text
Abstract
Introduction Hepatocyte growth factor (HGF) is elevated in patients with cancer and is a predictor for prognosis. We investigated whether slight elevation of serum HGF level was a marker for subclinical cancer and death in a general population.
Methods Apparently healthy 1492 subjects had a health examination in 1999. Subjects with a history of liver disease or malignancies were excluded by a questionnaire. Finally, we measured plasma HGF levels in 1470 subjects. They were followed-up periodically for 10 years. The follow-up rate was 99.3%. We calculated mortalities from cancer by multivariate proportional hazards model.
Results At follow-up, 169 subjects had died (61 from cancer, 32 from cerebro-cardiovascular disease and 76 from others). The mean HGF level at baseline was significantly (p<0.01) higher among subjects who died than those who survived (0.26±0.11 vs 0.23±0.09 ng/ml). In a Cox proportional hazard model, age, systolic blood pressure, HGF (HR 1.270; 95% CI 1.059 to 1.523; p=0.009), low albumin and smoking were independent predictors for death from all causes. Age, HGF (HR 1.309; 95% CI 1.042 to 1.654; p=0.02) and low cholesterol were independent predictors for cancer death.
Conclusion Slight elevation of HGF may be an early marker of subclinical cancer.