To investigate the secular change in the incidence rate of drug-induced hepatitis (DIH) due to anti-tuberculosis chemotherapy including isoniazid (INH) and rifampicin (RFP), but not including pyrazinamide (PZA), we retrospectively studied the incidence rates of DIH in patients treated with chemotherapy including INH and RFP in four periods 1980-83, 87-88, 91-92, and 1998-2000. The criteria for selection of the patients were as follows.<BR>1. The serum aspartate aminotransferase (AST) and alanine aminotransferase (ALT) were measured before, and one month (20-40 days) and 2 months (45-75 days) after starting anti-tuberculosis chemotherapy. When the serum AST and ALT were measured twice or more during period 20-40 days or 45-75 days after starting anti-tuberculosis chemotherapy, the data obtained nearest to 30 or 60 days after were chosen as those of one or two months after starting chemotherapy, respectively.<BR>2. The serum AST and ALT were within normal range before starting anti-tuberculosis chemotherapy. The normal range of serum AST and ALT were ≤ 40 K-A and ≤ 35 K-A (in 1980-83) or ≤ 31 IU/l and 34 IU/l (in 1987-2000), respectively.<BR>3. Chronic active hepatitis and cirrhosis patients were excluded.<BR>4. All alive after completion of anti-tuberculosis chemotherapy.<BR>The numbers of the subjects who fulfilled the above criteria were 113, 135, 128 and 154 in 1980-83, 1987-88, 1991-92 and 1998-2000, respectively.<BR>DIH was defined serologically by serum AST ≥ 40 K-A and/or ALT ≥ 35 K-A (in 1980-83), or AST ≥ 40 IU/l and/or ALT ≥ 40 IU/l (1987-2000). The DIH incidence rate of the subjects classified by the year of treatment and age were examined, and the contributions of the risk factors for DIH, such as age, sex, alcoholics, previous liver disease history, HBs ag positivity, anti-HCV ab positivity, and hypoalbuminenia were studied, and none except the age over 80 y.o. was found to be a risk factor to DIH, in our subjects. In patients with the age over 80 y.o., daily doses of antituberculosis drugs RFP, INH and ethambutol (EB) were significantly higher in patients with DIH than those without DIH, but body weight and serum albumin level were not significantly different between these two groups. There was no risk factor to DIH in our patients less than 80 y.o. and this could be explained by the above-mentioned criteria of study patients selection.<BR>To exclude the age dependence of the incidence rate of DIH in our subjects, the incidence rates of DIH were calculated in patients less than 80 y.o. by the period of treatment, and they were 10/111 (9.0%), 23/131 (17.6%), 26/123 (21.1%) and 32/117 (27.4%) in 1980-83, 87-88, 91-92, and 1998-2000, respectively. The secular increase of the incidence rate of DIH was statistically significant (p=0.01). It is quite clear that this secular increase was not at all attributable to the above-mentioned risk factors. It is suspected that human liver has become more easily affected with INH and RFP in recent years. It is suggested that the new chemical compounds present in our increasingly complicated human milieu give heavier burdens on human liver, weaken the liver function, and enhance the capacity of INH and RFP to cause DIH.