[Genesis of methicillin-resistant Staphylococcus aureus and selective antibacterial injection pressure]

We reported previously that cefazolin was related to an increase in methicillin-resistant Staphylococcus aureus (MRSA) between 1998 and 2000 at Yamagata University Hospital. The incidence of MRSA decreased, however, between 2001 and 2003. We examined the relationship between the use of antibacterial injection and MRSA incidence. MRSA increased in surgical and medical wards between 1998 and 2000, but decreased mainly in surgical wards between 2001 and 2003. We statistically analyed the number of inpatients detected with MRSA (MRSA patients) and the use of antibiotics per month. MRSA patients positively correlated with cephems and carbapenems, the highest positive correlation with cefazolin at a correlation coefficient (R) of 0.45 (p<0.001). In multiple linear regression analysis, cefazolin and carbapenems were chosen as independent variables of a regression equation predicting MRSA patients, during the MRSA decrease and in surgical wards. We thought, it was because carbapenems increased with cefazolin that only cefazolin was chosen as an independent variable during the MRSA increase. We found that the antibiotics as independent variables were associated with about 50% of MRSA by the multiple regression model contribution ratio. Cefazolin was used most for presurgical prophylaxis in surgical wards, and about 20% of surgical wards in medical wards. Carbapenems were use almost equally in surgical and medical wards, but the correlation with MRSA in medical wards was 0.21 (p<0.1) and in surgical wards 0.38 (p<0.005), showing a significant correlation with carbapenems in surgical wards. In conclusion, cefazolin and carbapenems were related to the incidence of MRSA, and carbapenems showed a significant correlation in the presence of cefazolin. This strongly suggests that MRSA is significantly generated when inpatients are given carbapenems after administration of cefazolin.