Predicting the long-term toxicity of five-antibiotic mixtures to Vibrio qinghaiensis sp. Q67.

Concentration addition (CA) is commonly used as a standard additive reference model to predict the short-term toxicity for most chemical mixtures. Whether CA can predict the long-term toxicity of antibiotic mixtures was investigated. The long-term toxicity of five antibiotics including apramycin sulfate, paromomycin sulfate, tetracycline hydrochloride, chloramphenicol and streptomycin sulfate and their mixtures to a photo bacterium Q67 were detected by the long-term toxicity microplate analysis procedure. Seven five-antibiotic mixtures with various concentration ratios and concentration levels were designed by employing uniform design ray method. The long-term mixture toxicity was predicted by CA based on the toxicity data of single antibiotics. The results showed that Weibull or Logit function fit well with the long-term toxicity data of all the components and their mixtures (R>0.98 and RMSE<0.07). According the toxicity index, the negative logarithm of mean effect concentration, the long-term toxicity of the five antibiotics differs greatly and is higher than their short-term toxicity. The predicted values by CA model conformed to the experimental values of mixtures, which implies CA can predict reliable results for the long-term toxicity of antibiotic mixtures.