Recently, collateral sensitivity networks have created excitement for designing new therapies that could reduce rates of evolution of antibiotic resistance in pathogens. To explore this idea, I tested the key assumption that collateral sensitivity should reduce the frequency of mutation to resistance. Ceftazidime resistant strains showed susceptibility to chloramphenicol. It was therefore expected that growth in sub-MIC concentrations of chloramphenicol would reduce mutation rate to ceftazidime resistance, however, an increase in mutation rate was observed. There are two likely hypotheses, the first is that the increase in mutation is due to antibiotic related SOS response, and the second is that multi-drug-resistance was selected for in both cases. This highlights the need for evolutionary considerations in designing new drug therapies, as apparent patterns of collateral sensitivity may not be a sufficient criterion.