Through statistically designed experiments, lysis agents were optimized to effectively disrupt bacterial cells in a microfluidic device. Most surfactants caused the efficient lysis of Gram-positive microbes, but not of Gram-negative bacteria. A Plackett-Burman design was used to select the components that increase the efficiency of the lysis of the Gram-negative bacteria Escherichia coli. Using this experimental design, both lysozyme and benzalkonium chloride were shown to significantly increase the cell lysis efficiency, and ATP was extracted in proportion to the lysis efficiency. Benzalkonium chloride affected the cell membrane physically, while lysozyme destroyed the cell wall, and the amount of ATP extracted increased through the synergistic interaction of these two components. The two-factor response-surface design method was used to determine the optimum concentrations of lysozyme and benzalkonium chloride, which were found to be 202 and 99 ppm, respectively. The lysis effect was further verified by microscopic observations in the microchannels. These results indicate that Gram-negative cells can be lysed efficiently in a microfluidic device, thereby allowing the rapid detection of bacterial cells using a bioluminescence-based assay of the released ATP.