Cell membrane damage induced by pulsed or continuous ultrasound (US) activation of attached contrast agent gas bodies was examined in an in vitro model system. Monolayers of mouse macrophage-like cells were cultured on the inside of one window of an exposure chamber. The monolayers were incubated with Optison (Amersham Health Inc., Princeton, NJ) or Definity (Bristol-Myers Squib Medical Imaging, North Billerica, MA) and then rinsed to remove unattached gas bodies. A 3.5-MHz focused transducer was aimed at the chamber 3.7 cm away in a 37 degrees C water bath. The cells were scored for Trypan blue dye exclusion, with stained nuclei indicative of cell membrane damage. Exposure-response functions were approximated with exposure levels spaced 3-dB apart. Thresholds were located between the lowest exposure with statistically significant counts of blue-stained cells relative to sham exposures, and the next lower level. Thresholds with Optison included 0.05 MPa for 60-s continuous exposure duration, and 0.21 MPa for 0.6-micros pulses with 60-micros repetition period for 60-s pulsed exposure duration. Results were similar for Definity. Thresholds changed slowly with changes in timing parameters; for example, the threshold for a 0.6-micros continuous exposure (i.e., one pulse) was 0.84 MPa. Compared to 60-s exposure, this represents a factor of 16.8 increase in threshold for a factor of 10(8) decrease in exposure duration. The thresholds are less than the pressure amplitudes needed for nucleation of inertial cavitation, which suggests classification of the phenomenon as a form of gas body activation.