OBJECTIVE: To evaluate the diagnostic capability of a hand-carried ultrasound device (HCU) in critically ill patients when using conventional transthoracic echocardiography (TTE) as a reference. DESIGN: Prospective, descriptive study. SETTING: Medical-surgical intensive care unit of a teaching hospital. PATIENTS: All patients requiring a TTE study were eligible. INTERVENTIONS: Each patient underwent an echocardiographic examination using a full-feature echocardiographic platform (Sonos 5500, Philips Medical Systems, Andover, MA) and a small battery-operated device (SonoHeart Elite, SonoSite, Bothell, WA). The operators (level III training in echocardiography) were randomized (HCU vs. TTE) and they independently interpreted the echocardiograms at the patient bedside. RESULTS: During a 2-month period, 55 consecutive patients (age: 61+/-16 years, simplified acute physiology score 46+/-15, body mass index 26+/-7) were studied, 40 of them being mechanically ventilated (73%). The number of acoustic windows was comparable using HCU and TTE (2.3+/-0.8 vs. 2.4+/-0.8: P=0.24). The overall diagnostic accuracy of HCU was lower compared with conventional TTE (137/171 vs. 158/171 clinical questions solved: P=0.002), reaching 80% and 92%, respectively. Despite its spectral Doppler capability, HCU missed diagnoses that were adequately identified by TTE: elevated left ventricular pressure ( n=2), relevant valvulopathy ( n=2) and moderate ( n=4) or severe ( n=2) pulmonary hypertension. Acute management was altered by HCU and TTE findings in 27 patients (49%) and 28 patients (51%), respectively. CONCLUSIONS: In this study, HCU had a lower diagnostic accuracy compared with conventional TTE, despite its spectral Doppler capability. Further studies are needed to validate these evolving diagnostic tools in critical care settings.