OBJECTIVE--To measure the noise produced and related subjective complaints after implantation of four different mechanical heart valve prostheses and to identify further factors related to the patient and prosthesis that influence noise generation and complaints. DESIGN--Sound pressure was measured 5 and 10 cm and 1 m from the point of maximal impulse on the body surface by a calibrated meter in quiet rooms with either a decibel(A) filter or octave filters. The patients were asked about their complaints and examined physically. SETTING--The measurements were conducted in silent rooms of ear, nose, and throat departments. The patients had been operated on either in a university hospital or a community hospital. MAIN OUTCOME MEASURES--Sound pressures of frequency bands and sound pressures measured in dB(A) at various distances. Complaints registerd were: sleep disturbance, disturbance during daytime, "wants a less noisy prosthesis," and "can hear the closing click". PATIENTS--143 patients after heart valve replacement with St Jude Medical (n = 35), Duromedics Edwards (n = 38), Carbomedics (n = 34) and Björk-Shiley Monostrut (n = 36) prostheses operated on between 1984 and 1988 were matched for valve position, ring size, and body surface area. RESULTS--Duromedics Edwards (33.5 (6) dB(A)) and Björk-Shiley Monostrut valves (31 (4) dB(A)) were significantly louder than St Jude Medical (24 (4) dB(A)) and Carbomedics (25 (6) dB(A)) prostheses (p = 0.0001) (mean (SD)). The louder valves were significantly more often heard by the patients (p = 0.0012) and caused more complaints both during sleep (p = 0.024) and during the daytime (p = 0.07). Patients with these valves were more likely to want a less noisy valve (p = 0.0047). Patients with symptoms were younger, had better hearing, and were more likely to be in sinus rhythm. As well as the type of prostheses, the valve diameter and body height also had an effect on sound emission. CONCLUSIONS--The intensity of the closing click of mechanical valve prostheses was significantly different for various designs. Patient complaints were related to the objectively measured sound pressure. Noise production should be considered when a mechanical valve is selected.