Percutaneous balloon mitral valvuloplasty has been shown to be an effective means of reducing mitral valve gradient and increasing mitral valve area in patients with mitral stenosis. Most techniques currently employed for performing this procedure involve delivery of one or two balloon valvuloplasty catheters through the interatrial septum en route to the mitral valve orifice. To determine the morphology of the resultant atrial septal defect (ASD), particularly as a function of the technique employed, we performed a series of in vitro experiments designed to simulate a variety of technical approaches. Ninety-eight experiments in total were performed in 19 normal adult hearts obtained in the fresh, nonpreserved state at necropsy. Transseptal delivery and withdrawal of two conventional, elliptical balloon catheters through two, individual septostomy sites was found to produce the largest ASD (combined area of two defects = 21.4 +/- 2.2 mm2). The defect resulting from transseptal delivery and tandem withdrawal of two elliptical balloon catheters through a single septostomy site measured 14.8 +/- 1.1 mm2, significantly (p = 0.0043) smaller than that produced by two septostomies. Transseptal delivery and withdrawal of a single, segmentally inflating (Inoue) balloon catheter produced a defect of intermediate size (17.5 +/- 1.2 mm2). ASD size was exacerbated by improper balloon withdrawal compared with tandem withdrawal of two completely deflated balloon catheters. Simultaneous withdrawal of two completely deflated balloon catheters through the same septostomy site increased ASD size from 14.8 +/- 1.1 mm2 to 23.6 +/- 2.3 mm2 (p = 0.0004). Simultaneous withdrawal of two incompletely deflated balloon catheters further increased ASD size to 45.8 +/- 2.6 mm2 (p less than 0.0001).(ABSTRACT TRUNCATED AT 250 WORDS)