Objectives This study was conducted to determine the proper hand position on the sternum for external chest compression to generate a maximal haemodynamic effect during cardiopulmonary resuscitation (CPR).

Methods 114 patients with cardiac arrest who underwent chest CT after successful resuscitation from January 2006 to August 2009 were included in the study. To evaluate the area of the cardiac chambers subjected to external chest compression, the area of each cardiac chamber under the sternum was measured using cross-sectional CT at three different locations: the internipple line on the sternum (point A), halfway between point A and the sternoxiphoid junction (point B) and at the sternoxiphoid junction (point C).

Results The widest total heart area, total ventricular area and left ventricular area (LVA) were observed most frequently at point C (58%, 85% and 78% of all cases, respectively). Few cases (six in total heart area, one in total ventricular area and one in LVA) were observed as the widest at point A. Predicted compressed areas of the right and left ventricle were wider at point C than at points A or B (right ventricular area: 366±536 mm² at point A, 961±653 mm² at point B and 1383±689 mm² at point C, p<0.001; LVA: 65±236 mm² at point A, 365±506 mm² at point B and 1099±817 mm² at point C, p<0.001).

Conclusions Only a small proportion of the ventricle is subjected to external chest compression when CPR is performed according to the current guidelines. Compression of the sternum at the sternoxiphoid junction might be more effective to compress the ventricles.