To obtain normal values for maternal stroke volume and cardiac output during pregnancy, a non-invasive, accurate and reproducible method is required. The thoracic electrical bioimpedance (TEB) method may be suitable. However, this method is as yet only qualified for short-term trend recordings, since it assumes that body dimensions such as height, weight and thoracic circumference remain constant during the study. This may not be the case in long-term studies, especially during pregnancy. In this paper it is argued that changes in stroke volume (SV) during pregnancy are reflected most strongly when using the formula: SV = PVET (dZdt) m a x /Z 0 , where P is a personal factor to be determined at the beginning of pregnancy; VET the ventricular ejection time; (dZdt) m a x the maximum of the first derivative of the thoracic impedance during the cardiac cycle and Z 0 the time average of this impedance during the cardiac cycle. Indexed parameters should not be used as this reduces sensitivity. Commercial equipment, based upon other algorithms, can be used by feeding the right parameters for each series of measurements. This enables calculation, trends in stroke volume and cardiac output for longitudinal studies for instance during pregnancy.