The risk of new-onset arrhythmia during pregnancy is high, presumably relating to changes in both haemodynamic and cardiac autonomic function. The ability to non-invasively assess an individual's risk of developing arrhythmia during pregnancy would therefore be clinically significant. We aimed to quantify electrocardiographic temporal characteristics during the first trimester of pregnancy and to compare these with non-pregnant controls. Ninety-nine pregnant women and sixty-three non-pregnant women underwent non-invasive cardiovascular and haemodynamic assessment during a protocol consisting of various physiological states (postural manoeurvres, light exercise and metronomic breathing). Variables measured included stroke volume, cardiac output, heart rate, heart rate variability, QT and QT variability and QTVI (a measure of the variability of QT relative to that of RR). Heart rate (p < 0.0005, p < 0.0005, p < 0.0005) and cardiac output (p = 0.043, p < 0.0005, p < 0.0005) were greater in pregnant women in all physiological states (respectively for the supine position, light exercise and metronomic breathing state), whilst stroke volume was lower in pregnancy only during the supine position (p < 0.0005). QTe (Q wave onset to T wave end) and QTa (T wave apex) were significantly shortened (p < 0.05) and QTeVI and QTaVI were increased in pregnancy in all physiological states (p < 0.0005). QT variability (p < 0.002) was greater in pregnant women during the supine position, whilst heart rate variability was reduced in pregnancy in all states (p < 0.0005). Early pregnancy is associated with substantial changes in heart rate variability, reflecting a reduction in parasympathetic tone and an increase in sympathetic activity. QTVI shifted to a less favourable value, reflecting a greater than normal amount of QT variability. QTVI appears to be a useful method for quantifying changes in QT variability relative to RR (or heart rate) variability, being sensitive not only to physiological state but also to gestational age. We support the use of non-invasive markers of cardiac electrical variability to evaluate the risk of arrhythmic events in pregnancy, and we recommend the use of multiple physiological states during the assessment protocol.