Maternal heart rate variability was evaluated by means of 24-hour Holter ECG monitoring in three groups of patients (15 preeclamptic, 15 women hospitalized due to other complications and 15 normal ambulatory pregnant women) in the 28th-33rd week of gestation. Heart rate variability was estimated by time and frequency domain measures. Patients with preeclampsia had significantly longer NN intervals during daytime compared to the other groups studied. With regard to frequency domain measures such as very-low-frequency, low-frequency and high-frequency power, the groups did not differ. Nor did they differ with regard to day and nighttime. We conclude that during pregnancy, the power of the maternal heart rate spectrum is markedly depressed, which affects the power of the method to separate normal pregnancies from pregnancies affected by preeclampsia.