Heart rate variability (HRV) is increasingly being used to assess the autonomic regulation of the heart. Altered HRV is a risk factor for adverse cardiovascular events, including sudden cardiac death [1]. One of our concerns was investigators using one five-minute ECG recording to measure baseline HRV and then comparing it with an intervention [2, 3]. We have previously shown a significant decrease in heart rate over three occasions [4]. This finding has gained importance in light of recent literature showing that heart rate can affect HRV; its influence can be removed through normalisation [5]. Our aim was to investigate whether there was any effect of repetitive testing on normalised HRV in healthy subjects. 65 healthy subjects volunteered to undergo three consecutive five-minute HRV measurements. Subjects were all staff and students of the Arabian Gulf University (AGU). The study was given ethical approval by the Research and Ethics committee, College of Medicine & Medical Sciences, AGU. All participating subjects gave informed and signed consent. Subjects were supine and each test was separated by a three minute time period. An ECG lead was attached to each limb. LabChart software and a PowerLab were used for data acquisition. The results were analysed by repeated measures ANOVA. P < 0.017 was considered as significant. The subjects’ mean (± S.D.) age was 25.9 ± 11.1 years. Over the three tests, mean heart rate showed a significant decrease for the 1st vs. 2nd (p=0.00004) and 1st vs. 3rd test (p=0.00003). There was no difference between the 2nd vs. 3rd test (72.88 ± 10.1 beats min-1 vs. 72.37 ± 10.0 beats min-1; p=0.35). With respect to non-normalised data, two parameters were significantly different. These were an increase in the low frequency power in normalised units over all occasions (LFnu; p=0.0017) and an increase in the non-linear parameter, standard deviation along the line of identity from the Poincare plot (SD2; p=0.009). After normalisation of the complete data and re-analysis, only one frequency domain, high frequency power in normalised units over all occasions (HFnu; p=0.0037) showed a decrease over all occasions. The most important finding from this study is that normalisation of our data radically alters the study results. The only possible reason for this is the fall in heart rate over these three occasions. Most likely, our subjects had a lower heart rate on the 2nd and 3rd occasion as they were apprehensive during the 1st test. Initially, before normalisation, SD2, a marker of long term variability, showed an increase along with LFnu, which reflects parasympathetic and sympathetic stimulation. However, after normalisation, only HFnu showed a fall, reflecting a decrease in parasympathetic stimulation. We suggest that HRV investigators may need to make sure they have a stable baseline before giving subjects any interventions.