The purpose of this study was to assess heart rate variability (HRV) in patients with coronary artery disease (CAD): reproducibility, circadian variability and the effects of stress (coronary angiogram). Sixty-one patients who had a coronary angiogram underwent 48-hour Holter monitoring during a period of high stress beginning 4-hours post-angiography (Day 1 and 2), and again two weeks later during a period of low stress (Day 3 and 4); both 24-hour time domain and power spectral measures were computed. To determine reproducibility, intraclass correlation coefficients were calculated for both time and frequency domain indices on Days 3 and 4. The intraclass correlation coefficient for the standard deviation of normal RR-intervals over 24-hours (SDNN) was 0.91, while the standard deviation of the mean of all 5-minute segments of normal RR-intervals for 24-hours (SDANN) was 0.85. The most reproducible time domain measure was pNN50 (defined as the percentage of differences between adjacent normal RR-intervals that are greater than 50 ms computed over 24-hours) with an intraclass correlation coefficient of 0.95. As for the frequency domain measures including low frequency (LF) area, high frequency (HF) area, low frequency to high frequency area ratio (LF:HF area), LF central frequency (cf), and total area (TA), intraclass correlations were found to be the best at 0300-hours and the worst at 1500-hours. Circadian pattern was determined on Day 4 of Holter monitoring. A main effect for time was found for heart rate (HR), LF area, HF area, and TA of the power spectra. Over a 24-hour period, HR, LF area, HF area and TA were the lowest at 0300-hours compared to all other times. There was also a main effect for myocardial infarction (MI) for the frequency domain indices LF area, HF area, LF:HF area ratio, and LFcf. Low frequency area and LF:HF area ratio were significantly elevated (both p<0.05), while both HF area (p<0.05) and LFcf (p<0.01) were reduced in patients with a prior Ml compared to those with no MI. As well, a significant interaction between time of day and Ml, and time of day and beta-blocker therapy was observed. Patients with coronary artery disease and a prior Ml demonstrated a reduced circadian pattern over 24-hours for HR. Likewise, the circadian pattern of HR for those on beta-blockers was also attenuated. The effects of stress was determined by comparing Day 1 to Day 4 of ambulatory Holter monitoring. The time domain measure SDNN was found to be significantly reduced during Day 1 (mean ± SEM; 111.67 ± 6.13 ms) compared to Day 4 (121.54 ± 6.94 ms; p<0.05). Patients with normal left ventricular function (LVF) showed a significant increase from Day 1 to 4 for both SDNN (p<0.01) and SDANN (p<0.05). In contrast, those with LV dysfunction had an attenuated response. Similarly, those CAD patients on beta-blockers demonstrated a significant increase for the time domain measure SDNN (p<0.05), unlike those not on beta-blockers. In the frequency domain, LF:HF area was significantly greater on Day 1 (1.74 ± 0.09) compared to Day 4 (1.64 ± 0.09; p<0.05). These findings suggest that HRV measures are reproducible, that a circadian pattern for HRV exists in patients with CAD, and that stress induced by an invasive procedure such as a coronary angiogram enhances sympathetic input to the SA node in the heart and thereby alters the sympathovagal balance, which is restored two weeks later. / Thesis / Master of Science (MSc)