Associations between PM2.5 and heart rate variability are modified by particle composition and beta-blocker use in patients with coronary heart disease.

BACKGROUND: It has been hypothesized that ambient particulate air pollution is able to modify the autonomic nervous control of the heart, measured as heart rate variability (HRV). Previously we reported heterogeneous associations between particulate matter with aerodynamic diameter < 2.5 microm (PM2.5) and HRV across three study centers. OBJECTIVES: We evaluated whether exposure misclassification, effect modification by medication, or differences in particle composition could explain the inconsistencies. METHODS: Subjects with coronary heart disease visited clinics biweekly in Amsterdam, the Netherlands; Erfurt, Germany; and Helsinki, Finland for 6-8 months. The standard deviation (SD) of NN intervals on an electrocardiogram (ECG; SDNN) and high frequency (HF) power of HRV was measured with ambulatory ECG during paced breathing. Outdoor levels of PM2.5 were measured at a central site. In Amsterdam and Helsinki, indoor and personal PM2.5 were measured during the 24 hr preceding the clinic visit. PM2.5 was apportioned between sources using principal component analyses. We analyzed associations of indoor/personal PM2.5, elements of PM2.5, and source-specific PM2.5 with HRV using linear regression. RESULTS: Indoor and personal PM2.5 were not associated with HRV. Increased outdoor PM2.5 was associated with decreased SDNN and HF at lags of 2 and 3 days only among persons not using beta-blocker medication. Traffic-related PM2.5 was associated with decreased SDNN, and long-range transported PM2.5 with decreased SDNN and HF, most strongly among persons not using beta blockers. Indicators for PM2.5 from traffic and long-range transport were also associated with decreased HRV. CONCLUSIONS: Our results suggest that differences in the composition of particles, beta-blocker use, and obesity of study subjects may explain some inconsistencies among previous studies on HRV.

Preoperative alterations in correlation properties and complexity of R-R interval dynamics predict the risk of atrial fibrillation after coronary artery bypass grafting in patients with preserved left ventricular function.

INTRODUCTION: We evaluated whether there are constant preoperative alterations in nonlinear R-R interval dynamics that associate with the risk of postoperative atrial fibrillation in patients with preserved left ventricular function. METHODS: We analyzed mean normal-to-normal R-R intervals, short-term scaling exponent of detrended fluctuation analysis (DFA alpha(1)), approximate entropy and entropy of symbolic dynamics (SymDyn En) from 10-minute ECG recordings during rest, paced breathing, and passive tilt performed 1 day before surgery in 67 elective coronary artery bypass grafting patients. RESULTS: Nineteen patients developed postoperative atrial fibrillation. The preoperative DFA alpha(1) was constantly lower in patients developing postoperative atrial fibrillation than in patients remaining in sinus rhythm (P = 0.016); during spontaneous breathing, the DFA alpha(1) was 0.93 +/- 0.33 in patients with atrial fibrillation and 1.13 +/- 0.24 in patients with sinus rhythm. The entropy of symbolic dynamics was higher during the spontaneous breathing in patients with atrial fibrillation than in patients with sinus rhythm (4.72 +/- 0.51 vs 4.36 +/- 0.51, P = 0.012). Higher short-term scaling exponent of detrended fluctuation analysis during the spontaneous breathing period reduced the risk of postoperative atrial fibrillation (OR 0.31 for an interquartile increase in DFA alpha(1), 95% CI 0.13-0.78), while higher entropy of symbolic dynamics increased it (OR 3.16 for an interquartile increase in SymDyn En, 95% CI 1.23-8.10), independently of age and clinical risk factors. CONCLUSION: The preoperatively altered nonlinear R-R interval dynamics were independent predictors of postoperative atrial fibrillation and might become a useful tool for the risk assessment of atrial fibrillation.

Comparison of methods for editing of ectopic beats in measurements of short-term non-linear heart rate dynamics.

Non-linear heart rate (HR) dynamics characterizes the fractal properties and complexity of the variations in HR. Ventricular and supraventricular ectopic beats might introduce a mathematical artefact to the analyses on sinus rhythm. We therefore evaluated the effects of different editing practices for ectopic beats such that 753 40-min ECG recordings were (i) not edited for the ectopic beats, or the ectopic beats were edited with (ii) an interpolation or with (iii) a deletion method before the analyses of non-linear HR dynamics. The non-linear HR dynamics analyses included detrended fluctuation analysis (DFA), approximate entropy, symbolic dynamics (SymDyn), fractal dimension and return map (RM). We found that the short-term scaling exponent (alpha1) of DFA, forbidden words of SymDyn and RM were sensitive measurements to the ectopic beats and there were strong correlations between these measurements and the number of ectopic beats. In addition, the unedited ectopic beats significantly lowered the stability of these measurements. However, the editing either with interpolation or deletion method corrected the measurements for the bias caused by the ectopic beats. On the contrary, the entropy measurements were not as sensitive to the ectopic beats. In conclusion, the ectopic beats affect the non-linear HR dynamics of sinus rhythm differently, causing a more marked bias in fractal than in complexity measurements of non-linear HR dynamics. This erroneous effect of ectopic beats can be corrected with a proper editing of these measurements. Therefore, there is an obvious need for standardized editing practices for ectopic beats before the analysis of non-linear HR dynamics.

Effect of acute carbon monoxide exposure on heart rate variability in patients with coronary artery disease.

Epidemiological studies have shown that air pollution is associated with increased cardiovascular mortality. Although the pathophysiological mechanisms behind this association have remained largely unknown, it has been suggested that changes in cardiac autonomic function may play a role. In this study, we investigated the association between acute carbon monoxide (CO) exposure and cardiac autonomic function as measured by heart rate variability (HRV) in subjects with stable coronary artery disease. Twenty-four hour ambulatory electrocardiographic recording with simultaneous continuous personal CO concentration monitoring was performed in six male patients with angiographically verified coronary artery disease three times with 1-week intervals. Time domain measures of HRV were calculated for 5-min segments before and during the CO exposure periods. For further analysis CO exposures were divided into low (2.7 p.p.m.) CO exposure periods. The mean of maximum CO levels during 61 CO exposure periods was 4.6 p.p.m. (SD 5.0 p.p.m.). High CO exposure was associated with an increase in the square root of the mean of the sum of the squares of differences between adjacent RR intervals (r-MSSD) (P = 0.034). Heart rate remained unchanged during the CO exposure. In conclusion, acute CO exposure which represented most likely exposure derived from traffic seems to modify cardiac autonomic control in patients with stable coronary artery disease.

Stability over time of short-term heart rate variability.

Heart rate variability (HRV) is a widely used method to assess cardiac autonomic control. However, the reproducibility of especially short-term HRV has not been properly evaluated. Therefore, we assessed the stability of short-term HRV over a three to four month period. We had seven consecutive electrocardiographic (ECG) recordings from 89 subjects with stable coronary artery disease obtained during a large multicenter study. The HRV assessments were performed from these 40-minute ECG-recordings simulating normal daily activities, i. e., recordings consisting of 5 to 10 minute periods of rest, paced breathing, standing, submaximal exercise and recovery. Both time and frequency domain HRV analyses were conducted from the whole 40-minute recordings and from the 5-minute periods of rest and paced breathing. The coefficient of variation (CV) varied between 5.1-16.7% for the 40-minute and 6.0-37.1% for the 5-minute time domain and 4.4-11.0 % for the 40-minute and 7.2-16.5 % for the 5-minute frequency domain measurements. The mean of the RR intervals and the total power showed the highest stability over time. The most unstable measure was the standard deviation of all NN intervals (SDNN). In conclusion, most short-term HRV measures were highly stable over time indicating low physiological variation. However, the SDNN showed large variability in consecutive recordings.