Author + information
- Received March 18, 2010
- Revision received July 28, 2010
- Accepted August 10, 2010
- Published online January 11, 2011.
- Gopal Sivagangabalan, MBBS⁎,
- Danna Spears, MD⁎,
- Stephane Masse, MSc⁎,
- Bruce Urch, PhD†,‡,
- Robert D. Brook, MD§,
- Frances Silverman, PhD†,‡∥,¶,
- Diane R. Gold, MD#,
- Karl Z. Lukic, MD†,
- Mary Speck, MLT†,
- Marjan Kusha, BSc⁎,
- Talha Farid, MBBS⁎,
- Kwaku Poku, BSc⁎,
- Evelyn Shi, BSc⁎,
- John Floras, MD⁎ and
- Kumaraswamy Nanthakumar, MD⁎,⁎ ()
- ↵⁎Reprint requests and correspondence:
Dr. Kumaraswamy Nanthakumar, Hull Family Cardiac Fibrillation Management Laboratory, Toronto General Hospital, GW 3-522, 150 Gerrard Street West, Toronto, Ontario M5G 2C4, Canada
Objectives We tested the hypothesis that exposure to concentrated ambient particles (CAP) and/or ozone (O3) would increase dispersion of ventricular repolarization.
Background Elevated levels of air pollution are associated with cardiac arrhythmias through mechanisms yet to be elucidated.
Methods Each of 25 volunteers (18 to 50 years of age) had four 2-h exposures to 150 μg/m3 CAP; 120 parts per billion O3; CAP + O3; and filtered air (FA). Exposure-induced changes (Δ = 5-min epochs at end-start) in spatial dispersion of repolarization were determined from continuous 12-lead electrocardiographic recording.
Results Spatial dispersion of repolarization assessed by corrected ΔT-wave peak to T-wave end interval increased significantly for CAP + O3 (0.17 ± 0.03, p < 0.0001) exposure only, remaining significant when factoring FA (CAP + O3 − FA) as control (0.11 ± 0.04, p = 0.013). The influence on repolarization was further verified by a significant increase in ΔQT dispersion (for CAP + O3 compared with FA (5.7 ± 1.4, p = 0.0002). When the low-frequency to high-frequency ratio of heart rate variability (a conventional representation of sympathetic-parasympathetic balances) was included as a covariate, the effect estimate was positive for both corrected ΔT-wave peak to T-wave end interval (p = 0.002) and ΔQT dispersion (p = 0.038). When the high-frequency component (parasympathetic heart rate modulation) was included as a covariate with corrected ΔT-wave peak to T-wave end interval, the effect estimate for high frequency was inverse (p = 0.02).
Conclusions CAP + O3 exposure alters dispersion of ventricular repolarization in part by increasing sympathetic and decreasing parasympathetic heart rate modulation. Detection of changes in repolarization parameters, even in this small cohort of healthy individuals, suggests an underappreciated role for air pollutants in urban arrhythmogenesis.
Recently, Pope et al. (1) demonstrated that decreases in ambient fine particulate air pollution correlate with improvement in life expectancy. This report followed several earlier studies that linked air pollution to mortality (2–5). Cardiovascular mortality associated with air pollution exposure may be sudden or arrhythmic in nature (6,7), a particular concern with the increasing number of people surviving with structural heart disease who are at increased risk of sudden death. Several observational studies have demonstrated a relationship between daily pollution levels and episodes of arrhythmia (8–13). However, the specific polluting agents and mechanisms responsible for such disturbances of rhythm were not characterized.
Spatial dispersion of repolarization, measured by noninvasive electrocardiogram (ECG)-derived methods, reflects alterations in repolarizing currents at the myocardial level and is a well-characterized contributor to ventricular arrhythmias. A definitive link between air pollution exposure and alterations in dispersion of repolarization has not been established. To test the effects of controlled exposure to air pollution in the form of concentrated ambient particles (CAP), ozone (O3), or both, we prospectively studied healthy human volunteers with the hypothesis that specific exposures would increase ECG-derived measures of spatial dispersion of repolarization. Testing the hypothesis would provide insights into the possible mechanisms of arrhythmogenesis related to air pollution.
The study was approved by the Research Ethics Committees of St. Michael's Hospital Toronto and the University of Toronto, Toronto, Ontario, Canada. All subjects provided written informed consent. Participants were recruited by placing posters around the University of Toronto campus. Subjects were healthy 18- to 50-year-old nonsmokers, without any preexisting cardiovascular disease or risk factors and not taking any prescribed medications. Screening exclusion criteria included a fasting total cholesterol >240 mg/dl or glucose >126 mg/dl, hypotension (resting blood pressure [BP] <100/50 mm Hg), hypertension (resting BP >140/90 mm Hg), airway disease (chronic obstructive pulmonary disease/asthma), pregnancy, or lactation.
Study design and exposure protocol
The study design was a 2 × 2 factorial with randomized block allocation of subjects to 4 exposure conditions and included 25 volunteers. On 4 occasions, at least 2 weeks apart, each participant was exposed at rest for 2 h to the following exposures: 1) CAP alone, 150 μg/m3; 2) O3 alone, 120 parts per billion; 3) CAP + O3; and 4) a filtered air (FA) control. On the first study day, each participant was randomly assigned 1 of 24 possible exposure-order-sequences (e.g., 1, 2, 3, 4) balanced for the 4 exposure conditions. Thus, each person acted as his or her own control. To reduce the impact of circadian variation and external ambient exposure, each study was scheduled to start at 9:00 am.
Baseline evaluation at each session included a 12-lead ECG, 10-min resting Holter recording, and resting BP. The latter 3 variables were acquired again after the exposure. During exposures, BP and respiratory minute ventilation were recorded at 30-min intervals. Immediately before entering the exposure enclosure (Fig. 1), participants were connected to the 12-lead ECG machine. An automated oscillometric BP device (Oscar-1 or Oscar-2, SunTech Medical Instruments, Inc., Raleigh, North Carolina) cuff was applied to the left upper arm before subjects entered the exposure enclosure. Data from the 12-lead ECG (PC-ECG 1200, Norav Medical Ltd., Kiryat Bialik, Israel) was logged continuously and displayed on a computer terminal outside the enclosure. Although participants were informed of the exposures that they would receive, they were not told the order in which they would be exposed. The technologist carrying out the exposure was not blinded; however, all data were coded so that the persons performing data entry and data analysis was blinded to the exposure condition.
CAP exposures were produced with a high flow (1,100 l/min) 2-stage virtual impactor system, using ambient fine particulate matter with an aerodynamic diameter <2.5 μm (PM2.5) drawn from outside the laboratory. O3 was produced by an arc generator that was added upstream to the particle concentrator. During FA exposures, a high-efficiency particle arrestor filter was inserted inline downstream to the particle concentrator. The subject exposure enclosure was a modified airtight body plethysmograph with exposure air flows (15 to 20 l/min) entering the enclosure via ducting ending in a face mask. PM2.5 levels were monitored during exposures by a tapered element oscillating microbalance (model 1400a, Rupprecht & Patashnick, Albany, New York). A PM2.5 filter sample was collected immediately upstream of the enclosure on a 47-mm Gelman Teflon filter with a 2-μm pore size (Pall Corp., Ann Arbor, Michigan) at an air flow of 8 l/min. The sample was analyzed gravimetrically for total mass on conditioned filters using a climate-controlled clean room. The human exposure facility and exposure characterization were described in detail previously (14,15), and a schematic of the exposure enclosure is shown in Figure 1. Because changes in respiratory pattern can also modify cardiac autonomic and repolarization parameters, changes in respiratory minute ventilation, tidal volume, and respiratory rate during exposure were recorded.
Measures of spatial dispersion of repolarization
Two independent methods were selected for assessment of spatial dispersion of repolarization: the interval from the T-wave peak to T-wave end (Tp-e) and QT dispersion (QTd). Two different methods for quantifying spatial dispersion of repolarization were used on the same dataset to confirm exposure effects.
The electrocardiographic data were processed using Stress PC ECG Application software, version 4.5.6 (Norav Medical Ltd.) and stored electronically. The Stress PC ECG Application software file was later converted to a text file as raw electrocardiographic data. Electrocardiographic data were available for 25 subjects who had electrocardiographic files for the FA and at least 1 pollutant exposure or a total of 91 exposures.
Analysis of ECG-derived data
The raw electrocardiographic data were initially verified by 1 investigator to confirm the quality of the data recordings. All data analyses were performed by investigators blinded to the subject and exposure. Lead V3 was used for Tp-e analysis by convention, with concurrent measurement in lead II to ensure optimal data quality and stability. Using a custom-designed cardiac mapping program developed with Matlab V7.5 (Mathworks, Natick, Massachusetts), the onset of the QRS complex as well as the peak and end points of each recorded T-wave were marked. The cardiac mapping software was used to extract the R-R intervals and Tp-e duration, with confirmatory visual verification by an independent observer. Because an ECG-derived interval (Tp-e) was continuously being measured with potential variation in heart rate (HR), the Tp-e interval was corrected for HR to normalize the data. The corrected Tp-e (cTp-e) was calculated by dividing Tp-e by the √R-R, a modified application of Bazett's formula more commonly applied to correct QT interval for HR. To evaluate the effect of each exposure on cTp-e, the change over time was determined by comparing the 5-min start and end periods (Δ = 5-min epochs at end-start). None of these periods contained significant baseline artifact or ectopic complexes. QTd was calculated as the maximum minus minimum QT interval across a minimum of 10 ECG leads. As with Tp-e, the ΔQTd during each exposure was calculated as the difference between the 5-min start and end epochs.
Changes within heart rate variability (HRV) spectral analysis was used to identify autonomic changes during each exposure, as previously described (16). A tachogram was built using the R-R interval time-series for each exposure. Linear interpolation was used to create equally spaced 100-Hz interval data. Power spectral density was computed using a Welch periodogram on the interpolated data using a 16,384-point window and 1,024-point overlap. To reduce edge effect, a Hamming window was used. Power for low frequencies (LFs) (from 0.04 to 0.15 Hz) and high frequencies (HFs) (from 0.15 to 0.4 Hz) was calculated by integrating the spectral density for each frequency band. The LF/HF ratio of HRV has been proposed as a power spectral representation of the relative relationship between sympathetic and parasympathetic HR modulation (16). The HF component of HRV has been proposed as a measure of parasympathetic HR modulation (16).
Data were analyzed using the statistical package SAS version 9.1 (SAS Institute, Inc., Cary, North Carolina). To examine the exposure-induced change of the electrocardiographic variables, a linear difference (Δ) was calculated for the end average 5-min segment minus the start average 5-min segment of each exposure. The linear difference was then used as the dependent variable in a mixed model. The 2 main effects, CAP and O3, and their interaction were tested. The model included a random-subject effect. Both CAP and O3 were included as dichotomous (yes/no) fixed effects. Least-squares means were computed for fixed effects. Interexposure comparisons were made by computing least-squares means differences. Intraexposure correlations between variables (e.g., ΔHR and ΔTp-e) are reported by linear regression. Statistical significance was reported as p < 0.05.
Exposure-related influence on HR
There was a significant increase in HR over the course of exposure for CAP alone (2.3 ± 1.0 beats/min, p = 0.029) and CAP + O3 (3.5 ± 1.3 beats/min, p = 0.015), and a trend toward an increase in HR for FA (2.4 ± 1.4 beats/min, p = 0.099) and O3 alone (2.0 ± 1.2 beats/min, p = 0.102). However, the increase in HR was not associated with exposure (p = 0.70 for O3 effect and p = 0.46 for CAP effect). There was no significant relationship between ΔHR and ΔTp-e or ΔHR and ΔQTd for any exposure (p > 0.08).
Exposure-related changes in cTp-e
The changes in ΔcTp-e over each exposure are shown in Table 3 (model 1) and Figure 2A. There was a significant increase in ΔcTp-e for the combined CAP + O3 exposure (p < 0.0001). To further clarify the significance of this change, we took the FA exposures as a control into consideration (i.e., CAP + O3 − FA) and again demonstrated significance (p = 0.01).
Exposure-related changes in QTd
The results of QTd changes for each exposure are shown in Table 4 (model 1) and Figure 2B. QTd presents a more widely accepted method of quantifying dispersion of repolarization and therefore a means of validating the results of cTp-e analysis. Analogous to the ΔcTp-e findings, a significant increase in ΔQTd was noted for CAP + O3 compared with FA 5.7 ± 1.4 (95% confidence interval: 2.8 to 8.5; p = 0.0002). A significant increase in ΔQTd was also noted for the CAP alone exposure compared with FA 3.8 ± 1.4 (95% confidence interval: 1.0 to 6.7; p = 0.008).
HRV and spatial dispersion of repolarization during exposures
The change from baseline in the LF/HF ratio was a significant covariate when included in the mixed model with both ΔcTp-e (p = 0.002) (Table 3, model 2) and ΔQTd (p = 0.038) (Table 4, model 2). The effect estimate for ΔLF/HF was positive for both ΔcTp-e and ΔQTd. Thus, an increase in the LF/HF ratio (increase in sympathetic HR modulation) during exposure was associated with a corresponding increase in both measures of spatial dispersion of repolarization, ΔcTp-e and ΔQTd.
The change in spectral power of the HF domain of HRV was a significant covariate (p = 0.02) when included in the mixed model with ΔcTp-e change as the dependent variable. The effect estimate for ΔHF was negative (i.e., a decrease in vagal HR modulation during exposure was associated with a corresponding increase in Tp-e). No significant effect was observed for the LF (p = 0.89) or very low frequency (p = 0.26) domains. For ΔQTd, no significant effect was noted for HF (p = 0.29), LF (p = 0.48) or very low frequency (p = 0.83) domains.
Because changes in respiratory pattern can modify HRV changes, we examined the changes in respiratory minute ventilation, tidal volume, and respiratory rate during exposures, but did not observe any effect modification of breathing pattern on the latter association. Further, there was no significant exposure-induced changes in the breathing parameters.
CAP increased diastolic BP (p = 0.0495) but not systolic BP (p = 0.58) (Fig. 3). There was, however, no effect modification of BP (diastolic or systolic) on the association between exposure and cTp-e (p ≥ 0.19) or QTd (p > 0.11). Further, there was no significant correlation (p > 0.2) between the change in cTp-e or QTd and the change in diastolic BP or systolic BP. Thus, the effects on BP and measures of spatial dispersion of repolarization are most likely independent responses.
In healthy volunteers, spatial dispersion of cardiac repolarization was most affected by the combined pollutant exposure of CAP and O3 compared with the FA control. The modulation of dispersion of repolarization is partly mediated by an increase in sympathetic tone and a decrease in parasympathetic HR modulation. The strength of the current study is that the subjects' exposure levels were fully characterized, a feature that is lacking in most observational and epidemiologic studies on the cardiovascular effects of air pollution where central site monitoring data are usually used to estimate personal exposures. Thus, the controlled nature of the study combined with the use of 2 independent ECG-derived measures of spatial dispersion of repolarization allowed rigorous analysis of the selected pollutant effects.
The study of pollution and the issues surrounding their effects on human myocardium have been difficult to study and quantify due to conflicting factors such as lack of exposure characterization and quantification and adequate controls. These difficulties mirror the early investigations linking lung cancer with cigarette smoking, in which, although an association was observed, the scientific evidence did not support a causal mechanism. The recent key study by Pope et al. (1) identified increases in life expectancy with decreases in fine particulate air pollution across U.S. metropolitan populations older than 45 years of age. We postulate that the estimated increase in life expectancy of 0.61 ± 0.20 years (p = 0.004) is likely to be higher in people who are at increased risk of sudden death due to the associations between air pollution and arrhythmia events. This susceptible group predominantly comprises people with ischemic and nonischemic structural heart disease who have an underlying electrophysiological substrate for arrhythmia. The rationale for studying healthy volunteers in the current study was to decrease the potential adverse risks of the experimental procedure and to establish the physiological relationship between air pollution and repolarization parameters in healthy humans without the confounding effects of structural abnormality and ongoing treatment (antiarrhythmic or heart failure medication, cardiac resynchronization therapy).
In isolated ventricular wedge preparations, the Tp-e interval has been correlated with transmural dispersion of repolarization (17,18). More recently in whole-heart studies and modeling studies, Tp-e has been found to correlate with global rather than transmural dispersion of repolarization (19–21). However, regardless of whether it measures transmural or global dispersion, Tp-e has been associated with arrhythmic risk (22–27). An alternative and more widely accepted ECG-derived measure of spatial dispersion of repolarization is QTd. Although equally controversial, an association between QTd and arrhythmia risk has been well described (28–31).
The magnitude of changes observed in spatial dispersion of repolarization in the current study was significant although small. Small changes in surface ECG-derived cTp-e and QTd represent global dispersion changes of a larger and more significant magnitude at the myocardial level. The positive association between ΔLF/HF and both ΔcTp-e and ΔQTd provide a mechanistic association with increased sympathetic tone possibly contributing to the increased spatial dispersion of repolarization. The negative association between ΔHF and ΔcTp-e suggests withdrawal of parasympathetic HR modulation possibly contributing to the increased dispersion of repolarization observed with this measure. Several factors have been proposed as mechanisms linking long-term exposure to air pollution and adverse cardiovascular outcomes. These include direct reflex activity from airways via stimulation of pulmonary vagal afferents (32), autonomic modulation occurring secondary to exposure (33), pulmonary oxidative stress resulting in systemic inflammation and the production of reactive oxygen species (34), and perhaps direct myocardial toxicity by circulating inhaled particles or inflammatory mediators (35,36). Interestingly, in the current study, the changes in diastolic BP, also likely autonomic mediated (37), and spatial dispersion of repolarization were independent of each other, suggesting further nonautonomic-mediated mechanisms for the increases in spatial dispersion of repolarization.
Fine particles account for most of the atmospheric particulate matter mass and are the least scavenged by atmospheric processes. Therefore, they have the longest atmospheric residence time and so have the greatest potential for human exposure under typical ambient conditions. Although the PM2.5 concentration used in this study was higher than typically observed over 24 h, levels exceeding 150 μg/m3 can occur for 1- to 2-h periods in many North American locations and are commonly encountered over even longer durations throughout developing nations (38). The greatest changes in measures of dispersion of repolarization were observed for the combined CAP + O3 exposure. The combined CAP + O3 exposure included in this study is clinically relevant because both pollutants are major contributors to urban smog and are typically present together. Strategies to decrease exposure to particulate matter or O3 independently may reduce the effects on dispersion of repolarization because both were required to produce the changes observed.
This study found measurable alteration in cardiac electrophysiological properties in a relatively small cohort of healthy volunteers to levels of pollution commonly encountered across the developed and developing world. The fact that in this small cohort of “normal” subjects, changes in repolarization parameters were measurable suggests an underappreciated role for pollutants in urban arrhythmogenesis. The magnitude of change in global dispersion of repolarization is unlikely to pose an immediate risk to healthy individuals with normal cardiovascular reserve. However, pollution-induced changes in repolarization dynamics may pose a risk to those with preexisting heart disease and baseline abnormalities of repolarization, especially during periods of acute or prolonged smog episodes. Further coordinated research into the proarrhythmic role of pollutants is urgently required, with concurrent physician awareness and participation in global efforts to increase education (e.g., smog alerts) and to reduce pollutant emissions.
Comparisons or extrapolations from this sample of healthy volunteers to long-term exposures in the general population must be made with caution. A possible explanation for the increase in HR across all exposures is that the testing protocol and confinement to the exposure chamber heightened sympathetic tone. This may have had an independent effect on dispersion of repolarization unrelated to the individual exposures and highlights the importance of our rigorous methodology of assessing the pollutant effects relative to the FA control effect.
Healthy human volunteers exposed to air pollution in the form of concentrated ambient particles and O3 demonstrate detectable increases in cTp-e and QTd, measures of spatial dispersion of myocardial repolarization. This increase is mediated by a relative increase in sympathetic and a relative decrease in parasympathetic HR modulation.
The authors thank the Gage team for their contributions, Paul Corey for statistical advice, and Jeffrey R. Brook at Environment Canada for filter analyses, facility/monitoring equipment, and technical advice.
Dr. Sivagangabalan is a recipient of a Heart and Stroke Foundation (HSF) of Canada Research Fellowship. Dr. Floras currently holds the position of Tier 1 Canada Research Chair in Integrative Cardiovascular Biology and is the recipient of an HSF of Ontario Career Investigator Award. This research was funded by grants by CIHR and HSF to Dr. Nanthakumar. This research was also funded in part by Natural Resources Canada and the U.S. Environmental Protection Agency (EPA) through CR830837 to the University of Michigan. Additional support for this study was provided by the Air Quality Health Effects Research Section, Government of Canada; Ontario Thoracic Society; AllerGen NCE Inc., and U.S. EPA (R832416) to Harvard University. It has not been subjected to the EPA's required peer and policy review and therefore does not necessarily reflect views of the Agency and no official endorsement should be inferred. All other authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- blood pressure
- concentrated ambient particle(s)
- corrected T-wave peak to T-wave end interval
- filtered air
- heart rate
- heart rate variability
- particulate matter with a diameter <2.5 μm
- QT dispersion
- T-wave peak to T-wave end interval
- Received March 18, 2010.
- Revision received July 28, 2010.
- Accepted August 10, 2010.
- American College of Cardiology Foundation
- Lave L.B.,
- Seskin E.P.
- Pope C.A. 3rd.,
- Burnett R.T.,
- Thurston G.D.,
- et al.
- Routledge H.C.,
- Ayres J.G.
- Rich D.Q.,
- Schwartz J.,
- Mittleman M.A.,
- et al.
- Santos U.P.,
- Terra-Filho M.,
- Lin C.A.,
- et al.
- Ljungman P.L.,
- Berglind N.,
- Holmgren C.,
- et al.
- ↵(1996) Heart rate variability: standards of measurement, physiological interpretation and clinical use: Task Force of the European Society of Cardiology and the North American Society of Pacing and Electrophysiology. Circulation 93:1043–1065.
- Yan G.X.,
- Antzelevitch C.
- Antzelevitch C.
- Takenaka K.,
- Ai T.,
- Shimizu W.,
- et al.
- Topilski I.,
- Rogowski O.,
- Rosso R.,
- et al.
- Castro Hevia J.,
- Antzelevitch C.,
- Tornes Barzaga F.,
- et al.
- Malik M.,
- Batchvarov V.N.
- Okin P.M.,
- Devereux R.B.,
- Howard B.V.,
- Fabsitz R.R.,
- Lee E.T.,
- Welty T.K.
- Chalil S.,
- Yousef Z.R.,
- Muyhaldeen S.A.,
- et al.
- Peters A.,
- Dockery D.W.,
- Muller J.E.,
- Mittleman M.A.
- Brook R.D.,
- Franklin B.,
- Cascio W.,
- et al.
- Nemmar A.,
- Hoet P.H.,
- Vanquickenborne B.,
- et al.