Increased Nonconducted P-Wave Arrhythmias after a Single Oil Fly Ash Inhalation Exposure in Hypertensive Rats

Background Exposure to combustion-derived fine particulate matter (PM) is associated with increased cardiovascular morbidity and mortality especially in individuals with cardiovascular disease, including hypertension. PM inhalation causes several adverse changes in cardiac function that are reflected in the electrocardiogram (ECG), including altered cardiac rhythm, myocardial ischemia, and reduced heart rate variability (HRV). The sensitivity and reliability of ECG-derived parameters as indicators of the cardiovascular toxicity of PM in rats are unclear. Objective We hypothesized that spontaneously hypertensive (SH) rats are more susceptible to the development of PM-induced arrhythmia, altered ECG morphology, and reduced HRV than are Wistar Kyoto (WKY) rats, a related strain with normal blood pressure. Methods We exposed rats once by nose-only inhalation for 4 hr to residual oil fly ash (ROFA), an emission source particle rich in transition metals, or to air and then sacrificed them 1 or 48 hr later. Results ROFA-exposed SH rats developed nonconducted P-wave arrhythmias but no changes in ECG morphology or HRV. We found no ECG effects in ROFA-exposed WKY rats. ROFA-exposed SH rats also had greater pulmonary injury, neutrophil infiltration, and serum C-reactive protein than did ROFA-exposed WKY rats. Conclusions These results suggest that cardiac arrhythmias may be an early sensitive indicator of the propensity for PM inhalation to modify cardiovascular function.

Daytoday variations in airborne particulate matter (PM) associated with air pollution have been linked to increases in cardiovascular morbidity and mortality (Dockery et al. 1993;Pope et al. 1995;Samet et al. 2000;Schwartz et al. 1996), especially in individuals who have preexisting cardiovascular disease (Dockery et al. 2001;Miller et al. 2007;Pope 2000). PM exposure likely exacerbates the cardiac and vascular pathophysiology attendant to hypertension, ischemia, heart failure, dia betes, coronary artery disease, and other dysfunc tions of the cardiovascular system. Several mechanisms of PMinduced cardiac dysfunc tion in humans have been postulated, includ ing autonomic modulation, direct effects of PM constituents on ion channels, myocardial ischemia, and vascular dysfunction related to systemic inflammation (Brook et al. 2003;Schulz et al. 2005;Utell et al. 2002).
The electrocardiogram (ECG) is an impor tant clinical tool that has great prognostic value in the assessment of cardio vascular function. Electrocardiographic recordings in humans have been used to detect abnormal myo cardial impulse conduction, cardiac rhythm disturbances, and altered autonomic regu lation of heart rate (HR) via the assessment of HR variability (HRV). PM exposure has been linked to changes in ECG morphology, including STsegment depression (Mills et al. 2007; Pekkanen et al. 2002), altered Twave amplitude (Henneberger et al. 2005), and QT prolongation (Yue et al. 2007). PM exposure has also been associated with cardiac rhythm changes, including increases in atrial ecto pic beats (Sarnat et al. 2006) and ventricu lar tachyarrhythmias (Riediker et al. 2004). Additionally, fine PM exposures have been associated with low HRV, suggesting an inap propriate increase in sympathetic tone to the heart (Rowan et al. 2007;Zareba et al. 2001). Studies in animal models have yielded a simi lar spectrum of ECG effects (Campen et al. 2001;Watkinson et al. 1998;Wellenius et al. 2002). However, the utility of any or all of these electrocardiographic end points as sensi tive or reproducible toxicologic indicators for targeted pathophysiologic studies of PM and its components in rats is unclear.
Recently, Ostro et al. (2007) demon strated an association between increased cardio vascular mortality and the finePMassociated metals nickel and iron in ambi ent air in nine California counties. Residual oil fly ash (ROFA), a waste product of fossil fuel combustion from boilers, is rich in the transition metals iron, nickel, and vanadium (Dreher et al. 1997) and, when released as a fugitive particle, is an important contribu tor to ambient fine PM air pollution (Ghio et al. 2002;Lippmann et al. 2006). Among boiler workers, exposure to ROFA deposits correlated strongly with nocturnal depres sion of HRV, suggesting a significant influ ence of transition metals on cardiac function (Cavallari et al. 2008). Kodavanti et al. (2000) demonstrated that hypertensive rats that inhale ROFA develop ST depression during expo sure. This sensitivity of hypertensive rats to ST depression suggests that they may also be susceptible to cardiac arrhythmias and changes in HRV. In the present study, we examined the effects of a single acute ROFA inhalation exposure on ECG morphology, HRV, and arrhythmia development in hypertensive and normal rats. Our goal in this study was to establish a reproducible exposure model in sensitive rats to further elucidate the underly ing mechanisms of cardiopulmonary responses to PM. We used a welldefined and widely used PM to dissect the various electrocardia graphic lesions induced with a single inhala tion challenge.

Animals.
We obtained adult (12weekold) male Wistar Kyoto (WKY; background strain for SH rats) and spontaneously hypertensive (SH) rats (SHR/NCrIBR) from Charles River Laboratory (Raleigh, NC). Rats were housed in plastic cages (one per cage), maintained on Background: Exposure to combustion-derived fine particulate matter (PM) is associated with increased cardiovascular morbidity and mortality especially in individuals with cardiovascular disease, including hypertension. PM inhalation causes several adverse changes in cardiac function that are reflected in the electrocardiogram (ECG), including altered cardiac rhythm, myocardial ischemia, and reduced heart rate variability (HRV). The sensitivity and reliability of ECG-derived parameters as indicators of the cardiovascular toxicity of PM in rats are unclear. oBjective: We hypothesized that spontaneously hypertensive (SH) rats are more susceptible to the development of PM-induced arrhythmia, altered ECG morphology, and reduced HRV than are Wistar Kyoto (WKY) rats, a related strain with normal blood pressure. Methods: We exposed rats once by nose-only inhalation for 4 hr to residual oil fly ash (ROFA), an emission source particle rich in transition metals, or to air and then sacrificed them 1 or 48 hr later. results: ROFA-exposed SH rats developed nonconducted P-wave arrhythmias but no changes in ECG morphology or HRV. We found no ECG effects in ROFA-exposed WKY rats. ROFA-exposed SH rats also had greater pulmonary injury, neutrophil infiltration, and serum C-reactive protein than did ROFA-exposed WKY rats. conclusions: These results suggest that cardiac arrhythmias may be an early sensitive indicator of the propensity for PM inhalation to modify cardiovascular function. key words : arrhythmia, electrocardiogram, heart rate variability, hypertension, inhala tion, non conducted P-waves, particulate matter, rats, residual oil fly ash. a 12/12hr light/dark cycle at approximately 22°C and 50% relative humidity in our Association for Assessment and Accreditation of Laboratory Animal Care-approved facility, and held for a minimum of 1 week before implan tation. All protocols were approved by the Institutional Animal Care and Use Committee of the U.S. Environmental Protection Agency. Food (Prolab RMH 3000; PMI Nutrition International, St. Louis, MO) and water were provided ad libitum. The animals were treated humanely and with regard for alleviation of suf fering. We randomized all rats by weight.
Surgical implantation of telemeters. We anesthetized animals (WKYair, n = 7; WKYROFA, n = 8; SHair, n = 8; SHROFA, n = 8) with an intraperitoneal injection of 1 mL/kg ketaminexylazine solution (80 mg/ mL ketamine HCl, 12 mg/mL xylazine HCl; Sigma Chemical, St. Louis, MO) and implanted them with a bio potential radio telemetry transmitter [model TA11CTAF40; Data Sciences International, Inc. (DSI), St. Paul, MN] using aseptic surgical procedures to obtain an ECG signal similar to that of lead II from the standard ECG and to allow measure ment of core body temperature (Watkinson et al. 1998). The WKYair group had one less animal because of postsurgical transmitter malfunction. We allowed animals 2 weeks for recovery from surgery preexposure to ROFA.
Exposure to ROFA. The ROFA used in this study (Southern Research Institute, Birmingham, AL) was collected as a fugi tive stack emission postemission control at a Florida Power & Light plant that burned no. 6 grade residual oil containing 1% sulfur (Hatch et al. 1985) and was the same ROFA sample used by Kodavanti et al (2000). The compo sition of ROFA included vanadium, nickel, and iron sulfates and has been previously described (Dreher et al. 1997). We selected the target concentration (15 mg/m 3 ) for this study to mirror that used by Kodavanti et al. (2000) in an effort to reproduce the effect of ROFA on the ST interval in SH rats, and to assess induction of altered HRV and cardiac arrhythmia. After 2 days of acclimatization, we exposed rats once via noseonly inhalation for 4 hr to 13.4 mg/m 3 of the PM 2.5 fraction (mass median aerodynamic diameter, 2.09 µm; geometric SD, 3.37) of ROFA or to filtered air in two separate 24port noseonly inhalation chambers (Lab Products, Seaford, DE), as pre viously described (Ledbetter et al. 1998).
Radiotelemetry data acquisition and analysis. Radiotelemetry allowed continuous moni toring and collection of electrocardiographic data in unanesthetized rats from implantation of the transmitters until sacrifice. We positioned receivers (DataART3.01; DSI) beneath home cages (when the rats were unrestrained) or beneath noseonly inhalation chambers (when rats were restrained in tubes) during exposure. Sixty second segments of ECG waveforms were acquired and saved at 60min intervals, from surgical recovery through euthanasia; values were obtained sequentially by animal and rep resent averages of 60 sec of data/animal for each hour. HR was obtained from the ECG wave form. Preexposure data permitted each animal to serve as its own control, and animals exposed to air provided timepaired control data. We obtained preexposure baseline data, that is, one 1min segment of the ECG waveform/hour, while the animals were unrestrained and in their home cages before noseonly inhalation exposure. We also obtained preexposure base line data at the same sampling rate while the rats were in the noseonly restraint just before the beginning of exposure. Data were also col lected during the noseonly inhalation exposure period once/hour for the duration of the 4hr exposure period. The rats were then returned to their home cages and postexposure data were collected until euthanasia approximately 48 hr after the beginning of exposure.
Electrocardiogram. We analyzed each signal using eMOUSE (Boston, MA), an Internetbased physiologic waveform analy sis portal. eMOUSE incorporates Fourier analyses and linear timeinvariant digital fil tering of frequencies < 2 Hz and > 100 Hz to minimize environmental signal disturbances, as previously described (Chu et al. 2001). The software uses a peak detection algorithm to find the peak of the Rwaves and to calculate HR. We calculated HRV as the mean of the differences between sequential HRs for the complete set of ECG signals. The software plots its interpretation of P, Q, R, S, and T for each beat so that spurious data resulting from unfiltered noise or motion artifacts may be rejected. It then calculates the means of the ECG intervals [i.e., PQ (peak of Pwave to beginning of QRS), PR (peak of Pwave to peak of QRS), QRS, QT, ST, and RR inter vals] for each set of waveforms. The QT inter vals were rate corrected (QTc) by application of the equation recommended by Mitchell et al. (1998) for use in mice and is applicable to rats. We used ECGAuto software (EMKA Technologies USA, Falls Church, VA) to calculate ST and Twave areas. (We calcu lated ST area as the area below the isoelectric line between the S point of the ECG and the beginning of the Twave, and Twave area as the area above the isoelectric line between the beginning of the Twave and the end of the Twave.) For each 1min stream of ECG  waveforms, we acquired mean time between successive QRS complex peaks (RR interval), mean HR, and mean HRVanalysis-generated timedomain measures. The timedomain measures included standard deviation of the time between normaltonormal beats (SDNN), and root mean squared successive differences (RMSSD). HRV analysis was also conducted in the frequency domain using a fastFourier transform. The spectral power obtained from this transformation represents the total harmonic variability for the fre quency range being analyzed. In this study, we divided the spectrum into lowfrequency (LF) and highfrequency (HF) regions. The ratio of these two frequency domains (LF/ HF) was calculated as an estimate of the rela tive balance between sympathetic (LF) and vagal (HF) activity. To account for potential effects of normal circadian rhythm, we quan tified both ECG and HRV parameters for timematched comparisons and compared them at different times of the day during the preexposure and postexposure periods while rats were unrestrained in their home cages. We analyzed only some of the total data col lected during the monitoring period: morning (1000 hr), afternoon (1600 hr), and evening (2200 hr). ECG and HRV parameters during exposure were analyzed for baseline (record ings while in noseonly restraint immediately before the beginning of exposure) and for hours 1-4 (constituting the entire exposure period between 0830 hr and 1230 hr). ECGAuto software (EMKA Technologies) was used to visualize individual ECG signals for identification and enumeration of spe cific cardiac arrhythmia events. We used the Lambeth conventions (Walker et al. 1988) as guidelines for the identification of cardiac arrhythmic events in rats. Arrhythmias were identified as nonconducted Pwaves, atrial and ventricular premature beats, ventricular bigeminy, ventricular couplets, or ventricular tachycardia. Figure 1A lists examples of these arrhythmias found in the rats used in this study. Arrhythmias are generally infrequent, and were therefore quantified and totaled over a 24hr period (24 segments of 1 min each) preexposure (corresponding to the same times assessed postexposure), during the 4hr expo sure period (four segments of 1 min each), or during the 24hr period (24 segments of 1 min each) beginning immediately post exposure.
Histopathology. We immersed each heart in a large volume of 10% acetatebuffered formalin. After fixation, hearts were trimmed and then processed to paraffin blocks, sec tioned at a thickness of 5 µm, placed on glass slides, and stained with hematoxylin and eosin or BarbeitoLopez trichrome stain for routine diagnosis of myocardial degeneration or necro sis (Milei and Bolomo 1983). Histopathologic changes were assessed by the study pathologist (A. Nyska, Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel).
Statistics. We performed statistical analy ses for all data in this study using SAS soft ware, version 9.1.3 (SAS Institute, Inc., Cary, NC). PROC MIXED and PROC GLIMMIX procedures were used to analyze all ECG and HRVgenerated data. We used a linear mixed model with restricted maximumlikelihood estimation analysis (SAS) and leastsquare means post hoc test to determine statistical differences for all data. The arrhythmia data, however, were not normally distributed, so for these we used PROC GENMOD for count data using Poisson regression. All arrhythmia values had one added to them to prevent zero values and to allow model convergence. All the biochemical and differential data were analyzed using a one, two, or threeway analysis of variance (ANOVA) model examin ing the main effects of each model as well as the interactive effects of two and threeway ANOVA models. We considered p < 0.05 sta tistically significant. A significant interaction in biochemical or differential data resulted in pairwise comparisons performed using Tukey's post hoc test.

HR and HRV.
We found no significant dif ferences in HR among any of the groups pre or postexposure (Table 1)  (data not shown). There were no significant differences in SDNN, RMSSD, LF, HF, or LF/HF among any of the groups pre or postexposure (Table 1) or during exposure (data not shown).

ECG interval duration and morphology.
We found no significant differences among any of the groups pre or postexposure in any of the ECG intervals (Table 2; data for PQ and ST intervals were not statistically signifi cantly different and are not shown) or in ST and Twave areas (Table 2).
Cardiac arrhythmia events. Both strains of rats had random and infrequent arrhyth mias preexposure, including atrial and ven tricular premature beats. We observed very few instances of bigeminy, ventricular cou plets, or ventricular tachycardia in either strain pre or postexposure. There were very few arrhythmias of any type evident during the 4hr exposure period in any of the rats (data not shown). ROFA exposure caused a 400% increase (p < 0.05) in the episodes of nonconducted Pwaves in SH rats post exposure (24hr period) compared with the number of events in the same rats pre exposure (timematched; Figure 1B). The ROFA induced increase in nonconducted Pwaves in SH rats was also > 327% than the aver age level in airexposed SH rats postexposure (p < 0.05). In addition, the ROFAinduced increase in nonconducted Pwaves in SH rats was > 133% than the average level in WKY rats exposed to ROFA postexposure (p < 0.05). Most of the nonconducted Pwaves took place in the first 12 hr postexposure to ROFA in the SH rats. Mean postexposure counts (± SE) for atrial premature beats were as follows: WKYair, 1.14 ± 0.14; WKYROFA, 2.14 ± 0.40; SHair, 1.17 ± 0.17; SHROFA, 1.57 ± 0.57. Mean postexposure counts for ven tricular premature beats were as follows: WKY air, 1.71 ± 0.17; WKYROFA,1.43 ± 0.20; SHair, 2.0 ± 0.52; SHROFA, 1.86 ± 0.46. However, we found no statistically significant exposure or timerelated differences in either type of premature beat (p > 0.05).
Inflammatory cell infiltration into the lung. ROFAexposed SH rats had significantly greater BAL neutrophils than did airexposed SH rats (6,376% difference; p < 0.05) or ROFAexposed WKY rats (659% difference; p < 0.05) 1 hr postexposure (Figure 2A). We found no significant difference in BAL neutro phils between air and ROFAexposed WKY rats, or in BAL macrophages between ROFA exposed SH and WKY rats, at 1 hr postexpo sure (data not shown). There was no significant difference among any of the treatment groups in macrophages and neutrophils 48 hr postex posure, suggesting that the inflammatory influx into the lung may peak soon after exposure.

Markers of injury and edema in lung.
ROFAexposed SH rats had significantly greater BAL LDH levels than did airex posed SH rats (211% difference; p < 0.05) or ROFAexposed WKY rats (54% difference; p < 0.05) 48 hr postexposure ( Figure 2B). We found no significant differences in any of the exposure groups in BAL LDH levels 1 hr postexposure. ROFA exposure did not signifi cantly increase BAL LDH levels in WKY rats 48 hr postexposure.
We found no significant effect of ROFA exposure in either strain relative to air controls 1 hr postexposure in BAL albumin or protein levels. Unlike the inflammatory influx, injury to the lung required more time to develop. Serum CRP. SH rats exposed to ROFA had significantly higher CRP levels than did airexposed SH rats (18% difference; p < 0.05) or ROFAexposed WKY rats (32% difference; p < 0.05) 48 hr postexposure ( Figure 2D). We found no significant differences among any of the treatment groups relative to their respective airexposed controls in serum CRP levels 1 hr postexposure. There was no significant differ ence between WKY rats exposed to ROFA and WKY rats exposed to air 48 hr postexposure. Like the markers of injury in the lung, elevation in serum CRP required more time to develop than did the inflammatory influx into the lung.
Heart weight. Mean hearttobody weight ratios of control rats exposed to air 48 hr postexposure were, for WKY, 3.2 ± 0.2 mg/g, and for SH, 3.53 ± 0.07 mg/g. Thus, SH rats had a > 10.3% hearttobody weight ratio than did WKY rats, which was statistically significant (p < 0.05). We found no statistically signifi cant effect of ROFA inhalation on heartto bodyweight ratios in either strain.
Heart histopathology and temperature. There were no PMexposure-related or strain effects on cardiac tissue morphology or body temperature (data not shown).

Discussion
In the present study, cardiac arrhythmias were more sensitive indicators of the effects of acute oil fly ash inhalation than were changes in ECG morphology and HRV in hyper tensive rats. ROFA inhalation increased the frequency of nonconducted Pwaves in hyper tensive but not in normal rats during the first 24 hr postexposure. Most of these arrhythmic events occurred during the first 12 hr after ROFA exposure. The non conducted Pwaves in the ROFAexposed SH rats in this study were preceded by unchanging PR intervals, and thus were similar to Mobitz type II sec onddegree atrioventricular block observed in humans, which is usually associated with a lesion in the bundle branch conduction pathway (Scheidt 1986). It is unclear, how ever, whether such lesions exist in rats and whether a single ROFA exposure would be sufficient to elicit such injury. Additionally, the role of autonomic, electrolyte, or other influences in the generation of these car diac events is uncertain, so the etiology of the nonconducted Pwaves that were pres ent in ROFAexposed SH rats needs to be further studied. Nevertheless, these abnormal rat rhythms are in line with multiple stud ies that have linked fine PM exposure with the development of cardiac arrhythmias. Fine PM exposure has also been associated with increases in ventricular and supra ventricular tachycardia in patients with coronary heart disease (Berger et al. 2006;Dockery et al. 2005), atrial premature beats in patrol officers (Riediker et al. 2004), and atrial and ven tricular ectopy in elderly subjects (Sarnat et al. 2006). Analogously, ROFA or its transition metals have been linked to the development of cardiac arrhythmia in rats (Campen et al. 2001;Watkinson et al. 1998;Wellenius et al. 2002). Collectively, these studies suggest that cardiac arrhythmias may be sensitive and quantifiable indicators of the cardiovascular effects of PM in susceptible individuals. ROFA exposure in both strains of rats failed to cause a decrement in HRV, suggest ing that HRV parameters are not as sensitive as cardiac arrhythmias to the effects of acute ROFA inhalation in this model. HRV is the degree of difference in the interbeat intervals of successive heartbeats and is an indicator of the balance between the sympathetic and parasympathetic arms of the autonomic ner vous system (Rowan et al. 2007). Low HRV, reflecting inappropriately increased sympa thetic tone (Rowan et al. 2007), is associated with an increased risk of cardiac arrhythmia (Corey et al. 2006) and an increased mortality rate in people with heart disease (Bigger et al. 1993;Fauchier et al. 2004). An association between high ambient PM and low HRV has been observed in several different studies in both healthy and diseased individuals (Chuang et al. 2005;Muggenburg et al. 2000;Schulz et al. 2005;Vallejo et al. 2006). Other inves tigators, however, have reported increases in HRV with exposure to vanadium in ambient PM (Magari et al. 2002) and fine PM mass itself (Timonen et al. 2006). The data in this study and others suggest that the effects of PM pollution on HRV and the prognostic value of HRV in air pollution studies are uncertain.
ROFA exposure in both strains of rats failed to cause any changes in ECG interval duration, including RR and QT intervals, sug gesting that such end points may not be sensi tive to acute PM exposure in this model of hypertension. In addition to interval time, the morphology of the ECG waveform can also provide information about cardiac physiol ogy. Air pollution exposure has been associ ated with changes in ventricular repolarization as indicated by altered Twave morphology, or Twave alternans, in patients with isch emic heart disease (Henneberger et al. 2005). Changes in repolarization may trigger arrhyth mogenesis and thus have been used to identify patients at risk for cardiac death (Henneberger et al. 2005). The ST segment of the ECG is most sensitive to drugs or disease (DSI 2002), and ST depression is temporally associated with myocardial ischemia (Scheidt 1986). Ischemic episodes are negative prognostic indicators that point to an increase in the probability of future cardiac events, including myocardial infarction (Tzivoni et al. 1988). Trafficrelated PM has been linked with STsegment depression in elderly patients  and patients with coronary heart disease (Pekkanen et al. 2002). Kodavanti et al. (2000) performed a qualitative assessment of ECG signals derived from SH rats and found that ROFA caused an ST depression during the first two of three inhalation exposures (Kodavanti et al. 2000). Rats lack an equivalent human STsegment because of the fast rate at which their ven tricular myocytes repolarize (DSI 2002). Notwithstanding this difference, perturbations that result in STsegment changes in species with ST segments produce a similar shift in the corresponding portion of the ECG in rats (DSI 2002). In our study, we found no significant differences in ST or Twave areas among any of the exposure groups or strains, suggesting no evidence of ST depression or altered ventricu lar repolarization. The inability to clearly repro duce the findings of Kodavanti et al. (2000) may be related to subtle differences between the studies (duration or concentration). On the other hand, these findings suggest that such a change may not be consistent because of con founders affecting sensitivity to challenge at rest. An analysis of STsegment changes may be more appropriate when HR is elevated, as in exercise (as found in humans; Pekkanen et al. 2002;Mills et al. 2007), because oxygen demand is elevated, thus increasing the likeli hood of ischemic episodes, which in turn may be reflected in ST changes in the ECG (Scheidt 1986). Such tests may be adaptable to animal bioassays to better assess the effects of PM or other pollutants on cardiac ischemic changes or changes in repolarization reflected in the ECG.
The inhalation of fine PM has been linked to both pulmonary and systemic inflammation (Brook et al. 2003;Utell et al. 2002), includ ing increases in the circulating acutephase Episodes of cardiac dysfunction, including cardiac arrhythmias, may result from sys temic inflammation via altered coagulative processes, destabilization of arterial plaques, and/or exacerbation of atherosclerotic plaque formation (Brook et al. 2003). Elevated CRP paralleling ECG changes, including QT pro longation and decreased Twave amplitude, has also been reported in coronary artery dis ease patients exposed to ambient fine PM (Yue et al. 2007). Furthermore, systemic inflamma tion may actually trigger cardiac arrhythmia as evidenced by the finding that elevated serum CRP was predictive of the recurrence rate of atrial fibrillation, another type of cardiac arrhythmia (HatzinikolaouKotsakou et al. 2006). SH rats exposed to ROFA in the pres ent study had significantly higher pulmonary influx of neutrophils, markers of pulmonary edema and injury, including LDH, protein, and albumin, and higher serum CRP levels than did similarly exposed WKY rats with nor mal blood pressure; these pulmonary responses were quite similar to the findings of Kodavanti et al. (2000) after three exposures. The strong, straindependent inflammatory response in SH rats exposed to ROFA in this study may have contributed to the development of the nonconducted Pwave arrhythmias. The incidence of nonconducted Pwave arrhythmias increased only in SH rats exposed to ROFA, suggesting that hypertension con ditions the heart, creating a substrate that increases the susceptibility to PMinduced arrhythmia. Although we did not measure blood pressure in this study, SH rats have sig nificantly higher mean arterial pressure relative to their related background strain, the WKY rats (Yamori et al. 1979). Hypertension over time leads to hypertrophy of the left ventricle (American Heart Association 2008), a condi tion present in the SH rats of this study as evidenced by their higher hearttobody weight ratio compared with WKY rats. Several remod eling events take place in the hypertrophying heart that may heighten myocardial sensitivity to the effects of PM exposure, including a) pro longed QT intervals and action potential duration resulting from changes in calcium ion channel density, b) aberrant expression of proteins/currents specific to the sinoatrial node, suggesting that the ventricles can acquire automaticity, and c) an increase in the den sity of connexin 40 (a gapjunction protein involved in celltocell electrical coupling), which is usually confined to the atria and has greater conductance (Swynghedauw 1999). That SH rats undergo left ventricular hypertro phy is well documented (Yamori et al. 1979), suggesting that the SH rats in this study may have undergone some of these remodeling events during the course of disease progres sion that predisposed them to the develop ment of ROFAinduced nonconducted Pwave arrhythmias. The absence of any clear histo logic distinction between SH and WKY rat hearts in this study suggests that the differences may be at the ultrastructural and molecular levels. Nevertheless, these findings clearly point to a compromising effect of hypertension on the responsiveness to ROFA in SH rats.
In summary, compared with HRV and ECG morphology in this rat model of hyper tension, cardiac arrhythmias, particularly non conducted Pwaves, were the most sensitive indicators of the adverse cardiovascular effects of ROFA inhalation, suggesting that cardiac arrhythmias may have predictive value in the assessments of the effects of air pollution. This model may in turn be used to compare the capacity of other PM samples to elicit cardiac arrhythmias in the context of hypertension, potentially facilitating comparisons of the cardiovascular toxicity of PM from different air sheds. In addition, PMinduced cardiac arrhythmias found in this model may help elu cidate the mechanism of action. One additional mechanism that requires further study involves direct injury to cardiac myocytes resulting from the translocation of metals to the heart after inhalation. ROFA metals, including nickel and vanadium, leach into the blood from the lung and enter other organs, including the heart (Wallenborn et al. 2007), and thus may have had a direct role in the development of noncon ducted Pwave arrhythmias in this study. This study also demonstrated that normal animals have very minimal responses to very high acute PM exposure, with only a mild inflammatory response and no evidence of cardiac arrhythmia. Healthy humans or animals may thus be more resistant to the effects of air pollution compared with diseased cohorts, suggesting that the effects of air pollution should more appropriately be assessed in the context of preexisting disease. Factors that may influence the sensitivity of ECG end points that we did not address in this study include repeated or chronic exposures to ROFA, exposure to lower concentrations of ROFA, exposure to other PM samples, or the exposure of animals with other cardiovascular diseases. Taken together, the data suggest, how ever, that cardiac arrhythmia may be an early sensitive ECG indicator of the propensity for PM inhalation to modify cardiovascular func tion and trigger acute cardiac events.