Kim, Jee Young ScD; Prouty, Lacey A. MS; Fang, Shona C. ScD; Rodrigues, Ema G. PhD; Magari, Shannon R. ScD; Modest, Geoffrey A. MD; Christiani, David C. MD, MPH
The adverse health effects of fine particulate matter, or PM2.5 (particulate matter with an aerodynamic mass median diameter ≤2.5 μm), have been widely noted in the epidemiology and toxicology literature. Short- and long-term exposure to PM2.5 has been associated with various health effects, including, but not limited to, decreased lung function and increased respiratory symptoms1,2; altered heart rate and heart rate variability3,4; pulmonary and systemic inflammation5,6; cardiopulmonary hospitalizations and mortality7,8; and lung cancer mortality.4
Epidemiologic studies have observed associations between personal PM2.5 exposure and various biomarkers of oxidative DNA damage in different cohorts with a wide range of exposures,9 eg, university students,10 urban area bus drivers,11 and boilermakers exposed to high levels of residual oil fly ash.12 PM2.5 is a complex mixture of compounds, including transition metals, which are capable of producing reactive oxygen species (ROS) through the Fenton reaction,13 and polycyclic aromatic hydrocarbons, which are metabolically activated to o-quinones that can generate ROS through redox cycling.14 These ROS can cause oxidative damage to DNA, which can further lead to mutations that are associated with the initiation and progression of human cancers.15 Oxidative stress and associated oxidative damage also have been found to mediate vascular injury and inflammation in many cardiovascular diseases, including hypertension, hyperlipidemia, and diabetes.16
Increased susceptibility to various PM2.5 health effects has been observed in individuals with preexisting cardiovascular and respiratory diseases.17 Nevertheless, the susceptibility of these individuals specifically to PM2.5-induced oxidative DNA damage has not been examined previously. Given their diseased state, individuals with cardiovascular disease may be more sensitive to oxidative DNA damage associated with PM2.5 exposure.
In this study, we investigated the association between personal exposure to PM2.5 and oxidative DNA damage and repair, as indicated by urinary 8-hydroxy-2′-deoxyguanosine (8-OHdG), in non-hypertensive and hypertensive individuals. Previous studies have found urinary 8-OHdG to be a useful biomarker in assessing ROS-induced DNA damage in both the clinical and epidemiological setting.18–23 The purpose of this study was to examine whether individuals with hypertension were more susceptible to PM2.5 effects on oxidative DNA damage.
The study was approved by the Institutional Review Board of the Harvard School of Public Health (Boston, MA). Written informed consent was obtained from each study participant. The study cohort consisted of an inner-city population living in close proximity to a bus terminal (one of the busiest Massachusetts Bay Transportation Authority bus stops in the Boston area). The study population was recruited from a local health clinic. A list of the adult clients living in four zip codes (02,124,02,122,02,121,02,119) in Roxbury, MA and North Dorchester, MA was generated and cleaned to remove patients with duplicate, incomplete, or invalid contact information. Lists were separated by “healthy” and “compromised” health status by excluding and selecting, respectively, relevant International Classification of Diseases-9 (ICD-9) codes in the health center patient database (Diabetes; Chronic Bronchitis [491, 491.9]; Asthma; Emphysema [492, 492.8]; and Hypertension). Twenty-seven healthy individuals and 40 compromised individuals were contacted, of which 22 agreed to participate in the study. The individuals were monitored during the period of March to August 2004. Of the 22 monitored individuals, one individual was excluded from the analysis because she withdrew from the study before the first 12-hour monitoring period was completed, leaving 21 individuals with relevant data for inclusion in this analysis. Six of the 21 individuals were “healthy,” whereas 15 individuals were “compromised,” with one or additional disease diagnoses as stated above. Eighteen of the 21 individuals were monitored for the entire 36-hour period; two individuals were monitored for a 24-hour period; and one individual was monitored for a 12-hour period.
Each participant completed a modified American Thoracic Society questionnaire.24 The self-administered questionnaires were used to obtain information on personal demographic information, smoking history, medical history, including cardiopulmonary symptoms and diseases, and medication use. In cases where the participant was not fluent in English, the questionnaire was administered by a translator in Spanish or Portuguese. Data from questionnaires were used to identify individuals with hypertension, inflammatory lung diseases (ie, chronic bronchitis, asthma, and emphysema), and diabetes.
PM2.5 Sample Collection
The TSI Incorporated model AM510 SIDEPAK personal aerosol monitor (TSI Incorporated, Shoreview, MN), a light scattering laser photometer, with a 2.5-μm cutsize impactor was used to assess personal exposure to PM2.5 during a 36-hour period. The air sample was drawn through a 4-feet long Tygon tube into the impaction inlet at a flow rate of 1.7 L/min. PM2.5 concentrations were logged every minute, an average of a series of measurements taken every 10 seconds. The minute-to-minute PM2.5 concentrations were averaged during an approximately 12-hour interval, which coincided with the time when spot urine samples were collected.
The SIDEPAK was factory calibrated to the respirable fraction of the International Organization for Standardization (ISO) 12,103-1, A1 Arizona test dust.25 Arizona test dust is a polydispersed test aerosol commonly used because it is representative of a wide variety of ambient aerosols. Previous studies have shown that measurements from similar light scattering laser photometers overestimate PM2.5 concentrations by a factor of two to three, when compared with measurements taken using a gravimetric method.26–28 To examine the possible overestimation of PM2.5 concentrations when using the SIDEPAK, comparisons were made with PM2.5 samples collected on a 37-mm polytetrafluoroethylene membrane filter (Gelman Laboratories, Ann Arbor, MI) using KTL cyclones (GK2.05SH, BGI Incorporated, Waltham, MA) with a 50% aerodynamic diameter cut point of 2.5 μm. The cyclone was collocated with the SIDEPAK and PM2.5 was collected during the entire monitoring period for all subjects, resulting in an approximately 36-hour average concentration. PM2.5 concentrations from the SIDEPAK and the cyclone gravimetric method were compared using analyses described previously.29 Eighteen of the 21 collocated cyclone samples provided data useful to this analysis. For the 18 samples, the geometric mean of the PM2.5 concentrations obtained using the gravimetric method was 0.026 mg/m3 (geometric standard deviation, GSD, 2.032) whereas the corresponding geometric mean using the Sidepak method was 0.059 mg/m3 (GSD 2.360). Among the non-smokers, the geometric means for the gravimetric and Sidepak concentrations were 0.020 mg/m3 (GSD 1.410) and 0.042 mg/m3 (GSD 1.379), respectively. Among the smokers, the geometric means of the gravimetric and Sidepak concentrations were 0.094 mg/m3 (GSD 1.922) and 0.317 mg/m3 (GSD 2.020), respectively. The gravimetric PM2.5 concentrations were regressed on SIDEPAK PM2.5 concentrations and an interaction term was included to study the effect of smoking status on the association between gravimetric and SIDEPAK PM2.5 concentrations. Comparisons indicated that the SIDEPAK overestimated PM2.5 concentrations and the degree of overestimation depended on the smoking status of the individual. To correct for the overestimation, a factor of 0.45 for smokers and 0.29 for nonsmokers was applied to the SIDEPAK PM2.5 concentrations.
Urine Collection and Analysis for 8-OHdG
Spot urine samples were collected at baseline and at approximately 12-hour intervals during the 36-hour monitoring period, resulting in the collection of up to four samples for each individual. The baseline and 24-hour spot urine samples were generally collected between 7:00 am and 9:00 am, whereas the 12-hour and 36-hour spot urine samples were collected between 7:00 pm and 9:00 pm Spot urine samples were collected instead of 12-hour pooled urine samples because of concerns of subject compliance. A study by Miwa et al30 observed significant correlations between 8-OHdG levels in a morning spot urine and in a 24-hour pooled urine, and concluded that a morning spot urine sample can be used for the measurement of 8-OHdG instead of 24-hour sampling, which can be cumbersome and inconvenient for the subjects.
After samples were collected in sterile 120-mL urine collection cups, they were aliquoted into 15-mL polypropylene tubes. One 5 mL aliquot was sent to Path Lab Incorporated (Portsmouth, NH) for creatinine analysis. The remaining specimen was stored at –20°C until further analysis.
Urine analysis for 8-OHdG was performed by Genox Corporation (Baltimore, MD). Urinary 8-OHdG was determined using a competitive enzyme-linked immunosorbent assay.18,31 Briefly, 50 μL of urine samples and standards were added to precoated 8-OHdG protein conjugate microtiter plates, followed by 50 μL of the primary antibody, anti-8-OHdG monoclonal antibody solution, and incubated for 1 hour at 37°C. The plates were washed and the enzyme-labeled horseradish peroxidase–conjugated secondary antibody (100 μL) was applied for 1 hour at 37°C. After washing, 100 μL of the chromatic substrate, (3,3′,5,5′) tetramethylbenzidene, was added to the plate and allowed to react at room temperature for 15 minutes. The intensity of color produced for each sample was measured at an optical density of 490 nm. Pooled urine samples from several healthy adults were used as the quality control samples. For each standard 96-well microplate, six to nine quality control samples were randomly placed among the unknown samples. The measured quality control values were averaged and compared with previously established values. Acceptable quality control values were defined as mean ± 2 SDs. The limit of detection for 8-OHdG was 0.64 ng/mL. Seven of 78 available samples (9%) had concentrations below the limit of detection and were given the value of 0.32 ng/mL (½ the limit of detection). For each sample, either a duplicate or triplicate measurement was performed. The mean, SD, and coefficient of variation (percent) were calculated. Any sample with a coefficient of variation equal to or greater than 20% was retested. Urinary 8-OHdG concentrations are dependent on kidney function. Creatinine levels are a standard indicator of kidney function; therefore, we adjusted the urinary 8-OHdG levels for urinary creatinine levels in all our analyses.
Statistical analyses were performed using SAS version 9.1 (SAS Institute Incorporated, Cary, NC). Urinary creatinine-adjusted 8-OHdG concentrations and PM2.5 concentrations with the correction factors applied were used in all analyses. Baseline urinary 8-OHdG concentrations were excluded from the analysis as PM2.5 concentration data were not available before baseline urine sample collection. As multiple samples were collected from each individual, linear mixed models were used to account for the correlation of repeated measurements within individuals.
These models were used first to examine whether urinary 8-OHdG and PM2.5 concentrations differed by age, gender, smoking status, and hypertension status. The 12-hour average PM2.5 concentrations were found to be positively skewed (skewness of 4.3); as PM2.5 concentrations were an outcome variable in these analyses, it was necessary to log-transform the PM2.5 data to improve skewness. Then, multivariate regression models were constructed to investigate the association between urinary 8-OHdG concentrations and average PM2.5 concentrations over the prior 12-hour interval. In these models, the PM2.5 concentration was a predictor, not an outcome; therefore, the non-transformed PM2.5 concentrations were used. Only a small percentage of urinary 8-OHdG and PM2.5 exposure data were missing (2% and 7%, respectively) and the missing data generally appeared random; therefore, all analyses were conducted using the available data only. A compound symmetry correlation structure was used because it resulted in the best Akaike’s Information Criterion compared with models with other covariance structures, including a first-order autoregressive structure.32 Restricted maximum likelihood was used to estimate the covariance parameters. To investigate effect modification of the association between urinary 8-OHdG and PM2.5 by hypertension status, an interaction term between PM2.5 and an indicator variable for hypertension was included in the model. Potential confounding by age, gender, and tobacco smoking (dichotomized as current smoker or non-current smoker) also was examined. Because of possible circadian variation of urinary 8-OHdG levels, all models also were adjusted for time of day (dichotomized as day or evening).33 As concerns remain regarding residual confounding by tobacco smoking and circadian variation even after adjustment using a dichotomized variable in the statistical models, data were also reanalyzed after restricting to morning urinary 8-OHdG samples from nonsmokers. In addition, as study subjects were monitored over two seasons, from March to August, a sensitivity analysis adjusting for seasonal variation was also conducted.
Study Population Demographics
The study population consisted of 21 individuals, five men and 16 women, with a mean age of 45 years (range, 21 to 70) (Table 1). Twelve individuals had physician-diagnosed hypertension (≥140 mm Hg systolic pressure or ≤90 mm Hg diastolic pressure). Eleven of the 12 individuals indicated that they were on antihypertensive medications; one individual did not provide that information. The antihypertensive medications currently taken by the individuals included angiotensin-converting enzyme (ACE) inhibitors (lisinopril, enalapril), angiotensin II receptor blockers (losartan), calcium channel blockers (amlodipine), diuretics (hydrochlorothiazide, triamterene), and beta-adrenoceptor antagonists (atenolol, metoprolol).
Individuals with hypertension were significantly older than those without hypertension (mean age 54 years vs 33 years, respectively [P = 0.001]). In addition, a greater percentage of individuals with hypertension compared with those without hypertension were current smokers (42% vs 11%, respectively [P = 0.14]). Two of the individuals with hypertension also had diabetes and five also had inflammatory lung disease (asthma or chronic bronchitis). Three of the nine individuals without hypertension had asthma; none had diabetes or chronic bronchitis.
PM2.5 Exposure Assessment
The personal PM2.5 exposure data are summarized in Table 2. Fifty-five 12-hour-average PM2.5 measurements were collected from the 21 participants. The number of consecutive 12-hour intervals for which an individual had PM2.5 data ranged from 1 to 3. The coefficient of variance (COV) of the PM2.5 concentrations was calculated for each individual subject. The COVs varied widely, from a minimum value of 5% to 111% across the subjects, with a median value of 55%. These results suggest that though autocorrelation may be of concern for a small number of subjects (only four subjects had a COV of less than 25%), there appears to be considerable variation in the PM2.5 data within an individual.
The 12-hour-average PM2.5 concentrations were found to be positively skewed. Linear mixed models indicated that the log-transformed 12-hour-average PM2.5 concentrations were not significantly different by hypertension status (P = 0.24). Additional analyses indicated that the PM2.5 exposures significantly differed by gender (P = 0.04) and marginally differed by smoking status (P = 0.06), but not by age (P = 0.11). Men and current smokers were found to have higher 12-hour-average PM2.5 concentrations compared with women and non-smokers, respectively.
Urinary 8-OHdG Analysis
A total of 78 spot urine samples were collected from the 21 individuals, with each individual providing two to four samples. The baseline urine samples (n = 20, one subject did not provide a baseline sample) were excluded from the analysis as PM2.5 concentration data were not available before baseline urine sample collection. The Shapiro-Wilk test indicated that the urinary 8-OHdG concentrations were normally distributed (P = 0.17). The distribution of urinary 8-OHdG levels for the remaining samples (n = 58) is presented in Table 3. In crude analyses, urinary 8-OHdG concentrations were not significantly different in individuals with hypertension compared to those without hypertension (P = 0.30). Urinary 8-OHdG concentrations did not significantly differ by gender (P = 0.39), smoking status (P = 0.79), or age (P = 0.60). Examination of the urinary 8-OHdG concentrations by time of day indicated that the levels were higher in the evening (7:00 pm to 9:00 pm) compared with the morning (7:00 am to 9:00 am) (P = 0.01).
Regression Analysis of Urinary 8-OHdG and PM2.5 Exposure
The 12-hour-average PM2.5 concentrations were positively skewed largely because of two outlying data (196 μg/m3 and 262 μg/m3). By excluding the outlying data, the positive skew reduced to 1.1. Regression models, excluding the two outlying data points, were conducted as sensitivity analyses.
Linear mixed regression models were constructed to investigate the association between urinary 8-OHdG and PM2.5 (Table 4). In models individually adjusting for age, gender, smoking status, and time of day, all but age were found to affect the PM2.5 effect estimate. Nevertheless, as previous studies have indicated that the age may influence urinary 8-OHdG levels,30 all four variables were included in the adjusted model. In the adjusted model, an inverse association between urinary 8-OHdG and PM2.5 was observed. In the model excluding the two outlying data points, similar results were observed; however, the effect estimate was not statistically significant. An additional sensitivity analysis was conducted to consider possible influences of seasonal variation. In models adjusted for seasonal variation, similar results were once again observed, with only slight changes in the effect estimates.
Before performing statistical analyses to investigate effect modification of the association between urinary 8-OHdG levels and PM2.5 by hypertension status, crude scatter plots stratified by hypertension status were examined (Fig. 1). The scatter plots suggested that the relationship between urinary 8-OHdG and PM2.5 differed by hypertension status. Regression models including an interaction term between PM2.5 and hypertension were used to formally investigate modification by hypertensive status. In the adjusted model, the interaction term between PM2.5 and hypertension status was marginally significant when using all available data (P = 0.06). The interaction term indicated that each 10 μg/m3 increase in PM2.5 was associated with a 1.71 μg/g creatinine (95% CI: −0.10 to 3.52) greater increase in urinary 8-OHdG in individuals without hypertension compared to those with hypertension. Similar results were observed in models excluding the outlying data points (2.44 μg/g creatinine [95% CI: 0.34 to 4.54] greater increase in urinary 8-OHdG with each 10 μg/m3 increase in PM2.5). Effect estimates for those with and without hypertension were calculated from the adjusted model with the interaction term for hypertension status. The effect estimate for PM2.5 was positive, though not significant for individuals without hypertension, while an inverse association was observed in those with hypertension.
Model assumptions were tested by examining a plot of the residuals against the predicted values and the normality of the residuals. In models using all available data, as well as those excluding outlying data, model assumptions were found to be met. The regression residuals were normally distributed (Shapiro-Wilk test P > 0.90), and the plots of the residuals against the predicted values indicated good model fit.
Because of concerns regarding residual confounding by tobacco smoking and circadian variation, a restricted analysis using only morning urinary 8-OHdG samples from non-smoking individuals was conducted (Table 4). For this analysis, a total of 15 spot urine samples were available from 15 non-smoking individuals. The same trend in results was observed in this restricted analysis—a positive association between PM2.5 and urinary 8-OHdG levels in individuals without hypertension, while an inverse association was observed in individuals with hypertension. Both effect estimates were not statistically significant.
Discussion and Conclusions
Exposure to PM2.5 has been linked to the induction of oxidative DNA damage in toxicologic studies.34 Particles may contain soluble transition metals, including iron, copper, chromium, and vanadium, which can generate ROS through Fenton-type reactions. Polycyclic aromatic hydrocarbons absorbed onto the surface of carbonaceous particles also may be metabolically activated to quinone radicals that may undergo redox-cycling to produce ROS. Oxidative DNA damage resulting from the ROS may be implicated in cancer risk and possibly serve as a marker for oxidative stress relevant for other health conditions, including cardiovascular diseases, caused by PM2.5 or other air pollutants.34
Recent interest has focused on 8-OHdG, measured either in white blood cells or urine, as an indicator of oxidative DNA damage and repair.35 8-OHdG is formed from a hydroxyl radical attack at the C-8 position of deoxyguanosine in DNA.36 The repair mechanism after 8-OHdG is incorporated into DNA involves base and nucleotide excision with DNA-specific nucleosides excreted into urine. Pilger and Rudiger37 noted that the 8-OHdG is the most studied oxidative DNA lesion because of its significance as an endogenous mutagen and its likely role in the process of carcinogenesis.
In this study, the association between personal PM2.5 exposure and oxidative DNA damage, as indicated by urinary 8-OHdG concentrations, was examined in a cohort of inner-city adults living in close proximity to a bus terminal. In the analysis including all subjects, PM2.5 concentration was associated with a decrease in urinary 8-OHdG after adjusting for age, gender, smoking status, and time of day. However, results from the analysis including an interaction term between PM2.5 and hypertension status suggested that the relationship between PM2.5 exposure and urinary 8-OHdG concentrations is modified by hypertension status. PM2.5 concentration was associated with a decrease in urinary 8-OHdG in individuals with hypertension compared to an increase in those without hypertension after adjusting for age, gender, smoking status, and time of day.
The positive, though not statistically significant, association between PM2.5 exposure and urinary 8-OHdG observed in those without hypertension is consistent with results from previous epidemiologic studies. In a recent study, Abder-Rahman and Nusair38 observed that the urinary 8-OHdG concentrations in a variety of occupationally and environmentally exposed populations were higher compared with the control population, suggesting that environmental toxins can induce DNA damage. In a group of non-smoking bus drivers from the greater Copenhagen area, higher urinary 8-oxo-2′-deoxyguanosine (8-oxodG) levels, another biomarker of oxidative DNA damage were observed in bus drivers from the urban area compared with those from rural areas, suggesting that exposure to ambient air pollution was associated with oxidative damage to DNA.11 Sørensen et al10 examined the association between personal PM2.5 exposures and biomarkers of oxidative DNA damage in students living in central Copenhagen, Denmark, and observed that each 10 μg/m3 increase in PM2.5 exposure was associated with an 11% increase in lymphocyte 8- oxodG levels (P = 0.007), though not with urinary 8-oxodG concentrations.
Oxidative stress may constitute a major pathogenic factor in the development of hypertension.39 Experimental studies by de Champlain et al39 indicated that ROS, mainly through the production of superoxide anion, could cause important alterations in the cellular signal transduction systems, ultimately leading to vasoconstriction. Further, hypertension is also associated with an impairment of endogenous antioxidant mechanisms.40 Nevertheless, antihypertensive medications have been shown to have antioxidant activities, including calcium channel blockers (amlodipine), ACE inhibitors (enalapril and lisinopril), and angiotensin II receptor blockers (losartan).41–44 In this study, eleven of the 12 hypertensives indicated that they were taking antihypertensive medication and eight of them were on treatments with some antioxidant capability. The antioxidant activity present in some antihypertensive medications may play a role in reducing the effects of PM2.5 on oxidative stress; however, it does not fully explain the inverse association observed between PM2.5 exposure and urinary 8-OHdG levels among individuals with hypertension. Further studies may consider measuring antioxidant activity to examine the effects of hypertensive medication use on the relationship between PM2.5 and 8-OHdG levels.
Another possible hypothesis for the observed inverse association in hypertensive individuals may be that PM2.5 exposure does not decrease oxidative damage, but rather reduces capacity to repair DNA damage. DNA damage that is not repaired tends to promote mutagenesis. As mentioned earlier, urinary 8-OHdG is a specific DNA repair product in the urine and although it is often considered a marker of oxidative damage, perhaps, it also reflects repair of that damage. Gackowski et al45 examined 8-OHdG levels in leukocytes and urine in patients with lung cancer and two control groups. They observed that the level of 8-OHdG in DNA isolated from leukocytes of cancer patients was significantly higher than that in DNA from the two control groups. Nevertheless, urinary 8-OHdG levels were similar in the cancer patients and control group. The authors concluded that the higher rate of oxidative damage in the cellular DNA of lung cancer patients was a result of a deficiency of repair mechanisms in this group. Similarly, the inverse association between PM2.5 concentrations and urinary 8-OHdG levels observed in the hypertensive individuals in our study may suggest a reduced capacity for DNA repair with increased PM2.5 exposure in this potentially susceptible population.
We acknowledge limitations to our study. In this panel study, we investigated the association between PM2.5 concentrations and urinary 8-OHdG levels in 21 individuals, 12 hypertensives, and nine non-hypertensives, monitored repeatedly during a 36-hour period. Although the small sample size is a limitation, the use of repeated measurements increased the power of the study, allowing us to examine this relationship.
We were unable to examine the modifying effect of PM2.5 on oxidative DNA damage in hypertensives untreated by medication compared with non-hypertensives. Both hypertensive and non-hypertensive individuals were recruited from a health center and selected using ICD-9 codes. Generally, individuals diagnosed with hypertension at the health center were prescribed antihypertensive medications for treatment. For ethical reasons, we could not request that hypertensive individuals refrain from taking their medications during the monitoring period. Because of the variety of medications taken by the hypertensive individuals (eg, ACE inhibitors, angiotensin II receptor blockers, calcium channel blockers, diuretics, and beta-adrenoceptor antagonists) and the small sample size of our study, we could not examine the effects of the different types of medications in our analyses.
Another limitation of this study is possible confounding of the association between PM2.5 and urinary 8-OHdG by copollutants, though the use of personal PM2.5 measurements rather than ambient PM2.5 concentrations from a community monitor likely reduced this risk. One study by Sarnat et al46 observed that personal PM2.5 exposures were not highly correlated with personal measurements of copollutants, particularly ozone, nitrogen dioxide, and sulfur dioxide. It should be noted, however, that in another study by Sarnat et al,47 personal exposure to PM2.5 was found to be associated with personal exposures to ozone and nitrogen dioxide in the summer.
In this study, associations between total exposure to PM2.5 and urinary 8-OHdG were examined. We were unable to distinguish between health effects associated with PM2.5 from ambient and non-ambient origins. Of particular concern in our study is PM2.5 exposure from tobacco smoking. Tobacco smoke has a particle size range of <0.1 μm to just more than 1 μm in particle diameter,48 thus, it has an impact on total PM2.5 exposure, as was observed in our study. A recent study by Ebelt et al49 examined cardiovascular and respiratory health effects after separating total personal particle exposure into that of ambient and non-ambient origins. The strongest associations were observed with personal exposure to PM2.5 of ambient origin compared with ambient PM2.5 concentrations, total personal exposure to PM2.5, or personal exposure to PM2.5 of non-ambient origin. We were unable to examine the associations between ambient PM2.5 and oxidative DNA damage because we did not collect the necessary information (time-activity data, measurements of pm infiltration) to estimate the ambient component of PM2.5 from our personal exposure samples.
Another limitation of this study is the possibility of confounding from age, gender, tobacco smoking, and circadian variation. Although these variables were controlled for in the multivariate regression models, the possibility of residual confounding remains, particularly for tobacco smoking and circadian variation as the biologically relevant surrogate measures for these variables in assessing the relationship between PM2.5 and urinary 8-OHdG were difficult to determine and adjusted for in the models. Concerns of residual confounding led us to conduct an analysis restricting the data to morning urinary 8-OHdG samples from non-smoking individuals. In this restricted analysis, the same trend was observed, though the 95% CIs were wider because of reduced power.
In conclusion, this study shows that the association between personal exposure to PM2.5 and urinary 8-OHdG, a biomarker of oxidative DNA damage and repair, was modified by hypertension status. A positive, but not statistically significant, exposure-response relationship was observed between urinary 8-OHdG concentrations and PM2.5 exposure in individuals without hypertension. In contrast, a statistically significant decrease in urinary 8-OHdG levels in relation to PM2.5 exposure was observed in those with hypertension. Because it is highly unlikely that PM2.5 has healthful effects on hypertensive individuals, these counterintuitive results require confirmation in further studies and exploration of possible mechanisms.
The authors thank Li Su, Meredith Jones, and Kim Weatherbee for their assistance with the field work and data entry, and Katio Depina for his assistance with translations. The authors thank Dr Daniel Guerra of the Duke University Medical Center, Division of Cardiology for his technical expertise. The authors also thank the staff of the Upham’s Corner Health Center (Dorchester, MA).
This study was supported by U.S. EPA STAR grant RD-83083801 and NIH grant ES00002.
J.Y.K., S.R.M, and L.A.P. were supported by T32 ES007069 from NIEHS, and S.C.F. was supported by T42 OH008416 from NIOSH.
1. Schwartz J, Neas LM. Fine particles are more strongly associated than coarse particles with acute respiratory health effects in school children. Epidemiology. 2000;11:6–10.
2. Gauderman WJ, Avol E, Gilliland F, et al. The effect of air pollution on lung development from 10 to 18 years of age. N Engl J Med. 2004;351:1057–1067.
3. Liao D, Creason J, Shy C, Williams R, Watts R, Zweidinger R. Daily variation of particulate air pollution and poor cardiac autonomic control in the elderly. Environ Health Perspect. 1999;107:521–525.
4. Pope CA III, Hansen ML, Long RW, et al. Ambient particulate air pollution, heart rate variability, and blood markers of inflammation in a panel of elderly subjects. Environ Health Perspect. 2004;112:339–345.
5. Adamkiewicz G, Ebelt S, Syring M, et al. Association between air pollution exposure and exhaled nitric oxide in an elderly population. Thorax. 2004;59:204–209.
6. Dubowsky SD, Suh H, Schwartz J, Coull BA, Gold DR. Diabetes, obesity, and hypertension may enhance associations between air pollution and markers of systemic inflammation. Environ Health Perspect. 2006;114:992–998.
7. Dominici F, Peng RD, Bell ML, et al. Fine particulate air pollution and hospital admission for cardiovascular and respiratory diseases. JAMA. 2006;295:1127–1134.
8. Ostro B, Broadwin R, Green S, Feng WY, Lipsett M. Fine particulate air pollution and mortality in nine California counties: results from CALFINE. Environ Health Perspect. 2006;114:29–33.
9. Valavanidis A, Vlachogianni T, Fiotakis C. 8-hydroxy-2′-deoxyguanosine(8-HOdG): a critical biomarker of oxidative stress and carcinogenesis. J Environ Sci Health C Environ Carcinog Ecotoxicol Rev. 2009;27:120–139.
10. Sørensen M, Autrup H, Hertel O, Wallin H, Knudsen LE, Loft S. Personal exposure to PM2.5 and biomarkers of DNA damage. Cancer Epidemiol Biomarkers Prev. 2003;12:191–196.
11. Loft S, Poulsen HE, Vistisen K, Knudsen LE. Increased urinary excretion of 8-oxo-2′-deoxyguanosine, a biomarker of oxidative DNA damage, in urban bus drivers. Mutat Res. 1999;441:11–19.
12. Kim JY, Mukherjee S, Ngo LC, Christiani DC. Urinary 8-hydroxy-2′-deoxyguanosine as a biomarker of oxidative DNA damage in workers exposed to fine particulates. Environ Health Perspect. 2004;112:666–671.
13. Pritchard RJ, Ghio AJ, Lehmann JR, et al. Oxidant generation and lung injury after particulate air pollution exposure increase with the concentrations of associated metals. Inhal Toxicol. 1996;8:457–477.
14. Bolton JL, Trush MA, Penning TM, Dryhurst G, Monks TJ. Role of quinones in toxicology. Chem Res Toxicol. 2000;13:135–160.
15. Feig DI, Reid TM, Loeb LA. Reactive oxygen species in tumorigenesis. Cancer Res. 1994;54:1890s–1894s.
16. Touyz RM, Schiffrin EL. Reactive oxygen species in vascular biology: implications in hypertension. Histochem Cell Biol. 2004;122:339–352.
17. US EPA. Air Quality Criteria for Particulate Matter. Washington, DC: U.S. Environmental Protection Agency, Office of Research and Development, National Center for Environmental Assessment; 2004.
18. Erhola M, Toyokuni S, Okada K, et al. Biomarker evidence of DNA oxidation in lung cancer patients: association of urinary 8-hydroxy-2′-deoxyguanosine excretion with radiotherapy, chemotherapy, and response to treatment. FEBS Lett. 1997;409:287–291.
19. Honda M, Yamada Y, Tomonaga M, Ichinose H, Kamihira S. Correlation of urinary 8-hydroxy-2′-deoxyguanosine (8-OHdG), a biomarker of oxidative DNA damage, and clinical features of hematological disorders: a pilot study. Leuk Res. 2000;24:461–468.
20. Lagorio S, Tagesson C, Forastiere F, Iavarone I, Axelson O, Carere A. Exposure to benzene and urinary concentrations of 8-hydroxydeoxyguanosine, a biological marker of oxidative damage to DNA. Occup Environ Med. 1994;51:739–743.
21. Pilger A, Germadnik D, Schaffer A, et al. 8-Hydroxydeoxyguanosine in leukocyte DNA and urine of quartz-exposed workers and patients with silicosis. Int Arch Occup Environ Health. 2000;73:305–310.
22. Tagesson C, Chabiuk D, Axelson O, Baranski B, Palus J, Wyszynska K. Increased urinary excretion of the oxidative DNA adduct, 8-hydroxydeoxyguanosine, as a possible early indicator of occupational cancer hazards in the asbestos, rubber, and azo-dye industries. Pol J Occup Med Environ Health. 1993;6:357–368.
23. Toraason M, Hayden C, Marlow D, et al. DNA strand breaks, oxidative damage, and 1-OH pyrene in roofers with coal-tar pitch dust and/or asphalt fume exposure. Int Arch Occup Environ Health. 2001;74:396–404.
24. Ferris BG. Epidemiology standardization project (American Thoracic Society). Am Rev Respir Dis. 1978;118(6 Pt 2):1–120.
25. International Organization for Standardization (ISO). Road Vehicles—Test Dust for Filter Evaluation—Part 1: Arizona Test Dust (ISO 12103–1) [Standard]. Geneva: ISO; 1997.
26. Chang LT, Suh HH, Wolfson JM, et al. Laboratory and field evaluation of measurement methods for one-hour exposures to O3, PM2.5, and CO. J Air Waste Manage Assoc. 2001;51:1414–1422.
27. Ramachandran G, Adgate JL, Hill N, Sexton K, Pratt GC, Bock D. Comparison of short-term variations (15-minute averages) in outdoor and indoor PM2.5 concentrations. J Air Waste Manage Assoc. 2000;50:1157–1166.
28. Yanosky, JD, Williams PL, MacIntosh DL. A comparison of two direct-reading aerosol monitors with the federal reference method for PM2.5 in indoor air. Atmos Environ. 2002;36:107–113.
29. Kim JY, Magari SR, Herrick RF, Smith TJ, Christiani DC. Comparison of fine particle measurements from a direct-reading instrument and a gravimetric sampling method. J Occup Environ Hyg. 2004;1:707–715.
30. Miwa M, Matsumaru H, Akimoto Y, Naito S, Ochi H. Quantitative determination of urinary 8-hydroxy-2′-deoxyguanosine level in healthy Japanese volunteers. Biofactors. 2004;22:249–253.
31. Leinonen J, Lehtimaki T, Toyokuni S, et al. New biomarker evidence of oxidative DNA damage in patients with non-insulin-dependent diabetes mellitus. FEBS Lett. 1997;417:150–152.
32. Verbeke G, Molenberghs G. Linear Mixed Models in Practice: A SAS-Oriented Approach. New York, NY: Springer; 1997.
33. Kanabrocki EL, Murray D, Hermida RC, et al. Circadian variation in oxidative stress markers in healthy and type II diabetic men. Chronobiol Int. 2002;19:423–439.
34. Risom L, Moller P, Loft S. Oxidative stress-induced DNA damage by particulate air pollution. Mutat Res. 2005;592:119–137.
35. Valavanidis A, Fiotakis K, Vlachogianni T. Airborne particulate matter and human health: toxicological assessment and importance of size and composition of particles for oxidative damage and carcinogenic mechanisms. J Environ Sci Health C Environ Carcinog Ecotoxicol Rev. 2008;26:339–362.
36. Kasai H, Crain PF, Kuchino Y, Nishimura S, Ootsuyama A, Tanooka H. Formation of 8-hydroxyguanine moiety in cellular DNA by agents producing oxygen radicals and evidence for its repair. Carcinogenesis. 1986;7:1849–1851.
37. Pilger A, Rudiger HW. 8-Hydroxy-2′-deoxyguanosine as a marker of oxidative DNA damage related to occupational and environmental exposures. Int Arch Occup Environ Health. 2006;80:1–15.
38. Abder-Rahman HA, Nusair S. 8- Hydroxy-2′-deoxyguanosine (8-OHdG) as a short-term predictor of regional and occupational health problems. J UOEH. 2007;29:247–258.
39. de Champlain J, Wu R, Girouard H, et al. Oxidative stress in hypertension. Clin Exp Hypertens. 2004;26:593–601.
40. Lassegue B, Griendling KK. Reactive oxygen species in hypertension: an update. Am J Hypertens. 2004;17:852–860.
41. Djordjevic VB, Pavlovic D, Pejovic M, Cvetkovic T, Lecic N, Deljanin-Ilic M. Changes of lipid peroxides and antioxidative factors levels in blood of patients treated with ACE inhibitors. Clin Nephrol. 1997;47:243–247.
42. Fiordaliso F, Cuccovillo I, Bianchi R, et al. Cardiovascular oxidative stress is reduced by an ACE inhibitor in a rat model of streptozotocin-induced diabetes. Life Sci. 2006;79:121–129.
43. Hornig B, Landmesser U, Kohler C, et al. Comparative effect of ace inhibition and angiotensin II type 1 receptor antagonism on bioavailability of nitric oxide in patients with coronary artery disease: role of superoxide dismutase. Circulation. 2001;103:799–805.
44. Rosenkranz AC, Lob H, Breitenbach T, Berkels R, Roesen R. Endothelial antioxidant actions of dihydropyridines and angiotensin converting enzyme inhibitors. Eur J Pharmacol. 2006;529:55–62.
45. Gackowski D, Speina E, Zielinska M, et al. Products of oxidative DNA damage and repair as possible biomarkers of susceptibility to lung cancer. Cancer Res. 2003;63:4899–4902.
46. Sarnat JA, Koutrakis P, Suh HH. Assessing the relationship between personal particulate and gaseous exposures of senior citizens living in Baltimore, MD. J Air Waste Manag Assoc. 2000;50:1184–1198.
47. Sarnat JA, Brown KW, Schwartz J, Coull BA, Koutrakis P. Ambient gas concentrations and personal particulate matter exposures: implications for studying the health effects of particles. Epidemiology. 2005;16:385–395.
48. Hinds, WC. Aerosol Technology: Properties, Behavior, and Measurement of Airborne Particles. 2nd ed. New York, NY: John Wiley & Sons, Inc; 1999.
49. Ebelt ST, Wilson WE, Brauer M. Exposure to ambient and nonambient components of particulate matter: a comparison of health effects. Epidemiology. 2005;16:396–405.