Grantee Research Project Results

The main objective of this project is to carry out controlled human exposures to both fine and coarse ambient PM size fractions and examine acute changes in cardiovascular and respiratory outcomes.  We hope to better understand how specific particle characteristics (size, composition, sources) affect health responses.  Exposures are 2 hours in duration and follow a randomized block design with each subject having four exposures. The exposures are to concentrated ambient particles (CAPs) using our newly constructed controlled particle exposure facility, in Toronto, ON, Canada.  Exposures include fine CAPs, two coarse CAPs and filtered air (FA) assigned in random order and with a minimum two-week wash-out period between. In the design phase of the project it was agreed to focus on coarse and fine CAPs and not to include ultrafine CAPs exposures. The target and maximum levels are as follows: fine CAPs (target 250 µg/m³; max 500 µg/m³); and coarse CAPs (target 200 µg/m³; max 400 µg/m³).  Biologic outcomes are measured before, during and following exposure. Cardiovascular outcomes include: i) vascular dysfunction (brachial artery diameter and reactivity); ii) cardiac output; iii) blood pressure; iv) markers of systemic inflammation (CBCs and blood IL-6, CRP & endothelins); and v) markers of oxidative stress in blood and urine. Respiratory outcomes include: i) pulmonary function (spirometry); ii) respiratory inflammation (induced sputum); iii) respiratory and nasal symptomotology; and iv) nasal inflammation (nasal smears in a subset of allergic subjects).  In addition, we will test for susceptibility genes (DNA methylation) of PM-induced oxidative stress using blood samples. During exposure, end-tidal CO₂ (nasal prongs) and beat-to-beat arterial BP are measured continuously for 10-min periods every 30 min using a Finometer finger cuff (pulse pressure) that includes calculated determinations of cardiac output, stroke volume and systemic vascular resistance. Mortara Holter (ECG) monitors are worn by the subjects over 24 hours, as a measure of cardiac autonomic dysfunction (HRV analyses). Exposures are characterized using continuous measures of particle mass and black carbon as well as integrated measurements of particle mass, sulfate, nitrate, ammonium, trace elements, organic and elemental carbon and biological material including airborne endotoxin and markers of fungi (β-glucan). On-site daily measures include vehicle counts, meteorological data, TEOM PM_2.5, gaseous criteria air pollutants, as well as pollen characterization (GRIPST-2000 pollen sampler) and fungal spores/pollen (Burkard sampler). Daily stationary central site monitoring data (gaseous and PM criteria pollutants) are obtained to statistically adjust for potential affects on baseline pre-exposure data.

 

Controlled Human Exposures: The first human exposures began late in November 2007.  As of July 31, 2009 a total of 77 exposures have been completed.  This includes 32 coarse CAPs and 16 fine CAPs, the rest being filtered air.  A total of 24 subjects have been enrolled of which 3 did not qualify for the exposure phase of the study.  Three subjects dropped out after completing 1 or 2 exposures. Each exposure involves a 2-day protocol and we attempt to have 2 exposures each week, depending on subject availability and rescheduling due to upper respiratory tract infections (must be symptom free for 3 weeks), cancellations and holidays. Fourteen subjects have completed the 4 exposures (FA, fine CAP and 2 coarse CAPs). An additional 5^th (control) exposure of HEPA-filtered medical air was added November 13, 2008, as described in detail below. Ten filtered medical air exposures have been carried out. Subjects continue to be recruited and exposures carried out.

 

Addition of HEPA-Filtered Medical Air as a Control Exposure: In all of our controlled human exposure studies we have included a filtered air (FA) control/sham exposure, as do other researchers who carry out controlled human exposures.  The FA exposure is designed to simulate experimental exposure conditions but without the exposure pollutant (e.g., CAPs). Statistical models then include the FA results along with the CAPs results, thus accounting for any health effects that are a result of the experimental conditions. Thus the FA exposure is critical in determining if significant exposure (e.g., CAPs)-induced changes have occurred.  For the current project, the FA exposure is carried out by running the coarse concentrator as usual, but a HEPA filter is inserted in-line before delivery to the subject, simulating the CAPs exposure experimental conditions but removing the particles. Since 50% of the exposures were coarse CAPs, it was decided to run the coarse CAPs concentrator for the FA exposures and not to use the fine concentrator. We carried out an interim analysis of the data in October, 2008 and for FA exposures there were small changes (pre to post and during exposure) for the main response variables (BP, blood neutrophils, FMD, BAD) of similar magnitude compared to the CAPs exposures. These FA changes were larger than had been observed in FA exposures from our previous fine CAPs study using another CAPs exposure facility. The most significant change between the two facilities (both located in the same building) was that the PM₁₀ inlet for the new CAPs facility (fine and coarse CAPs) is located at ground-level (~1.5 m high) adjacent to the street (closer to traffic sources) compared to the previous fine CAPs facility in which the inlet was located on the 2^nd floor (~10 m high). Furthermore, although the HEPA filter removes particles (confirmed by particle count and mass measures), volatile organic compounds (VOCs) and gaseous pollutants can penetrate through the HEPA filter and potentially contribute to acute physiological effects. The gaseous criteria pollutant levels were no higher than previously measured on the 2^nd floor, thus not likely responsible, however, no data was available (collected) for VOCs.  Alternatively, other differences in the experimental setup or in the subjects, account for the observed FA responses. The latter findings were discussed in group meetings that included the study PIs (Frances Silverman, Diane Gold), study coordinator (Bruce Urch), lab manager (Mary Speck), statisticians (Paul Corey, Brent Coull), Joel Schwartz, John Godleski and Petros Koutrakis. A consensus decision was made to add a 5^th exposure using humidified HEPA-filtered medical air (FA-MA), after all four exposures had been completed. In November 2008, the first FA-MA exposure was carried out in subjects that had completed all four exposures. The medical air (breathing quality grade) is supplied from a compressed air gas cylinder using a regulator and flow measuring device. Briefly, during FA-MA exposures, the medical air is first passed into a humidifier system (distilled ozonized water) regulated to maintain 30% relative humidity. The humidified air is then passed through a HEPA filter to remove particles and is delivered to the exposure facility at the same flow rate and using the same inlet and face mask as with the FA/CAPs exposures. During the FA-MA exposures, the subjects are not able to detect any visual cues that the exposure is any different from the FA or CAPs exposures. An interim statistical analysis of FA-MA results up to June, 2009 was carried out on the main health effect outcome measures and revealed differences in the magnitude of response between the FA and FA-MA exposures. As a result, on consultation with Harvard’s statistical core (Brent Coull) it was decided henceforth to include the FA-MA in the study randomization of the five exposures (FA, FA-MA, 2 x coarse CAPs, fine CAPs). In the group meetings it was also agreed that it would be beneficial to measure VOCs during FA and CAPs exposures. Harvard was able to supply multi-bed thermal desorption tubes for VOC sampling and to carry out the analyses for 23 VOCs. Starting December, 2008 VOC sampling was commenced with measures of both outdoor samples (concentrator inlet) and in the airstream delivered to subjects. The analyzed VOC results were sent to GOEHU for linkage with health effects data. The initial results showed variability in total VOC levels within each exposure type, including FA. Integrated exposure total VOC levels were associated with outdoor total VOC levels, usually slightly lower. Linkages with health effects data have not been carried out as of yet. A Masters of Public Health student at GOEHU will be working with the VOC and health data as a term project (starting January 2010). 

 

Preliminary Results: The most recent interim statistical analysis was November, 2009 and 16 subjects had completed four exposures (FA, fine CAPs and 2 coarse CAPs) and among those, 10 had completed 5 exposures including the FA-MA. Findings are summarized as follows:  Blood pressure (BP) was measured during the exposure at 30-min intervals, using an automated BP cuff. Diastolic BP increased over the 2-hr CAPs exposures with the largest median increase for coarse CAPs (2.63 mm Hg, p<0.0001, n=37, Wilcoxin Signed Rank) compared to fine CAPs (2.01 mm Hg, p=0.018, n=19). The median increase for FA was 3.29 mm Hg (p=0.039, n=18) with the smallest increase for FA-MA (1.31 mm Hg, p=0.36, n=14). The coarse CAPs response was not significantly different from FA-MA (p=0.34). Systolic BP showed a different pattern with the largest CAPs increase for fine CAPs (2.48 mm Hg, p=0.096, n=19) compared to coarse CAPs (2.04 mm Hg, p=0.025, n=37). The median increase for FA was 3.73 mm Hg (p=0.009, n=18) with the smallest increase for FA-MA (1.56 mm Hg, p=0.54, n=14). Neither fine CAPs nor coarse CAPs responses were significantly different from FA-MA (p=0.34 for both).

 

DNA Methylation: We plan to analyze more blood samples from the coarse and fine CAPs exposure studies to assess DNA methylation changes, following from our promising pilot study data showing decreased global methylation and a suggestion of resolution with iNOS 24-hrs after CAPs exposures. Sputum neutrophils, measured 24 hrs after exposure were higher following coarse CAPs (median 0.40 x 10⁶ cells/g, n=17) compared to fine CAPs (0.13 x 10⁶ cells/g, n=8), and both FA (median 0.22 x 10⁶ cells/g, n=7) and FA-MA (median 0.25 x 10⁶ cells/g, n=6). Not all subjects were able to produce sputum plugs, resulting in less data points. Even so, the coarse CAPs change vs FA change approached significance (p=0.069). An abstract and poster of the induced sputum results was presented at the American Thoracic Society meeting in May 2009. 

 

Blood analysis: Blood was taken before exposure (pre) and 1 and 20 hrs after completion of the exposure (post). Blood neutrophils, a gross marker of systemic inflammation, showed the greatest median post minus pre-exposure increase at 1 hr after fine CAPS, a 17.2% increase (p=0.008, n=19).  For the other exposure types, the median neutrophil changes at 1 hr after were a 5.7% increase for coarse CAPs (p=0.003, n=36), a 10.6% increase for FA (p=0.009, n=17) and a slight decrease of -0.1% for FA-MA (p=0.58, n=14). However, none of the 1-hr neutrophil changes were significantly different from each other. Neutrophil responses 20 hrs post presented as a small decrease from pre-exposure for all exposure types.

Brachial artery diameter (BAD) and flow-mediated dilation (FMD) were measured before (pre-exposure) and 1 and 20 hrs after exposure (post). Intra-exposure changes in BAD and FMD were assessed as the post minus pre-exposure values for each exposure type.  Inter-exposure differences were assessed using the SAS Mixed Model differences of least square means procedure. There were no significant adverse effects of CAPs exposure on BAD (i.e., a decrease in diameter) compared to FA or FA-MA.  FMD showed a median decrease of 2.3% (n=16) 1 hr after coarse CAPs and a decrease of 0.8% (n=9) 1 hr after fine CAPs.  The corresponding median change for FA was a 3.3% decrease (n=8) and a 3.3% increase (n=8) after FA-MA. There was a significant difference between the coarse CAPs and FA-MA changes (p=0.03) and the FA and FA-MA changes (p=0.02).  No significant changes were seen 24 hrs post for FMD. 

 

Pulmonary function changes were small with no significant adverse effects following CAPs exposures compared to FA.

 

Traffic Density: A video camera was set up to record traffic flow adjacent to the CAP inlet on College Street. Thirty recordings have been analyzed from April 29, 2008 to March 25, 2009.  Over this period, the mean ± SD counts were 2,386 ± 62 (range: 2,202 - 2,550 vehicles) over the 2 hr 10 minute exposure period.  The mean percentage of diesel vehicles (trucks) was 6.2 ± 1.0% (range: 4.2-8.9%). The traffic counts will be used as predictors of health outcomes in regression analyses and to examine potential relationships between traffic flow and changes in PM₁₀ mass concentrations at the concentrator inlet.  

 

Finometer data have been sent to an expert (Gianfranco Parati) with customized software analyses, and we are awaiting analyses. Echocardiography proved to be difficult to obtain in some patients and was time consuming. A preliminary analysis of the echo data showed no differences between FA and CAPs exposures. After consultation with Rob Brook it was agreed to drop the measure and replace it with a SphygmoCor device (large arterial compliance, central aortic BP and hemodynamics). The advantage of the SphygmoCor is that Rob Brook will also be using this measure in his recently funded EPA study of coarse CAPs, so it provides a common endpoint for inter comparisons. The Finometer tests also give an estimate of cardiac output, to replace the echocardiography measure.

Nasal scrapings: Nasal scrapings and RNA gene expression (micro-array) have been included as an outcome measure for all newly recruited subjects with allergic nasal rhinitis (at least 1 positive skin prick test). This test is a component of another study (AllerGen NCE, Dr. Jeremy Scott) in subjects with allergic rhinitis, recruited from the PM Center study subjects.  Biological samples are sent to Health Canada for analysis of markers of oxidative stress in blood (8-OHdG, IL-6, ET-1, VEGF, TBARS, oxidative potential, isoprostane-8 and conjugated diene) and urine (8-OHdG, isoprostane-8, D-Glucaric acid, VEGF 8-OHdG and TBARS). 

 

Heart Rate Variability: Subjects wear a Mortara Holter ECG recorder during the entire exposure day and through to the next day (24-hr recording). ECG digital data are sent to Harvard for analyses.  No results are available at this time. 

 

Exposure Characterization: Teflon filters (47 mm) are sent to Environment Canada (EC) for pre- and post-exposure conditioning and weighing. The median integrated mass concentrations for 38 coarse CAPs exposures was 197 µg/m³ (interquartile range: 187-216 µg/m³) and for 20 fine CAPs was 257 µg/m³ (interquartile range: 219-286 µg/m³). After gravimetric measurements are completed, filters are analyzed at EC for sulphate, nitrate and ammonium by IC. Quartz filters (25 mm) are pre-fired to remove organics and after exposure sent back for OC/EC analyses by TOT using a 1.45 cm² punch. A punch will also be taken for selected organics, including a motor vehicle tracer (engine lubricant), by GC-MS. Teflon filters (37 mm) are collected for trace element analyses and are sent to the DRI lab for XRF. Biologic components are collected on 25 mm polycarbonate filters for endotoxin and β-glucans, both exposure and outdoor samples. Continuous data are also collected for PM_2.5 mass (TEOM, DustTrak), black carbon (PSAP), temperature, RH% and criteria gases.

 

Biologic Characterization: We have collected filter samples during all FA (control) coarse and fine CAP exposures for determination of biological material including airborne endotoxin and markers of fungi. Samples will be batch analyzed at a later date. In addition, we have recently started to collect filter samples from pre-exposure calibration runs of coarse and fine CAPs for the specific objective of adapting DGGE and DNA macro-array technologies for the characterization of airborne and dust-borne fungal contaminants and to investigate the value of these data in the context of studying indoor/outdoor exposures and health outcome measures. The latter information will inform the Canadian Healthy Infant Longitudinal Development study (a multi-disciplinary, longitudinal, birth-cohort study).

 

Meetings: There have been regular scientific communications/meetings between the study investigators (GOEHU, Michigan & Harvard). A weekly 1-hr teleconference between Toronto and Harvard was initiated in June 2009 to discuss project-related issues. These meetings have proven to be both fruitful and beneficial to the development/progress of this project, thus will continue throughout to its completion. 

 

Health Canada: Oxidative Stress Collaboration: Our collaboration with Health Canada on oxidative stress markers was extended to include a total of 50 subjects.  Specifically, measures in urine include: 8-hydroxy-2'-deoxyguanosine (8-OHdG) for oxidative DNA damage, isoprostane-8 and thiobarbituric acid reactive substances (TBARS) for oxidative stress; and D-Glucaric acid for liver response and enzyme stimulation.  Measures in blood will include: 8-OHdG, VEGF an angiogenesis factor, ET-1, IL-6 and TNF-α for inflammation; and oxidative potential, isoprostane-8, TBARS and conjugated diene for oxidative stress.

EPA-STAR Coarse CAP Study Collaboration: University of Michigan (Ann Arbor) collaborator Dr. Rob Brook will be ready to start human exposures in his new coarse CAPs facility in 2010. Common coarse CAPs outcome measures will be compared between our Toronto site and Ann Arbor, including: ultrasound baseline diameter and reactivity; Finometer continuous BP; arm cuff BP; SphygmoCor; and biological constituents collected on filters, with the characterization coordinated by Diane Gold at Harvard for measures of airborne endotoxin and markers of fungi.