Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Assessing the impacts of traffic-related and woodsmoke particulate matter on subclinical measures of… Kajbafzadeh, Majid 2014

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


24-ubc_2014_spring_kajbafzadeh_majid.pdf [ 4.59MB ]
JSON: 24-1.0107211.json
JSON-LD: 24-1.0107211-ld.json
RDF/XML (Pretty): 24-1.0107211-rdf.xml
RDF/JSON: 24-1.0107211-rdf.json
Turtle: 24-1.0107211-turtle.txt
N-Triples: 24-1.0107211-rdf-ntriples.txt
Original Record: 24-1.0107211-source.json
Full Text

Full Text

ASSESSING	  THE	  IMPACTS	  OF	  TRAFFIC-­‐RELATED	  AND	  WOODSMOKE	  PARTICULATE	  MATTER	  ON	  SUBCLINICAL	  MEASURES	  OF	  CARDIOVASCULAR	  HEALTH:	  A	  HEPA	  FILTER	  INTERVENTION	  STUDY	  	   by	  Majid	  Kajbafzadeh	  	  B.Sc.,	  Simon	  Fraser	  University,	  2011	  	  A	  THESIS	  SUBMITTED	  IN	  PARTIAL	  FULFILLMENT	  OF	  THE	  REQUIREMENTS	  FOR	  THE	  DEGREE	  OF	  	  MASTER	  OF	  SCIENCE	  in	  THE	  FACULTY	  OF	  GRADUATE	  AND	  POSTDOCTORAL	  STUDIES	  (Occupational	  and	  Environmental	  Hygiene)	  	  THE	  UNIVERSITY	  OF	  BRITISH	  COLUMBIA	  (Vancouver)	  	  	  April	  2014	  	  ©	  Majid	  Kajbafzadeh,	  2014	  	   ii	  Abstract	  	  Fine	   particulate	   matter	   (PM2.5)	   plays	   an	   important	   role	   in	   the	   link	   between	   air	  pollution	   and	   a	   range	   of	   health	   effects	   including	   respiratory	   and	   cardiovascular	  morbidity	  and	  mortality.	  The	  specific	  sources	  of	  PM2.5	  responsible	  for	  these	  effects	  have	  not	  been	  definitively	   identified.	  With	   traffic-­‐related	  air	  pollution	   (TRAP)	  and	  woodsmoke	  (WS)	  as	  two	  of	  the	  major	  contributors	  to	  ambient	  PM2.5	  concentrations,	  this	   study	  was	   the	   first	   to	   investigate	   the	   difference	   in	   health	   outcomes	   between	  these	   two	   sources.	   The	   purpose	   of	   this	   study	   was	   to	   compare	   cardiovascular	  exposure-­‐response	   relationships	   for	   TRAP	   and	  WS	   and	   to	   evaluate	   the	   impact	   of	  HEPA	  filtration	  on	  indoor	  TRAP	  and	  WS	  PM2.5	  levels.	  	  	  In	  this	  single-­‐blind	  randomized	  crossover	  study,	  83	  healthy	  adults	  (54	  living	  in	  high	  TRAP	  and	  29	  living	  in	  high	  WS	  areas)	  between	  the	  ages	  of	  19	  and	  72	  living	  in	  Metro	  Vancouver	  were	  recruited.	  Areas	  with	  high	  TRAP	  or	  high	  WS	  were	  identified	  using	  previously	  developed	  spatial	  models	  and	  subjects	  were	  recruited	  by	  letters	  sent	  to	  households	  in	  these	  areas.	  Sampling	  was	  conducted	  over	  two	  consecutive	  one-­‐week	  periods,	  one	  with	   filtration	  and	  one	  with	  no	   filtration.	  Two	   filtration	  devices	  were	  used,	  one	   in	   the	  main	   living	   room	  and	  one	   in	  main	  bedroom.	  Endothelial	   function	  was	  measured	  at	  the	  end	  of	  each	  week	  and	  blood	  was	  drawn	  at	  baseline	  and	  at	  the	  end	   of	   each	   week.	   Mixed	   effect	   models	   were	   used	   to	   investigate	   the	   relationship	  between	  exposure	  and	  outcome	  variables.	  	  Overall,	   HEPA	   filtration	   was	   associated	   with	   a	   40%	   decrease	   in	   indoor	   PM2.5	  concentrations.	   There	   was	   inconclusive	   evidence	   on	   the	   potential	   relationship	  between	  TRAP	  or	  WS	  PM2.5	  exposure	  and	  endothelial	  function.	  However,	  there	  was	  some	   suggestion	   of	   an	   association	   between	   PM2.5	   exposure	   and	   CRP	   specifically	  among	  male	  participants	  in	  high-­‐TRAP	  locations	  (20.6%	  increase	  in	  CRP	  levels	  per	  unit	  median	   increase	   in	  PM2.5,	  95%	  CI,	  2.62%	  –	  41.7%).	  There	  was	  no	  association	  between	  any	  exposure	  indicators	  and	  IL-­‐6	  or	  BCC.	  In	  summary,	  the	  results	  support	  the	   hypothesis	   that	   HEPA	   filtration	   can	   be	   effective	   in	   reducing	   indoor	   PM2.5	  	   iii	  concentrations	  with	  some	  support	  for	  the	  a	  priori	  hypothesis	  of	  a	  greater	  impact	  on	  markers	  of	  inflammation	  in	  areas	  of	  high	  TRAP.	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   iv	  Preface	  	  This	  study	  was	  conducted	  as	  a	  collaborative	  effort	  between	  the	  University	  of	  British	  Columbia	   and	   Simon	   Fraser	  University.	   All	   research	   described	   in	   this	   dissertation	  was	   conducted	   under	   the	   approval	   of	   Simon	   Fraser	   University	   Research	   Ethics	  Boards	   (2011a0431).	   As	   per	   the	   evaluation	   by	   UBC	   Research	   Ethics	   Boards,	   the	  approval	  obtained	  from	  SFU	  was	  recognized	  by	  UBC	  and	  sufficient	  for	  the	  purposes	  of	  this	  study.	  	  	  	   v	  Table	  of	  Contents	  Abstract	  .......................................................................................................................................	  ii	  Preface	  ........................................................................................................................................	  iv	  Table	  of	  Contents	  .....................................................................................................................	  v	  List	  of	  Tables	  ...........................................................................................................................	  vii	  List	  of	  Figures	  ........................................................................................................................	  viii	  List	  of	  Abbreviations	  .............................................................................................................	  ix	  Acknowledgements	  ................................................................................................................	  xi	  1	   Introduction	  .......................................................................................................................	  1	  1.1	   Literature	  Review	  ...................................................................................................................	  3	  1.1.1	   Components	  of	  Air	  Pollution	  .......................................................................................................	  3	  1.1.2	   Physical	  and	  Chemical	  Composition	  of	  PM	  ...........................................................................	  4	  1.1.3	   PM	  Exposure	  Health	  Effects	  ........................................................................................................	  5	  1.2	   Mechanisms	  of	  Action	  ........................................................................................................	  10	  1.3	   Two	  Major	  Sources	  of	  Combustion-­‐Derived	  PM	  .......................................................	  12	  1.3.1	   Traffic-­‐Related	  PM	  ........................................................................................................................	  12	  1.3.2	   WS	  PM	  ................................................................................................................................................	  13	  1.4	   Surrogates	  of	  TRAP	  and	  WS	  PM	  ......................................................................................	  14	  1.4.1	   Hopanes	  .............................................................................................................................................	  14	  1.4.2	   Optical	  Absorbance	  ......................................................................................................................	  15	  1.4.3	   Levoglucosan	  (LG)	  ........................................................................................................................	  15	  1.5	   Study	  Rationale	  ....................................................................................................................	  15	  1.6	   Objectives	  and	  Hypotheses	  ..............................................................................................	  19	  2	   Methods	  .............................................................................................................................	  21	  2.1	   Study	  Design	  ..........................................................................................................................	  21	  2.2	   TRAP	  and	  WS	  Region	  Selection	  in	  Greater	  Vancouver	  ...........................................	  22	  2.2.1	   Home	  Categorization	  Verification	  ..........................................................................................	  24	  2.3	   Participant	  Recruitment	  Campaign	  ...............................................................................	  24	  2.4	   Study	  Participant	  Characteristics	  ..................................................................................	  24	  2.5	   Participant	  Preparation	  Before	  Sampling	  ..................................................................	  25	  2.6	   Study	  Period	  ..........................................................................................................................	  26	  2.7	   HEPA	  Filtration	  ....................................................................................................................	  26	  2.7.1	   Electricity	  Use	  Meter	  ...................................................................................................................	  27	  2.8	   Health	  and	  Exposure	  Measures	  ......................................................................................	  27	  2.9	   Exposure	  Measurements	  ..................................................................................................	  28	  2.9.1	   PM2.5	  ....................................................................................................................................................	  28	  2.9.2	   Indoor	  Set	  Up	  ..................................................................................................................................	  29	  2.9.3	   Gravimetric	  Analysis	  of	  the	  Filters	  ........................................................................................	  29	  2.9.4	   Harvard	  Impactor	  Maintenance	  .............................................................................................	  31	  2.9.5	   Optical	  Reflectance	  Analysis	  ....................................................................................................	  31	  2.9.6	   Optical	  Reflectance	  and	  Absorbance	  ....................................................................................	  31	  2.9.7	   Hopanes	  Laboratory	  Analysis	  ..................................................................................................	  32	  	   vi	  2.9.8	   LG	  Laboratory	  Analysis	  ..............................................................................................................	  32	  2.9.9	   Indoor	  Temperature	  and	  Relative	  Humidity	  (RH)	  .........................................................	  33	  2.9.10	  	  Time-­‐Location-­‐Activity	  Log	  ......................................................................................................	  33	  2.10	   Health	  Measurements	  .....................................................................................................	  33	  2.10.1	   Microvascular	  Endothelial	  Function	  ..................................................................................	  34	  2.10.2	   Blood	  Collection,	  Processing,	  and	  Analysis	  .....................................................................	  35	  2.11	   Statistical	  Analysis	  ............................................................................................................	  37	  2.11.1	   Data	  Cleaning	  and	  Descriptive	  Statistics	  ..........................................................................	  37	  2.11.2	   Distribution	  of	  Exposure	  and	  Health	  Variables	  .............................................................	  38	  2.11.3	   Extreme	  Values	  ............................................................................................................................	  38	  2.11.4	   Scaling	  Exposure	  Contrasts	  ....................................................................................................	  38	  2.11.5	   Evaluation	  of	  Difference	  Between	  Home	  Types	  and	  Filtration	  Status	  ................	  39	  2.11.6	   Mixed	  Model	  Analysis	  ...............................................................................................................	  39	  2.11.7	   Effect	  Modification	  by	  Age,	  BMI,	  and	  Gender	  .................................................................	  41	  2.11.8	   Interpreting	  Effect	  Estimates	  from	  Log	  Transformed	  Models	  ................................	  41	  3	   Results	  ...............................................................................................................................	  43	  3.1	   Summary	  Statistics	  –	  Participants	  with	  Complete	  Data	  .........................................	  43	  3.2	   Summary	  Statistics	  –	  Participants	  with	  Incomplete	  Data	  ......................................	  45	  3.3	   Exposure	  Characteristics	  ..................................................................................................	  46	  3.4	   Health	  Outcomes	  ..................................................................................................................	  54	  3.5	   Mixed	  Model	  Results	  ...........................................................................................................	  55	  3.6	   HEPA	  Filtration	  ....................................................................................................................	  56	  3.7	   PM2.5	  .........................................................................................................................................	  56	  3.8	   Absorbance	  ............................................................................................................................	  57	  3.9	   LG	  ..............................................................................................................................................	  58	  3.10	  	  Effect	  Modification	  by	  BMI,	  Age,	  and	  Sex	  .....................................................................	  59	  4	   Discussion	  ........................................................................................................................	  61	  4.1	   HEPA	  Filtration	  Efficiency	  ................................................................................................	  62	  4.2	   Mixed	  Model	  Conclusions	  .................................................................................................	  62	  4.2.1	   RHI	  .......................................................................................................................................................	  63	  4.2.2	   CRP	  ......................................................................................................................................................	  65	  4.2.3	   IL-­‐6	  ......................................................................................................................................................	  68	  4.2.4	   BCC	  and	  %PMN	  ..............................................................................................................................	  70	  4.3	   Conclusions	  ............................................................................................................................	  71	  4.4	   Significance	  of	  Study	  ...........................................................................................................	  72	  4.5	   Strengths	  ................................................................................................................................	  72	  4.6	  	  	  	  	  Limitations	  .............................................................................................................................	  73	  4.7	   Suggested	  Future	  Research	  ..............................................................................................	  76	  4.8	   Implications	  on	  Public	  Health	  .........................................................................................	  76	  References	  ..............................................................................................................................	  78	  APPENDICES	  .........................................................................................................................	  107	  APPENDIX	  I:	  INVITATION	  LETTERS	  ........................................................................................	  108	  APPENDIX	  II:	  PARTICIPANT	  SCREENING	  ..............................................................................	  110	  APPENDIX	  III:	  INFORMED	  CONSENT	  ......................................................................................	  115	  APPENDIX	  IV:	  HEALTH	  AND	  EXPOSURE	  REPORT	  ..............................................................	  122	  APPENDIX	  V:	  DWELLING	  INFORMATION	  ..............................................................................	  128	  APPENDIX	  VI:	  TIME-­‐LOCATION-­‐ACTIVITY	  DIARY	  .............................................................	  132	  APPENDIX	  VIII:	  HEALTH	  LOG	  ...................................................................................................	  138	  	   vii	  	  List	  of	  Tables	  	  Table	  2-­‐1	  Health	  and	  Exposure	  Measures	  ....................................................................................	  28	  Table	  2-­‐2	  Summary	  of	  the	  mixed	  effect	  models	  used	  in	  the	  analysis	  ..............................	  41	  Table	  3-­‐1	  Summary	  of	  study	  population	  characteristics	  with	  complete	  data	  ..............	  44	  Table	  3-­‐2	  Comparison	  of	  study	  population	  characteristics	  between	  participants	  with	  complete	  and	  incomplete	  data	  for	  RHI	  ................................................................................	  46	  Table	  3-­‐3	  Summary	  of	  exposure	  characteristics	  by	  HEPA	  filtration	  status	  ..................	  48	  Table	  3-­‐4	  Summary	  of	  exposure	  characteristics	  by	  home	  type	  (No	  HEPA)	  ..................	  51	  Table	  3-­‐5	  Summary	  of	  exposure	  characteristics	  by	  HEPA	  filtration	  status	  and	  home	  type	  ......................................................................................................................................................	  53	  Table	  3-­‐6	  Summary	  statistics	  to	  confirm	  categorization	  of	  TRAP	  and	  WS	  homes	  .....	  54	  Table	  3-­‐7	  Summary	  of	  health	  outcomes	  by	  HEPA	  filtration	  status	  ...................................	  55	  Table	  3-­‐8	  Change	  in	  health	  outcomes	  with	  HEPA	  filtration	  .................................................	  56	  Table	  3-­‐9	  Change	  in	  health	  outcomes	  per	  unit	  median	  change	  in	  PM2.5	  .........................	  57	  Table	  3-­‐10	  Change	  in	  health	  outcomes	  per	  unit	  median	  change	  in	  absorbance	  .........	  58	  Table	  3-­‐11	  Change	  in	  health	  outcomes	  per	  unit	  median	  change	  in	  LG	  ...........................	  59	  	  	  	  	  	  	  	  	  	  	  	  	  	   viii	  List	  of	  Figures	  	  Figure	  2-­‐1	  Summary	  of	  the	  Study	  Design	  .....................................................................................	  22	  Figure	  2-­‐2	  a.	  modeled	  TRAP	  concentrations	  (Henderson	  et	  al.,	  2007);	  b.	  modeled	  WS	  tertiles	  (Larson	  et	  al.,	  2007);	  c.	  high	  TRAP	  and	  low	  WS	  postal	  codes;	  d.	  high	  WS	  and	  low	  TRAP	  postal	  codes	  .......................................................................................................	  23	  Figure	  2-­‐3	  HEPA	  Filtration	  Devices	  Used	  –	  Honeywell	  50250/50300	  (Left)	  &	  Honeywell	  18150	  (Right)	  ..........................................................................................................	  26	  Figure	  3-­‐1	  Effect	  modification	  by	  sex	  in	  the	  relationship	  between	  CRP	  and	  PM2.5	  ....	  59	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   ix	  List	  of	  Abbreviations	  	  BCC	  –	  Band	  cell	  counts	  BMI	  –	  Body	  mass	  index	  CHD	  –	  Coronary	  heart	  disease	  CO	  -­‐	  Carbon	  monoxide	  CRP	  –	  C-­‐reactive	  protein	  CVD	  –	  Cardiovascular	  Disease	  DALY	  –	  Disability	  adjusted	  life	  years	  EC	  –	  Elemental	  carbon	  GC-­‐MS	  –	  Gas	  chromatography	  mass	  spectroscopy	  HEPA	  –	  High	  efficiency	  particulate	  air	  filter	  hs-­‐CRP	  –	  High	  sensitivity	  C-­‐reactive	  protein	  IARC	  –	  International	  Agency	  for	  Research	  on	  Cancer	  IL-­‐6	  –	  Interleukin	  6	  Kg/m2	  –	  Kilograms	  per	  square	  meter	  L	  -­‐	  Liter	  LG	  –	  Levoglucosan	  LOD	  –	  Limit	  of	  detection	  m2	  –	  Meter	  square	  mmHg	  –	  Millimeters	  mercury	  PAT	  –	  Peripheral	  arterial	  tonometry	  PM	  –	  Particulate	  Matter	  PM2.5	  –	  Particulate	  matter	  smaller	  than	  2.5	  microns	  PMN	  –	  Polymorphonuclear	  leukocytes	  %PMN	  –	  Percent	  polymorphonuclear	  leukocytes	  PWA	  –	  Pulse	  wave	  amplitude	  QC	  –	  Quality	  control	  RH	  –	  Relative	  humidity	  RHI	  –	  Reactive	  hyperemia	  index	  RWC	  –	  Residential	  wood	  combustion	  	   x	  SD	  –	  Standard	  deviation	  SOP	  –	  Standard	  operating	  procedure	  SPPH	  –	  School	  of	  Population	  and	  Public	  Health	  TRAP	  –	  Traffic-­‐related	  air	  pollution	  WS	  –	  Woodsmoke	  µg	  –	  Micro	  grams	  µg/m3	  –	  Micro	  grams	  per	  cubic	  meter	  µm	  –	  Micro	  meter	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   xi	  Acknowledgements	  	  This	   thesis	   would	   not	   have	   been	   possible	   without	   the	   selfless	   guidance	   of	   a	   few	  individuals.	   First	   of	   all,	   I	   would	   like	   to	   thank	  my	   supervisors,	   Dr.	  Michael	   Brauer	  (UBC)	  and	  Dr.	  Ryan	  Allen	  (SFU)	  who	  both	  guided	  me	  and	  at	  the	  same	  time	  pushed	  me	  to	  think	  like	  an	  a	  experienced	  researcher;	  I	  am	  honored	  to	  have	  been	  one	  of	  your	  students.	  With	  your	  help	  I	  have	  been	  able	  to	  learn	  more	  than	  I	  ever	  thought	  I	  would	  in	  these	  two	  years.	  	  	  I	   would	   also	   like	   to	   thank	   Dr.	   Christopher	   Carlsten,	   my	   committee	   member,	   for	  assisting	   me	   with	   widening	   my	   view	   of	   the	   research	   I	   was	   involved	   in.	   I	   really	  appreciate	  your	  patience	  and	  help	  with	  ensuring	  that	  I	  also	  understand	  the	  clinical	  aspects	  of	  this	  study	  and	  air	  pollution	  in	  general.	  Moreover,	  this	  project	  would	  not	  have	  been	  possible	  without	  the	  coordination	  and	  management	  skills	  of	  our	  research	  project	   coordinator,	   Barbara	   Karlen	   who	   was	   very	   supportive	   and	   was	   always	  available	  to	  help.	  Also,	  thanks	  to	  all	  of	  the	  occupational	  and	  environmental	  hygiene	  faculty	  specially	  Dr.	  Hugh	  Davies,	  Dr.	  Karen	  Bartlett,	  and	  Dr.	  George	  Astrakianakis	  for	  their	  invaluable	  practical	  advice	  and	  support.	  	  I	  would	  like	  to	  express	  my	  deepest	  thanks	  to	  my	  parents	  for	  their	  much	  appreciated	  love	  and	  encouragement	  throughout	  this	  short,	  yet	  tough	  journey.	  Thank	  you	  to	  my	  awesome	   siblings	   and	   friends	   for	   keeping	  me	   upbeat	   and	  motivated	   through	   the	  tough	  days	  of	  this	  research	  project.	  	  	  Last	  but	  definitely	  not	  least,	  I	  would	  like	  to	  thank	  the	  Canadian	  Institutes	  for	  Health	  Research	  (CIHR)	  (Funding	  Reference	  Number	  111042)	  and	  the	  Collaborative	  Research	  and	  Training	  Experience	  –	  Atmospheric	  Aerosol	  Program	  (CREATE-­‐APP)	  for	  funding	  this	  study	  and	  my	  graduate	  work.	  	  	  	   xii	  Dedication	  	  “The	  family	  is	  one	  of	  nature’s	  masterpieces”	  –	  George	  Santayana	  	  To	  my	  lovely	  parents,	  Mohammad	  and	  Shohreh,	  and	  awesome	  siblings,	  Behrad	  and	  Kiana.	  	  	  	  	  	   1	  1 Introduction	  	  A	  number	  of	  events	  in	  the	  20th	  century	  have	  revealed	  the	  significance	  of	  exposure	  to	  air	  pollution	  among	  the	  general	  public	  and	  the	  scientific	  community.	  One	  of	  the	  first	  well-­‐documented	  events	  occurred	  in	  December	  of	  1930	  over	  the	  span	  of	  five	  days	  in	  the	  Meuse	  Valley	  of	  Belgium,	  at	  the	  time	  the	  most	  industrialized	  region	  of	  Europe.	  As	  a	   result	   of	   stable	   atmospheric	   conditions	   and	   industrial	   emissions,	   a	   thick	   fog	  covered	  a	  great	  area	  of	  the	  region.	  Starting	  from	  the	  third	  day	  of	  the	  fog,	  scores	  of	  people	  exhibited	  severe	  respiratory	  symptoms.	  Over	  60	  people	  died	   in	  a	  period	  of	  three	  days,	  which	  was	  more	  than	  10	  times	  the	  normal	  mortality	  rate	  for	  the	  region	  (Nemery,	  Hoet,	  &	  Nemmar,	  2001).	  	  	  In	  1948,	   the	   first	  publicized	  air	  pollution	  event	   in	   the	  United	  States	  was	  recorded.	  Due	   to	   residential	   and	   industrial	   emissions	   in	   Donora,	   Pennsylvania	   and	   a	  temperature	   inversion,	  about	  half	  of	   the	  14,000	  population	  of	   the	   town	  became	   ill	  and	   20	   died,	   6	   times	   the	   normal	   mortality	   rate	   (Helfand,	   Lazarus,	   &	   Theerman,	  2001).	  Finally,	   in	  the	  best	  known	  and	  most	  severe	  air	  pollution	  episode	  of	  the	  20th	  century,	   over	   4,000	   excess	   deaths	   were	   reported	   when	   London,	   England	   was	  covered	  with	  dense	  smog	  from	  December	  5th	   to	  December	  9th	  in	  1952	  (Ministry	  of	  Health,	  1954).	  Subsequent	  analyses	  of	  the	  London	  smog	  episode	  have	  reported	  over	  12,000	  excess	  deaths	  as	  a	  result	  of	  acute	  and	  longer	  term	  exposures	  to	  high	  levels	  of	  air	  pollution	  (Bell	  &	  Davis,	  2001).	  	  With	  the	  extensive	  news	  coverage	  of	  each	  of	  these	  air	  pollution	  episodes,	  it	  became	  clear	   to	   the	   public	   that	   air	   pollution	  was	   responsible	   for	   these	   deaths.	  Moreover,	  many	   individuals	   realized	   that	   even	   short-­‐term	   exposures	   could	   have	   detrimental	  effects	   on	   their	   health.	   In	   all	   three	   of	   the	   events	   mentioned	   above,	   immediate	  investigations	   were	   carried	   out,	   which	   confirmed	   the	   association	   between	   short-­‐term	  exposure	  to	  high	  levels	  of	  air	  pollution	  and	  increased	  morbidity	  and	  mortality	  at	  a	  population	  level	  (Holgate,	  Koren,	  Samet,	  &	  Maynard,	  1999).	  	  	   2	  It	   was	   only	   after	   the	   London	   smog	   that	   significant	   efforts	   were	   initiated	   to	  understand	   the	   relationship	   between	   exposure	   to	   various	   components	   of	   air	  pollution	  and	  their	  potential	  effect	  on	  health.	  More	  recently,	  among	  other	  pollutants,	  most	  of	  the	  focus	  has	  been	  on	  exposure	  to	  particulate	  matter	  (PM)	  as	  more	  evidence	  has	  unveiled	  the	  significant	  contribution	  of	  PM	  to	  various	  serious	  health	  effects	  with	  acute	   and	   chronic	   exposures	   (Pope	   &	   Dockery,	   2006).	   Given	   the	   fact	   that	   air	  pollution	  exposure	  is	  ubiquitous	  with	  no	  safe	  exposure	  threshold	  (Pope	  &	  Dockery,	  2006),	  it	  can	  be	  considered	  as	  a	  serious	  public	  health	  issue	  that	  could	  affect	  anyone	  regardless	  of	   their	   socioeconomic	   status,	   age,	   sex,	   etc.	  Over	   the	  past	   four	  decades,	  various	  measures	   have	   been	   introduced	   to	   reduce	   PM	   emissions	   including	  motor	  vehicle	  engine	  re-­‐design,	  more	  stringent	  emission	  regulations,	  and	  improvements	  in	  fuel	  quality.	  These	  measures	  have	  significantly	  reduced	  air	  pollution/PM	  exposure	  in	  North	  America	  and	  Western	  Europe	  (van	  Erp,	  O’Keefe,	  Cohen,	  &	  Warren,	  2008).	  However,	   high	   air	  pollution	   levels	   are	   still	  widely	  present	   in	  developing	   countries	  such	   as	   China	   and	   India.	   In	   China	  with	   a	   rapidly	   expanding	   economy,	   there	   have	  been	  extreme	  increases	  in	  air	  pollutant	  emissions	  across	  much	  of	  the	  country.	  With	  a	  growing	  need	  for	  energy	  and	  reliance	  on	  coal	  for	  meeting	  75%	  of	  the	  energy	  needs	  and	   increasing	   number	   of	   cars,	   coal	   and	   fuel	   combustion	   PM	   and	   sulfur	   dioxide	  (SO2)	  have	  spiked	  to	  record	  levels	  in	  recent	  years	  (Watts,	  2005).	  This	  has	  left	  major	  metropolitan	   areas	   in	   China	   such	   as	   Beijing	   and	   Shanghai	   among	   the	   cities	   with	  highest	   levels	   of	   air	   pollutants	   globally	   (Kan,	   Chen,	   &	   Tong,	   2012).	   Similarly,	   air	  pollution	   exposure	   is	   more	   severe	   than	   ever	   in	   India	   with	   sources	   ranging	   from	  traffic	  and	  industry	  to	  residential	  biomass	  burning	  and	  open	  burning	  of	  solid	  waste	  (Saraswat	  et	  al.,	  2013). 	  	  	  It	   is	   important	   to	   note	   that	   the	   first	   line	   of	   intervention	   to	   reduce	   air	   pollution	  exposure	  at	  a	  population	  level	  should	  be	  focused	  on	  reducing	  emissions	  at	  source	  as	  described	   above.	  Meanwhile,	   the	   effort	   should	   be	   continued	   to	   better	   understand	  the	   potential	   effects	   of	   air	   pollution	   on	   different	   body	   systems	   including	   the	  cardiovascular	  system.	  One	  of	   the	  more	   innovative	  methods	  to	  evaluate	  the	  health	  effects	  of	  higher	  air	  pollution	  exposure	   is	   the	  use	  of	  high	  efficiency	  particulate	  air	  	   3	  (HEPA)	  filtration	  to	  introduce	  an	  exposure	  concentration	  gradient	  while	  evaluating	  health	  outcomes,	  which	  will	  be	  discussed	  extensively	  in	  the	  following	  sections.	  	  	  It	   has	   been	   shown	   that	   HEPA	   filters	   can	   remove	   a	   great	   portion	   of	   respirable	  particles	  (99.97%	  of	  0.3	  µm	  particles)	  (Yamada	  et	  al.,	  1984)	  and	  can	  reduce	  indoor	  PM	  concentrations	  by	  over	  50%,	  which	  is	  significant	  considering	  the	  fact	  that	  people	  generally	   spend	   the	   majority	   of	   their	   time	   indoors	   (Batterman	   et	   al.,	   2011).	   It	   is	  important	   to	  note	   that	  unlike	  other	  air	   cleaners	   such	  as	  electrostatic	  precipitators	  and	  ion	  generators,	  HEPA	  filtration	  devices	  do	  not	  emit	  ozone	  particles,	  which	  are	  a	  health	  concern	  (Waring,	  Siegel,	  &	  Corsi,	  2008).	  In	  addition,	  HEPA	  filters	  do	  not	  have	  any	  effect	  on	  other	  gases	  such	  as	  nitrogen	  oxides	  thus	  focusing	  the	  intervention	  on	  PM	  only.	  	  An	   additional	   benefit	   of	   using	   HEPA	   filtration	   is	   that	   it	   might	   be	   beneficial	   in	  reducing	   individual-­‐level	   PM	   exposure	   and	   potentially,	   alleviating	   adverse	   health	  effects.	  To	  evaluate	  the	  health	  effects	  of	  exposure	  to	  air	  pollution	  and	  to	  characterize	  the	  effectiveness	  of	   such	   individual-­‐level	   interventions,	   there	   is	   a	  need	   for	   further	  research.	  	  1.1 Literature	  Review	  1.1.1 Components	  of	  Air	  Pollution	  	  There	   are	   six	   common	   air	   pollutants,	   which	   are	   commonly	   known	   as	   “criteria	  pollutants”;	  these	  include	  ozone,	  PM,	  carbon	  monoxide	  (CO),	  nitrogen	  oxides	  (NOx),	  SO2,	   and	   lead.	   There	   have	   been	   a	   number	   of	   studies	   evaluating	   the	   relationship	  between	  these	  pollutants	  and	  health.	  There	  is	  some	  evidence	  linking	  sulfur	  dioxide,	  nitrogen	   dioxide,	   and	   CO	   exposure	   to	   adverse	   cardiopulmonary	   health	   outcomes;	  however,	   there	   is	   still	   uncertainty	   surrounding	   these	   relationships	   (Health	  Effects	  Institute,	  2010a;	  Koken	  et	  al.,	  2003;	  Tarlo	  et	  al.,	  2001;	  Tsai,	  Goggins,	  Chiu,	  &	  Yang,	  2003).	   There	   is	   also	   extensive	   evidence	   linking	   ozone	   exposure	   to	   adverse	  respiratory	   symptoms	   and	   mortality	   from	   respiratory	   causes	   (Ito,	   De	   Leon,	   &	  	   4	  Lippmann,	  2005;	  Jerrett	  et	  al.,	  2009).	  More	  recent	   literature	  indicates	  that	   fine	  PM	  plays	  a	  major	  role	   in	  the	  development	  of	  adverse	  health	  effects	   in	  a	  range	  of	  body	  systems	  (e.g.	  cardiovascular	  and	  potentially	  reproductive)	  (Pope	  &	  Dockery,	  2006).	  With	  increasing	  robust	  indications	  of	  the	  effects	  of	  PM	  on	  health,	  this	  study	  focuses	  on	  PM	  as	  the	  air	  pollutant	  of	  interest	  and	  its	  potential	  effects	  on	  endothelial	  function	  and	  systematic	  inflammation.	  	  	  1.1.2 Physical	  and	  Chemical	  Composition	  of	  PM	  	  PM	   is	  mixture	  of	   solid	  and	   liquid	  particles	  with	  a	  variety	  of	  physical	  and	  chemical	  characteristics.	  Particles	  are	  categorized	  based	  on	  aerodynamic	  diameter,	  which	  can	  be	  related	  to	   their	  deposition	  patterns,	  sources,	  and	  composition	  (Chow,	  1995).	   In	  general,	  particles	  are	  divided	  into	  those	  with	  aerodynamic	  diameters	  of:	  1.	  Between	  2.5	  µm	  and	  10	  µm	  (Coarse),	  2.	  Less	  than	  or	  equal	  to	  2.5	  µm	  (Fine	  or	  PM2.5),	  and	  3.	  Less	  than	  0.1	  µm	  (Ultrafine).	  These	  categories	  are	  significant	  in	  that	  the	  smaller	  the	  aerodynamic	  diameter,	   the	   larger	   the	  surface	  area	  per	  mass,	  which	  could	   increase	  lung’s	  exposure	  to	  various	  compounds	  such	  as	  transition	  metals	  and	  free	  radicals.	  	  Coarse	   particles	   are emitted	   into	   our	   environment	   in	   various	   ways	   including	   the	  suspension	  of	  dust	  and	  soil	  and	  through	  mining	  and	  farming,	  as	  well	  as	  the	  airborne	  release	   of	   pollen	   and	  mold.	   Fine	   particles	   are	  mostly	   formed	   by	   fuel	   combustion,	  wood	   and	   coal	   burning,	   and	   industrial	   processes.	   Similarly,	   ultrafine	   particles	   are	  formed	   from	   sources	   of	   combustion,	   their	   coagulation	   to	   form	   fine	   particles,	   and	  atmospheric	  reactions	  (Pope	  &	  Dockery,	  2006).	  	  	  Even	  in	  the	  same	  size	  categories,	  PM	  particles	  differ	  in	  morphology,	  surface	  charge,	  and	   most	   importantly	   chemical	   composition	   based	   on	   their	   source,	   making	   the	  study	  of	  their	  potential	  health	  effects	  based	  on	  their	  physicochemical	  characteristics	  very	   challenging	   (Schwarze	   et	   al.,	   2006).	   These	   characteristics	   can	   affect	   where	  particles	   will	   be	   deposited	   and	   what	   effects	   may	   result;	   morphology	   and	   surface	  charge,	  among	  other	  factors,	  can	  determine	  if	  the	  particles	  can	  penetrate	  into	  the	  air	  	   5	  exchange	   region	   of	   the	   lungs,	   which	   makes	   their	   clearance	   more	   difficult	   while	  chemical	  composition	  (e.g.	  presence	  of	  transition	  metals)	  can	  have	  an	  effect	  on	  the	  level	   of	   toxicity.	   Very	   few	   studies	   have	   specifically	   compared	   the	   physicochemical	  characteristics	  of	   traffic-­‐related	  PM	  and	  woodsmoke	   (WS)	  as	   two	  main	   sources	  of	  combustion-­‐derived	   PM.	   The	   available	   literature	   reports	   some	   differences	   in	  polyaromatic	   hydrocarbon	   composition,	   particle	   size	   distribution,	   and	   elemental	  composition	   variation	   between	   these	   two	   sources	   (Hedberg	   et	   al.,	   2002).	   These	  findings	   might	   have	   implications	   for	   potential	   differences	   in	   the	   adverse	   health	  effects	  (e.g.	  systemic	  inflammation)	  caused	  by	  different	  sources	  of	  PM.	  	  1.1.3 PM	  Exposure	  Health	  Effects	  	  Indoor	   and	   outdoor	   air	   pollution	   is	   one	   of	   the	   main	   contributors	   to	   the	   global	  burden	  of	  disease	  (Pope	  &	  Dockery,	  2006).	  Among	  the	  different	  components	  of	  air	  pollution,	  PM	  is	  unique	  in	  the	  sense	  that	  numerous	  studies	  have	  demonstrated	  well-­‐defined	   and	   robust	   impacts	   on	   a	  multitude	   of	   body	   systems.	   There	   is	   also	   strong	  toxicological	   evidence	   linking	  PM	   from	  different	   sources	   to	   adverse	  health	   effects.	  While	   there	   are	   different	   sources	   of	   combustion-­‐derived	   PM	   (e.g.	   traffic	   and	  WS),	  not	  all	  of	   them	  have	  been	  studied	  sufficiently	   to	  evaluate	  related	  health	  effects.	  As	  discussed	  in	  the	  following	  sections,	  currently	  there	  are	  strong	  suggestions	  of	  a	  link	  between	   exposure	   to	   traffic	   PM	   and	   different	   health	   outcomes	   (including	  cardiovascular	  health	  outcomes);	  however,	  other	  combustion	  sources	  have	  not	  been	  studied	   as	   extensively	   or	   compared	   to	   health	   effects	   involved	   with	   traffic	   PM	  exposure.	  	  Numerous	  studies	  have	  investigated	  the	  potential	  adverse	  health	  effects	  of	  exposure	  to	  PM	  around	   the	  world.	  The	  early	   research	  on	  air	  pollution	  and	   its	  health	  effects	  predominantly	   focused	   on	  mortality	   and	   adverse	   respiratory	   health	   outcomes.	   In	  the	   past	   few	   decades,	   the	   focus	   has	   shifted	   towards	   other	   outcomes	   such	   as	  cardiovascular,	  reproductive,	  and	  cognitive	  effects.	  	  Previous	  studies	  have	  reported	  links	   between	   PM	   and	   morbidity	   and	   mortality	   attributed	   to	   short-­‐term	   PM	  exposure	   (Crouse	   et	   al.,	   2012;	  Dominici	   F,	   Peng	  RD,	  Bell	  ML,	  &	   et	   al,	   2006;	  Kloog,	  	   6	  Ridgway,	  Koutrakis,	  Coull,	  &	  Schwartz,	  2013)	  or	  long-­‐term	  PM	  exposure	   	  (Dockery	  et	   al.,	   1993;	   Pope	   et	   al.,	   1995;	   Pope	   et	   al.,	   2004).	   Furthermore,	   there	   have	   been	  reports	  of	  health	  effects	  from	  long-­‐term	  exposure	  to	  relatively	  low	  concentrations	  of	  ambient	   PM	   including	   the	   Harvard	   Six	   Cities	   and	   American	   Cancer	   Society	   (ACS)	  studies.	  In	  a	  recent	  large	  cohort	  study	  in	  Canada,	  researchers	  found	  an	  elevated	  risk	  of	   death	   from	   non-­‐accidental	   and	   cardiovascular	   causes	   at	   PM	   levels	   significantly	  lower	   than	   those	   in	   the	   Harvard	   Six	   Cities	   and	   ACS	   studies	   (Crouse	   et	   al.,	   2012).	  Studies	  of	   this	  sort	   further	  support	   the	  hypothesis	   that	  no	   threshold	  exists	   for	  PM	  exposures	  in	  the	  development	  of	  adverse	  health	  outcomes	  (Pope	  &	  Dockery,	  2006).	  	  	  It	   is	   important	   to	   note	   that	   different	   sources	   of	   PM	   could	   potentially	   result	   in	  different	   morbidity	   and	   mortality	   outcomes.	   Recent	   research	   has	   shown	   that	  combustion-­‐derived	  fine	  PM	  exposure	  is	  one	  of	  the	  most	  detrimental	  components	  of	  air	  pollution	  (Kelly	  &	  Fussell,	  2012).	  More	  specifically,	  some	  studies	  have	  identified	  PM2.5	   to	   have	   significantly	   greater	   effects	   on	   human	   health	   compared	   to	   larger	  particles	   potentially	   due	   to	   the	   fact	   that	   they	   can	  penetrate	   deeper	   into	   the	   lungs	  (Burnett	   et	   al.,	   2000;	   Cifuentes,	   Vega,	  Köpfer,	  &	  Lave,	   2000;	   Schwartz,	  Dockery,	  &	  Neas,	   1996).	   However,	   this	   is	   not	   to	   undermine	   numerous	   studies	   that	   have	  reported	  morbidity	  and	  mortality	  with	  exposure	  to	  coarse	  particles	  (Brunekreef	  &	  Forsberg,	  2005). PM	  Exposure	  and	  Mortality	  	  Annually,	   3.2	   million	   premature	   deaths	   are	   attributable	   to	   ambient	   PM	   exposure	  around	  the	  world;	  PM	  also	  accounts	  for	  3.1%	  of	  global	  disability-­‐adjusted	  life	  years	  (DALYs)	  (Lim	  et	  al.,	  2012).	  It	  has	  also	  been	  estimated	  that	  exposure	  to	  solid	  fuels	  air	  pollutions	   is	   accounts	   for	   3.5	   million	   premature	   deaths	   and	   4.5%	   of	   the	   global	  DALYs	   in	   2010.	   In	   Canada,	   it	   is	   estimated	   that	   over	   21,000	   premature	   deaths	   are	  attributable	  to	  air	  pollution	  each	  year,	  which	  is	  9	  times	  greater	  than	  the	  number	  of	  deaths	  from	  traffic	  accidents	  (Canadian	  Medical	  Association,	  2008).	  	  	  	  	   7	  Exposure	   to	   PM	   has	   been	   linked	   with	   reduced	   life	   expectancy	   and	   increased	  mortality.	  In	  a	  study	  conducted	  in	  the	  United	  States,	  Pope	  et	  al.	  (2009)	  found	  that	  a	  decrease	   of	   10	   µg	   per	   cubic	   meter	   (µg/m3)	   in	   the	   concentration	   of	   fine	   PM	   was	  associated	   with	   an	   increase	   of	   0.61	   year	   in	   life	   expectancy.	   They	   concluded	   that	  reductions	   in	   fine	   PM	   could	   lead	   to	   significant	   improvements	   in	   life	   expectancy	   (	  Pope,	  Ezzati,	  &	  Dockery,	  2009).	  	  In	   a	   landmark	   study,	   the	   association	   between	   air	   pollution	   with	   mortality	   was	  investigated	   in	  six	  U.S.	  cities	  (Dockery	  et	  al.,	  1993).	  As	  a	  prospective	  cohort	  study,	  8,000	  adults	  were	  followed	  for	  14	  to	  16	  years,	  linking	  mortality	  to	  different	  levels	  of	  air	  pollution	  in	  various	  cities.	  The	  authors	  found	  mortality	  rates	  to	  be	  26%	  higher	  in	  the	   most	   polluted	   cities	   compared	   to	   the	   least	   polluted	   one	   (approximately	   19	  µg/m3	   difference	   in	   PM2.5	   concentrations).	   Furthermore,	   mortality	   was	   most	  associated	   with	   fine	   PM	   air	   pollution	   exposure	   compared	   to	   total	   particle	  concentration.	  	  In	  an	  extension	  of	  the	  “U.S.	  six	  cities”	  study,	  Laden	  et	  al.	  (2000)	  directly	  investigated	  the	   association	   of	   traffic	   combustion-­‐derived	   PM2.5	   and	   change	   in	  mortality	   rates.	  The	   results	   indicated	   a	  3.4%	   increase	   in	  daily	  mortality	  per	  10	  µg/m3	   increase	   in	  mobile	  combustion	  source	  PM2.5	   levels.	  They	  concluded	  that	   fine	  particulates	   from	  mobile	   combustion	   sources	  were	   significantly	   associated	  with	   increased	  mortality	  (Laden,	  Neas,	  Dockery,	  &	  Schwartz,	  2000).	  	  	  	  	  	  In	  a	  more	  recent	  cohort	  study	  in	  Canada,	  Crouse	  et	  al.	  (2012)	  looked	  at	  2.1	  million	  individuals	  between	  1991	  and	  2001	  and	  identified	  the	  deaths	  occurring	  during	  this	  period.	   After	   assigning	   PM2.5	   concentrations	   using	   ground-­‐base	   station	   data,	   they	  found	   significant	   increases	   in	   nonaccidental	   and	   cardiovascular	   mortality	   with	  hazard	  ratios	  of	  1.15	  (95%	  CI,	  1.13,	  1.16)	  and	  1.31	  (95%	  CI,	  1.27,	  1.35),	  respectively.	  Their	  results	  are	  similar	  in	  magnitude	  to	  those	  of	  the	  ACS	  study	  where	  over	  550,000	  individuals	  were	  followed	  for	  7	  years	  and	  the	  authors	  found	  significant	  increase	  in	  all-­‐cause,	  cariopulmonary,	  and	  cancer	  mortality	  (Pope	  et	  al.,	  1995).	  	  	   8 PM	  Exposure	  and	  Morbidity	  	  Along	   with	   mortality	   studies,	   numerous	   studies	   have	   evaluated	   the	   morbidity	  related	  to	  exposure	  to	  PM	  air	  pollution.	  With	  the	  growing	  body	  of	  literature,	  we	  are	  gaining	  more	   insight	   into	   the	   diverse	   health	   impacts	   associated	  with	   exposure	   to	  PM.	   Exposure	   to	   PM	   has	   been	   associated	   with	   various	   effects	   on	   physiological	  functions	   and	   subclinical	   symptoms,	   a	   range	   of	   cardiovascular	   conditions	   (e.g.	  myocardial	   infarction	   and	   cardiac	   arrhythmia)	   (Dockery,	   2001),	   exacerbation	   of	  asthma	   (Anderson,	   Favarato,	   &	   Atkinson,	   2013),	   chronic	   obstructive	   pulmonary	  disease	  (Ko	  &	  Hui,	  2012),	  stroke	  (Mateen	  &	  Brook	   ,	  2011),	   lung	  cancer	  (Raaschou-­‐Nielsen	  et	  al.,	  2010,	  2011),	  and	  premature	  births	  (Brauer	  et	  al.,	  2008).	  Moreover,	  the	  International	  Agency	  for	  Research	  on	  Cancer	  (IARC)	  recently	  classified	  outdoor	  air	  pollution	   and	   PM	   as	   known	   human	   carcinogens	   (IARC	   Group	   1)	   due	   to	   the	   link	  between	  exposure	  to	  PM	  and	  increased	  risk	  of	  lung	  cancer	  (Loomis	  et	  al.,	  2013). Air	  Pollution	  and	  Cardiovascular	  Health	  	  Numerous	   studies	   have	   successfully	   linked	   exposure	   to	   air	   pollution	   to	   adverse	  cardiovascular	  outcomes	  and	  mortality.	  Kunzli	  et	  al.	  (2010)	  evaluated	  the	  potential	  cardiovascular	   effects	   of	   living	   in	   close	   proximity	   to	   traffic;	   they	   concluded	   that	  individuals	   living	   within	   100	   meters	   of	   a	   freeway	   have	   a	   rate	   of	   atherosclerosis	  progression	   twice	   that	   of	   the	   general	   population	   (Künzli	   et	   al.,	   2010).	   There	   is	   an	  increasing	  body	  of	  evidence	  linking	  long-­‐term	  exposure	  to	  higher	  levels	  of	  PM	  and	  the	  progression	  of	  atherosclerosis,	  however,	  there	  is	  still	  no	  conclusive	  evidence	  on	  the	  relationship	  (Brook	  et	  al.,	  2010).	  	  There	   have	   been	   a	   number	   of	   Canadian	   studies	   demonstrating	   associations	   with	  cardiovascular	   outcomes;	   in	   a	   cohort	   of	   over	   452,000	   individuals,	   proximity	   to	  major	   roads	  was	   linked	   to	   an	   increase	   in	   coronary	   heart	   disease	   (CHD)	  mortality	  (Gan	  et	  al.,	  2011).	  In	  a	  separate	  study,	  Gan	  et	  al	  (2010)	  found	  that	  moving	  away	  from	  high	  traffic	  roads	  and	  freeways	  reduced	  the	  risk	  of	  CHD	  mortality	  (Gan	  et	  al.,	  2010).	  Moreover,	   several	   other	   studies	   have	   reported	   a	   link	   between	   traffic-­‐related	   air	  	   9	  pollution	   (TRAP)	   exposure	   and	   hypertension	   and	   stroke	   (Baccarelli	   et	   al.,	   2009;	  Fuks	   et	   al.,	   2011).	   There	   is	   also	   evidence	   reporting	   associations	   between	   TRAP	  exposure	  and	  myocardial	   infarction	  (von	  Klot	  et	  al.,	  2011).	  Finally,	  a	  review	  of	   the	  literature	   on	   TRAP	   and	   cardiovascular	   health	   by	   the	   American	   Heart	   Association	  concluded	   that	   there	   seems	   to	   be	   an	   association	   between	   TRAP	   and	   adverse	  cardiovascular	  event	  risk,	  however,	  there	  is	  still	  a	  significant	  need	  for	  more	  research	  to	  identify	  potential	  links	  (Brook	  et	  al.,	  2010).	  	  Despite	  being	  a	  major	  source	  of	  particulate	  air	  pollution,	  limited	  data	  is	  available	  on	  the	   toxicity	   of	   residential	   wood	   combustion	   (RWC)	   PM,	   leading	   to	   the	   false	  impression	   that	   RWC	   is	   not	   an	   important	   source	   of	   toxicity	   compared	   to	   TRAP	  (Kocbach	  Bølling	   et	   al.,	   2009;	  Naeher	   et	   al.,	   2007).	  The	   IARC	  has	   classified	   indoor	  emissions	  from	  household	  combustion	  of	  biomass	  (primarily	  wood)	  as	  a	  group	  2A	  carcinogen	   (i.e.	   probably	   carcinogenic	   to	   humans)	   (International	   Agency	   for	  Research	   on	   Cancer,	   2012).	   There	   has	   been	   some	   evidence	   that	   the	   potential	  adverse	   health	   effects	   as	   a	   result	   of	   WS	   exposure	   in	   developed	   countries	   is	   not	  significantly	   different	   than	   for	   ambient	   particles	   from	   other	   sources	   (Boman,	  Forsberg,	   &	   Järvholm,	   2003;	   Naeher	   et	   al.,	   2007).	   Adding	   to	   the	   complexity	   of	  investigating	  WS	  PM	  is	  the	  fact	  that	  emission	  characteristics	  depend	  on	  both	  the	  fuel	  and	   burn	   conditions	   (Guillén	  &	   Ibargoitia,	   1999).	  With	   regards	   to	   the	   respiratory	  effects	  due	  to	  WS	  particles,	  Naeher	  et	  al.	  (2007)	  point	  out	  that	  at	  the	  present	  time,	  there	  is	  insufficient	  evidence	  indicating	  that	  WS	  particles	  have	  less	  effect	  than	  other	  major	   categories	   of	   combustion	   derived	   particles	   of	   the	   same	   size	   range.	   Various	  studies	  have	  linked	  exposure	  to	  WS	  PM	  with	  respiratory	  health	  outcomes	  including	  respiratory	   symptoms	   in	   children	   (Browning,	   Koenig,	   Checkoway,	   Larson,	   &	  Pierson,	  1990),	  decreases	  in	  lung	  function	  in	  asthmatic	  children	  (Allen	  et	  al.,	  2008;	  Koenig	   et	   al.,	   1993)	   and	   adults	   with	   COPD	   (Trenga	   et	   al.,	   2006),	   and	   increased	  asthma	  emergency	   room	  visits	   (Schwartz,	   Slater,	  Larson,	  Pierson,	  &	  Koenig,	  1993;	  Sheppard,	  Levy,	  Norris,	  Larson,	  &	  Koenig,	  1999).	  Moreover,	  there	  is	  evidence	  of	  an	  association	   with	   low	   birth	   weight	   (Gehring,	   Tamburic,	   Sbihi,	   Davies,	   &	   Brauer,	  	   10	  2014),	   otitis	   media	   (MacIntyre	   et	   al.,	   2011),	   and	   COPD	   hospitalizations	   (Gan,	  FitzGerald,	  Carlsten,	  Sadatsafavi,	  &	  Brauer,	  2013).	  	  However,	  we	  still	  lack	  the	  necessary	  evidence	  to	  comment	  on	  the	  potential	  adverse	  health	  effects	  of	  WS	  particles	  on	  the	  cardiovascular	  system	  and	  cancer	  (Naeher	  et	  al.,	  2007).	  There	  have	  been	  suggestions	  of	   a	  potential	   link	  between	  WS	  exposure	  and	  cardiovascular	  morbidity	  and	  mortality.	  A	  study	  conducted	  in	  a	  predominantly	  WS	  affected	  city	  in	  Australia	  investigated	  the	  relationship	  between	  exposure	  to	  WS	  PM	  and	   cardio-­‐respiratory	   illness	   hospital	   admissions.	   The	   researchers	   found	  significant	   increases	   in	   hospital	   admissions	   with	   increases	   in	  WS	   PM	   (McGowan,	  Hider,	  Chacko,	  &	  Town,	  2002).	  	  In	  another	  study	  conducted	  in	  one	  of	  the	  most	  highly	  WS	  polluted	  cities	  in	  the	  world	  in	  Peru,	  the	  authors	  observed	  significant	  associations	  between	   ambient	   PM	   levels	   and	  hospital	   admissions	   and	   cardiovascular	  mortality	  (Sanhueza	  et	  al.,	  2009).	  	  	  Based	   on	   the	   available	   literature,	   there	   is	   a	   potential	   for	   combustion	   related	   air	  pollution	   to	   be	   responsible	   for	   adverse	   cardiovascular	   health	   outcomes	   in	   the	  general	   population.	  With	   current	   knowledge	   gap	   in	   this	   topic,	   there	   is	   a	   need	   for	  more	  research	  on	  the	  effects	  of	  exposure	  to	  WS	  PM	  and	  cardiovascular	  health.	  1.2 Mechanisms	  of	  Action	  	  The	  human	  lung	  has	  a	  surface	  area	  of	  40-­‐120	  m2,	  which	  is	  exposed	  to	  10-­‐20,000	  L	  of	  ambient	  air	  every	  day.	  The	  lung	  employs	  various	  defense	  mechanisms	  to	  deal	  with	  the	   particles	   that	   deposit	   on	   its	   surface,	   which	   include	   mechanical	   removal	   and	  biochemical	  neutralization	  of	  harmful	  particles.	  Mucociliary	  clearance	  mechanisms	  remove	  the	  larger	  particles	  that	  have	  deposited	  in	  regions	  such	  as	  the	  trachea	  and	  bronchi	   into	   the	   throat,	  which	   can	   then	  be	   swallowed.	  However,	   smaller	   particles	  can	   penetrate	   to	   the	   lower	   airways	   and	   into	   the	   oxygen	   exchange	   region	   of	   the	  lungs;	   they	   are	   then	   engulfed	   by	  macrophages,	   which	   are	   either	   removed	   by	   the	  phagocytosed	  particles	  through	  mucociliary	  clearance	  system	  or	  are	  passed	  through	  	   11	  the	   alveolar	  wall	   into	   the	   lymphatic	   vessels.	   The	   defense	  mechanisms	   of	   the	   lung	  may	  be	  overwhelmed	  by	  the	  toxicity	  of	  the	  particles	  and	  particle	  number	  overload	  (Salvi	  &	  Holgate,	  2001).	  It	  has	  been	  shown	  that	  smaller	  particles	  of	  a	  diameter	  of	  2.5	  µm	   are	   not	   cleared	   effectively	   and	   remain	   in	   the	   lungs	   to	   a	   greater	   extent	   (77%	  retention	  rate)	  compared	  to	  particles	  of	  8.2	  µm	  (15%	  retention	  rate)	  after	  24	  hours	  (Svartengren,	  Linnman,	  Philipson,	  &	  Camner,	  1987).	  Moreover,	  with	   the	  small	  size	  of	   the	   PM2.5,	   these	   particles	   will	   have	   a	   greater	   surface	   area	   increasing	   the	   area	  available	   for	   the	  adsorption	  of	   toxic	   compounds.	  The	   size	  and	   surface	  area	  of	   fine	  particles	  could	  provide	  an	  opportunity	  for	  the	  particles	  to	  impact	  the	  lungs	  and	  the	  whole	  body	  (Kelly	  &	  Fussell,	  2012).	  	  	  Another	  proposed	  mechanism	  indicates	  that	  with	  increased	  particle	  presence	  in	  the	  airways,	  the	  phagocytic	  capacity	  decreases	  which	  leads	  to	  PM	  presence	  in	  the	  lungs	  for	  an	  extended	  time	  and	  hence,	  increase	  the	  interaction	  with	  various	  cell	  types	  such	  as	   epithelial	   cells	   (MacNee,	   Li,	   Gilmour,	   &	   Donaldson,	   1997).	   Particles	   containing	  transition	   metals	   and	   the	   free	   radicals	   produced	   by	   them	   can	   lead	   to	   oxidative	  damage	   and	  macrophage	   activation	   in	   the	   lungs.	   This	   could	   lead	   to	   acute	   cellular	  and	   mediator	   inflammatory	   response	   in	   the	   airways	   through	   the	   release	   of	   pro-­‐inflammatory	  mediators.	  Moreover,	   the	   epithelial	   cells	   phagocytose	   fine	   PM	   upon	  contact	   leading	   to	   an	   increase	   in	   the	   production	   and	   release	   of	   pro-­‐inflammatory	  cytokines,	  strengthening	  the	  inflammatory	  response.	  	  	  Brook	   et	   al.	   (2009)	   propose	   another	   plausible	   mechanism	   for	   the	   cardiovascular	  health	  outcomes	  of	  exposure	  to	  PM2.5.	  According	  to	  their	  hypothesis,	   inhaled	  PM2.5	  interacts	  with	  pulmonary	  irritant	  receptors	  lining	  human	  airways;	  this	  could	  result	  in	   impacts	  via	  the	  autonomic	  nervous	  system	  leading	  to	  prohypertensive	  response	  and	   cardiovascular	   symptoms	   such	  as	   variations	   in	  heart	   rate	   and	  blood	  pressure	  (Brook	  et	  al.,	  2009).	  	  	  A	  third	  hypothesized	  mechanism	  of	  action	  of	  fine	  PM	  is	  through	  penetration	  into	  the	  interstitium.	   After	   entering	   the	   interstitium,	   inflammatory	  mediators	   are	   released	  	   12	  which	   can	   result	   in	   a	   low-­‐grade	   systemic	   inflammatory	   response	   affecting	   blood	  platelets	  and	  clotting	   factors.	  This	  could	  potentially	   lead	  to	  adverse	  cardiovascular	  effects,	   especially	   in	   those	   with	   pre-­‐existing	   cardiovascular	   conditions	   (Salvi	   &	  Holgate,	  2001).	  For	  this	  mechanism,	   it	  has	  been	  hypothesized	  that	  oxidative	  stress	  links	   PM	   exposure	   to	   inflammation,	   endothelial	   dysfunction,	   and	   atherosclerosis	  (Brook	  et	  al.,	  2010).	  TRAP	  particles	  contain	  transition	  metals,	  which	  in	  turn	  produce	  hydroxyl	   radical	   that	   can	   cause	   oxidative	   stress	   and	   inflammation	   (Ghio,	  Stonehuerner,	  Dailey,	  &	  Carter,	  1999;	  Valavanidis,	  Vlahoyianni,	  &	  Fiotakis,	  2005).	  	  1.3 Two	  Major	  Sources	  of	  Combustion-­‐Derived	  PM	  There	   is	   evidence	   of	   both	   combustion	   and	   non-­‐combustion	   emission	   sources	  contributing	  to	  ambient	  PM2.5	  in	  our	  cities.	  The	  major	  sources	  of	  combustion	  derived	  PM	  vary	  in	  different	  parts	  of	  the	  world.	  For	  example,	  in	  China,	  coal	  combustion,	  dust,	  biomass	   aerosol,	   and	   car	   exhaust	   are	   the	   main	   contributors	   of	   PM2.5	   with	   coal	  combustion	  being	   the	   greatest	   emission	   source	   (Meng	  et	   al.,	   2007;	  H.	  Wang	  et	   al.,	  2008;	  Xinhua	  Wang,	  Bi,	   Sheng,	  &	  Fu,	   2006;	   Zheng,	   Salmon,	   Schauer,	   Zeng,	  &	  Kian,	  2005).	  In	  fact,	  in	  many	  parts	  of	  the	  developing	  world	  it	  is	  the	  combustion	  of	  coal	  and	  biomass	   that	   contribute	   extensively	   to	   suspended	   PM2.5.	   In	   North	   America,	   the	  situation	   is	   different,	   with	   traffic-­‐related	   and	   wood	   combustion	   being	   two	   of	   the	  major	  sources	  of	  combustion-­‐derived	  PM.	  Wood	  combustion	  PM	  emissions	  increase	  moving	  north	  in	  North	  America	  (Maykut,	  Lewtas,	  Kim,	  &	  Larson,	  2003;	  Zheng	  et	  al.,	  2005).	   For	   the	   purposes	   of	   this	   study,	   the	   focus	  will	   be	   on	   the	   two	   predominant	  sources	  of	  combustion	  derived	  PM2.5	   in	  North	  America:	  Traffic-­‐related	  PM	  and	  WS	  PM.	  	  1.3.1 Traffic-­‐Related	  PM	  	  More	   and	   more	   scientific	   evidence	   is	   linking	   PM2.5	   exposure	   from	   combustion	  sources,	   which	   include	   traffic,	   industries	   and	   domestic	   heating	   and	   cooking,	   to	  adverse	   cardiovascular	   outcomes	   (Goldberg	   et	   al.,	   2001;	   Han	   &	   Naeher,	   2006).	  Traffic-­‐generated	   emissions	   are	   one	   of	   the	   major	   sources	   of	   exposure	   to	   PM.	   In	  Metro	  Vancouver,	  TRAP	  accounts	  for	  approximately	  12%	  of	  the	  total	  PM	  emissions	  	   13	  in	  the	  region	  (Metro	  Vancouver,	  2010).	  Furthermore,	  approximately	  90%	  of	  the	  PM	  associated	  with	  motor	  vehicles	  is	  in	  the	  PM2.5	  size	  range	  increasing	  the	  possibility	  of	  deeper	   penetration	   into	   the	   lungs	   (Health	   Effects	   Institute,	   2010a).	   It	   has	   been	  estimated	   that	   approximately	   30-­‐45%	   of	   typical	   urban	   populations	   in	   North	  America	   live	   in	  close	  proximity	  to	  highway	  or	  major	  road	  (Health	  Effects	  Institute,	  2010a).	   Estimates	   in	   Canada	   indicate	   that	   about	   32%	   of	   the	   Canadian	   population	  lives	   in	  areas	  of	  high	  exposure	  to	  TRAP	  (Brauer,	  Reynolds,	  &	  Hystad,	  2012;	  Health	  Effects	   Institute,	   2010b).	   Furthermore,	   the	   United	   Nations	   has	   estimated	   that	  approximately	   600	   million	   people	   in	   urban	   areas	   worldwide	   are	   exposed	   to	  dangerous	  levels	  of	  traffic-­‐generated	  air	  pollutants	  (Cacciola,	  Sarvà,	  &	  Polosa,	  2002).	  With	   such	   a	   great	   proportion	   of	   the	   population	   exposed	   to	   air	   pollution	   and	  considering	   the	   overwhelming	   evidence	   on	   potential	   health	   effects,	   TRAP	   is	   an	  important	  public	  health	  issue.	  	  1.3.2 WS	  PM	  	  Based	  on	  the	  available	  estimates,	  over	  half	  of	  the	  world’s	  households	  use	  solid	  fuel	  every	  day.	  Approximately	  95%	  of	  this	  consists	  of	  wood	  and	  agricultural	  residues.	  In	  spite	   of	   its	   more	   prevalent	   use	   in	   developing	   countries,	   residential	   wood	  combustion	  (RWC)	  is	  also	  impacting	  air	  quality	  in	  Canada.	  RWC	  is	  a	  major	  source	  of	  PM	  emissions	   in	  higher	   latitudes	  accounting	   for	  over	  25%	  of	   fine	  PM	  emissions	   in	  Canada	   (Andrea	   Careless,	   2004;	   Barregard	   et	   al.,	   2006b;	   Swiston	   et	   al.,	   2008).	   In	  Metro	   Vancouver,	   WS	   accounts	   for	   25%	   of	   PM2.5	   emissions	   in	   the	   region	   (Metro	  Vancouver,	  2013).	  With	  the	  rising	  costs	  of	  other	  fuel	  options	  and	  having	  wood	  as	  a	  cheaper	   and	   readily	   available	   alternative,	   the	   significance	   of	   RWC	   emissions	   is	  expected	   to	   increase	   in	   the	   future	   (deB.	   Richter	   et	   al.,	   2009;	   Zezima,	   2008).	   The	  available	  data	  on	  RWC	  indicate	  that	  WS	  is	  not	  exclusively	  an	  issue	  in	  the	  developing	  world;	   the	   use	   of	   wood	   and	   other	   biomass	   increased	   at	   an	   annual	   rate	   of	   2.4%	  during	  the	  1990s	  in	  North	  America	  (Naeher	  et	  al.,	  2007).	  	  	  	   14	  Despite	   the	   increasing	   contributions	   of	  WS	   PM	   to	   the	   general	   air	   pollution,	   there	  have	   been	   very	   few	   studies	   comparing	   the	   health	   effects	   of	   TRAP	   with	   WS	   air	  pollution,	   leaving	  a	  wide	  knowledge	  gap	   for	   future	  policymaking.	  This	   is	  primarily	  due	   to	   the	   fact	   that	  many	  pollutants	   are	   shared	  between	  various	   sources.	   In	  most	  cases,	   it	   is	   difficult	   to	   directly	   determine	   the	   degree	   to	   which	   residential	   WS	  contributes	  to	  indoor	  and	  outdoor	  particle	  exposure	  (Kocbach	  Bølling	  et	  al.,	  2009).	  1.4 Surrogates	  of	  TRAP	  and	  WS	  PM	  1.4.1 Hopanes	  	  Hopanes	  are	  specific	  compounds	  contained	  mainly	   in	  the	  hydrocarbon	  fractions	  of	  petroleum	   products	   and	   are	   relatively	   stable	   in	   the	   ambient	   environment.	   The	  presence	   of	   hopane	   biomarkers	   in	   aerosols	   indicates	   a	   fossil	   fuel	   source	   (e.g.	  gasoline	   and	   diesel)	   (Simoneit,	   1999).	   Being	   specific	   to	   fossil	   fuels	   and	   oils,	   these	  tracers	   can	   be	   used	   as	   indicators	   of	   primary	   particle	   emissions	   from	   diesel	   and	  gasoline	   engines	   (Cass,	   1998).	   Schauer	   et	   al.	   demonstrated	   and	   validated	   that	  hopanes	   can	   be	   used	   as	   organic	   tracers	   for	   vehicle	   emissions	   exclusively	   (Rogge,	  Hildemann,	  Mazurek,	  Cass,	  &	  Simoneit,	  1993).	  Hopanes	  have	  also	  been	  used	  in	  some	  validated	  source	  apportionment	  studies	   further	   indicating	  their	  relevance	  to	  TRAP	  (Lin,	  Lee,	  &	  Eatough,	  2010).	  There	  has	  also	  been	  some	  evidence	  on	  the	  similarity	  of	  hopane	   and	   TRAP	   PM	   relationship	  with	   distance	   from	  major	   roads	   and	   highways	  (Olson	   &	   McDow,	   2009).	   The	   authors	   measured	   the	   concentration	   of	   hopanes	   at	  varying	   distances	   from	   a	   highway	   and	   found	   that	   hopanes	   concentration	   was	  greater	  closer	  to	  the	  highway	  and	  decreased	  moving	  away,	  as	  one	  would	  expect	  with	  TRAP	  PM	  concentrations.	  Another	   study	   in	  Texas	   identified	   the	   same	   relationship	  with	   decreasing	   hopanes	   concentration	   with	   increasing	   downwind	   distance	   from	  highway.	   Considering	   the	   available	   evidence,	   hopanes	   can	   be	   considered	   suitable	  organic	   markers	   for	   characterizing	   the	   proportion	   of	   TRAP	   in	   indoor	   PM2.5	  concentrations.	  	  	  	   15	  1.4.2 Optical	  Absorbance	  	  In	   the	   current	   study,	   optical	   absorbance	   was	   measured	   to	   represent	   elemental	  carbon	  (EC),	  which	  is	  a	  component	  of	  traffic	  combustion	  emissions.	  As surrogates of combustion derived PM, studies	   have	   shown	   a	   high	   correlation	   between	   EC	   and	  absorbance	  (Janssen	  et	  al.,	  2000;	  Kinney,	  Aggarwal,	  Northridge,	  Janssen,	  &	  Shepard,	  2000).	  Various	  studies	  have	  also	  found	  an	  association	  between	  exposure	  to	  EC	  and	  morbidity	   and	   mortality	   in	   the	   general	   population	   and	   different	   vulnerable	   sub-­‐populations	  (Fang	  et	  al.,	  2012;	  Geng	  et	  al.,	  2013;	  Janssen	  et	  al.,	  2011;	  Nichols,	  Owens,	  Dutton,	  &	  Luben,	  2013;	  Xi	  Wang	  et	  al.,	  2013).	  Moreover,	  there	  are	  some	  suggestions	  that	  EC	  may	  be	  a	  more	   sensitive	   indicator	  of	   traffic-­‐related	   impacts	  on	  air	  quality	  compared	  to	  PM	  mass	  (Keuken,	  Jonkers,	  Zandveld,	  Voogt,	  &	  Elshout	  van	  den,	  2012).	  	  1.4.3 Levoglucosan	  (LG)	  	  Cellulose	   is	   one	   of	   the	   main	   polymers	   found	   in	   wood	   (50%-­‐70%	   by	   weight)	  (Simoneit,	   1999).	   LG	   is	   a	   by-­‐product	   of	   cellulose	   combustion,	  which	   is	   specific	   to	  combustion	  of	  biomass.	  It	  is	  a	  favorable	  organic	  tracer	  for	  WS	  since	  it	  is	  highly	  stable	  in	  the	  environment	  (Fraser	  &	  Lakshmanan,	  2000).	  It	  has	  also	  been	  used	  in	  various	  studies	  of	  WS	  exposure	  in	  humans	  (Allen	  et	  al.,	  2008;	  Allen,	  Leckie,	  Millar,	  &	  Brauer,	  2009;	  Naeher	  et	  al.,	  2007).	  The	  significance	  of	  this	  organic	  marker	  is	  that	  it	  enables	  us	  to	  determine	  the	  extent	  of	  WS	  influence	  on	  indoor	  PM2.5	  concentrations.	  1.5 Study	  Rationale	  	  Considering	  the	  available	  evidence	  on	  the	  causal	  relationship	  between	  air	  pollution	  and	  morbidity	  and	  mortality,	  we	  need	  to	  evaluate	  potential	  interventions,	  which	  can	  be	   used	   to	  minimize	   exposure	   and	   associated	   health	   risks	   (van	   Erp	   et	   al.,	   2008).	  Among	  various	  adverse	  health	  effects	  related	  to	  exposure	  to	  air	  pollution,	  CVD	  has	  a	  major	  population	  health	  impact.	  Despite	  various	  efforts	  to	  identify	  the	  sources	  of	  PM	  responsible	   for	   cardiovascular	   health	   effects,	   there	   is	   still	   a	   knowledge	   gap	   in	   the	  literature.	  Moreover,	  no	  studies	  have	  directly	  compared	  two	  sources	  of	  PM	  and	  their	  	   16	  potential	   differential	   effects	   on	   cardiovascular	  health,	  which	   further	  demonstrates	  the	  need	  for	  studies	  of	  this	  kind	  with	  regards	  to	  CVD.	  	  	  A	  variety	  of	  successful	  interventions	  have	  been	  used	  during	  the	  past	  few	  decades	  to	  reduce	  TRAP	  and	  RWC	  PM	  emissions	  and	  exposures.	  With	   regards	   to	  TRAP,	   there	  have	   been	   efforts	   to	   change	   individual	   behavior,	   improvements	   to	   combustion	  technology,	   fuel	   standards,	   more	   stringent	   emission	   requirements,	   and	   land-­‐use	  planning	  and	  transportation	  management	  aimed	  to	  decrease	  exposure	  and	  potential	  health	   impacts	   (Brauer	   et	   al.,	   2012;	   Giles	   et	   al.,	   2011).	   Examples	   of	   such	  interventions	  include	  London’s	  congestion	  charge	  scheme	  restricting	  the	  number	  of	  vehicles	  entering	  central	  London	  daily,	  which	  resulted	  in	  a	  reduction	  in	  air	  pollution	  and	  mortality	  (Tonne,	  Beevers,	  Armstrong,	  Kelly,	  &	  Wilkinson,	  2008).	  As	  the	  other	  major	   source	   of	   combustion-­‐derived	   PM2.5	   emissions,	   there	   have	   been	   efforts	   to	  identify	  interventions	  to	  reduce	  exposure	  to	  WS	  PM.	  One	  of	  the	  effective	  measures	  has	   been	   the	   implementation	   of	   woodstove	   exchange	   programs	   to	   replace	   old,	  inefficient	   stoves	   with	   cleaner	   burning	   new	   stoves.	   A	   successful	   wood	   stove	  exchange	   program	   was	   implemented	   in	   Libby,	   Montana	   which	   resulted	   in	   the	  replacement	  of	  over	  90%	  of	  non-­‐certified	  stoves;	  A	  reduction	  of	  20%	  was	  witnessed	  in	   ambient	   fine	   PM	   concentrations	   while	   the	   levels	   of	   LG	   decreased	   by	   50%	  (Bergauff,	   Ward,	   Noonan,	   &	   Palmer,	   2009).	   In	   a	   study	   conducted	   in	   Tasmania,	  Australia,	   wood	   heaters	  were	   gradually	   replaced	  with	   electrical	   heating,	   reducing	  the	   prevalence	   of	   woodstove	   use	   from	   66%	   to	   30%	   in	   all	   households	   (Johnston,	  Hanigan,	  Henderson,	  &	  Morgan,	  2013).	  This	  intervention	  resulted	  in	  a	  40%	  drop	  in	  PM10	   levels	   and	   over	   11%	   decrease	   in	   all-­‐cause	   mortality	   among	   males.	   The	  discussed	  examples	   illustrate	   the	  benefits	   that	   targeted	   interventions	   can	  have	  on	  traffic	  and	  WS	  air	  pollution	  and	  related	  adverse	  health	  outcomes.	  	  Household-­‐level	  exposure	  interventions	  can	  also	  prove	  to	  be	  effective	  with	  regards	  to	  reducing	  air	  pollution	  exposures	  and	  health	  risks.	  Portable	  HEPA	  filters	  appear	  to	  be	  a	  viable	  approach	  to	  reduce	  indoor	  respirable	  particle	  concentrations	  (Barn	  et	  al.,	  2008;	   Sublett	   et	   al.,	   2010;	   Yamada,	   Miyamoto,	   Mori,	   &	   Koizumi,	   1984).	   Some	  	   17	  research	  evidence	  suggests	  that	  HEPA	  filters	  can	  significantly	  reduce	  PM	  exposure	  in	  subjects	  exposed	  to	  high	  TRAP	  (Bräuner,	  Forchhammer,	  et	  al.,	  2008).	  HEPA	  filters	  can	  be	  considered	  a	  practical	  intervention	  for	  reducing	  PM	  exposure	  since	  they	  can	  be	   used	   to	   impact	   the	  most	   susceptible	   populations	   exposed	   to	   high	   levels	   of	   air	  pollution	   within	   a	   community.	   In	   addition,	   they	   are	   relatively	   inexpensive	   to	  purchase	   and	   maintain,	   making	   them	   accessible	   for	   the	   general	   population	   and	  potentially	   low-­‐income	  households	   through	   inexpensive	   subsidies	   (Fisk,	   Faulkner,	  Palonen,	  &	  Seppanen,	  2002).	  	  	  	  Most	   previous	   research	   on	   HEPA	   filter	   interventions	   has	   focused	   on	   indoor	  aeroallergens	   and	   asthma	   aggravation	   (Ezzati	   &	   Kammen,	   2002).	   In	   a	   recent	  systematic	  review	  of	  the	  available	  literature,	  Fisk	  (2013)	  looked	  at	  9	  HEPA	  filtration	  intervention	   studies	   and	   their	   effects	   on	   allergy	   and	   asthma	   symptoms.	   Based	   on	  this	  review,	  there	  were	  significant	  measurable	  drops	  in	  aeroallergens	  with	  filtration.	  The	   author	   concluded	   that	   there	   is	   persuasive	   evidence	   suggesting	   that	   filtration	  systems	  might	  be	  modestly	   effective	   in	   reducing	  adverse	  health	   symptoms	  among	  those	  with	  allergies	  and	  asthma,	  especially	  in	  the	  presence	  of	  large	  allergen	  sources.	  It	  should	  be	  noted	  that	  only	  a	  small	  proportion	  of	  these	  studies	  reported	  statistically	  significant	  improvements	  in	  health	  outcomes,	  with	  most	  being	  small	  to	  moderate	  in	  magnitude.	  Despite	  having	  multiple	  studies	  evaluating	  the	  effect	  of	  HEPA	  filtration	  on	   allergies	   and	   asthma,	   there	   have	   been	   very	   few	   studies	   investigating	   the	  effectiveness	  of	  HEPA	  filter	  for	  improving	  cardiovascular	  health.	  	  	  Current	   available	   evidence	   indicates	   that	   HEPA	   filtration	   inside	   participants’	  residences	  could	  improve	  microvascular	  endothelial	  function	  and	  decrease	  systemic	  inflammation.	  In	  a	  HEPA	  filter	  intervention	  study	  in	  Denmark,	  Brauner	  et	  al.	  (2008)	  reported	   improvements	   in	   endothelial	   function	   after	   reductions	   in	  TRAP	  PM	  with	  indoor	   filtration.	   In	   a	   similar	   study,	  Allen	   et	   al.	   (2011)	   used	  HEPA	   filtration	   as	   an	  intervention	   in	  a	  WS	   impacted	  community	   in	  Canada,	   and	   found	   improvements	   in	  both	  endothelial	   function	  and	  marker	  of	  CVD	   risk	   (C-­‐reactive	  protein	   (CRP))	   after	  reductions	  in	  WS	  PM	  with	  indoor	  filtration.	  The	  similar-­‐sized	  effects	  among	  TRAP-­‐	  	   18	  and	   WS-­‐exposed	   individuals	   was	   the	   initial	   source	   of	   motivation	   for	   the	   current	  study	  to	  evaluate	  TRAP	  and	  WS	  PM	  in	  the	  same	  study.	  	  	  Furthermore,	  it	  has	  been	  shown	  that	  particles	  containing	  transition	  metals	  and	  the	  free	   radicals	   produced	   by	   them	   can	   lead	   to	   oxidative	   damage	   and	   macrophage	  activation	   in	   the	   lungs.	   This	   could	   result	   in	   acute	   cellular	   and	   mediator	  inflammatory	   response	   in	   the	   airways	   through	   the	   release	   of	   pro-­‐inflammatory	  mediators.	  Some	  of	  these	  particles	  can	  penetrate	  through	  into	  the	  interstitium.	  After	  entering	  the	  interstitium,	  inflammatory	  mediators	  are	  released	  which	  can	  result	  in	  a	  low-­‐grade	   systemic	   inflammatory	   response,	   which	   could	   lead	   to	   adverse	  cardiovascular	  effects	  (Salvi	  &	  Holgate,	  2001).	  There	  appears	  to	  be	  more	  transition	  metals	  in	  TRAP	  PM	  compared	  to	  WS	  PM,	  which	  can	  generate	  hydroxyl	  radicals	  and	  lead	  to	  oxidative	  stress	  (Ghio	  et	  al.,	  1999;	  Verma	  et	  al.,	  2009).	  This	  chain	  of	  reactions	  is	   considered	   as	   one	   of	   the	   pathways	   that	   links	   PM	   exposure	   with	   inflammation	  (Brook	  et	  al.,	  2010).	  	  	  In	   addition,	   there	   has	   been	   some	   evidence	   indicative	   of	   differential	   depositional	  patterns	   between	   TRAP	   and	   WS	   PM	   in	   lungs.	   Two	   recent	   human	   experimental	  studies	   concluded	   that	   the	   respiratory	   tract	   deposition	   of	   traffic	   particles	  was	   16	  times	  higher	  than	  residential	  WS	  particles	  in	  addition	  to	  a	  total	  surface	  area	  3	  times	  that	  of	  WS	  particles	  (Löndahl	  et	  al.,	  2008;	  Löndahl	  et	  al.,	  2009);	  there	  is	  also	  in	  vitro	  evidence	   that	   traffic	  particles	   can	  have	  a	   greater	   inflammatory	  effect	  on	   the	   lungs	  (Kocbach,	   Herseth,	   Låg,	   Refsnes,	   &	   Schwarze,	   2008).	   These	   studies	   support	   the	  potential	   for	  differential	   effects	  of	  TRAP	  and	  WS	  PM	  on	  health	  and	   the	  need	   for	   a	  comparison	  between	  these	  two	  sources	  of	  combustion-­‐derived	  PM.	  	  Although	   there	   are	   a	   number	   of	   studies	   that	   have	   evaluated	   the	   relationship	  between	  TRAP	  or	  WS	  PM	  and	   various	   health	   effects,	   it	   is	   very	   difficult	   to	   directly	  compare	   their	   conclusions.	   Different	   studies	   differ	   in	   their	   exposure/outcome	  assessment	   methodologies	   and	   statistical	   analysis,	   hence	   differences	   in	   effect	  estimates	   between	   TRAP	   and	   WS	   exposed	   could	   be	   due	   to	   more	   or	   less	  	   19	  misclassification	  error	  for	  either	  of	  the	  sources.	  To	  further	  complicate	  comparisons,	  the	  size	  of	  exposure	  gradients	  differ	   in	  different	  studies	  making	   it	  very	  difficult	   to	  directly	   compare	   the	   results.	  By	  evaluating	  both	  TRAP	  and	  WS	   in	   the	   same	  study,	  this	   study	   addressed	   the	   first	   issue	   by	   utilizing	   detailed	   in-­‐home	   exposure	  measurements	  for	  both	  groups.	  As	  for	  the	  second	  obstacle,	  HEPA	  filtration	  devices	  were	  used	  to	  produce	  an	  exposure	  gradient	  of	  similar	  magnitude	  in	  both	  groups.	  	  Overall,	  this	  study	  aims	  to	  help	  identify	  the	  most	  important	  sources	  and	  components	  of	   air	   pollution	   mixture	   relevant	   to	   cardiovascular	   health	   and	   to	   assist	   policy-­‐making	  bodies	  target	  the	  key	  components	  to	  more	  effectively	  protect	  public	  health.	  Considering	   this	   overarching	   rationale,	   this	   study	   aims	   to	  determine	   the	  potential	  difference	   in	   subclinical	   cardiovascular	   health	   effects	   between	   WS-­‐	   and	   TRAP-­‐related	   particulate	   exposures,	   using	   HEPA	   filtration	   to	   generate	   a	   PM2.5	  concentration	  gradient,	  and	  to	  evaluate	  the	  use	  of	  HEPA	  filters	  as	  an	  intervention	  for	  TRAP	  and	  WS	  PM	  concentration	  reduction.	  	  1.6 Objectives	  and	  Hypotheses	  	  In	  order	  to	  evaluate	  the	  cardiovascular	  health	  risks	  of	  RWC	  and	  TRAP-­‐generated	  PM	  and	  the	  health	  benefits	  of	  HEPA	  filters	  as	  an	  intervention,	  a	  randomized	  single-­‐blind	  crossover	  study	  was	  designed	  with	  the	  following	  objectives:	  	  	   a) To	  evaluate	  the	  effectiveness	  of	  HEPA	  filters	  in	  reducing	  indoor	  PM	  in	  areas	  where	  TRAP	  and	  RWC	  are	  important	  sources	  	   b) To	  quantify	  the	  relationship	  between	  exposures	  to	  RWC-­‐	  and	  TRAP-­‐derived	  PM	  air	   pollution	   and	   subclinical	   indicators	   of	   cardiovascular	   disease	   (CVD)	  risk	   including	   endothelial	   function	   and	   high	   sensitivity	   C-­‐reactive	   protein	  (hs-­‐CRP),	   interleukin-­‐6	   (IL-­‐6),	   and	   band	   cell	   counts	   (BCC)	   as	   indicators	   of	  systematic	  inflammation;	  	  	   20	  c) To	   compare	   the	   relative	   impact	   of	   HEPA	   filtration	   for	   TRAP	   and	   RWC	   PM	  sources	  on	  microvascular	  endothelial	   function	  and	  hs-­‐CRP,	  IL-­‐6,	  and	  BCC	  as	  indicators	  of	  systemic	  inflammation	  among	  healthy	  adult	  participants.	  	  In	  this	  study,	  we	  hypothesize	  that:	  	   a) The	  use	  of	  indoor	  HEPA	  filtration	  for	  seven	  days	  will	  result	  in	  reductions	  in	  indoor	  concentrations	  of	  PM2.5	  compared	  to	  seven	  days	  of	  no	  filtration.	  	  	   b) A	  reduction	  in	  combustion	  PM2.5	  exposure	  in	  healthy	  adults	  by	  implementing	  indoor	  HEPA	   filtration	  will	   lead	   to	   decreases	   in	   hs-­‐CRP,	   IL-­‐6,	   and	   BCC	   and	  improvements	  in	  microvascular	  function	  	  	   c) There	  will	  be	  a	  greater	  impact	  on	  the	  two	  outcomes	  of	  interest	  (measures	  of	  systemic	   inflammation	   and	   microvascular	   endothelial	   dysfunction)	   among	  those	   living	   in	   TRAP-­‐impacted	   locations	   than	   among	   those	   living	   in	   WS-­‐impacted	  locations.	  	  The	   proposed	   hypotheses	   are	   supported	   by	   the	   available	   literature	   on	   the	  effectiveness	  of	  HEPA	  filters	  in	  reducing	  indoor	  PM	  levels	  (Allen	  et	  al.,	  2011;	  Barn	  et	  al.,	   2008;	   Batterman	   et	   al.,	   2011;	   Bräuner,	  Møller,	   et	   al.,	   2008;	  Weichenthal	   et	   al.,	  2013)	   and	   the	   potential	   for	   the	   greater	   toxicity	   of	   TRAP	  PM	   compared	   to	  WS	  PM	  (Kocbach	  et	  al.,	  2008;	  Löndahl	  et	  al.,	  2008;	  Löndahl	  et	  al.,	  2009).	  	  	  	  	  	  	  	  	   21	  2 Methods	  2.1 Study	  Design	  	  This	  study	  had	  a	  single-­‐blind	  randomized	  crossover	  design	  where	  participants	  were	  blind	   to	   the	   status	   of	   HEPA	   filtration.	   It	   should	   be	   noted	   that	   the	   laboratory	  technicians	  performing	  blood	  analysis	  were	  blinded	  to	  the	  HEPA	  filter	  status	  as	  well.	  HEPA	  filters	  were	  used	  as	  an	  intervention	  to	  reduce	  PM2.5	  exposure	  and	  produce	  an	  exposure	  gradient	  for	  the	  duration	  of	  the	  study.	  For	  this	  study,	  we	  aimed	  to	  recruit	  a	  total	   of	   100	   participants,	   50	   from	   each	   exposure	   category	   (i.e.	   TRAP	   or	   WS).	  Individuals	   were	   recruited	   from	   homes	   in	   high	   TRAP	   or	   high	   RWC	   in	   Metro	  Vancouver,	  which	  were	   identified	  using	  previously	  developed	   air	   pollution	   spatial	  models	  for	  the	  region.	  	  Each	  home	  was	  monitored	  during	  two	  consecutive	  weeks	  for	  a	  total	  of	  14	  days	  with	  one	  HEPA	  filtration	  device	  in	  the	  living	  room	  and	  one	  in	  the	  main	  bedroom.	  During	  one	   7-­‐day	   period	   the	   HEPA	   units	   were	   operated	   with	   a	   HEPA	   filter	   in	   them	   and	  during	   the	   other	   7-­‐day	   period	   there	   was	   no	   HEPA	   filter	   in	   the	   unit	   (i.e.	   placebo	  filtration).	  As	  illustrated	  in	  Fig.	  2-­‐1,	  each	  home	  was	  randomly	  assigned	  to	  one	  of	  the	  two	  possible	  orders	  of	  HEPA/no	  HEPA	  by	  coin	  flip	  on	  the	  first	  day	  of	  sampling:	  one	  group	  where	  there	  was	  HEPA	  filtration	  in	  effect	  during	  the	  first	  week	  and	  no	  HEPA	  in	  the	  second	  week,	  or	  the	  other	  group	  where	  there	  was	  no	  HEPA	  filtration	  during	  the	   first	  week	   and	  HEPA	   filtration	   in	   the	   second	  week.	   In	   each	   home,	   there	  were	  between	  1	  to	  3	  participants	  being	  evaluated	  at	  the	  same	  time.	  	  	   22	  	  Figure	  2-­‐1	  Summary	  of	  the	  Study	  Design	  Air	  pollution	  concentrations	  were	  measured	  both	  indoors	  and	  outdoors	  in	  addition	  to	   the	   study	   subjects’	   time-­‐location	  patterns	   for	   the	   full	   duration	  of	   the	   study	   (i.e.	  two	  seven-­‐day	  samples).	  Furthermore,	  biological	  indicators	  of	  endothelial	  function,	  as	   the	   primary	   outcome	   variable	   of	   this	   study,	   and	   systemic	   inflammation	   were	  assessed	   after	   each	   of	   the	   7-­‐day	   periods,	   in	   addition	   to	   baseline	   for	   systematic	  inflammation.	   The	   data	   from	   these	   assessments	   was	   used	   to	   compare	   each	  individual’s	  assessment	  between	  filtration	  and	  placebo	  sessions.	  	  2.2 TRAP	  and	  WS	  Region	  Selection	  in	  Greater	  Vancouver	  	  	  In	  order	  to	  accurately	  categorize	  postal	  codes	  into	  either	  high	  TRAP/low	  WS	  or	  low	  TRAP/high	  WS,	  we	   needed	   to	   identify	   such	   regions	   in	   Greater	   Vancouver	   Region.	  Spatial	  models	  of	  TRAP	  have	  been	  previously	  developed	   for	   several	   traffic-­‐related	  air	   pollutants	   in	   greater	   Vancouver	   (Henderson,	   Beckerman,	   Jerrett,	   &	   Brauer,	  2007).	   Henderson	   et	   al.	   (2007)	   used	   measurements	   of	   NOx	   and	   light	   absorbing	  carbon	  at	  116	  and	  25	  locations,	  respectively,	   in	  two	  seasons	  to	  develop	  TRAP	  land	  use	  regression	  models	  for	  Metro	  Vancouver.	  Larson	  et	  al.	  (2007)	  developed	  a	  spatial	  model	   for	  WS	   dividing	   the	   region	   into	   three	   tertiles	   of	  WS	   concentrations	   during	  hypothesized	   highest	   emission	   periods.	   They	   used	   nighttime	   PM2.5	   mobile	  monitoring	   in	   winter	   and	   a	   two-­‐week	   average	   PM2.5	   and	   LG	   concentration	  measurements	   to	  develop	   land	  use	  regression	  models	   for	  WS	   in	  Metro	  Vancouver.	  	   23	  Using	  these	  two	  spatial	  models	  enabled	  the	  identification	  of	  regions	  and	  populations	  with	  exposure	  levels	  of	  interest	  and	  minimized	  exposure	  miscategorization	  (Fig.	  2-­‐1).	  	  	   	  	   	  	  Figure	  2-­‐2	  a.	  modeled	  TRAP	  concentrations	  (Henderson	  et	  al.,	  2007);	  b.	  modeled	  WS	  tertiles	  (Larson	  et	  al.,	  2007);	  c.	  high	  TRAP	  and	  low	  WS	  postal	  codes;	  d.	  high	  WS	  and	  low	  TRAP	  postal	  codes	  For	  the	  purposes	  of	  this	  study,	  “traffic	  area”	  was	  defined	  as	  regions	  with	  an	  annual	  average	  NO	  concentration	  of	  >50	  ppb	  and	  WS	  concentration	   in	   the	  bottom	   tertile.	  The	   areas	   meeting	   these	   criteria	   included	   730	   postal	   codes	   and	   approximately	  13,000	  residents.	  The	  “WS	  area”	  was	  defined	  as	  regions	  with	  an	  annual	  NO	  average	  of	  <15	  ppb	  and	  WS	  concentration	  in	  the	  top	  tertile.	  The	  WS	  area	  included	  800	  postal	  codes	   and	  approximately	  23,000	   residents.	  The	   extracted	  postal	   codes	   from	   these	  two	  models	  were	  then	  used	  to	  mail	  out	  informational	  letters	  to	  recruit	  participants.	  	  	  	  a) b) c) d) 	   24	  2.2.1 Home	  Categorization	  Verification	  	  In	   order	   to	   further	   verify	   traffic	   and	   WS	   home	   categorization	   that	   were	   used	   to	  recruit	  participants,	  some	  additional	  analysis	  was	  performed.	  Using	  ArcGIS	  Desktop	  10	   (ESRI,	   Redlands,	   CA:	   Environmental	   Systems	   Research	   Institute)	   and	   maps	  provided	   by	   DMTI	   Spatial	   Inc.	   (Markham,	   Ontario,	   Canada),	   the	   distance	   of	  individual	  homes	  to	  major	  roads	  or	  highways	  were	  extracted.	  In	  addition,	  highway	  or	  major	  road	  length	  at	  varying	  buffers	  was	  calculated.	   In	  general,	   it	  was	  expected	  that	  TRAP	  homes	  be	  closer	  to	  major	  road	  or	  highway	  with	  a	  greater	  length	  of	  such	  roads	  in	  their	  immediate	  vicinity	  compared	  to	  WS	  homes.	  	  2.3 Participant	  Recruitment	  Campaign	  	  In	  order	  to	  recruit	  participants, invitation	  letters	  with	  study	  information	  and	  contact	  information	  were	  mailed	   to	  postal	   codes	   that	  were	   identified	  using	  TRAP	  and	  WS	  models	   described	   earlier	   (Appendix	   I).	   Subjects	   were	   then	   recruited	   after	   a	  telephone	   interview	  (Appendix	  II)	  and	  based	  on	  our	   inclusion	  criteria.	   In	  addition,	  written	  informed	  consent	  was	  obtained	  from	  each	  participant	  prior	  to	  starting	  data	  collection	   (Appendix	   III).	   	  An	  honorarium	  of	   $250	  was	  offered	   to	   each	  participant	  who	  completed	  the	  study.	  In	  addition,	  participants	  were	  promised	  to	  be	  sent	  all	  the	  results	  for	  exposure	  and	  biological	  measurements	  at	  the	  end	  of	  the	  study	  (Appendix	  IV).	  2.4 Study	  Participant	  Characteristics	  	  The	  basic	   inclusion	  criteria	  of	   this	  study	  were	   to	  recruit	   individuals	  over	  19	  years	  old,	   non-­‐smokers,	   and	   those	   living	   in	   the	   target	   areas.	   To	   minimize	   occupational	  exposure	   and	   maximize	   the	   exposure	   reduction	   potential	   of	   the	   HEPA	   filters,	  participant	  priority	  was	  given	  to	   those	  who	  did	  not	  work	  or	  volunteer	  outside	  the	  home.	   In	   order	   to	   meet	   the	   study	   requirement	   of	   recruiting	   healthy	   adults	   and	  having	   a	   homogenous	   study	   population,	   a	   number	   of	   exclusion	   criteria	   were	  followed.	  	  	   25	  Individuals	   were	   excluded	   if	   they	   had	   any	   condition	   that	   might	   affect	   exposure,	  cardiovascular	   health	   outcomes,	   and/or	   microvascular	   endothelial	   function.	  Subjects	   with	   recent	   surgeries,	   diabetes,	   heart	   disease,	   hypertension,	   metabolic	  syndrome,	   asthma,	   and	   COPD	   in	   addition	   to	   pregnant	   women	   were	   excluded	   as	  these	  conditions	  might	  have	  an	  effect	  on	  systemic	   inflammation.	  Moreover,	  due	   to	  the	   fact	   that	   we	   measured	   endothelial	   function	   using	   probes	   on	   fingertips	  (explained	   in	   more	   detail	   in	   the	   following	   sections),	   individuals	   with	   Reynaud’s	  syndrome	   were	   excluded	   as	   this	   condition	   might	   interfere	   with	   accurate	  measurements	   and	   was	   excluded.	   In	   addition,	   those	   in	   occupations	   with	   high	  exposures	  to	  air	  pollution	  such	  as	  bus	  drivers	  and	  mechanics	  were	  excluded.	  Finally,	  individuals	  taking	  anti-­‐inflammatory	  medications	  were	  also	  excluded.	  	  The	   goal	   of this study was	   to	   recruit	   a	   homogenous	   population	   of	   healthy	   adults.	  Moreover,	   despite	  having	   some	  of	   the	  highest	  TRAP	   concentrations	  of	   the	   greater	  Vancouver	  region	  in	  Downtown	  Vancouver,	   this	  area	  was	  excluded	  from	  the	  study	  since	   most	   residences	   are	   in	   high-­‐rise	   buildings.	   There	   is	   some	   evidence	   that	  indicates	   ground-­‐level	   exposure	   might	   not	   be	   well	   correlated	   with	   levels	  determined	  in	  high-­‐rise	  buildings	  due	  to	  vertical	  concentration	  gradients	  (Restrepo	  et	  al.,	  2004;	  Villena	  et	  al.,	  2011).	  Based	  on	  the	  same	  evidence,	  residences	  above	  the	  third	  floor	  were	  also	  excluded.	  	  2.5 Participant	  Preparation	  Before	  Sampling	  	  At	  the	  beginning	  of	  each	  sampling	  session,	   the	  sampling	  plan	  was	  explained	  to	  the	  residents	   in	   detail	   by	   both	   the	   environmental	   and	   the	   health	   technician.	   The	  environmental	   technician	   explained	   the	   devices	   being	   placed	   in	   the	   home	   and	  discussed	   the	   location	   of	   sampling	   equipment	   with	   the	   participants.	   The	   health	  technician	   reviewed	   the	   consent	   form	   with	   the	   participants	   and	   obtained	   their	  signature.	   Afterwards,	   the	   dwelling	   information	   form	   was	   completed	   with	  participants	   (Appendix	   V).	   Furthermore,	   time-­‐location-­‐activity	   diaries	   were	  explained	   to	   the	   participants	   ensuring	   that	   they	   understand	   how	   they	   should	   be	  filled	   out	   on	   a	   daily	   basis	   for	   the	   duration	   of	   the	   study	   (Appendix	   VI).	   The	   study	  	   26	  protocol	   was	   approved	   by	   the	   ethics	   committee	   at	   Simon	   Fraser	   University	  (Appendix	  VI,	  File	  #	  2011S0431).	  2.6 Study	  Period	  All	  exposure	  and	  health	  measurements	  were	  conduced	  between	  December	  5th,	  2011	  to	  August	  21st,	  2012.	  WS	  homes	  were	  sampled	  from	  December	  to	  April	  while	  TRAP	  homes	  were	  sampled	  from	  December	  to	  August.	  2.7 HEPA	  Filtration	  One	   of	   the	  most	   important	   factors	   in	  maximizing	   the	   efficiency	   of	  HEPA	   filtration	  devices	  is	  choosing	  a	  device	  with	  sufficient	  filtered	  air	  delivery	  for	  the	  room	  size	  of	  interest.	  Hence,	  we	  used	  two	  different	  models	  of	  filtration	  devices	  suitable	  for	  areas	  they	  were	  placed	  in.	  In	  the	  bedroom,	  smaller	  devices	  (18150,	  Honeywell,	  Tennessee,	  US)	  were	  used,	  which	  are	  effective	  for	  rooms	  up	  to	  197	  square	  feet	  in	  size.	  For	  the	  main	   living	   room,	   larger	   devices	   (50250/50300,	   Honeywell,	   Tennessee,	   US)	  were	  operated,	  which	  are	  effective	  for	  rooms	  up	  to	  390	  square	  feet	  in	  size	  (Figure	  2-­‐3).	  	  	   	  	  	  	  	  	  	  	  	  	  	   	  Figure	  2-­‐3	  HEPA	  Filtration	  Devices	  Used	  –	  Honeywell	  50250/50300	  (Left)	  &	  Honeywell	  18150	  (Right)	  	  In	   the	   current	   study,	   HEPA	   filtration	   devices	   were	   mainly	   used	   to	   introduce	   a	  concentration	   gradient	   of	   PM2.5	   between	   filtration	   and	   no	   filtration	   weeks.	   The	  concentration	  gradient	  was	  then	  used	  to	  characterize	  the	  potential	  change	  in	  health	  effects	  with	  a	  reduction	  in	  PM2.5	  exposure.	  In	  addition,	  we	  investigated	  the	  potential	  	   27	  difference	  in	  efficiency	  of	  HEPA	  filtration	  for	  removal	  of	  PM	  from	  two	  major	  sources	  of	  combustion-­‐derived	  PM2.5	  (i.e.	  TRAP	  and	  WS).	  	  2.7.1 Electricity	  Use	  Meter	  	  An	   electricity	  meter (Kill A Watt P4400, P3 International, NY, USA) was	   installed	  between	  the	  HEPA	  filtration	  device	  and	  the	  power	  outlet.	  The	  purpose	  of	  this	  device	  was	  to	  assist	  with	  estimating	  the	  usage	  of	  the	  device	  during	  the	  study	  period.	  Prior	  to	  the	  study,	  each	  of	  the	  different	  filtration	  devices	  was	  operated	  for	  one	  week	  in	  a	  laboratory	   setting.	   The	   amount	   of	   electricity	   used	   was	   recorded	   for	   three	   output	  setting	  on	  each	  device	  (i.e.	  low,	  medium,	  and	  high).	  After	  each	  sampling	  session,	  the	  amount	  of	  electricity	  used	  was	  compared	  to	  previously	  estimated	  electricity	  usage	  at	  each	  setting.	  The	  homes	  had	  to	  meet	  at	  least	  the	  approximated	  electricity	  use	  at	  the	  low	  setting	  for	  each	  type	  of	  device	  to	  verify	  that	  filtration	  devices	  were	  used	  for	  the	  duration	  of	  this	  study.	  	  2.8 Health	  and	  Exposure	  Measures	  	  Various	  health	  and	  exposure	  measures	  were	  included	  in	  this	  study.	  At	  the	  beginning	  of	  each	  sampling	  session,	  the	  sampling	  plan	  was	  explained	  to	  the	  residents	  in	  detail	  by	   both	   the	   environmental	   and	   the	   health	   technician.	  Moreover,	   a	   health	   log	  was	  completed	  with	  the	  health	  technician’s	  assistance	  at	  the	  end	  of	  each	  sampling	  week	  (Appendix	   VIII).	   Table	   2-­‐1	   presents	   the	   health	   and	   exposure	   measures	   used,	  rationale	  for	  use,	  and	  measurement	  methods.	  	  	  	   28	  Table	  2-­‐1	  Health	  and	  Exposure	  Measures	  	  	  2.9 Exposure	  Measurements	  2.9.1 PM2.5	  	  Gravimetric	  indoor	  and	  outdoor	  samples	  of	  PM2.5	  were	  collected	  for	  the	  duration	  of	  each	  7-­‐day	  session	  for	  all	  homes.	  Sampling	  was	  conducted	  using	  a	  Harvard	  Impactor	  (Air	  Diagnostics	  and	  Engineering	  Inc.,	  Naples,	  Maine,	  USA)	  attached	  to	  a	  SKC	  Leland	  Legacy	  pump	  operated	  at	  a	  flow	  rate	  of	  10	  liter/minute.	  Each	  pump	  was	  calibrated	  using	  a	  BIOS	  DryCal	  Defender	  520	  (Mesa	  Laboratories	  Inc.,	  Colorado,	  USA)	  at	  each	  sampling	   location	   after	   a	   five-­‐minute	   warm	   up	   period	   and	   using	   the	   assembled	  sampling	   train	  with	   all	   components	   in	   place.	   In	   short,	  DryCal	  was	   attached	   to	   the	  Harvard	  Impactor	  with	  a	  calibration	  cap	  while	  the	  pump	  was	  running.	  If	  necessary,	  the	  pump’s	  flow	  rate	  was	  adjusted	  to	  reflect	  10	  liters/min	  on	  the	  calibrator.	  	  	  	  	   29	  Samples	  were	  collected	  on	  37	  mm	  2µm	  pore	  size	  Pall	  Teflo	  membrane	  filters	  with	  ring.	  Each	  filter	  was	  placed	  in	  a	  filter	  cassette	  holder	  in	  a	  laboratory	  setting	  and	  then	  placed	  in	  a	  sealed	  petri	  dish	  for	  transport.	  The	  sampling	  train	  was	  assembled	  upon	  arrival	   in	   each	   home	   and	   a	   filter	   (in	   its	   cassette)	   was	   placed	   in	   the	   Harvard	  Impactor.	   After	   each	   seven-­‐day	   sampling	   period	   and	   prior	   to	   removing	   the	   filter,	  post-­‐flow	   of	   the	   pump	   was	   measured	   using	   BIOS	   DryCal	   Defender	   (Mesa	  Laboratories	   Inc.,	   Colorado,	   USA)	   and	   its	   calibration	   cap.	   Finally,	   the	   filter	   was	  removed	   from	   the	  Harvard	   Impactor	   and	   secured	   in	   its	   respective	   petri	   dish.	   The	  petri	  dish	  was	  placed	  in	  a	  Ziploc	  bag	  and	  in	  a	  foam	  padded	  box	  for	  transport	  to	  the	  OEH	  laboratory	  at	  UBC	  (Allen,	  Karlen,	  &	  Nichol,	  2014).	  2.9.2 Indoor	  Set	  Up	  	  All	  indoor	  equipment	  was	  placed	  in	  the	  living	  room	  where	  there	  would	  be	  minimal	  disruption	  to	  the	  residents	  and	  the	  sampling	  equipment.	  Due	  to	  the	  amount	  of	  noise	  produced	   by	   the	   sampling	   equipment	   and	   despite	   the	   fact	   that	   individuals	   spend	  more	   time	   in	   the	   bedroom,	   exposure	  measurements	  were	   conducted	   in	   the	   living	  room.	   Instruments	   were	   placed	   a	   meter	   away	   from	   walls,	   corners,	   windows,	   air	  conditioners,	  and	  any	  other	  ventilation	  outlet/inlets.	  	  The	  HEPA	   filtration	   devices	  were	   also	   placed	   in	   areas	   of	   the	   living	   room	   and	   the	  main	   bedroom	   where	   there	   would	   be	   minimal	   interference	   and	   least	   direct	   air	  current	   to	   occupants.	  Moreover,	   special	   care	  was	   taken	   to	  maximize	   the	   distance	  between	   air	   sampling	   instruments	   and	   HEPA	   filtration	   devices.	   A	   detailed	  description	   of	   the	   indoor	   set	   up	   is	   included	   in	   the	   supplementary	   standard	  operating	  procedure	  (SOP)	  document	  (Study	  Protocol	  III)	  (Allen	  et	  al.,	  2014).	  	  2.9.3 Gravimetric	  Analysis	  of	  the	  Filters	  	  The	   filters	   were	   equilibrated	   for	   at	   least	   48	   hours	   in	   humidity	   and	   temperature	  controlled	  room	  (35±5%	  RH,	  22±3℃)	  at	  the	  School	  of	  Population	  and	  Public	  Health	  (SPPH)	   Occupational	   and	   Environmental	   Health	   laboratory	   in	   the	   University	   of	  	   30	  British	  Columbia	  prior	   to	  measuring	  pre-­‐	  and	  post-­‐weights.	  After	   the	  equilibration	  period,	   a	   radioactive	   alpha	   emitter	   neutralizer	   (NRD	   LLC.,	   Grand	   Island	   NY)	   was	  used	  for	  approximately	  2	  seconds	  to	  remove	  any	  potential	  static	  charge	  present.	  The	  filters	  were	   then	  weighed	   in	   triplicate	   using	   a	   Sartorius	  M3P	  microbalance	   (1	   µg	  resolution,	  ±	  2	  µg	  sensitivity).	  Triplicate	  weights	  were	  required	  to	  be	  within	  10	  µg	  of	  each	  other	  otherwise	  the	  filters	  had	  to	  be	  reweighed	  until	  the	  criterion	  was	  met.	  The	  average	   of	   each	   triplicate	   was	   used	   as	   the	   final	   weight.	   Pre-­‐weighed	   filters	   were	  placed	   in	   clean	   petri	   dishes,	   which	   were	   labeled	   with	   a	   letter	   followed	   by	   three	  digits	   (e.g.	   A100).	   All	   pre-­‐weight	   filters	   were	   stored	   in	   the	   humidity	   and	  temperature	   controlled	   room	   and	   removed	   only	  when	  needed	   for	   sampling.	   After	  sampling,	   the	   filters	  were	   removed	   from	  Harvard	   Impactors	   and	  placed	   into	   their	  respective	   petri	   dishes,	   and	   equilibrated	   for	   at	   least	   48	   hours	   prior	   to	   weighing,	  following	  the	  same	  procedure	  as	  used	  for	  pre-­‐weights	  (Allen	  et	  al.,	  2014). Quality	  Control	  (QC)	  Filters	  	  Prior	  to	  each	  filter	  weighting	  session,	  three	  previously	  weighed,	  unused	  filters	  were	  used	   as	  QC	   filters.	   The	  weight	   of	   each	  QC	   filter	  was	   then	   compared	   to	   previously	  developed	  mean,	  warning	  (mean	  ±	  2SD),	  and	  control	  (mean	  ±	  3SD)	  of	  all	  previous	  repeated	  weightings	  of	  the	  same	  filters.	  Upon	  verifying	  that	  all	  triplicate	  pre-­‐weights	  of	  the	  QC	  filters	  are	  with	  the	  control	  limits,	  the	  gravimetric	  analysis	  of	  filters	  would	  commence	  (Allen	  et	  al.,	  2014). Field	  and	  Lab	  Blanks	  	  Field	  blanks	  were	  collected	  at	  the	  first	  home	  and	  then	  every	  second	  home	  visit	  after	  that	  (i.e.	  Home	  1,	  3,	  5,	  etc.).	  Field	  blanks	  were	  prepared	  by	  inserting	  an	  unused	  filter	  into	   a	   filter	   cassette	   and	   then	   into	   a	   clean	   and	   assembled	  Harvard	   Impactor.	   The	  Harvard	  Impactor	  was	  closed	  without	  turning	  the	  sampling	  pump	  on.	  After	  waiting	  for	   approximately	   30	   seconds,	   the	   filter	  was	   removed	   from	   the	  Harvard	   Impactor	  and	  filter	  cassette	  and	  placed	  back	  into	  a	  labeled	  and	  clean	  petri	  dish.	  The	  field	  blank	  was	  treated	  as	  a	  regular	  sample	  filter	  after	  this	  point	  (Allen	  et	  al.,	  2014).	  	   31	  Mean	  weight	  of	  field	  blanks	  was	  deducted	  from	  all	  filter	  weights	  prior	  to	  performing	  further	   calculations.	   The	   same	   procedure	   was	   followed	   for	   absorbance,	   hopanes,	  and	  LG	  concentrations.	  	  2.9.4 Harvard	  Impactor	  Maintenance	  	  Subsequent	   to	   each	   sampling	   session,	   the	   Harvard	   Impactors	   were	   completely	  disassembled.	  All	  parts	  of	   the	   inside	  and	  outside	  were	  cleaned	  with	  Kimwipes	  and	  alcohol.	   The	   impaction	   plates	   were	   cleaned	   using	   soap	   distilled	   water	   and	   an	  ultrasonic	   cleaner	   for	   a	  minimum	  of	  15	  minutes.	  After	   sonication,	   the	  plates	  were	  rinsed	  thoroughly	  using	  distilled	  water	  (3	  times)	  (Allen	  et	  al.,	  2014).	  2.9.5 Optical	  Reflectance	  Analysis	  	  All	   filters	   were	   analyzed	   at	   SPPH	   Occupational	   and	   Environmental	   Health	  Laboratory	   at	   the	   University	   of	   British	   Columbia	   for	   reflectance	   using	   a	   Diffusion	  Systems	  Ltd.	  Smoke	  Stain	  Reflectometer	  (M43D).	  After	  weighing	  the	  filters	  and	  prior	  to	   hopanes/LG	   analysis,	   the	   reflectance	  of	   all	   filters	  was	  measured.	  After	   cleaning	  the	  measuring	  head,	  mask,	  and	  standard	  plate	  with	  alcohol,	   the	  reflectometer	  was	  calibrated	  using	  five	  control	  filters.	  One	  filter	  was	  used	  to	  adjust	  the	  reflectometer	  to	  100.0	   followed	   by	  measuring.	   The	   other	   four	   control	   filters	  without	   adjusting	   the	  reflectometer	   reading.	   The	   control	   filter	   having	   the	  median	   reflectance	   value	  was	  selected	  as	  the	  primary	  control	  filter	  to	  be	  used	  to	  recalibrate	  the	  device	  after	  every	  25	   filters.	   After	   calibration,	   reflectance	   measurements	   were	   taken	   at	   five	   points	  across	  each	  filter	  (location	  previously	  decided).	  If	  the	  standard	  deviation	  (SD)	  of	  the	  collected	   data	   was	   over	   0.5	   units,	   the	   process	   was	   repeated	   again.	   A	   detailed	  protocol	  for	  measuring	  absorbance	  can	  be	  found	  in	  the	  supplemental	  SOP	  document	  (Allen	  et	  al.,	  2014).	  2.9.6 Optical	  Reflectance	  and	  Absorbance	  Reflectance	  values	  were	  converted	  to	  absorbance.	  The	  following	  formula	  was	  used	  to	  calculate	  absorbance	  before	  analysis:	  	  	   32	  α	  =	  [A/2V]	  .	  ln[Rf/Rs]	  	  where	  α	  is	  absorbance	  in	  10-­‐5	  m-­‐1	  ,	  A	  is	  area	  of	  the	  filter	  in	  m2,	  V	  is	  volume	  sampled	  in	  m3,	   Rf	   	   is	   average	   reflectance	   of	   field	   blank	   filters,	   and	   Rs	   is	   reflectance	   of	   the	  sample	  filter	  as	  a	  percentage	  of	  100.0.	  	  2.9.7 Hopanes	  Laboratory	  Analysis	  	  	  All	   filters	   were	   analyzed	   for	   Hopanes	   at	   SPPH	   Occupational	   and	   Environmental	  Health	  Laboratory	  at	  the	  University	  of	  British	  Columbia.	  Initially,	  each	  filter	  was	  cut	  in	   half	   using	   scissors,	   with	   each	   half	   placed	   in	   a	   separate	   clean	   Petri	   dish.	  Afterwards,	   the	  plastic	   ring	  was	   removed	   from	  one	  half	   of	   the	   filter	   and	   the	   filter	  was	   transferred	   to	   an	   extraction	   vessel	   (one	   for	   Hopane	   analysis	   and	   one	   for	   LG	  analysis).	  Approximately	  1	  mL	  of	  isooctane	  was	  added	  to	  the	  extraction	  vessel	  and	  ultrasonicated	  for	  20	  minutes.	  After	  spiking	  all	  samples	  with	  50	  µL	  of	  pyrene	  stock	  solution,	  each	  sample	  was	  transferred	  to	  gas	  chromatography	  (GC)	  vials	  for	  analysis.	  The	   analyses	  were	   conducted	   using	   gas	   chromatography	  mass	   spectrometry	   (GC-­‐MS)	   as	   per	   study	   protocol	   XI	   in	   the	   supplementary	   SOP	   document	   (Allen	   et	   al.,	  2014).	   Finally,	   in	   order	   to	   reduce	   the	   number	   of	   missing	   values	   ,	   hopane	  measurements	  <LOD	  were	  replaced	  with	  LOD/2	  values	  for	  further	  analysis.	  2.9.8 LG	  Laboratory	  Analysis	  	  All	   filters	   were	   analyzed	   for	   LG	   at	   SPPH	   Occupational	   and	   Environmental	   Health	  Laboratory	   at	   the	  University	   of	   British	   Columbia.	   The	   previously	   prepared	   half	   of	  the	  filter	  was	  transferred	  to	  an	  extraction	  vessel.	  2	  mL	  of	  ethyl	  acetate	  was	  added	  to	  the	  vessel	  and	  it	  was	  then	  ultrasonicated	  for	  30	  minutes.	  100	  µL	  of	  the	  extract	  was	  transferred	   to	   GC	   vials;	   subsequently,	   pyridine	   (15	   µL)	   and	   MSTFA	   in	   1%	   TMCS	  solution	   (30	   µL)	   was	   added	   to	   the	   vial.	   After	   vortexing	   the	   mixture	   for	   10-­‐20	  seconds,	  the	  samples	  were	  placed	  in	  a	  dark	  location	  for	  at	  least	  6	  hours.	  As	  the	  last	  step	   before	   analyzing	   the	   samples,	   each	   vial	  was	   spiked	   by	   10	   µL	   of	   tri-­‐isopropyl	  benzene.	  The	  analyses	  were	  conducted	  using	  GC-­‐MS	  as	  per	  study	  protocol	  XII	  in	  the	  supplementary	   SOP	   document	   (Allen	   et	   al.,	   2014).	   Finally,	   in	   order	   to	   reduce	   the	  	   33	  number	   of	   missing	   values	   ,	   LG	   measurements	   <LOD	   were	   replaced	   with	   LOD/2	  values	  for	  further	  analysis.	  2.9.9 Indoor	  Temperature	  and	  Relative	  Humidity	  (RH)	  	  Indoor	   temperature	   and	  RH	  were	   logged	   continuously	   for	   both	   one-­‐week	   periods	  using	   HOBO	   data	   loggers (UX100, Bourne, MA).	   One-­‐minute	   averages	   of	  temperature	  and	  RH	  were	  collected.	  The	  test	  data	  were	  downloaded	  after	  each	  two-­‐week	   period	   and	   measurements	   from	   the	   appropriate	   time	   were	   extracted	   for	  further	  analysis.	  	  2.9.10 Time-­‐Location-­‐Activity	  Log	  	  Participants	  were	  asked	  to	  complete	  a	  self-­‐reported	  time-­‐location-­‐activity	  log	  at	  30-­‐minute	  intervals.	  The	  items	  in	  this	  log	  included	  participants’	  locations	  (e.g.	  at	  home,	  commuting,	  etc.),	  mode	  of	  transport,	  proximity	  to	  potential	  sources	  of	  PM	  during	  the	  day	   (e.g.	   candles,	   cooking,	   tobacco,	   etc.),	  window	   and	   air	   conditioning	   status,	   and	  activities	  such	  as	  using	  a	  wood-­‐burning	  stove.	  	  	  This	  diary	  was	  used	  to	  identify	  other	  potential	  sources	  of	  exposure	  to	  PM	  that	  could	  affect	  the	  efficiency	  of	  HEPA	  filtration	  in	  addition	  to	  unusual	  peaks	  in	  results.	  It	  was	  expected	  that	  those	  who	  spend	  most	  of	  their	  time	  indoors	  at	  home	  would	  have	  the	  greatest	  impact	  from	  filtration	  especially	  since	  commuting	  can	  have	  a	  large	  impact	  on	  exposure	  to	  TRAP.	  Moreover,	  some	  medications	  and	  illnesses	  have	  been	  shown	  to	   affect	   systematic	   inflammation	   and	   endothelial	   function	   (Delfino	   et	   al.,	   2009;	  Pearson	  et	  al.,	  2003;	  Widlansky,	  Gokce,	  Keaney	  Jr,	  &	  Vita,	  2003),	  hence	  participants	  were	   requested	   to	   complete	   a	  weekly	   health	   diary	   recording	   any	  medication	   and	  supplement	  intake	  in	  addition	  to	  any	  health	  symptoms.	  	  2.10 Health	  Measurements	  	  Various	  markers	  of	  inflammation	  were	  chosen	  based	  on	  previous	  research	  and	  the	  available	   literature.	   There	   are	   several	   cytokines	   involved	   in	   initiating	   an	  	   34	  inflammatory	  response	   including	   IL-­‐6	  (Gabay	  &	  Kushner,	  1999;	  van	  Eeden,	  Yeung,	  Quinlam,	   &	  Hogg,	   2005).	   This	   response	   involves	   other	   downstream	   inflammatory	  proteins	   such	   as	   CRP.	   Moreover,	   increased	   number	   of	   band	   cells	   in	   the	   blood	  suggests	  an	  increase	  in	  the	  activity	  of	  the	  bone	  marrow	  to	  initiate	  an	  inflammatory	  response	   (Swiston	   et	   al.,	   2008;	   Tan	   et	   al.,	   2000).	   Hence,	   CRP,	   IL-­‐6,	   and	   BCC	  were	  included	  in	  this	  study	  to	  cover	  different	  components	  of	  the	  inflammatory	  response.	  	  In	  addition	  to	  the	  main	  health	  measurements	  discussed	  below,	  general	  physiological	  measurements	   and	   variables	   were	   also	   collected.	   Self-­‐reported	   age,	   height,	   and	  weight	  were	  recorded;	  BMI	  was	  calculated	  using	  the	  obtained	  information.	  For	  each	  study	   subject	   heart	   rate,	   systolic	   and	   diastolic	   blood	   pressures	  were	  measured	   at	  least	  10	  minutes	  before	  measuring	  endothelial	  function	  to	  avoid	  potential	  effects	  on	  microvascular	   endothelial	   function.	   Finally,	   indoor	   temperature	   at	   the	   time	   of	  conducting	   the	  EndoPAT	  was	   recorded	  since	   temperature	  might	  affect	  endothelial	  function	  measurements.	  	  2.10.1 Microvascular	  Endothelial	  Function	  	  	  As	   our	   primary	   endpoint,	   vascular	   function	   was	   evaluated	   at	   the	   end	   of	   the	   first	  week	   and	   at	   the	   end	   of	   the	   second	   week.	   In	   order	   to	   evaluate	   microvascular	  endothelial	   function,	   peripheral	   arterial	   tonometry	   (PAT)	   was	   performed	   using	   a	  portable	  EndoPAT	  2000	  device	  (Itamar	  Medical	  Ltd.,	  Cesari,	  Israel).	  This	  device	  uses	  finger	  pneumatic	  sensors	  to	  detect	  changes	  in	  pulse	  wave	  amplitude	  (PWA)	  during	  a	  period	   of	   induced	   reactive	   hyperemia.	   EndoPAT	   2000	   was	   used	   as	   per	   the	   SOP	  provided	  by	  Itamar	  Medical	  Ltd.	  In	  short,	  the	  subjects	  were	  laid	  down	  in	  a	  dark	  and	  quiet	  room	  where	  pneumatic	  finger	  probes	  were	  attached	  to	  the	  index	  finger	  of	  each	  hand.	   Recordings	   were	   conducted	   by	   the	   device	   for	   15	   minutes,	   including	   a	   5-­‐minute	   baseline	   period,	   a	   5-­‐minute	   period	   with	   an	   inflated	   blood	   pressure	   cuff	  restricting	  blood	   flow,	   and	  a	  5-­‐min	  post-­‐occlusion	  period.	  The	   reactive	  hyperemia	  index	   (RHI)	   score	  was	   then	   automatically	   calculated	   by	   built	   in	   algorithms	   in	   the	  EndoPAT	  2000	  software.	  	   35 Post-­‐Measurement	  Processing	  of	  EndoPAT	  Data	  	  Generally,	   the	   algorithm	   embedded	   in	   the	   EndoPAT	   2000	   software	   automatically	  calculates	  an	  RHI	  score	  after	  each	  complete	  session.	  Data	  were	  screened	  to	  identify	  incomplete	  occlusion,	  noisy	  signal,	  and	  non-­‐standard	  occlusion	  length	  and	  produce	  a	   	  “corrected”	  RHI	  score,	   if	  possible.	  Otherwise,	  the	  RHI	  score	  was	  used	  directly	  as	  calculated	  by	  the	  software.	  	  For	   all	   participants	   and	   their	   EndoPAT	   evaluations,	   the	   PWA	   and	   software	   errors	  were	  manually	   inspected	   to	   find	   potential	   issues	   with	   the	   scores.	   For	   the	   results	  with	  non-­‐standard	  occlusion	  length,	  the	  occlusion	  duration	  was	  manually	  set	  at	  5:00	  minutes	  by	  moving	  the	  end	  border	  in	  the	  software.	  In	  the	  case	  of	  having	  occlusion	  times	   >5:30	   minutes,	   occlusion	   time	   was	   manually	   changed	   to	   5:00	   minutes	   (3	  subjects	  only).	  The	  results	  were	  then	  re-­‐analyzed	  by	  the	  software	  and	  an	  RHI	  score	  was	  produced.	  2.10.2 Blood	  Collection,	  Processing,	  and	  Analysis	  	  Blood	   samples	   were	   collected	   after	   EndoPAT	   measurements	   by	   a	   trained	  phlebotomist	   (1x10	  ml	  gold	  SST	  and	  1x6	  ml	   lavender	  EDTA	   tube	  per	  participant).	  All	   blood	   samples	   were	   kept	   in	   a	   cooler	   with	   ice	   packs	   until	   transferred	   to	   the	  iCAPTURE	  laboratory	  at	  St.	  Paul’s	  Hospital	  for	  storage	  at	  -­‐80°	  	  prior	  to	  analysis.	  	  	  All	   blood	   samples	   were	   analyzed	   by	   the	   UBC	   James	   Hogg	   Research	   Centre	   at	   St.	  Paul’s	   Hospital.	   Blood	   samples	   were	   processed	   within	   4-­‐6	   hours	   after	   collection.	  Blood	  sample	  processing	  involved	  obtaining	  a	  complete	  blood	  count	  report	  from	  the	  EDTA	  blood	  tube,	  preparation	  of	  Wright’s	  stain	  (EDTA	  tube),	  removal	  of	  SST	  serum	  (gold	  SST	  tube),	  removal	  of	  the	  EDTA	  plasma	  (EDTA	  tube),	  and	  the	  removal	  of	  the	  Buffy	   coat	   from	   EDTA	   tubes.	   Each	   of	   these	   fractions	   were	   then	   aliquoted	   into	  Eppendorf	   tubes	   and	   stored.	   	   The	   blood	  was	   analyzed	   for	   CRP,	   IL-­‐6,	   and	   BCC,	   as	  markers	   of	   systematic	   inflammation.	   Blood	   sample	   collection,	   processing,	   and	  	   36	  analysis	  have	  been	  discussed	   in	   study	  protocols	  XIV	  and	  XV	   in	   the	   supplementary	  SOP	  document	  (Allen	  et	  al.,	  2014). CRP	  	  Samples	  were	   analyzed	   for	   CRP	   on	   a	   fluorescent-­‐coded	   bead-­‐based	   immunoassay	  testing	   platform,	   following	   the	   manufacturer’s	   protocol	   (R&D	   Systems,	   Catalog	  Number	  LUCB000)	  with	  the	  following	  modifications:	  	   • Study	  plasma	  was	  used	  in	  the	  assay	  • Samples	   were	   diluted	   200-­‐fold	   rather	   than	   the	   manufacturer’s	   suggested	  100-­‐fold	  Prior	   to	   plasma	   use	   in	   the	   assay,	   the	   samples	   were	   vortexed	   and	   centrifuged	   at	  10,000x	  g	  for	  1	  minute	  in	  order	  to	  remove	  viscous	  material	  that	  might	  clog	  the	  filter	  plate	  and	  machine	  probe.	  Luminex	  100	  was	  used	  for	  all	  analysis. IL-­‐6	  	  Similar	   to	   CRP	   analysis,	   the	   samples	  were	   analyzed	   on	   a	   fluorescent-­‐coded	   bead-­‐based	   immunoassay	   testing	   platform.	   The	   manufacturer’s	   protocol	   (Millipore	  HSCYTO-­‐60SK)	  was	  used	  for	  IL-­‐6	  measurements	  with	  the	  following	  modification:	  	   • Undiluted	  study	  plasma	  was	  used	  in	  the	  assay	  	  Prior	   to	   plasma	   use	   in	   the	   assay,	   the	   samples	   were	   vortexed	   and	   centrifuged	   at	  10,000x	  g	  for	  1	  minute	  in	  order	  to	  remove	  viscous	  material	  that	  might	  clog	  the	  filter	  plate	  and	  machine	  probe.	  Luminex	  100	  was	  used	  for	  all	  analysis. BCC	  	  Manual	   BCC	   was	   performed	   by	   a	   trained	   laboratory	   technician	   at	   the	   iCAPTURE	  laboratory	  at	  St.	  Paul’s	  Hospital,	  Vancouver,	  BC.	  Thin	  blood	  smears	  were	  prepared,	  dipped	   into	  methanol	   and	   air	   dried,	   and	   stained	  with	  Wright-­‐Giemsa	   stain	   (Bayer	  	   37	  HEMA-­‐TEK	  2000	   Slide	   Stainer,	   Leverkusen,	  Germany)	   and	   counted	  under	   the	  40x	  objective	  of	  a	  light	  microscope.	  2.11 Statistical	  Analysis	  	  For	  the	  purposes	  of	  this	  study,	  Excel	  2011	  was	  used	  to	  prepare	  the	  initial	  and	  final	  database.	   JMP®	  10	   (SAS	   Institute	   Inc.,	   Cary,	  NC)	  was	   used	   to	   conduct	   all	   required	  statistical	  analyses.	  2.11.1 Data	  Cleaning	  and	  Descriptive	  Statistics	  	  Prior	   to	   data	   analysis,	   all	   datasets	   were	   reviewed	   to	   ensure	   quality.	   In	   general,	  individuals	   with	   incomplete	   or	   poor	   quality	   data	   for	   one	   or	   both	   of	   the	   sessions	  were	  excluded	  from	  final	  analysis.	  This	  approach	  enabled	  a	  “cleaner”	  data	  analysis	  with	  matched	   data	   and	  without	   having	   to	   account	   for	   characteristics	   such	   as	   age,	  BMI,	  and	  sex. Exposure	  Data	  	  First,	   any	   homes	   with	   incomplete	   exposure	   data	   collections	   were	   marked	   as	  “excluded”.	  PM2.5	  sampling	  data	  were	  then	  reviewed	  to	  ensure	  appropriate	  sampling	  times	  and	  flow	  rates	  for	  the	  duration	  of	  the	  study.	  Technician’s	  sampling	  logs	  were	  inspected	  to	  detect	  any	  irregularities	  in	  sampling	  which	  warranted	  the	  exclusion	  of	  data	  points	  from	  the	  analysis.	  Samples	  were	  excluded	  if	  any	  of	  the	  following	  applied:	  	   • The	  tube	  was	  disconnected	  from	  the	  pump	  at	  any	  point	  during	  the	  study,	  	  • The	  sampling	  pump	  had	  failed,	  • The	  sampling	  period	  was	  less	  than	  9,000	  minutes	  per	  week	  of	  sampling	  (90%	  of	  a	  week),	  • The	   pump	   flow	   rate	  was	   found	   to	   be	   10%	  higher	   or	   lower	   than	   the	   initial	  10L/min,	  	  (e.g.	  disconnected	  hose,	  pump	  failure,	  and	  short	  sampling	  period)	  	  	  	   38 Health	  Data	  	  As	   with	   the	   exposure	   data,	   any	   health	   variable	   with	   incomplete	   information	   for	  either	  or	  both	  sessions	  was	  marked	  as	  “excluded”	  for	  that	  particular	  health	  endpoint	  analysis.	  Moreover,	  the	  EndoPAT	  records	  were	  reviewed	  to	  ensure	  proper	  occlusion	  time	  and	  the	  absence	  of	  background	  PWA	  noise.	  2.11.2 Distribution	  of	  Exposure	  and	  Health	  Variables	  	  Box	   plots	   of	   all	   exposure	   and	   health	   variables	   were	   produced	   using	   JMP®	   10	   to	  examine	  the	  normality	  of	  the	  distribution	  and	  also	  the	  presence	  of	  outliers.	  For	  log-­‐normally	   distributed	   health	   variables	   (i.e.	   outcome	   variables),	   the	   data	   were	   log-­‐transformed	  before	  any	  subsequent	  analysis.	  Moreover,	  descriptive	  statistics	  were	  tabulated	  for	  exposure	  and	  health	  variables.	  2.11.3 Extreme	  Values	  	  For	  outliers	  lying	  outside	  mean	  ±	  2SD	  of	  all	  relevant	  variable	  values	  in	  the	  dataset,	  the	  spreadsheets	  were	  reviewed	  and	  compared	  to	  field	  notes	  to	  ensure	  the	  accuracy	  of	  each	  data	  point.	  Weekly	  diaries	  were	  also	  checked	  for	  other	  potential	  sources	  that	  could	  have	   affected	   values	   (e.g.	   indoor	   smoking,	  wood	  burning,	   infections,	   etc.).	   If	  feasible,	   gravimetric	   and	   laboratory	   analyses	   were	   re-­‐done	   to	   ensure	   that	   the	  results	  of	  both	  analyses	  match.	  Finally,	  if	  no	  reasonable	  and	  logical	  explanation	  for	  extreme	  values	  was	  available,	  they	  were	  included	  in	  the	  analysis	  but	  were	  marked	  for	  potential	  sensitivity	  analysis.	  2.11.4 Scaling	  Exposure	  Contrasts	  	  In	   order	   to	   compare	   the	   magnitude	   of	   the	   effects	   across	   different	   exposure	   (i.e.	  PM2.5,	   hopanes,	   LG,	   and	   absorbance),	   the	   exposure	   contrasts	   were	   scaled	   to	   the	  median	   within-­‐home	   change	   between	   filter	   and	   no	   filter	   conditions.	   In	   order	   to	  achieve	  this,	  change	  in	  exposure	  from	  filter	  to	  no	  filter	  was	  calculated	  for	  all	  homes	  with	  complete	  exposure	  and	  health	  data.	  Subsequently,	  all	  respective	  measurements	  were	  divided	  by	  the	  median	  value	  of	  change	  in	  exposure	  among	  all	  homes.	  	  	  	  	  	   39	  2.11.5 Evaluation	  of	  Difference	  Between	  Home	  Types	  and	  Filtration	  Status	  	  To	   further	   characterize	   the	   significance	   of	   differences	   in	   various	   exposure	   and	  health	  variables	  between	  TRAP	  and	  WS	  homes	  and	  filtration/no	  filtration	  sessions,	  paired	   t-­‐tests	   were	   performed.	   To	   illustrate	   the	   general	   characteristics	   and	  homogeneity	   of	   the	   participants	   in	   each	   home	   type,	   summary	   statistics	   of	   study	  population	  characteristics	  (e.g.	  age,	  BMI,	  blood	  pressure,	  %	  time	  at	  home,	  etc.)	  were	  calculated.	  Moreover,	   summary	  of	  exposure	  characteristics	   (e.g.	  PM2.5,	  absorbance,	  LG,	   hopanes,	   etc.)	   were	   calculated	   by	   HEPA	   filtration	   status	   and	   home	   type.	   The	  same	  procedure	  was	   followed	   for	   the	  health	  outcomes	   included	   in	   this	   study	   (e.g.	  RHI,	  CRP,	  IL-­‐6,	  BCC,	  and	  %PMN).	  2.11.6 Mixed	  Model	  Analysis	  	  Mixed	   models	   were used to properly	   account for	   measurements	   clustered	   within	  participants	  and	  participants	  clustered	  within	  homes.	  Due	  to	  the	  crossover	  design	  of	  this	   study	   and	   the	   fact	   that	   only	   individuals	  with	   complete	   data	   for	   both	   sessions	  (i.e.	  paired	  data)	  were	   included	   in	   the	   final	  analysis,	   there	  was	  no	  need	   to	   include	  time	  invariate	  variables	  (e.g.	  age,	  sex	  and	  BMI)	  in	  the	  model	  to	  account	  for	  potential	  confounding.	  Thus,	  we	  included	  only	  time-­‐varying	  adjustment	  variables	  (Subject	  ID,	  Home	  ID,	  and	  average	  indoor	  temperature)	  in	  the	  model.	  The	  general	  mixed	  model	  for	  measurement	  i	  on	  participant	  j	  living	  in	  home	  k	  was:	  	  	   logYijk	  =	  αj	  +	  ɣk	  +	  β0	  +	  β1HEPAijk	  +	  β2	  	  Ind.Tempijk	  +	  eijk	  	  where	  αj	  and	  ɣk	  are	  random	  participant-­‐	  and	  home-­‐specific	  intercepts,	  respectively,	  and	  β1	  represents	  the	  fixed	  effect	  of	  HEPA	  filtration	  on	  the	  log-­‐transformed	  outcome	  variable,	   logYijk.	   The	  HEPA	   variable	  was	   replaced	  with	   other	   continuous	   exposure	  variables	   (e.g.	   PM2.5	   concentration,	   LG	   concentration,	   etc.)	   for	   different	  models.	   β2	  represents	   the	   fixed	   effect	   of	   average	   indoor	   temperature	   on	   the	   log-­‐transformed	  outcome	  variable,	  logYijk	  .	  	  	  	   40	  Mixed	  models	  were	  prepared	  for	  the	  a	  priori	  primary	  outcome	  variable	  of	  this	  study,	  which	  was	  RHI.	  Several	  other	  mixed	  models	  were	  created	  using	  secondary	  outcome	  variables	   including	   CRP,	   IL-­‐6,	   BCC,	   and	   %PMN.	   RHI,	   CRP,	   and	   IL-­‐6	   were	   log-­‐transformed	   since	   they	   were	   skewed.	   For	   each	   of	   these	   outcome	   variables,	   four	  exposure	  variables	  were	  included	  in	  the	  models:	  HEPA	  filtration	  status	  (binary),	  and	  absorbance,	  LG,	  and	  PM2.5	  (continuous).	  	  In	  order	  to	  evaluate	  the	  differences	  in	  health	  effects	  of	  exposures	  between	  TRAP	  and	  WS	  homes,	  the	  following	  model	  was	  used:	  	  	   logYijk	  =	  αj	  +	  ɣk	  +	  β0	  +	  β1HEPAijk	  +	  β2Groupijk	  +	  β3(HEPA	  x	  Group)ijk	  +	  β4	  	  Ind.Tempijk	  +	  eijk	  	  where	  “Group”	  is	  a	  binary	  variable	  representing	  WS	  or	  traffic	  exposed	  homes.	  Table	  2-­‐2	  below	  summarizes	   the	  main	  mixed	  effect	  models	   that	  were	  used	   in	   this	  study.	   For	   each	   of	   the	   listed	   explanatory	   variables,	   a	   separate	   mixed	   model	   was	  prepared.	  	  	   41	  Table	  2-­‐2	  Summary	  of	  the	  mixed	  effect	  models	  used	  in	  the	  analysis	  	  2.11.7 Effect	  Modification	  by	  Age,	  BMI,	  and	  Gender	  The	   possibility	   of	   effect	   modification	   by	   age,	   BMI,	   and	   sex	   was	   also	   explored	   in	  stratified	  analyses.	  Each	  of	  these	  variables	  were	  converted	  to	  binary	  variables	  using	  median	   age	   (above	   and	   below	   42	   years	   old),	   normal	   BMI	   (above	   and	   below	   25	  kg/m2),	  and	  sex.	  	  2.11.8 Interpreting	  Effect	  Estimates	  from	  Log	  Transformed	  Models	  	  Interpreting	   the	   results	   of	   a	   regression	  model	   when	   the	   outcome	   variable	   is	   log-­‐transformed	   is	   not	   as	   simple	   as	   looking	   at	   the	   effect	   estimates.	   Interpretation	   of	  such	  models	   can	  be	   challenging	   since	  back	   transforming	   the	  estimates	  will	   not	  be	  sufficient	  either.	   In	  order	   to	   interpret	   the	   log-­‐transformed	  outcomes	   the	   following	  formula	  was	  used:	  	  	   %	  change	  in	  outcome	  per	  1-­‐unit	  change	  in	  predictor	  =	  100*(eβ	  –	  1)	  	   42	  where	   β	   is	   the	   effect	   estimate	   for	   the	   exposure	   of	   interest	   (e.g.	   PM2.5).	   Using	   this	  formula	   and	   by	   scaling	   exposure	   contrasts	   as	   discussed	   above,	  we	   can	   report	   the	  results	  of	  the	  model	  as	  %	  change	  in	  outcome	  per	  unit	  median	  change	  in	  exposure.	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   43	  3 Results	  Overall,	   13,200	   invitation	   letters	   were	   mailed	   out	   to	   residences	   in	   the	   targeted	  postal	   codes	   across	   Greater	   Vancouver	   Region.	   From	   those	   who	   called	   back	  requesting	   information	   from	   the	   study	   coordinator	   and	   responding	   to	   screening	  interview	  questions,	  83	  subjects	  between	  the	  ages	  of	  19-­‐72	  years	  old	  (54	  from	  TRAP	  homes	  and	  29	  from	  WS	  homes)	  living	  in	  44	  different	  residences,	  met	  the	  inclusion	  criteria	   for	   this	   study.	   Twenty-­‐eight	   homes	  were	   located	   in	   high-­‐TRAP	   areas	   and	  sixteen	  were	  in	  high-­‐WS	  areas.	  	  3.1 Summary	  Statistics	  –	  Participants	  with	  Complete	  Data	  	  Overall,	   out	  of	  83	  participants,	   68	   individuals	  with	   complete	  data	   for	   all	   exposure	  and	  health	  outcome	  variables	  of	   interest	  were	   included	   in	   the	  RHI	  analysis;	   forty-­‐eight	  of	  these	  subjects	   lived	  in	  high	  TRAP	  homes	  while	  20	  lived	  in	  high	  WS	  homes	  (Figure	   3-­‐1).	   Of	   the	   15	   excluded	   participants,	   11	   were	   excluded	   due	   to	   sampling	  duration	  not	  meeting	  the	  criteria	  of	  this	  study	  (i.e.	  90%	  of	  the	  week),	  2	  due	  to	  filter	  damage,	  1	  due	  to	  participant	  drop	  out,	  and	  1	  due	  to	  missing	  pump	  data	  (exposure	  calculation	  not	  possible).	  There	  were	  also	  2	  individuals	  with	  missing	  RHI	  measures,	  which	   overlapped	   with	   exposure	   exclusions.	   For	   CRP,	   IL-­‐6,	   and	   BCC	   models,	   15	  participants	  were	  excluded	  due	  to	  exposure	  data	  issues	  (mentioned	  above)	  and	  an	  additional	   16	   participants	   due	   to	   the	   lack	   of	   blood	   work	   data	   for	   either	   or	   both	  sampling	   sessions.	   Of	   the	   16	   individuals	   with	   missing	   CRP,	   IL-­‐6,	   and	   BCC,	   4	  participants	   were	   excluded	   due	   to	   inability	   to	   process	   the	   samples	   (laboratory	  closure),	   10	   participants	   excluded	   for	   the	   lack	   of	   blood	   samples	   (no	   samples	  collected),	   and	   2	   participants	   excluded	   since	   insufficient	   blood	  was	   collected.	   	   All	  participants	  had	  a	  mean	  (±SD)	  age	  of	  43.8	  ±	  12.8	  years	  and	  53%	  were	  female.	  	  There	  were	   also	   approximately	   equal	   numbers	   of	   male	   and	   female	   participants	   in	   both	  TRAP	  and	  WS	  homes.	  The	  WS	  exposure	  participants	  were	  significantly	  older	  (mean	  age	  40.8	  years	  vs.	  51.3	  years)	  and	  spent	  slightly	  more	  time	  indoors	  at	  home	  but	  the	  two	  groups	  did	  not	  differ	  in	  any	  other	  measured	  characteristics	  (Table	  3-­‐1).	  The	  age	  range	  for	  all	  participants	  was	  between	  19	  to	  72	  years	  old.	  	   44	  	  Table	  3-­‐1	  summarizes	  study	  population	  characteristics	  for	  all	  participants	  and	  also	  participants	  stratified	  by	  home	  type.	  P-­‐values	  reflect	  t-­‐test	  for	  TRAP	  versus	  WS.	  	  Table	  3-­‐1	  Summary	  of	  study	  population	  characteristics	  with	  complete	  data	  	  	  Mean	   BMI	   was	   approximately	   25	   for	   each	   group,	   which	   reflects	   the	   high	   end	   of	  normal	  BMI.	  Mean	  systolic	  and	  diastolic	  blood	  pressures	  were	  equal	  and	  within	  the	  range	   of	   normal	   blood	   pressure	   at	   <120	  mmHg	   and	   <80	  mmHg,	   respectively.	   On	  average,	   all	   participants	   spent	   about	   three-­‐fourth	  of	   their	   time	  at	  home	  with	  both	  groups	  being	  similar;	  this	  is	  particularly	  important	  since	  the	  greater	  amount	  of	  time	  participants	   spent	   at	   home,	   the	   greater	   the	   potential	   effect	   of	   HEPA	   filtration	   on	  their	  health.	  Furthermore,	  it	  should	  be	  noted	  that	  the	  baseline	  CRP	  levels	  were	  not	  significantly	  different	  between	  the	  two	  groups	  	  	   45	  3.2 Summary	  Statistics	  –	  Participants	  with	  Incomplete	  Data	  	  After	   reviewing	   all	   the	   required	   data	   for	   running	   mixed	   statistical	   models,	   15	  participants	  had	  incomplete	  data	  for	  RHI	  models,	  missing	  values	  for	  one	  or	  more	  of	  the	  variables	  needed.	  It	  is	  important	  to	  note	  that	  with	  regards	  to	  outcome	  variables,	  if	   a	   subject	   was	   missing	   a	   value,	   that	   person	   was	   excluded	   only	   for	   the	   model	  including	   that	   specific	   outcome	  variable.	   For	   example,	   if	   subject	  1	  was	  missing	   an	  RHI	   measurement,	   he	   was	   only	   excluded	   from	   the	   RHI	   model	   and	   not	   models	  including	   CRP	   or	   IL-­‐6.	   The	   same	   analogy	   was	   used	   for	   deciding	   to	   exclude	  participants	  for	  the	  lack	  of	  exposure	  variables.	  In	  order	  to	  compare	  the	  population	  characteristics	   of	   those	   with	   incomplete	   data	   with	   those	   with	   complete	   data,	  summary	   statistics	   were	   prepared	   for	   those	   with	   incomplete	   data.	   Table	   3-­‐2,	  provides	   a	   summary	   of	   the	   differences	   in	   population	   characteristics	   between	  participants	   with	   complete	   and	   incomplete	   data.	   As	   seen	   in	   this	   table,	   the	   study	  population	   with	   incomplete	   data	   is	   similar	   to	   those	   with	   complete	   data,	   with	   no	  statistically	  significant	  differences	  between	  the	  two	  groups.	   It	   is	  worth	  mentioning	  that	  BMI	  and	  %	  time	  at	  home	  are	  borderline	  significantly	  different	  between	  the	  two	  groups.	   Overall,	  we	   can	   assume	   that	  we	   have	   not	   selectively	   excluded	   individuals	  with	  different	  characteristics	  and	  potentially	  affecting	  the	  outcomes.	  	  	   46	  Table	  3-­‐2	  Comparison	  of	  study	  population	  characteristics	  between	  participants	  with	  complete	  and	  incomplete	  data	  for	  RHI	  	  3.3 Exposure	  Characteristics	  Table	   3-­‐3	   provides	   a	   summary	   of	   different	   exposure	   characteristics	   stratified	   by	  HEPA	   filtration	   status.	   General	   outdoor	   PM2.5	  levels	  were	   similar	   in	   two	   sampling	  sessions	  confirming	  the	  absence	  of	  variability	  between	  sampling	  periods.	  With	  the	  implementation	  of	   indoor	  HEPA	   filters,	  85%	  of	   the	  homes	  had	  reductions	   in	  PM2.5	  concentrations	  and	  15%	  of	  the	  homes	  having	  increase	  in	  PM2.5	  levels;	  all	  WS	  homes	  had	  reduced	  levels	  while	  79%	  of	  TRAP	  homes	  had	  dropped	  PM2.5	  levels.	  There	  was	  a	  statistically	  significant	  reduction	  of	  over	  40%	  in	  PM2.5	  concentrations	  with	  a	  drop	  of	  36%	  in	  TRAP	  homes	  and	  48%	  in	  WS	  homes.	  Similar	  to	  PM2.5,	  an	  approximately	  8%	  reduction	  in	  absorbance	  was	  found	  which	  was	  also	  statistically	  significant.	  No	  such	  relationship	  was	  found	  with	  regards	  to	  hopanes	  and	  LG	  as	  the	  markers	  of	  TRAP	  and	  WS,	  respectively.	  However,	  considering	   the	  median	  values	   for	   indoor	  hopanes	  and	  LG,	  there	  was	  a	  clear	  decreasing	  trend	  in	  concentrations	  with	  HEPA	  filtration.	  	  Variations	   in	   outdoor	   temperatures	   can	   potentially	   affect	   air	   pollution	  concentrations	  indoors	  as	  a	  result	  of	  increased	  infiltration	  into	  residences	  (Allen	  et	  	   47	  al.,	   2012).	   However,	   the	   results	   of	   the	   current	   study	   indicate	   that	   outdoor	  temperatures	  were	  quite	  stable	  and	  similar	  between	  the	  filtration	  and	  no	  filtration	  weeks	  reducing	  the	  potential	  for	  variations	  in	  infiltration	  between	  the	  two	  sessions.	  Moreover,	   %	   time	   spent	   at	   home	   or	   in	   transit	   was	   statistically	   similar	   during	  filtration	  and	  no	   filtration	  weeks	   indicating	   similar	   filtration	  exposure	  and	   similar	  exposure	  to	  other	  sources	  of	  PM2.5	  outside	  home.	  	   48	  Table	  3-­‐3	  Summary	  of	  exposure	  characteristics	  by	  HEPA	  filtration	  status	  	  	  	   49	  The	  SDs	  of	  indoor	  and	  outdoor	  concentrations	  of	  hopanes	  and	  LG	  were	  quite	  wide.	  This	  observation	  was	  due	  to	  the	  presence	  of	  large	  outliers	  in	  each	  of	  these	  variables	  for	  which	  there	  was	  no	  reasonable	  explanation	  to	  exclude	  from	  the	  analysis.	  As	  with	  any	   other	   laboratory	   analysis	   procedure,	   hopane	   and	   LG	   concentration	  measurement	   was	   limited	   by	   a	   method	   limit	   of	   detection	   (LOD).	   Overall,	   the	  concentration	   of	   hopanes	  was	   less	   than	   the	   LOD	   for	   38.2%	   of	   the	   filters.	   For	   LG,	  16.5%	  of	  the	  filters	  had	  concentrations	  less	  than	  the	  method	  LOD.	  	  	  This	   might	   have	   lead	   to	   biased	   values	   undermining	   the	   usefulness	   of	   such	   data.	  Moreover,	   there	   was	   clear	   relationship	   between	   temperature	   and	   hopane	  concentrations,	   which	   further	   complicated	   interpretation	   of	   the	   hopane	  measurements.	  Hence,	  we	  decided	  to	  use	  hopane	  data	  exclusively	  for	  confirming	  the	  categorization	  of	  postal	  codes	  into	  TRAP	  or	  WS	  for	  participant	  recruitment.	  	  	  To	  better	  characterize	  different	  exposures,	  exposures	  were	  stratified	  by	  home	  type	  and	  HEPA	   filtration.	   Table	   3-­‐4	   summarizes	   exposure	   characteristics	   in	  TRAP-­‐	   and	  WS-­‐exposed	   homes.	   Mean	   indoor	   and	   outdoor	   PM2.5	   levels	   were	   not	   significantly	  different	  between	  TRAP	  and	  WS	  homes.	  Indoor	  absorbance	  was	  significantly	  higher	  in	   high-­‐TRAP	   homes	   (p-­‐value	   <0.05)	   compared	   to	   those	   in	   the	   high-­‐WS	   areas.	  Although	   not	   statistically	   significant,	   both	   indoor	   and	   outdoor	   LG	   levels	   were	  greater	   in	  WS	   homes.	   Median	   outdoor	   LG	   levels	   were	   approximately	   three	   times	  greater	   in	  WS	   areas	   compared	   to	   TRAP	   areas.	   Hopanes	   measurements	   also	   have	  large	  SD	  with	  non-­‐significant	  differences	  between	  TRAP	  and	  WS	  homes.	  However,	  both	  indoor	  and	  outdoor	  median	  hopane	  levels	  in	  TRAP	  homes	  were	  over	  twice	  that	  of	  WS	  homes.	  	  Both	  mean	  indoor	  and	  outdoor	  temperatures	  are	  different	  between	  the	  two	  groups	  (P-­‐value	  =	  0.0001).	  Mean	  outdoor	  temperature	  was	  lower	  in	  WS	  homes	  due	  to	  the	  fact	   that	  wood	   is	  primarily	  used	  during	   the	  colder	  months	  of	   the	  year	   for	  heating,	  which	  is	  when	  sampling	  was	  conducted.	  The	  same	  reasoning	  can	  be	  applied	  to	  the	  small	  drop	  in	  mean	  indoor	  temperature	  for	  WS	  homes.	  	  	   50	  	  Based	  on	  the	  results	  from	  paired	  t-­‐tests,	  mean	  %	  time	  spent	  at	  home	  and	  mean	  %	  time	  spent	  in	  transit	  are	  equal	  between	  TRAP	  and	  WS	  homes	  with	  p-­‐values	  greater	  than	  0.05.	  It	  is	  also	  worth	  mentioning	  that	  participants	  in	  both	  TRAP	  and	  WS	  homes	  spent	   approximately	   75%	   of	   their	   time	   at	   home;	   this	   implies	   similar	   exposure	   to	  HEPA	  filtration	  at	  home.	  	   51	  Table	  3-­‐4	  Summary	  of	  exposure	  characteristics	  by	  home	  type	  (No	  HEPA)	  	  	  	  	   52	  Finally,	   Table	   3-­‐5	   provides	   a	   summary	   of	   the	   effect	   of	   HEPA	   filtration	   on	   various	  exposures	  in	  both	  TRAP-­‐	  and	  WS-­‐exposed	  homes.	  Indoor	  PM2.5	  concentrations	  were	  reduced	   in	   both	   TRAP	   and	  WS	   homes	   to	   different	   extents	  with	   an	   approximately	  36%	  drop	  in	  PM2.5	  in	  TRAP	  homes	  and	  48%	  in	  WS	  homes.	  Mean	  outdoor	  PM2.5	  levels	  were	  very	  close	  in	  TRAP	  and	  WS	  homes	  for	  both	  HEPA	  versus	  no	  HEPA	  weeks.	  With	  wide	  confidence	  intervals,	  there	  are	  no	  clear	  trends	  for	  HEPA	  filtration	  effectiveness	  with	   regards	   to	  hopanes	  and	  LG	   in	  either	  home	   type.	  There	  appears	   to	  be	  a	   small	  decrease	  in	  indoor	  absorbance	  in	  TRAP	  homes.	  	  	   53	  Table	  3-­‐5	  Summary	  of	  exposure	  characteristics	  by	  HEPA	  filtration	  status	  and	  home	  type	  	  	  As	  expected,	  there	  was	  higher	  indoor	  absorbance	  in	  TRAP	  homes	  compared	  to	  WS	  homes,	  with	  small	  reductions	  in	  both	  home	  types	  with	  HEPA	  filtration.	  With	  regards	  	   54	  to	  absorbance,	  there	  seems	  to	  be	  a	  variation	  between	  HEPA	  and	  no	  HEPA	  weeks	  but	  it	   should	   be	   noted	   that	   during	   the	   week	  with	   no	   filtration,	   there	   is	   a	   fairly	   large	  confidence	  interval	  for	  both	  home	  types	  compared	  to	  the	  week	  with	  filtration.	  	  As	   predicted,	  WS	   homes	  were	   over	   7	   times	   further	   away	   from	   highway	   or	  major	  road	   compared	   to	   TRAP	   homes.	   Moreover,	   TRAP	   homes	   had	   on	   average	   8	   times	  more	   road	   length	   in	   their	   100-­‐meter	   buffer	   (Table	   3-­‐6).	   Overall,	   higher	  concentrations	   of	   indoor	   and	   outdoor	   LG	  were	   present	   in	  WS	   homes,	   though	   not	  statistically	   significant.	   All	   together,	   we	   were	   able	   to	   verify	   that	   homes	   were	  categorized	  into	  their	  respective	  categories	  correctly.	  	  	   Table	  3-­‐6	  Summary	  statistics	  to	  confirm	  categorization	  of	  TRAP	  and	  WS	  homes	  	  3.4 Health	  Outcomes	  	  As	   with	   exposure	   characteristics,	   summary	   statistics	   were	   also	   prepared	   for	  different	   health	   outcomes	   included	   in	   this	   study	   (Tables	   3-­‐7	   and	   3-­‐8).	   Table	   3-­‐7	  summarizes	   mean	   (±SD)	   and	   median	   of	   health	   outcomes	   stratified	   by	   HEPA	  filtration	   status.	  Based	  on	   the	  basic	   statistics	   and	  p-­‐values	  presented	   in	  Table	  3-­‐7	  	   55	  below,	   there	   was	   no	   statistically	   significant	   difference	   between	   any	   of	   the	   health	  outcomes	  with	  HEPA	  filtration.	  	   	  Table	  3-­‐7	  Summary	  of	  health	  outcomes	  by	  HEPA	  filtration	  status	  	  	  3.5 Mixed	  Model	  Results	  	  After	   reviewing	   all	   collected	  data,	   68	   participants	  were	   included	   in	  mixed	  models	  including	   RHI	   as	   their	   outcome	   variable,	   forty-­‐eight	   of	   whom	   lived	   in	   high-­‐TRAP	  homes	   and	   twenty	   lived	   in	   high-­‐WS	   homes.	   With	   regards	   to	   CRP	   models,	   52	  participants	   had	   complete	   data	   required	   for	   inclusion	   in	   the	  models.	   It	   should	   be	  noted	   that	   all	   effect	   estimates	   of	   log-­‐transformed	  models	   (i.e.	   RHI,	   CRP,	   and	   IL-­‐6)	  were	   converted	   to	   %	   change	   to	   facilitate	   interpretation	   of	   results.	   Moreover,	  average	   indoor	   temperature	   (fixed	   effect),	   and	   home	   and	   subject	   ID	   (random	  effects)	  were	  included	  in	  all	  models.	  	   56	  3.6 HEPA	  Filtration	  	  Table	  3-­‐9	  summarizes	  the	  model	  estimates	  of	  change	  in	  health	  outcomes	  with	  HEPA	  filtration.	   Results	   are	   presented	   for	   all	   homes,	   TRAP	   homes	   only,	   and	  WS	   homes	  only.	  	  	   Table	  3-­‐8	  Change	  in	  health	  outcomes	  with	  HEPA	  filtration	  	  	  The	   initial	   hypothesis	  was	   that	  with	   the	   implementation	   of	   HEPA	   filtration,	   there	  would	  be	  an	  improvement	  in	  endothelial	  function	  and	  systematic	  inflammation	  (i.e.	  an	  increase	  in	  RHI	  score	  and	  reductions	  in	  markers	  of	  inflammation).	  For	  RHI,	  there	  was	  no	  association	  with	  HEPA	  filtration.	  There	  was	  also	  no	  association	  between	  CRP	  and	  IL-­‐6	  and	  HEPA	  filtration	  in	  either	  of	  the	  groups.	  Neither	  BCC	  nor	  %PMN	  showed	  statistically	  significant	  associations	  with	  HEPA	  filtration.	  The	  effect	  estimates	  were	  extremely	   small	   with	   very	   large	   confidence	   intervals.	   Overall,	   there	   were	   no	  significant	  associations	  between	  HEPA	  filtration	  and	  RHI,	  CRP,	  IL-­‐6,	  BCC,	  or	  %PMN.	  3.7 PM2.5	  	  Table	  3-­‐10	  summarizes	  the	  model	  estimates	  of	  change	  in	  health	  outcomes	  per	  unit	  median	   increase	   in	   PM2.5	   concentration.	   As	   described	   in	   the	   Methods	   section,	   all	  exposure	  contrasts	  were	  scaled	   to	   the	  within-­‐home	  change	  between	  HEPA	  and	  no	  HEPA	  conditions.	  Hence,	  all	  changes	  in	  health	  outcomes	  are	  considered	  to	  be	  change	  per	   unit	  median	   increase	   in	   exposure.	   Results	   are	   presented	   for	   all	   homes,	   TRAP	  homes	  only,	  and	  WS	  homes	  only.	  	  	  	   57	  Table	  3-­‐9	  Change	  in	  health	  outcomes	  per	  unit	  median	  change	  in	  PM2.5	  	  	  With	   regards	   to	   PM2.5	   exposure,	   we	   expected	   the	   deterioration	   of	   microvascular	  endothelial	   function	  and	  systematic	   inflammation	  (i.e.	  a	  decrease	   in	  RHI	  score	  and	  increases	   in	   markers	   of	   inflammation).	   With	   regards	   to	   RHI,	   there	   was	   no	  association	  found	  with	  PM2.5	  concentrations.	  However,	  a	  non-­‐significant	  increase	  in	  CRP	  was	  observed	  per	  unit	  median	  increase	  in	  PM2.5	  for	  all	  homes.	  After	  stratifying	  by	  home	  type,	  there	  was	  an	  18.4%	  increase	  in	  CRP	  levels	  per	  unit	  median	  increase	  in	  PM2.5,	  which	  had	  a	  borderline	  statistical	  significance.	  No	  such	  effect	  was	  found	  in	  WS	  homes	  with	  the	  effect	  estimate	  in	  the	  opposite	  direction	  of	  TRAP	  homes.	  	  	  For	  IL-­‐6,	  there	  were	  no	  significant	  associations	  with	  PM2.5	  concentrations	  in	  either	  TRAP	  or	  WS	  homes.	  Finally,	  there	  was	  no	  significant	  increase	  in	  BCC	  or	  %PMN	  with	  increases	   in	   PM2.5	   levels.	   Overall,	   there	   was	   some	   suggestion	   of	   an	   association	  between	  PM2.5	  exposure	  and	  CRP	  levels	  among	  individuals	  residing	  in	  TRAP	  homes.	  3.8 Absorbance	  	  Table	  3-­‐11	  summarizes	  the	  model	  estimates	  of	  change	  in	  health	  outcomes	  per	  unit	  median	   increase	   in	   absorbance.	  Results	   are	  presented	   for	   all	   homes,	  TRAP	  homes	  only,	  and	  WS	  homes	  only.	  	  	  	   58	  Table	  3-­‐10	  Change	  in	  health	  outcomes	  per	  unit	  median	  change	  in	  absorbance	  	  	  A	  non-­‐significant	  improvement	  in	  RHI	  was	  found	  with	  absorbance	  in	  both	  TRAP	  and	  WS	   homes.	   TRAP	   homes	   demonstrated	   a	   borderline	   statistical	   significance.	  However,	   there	  was	  no	  significant	  association	  between	  CRP	   levels	  and	  absorbance	  in	  any	  of	  the	  subgroups.	  With	  regards	  to	  IL-­‐6,	  there	  was	  no	  significant	  relationship	  with	  absorbance.	  Finally,	  there	  was	  no	  significant	  association	  between	  BCC/%PMN	  and	   absorbance,	   with	   very	   small	   non-­‐significant	   effect	   estimates	   for	   both	   health	  endpoints.	  Overall,	   there	  were	  no	  significant	  associations	  between	  absorbance	  and	  RHI,	   CRP,	   IL-­‐6,	   BCC,	   or	   %PMN.	   However,	   there	   were	   some	   suggestions	   of	   an	  association	  between	  RHI	  and	  TRAP	  absorbance.	  	  3.9 LG	  	  Table	  3-­‐12	  summarizes	  the	  model	  estimates	  of	  change	  in	  health	  outcomes	  per	  unit	  median	  increase	  in	  LG.	  Results	  are	  presented	  for	  all	  homes,	  TRAP	  homes	  only,	  and	  WS	  homes	  only.	  	  	  	   59	  Table	  3-­‐11	  Change	  in	  health	  outcomes	  per	  unit	  median	  change	  in	  LG	  	  	  There	   was	   no	   significant	   association	   between	   any	   of	   the	   health	   endpoints	   and	  exposure	  to	  LG.	  In	  fact,	  all	  effect	  estimates	  were	  extremely	  small	  and	  with	  very	  wide	  confidence	  intervals.	  Overall,	  there	  was	  no	  evidence	  of	  an	  association	  between	  RHI,	  CRP,	  IL-­‐6,	  BCC,	  or	  %PMN	  and	  exposure	  to	  LG.	  3.10 Effect	  Modification	  by	  BMI,	  Age,	  and	  Sex	  	   	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   	  Figure	  3-­‐1	  Effect	  modification	  by	  sex	  in	  the	  relationship	  between	  CRP	  and	  PM2.5	  In	  order	  to	  evaluate	  whether	  effects	  differed	  by	  BMI	  (⪯25	  and	  >25),	  age	  (⪯42	  and	  >42),	  and	  sex,	  stratified	  analyses	  were	  conducted.	  After	  stratifying	  by	  sex	  in	  the	  CRP	  and	  PM2.5	  mixed	  model	  results,	  it	  was	  observed	  that	  the	  effect	  of	  PM2.5	  exposure	  was	  only	   present	   in	   males	   (Figure	   3-­‐1).	   There	   was	   a	   20.6%	   (95%	   CI,	   2.62%,	   41.7%)	  increase	  in	  CRP	  levels	  in	  men	  per	  unit	  median	  increase	  in	  PM	  while	  in	  women,	  there	  	   60	  was	  no	  evidence	  of	  an	  association	  (-­‐13.7%	  (95%	  CI,	  -­‐26.5%,	  1.42%)).	  This	  analysis	  revealed	   that	   the	  association	  observed	  between	  TRAP	  and	  CRP	  was	  driven	  by	   the	  effect	   in	  males	  only.	  All	  other	  exposure	  and	  outcome	  pairs	  were	  also	  evaluated	  for	  effect	  modification	  by	  BMI,	  age,	  and	  sex;	  however,	   there	  was	  no	  evidence	  of	  effect	  modification	  in	  any	  of	  the	  mixed	  models.	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   61	  4 Discussion	  	  Traffic-­‐related	   and	   wood	   combustion	   PM	   are	   considered	   to	   be	   two	   major	  contributors	   of	   ambient	   particle	   levels	   in	   many	   industrialized	   countries	   (de	   Kok,	  Driece,	  Hogervorst,	  &	  Briedé,	  2006;	  Saarikoski	  et	  al.,	  2008;	  Song	  et	  al.,	  2007;	  Wu	  et	  al.,	  2007).	  Exposure	  to	  TRAP	  PM	  occurs	  throughout	  the	  year	  while	  WS	  exposure	  is	  greater	  during	  colder	  months.	  However,	  it	  is	  believed	  that	  during	  cold	  months,	  both	  TRAP	  and	  WS	  contribute	  to	  a	  similar	  magnitude	  to	  the	  ambient	  PM	  concentrations	  (Glasius	   et	   al.,	   2006;	   Saarikoski	   et	   al.,	   2008;	   Song	   et	   al.,	   2007;	   Wu	   et	   al.,	   2007).	  Exposure	   to	  both	   sources	  has	  been	   shown	   to	  be	   associated	  with	   increased	   risk	  of	  adverse	  health	  effects	  including	  adverse	  cardiovascular	  health	  outcomes	  (Naeher	  et	  al.,	   2007;	   Rückerl,	   Schneider,	   Breitner,	   Cyrys,	   &	   Peters,	   2011).	   Despite	   the	  overwhelming	  evidence	  on	  cardiovascular	  effects	  of	  PM,	   there	  have	  been	  very	   few	  studies	  evaluating	  potential	  differences	  between	  cardiovascular	  effects	  of	  PM	  from	  TRAP	  and	  WS	  (Brook	  et	  al.,	  2010).	  This	  study	  is	  the	  first	  study	  of	  its	  kind	  to	  directly	  compare	   the	   cardiovascular	   effects	   of	   PM	   from	   these	   two	   dominant	   sources	   of	  combustion	  PM.	  	  	  Overall,	   HEPA	   filtration	   was	   associated	   with	   a	   40%	   decrease	   in	   indoor	   PM2.5	  concentration.	   However,	   there	   was	   no	   conclusive	   evidence	   on	   the	   relationship	  between	   TRAP	   or	   WS	   PM2.5	   exposure	   and	   endothelial	   function	   or	   any	   of	   the	  biomarkers	   of	   inflammation.	   There	   was	   some	   suggestion	   of	   an	   association	  (borderline	   significance)	   between	   exposure	   to	   TRAP	   PM2.5	   and	   CRP	   levels	   as	   the	  primary	  endpoint	   for	   systematic	   inflammation	   in	   this	   study.	  After	   stratification	  by	  sex,	   the	   effect	  was	   only	  present	   in	  males	   (statistically	   significant).	   There	  was	   also	  some	   evidence	   of	   an	   association	   between	   RHI	   and	   absorbance	   in	   TRAP	   homes.	  Moreover,	  there	  was	  no	  evidence	  of	  an	  association	  between	  IL-­‐6	  or	  BCC	  and	  HEPA	  filtration	  or	  PM2.5	  exposure.	  All	  in	  all,	  there	  was	  a	  lack	  of	  conclusive	  evidence	  on	  the	  potential	   differences	   in	   cardiovascular	   effects	   between	   TRAP	   and	   WS	   particles.	  	   62	  However,	   the	   CRP	   results	   may	   suggest	   stronger	   inflammatory	   effects	   for	   TRAP	  particles	  and	  in	  males.	  	  	  This	  is	  the	  first	  study	  of	  its	  kind	  that	  has	  directly	  compared	  the	  potential	  impact	  of	  two	   combustion	   sources	   of	   PM	   on	   cardiovascular	   health.	   This	   study	   was	   able	   to	  effectively	   identify	   individuals	   exposed	   to	   either	   TRAP	   or	   WS	   PM	   using	   detailed	  spatial	   air	   pollution	   modeling	   across	   the	   sampling	   region.	   Comparison	   of	   the	  differences	   in	   effect	   between	   the	   two	   sources	   provides	   a	   unique	   insight	   into	   the	  similarities	  and	  the	  differences	  of	  TRAP	  and	  WS	  PM.	  	  4.1 HEPA	  Filtration	  Efficiency	  	  With	  the	  implementation	  of	  indoor	  HEPA	  filters,	  there	  was	  a	  statistically	  significant	  reduction	  of	  over	  40%	  in	  PM2.5	  concentrations	  with	  a	  drop	  of	  36%	  in	  TRAP	  homes	  and	  48%	  in	  WS	  homes.	  The	  results	  in	  this	  study	  were	  in	  agreement	  with	  a	  number	  of	  other	  North	  American	  studies	  confirming	  the	  effectiveness	  of	  HEPA	  filters.	  A	  2013	  review	   of	   the	   literature,	   the	   author	   concluded	   that	   overall,	   studies	   observed	   a	  reduction	  of	  over	  50%	  on	  PM	  concentrations	  with	  HEPA	  filtration	  (Fisk,	  2013).	  In	  a	  study	  of	  a	  WS	  impacted	  community	  in	  Canada,	  Allen	  et	  al.	  (2011)	  found	  a	  reduction	  of	  over	  60%	  with	  the	  utilization	  of	  HEPA	  filtration.	  Similarly,	  in	  another	  study	  of	  WS	  impacted	  homes,	  there	  was	  a	  55-­‐65%	  drop	  in	  PM	  concentration	  with	  filtration	  (Barn	  et	  al.,	  2008)	  	  4.2 Mixed	  Model	  Conclusions	  	  It	  should	  be	  noted	  that	  initially,	  we	  intended	  to	  include	  %	  time	  at	  home	  and	  %	  time	  commuting	  in	  the	  mixed	  models	  to	  account	  for	  potential	  confounding.	  However,	  due	  to	   lack	   of	   data	   for	  many	   individuals	   and	   the	   insensitivity	   of	   the	   analysis	   to	   their	  inclusion,	   it	   was	   decided	   to	   remove	   these	   variables	   from	   the	   mixed	   models.	  Moreover,	  baseline	  blood	  measures	  were	  not	  included	  or	  used	  in	  any	  of	  the	  mixed	  models.	   This	   decision	   was	   made	   primarily	   due	   to	   the	   fact	   that	   there	   was	   no	  information	   on	   exposure	   metrics	   prior	   to	   the	   initiation	   of	   the	   study.	   As	   a	   result,	  	   63	  blood	  measures	  might	  have	  not	  been	  representative	  of	  what	  background	  subclinical	  indicator	  levels	  would	  have	  been.	  4.2.1 RHI	  	  With	  RHI	  as	  the	  primary	  endpoint,	  mixed	  model	  results	  did	  not	  show	  any	  significant	  association	  between	  HEPA	  filtration	  and	  endothelial	   function.	   In	   fact,	  we	   found	  no	  significant	   relationship	   between	   any	   of	   PM2.5,	   LG,	   or	   absorbance	   and	   RHI.	  Furthermore,	   the	   results	   of	   the	   analysis	   after	   stratification	   by	   home	   type	   did	   not	  unveil	   any	   differences	   in	   the	   effect	   of	   filtration	   on	   endothelial	   function	   between	  TRAP	  and	  WS	  homes,	   except	   for	  a	   small	   association	  between	  RHI	  and	  absorbance	  (borderline	  significance).	  	  	  Similar	   to	   the	   current	   study,	   there	   have	   been	   a	   number	   of	   studies	   that	   have	  evaluated	   the	   effect	   of	   HEPA	   filtration	   on	   endothelial	   function	   with	   different	  conclusions.	   In	   a	   study	   conducted	   in	   a	   WS-­‐impacted	   community	   in	   2008-­‐09,	  members	   of	   the	   current	   study’s	   research	   group	   used	   a	   HEPA	   filter	   intervention	  design	  to	  investigate	  the	  effects	  of	  WS	  PM2.5	  on	  endothelial	  function	  among	  healthy	  adults	  (Allen	  et	  al.,	  2011).	  With	  filtration,	  the	  PM2.5	  concentration	  dropped	  from	  11.2	  µg/m3	   to	   4.6	   µg/m3,	   which	   is	   a	   60%	   reduction.	   The	   authors	   found	   a	   statistically	  significant	   association	   between	   filtration	   of	   WS	   PM2.5	   and	   improvements	   in	  endothelial	   function;	   the	   use	   of	   air	   filters	  was	   associated	  with	   a	   9.4%	   increase	   in	  RHI.	  In	  another	  study	  conducted	  in	  Denmark,	  HEPA	  filtration	  was	  used	  to	  evaluate	  the	   effects	   traffic-­‐related	   PM	   on	   various	  markers	   of	   cardiovascular	   health	   among	  healthy	  elderly	  couples	  (Bräuner,	  Forchhammer,	  et	  al.,	  2008).	  The	  authors	  observed	  a	  reduction	  from	  12.6	  µg/m3	  to	  4.7	  µg/m3	  with	  filtration,	  a	  drop	  of	  over	  60%.	  They	  concluded	   that	   indoor	   air	   filtration	   was	   significantly	   associated	   with	   an	  improvement	  of	  8.1%	  (95%	  CI,	  0.4%-­‐16.3%)	  in	  microvascular	  function.	  	  	  In	   contrast	   to	   the	   above	   studies,	   which	   reported	   a	   significant	   improvement	   in	  endothelial	   function	   with	   HEPA	   filtration,	   some	   other	   studies	   reported	   no	  improvement	   in	   endothelial	   function.	   In	   a	   randomized	   double-­‐blind	   crossover	  	   64	  intervention	   study,	   HEPA	   filtration	   was	   used	   to	   reduce	   PM2.5	   levels	   in	   mostly	  smoking	  First	  Nations	  homes	  to	  evaluate	  acute	  changes	  in	  cardiorespiratory	  health	  outcomes	   (Weichenthal	   et	   al.,	   2013).	   Despite	   drastic	   reductions	   in	   indoor	   PM2.5	  levels	   from	   42.5	   µg/m3	   to	   22.0	   µg/m3	   by	   filtration,	   the	   authors	   did	   not	   find	   any	  significant	  association	  between	  HEPA	  filtration	  and	  endothelial	   function.	  The	  same	  group	  of	  researchers	  from	  the	  study	  conducted	  in	  Denmark	  recently	  completed	  an	  indoor	  air	  filtration	  study	  among	  48	  elderly	  subjects	  (51	  to	  81	  years)	  and	  evaluated	  its	   effect	   of	   microvascular	   endothelial	   dysfunction	   (Karottki	   et	   al.,	   2013).	   The	  authors	  did	  not	  find	  any	  significant	  relationship	  looking	  at	  filtration	  as	  a	  categorical	  variable	   and	  microvascular	   function;	   however,	   they	   found	   an	   association	  between	  PM2.5	  as	  a	  continuous	  variable	  and	  RHI.	  	  	  There	  might	  be	   a	  number	  of	   explanations	   for	   the	   inability	   to	  observe	  an	  effect	  on	  endothelial	   function	   by	   HEPA	   filtration	   and	   other	   aforementioned	   exposure	  variables.	  The	  EndoPAT	  device	  used	  to	  measure	  RHI	  is	  very	  particular	  with	  regards	  to	  exactly	  following	  the	  operating	  protocol.	  There	  were	  some	  inconsistencies	  in	  the	  occlusion	   time	   after	   inspecting	   the	  PWA	  output	   of	   test,	  which	   could	  have	   affected	  the	  outcome	  (i.e.	  under-­‐	  or	  over-­‐estimation	  of	  RHI	  depending	  on	   the	  PWA).	  There	  has	   also	   been	   some	   evidence	   of	   poor	   inter-­‐day	   reproducibility,	   which	   might	   be	  important	  when	  comparing	  RHI	  data	  of	  the	  same	  individual	  from	  different	  days	  (Liu,	  Wang,	   Jin,	  Roethig,	  &	  Unverdorben,	  2009).	  The	  combination	  of	  these	  two	  potential	  sources	  of	  error	  might	  have	  affected	  the	  results	  obtained	  using	  EndoPAT	  by	  either	  under-­‐	   or	   over-­‐estimation	   of	   RHI.	   Another	   potential	   explanation	   for	   the	   lack	   of	   a	  significant	   relationship	   between	   RHI	   and	   filtration,	   PM2.5,	   or	   absorbance	   in	   this	  study	   could	   be	   the	   relatively	   low	   exposure	   concentrations	   and	   small	   exposure	  gradients.	  The	  indoor	  concentrations	  might	  have	  been	  too	  low	  to	  be	  able	  to	  see	  any	  measurable	  improvements	  in	  endothelial	  function	  by	  filtration.	  The	  small	  change	  in	  PM2.5	  concentration	  with	  filtration	  could	  have	  resulted	  in	  a	  low	  signal	  to	  noise	  ratio	  in	  EndoPAT	  output,	  masking	  small	  changes	  in	  endothelial	  function.	  	  	   65	  There	  have	  been	  a	  number	  of	  controlled	  human	  WS	  exposure	  studies	  conducted	  to	  evaluate	   the	   relationship	  between	  WS	  PM	  and	   endothelial	   function.	   Similar	   to	   the	  results	  of	   the	  current	  study,	  most	  recent	  studies	  have	   failed	  to	   find	  any	  significant	  effects	  on	  endothelial	   function.	   In	  a	  randomized	  double-­‐blind,	  cross-­‐over	  study,	  20	  non-­‐smoking	  participants	  were	  exposed	  to	  between	  14	  and	  354	  µg/m3	  of	  particles	  from	   a	  well-­‐burning	  modern	  woodstove	   and	   clean	   air	   for	   3	   hours	   each	   session,	   7	  days	  apart	  (Forchhammer	  et	  al.,	  2012).	  The	  authors	  measured	  endothelial	  function	  using	  endoPAT	  six	  hours	  after	  exposure	  but	  found	  no	  effect	  on	  endothelial	  function	  by	  exposure	  to	  WS	  PM.	  In	  another	  study,	  26	  healthy	  nonsmoking	  young	  participants	  were	  exposed	  to	  150-­‐200	  µg/m3	  of	  fine	  wood	  combustion	  particles	  generated	  by	  a	  standard	   woodstove	   and	   clean	   air	   for	   three	   hours	   each	   (Pope	   et	   al.,	   2011).	  Endothelial	   function	   was	   measured	   using	   EndoPAT	   after	   WS	   exposure	   and	   after	  exposure	  to	  clean	  air.	  The	  authors	  did	  not	  find	  any	  evidence	  of	  change	  in	  endothelial	  function	   after	   exposure	   to	  WS	   PM.	   It	   should	   be	   noted	   that	   these	   exposures	  were	  between	   twice	   and	   over	   50	   times	   the	   exposures	   our	   participants	   experienced	   in	  their	   homes.	   In	   addition,	   these	   studies	   introduced	   much	   larger	   concentration	  gradient	  between	  treatment	  and	  control	  session	  compared	  to	  this	  study,	  yet	  failing	  to	  find	  any	  association	  between	  exposure	  to	  WS	  PM	  and	  endothelial	  dysfunction.	  	  4.2.2 CRP	  	  In	   the	  current	  study,	  we	  used	  hs-­‐CRP	  as	  a	  marker	  of	   systematic	   inflammation	  and	  cardiovascular	  health.	  Numerous	  studies	  have	  shown	  that	  hs-­‐CRP	  can	  be	  used	  as	  a	  predictor	  of	   future	  cardiovascular	  events	  as	  a	  proinflammatory	  response	  mediator	  (Ridker,	   2007).	   Hence,	   it	   is	   recommended	   as	   a	   biomarker	   for	  measuring	   adverse	  future	  cardiovascular	  risk	  (Jialal,	  Devaraj,	  &	  Venugopal,	  2004).	  	  	  	  The	   available	   evidence	   in	   the	   literature	   is	   not	   entirely	   consistent	   regarding	   the	  relationship	   between	   short-­‐term	   PM	   exposure	   and	   systematic	   inflammation	  (Bräuner,	  Møller,	  et	  al.,	  2008;	  Rudez	  et	  al.,	  2009).	  However,	  there	  is	  a	  growing	  body	  of	   evidence	   linking	   PM	   exposure	   with	   systematic	   inflammation	   and	   adverse	  cardiovascular	  health	  outcomes	  in	  general	  (Brook	  et	  al.,	  2010).	  In	  this	  study,	  there	  	   66	  was	  no	  significant	  relationship	  between	  HEPA	  filtration	  and	  CRP	  levels	  in	  either	  the	  overall	  population	  or	  when	  stratified	  by	  home	  type.	  When	  evaluating	  the	  effects	  of	  PM2.5	  levels	  among	  all	  participants	  (both	  TRAP	  and	  WS),	  we	  found	  a	  non-­‐significant	  increase	  in	  CRP	  levels	  per	  unit	  median	  increase	  in	  PM2.5	  in	  all	  homes.	  However,	  after	  stratifying	   by	   home	   type,	   we	   found	   a	   borderline	   statistical	   significance	   (P-­‐value	  0.056)	   for	   association	   between	   CRP	   and	   PM2.5	  levels	   in	   TRAP	   homes	   only.	   In	   fact,	  there	  was	  an	  18.4%	  (95%	  CI,	  -­‐0.24%,	  40.6%)	  increase	  in	  CRP	  concentration	  per	  unit	  median	  increase	  in	  indoor	  PM2.5	  levels	  in	  this	  subset	  of	  homes.	  After	  stratification	  by	  sex,	   the	   effect	   was	   only	   present	   in	   males	   with	   a	   20.6%	   (95%	   CI,	   2.62%-­‐41.7%)	  increase	  in	  CRP	  per	  unit	  median	  increase	  in	  indoor	  PM2.5	  levels.	  	  However,	  there	  was	  no	   association	   with	   WS	   PM2.5.	   There	   was	   also	   no	   association	   between	   CRP	   and	  absorbance	  in	  this	  study.	  	  So	  overall,	  based	  on	  the	  results	  from	  PM2.5	  models,	  there	  was	  some	  suggestion	  of	  an	  association	   between	   exposure	   to	   TRAP	   particles	   and	   systematic	   inflammation	  amongst	  male	  participants	  of	  this	  study.	  The	  results	  support	  the	  a	  priori	  hypothesis	  of	  a	  greater	  impact	  of	  TRAP	  on	  systematic	  inflammation	  compared	  to	  WS.	  	  In	   order	   to	   investigate	   potential	   gender	   differences	   in	   CRP	   effect	   estimates,	   the	  mixed	  models	  were	  analyzed	  by	  including	  a	  gender	  term	  in	  the	  model.	  The	  results	  indicate	  that	  statistically	  significant	  CRP	  effect	  was	  restricted	  to	  the	  male	  subjects	  in	  this	   study.	   The	   more	   pronounced	   association	   between	   CRP	   and	   PM	   in	   men	   is	  consistent	   with	   other	   studies	   suggesting	   a	   link	   between	   short-­‐term	   air	   pollution	  exposure	  and	  inflammation	  among	  male	  subjects	  (Briet	  et	  al.,	  2007;	  Riediker	  et	  al.,	  2004;	  Rundell,	  Hoffman,	  Caviston,	  Bulbulian,	  &	  Hollenbach,	  2007;	  Tan	  et	  al.,	  2000;	  Törnqvist	  et	  al.,	  2007).	  There	  have	  also	  been	  some	  studies	  showing	  an	  association	  between	  CRP	  and	  PM2.5	  in	   females	  but	   the	  results	  were	   less	  consistent.	   In	  a	  recent	  systematic	   review	   of	   the	   literature,	   Kaptoge	   et	   al.	   (2012)	   evaluated	   the	   evidence	  from	  38	  prospective	  cohorts,	  with	  over	  166,000	  participants,	   investigating	  the	  link	  between	  CRP	  levels	  and	  future	  CHD	  events.	  In	  the	  pooled	  analysis	  of	  the	  data	  from	  these	  studies,	  the	  author	  found	  a	  statistically	  significant	  increase	  in	  the	  risk	  of	  future	  	   67	  CVD	  events	  associated	  with	  CRP	   levels	   in	  males	  only	   (Kaptoge	  et	  al.,	  2012).	  There	  was	  also	  an	  increase	  in	  CVD	  risk	  in	  women	  however,	  the	  effect	  estimate	  was	  much	  smaller	  than	  men’s	  and	  non-­‐significant.	  	  Overall,	  there	  is	  a	  more	  extensive	  body	  of	  literature	  investigating	  the	  effects	  of	  TRAP	  PM	  on	   inflammation	  as	  compared	  to	  WS	  studies.	   In	  a	  study	  of	  healthy	  young	  men,	  (Riediker	   et	   al.,	   2004)	  monitored	   ten	   highway	   patrol	   troopers	   and	   PM2.5	  levels	   in	  their	  cars.	   In	   their	  analysis,	   there	  was	  a	  32%	   increase	   in	  CRP	   levels	  per	  10	  µg/m3	  increase	  in	  PM2.5.	  Dubowsky	  and	  colleagues	  (2006)	  studied	  44	  senior	  citizens	  with	  and	  without	  conditions	  related	  to	  chronic	  inflammation	  (Dubowsky,	  Suh,	  Schwartz,	  Coull,	  &	  Gold,	  2006).	  They	  found	  significant	  associations	  between	  exposure	  to	  TRAP	  PM	  and	  various	  markers	  of	  inflammation	  including	  CRP.	  Moreover,	  a	  recent	  study	  of	  110	  traffic	  policemen	  in	  China	  also	  found	  a	  significant	  association	  between	  exposure	  to	  TRAP	  PM	  and	  CRP	  (Zhao	  et	  al.,	  2013).	  The	  current	  study	  results	  are	  in	  agreement	  with	   the	   growing	   body	   of	   evidence	   showing	   the	   link	   between	  TRAP	  PM	   exposure	  and	  CRP	  levels.	  	  The	  available	   literature	   investigating	  the	  effects	  of	  WS	  PM	  on	  CRP	  levels	   is	   limited	  and	  inconclusive	  at	  best.	  There	  have	  been	  very	  few	  studies	  that	  have	  linked	  WS	  PM	  to	   CRP	   levels.	   A	  HEPA	   filter	   intervention	   study	   by	   the	   same	   research	   team	   as	   the	  present	  study	  in	  a	  WS	  impacted	  community	  found	  a	  32.6%	  (4.4-­‐60.9%)	  decrease	  in	  CRP	  levels	  with	  HEPA	  filtration.	  However,	  they	  did	  not	  observe	  an	  association	  with	  WS	   PM2.5	   levels	   in	   homes	   (Allen	   et	   al.,	   2011).	   A	   recent	   study	   investigated	   the	  inflammatory	  effects	  of	  WS	  exposure	  among	  wildfire	  firefighters	  (Hejl	  et	  al.,	  2013).	  They	  found	  borderline	  statistical	  significance	  increase	  in	  CRP	  levels	  with	  exposure.	  In	   contrast,	   there	   have	   also	   been	   a	   number	   of	   other	   studies,	   which	   found	   no	  association	  at	  all.	  A	  short-­‐term	  controlled	  exposure	  study	  to	  WS	  was	  conducted	  on	  16	  healthy	  non-­‐smokers	  (20-­‐57	  years	  old)	  evaluating	  the	  effect	  of	  WS	  on	  systematic	  inflammation	  (Stockfelt,	  Sallsten,	  Almerud,	  Basu,	  &	  Barregard,	  2013).	  They	  did	  not	  find	  any	  significant	  association	  between	  WS	  PM	   levels	  of	  146-­‐295	  µg/m3	  and	  CRP	  concentrations.	  	  	   68	  The	   current	   study	   results	   further	   support	   the	   a	   priori	   hypothesis	   of	   greater	  inflammatory	   effects	   for	   CRP	   in	   those	   exposed	   to	   TRAP	   PM	   compared	   to	   those	  exposed	  to	  WS	  PM.	  There	  are	  at	  least	  two	  possible	  explanations	  for	  this	  observation.	  First,	  the	  small	  size	  of	  the	  PM2.5	  and	  their	  large	  surface	  area	  enables	  them	  to	  carry	  a	  large	  number	  of	  toxic	  substance	  and	  free	  radicals	  present	  mostly	  in	  TRAP	  PM.	  This	  provides	  an	  opportunity	  for	  the	  particles	  to	  have	  a	  biological	  impact	  on	  the	  cells	  in	  the	  lungs.	  With	  increased	  particle	  presence	  in	  the	  airways,	  the	  phagocytic	  capacity	  decreases	  which	  leads	  to	  their	  presence	  for	  an	  extended	  time	  in	  the	  lungs	  and	  hence,	  greater	  interaction	  with	  various	  cell	  lines	  (e.g.	  epithelial	  cells)	  (MacNee	  et	  al.,	  1997).	  Second,	  as	  discussed	  earlier,	  the	  greater	  abundance	  of	  transition	  metals	  in	  TRAP	  PM	  compared	  to	  WS	  PM,	  which	  can	  lead	  to	  the	  generation	  of	  hydroxyl	  radicals,	  oxidative	  stress,	   and	   inflammation	   (Ghio	   et	   al.,	   1999;	   Verma	   et	   al.,	   2009)	   and	   the	   greater	  deposition	   of	   traffic	   particles	   in	   the	   lungs	   could	   further	   explain	   these	   results	  (Löndahl	  et	  al.,	  2008;	  Löndahl	  et	  al.,	  2009).	  	  Finally,	  as	  mentioned	  above,	  there	  was	  no	  relationship	  between	  HEPA	  filtration	  and	  CRP	   while	   there	   was	   a	   suggestion	   of	   an	   association	   between	   TRAP	   PM2.5,	   as	   a	  continuous	  variable,	  and	  CRP	  levels.	  These	  findings	  could	  be	  possibly	  explained	  by	  the	   fact	   that	   some	   homes	   were	   included	   in	   the	   study	   during	   warmer	  month	   and	  hence,	  there	  was	  a	  greater	  probability	  of	  having	  open	  windows	  while	  sampling.	  As	  a	  result,	   there	  might	   have	   been	   reduced	  HEPA	   filter	   effectiveness,	  which	   cannot	   be	  captured	  by	  the	  binary	  intervention	  variable	  (i.e.	  HEPA/no	  HEPA).	  However,	  such	  a	  discrepancy	   in	   filtration	   efficiency	   can	   be	   captured	   by	   the	   continuous	   PM2.5	  concentration	  variable.	  As	   a	   result,	   this	  might	   explain	   the	   reason	  behind	   finding	  a	  suggestion	  of	  an	  association	  between	  PM2.5	  and	  CRP	  but	  not	  HEPA	  and	  CRP.	  4.2.3 IL-­‐6	  	  As	   one	   of	   the	   secondary	   endpoints	   of	   systematic	   inflammation,	   IL-­‐6	   was	   also	  included	   in	   the	   analysis.	   IL-­‐6	   is	   known	   to	   be	   one	   of	   the	   cytokines	   involved	   in	  initiating	   acute	   phase	   inflammatory	   response	   in	   the	   body	   by	   promoting	   the	  	   69	  synthesis	  of	  various	  proteins	  including	  CRP	  (Gabay	  &	  Kushner,	  1999;	  van	  Eeden	  et	  al.,	  2005).	  	  	  There	  have	  been	  conflicting	  studies	  investigating	  an	  association	  between	  IL-­‐6	  as	  an	  inflammatory	  marker	  and	  the	  risk	  of	  cardiovascular	  events.	  In	  a	  prospective	  cohort	  study,	  2,225	  elderly	  participants	  without	  baseline	  CVD	  were	  enrolled	   in	   the	  study.	  Upon	   follow	   up	   and	   adjusting	   for	   potential	   confounders,	   IL-­‐6	   was	   found	   to	   be	  significantly	  associated	  with	  increased	  risk	  of	  CHD	  events	  (27%-­‐86%	  increase)	  per	  IL-­‐6	  SD	  increase	  (Cesari	  et	  al.,	  2003).	   In	  a	  nested	  case-­‐control	  on	  317	  participants,	  (Luc	  et	  al.,	  2003)	  followed	  subjects	  for	  five	  years,	  recording	  initial	  CHD	  events.	  They	  concluded	   that	   IL-­‐6	   levels	   were	   significantly	   associated	   with	   adverse	   coronary	  outcomes.	   However,	   a	   recent	   large-­‐scale	   prospective	   cohort	   study	   followed	   over	  51,000	   individuals	  with	   no	   baseline	   cardiovascular	   health	   issue	   for	   up	   to	   8	   years	  (Pai	  et	  al.,	  2004).	  During	  the	  follow-­‐up	  period,	  they	  identified	  249	  women	  and	  266	  men	  who	  developed	  cardiovascular	  events.	  Blood	  samples	  were	  collected	  biennially	  during	   the	   follow-­‐up	   period	   and	   analyzed	   for	   various	   inflammatory	   biomarkers.	  After	  adjusting	  for	  potential	  confounders,	  the	  authors	  concluded	  that	  there	  was	  no	  association	  between	  adverse	  cardiovascular	  events	  and	  IL-­‐6	  levels.	  	  	  	  The	   results	  of	   the	  analysis	  were	  not	  as	  hypothesized	   for	   the	   relationship	  between	  HEPA	  filtration	  and	  IL-­‐6	  levels.	  There	  was	  no	  significant	  relationship	  between	  IL-­‐6	  and	  any	  of	  the	  exposure	  variables	  that	  were	  evaluated.	  Most	  effect	  estimates	  were	  very	  small	  with	  wide	  confidence	  intervals	  rejecting	  an	  association.	  	  	  Despite	   the	   expectation	   of	   the	   same	   level	   of	   association	   between	   CRP	   and	   PM2.5	  exposure	  due	  to	  the	  physiological	  link	  between	  IL-­‐6	  and	  CRP,	  the	  results	  seem	  to	  be	  in	   agreement	  with	   a	   number	   of	   other	   studies	  which	   failed	   to	   find	   any	   significant	  effects	   of	   TRAP	   and	   WS	   PM	   on	   IL-­‐6	   levels	   in	   the	   body.	   In	   a	   recent	   controlled	  exposure	  study,	  participants	  were	  exposed	   to	  146-­‐295	  µg/m3	  WS	  PM	  (Stockfelt	  et	  al.,	  2013).	  The	  authors	   found	  no	  relationship	  between	  exposure	  to	  WS	  and	  CRP	  or	  IL-­‐6	  in	  participants.	  This	  was	  contrary	  to	  their	  previous	  exposure	  study	  where	  they	  	   70	  used	  higher	  doses	  of	  WS	  PM	  and	  found	  a	  decrease	  in	  IL-­‐6	  with	  exposure	  to	  filtered	  air	  (Barregard	  et	  al.,	  2006a).	  In	  a	  filtration	  based	  intervention	  study	  in	  Denmark,	  21	  elderly	   couples	   participated	   in	   a	   randomized,	   double-­‐blind,	   crossover	   study	  investigating	   the	   relationship	   between	   HEPA	   filtration	   and	   various	   markers	   of	  inflammation	   (Bräuner,	   Forchhammer,	   et	   al.,	   2008).	   They	   found	   no	   significant	  change	   in	   IL-­‐6	   levels	   with	   HEPA	   filtration	   of	   TRAP	   PM.	   Moreover,	   (Jacobs	   et	   al.,	  2010)	   recruited	   38	   volunteers	   that	   cycled	   in	   real	   traffic	   and	   in	   a	   laboratory	  with	  filtered	  air.	  Their	  results	  were	  inconclusive	  with	  regards	  to	  exposure	  to	  TRAP	  PM2.5	  and	  IL-­‐6	  levels.	  	  4.2.4 BCC	  and	  %PMN	  	  Band	   cells	   are	   immature	   polymorphonuclear	   leukocytes	   (PMN)	   produced	   in	   the	  bone	  marrow.	  Increases	  in	  BCC	  is	  an	  indication	  of	  stimulation	  of	  the	  immune	  system	  leading	   to	   an	   increased	   release	   of	   granulocytes	   in	   the	   bone	   marrow	   (Tan	   et	   al.,	  2000).	  There	  is	  evidence	  that	  exposure	  to	  high	  levels	  of	  ambient	  PM	  can	  potentially	  lead	  to	  a	  spike	  in	  the	  release	  of	  band	  cells	  (van	  Eeden	  et	  al.,	  2005).	  In	  a	  study	  of	  high	  level	   WS	   exposure,	   healthy	   non-­‐smoking	   firefighters	   (17-­‐60	   years	   old)	   with	   no	  chronic	  medical	  condition	  or	  prescription	  medication	  use	  were	  recruited	  (Swiston	  et	  al.,	  2008).	  During	  wildfire	  episodes,	  the	  firefighters	  were	  exposed	  to	  PM	  levels	  as	  high	  as	  2,000	  µg/m3.	  This	  study	  found	  a	  statistically	  significant	  increase	  in	  BCC	  and	  PMN	  following	  fire	  fighting,	  which	  could	  be	  indicative	  of	  systematic	  inflammation.	  In	  another	  study,	   (Sakai	  et	  al.,	  2004)	   followed	  39	  research	  expedition	  members	   from	  Japan	  to	  Antarctica	  for	  several	  months	  and	  then	  back	  to	  Japan.	  They	  found	  PM	  levels	  to	  be	  <1%	  of	  that	  measured	  in	  Japan	  because	  of	  less	  use	  of	  fossil	  fuels	  in	  Antarctica.	  The	   authors	   found	   a	   significant	   association	   between	   PM	   exposure	   and	   BCC/PMN	  with	  reductions	  after	  moving	  to	  Antarctica	  and	  increases	  after	  moving	  back	  to	  Japan.	  Despite	   the	   available	   evidence	   in	   the	   literature,	   there	   was	   no	   evidence	   of	   an	  association	   between	   HEPA	   filtration	   and	   BCC	   as	   %PMN	   in	   the	   current	   study.	  Furthermore,	   there	   was	   no	   significant	   relationship	   between	   PM2.5	   and	   %PMN.	  However,	  the	  effect	  estimates	  were	  in	  the	  hypothesized	  direction	  with	  increases	  in	  %PMN	  per	  unit	  median	   increase	   in	  PM2.5	  levels,	   though	  not	   statistically	   significant	  	   71	  and	   very	   small.	   The	   same	   trend	   was	   true	   when	   looking	   at	   the	   relationship	   with	  absorbance	  and	  LG.	  	  	  Overall,	   the	   results	   of	   this	   current	   study	   are	   not	   in	   agreement	  with	   the	   available	  literature	   discussed	   earlier.	   There	   was	   no	   significant	   relationship	   between	  BCC/%PMN	   and	   various	   exposure	   measures.	   The	   subjects	   in	   most	   studies	   that	  found	   an	   association	   between	   the	   two	   variables	  were	   exposed	   to	   PM	   levels	   up	   to	  200	  times	  greater	  than	  what	  was	  measured	  in	  the	  current	  study	  (Sakai	  et	  al.,	  2004;	  Swiston	  et	  al.,	  2008).	  The	  fact	  that	  the	  exposures	  were	  very	  low	  in	  this	  study	  could	  be	   a	   valid	   explanation	   for	   the	   inability	   to	   observe	   an	   effect	   by	   TRAP	   and	   WS	  exposure	  variables	  on	  BCC/%PMN.	  4.3 Conclusions	  	  All	  in	  all,	  this	  study	  suggests	  an	  association	  between	  TRAP	  PM2.5	  exposure	  and	  CRP	  levels	  in	  male	  participants	  but	  no	  associations	  with	  IL-­‐6,	  BCC,	  or	  %PMN.	  The	  value	  of	   “negative”	   results	   of	   this	   study	   should	  not	  be	  underestimated,	   as	   they	   could	  be	  valuable	   for	   determining	   future	   research	   direction	   evaluating	   the	   effects	   of	   TRAP	  and	  WS	  PM	  and	  also	  interventions	  on	  cardiovascular	  health.	  It	  should	  be	  noted	  that	  the	   baseline	   PM	   concentration	   was	   relatively	   low	   with	   small	   exposure	   gradients	  after	  HEPA	  filtration.	  Moreover,	  the	  population	  evaluated	  was	  relatively	  young	  and	  healthy	   without	   any	   type	   of	   CVD	   risk	   factors,	   medications	   or	   health	   conditions	  affecting	   inflammation.	   In	  such	  a	  context,	   the	   findings	  with	  regards	  to	  the	  effect	  of	  PM	  exposure	  on	  CRP	   in	  TRAP	  homes	  can	  be	  considered	  a	  valuable	  addition	   to	   the	  scientific	  literature.	  In	  addition,	  the	  results	  are	  partially	  consistent	  with	  the	  a	  priori	  hypothesis	   that	   there	   will	   be	   a	   greater	   impact	   on	   cardiovascular	   health	   in	   those	  exposed	   to	   traffic-­‐related	   PM2.5,	   	  however,	   we	   observed	   this	   impact	   only	   amongst	  male	  participants.	  The	  results	  partially	  support	  the	  hypothesis	  that	  HEPA	  filtration	  can	  be	  effective	  in	  reducing	  indoor	  PM2.5	  concentrations	  with	  some	  suggestion	  of	  an	  effect	   on	   systematic	   inflammation	   marker	   concentrations	   in	   the	   body	   in	   TRAP	  homes,	  among	  males	  only.	  	  	  	   72	  4.4 Significance	  of	  Study	  	  To	   our	   knowledge,	   this	   was	   the	   first	   study	   to	   directly	   compare	   the	   potential	  cardiovascular	   effects	   of	   WS	   and	   TRAP	   particles	   directly.	   There	   has	   been	   some	  evidence	  on	  the	  effectiveness	  of	  HEPA	  filters	  as	  an	  intervention	  for	  traffic-­‐generated	  particles	  but	  they	  are	  from	  studies	  conducted	  in	  Europe.	  Since	  a	  greater	  proportion	  of	  vehicles	  use	  diesel	  as	  fuel	  in	  Europe	  compared	  to	  North	  America,	  it	  might	  not	  be	  practical	   to	   extrapolate	   their	   results	   into	   a	   North	   American	   setting.	   This	   study	  provides	   a	   better	   insight	   into	   the	   use	   of	   HEPA	   filters	   in	   North	   America	   as	   an	  intervention	   for	   traffic-­‐generated	  PM	  exposure.	  Furthermore,	   the	  crossover	  design	  of	  this	  study	  neutralized	  potential	  confounding	  bias	  since	  individuals	  serve	  as	  their	  own	  control.	  	  Being	  the	  first	  study	  comparing	  the	  cardiovascular	  effects	  of	  PM	  exposure	  produced	  from	  TRAP	  and	  WS,	  this	  study	  is	  valuable	  for	  risk	  assessment	  of	  these	  two	  common	  sources	  of	  PM.	  The	  quantitative	  evaluation	  of	  the	  effectiveness	  of	  HEPA	  filtration	  as	  an	  intervention	  for	  the	  reduction	  of	  PM	  exposure	  from	  these	  two	  sources	  can	  assist	  us	   with	   future	   decision	   and	   policymaking	   for	   reducing	   exposure	   to	   PM.	   Future	  research	  on	  the	  health	  effects	  of	  PM	  from	  RWC	  will	  better	  enable	  us	  to	  develop	  air	  quality	  management	  strategies.	  4.5 Strengths	  	  In	   this	   study,	   several	   recommended approaches in the understanding of PM source impacts were	   addressed	   (WHO, 2007). The	   effect	   of	   PM	   from	   sources	   other	   than	  TRAP,	   assessed the	   contribution	   of	   different	   exposure	   sources	   to	   population	  exposure,	   using	   specific	   health	   markers	   other	   than	   general	   mortality	   and	   lung	  function,	   and	   finally,	   evaluating	   the	   effects	   of	   air	   pollution	   reduction	   on	   health	  outcomes	   were	   addressed	   by	   using	   a	   novel	   semi-­‐experimental	   study	   design.	  Controlled	   laboratory	   experimental	   exposure	   studies	   generally	   restrict	   exposure	  parameters	  to	  a	  specific	  air	  pollutant	  at	  very	  high	  concentrations	  or	  certain	  sources	  such	  as	  gasoline	  exhaust,	  while	  this	   is	  not	  the	  case	  in	  real	  world	  exposure	  settings	  	   73	  (Strak	  et	  al.,	  2012).	  With	  a	  semi-­‐experimental	  design,	  the	  test	  subjects	  were	  exposed	  to	   real-­‐life	   ambient	   air	   pollution	   with	   contrasting	   PM	   exposure	   sources	   and	  concentrations.	  This	  advantage	  is	  valuable	  for	  the	  applicability	  of	  this	  study’s	  results	  to	  the	  general	  population.	  4.6 Limitations	  	  As	  with	  any	  other	  study,	  the	  limitations	  of	  the	  current	  study	  should	  be	  considered.	  As	  shown	  in	  this	  study,	  PM2.5	  concentrations	  were	  very	  low	  in	  the	  sampling	  region.	  With	  such	  a	   low	  PM2.5	  concentration,	   the	  gradient	   introduced	  with	  HEPA	  filtration	  was	  also	  very	  small	  compared	  to	  other	  similar	  studies.	  As	  a	  result,	  the	  concentration	  gradient	   might	   have	   been	   too	   low	   to	   see	   any	   measurable	   change	   in	   the	   health	  variables	  of	  interest.	  	  In	  this	  study,	  RHI	  was	  used	  as	  the	  primary	  outcome	  for	  assessing	  cardiovascular	  risk	  of	   exposure	   to	   PM2.5.	   However,	   the	   available	   literature	   has	   not	   been	   entirely	  consistent	   on	   the	   extent	   that	   EndoPAT	   results	   agree	   with	   other	   well	   developed,	  clinically	  used	  methods	  (e.g.	  endothelium-­‐dependent	  brachial	  artery	  flow-­‐mediated	  dilation).	   Of	   the	   nine	   available	   studies	   comparing	   the	   two	   methods,	   six	   found	  statistically	   significant	   weak	   to	   moderate	   correlation	   (Dhindsa	   et	   al.,	   2008;	  Heffernan,	  Karas,	  Mooney,	  Patel,	  &	  Kuvin,	  2010;	  Kuvin	  et	  al.,	  2003;	  Kuvin,	  Mammen,	  Mooney,	  Alsheikh-­‐Ali,	  &	  Karas,	   2007;	  Onkelinx	   et	   al.,	   2012;	   Schnabel	   et	   al.,	   2011),	  while	   three	   found	   no	   significant	   correlation	   between	   endoPAT	   results	   and	   flow-­‐mediated	   dilation	   method	   (Aizer	   et	   al.,	   2009;	   Dickinson,	   Clifton,	   &	   Keogh,	   2011;	  Hamburg	  et	  al.,	  2011).	  Despite	   these	  differences,	   there	  have	  been	  suggestions	   that	  RHI	   is	   able	   to	  predict	   adverse	   cardiovascular	  health	   outcomes	   (Rubinshtein	   et	   al.,	  2010).	   It	   should	  be	  noted	   that	   these	   studies	  used	  different	   flow	  mediated	  dilation	  protocols,	  which	  makes	  drawing	  any	  sort	  of	  conclusion	  on	  the	  reliability	  of	  endoPAT	  very	   difficult;	   there	   is	   a	   need	   for	   more	   standardized	   research	   to	   enable	   us	   to	  conclusively	   determine	   the	   correlation	   between	   clinical	   methods	   and	   endoPAT.	  Meanwhile,	   there	   is	   a	   chance	   that	   such	   inconsistencies	   might	   have	   affected	   the	  evaluation	  of	  endothelial	  function	  in	  participants.	  	   74	  It	  should	  also	  be	  noted	  that	  many	  different	  mixed	  models	  were	  used	  in	  this	  study.	  In	  addition,	  each	  of	  the	  mixed	  models	  were	  also	  analyzed	  after	  stratifying	  for	  various	  variables.	   As	   a	   result,	   such	   a	   great	   number	   of	   models	   might	   have	   increased	   the	  chances	   of	   finding	   an	   association	   due	   to	   chance	   only	   (e.g.	   PM	   and	   CRP	   in	   TRAP	  homes	  among	  males).	  	  	  Moreover,	   2	   to	  31	  participants	  were	   excluded	   from	  different	  mixed	  models	   in	   the	  study	  due	  to	  missing	  or	  flawed	  data	  for	  different	  variables	  (e.g.	  RHI,	  CRP,	  PM2.5,	  etc.).	  However,	  it	  should	  be	  noted	  that	  the	  comparison	  of	  health	  and	  exposure	  outcomes	  between	  the	  “included”	  and	  the	  “excluded”	  participants	  did	  not	  show	  any	  significant	  differences	  between	  the	  two	  groups.	  	  	  	  	  In	  order	  to	  identify	  homes	  in	  high	  TRAP	  PM	  areas,	  the	  spatial	  model	  used	  NOx,	  which	  is	   a	   gas	   and	   not	   a	   particle,	   as	   a	   surrogate	   for	   TRAP.	   Due	   to	   the	   fact	   the	   gas	  concentration	  gradient	  as	  one	  moves	  away	  from	  traffic	  source	  is	  not	  the	  same	  as	  PM	  gradient,	  these	  areas	  may	  not	  actually	  be	  high	  in	  traffic-­‐related	  PM.	  While	  there	  are	  not	  big	   gradients	   in	  PM	  mass,	   there	   are	   somewhat	  more	  pronounced	  gradients	   in	  some	  PM	  measures	  such	  as	  black	  carbon.	  In	  addition,	  since	  most	  of	  the	  TRAP	  homes	  were	  in	  close	  proximity	  to	  major	  roads	  or	  highways	  this	  limitation	  might	  have	  had	  minimal	  effect	  on	  the	  categorization.	  	  	  	  Despite	  the	  benefits	  of	  the	  crossover	  design	  of	  this	  study	  in	  minimizing	  the	  effects	  of	  within-­‐participant	  characteristics	  of	   the	  analysis,	   treatment	  carryover	  effects	  were	  of	   concern.	  However,	   the	   carryover	   effect	   between	  HEPA	   and	  no	  HEPA	   treatment	  periods	  was	  eliminated	  by	  using	  a	  7-­‐day	  period	  for	  each	  treatment.	  According	  to	  the	  available	   literature,	   the	   half-­‐life	   of	   CRP	   is	   approximately	   19	   hours	   (Gabay	   &	  Kushner,	  1999).	  There	  have	  also	  been	  reports	  of	  a	  lag	  time	  of	  2	  days	  for	  seeing	  the	  effects	   of	   HEPA	   filtration	   on	   microvascular	   endothelial	   function	   (Bräuner	   et	   al.,	  2008;	   Karottki	   et	   al.,	   2013).	   Considering	   these	   findings,	   it	   is	   expected	   that	   there	  were	  minimal	  carryover	  effects	  at	  the	  end	  of	  each	  7-­‐day	  treatment	  session.	  	   75	  The	  initial	  goal	  was	  to	  enroll	  a	  total	  of	  100	  participants,	  50	  in	  each	  exposure	  group.	  However,	   due	   to	   the	   shorter	   length	   of	   the	   wood	   use	   season	   during	   the	   colder	  months	  of	  the	  year	  compared	  to	  traffic-­‐related	  sampling,	  only	  83	  participants	  were	  enrolled	  with	  29	  from	  WS	  homes	  and	  54	  from	  TRAP	  exposed	  homes.	  As	  a	  result,	  this	  study	  did	  not	  have	  sufficient	  statistical	  power	  to	  test	  for	  small	  differences	  in	  HEPA	  filtration	   effects	   between	   the	   two	   exposure	   groups.	   With	   the	   initial	   expected	  number	  of	  participants	  and	  an	  expectation	  to	  find	  a	  large	  difference	  between	  TRAP	  and	   WS	   PM	   exposure	   effects,	   the	   mixed	   models	   had	   high	   statistical	   power.	  Considering	   the	   fact	   that	   the	   expected	   number	   of	   participants	   was	   not	   met,	   this	  study	  might	  have	  failed	  to	  measure	  small	  differences	  between	  the	  two	  groups	  due	  to	  lack	  of	  sufficient	  statistical	  power.	  However,	  it	  should	  be	  noted	  that	  the	  probability	  of	   change	   in	   the	   direction	   of	   effect	   is	   minimal	   considering	   the	   current	   trends	  presented	  in	  the	  results	  section.	  	  Although	   electrical	   current	   meters	   were	   utilized	   to	   ensure	   HEPA	   filtration	   use	  during	   this	   study,	   there	  are	   still	   some	  uncertainties	  with	   regards	   to	  proper	  use	  of	  these	  devices	  due	  to	  the	  different	  output	  settings	  present	  on	  the	  device.	  Despite	  this	  shortfall,	   most	   homes	   had	   reductions	   in	   PM2.5	   levels	   with	   filtration	   reducing	   the	  likelihood	   of	   a	   large	   effect	   on	   the	   study.	   Moreover,	   there	   was	   a	   possibility	   for	  individuals	   to	   open	   the	   devices,	   check	   for	   filter	   presence,	   and	   disturb	   blinding;	  however,	  since	  participants	  were	  not	  aware	  that	  there	  will	  only	  be	  placebo	  filtration	  during	  one	  of	  the	  weeks,	  there	  was	  no	  reason	  to	  be	  concerned	  about	  this	  issue.	  	  Finally,	   while	   it	   was	   not	   possible	   to	   record	   air	   pollution	   exposure	   in	   other	  microenvironments	  outside	  the	  homes,	  where	  the	  participants	  spend	  about	  25%	  of	  their	  time.	  This	  should	  not	  be	  a	  major	  source	  of	  error	  affecting	  the	  effectiveness	  of	  HEPA	   filtration,	   due	   to	   the	   crossover	   nature	   of	   this	   study.	   In	   addition,	   the	   time-­‐activity	  patterns	  were	  very	  similar	  between	  the	  HEPA	  and	  no	  HEPA	  weeks	  further	  alleviating	  any	  concern.	  	  	   76	  4.7 Suggested	  Future	  Research	  	  To	   further	   evaluate	   the	   results	   of	   this	   study	   and	   to	   extend	   the	   conclusions,	  more	  research	   is	   needed.	   Future	   studies	   investigating	   the	   effects	   of	   HEPA	   filtration	   on	  cardiovascular	  health	  could	  focus	  on	  homes	  with	  higher	  levels	  of	  PM2.5.	  Higher	  levels	  will	  assist	  with	  producing	  a	  greater	  concentration	  gradient	  using	  filtration	  devices,	  which	  might	   lead	   to	  more	  measurable	  differences.	  Using	   the	  same	  novel	  approach	  employed	  in	  this	  study	  by	  utilizing	  filtration	  devices	  and	  evaluating	  more	  than	  one	  source	  of	  PM	  in	  the	  same	  study	  population	  could	  be	  useful	  in	  testing	  the	  differential	  impacts	  of	  specific	  PM	  sources	  without	  removing	  participants	  from	  their	  residences.	  	  At	  such	  low	  PM	  concentrations	  in	  the	  developed	  world,	  the	  change	  in	  various	  health	  outcomes	   could	   be	   very	   small.	   Hence,	   the	   recruitment	   of	  more	   participants	   could	  prove	   useful	   in	   increasing	   the	   statistical	   power	   of	   the	   study	   to	   detect	   smaller	  differences	   in	   effect	   between	   TRAP	   and	  WS	   PM2.5.	   Finally,	   the	   participants	   in	   this	  study	  were	  relatively	  young	  and	  healthy.	  This	  might	  limit	  the	  generalizability	  of	  this	  study	   to	   the	  general	  population.	   It	  might	  be	  valuable	   to	   study	   the	  effects	  of	  HEPA	  filtration	  among	  vulnerable	  populations	  (e.g.	  diabetics)	  as	  they	  are	  the	  group	  at	  the	  greatest	  risk	  of	  developing	  adverse	  health	  effects	   from	  exposure	   to	  PM2.5	  (Sacks	  et	  al.,	  2011).	  4.8 Implications	  on	  Public	  Health	  	  With	   the	  mounting	   evidence	   on	   the	   effects	   of	   exposure	   to	   PM2.5	   from	   combustion	  sources	   on	   cardiovascular	   health,	   there	   is	   an	   immediate	   need	   to	   reduce	   public	  exposure.	  To	  achieve	  such	  a	  goal,	   the	  decision	  making	  bodies	  will	  need	  to	  have	  an	  understanding	  of	  the	  sources	  of	  exposure	  and	  the	  potential	  differential	  effects	  that	  they	  might	   have	   on	  health.	  While	   changes	   are	   being	  brought	   about	   at	   the	   general	  population	   and	   policy	   levels,	   steps	   could	   be	   taken	   at	   individual	   level	   to	   reduce	  exposure	   to	   PM2.5.	   Although	   there	   was	   no	   evidence	   that	   HEPA	   filtration	   reduced	  health	   impacts	   in	   the	   current	   study,	   there	   was	   some	   suggestion	   of	   worsening	  systematic	  inflammation	  with	  TRAP	  PM2.5	  exposure.	  Moreover,	  this	  study	  along	  with	  	   77	  other	   available	   literature	  have	   suggested	   a	  potential	   for	  HEPA	   filtration	   to	   reduce	  indoor	   TRAP	   and	   WS	   PM2.5	   concentrations.	   Considering	   the	   toxicological	   and	  biological	   plausibility	   of	   a	   link	   between	   PM2.5	   and	   adverse	   health	   effects,	   the	  utilization	  of	  HEPA	   filtration	  at	  homes	  and	  workplaces	   can	  be	  an	   inexpensive	  and	  yet	  effective	  means	  of	  reducing	  the	  effects	  of	  air	  pollution	  on	  cardiovascular	  health	  and	  other	  health	  effects.	  The	  effect	  of	  HEPA	  filtration	  might	  be	  relatively	  small	  at	  an	  individual	   level	   but	   when	   evaluating	   the	   effects	   at	   a	   population	   level,	   there	   is	   a	  potential	   for	   substantially	   reducing	   risks	   associated	   with	   exposure	   to	   PM2.5.	  Nevertheless,	  it	  should	  be	  kept	  in	  mind	  that	  the	  most	  important	  aspects	  of	  reducing	  air	   pollution	   exposure	   is	   through	   targeting	   the	   source	   of	   pollution	   and	   designing	  interventions	   at	   the	   source.	   Meanwhile,	   HEPA	   filtration	   could	   be	   evaluated	   as	   a	  potentially	  viable	  option.	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   78	  References	  	  Aizer,	  J.,	  Karlson,	  E.,	  Chibnik,	  L.,	  Costenbader,	  K.,	  Post,	  D.,	  Liang,	  M.,	  …	  Gerhard-­‐Herman,	  M.	  (2009).	  A	  controlled	  comparison	  of	  brachial	  artery	  flow	  mediated	  dilation	  (FMD)	  and	  digital	  pulse	  amplitude	  tonometry	  (PAT)	  in	  the	  assessment	  of	  endothelial	  function	  in	  systemic	  lupus	  erythematosus.	  Lupus,	  18(3),	  235–242.	  doi:10.1177/0961203308096663	  Allen,	  R.,	  Karlen,	  B.,	  &	  Nichol,	  J.	  (2014).	  Standard	  Operating	  Procedures:	  CardiovascuLar	  Effects	  of	  Aerosols	  in	  Residences	  (CLEAR)	  Study.	  Retrieved	  from	  Allen,	  R.	  W.,	  Adar,	  S.	  D.,	  Avol,	  E.,	  Cohen,	  M.,	  Curl,	  C.	  L.,	  Larson,	  T.,	  …	  Kaufman,	  J.	  D.	  (2012).	  Modeling	  the	  Residential	  Infiltration	  of	  Outdoor	  PM2.5	  in	  the	  Multi-­‐Ethnic	  Study	  of	  Atherosclerosis	  and	  Air	  Pollution	  (MESA	  Air).	  Environmental	  Health	  Perspectives,	  120(6),	  824–830.	  doi:10.1289/ehp.1104447	  Allen,	  R.	  W.,	  Carlsten,	  C.,	  Karlen,	  B.,	  Leckie,	  S.,	  Eeden,	  S.	  V.,	  Vedal,	  S.,	  …	  Brauer,	  M.	  (2011).	  An	  Air	  Filter	  Intervention	  Study	  of	  Endothelial	  Function	  Among	  Healthy	  Adults	  in	  a	  Woodsmoke-­‐Impacted	  Community.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  183(9),	  1222–1230.	  doi:10.1164/rccm.201010-­‐1572OC	  Allen,	  R.	  W.,	  Leckie,	  S.,	  Millar,	  G.,	  &	  Brauer,	  M.	  (2009).	  The	  impact	  of	  wood	  stove	  technology	  upgrades	  on	  indoor	  residential	  air	  quality.	  Atmospheric	  Environment,	  43(37),	  5908–5915.	  doi:10.1016/j.atmosenv.2009.08.016	  Allen,	  R.	  W.,	  Mar,	  T.,	  Koenig,	  J.,	  Liu,	  L.-­‐J.	  S.,	  Gould,	  T.,	  Simpson,	  C.,	  &	  Larson,	  T.	  (2008).	  Changes	  in	  lung	  function	  and	  airway	  inflammation	  among	  asthmatic	  children	  	   79	  residing	  in	  a	  woodsmoke-­‐impacted	  urban	  area.	  Inhalation	  Toxicology,	  20(4),	  423–433.	  doi:10.1080/08958370801903826	  Anderson,	  H.	  R.,	  Favarato,	  G.,	  &	  Atkinson,	  R.	  W.	  (2013).	  Long-­‐term	  exposure	  to	  air	  pollution	  and	  the	  incidence	  of	  asthma:	  meta-­‐analysis	  of	  cohort	  studies.	  Air	  Quality,	  Atmosphere	  &	  Health,	  6(1),	  47–56.	  doi:10.1007/s11869-­‐011-­‐0144-­‐5	  Andrea	  Careless,	  W.	  M.	  B.	  C.	  M.	  of	  H.	  L.	  and	  S.	  (2004).	  Provincial	  Health	  Officer’s	  Annual	  Report	  2003:	  Air	  Quality	  in	  British	  Columbia,	  a	  Public	  Health	  Perspective.	  Retrieved	  February	  16,	  2012,	  from	  Baccarelli,	  A.,	  Martinelli,	  I.,	  Pegoraro,	  V.,	  Melly,	  S.,	  Grillo,	  P.,	  Zanobetti,	  A.,	  …	  Schwartz,	  J.	  (2009).	  Living	  near	  major	  traffic	  roads	  and	  risk	  of	  deep	  vein	  thrombosis.	  Circulation,	  119(24),	  3118–3124.	  doi:10.1161/CIRCULATIONAHA.108.836163	  Barn,	  P.,	  Larson,	  T.,	  Noullett,	  M.,	  Kennedy,	  S.,	  Copes,	  R.,	  &	  Brauer,	  M.	  (2008).	  Infiltration	  of	  forest	  fire	  and	  residential	  wood	  smoke:	  an	  evaluation	  of	  air	  cleaner	  effectiveness.	  Journal	  of	  Exposure	  Science	  &	  Environmental	  Epidemiology,	  18(5),	  503–511.	  doi:10.1038/sj.jes.7500640	  Barregard,	  L.,	  Sällsten,	  G.,	  Gustafson,	  P.,	  Andersson,	  L.,	  Johansson,	  L.,	  Basu,	  S.,	  &	  Stigendal,	  L.	  (2006a).	  Experimental	  exposure	  to	  wood-­‐smoke	  particles	  in	  healthy	  humans:	  effects	  on	  markers	  of	  inflammation,	  coagulation,	  and	  lipid	  peroxidation.	  Inhalation	  Toxicology,	  18(11),	  845–853.	  doi:10.1080/08958370600685798	  	   80	  Barregard,	  L.,	  Sällsten,	  G.,	  Gustafson,	  P.,	  Andersson,	  L.,	  Johansson,	  L.,	  Basu,	  S.,	  &	  Stigendal,	  L.	  (2006b).	  Experimental	  exposure	  to	  wood-­‐smoke	  particles	  in	  healthy	  humans:	  effects	  on	  markers	  of	  inflammation,	  coagulation,	  and	  lipid	  peroxidation.	  Inhalation	  Toxicology,	  18(11),	  845–853.	  doi:10.1080/08958370600685798	  Batterman,	  S.,	  Du,	  L.,	  Mentz,	  G.,	  Mukherjee,	  B.,	  Parker,	  E.,	  Godwin,	  C.,	  …	  Lewis,	  T.	  (2011).	  Particulate	  matter	  concentrations	  in	  residences:	  an	  intervention	  study	  evaluating	  stand-­‐alone	  filters	  and	  air	  conditioners.	  Indoor	  Air.	  doi:10.1111/j.1600-­‐0668.2011.00761.x	  Bell,	  M.	  L.,	  &	  Davis,	  D.	  L.	  (2001).	  Reassessment	  of	  the	  lethal	  London	  fog	  of	  1952:	  novel	  indicators	  of	  acute	  and	  chronic	  consequences	  of	  acute	  exposure	  to	  air	  pollution.	  Environmental	  Health	  Perspectives,	  109(Suppl	  3),	  389–394.	  Bergauff,	  M.	  A.,	  Ward,	  T.	  J.,	  Noonan,	  C.	  W.,	  &	  Palmer,	  C.	  P.	  (2009).	  The	  effect	  of	  a	  woodstove	  changeout	  on	  ambient	  levels	  of	  PM2.5	  and	  chemical	  tracers	  for	  woodsmoke	  in	  Libby,	  Montana.	  Atmospheric	  Environment,	  43(18),	  2938–2943.	  doi:10.1016/j.atmosenv.2009.02.055	  Boman,	  B.	  C.,	  Forsberg,	  A.	  B.,	  &	  Järvholm,	  B.	  G.	  (2003).	  Adverse	  health	  effects	  from	  ambient	  air	  pollution	  in	  relation	  to	  residential	  wood	  combustion	  in	  modern	  society.	  Scandinavian	  Journal	  of	  Work,	  Environment	  &	  Health,	  29(4),	  251–260.	  Brauer,	  M.,	  Lencar,	  C.,	  Tamburic,	  L.,	  Koehoorn,	  M.,	  Demers,	  P.,	  &	  Karr,	  C.	  (2008).	  A	  Cohort	  Study	  of	  Traffic-­‐Related	  Air	  Pollution	  Impacts	  on	  Birth	  Outcomes.	  Environmental	  Health	  Perspectives,	  116(5),	  680–686.	  doi:10.1289/ehp.10952	  	   81	  Brauer,	  M.,	  Reynolds,	  C.,	  &	  Hystad,	  P.	  (2012).	  Traffic-­‐related	  air	  pollution	  and	  health :	  a	  Canadian	  perspective	  on	  scientific	  evidence	  and	  potential	  exposure-­‐mitigation	  strategies.	  Retrieved	  from	  Bräuner,	  E.,	  Forchhammer,	  L.,	  Møller,	  P.,	  Barregard,	  L.,	  Gunnarsen,	  L.,	  Afshari,	  A.,	  …	  Loft,	  S.	  (2008).	  Indoor	  particles	  affect	  vascular	  function	  in	  the	  aged:	  an	  air	  filtration-­‐based	  intervention	  study.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  177(4),	  419–425.	  doi:10.1164/rccm.200704-­‐632OC	  Bräuner,	  E.,	  Møller,	  P.,	  Barregard,	  L.,	  Dragsted,	  L.	  O.,	  Glasius,	  M.,	  Wåhlin,	  P.,	  …	  Loft,	  S.	  (2008).	  Exposure	  to	  ambient	  concentrations	  of	  particulate	  air	  pollution	  does	  not	  influence	  vascular	  function	  or	  inflammatory	  pathways	  in	  young	  healthy	  individuals.	  Particle	  and	  Fibre	  Toxicology,	  5,	  13.	  doi:10.1186/1743-­‐8977-­‐5-­‐13	  Briet,	  M.,	  Collin,	  C.,	  Laurent,	  S.,	  Tan,	  A.,	  Azizi,	  M.,	  Agharazii,	  M.,	  …	  Boutouyrie,	  P.	  (2007).	  Endothelial	  Function	  and	  Chronic	  Exposure	  to	  Air	  Pollution	  in	  Normal	  Male	  Subjects.	  Hypertension,	  50(5),	  970–976.	  doi:10.1161/HYPERTENSIONAHA.107.095844	  Brook,	  R.	  D.,	  Rajagopalan,	  S.,	  Pope,	  C.	  A.,	  Brook,	  J.	  R.,	  Bhatnagar,	  A.,	  Diez-­‐Roux,	  A.	  V.,	  …	  Kaufman,	  J.	  D.	  (2010).	  Particulate	  Matter	  Air	  Pollution	  and	  Cardiovascular	  Disease	  An	  Update	  to	  the	  Scientific	  Statement	  From	  the	  American	  Heart	  Association.	  Circulation,	  121(21),	  2331–2378.	  doi:10.1161/CIR.0b013e3181dbece1	  	   82	  Brook,	  R.	  D.,	  Urch,	  B.,	  Dvonch,	  J.	  T.,	  Bard,	  R.	  L.,	  Speck,	  M.,	  Keeler,	  G.,	  …	  Brook,	  J.	  R.	  (2009).	  Insights	  into	  the	  mechanisms	  and	  mediators	  of	  the	  effects	  of	  air	  pollution	  exposure	  on	  blood	  pressure	  and	  vascular	  function	  in	  healthy	  humans.	  Hypertension,	  54(3),	  659–667.	  doi:10.1161/HYPERTENSIONAHA.109.130237	  Browning,	  K.	  G.,	  Koenig,	  J.	  Q.,	  Checkoway,	  H.,	  Larson,	  T.	  V.,	  &	  Pierson,	  W.	  E.	  (1990).	  A	  Questionnaire	  Study	  of	  Respiratory	  Health	  in	  Areas	  of	  High	  and	  Low	  Ambient	  Wood	  Smoke	  Pollution.	  Pediatric	  Asthma,	  Allergy	  &	  Immunology,	  4(3),	  183–191.	  doi:10.1089/pai.1990.4.183	  Brunekreef,	  B.,	  &	  Forsberg,	  B.	  (2005).	  Epidemiological	  evidence	  of	  effects	  of	  coarse	  airborne	  particles	  on	  health.	  The	  European	  Respiratory	  Journal,	  26(2),	  309–318.	  doi:10.1183/09031936.05.00001805	  Burnett,	  R.	  T.,	  Brook,	  J.,	  Dann,	  T.,	  Delocla,	  C.,	  Philips,	  O.,	  Cakmak,	  S.,	  …	  Krewski,	  D.	  (2000).	  Association	  between	  particulate-­‐	  and	  gas-­‐phase	  components	  of	  urban	  air	  pollution	  and	  daily	  mortality	  in	  eight	  Canadian	  cities.	  Inhalation	  Toxicology,	  12	  Suppl	  4,	  15–39.	  doi:10.1080/08958370050164851	  Cacciola,	  R.	  R.,	  Sarvà,	  M.,	  &	  Polosa,	  R.	  (2002).	  Adverse	  respiratory	  effects	  and	  allergic	  susceptibility	  in	  relation	  to	  particulate	  air	  pollution:	  flirting	  with	  disaster.	  Allergy,	  57(4),	  281–286.	  Canadian	  Medical	  Association.	  (2008).	  No	  Breathing	  Room:	  National	  Illness	  Costs	  of	  Air	  Pollution.	  	   83	  Cass,	  G.	  R.	  (1998).	  Organic	  molecular	  tracers	  for	  particulate	  air	  pollution	  sources.	  TrAC	  Trends	  in	  Analytical	  Chemistry,	  17(6),	  356–366.	  doi:10.1016/S0165-­‐9936(98)00040-­‐5	  Cesari,	  M.,	  Penninx,	  B.	  W.	  J.	  H.,	  Newman,	  A.	  B.,	  Kritchevsky,	  S.	  B.,	  Nicklas,	  B.	  J.,	  Sutton-­‐Tyrrell,	  K.,	  …	  Pahor,	  M.	  (2003).	  Inflammatory	  markers	  and	  onset	  of	  cardiovascular	  events:	  results	  from	  the	  Health	  ABC	  study.	  Circulation,	  108(19),	  2317–2322.	  doi:10.1161/01.CIR.0000097109.90783.FC	  Chow,	  J.	  C.	  (1995).	  Measurement	  methods	  to	  determine	  compliance	  with	  ambient	  air	  quality	  standards	  for	  suspended	  particles.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association	  (1995),	  45(5),	  320–382.	  Cifuentes,	  L.	  A.,	  Vega,	  J.,	  Köpfer,	  K.,	  &	  Lave,	  L.	  B.	  (2000).	  Effect	  of	  the	  fine	  fraction	  of	  particulate	  matter	  versus	  the	  coarse	  mass	  and	  other	  pollutants	  on	  daily	  mortality	  in	  Santiago,	  Chile.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association	  (1995),	  50(8),	  1287–1298.	  Crouse,	  D.	  L.,	  Peters,	  P.	  A.,	  van	  Donkelaar,	  A.,	  Goldberg,	  M.	  S.,	  Villeneuve,	  P.	  J.,	  Brion,	  O.,	  …	  Burnett,	  R.	  T.	  (2012).	  Risk	  of	  nonaccidental	  and	  cardiovascular	  mortality	  in	  relation	  to	  long-­‐term	  exposure	  to	  low	  concentrations	  of	  fine	  particulate	  matter:	  a	  Canadian	  national-­‐level	  cohort	  study.	  Environmental	  Health	  Perspectives,	  120(5),	  708–714.	  doi:10.1289/ehp.1104049	  De	  Kok,	  T.	  M.	  C.	  M.,	  Driece,	  H.	  A.	  L.,	  Hogervorst,	  J.	  G.	  F.,	  &	  Briedé,	  J.	  J.	  (2006).	  Toxicological	  assessment	  of	  ambient	  and	  traffic-­‐related	  particulate	  matter:	  a	  review	  of	  recent	  studies.	  Mutation	  Research,	  613(2-­‐3),	  103–122.	  doi:10.1016/j.mrrev.2006.07.001	  	   84	  deB.	  Richter,	  D.,	  Jenkins,	  D.	  H.,	  Karakash,	  J.	  T.,	  Knight,	  J.,	  McCreery,	  L.	  R.,	  &	  Nemestothy,	  K.	  P.	  (2009).	  Wood	  Energy	  in	  America.	  Science,	  323(5920),	  1432–1433.	  doi:10.1126/science.1166214	  Dhindsa,	  M.,	  Sommerlad,	  S.	  M.,	  DeVan,	  A.	  E.,	  Barnes,	  J.	  N.,	  Sugawara,	  J.,	  Ley,	  O.,	  &	  Tanaka,	  H.	  (2008).	  Interrelationships	  among	  noninvasive	  measures	  of	  postischemic	  macro-­‐	  and	  microvascular	  reactivity.	  Journal	  of	  Applied	  Physiology	  (Bethesda,	  Md.:	  1985),	  105(2),	  427–432.	  doi:10.1152/japplphysiol.90431.2008	  Dickinson,	  K.	  M.,	  Clifton,	  P.	  M.,	  &	  Keogh,	  J.	  B.	  (2011).	  Endothelial	  function	  is	  impaired	  after	  a	  high-­‐salt	  meal	  in	  healthy	  subjects.	  The	  American	  Journal	  of	  Clinical	  Nutrition,	  93(3),	  500–505.	  doi:10.3945/ajcn.110.006155	  Dockery,	  D.	  W.	  (2001).	  Epidemiologic	  evidence	  of	  cardiovascular	  effects	  of	  particulate	  air	  pollution.	  Environmental	  Health	  Perspectives,	  109(Suppl	  4),	  483–486.	  Dockery,	  D.	  W.,	  Pope,	  C.	  A.,	  3rd,	  Xu,	  X.,	  Spengler,	  J.	  D.,	  Ware,	  J.	  H.,	  Fay,	  M.	  E.,	  …	  Speizer,	  F.	  E.	  (1993).	  An	  association	  between	  air	  pollution	  and	  mortality	  in	  six	  U.S.	  cities.	  The	  New	  England	  Journal	  of	  Medicine,	  329(24),	  1753–1759.	  doi:10.1056/NEJM199312093292401	  Dominici	  F,	  Peng	  RD,	  Bell	  ML,	  &	  et	  al.	  (2006).	  FIne	  particulate	  air	  pollution	  and	  hospital	  admission	  for	  cardiovascular	  and	  respiratory	  diseases.	  JAMA,	  295(10),	  1127–1134.	  doi:10.1001/jama.295.10.1127	  Dubowsky,	  S.	  D.,	  Suh,	  H.,	  Schwartz,	  J.,	  Coull,	  B.	  A.,	  &	  Gold,	  D.	  R.	  (2006).	  Diabetes,	  obesity,	  and	  hypertension	  may	  enhance	  associations	  between	  air	  pollution	  	   85	  and	  markers	  of	  systemic	  inflammation.	  Environmental	  Health	  Perspectives,	  114(7),	  992–998.	  Ezzati,	  M.,	  &	  Kammen,	  D.	  M.	  (2002).	  The	  health	  impacts	  of	  exposure	  to	  indoor	  air	  pollution	  from	  solid	  fuels	  in	  developing	  countries:	  knowledge,	  gaps,	  and	  data	  needs.	  Environmental	  Health	  Perspectives,	  110(11),	  1057–1068.	  Fang,	  S.	  C.,	  Mehta,	  A.	  J.,	  Alexeeff,	  S.	  E.,	  Gryparis,	  A.,	  Coull,	  B.,	  Vokonas,	  P.,	  …	  Schwartz,	  J.	  (2012).	  Residential	  black	  carbon	  exposure	  and	  circulating	  markers	  of	  systemic	  inflammation	  in	  elderly	  males:	  the	  normative	  aging	  study.	  Environmental	  Health	  Perspectives,	  120(5),	  674–680.	  doi:10.1289/ehp.1103982	  Fisk,	  W.	  J.	  (2013).	  Health	  benefits	  of	  particle	  filtration.	  Indoor	  Air,	  23(5),	  357–368.	  doi:10.1111/ina.12036	  Fisk,	  W.	  J.,	  Faulkner,	  D.,	  Palonen,	  J.,	  &	  Seppanen,	  O.	  (2002).	  Performance	  and	  costs	  of	  particle	  air	  filtration	  technologies.	  Indoor	  Air,	  12(4),	  223–234.	  Forchhammer,	  L.,	  Møller,	  P.,	  Riddervold,	  I.	  S.,	  Bønløkke,	  J.,	  Massling,	  A.,	  Sigsgaard,	  T.,	  &	  Loft,	  S.	  (2012).	  Controlled	  human	  wood	  smoke	  exposure:	  oxidative	  stress,	  inflammation	  and	  microvascular	  function.	  Particle	  and	  Fibre	  Toxicology,	  9,	  7.	  doi:10.1186/1743-­‐8977-­‐9-­‐7	  Fraser,	  M.	  P.,	  &	  Lakshmanan,	  K.	  (2000).	  Using	  Levoglucosan	  as	  a	  Molecular	  Marker	  for	  the	  Long-­‐Range	  Transport	  of	  Biomass	  Combustion	  Aerosols.	  Environ.	  Sci.	  Technol.,	  34(21),	  4560–4564.	  doi:10.1021/es991229l	  Fuks,	  K.,	  Moebus,	  S.,	  Hertel,	  S.,	  Viehmann,	  A.,	  Nonnemacher,	  M.,	  Dragano,	  N.,	  …	  Heinz	  Nixdorf	  Recall	  Study	  Investigative	  Group.	  (2011).	  Long-­‐term	  urban	  	   86	  particulate	  air	  pollution,	  traffic	  noise,	  and	  arterial	  blood	  pressure.	  Environmental	  Health	  Perspectives,	  119(12),	  1706–1711.	  doi:10.1289/ehp.1103564	  Gabay,	  C.,	  &	  Kushner,	  I.	  (1999).	  Acute-­‐phase	  proteins	  and	  other	  systemic	  responses	  to	  inflammation.	  The	  New	  England	  Journal	  of	  Medicine,	  340(6),	  448–454.	  doi:10.1056/NEJM199902113400607	  Gan,	  W.	  Q.,	  FitzGerald,	  J.	  M.,	  Carlsten,	  C.,	  Sadatsafavi,	  M.,	  &	  Brauer,	  M.	  (2013).	  Associations	  of	  ambient	  air	  pollution	  with	  chronic	  obstructive	  pulmonary	  disease	  hospitalization	  and	  mortality.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  187(7),	  721–727.	  doi:10.1164/rccm.201211-­‐2004OC	  Gan,	  W.	  Q.,	  Koehoorn,	  M.,	  Davies,	  H.	  W.,	  Demers,	  P.	  A.,	  Tamburic,	  L.,	  &	  Brauer,	  M.	  (2011).	  Long-­‐term	  exposure	  to	  traffic-­‐related	  air	  pollution	  and	  the	  risk	  of	  coronary	  heart	  disease	  hospitalization	  and	  mortality.	  Environmental	  Health	  Perspectives,	  119(4),	  501–507.	  doi:10.1289/ehp.1002511	  Gan,	  W.	  Q.,	  Tamburic,	  L.,	  Davies,	  H.	  W.,	  Demers,	  P.	  A.,	  Koehoorn,	  M.,	  &	  Brauer,	  M.	  (2010).	  Changes	  in	  residential	  proximity	  to	  road	  traffic	  and	  the	  risk	  of	  death	  from	  coronary	  heart	  disease.	  Epidemiology	  (Cambridge,	  Mass.),	  21(5),	  642–649.	  doi:10.1097/EDE.0b013e3181e89f19	  Gehring,	  U.,	  Tamburic,	  L.,	  Sbihi,	  H.,	  Davies,	  H.,	  &	  Brauer,	  M.	  (2014).	  Environment	  and	  Health	  Abstracts	  –	  The	  impact	  of	  noise	  and	  air	  pollution	  exposure	  on	  pregnancy	  outcomes,	  In	  Press.	  Retrieved	  from­‐4-­‐28-­‐03/	  	   87	  Geng,	  F.,	  Hua,	  J.,	  Mu,	  Z.,	  Peng,	  L.,	  Xu,	  X.,	  Chen,	  R.,	  &	  Kan,	  H.	  (2013).	  Differentiating	  the	  associations	  of	  black	  carbon	  and	  fine	  particle	  with	  daily	  mortality	  in	  a	  Chinese	  city.	  Environmental	  Research,	  120,	  27–32.	  doi:10.1016/j.envres.2012.08.007	  Ghio,	  A.	  J.,	  Stonehuerner,	  J.,	  Dailey,	  L.	  A.,	  &	  Carter,	  J.	  D.	  (1999).	  Metals	  associated	  with	  both	  the	  water-­‐soluble	  and	  insoluble	  fractions	  of	  an	  ambient	  air	  pollution	  particle	  catalyze	  an	  oxidative	  stress.	  Inhalation	  Toxicology,	  11(1),	  37–49.	  doi:10.1080/089583799197258	  Giles,	  L.	  V.,	  Barn,	  P.,	  Kunzli,	  N.,	  Romieu,	  I.,	  Mittleman,	  M.	  A.,	  van	  Eeden,	  S.,	  …	  Brauer,	  M.	  (2011).	  From	  Good	  Intentions	  to	  Proven	  Interventions:	  Effectiveness	  of	  Actions	  to	  Reduce	  the	  Health	  Impacts	  of	  Air	  Pollution.	  Environmental	  Health	  Perspectives,	  119(1),	  29–36.	  doi:10.1289/ehp.1002246	  Glasius,	  M.,	  Ketzel,	  M.,	  Wåhlin,	  P.,	  Jensen,	  B.,	  Mønster,	  J.,	  Berkowicz,	  R.,	  &	  Palmgren,	  F.	  (2006).	  Impact	  of	  wood	  combustion	  on	  particle	  levels	  in	  a	  residential	  area	  in	  Denmark.	  Atmospheric	  Environment,	  40(37),	  7115–7124.	  doi:10.1016/j.atmosenv.2006.06.047	  Goldberg,	  M.	  S.,	  Burnett,	  R.	  T.,	  Bailar,	  J.	  C.,	  3rd,	  Brook,	  J.,	  Bonvalot,	  Y.,	  Tamblyn,	  R.,	  …	  Vincent,	  R.	  (2001).	  The	  association	  between	  daily	  mortality	  and	  ambient	  air	  particle	  pollution	  in	  Montreal,	  Quebec.	  2.	  Cause-­‐specific	  mortality.	  Environmental	  Research,	  86(1),	  26–36.	  doi:10.1006/enrs.2001.4243	  Guillén,	  M.	  D.,	  &	  Ibargoitia,	  M.	  L.	  (1999).	  Influence	  of	  the	  moisture	  content	  on	  the	  composition	  of	  the	  liquid	  smoke	  produced	  in	  the	  pyrolysis	  process	  of	  Fagus	  	   88	  sylvatica	  L.	  wood.	  Journal	  of	  Agricultural	  and	  Food	  Chemistry,	  47(10),	  4126–4136.	  Hamburg,	  N.	  M.,	  Palmisano,	  J.,	  Larson,	  M.	  G.,	  Sullivan,	  L.	  M.,	  Lehman,	  B.	  T.,	  Vasan,	  R.	  S.,	  …	  Benjamin,	  E.	  J.	  (2011).	  Relation	  of	  brachial	  and	  digital	  measures	  of	  vascular	  function	  in	  the	  community:	  the	  Framingham	  heart	  study.	  Hypertension,	  57(3),	  390–396.	  doi:10.1161/HYPERTENSIONAHA.110.160812	  Han,	  X.,	  &	  Naeher,	  L.	  P.	  (2006).	  A	  review	  of	  traffic-­‐related	  air	  pollution	  exposure	  assessment	  studies	  in	  the	  developing	  world.	  Environment	  International,	  32(1),	  106–120.	  doi:10.1016/j.envint.2005.05.020	  Health	  Effects	  Institute.	  (2010a).	  Traffic-­‐Related	  Air	  Pollution:	  A	  Critical	  Review	  of	  the	  Literature	  on	  Emissions,	  Exposure,	  and	  Health	  Effects.	  Retrieved	  from	  Health	  Effects	  Institute.	  (2010b).	  Traffic-­‐Related	  Air	  Pollution:	  A	  Critical	  Review	  of	  the	  Literature	  on	  Emissions,	  Exposure,	  and	  Health	  Effects	  (Special	  Reports	  No.	  17).	  Hedberg,	  E.,	  Kristensson,	  A.,	  Ohlsson,	  M.,	  Johansson,	  C.,	  Johansson,	  P.-­‐Å.,	  Swietlicki,	  E.,	  …	  Westerholm,	  R.	  (2002).	  Chemical	  and	  physical	  characterization	  of	  emissions	  from	  birch	  wood	  combustion	  in	  a	  wood	  stove.	  Atmospheric	  Environment,	  36(30),	  4823–4837.	  doi:10.1016/S1352-­‐2310(02)00417-­‐X	  Heffernan,	  K.	  S.,	  Karas,	  R.	  H.,	  Mooney,	  P.	  J.,	  Patel,	  A.	  R.,	  &	  Kuvin,	  J.	  T.	  (2010).	  Pulse	  wave	  amplitude	  is	  associated	  with	  brachial	  artery	  diameter:	  implications	  for	  gender	  differences	  in	  microvascular	  function.	  Vascular	  Medicine	  (London,	  England),	  15(1),	  39–45.	  doi:10.1177/1358863X09349523	  	   89	  Hejl,	  A.	  M.,	  Adetona,	  O.,	  Diaz-­‐Sanchez,	  D.,	  Carter,	  J.	  D.,	  Commodore,	  A.	  A.,	  Rathbun,	  S.	  L.,	  &	  Naeher,	  L.	  P.	  (2013).	  Inflammatory	  effects	  of	  woodsmoke	  exposure	  among	  wildland	  firefighters	  working	  at	  prescribed	  burns	  at	  the	  Savannah	  River	  Site,	  SC.	  Journal	  of	  Occupational	  and	  Environmental	  Hygiene,	  10(4),	  173–180.	  doi:10.1080/15459624.2012.760064	  Helfand,	  W.	  H.,	  Lazarus,	  J.,	  &	  Theerman,	  P.	  (2001).	  Donora,	  Pennsylvania:	  an	  environmental	  disaster	  of	  the	  20th	  century.	  American	  Journal	  of	  Public	  Health,	  91(4),	  553.	  Henderson,	  S.	  B.,	  Beckerman,	  B.,	  Jerrett,	  M.,	  &	  Brauer,	  M.	  (2007).	  Application	  of	  land	  use	  regression	  to	  estimate	  long-­‐term	  concentrations	  of	  traffic-­‐related	  nitrogen	  oxides	  and	  fine	  particulate	  matter.	  Environmental	  Science	  &	  Technology,	  41(7),	  2422–2428.	  Holgate,	  S.	  .,	  Koren,	  H.	  .,	  Samet,	  J.	  .,	  &	  Maynard,	  R.	  .	  (1999).	  Air	  pollution	  and	  health.	  San	  Diego:	  Academic	  Press.	  International	  Agency	  for	  Research	  on	  Cancer.	  (2012).	  IARC	  Monographs-­‐	  Classifications.	  Retrieved	  June	  3,	  2012,	  from	  Ito,	  K.,	  De	  Leon,	  S.	  F.,	  &	  Lippmann,	  M.	  (2005).	  Associations	  between	  ozone	  and	  daily	  mortality:	  analysis	  and	  meta-­‐analysis.	  Epidemiology	  (Cambridge,	  Mass.),	  16(4),	  446–457.	  Jacobs,	  L.,	  Nawrot,	  T.	  S.,	  Geus,	  B.	  de,	  Meeusen,	  R.,	  Degraeuwe,	  B.,	  Bernard,	  A.,	  …	  Panis,	  L.	  I.	  (2010).	  Subclinical	  responses	  in	  healthy	  cyclists	  briefly	  exposed	  to	  	   90	  traffic-­‐related	  air	  pollution:	  an	  intervention	  study.	  Environmental	  Health,	  9(1),	  64.	  doi:10.1186/1476-­‐069X-­‐9-­‐64	  Janssen,	  N.	  A.,	  de	  Hartog,	  J.	  J.,	  Hoek,	  G.,	  Brunekreef,	  B.,	  Lanki,	  T.,	  Timonen,	  K.	  L.,	  &	  Pekkanen,	  J.	  (2000).	  Personal	  exposure	  to	  fine	  particulate	  matter	  in	  elderly	  subjects:	  relation	  between	  personal,	  indoor,	  and	  outdoor	  concentrations.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association	  (1995),	  50(7),	  1133–1143.	  Janssen,	  N.	  A.	  H.,	  Hoek,	  G.,	  Simic-­‐Lawson,	  M.,	  Fischer,	  P.,	  van	  Bree,	  L.,	  ten	  Brink,	  H.,	  …	  Cassee,	  F.	  R.	  (2011).	  Black	  carbon	  as	  an	  additional	  indicator	  of	  the	  adverse	  health	  effects	  of	  airborne	  particles	  compared	  with	  PM10	  and	  PM2.5.	  Environmental	  Health	  Perspectives,	  119(12),	  1691–1699.	  doi:10.1289/ehp.1003369	  Jerrett,	  M.,	  Burnett,	  R.	  T.,	  Pope,	  C.	  A.,	  Ito,	  K.,	  Thurston,	  G.,	  Krewski,	  D.,	  …	  Thun,	  M.	  (2009).	  Long-­‐Term	  Ozone	  Exposure	  and	  Mortality.	  New	  England	  Journal	  of	  Medicine,	  360(11),	  1085–1095.	  doi:10.1056/NEJMoa0803894	  Jialal,	  I.,	  Devaraj,	  S.,	  &	  Venugopal,	  S.	  K.	  (2004).	  C-­‐reactive	  protein:	  risk	  marker	  or	  mediator	  in	  atherothrombosis?	  Hypertension,	  44(1),	  6–11.	  doi:10.1161/01.HYP.0000130484.20501.df	  Johnston,	  F.	  H.,	  Hanigan,	  I.	  C.,	  Henderson,	  S.	  B.,	  &	  Morgan,	  G.	  G.	  (2013).	  Evaluation	  of	  interventions	  to	  reduce	  air	  pollution	  from	  biomass	  smoke	  on	  mortality	  in	  Launceston,	  Australia:	  retrospective	  analysis	  of	  daily	  mortality,	  1994-­‐2007.	  BMJ,	  346(jan08	  12),	  e8446–e8446.	  doi:10.1136/bmj.e8446	  	   91	  Kan,	  H.,	  Chen,	  R.,	  &	  Tong,	  S.	  (2012).	  Ambient	  air	  pollution,	  climate	  change,	  and	  population	  health	  in	  China.	  Environment	  International,	  42,	  10–19.	  doi:10.1016/j.envint.2011.03.003	  Kaptoge,	  S.,	  Pennells,	  L.,	  Wood,	  A.	  M.,	  White,	  I.	  R.,	  Gao,	  P.,	  Walker,	  M.,	  …	  Danesh,	  J.	  (2012).	  C-­‐reactive	  protein,	  fibrinogen,	  and	  cardiovascular	  disease	  prediction.	  The	  New	  England	  Journal	  of	  Medicine,	  367(14),	  1310–1320.	  doi:10.1056/NEJMoa1107477	  Karottki,	  D.	  G.,	  Spilak,	  M.,	  Frederiksen,	  M.,	  Gunnarsen,	  L.,	  Brauner,	  E.	  V.,	  Kolarik,	  B.,	  …	  Loft,	  S.	  (2013).	  An	  indoor	  air	  filtration	  study	  in	  homes	  of	  elderly:	  cardiovascular	  and	  respiratory	  effects	  of	  exposure	  to	  particulate	  matter.	  Environmental	  Health,	  12(1),	  116.	  doi:10.1186/1476-­‐069X-­‐12-­‐116	  Kelly,	  F.	  J.,	  &	  Fussell,	  J.	  C.	  (2012).	  Size,	  source	  and	  chemical	  composition	  as	  determinants	  of	  toxicity	  attributable	  to	  ambient	  particulate	  matter.	  Atmospheric	  Environment,	  60,	  504–526.	  doi:10.1016/j.atmosenv.2012.06.039	  Keuken,	  M.	  P.,	  Jonkers,	  S.,	  Zandveld,	  P.,	  Voogt,	  M.,	  &	  Elshout	  van	  den,	  S.	  (2012).	  Elemental	  carbon	  as	  an	  indicator	  for	  evaluating	  the	  impact	  of	  traffic	  measures	  on	  air	  quality	  and	  health.	  Atmospheric	  Environment,	  61,	  1–8.	  doi:10.1016/j.atmosenv.2012.07.009	  Kinney,	  P.	  L.,	  Aggarwal,	  M.,	  Northridge,	  M.	  E.,	  Janssen,	  N.	  A.,	  &	  Shepard,	  P.	  (2000).	  Airborne	  concentrations	  of	  PM(2.5)	  and	  diesel	  exhaust	  particles	  on	  Harlem	  sidewalks:	  a	  community-­‐based	  pilot	  study.	  Environmental	  Health	  Perspectives,	  108(3),	  213–218.	  	   92	  Kloog,	  I.,	  Ridgway,	  B.,	  Koutrakis,	  P.,	  Coull,	  B.	  A.,	  &	  Schwartz,	  J.	  D.	  (2013).	  Long-­‐	  and	  short-­‐term	  exposure	  to	  PM2.5	  and	  mortality:	  using	  novel	  exposure	  models.	  Epidemiology	  (Cambridge,	  Mass.),	  24(4),	  555–561.	  doi:10.1097/EDE.0b013e318294beaa	  Ko,	  F.	  W.	  S.,	  &	  Hui,	  D.	  S.	  C.	  (2012).	  Air	  pollution	  and	  chronic	  obstructive	  pulmonary	  disease.	  Respirology	  (Carlton,	  Vic.),	  17(3),	  395–401.	  doi:10.1111/j.1440-­‐1843.2011.02112.x	  Kocbach,	  A.,	  Herseth,	  J.	  I.,	  Låg,	  M.,	  Refsnes,	  M.,	  &	  Schwarze,	  P.	  E.	  (2008).	  Particles	  from	  wood	  smoke	  and	  traffic	  induce	  differential	  pro-­‐inflammatory	  response	  patterns	  in	  co-­‐cultures.	  Toxicology	  and	  Applied	  Pharmacology,	  232(2),	  317–326.	  doi:10.1016/j.taap.2008.07.002	  Kocbach	  Bølling,	  A.,	  Pagels,	  J.,	  Yttri,	  K.	  E.,	  Barregard,	  L.,	  Sallsten,	  G.,	  Schwarze,	  P.	  E.,	  &	  Boman,	  C.	  (2009).	  Health	  effects	  of	  residential	  wood	  smoke	  particles:	  the	  importance	  of	  combustion	  conditions	  and	  physicochemical	  particle	  properties.	  Particle	  and	  Fibre	  Toxicology,	  6(1),	  29.	  doi:10.1186/1743-­‐8977-­‐6-­‐29	  Koenig,	  J.	  Q.,	  Larson,	  T.	  V.,	  Hanley,	  Q.	  S.,	  Rebolledo,	  V.,	  Dumler,	  K.,	  Checkoway,	  H.,	  …	  Pierson,	  W.	  E.	  (1993).	  Pulmonary	  function	  changes	  in	  children	  associated	  with	  fine	  particulate	  matter.	  Environmental	  Research,	  63(1),	  26–38.	  Koken,	  P.	  J.	  M.,	  Piver,	  W.	  T.,	  Ye,	  F.,	  Elixhauser,	  A.,	  Olsen,	  L.	  M.,	  &	  Portier,	  C.	  J.	  (2003).	  Temperature,	  air	  pollution,	  and	  hospitalization	  for	  cardiovascular	  diseases	  among	  elderly	  people	  in	  Denver.	  Environmental	  Health	  Perspectives,	  111(10),	  1312–1317.	  	   93	  Künzli,	  N.,	  Jerrett,	  M.,	  Garcia-­‐Esteban,	  R.,	  Basagaña,	  X.,	  Beckermann,	  B.,	  Gilliland,	  F.,	  …	  Mack,	  W.	  J.	  (2010).	  Ambient	  Air	  Pollution	  and	  the	  Progression	  of	  Atherosclerosis	  in	  Adults.	  PLoS	  ONE,	  5(2),	  e9096.	  doi:10.1371/journal.pone.0009096	  Kuvin,	  J.	  T.,	  Mammen,	  A.,	  Mooney,	  P.,	  Alsheikh-­‐Ali,	  A.	  A.,	  &	  Karas,	  R.	  H.	  (2007).	  Assessment	  of	  peripheral	  vascular	  endothelial	  function	  in	  the	  ambulatory	  setting.	  Vascular	  Medicine	  (London,	  England),	  12(1),	  13–16.	  Kuvin,	  J.	  T.,	  Patel,	  A.	  R.,	  Sliney,	  K.	  A.,	  Pandian,	  N.	  G.,	  Sheffy,	  J.,	  Schnall,	  R.	  P.,	  …	  Udelson,	  J.	  E.	  (2003).	  Assessment	  of	  peripheral	  vascular	  endothelial	  function	  with	  finger	  arterial	  pulse	  wave	  amplitude.	  American	  Heart	  Journal,	  146(1),	  168–174.	  doi:10.1016/S0002-­‐8703(03)00094-­‐2	  Laden,	  F.,	  Neas,	  L.	  M.,	  Dockery,	  D.	  W.,	  &	  Schwartz,	  J.	  (2000).	  Association	  of	  fine	  particulate	  matter	  from	  different	  sources	  with	  daily	  mortality	  in	  six	  U.S.	  cities.	  Environmental	  Health	  Perspectives,	  108(10),	  941–947.	  Larson,	  T.,	  Su,	  J.,	  Baribeau,	  A.-­‐M.,	  Buzzelli,	  M.,	  Setton,	  E.,	  &	  Brauer,	  M.	  (2007).	  A	  spatial	  model	  of	  urban	  winter	  woodsmoke	  concentrations.	  Environmental	  Science	  &	  Technology,	  41(7),	  2429–2436.	  Lim,	  S.	  S.,	  Vos,	  T.,	  Flaxman,	  A.	  D.,	  Danaei,	  G.,	  Shibuya,	  K.,	  Adair-­‐Rohani,	  H.,	  …	  Ezzati,	  M.	  (2012).	  A	  comparative	  risk	  assessment	  of	  burden	  of	  disease	  and	  injury	  attributable	  to	  67	  risk	  factors	  and	  risk	  factor	  clusters	  in	  21	  regions,	  1990–2010:	  a	  systematic	  analysis	  for	  the	  Global	  Burden	  of	  Disease	  Study	  2010.	  The	  Lancet,	  380(9859),	  2224–2260.	  doi:10.1016/S0140-­‐6736(12)61766-­‐8	  	   94	  Lin,	  L.,	  Lee,	  M.	  L.,	  &	  Eatough,	  D.	  J.	  (2010).	  Review	  of	  Recent	  Advances	  in	  Detection	  of	  Organic	  Markers	  in	  Fine	  Particulate	  Matter	  and	  Their	  Use	  for	  Source	  Apportionment.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association,	  60(1),	  3–25.	  doi:10.3155/1047-­‐3289.60.1.3	  Liu,	  J.,	  Wang,	  J.,	  Jin,	  Y.,	  Roethig,	  H.	  J.,	  &	  Unverdorben,	  M.	  (2009).	  Variability	  of	  peripheral	  arterial	  tonometry	  in	  the	  measurement	  of	  endothelial	  function	  in	  healthy	  men.	  Clinical	  Cardiology,	  32(12),	  700–704.	  doi:10.1002/clc.20668	  Löndahl,	  J.,	  Massling,	  A.,	  Swietlicki,	  E.,	  Bräuner,	  E.	  V.,	  Ketzel,	  M.,	  Pagels,	  J.,	  &	  Loft,	  S.	  (2009).	  Experimentally	  Determined	  Human	  Respiratory	  Tract	  Deposition	  of	  Airborne	  Particles	  at	  a	  Busy	  Street.	  Environ.	  Sci.	  Technol.,	  43(13),	  4659–4664.	  doi:10.1021/es803029b	  Löndahl,	  J.,	  Pagels,	  J.,	  Boman,	  C.,	  Swietlicki,	  E.,	  Massling,	  A.,	  Rissler,	  J.,	  …	  Sandström,	  T.	  (2008).	  Deposition	  of	  biomass	  combustion	  aerosol	  particles	  in	  the	  human	  respiratory	  tract.	  Inhalation	  Toxicology,	  20(10),	  923–933.	  doi:10.1080/08958370802087124	  Loomis,	  D.,	  Grosse,	  Y.,	  Lauby-­‐Secretan,	  B.,	  Ghissassi,	  F.	  E.,	  Bouvard,	  V.,	  Benbrahim-­‐Tallaa,	  L.,	  …	  Straif,	  K.	  (2013).	  The	  carcinogenicity	  of	  outdoor	  air	  pollution.	  The	  Lancet	  Oncology,	  14(13),	  1262–1263.	  doi:10.1016/S1470-­‐2045(13)70487-­‐X	  Luc,	  G.,	  Bard,	  J.-­‐M.,	  Juhan-­‐Vague,	  I.,	  Ferrieres,	  J.,	  Evans,	  A.,	  Amouyel,	  P.,	  …	  PRIME	  Study	  Group.	  (2003).	  C-­‐reactive	  protein,	  interleukin-­‐6,	  and	  fibrinogen	  as	  predictors	  of	  coronary	  heart	  disease:	  the	  PRIME	  Study.	  Arteriosclerosis,	  Thrombosis,	  and	  Vascular	  Biology,	  23(7),	  1255–1261.	  doi:10.1161/01.ATV.0000079512.66448.1D	  	   95	  MacIntyre,	  E.	  A.,	  Karr,	  C.	  J.,	  Koehoorn,	  M.,	  Demers,	  P.	  A.,	  Tamburic,	  L.,	  Lencar,	  C.,	  &	  Brauer,	  M.	  (2011).	  Residential	  air	  pollution	  and	  otitis	  media	  during	  the	  first	  two	  years	  of	  life.	  Epidemiology	  (Cambridge,	  Mass.),	  22(1),	  81–89.	  doi:10.1097/EDE.0b013e3181fdb60f	  MacNee,	  W.,	  Li,	  X.	  Y.,	  Gilmour,	  P.	  S.,	  &	  Donaldson,	  K.	  (1997).	  Pro-­‐Inflammatory	  Effect	  of	  Particulate	  Air	  Pollution	  (PM10)	  in	  vivo	  and	  in	  vitro.	  Annals	  of	  Occupational	  Hygiene,	  41(inhaled	  particles	  VIII),	  7–13.	  doi:10.1093/annhyg/41.inhaled_particles_VIII.7	  Mateen	  FJ,	  &	  Brook	  RD.	  (2011).	  AIr	  pollution	  as	  an	  emerging	  global	  risk	  factor	  for	  stroke.	  JAMA,	  305(12),	  1240–1241.	  doi:10.1001/jama.2011.352	  Maykut,	  N.	  N.,	  Lewtas,	  J.,	  Kim,	  E.,	  &	  Larson,	  T.	  V.	  (2003).	  Source	  apportionment	  of	  PM2.5	  at	  an	  urban	  IMPROVE	  site	  in	  Seattle,	  Washington.	  Environmental	  Science	  &	  Technology,	  37(22),	  5135–5142.	  McGowan,	  J.	  a.,	  Hider,	  P.	  n.,	  Chacko,	  E.,	  &	  Town,	  G.	  i.	  (2002).	  Particulate	  air	  pollution	  and	  hospital	  admissions	  in	  Christchurch,	  New	  Zealand.	  Australian	  and	  New	  Zealand	  Journal	  of	  Public	  Health,	  26(1),	  23–29.	  doi:10.1111/j.1467-­‐842X.2002.tb00266.x	  Meng,	  Z.	  Y.,	  Jiang,	  X.	  M.,	  Yan,	  P.,	  Lin,	  W.	  L.,	  Zhang,	  H.	  D.,	  &	  Wang,	  Y.	  (2007).	  Characteristics	  and	  sources	  of	  PM2.5	  and	  carbonaceous	  species	  during	  winter	  in	  Taiyuan,	  China.	  Atmospheric	  Environment,	  41(32),	  6901–6908.	  doi:10.1016/j.atmosenv.2007.07.049	  Metro	  Vancouver.	  (2010).	  Metro	  Vancouver	  Emission	  Inventories	  and	  Forecasts.	  Retrieved	  February	  14,	  2014,	  from	  	   96	  Metro	  Vancouver.	  (2013).	  PM	  Fact	  Sheet.	  Ministry	  of	  Health.	  (1954).	  Mortality	  and	  Morbidity	  during	  the	  London	  Fog	  of	  December	  1952.	  Report	  by	  a	  Committee,	  etc	  (Ministry	  of	  Health.	  Reports	  on	  Public	  Health	  and	  Medical	  Subjects.	  no.	  95.).	  Naeher,	  L.	  P.,	  Brauer,	  M.,	  Lipsett,	  M.,	  Zelikoff,	  J.	  T.,	  Simpson,	  C.	  D.,	  Koenig,	  J.	  Q.,	  &	  Smith,	  K.	  R.	  (2007).	  Woodsmoke	  health	  effects:	  a	  review.	  Inhalation	  Toxicology,	  19(1),	  67–106.	  doi:10.1080/08958370600985875	  Nemery,	  B.,	  Hoet,	  P.	  H.,	  &	  Nemmar,	  A.	  (2001).	  The	  Meuse	  Valley	  fog	  of	  1930:	  an	  air	  pollution	  disaster.	  The	  Lancet,	  357(9257),	  704–708.	  doi:10.1016/S0140-­‐6736(00)04135-­‐0	  Nichols,	  J.	  L.,	  Owens,	  E.	  O.,	  Dutton,	  S.	  J.,	  &	  Luben,	  T.	  J.	  (2013).	  Systematic	  review	  of	  the	  effects	  of	  black	  carbon	  on	  cardiovascular	  disease	  among	  individuals	  with	  pre-­‐existing	  disease.	  International	  Journal	  of	  Public	  Health,	  58(5),	  707–724.	  doi:10.1007/s00038-­‐013-­‐0492-­‐z	  Olson,	  D.	  A.,	  &	  McDow,	  S.	  R.	  (2009).	  Near	  roadway	  concentrations	  of	  organic	  source	  markers.	  Atmospheric	  Environment,	  43(18),	  2862–2867.	  doi:10.1016/j.atmosenv.2009.03.016	  Onkelinx,	  S.,	  Cornelissen,	  V.,	  Goetschalckx,	  K.,	  Thomaes,	  T.,	  Verhamme,	  P.,	  &	  Vanhees,	  L.	  (2012).	  Reproducibility	  of	  different	  methods	  to	  measure	  the	  endothelial	  function.	  Vascular	  Medicine	  (London,	  England),	  17(2),	  79–84.	  doi:10.1177/1358863X12436708	  	   97	  Pai,	  J.	  K.,	  Pischon,	  T.,	  Ma,	  J.,	  Manson,	  J.	  E.,	  Hankinson,	  S.	  E.,	  Joshipura,	  K.,	  …	  Rimm,	  E.	  B.	  (2004).	  Inflammatory	  markers	  and	  the	  risk	  of	  coronary	  heart	  disease	  in	  men	  and	  women.	  The	  New	  England	  Journal	  of	  Medicine,	  351(25),	  2599–2610.	  doi:10.1056/NEJMoa040967	  Pope,	  C.	  .,	  Ezzati,	  M.,	  &	  Dockery,	  D.	  W.	  (2009).	  Fine-­‐particulate	  air	  pollution	  and	  life	  expectancy	  in	  the	  United	  States.	  The	  New	  England	  Journal	  of	  Medicine,	  360(4),	  376–386.	  doi:10.1056/NEJMsa0805646	  Pope,	  C.	  .,	  Hansen,	  J.	  C.,	  Kuprov,	  R.,	  Sanders,	  M.	  D.,	  Anderson,	  M.	  N.,	  &	  Eatough,	  D.	  J.	  (2011).	  Vascular	  Function	  and	  Short-­‐Term	  Exposure	  to	  Fine	  Particulate	  Air	  Pollution.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association,	  61(8),	  858–863.	  doi:10.3155/1047-­‐3289.61.8.858	  Pope,	  C.	  A.,	  3rd,	  Burnett,	  R.	  T.,	  Thurston,	  G.	  D.,	  Thun,	  M.	  J.,	  Calle,	  E.	  E.,	  Krewski,	  D.,	  &	  Godleski,	  J.	  J.	  (2004).	  Cardiovascular	  mortality	  and	  long-­‐term	  exposure	  to	  particulate	  air	  pollution:	  epidemiological	  evidence	  of	  general	  pathophysiological	  pathways	  of	  disease.	  Circulation,	  109(1),	  71–77.	  doi:10.1161/01.CIR.0000108927.80044.7F	  Pope,	  C.	  A.,	  3rd,	  Thun,	  M.	  J.,	  Namboodiri,	  M.	  M.,	  Dockery,	  D.	  W.,	  Evans,	  J.	  S.,	  Speizer,	  F.	  E.,	  &	  Heath,	  C.	  W.,	  Jr.	  (1995).	  Particulate	  air	  pollution	  as	  a	  predictor	  of	  mortality	  in	  a	  prospective	  study	  of	  U.S.	  adults.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  151(3	  Pt	  1),	  669–674.	  doi:10.1164/ajrccm/151.3_Pt_1.669	  	   98	  Pope,	  &	  Dockery,	  D.	  W.	  (2006).	  Health	  effects	  of	  fine	  particulate	  air	  pollution:	  lines	  that	  connect.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association	  (1995),	  56(6),	  709–742.	  Raaschou-­‐Nielsen,	  O.,	  Andersen,	  Z.	  J.,	  Hvidberg,	  M.,	  Jensen,	  S.	  S.,	  Ketzel,	  M.,	  Sørensen,	  M.,	  …	  Tjønneland,	  A.	  (2011).	  Lung	  cancer	  incidence	  and	  long-­‐term	  exposure	  to	  air	  pollution	  from	  traffic.	  Environmental	  Health	  Perspectives,	  119(6),	  860–865.	  doi:10.1289/ehp.1002353	  Raaschou-­‐Nielsen,	  O.,	  Bak,	  H.,	  Sørensen,	  M.,	  Jensen,	  S.	  S.,	  Ketzel,	  M.,	  Hvidberg,	  M.,	  …	  Loft,	  S.	  (2010).	  Air	  pollution	  from	  traffic	  and	  risk	  for	  lung	  cancer	  in	  three	  Danish	  cohorts.	  Cancer	  Epidemiology,	  Biomarkers	  &	  Prevention:	  A	  Publication	  of	  the	  American	  Association	  for	  Cancer	  Research,	  Cosponsored	  by	  the	  American	  Society	  of	  Preventive	  Oncology,	  19(5),	  1284–1291.	  doi:10.1158/1055-­‐9965.EPI-­‐10-­‐0036	  Restrepo,	  C.,	  Zimmerman,	  R.,	  Thurston,	  G.,	  Clemente,	  J.,	  Gorczynski,	  J.,	  Zhong,	  M.,	  …	  Chi	  Chen,	  L.	  (2004).	  A	  comparison	  of	  ground-­‐level	  air	  quality	  data	  with	  New	  York	  State	  Department	  of	  Environmental	  Conservation	  monitoring	  stations	  data	  in	  South	  Bronx,	  New	  York.	  Atmospheric	  Environment,	  38(31),	  5295–5304.	  doi:10.1016/j.atmosenv.2004.06.004	  Ridker,	  P.	  M.	  (2007).	  C-­‐reactive	  protein	  and	  the	  prediction	  of	  cardiovascular	  events	  among	  those	  at	  intermediate	  risk:	  moving	  an	  inflammatory	  hypothesis	  toward	  consensus.	  Journal	  of	  the	  American	  College	  of	  Cardiology,	  49(21),	  2129–2138.	  doi:10.1016/j.jacc.2007.02.052	  	   99	  Riediker,	  M.,	  Cascio,	  W.	  E.,	  Griggs,	  T.	  R.,	  Herbst,	  M.	  C.,	  Bromberg,	  P.	  A.,	  Neas,	  L.,	  …	  Devlin,	  R.	  B.	  (2004).	  Particulate	  matter	  exposure	  in	  cars	  is	  associated	  with	  cardiovascular	  effects	  in	  healthy	  young	  men.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  169(8),	  934–940.	  doi:10.1164/rccm.200310-­‐1463OC	  Rogge,	  W.	  F.,	  Hildemann,	  L.	  M.,	  Mazurek,	  M.	  A.,	  Cass,	  G.	  R.,	  &	  Simoneit,	  B.	  R.	  T.	  (1993).	  Sources	  of	  fine	  organic	  aerosol.	  2.	  Noncatalyst	  and	  catalyst-­‐equipped	  automobiles	  and	  heavy-­‐duty	  diesel	  trucks.	  Environmental	  Science	  &	  Technology,	  27(4),	  636–651.	  doi:10.1021/es00041a007	  Rubinshtein,	  R.,	  Kuvin,	  J.	  T.,	  Soffler,	  M.,	  Lennon,	  R.	  J.,	  Lavi,	  S.,	  Nelson,	  R.	  E.,	  …	  Lerman,	  A.	  (2010).	  Assessment	  of	  endothelial	  function	  by	  non-­‐invasive	  peripheral	  arterial	  tonometry	  predicts	  late	  cardiovascular	  adverse	  events.	  European	  Heart	  Journal,	  31(9),	  1142–1148.	  doi:10.1093/eurheartj/ehq010	  Rückerl,	  R.,	  Schneider,	  A.,	  Breitner,	  S.,	  Cyrys,	  J.,	  &	  Peters,	  A.	  (2011).	  Health	  effects	  of	  particulate	  air	  pollution:	  A	  review	  of	  epidemiological	  evidence.	  Inhalation	  Toxicology,	  23(10),	  555–592.	  doi:10.3109/08958378.2011.593587	  Rudez,	  G.,	  Janssen,	  N.	  A.	  H.,	  Kilinc,	  E.,	  Leebeek,	  F.	  W.	  G.,	  Gerlofs-­‐Nijland,	  M.	  E.,	  Spronk,	  H.	  M.	  H.,	  …	  de	  Maat,	  M.	  P.	  M.	  (2009).	  Effects	  of	  Ambient	  Air	  Pollution	  on	  Hemostasis	  and	  Inflammation.	  Environmental	  Health	  Perspectives,	  117(6),	  995–1001.	  doi:10.1289/ehp.0800437	  Rundell,	  K.	  W.,	  Hoffman,	  J.	  R.,	  Caviston,	  R.,	  Bulbulian,	  R.,	  &	  Hollenbach,	  A.	  M.	  (2007).	  Inhalation	  of	  ultrafine	  and	  fine	  particulate	  matter	  disrupts	  systemic	  vascular	  	   100	  function.	  Inhalation	  Toxicology,	  19(2),	  133–140.	  doi:10.1080/08958370601051727	  Saarikoski,	  S.	  K.,	  Sillanpää,	  M.	  K.,	  Saarnio,	  K.	  M.,	  Hillamo,	  R.	  E.,	  Pennanen,	  A.	  S.,	  &	  Salonen,	  R.	  O.	  (2008).	  Impact	  of	  Biomass	  Combustion	  on	  Urban	  Fine	  Particulate	  Matter	  in	  Central	  and	  Northern	  Europe.	  Water,	  Air,	  and	  Soil	  Pollution,	  191(1-­‐4),	  265–277.	  doi:10.1007/s11270-­‐008-­‐9623-­‐1	  Sacks,	  J.	  D.,	  Stanek,	  L.	  W.,	  Luben,	  T.	  J.,	  Johns,	  D.	  O.,	  Buckley,	  B.	  J.,	  Brown,	  J.	  S.,	  &	  Ross,	  M.	  (2011).	  Particulate	  matter-­‐induced	  health	  effects:	  who	  is	  susceptible?	  Environmental	  Health	  Perspectives,	  119(4),	  446–454.	  doi:10.1289/ehp.1002255	  Sakai,	  M.,	  Sato,	  Y.,	  Sato,	  S.,	  Ihara,	  S.,	  Onizuka,	  M.,	  Sakakibara,	  Y.,	  &	  Takahashi,	  H.	  (2004).	  Effect	  of	  relocating	  to	  areas	  of	  reduced	  atmospheric	  particulate	  matter	  levels	  on	  the	  human	  circulating	  leukocyte	  count.	  Journal	  of	  Applied	  Physiology	  (Bethesda,	  Md.:	  1985),	  97(5),	  1774–1780.	  doi:10.1152/japplphysiol.00024.2004	  Salvi,	  &	  Holgate.	  (2001).	  Mechanisms	  of	  particulate	  matter	  toxicity.	  Clinical	  &	  Experimental	  Allergy,	  29(9),	  1187–1194.	  doi:10.1046/j.1365-­‐2222.1999.00576.x	  Sanhueza,	  P.	  A.,	  Torreblanca,	  M.	  A.,	  Diaz-­‐Robles,	  L.	  A.,	  Schiappacasse,	  L.	  N.,	  Silva,	  M.	  P.,	  &	  Astete,	  T.	  D.	  (2009).	  Particulate	  air	  pollution	  and	  health	  effects	  for	  cardiovascular	  and	  respiratory	  causes	  in	  Temuco,	  Chile:	  a	  wood-­‐smoke-­‐polluted	  urban	  area.	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association	  (1995),	  59(12),	  1481–1488.	  	   101	  Saraswat,	  A.,	  Apte,	  J.	  S.,	  Kandlikar,	  M.,	  Brauer,	  M.,	  Henderson,	  S.	  B.,	  &	  Marshall,	  J.	  D.	  (2013).	  Spatiotemporal	  land	  use	  regression	  models	  of	  fine,	  ultrafine,	  and	  black	  carbon	  particulate	  matter	  in	  New	  Delhi,	  India.	  Environmental	  Science	  &	  Technology,	  47(22),	  12903–12911.	  doi:10.1021/es401489h	  Schnabel,	  R.	  B.,	  Schulz,	  A.,	  Wild,	  P.	  S.,	  Sinning,	  C.	  R.,	  Wilde,	  S.,	  Eleftheriadis,	  M.,	  …	  Münzel,	  T.	  (2011).	  Noninvasive	  Vascular	  Function	  Measurement	  in	  the	  Community	  Cross-­‐Sectional	  Relations	  and	  Comparison	  of	  Methods.	  Circulation:	  Cardiovascular	  Imaging,	  4(4),	  371–380.	  doi:10.1161/CIRCIMAGING.110.961557	  Schwartz,	  J.,	  Dockery,	  D.	  W.,	  &	  Neas,	  L.	  M.	  (1996).	  Is	  daily	  mortality	  associated	  specifically	  with	  fine	  particles?	  Journal	  of	  the	  Air	  &	  Waste	  Management	  Association	  (1995),	  46(10),	  927–939.	  Schwartz,	  J.,	  Slater,	  D.,	  Larson,	  T.	  V.,	  Pierson,	  W.	  E.,	  &	  Koenig,	  J.	  Q.	  (1993).	  Particulate	  air	  pollution	  and	  hospital	  emergency	  room	  visits	  for	  asthma	  in	  Seattle.	  The	  American	  Review	  of	  Respiratory	  Disease,	  147(4),	  826–831.	  doi:10.1164/ajrccm/147.4.826	  Schwarze,	  P.	  E.,	  Ovrevik,	  J.,	  Låg,	  M.,	  Refsnes,	  M.,	  Nafstad,	  P.,	  Hetland,	  R.	  B.,	  &	  Dybing,	  E.	  (2006).	  Particulate	  matter	  properties	  and	  health	  effects:	  consistency	  of	  epidemiological	  and	  toxicological	  studies.	  Human	  &	  Experimental	  Toxicology,	  25(10),	  559–579.	  Sheppard,	  L.,	  Levy,	  D.,	  Norris,	  G.,	  Larson,	  T.	  V.,	  &	  Koenig,	  J.	  Q.	  (1999).	  Effects	  of	  ambient	  air	  pollution	  on	  nonelderly	  asthma	  hospital	  admissions	  in	  Seattle,	  Washington,	  1987-­‐1994.	  Epidemiology	  (Cambridge,	  Mass.),	  10(1),	  23–30.	  	   102	  Simoneit,	  B.	  (1999).	  A	  review	  of	  biomarker	  compounds	  as	  source	  indicators	  and	  tracers	  for	  air	  pollution.	  Environmental	  Science	  and	  Pollution	  Research,	  6(3),	  159–169.	  doi:10.1007/BF02987621	  Song,	  Y.,	  Tang,	  X.,	  Xie,	  S.,	  Zhang,	  Y.,	  Wei,	  Y.,	  Zhang,	  M.,	  …	  Lu,	  S.	  (2007).	  Source	  apportionment	  of	  PM2.5	  in	  Beijing	  in	  2004.	  Journal	  of	  Hazardous	  Materials,	  146(1-­‐2),	  124–130.	  doi:10.1016/j.jhazmat.2006.11.058	  Stockfelt,	  L.,	  Sallsten,	  G.,	  Almerud,	  P.,	  Basu,	  S.,	  &	  Barregard,	  L.	  (2013).	  Short-­‐term	  chamber	  exposure	  to	  low	  doses	  of	  two	  kinds	  of	  wood	  smoke	  does	  not	  induce	  systemic	  inflammation,	  coagulation	  or	  oxidative	  stress	  in	  healthy	  humans.	  Inhalation	  Toxicology,	  25(8),	  417–425.	  doi:10.3109/08958378.2013.798387	  Strak,	  M.,	  Janssen,	  N.	  A.	  H.,	  Godri,	  K.	  J.,	  Gosens,	  I.,	  Mudway,	  I.	  S.,	  Cassee,	  F.	  R.,	  …	  Hoek,	  G.	  (2012).	  Respiratory	  health	  effects	  of	  airborne	  particulate	  matter:	  the	  role	  of	  particle	  size,	  composition,	  and	  oxidative	  potential-­‐the	  RAPTES	  project.	  Environmental	  Health	  Perspectives,	  120(8),	  1183–1189.	  doi:10.1289/ehp.1104389	  Sublett,	  J.	  L.,	  Seltzer,	  J.,	  Burkhead,	  R.,	  Williams,	  P.	  B.,	  Wedner,	  H.	  J.,	  &	  Phipatanakul,	  W.	  (2010).	  Air	  filters	  and	  air	  cleaners:	  rostrum	  by	  the	  American	  Academy	  of	  Allergy,	  Asthma	  &	  Immunology	  Indoor	  Allergen	  Committee.	  The	  Journal	  of	  Allergy	  and	  Clinical	  Immunology,	  125(1),	  32–38.	  doi:10.1016/j.jaci.2009.08.036	  Svartengren,	  M.,	  Linnman,	  L.,	  Philipson,	  K.,	  &	  Camner,	  P.	  (1987).	  Regional	  Deposition	  in	  Human	  Lung	  of	  2.5	  μM	  Particles,	  Experimental	  Lung	  Research,	  Informa	  	   103	  Healthcare.	  Retrieved	  June	  3,	  2012,	  from	  Swiston,	  J.	  R.,	  Davidson,	  W.,	  Attridge,	  S.,	  Li,	  G.	  T.,	  Brauer,	  M.,	  &	  van	  Eeden,	  S.	  F.	  (2008).	  Wood	  smoke	  exposure	  induces	  a	  pulmonary	  and	  systemic	  inflammatory	  response	  in	  firefighters.	  European	  Respiratory	  Journal,	  32(1),	  129–138.	  doi:10.1183/09031936.00097707	  Tan,	  W.	  C.,	  Qiu,	  D.,	  Liam,	  B.	  L.,	  Ng,	  T.	  P.,	  Lee,	  S.	  H.,	  van	  Eeden,	  S.	  F.,	  …	  Hogg,	  J.	  C.	  (2000).	  The	  human	  bone	  marrow	  response	  to	  acute	  air	  pollution	  caused	  by	  forest	  fires.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  161(4	  Pt	  1),	  1213–1217.	  doi:10.1164/ajrccm.161.4.9904084	  Tarlo,	  S.	  M.,	  Broder,	  I.,	  Corey,	  P.,	  Chan-­‐Yeung,	  M.,	  Ferguson,	  A.,	  Becker,	  A.,	  …	  Manfreda,	  J.	  (2001).	  The	  role	  of	  symptomatic	  colds	  in	  asthma	  exacerbations:	  Influence	  of	  outdoor	  allergens	  and	  air	  pollutants.	  The	  Journal	  of	  Allergy	  and	  Clinical	  Immunology,	  108(1),	  52–58.	  doi:10.1067/mai.2001.116574	  Tonne,	  C.,	  Beevers,	  S.,	  Armstrong,	  B.,	  Kelly,	  F.,	  &	  Wilkinson,	  P.	  (2008).	  Air	  pollution	  and	  mortality	  benefits	  of	  the	  London	  Congestion	  Charge:	  spatial	  and	  socioeconomic	  inequalities.	  Occupational	  and	  Environmental	  Medicine,	  65(9),	  620–627.	  doi:10.1136/oem.2007.036533	  Törnqvist,	  H.,	  Mills,	  N.	  L.,	  Gonzalez,	  M.,	  Miller,	  M.	  R.,	  Robinson,	  S.	  D.,	  Megson,	  I.	  L.,	  …	  Blomberg,	  A.	  (2007).	  Persistent	  endothelial	  dysfunction	  in	  humans	  after	  diesel	  exhaust	  inhalation.	  American	  Journal	  of	  Respiratory	  and	  Critical	  Care	  Medicine,	  176(4),	  395–400.	  doi:10.1164/rccm.200606-­‐872OC	  	   104	  Trenga,	  C.	  A.,	  Sullivan,	  J.	  H.,	  Schildcrout,	  J.	  S.,	  Shepherd,	  K.	  P.,	  Shapiro,	  G.	  G.,	  Liu,	  L.-­‐J.	  S.,	  …	  Koenig,	  J.	  Q.	  (2006).	  Effect	  of	  particulate	  air	  pollution	  on	  lung	  function	  in	  adult	  and	  pediatric	  subjects	  in	  a	  Seattle	  panel	  study.	  Chest,	  129(6),	  1614–1622.	  doi:10.1378/chest.129.6.1614	  Tsai,	  S.-­‐S.,	  Goggins,	  W.	  B.,	  Chiu,	  H.-­‐F.,	  &	  Yang,	  C.-­‐Y.	  (2003).	  Evidence	  for	  an	  association	  between	  air	  pollution	  and	  daily	  stroke	  admissions	  in	  Kaohsiung,	  Taiwan.	  Stroke;	  a	  Journal	  of	  Cerebral	  Circulation,	  34(11),	  2612–2616.	  doi:10.1161/01.STR.0000095564.33543.64	  Valavanidis,	  A.,	  Vlahoyianni,	  T.,	  &	  Fiotakis,	  K.	  (2005).	  Comparative	  study	  of	  the	  formation	  of	  oxidative	  damage	  marker	  8-­‐hydroxy-­‐2’-­‐deoxyguanosine	  (8-­‐OHdG)	  adduct	  from	  the	  nucleoside	  2’-­‐deoxyguanosine	  by	  transition	  metals	  and	  suspensions	  of	  particulate	  matter	  in	  relation	  to	  metal	  content	  and	  redox	  reactivity.	  Free	  Radical	  Research,	  39(10),	  1071–1081.	  doi:10.1080/10715760500188671	  Van	  Eeden,	  S.	  F.,	  Yeung,	  A.,	  Quinlam,	  K.,	  &	  Hogg,	  J.	  C.	  (2005).	  Systemic	  response	  to	  ambient	  particulate	  matter:	  relevance	  to	  chronic	  obstructive	  pulmonary	  disease.	  Proceedings	  of	  the	  American	  Thoracic	  Society,	  2(1),	  61–67.	  doi:10.1513/pats.200406-­‐035MS	  Van	  Erp,	  A.	  M.	  M.,	  O’Keefe,	  R.,	  Cohen,	  A.	  J.,	  &	  Warren,	  J.	  (2008).	  Evaluating	  the	  effectiveness	  of	  air	  quality	  interventions.	  Journal	  of	  Toxicology	  and	  Environmental	  Health.	  Part	  A,	  71(9-­‐10),	  583–587.	  doi:10.1080/15287390801997708	  	   105	  Verma,	  V.,	  Polidori,	  A.,	  Schauer,	  J.	  J.,	  Shafer,	  M.	  M.,	  Cassee,	  F.	  R.,	  &	  Sioutas,	  C.	  (2009).	  Physicochemical	  and	  toxicological	  profiles	  of	  particulate	  matter	  in	  Los	  Angeles	  during	  the	  October	  2007	  southern	  California	  wildfires.	  Environmental	  Science	  &	  Technology,	  43(3),	  954–960.	  Villena,	  G.,	  Kleffmann,	  J.,	  Kurtenbach,	  R.,	  Wiesen,	  P.,	  Lissi,	  E.,	  Rubio,	  M.	  A.,	  …	  Rappenglück,	  B.	  (2011).	  Vertical	  gradients	  of	  HONO,	  NOx	  and	  O3	  in	  Santiago	  de	  Chile.	  Atmospheric	  Environment,	  45(23),	  3867–3873.	  doi:10.1016/j.atmosenv.2011.01.073	  Von	  Klot,	  S.,	  Cyrys,	  J.,	  Hoek,	  G.,	  Kühnel,	  B.,	  Pitz,	  M.,	  Kuhn,	  U.,	  …	  Peters,	  A.	  (2011).	  Estimated	  personal	  soot	  exposure	  is	  associated	  with	  acute	  myocardial	  infarction	  onset	  in	  a	  case-­‐crossover	  study.	  Progress	  in	  Cardiovascular	  Diseases,	  53(5),	  361–368.	  doi:10.1016/j.pcad.2011.01.002	  Wang,	  H.,	  Zhuang,	  Y.,	  Wang,	  Y.,	  Sun,	  Y.,	  Yuan,	  H.,	  Zhuang,	  G.,	  &	  Hao,	  Z.	  (2008).	  Long-­‐term	  monitoring	  and	  source	  apportionment	  of	  PM2.5/PM10	  in	  Beijing,	  China.	  Journal	  of	  Environmental	  Sciences	  (China),	  20(11),	  1323–1327.	  Wang,	  X.,	  Bi,	  X.,	  Sheng,	  G.,	  &	  Fu,	  J.	  (2006).	  Chemical	  Composition	  and	  Sources	  of	  PM10	  and	  PM2.5	  Aerosols	  in	  Guangzhou,	  China.	  Environmental	  Monitoring	  and	  Assessment,	  119(1-­‐3),	  425–439.	  doi:10.1007/s10661-­‐005-­‐9034-­‐3	  Wang,	  X.,	  Chen,	  R.,	  Meng,	  X.,	  Geng,	  F.,	  Wang,	  C.,	  &	  Kan,	  H.	  (2013).	  Associations	  between	  fine	  particle,	  coarse	  particle,	  black	  carbon	  and	  hospital	  visits	  in	  a	  Chinese	  city.	  The	  Science	  of	  the	  Total	  Environment,	  458-­‐460,	  1–6.	  doi:10.1016/j.scitotenv.2013.04.008	  	   106	  Watts,	  J.	  (2005).	  China:	  the	  air	  pollution	  capital	  of	  the	  world.	  Lancet,	  366(9499),	  1761–1762.	  Weichenthal,	  S.,	  Mallach,	  G.,	  Kulka,	  R.,	  Black,	  A.,	  Wheeler,	  A.,	  You,	  H.,	  …	  Sharp,	  D.	  (2013).	  A	  randomized	  double-­‐blind	  crossover	  study	  of	  indoor	  air	  filtration	  and	  acute	  changes	  in	  cardiorespiratory	  health	  in	  a	  First	  Nations	  community.	  Indoor	  Air,	  23(3),	  175–184.	  doi:10.1111/ina.12019	  Wu,	  C.,	  Larson,	  T.	  V.,	  Wu,	  S.-­‐Y.,	  Williamson,	  J.,	  Westberg,	  H.	  H.,	  &	  Liu,	  L.-­‐J.	  S.	  (2007).	  Source	  apportionment	  of	  PM(2.5)	  and	  selected	  hazardous	  air	  pollutants	  in	  Seattle.	  The	  Science	  of	  the	  Total	  Environment,	  386(1-­‐3),	  42–52.	  doi:10.1016/j.scitotenv.2007.07.042	  Yamada,	  Y.,	  Miyamoto,	  K.,	  Mori,	  T.,	  &	  Koizumi,	  A.	  (1984).	  Penetration	  of	  submicron	  aerosols	  through	  high-­‐efficiency	  air	  filters.	  Health	  Physics,	  46(3),	  543–547.	  Zezima,	  K.	  (2008,	  February	  19).	  With	  Oil	  Prices	  Rising,	  Wood	  Makes	  a	  Comeback.	  The	  New	  York	  Times.	  Retrieved	  from	  Zhao,	  J.,	  Gao,	  Z.,	  Tian,	  Z.,	  Xie,	  Y.,	  Xin,	  F.,	  Jiang,	  R.,	  …	  Song,	  W.	  (2013).	  The	  biological	  effects	  of	  individual-­‐level	  PM(2.5)	  exposure	  on	  systemic	  immunity	  and	  inflammatory	  response	  in	  traffic	  policemen.	  Occupational	  and	  Environmental	  Medicine,	  70(6),	  426–431.	  doi:10.1136/oemed-­‐2012-­‐100864	  Zheng,	  M.,	  Salmon,	  L.	  G.,	  Schauer,	  J.	  J.,	  Zeng,	  L.,	  &	  Kian.	  (2005).	  Seasonal	  trends	  in	  PM2.5	  source	  contributions	  in	  Beijing,	  China.	  Atmospheric	  Environment,	  39,	  3967–3976.	  	  	   107	  	  	  	  	  	  	  	  	   	  	  	  APPENDICES	  	  	  	  	  	  	  	  	  	  	  	  	  	   108	  	  	  	  	  	  	  	  	  	  	  APPENDIX	  I:	  INVITATION	  LETTERS	  	  	  	  	  	  	  	  	  	  	  	   107	  	  	  	  	  Cardiovascular	  Effects	  of	  Aerosols	  in	  Residences	  Study	  (CLEAR)	  	  Dear	  Resident:	  	  I	  am	  working	  with	  a	  team	  of	  researchers	  from	  Simon	  Fraser	  University	  (SFU)	  and	  the	  University	  of	  British	  Columbia	  (UBC)	  who	  are	  investigating	  the	  health	  effects	  of	  air	  pollution.	  I	  am	  writing	  to	  invite	  you	  to	  participate	  in	  a	  study	  currently	  underway.	  Adults	  (19	  years	  or	  older)	  residing	  in	  Greater	  Vancouver,	  BC	  may	  be	  eligible	  to	  participate	  in	  the	  study.	  Unfortunately,	  homes	  with	  residents	  who	  smoke	  are	  not	  able	  to	  be	  considered	  for	  participation.	  	  The	  study	  involves	  air	  monitoring	  inside	  and	  outside	  your	  home	  and	  the	  use	  of	  portable	  air	  cleaners	  to	  determine	  their	  effectiveness	  regarding	  improved	  indoor	  air	  quality	  and	  improved	  health.	  In	  order	  to	  measure	  health	  improvements	  we	  will	  be	  collecting	  blood	  samples	  (three	  times)	  and	  conducting	  a	  non-­‐invasive	  test	  of	  vascular	  function	  (twice).	  All	  air	  sampling	  will	  be	  conducted	  over	  2	  consecutive	  7-­‐day	  periods;	  health	  measurements	  will	  be	  done	  at	  the	  beginning	  of	  sampling,	  halfway	  through	  sampling	  and	  again	  at	  the	  end	  of	  the	  sampling	  period.	  At	  the	  initial	  visit,	  2	  technicians	  will	  visit	  your	  home.	  The	  environmental	  technician	  will	  setup	  the	  air	  monitoring	  equipment	  inside	  and	  outside	  your	  home	  and	  gather	  some	  information	  on	  your	  home	  (age,	  size,	  type	  of	  heating	  system,	  etc.);	  the	  health	  technician	  will	  complete	  a	  short	  questionnaire	  on	  your	  health	  and	  perform	  the	  first	  blood	  	   108	  draw.	  In	  addition	  to	  the	  air	  monitoring	  equipment,	  2	  portable	  high	  efficiency	  particulate	  air	  (HEPA)	  filters	  will	  be	  installed	  in	  your	  home,	  one	  in	  your	  bedroom	  and	  one	  in	  your	  main	  living	  room.	  These	  devices	  are	  quiet,	  require	  no	  maintenance,	  and	  will	  not	  affect	  your	  daily	  routine.	  	  After	  the	  initial	  visit,	  the	  air	  monitoring	  equipment	  will	  run	  unaided	  until	  the	  end	  of	  the	  first	  7-­‐day	  period	  when	  both	  technicians	  will	  visit	  your	  home	  again.	  At	  this	  second	  visit,	  the	  environmental	  technician	  will	  perform	  maintenance	  checks	  on	  the	  equipment,	  and	  the	  health	  technician	  will	  perform	  a	  second	  blood	  draw	  and	  also	  perform	  a	  non-­‐invasive	  test	  of	  your	  vascular	  function.	  The	  vascular	  function	  test	  involves	  having	  a	  blood-­‐pressure	  cuff	  on	  one	  arm,	  and	  a	  small	  sensor	  on	  two	  fingers.	  The	  test	  lasts	  about	  20	  minutes	  and	  should	  not	  cause	  any	  discomfort.	  We	  will	  also	  ask	  you	  questions	  about	  any	  health	  symptoms	  you	  may	  have	  experienced	  during	  the	  first	  week	  of	  sampling.	  	  Data	  collection	  is	  currently	  underway	  and	  will	  continue	  until	  approximately	  August,	  2012.	  Your	  commitment	  will	  be	  approximately	  4	  hours	  in	  total.	  Sampling	  will	  be	  conducted	  by	  fully	  trained	  study	  technicians	  and	  will	  be	  scheduled	  to	  start	  and	  end	  on	  a	  weekday	  morning.	  All	  3	  visits	  will	  be	  conducted	  in	  the	  morning,	  as	  it	  is	  important	  that	  the	  health	  measurements	  be	  performed	  before	  you’ve	  had	  breakfast	  or	  any	  caffeine.	  The	  technicians	  will	  work	  with	  you	  to	  find	  a	  time	  convenient	  for	  you	  to	  install	  the	  equipment	  and	  perform	  the	  health	  sampling.	  	  Finally,	  while	  the	  equipment	  is	  running,	  we	  ask	  that	  you	  keep	  a	  simple	  diary	  of	  time	  spent	  at	  home	  and	  of	  any	  possible	  pollution-­‐generating	  activities	  (cooking,	  vacuuming,	  etc.).	  A	  diary	  will	  be	  supplied	  for	  this	  recording.	   109	  The	  sampling	  instruments	  collect	  information	  only	  on	  air	  quality	  and	  run	  unaided	  but	  do	  require	  a	  small	  amount	  of	  power	  to	  run.	  Beyond	  the	  activities	  described	  above,	  your	  participation	  in	  this	  study	  should	  not	  interfere	  with	  your	  daily	  routine.	  All	  personal	  information	  collected	  will	  remain	  completely	  confidential	  and	  will	  be	  securely	  stored	  according	  to	  University	  policy.	  No	  one	  other	  than	  the	  study	  team	  members	  will	  have	  access	  to	  the	  data	  collected,	  and	  individuals	  are	  never	  named	  or	  identifiable	  in	  any	  reports	  or	  publications.	  	  In	  order	  to	  assist	  with	  the	  inconvenience	  of	  participating	  in	  this	  study,	  you	  will	  receive	  an	  honorarium.	  We	  are	  able	  to	  enroll	  1	  or	  2	  subjects	  per	  home,	  so	  if	  there	  is	  a	  second	  adult	  in	  your	  home	  who	  is	  interested	  and	  eligible	  to	  participate,	  they	  would	  also	  receive	  the	  honorarium.	  	  If	  you	  are	  interested	  in	  participating,	  or	  would	  like	  further	  details	  about	  the	  study,	  equipment	  used	  or	  any	  of	  the	  procedures,	  please	  call	  Barbara	  Karlen	  at	  778-­‐782-­‐9324	  or	  email	  Further	  information	  is	  also	  available	  on	  the	  study	  website:	  Alternately	  you	  can	  contact	  me	  at	  the	  number	  or	  e-­‐mail	  address	  listed	  below.	  	  Thank	  you	  for	  your	  time	  and	  for	  considering	  participating	  in	  this	  important	  research.	  Sincerely,	  	  	  	  	  	  Dr.	  Ryan	  Allen	  Assistant	  Professor	  Faculty	  of	  Health	  Sciences	  Simon	  Fraser	  University	  (778)	  782-­‐7631	  	   110	  	   	  	  	  	  	  	  	  	  	  	  APPENDIX	  II:	  PARTICIPANT	  SCREENING	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   111	  	  Participant	  screening	  (to	  be	  done	  by	  phone)	  	  	   Name:    Date:    	   First, give a description of the research study and what is involved in participating, specifically: 	   - early morning, weekday sampling session start and end; - 3 blood samples; - fasting before all 3 blood samples and vascular function measurement – first thing in morning; - instruments running indoors and outdoors for two consecutive 7-day sessions; - home and symptom questionnaires; - activity diary. 	  If	  potentially	  interested,	  ask	  the	  following	  eligibility	  screening	  questions:	  	   1. Home address (with postal code):   	   2. What type of building do you live in?    	   a. If apartment or condo, what floor do you live on (1st floor = ground level)?    	   3. Is there a secure outdoor location close to your home (such as a yard, patio, or balcony) where we could place our outdoor air pollution sampling equipment?   Yes No 	   a. If yes, does this location have an accessible electrical outlet?  Yes No 	   4. Do you smoke?  Yes No 	   5. Are there any smokers residing in your home?  Yes No 	   6. Have you been diagnosed by a physician/dentist with any of the following? Check if yes, and for those that apply, ask if they currently take any medication for the condition and which medication is taken. 	     COPD (chronic obstructive pulmonary disease/chronic bronchitis/emphysema) Medication:    	     asthma Medication:    	     diabetes Medication:    	     heart disease (including coronary heart disease, stroke, myocardial infarction (heart attack), heart failure, angina, arrhythmia) Medication:    	   112	  	     hypertension Medication:      arthritis Medication:    	     gum disease Medication:    	   7. Have you been diagnosed with any other current conditions? 	   Describe:    	   8. Have you had surgery in the last 6 months? Yes No Date:    	   9. Are you pregnant? Yes No 	   10. Do you currently have an infection of any kind (including colds)? Yes No Describe:    11. Do you currently take any other regular medications?   Yes No 	   Please list:    	   12. What is your current weight?    	   13. What is your height?    	   14. Have you traveled out of the region recently?  Dates (from – to) :   	   Where:   	   15. Do you work or regularly volunteer outside the home?   Yes No (IF NO SKIP TO 16) 	   a. If yes, how many hours per week do you typically work/volunteer outside the home?    	   b. what is your occupation / volunteer activity?    	   c. are you regularly exposed to dust, exhaust, or smoke at work/volunteer? Yes No 	   i. If yes, please describe:    	  	   113	  d. how do you usually commute to work / volunteer location?    	   e. how much time per day do you typically spend commuting to and from work/volunteer?   (enter total round trip commuting time) 	   16. Do you currently use a wood-burning appliance in your home? Yes No (IF NO SKIP TO 17) 	   a. If yes, what type of appliance do you have? i. Fireplace ii. Conventional wood stove iii. Insert iv. Certified wood stove v. Pellet stove 	   b. do you tend the fire? Yes No 	   c. approximate proportion of fire-tending that you do:    	  	   17. What is your year of birth?    	   18. Is there another adult in your household who may be interested in participating?  Yes No 	   Would you ask them whether they are interested in speaking to me about this study and whether they are available to speak to me now? 	   If they are not available now, would you please pass on the intro letter to them and ask them to phone me if they are interested in participating or would like more information. 	   do you tend the fire? Yes No 	   a. approximate proportion of fire-tending that you do:    	  	   19. What is your year of birth?    	   20. Is there another adult in your household who may be interested in participating?  Yes No 	   Would you ask them whether they are interested in speaking to me about this study and whether they are available to speak to me now? 	   If they are not available now, would you please pass on the intro letter to them and ask them to phone me if they are interested in participating or would like more information. 	   114	  	  	  Contact	  information:	  	   Phone number: (H)    (C)    	   Best time to contact:    	   Best to call:  home / cell  Email:    	  	   End the call with: 	   “Thank you for your interest and your time; we'll be reviewing your eligibility for the study and will be back in touch with you within 2 weeks to let you know whether we are able to enroll you in the study”. 	  	  	  	   OFFICE USE ONLY: 	   □	   	  	  Acceptable	   Not	  Acceptable	  □	   	  	  Priority	  1	  recruit	   or,	   	  Priority	  2	  recruit	  	  	  	  Call	  back	  date:	  	  	  	   	  	  Scheduled	  for:	  	  	  	   	  	  	  	   115	  	  	  	  	  	  	  	  	  	  	  	  APPENDIX	  III:	  INFORMED	  CONSENT	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   116	  	  	  	  	  	  	  	  	  	  	  	   SUBJECT INFORMATION AND CONSENT FORM 	  	   Cardiovascular Effects of Aerosols in Residences Study (CLEAR) 	  	  Principal Investigator: 	  Dr. Ryan Allen Faculty of Health Sciences Simon Fraser University Phone: (778) 782-7631 E-mail: 	  Co-Investigators: 	  Dr. Michael Brauer School of Population and Public Health The University of British Columbia 	  Dr. Christopher Carlsten School of Population and Public Health The University of British Columbia 	  	  Dr. Stephan van Eeden Dept. of Internal Medicine & Respirology University of British Columbia 	  	  The purpose of this study is to evaluate the health benefits of high efficiency particulate air (HEPA) filters and to compare the cardiovascular health effects of two major particulate matter sources (traffic-related air pollution & residential woodsmoke emissions). 	  This study will be conducted in the Vancouver and surrounding area commencing at the start of the 2011/12 winter heating season and will continue until approximately 100 people are enrolled. 	  Introduction 	  You are being invited to participate in this study because you live in a community in greater Vancouver, BC that has air pollution originating from traffic emissions and/or from residential woodsmoke. 	  Your participation is voluntary 	  Your participation in this study is entirely voluntary. If you decide to take part in this study, you are still free to withdraw at any time without giving any reasons for your decision. We ask that you please take time to read the following information before you decide. 	  Who is conducting the study? 	  	  	   117	  This study is being conducted by researchers at Simon Fraser University, and the University of British Columbia and is being funded by the Canadian Institutes of Health Research. 	  Background 	  Air pollution from a variety of sources can negatively affect health. HEPA filters are expected to significantly reduce indoor levels of airborne particles and to improve health. 	  The objectives of this study are to: 1) quantify the relationship between low-level exposures to combustion-derived PM air pollution and subclinical indicators of cardiovascular disease risk; 2) evaluate and compare the impact of two exposure reduction approaches – HEPA filters and a community  woodstove  exchange  program  (underway  in  the  Smithers  and  Telkwa  areas)  –  on reductions in exposure and measurable improvements in sensitive subclinical cardiovascular health indicators among healthy adult participants; and 3) compare the relative impact of HEPA filtration for two major PM sources (traffic and residential wood combustion) on these indicators. 	  What is the purpose of the study? 	  This study will examine the health effects of exposure to different sources of air pollution in adults, and assess the effectiveness of portable HEPA filter air cleaners toward improving indoor air quality and improving health. 	  Who can participate in the study? 	  To participate in this study, you must be at least 19 years old and live in one of the targeted areas in the Vancouver and outlying areas. 	  Who should not participate in the study? 	  Smoking will affect the air quality inside the home, and therefore households with residents who smoke must be excluded from this study. We may exclude individuals with a prior diagnosis by their doctor  of  asthma,  chronic  obstructive  pulmonary  disease  (COPD),  heart  disease,  diabetes,  or hypertension. We may also exclude individuals living in neighbourhoods previously identified as being influenced by air pollution sources other than those of interest in this study. 	  What does the study involve? Overview The total sampling time is 14 days. Two back-to-back air pollution sampling sessions will take place in your home and will last 7 days each. The two consecutive 7-day sessions will start and end on a weekday morning, with a total of 3 visits during that time from the study technicians. Sampling sessions will be scheduled at a time convenient for you, when you have been free of any colds or other infections for at least 2 weeks. 	  For the duration of the sampling session, air monitoring instruments and HEPA filters will be running in your home. At the start of sampling (visit one), sampling instruments will be set up, and you will be interviewed about your health status and characteristics of your home; you will also be asked to provide a small blood sample. At visit 2 and visit 3 (the end of sampling) maintenance on the equipment (equipment removal at visit 3) will be performed. Also at visits 2 and 3, repeat blood samples will be drawn, and the blood vessel health test will be performed. In all, 3 blood draws will be done, and 2 blood vessel health tests will be performed. 	  	   118	  Here is a list of the specific procedures that will be included if you decide to join this study: 	  At visit one: 	  The environmental technician will set up a portable air cleaner in your bedroom and in the main common room (a room other than the kitchen) of your home. He or she will also set up air quality monitoring equipment in the main common room and in a secure outdoor area. This equipment will measure fine particulate matter (PM2.5) in the air as well as an odourless gaseous air pollutant, carbon monoxide (CO). The equipment is designed to operate quietly and unobtrusively and presents no known risks to the occupants of the home. In total, the equipment will take up a 1 meter x 1 meter space and technicians will work with you to locate it to minimize any disruptions. The equipment will need to be plugged into an electrical outlet in each of those areas; 	  The environmental technician will need access to your home for approximately one hour at visit one for the installation, and you will need to be present to answer some questions about your home’s characteristics (age, building materials, heating system etc.); 	  The medical technician will be present to ask you some questions about your health status and take a blood sample (approximately 15 ml, or 1 tablespoon) from your arm.  This visit will be made as early as convenient for you in the morning. 	  At visit two (7 days after visit one): 	  The technicians will need access to your home for approximately 45 minutes to perform maintenance on the air samplers; 	  During this visit, a non-invasive test of the health of your blood vessels will be performed (vascular function). This involves having a blood-pressure cuff on one arm and a small sensor on two fingers. The test takes approximately 20 minutes and should not cause any discomfort. Since this measurement requires you to fast for 5 hours before the measurement, this test will be made as early as convenient for you in the morning. The second blood sample will be collected at this time. 	  At visit three (7 days after visit 2): 	  The technicians will return to your home a final time to shut off and remove all equipment; 	  The final blood vessel health test and collection of the final blood sample will be performed. The visit will take approximately 45 minutes to 1 hour, and as with visit two, will be scheduled for as early as convenient in the morning. 	  With the exception of fasting prior to the blood draws and blood vessel tests, you and other residents should engage in normal activities during the monitoring session. You will be asked to keep a simple log of locations and activities (at home, at work, use of woodstove, cooking, dusting, etc.) for each sampling period. You will also be asked to complete a brief questionnaire about health irritation symptoms you may have experienced during the sampling session. You do not need to respond to any questions that you are not comfortable answering. 	  The total amount of direct contact time with study technicians required for you to participate in the study is about 3.5 to 4 hours. 	  	   119	  	  All blood samples and blood vessel tests will be identified by research code only; no personally identifying information will be used. 	  If you agree to being contacted in the future, we may invite you to participate in a follow-up study. You may indicate your willingness to be contacted for future studies at the end of this consent form. 	  What are the possible harms of participating? 	  There may be some minor discomfort or bruising on your arm at the site of the blood collection and you may experience some lightheadedness and/or fainting. As with any blood sample collection, there is a very small risk of infection at the site of the collection. There are no risks involved with the measurement of vascular function. As both the blood test and the vascular function measurement requires that you do not eat in the 5 hours before the test, this may cause some temporary discomfort due to hunger; we will schedule all procedures as early as convenient for you in order to minimize this disruption. 	  There are no risks involved with the air pollution sampling. All sampling involves measurement of compounds normally present in indoor and outdoor air. There are no risks involved with the air cleaner, which filters out fine particles from the air. 	  Participating in this study may cause some inconvenience for yourself and other members of your household, but technicians will work with you to minimize these impacts. 	  What are the possible benefits of participating in this study? 	  You will receive a summary of the measurements made in your home at the end of the study. You may also benefit from learning more about air quality and the effectiveness of HEPA air cleaners toward improving indoor air quality. 	  What happens if you decide to withdraw your consent to participate? 	  Your participation in this research is entirely voluntary. You may withdraw from this study at any time. If you decide to participate and then decide to withdraw at any time during the course of the study, your samples will have their ID number removed and they will be discarded according to standard laboratory procedures. There will be no penalty or loss of benefits to which you are otherwise entitled. If you wish to withdraw from this study, please contact the study coordinator, at the number listed at the end of this document. 	  If you choose to enter the study and then decide to withdraw after the study has been completed, all data collected about you during your enrolment in the study will be retained for analysis. All data and personal samples collected will be governed by the Personal Protection and Electronic Documents Act (PIPEDA) and also by the SFU Office of Research Ethics Policy R20.01 	  After the study is finished 	  At the end of the study, a summary of your personal results will be sent to you by mail to your home address. 	  Possible cost of participating in the study 	  The equipment will need to be plugged into an electrical source and will use a small amount of electricity. 	  	   120	  	  Payment to you for Participating 	  To assist with the inconvenience of your participation in this study, you will receive an honorarium of $250 upon completion of the two week sampling session. If you do not complete the sessions, the payment will be pro-rated to the amount of time you participated. 	  Will my taking part in this study be kept confidential? Your confidentiality will be respected.  No information that discloses your identity will be released or published without your specific consent to the disclosure. All samples and tests will be identified by a code only; personal identifiers will be removed. Participants Master File Key (the document with your name and individual code) are kept in a locked cabinet in a secure office at SFU. However, no records which identify you by name or initials will be allowed to leave the Investigators' offices. All documents are kept at SFU for the purpose of monitoring the research for a period of 5 years. 	  Whom do I contact if I have questions about the study during my participation? 	  If  you  have  any  questions  or  desire  further  information  about  this  study  before  or  during participation,  you  can  contact  the  study  technicians,  (T.B.D.),  the  study  coordinator,  Barbara Karlen  at  778.782.9324  (cell:  604.839.2341),  or  the  principal  investigator  Ryan  Allen  at 778.782.7631 	  Whom do I contact if I have any questions or concerns about my rights as a subject during the study? 	  If you have any concerns about your rights as a research subject and/or your experiences while participating in this study, contact: 	  Dr.  Hal  Weinberg,  Director,  Simon  Fraser  University  Office  of  Research  Ethics  at <> or 778-782-6593 	  Conflict of Interest 	  There are no known conflicts of interest on the part of the study investigators or the study sponsors (CIHR: Canadian Institutes of Health Research). 	  Subject consent to participate 	  I have read and understood this consent form. I have had sufficient time to consider the information provided and to ask for advice if necessary; I understand that my participation in this study is entirely voluntary and that I may refuse to participate or withdraw from the study at any time without any consequences; I understand that all of the information collected will be kept confidential and will only be used for scientific objectives; I understand that I am not waiving any of my legal rights by signing this consent form; I have been told that I will receive a dated and signed copy of this consent form. 	  My signature below indicates that I consent to participate in the Cardiovascular Effects of Aerosols in Residences (CLEAR) Study: 	  And, if checked below, that I consent to future contact for follow-up. 	  Study personnel may contact me about future follow-up research studies 	  	   121	  	  	  	  	  	  Printed Name of Subject Signature Date 	  	  	  	  	  Printed Name of Witness Signature Date 	  	  	  	  	  Printed Name of Signature Date Principal Investigator	  	  	  	  	   122	  	  	  	  	  	   	  	  	  	  	  APPENDIX	  IV:	  HEALTH	  AND	  EXPOSURE	  REPORT	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   123	  	  	  	  <first	  name>	  <last	  name>	  <address	  1>	  <address	  2>	  <city>,	  BC	  	   <postal	  code>	  	  Thank	  you	  for	  participating	  in	  the	  Cardiovascular	  Effects	  of	  Aerosols	  in	  Residences	  (CLEAR)	  Study	  October	  27th,	  2012	  	  Dear	  <first	  name>,	  	  Thank	  you	   for	   participating	   in	   the	   Cardiovascular	   Effects	   of	   Aerosols	   in	   Residences	   (CLEAR)	   Study.	  We	  sincerely	  appreciate	  your	  willingness	  to	  allow	  us	  into	  your	  home,	  permit	  air	  sampling,	  and	  participate	  in	  blood	  draws	   and	  blood	   vessel	  measurements.	  We	  are	   also	   grateful	   for	   your	  patience	   in	   awaiting	   these	  results;	  the	  numerous	  analyses	  in	  multiple	  laboratories	  take	  time	  and	  results	  are	  most	  meaningful	  when	  we	  can	  report	  your	  measurements	  in	  the	  context	  of	  all	  study	  participants’	  results.	  	  STUDY	  OVERVIEW:	  	  As	  you	  will	  recall,	  this	  survey	  had	  two	  components:	  a	  health	  component	  and	  an	  air	  monitoring	  component.	  We	   collected	   participants’	   blood	   on	   three	   occasions	   and	   measured	   blood	   pressure	   and	   the	   health	   of	  participants’	  blood	  vessels	  on	  two	  occasions;	  in	  addition,	  we	  measured	  the	  air	  (particulate	  matter)	  inside	  and	  outside	  participants’	  homes	  over	  a	   two-­‐week	  period.	  Together	   these	  measurements	   are	  helping	  us	  understand	   the	   cardiovascular	   health	   effects	   of	   air	   pollution	   and	   the	   potential	   benefits	   of	   using	   high	  efficiency	  particulate	  air	  (HEPA)	  filters	  indoors.	  	  THIS	  REPORT:	  	  This	   report	   shows	   your	  personal	   (physiological)	   results	   from	  both	   testing	   sessions	   and	   results	   from	  the	  environmental	  sampling	  conducted	  inside	  your	  home	  (with	  outdoor	  measurements	  used	  for	  a	  comparison).	  Physiological	   results	   reported	   include	   C-­‐reactive	   protein	   (CRP)	   in	   your	   blood,	   an	  indicator	   of	  inflammation,	  and	  endothelial	  function,	  an	  indicator	  of	  blood	  vessel	  health.	  For	  both	  measurements,	  we	  are	  also	  providing	  some	  information	  to	  help	  you	  interpret	  your	  results.	  Other	  measurements	  in	  blood	  are	  considered	  experimental;	  results	  for	  those	  tests	  are	  not	  included	  here	  as	  there	  are	  no	  expected	  “normal”	  values	  for	  comparison.	  Please	  feel	  free	  to	  contact	  us	  if	  you	  are	  particularly	  interested	  in	  any	  results	  not	  reported	  here.	  	  Environmental	  results	  reported	  include	  PM2.5	  (particulate	  matter	  of	  2.5	  microns	  or	  smaller	  in	  diameter),	  levoglucosan	  (a	  marker	  of	  wood	  smoke	  particulate	  matter),	  hopanes	  (an	  organic	  marker	  for	  traffic-­‐related	  particles),	  and	  infiltration	  efficiency	  (the	  fraction	  of	  the	  outdoor	  particulate	  matter	  that	  makes	  its	  way	  inside	  your	  home).	  All	  environmental	  measurements	  were	  taken	  over	  two	  1-­‐week	  (7-­‐day)	  periods.	  For	  one	  of	  the	  7-­‐day	  periods,	  a	  HEPA	  filter	  was	  in	  place	  in	  the	  air	  cleaner	  housing	  unit;	  for	  the	  other	  7-­‐day	  period	  the	  air	  cleaner	  was	  operated	  without	  a	  HEPA	  filter	  installed.	  Whether	  the	  HEPA	  filter	  was	  in	  place	  during	  the	  first	  or	  second	  week	  of	  sampling	  was	  determined	  randomly.	  	  At	   the	   same	   time	   as	   your	   indoor	   sampling,	  we	   collected	   environmental	   data	   outside	   of	   your	  home.	   In	  addition	  to	  giving	  us	  measurements	  of	  the	  amount	  of	  particulate	  matter	  in	  the	  atmosphere	  around	  your	  residence,	  this	  information	  is	  used	  to	  determine	  the	  infiltration	  rate	  for	  your	  home.	  	  	  GENERAL	  STUDY	  FINDINGS:	  	  	  	   124	  Our	  analysis	  confirmed	  that	  the	  HEPA	  filter	  air	  cleaners	  do	  improve	  overall	  air	  quality	  inside	  homes.	  On	  average,	  the	  use	  of	  HEPA	  filters	  resulted	  in	  a	  XX%	  reduction	  in	  levels	  of	  particulate	  air	  pollution	  (from	  a	  variety	  of	  sources	  including	  traffic,	  woodburning,	  cooking,	  etc.)	  indoors.	  In	  addition,	  we	  found	  that	  the	  use	  of	  HEPA	  filter	  air	  cleaners	  resulted	  in	  improved	  blood	  vessel	  health	  and	  reduced	  levels	  inflammation.	  Although	  the	  clinical	  significance	  of	  these	  relatively	  small	  changes	  is	  not	  clear,	  it	  does	  suggest	  that	  the	  use	   of	   HEPA	   filter	   air	   cleaners	   in	   this	   setting	   can	   reduce	   some	   negative	   impacts	   of	   exposure	   to	   air	  pollution.	  	  YOUR	  PERSONAL	  MEASUREMENTS:	  	  You	   supplied	   blood	   samples	  on	   three	   occasions	   and	   underwent	   a	   blood	   vessel	   health	   measurement	  (endothelial	   function)	   on	   two	   occasions.	   One	   7-­‐day	   sampling	   session	  measured	  with	   the	  HEPA	   filter	   in	  place	  and	  the	  other	  7-­‐day	  sampling	  session	  measured	  without	  the	  filter	  in	  place.	  	  As	  mentioned	  on	  the	  previous	  page,	  most	  of	  the	  analyses	  on	  the	  blood	  are	  experimental;	  therefore,	  only	  the	  tests	  for	  CRP	  in	  blood	  and	  your	  blood	  vessel	  (endothelial)	  function	  are	  reported	  here.	  Please	  note	  that	  for	  CRP	  a	  higher	  value	  indicates	  greater	  cardiovascular	  risk,	  while	  for	  the	  blood	  vessel	  measurements	  a	  higher	  value	  indicates	  healthier	  vessels.	  	  Cold/Flu	  symptoms	  present	  at	  the	  time	  of	  your	  tests:	  <cold/flu	  –	  yes/no>	  Blood	  Pressure	  at	  visit	  1:	  <bp	  visit	  X>	  Blood	  Pressure	  at	  visit	  2:	  <bp	  visit	  1>	  Blood	  Pressure	  at	  visit	  3:	  <bp	  visit	  2>	  	   	  	  TEST	   With	  HEPA	  Filter	  in	  place	   Without	  HEPA	  Filter	  in	  place	  Your	  result	  –	  refer	  to	  chart	  below	  1,2	   Range	  of	  values***	  (25th	  –	  75th	  percentile	  range	  for	  all	  CLEAR	  participants)	   Your	  result	  –	  refer	  to	  chart	  below	  1,2	   Range	  of	  values***	  (25th	  –	  75th	  percentile	  range	  for	  all	  CLEAR	  participants)	  	  CRP*	  (mg/l)	   	  <hepa	  1	  CRP>	   	  X.XXX	  -­‐	  X.XXX	   	  <hepa	  0	  CRP>	   	  X.XXX	  -­‐	  X.XXX	  	  Endothelial	  Function**	   	  <hepa	  1	  endo>	   	  X.XXX	  -­‐	  X.XXX	   	  <hepa	  0	  endo>	   	  X.XXX	  -­‐	  X.XXX	  	  *	  C-­‐Reactive	  Protein	  (CRP)	  measured	  in	  blood	  is	  a	  marker	  of	  inflammation	  in	  the	  body,	  and	  is	  used	  as	  a	  general	  indicator	  of	  cardiovascular	  disease	  risk3.	  Higher	  values	  indicate	  more	  inflammation.	  	  It	  is	  important	  to	  note	  that	  this	  measurement	  is	  sensitive	  to	  whether	  or	  not	  you	  had	  a	  cold,	  flu	  or	  other	  infection	  during	  this	  time	  and	  the	  results	  should	  therefore	  be	  interpreted	  cautiously.	  The	  following	  chart	  will	  help	  you	  to	  see	  how	  your	  results	  for	  CRP	  can	  be	  interpreted.	  We	  suggest	  that	  you	  discuss	  any	  elevated	  measurements	  with	  your	  physician:	  	  1	  mg/l	  =	  low	  risk	  1	  -­‐	  3	  mg/l	   =	  moderate	  risk	  3	  mg/l	   =	  high	  risk	  10	  mg/l	   =	  consult	  your	  physician	  for	  a	  repeat	  test	  	  **	  Endothelial	  Function	  is	  a	  test	  of	  the	  health	  of	  the	  endothelium	  -­‐	  a	  layer	  of	  cells	  that	  line	  blood	  vessels.	  It	  is	  important	  to	  keep	  in	  mind	  that	  this	  test	  is	  sensitive	  to	  many	  external	  factors,	  and	  is	  not	  meant	  to	  be	  used	  to	  diagnose	  disease	  but	  may	  indicate	  a	  general	  risk	  for	  atherosclerosis	  (cardiovascular	  disease).	  Higher	  values	  indicate	  healthier	  blood	  vessels.	  For	  Endothelial	  Function,	  any	  result	  of	  1.6	  or	  higher	  is	  considered	  “normal”.	  	  ***	  We	  have	  included	  the	  25th–75th	  percentile	  measurement	  range	  for	  all	  CLEAR	  study	  participants	  (N=XX),	  so	  you	  can	  compare	  your	  personal	  measurements.	  Note	  that	  50%	  of	  subjects	  will	  have	  values	  outside	  of	  this	  range.	  Please	  note	  that	  ranges	  are	  meant	  only	  for	  interest	  and	  should	  not	  be	  used	  to	  infer	  any	  disease.	  	  If	  you	  see	  a	  “.”,	  this	  means	  there	  was	  a	  sampling	  problem	  from	  that	  session	  or	  the	  session	  was	  not	  completed	  A	  note	  on	  units	  of	  measure:	  mg/l	  =	  milligrams	  per	  litre	  	  	  	   125	  YOUR	  HOME	  MEASUREMENTS:	  	   	  	  Average	  pollutant	  level	  over	  7	  days	  	  With	  HEPA	  Filter	  in	  place	   	  Without	  HEPA	  Filter	  in	  place	  	  Your	  home	  4,5	   	  Range	  of	  values*	  (25th	  –	  75th	  percentile	  range	  in	  all	  CLEAR	  study	  homes)	   	  Your	  home	  4,5	  	  Range	  of	  values*	  (25th	  –	  75th	  percentile	  range	  in	  all	  CLEAR	  study	  homes)	  Particulate	  matter6	  (PM2.5)	  [μg/m3]	   <hepa	  1	  home	  PM>	   	  X.XXX	  -­‐	  X.XXX	   <hepa	  0	  home	  PM>	   	  X.XXX	  -­‐	  X.XXX	  Levoglucosan7	  [ng/m3]	   <hepa	  1	  home	  levo>	   	  X.XXX	  -­‐	  X.XXX	   <hepa	  0	  home	  levo>	   	  X.XXX	  -­‐	  X.XXX	  Hopanes8	  [ng/m3]	   <hepa	  1	  home	  hopane>	   	  X.XXX	  -­‐	  X.XXX	   <hepa	  0	  home	  hopane>	   	  X.XXX	  -­‐	  X.XXX	  	  Infiltration9	  %	   <hepa	  1	  home	  infiltration>	   	  X.XXX	  -­‐	  X.XXX	   <hepa	  0	  home	  infiltration>	   	  X.XXX	  -­‐	  X.XXX	  	  *We	  have	  included	  the	  25th–75th	  percentile	  measurement	  range	  for	  the	  homes	  sampled	  in	  this	  study	  (N=XX),	  so	  you	  can	  compare	  the	  values	  from	  your	  home.	  Note	  that	  50%	  of	  homes	  will	  have	  values	  outside	  of	  this	  range.	  	  If	   you	   see	   a	   “.”	   in	   any	   cell,	   this	  means	   there	  was	   a	   sampling	   problem	   from	   that	   session	   or	   the	   session	  was	   not	  completed.	  A	  note	  on	  units:	  all	  pollutants	  are	  expressed	  as	  a	  concentration	  in	  air:	  “ppm”	  (parts	  per	  million)	  means	  that	  there	  was	  1	  unit	  of	  pollutant	  per	  million	  units	  of	  air.	  Measurements	  of	  fine	  particulate	  matter	  are	  expressed	  as	  a	  mass	  per	  volume	  of	  air	  (1	  μg	  =	  1	  part	  per	  1,000,000	  grams;	  1	  ng	  =	  1	  part	  per	  1,000,000	  grams).	  Particulate	  matter:	  Higher	  values	  indicate	  higher	  levels	  of	  particles	  inside	  the	  home.	  Levoglucosan:	  1,6-­‐anhydro-­‐β-­‐d-­‐glucopyranose,	  a	  cellulose	  combustion	  product,	   is	  a	  marker	  of	  wood	  smoke.	  Higher	  values	  indicate	  higher	  levels	  of	  woodsmoke	  inside	  the	  home.	  Hopane:	   17α(H),21β(H)	   –	   hopane	   an	   organic	   marker	   of	   traffic-­‐related	   air	   pollution.	   Higher	   levels	   indicate	   higher	  levels	  of	  traffic-­‐generated	  particles	  inside	  your	  home.	  Infiltration:	  Higher	  values	  indicate	  a	  greater	  contribution	  of	  air	  pollution	  from	  outdoor	  sources	  to	  indoor	  air.	  	  How	  do	  my	  results	  compare	  with	  air	  quality	  standards?	  	  There	  are	  no	  national	   indoor	  air	  quality	   standards	   and	   indoor	  concentrations	  of	   particulate	  matter	  are	  typically	   higher	   than	   those	   measured	   outdoors.	   For	   outdoor	   air,	   B.C.’s	   criteria	   for	  PM2.5	  of	   25	   µg/m3	  (averaged	  over	  24	  hours)	  established	  in	  2009,	  remains	  in	  effect.	  	  What	  can	  I	  do	  to	  minimize	  exposure	  to	  air	  pollution	  and	  improve	  the	  air	  quality	  in	  my	  home	  and	  community?	  	  There	   are	   a	   number	   of	   things	  we	   can	   do	   as	   individuals	   to	  minimize	   exposure	   to	   air	   pollution	  and	   to	  improve	  the	  air	  quality	  in	  our	  homes	  and	  communities.	  	  Personal	  exposure	  to	  air	  pollution	  can	  be	  reduced	  by	  altering	  behaviour.	  The	  BC	  Air	  Quality	  Health	  Index	  (	  is	  a	  scale	  designed	  to	  help	  you	  understand	  day-­‐to-­‐day	  changes	  in	  outdoor	  air	  quality	  and	  make	  decisions	  to	  protect	  your	  health	  by	  limiting	  short-­‐term	  exposure	  and	  adjusting	  your	  activity	   levels	   during	   increased	   levels	   of	   air	   pollution.	  Increased	   use	   of	   active	  transportation,	   public	  transportation,	  or	  car	  sharing	  can	  reduce	  reliance	  on	  personal	  vehicles,	  thus	  reducing	  vehicle	  emissions,	  a	  major	   source	   of	   air	   pollution	   in	   greater	   Vancouver.	  When	   traveling,	  your	   route	   can	   influence	   your	  exposure	  to	  traffic-­‐related	  air	  pollution.	   An	  online	  tool	  (	  has	  been	  developed	   for	  greater	  Vancouver	   that	   allows	  users	  to	  select	   the	  cycling/walking	   route	  with	   the	   lowest	  traffic-­‐related	  air	  pollution	  levels	  	  There	  are	  also	  options	  for	  reducing	  air	  pollution	  levels	  inside	  your	  home.	  As	  the	  results	  from	  this	  study	  show,	  HEPA	   filter	   air	   cleaners	   help	   reduce	   particulate	  matter	   concentrations	   inside	   homes	  	  	   126	  by	  XX%,	  on	  average.	   You	  can	  also	  minimize	   the	   infiltration	  of	  outdoor	  pollution	  by	  keeping	  windows	  closed	  during	  high	  pollution	  periods	  (e.g.	  rush	  hour).	  	  If	  you	  own	  a	  woodstove,	  both	  the	  type	  of	  stove	  and	  its	  operation	  are	  important	  for	  air	  quality.	  Newer,	  EPA	   certified	  woodstoves	   emit	   less	   pollution	   than	  older,	   non-­‐certified	   stoves.	  Also,	  we	  urge	   all	  wood	  stove	  owners	  to	  take	  a	  training	  course	  to	  learn	  how	  to	  use	  your	  stove	  most	  efficiently.	  	  The	  study	  website	  will	  be	  updated	  with	  publications	  related	  to	  this	  study	  as	   they	  become	  available.	  No	   individual	   results	  will	   ever	  be	  posted	  on	   the	  website	  or	  reported	   in	   any	  publications.	  We	   have	   included	   a	   list	   of	   helpful	  web	   links	   on	   the	   last	   page	   of	  this	   letter,	   but	   please	  contact	  us	   should	  you	  have	  any	   further	  questions	  regarding	   the	  study	  or	  results	   (	  or	  Dr.	  Ryan	  Allen,	  	  Thank	  you	  again	   for	  contributing	  to	  a	  better	  understanding	  of	   the	   impacts	  of	  wood	  burning	  and	  traffic	  emissions	  on	  air	  quality	  and	  health.	  	  Sincerely,  	  Dr Ryan Allen Principal Investigator	  	  	  	  	  	  	  	  	  	   127	  Helpful Resources and Links: 	   1. SFU Cardiovascular Effects of Aerosols in Residences Study (CLEAR) 	   2. The Lung Association Great resource about pollution and air quality, and what you can do to recognize and solve problems. Excellent information about air cleaning devices including HEPA filters under the Indoor Air section: 	   3. BC Air Quality Health Index 	   4. BC Health Files For information on air quality and pollution. 	   5. BC Air Quality Provincial site devoted to air quality in BC. 	   6. Canada Mortgage and Housing For an on-line guide to residential wood heating. 	   7. Environment Canada Clean Air section has information about wood heating and resources to help Canadians make informed decisions and take action to reduce air pollution. 	   8. Health Canada Air Quality Information 	  Provides information about Canadian Research and guidelines covering air quality: 	   9. Excellent Information on HEPA Filters can be found at: 	  The Lung Association: U.S. Environmental Protection Agency: 	   10. Recent publications on Traffic-Related Air Pollution and Health: 	   11. If you are interested in learning more about cycling in urban areas: 	  	   	  	  	  	  	  	   128	  	  	  	  	  	  	  	  	  	  	  	  APPENDIX	  V:	  DWELLING	  INFORMATION	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   129	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  Home	  and	  heating	  questionnaire	  (to	  be	  completed	  by	  technician)	  	  	  Home ID:    Date :    	  Participant ID:    	  Second participant ID (if applicable)    	  Sampling dates: and . DD/MM/YY   -  DD/MM/YY DD/MM/YY   -   DD/MM/YY 	   1. Dwelling: a. Postal Code:    	   b. GPS Accuracy: 	   c. Latitude:    	   d. Longitude:    	   e. Elevation:    	   f. Age of home:    (year built: ) 	   g. Substantial renovations (frame/windows/exterior) year: 	   	  	  h. 	  Proximity to major roads (major road = 4 lanes): i.  On a major road? Yes 	  	  	  No 	   ii.  If not, within 50m of a major road? Yes No 	  	  i. 	  Type of building: i.  single family mobile/trailer duplex 	  	  townhouse 	  	  apartment building ii.  other:   	   j. Floor (ground floor = 1st floor): 	  (single family home or townhouse = 1) 	   k. Size of home: i. square footage:    	   ii. ceiling height:    	   iii. approximate volume:    	  	   l. Number of levels: (including basement) 	   m. Number of windows:    Number of windows that open:    	  	   130	  2. Heating system: a. Primary heating system: i. Woodstove ii. Electrical iii. Gas iv. Forced Air/Furnace 	  	  	   v. Hot Water/Radiator vi. Propane vii. Oil viii. Other:    	   b. Secondary heating system: i. Woodstove ii. Electrical iii. Gas iv. Forced Air/Furnace 	  	   v. Hot Water/Radiator vi. Propane vii. Oil viii. Other:    	   3. Wood stove (if applicable): 	   a. Type of wood stove (if applicable): i. Fireplace ii. Conventional wood stove iii. Insert 	  	  	  	   iv. Certified wood stove v. Pellet stove 	   b. Wood stove brand and model:   	   c. Approximate age of wood stove (years): (year of stove: ) 	   d. Location of wood stove in house (describe):   	   e. Fresh air intake installed on stove: yes no 	   f. Chimney: i. Masonry Interior/Exterior Condition:   	   	   ii.Class A Interior/Exterior Condition:   	   g. Approximate proportion of household heating from wood: 	  i.  >90% ii.   50-90% iii.   20-50% iv.   <20% 	  	  	   h. Moisture content of two representative logs:    	  	  	  	  	  	  	  	  	   131	  4. Other house characteristics and emission sources: 	   a. Estimate percentage of floor space covered with carpets (for the entire house):    	   b. Do any you have any pets that reside INSIDE the home? Yes No  Type: 	  c. 	  Are there any cigarette smokers residing at this residence? 	  Yes 	  No 	   d. Kitchen: 	  	   	   i.  Stove type? Gas Electric 	  	   	   ii.  Range hood? Yes No 	  	   	   1.  If yes, is it used? Yes No 	  	   	   a.  If yes, how often? Always Sometimes Never 	   e. Fireplace (in addition to wood burning appliance specified above): i. Present: Yes No ii. Type and number: Wood: Gas:   	   f. Chimney: 	  	   i. Masonry Interior/Exterior Condition:   	   ii. Class A Interior/Exterior Condition:   	   g. Type of Garage: # of cars    i. Attached (used for parking) ii. Attached (not used for parking) iii. Underneath building  iv. Not attached to building v. No garage	  	   132	  	  	  	  	  	  	  	  	  	  	  	  APPENDIX	  VI:	  TIME-­‐LOCATION-­‐ACTIVITY	  DIARY	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   133	  	  	  	  	  	  	  	  	  	  	  	  	  	  Pg __ of 7KEY: !"#$% Time YES NO YES NO YES NO YES NO YES NO YES NO YES NO YES NO YES NO YES NO&%''()%''*+, - . - . - . - . - . - . - . - . - . - .)%''(/%''*+, - . - . - . - . - . - . - . - . - . - ./%''(0'%''*+, - . - . - . - . - . - . - . - . - . - .0'%''(00%''*+, - . - . - . - . - . - . - . - . - . - .00%''(01%''*2, - . - . - . - . - . - . - . - . - . - .01%''(0%''*2, - . - . - . - . - . - . - . - . - . - .0%''(1%''2, - . - . - . - . - . - . - . - . - . - .1%''(3%''*2, - . - . - . - . - . - . - . - . - . - .3%''(4%''*2, - . - . - . - . - . - . - . - . - . - .4%''(5%''*2, - . - . - . - . - . - . - . - . - . - .5%''(6%''*2, - . - . - . - . - . - . - . - . - . - .6%''(&%''*2, - . - . - . - . - . - . - . - . - . - .&%''()%''*2, - . - . - . - . - . - . - . - . - . - .)%''(/%''*2, - . - . - . - . - . - . - . - . - . - ./%''(0'%''*2, - . - . - . - . - . - . - . - . - . - .0'%''(00%''*2, - . - . - . - . - . - . - . - . - . - .00%''(01%''*+, - . - . - . - . - . - . - . - . - . - .01%''(0%''*+, - . - . - . - . - . - . - . - . - . - .0%''(1%''*+, - . - . - . - . - . - . - . - . - . - .1%''(3%''*+, - . - . - . - . - . - . - . - . - . - .3%''(4%''*+, - . - . - . - . - . - . - . - . - . - .4%''(5%''*+, - . - . - . - . - . - . - . - . - . - .5%''(6%''*+, - . - . - . - . - . - . - . - . - . - .6%''(&%''*+, - . - . - . - . - . - . - . - . - . - .Location: H = Home, W = Work, Oth = Other indoor location, Out = any outdoor location, Tr = in transitPlease circle Y or N for all other questions for each time period. Please use 1 sheet per day. Include any locations/activites of 5 minutes or more in that hour.  OK to mark multiple locations/activities in an hour.!"#$%&#'%#(%)#*+,-#+#.%%'#/)%,-0#12-+/-#2-+,-#)*-/-#/-3)4%(/#52+(67H, W, Oth, Out, TrSubject A ID:                               Subject B ID: Session: Filter ID: Stove door opened? Other heat source on? Cooking? Dusting or  Vacuuming? Candles or Incense? Window Open?Subject A location ________ Subject B location ________Session Start Time: Subject A Tobacco Smoke? Subject B Tobacco Smoke? Wood stove burning? Wood added to stove? Notes/ Comments	  	   134	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  Weekly	  Medication	  &	  Activity	  Questionnaire	  	  	  	  Participant ID:    Session:   1 or  2 Questionnaire Date:    	  Sampling session start and end date:    dd/mm/yyyy - dd/mm/yyyy 	  Please give the name, strength, and frequency of use for all medications, vitamins, and supplements that you have used in the past 7 days (since the last technicians’ visit). 	  Prescription medication use: (Those that were prescribed by a doctor and filled by a pharmacist; including: pills, dermal patches, eye drops, creams, salves, nasal sprays, metered dose inhalers, suppositories, and injections). The lists below should include medications you informed us of during the initial screening. Check/confirm strengths, etc. with participant screening questionnaire if applicable. 	   	  	  Data Entry Code 	   Medication Name 	   (Please include any use of „rescue‟ medications such as Nitroglycerin, Asthma Inhalers, etc.) 	  Strength 	  	  	  (mg, mcg) Frequency How many of these pills, applications, or inhalations did you take (use) in the past 7 days? If none, please write in “0”. 	   	   	   	  	   	   	   	  	   	   	   	  	   	   	   	  	   	   	   	  	  Use of non-prescription medications, vitamins, or supplements: 	  	  Data Entry Code 	  Substance or Brand Name Type of Product 	  (pill, cream, spray, etc.) 	  Strength 	  	  	  (mg, mcg) Frequency How many of these pills, applications, or inhalations did you take (use) in the past 7 days? If none, please write in “0”. 	   	   	   	   	  	   	   	   	   	  	   	   	   	   	  	   	   	   	   	  	   	   	   	   	  	   	   	   	   	  	  Notes or other comments: (please use the last sheet if you need additional space) 	  	   135	  The	  following	  questions	  are	  intended	  to	  obtain	  more	  specific	   information	  about	  your	  activities	  and	  the	  characteristics	  of	  your	  home	  over	  the	  past	  7	  days	  (since	  our	  last	  visit	  to	  your	  home).	  	   1. During the past 7 days, how much time did you spend doing the following during your travel time (including commute)? 	   a. walking or biking  hours  minutes 	   b. in a private car or taxi  hours  minutes 	   c. on a bus 	    hours  minutes 	   d. on a train (inc. SkyTrain)  hours  minutes 	   e. other 	    hours  minutes please specify:   	   2. What road/traffic condition best describes the roads on which you spent the majority of your travel time during the past 7 days? 	   a. Side roads / neighbourhood streets 	   b. Highways / major roads / arterials with: 	   i. free-flowing traffic moving at the speed limit 	   ii. heavy traffic moving below the speed limit 	   iii. very heavy “stop and go” traffic 	   c. N/A, I travel primarily by train or SeaBus 	  	   3. During the past 7 days, how much of the time were windows open in your home (include any open windows)? 	    hours  minutes OR  percent of the 7-day period 	   4. Did anyone (including yourself) smoke in your home in the past 7 days? 	   □ Yes 	  	  	  4a.  How many hours total did smoking occur inside your home in the past 7 days? 	   □ No (skip to question 5)  hour(s) □ Don‟t Know (skip to question 5) 	  	  	   5. During the past 7 days, has your home been smoky from cooking (e.g. burnt toast, barbecue, stir fry, etc.) at any time? 	   □ Yes 	   □ No (skip to question 6) 	   □ Don‟t Know (skip to question 6) 	  	  5a.  How many hours total was your home smoky from cooking in the past 7 days? 	    hour(s) 	  	   136	  6. During the past 7 days, in addition to the portable HEPA filters that we placed in your home as part of this study, was any other air cleaner/filter (stand-alone/portable or central) used in your home? 	   □ Yes 	   □ No (skip to question 7) 	   □ Don‟t Know (skip to question 7) 	  	  	   6a.  What kind of air cleaner was used? 	   a. HEPA filter 	   b. Electrostatic precipitator 	   c. Negative ion generator 	   d. Ozone generator 	   e. Don‟t know 	   f. Other, specify:    	  6b.  Where is the air cleaner located? 	  	  	  6c.  During the past 7 days, how much of the time was the air cleaner used? 	    hours  minutes 	  OR	  	    percent of the 7-day period 	   □ Don‟t know 	  	   7. [If the home has a woodstove] How much was the woodstove used during the past 7 days? 	    hours  minutes OR  percent of the 7-day period (if 0 skip to question 8) 	  	  	  7a.  Estimated # of pieces of split firewood burned in the past 7 days? 	  	  7b.  Type of wood burned (and proportions if more than 1 type burned)? 	   a. Pine    	   b. Spruce    	   c. Alder    	   d. Fir   	   e. Hemlock    f. Maple    	   g. Scrap lumber     	   h. Other:    	  	   i. Don‟t know    	  	   8. How much time were the following used during the past 7 days in your home? 	   a. Air conditioning  hours minutes OR  percent of the 7-day period 	   b. Woodburning fireplace hours minutes OR  percent of the 7-day period 	   c. Gas fireplace  hours minutes OR  percent of the 7-day period 	  	   137	  TO	  BE	  COMPLETED	  BY	  PARTICIPANT	  	  For each item below, please make an X at the place on the line that best represents how you feel now compared to how you usually feel. If you feel no different than usual, mark the place below “As Usual,” even if you usually do not have that symptom. 	  Date:    	  	  	  Much Better Better 	  	  Slightly Better Irritated Eyes 	   As Usual 	  	  Slightly Worse Worse 	  	  Much Worse 	  	  	  	  Much Better Better 	  	  Slightly Better Itchy Eyes 	   As Usual 	  	  Slightly Worse Worse 	  	  Much Worse 	  	  	  	  Much Better Better 	  	  Slightly Better Stuffy Nose 	   As Usual 	  	  Slightly Worse Worse 	  	  Much Worse 	  	  	  	  Much Better Better 	  	  Slightly Better Runny Nose 	   As Usual 	  	  Slightly Worse Worse 	  	  Much Worse 	  	   Scratchy Throat or Sore Throat 	  Much Better Better 	  Slightly Better 	  As Usual 	  Slightly Worse Worse 	  Much Worse 	  	  	  	  Much Better Better 	  	  Slightly Better “Hoarse” Throat 	   As Usual 	  	  Slightly Worse Worse 	  	  Much Worse 	  	  	  	  Much Better Better 	  	  Slightly Better Cough 	   As Usual 	  	  Slightly Worse Worse 	  	  Much Worse 	  	   138	  	  	   	  	  	  	  	  	  	  	  	  	  	  	  	  	   	  	  	  APPENDIX	  VIII:	  HEALTH	  LOG	  	  	  	  	  	  	  	  	  	  	  	  	  	  	  	   139	   Health log sheet (to be completed by technician & subject)   Session X:  Home ID: _______________ Participant ID: ____________     Measurement Date: _______________ Technician: ____________                            (today’s date)   dd/mm/yyyy        For Participant to answer (1 & 2):   1. Please indicate which of the following best describes your racial heritage (you may indicate more than one group):   a. White b. Chinese c. South Asian (e.g., East Indian,  Pakistani, Sri Lankan, etc.) d. Filipino  e. Korea f. Aboriginal (First Nations) Including North American Indian, Metis, Inuit [Eskimo] g. Southeast Asian (e.g., Vietnamese, Cambodian, Malaysian, Laotian, etc.) h. West Asian (e.g., Iranian, Afghan, etc.) i. Japanese j. Latin American k. Black l. Arab m. Other — Please specify:     _________________________ 2. Do you currently have a cold or any other known infection?   _________________________    For Technician:   Session X Blood Draw Details  Confirm:  1.   Has subject had anything to eat or drink (including water)? ____________________ 2. Does subject have a history of fainting/troubles with blood draws? ___________________________________________________________________ 3. Blood sample collected date: _____________________________ i. Yellow tube collected  yes / no   # of tubes: _____ ii. Lavender tube collected  yes / no  # of tubes: _____  Before leaving Confirm:  o Consent signed o Home and Heating Questionnaire completed o Health Symptom Questionnaire reviewed o Activity log completion reviewed o Overview of next visit offered  	  	   140	  Session 1:  Home ID: _______________ Participant ID: ____________     Measurement Date: _______________ Technician: ____________                            (today’s date)   dd/mm/yyyy        ENDOPAT:   Subject’s dominant arm:  (LEFT / RIGHT)  Describe setup: room  ____________ lying/sitting/reclining ___________  on bed/chair/couch ____________ lighting (type:low/med) ___________ Room temperature:________   Has subject had anything to eat or drink (including water)? __________________________   Blood Pressure (dominant arm) Average Systolic Average Diastolic Heart rate Notes       EndoPAT (non-dominant arm) Target cuff pressure (systolic + 60; min. 200) Patient ID (Subject ID_session) RHI result AI result HR result Notes           Session 1 BLOOD Draw  Blood sample collected date: _____________________________ i. Yellow tube collected      yes / no   # of tubes: _____ ii. Lavender tube collected  yes / no  # of tubes: _____  Problems with blood collection? _______________________________________________________  _________________________________________________________________________________	  	   141	  Session 2:  Home ID: ______________ Participant ID: ____________    Measurement Date: ______________     Technician:  ____________                            (today’s date)   dd/mm/yyyy            Does subject currently have a cold or any other known infection? __________________________  Subject’s dominant arm:  (LEFT / RIGHT)  Describe setup: room   ____________   lying/sitting   ____________  on bed/chair/couch    ____________  lighting (type:low/med) _____________________________ Room temperature:________   Has subject had anything to eat or drink (including water)? __________________________  Blood Pressure (dominant arm) Average Systolic Average Diastolic Heart rate Notes       EndoPAT (non-dominant arm) Target cuff pressure (systolic + 60; min. 200) Patient ID (Subject ID_session) RHI result AI result HR result Notes          Session 2 BLOOD Draw  Blood sample collected date: _____________________________ i. Yellow tube collected      yes / no   # of tubes: _____ ii. Lavender tube collected  yes / no  # of tubes: _____  Problems with blood collection? _______________________________________________________  _________________________________________________________________________________ 	  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items