|
|
|
|
||||||||||||||||||||||||||||||||||||||||||||||||
|
||||||||||||||||||||||||||||||||||||||||||||||||
|
 |
|
||||||||||||||||||||||||||||||
|
Committee on the medical effects of air pollutantsMinutes of the meeting held on Friday 23rd February 2007, The Old Library, Department of Health, Richmond House, 79 Whitehall, London. SW1A
ITEM 1. OPENING REMARKS, APOLOGIES FOR ABSENCE AND ANNOUCEMENTS 1. Apologies for absence were received from Prof. V Stone, Prof. S Holgate, Prof. Poole-Wilson, Mr Fintan Hurley, Mr R Alexander (Welsh Assembly) 2. The Chairman welcomed Prof. Virginia Murray, from HPA, London, who had come to hear the Committee's discussion of paper 2007/01; Mr Richard Daniels from the Department for Education and Skills and Doctor Philip Penney, Medical Advisor to the Pool Water Treatment Advisory Group who had come to hear the Committee's discussion of Paper 2007/02; Dr John Rea who was attending as the Defra representative and Mr Andrew Whitcombe, policy co-ordinator at the Department of Health. 3. The Chairman reminded Members that they should declare any relevant interests. 4. The Chairman reminded Members to submit their expense claim forms. ITEM 2. MINUTES OF THE MEETING HELD ON 13th OCTOBER 2006 5. The Minutes of the last meeting were considered, a typographical error was noted and members agreed that, with a correction, the minutes were an accurate record of the meeting. ITEM 3. MATTERS ARISING Sub-group on Air Pollution and Child Health 6. Members were told that this sub-group was now working via correspondence: arranging a date for a meeting had not been possible. Members of the sub-group had been asked to review the WHO report on Childrens' Health and to identify gaps in that report and any advance made within their fields of expertise since it had been published. Members of the sub-group had been asked to forward their comments to the Secretariat on a chapter by chapter basis. The Secretariat will draft a short report which will be discussed at the June meeting of COMEAP. Strategy Meeting 7. The Chairman reported that he and the Secretariat had met to discuss the COMEAP work plan. He noted the following work items: Report: Does air pollution cause asthma? Action: Secretariat 9. It was noted that there had been some interesting work incorporating
GIS information into epidemiological studies of asthma, and it was agreed
that information about this should be included in the report. Action: Secretariat Report - Is there a threshold for Ozone? 10. This report will also be published in sections on the COMEAP website. Part One of the report will be published towards the end of this year and Part Two early in 2009. 11. It was noted that Chapter 8, on the effects of long-term exposure to ozone, would be published, initially, as a stand alone working paper accompanying the QUARK II mortality report. 12. It was noted that the Royal Society is studying ozone and that some
Members have been asked to contribute to this work. It was agreed that
the Secretariat would circulate the Royal Society's request for the submission
of evidence to Members. Members were encouraged to comment on an individual
basis if they so wished. It was also noted that the Air Quality Expert
Group (AQEG) was drafting a report on ozone which would be available in
the summer of 2007. Action: Secretariat 5 Year Review 13. Dr Roy Hamlet of the Radiation Protection Division of HPA, had agreed to undertake a five year review of COMEAP. The Secretariat agreed to keep Members informed of progress. Action: Secretariat Matters arising from the last meeting. 14. The Chairman told Members that the Secretariat had received comments
from Professor Paul Elliot (SAHSU) on the paper published by George Knox.
Professor Elliot had highlighted some issues regarding the linking of
recent mapping information with historical dates of death. The Secretariat
agreed to circulate Professor Elliot's comments to Members. 15. Members were informed that the Secretariat had incorporated comments from Members. The final editing of the Report by the Secretariat and the Chairman of the sub-group, Mr Fintan Hurley, would take place prior to the report being sent to the Chairman of COMEAP for final clearance. The report would then be placed on the COMEAP website for a period of six weeks for public comment. 16. Members were reminded that the report would be peer-reviewed by Doctor Bart Ostro, Professor David Coggon and Doctor Michal Krzyzanowski. Members were asked to provide suggestions for a fourth reviewer. It was agreed that Members would provide names for the Secretariat. Action: Members 17. Members were told that, once all comments had been incorporated and the necessary changes made, the report would be published in hard copy and on the COMEAP website. Total impact calculations 18. The Chairman reminded Members of the work being carried out by Doctor Heather Walton, Mr John Stedman, Dr Brian Miller and Mr Fintan Hurley to calculate the impact on mortality and life expectancy of changes in levels of air pollutants. An initial report would be presented to the QUARK II working group in May/June 2007 and to COMEAP before the end of the year. Elicitation exercise 19. Members asked whether the median results from the elicitation exercise carried out by Members of the sub-group would be included in the report. Members were informed that median results had been mentioned within the report and noted as another method which could be used, although the mean results from the exercise had been utilised in estimating the probability distribution for a range of possible coefficients. 20. A member noted that a paper had been published by the US EPA which showed the results of another elicitation exercise. The results of the US exercise differed from those produced by the COMEAP sub-group: the estimates from the US being generally higher. However, it was noted that it was hard to integrate the results from two exercises (QUARK II and the US EPA) as they differed in a number of ways, primarily in that a synthesis was not made of individual responses in the EPA exercise. It was agreed that the executive summary of the US EPA report should be circulated for information to members and the full report would be made available on request. Action:Secretariat
21. Members were informed that the first meeting of this group had been held on 18th December 2006. The purpose of the meeting had been to plan the work of the group and to set out the approach to be taken in considering the available evidence and in the preparation of the report. It was decided that the report would deal with both the effects of long-term and short-term exposure to air pollutants. It was acknowledged that the effects of short-term exposure, as reflected by time-series and panel studies were well established and it was agreed that the group should investigate whether studies which examined the effects of long-term exposure added significantly to the findings of the short-term studies. 22. It was confirmed that the Department of Health had agreed to fund an extension to the St George's database. This would deal with studies relating long-term exposure to air pollutants and effects on morbidity. This will be the key to much of QUARK's work. 23. The next meeting of the QUARK II working group was scheduled for 15th March 2007. Members were informed that the final report would be published in December 2008. ITEM 5a BUNCEFIELD FIRE (COMEAP/2007/01) 24. The Chairman welcomed Professor Virginia Murray to the meeting. Members were reminded that at the COMEAP meeting of June 2006 they had asked that the Met Office provide analyses of a number of scenarios showing what might have happened had the Buncefield fire occurred under different meteorological conditions. Members had made this request to help determine whether larger effects on health might have occurred. Members' attention was drawn to the report: Dispersion Modelling Studies of the Buncefield oil depot incident
25. The Chairman noted that the report showed that meteorological conditions conducive to better mixing within the boundary layer might have caused higher PM10 levels than those recorded. It was noted that the paper produced by the Secretariat had suggested that PM10 was unlikely to have exceeded 300µg/m3 (24 hour average). Members sought clarification of the latter figure. 26. Members' attention was drawn to the abstract of the Met Office report. It was noted that modelling the actual weather conditions during the incident predicted a maximum hourly average concentration of 151µg/m3 (PM10). The last paragraph of the abstract indicated that, under worst case conditions, concentrations might have reached twice this level. It was noted that these figures referred to predicted peak hourly average concentrations. Twenty four hour average concentrations would be likely to be significantly lower than these peak values. However, in the absence of direct information on predicted 24 hour average concentrations, the Secretariat had chosen to suggest that a worst case 24 hour average concentration of 300µg/m3 might have been achieved. It was agreed that this was very much a worst case estimate and probably unrealistic. 27. Members asked for the locations of measurements of pollutant concentrations to be identified. Professor Murray informed Members that measurements had been taken very close to the fires. 28. Members agreed that the model used by the Met Office was state of the art but noted that such modelling had inherent weaknesses. These included the lack of emission factors and uncertainties regarding estimation of the height of the plume. It was noted that assumed emission rates remained uncertain. Members accepted that these weaknesses could not be rectified without extensive further work. 29. In discussion of the characteristics of the boundary layer, (the lowest layer of the atmosphere, in which mixing of the air occurs) it was noted that, under stable conditions mixing within the boundary layer is poor. In unstable conditions there is good mixing by convection and in windy conditions good mechanical mixing is achieved. It was noted that modelling average boundary layer concentrations under poor mixing conditions might not provide an accurate representation of ground level concentration. 30. It was also noted that if the plume did not make contact with the ground then there would be no dry deposition. Grounding would lead to removal of both small and large particles: this process would therefore have an effect on the inhalable fraction left within the plume. 31. Some Members felt that the figure of 300µg/m3 PM10 24 hour average was a gross over estimate of likely peak concentrations even under worst case conditions. Others felt that, though the predicted level was unlikely, it was not inconceivable and therefore was an acceptable figure for worst case calculations. It was also pointed out that the comparison of worst case weather conditions where the conditions which obtained during the actual incident had involved the use of relative concentrations and not absolute concentrations. 32. It was agreed that some refinement of the estimate was required and the Committee requested that the Met Office ran an hourly simulation, 24 times, to provide more information on the distribution of concentrations within the 24 hour average. It was felt that the simulation could be used to provide finer spatial resolution of concentrations and that these could be mapped against population density. It was suggested that finer gradations of concentration than those used in the Met Office report might be applied. It was also suggested that any health impact calculations should be put into context. It was noted that current calculations of impact during worse case conditions suggested about the same increase in daily death rate as might be expected on a moderately cold day in London. Action: Professor Murray and Secretariat 33. Members asked whether particle samples had been taken from the plume and whether studies on the toxicity of particles were underway. It was noted that samples had been taken during a fly through of the plume and that on analysis only a very small amount of metal and no PAH compounds had been found. Particle number concentration measured in the plume was not especially high although it was noted that the measurement method used was insensitive to particles of less than 0.1µm diameter. Members felt that the clustering of small soot particles could have implications: this would reduce the particle count but surface area and thus, perhaps, free radical generating activity might still remain high. Professor Murray agreed to ascertain whether any samples were available for further analysis. Action: Professor Murray 34. Professor Murray told members that the Met Office report and their advice would be important in providing guidance on how to manage incidents such as the Buncefield fire. One point of a particular concern: it was often asked whether such a fire should be left to burn out or should be extinguished. Either option has implications for health. It was agreed that many of the concerns raised by scientists working on different aspects of the Buncefield incident might be discussed with Members of COMEAP at a future date. The Chairman agreed that it was important that Members should stay in touch with such important work. 35. Professor Murray told Members that further work on the Buncefield
incident was being carried out by the Health Protection Agency. It was
agreed that Professor Murray would bring results of this work to the June
meeting of COMEAP. ITEM 5b. SWIMMING POOLS AND ASTHMA (COMEAP/2007/02) 36. The Chairman welcomed Mr Daniels and Dr Penney to the meeting. The Chairman reminded Members of the statement they had agreed in 2003 on the possible association between the use of swimming pools and the prevalence of asthma. Consideration of this point had arisen from a previous paper published by Bernard et. al. New papers by Bernard et. al. had been characterised by firmer conclusions and Members were asked whether, in the light of further research in this area, their original statement required revision. 37. One Member noted that the papers did suggest an adverse effect but that this was less strong than claimed by the author. Information had been provided on study design and it was noted that the data were clustered (groups of schools or countries). In many cases the handling of clustered data had been appropriate in some analyses although the headline results quoted were not always from the analyses that took clustering into account. 38. Members discussed evidence regarding possible changes in lung permeability as a result of exposure to swimming pool atmospheres. Results relating to CC16 and albumin concentrations were not regarded as entirely persuasive. 39. Members considered it possible that asthma symptoms were exacerbated by exposure to swimming pool atmospheres. This they felt rather more likely than asthma being induced, de novo, as a result of exposure to chlorine and its by products in concentrations likely to be encountered at swimming pools. It was noted that these concentrations were markedly lower than those which had been associated with Reactive Airways Dysfunction Syndrome (RADS). It was noted that "swimming pool asthma" had been linked with dust in changing rooms at swimming pool facilities. 40. Members were reminded that chlorine is a very soluble gas and therefore not found, as such, in the atmosphere of swimming pools. Reaction products such as nitrogen trichloride (trichloramine) are, however, found in the air. It was questioned whether the irritative effects of nitrogen trichloride could induce an inflammatory process in the lung. It was noted that irritation could be as a result of direct effects on tissue or perhaps, as a result of neurological reflex activity. It was agreed that these mechanisms could only be separated by research. 41. In considering nitrogen trichloride Members were told that: a) At a concentration of 1mg/m3 there was no smell of chlorine (the odour of nitrogen trichloride is often confused with that of chlorine) and that this concentration was considered tolerable. b) At 3mg/m3 the usual smell associated with swimming pools was detected. c) At 5mg/m3 levels "objectionable". 42. It was noted that some of the levels of nitrogen trichloride in the
studies reported were considerably higher than 5mg/m3. 44. Members were told of a pilot study funded by PWTAG which was being led by the Amateur Swimming Association and was investigating the possibility of a link between swimming pools and asthma. Results were expected shortly and might feed into a larger epidemiological study if funding could be obtained. Members expressed interest in this study and asked to be kept informed. Members noted that chlorination is almost universally used at swimming pools and thus finding "control pools" would be difficult. The study would separate pools into those with non-irritant NCl3 levels and those with higher concentrations. Conclusions 45. Members noted that a balance of health outcomes should be considered. a) Reducing levels of swimming amongst children would incur a significant
health cost (less exercise). 46. It was agreed that it was no part of the remit of COMEAP to decide whether swimming was an activity which should or should not be undertaken by people with asthma. Members agreed that a precautionary approach should be taken and that chlorine levels should be kept at a minimum compatible with adequate control of bacterial levels. 47. It was agreed that the revised statement would be produced in the light of the above discussion. The statement would highlight the need for research as uncertainties were apparent, and would also recommend types of research which would help to eliminate such uncertainties. It was agreed that a revised statement would be circulated and discussed at the June meeting of COMEAP. 48. It was agreed that, ideally, a new prospective cohort study would be more appropriate than a retrospective study using an existing cohort of children. Members also noted that the use of swimming pools by babies might require some consideration. 49. Members were told that the advice to maintenance staff at swimming pools was likely to be adequate though excessive use of chlorination products is not uncommon. This highlighted the requirement for establishing and maintaining good conditions without over chlorinating swimming pools. ITEM 5c. GAUDERMAN ET AL 2007. EFFECT OF EXPOSURE TO TRAFFIC ON LUNG DEVELOPMENT FROM 10 - 18 YEARS OF AGE: A COHORT STUDY (COMEAP/2007/03) 50. The Chairman reminded Members of an earlier paper from Gauderman et al reporting results from the Southern Californian Children's Health Study. The most recent paper from this study was published in January 2007 and attracted much press attention. 51. The study reported an effect on lung development (reflected by indices of lung function, FEV1 and MMEF) from living close to motorways and from exposure to 'regional air pollution'. 52. Members were asked for their comments. 53. One member was particularly concerned with how the study related emissions from roads to the effects on lung development. The member questioned how confounders of the effect of distance from roads were taken into account. It was noted that Table 1 lacked correlation in the results between 'distance from freeway' (significant associations with lung development) and 'model based pollution from freeway' (no significant associations with lung development). This might suggest that a factor other than air pollution that was related to proximity to roads might be involved. Another Member noted, however, that there was an increasing deficit in lung function growth with increasing quartiles of 'model-based pollution from freeway' (second set of results in Table 1) even if the results for each quartile were not statistically significant. Trends had not been formally assessed in Table 1 but some might be nominally statistically significant. 54. Even though the effects were not statistically significant, deficits in lung function growth were seen at distances from the freeway of >1000m. In the UK, air pollution returns to background levels 1-200m from roads so this possible effect at >1000m was puzzling if air pollution was proposed as the responsible factor. It was noted that due to ultrafine particulate matter 'falling away' with distance from roads more than larger PM metrics, it was unlikely that an ultrafine effect was occurring. Members were not clear as to what the gradient was representing in Table 1 and Members thought that there might be further data which had not been published in the actual paper which might provide further clarification. 55. Another member was also concerned that the paper did not clearly indicate whether it had accounted for all confounding by social factors. Adjustment for socio-economic status was not mentioned in the penultimate paragraph of the statistical methods section of the paper. It was noted that socio-economic status had been accounted for in the sensitivity analysis results but these were not as would be expected if residual confounding by socio-economic status had been occurring in the base model. (Some loss of signal would have been expected with residual confounding but the size and statistical significance of the deficit in lung growth increased with addition of socio-economic status to the model (Table 2)). 56. Another member provided further information on lung development of the child, noting that a child from a low socioeconomic background was likely to be small. The range in lung size is greater at 18 years of age than at 11 years. Absolute growth is less if a child is small (rather than large) and the results in the tables were given in ml i.e. an absolute measure of growth. It was acknowledged that corrections for height and other measurements of stature had been taken, but it was noted that possibly a better measure was percentage change in lung growth. This discussion led to members to question how the authors overcame the fact that lung growth is not linear and whether height was only measured at the beginning and after 8 years. 57. It was discussed whether the findings in this paper were transferable to the UK. Members remembered previous discussions which had taken place on transferability of findings in the US to the UK and noted again that if PM was the pollutant causing the effect, then the data should be transferable. There were however higher concentrations of nitric acid in California compared with the UK. Ozone concentrations were also higher in California but it was noted that in Table 3, on the effects of regional air quality, an effect of ozone was not found. Members were reminded that NO2 was the most transferable of all the air pollutants. 58. Members agreed that for a number of reasons, it was necessary to look at other work by the same authors, as well as studies which also show reversibility of the effect in children moving away from polluted areas. It was confirmed that toxicology studies had been carried out in newborn rats which would provide further insight into mechanisms behind the results of such epidemiological studies. These further considerations would be taken forward in the context of the work on morbidity by the Quantification sub-group. 59. It was agreed that a letter would be written to the Department of
Health on this paper noting the need to put its findings in a wider context.
Duncan Laxen volunteered to look at the web appendices to the paper and
report back to the Secretariat.
60. Members' attention was drawn to this paper which reported an association between air pollution levels and the occurrence of cardiovascular disease in a cohort of 65,893 women in the USA, followed up between 1994 and 1998. 61. The paper reported a clear association between air pollution levels close to the individual's place of residence and the risk of illness. An editorial on the paper had been tabled. 62. It was noted that a statement on the paper would not be required, but that the paper was an important one to consider. 63. The Chairman drew Members' attention to a surprisingly large effect that was noted: each 10µg/m3 increase in PM2.5 was associated with a 76% increase in the risk of illness - an increase which was statistically significant. 64. Members discussed why the effect size might be so large and questioned
the possibility of the effect size increasing further if the study focussed
on an increasingly smaller group. This encouraged Members to discuss the
group which had been studied and in particular, to note that this cohort
were not taking drugs for cardiovascular conditions. It was questioned
whether members of the cohort were taking HRT as this is believed to protect
against coronary artery disease, but possibly increase thrombosis. It
was agreed that the Chairman would write to the paper author to question
whether a therapeutic effect might be occurring i.e. general population
studies including treated patients with cardiovascular disease were giving
smaller results than this study which excluded those treated for cardiovascular
disease (at least at the start of the study) . 65. It was agreed that the Secretariat would contact Philip Poole-Wilson, who had sent his apologies for this meeting, to establish further information on the Euro-heart study and the proposed theory regarding increased risk for women as referred to in the tabled editorial. Action: Secretariat 66. Other indicators of health of the cohort were questioned, such as BMI. It was noted that half of the cohort had a BMI greater than 27.3 and it was questioned whether this was considered 'healthy'. 67. It was established that there were many caveats which could explain the size of the coefficient, other than spatial scale, and in particular, the different types of confounding associated with between city studies and within city studies. The ACS study controlled for spatial autocorrelation, but this study appeared not to do this to such a rigorous degree. It was agreed that a greater degree of controlling for spatial autocorrelation would influence the size of the effect estimate. It was suggested that a list of reasons suggesting why the coefficient was so large should be produced, which would include susceptible group, measurement error, better design / control of confounding etc and that this list would be ranked in terms of likelihood. Action: Secretariat 68. The Committee agreed that this was a most striking paper.
69. The Secretariat had supplied members with papers which may have been of interest. The Chairman provided members with the opportunity to raise interesting points which members had regarding these papers, although no formal comment was required. 70. A paper by Yamasaki et al (2007) on stroke encouraged members to agree to keep air pollution and stroke on the list of subjects the Committee should consider at some point in the future. 71. The paper suggested that deaths were occurring as a direct result of having a stroke. Members agreed that death as a direct result of a stroke was unlikely as stroke victims usually die from something other than the actual event of a stroke (although it was agreed that the stroke may have precipitated the event which caused death). Members questioned whether it was plausible that an elderly person with high blood pressure might suffer from a stroke due to exposure to air pollution one hour prior to the stroke taking place? Members noted that 1 hour was too short for an inflammatory mechanism to be involved. 72. Members noticed that the data in this paper did not cohere with the data from the Committee's report 'Health Effects of Air Pollutants on Cardiovascular Disease'. 73. It was agreed that with more work in this area being likely, that the Committee required an additional cardiovascular specialist to support the role of Philip Poole-Wilson. It was agreed that a cardiovascular physician would be most appropriate. Action: Secretariat ITEM 7. DATES OF NEXT MEETINGS 74. 1st and 2nd May 2007 - Annual Air Pollution Meeting, Cranfield University. 75. 22nd June 2007 at HPA, Chilton. Secretariat
|
|
|
|||||||||||||||||||||||||||||
You
are here: 
