Understanding Premature Deaths from NOx
Sunday, July 23, 2017

.....
To arrive at this time figure of premature loss of life of between half-a-day and 3.5 hours directly attributable to NOx pollution all I’ve done here is look at the tables in the EEA report and check the numbers. The 40,000 figure that’s now widely reported, broadly unquestioned, across the media is because of use of the word premature and if we’re not careful this misunderstanding of the actual facts of NOx pollution will cost us trillions in transport policy and legislation changes and improve our life expectancy by only a relatively tiny amount of time. We absolutely need to clean our air but we need to apply real numbers and real science to this debate. We’ve taken a badly phrased headline figure from an official report and allowed it to terrify millions of consumers. If we can’t even accurately understand the numbers what hope have we to really improve our air quality? Bad science won’t help us clean our air.

Quentin Willson
Lead Spokesman for FairFuelUK and TV Motoring Journalist and Broadcaster

http://www.fairmotoring.com/index.php?entry_id=1500786900

Press release

Plan for roadside NO2 concentrations published

From:Department for Environment, Food & Rural Affairs, Department for Transport, Thérèse Coffey MP, and The Rt Hon Michael Gove MP Part of:Environmental quality Published:26 July 2017Last updated: 26 July 2017, see all updates

Plan includes an end to the sale of all new conventional petrol and diesel cars and vans by 2040 and a new Clean Air Fund

https://www.gov.uk/government/news/plan-for-roadside-no2-concentrations-published

Stewgreen re: the impacts of NO2 on mortality - below is the latest thinking from the UK's COMEAP Committee (link ref at end)

This was published after the Speigelhalter piece you linked. AFAIK the full COMEAP report on NO2 has not yet been published:

//
UK Plan for tackling roadside nitrogen dioxide concentrations
Technical report
July 2017
Annex A – Refined COMEAP recommendations letter

c/o COMEAP Secretariat
Air Quality and Public Health Group
Public Health England
Chilton
Didcot
Oxfordshire
Dr Thérèse Coffey MP
Parliamentary Under Secretary of State for the Environment and Rural Life Opportunities
Department for Environment, Food & Rural Affairs
Nobel House
17 Smith Square
London
SW1P 3JR

14 July 2017

Dear Dr Coffey

Refined COMEAP recommendations for quantifying mortality effects on the basis of long-term average concentrations of nitrogen dioxide (NO2)

This letter summarises current thinking of the Committee on the Medical Effects of Air Pollutants (COMEAP) on the association between long-term average concentrations of nitrogen dioxide (NO2) and mortality risk. We previously (July 2015) provided interim recommendations to assist Defra when developing plans to improve air quality. This letter presents updated recommendations that arose from discussion at the COMEAP meeting held on 24 February 2017.

Summary
1. Population-based studies following people’s health over several years show statistical associations between higher long-term average concentrations of ambient NO2 where people live and increased mortality risk. It is likely that some of this effect is due to NO2 itself. However, as other co-emitted pollutants, e.g. from traffic, are also high in the same places, these could also be responsible to some extent. In our view, the available evidence and methods do not allow us to make a reliable assessment of the size of the effect which is attributable to NO2 itself.

2. We therefore recommend two different approaches for assessing the mortality benefits of interventions intended to reduce NOx emissions from traffic:
• For interventions which reduce all traffic-related air pollutants, use the statistical association obtained from population studies. In this case, NO2 is regarded as acting as a marker for the effects of the traffic pollutant mixture overall, including NO2.
• For interventions which primarily target emissions of NOx, use 25-55% of the statistical association obtained from population studies. This is, in our judgement, the likely extent to which this association represents effects causally related to NO2. This is more uncertain than assessing traffic pollutants as a mixture.

Background

3. In our March 2015 Statement on the evidence for the effects of nitrogen dioxide on health (81) we noted the strengthening evidence linking NO2 with health effects. We concluded that:

81 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/411756/COMEAP_The _evidence_for_the_effects_of_nitrogen_dioxide.pdf

i. Evidence of associations of ambient concentrations of NO2 with a range of effects on health has strengthened in recent years. These associations have been shown to be robust to adjustment for other pollutants including some particle metrics.
ii. Although it is possible that, to some extent, NO2 acts as a marker of the effects of other traffic-related pollutants, the epidemiological and mechanistic evidence now suggests that it would be sensible to regard NO2 as causing some of the health impact found to be associated with it in epidemiological studies.

4. At that stage, we did not draw conclusions on specific health outcomes nor look in detail at the methodological issues relevant to quantification of effects associated with ambient NO2.

5. We were subsequently asked to propose approaches to quantifying mortality associated with long-term average concentrations of NO2. This was primarily needed to provide Defra with a method for assessing the potential mortality benefits of measures to reduce NO2 concentrations, to assist with the development of plans to improve air quality. In July 2015, a COMEAP working group provided interim recommendations to your predecessor as Parliamentary Under Secretary, Mr Rory Stewart. As well as recommending an interim recommendation for a coefficient, the letter explained that:
“…there is uncertainty in the extent to which the association between long-term average concentrations of NO2 and mortality is causal. It is likely that some of the effect is due to NO2, but other co-emitted pollutants could also be responsible to some extent. Therefore, the uncertainty in applying a coefficient to assess the health benefit of measures to reducing NO2 will depend on the extent to which the measure is specific to NO2, or also reduces concentrations of other coemitted pollutants. There is likely to be more uncertainty when the measure is specific for a reduction in NO2, compared to when an intervention aims to reduce the whole mixture of air pollutants.”

6. In our Interim statement on quantifying the association of long-term average concentrations of nitrogen dioxide and mortality published in December 2015 (82) , we

82 https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/485373/COMEAP_NO 2_Mortality_Interim_Statement.pdf

explained the additional work that we were undertaking to refine our recommendations. These included carrying out a systematic review and meta-analysis of epidemiological studies of long-term average concentrations of NO2 and all-cause mortality. We noted that, in interpreting these, a number of scientific and methodological challenges needed to be considered, including the extent of independence of the associations of mortality with NO2 and PM2.5. We also noted the uncertainty in the extent to which the association between long-term average concentrations of NO2 and mortality was causal.

Recommendations arising from COMEAP meeting held on 24 February 2017

7. We have discussed these complex scientific and statistical issues on several occasions since the publication of our interim statement. The Committee has not been able to come to a consensus view on how the epidemiological associations between NO2 and mortality can be used to either predict the benefits of interventions to improve air quality or to estimate the current mortality burden imposed on the UK population by air pollution.

8. Some Members are doubtful that the evidence is sufficient to allow a robust recommendation for quantification to be made. This is particularly the case for effects likely to be caused by NO2 itself. Others think it important to make an estimate of the possible mortality benefit from reducing NO2 concentrations. They note that to recommend against undertaking quantification would have the same consequence, for policy development, as assuming that there would be no mortality benefit, which they do not consider to be likely.

9. We last discussed this issue as a Committee at a meeting held on 24 February 2017. We considered possible uses of the epidemiological associations (coefficients) between long-term average concentrations of NO2 and mortality effects. For assessment of the benefits (impacts) of interventions to reduce emissions, we discussed the use of these coefficients to:
a. Predict the benefits of interventions which reduce all traffic-related air pollutants
b. Predict the benefits of interventions which primarily target emissions of nitrogen oxides (NOx)

10. For both these types of intervention, we also discussed (c) assessing the benefits associated with reductions in secondary nitrate concentrations arising from reductions in NOx emissions.

11. We also discussed
d. Estimating the mortality burden attributable to current concentrations of air pollutants

a. Predicting the benefits of interventions which reduce all traffic-related air pollutants

12. We have derived a summary coefficient linking long-term average concentrations of NO2 with all-cause mortality, by undertaking a meta-analysis of associations reported in cohort studies. We used associations from single-pollutant models (ie with no attempt to adjust for effects associated with correlated pollutants) in this analysis. This summary 82
coefficient reflects the effect of NO2 and also, to some extent, other pollutants with which NO2 is correlated. These include PM2.5, other fractions of particulate matter (PM), and other components of the air pollution mixture. In particular, associations with NO2 are likely to reflect the effects of the mixture of traffic-related pollutants.

13. Interventions that would reduce traffic movements, or remove traffic altogether, would reduce the whole mixture of traffic-related pollutants. Some other interventions, such as replacing Euro 3/III vehicles by Euro 6/VI, would also reduce emissions of a number of other potentially causal pollutants/metrics (eg volatile organic compounds, aldehydes, organic compounds bound to primary PM) as well as reducing NOx emissions.

14. We recommend that the summary unadjusted NO2 coefficient of 1.023 (95% CI: 1.008, 1.037) per 10 μg/m3 annual average NO2 is used to estimate the effect on mortality of reductions in the whole pollution mixture.

15. Furthermore: as these measures will also reduce PM concentrations, an alternative calculation of benefits associated with this reduction, using an unadjusted PM2.5 coefficient can also be undertaken. Discussion at the meeting held on 7 June 2017 confirmed that our recommendation is to use a coefficient of 1.06 (95%CI: 1.04 - 1.08) per 10μg/m3 annual average PM2.5 derived from a meta-analysis of single pollutant studies (Hoek et al, 2013) (83). As either of these calculations is likely to underestimate the likely benefits of interventions, the higher of the two values calculated from these two approaches can be used as the most appropriate estimate of the predicted benefits.

83 Hoek, G., Krishnan, R. M., Beelen, R., Peters, A., Ostro, B., Brunekreef, B. & Kaufman, J. D. 2013.
Long-term air pollution exposure and cardio- respiratory mortality: a review. Environ Health, 12, 43.

b. Predicting the benefits of interventions which primarily target emissions of NOx

16. Some interventions are primarily targeted at reducing NOx emissions, and would have little impact on emissions of other traffic-related pollutants. Using the unadjusted coefficient to predict the mortality benefits of these interventions would produce an over-estimate.

17. We have discussed whether it would be possible to use epidemiological associations for NO2 reported from two-pollutant models (with PM) to refine the summary coefficient. These coefficients would be adjusted, as far as possible, for effects associated with PM, especially PM2.5. In our view, the available evidence and methods do not provide a satisfactory basis on which to reliably propose an adjusted coefficient. It should be noted that even a coefficient adjusted for effects more closely associated with PM2.5 concentrations than with NO2 would not reliably reflect the size of the causal effect of NO2 itself: the adjusted coefficient would also reflect effects of other pollutants which are more closely spatially correlated with NO2 than with PM2.5, such as ultrafine particles, primary combustion particles, volatile organic compounds etc.

18. The majority of Members therefore considered it preferable to use expert judgement to make a recommendation as to how the benefits of interventions that primarily target NOx could be estimated. We considered it likely that the effect of NO2, itself, on mortality was likely to be in the range of 25 – 55 % (mid-point of range 40%) of the unadjusted coefficient of 1.023 (95% CI: 1.008, 1.037) per 10μg/m3 annual average NO2, and recommend that this be used in assessments of interventions that primarily target NOx emissions. This is equivalent to reducing the unadjusted coefficient by 20% (an approximate adjustment for effects associated with PM2.5 concentrations, based on two-pollutant models) and applying expert judgement, inferred from other types of evidence, suggesting that 30-70% of this adjusted coefficient may be caused by NO2 itself, rather than other correlated (e.g. co-emitted) pollutants.

c. Assessment of effects associated with secondary nitrate

19. For both types of intervention, we consider it appropriate to, additionally, assess mortality benefits associated with reductions in secondary nitrate concentrations arising from the reductions in NOx emissions. Because secondary nitrate concentrations occur some distance from the source of NOx emissions, effects associated with them would not be represented by the NO2 coefficient.

20. We recommend using the unadjusted coefficient 1.06 (95%CI: 1.04 - 1.08) per 10μg/m3 annual average PM2.5.

d. Estimating the mortality burden attributable to current concentrations of air pollutants

21. We do not think it appropriate to try to estimate the mortality burden attributable to current concentrations of NO2 alone. In numerical terms, the same coefficient could be applied to impact or burden calculations. However, several Members felt that there were differences in terms of how the results are used. Burden estimates may include estimation of effects at low concentrations (typically, impact estimates do not) where there is a lack of certainty whether NO2 increases mortality and/or over the shape of the concentration-response relationship. Impact calculations typically involve comparisons across policies. The uncertainties may not affect the relative comparisons whereas burden has a more absolute status. Burden calculations may be publicised in the media without the associated uncertainties. Finally, the main interest is in the overall burden associated with air pollution as a whole, as a general impetus for action, rather than the effects of particular pollutants.

22. Some Members do not think it appropriate to try to calculate an overall burden of the mortality associated with the air pollution mixture. Others are of the view that an attempt can be made based on associations with NO2 and PM2.5, and using information from two-pollutant models. This could be presented as a range of central estimates, but methods to represent the full statistical uncertainty are unlikely to be available.
Provision of these recommendations to Defra

23. Summaries of these draft recommendations were provided to Defra officials following the COMEAP meeting held on 24 February 2017. It was noted that they were subject to confirmation by the Committee, and that a number of caveats would need to be borne in mind when any calculations were undertaken.

Next steps

24. These recommendations remain draft until they are formally signed off by Members during finalisation of our report. We are currently working to develop a version of the report which will present the recommendations agreed by the majority of Members. It will also reflect the full range of contrasting views held across the Committee.

We hope these draft recommendations are useful for your revised cost-benefit analyses of measures to reduce NO2 concentrations.

Yours sincerely

Professor Frank Kelly, COMEAP Chair

Professor Roy Harrison, COMEAP NO2 Working Group Chair

Dr Heather Walton, COMEAP's subgroup on Quantification of Air Pollution Risks Chair

CC: Mr Jesse Norman MP, Parliamentary Under Secretary of State for Roads, Local
Transport and Devolution
//

Source - Annex A:
https://www.gov.uk/government/uploads/system/uploads/attachment_data/file/632916/air-quality-plan-technical-report.pdf