The Alliance thanks Professor Kirk Smith for the following breakdown of the HAP-related findings from the Global Burden of Disease 2010 study. The Alliance's response to the study can be found here.

The papers describing the new Global Burden of Disease have just been published in The Lancet. See The last of the seven major papers reports the comparative risk assessment among about 60 risk factors, including HAP. It can be cited as

Lim S.S and many others, 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, Lancet, 380: 2224-60.

As chair of the skilled and dedicated expert group on HAP, I here summarize four aspects of the assessment related to HAP for your possible interest:

First, the numbers. The scale of HAP's health impact is quite large.

  • 3.5 million direct premature deaths annually in 2010 – compared to 3.3 for outdoor air pollution (particles and ozone) and 6.3 for active and passive smoking
  • 0.5 million more from the outdoor air pollution due to household fuels -- what one might call "secondhand cookfire smoke".
  • This makes ~4 million total attributable to HAP from cooking fuel
  • 0.5 million of the total are child pneumonia deaths
  • Rest are adult deaths (men and women) from lung cancer, cardiovascular disease (CVD), and COPD. Cataracts are also included, but they cause very few deaths.
  • In terms of absolute impacts, men are more affected than women. This counterintuitive result is because men have so much higher background rates of the major diseases. Thus, although women have higher exposures and higher elevations in risk for these diseases, men end up with the larger burden. In relative terms, as indicated below, however, women are more affected by HAP.
  • In terms of DALYs (lost healthy life years), HAP is the 2nd most important risk factor for women and girls globally among those examined 5th for men and boys. 1st for both sexes in South Asia and for women and girls in most of Sub-Saharan Africa. 6th for both sexes in East Asia.
  • Most important single environmental risk factor globally and in poor regions. Behind outdoor air pollution (OAP) in richer countries, of course, and in China, where OAP ranks fourth among all risk factors examined.
  • Note that the portion of OAP ill-health due to secondhand smoke from cookfires (~500k deaths) is not included in the summary figures in the article, however, but only in the text. When we have access to the raw numbers so we can make this adjustment ourselves, the HAP ranking may change in a few regions. In particular, it may move ahead of alcohol and OAP in East Asia.
  • Given the uncertainties involved, of course, one should not make a big deal of minor differences. Suffice to say OAP and HAP are about equal in East Asia.
  • 2.8 billion people rely on solid fuels for their main cooking fuel in 2010, a number that seems to have been roughly stable globally for the last decade or so.  This is more people than anytime in human history relying on solid fuels for cooking.

Second, the bottom lines:

  • Most of the impact is now understood to occur in adults
  • But still ALRI (pneumonia)) deaths in children are significant -- 500k globally -- even with much lower background ALRI death rate globally
  • Still few data about school age children -- thus, no estimates possible -- but they have low death rates in general
  • Household cooking fuel emissions contribute substantially to outdoor air pollution in many countries -- 25-30% in India, for example -- "secondhand cookfire smoke."
  • The new integrated exposure-response (IER) analyses provide excellent new cross-risk-factor validation of the effects we find for a range of diseases -- lung cancer, COPD, ALRI (IERs link epi evidence across the four particle categories - outdoor air pollution, secondhand tobacco smoke, HAP, and active smoking)
  • The IERs provide a way to estimate effects for cardiovascular disease from HAP
  • But we still need direct evidence to verify CVD itself from HAP, although we do have evidence of associated CVD risk factors (blood pressure, EKG, etc)
  • Large role HAP plays in non-communicable diseases (NCDs) which are now understood to be so important in India and other low and mid-income countries.
  • Not all diseases for which we have evidence were included as the data are still incomplete -- growing databases, however, on TB, adult ALRI, cervical cancer, LBW, cognitive effects, etc.
  • It is important to remember that interventions that do not reduce all the way to the clean levels used for comparison (see next section) will not produce the full health benefits.  Based on studies across combustion particle sources in the GBD, it is now believed that the impacts are highly non-linear at the levels commonly experienced in households cooking with solid fuels.  Thus, for example, a reduction of a factor of two in smoke exposures at these high levels will produce far less than a reduction of 50% in health impacts.  Need to reduce exposures down to levels in the range of the WHO Air Quality Guidelines to do so.
  • Perhaps the most striking result overall in the report is the very low burden from poor water and poor sanitation (~300k premature deaths for the world for both risks). I urge you to read those sections in the paper if you are interested to know why.

Third, many will wonder why the results show so much larger effects for HAP compared to what was found in the previous CRA for 2000 published in 2004 (1.6 million premature deaths)

There were actually factors that tended to decrease the attributable burden in the new estimates

  • ~40% of the world use solid fuels for cooking in 2010, down from the ~50% estimated for 2000. This is based on much more robust estimate of global household solid fuel use for cooking informed by ~600 nationally representative household fuel use surveys compared to ~50 last time.
  • We were able to be more careful in separating out cooking fuel only -- much less confusion with heating fuels than before
  • The risks for COPD were less than before, based on the new systematic reviews and meta-analyses.
  • The background child ALRI in the world is much less, leading to less impact from HAP (500k deaths instead of 1000k before)
  • There were other differences, however, that led to higher estimates:
  • Evidence now allows us to add estimates for two new categories of disease: cardiovascular disease (CVD) and cataracts.
  • As CVD is the chief cause of mortality in nearly all countries, the additional burden is substantial
  • Evidence of effects in men now for all adult endpoints except cataracts
  • We now using a low counterfactual (~7 ug/m3 annual PM2.5 -- same as the OAP group used) -- equivalent to cooking with gas.
  • This allows the HAP results to reflect the full benefit that could be expected from moving the 40% of households with solid fuel to the low pollution experienced by 60% of world population using gas or electricity for cooking.

The net result was a major increase in the estimated burden.

Fourth, although a major ground breaking effort in global health that will have major influences on policy discussions, there are issues about the overall GBD effort itself, unfortunately, that detract from our ability to put the results into full policy context:

  • Unlike the previous WHO version published in 2004 and making estimates for 2000, the management of this new version did not include any representation from the WHO in the end. It now should be considered an international effort organized by Institute of Health Metrics and Evaluation of the University of Washington.  I urge you to read the commentary in the same Lancet issue from Dr. Margaret Chan, Director General of WHO, which explains the reasons they withdrew.
  • The choice of risk factors (RFs) to examine is incomplete. There are none that relate to malaria, HIV, child vaccinatable diseases, or road traffic accidents, for example, which are huge burdens and interventions for which should be considered for any health policy agenda.  Thus, it is misleading to imply that the risk factor list is complete for policy.   See the excellent commentary by Charlotte Watts and Sandy Cairncross in the same issue, which goes into these issues more.
  • This is why I try to say something to the effect that "Among the risk factors examined, HAP is the second most important globally for women and girls"
  • There are also some rather strange RFs that were included, in particular lack of fruit is the 5th RF in the world among those examined and lack of vegetables, lack of nuts, and lack of whole grains are each in the top 20 RFs. It is difficult to understand exactly how to think of these type of ("lack of") risk factors in comparison to air pollution, high blood pressure, etc.   And why “lack of bednets” and ”lack of condoms”, for example, were not also included.
  • There are also some strange aggregations or lack of them. Why are poor water and sanitation not combined in the major figures in the paper, as they were last time? And why are all the occupational risks presented separately? Combining would make both WASH and occupational health appear more important.
  • Figure 6 in the article (page 2250 – it is not labeled)  is incorrect and worse,  misleading – as can be seen by the vertical axis, which implies that more than 100% of ill-health can be averted by these interventions in some regions.  It is scientifically incorrect to add attributable risks in this way.  Unfortunately, this was caught too late in the publication process to change.
  • Some of these things we can fix in our own presentations after we have access to the raw results, but not at present.
  • All these issues are reasons not to put too much emphasis on the rankings for the time being and focus on the HAP results themselves in the short term. These are what we had control over and can have some confidence in.


Kirk R. Smith, MPH, PhD
2012 Tyler Laureate
Professor of Global Environmental Health
Director of the Global Health and Environment Program
School of Public Health
747 University Hall
University of California
Berkeley, California, 94720-7360
510-643-0793; fax 642-5815