January 2003, pages 4 - 7 Despite improvements in air quality over the last few decades, air pollution continues to seriously affect people's health. A study in the Bristol area argues that further research is needed to measure the long-term impacts

Since the 1950s, a significant body of evidence has accumulated showing that air pollution has a damaging effect on health and is associated with increases in outpatient visits due to respiratory and cardiovascular diseases, in hospital admissions and in daily mortality.¹

Ozone is a powerful oxidant which causes inflammation of the respiratory tract and exacerbates existing lung disease. Particulates, particularly fine particles of diameter less than 10µ (PM10) which can penetrate deep into the lungs, are linked with heart and lung disease. Volatile organic compounds (VOCs) (such as benzene and 1,3-butadiene) originate predominantly from traffic exhausts and many are carcinogenic, irritant or toxic, although this is dependent on the specific compound. In the UK, only benzene and 1,3-butadiene are currently classified as particularly harmful in the National Air Quality Strategy (NAQS) because of their status as carcinogens.²

Although most data on the toxicity of VOCs stems from very high dose studies, low background levels may still affect human health. Studies of indoor air pollution and "sick building syndrome" have found connections between the mixtures of VOCs present in these environments and adverse health effects such as headaches and sensory irritation.³ It should be noted that many compounds classed as pollutants do occur naturally in the atmosphere and our cardio-respiratory systems have evolved to cope with these background levels. However, any chemical whose level is elevated above normal may be implicated in causing an adverse health effect.

The complex mix of chemicals present in urban air makes it difficult to attribute specific health effects to any one compound and they may stem from cumulative or synergistic effects.⁴ It is possible that the cocktail of VOCs present in the urban atmosphere may be adding to the health impact of the criteria pollutants such as ozone and PM10, hence the inclusion of benzene, 1,3-butadiene and total volatile organic compound concentration (TVOC) in this study.

Air pollution and health

Studies of air pollution and health in the UK have mostly focused on London, although some have also been carried out in other large cities such as Birmingham and Edinburgh. The main associations found have been between particulate matter and all-cause mortality, including specific cardiovascular and respiratory mortality. The APHEA project (Air pollution and health: a European approach) has attempted to provide quantitative estimates of the acute effects of air pollution on health in 11 cities across Europe.

The study found that particulates, sulphur dioxide, nitrogen dioxide and ozone were all associated with increased all-cause or cause-specific mortality to a greater or lesser extent.⁵ A report commissioned by the Department of Health and produced by the Committee on the Medical Effects of Air Pollutants (COMEAP) estimated that PM10 and SO2 were responsible for the advancement of over 11,000 deaths in Great Britain every year.⁶

Despite the large number of studies that have been carried out over the years it is hard to find repeatable results which can be automatically applied to many situations. Even in the same country, no cities have exactly the same pollutant composition or population with respect to age distribution, social status, underlying health and exposure to other risks such as climate or weather. It is therefore difficult to extrapolate findings from a study in one area to a different place. A study using data from Bristol must be undertaken to ascertain whether current levels of air pollution are having a significant effect on the population of the city.

In Bristol, Sarah McMahon and colleagues working at the University of the West of England (UWE) and Bristol City Council have undertaken several projects to quantify the effects of air pollution in Bristol on the local population. The first study examined the effect of particulate concentrations on emergency hospital admission for respiratory disease between December 1992 and June 1993.⁷ Positive correlations were found in both the summer and winter.

Further studies specifically investigated the effects on children of primary school age who suffered from asthma. The use of ventolin inhalers (those used for the relief of symptoms) was found to be associated with fine particles and NO2. These effects were seen in children across the region, responding to the same high pollution episodes whose effects could be felt across the entire study area.^8,9

Analysing data

The Avon Health Authority (AHA) provided information on daily cardiovascular mortality, respiratory mortality, all cause mortality and population in the Bristol area (BS1 to BS40 inclusive) for the period of January 1996 to December 1999. Causes of death were classified according to the International Classification of Diseases (Revision 9). Circulatory and cardiovascular diseases are classified as ICD-9 codes 390-459 and respiratory diseases as 460-519. The data was broken down into four age groups: under 17, 17-44, 45-64 and over 65. Pollution data was obtained from the National Air Quality Information Archive for the two continuous air quality monitors in the city centre which are part of the National Automatic Monitoring Network and operated by the National Environmental Technology Centre (NETCEN). Weather data was provided by the British Atmospheric Data Centre for a weather station located in central Bristol.

The STATA statistics package was used to analyse the data set with a multiple linear regression model. A multi-factor analysis can begin to account for confounding issues, such as the fact that all of the pollutants correlate with each other and with the weather, as well as potentially with the health outcomes. The daily health outcome data was compared with daily average levels of ozone, PM10, benzene, 1,3-butadiene and TVOC. Confounding by weather was accounted for by the inclusion of average daily temperature and wind speed in the model. Lags of up to five days were applied to the data and the results were tested at the 95 per cent confidence level.

Statistical correlations

A number of statistically significant correlations were found in the data for Bristol. The associations were generally found in the oldest age group (over 65).

Weather: temperature showed a consistently strong negative correlation with mortality, especially in the older age groups (ages 45-65 and over 65). Wind speed exhibited no significant correlations.

The medical and epidemiological community has long known of the link between weather and adverse health outcomes. In this country, the greatest effect of temperature is during winter when the number of cases of influenza and other diseases of the respiratory system rise. The spread of these infectious illnesses is exacerbated during colder weather. It is encouraging that this expected result was consistently found in the analyses, as it demonstrates that effects can be observed in the data using this statistical method.

Particulate matter: levels of PM10 were consistently associated with increases in all-cause and cause specific mortality in the over 65 age group. These connections were found for each lag period, although the strongest association was on the same day. The effect was strongest for cardiovascular deaths, less strong for respiratory deaths, and somewhere in between for all-cause mortality. A similar order of effect for PM10 was also found in the APHEA study of Western European cities.⁵ A quantifiable effect of PM10 on respiratory disease in Bristol was also found by McMahon,⁷ albeit in daily admissions, but it would be expected that incidences of admissions and mortality are closely linked.

Ozone: ozone showed a positive correlation with cardiovascular mortality in the over 65 age group for a one-day lag. Ozone has found to be associated specifically with cardiovascular disease in several other UK studies as well as in the APHEA project. ^10,11,5 In the APHEA project however, ozone was more strongly correlated with all cause and respiratory mortality, effects which were not found in this analysis.

TVOC: concentrations of TVOC showed a positive association with next-day respiratory deaths in the over 65 age group. Very few studies of outdoor air pollution have investigated the influence of VOCs on health but those that have done so have found VOC concentrations to be associated with respiratory illnesses. One study of school children in West Virginia found a positive correlation between chronic respiratory symptoms and VOC concentrations.¹²

A study at a London accident and emergency department found that levels of benzene, propane and isoprene had a significant positive linear relationship with attendances of very young children for acute wheeze.¹³ In a Norwegian study benzene was found to be more strongly associated with emergency hospital admissions for respiratory disease than PM10.14 In the Bristol data the correlation coefficients for TVOC were very similar to those for PM10, suggesting a similar magnitude of effect.

Limitations of methods

The limitations of the methods used in this study must be borne in mind when deciding on the significance of the results. It must be remembered that the numbers of deaths and hospital admissions that are being dealt with are very low. The average number of deaths (from all causes) in the Bristol area is 23 a day, in a total population of almost 900,000. Around 45 per cent of these deaths are from cardiovascular disease and 15 per cent from respiratory illness.

Identifying and quantifying small fluctuations in a low number of incidences is inevitably hard. The statistical method used - multiple linear regression - assumes that the relationship between the variables is linear. The complex interplay of pollutants and weather means this is not always the case. Other factors which must be considered include the spatial variation of pollutant concentrations across a large region and confounding issues such as bed availability. Nevertheless, this analysis has produced similar results to those found by more complex investigations.

Levels of pollutants in the Bristol area are not generally considered dangerous to human health according to current government standards. The Government's air quality index is used to describe the severity of air pollution using a banding system.² During the entire period of this study, air pollution in Bristol was generally "low" and occasionally "moderate", never "high" or "very high" according to the index. Despite this, an association has still been identified between daily concentrations of pollutants, notably PM10, and mortality.

These effects were overwhelmingly seen in people aged over 65. This is unsurprising, given the increased vulnerability of this age group. Cold weather and "flu" epidemics are known to hasten mortality in the sick and elderly and the COMEAP report concluded that it was likely that air pollutants could act in a similar manner, bringing forward death by a few weeks or days.⁶

If air pollution is advancing mortality then it could also be expected that it is adversely affecting hospital admissions, attendances at hospitals and doctors surgeries and the occurrence of symptoms of respiratory or cardiovascular disease. Further research could elucidate the effect of air pollution on the quality of life of the wider population and its impact on economic factors such as numbers of sick days or early retirements. The chronic, long-term impact of poor air quality on human health is a subject which is only just beginning to be explored. This is potentially more important in public health terms than the acute (short-term) effects. Two American studies have indicated that life expectancy may be two or three years shorter in communities with high particulate levels compared with those communities which experience lower levels.^15,16

The linear regression models used in this study have identified an association between daily levels of pollutants (particularly PM10) and increased mortality in the Bristol area. These findings are in keeping with the results of similar studies from other cities in the UK and indicate that further investigation into the acute and chronic effects of poor air quality is warranted.

Alison Rivett, Dudley Shallcross and Ruth Wood, are from the School of Chemistry at University of Bristol, and Ben Wheeler and David Gunnell are from the Department of Social Medicine. For further details contact Alison Rivett by e-mail: a.c.rivett@bristol.ac.uk

Thanks to Sarah McMahon at Bristol City Council for many useful discussions and information on this topic. Thanks also to David Prothero and Angela Chung at the Avon Health Authority for making available the health data.

References

1. WHO (2000). "Guidelines for air quality", World Health Organisation, Geneva.
2. DEFRA (2002). National Air Quality Strategy (NAQS) environmental protection website. Department for Environment, Food and Rural Affairs, London: www.defra.gov.uk/environment/airquality/strategy/index.htm
3. Jones A P, (1999). "Indoor air quality and health". Atmospheric Environment, 33, 4535-4564.
4. Leikauf G D, (2002). Hazardous air pollutants and asthma. Environmental Health Perspectives, 110, 505-526.
5. APHEA (2002). APHEA project: brief summary of results. AIRNET - A thematic network on air pollution and health: www.airnet.iras.uu.nl/reports_and_annexes
6. COMEAP (1998). Quantification of the effects of air pollution on health in the United Kingdom. Committee on the Medical Effects of Air Pollution, Department of Health, London.
7. McMahon S K, (1993). Particulate air pollution and respiratory health in Bristol. MSc Thesis, University of the West of England, Bristol.
8. McMahon S K, Fryer P, (1995). "Asthma in Children". Environmental Health Journal, July, 161-162.
9. McMahon S K, (2001). "Accelerating the symptoms: A study of traffic pollution and respiratory health of primary school children living in the Bristol area". Environmental Health Journal, June, 176-178.
10. Anderson H R, de Leon A P, Bland J M, Bower JS, Strachan DP, (1996). "Air pollution and daily mortality in London": 1987-92. British Medical Journal, 312, 665-669.
11. Bremner S A, Anderson H R, Atkinson R W, McMichael A J, Strachan D P, Bland J M, Bower J S, (1999). "Short-term associations between outdoor air pollution and mortality in London 1992-4". Occupational Environmental Medicine, 56, 237-244.
12. Ware J H, Spengler J D, Neas L M, Samet J M, Wagner G R, Coultas D, Ozkaynak H, Schwab M, (1993). "Respiratory and irritant health-effects of ambient volatile compounds - the Kanawha County health study". American Journal of Epidemiology, 137, 1287-1301.
13. Buchdahl R, Willems C D, Vander M, Babiker A, (2000). "Associations between ambient ozone, hydrocarbons and childhood wheezy episodes: a prospective observational study in south east London". Occupational Environmental Medicine, 57, 86-93.
14. Hagen J A, Nafstad P, Skrondal A, Bjorkly S, Magnus P, (2000)." Associations between outdoor air pollutants and hospitalisation for respiratory diseases". Epidemiology, 11, 136-140.
15. Dockery D W, Pope C A III, Xu X P, Spengler J D, Ware J H, Fay M E, Ferris B G, Speizer F E, (1993). "An association between air pollution and mortality in six United States cities". New England Journal of Medicine, 329, 1753-1759.
16. Pope C A III, Thun M J, Namboodiri M M, Dockery D W, Evans J S, Speizer F E, Heath C W, (1995). "Particulate air-pollution as a predictor of mortality in a prospective-study of US adults". American Journal of Respiratory and Critical Care Medicine, 151, 669-674.