Do people living near inner city main roads have more asthma needing treatment? Case-control study

BMJ 1996; 312 doi: (Published 16 March 1996) Cite this as: BMJ 1996;312:676
  1. Anna Eleri Livingstone,
  2. Gavin Shaddick,
  3. Christopher Grundy,
  4. Paul Elliott
  1. Environmental Epidemiology Unit, London School of Hygiene and Tropical Medicine, London WC1E 7HT
  2. G Shaddick, statistician C Grundy, research assistant P Elliott, reader. Limehouse Practice, Gill Street Health Centre, London E14 8HQ
  3. A E Livingstone, general practitioner.
  1. Correspondence to: Dr Livingstone.
  • Accepted 9 January 1996

Hospital admissions for asthma in east London are 80% above the national rates. This may reflect the high incidence of acute asthma. Recent reports of a higher prevalence of wheeze1 2 or hospital admissions in children in association with traffic flow or proximity of residence to roads3 have highlighted concerns about the possible health effects of road traffic in the London Borough of Tower Hamlets.

In each of two computerised general practices in Tower Hamlets around 20% of the population have received computer prescriptions for bronchodilators, inhaled steroids, or inhaled anti-inflammatory drugs since 1990. The diagnostic computer coding for asthma showed a prevalence of treated asthma of 9% in one practice and 17% in the other (unpublished observation). We examined whether the proximity of residence to main roads was associated with these high prescribing rates for asthma in the two inner city practices.

Subjects, methods, and results

This case-control study took place in June 1994 in two adjacent general practices located near major roads; they have a combined list of 15 300 patients and 10 principals. Since 1990 both practices have routinely used EMIS (Egton Medical Information Systems) consulting room computers. We included all eligible patients aged 2 to 64 who had had computer consultations in the preceding year. Cases had a computer record of prescriptions for asthma drugs in the previous year and a computer diagnosis of asthma. Control subjects had no computer record of asthma drugs and no recorded diagnosis of asthma. We collected data on age, sex, practice, smoking, and residential postcode from the notes. Data were linked by postcode to information on deprivation from the 1991 census and using the ARC geographic information system, with the shortest distance of residential postcode to a busy road carrying more than 1000 vehicles an hour at peak times. Postcode grid references within 10 m were available for 92% of cases (978) and 91% of controls (5685). The remainder were within 100 m.

Table 1 summarises baseline characteristics and results. Overall, 51% of subjects (3172/7299) in the study came from practice 2, whose patient population made up 43% of the whole (6558/15 393 registered patients). Most cases and controls lived in districts rated to be in the bottom fifth of the Carstairs deprivation scores (98% (1047/1066) v 97% (6032/6233)). Cases tended to be younger and male, and proportionately more of the adult cases were current smokers. Smoking habit did not differ significantly in relation to distance of home from roads.

Table 1

Details of cases with active asthma and controls. Values are numbers (percentages) of patients unless stated otherwise

View this table:

In children under 16 the unadjusted odds ratio for being treated for asthma when living 150 m or less from busy roads compared with more than 150 m from them was 0.94 (95% confidence interval 0.75 to 1.19); in adults the odds ratio was 0.81 (0.68 to 0.97) (data not shown). The odds ratios did not differ significantly from unity after adjustment for age, age group (in adults) sex, and practice (table 1). Point process methods to analyse the effect of living at different distances from roads also showed no evidence of spatial clustering.5


This study shows no increase in risk of asthma with living close to busy roads. Although table 1 shows analyses based on residence 150 m or less from busy roads or more than 150 m away from them, we also examined risk with distance measured as a continuous variable; spatial point process methods fit an exponential function for the decline in risk with distance.5 With this regression approach there was also no evidence of spatial clustering of cases of asthma near roads when we adjusted for age, sex, and smoking habit. These negative findings are consistent with an ecological study of the prevalence of wheeze near the M25 motorway, but they contradict other reports.1 2 3 The use of routine data from general practice computers allowed basic covariates to be adjusted in a way that is impossible when using many routine data sources. The definition of cases and controls made misclassification between them unlikely. Possibly, some true cases were missed because of the requirement for computer recording of both treatment and asthma diagnosis. There is no reason to suspect that this was a source of bias. On the other hand, distance of residence from a busy road is a crude proxy for exposure to traffic related pollution, especially among adults for whom work and commuting patterns may be more relevant. Misclassification of exposure is therefore inevitable.

We thank the Chrisp Street and Gill Street Practices, London E14, the Environmental Health Department, the London Borough of Tower Hamlets, and the Department of Transport for their help.


  • Funding AEL was supported for her MSc by payments from the Department of Health for extended study leave from general practice.

  • Conflict of interest None.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract