Influence of Weather and Atmospheric Pollution on Physical Activity in Patients With COPD

Ayedh D. Alahmari; Alex J. Mackay; Anant R.C. Patel; Beverly S. Kowlessar; Richa Singh; Simon E. Brill; James P. Allinson; Jadwiga A. Wedzicha; Gavin C. Donaldson


Respiratory Research. 2015;16(71) 

In This Article


This study shows that day-of-the-week, meteorological factors and for the first time that high levels of atmospheric pollutants affect physical activity in COPD patients. The reduction in activity at weekends was not unexpected as this is typically a period of rest. Days that were warm, dry and sunny appeared to encouraged patients to go outside and walk more, whereas cold, rainy and overcast days reduced activity. A number of studies have observed that physical activity decreases in healthy adults during the colder, shorter winter months or increases on longer, sunny days.[25] Indeed, some studies have shown seasonal variation in activity in COPD patients[26–28] and is reduced at weekends compared to weekdays[29,30] but we extend these findings by showing that activity is primarily related to meteorological conditions irrespective of the season.

Our findings are important because COPD patients already have a reduced exercise capacity due to their airflow limitation. Any further reductions of activity due to the weather or day-of-the-week may worsen muscular de-conditioning which is common in inactive COPD patients. Muscle weakness and feelings of fraility may make the patients feel unable to leave their homes and once this behaviour is established may prove difficult to reverse. It might in part explain why health related quality of life is poorer in winter than spring or summer[31] and measures of anxiety and depression higher in winter.[32] The findings are also important because for the first time we show an effect of atmospheric pollution on physical activity which was only possible because we studied a group whose air-flow limitation is sufficient to make such effects apparent.

There are mechanisms by which outdoor atmospheric pollution might cause patients to be less active when outdoors. O3 above 200 ppb can affect peak expiratory flow in elite cyclists during maximal exercise[33] but may not cause problems at low levels.[34] We found that PEF was also reduced only at high levels of O3 during the weekdays (p = 0.040) though it just failed to reach significance over the whole week (p = 0.054). Atmospheric pollutants can also produce harmful effects on the airways, such as pulmonary and systemic inflammation,[35,36] reduction in airway ciliary activity,[37] increases in bronchial reactivity[38] and airway oxidative stress.[39] Exposure to O3 can also significantly increases heart rate and blood pressure, as well as causing mitochondrial damage.[40] However, whether patients are aware of these systemic and anatomic effects is not clear. We found that dyspnoea increased with higher O3 levels but did not find any effect on the CAT quality of life score. Some patients may not have gone outdoors when the pollution levels were high but it is not obvious how the patients knew not to go out. There is little evidence that people alter their behaviour in response to pollutant alerts in the news or from other advisory systems.[41] O3 is a colourless, odourless, gas which cannot be seen or smelt but its precursors are mainly motor vehicle exhaust fumes might be detected.[41] In London, a pollution haze can be seen on some days[42] but the patients might not live on hills or in high-rise buildings where these observations can easily be made. High levels of O3 are known to be associated with hot weather which might discourage patients from taking exercise. However, we excluded from the analysis the hottest days with mean temperatures over night and day exceeding 22.5 °C. Further studies are needed to determine if and how COPD patients can detect increased atmospheric pollution.

The limitations of this study should be discussed. We were not able to assess the intensity of the physical activity. This can be measured with accelerometers but would require weekly or fortnightly clinic visits by patients to download data which was not practical in this long term study. Pedometers can be inaccurate in slow walking individuals but this would be a consistent bias in a given patient and thus unlikely to alter how they respond to changes in pollution or the weather. Another limitation was that we did not collect pedometry data as fully as the PEF or dyspnoea data and we have no control group. Some patients did not wear their pedometer every day, some were lost and/or broken when inadvertently washed and a replacement only possible at their 3 monthly clinic visit. Some patients were excluded because too little data remained after excluding periods of exacerbation. These excluded patients may well have been frequent exacerbators, and thus our findings might not necessarily apply to this group though the exacerbation frequency in the studied group was similar to the 126 patients not included. We have also not examined other weather conditions such as snow or ice, when the risk of slipping might discourage excursions outdoors. It was not practical to monitor the pollution and climate exposure of each individual and thus we assumed that the pollution levels at the monitoring site in Bloomsbury and weather measured at Heathrow were indicative of that experienced by the patient. Previous studies have shown the data at Bloomsbury is correlated with outer suburban sites[43] and similar in temporal evolution to other sites in London.[44] Although, we did not use personal pollution monitors we did collect individual outcome data – and this semi-individual design is considered valid for air pollution studies.[45] In our analysis, we felt it necessary to analyse separately weekdays and weekends as well as the whole week because the "day-of-the-week" effect was very large and may have confounded the results. In many countries, O3 is significantly higher at weekends compared to weekdays[46–48] whereas PM10 is higher at weekdays.[46] We found similar effects in London. By analysing the data in this way, we reduced the statistical power and this could explain the absence of consistent effects during both weekdays and weekends.