During periods of exacerbation, coryzal and cough symptoms were significantly more likely in the cold seasons (P<.001), with a seasonal change of 35.3% and 11.5%, respectively. To illustrate these differences in symptom composition, Figure 2 shows the time-course in the prevalence of dyspnea, coryzal, and cough symptoms for exacerbations in the cold and warm seasons.

All stable-state respiratory symptoms, apart from dyspnea, were significantly higher in relative terms (P < .05) in the cold seasons compared with the warm seasons. However, the differences are small and unlikely to explain the longer recoveries in the cold seasons.

Figure 1. Average exacerbation recovery time, in days, from all symptoms at 1°C intervals of the average daily mean outdoor temperature. The line joins together the predicted values from the random effects Poisson regression model.

Table 2—Exacerbation and Stable Symptoms in Cold and Warm Seasons

Respiratory Symptom Cold Season Warm Season P value
Exacerbations, d 1,052 676
Dyspnea 71.2 74.9 .096
Sputum purulence, 36.8 39.1 .343
Sputum volume 54.1 54.4 .887
Coryza 42.5 31.4 <.001
Wheeze 46.0 46.5 .857
Sore throat 20.2 17.9 .247
Cough 50.6 45.4 .036
Stable (outside exacerbation), d 110,144 118,959
Dyspnea 9.5 9.6 .403
Sputum purulence 1.0 0.8 <.001
Sputum volume 2.7 2.4 <.001
Coryza 3.4 1.9 <.001
Wheeze 4.8 3.4 <.001
Sore throat 1.1 0.6 <.001
Cough 3.5 2.5 <.001

Data are given as % of d unless otherwise indicated.

Outdoor Activity

Throughout the year, patients changed their outdoor activities nonlinearly with temperature (Fig 3). Excluding periods of heat stress, when the average daytime and nighttime temperature was > 20.5°C, fewer patients go out as temperatures get colder; the OR was 1.028 per 1°C rise (95% CI, 1.024-1.032). At < 2.5°C, there was a significantly faster reduction in the likelihood of patients going outdoors with temperature (P < .001), with the OR at 1.13 per 1°C rise (95% CI, 1.09-1.18; P< .001). During heat stress, as temperatures exceeded 20.5°C, patients also reduced outdoor activity, with an OR at 0.96 per °C (95% CI, 0.93-0.999; P = .044).