Table 2

Baseline ECRHS I crude prevalences and ECRHS I inverse-propensity weighted prevalences with net change in prevalence of respiratory symptoms from ECRHS I to ECRHS II, from ECRHS II to ECRHS III, and from ECRHS I to ECRHS III

ECRHS I subjects*Number responding to ECRHS I, II and III
Symptom from postal surveyNumber of centresNumber of participantsECRHS I crude prevalence (%)SubjectsECRHS I prevalence (%)Net change ECRHS II-I (%)Net change ECRHS III-II (%)Net change ECRHS III-I (%) (95% CI)pHeterogeneity I2 (%)Heterogeneity p
Wheeze2914 73421.5773821.3−1.6−0.6 −2.4 (−3.5 to −1.3) 0.000015 0.00.71
Wheeze with breathlessness2914 69910.9763010.1−0.20.50.6 (−0.2 to 1.5)0.120.00.61
Wheeze without a cold2914 72013.4766212.7−0.5−0.8 −1.5 (−2.4 to −0.6) 0.0012 0.00.84
Woken with chest tightness2314 36514.3642012.6−0.5−0.7−1.3 (−3.0 to 0.3)0.157.60.00033
Woken with shortness of breath2814 3616.374595.50.30.40.9 (−0.1 to 2.0)0.07150.60.0013
Woken by attack of coughing2114 75228.6627927.8−0.40.50.2 (−2.3 to 2.6)0.9165.00.000019
Asthma attack2914 7453.976943.50.70.1 0.6 (0.1 to 1.1) 0.013 0.00.89
Current asthma medication2714 7494.070693.52.01.5 3.6 (3.0 to 4.2) 3.6×10−31 0.00.56
Diagnosed asthma2714 7115.570024.82.01.5 3.5 (2.9 to 4.2) 2.6×10−25 0.00.94
Nasal allergies2714 27924.8702023.22.8−0.0 2.7 (1.7 to 3.7) 1.2×10−7 0.00.8
Symptom from interviewer survey on a subsample
Chronic winter cough2614 6987.745707.1−0.31.41.1 (−0.3 to 2.5)0.1349.10.0027
Chronic winter phlegm2614 6656.345455.80.40.61.0 (−0.3 to 2.2)0.1240.10.019
  • All estimates for change during the first 10  years of follow-up were compatible with estimates published in Chinn et al 1 (which were unadjusted for non-response, and included some centres which did not take part in ECRHS III).

  • *Crude prevalence in all target sample participants who completed postal survey in ECRHS I.

  • Numbers are in bold if p<0.05.

  • ECRHS, European Community Respiratory Health Survey.

  • Note that  changes in estimates from ECRHS I to ECRHS III are not the sum of changes from ECRHS I to ECRHS II, and from ECRHS II to ECRHS III due to the use of inverse variance weighting required by DerSimonian and Laird meta-analysis.12