Table 4

Associations of the correlation exponent (α) and mean peak expiratory flow (PEF) with asthma control from multiple regression analyses.

OutcomeParameterStudy 1 (mild to moderate)Study 2 (severe)
RR/OR (95% CI)p ValueRR/OR (95% CI)p Value
GINA*
 Partly controlledα1.27 (1.01 to 1.60)0.040––
%predPEF0.65 (0.50 to 0.86)0.002––
 Uncontrolledα0.75 (0.45 to 1.26)0.2760.73 (0.55–0.98)0.034
%predPEF0.41 (0.24 to 0.72)0.0020.65 (0.53–0.81)<0.001
ACQ score*
 Mediumα––1.23 (0.91 to 1.66)0.178
%predPEF––0.64 (0.50 to 0.82)<0.001
 Highα––1.08 (0.77 to 1.52)0.643
%predPEF––0.44 (0.32 to 0.60)<0.001
Exacerbation†α1.52 (1.20 to 1.94)0.0011.35 (1.02 to 1.80)0.039
%predPEF0.87 (0.68 to 1.10)0.2390.77 (0.63 to 0.94)0.010
AQLQ score*
 Mediumα––0.96 (0.70 to 1.30)0.775
%predPEF––0.61 (0.47 to 0.78)<0.001
 Lowα––0.91 (0.68 to 1.22)0.525
%predPEF––0.78 (0.63 to 0.97)0.026

  • For all multiple regression models, RR ratios and ORs are expressed per 0.1 change in α and per 10% change in percentage of predicted PEF (%predPEF). For each outcome, the ‘best’ category was chosen as baseline for comparisons.

  • ↵* Multinomial logistic regression: RR ratio (95% CIs) reported. For GINA, the partly controlled and uncontrolled categories were compared with the controlled category as baseline, except for those with severe asthma where the comparison was made with the partly controlled category due to low numbers in the controlled category. For ACQ score, the medium and highest tertiles were compared with the lowest tertile as baseline (best asthma control). For AQLQ score, the medium and lowest tertiles were compared with the highest tertile as baseline (best quality of life).

  • ↵† Logistic regression: OR (95% CIs) reported, having an exacerbation during the observation period was compared with having none as baseline.

  • ACQ, Asthma Control Questionnaire; AQLQ, Asthma Quality of Life Questionnaire; GINA, Global Initiative for Asthma.