RESULTS
Demographic and clinical characteristics of all the participants,
stratified by country, are presented in Table 1.
Of all participants, nearly 40%
were males with a mean age of 44 (standard deviation, SD: 14) years,
15% of the participants were obese, 47% were smokers and nearly 33%
reported exposure to smoke, 78% of the participants had asthma as a
comorbidity, 54% had moderate-severe intermittent AR and 34% had
moderate-severe persistent AR. As for allergic sensitization, pollens
were the most prevalent allergen (89%) among the participants, followed
by house dust mites (57%). Concerning the HRQoL parameters, the
participants had a median (IQR) RHINASTHMA-total score of 76 (53, 91)
and CARAT-total score of 18 (14, 22). In the bivariate analysis, we
found that participants with both AR and asthma had significantly higher
RHINASTHMA (total and subdomain) scores than the participants with AR
alone (Figure 1 ). Moreover, CARAT (total and subdomain) scores
were significantly lower in AR with comorbid asthma than in AR alone
(Figure 2 ).
Table 1
Figure 1 & 2
In the multivariable analysis, we observed that, compared to AR alone,
AR with comorbid asthma was significantly associated with poorer quality
of life (regression coefficient [β] for RHINASTHMA-total score:
0.22; 95% confidence interval [CI]: 0.19, 0.25). The association
was persistent in RHINASTHMA subdomain scores; however, the magnitude of
the estimates was different, being the highest for lower airways (0.36;
0.31, 0.41) and the lowest for upper airways (0.09; 0.04, 0.14). Upon
stratifying asthma, based on the 2017 GINA grades, the magnitude of the
association was the highest in AR patients with severe persistent asthma
(β for RHINASTHMA-total score: 0.25; 95%CI: 0.22, 0.29), and the lowest
in AR patients with mild persistent asthma (0.15; 0.10, 0.20)
(Figure 3 and Supplementary Table 1 ). We did not find any
multicollinearity between the covariates (VIF<3).
Figure 3
We observed a poorer control of symptoms in AR patients with asthma
comorbidity than in patients with AR alone (β for CARAT-total score:
-0.20; 95%CI: -0.25, -0.15); the lower airway symptoms were more poorly
controlled (-0.23; -0.29, -0.17) than the upper airway symptoms (-0.11;
-0.20, -0.01). Upon stratifying asthma according to GINA grade, AR
patients with severe persistent asthma had the poorest control (β for
CARAT-total score: -0.25; 95%CI: -0.31, -0.19) than those with mild
persistent asthma (-0.06; -0.14, 0.03) (Figure 4 and
Supplementary Table 2 ).
Figure 4
In the sensitivity analysis for effect modification by obesity
(Supplementary Table 3 ), we found that the association between
AR+asthma and RHINASTHMA-total score was marginally higher among
non-obese participants (β: 0.23; 95%CI: 0.19, 0.26) than obese ones
(0.16; 0.07, 0.25) (p-value for interaction = 0.09). The
difference was more pronounced for upper airways, the association being
significantly higher among non-obese participants (β for
RHINASTHMA-upper: 0.10; 0.05, 0.15) than obese ones (0.01; -0.13, 0.15)
(p-value for interaction = 0.04). However, we did not observe
significant differences in other subdomains. We did not observe any
effect modification by obesity for CARAT scores (Supplementary
Table 4 ).
The association of AR+asthma with RHINASTHMA-total score was highly
heterogeneous (I2: 87%; p-value for heterogeneity =
<0.001) across the participating countries (Figure
5A ). While the association was the highest in Austria (β: 0.29; 95%CI:
0.24, 0.34), it tended towards null in France (0.07, -0.06, 0.19).
Similar heterogeneity was observed for CARAT-total score
(I2: 79%, p = 0.008) (Figure 5B ). However,
the overall estimates from the meta-analyses for the association between
AR+asthma, and RHINASTHMA-total and CARAT-total scores were similar to
the ones reported in the main analysis.
Figure 5