DISCUSSION
In this prospective study, the compatibility of the tools such as PACT, ACT, FeNO, PALQLQ and lung function with GINA criteria were evaluated in children. According to GINA criteria, 84.2% of the patients had “well-controlled” asthma.
FeNO and lung function were unsuccessful at revealing control status according to GINA criteria(Table 1, 2). PACT/ACT and PAQLQ were demonstrated to be able to determine “well controlled asthma” to some extent of consistency with GINA criteria(Table 2). The best consistency with the highest sum of sensitivity and specifity were obtained with the cut-off levels of 22 for PACT, 21 for ACT and 5.9 for PAQLQ(Table 2). Neverthless, there were only fair agreement with the kappa value of 0.221 and 0.150 for PACT and PAQLQ, respectively(p<0.001). Besides, moderate agreement was obtained between ACT and GINA(κ=0.473, p<0.001). Correctly classified patients with PACT, ACT and PALQLQ according to GINA were 93(64.1%), 63(75.9%) and 139(62.9%), respectively. Ultimately, ACT gives the best result for assessing asthma control in children > 12 year-old. Also, PAQLQ gives better compatibility for asthma control in children> 12 years old than those younger <12 years old.
In our study, girls with older age have more frequently “not well-controlled” asthma. This may be due to the adverse impact of sex hormones on the control status of asthma which is a particular problem for female adult patients. Frequency of allergic sensitisation is also lower in the group of “not well-controlled” asthma. All these may be a suggestion of the hypothesis that gender disparity in asthma control starts at puberty approximately at the age of 10 years.24
Although there is no approved or validated gold standard for determining asthma control, GINA criteria are accepted to be used such as worldwide in daily practice.5, 25 PACT and ACT are some of the tools validated and mostly used in clinical practice to assess asthma control in children. Although the studies comparing the compatibility of them and GINA criteria for the assessment of asthma control had challenging results.6-10
In the study of Koolen et al, they compared PACT and ACT with GINA criteria in assessing asthma control of children with similar study design.10 They found an AUC of ROC curve analyses for PACT 0.89 and ACT 0.92 which were higher than our study(0.79 and 0.86 in our study, respectively). In their study, the cut-off value with the highest sum of sensitivity and specifity for ACT was< 20 for uncontrolled asthma, which is the same as our study. The cut-off value for PACT in their study was< 22, which is one point higher than that in our study(< 21). The sensitivity and specifity of both PACT and ACT(sensitivity 82% and 76%; specifity 85% and 96%, respectively) with the calculated cut-off values were higher than our study(sensitivity 60.9% and 71.9%; specifity 88.2% and 89.5%, respectively). This may be due to that Koolen et al(n=145) had more patients with uncontrolled(13.1%) and partly controlled asthma(33.8%) than those in our study(6.1% and 9.6%, respectively). Another reason may be patients with uncontrolled and partly controlled asthma were evaluated together in the “not-well controlled” asthma group in our study.
In another study with similar design, Voorend-van Bergen et al(n=228) found the same best cut-off value for PACT(> 22) for “well-controlled” asthma with respect to GINA criteria.8 They proposed to determine “well-controlled” asthma which is the primary goal of clinical asthma management according to GINA criteria. Their AUC for PACT(0.81) was also similar to our results(0.79). On the other hand, their best cut-off value for ACT(> 23, AUC=0.91) was higher than ours(> 21, AUC=0.86).
In a recent study, Deschildre et al compared PACT performance of assessing asthma control with respect to GINA crtieria.7 Also, they grouped asthma level of patients as well controlled and not controlled(partially controlled or uncontrolled), as we did in our study. Using ROC curve analyses, they found the same cut-off value for PACT(< 21, for “not controlled” asthma) with a higher sensitivity(76%), but lower specifity(81.5%) than those found in our study(60.9% and 88.2%, respectively). However, the most important difference from our study is that the rate of “not controlled” asthma in their study (76.5%) was much higher than that in our study(15.7%).7
As in the study of Voorend-van Bergen et al, FeNO was not difference between the control groups of asthma as in our study.8As similar to earlier studies, we found no correlation between asthma control scores, GINA criteria and FeNO or FEV126-28. Symptoms, lung function and airway inflammation represent different domains of asthma phenotype and show poor agreement8. Many children with uncontrolled asthma have normal lung functions between exacerbations1. A low FEV1 percent predicted, particularly if it is<60%, identifies patients at risk of future asthma exacerbations independent of symptom levels1. If symptoms are few despite low FEV1 %predicted, limitation of lifestyle or poor perception of airflow limitation should be considered that would be a marker of untreated airway inflammation.29, 30 In our study, none of the participants had FEV1<60%. Besides, there were no difference for FEV1(expected %), FEV1/FVC(ratio) and FEF25-75(expected %) between control groups and no compatibility with GINA criteria(Tabel 1 and 2).
In our study, ACT had better results for agreement and compatibility with GINA criteria than PACT. This may be due to that older children replies ACT and they had better perception of symptoms and longer recall period than children younger than 12 years old who replies PACT with their parents. Parents have a longer recall period than children, who may recall only the last few days1. The fact that younger children would be more frequently asymptomatic except exacerbations may have an additional contribution to this result. In accordance, the correlation of PAQLQ and ACT was stronger than that of PAQLQ and PACT in our study. Voorend-van Bergen et al had similar results for this correlation8. Additionally, PAQLQ had higher agreement with GINA criteria in children > 12 years old than those of younger ones in our study (κ=0.326, p<0.001 and κ=0.151, p=0.014, respectively). There are several studies demonstrated the correlation between PAQLQ and ACT/PACT31-34. However, to the best of our knowledge, this is the first study to achieve a cut-off value for PAQLQ that has fair compatibility with GINA criteria.
The most important limitation of our study was lower number of patients with uncontrolled and partly controlled asthma. This may be a consequence of the study design that firstly the control status of asthma was evaluated, treatment adherence and inhaler technique were adjusted if needed. Four weeks later, patients were called for the main visit at which the adherence and technique of treatment was better than the first visit(data not shown). As a natural consequence, asthma control was better at the main visit. Besides, this was a real-world study which would provide more true information from daily clinical practice. Real-world studies presently start to become cause of choice because randomised controlled trials include selected patients populations that rarely represent the real situation35.
In conclusion, ACT seems to be better than PACT for compatibility with GINA in assessing control status in children. Probably, it is because older children have a longer recall period than younger ones(<12 yrs). The better correlation of PAQLQ and ACT and better compability of PAQLQ and GINA in older children may also be related to this reason. Neverthless, the cut-off levels for PACT and ACT differ from study to study. This may be because diffent study populations, varying study designs, care settings and rate of patients with uncontrolled asthma. So, it would be better to use these tools for individual patients by comparing their own tests, instead of determining control status according to cut-off values, until large scale studies are performed to determine cut-off values of the tools for each population individually.