Qualitative or discrete quantitative variables were summarized using absolute and relative frequencies, while continuous quantitative variables were described using central tendency and dispersion measurements. The degree of concordance between (a) patient-perceived asthma control and GEMA criteria, (b) physician-perceived asthma control and GEMA criteria and (c) physician-perceived and patient-perceived control was calculated. Differences between controlled patients and uncontrolled patients were evaluated with Chi-squared tests (categorical variables) and the Mann–Whitney–Wilcoxon non-parametric test (continuous variables). The variables with greatest association in the bivariate analysis were included in multivariate logistic regressions in order to determine the variables with a significant independent effect in the perception of asthma control according to patients or physicians. The choice of possible variables was conducted similarly to previous studies. Hosmer and Lemeshow's goodness-of-fit test was used to determine the total variation explained by the model.

A total of 343 patients with a mean age (standard deviation [SD]) of 48.1 (14.7) years were included, of whom one third were women. According to GEMA criteria, only 10.2% of patients were controlled, 27.7% had partial control and 62.1% were poorly controlled. Table 2 shows baseline demographic and clinical data by grades of control. As expected, patients controlled according to GEMA had significantly higher FVC and FEV1 values (Table 2). Moreover, they had higher levels of total IgE (P=.0409). Patients with uncontrolled asthma according to GEMA had a higher body mass index (BMI) (Fig. 1, P=.006) and a more sedentary lifestyle (Fig. 2, P=.016), as well as a tendency toward a higher mean age (48.5 [14.7] vs 44.1 [14.1] years in the controlled asthma subgroup, P=.079).

Fig. 1.

BMI in subgroups defined as asthma control according to GEMA and according to the patient him/herself (controlled/uncontrolled [partially controlled/poorly controlled]). Differences between subgroups according to GEMA criteria were significant (P=.0065) and non-significant according to patient perception (P=.5088).

(0.07MB).

Fig. 2.

Level of physical activity in subgroups defined by asthma control according to GEMA (top figure) and the patient (lower figure) (controlled/uncontrolled [partially controlled/poorly controlled]).

(0.23MB).

Omalizumab use, prescribed in 193 of 342 patients, was significantly more prevalent in patients with controlled asthma according to GEMA (71.4%) than in partially controlled patients (63.2%) and those with poor control (49.5%) (P<.005).

No significant association between smoking and asthma control according to GEMA was observed (P=.464).

According to patients’ perceptions, 42.3% of the population were well controlled, 36.1% had partial control and 21.6% were poorly controlled. According to the physicians, these percentages were 25.7%. 39.7% and 34.6% of patients, respectively.

When the different asthma control criteria were compared, significant differences were observed for all the comparisons made: patient vs GEMA (P<.0001); physician vs GEMA (P<.0001) and patient vs physician (P<.0001). Both patients and physicians overestimated asthma control: in the subgroup of patients with uncontrolled asthma according to GEMA, 35.7% considered themselves to be well controlled (Fig. 3A) and 16.7% were well controlled according to their physician (P<.0001) (Fig. 3B). Of the total sample, 75.8% and 59.3% of patients controlled from the perspective of the patient and the physician, respectively, did not meet the criteria for control according to GEMA.

Fig. 3.

Concordance between grade of asthma control according to different criteria. (A) Perception of patients in subgroups defined by GEMA criteria. (B) Perception of physicians defined by GEMA criteria. C. Perception of physicians in subgroups defined by patient perception.

(0.35MB).

Patients overestimated asthma control more than physicians: 41.3% of patients who considered themselves controlled were not well controlled according to their physicians (P<.0001) (Fig. 3C).

The main factors associated with patient- and physician-perceived lack of control are shown in Tables 3 and 4, respectively (multivariate logistic regression analysis). Other variables tested but not included in either of the two final models were age, sex, level of studies, BMI, time since diagnosis, morning symptoms and treatment with theophyllines.

Five independent factors significantly associated with patient-perceived lack of control were found (Table 3): nocturnal awakenings (P<.0001), a greater perception of breathlessness evaluated according to the Nijmegen questionnaire (P<.0001), more frequent use of rescue medication (P<.0001), more emergency room visits in the last year (P=.0027) and greater limitation in activities of daily living (P=.0265). The model was significant and the Hosmer and Lemeshow goodness-of-fit adjustment gave a P-value of .5080, showing that the analysis explains more than 50% of the variation in patient-perceived asthma control.

In the model for predicting physician-perceived lack of control (Table 4), a total of 6 significant predictive variables were obtained: more frequent use of rescue medication (P<.0001), greater perceived limitation in activities of daily living (P=.0002), FEV1 lower than 80% (P=.0031), no omalizumab treatment (P=.0115), nocturnal awakenings (P=.0277) and misunderstanding of the biological cause of asthma shown in the asthma knowledge evaluation questionnaire (P=.0435). This second model was also highly significant (Hosmer and Lemeshow goodness-of-fit test: P=.8776) and explained most of the variation associated with physician-perceived lack of control (around 88% variation).

This study, carried out in a large Spanish cohort of patients with persistent severe asthma, shows that both patients and physicians overestimate asthma control as defined by the clinical criteria recommended by GEMA. Moreover, it shows that the main factors associated with lack of control according to GEMA are sedentary lifestyle and higher BMI. In our cohort, the degree of disease control according to the guidelines was very low: only 1 in 10 patients was well controlled. Accordingly, the prevalence of uncontrolled asthma (around 90%) was clearly higher than that of other studies which record a range of 55.0%–71.6%.2,16,17 These differences may be due to the use of other definitions of control and the lack of objective and reliable indicators to complement the information on patient-reported symptoms. A recent study showed that even asthmatics who are well controlled according to asthma control questionnaires (e.g., ACQ) present persistent corticosteroid-sensitive inflammation.18

Although asthma control is overestimated by both patients and physicians, patients showed higher subjective perception of control than physicians. This discordance between physician- and patient-perceived control has already been observed in previous studies.5 A British study found that only 52% of classifications were consistent between physicians and patients.19 Similarly, it is interesting to note the results of a recent study conducted in Spain20 in the same healthcare setting as ours, that showed a considerable degree of discordance between patient and the physician perception of aspects such as treatment satisfaction and impact of asthma on daily life (for both variables, in 26% of the cases the physicians’ perception was lower than the patients’). The physician-patient discordance was shown to increase as asthma control worsened (being about 30% in the group of controlled patients, but over 73% in the case of patients with poor asthma control). Although these results underline the importance of the perception of control in the asthma patient, current Spanish guidelines do not take this opinion into account.

In this study, no differences were reported in asthma control with respect to age, sex or smoking habit. In contrast, sedentary lifestyle, higher BMI and poorer knowledge of the disease were associated with a poorer degree of control in severe asthma. These results coincide partially with those of other studies. In the REACT study,16 no significant differences were found in control with respect to age, ethnic origin, lifestyle (urban, suburban or rural) or the type of health insurance. However, lack of control was more common in women and in patients from lower income groups and lower educational level. Díez et al.2 identified sex, ethnic group, BMI, smoking habit, educational level, physical activity and medical history as factors associated with asthma control.

The association between poor asthma control and higher BMI and/or sedentary lifestyle has also been described in other publications.21–24 A recent review concluded that obesity increases the prevalence and incidence of asthma, although this effect appears to be modest.25 Although no causal mechanism has been described, lack of control in obese patients may be explained by mechanical and inflammatory phenomena. In particular, obesity affects pulmonary and airway anatomy and increases abdominal wall and chest mass. This leads to reduced bronchial diameter and residual functional capacity and produces a state of chronic inflammation that contributes significantly to hyperresponsiveness.26 In contrast to another review,27 no association was found between control and smoking, probably because of differences in the definition of asthma control.

In multivariate analyses, common factors associated with both physician- and patient-perceived lack of control included nocturnal awakenings, frequent use of rescue medication and significant limitation of activities. On the other hand, 4 factors were perceived differently by physicians and patients. When patients’ perception was analyzed, a marked perception of breathlessness and visits to the emergency room in the last year were significant. When physicians’ perception was analyzed, low FEV1 and poor understanding of the biological cause of asthma were significant. Thus, the most important variables in determining the perception of poor asthma control, particularly for patients, are generally related with symptoms and interference in daily life and sleep, while physicians also take lung function into account. With regard to these results, a recent study28 showed that uncontrolled patients had more catastrophic thinking in the absence of exacerbation, as well as more activity limitation. Thus, the mental component appears to be important in the effective management of uncontrolled asthma.19 It is interesting to note that neither morning symptoms nor previous exacerbations were taken by the physicians as predictive factors for lack of control. These 2 variables are taken into account in the GEMA classification, but our results suggest that physicians either do not use them in their clinical practice or do not give them the importance they deserve.

Since this was an observational, uncontrolled, non-interventional study, there are several limitations inherent in the design, and selection or data bias cannot be ruled out. The amount of information available for each patient and the sample size were limited. Accordingly, it would be of interest to replicate the results in a larger, more diverse group of patients with severe asthma.

Control of severe asthma is considered the principal objective of clinical guidelines. In our opinion, the findings of this study have important implications, since they reveal lack of control of asthma in clinical practice and the direct effect this has on patients’ quality of life. Our results also show that, from the patients’ point of view, symptoms should be considered important when evaluating control of severe asthma. In this respect, guidelines should consider the perspective of the patient in order to facilitate physician–patient consensus. Overestimation by physicians must be reduced if better disease control is to be achieved, and raising awareness of the importance of correctly questioning patients about their morning symptoms and exacerbations in the weeks or months before the visit could improve detection of poorly controlled patients. Future studies must also address the cost-efficacy ratio of treatment strategies, based on differences in control achieved by each of those strategies.