- Email: [email protected]
A new criterion by which to discriminate between patients with moderate allergic rhinitis and patients with severe allergic rhinitis based on the Allergic Rhinitis and its Impact on Asthma severity items
Rhinitis, sinusitis, and ocular diseases
Antonio Valero, MD, PhD,a,m Montse Ferrer, MD, PhD,b Joaquı´n Sastre, MD, PhD,c Ana M. Navarro, MD, PhD,d Laura Monclu´s, MD,e Enrique Martı´-Guadan˜o, MD, PhD,f Michael Herdman, MSc,b,g Ignacio Da´vila, MD, PhD,h Alfonso del Cuvillo, MD, PhD,i Carlos Cola´s, MD, PhD,j Eva Baro´, BSc,g Ignacio Ante´para, MD, PhD,k Jordi Alonso, MD, PhD,b and Joaquim Mullol, MD, PhDl,m Barcelona, Madrid, Seville, Badalona, Salamanca, Ca´diz, Zaragoza, and Bilbao, Spain
Background: Allergic Rhinitis and its Impact on Asthma (ARIA) differentiates mild from moderate/severe patients on the basis of 4 severity items. The high prevalence of moderate/ severe patients suggests the need to differentiate between them. Objectives: To identify the categorization that maximizes discrimination between moderate and severe allergic rhinitis (AR) by using ARIA guidelines. Methods: Observational, cross-sectional study. Clinical characteristics, nasal symptoms (Total Symptom Score 4), and health-related quality of life (HRQL; Rhinoconjunctivitis Quality of Life Questionnaire and Short Form 12) were assessed. The association of severity items (sleep, daily activities/sport, work/ school, and troublesome symptoms) with symptoms and HRQL was analyzed using linear regression models. ANOVA and effect sizes were used to assess differences in symptoms and HRQL among groups defined by the number of affected ARIA items. From athe Unidad de Alergia, Servicio de Neumologı´a y Alergia Respiratoria, Hospital Clı´nic, Barcelona; bthe Health Services Research Unit, Institut Municipal d’Investigacio´ Me`dica, Barcelona; cthe Servicio de Alergia, Fundacio´n Jime´nez Dı´az, Madrid; dthe Unidad de Alergia, Hospital El Tomillar, Dos Hermanas, Sevilla; eJ Uriach & Co, Barcelona; fthe Unidad de Alergia, Hospital Germans Trias i Pujol, Badalona; g3D Health Research, Barcelona; hthe Servicio de Alergia, Hospital Clı´nico, Salamanca; ithe Clı´nica Dr Lobato´n, Ca´diz; jthe Servicio de Alergia, Hospital Clı´nico ‘‘Lozano Blesa,’’ Zaragoza; kthe Servicio de Alergia, Hospital de Basurto, Bilbao; lthe Unitat de Rinologia i Clinica de l’Olfacte, Servei d’Otorinolaringologia, Hospital Clı´nic, Institut d’Investigacions Biome`diques August Pi i Sunyer (IDIBAPS), Barcelona; and mthe Centro de Investigacio´n Biome´dica en Red (CIBER) de Enfermedades Respiratorias, Barcelona. Supported by an unrestricted educational grant from Uriach & Co, Barcelona, Spain. Disclosure of potential conflict of interest: The authors have declared that they have no conflict of interest. Received for publication December 4, 2006; revised March 30, 2007; accepted for publication April 5, 2007. Available online May 26, 2007. Reprint requests: Antonio Valero, MD, PhD, Unidad de Alergia, Servicio de Neumologı´a y Alergia Respiratoria, ICT, Hospital Clı´nic, C Villarroel, 170, 08036 Barcelona, Spain. E-mail: [email protected]. 0091-6749/$32.00 Ó 2007 American Academy of Allergy, Asthma & Immunology doi:10.1016/j.jaci.2007.04.006
Results: Nontreated patients (N 5 141) with moderate/severe AR were studied. All severity items showed a similar independent association with symptoms and HRQL scores, and there were no interaction effects, indicating that categorization of patients into moderate and severe could be based only on the number of items affected. Effect sizes were highest between patients with 4 affected ARIA items and those with 3, 2, or 1 affected item (effect sizes greater than 0.8 in all comparisons using Rhinoconjunctivitis Quality of Life Questionnaire and Short Form 12 Physical Composite Summary, and greater than 0.5 using the Total Symptom Score 4; P < .001). Conclusion: Using ARIA severity items, the criterion that best discriminates AR severity is considering moderate those with 1 to 3 affected items and severe those with 4. Clinical implications: Discrimination between patients with moderate and severe AR should help to obtain homogeneous populations for both research and clinical purposes. (J Allergy Clin Immunol 2007;120:359-65.) Key words: Allergic rhinitis, severity criteria, ARIA, quality of life
Allergic rhinitis (AR) is a highly common and increasingly prevalent disease that generates high social and health care costs and has a significant effect on quality of life (QOL).1-3 Allergic rhinitis has traditionally been classified as perennial or seasonal depending on the allergen responsible for symptoms. This classification has been found to be clinically unsatisfactory for at least 2 reasons: (1) the frequency of allergen polysensitization is increasing among patients with AR, and (2) pollens and mold may cause perennial symptoms whereas house dust mite may provoke seasonal symptoms in some patients.4,5 The Allergic Rhinitis and its Impact on Asthma (ARIA) document developed a new classification, intermittent allergic rhinitis (IAR) and persistent allergic rhinitis (PAR), on the basis of symptoms duration. The ARIA classification also introduced a system for assessing AR severity on the basis of the presence or absence of 359
Rhinitis, sinusitis, and ocular diseases
A new criterion by which to discriminate between patients with moderate allergic rhinitis and patients with severe allergic rhinitis based on the Allergic Rhinitis and its Impact on Asthma severity items
360 Valero et al
Abbreviations used AR: Allergic rhinitis ARIA: Allergic Rhinitis and its Impact on Asthma COPD: Chronic obstructive pulmonary disease HRQL: Health-related quality of life IAR: Intermittent allergic rhinitis MCS: Mental composite summary PAR: Persistent allergic rhinitis PCS: Physical composite summary QOL: Quality of life RQLQ: Rhinoconjunctivitis Quality of Life Questionnaire SF-12: Medical Outcomes Study 12-Item Short-Form Health Survey TSS: Total Symptom Score
Rhinitis, sinusitis, and ocular diseases
impairment in any of 4 health-related quality of life (HRQL) items: sleep, daily activities/sport, work/school, and troublesome symptoms. According to the ARIA workshop group, AR was defined as mild when there was no impairment in any of these items, and moderate/severe when there was impairment in 1 or more areas. One of the merits of this classification is that, for the first time, it explicitly used HRQL criteria to assess severity.3 These new classifications (symptoms duration and disease severity) were based on expert consensus and required further validation. In recent years, the symptoms duration classification (IAR vs PAR) was shown to be valid in several large epidemiologic studies.6-10 However, the validity of the severity classification has not been established yet. Moreover, the large prevalence of moderate/severe rhinitis (69% of patients with rhinitis consulting at otolaryngology and allergy clinics and 90% of those consulting general practitioners)4,8-10 suggests an important heterogeneity in this disease severity group. In addition, some concerns have been raised on the grading of severity proposed in the ARIA guidelines.4,11 Bousquet et al4 proposed the replacement of the term ‘‘moderate/severe’’ by ‘‘severe’’ because of the high proportion of patients reporting impaired activities. Recent recommendations of differentiating between moderate and severe rhinitis and therefore modifying the current ARIA severity criteria have been made.11 The objectives of the study were (1) to estimate the heterogeneity of symptoms and HRQL scores in patients classified as moderate/severe according to current ARIA severity classification criteria, (2) to assess the degree of association of each of the ARIA severity items with symptoms and HRQL, and (3) to identify the categorization of the ARIA items that maximized the differentiation between moderate and severe AR cases in terms of symptoms and HRQL.
METHODS Study design and participants An observational, cross-sectional study was performed between March and August 2004 in the allergy and otorhinolaryngology
J ALLERGY CLIN IMMUNOL AUGUST 2007
departments of 27 Spanish hospitals. This period of 6 months was planned to include 2 different seasons, spring and summer, because in Spain, exacerbations of rhinitis induced by some allergens mainly appear in spring and fall (ie, house dust mites, pollens). Each center consecutively included patients with a diagnosis of AR based on any clinically relevant allergen sensitization who met inclusion criteria and agreed to participate. Inclusion criteria were (1) age older than 18 years, (2) diagnosis of IAR or PAR, (3) classified as moderate/severe according to the ARIA guidelines,3 and (4) nontreated at the time of inclusion. The period without treatment had to be at least 8 and 4 weeks in the case of oral and intranasal corticosteroids, respectively, and 2 weeks for oral and topical antihistamines and antileukotrienes. None of the patients included in the study were initiating specific immunotherapy. All patients provided informed consent to participate in the study, and the study was approved by the Ethics Committee of the Hospital Clı´nic (Barcelona).
Patient evaluation Patient assessments were performed on inclusion in the study. Sociodemographic data collected included sex, age, and educational level, and clinical data collected included duration of AR (days with symptoms per week and number of consecutive weeks with symptoms), AR severity (number of impaired items on the ARIA classification), time since diagnosis, comorbidities, presence of treatment for AR, and symptom severity. Allergic rhinitis symptoms were assessed using the Total Symptom Score (TSS) 4 by calculating the sum of scores for nasal obstruction, rhinorrhea, nasal itching, and sneezing. Each nasal symptom was scored on a scale from 0 to 3 (0, no symptoms; 1, mild; 2, moderate; 3, severe). resulting in a TSS4 score ranging from 0 to 12. Health-related quality of life assessment included a diseasespecific (the Rhinoconjunctivitis Quality of Life Questionnaire [RQLQ])12 and a generic (the Medical Outcomes Study 12-Item Short-Form Health Survey [SF-12])13 instrument. Both of them have been adapted and validated for their use in the Spanish population.14,15 The RQLQ consists of 28 items distributed in 7 dimensions: sleep (3 items), non–hay fever symptoms (7 items), practical problems (3 items), nasal symptoms (4 items), eye symptoms (4 items), activities (3 items), and emotions (4 items). Responses to the items are scored on a 7-point Likert scale, whereas dimensions and overall scores are scored on a 0 to 6 scale. In both cases, the lower the score, the better the HRQL. The SF-1214 is a shortened version of the 36-item health survey. The SF-12 provides physical (PCS) and mental (MCS) composite summaries, with lower scores indicating worse health status. All scores were calculated by using standard scoring algorithms recommended by the authors, which yield a mean score of 50 and a SD of 10 in the Spanish general population.15
ARIA severity classification The 4 ARIA severity items (sleep, daily activities/sport, work/ school, and troublesome symptoms) provide various options for discriminating between patients with moderate and severe AR. For example, moderate patients could be defined as those in whom only 2 of the 4 aspects are affected, whereas severe patients would be those in whom 3 or 4 aspects are affected, although the ratio might be 1 aspect versus 2, 3, or 4 aspects, and so forth. Likewise, it might be that different domains carry different weight in terms of the impact on QOL; for example, the presence of problems in the ARIA sleep item might represent more of a negative effect on QOL than the presence of problems on the ARIA work and school domain. The analysis used in the current study was designed to determine how best to define moderate and severe AR using the ARIA criteria by taking into
Valero et al 361
J ALLERGY CLIN IMMUNOL VOLUME 120, NUMBER 2
TABLE I. Baseline characteristics of the sample of patients with AR (N 5 141)
85 (60.3) 32.3 (10.7) 29 62 49 7
(20.7) (44.3) (35.0) (6.9) [0-30]
48 (34) 69 (49) 51 (36) 90 (64) 24 32 48 37
(17) (23) (34) (26)
70 90 83 137 7.64
(50) (64) (60) (97) (2.2)
Rhinitis, sinusitis, and ocular diseases
Sex, n (%) Women Age, mean y (SD) Educational level, n (%) Primary education Secondary education University education Time from diagnosis, mean y (SD) [range] Comorbidities, n (%) Asthma Conjunctivitis Type of allergic rhinitis, n (%) Persistent Intermittent No. of QOL affected items, n (%) 1 item 2 items 3 items 4 items Affected item, n (%) Sleep Daily activities Work and school Symptoms TSS4,* mean (SD) RQLQ, mean (SD) Global score SF-12,à mean (SD) PCS MCS
2.62 (1.1) 47.5 (9.3) 47.2 (10.2)
*Scores range from 3 (mild symptoms) to 12 (severe symptoms). RQLQ global score range from 0 (no impairment) to 6 (greatest impairment). àSF-12 PCS and MCS have a mean of 50 and a SD of 10 in the Spanish general population. All scores above or below 50 are better or worse, respectively.
account both the number and type of ARIA domains affected and by examining their capacity to discriminate according to the HRQL. For the RQLQ questionnaire global score, a difference of 0.5 has been established as the minimal important difference.16 Assuming a SD of 1,16 a risk of 5%, statistical power of 80%, and an unbalanced number of patients in the moderate and severe groups (70% and 30%), a total of 152 was considered necessary to detect this minimal important difference of 0.5 points on the RQLQ global score between both independent groups.
Statistical analysis The statistical analysis consisted of several steps. First, the distribution of symptoms (TSS4) and HRQL (RQLQ global, SF-12 PCS and MCS) scores was examined using box plots. Second, a series of linear regression models were built to evaluate the degree of association of each of the 4 ARIA severity items with the symptoms and the HRQL scores. In these models, each ARIA severity item was included as an independent variable and the symptom of the HRQL score as the dependent variable. We specifically tested 2 issues: (1) whether the association of all the 4 ARIA severity items with symptoms and HRQL was similar, and (2) whether particular combinations of ARIA items implied some effect modification on symptoms and HRQL. To assess the latter, the significance of first, second, and third level interactions between the 4 ARIA items was
FIG 1. Distribution of patients with moderate/severe allergic rhinitis (N 5 141) on symptoms (TSS4) and HRQL (RQLQ global score, SF-12 MCS and PCS).
tested. Third, mean scores on the RQLQ, SF-12 PCS, SF-12 MCS, and TSS4 for groups of patients defined by the number of affected ARIA items were compared by using the ANOVA test for linear trend. If a statistically significant difference was found among groups, post hoc tests for pairwise comparisons (Tukey procedure) were performed. The effect size, a standardized difference of means, was calculated to assess the magnitude of the difference in scores between each pair of groups. Effect sizes were calculated as the difference between means divided by the pooled SD. Generally accepted guidelines define an effect size around 0.5 as moderate and 0.8 as large.17 Finally, box plots were produced to show the distribution of QOL scores in moderate and in severe patients with AR, and their means were compared by using the t test. Analyses were performed by using the Statistical Package for the Social Sciences (version 13.0 for Windows, Chicago, Ill). A P value of < .05 was considered statistically significant.
RESULTS The study sample included 141 nontreated patients with AR, and they are described in Table I. Mean age was 32.3 years old, 60.3% were women, and 79% of patients had completed secondary or university education. The AR was persistent in 36% and intermittent in 64% of patients. Main comorbidities were asthma (34%) and conjunctivitis (49%).
362 Valero et al
J ALLERGY CLIN IMMUNOL AUGUST 2007
TABLE II. Linear regression model to assess the association of the 4 ARIA severity items with quality of life (N 5 141)* RQLQ global score
Intercept Sleep Daily activities Work and school Symptoms R2 adjusted
SF-12 PCS
SF-12 MCS
b Coefficient
P value
b Coefficient
P value
b Coefficient
0.57 0.55 0.42 0.54 1.25 0.290
NS <.001 .02 .002 .05
56.45 –4.01 –4.36 –3.78 –2.16 0.199
<.001 .008 .01 .02 NS
50.2 –0.61 –1.31 –1.42 –0.95 0.013
TSS4
P value
b Coefficient
P value
3.98 0.94 0.44 0.89 2.45 0.161
<.001 .009 NS .02 .02
<.001 NS NS NS NS
*Model significance was P < .001.
TABLE III. HRQL scores (RQLQ, MCS and PCS SF-12) and symptoms (TSS4) of patients with AR according to the number of ARIA affected items (N 5 141) 1 Affected item
Rhinitis, sinusitis, and ocular diseases
No. of patients (%) RQLQ Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items) SF-12 PCS Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items) SF-12 MCS Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items) TSS4 Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items)
2 Affected items
3 Affected items
4 Affected items
P value* (post hoc pairwise comparisons with P < .05)
24 (17%)
32 (23%)
48 (34%)
37 (26%)
1.89 (1.10) — — —
2.27 (0.88) 0.38 — —
2.71 (0.76) 0.86 0.53 —
3.44 (0.85) 1.57 1.35 0.90
<.0001 (3 vs 1; 4 vs 1) (4 vs 2) (4 vs 3)
51.95 (7.82) — — —
51.50 (6.41) 0.06 — —
47.57 (8.41) 0.53 0.53 —
40.46 (9.32) 1.33 1.38 0.80
<.0001 (4 vs 1) (4 vs 2) (4 vs 3)
48.66 (8.89) — — —
48.68 (10.10) 0 — —
47.06 (10.29) 0.16 0.15 —
45.59 (11.00) 0.30 0.29 0.13
.5665
6.16 (2.35) — — —
7.37 (1.96) 0.55 — —
7.72 (2.01) 0.71 0.18 —
8.72 (1.95) 1.18 0.69 0.50
<.0001 (3 vs 1; 4 vs 1) (4 vs 2)
ES, Effect size. *P value of ANOVA test for linear trend. Post hoc statistically significant (P < .05) pairwise comparisons are shown in parentheses.
The RQLQ global scores of the study patients covered most of their maximum theoretical range, from 0 to 6 (Fig 1). The interquartile range of the RQLQ global score was 2.0 to 3.4, and the other 50% of patients were distributed between these limits and the observed minimum and maximum scores (0.35 and 5.5, respectively). On the TSS4, there was a slight tendency to score toward the higher end of the scale, with some patients having the maximum score of 12. The linear regression models constructed to evaluate the impact on QOL of the 4 items used to define AR severity, with RQLQ (global score), SF-12 (PCS and MCS), and symptoms (TSS4) as dependent variables, are shown in Table II. Interactions among the 4 AR severity items were nonsignificant and were therefore excluded from the final models. The 4 AR severity items were significantly associated with the RQLQ global score, whereas 3 of the items were also significantly associated with the SF-12 PCS and TSS4 in the respective regression model.
No item was significantly associated when the dependent variable was the MCS SF-12 score. The degree of association as measured by the b coefficients for each model was very similar for all items (ie, around 0.5 of the RQLQ global score) except for troublesome symptoms item, which was more strongly associated with RQLQ and TSS4 scores, and less strongly with the SF-12 PCS. The similarity of the b coefficients for each AR severity item would indicate a similar strength of association with QOL. None of the first-order, second-order, and third-order interactions were statistically significant, suggesting that the degree of association of each individual ARIA severity item is not modified when they are present in combination with any of the other severity items. As a result, our definition of moderate and severe AR was based only on the number of affected items. Significant differences were found in RQLQ, SF-12 PCS, and TSS4 scores among groups of patients with AR defined by the number of affected ARIA items (Table III).
J ALLERGY CLIN IMMUNOL VOLUME 120, NUMBER 2
Valero et al 363
Rhinitis, sinusitis, and ocular diseases
Post hoc tests for pairwise comparisons showed statistically significant differences between patients with 4 affected items and almost all other groups. The highest effect sizes were observed when comparing patients with 4 affected items with those with 3, 2, or 1 affected item: all effect sizes were greater than 0.8 when comparing groups using the RQLQ and SF-12 PCS, and greater than 0.5 when using the TSS4. Thus, the clearest distinction is between patients with 4 impaired items (severe), and those with 1, 2, or 3 impaired items (moderate). Following this definition, the final distribution of HRQL scores among moderate and severe patients with AR is shown in Fig 2. For instance, on the RQLQ global score, 75% of patients with moderate AR scored below 3 (median, 2.3; interquartile range, 1.6-3.0), whereas 75% of those with severe AR scored greater than 2.7 (median, 3.5; interquartile range, 2.7-4.0). A similar pattern was observed for the SF-12 PCS and TSS4.
DISCUSSION The main findings of the current study were (1) substantial heterogeneity in terms of symptom and QOL impairment (RQLQ and SF-12) in patients with AR classified as having moderate-severe disease; (2) all ARIA severity items used to define moderate/severe AR had a similar independent association with QOL, making it appropriate to establish a simple cutoff point on the number of affected items to divide between moderate and severe disease; and (3) there were significant differences on HRQL and symptoms measures, and the largest pairwise effect sizes were seen, when patients with 4 affected items (severe) were compared with patients with 1, 2, or 3 (moderate) affected items. The disproportionate relative size of the patients with AR classified under moderate/severe (69% to 90%)8-10,11 clearly supports the theoretical advantage of differentiating between moderate and severe rhinitis. However, it was also necessary a sufficient heterogeneity to justify the definition of 2 different severity grades from this group of patients. In our sample, this heterogeneity/dispersion has been demonstrated by evaluating the score of symptoms (TSS4) and the impairment of QOL (RQLQ and SF-12), and it supports the need of discrimination between moderate and severe patients rather than including both in a single moderate/severe group as recommended by the ARIA document.3 The discrimination between moderate and severe patients would be useful to obtain more homogeneous populations of patients, according to their severity, to design better epidemiological studies and clinical trials and develop disease management strategies.3,4,18-20 The QOL impairment has been useful to validate the clinical classification of severity in other chronic diseases such as asthma21 and chronic obstructive pulmonary disease (COPD).22,23 An excellent correlation has been found between both clinical classification and impairment of QOL in asthma (Global Initiative for Asthma) and COPD (Global Initiative for Chronic Obstructive Lung Disease).21,22,24
FIG 2. Distribution of patients when defined by the number of affected ARIA items in severe (4 impaired items) and moderate (those with 1, 2, or 3 impaired items). Mean 6 SD and P value of t test comparing moderate and severe groups were as follows: for TSS4, symptom score 7.3 6 2.2 vs 8.7 6 2.0 (P < .0004); for RQLQ, global score 2.4 6 1.0 vs 3.5 6 0.9 (P < .0001); for SF-12 PCS, score 49.8 6 7.9 vs 40.5 6 9.3 (P < .0001); and for SF-12 MCS, score 47.9 6 9.9 vs 45.6 6 11.0 (NS).
A good correlation between the severity of persistent AR evaluated by symptoms score and by QOL assessment using the RQLQ and SF-12 has been reported recently.25 When comparing HRQL among the 4 groups of patients with AR defined by the number of affected ARIA items, significant differences between some groups were observed in pairwise analysis. However, given the higher effect sizes found when comparing patients with 4 affected items to all other groups (ie, those with 1, 2, or 3 affected items), and given the fact that no interaction effect between the ARIA severity items was observed, we considered the most appropriate criterion for classifying patients with AR would be in terms of the number of items
364 Valero et al
Rhinitis, sinusitis, and ocular diseases
affected, with moderate patients defined as those with 1, 2, or 3 affected items and severe patients as those with all 4 items affected. Regarding the discriminative capacity of this cutoff point, it is notable that the between group effect sizes obtained when comparing groups using the RQLQ and SF-12 PCS were larger than 0.8, and would therefore be considered large. Furthermore, taking into account that the minimal important difference for the RQLQ has been established on 0.5, the mean difference of 1.1 obtained between moderate and severe groups defined by applying this cutoff point seems striking from a quantitative point of view. On the other hand, using this criterion, 74% and 26% of the population studied would be classified as having moderate or severe AR, respectively. These severity percentages are in agreement with those reported for other chronic respiratory diseases such as asthma and COPD, for which the proportion of patients with severe disease is always lower than the proportion with moderate disease.26,27 The modification of the ARIA severity criteria proposed by van Hoecke et al11 has raised a new empirical model to define moderate/severe rhinitis by removing the item of troublesome symptoms and recombining the other three items in 2 questions: sleep disturbance and impairment of daily personal (daily activities, leisure, sports) and/or professional (school, work) life. In their model, AR severity is finally classified into 3 levels: mild (for patients answering no to both questions), moderate (for patients answering yes to some of them), and severe (for those answering yes to both questions). In their study, the claim of those criteria in a sample of 804 patients with AR resulted in 20.5% with mild, 45.9% with moderate, and 33.6% with severe disease. The current study and the study by van Hoecke et al11 match both in the scientific basis and in the proposal of classification by sharing the need to discriminate between moderate and severe AR. However, no current ARIA severity item would be deleted or modified in our proposal. In spite of deleting troublesome symptoms, the less discriminative item between moderate and severe AR because of its high prevalence (86.9% of the patients with moderate-severe AR), we propose that this item should remain to discriminate moderate from mild AR. Our study had some limitations. First, because no information on patients who refused to participate in the study was recorded, the response rate could vary according to severity stage. For example, if patients with fewer or milder symptoms are more likely to refuse, the prevalence of patients with moderate disease would be underestimated. However, this would not affect the strength of the association between ARIA severity items and HRQL on which the selection of the cutoff point is based. Second, the sample of patients (N 5 141) in our study did not reach the calculated sample size of 152 patients. However, the shortfall of 11 patients did not affect the statistical significance of the results, given that the observed mean difference between moderate and severe groups (2.4 vs 3.5; P < .0001) was higher than the minimal important
J ALLERGY CLIN IMMUNOL AUGUST 2007
difference of the RQLQ global score (0.5) used to estimate the sample size. This is the first study to estimate the relative effect of each severity ARIA item on the QOL of patients with AR, and to assess the strengths of different possible cutoff points in discriminating between patients with moderate and severe AR. The similarity in the QOL effect of the 4 severity items as well as the nonsignificant interactions found in our study support the validity and usefulness of the 4 current ARIA items for the classification of AR severity. In conclusion, we propose a new criterion to discriminate moderate from severe allergic rhinitis, based on the number of affected items following the ARIA severity classification: mild (not affected items), moderate (1, 2, or 3 affected items), and severe (4 affected items). Although further studies are required, we believe that the discrimination between moderate and severe AR is necessary and could have important in developing disease management strategies. We thank all of the investigators participating in the different phases of the EStudio de calidad de vida en Pacientes con RINoconjuncTivitis (ESPRINT) project. REFERENCES 1. Bauchau V, Durham SR. Prevalence and rate of diagnosis of allergic rhinitis in Europe. Eur Respir J 2004;24:758-64. 2. Bousquet J, Demarteau N, Mullol J, van den Akker-van Marle ME, van Ganse E, Bachert C. Costs associated with persistent allergic rhinitis are reduced by levocetirizine. Allergy 2005;60:788-94. 3. Bousquet J, van Cauwenberge P, Khaltaev N. ARIA Workshop Group, World Health Organization. Allergic rhinitis and its impact on asthma. J Allergy Clin Immunol 2001;108:S147-334. 4. Bousquet J, Neukirch F, Bousquet PJ, Gehano P, Klossek JM, Le Gal M, et al. Severity and impairment of allergic rhinitis in patients consulting in primary care. J Allergy Clin Immunol 2006;117:158-62. 5. Ciprandi G, Cirillo I, Vizzaccaro A, Tosca M, Passalacqua G, Pallestrini E, et al. Seasonal and perennial allergic rhinitis: is this classification adherent to real life? Allergy 2005;60:882-7. 6. Demoly P, Allaert FA, Lecasble M, Bousquet J. Validation of the classification of ARIA (allergic rhinitis and its impact on asthma). Allergy 2003;58:672-5. 7. Bauchau V, Durham SR. Epidemiological characterization of the intermittent and persistent types of allergic rhinitis. Allergy 2005;60:350-3. 8. Bousquet J, Annesi-Maesano I, Carat F, Leger D, Rugina M, Pribil C, et al. Characteristics of intermittent and persistent allergic rhinitis: DREAMS study group. Clin Exp Allergy 2005;35:728-32. 9. Bachert C, van Cauwenberge P, Olbrecht J, van Schoor J. Prevalence, classification and perception of allergic and nonallergic rhinitis in Belgium. Allergy 2006;61:693-8. 10. Pereira C, Valero A, Loureiro C, Davila I, Martinez-Cocera C, Murio C, et al. Iberian study of aeroallergens sensitization in allergic rhinitis. Allerg Immunol (Paris) 2006;38:186-94. 11. van Hoecke H, Vastesaeger N, Dewulf L, De Bacquer D, van Cauwenberge P. Is the allergic rhinitis and its impact on asthma classification useful in daily primary care practice? J Allergy Clin Immunol 2006;118:758-9. 12. Juniper EF, Guyatt GH. Development and testing of a new measure of health status for clinical trials in rhinoconjunctivitis. Clin Exp Allergy 1991;21:77-83. 13. Ware JE, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care 1996;34:220-33. 14. Soler R, de la Hoz B, Badia X, Mercadal J, Lozano R, Benavides A, et al. Validation of the Spanish version of the Rhinoconjunctivitis Quality of Life Questionnaire (RQLQ). Rev Clin Esp 2004;204:131-8.
Valero et al 365
J ALLERGY CLIN IMMUNOL VOLUME 120, NUMBER 2
23.
24.
25.
26.
27.
EuroQol five-dimension questionnaire differentiate between COPD severity stages? Chest 2006;130:117-28. Ferrer M, Alonso J, Morera J, Marrades R, Khalaf A, Aguar MC, et al. Chronic obstructive pulmonary disease stage and health-related quality of life. The Quality of Life of Chronic Obstructive Pulmonary Disease Study Group. Ann Intern Med 1997;127:1072-9. Workshop Report, Global Strategy for Asthma Management and Prevention (GINA). Initiative World Health Organization, WHO. Updated November 2006. Available at: http://www.ginasthma.com/. Accessed March 30, 2007. Canonica GW, Bousquet J, van Hammee G, Bachert C, Durham SR, Klimek L, et al. Levoceterizine improves health-related quality of life and health status in persistent allergic rhinitis. Respir Med 2006;100: 1706-15. Liard R, Leynaert B, Zureik M, Beguin FX, Neukirch F. Using Global Initiative for Asthma Guidelines to assess asthma severity in populations. Eur Respir J 2000;16:615-20. Hoogendoorn M, Feenstra TL, Schermer TR, Hesselink AE, Rutten-van Molken MP. Severity distribution of chronic obstructive pulmonary disease (COPD) in Dutch general practice. Respir Med 2006;100: 83-6.
Rhinitis, sinusitis, and ocular diseases
15. Vilagut G, Ferrer M, Rajmil L, Rebollo P, Permanyer-Miralda G, Quintana JM, et al. The Spanish version of the Short Form 36 Health Survey: a decade of experience and new developments. Gac Sanit 2005;19:135-50. 16. Juniper EF, Guyatt GH, Griffith LE, Ferrie PJ. Interpretation of rhinoconjunctivitis quality of life questionnaire data. J Allergy Clin Immunol 1996;98:843-5. 17. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale (NJ): L. Erlbaum Associates; 1988. 18. Bousquet J, van Cauwenberge P, Ait Khaled N, Bachert C, Baena-Cagnani CE, Bouchard J, et al. Pharmacologic and anti-IgE treatment of allergic rhinitis ARIA update (in colaboration with GA2LEN). Allergy 2006;61:1086-96. 19. Prenner BM, Schenkel E. Allergic rhinitis: treatment based on patients profiles. Am J Med 2006;119:230-7. 20. Hoecke H, Vastesaeger N, Dewulf L, Sys L, van Cauwenberge P. Classification and management of allergic rhinitis patients in general practice during pollen season. Allergy 2006;61:705-11. 21. Fonseca JA, Delgado L, Costa-Pereira A, Tavares C, Moreira A, Morete A, et al. Evaluation of the Asthma Life Quality test for the screening and severity assessment of asthma. Allergy 2004;59:1198-204. 22. Rutten-vanMolken MP, Oostenbrink JB, Tashkin DP, Burkhart D, Monz BU. Does quality of life of COPD patients as measured by the generic
Correction With regard to the January 2005 article entitled ‘‘The September epidemic of asthma exacerbations in children: A search for etiology’’ (2005;115:132-8), the legend for Figure 2 gives an incorrect date range. The correct dates are April 2001 to March 2002.
Antonio Valero, MD, PhD,a,m Montse Ferrer, MD, PhD,b Joaquı´n Sastre, MD, PhD,c Ana M. Navarro, MD, PhD,d Laura Monclu´s, MD,e Enrique Martı´-Guadan˜o, MD, PhD,f Michael Herdman, MSc,b,g Ignacio Da´vila, MD, PhD,h Alfonso del Cuvillo, MD, PhD,i Carlos Cola´s, MD, PhD,j Eva Baro´, BSc,g Ignacio Ante´para, MD, PhD,k Jordi Alonso, MD, PhD,b and Joaquim Mullol, MD, PhDl,m Barcelona, Madrid, Seville, Badalona, Salamanca, Ca´diz, Zaragoza, and Bilbao, Spain
Background: Allergic Rhinitis and its Impact on Asthma (ARIA) differentiates mild from moderate/severe patients on the basis of 4 severity items. The high prevalence of moderate/ severe patients suggests the need to differentiate between them. Objectives: To identify the categorization that maximizes discrimination between moderate and severe allergic rhinitis (AR) by using ARIA guidelines. Methods: Observational, cross-sectional study. Clinical characteristics, nasal symptoms (Total Symptom Score 4), and health-related quality of life (HRQL; Rhinoconjunctivitis Quality of Life Questionnaire and Short Form 12) were assessed. The association of severity items (sleep, daily activities/sport, work/ school, and troublesome symptoms) with symptoms and HRQL was analyzed using linear regression models. ANOVA and effect sizes were used to assess differences in symptoms and HRQL among groups defined by the number of affected ARIA items. From athe Unidad de Alergia, Servicio de Neumologı´a y Alergia Respiratoria, Hospital Clı´nic, Barcelona; bthe Health Services Research Unit, Institut Municipal d’Investigacio´ Me`dica, Barcelona; cthe Servicio de Alergia, Fundacio´n Jime´nez Dı´az, Madrid; dthe Unidad de Alergia, Hospital El Tomillar, Dos Hermanas, Sevilla; eJ Uriach & Co, Barcelona; fthe Unidad de Alergia, Hospital Germans Trias i Pujol, Badalona; g3D Health Research, Barcelona; hthe Servicio de Alergia, Hospital Clı´nico, Salamanca; ithe Clı´nica Dr Lobato´n, Ca´diz; jthe Servicio de Alergia, Hospital Clı´nico ‘‘Lozano Blesa,’’ Zaragoza; kthe Servicio de Alergia, Hospital de Basurto, Bilbao; lthe Unitat de Rinologia i Clinica de l’Olfacte, Servei d’Otorinolaringologia, Hospital Clı´nic, Institut d’Investigacions Biome`diques August Pi i Sunyer (IDIBAPS), Barcelona; and mthe Centro de Investigacio´n Biome´dica en Red (CIBER) de Enfermedades Respiratorias, Barcelona. Supported by an unrestricted educational grant from Uriach & Co, Barcelona, Spain. Disclosure of potential conflict of interest: The authors have declared that they have no conflict of interest. Received for publication December 4, 2006; revised March 30, 2007; accepted for publication April 5, 2007. Available online May 26, 2007. Reprint requests: Antonio Valero, MD, PhD, Unidad de Alergia, Servicio de Neumologı´a y Alergia Respiratoria, ICT, Hospital Clı´nic, C Villarroel, 170, 08036 Barcelona, Spain. E-mail: [email protected]. 0091-6749/$32.00 Ó 2007 American Academy of Allergy, Asthma & Immunology doi:10.1016/j.jaci.2007.04.006
Results: Nontreated patients (N 5 141) with moderate/severe AR were studied. All severity items showed a similar independent association with symptoms and HRQL scores, and there were no interaction effects, indicating that categorization of patients into moderate and severe could be based only on the number of items affected. Effect sizes were highest between patients with 4 affected ARIA items and those with 3, 2, or 1 affected item (effect sizes greater than 0.8 in all comparisons using Rhinoconjunctivitis Quality of Life Questionnaire and Short Form 12 Physical Composite Summary, and greater than 0.5 using the Total Symptom Score 4; P < .001). Conclusion: Using ARIA severity items, the criterion that best discriminates AR severity is considering moderate those with 1 to 3 affected items and severe those with 4. Clinical implications: Discrimination between patients with moderate and severe AR should help to obtain homogeneous populations for both research and clinical purposes. (J Allergy Clin Immunol 2007;120:359-65.) Key words: Allergic rhinitis, severity criteria, ARIA, quality of life
Allergic rhinitis (AR) is a highly common and increasingly prevalent disease that generates high social and health care costs and has a significant effect on quality of life (QOL).1-3 Allergic rhinitis has traditionally been classified as perennial or seasonal depending on the allergen responsible for symptoms. This classification has been found to be clinically unsatisfactory for at least 2 reasons: (1) the frequency of allergen polysensitization is increasing among patients with AR, and (2) pollens and mold may cause perennial symptoms whereas house dust mite may provoke seasonal symptoms in some patients.4,5 The Allergic Rhinitis and its Impact on Asthma (ARIA) document developed a new classification, intermittent allergic rhinitis (IAR) and persistent allergic rhinitis (PAR), on the basis of symptoms duration. The ARIA classification also introduced a system for assessing AR severity on the basis of the presence or absence of 359
Rhinitis, sinusitis, and ocular diseases
A new criterion by which to discriminate between patients with moderate allergic rhinitis and patients with severe allergic rhinitis based on the Allergic Rhinitis and its Impact on Asthma severity items
360 Valero et al
Abbreviations used AR: Allergic rhinitis ARIA: Allergic Rhinitis and its Impact on Asthma COPD: Chronic obstructive pulmonary disease HRQL: Health-related quality of life IAR: Intermittent allergic rhinitis MCS: Mental composite summary PAR: Persistent allergic rhinitis PCS: Physical composite summary QOL: Quality of life RQLQ: Rhinoconjunctivitis Quality of Life Questionnaire SF-12: Medical Outcomes Study 12-Item Short-Form Health Survey TSS: Total Symptom Score
Rhinitis, sinusitis, and ocular diseases
impairment in any of 4 health-related quality of life (HRQL) items: sleep, daily activities/sport, work/school, and troublesome symptoms. According to the ARIA workshop group, AR was defined as mild when there was no impairment in any of these items, and moderate/severe when there was impairment in 1 or more areas. One of the merits of this classification is that, for the first time, it explicitly used HRQL criteria to assess severity.3 These new classifications (symptoms duration and disease severity) were based on expert consensus and required further validation. In recent years, the symptoms duration classification (IAR vs PAR) was shown to be valid in several large epidemiologic studies.6-10 However, the validity of the severity classification has not been established yet. Moreover, the large prevalence of moderate/severe rhinitis (69% of patients with rhinitis consulting at otolaryngology and allergy clinics and 90% of those consulting general practitioners)4,8-10 suggests an important heterogeneity in this disease severity group. In addition, some concerns have been raised on the grading of severity proposed in the ARIA guidelines.4,11 Bousquet et al4 proposed the replacement of the term ‘‘moderate/severe’’ by ‘‘severe’’ because of the high proportion of patients reporting impaired activities. Recent recommendations of differentiating between moderate and severe rhinitis and therefore modifying the current ARIA severity criteria have been made.11 The objectives of the study were (1) to estimate the heterogeneity of symptoms and HRQL scores in patients classified as moderate/severe according to current ARIA severity classification criteria, (2) to assess the degree of association of each of the ARIA severity items with symptoms and HRQL, and (3) to identify the categorization of the ARIA items that maximized the differentiation between moderate and severe AR cases in terms of symptoms and HRQL.
METHODS Study design and participants An observational, cross-sectional study was performed between March and August 2004 in the allergy and otorhinolaryngology
J ALLERGY CLIN IMMUNOL AUGUST 2007
departments of 27 Spanish hospitals. This period of 6 months was planned to include 2 different seasons, spring and summer, because in Spain, exacerbations of rhinitis induced by some allergens mainly appear in spring and fall (ie, house dust mites, pollens). Each center consecutively included patients with a diagnosis of AR based on any clinically relevant allergen sensitization who met inclusion criteria and agreed to participate. Inclusion criteria were (1) age older than 18 years, (2) diagnosis of IAR or PAR, (3) classified as moderate/severe according to the ARIA guidelines,3 and (4) nontreated at the time of inclusion. The period without treatment had to be at least 8 and 4 weeks in the case of oral and intranasal corticosteroids, respectively, and 2 weeks for oral and topical antihistamines and antileukotrienes. None of the patients included in the study were initiating specific immunotherapy. All patients provided informed consent to participate in the study, and the study was approved by the Ethics Committee of the Hospital Clı´nic (Barcelona).
Patient evaluation Patient assessments were performed on inclusion in the study. Sociodemographic data collected included sex, age, and educational level, and clinical data collected included duration of AR (days with symptoms per week and number of consecutive weeks with symptoms), AR severity (number of impaired items on the ARIA classification), time since diagnosis, comorbidities, presence of treatment for AR, and symptom severity. Allergic rhinitis symptoms were assessed using the Total Symptom Score (TSS) 4 by calculating the sum of scores for nasal obstruction, rhinorrhea, nasal itching, and sneezing. Each nasal symptom was scored on a scale from 0 to 3 (0, no symptoms; 1, mild; 2, moderate; 3, severe). resulting in a TSS4 score ranging from 0 to 12. Health-related quality of life assessment included a diseasespecific (the Rhinoconjunctivitis Quality of Life Questionnaire [RQLQ])12 and a generic (the Medical Outcomes Study 12-Item Short-Form Health Survey [SF-12])13 instrument. Both of them have been adapted and validated for their use in the Spanish population.14,15 The RQLQ consists of 28 items distributed in 7 dimensions: sleep (3 items), non–hay fever symptoms (7 items), practical problems (3 items), nasal symptoms (4 items), eye symptoms (4 items), activities (3 items), and emotions (4 items). Responses to the items are scored on a 7-point Likert scale, whereas dimensions and overall scores are scored on a 0 to 6 scale. In both cases, the lower the score, the better the HRQL. The SF-1214 is a shortened version of the 36-item health survey. The SF-12 provides physical (PCS) and mental (MCS) composite summaries, with lower scores indicating worse health status. All scores were calculated by using standard scoring algorithms recommended by the authors, which yield a mean score of 50 and a SD of 10 in the Spanish general population.15
ARIA severity classification The 4 ARIA severity items (sleep, daily activities/sport, work/ school, and troublesome symptoms) provide various options for discriminating between patients with moderate and severe AR. For example, moderate patients could be defined as those in whom only 2 of the 4 aspects are affected, whereas severe patients would be those in whom 3 or 4 aspects are affected, although the ratio might be 1 aspect versus 2, 3, or 4 aspects, and so forth. Likewise, it might be that different domains carry different weight in terms of the impact on QOL; for example, the presence of problems in the ARIA sleep item might represent more of a negative effect on QOL than the presence of problems on the ARIA work and school domain. The analysis used in the current study was designed to determine how best to define moderate and severe AR using the ARIA criteria by taking into
Valero et al 361
J ALLERGY CLIN IMMUNOL VOLUME 120, NUMBER 2
TABLE I. Baseline characteristics of the sample of patients with AR (N 5 141)
85 (60.3) 32.3 (10.7) 29 62 49 7
(20.7) (44.3) (35.0) (6.9) [0-30]
48 (34) 69 (49) 51 (36) 90 (64) 24 32 48 37
(17) (23) (34) (26)
70 90 83 137 7.64
(50) (64) (60) (97) (2.2)
Rhinitis, sinusitis, and ocular diseases
Sex, n (%) Women Age, mean y (SD) Educational level, n (%) Primary education Secondary education University education Time from diagnosis, mean y (SD) [range] Comorbidities, n (%) Asthma Conjunctivitis Type of allergic rhinitis, n (%) Persistent Intermittent No. of QOL affected items, n (%) 1 item 2 items 3 items 4 items Affected item, n (%) Sleep Daily activities Work and school Symptoms TSS4,* mean (SD) RQLQ, mean (SD) Global score SF-12,à mean (SD) PCS MCS
2.62 (1.1) 47.5 (9.3) 47.2 (10.2)
*Scores range from 3 (mild symptoms) to 12 (severe symptoms). RQLQ global score range from 0 (no impairment) to 6 (greatest impairment). àSF-12 PCS and MCS have a mean of 50 and a SD of 10 in the Spanish general population. All scores above or below 50 are better or worse, respectively.
account both the number and type of ARIA domains affected and by examining their capacity to discriminate according to the HRQL. For the RQLQ questionnaire global score, a difference of 0.5 has been established as the minimal important difference.16 Assuming a SD of 1,16 a risk of 5%, statistical power of 80%, and an unbalanced number of patients in the moderate and severe groups (70% and 30%), a total of 152 was considered necessary to detect this minimal important difference of 0.5 points on the RQLQ global score between both independent groups.
Statistical analysis The statistical analysis consisted of several steps. First, the distribution of symptoms (TSS4) and HRQL (RQLQ global, SF-12 PCS and MCS) scores was examined using box plots. Second, a series of linear regression models were built to evaluate the degree of association of each of the 4 ARIA severity items with the symptoms and the HRQL scores. In these models, each ARIA severity item was included as an independent variable and the symptom of the HRQL score as the dependent variable. We specifically tested 2 issues: (1) whether the association of all the 4 ARIA severity items with symptoms and HRQL was similar, and (2) whether particular combinations of ARIA items implied some effect modification on symptoms and HRQL. To assess the latter, the significance of first, second, and third level interactions between the 4 ARIA items was
FIG 1. Distribution of patients with moderate/severe allergic rhinitis (N 5 141) on symptoms (TSS4) and HRQL (RQLQ global score, SF-12 MCS and PCS).
tested. Third, mean scores on the RQLQ, SF-12 PCS, SF-12 MCS, and TSS4 for groups of patients defined by the number of affected ARIA items were compared by using the ANOVA test for linear trend. If a statistically significant difference was found among groups, post hoc tests for pairwise comparisons (Tukey procedure) were performed. The effect size, a standardized difference of means, was calculated to assess the magnitude of the difference in scores between each pair of groups. Effect sizes were calculated as the difference between means divided by the pooled SD. Generally accepted guidelines define an effect size around 0.5 as moderate and 0.8 as large.17 Finally, box plots were produced to show the distribution of QOL scores in moderate and in severe patients with AR, and their means were compared by using the t test. Analyses were performed by using the Statistical Package for the Social Sciences (version 13.0 for Windows, Chicago, Ill). A P value of < .05 was considered statistically significant.
RESULTS The study sample included 141 nontreated patients with AR, and they are described in Table I. Mean age was 32.3 years old, 60.3% were women, and 79% of patients had completed secondary or university education. The AR was persistent in 36% and intermittent in 64% of patients. Main comorbidities were asthma (34%) and conjunctivitis (49%).
362 Valero et al
J ALLERGY CLIN IMMUNOL AUGUST 2007
TABLE II. Linear regression model to assess the association of the 4 ARIA severity items with quality of life (N 5 141)* RQLQ global score
Intercept Sleep Daily activities Work and school Symptoms R2 adjusted
SF-12 PCS
SF-12 MCS
b Coefficient
P value
b Coefficient
P value
b Coefficient
0.57 0.55 0.42 0.54 1.25 0.290
NS <.001 .02 .002 .05
56.45 –4.01 –4.36 –3.78 –2.16 0.199
<.001 .008 .01 .02 NS
50.2 –0.61 –1.31 –1.42 –0.95 0.013
TSS4
P value
b Coefficient
P value
3.98 0.94 0.44 0.89 2.45 0.161
<.001 .009 NS .02 .02
<.001 NS NS NS NS
*Model significance was P < .001.
TABLE III. HRQL scores (RQLQ, MCS and PCS SF-12) and symptoms (TSS4) of patients with AR according to the number of ARIA affected items (N 5 141) 1 Affected item
Rhinitis, sinusitis, and ocular diseases
No. of patients (%) RQLQ Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items) SF-12 PCS Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items) SF-12 MCS Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items) TSS4 Mean (SD) ES (vs 1 item) ES (vs 2 items) ES (vs 3 items)
2 Affected items
3 Affected items
4 Affected items
P value* (post hoc pairwise comparisons with P < .05)
24 (17%)
32 (23%)
48 (34%)
37 (26%)
1.89 (1.10) — — —
2.27 (0.88) 0.38 — —
2.71 (0.76) 0.86 0.53 —
3.44 (0.85) 1.57 1.35 0.90
<.0001 (3 vs 1; 4 vs 1) (4 vs 2) (4 vs 3)
51.95 (7.82) — — —
51.50 (6.41) 0.06 — —
47.57 (8.41) 0.53 0.53 —
40.46 (9.32) 1.33 1.38 0.80
<.0001 (4 vs 1) (4 vs 2) (4 vs 3)
48.66 (8.89) — — —
48.68 (10.10) 0 — —
47.06 (10.29) 0.16 0.15 —
45.59 (11.00) 0.30 0.29 0.13
.5665
6.16 (2.35) — — —
7.37 (1.96) 0.55 — —
7.72 (2.01) 0.71 0.18 —
8.72 (1.95) 1.18 0.69 0.50
<.0001 (3 vs 1; 4 vs 1) (4 vs 2)
ES, Effect size. *P value of ANOVA test for linear trend. Post hoc statistically significant (P < .05) pairwise comparisons are shown in parentheses.
The RQLQ global scores of the study patients covered most of their maximum theoretical range, from 0 to 6 (Fig 1). The interquartile range of the RQLQ global score was 2.0 to 3.4, and the other 50% of patients were distributed between these limits and the observed minimum and maximum scores (0.35 and 5.5, respectively). On the TSS4, there was a slight tendency to score toward the higher end of the scale, with some patients having the maximum score of 12. The linear regression models constructed to evaluate the impact on QOL of the 4 items used to define AR severity, with RQLQ (global score), SF-12 (PCS and MCS), and symptoms (TSS4) as dependent variables, are shown in Table II. Interactions among the 4 AR severity items were nonsignificant and were therefore excluded from the final models. The 4 AR severity items were significantly associated with the RQLQ global score, whereas 3 of the items were also significantly associated with the SF-12 PCS and TSS4 in the respective regression model.
No item was significantly associated when the dependent variable was the MCS SF-12 score. The degree of association as measured by the b coefficients for each model was very similar for all items (ie, around 0.5 of the RQLQ global score) except for troublesome symptoms item, which was more strongly associated with RQLQ and TSS4 scores, and less strongly with the SF-12 PCS. The similarity of the b coefficients for each AR severity item would indicate a similar strength of association with QOL. None of the first-order, second-order, and third-order interactions were statistically significant, suggesting that the degree of association of each individual ARIA severity item is not modified when they are present in combination with any of the other severity items. As a result, our definition of moderate and severe AR was based only on the number of affected items. Significant differences were found in RQLQ, SF-12 PCS, and TSS4 scores among groups of patients with AR defined by the number of affected ARIA items (Table III).
J ALLERGY CLIN IMMUNOL VOLUME 120, NUMBER 2
Valero et al 363
Rhinitis, sinusitis, and ocular diseases
Post hoc tests for pairwise comparisons showed statistically significant differences between patients with 4 affected items and almost all other groups. The highest effect sizes were observed when comparing patients with 4 affected items with those with 3, 2, or 1 affected item: all effect sizes were greater than 0.8 when comparing groups using the RQLQ and SF-12 PCS, and greater than 0.5 when using the TSS4. Thus, the clearest distinction is between patients with 4 impaired items (severe), and those with 1, 2, or 3 impaired items (moderate). Following this definition, the final distribution of HRQL scores among moderate and severe patients with AR is shown in Fig 2. For instance, on the RQLQ global score, 75% of patients with moderate AR scored below 3 (median, 2.3; interquartile range, 1.6-3.0), whereas 75% of those with severe AR scored greater than 2.7 (median, 3.5; interquartile range, 2.7-4.0). A similar pattern was observed for the SF-12 PCS and TSS4.
DISCUSSION The main findings of the current study were (1) substantial heterogeneity in terms of symptom and QOL impairment (RQLQ and SF-12) in patients with AR classified as having moderate-severe disease; (2) all ARIA severity items used to define moderate/severe AR had a similar independent association with QOL, making it appropriate to establish a simple cutoff point on the number of affected items to divide between moderate and severe disease; and (3) there were significant differences on HRQL and symptoms measures, and the largest pairwise effect sizes were seen, when patients with 4 affected items (severe) were compared with patients with 1, 2, or 3 (moderate) affected items. The disproportionate relative size of the patients with AR classified under moderate/severe (69% to 90%)8-10,11 clearly supports the theoretical advantage of differentiating between moderate and severe rhinitis. However, it was also necessary a sufficient heterogeneity to justify the definition of 2 different severity grades from this group of patients. In our sample, this heterogeneity/dispersion has been demonstrated by evaluating the score of symptoms (TSS4) and the impairment of QOL (RQLQ and SF-12), and it supports the need of discrimination between moderate and severe patients rather than including both in a single moderate/severe group as recommended by the ARIA document.3 The discrimination between moderate and severe patients would be useful to obtain more homogeneous populations of patients, according to their severity, to design better epidemiological studies and clinical trials and develop disease management strategies.3,4,18-20 The QOL impairment has been useful to validate the clinical classification of severity in other chronic diseases such as asthma21 and chronic obstructive pulmonary disease (COPD).22,23 An excellent correlation has been found between both clinical classification and impairment of QOL in asthma (Global Initiative for Asthma) and COPD (Global Initiative for Chronic Obstructive Lung Disease).21,22,24
FIG 2. Distribution of patients when defined by the number of affected ARIA items in severe (4 impaired items) and moderate (those with 1, 2, or 3 impaired items). Mean 6 SD and P value of t test comparing moderate and severe groups were as follows: for TSS4, symptom score 7.3 6 2.2 vs 8.7 6 2.0 (P < .0004); for RQLQ, global score 2.4 6 1.0 vs 3.5 6 0.9 (P < .0001); for SF-12 PCS, score 49.8 6 7.9 vs 40.5 6 9.3 (P < .0001); and for SF-12 MCS, score 47.9 6 9.9 vs 45.6 6 11.0 (NS).
A good correlation between the severity of persistent AR evaluated by symptoms score and by QOL assessment using the RQLQ and SF-12 has been reported recently.25 When comparing HRQL among the 4 groups of patients with AR defined by the number of affected ARIA items, significant differences between some groups were observed in pairwise analysis. However, given the higher effect sizes found when comparing patients with 4 affected items to all other groups (ie, those with 1, 2, or 3 affected items), and given the fact that no interaction effect between the ARIA severity items was observed, we considered the most appropriate criterion for classifying patients with AR would be in terms of the number of items
364 Valero et al
Rhinitis, sinusitis, and ocular diseases
affected, with moderate patients defined as those with 1, 2, or 3 affected items and severe patients as those with all 4 items affected. Regarding the discriminative capacity of this cutoff point, it is notable that the between group effect sizes obtained when comparing groups using the RQLQ and SF-12 PCS were larger than 0.8, and would therefore be considered large. Furthermore, taking into account that the minimal important difference for the RQLQ has been established on 0.5, the mean difference of 1.1 obtained between moderate and severe groups defined by applying this cutoff point seems striking from a quantitative point of view. On the other hand, using this criterion, 74% and 26% of the population studied would be classified as having moderate or severe AR, respectively. These severity percentages are in agreement with those reported for other chronic respiratory diseases such as asthma and COPD, for which the proportion of patients with severe disease is always lower than the proportion with moderate disease.26,27 The modification of the ARIA severity criteria proposed by van Hoecke et al11 has raised a new empirical model to define moderate/severe rhinitis by removing the item of troublesome symptoms and recombining the other three items in 2 questions: sleep disturbance and impairment of daily personal (daily activities, leisure, sports) and/or professional (school, work) life. In their model, AR severity is finally classified into 3 levels: mild (for patients answering no to both questions), moderate (for patients answering yes to some of them), and severe (for those answering yes to both questions). In their study, the claim of those criteria in a sample of 804 patients with AR resulted in 20.5% with mild, 45.9% with moderate, and 33.6% with severe disease. The current study and the study by van Hoecke et al11 match both in the scientific basis and in the proposal of classification by sharing the need to discriminate between moderate and severe AR. However, no current ARIA severity item would be deleted or modified in our proposal. In spite of deleting troublesome symptoms, the less discriminative item between moderate and severe AR because of its high prevalence (86.9% of the patients with moderate-severe AR), we propose that this item should remain to discriminate moderate from mild AR. Our study had some limitations. First, because no information on patients who refused to participate in the study was recorded, the response rate could vary according to severity stage. For example, if patients with fewer or milder symptoms are more likely to refuse, the prevalence of patients with moderate disease would be underestimated. However, this would not affect the strength of the association between ARIA severity items and HRQL on which the selection of the cutoff point is based. Second, the sample of patients (N 5 141) in our study did not reach the calculated sample size of 152 patients. However, the shortfall of 11 patients did not affect the statistical significance of the results, given that the observed mean difference between moderate and severe groups (2.4 vs 3.5; P < .0001) was higher than the minimal important
J ALLERGY CLIN IMMUNOL AUGUST 2007
difference of the RQLQ global score (0.5) used to estimate the sample size. This is the first study to estimate the relative effect of each severity ARIA item on the QOL of patients with AR, and to assess the strengths of different possible cutoff points in discriminating between patients with moderate and severe AR. The similarity in the QOL effect of the 4 severity items as well as the nonsignificant interactions found in our study support the validity and usefulness of the 4 current ARIA items for the classification of AR severity. In conclusion, we propose a new criterion to discriminate moderate from severe allergic rhinitis, based on the number of affected items following the ARIA severity classification: mild (not affected items), moderate (1, 2, or 3 affected items), and severe (4 affected items). Although further studies are required, we believe that the discrimination between moderate and severe AR is necessary and could have important in developing disease management strategies. We thank all of the investigators participating in the different phases of the EStudio de calidad de vida en Pacientes con RINoconjuncTivitis (ESPRINT) project. REFERENCES 1. Bauchau V, Durham SR. Prevalence and rate of diagnosis of allergic rhinitis in Europe. Eur Respir J 2004;24:758-64. 2. Bousquet J, Demarteau N, Mullol J, van den Akker-van Marle ME, van Ganse E, Bachert C. Costs associated with persistent allergic rhinitis are reduced by levocetirizine. Allergy 2005;60:788-94. 3. Bousquet J, van Cauwenberge P, Khaltaev N. ARIA Workshop Group, World Health Organization. Allergic rhinitis and its impact on asthma. J Allergy Clin Immunol 2001;108:S147-334. 4. Bousquet J, Neukirch F, Bousquet PJ, Gehano P, Klossek JM, Le Gal M, et al. Severity and impairment of allergic rhinitis in patients consulting in primary care. J Allergy Clin Immunol 2006;117:158-62. 5. Ciprandi G, Cirillo I, Vizzaccaro A, Tosca M, Passalacqua G, Pallestrini E, et al. Seasonal and perennial allergic rhinitis: is this classification adherent to real life? Allergy 2005;60:882-7. 6. Demoly P, Allaert FA, Lecasble M, Bousquet J. Validation of the classification of ARIA (allergic rhinitis and its impact on asthma). Allergy 2003;58:672-5. 7. Bauchau V, Durham SR. Epidemiological characterization of the intermittent and persistent types of allergic rhinitis. Allergy 2005;60:350-3. 8. Bousquet J, Annesi-Maesano I, Carat F, Leger D, Rugina M, Pribil C, et al. Characteristics of intermittent and persistent allergic rhinitis: DREAMS study group. Clin Exp Allergy 2005;35:728-32. 9. Bachert C, van Cauwenberge P, Olbrecht J, van Schoor J. Prevalence, classification and perception of allergic and nonallergic rhinitis in Belgium. Allergy 2006;61:693-8. 10. Pereira C, Valero A, Loureiro C, Davila I, Martinez-Cocera C, Murio C, et al. Iberian study of aeroallergens sensitization in allergic rhinitis. Allerg Immunol (Paris) 2006;38:186-94. 11. van Hoecke H, Vastesaeger N, Dewulf L, De Bacquer D, van Cauwenberge P. Is the allergic rhinitis and its impact on asthma classification useful in daily primary care practice? J Allergy Clin Immunol 2006;118:758-9. 12. Juniper EF, Guyatt GH. Development and testing of a new measure of health status for clinical trials in rhinoconjunctivitis. Clin Exp Allergy 1991;21:77-83. 13. Ware JE, Kosinski M, Keller SD. A 12-Item Short-Form Health Survey: construction of scales and preliminary tests of reliability and validity. Med Care 1996;34:220-33. 14. Soler R, de la Hoz B, Badia X, Mercadal J, Lozano R, Benavides A, et al. Validation of the Spanish version of the Rhinoconjunctivitis Quality of Life Questionnaire (RQLQ). Rev Clin Esp 2004;204:131-8.
Valero et al 365
J ALLERGY CLIN IMMUNOL VOLUME 120, NUMBER 2
23.
24.
25.
26.
27.
EuroQol five-dimension questionnaire differentiate between COPD severity stages? Chest 2006;130:117-28. Ferrer M, Alonso J, Morera J, Marrades R, Khalaf A, Aguar MC, et al. Chronic obstructive pulmonary disease stage and health-related quality of life. The Quality of Life of Chronic Obstructive Pulmonary Disease Study Group. Ann Intern Med 1997;127:1072-9. Workshop Report, Global Strategy for Asthma Management and Prevention (GINA). Initiative World Health Organization, WHO. Updated November 2006. Available at: http://www.ginasthma.com/. Accessed March 30, 2007. Canonica GW, Bousquet J, van Hammee G, Bachert C, Durham SR, Klimek L, et al. Levoceterizine improves health-related quality of life and health status in persistent allergic rhinitis. Respir Med 2006;100: 1706-15. Liard R, Leynaert B, Zureik M, Beguin FX, Neukirch F. Using Global Initiative for Asthma Guidelines to assess asthma severity in populations. Eur Respir J 2000;16:615-20. Hoogendoorn M, Feenstra TL, Schermer TR, Hesselink AE, Rutten-van Molken MP. Severity distribution of chronic obstructive pulmonary disease (COPD) in Dutch general practice. Respir Med 2006;100: 83-6.
Rhinitis, sinusitis, and ocular diseases
15. Vilagut G, Ferrer M, Rajmil L, Rebollo P, Permanyer-Miralda G, Quintana JM, et al. The Spanish version of the Short Form 36 Health Survey: a decade of experience and new developments. Gac Sanit 2005;19:135-50. 16. Juniper EF, Guyatt GH, Griffith LE, Ferrie PJ. Interpretation of rhinoconjunctivitis quality of life questionnaire data. J Allergy Clin Immunol 1996;98:843-5. 17. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale (NJ): L. Erlbaum Associates; 1988. 18. Bousquet J, van Cauwenberge P, Ait Khaled N, Bachert C, Baena-Cagnani CE, Bouchard J, et al. Pharmacologic and anti-IgE treatment of allergic rhinitis ARIA update (in colaboration with GA2LEN). Allergy 2006;61:1086-96. 19. Prenner BM, Schenkel E. Allergic rhinitis: treatment based on patients profiles. Am J Med 2006;119:230-7. 20. Hoecke H, Vastesaeger N, Dewulf L, Sys L, van Cauwenberge P. Classification and management of allergic rhinitis patients in general practice during pollen season. Allergy 2006;61:705-11. 21. Fonseca JA, Delgado L, Costa-Pereira A, Tavares C, Moreira A, Morete A, et al. Evaluation of the Asthma Life Quality test for the screening and severity assessment of asthma. Allergy 2004;59:1198-204. 22. Rutten-vanMolken MP, Oostenbrink JB, Tashkin DP, Burkhart D, Monz BU. Does quality of life of COPD patients as measured by the generic
Correction With regard to the January 2005 article entitled ‘‘The September epidemic of asthma exacerbations in children: A search for etiology’’ (2005;115:132-8), the legend for Figure 2 gives an incorrect date range. The correct dates are April 2001 to March 2002.