Burden of allergic respiratory disease: a systematic review

In Europe, approximately 23 % of the population is affected by allergic rhinitis (AR) [1]. The most common symptoms are sneezing, itchy nose, rhinorrhea, and/or nasal congestion [2], with many patients experiencing symptoms of at least moderate severity; one study estimated that 93 % of patients with AR who consulted general practitioners had symptoms that were moderate-to-severe [3]. Studies of patients consulting general practitioners for AR reported that 18–48 % had symptoms that were not controlled by pharmacotherapy [4, 5]. Despite the bothersome nature of symptoms, AR is often trivialized by the patient; less than half (45 %) seek medical advice or treatment for their condition [4], which results in under-treatment and poor control of symptoms. A UK study found that only 18 % of patients with AR had consulted their doctor about the condition in the preceding 2 years [6]. In a sample of 230 French patients with typical symptoms of AR, 19 % had never consulted their doctor about the problem, despite 90 % reporting that their nasal symptoms affected their daily lives, over half complaining of sleepiness and headaches, and 8 % reporting that they had taken time off work because of their rhinitis [7]. Allergic rhinitis is frequently accompanied by allergic asthma (AA), the prevalence of which increases from less than 2 % in individuals without AR to 10–40 % in those with AR; this combination of upper and lower respiratory symptoms increases the overall impact on the patient [8‐10].

There is now considerable evidence that the symptoms of AR and AA negatively affect patients’ health-related quality of life (HRQL) [11‐20]. This is supported by evidence that successful treatment of symptoms with allergy immunotherapy (AIT) [21‐23] or symptomatic treatments [24‐31] improves the HRQL of patients with AR. These improvements generally mirror those of conventional clinical outcome measures, such as symptom and medication scores, and lung function tests [32]. Although clinical measures provide information on the affected organ systems, they do not capture the patient’s overall perception of the disease burden caused by the physical, emotional, and social impairments in everyday life [33]. The wider impact of AR on patients’ lives is now acknowledged in the allergic rhinitis and its impact on asthma (ARIA) guidelines, which recognize the impact of HRQL on AR, and classify its severity based both on symptoms and its effect on HRQL [32].

There is evidence that AR represents a considerable economic burden [34, 35], with annual costs in the US in 2003 estimated at $2–$5 billion USD [34]. Indirect costs resulting from symptoms causing under-performance or loss of productivity in the workplace were particularly high [36‐39]. Data published after the present study was completed show that patients with mild AR have less impact on the health economy, with costs of around a quarter of those with moderate-to-severe disease [39].

To date, there has been no systematic review of HRQL data in patients with perennial or seasonal allergies. The objective of this systematic review and meta-analysis was, therefore, to compare HRQL in individuals with AR and/or AA with perennial house dust mite (HDM) or seasonal AR caused by exposure to pollen, or both. As HRQL instruments/questionnaires are included in many clinical studies as a matter of course, HRQL measurements enable comparisons with data from individual studies. In addition, data on the economic burden of AR and AA in selected European countries were systematically reviewed in order to compare direct and indirect costs in patients with perennial HDM and seasonal AR.

Grouping studies by seasonal or perennial symptoms gave an acceptable level of heterogeneity (I² ≤ 63 %) that allowed summary estimates of HRQL from RQLQ and SF-36 instruments to be determined. Heterogeneity was generally higher for the SAR/pollen phenotype than for PAR/HDM, which could reflect the different types of baseline measurements in the selected studies, or the large heterogeneity of SAR. Patients with PAR/HDM allergy had poorer HRQL than those with SAR/pollen allergy when measured by RQLQ, SF-36 total score, and SF-36 PCS. The underlying reason for this difference is unclear, and could be due to differences in symptom frequency, duration, or severity. In Northern Europe, the pollen season is largely confined to the summer months, with the season for grass pollen, which is by far the most frequent cause of pollinosis, usually lasting from mid-May to July [74]; at the height of the pollen season, patients with SAR suffer symptoms approximately 50 % of the time and report an impact on their HRQL [75]. It can be hypothesized that the year-round nature of symptoms might be responsible for the observed poorer HRQL in PAR/HDM allergy; however, published studies report varying results. Bousquet found that the severity of AR was more important in determining HRQL, sleep, daily activities, and work performance than symptom duration [3]. Similar findings were reported by Delgado and colleagues, who observed that HRQL was lower in Spanish patients with olive or grass pollen allergy vs. HDM allergy. This was attributed to the greater severity of symptoms caused by the explosive nature of olive pollination, resulting in sudden and elevated pollen concentrations, in contrast to the more constant concentrations of perennial allergens that were more tolerable for patients [76]. Another study reported no significant differences in baseline RQLQ scores between patients with HDM, grass pollen, and Parietaria pollen allergy [48].

Analysis of SF-36 component scores showed that mental and physical component scores were affected in seasonal and perennial allergy, with HDM/PAR having a greater impact on physical vs. mental component scores. This supports the findings of other studies that report significant emotional problems and poor mental wellbeing in patients with AR, in addition to physical symptoms [13, 33, 77].

Meta-analyses were performed using HRQL measures from generic and disease-specific instruments. Generic questionnaires, like the SF-36, measure a wide range of physical, mental and psychosocial functions; they can be used in different health conditions and in the general population, and have the advantage of allowing comparisons of burden of illness across different disorders. Disease-specific instruments like the RQLQ more accurately reflect the patient’s inherent disease-associated problems, but scores cannot be directly compared with other populations. Both the generic and RQLQ instruments show improvements in HRQL in patients treated for AR; disease-specific questionnaires tend to show more important improvements, as they focus on the areas that are impacted by the disease [32]. Nowadays, HRQL questionnaires are used almost routinely in clinical studies of allergy treatments. Observed changes in HRQL generally reflect clinical outcomes; however, some studies report changes in HRQL that are not captured by symptom scores. Therefore, monitoring patients’ HRQL in clinical practice has been suggested as a possible strategy to improve the management of AR [78].

Our analysis does not permit comparison of the HRQL of AR patients with that of the general population; however, a Spanish cross-sectional study that investigated the overall HRQL of patients with AR using SF-12 reported that HRQL was 25 % lower than in the general population [79]. Poor HRQL in patients with AR is also supported by evidence that successful treatment of AR symptoms with AIT [21‐23] or symptomatic treatment [24‐31] improves patient’s HRQL scores. A comparison of published EQ-5D utility baseline scores indicated that perennial allergy and/or AA (13 patients with AR and 12 with AR+AA), on a typical day with allergy symptoms, had a score of 0.60 [23], indicating comparable HRQL to diabetes (0.83) [80], cardiovascular disorders (0.73), musculoskeletal disorders (0.63), and psychosomatic disorders (0.57), when measured with the same EQ-5D instrument [81]. Ranges of published EQ-5D scores for AR and/or AA and other diseases are summarized in Fig. 7, and indicate that the impact of AR and/or AA on generic HRQL on a day with allergen exposure is not trivial, with an effect that is comparable to chronic conditions, such as liver disease, epilepsy, migraine, and visual disorders [23, 82‐86].

There were insufficient data in the present analysis to compare HRQL in patients with AR vs. those with both AR and AA. Delgado and colleagues reported that HRQL associated with AR was worse in patients diagnosed with both AR and asthma than in those with only AR [76]. Using the SF-36 instrument, Leynaert reported that patients with asthma experienced greater impairments in physical functioning that those with AR alone, but asthma did not further impair mental wellbeing of patients with AR [13].

The cost analysis shows variations in direct and indirect AR costs between countries, which may result from true differences in AR costs or from differences in the health systems, diagnostic behavior, or reporting between countries. Indirect costs make up the majority of the cost burden of AR, in line with the outcomes of the most recent European [39] and US studies [34]. A few studies have reported higher direct vs. indirect costs [35, 87]; however, many studies estimate indirect costs by multiplying patients’ salaries by the amount of time they are absent from work due to AR. Health impairments associated with AR are often not severe enough to cause absence from work, although they can interfere with cognitive functioning, resulting in fatigue and an impaired ability to learn, concentrate, and make decisions [88], all of which can affect performance in the workplace. In one study, more than a third of patients with AR (36 %) reported reduced workplace performance, or ‘presenteeism’, which was higher than among patients with asthma (19 %) [89]. In the US, where up to 60 million people are affected by AR [90], annual productivity loss associated with a diagnosis of AR was estimated at $601 million in 1995. When the use of sedating allergy medications and workers’ self-assessments of their reduction work productivity due to AR were taken into account, the estimated annual at-work productivity losses increased to $2.4–$4.6 billion [37]. In 2006, the cost of lost productivity in the workplace due to AR was $593 per employee per year, and exceeded that of stress ($518), migraine ($277), depression ($273), and arthritis/rheumatism ($269) [36]. Comorbid AA resulted in higher costs than AR alone [91], or AR with non-allergic asthma [92], indicating its importance in driving costs. The cost of productivity loss is greater than the cost of treating AR symptoms; however, many cost analysis studies do not consider the cost to the patient of over-the-counter AR medications, which could lead to an underestimation of direct costs [34].

Some limitations of the study design should be acknowledged. Search terms were chosen from the RQLQ and SF-36 questionnaires, or were based on the suspected impact of AA and AR on patients’ HRQL. The articles identified by the search were mainly safety or efficacy studies that included HRQL as a secondary endpoint. Consequently, abstracts that did not include secondary outcomes would not be identified by the search. This could introduce a bias, such that only studies reporting a change in HRQL were identified. Inclusion of patients with comorbid AA in some groups could also affect HRQL scores, although we estimate any effect to be small. Moreover, the study only included HDM, and not pet allergens, as perennial allergens for perennial allergic rhinitis/asthma. Many of the studies from which data are collated were randomized controlled interventional studies, which may tend to include patients with more severe symptoms.

This review presents summary data on HRQL levels in patients with allergic respiratory disease that may be used for comparison with assessments from individual studies. HRQL in PAR/HDM patients was significantly worse than that of SAR/pollen patients, when measured by both disease-specific and generic HRQL instruments, and was reflected by an impact on both physical and mental health. SAR/pollen and PAR/HDM affected MCS to a similar degree, despite a greater impact of PAR/HDM on PCS. A comparison of published EQ-5D utility index scores showed that the impact of PAR, with or without AA, on patients’ HRQL was comparable to other chronic diseases. The economic burden of AR is considerable, with indirect costs due to absenteeism and presenteeism making up the majority of costs. Greater awareness of the detrimental effects of AR on HRQL and its cost burden might encourage early diagnosis and treatment in order to minimize the disease burden and ensure beneficial and cost-effective outcomes.