Chapter  67:  Management of Allergic Rhinitis in the Working-Age Population

Overview

Allergic rhinitis affects as many as 35 million people in the United States annually; of these, an estimated 19 million are employed adults. Overall, 10 to 30 percent of adults and up to 40 percent of children are affected, making it the sixth most common chronic illness in the United States. Approximately one-third to one-half of sufferers have seasonal rhinitis, with the remainder experiencing perennial disease or both seasonal and perennial forms of the disease. Other atopic conditions, such as atopic eczema, allergic conjunctivitis, and asthma, often co-occur.

Estimates of the annual direct medical costs of allergic rhinitis in the US range from $1.16 billion to $4.5 billion, rising to $7.7 billion when indirect costs are included. These estimates, however, are based on information that predates the increased use of non-sedating antihistamines and nasal glucocorticoids. Recent prescription claims data show that approximately two-thirds of patients with allergic rhinitis receive treatment with one or more medications from these two drug classes, with expenditures exceeding $3.0 billion for prescription antihistamines alone.

Rhinitis is typically classified etiologically into allergic and non-allergic causes. Non-allergic rhinitis is characterized by chronic nasal symptoms and the lack of identifiable allergic triggers. This report focuses on individuals with allergic rhinitis, including both seasonal and perennial allergic rhinitis. Seasonal allergic rhinitis is associated with sensitization to fungal, tree, grass, and weed pollens, and with symptoms that vary seasonally. Perennial allergic rhinitis is associated with sensitization to indoor allergens such as fungi, cockroaches, dust mites, and animal proteins (e.g., cat dander), and with year-round symptoms, with or without seasonal exacerbations.

The physical symptoms of allergic rhinitis, such as sneezing, rhinorrhea, and nasal congestion, may interfere with one's ability to carry out daily activities. Rhinitis symptoms may be associated with headache, irritability, poor concentration, loss of sleep, and resulting fatigue. The functional impact of these symptoms ranges from mild to seriously debilitating effects on social, physical, and emotional functioning. Allergic rhinitis may interfere with cognitive tasks, may impair work performance, and may cause work absences.

Because allergic rhinitis is so common in the population and allergens are ubiquitous, allergic rhinitis creates a significant burden in the workplace in terms of effects on work performance and health care costs. Although some occupational exposures to airborne allergens present in the workplace can cause occupational rhinitis, non-occupational allergic rhinitis represents a vastly greater burden in workplace settings overall.

The topic of this report was selected by the Agency for Healthcare Research and Quality (AHRQ) in response to a nomination by the American Association of Health Plans. The Duke Evidence-based Practice Center (EPC) conducted the research and developed the final report for AHRQ. The emphasis on the working-age population raises unique issues, including the relationship between symptoms or functional status and work performance, the effects of allergic rhinitis and its treatments on costs and work performance, and variability in management approaches and patient outcomes among patients treated by generalist physicians, allergy specialists, and otolaryngologists.

The general diagnostic and treatment issues relating to allergic rhinitis were summarized in an earlier evidence report, Management of Allergic and Nonallergic Rhinitis, prepared by the EPC at the New England Medical Center. However, the Duke evidence report prioritizes issues not addressed in the New England Medical Center report, including the effect of allergic rhinitis treatment on work performance and costs, and the effectiveness of combinations of pharmacological treatments, immunotherapy, and the use of strict environmental control measures. The Duke research team sought evidence on these issues, evidence that may be valuable not only to employers, policy decisionmakers, and guideline developers, but also to researchers who wish to identify and address gaps in evidence, and to clinicians who care for patients with allergic rhinitis.

Reporting the Evidence

The Duke EPC staff, in consultation with AHRQ and a multidisciplinary panel of experts, refined the key research questions addressed in this report:

1. How do currently available clinical treatments for allergic rhinitis affect costs and work performance?

2. What is the relationship between symptom outcomes or disease-specific quality-of-life measures and work performance among adults with allergic rhinitis? Can data on symptomatic outcome or quality of life be reliably translated into work performance measures?

3. How effective are (a) environmental measures, (b) immunotherapy, and (c) combined treatments, such as antihistamines and nasal steroids or antihistamines and oral decongestants, for relief of symptoms in adults with allergic rhinitis?

4. How do different types of health care providers (generalists, allergy specialists, and otolaryngologists) treat adults with allergic rhinitis, and how do treatment outcomes vary by provider?

5. In adult patients with symptoms of allergic rhinitis, does the prevalence, treatment patterns, or response to treatment vary according to a patient's race or ethnicity?

Methodology

The Duke EPC researchers systematically reviewed the literature for evidence addressing the above questions. They searched for English-language articles indexed in computerized bibliographic databases: MEDLINE®, CINAHL®, the Cochrane Database of Systematic Reviews, the Database of Abstracts of Reviews of Effectiveness, International Pharmaceutical Abstracts, EconLit, and EMBASE. Searches of these databases were supplemented by searching the reference lists of all included articles, especially review articles and meta-analyses, and by scanning current issues of relevant journals not yet indexed in the online databases.

The results of the literature searches were screened by two investigators according to inclusion and exclusion criteria. Empirical studies were included if: (a) the study population had allergic rhinitis; (b) the study provided data on at least one of the five key research questions; and (c) the study met minimal study-design criteria for the question being addressed. Minimal study design criteria for the key questions follow:

• Question 1 and 2: Costs and work performance. Any empirical study involving more than 20 patients with allergic rhinitis. Includes randomized controlled trials (RCTs), case series, cohort studies, non-randomized comparison studies, surveys, and secondary data analyses.

• Question 3a: Environmental measures. RCTs and non-randomized prospective cohort comparisons.

• Questions 3b and 3c: Immunotherapy and combination drug therapy. RCTs and pseudo-randomized placebo-controlled trials.

• Questions 4 and 5: Clinician specialty differences and racial and ethnic variation. Any empirical study involving more than 20 patients with allergic rhinitis. Includes RCTs, case series, cohort studies, non-randomized comparison studies, surveys, and secondary data analyses.

The full text of each article included at the screening stage was independently reviewed by two investigators. Articles found to meet inclusion criteria were selected for data abstraction. The EPC required patient-assessed symptom outcomes for efficacy questions,; and researchers also reported quality of life, functional status, adverse events, and patient global assessments for these questions. For all questions, they recorded work performance and cost outcomes.

The EPC's senior writer/editor began the data abstraction process with a partial abstraction, which included a description of the study design, intervention, number of subjects at the start of the study, and types of outcome data collected. One investigator then completed abstraction of details of the study population, results, and comments; a second investigator over-read the table for completeness and accuracy and performed quality scoring. They evaluated each article included in the evidence tables for methodological quality, grading the level of evidence and describing 13 factors affecting internal or external validity.

The EPC employed quality-monitoring checks at every phase of the literature search, review, and data abstraction process to reduce bias, enhance consistency, and check the accuracy of screening.

Findings

Future Research

The EPC assessment of the current evidence suggests that the following issues should be addressed in future research.

Updated estimates of the cost of allergic rhinitis could provide a more accurate assessment by:

• Estimating indirect costs using valid objective measures of productivity changes.

• Including over-the-counter medications in direct medical costs.

• Accounting for increased use of non-sedating antihistamines and nasal corticosteroids.

• Carefully defining allergic rhinitis, particularly when using administrative data sets.

Although environmental control measures are strongly endorsed by experts, studies of such interventions have been equivocal. More comprehensive environmental control measures, such as those recommended in the National Heart, Lung, and Blood Institute's Practical Guide for the Diagnosis and Management of Asthma should be tested in patients with allergic rhinitis and significant functional impairment. If comprehensive interventions prove effective, then future studies should identify critical components.

To better understand the role of immunotherapy in the treatment of allergic rhinitis, we need trials employing vaccines with most or all of the relevant allergens for each individual to assess immunotherapy as it is administered in most community settings. Additional future research objectives should focus on the following:

• Methods to identify patients likely to benefit from immunotherapy.

• Determination of whether immunotherapy alters the natural history of allergic rhinitis and reduces possible sequelae such as bacterial sinusitis and asthma.

• Comparisons of immunotherapy and the best available medical management and/or allergen avoidance.

• Clarifying the optimal duration of immunotherapy.

Certain combination pharmacologic treatments have been shown to be effective in relatively short-term trials, mostly in seasonal allergic rhinitis. Additional data are needed on:

• The effectiveness of combination treatment in perennial allergic rhinitis.

• Longer duration treatment in primary care populations with clinically diagnosed seasonal or perennial allergic rhinitis.

• Effectiveness trials that include outcomes such as health-related quality of life and cost-effectiveness.

• The effectiveness of combinations including mast cell stabilizers, ipratropium, and newer drugs such as leukotriene antagonists.

To understand the quality of current patient care by different clinical specialists, we need:

• Studies describing current practice patterns.

• Prospective studies that compare symptomatic treatment to allergen identification with specific treatment, two approaches commonly used in generalist and specialty practices.

• Observational studies that compare treatment patterns and outcomes across specialties that provide case-mix adjustment (a standardized and validated severity-of-illness scale would facilitate this research).

Finally, the research team did not identify any studies that described racial or ethnic differences in treatment patterns or treatment response, in part because study populations were often incompletely described. Future studies should provide more complete descriptions of patient populations, including racial and ethnic descriptors that might allow subgroup analyses to assess racial or ethnic differences in treatment or response.

Background

Allergic rhinitis, also known as hay fever, is one of the most common allergic diseases in the United States. The National Institute of Allergy and Infectious Diseases currently estimates that allergic rhinitis affects as many as 35 million Americans and accounts for 16.7 million office visits to healthcare providers each year (National Institute of Allergy and Infectious Diseases, 2002; National Institutes of Health, 2002). A recent report from the American Academy of Allergy, Asthma & Immunology estimates that about 19 million employed adults suffer from allergic rhinitis, and that approximately $4.5 billion in direct costs and 3.8 million lost work and school days are attributable to this disease annually (American Academy of Allergy, Asthma & Immunology, 2000).

Allergic rhinitis usually begins in childhood, adolescence, or early adulthood, and often wanes, but may persist, with increasing age. Rhinitis is defined as inflammation of the membranes lining the nose. The symptoms of allergic rhinitis usually include sneezing, rhinorrhea, itching and watery eyes, nasal congestion, and, in severe cases, facial pressure or pain. These symptoms may be associated with headache, irritability, poor concentration, loss of sleep, and fatigue. The functional impact of allergic rhinitis ranges from mild to seriously debilitating effects on social, physical, and emotional functioning, which may interfere with cognitive tasks, impair work performance, and cause work absences.

Because allergic rhinitis is so common and allergens are ubiquitous, allergic rhinitis creates a significant burden in the workplace in terms of work performance and healthcare costs. Although exposures to airborne allergies present in the workplace can cause occupational rhinitis, non-occupational rhinitis represents a vastly greater burden in workplace settings overall.

An evidence report on the topic of allergies and their effect on working-age populations was proposed to the Agency for Healthcare Research and Quality (AHRQ) by the American Association of Health Plans (AAHP), who became the Duke Evidence-based Practice Center's partner in developing this report. The specific research questions were refined in consultation with AHRQ, AAHP, and an advisory panel of eight experts convened especially for this study. The key research questions addressed in this report are:

1. How do currently clinically available treatments for allergic rhinitis affect costs and work performance?

2. What is the relationship between symptom outcomes or disease-specific quality-of-life measures and work performance among adults with allergic rhinitis? Can data on symptomatic outcome or quality of life be reliably translated into work performance measures?

3. How effective are (a) environmental measures, (b) immunotherapy, and (c) combined treatments, such as with antihistamines and nasal steroids or antihistamines and oral decongestants, for relief of symptoms in adults with allergic rhinitis?

4. How do different types of healthcare providers (generalists, allergy specialists, and otolaryngologists) treat adults with allergic rhinitis, and how do treatment outcomes vary by provider?

5. In adult patients with symptoms of allergic rhinitis, does the prevalence, treatment patterns, or response to treatment vary according to a patient's race or ethnicity?

Scope and Purpose

The purpose of this evidence report is to review the published evidence on strategies for managing the treatment of patients with allergic rhinitis, particularly those of employment age (18 to 64 years old). The report covers both seasonal and perennial allergic rhinitis. Seasonal allergic rhinitis is associated with sensitization to fungal, tree, grass, and weed pollens, and with symptoms that vary seasonally. Perennial allergic rhinitis is associated with sensitization to indoor allergens such as fungi, cockroaches, dust mites, and animal proteins (e.g., cat dander), and with year-round symptoms, with or without seasonal exacerbations.

Treatment options considered in this report are environmental measures ( allergen avoidance), immunotherapy, and combination therapies employing antihistamines and nasal steroids or antihistamines and oral decongestants.

Also considered in the present report are the unique issues raised by the emphasis on working-age populations, including the relationship between symptoms or functional status and work performance, and the effects of allergic rhinitis and its treatment on costs and work performance. In addition, the report reviews the evidence on variability in management approaches and patient outcomes by type of clinician ( generalist physician vs. allergy specialist vs. otolaryngologist), as well as by patient race and ethnicity.

Our goals were primarily to identify, review, and evaluate the published literature on these topics and, secondarily, where relevant evidence could not be identified or had important limitations, to describe the type of data that would be needed to more fully address the research questions. Ultimately, we hope to provide clinicians, policymakers, and patients with the evidence they need to decide for themselves on the best treatment and management options from among those considered here.

Epidemiology of Allergic Rhinitis

Allergic rhinitis affects 20 to 40 million people in the United States annually, including 10 to 30 percent of adults and up to 40 percent of children (Joint Task Force on Practice Parameters in Allergy, Asthma and Immunology, 1998). Approximately one-third to one-half of these patients suffer from seasonal allergic rhinitis, with the remainder experiencing perennial disease or both seasonal and perennial forms of the disease (Joint Task Force on Practice Parameters in Allergy, Asthma and Immunology, 1998). Other atopic conditions, such as atopic eczema, allergic conjunctivitis, and asthma, often co-occur with allergic rhinitis.

Allergic rhinitis may begin at any age, with most individuals developing symptoms as children or young adults. Risk factors include a family history of atopy, higher socioeconomic class, and exposure to indoor allergens such as animals and dust mites (Joint Task Force on Practice Parameters in Allergy, Asthma and Immunology, 1998). The risk of allergic rhinitis is 30 percent if one parent is atopic, at least 50 percent if both parents are atopic, and greater than 70 percent if both parents have the same allergic disease (Nimmagadda and Evans, 1999).

Table 1. 1996 US prevalence rates and numbers for hay (more...)

Table 1. 1996 US prevalence rates and numbers for hay fever/allergic rhinitis without asthma by age, sex, race, and family income

Prevalence rates	Age 18–44 (unless otherwise noted)	Age 45–64	Total population
Per 1,000 persons	109.4	104.8	89.8
By sex	Male, under 45: 86.3	Male: 85.6	Not available (NA)
	Female, under 45: 92.1	Female: 122.8
By race	White, under 45: 92.0	White: 111.0	NA
	Black, under 45: 66.2	Black: 64.6
	< $10,000, under 45: 82.7	< $10,000: 106.9
By family income	$10,000–19,999, under 45: 69.1	$10,000–19,999: 111.8	NA
	$20,000–34,999, under 45: 75.1	$20,000–34,999: 105.0
	$35,000 or more, under 45: 108.9	$35,000 or more: 109.2
Prevalence numbers, in thousands	Age 18–44 (unless otherwise noted)	Age 45–64	Total population
Number	11,809	5,572	23,721
By sex	Male, under 45: 7,751	Male: 2,198	NA
	Female, under 45: 8,248	Female: 3,374
By race	White, under 45: 13,404	White: 5,077	NA
	Black, under 45: 1,665	Black: 350
	< $10,000, under 45: 1,128	< $10,000: 290
By family income	$10,000–19,999, under 45: 1,673	$10,000–19,999: 621	NA
	$20,000–34,999, under 45: 2,797	$20,000–34,999: 983
	$35,000 or more, under 45: 8,406	$35,000 or more: 2,866

Source: Centers for Disease Control and Prevention, National Center for Health Statistics. Current estimates from the National Health Interview Survey, 1996. Vital and Health Statistics, Series 10, No. 200. DHHS Publication No. (PHS) 99–1528. Hyattsville, MD: US Department of Health and Human Services. October 1999.

Table 2. 1996 US prevalence rates and numbers for hay (more...)

Table 2. 1996 US prevalence rates and numbers for hay fever/allergic rhinitis without asthma, by geographic location and place of residence

Geographic location	Prevalence rates per 1,000 persons	Prevalence numbers, in thousands
US	89.8	23,721
Northeast	78.3	4,220
Midwest	85.5	5,424
South	94.9	8,593
West	97.3	5,484
Place of residence	Prevalence rates per 1,000 persons	Prevalence numbers, in thousands
All Metropolitan Statistical Areas (MSA)	90.6	18,887
Central city	86.3	6,742
Not central city	93.3	12,145
Not MSA	86.5	4,834

Source: Centers for Disease Control and Prevention, National Center for Health Statistics. Current estimates from the National Health Interview Survey, 1996. Vital and Health Statistics, Series 10, No. 200. DHHS Publication No. (PHS) 99–1528. Hyattsville, MD: US Department of Health and Human Services. October 1999.

Overview of Disease Biology

The symptoms of allergic rhinitis result from exposure to particulate allergens that are large enough to be filtered by the nose. In susceptible adults, allergen-specific T cell sensitization leads to B cell production of allergen-specific immunoglobulin E (IgE) antibodies after an initial allergen exposure (e.g., pollen) (American Academy of Allergy, Asthma & Immunology, 2000). Allergen-specific IgE then binds to the surface of mast cells in the nasal mucosa or to circulating basophils. With subsequent exposure, the allergen is recognized by its specific antibody, resulting in the activation of IgE-primed mast cells and basophils, with release of a variety of potent inflammatory mediators. These include granule-associated mediators (e.g., histamine), membrane-derived lipid mediators (e.g., leukotriene), as well as cytokines and chemokines that attract inflammatory cells from the peripheral circulation to the site of degranulation. These mediators cause immediate mucosal edema and vasodilation and the clinical features of allergic rhinitis. “Early-phase” symptoms occur within minutes of the allergen exposure and are due to release of preformed mediators; “late-phase” symptoms occur 4 to 12 hours after exposure and involve synthesis of newly formed mediators and infiltration of inflammatory white blood cells from the circulation (Bellanti and Wallerstedt, 2000; Parikh and Scadding, 1997; Skoner, 2001). The late phase has been observed with large exposure allergen challenges, but the clinical importance of this observation is uncertain. Symptoms affect about 30 to 40 percent of individuals during the “late-phase” time period. Nasal itching is prominent during the early phase. Sneezing, nasal congestion, and rhinorrhea are common to early and late phases, and nasal congestion dominates during the late-phase reaction.

Burden of Illness

The symptoms of allergic rhinitis, such as sneezing, rhinorrhea, and nasal congestion, may interfere with one's ability to carry out daily activities. Rhinitis symptoms may be associated with headache, irritability, poor concentration, loss of sleep, and resulting fatigue. The functional impact of these symptoms ranges from mild to seriously debilitating effects on social, physical, and emotional functioning (Blaiss, 1999; Thompson, Juniper, and Meltzer, 2000). In a study comparing 116 healthy subjects to 111 patients with moderate to severe perennial allergic rhinitis, patients with allergic rhinitis had significantly decreased functioning in eight domains; negative effects were particularly prominent for physical and emotional role limitations, social functioning, and general health perceptions (Bousquet, Bullinger, Fayol, et al., 1994). Allergic rhinitis may interfere with cognitive tasks, may impair work performance, and may cause work absences. In a pooled analysis of 1,948 patients with moderate to severe allergic rhinitis, over 90 percent reported that their classroom or work performance was affected negatively (Tanner, Reilly, Meltzer, et al., 1999).

In addition to direct symptom effects, allergic rhinitis may be related to the development of asthma, sinusitis, or otitis media (Bousquet, van Cauwenberge, Khaltaev, et al., 2001; Spector, 1997). Asthma symptoms occur in 17 to 19 percent of patients with allergic rhinitis, a prevalence that is significantly higher than the five percent prevalence observed in the general population (Blair, 1977; Moller, Dreborg, Ferdousi, et al., 2002; Pedersen and Weeke, 1983; Settipane, 1986). In a cohort of 7,225 children followed from birth to age 23, children with allergic rhinitis were 2.0 to 2.9 times more likely to develop asthma during followup (Anderson, Pottier, and Strachan, 1992). A similar cohort study of college students found that those with allergic rhinitis were three times more likely to develop asthma than non-atopic controls during the 23-year followup (Settipane, Hagy, and Settipane, 1994). In cross-sectional studies, allergic rhinitis is associated with acute and chronic bacterial sinusitis (Long, McFadden, DeVine, et al., 2002).

Adverse effects from therapies are an additional burden associated with this illness, since they may impact more significantly on functional status than the disease itself, especially for patients with very mild disease. For adults, the only life-threatening effect from commonly used treatments is anaphylaxis associated with immunotherapy, which occurs at a rate of about one fatality per two million doses (Cook and Farias, 1998). Non-fatal systemic reactions are more common; estimates of their frequency vary widely, from 0.3 percent to more than 30 percent (Cook and Farias, 1998). Minor adverse effects of somnolence, dry mouth, dizziness, and headache may occur in up to 50 percent of patients taking sedating antihistamines (Long, McFadden, DeVine, et al., 2002). Published experimental work suggests that adverse effects associated with some treatments, particularly sedating antihistamines, which cause somnolence and psychomotor impairment, have an adverse impact on driving performance and reaction time (Adelsberg, 1997; Weiler, Bloomfield, Woodworth, et al., 2000); these effects may also interfere with work productivity and increase on-the-job accidents. The most frequently reported adverse effects associated with nasal corticosteroids are epistaxis, headache, and pharyngitis; with cromolyn, nasal irritation and headache are the most commonly reported adverse effects.

Management Strategies and Treatment Options

Allergen avoidance, immunotherapy, and an array of pharmacotherapies are commonly used to treat allergic rhinitis. For clinicians, management begins with accurate diagnosis, distinguishing between allergic and non-allergic etiologies. The clinical evaluation may include radioallergosorbent testing (RAST) or allergy skin testing to confirm allergy sensitization. For patients with allergic rhinitis, relevant treatment issues are: the efficacy of individual treatments; monotherapy versus combinations of treatments; the most cost-effective sequencing of treatments; and the effectiveness of generalist versus specialist care. In working populations, relevant treatment outcomes are: symptom control; effects on health-related quality of life; cost-effectiveness; and effects on work performance.

The specific therapies covered in this evidence report are environmental measures, or allergen avoidance; immunotherapy; and combination therapies such as antihistamines and nasal steroids or antihistamines and oral decongestants. Given the variety of treatment options, the variability in acceptability and cost of treatments, and the lack of a previous focus on work-related outcomes, a systematic review that addresses these issues is timely.

Environmental Measures

Given the known biology of allergic rhinitis, environmental measures (allergen avoidance) represent a conceptually appealing treatment option. Such measures are recommended in the rhinitis clinical guidelines developed by the Joint Task Force on Practice Parameters in Allergy, Asthma, and Immunology (1998), and by the American Academy of Otolaryngic Allergy (Fornadley, Corey, Osguthorpe, et al., 1996); they have also been recognized by the American Academy of Allergy, Asthma & Immunology in its recent report (2000). Allergen avoidance measures range from relatively inexpensive measures, such as removing feather pillows and down comforters, to more intensive measures, such as high-flow air filtration units like a high efficiency particulate air (HEPA) cleaner, elimination of carpeting in favor of tile or hardwood floors, and acaricides or dust-proof covers for mattresses and bedding to control house dust mites. Allergen avoidance may be more difficult in the case of outdoor allergens and may have important life implications for individuals working outdoors or who experience occupational rhinitis.

Immunotherapy

Immunotherapy (allergen desensitization) is most often used by specialists for patients with more severe allergic rhinitis or for patients who do not tolerate or respond well to multiple medications. A program of immunotherapy requires once- or twice- weekly injections of escalating doses of allergen extracts over a period of months. This is followed by once- or twice- monthly maintenance injections, typically for a period of at least 2 to 3 years. Immunotherapy is costly and inconvenient to patients, but has the potential for continued efficacy after the treatment is discontinued (Durham, Walker, Varga, et al., 1999 ; Mosbech and Osterballe, 1988). Given the potential for long-term effectiveness, immunotherapy may be cost-effective compared to continuous treatment with medications for patients with more severe disease. In addition, immunotherapy has the potential to prevent the development of asthma (Ragusa, Passalacqua, Gambardella, et al., 1997).

Pharmacologic Therapy

Symptoms of allergic rhinitis may be treated with any of several different types of medication, including antihistamines, intranasal corticosteroids, decongestants, cromolyn sodium, and ipratropium. Each of these medications has a different mechanism of action and a different pattern of symptom relief. Clinically, these drugs are often used concurrently for improved symptom relief or for relief of multiple symptoms.

Antihistamines are the most commonly used medications for allergic rhinitis and are usually administered on an intermittent basis for patients with mild or seasonal symptoms. Oral antihistamines act in part by competitively inhibiting the binding of histamine to H1 receptors. Second generation oral antihistamines such as cetirizine, fexofenadine, loratadine, and desloratadine are more pharmacologically selective and less sedating than earlier antihistamines. A unique topical antihistamine, azelastine, is non-selective, but may be associated with less sedation and fewer other systemic adverse effects than oral non-selective antihistamines. Sedating and non-sedating antihistamines appear roughly equivalent for controlling symptoms of seasonal and perennial allergic rhinitis (Long, McFadden, DeVine, et al., 2002).

Intranasal corticosteroids are anti-inflammatory medications that require days to weeks for maximal symptom relief. Nasal steroids inhibit multiple steps in the inflammatory cascade of allergic rhinitis and provide excellent relief for numerous symptoms, including itching, sneezing, rhinorrhea, and nasal congestion. Multiple preparations are available: beclomethasone dipropionate (Beconase^® and Vancenase^®), budesonide (Rhinocort^®), flunisolide (Nasarel^® and Nasalide^®), fluticasone propionate (Flonase^®), mometasone (Nasonex^®), and triamcinolone acetonide (Nasacort^®). In head-to-head comparisons, nasal corticosteroids relieve allergic rhinitis symptoms more effectively than sedating or non-sedating antihistamines (Long, McFadden, DeVine, et al., 2002).

Nasal decongestants reduce nasal congestion through vasoconstriction. They are available in topical (phenylephrine, oxymetazoline) and oral (phenylephrine, pseudoephedrine) formulations. Oral agents are less likely to cause rebound vasodilation, accompanied by increased nasal congestion, than topical decongestants. Two studies have shown some benefit for nasal congestion but not for the other symptoms of allergic rhinitis (Long, McFadden, DeVine, et al., 2002).

Cromolyn sodium is postulated to prevent mast cell degranulation and is thus best used prophylactically. It requires four-times-per-day dosing and may require up to 2 weeks of continuous use for maximal benefit. In 32 randomized trials of cromolyn, all but two showed significant improvements in symptoms of allergic rhinitis. Cromolyn appeared to have higher efficacy for seasonal than perennial rhinitis. Dosing studies showed greater effect at higher doses (Long, McFadden, DeVine, et al., 2002). The anticholinergic ipratropium (Atrovent^® nasal) decreases rhinorrhea for non-allergic rhinitis and has the potential for similar benefits in allergic rhinitis (Long, McFadden, DeVine, et al., 2002).

Although drug treatments for allergic rhinitis are often used clinically in regimens that combine more than one drug from different classes, most clinical trials have focused on proving individual drugs superior to placebo (Long, McFadden, DeVine, et al., 2002). Combined drug treatments, compared with single-agent treatments, may work synergistically to provide greater efficacy, may complement one another to relieve a broader array of symptoms, and may allow lower dosing and, hence, reduce adverse effects.

Costs and Work Performance

The American Academy of Allergy, Asthma & Immunology estimates that approximately 19 million employed adults are affected by allergic rhinitis, resulting in several million lost work days each year and annual direct healthcare costs of $4.5 billion (American Academy of Allergy, Asthma & Immunology, 2000) . An evaluation of the evidence on costs and on work performance and symptoms requires the review of several types of literature. Determining the overall economic impact of allergic rhinitis requires a review of burden-of-illness studies. The effects of allergic rhinitis on work performance can be measured by studying employees' subjective estimates of their work performance and/or through the use of objective measurements of employee productivity. The impact of specific treatments can also be assessed by cost-effectiveness analysis , which estimates the costs associated with observed improvements in symptoms or quality of life, and by cost-benefit analysis, which considers the benefit of treatment in monetary terms, such as improvements in work productivity, balanced against the cost of treatment. There are few studies that directly associate allergic rhinitis symptoms and work performance, but studies of the treatment effects of various pharmacologic therapies, such as comparisons of sedating and non-sedating antihistamines, may be informative.

Treatment Outcomes by Clinician Specialty

The research question for this topic focuses on two issues: (a) whether different types of clinicians treat allergic rhinitis patients differently; and (b) whether treatment outcomes vary by type of clinician. Primary care clinicians are likely to be the first medical contact for someone with allergic rhinitis, and they have been shown to effectively treat a significant proportion of allergic rhinitis sufferers. On the other hand, allergy specialists and otolaryngologists tend to treat patients with more severe cases of allergic rhinitis (often referred by a primary care clinician), have more precise diagnostic tools available (e.g., nasal endoscopy), and are skilled in administering more specific and complex treatments (e.g., immunotherapy). Also at issue is whether there are variations in treatment and patient outcomes between specialists, i.e., between medically trained allergists and surgically trained otolaryngologists.

Prevalence and Patient Outcomes by Race and Ethnicity

Target Populations

We focused on patients with either seasonal or perennial allergic rhinitis. Given our focus on working populations, we prioritized studies in adults. Due to sparse data, we broadened the target population to include school-age children for questions with little relevant data in adults. Our rationale was that the clinical syndrome and underlying biology are similar in children and adults, and that effects on school performance may serve as a rough proxy for work productivity.

Subclinical or clinical asthma frequently co-exists with allergic rhinitis, and patients with co-occurring asthma were included in our review. Because data were extremely limited on the effects of environmental measures in adults with allergic rhinitis, we expanded our scope to patients with asthma. This decision is supported by the “unified airway” theory, according to which treatments for allergic rhinitis may affect asthma and, conversely, treatments for asthma may affect allergic rhinitis (Bousquet, van Cauwenberge, Khaltaev, et al., 2001).

We did not specifically target patients with occupational rhinitis. B y definition a work-related illness, occupational rhinitis has allergic and non-allergic mediators, but its prevalence is far lower than non-occupational allergic rhinitis.

Target Practice Settings

Because of the broad scope of this report, multiple practice settings were relevant. We were interested in primary care and specialty settings, where pharmacological and immunotherapy treatments are often initiated. Environmental control measures are usually prescribed in medical settings, but are typically carried out in the home. In addition, interventions aimed at increasing worker productivity may be designed for, or delivered in, the work setting.

Target Audience

Our principal audience is groups developing guidelines or educational documents on allergic rhinitis for healthcare professionals. In addition, we expect healthcare professionals who provide care to patients with allergic rhinitis will have a particular interest in the report. These include family physicians, internal medicine physicians, allergy specialists, otolaryngologists, occupational medicine physicians, nurse practitioners, and physician assistants. Secondary target audiences include employers, policymakers involved in payment decisions, agencies involved in funding research, media involved in dissemination and education about health issues, and patients interested in state-of-the-art medical literature.

Limitations of the Report

This report reviews published evidence relevant to the five key research questions listed above. It does not cover topics addressed in the evidence report on “Management of Allergic and Nonallergic Rhinitis” recently completed by the Evidence-based Practice Center at the New England Medical Center (Long, McFadden, DeVine, et al., 2002). The latter report includes comprehensive assessments of the literature on diagnosis of allergic and non-allergic rhinitis, efficacy of single-agent treatments for both conditions, and co-morbidity with asthma and acute rhinosinusitis.

Occupational rhinitis is much less common than non-occupational rhinitis, and includes both allergic and non-allergic causes. Because of its relatively high prevalence, non-occupational allergic rhinitis creates a greater burden in the workplace in terms of work performance and healthcare costs than does occupational rhinitis. Although occupational allergic rhinitis falls within the scope of this report, few data on this condition focus on the key questions addressed here, and thus nearly all the data reviewed concern allergic rhinitis associated with the most common allergens rather than workplace- specific exposures.

Finally, several agents are currently being evaluated in clinical trials, but are not yet in common use, and are thus not reviewed in this report. These agents include leukotriene inhibitors, anti-immunoglobulin E (anti-IgE) therapy, and cytokine antagonists.

Topic Assessment and Refinement

The American Association of Health Plans (AAHP) proposed the original topic for this report, “Seasonal Allergies, Effect on Working Populations.” An eight-member national advisory panel of technical experts, which included a representative of AAHP, was convened to work with the Duke research team to refine the key research questions and to review literature search strategies, inclusion and exclusion criteria, the causal pathway or evidence model, quality scoring criteria, interventions to be assessed, and specific outcomes to be reported in the evidence tables. The panel also assisted in identifying key research issues, advised on the scope of the project and methods, nominated peer reviewers, and reviewed preliminary drafts of research findings. Specialties represented on the panel included allergy and immunology, family medicine, general internal medicine, occupational medicine, otolaryngology, and pharmacology. Two meetings of the full panel were conducted via conference calls.

During its first conference call, the panel was presented with the five key research questions specified in the task order:

1. What is the appropriate treatment protocol for diagnosing and managing seasonal allergic rhinitis in a timely and cost-effective manner?

2. What measures can healthcare providers take to help prevent complications or reduce the severity of complications associated with chronic allergic rhinitis?

3. What is the role of new therapies such as anti-immunoglobulin E (anti-IgE) therapy and cytokine antagonists?

4. Can early interventions by allergy specialists reduce the rate of complications associated with chronic allergic rhinitis and lower costs?

5. Do treatment outcomes vary according to a patient's race or ethnicity?

Based on Duke's preliminary assessment of the literature and individual and group discussion with the advisory panel and the task order officer at the Agency for Healthcare Research and Quality (AHRQ), all parties agreed to refine the questions as follows:

1. How do currently clinically available treatments for allergic rhinitis affect costs and work performance?

2. What is the relationship between symptom outcomes or disease-specific quality-of-life measures and work performance among adults with allergic rhinitis? Can data on symptomatic outcome or quality of life be reliably translated into work performance measures?

3. How effective are (a) environmental measures, (b) immunotherapy, and (c) combined treatments, such as with antihistamines and nasal steroids or antihistamines and oral decongestants, for relief of symptoms in adults with allergic rhinitis?

4. How do different types of healthcare providers (generalists, allergy specialists, and otolaryngologists) treat adults with allergic rhinitis, and how do treatment outcomes vary by provider?

5. In adult patients with symptoms of allergic rhinitis, does the prevalence, treatment patterns or response to treatment vary according to a patient's race or ethnicity?

Given the changes in the research questions, after the second conference call and with the panel's agreement, we requested that the title of the task order be changed to “Management of Allergic Rhinitis in the Working-Age Population” to more accurately reflect the contents of the evidence report. This request was approved by AHRQ.

Causal Pathway

Figure 1. Causal Pathway

   Figure 1. Causal Pathway

Figure 1

Figure 1

Figure 1

Literature Search and Review

The comprehensive review of the literature, from identification of databases through abstraction of individual articles into evidence tables, was a multi-step, sequential process.

Search Strategy

We developed the basic search strategy using the National Library of Medicine's MeSH key word nomenclature developed for MEDLINE. The same strategy was used to search the other databases listed above. A Duke University Medical Center librarian checked the strategies and assisted with their translation to the key word structure used by EMBASE.

The initial searches, conducted in October 2001, were performed in MEDLINE, updated in MEDLINE in January 2002, and duplicated in additional databases in January 2002. All years of each database were searched - the periods covered by the searches are given above. The searches were limited to the English language and to human subjects. For topics concerning treatment efficacy, search terms focused on identifying randomized controlled trials, except in the case of the environmental measures topic, where the search strategy used additional, less restrictive, search terms, including “controlled trials” and “clinical trials.” Suggestions regarding search terms and specific articles were solicited from the advisory panel and resulted in additions to the literature database.

Table 3. Search strategy - preliminary general search, (more...)

Table 3. Search strategy - preliminary general search, MEDLINE, 1966 through September 2001

Set	Search term	Results
1	exp rhinitis/	12649
2	pollinosis.tw.	842
3	hay fever.tw.	1215
4	rhinitis.tw.	8000
5	or/1–4	15475
6	desensitization, immunologic/	4765
7	immunotherapy.tw.	15633
8	desensitization.tw.	11430
9	or/6–8	29720
10	and/5, 9	1679
11	limit 10 to human	1647
12	limit 11 to english language	1128
13	limit 12 to randomized controlled trial	159
14	exp filtration/	21390
15	air conditioning/	1546
16	air pollution, indoor/	2810
17	dust/	11250
18	“bedding and linens”/	2461
19	mites/	5942
20	environmental control.tw.	696
21	mite$.tw.	6141
22	or/14–21	45324
23	5 and 22	1312
24	limit 23 to human	1280
25	limit 24 to english language	930
26	limit 25 to randomized controlled trial	66
27	drug therapy, combination/	65666
28	5 and 27	142
29	limit 28 to human	138
30	limit 29 to english language	104
31	limit 30 to randomized controlled trial	54
32	exp psychology, industrial/	36848
33	exp “costs and cost analysis”/	110582
34	burden of illness.tw	188
35	or/32–34	144427
36	5 and 35	72
37	limit 36 to human	71
38	limit 37 to english language	68
39	leukotriene antagonists/tu	241
40	interleukin-4/tu	141
41	antibodies, anti-idiotypic/	9499
42	or/39–41	9879
43	5 and 42	106
44	limit 43 to human	103
45	limit 44 to english language	92
46	limit 45 to randomized controlled trial	17
47	quality of life/	28524
48	health status/	17994
49	karnofsky performance status/	404
50	activities of daily living/	21523
51	or/47–50	62587
52	5 and 51	117
53	limit 52 to human	117
54	limit 53 to english language	107
55	limit 54 to abstracts	94
56	exp anti-inflammatory agents, steroidal/tu	45608
57	5 and 56	619
58	limit 57 to human	614
59	limit 58 to english language	505
60	limit 59 to randomized controlled trial	190
61	cetirizine/tu	194
62	fexofenadine/tu	0
63	loratadine/tu	145
64	terfenadine/tu	168
65	or/61–64	441
66	exp histamine h1 antagonists/tu	7227
67	66 not 65	6786
68	5 and 65	225
69	limit 68 to human	223
70	limit 69 to english language	198
71	limit 70 to randomized controlled trial	127
72	limit 67 to human	6094
73	limit 72 to english language	4250
74	limit 73 to randomized controlled trial	787
75	71 or 74	914

Table 4. Search strategy - clinician specialty differences, (more...)

Table 4. Search strategy - clinician specialty differences, MEDLINE, 1966 to October Week 3 2001

Set	Search term	Results
1	physicians, family/	8358
2	exp physician's practice patterns/	11285
3	family practice/	38292
4	internal medicine/	9345
5	“referral and consultation”/	29576
6	specialties, medical/	11701
7	specialties, surgical/	935
8	surgery/	17749
9	exp attitude of health personnel/	55556
10	exp "outcome and process assessment (health	151936
11	“allergy and immunology”/	2635
12	or/1–11	310954
13	exp rhinitis/	12676
14	pollinosis.tw.	843
15	hay fever.tw.	1217
16	rhinitis.tw.	8034
17	or/13–16	15518
18	and/12, 17	450
19	from 18 keep 28, 43–44, 50, 52, 63, 66, 108, 110, 1	18
20	limit 18 to yr=1966-1998	289
21	limit 20 to yr=1966-1997	217
22	from 21 keep 30, 33, 40, 43, 88, 99, 107, 156, 205,	10

Table 5. Search strategy - environmental measures (1), (more...)

Table 5. Search strategy - environmental measures (1), MEDLINE, 1966 to October Week 1 2001

Set	Search term	Results
1	exp rhinitis/	12654
2	air pollutants, Environmental/ip	49
3	Allergens/ip	972
4	MITES/	5946
5	1 or 2 or 3 or 4	18967
6	Rhinitis/pc	64
7	air pollution/pc	2146
8	respiratory hypersensitivity/pc	206
9	dust/pc	288
10	Micropore Filters/	1779
11	FILTRATION/	11554
12	INSECTICIDES/	7545
13	Insect Control/	3225
14	air-cleaning.tw.	48
15	(air adj filter).tw.	96
16	(air adj cleaner$).tw.	48
17	acaricide.tw.	343
18	acardust.tw.	3
19	hepa.tw.	582
20	(allergen adj avoidance).mp. [mp=title, abstract, registry number word, mesh subject heading]	216
21	(allergen adj control).mp. [mp=title, abstract, registry number word, mesh subject heading]	27
22	(environmental adj control$).mp. [mp=title, abstract, registry number word, mesh subject heading]	811
23	6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19 or 20 or 21 or 22	27516
24	5 and 23	543
25	randomized-controlled-trial (pt)	151353
26	meta-analysis (pt)	5987
27	controlled-clinical-trial (pt)	58987
28	clinical-trial (pt)	319348
29	random$.ti, ab, sh.	254436
30	(meta-anal$ or metaanaly$ or meta analy$).ti, ab, sh.	9346
31	((doubl$ or singl$) and blind$).ti, ab, sh.	67067
32	exp Clinical trials/	127044
33	crossover.ti, ab, sh.	18070
34	25 or 26 or 27 or 28 or 29 or 30 or 31 or 32 or 33	501236
35	24 and 34	89

Table 6. Search strategy - environmental measures (2), (more...)

Table 6. Search strategy - environmental measures (2), MEDLINE, 1966 to October Week 1 2001

Set	Search term	Results
1	exp rhinitis/	12654
2	air pollutants, Environmental/ip	49
3	Allergens/ip	972
4	MITES/	5946
5	1 or 2 or 3 or 4	18967
6	Rhinitis/pc	64
7	air pollution/pc	2146
8	respiratory hypersensitivity/pc	206
9	dust/pc	288
10	Micropore Filters/	1779
11	FILTRATION/	11554
12	INSECTICIDES/	7545
13	Insect Control/	3225
14	air-cleaning.tw.	48
15	(air adj filter).tw.	96
16	(air adj cleaner$).tw.	48
17	acaricide.tw.	343
18	acardust.tw.	3
19	hepa.tw.	582
20	6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 or 19	26537
21	Randomized Controlled Trials/	20303
22	5 and 20	421
23	21 and 22	1
24	pollinosis.tw.	842
25	hay fever.tw.	1216
26	rhinitis.tw.	8011
27	mite$.tw.	6147
28	5 or 24 or 25 or 26 or 27	23563
29	exp filtration/	21404
30	air conditioning/	1548
31	air pollution, indoor/	2815
32	dust/	11255
33	“bedding and linens”/	2463
34	20 or 29 or 30 or 31 or 32 or 33	51223
35	randomized-controlled-trial (pt)	151353
36	meta-analysis (pt)	5987
37	controlled-clinical-trial (pt)	58987
38	clinical-trial (pt)	319348
39	random$.ti, ab, sh.	254436
40	(meta-anal$ or metaanaly$ or meta analy$).ti, ab, sh.	9346
41	((doubl$ or singl$) and blind$).ti, ab, sh.	67067
42	exp Clinical trials/	127044
43	crossover.ti, ab, sh.	18070
44	35 or 36 or 37 or 38 or 39 or 40 or 41 or 42 or 43	501236
45	28 and 34	2799
46	44 and 45	291
47	limit 46 to (human and english language)	224

The basic search strategies used are reproduced in Tables 3 to 6.

Screening Criteria

Inclusion and exclusion criteria were developed for the literature searches so that the yield of articles would be appropriately focused. Citations were excluded based on the following criteria:

• Article was not original research;

• Article did not address allergic rhinitis or was not applicable to the key research questions;

• The study design was a single case report;

• The study design was a small case series with 20 or fewer subjects.

Empirical studies were included based on the following criteria:

• The study population must address allergic rhinitis;

• All original research or relevant reviews must relate to at least one of the five key research questions;

• 

  Table 7. Included study designs, by key research question (more...)

  Table 7. Included study designs, by key research question

  Question	Topic	Included study designs
  1	Costs and work performance	Any empirical study involving more than 20 patients with allergic rhinitis. Includes randomized controlled trials (RCTs), case series, cohort studies, non-randomized comparison studies, surveys, and secondary data analyses.
  2	Relationship between symptom outcomes or disease-specific quality of life and work performance
  3a	Environmental measures	RCTs, non-randomized prospective cohort comparisons
  3b	Immunotherapy	RCTs, pseudo-randomized placebo-controlled trials
  3c	Combination drug therapy
  4	Clinician specialty differences	Any empirical study involving more than 20 patients with allergic rhinitis. Includes RCTs, case series, cohort studies, non-randomized comparison studies, surveys, and secondary data analyses.
  5	Racial and ethnic variation

  Included study designs varied depending on the key research question being addressed (Table 7). Randomized controlled trials (RCTs) were included for all questions. For question 3a (environmental measures), we also included non-randomized prospective cohort comparisons. For questions 3b (immunotherapy) and 3c (combined treatments), we included RCTs and pseudo-randomized placebo-controlled trials. We defined “pseudo-randomized” to mean using some unbiased but non-random method of allocation, such as enrollment order, identification number, or date of birth. For question 1 (costs and work performance), question 2 (relationship between symptom outcomes or disease-specific quality of life and work performance), question 4 (clinician specialty differences), and question 5 (racial and ethnic variation), we included RCTs, large case series (> 20 subjects), cohort studies, non-randomized comparison studies, and articles reporting data from surveys and secondary data analyses.

Table 8. Abstract and full-text screening criteria

Table 8. Abstract and full-text screening criteria

Key research questions:
How do currently clinically available treatments for allergic rhinitis affect costs and work performance? What is the relationship between symptom outcomes or disease-specific quality-of-life measures and work performance among adults with allergic rhinitis? Can data on symptomatic outcome or quality of life be reliably translated into work performance measures? How effective are (a) environmental measures, (b) immunotherapy, and (c) combined treatments, such as with antihistamines and nasal steroids or antihistamines and oral decongestants, for relief of symptoms in adults with allergic rhinitis? How do different types of healthcare providers (generalists, allergy specialists, and otolaryngologists) treat adults with allergic rhinitis, and how do treatment outcomes vary by provider? In adult patients with symptoms of allergic rhinitis, does the prevalence, treatment patterns or response to treatment vary according to a patient's race or ethnicity?
Inclusion/exclusion criteria:
Not original research or relevant review Not allergic rhinitis or allergic rhinitis not applicable to research questions Case report Small case series (≤ 20 patients, no controls) Large case series (> 20 patients, no controls) Non-randomized assignment to treatment (comparison group, but not randomly assigned) Randomized controlled trial Relevant review Original research on other aspects (for use as background or in model, e.g., prevalence, natural history, diagnostic testing) Basic science Survey and secondary data
Inclusion rules:
Question 1:	codes 5–9, 11:	Evidence tables for codes 5, 6 7, 11
Question 2:	codes 5–9, 11:	Evidence tables for codes 5, 6, 7, 11
Question 3a:	codes 6–9, 11:	Evidence tables for codes 6, 7
Question 3b:	codes 7–9, 11:	Evidence tables for code 7
Question 3c:	codes 7–9, 11:	Evidence tables for code 7
Question 4:	codes 5–9, 11:	Evidence tables for codes 5, 6, 7, 11
Question 5:	codes 5–9, 11:	Evidence tables for codes 5 , 6, 7, 11

The final version of the abstract and full-text screening criteria is shown in Table 8.

Screening Results

The literature search yielded 1, 593 articles. The titles and abstracts of these articles were reviewed against the inclusion/exclusion criteria by the investigators. Two investigators reviewed each abstract. When no abstract was available, the title, source, and keywords were screened. At this stage, articles were included if requested by one investigator. The full text of each article passing the title-and-abstract screen was retrieved from the library for further review.

At the full-text screening stage, each article was independently reviewed by two investigators, who forwarded their decisions to the task order manager for recording and comparison. If indicated, reviewers were asked to reconcile differences of opinion and return a reconciled final decision to the task order manager. Overall, the teams reconciled about 40 percent of their decisions. If team members had difficulty reaching agreement on decisions, or submitted indecisive codes, the principal investigator was the arbiter. This situation arose in about 10 percent of the reconciled decisions, largely when “include” or “exclude” decisions were at variance with the study design (e.g., an RCT coded as “exclude”).

Table 9. Summary of results of abstract and full-text (more...)

Table 9. Summary of results of abstract and full-text screening

Articles identified	1593
Abstracts:
Included	546
Excluded	1089
Full-text articles:
Included	258
Excluded	288

Table 10. Full-text screening results, by key research (more...)

Table 10. Full-text screening results, by key research question and by inclusion/exclusion criteria

INCLUDED ARTICLES
(ET = included in evidence tables)
Question 1 (Note: one article screened for this question reported results of both an RCT and a large case series)	54
5-large case series (> 20 patients, no controls): ET	14
6-non-randomized controlled trials: ET	0
7-randomized controlled trial: ET	7
8-relevant review	11
9-original research on other aspects for use in background or model	13
11-survey or secondary data: ET	11
Question 2 (screened with Question 1 articles)	6
5-large case series (> 20 patients, no controls): ET	0
6-non-randomized controlled trials: ET	0
7-randomized controlled trial: ET	3
8-relevant review	2
9-original research on other aspects for use in background or model	0
11-survey or secondary data: ET	1
Question 3a (environmental measures)	40
6-non-randomized controlled trials: ET	1
7-randomized controlled trial: ET	26
8-relevant review	9
9-original research on other aspects for use in background or model	0
11-survey or secondary data	4
Question 3b (immunotherapy)	80
7-randomized controlled trial: ET	62
8-relevant review	11
9-original research on other aspects for use in background or model	4
11-survey or secondary data	3
Question 3c (combination treatments)	32
7-randomized controlled trial: ET	31
8-relevant review	0
9-original research on other aspects for use in background or model	1
Question 4	26
5-large case series (> 20 patients, no controls): ET	4
6-non-randomized controlled trials: ET	0
7-randomized controlled trial: ET	0
8-relevant review	12
9-original research on other aspects for use in background or model	6
11-survey or secondary data: ET	1
Question 5	8
5-large case series (> 20 patients, no controls)	1
6-non-randomized controlled trials	0
7-randomized controlled trial	0
8-relevant review	3
9-original research on other aspects for use in background or model	0
11-survey or secondary data	4
EXCLUDED ARTICLES
Question 1	82
1-not original research or relevant review	24
2-not allergic rhinitis or not applicable to study questions	48
3-case report	0
4-small case series (≤ 20 patients, no controls)	1
10-basic science	0
Excluded during data abstraction (e.g., no relevant data reported)	9
Question 2 (screened with Question 1 articles)	15
1-not original research or relevant review	6
2-not allergic rhinitis or not applicable to study questions	5
3-case report	0
4-small case series (≤ 20 patients, no controls)	0
5-large case series (> 20 patients, no controls)	0
10-basic science	1
Excluded during data abstraction (no relevant data)	3
Question 3a (environmental measures)	41
1-not original research or relevant review	10
2-not allergic rhinitis or not applicable to study questions	10
3-case report	0
4-small case series (≤ 20 patients, no controls)	0
5-large case series (> 20 patients, no controls)	3
6-non-randomized controlled trials	0
10-basic science	11
Excluded during data abstraction (e.g., no relevant data, insufficient data, no symptom outcomes or other relevant outcomes, only atopic dermatitis)	7
Question 3b (immunotherapy):	87
1-not original research or relevant review	5
2-not allergic rhinitis or not applicable to study questions	71
3-case report	0
4-small case series (≤ 20 patients, no controls)	0
5-large case series (> 20 patients, no controls)	1
6-non-randomized controlled trials	4
10-basic science	2
Excluded during data abstraction (e.g., no separate results for allergic rhinitis, asthma data only, no symptom outcomes)	4
Question 3c (combination treatments)	25
1-not original research or relevant review	5
2-not allergic rhinitis or not applicable to study questions	14
3-case report	0
4-small case series (≤ 20 patients, no controls)	0
10-basic science	0
Excluded during data abstraction (no relevant data)	6
Question 4	30
1-not original research or relevant review	6
2-not allergic rhinitis or not applicable to study questions	9
3-case report	0
4-small case series (≤ 20 patients, no controls)	1
10-basic science	1
Excluded during data abstraction (e.g., no relevant allergic rhinitis data; no data on provider differences)	13
Question 5	21
1-not original research	2
2-not allergic rhinitis or not applicable to study questions	18
3-case report	0
4-small case series (≤ 20 patients, no controls)	0
10-basic science	0
Excluded during data abstraction (e.g., no relevant data)	1

The records in the literature database were coded at each screening stage. A summary of the results of the title-and-abstract and full-text screenings is provided in Table 9. A more detailed accounting of the screening process is provided in Table 10.

Data Abstraction

Table 11. Partial data abstraction - sample

Table 11. Partial data abstraction - sample

Study	Design and Interventions	Patient Population	Outcomes Reported	Results	Quality Score/Notes
Andri, Senna, Betteli, et al., 1992	Design: RCT, parallel-group, method of randomization not described	No. of subjects at start: 30	1) Investigator-assessed symptom severity	1) Investigator-assessed symptom severity: DO NOT ABSTRACT	[IF ARTICLE SHOULD BE EXCLUDED, PLEASE EXPLAIN WHY HERE]
		Dropouts/withdrawals:
			2) Patient-assessed symptom severity: nasal itching, nasal obstruction, sneezing, running nose, eye irritation, and eye watering graded daily by patients scale of 0 (none) to 3 (severe)	2) Patient-assessed symptom severity:
	Interventions:	No. of subjects at end:			Quality Scoring:
#210	1) Terfenadine 60 mg bid + nimesulide 100 mg bid (n = 15)	Inclusion criteria:
				3) Patient global assessment of efficacy:
	2) Terfenadine 60 mg bid + placebo (n = 15)	Exclusion criteria:			Notes:
		Age:			Local pollen counts conducted daily during trial.
	Duration of study treatment: 30 days	Sex:	3) Patient global assessment of efficacy: recorded once at end of trial - categorical scale keyed to perceived degree of improvement in symptoms (< 50%, 50–80%, > 80%)	4) Adverse events:
	No other drugs “likely to affect hay fever” permitted	Race:
	No pre-trial washout period described	[IF RESULTS ARE BROKEN DOWN BY RACE/ETHNICITY, PLEASE MAKE THIS CLEAR IN “RESULTS” COLUMN]
	Dates:
		Other:	4) Adverse events: Not clear how reported/ recorded
	Location:
	Setting:
	Type(s) of providers:

In the partial abstraction performed by the senior writer/editor, all outcomes reported were listed, and the outcomes meeting our criteria were selected for abstraction. We required patient-assessed symptom outcomes for efficacy questions; we also reported quality of life, functional status, adverse events, and patient global assessments for these questions. For all questions, we recorded work performance and cost outcomes. Specifically, outcomes abstracted for each key research question were as follows:

• Question 1:

  • Work performance

  • Costs (direct medical or non-medical)

  • Costs (indirect)

• Question 2:

  • Association between symptoms and work performance

  • Association between quality-of-life and work performance

• Question 3:

  • Symptoms, assessed by patients

  • Quality of life

  • Functional status

  • Global assessments by patients

  • Adverse events

• Question 4:

  • Practice patterns by provider specialty (referral, drug and other treatment use, case mix)

  • Drug and other treatment response by provider specialty

• Question 5:

  • Allergic rhinitis prevalence by racial/ethnic groups

  • Severity of allergic rhinitis by racial/ethnic groups

  • Provider consultation by racial/ethnic groups

  • Drug and other treatment use by racial/ethnic groups

  • Drug and other treatment response by racial/ethnic groups

Grading of Articles (Quality Scoring)

We evaluated each article included in the evidence tables for factors affecting internal and external validity. The quality scoring criteria are given below:

Internal validity:

1. 

   Table 12. Oxford Centre for Evidence-based Medicine (more...)

   Table 12. Oxford Centre for Evidence-based Medicine levels of evidence (May 2001) ¹

   Level	Therapy/prevention, aetiology/harm	Prognosis	Diagnosis	Differential diagnosis/symptom prevalence study	Economic and decision analyses
   1a	Systematic review (SR) (with homogeneity*) of RCTs	SR (with homogeneity*) of inception cohort studies; CDR† validated in different populations	SR (with homogeneity*) of Level 1 diagnostic studies; CDR† with 1b studies from different clinical centres	SR (with homogeneity*) of prospective cohort studies	SR (with homogeneity*) of Level 1 economic studies
   1b	Individual RCT (with narrow Confidence Interval‡)	Individual inception cohort study with ≥ 80% follow-up; CDR† validated in a single population	Validating** cohort study with good††† reference standards; or CDR† tested within one clinical centre	Prospective cohort study with good follow-up****	Analysis based on clinically sensible costs or alternatives; systematic review(s) of the evidence; and including multi-way sensitivity analyses
   1c	All or none§	All or none case-series	Absolute SpPins and SnNouts††	All or none case-series	Absolute better-value or worse-value analyses ††††
   2a	SR (with homogeneity* ) of cohort studies	SR (with homogeneity*) of either retrospective cohort studies or untreated control groups in RCTs	SR (with homogeneity*) of Level >2 diagnostic studies	SR (with homogeneity*) of 2b and better studies	SR (with homogeneity*) of Level >2 economic studies
   2b	Individual cohort study (including low quality RCT; e.g., <80% follow-up)	Retrospective cohort study or follow-up of untreated control patients in an RCT; Derivation of CDR† or validated on split-sample§§§ only	Exploratory** cohort study with good†††reference standards; CDR† after derivation, or validated only on split-sample§§§ or databases	Retrospective cohort study, or poor follow-up	Analysis based on clinically sensible costs or alternatives; limited review(s) of the evidence, or single studies; and including multi-way sensitivity analyses
   2c	“Outcomes” Research; Ecological studies	“Outcomes” Research		Ecological studies	Audit or outcomes research
   3a	SR (with homogeneity*) of case-control studies		SR (with homogeneity*) of 3b and better studies	SR (with homogeneity*) of 3b and better studies	SR (with homogeneity*) of 3b and better studies
   3b	Individual Case-Control Study		Non-consecutive study; or without consistently applied reference standards	Non-consecutive cohort study, or very limited population	Analysis based on limited alternatives or costs, poor quality estimates of data, but including sensitivity analyses incorporating clinically sensible variations.
   4	Case-series (and poor quality cohort and case-control studies§§ )	Case-series (and poor quality prognostic cohort studies***)	Case-control study, poor or non-independent reference standard	Case-series or superseded reference standards	Analysis with no sensitivity analysis
   5	Expert opinion without explicit critical appraisal, or based on physiology, bench research or “first principles”	Expert opinion without explicit critical appraisal, or based on physiology, bench research or “first principles”	Expert opinion without explicit critical appraisal, or based on physiology, bench research or “first principles”	Expert opinion without explicit critical appraisal, or based on physiology, bench research or “first principles”	Expert opinion without explicit critical appraisal, or based on economic theory or “first principles”

   1

   Produced by Bob Phillips, Chris Ball, Dave Sackett, Doug Badenoch, Sharon Straus, Brian Haynes, Martin Dawes since November 1998. Available at: http://163.1.96.10/docs/levels.html#levels. Accessed May 30, 2002.

   Users can add a minus-sign “-” to denote the level of that fails to provide a conclusive answer because of:

   · EITHER a single result with a wide Confidence Interval (such that, for example, an ARR in an RCT is not statistically significant but whose confidence intervals fail to exclude clinically important benefit or harm)

   · OR a Systematic Review with troublesome (and statistically significant) heterogeneity.

   · Such evidence is inconclusive, and therefore can only generate Grade D recommendations.

   *By homogeneity we mean a systematic review that is free of worrisome variations (heterogeneity) in the directions and degrees of results between individual studies. Not all systematic reviews with statistically significant heterogeneity need be worrisome, and not all worrisome heterogeneity need be statistically significant. As noted above, studies displaying worrisome heterogeneity should be tagged with a “-” at the end of their designated level.

   †Clinical Decision Rule. (These are algorithms or scoring systems which lead to a prognostic estimation or a diagnostic category. )

   ‡See note #2 for advice on how to understand, rate and use trials or other studies with wide confidence intervals.

   §Met when all patients died before the Rx became available, but some now survive on it; or when some patients died before the Rx became available, but none now die on it.

   §§By poor quality cohort study we mean one that failed to clearly define comparison groups and/or failed to measure exposures and outcomes in the same (preferably blinded), objective way in both exposed and non-exposed individuals and/or failed to identify or appropriately control known confounders and/or failed to carry out a sufficiently long and complete follow-up of patients. By poor quality case-control study we mean one that failed to clearly define comparison groups and/or failed to measure exposures and outcomes in the same (preferably blinded), objective way in both cases and controls and/or failed to identify or appropriately control known confounders.

   §§§Split-sample validation is achieved by collecting all the information in a single tranche, then artificially dividing this into “derivation” and “validation” samples.

   ††An “Absolute SpPin” is a diagnostic finding whose Specificity is so high that a Positive result rules-in the diagnosis. An “Absolute SnNout” is a diagnostic finding whose Sensitivity is so high that a Negative result rules-out the diagnosis.

   ††Good, better, bad and worse refer to the comparisons between treatments in terms of their clinical risks and benefits.

   ††† Good reference standards are independent of the test, and applied blindly or objectively to applied to all patients. Poor reference standards are haphazardly applied, but still independent of the test. Use of a non-independent reference standard (where the ‘test’ is included in the ‘reference’, or where the ‘testing’ affects the ‘reference’) implies a level 4 study.

   ††††Better-value treatments are clearly as good but cheaper, or better at the same or reduced cost. Worse-value treatments are as good and more expensive, or worse and the equally or more expensive.

   **Validating studies test the quality of a specific diagnostic test, based on prior evidence. An exploratory study collects information and trawls the data (e.g. using a regression analysis) to find which factors are ‘significant’.

   ***By poor quality prognostic cohort study we mean one in which sampling was biased in favour of patients who already had the target outcome, or the measurement of outcomes was accomplished in <80% of study patients, or outcomes were determined in an unblinded, non-objective way, or there was no correction for confounding factors.

   ****Good follow-up in a differential diagnosis study is >80%, with adequate time for alternative diagnoses to emerge (eg 1–6 months acute, 1 – 5 years chronic)

   Grades of Recommendation

   A consistent level 1 studies

   B consistent level 2 or 3 studies or extrapolations from level 1 studies

   C level 4 studies or extrapolations from level 2 or 3 studies

   D level 5 evidence or troublingly inconsistent or inconclusive studies of any level

   “Extrapolations” are where data is used in a situation which has potentially clinically important differences than the original study situation.

   “Extrapolations” are where data is used in a situation which has potentially clinically important differences than the original study situation.

   What is the level of evidence (Oxford Centre for Evidence-based Medicine, 2001; seeTable 12)?

2. Were the main outcomes of interest measured in a way that has been demonstrated empirically to be valid and reliable (e.g., using a standardized scale such as the Rhinoconjunctivitis Quality of Life Questionnaire [RQLQ] or the Medical Outcome Study Short-Form Health Survey [SF-36])?

   External validity:

3. Was the study population described and reasonably similar to an adult working US population? (Based mostly on age of study population.)

4. Were the intervention protocols referenced or described in sufficient detail to replicate?

5. Was the presence of comorbid asthma (or other upper respiratory conditions) described in the study population?

6. Was the diagnosis of allergic rhinitis based on physician diagnosis?

7. If physician-diagnosed, was the diagnosis supported by objective evidence of allergy (e.g. skin prick or serum IgE antibody testing)?

Additional quality criteria were applied to studies on environmental measures, immunotherapy, and combination therapy:

1. Was the study described as “randomized”?

2. If the method for concealing allocation from the investigators was described, was it adequate (table of random numbers, computer generated, coin toss, etc.) or inadequate (alternating, date of birth, hospital number, etc.)?

3. Was the study described as “double-blind”?

4. If the method of double-blinding was described, was it adequate (e.g., identical placebo, active placebo, injection vs. tablet with double dummy) or inadequate (e.g., tablet vs. injection with no double dummy)?

5. Did the study describe dropouts and withdrawals so that all patients entering the trial could be accounted for?

6. Was the analysis performed according to the intention-to-treat principle? (Did the analysis in some way consider all patients that were allocated to treatment, including dropouts and withdrawals?)

We did not aggregate these items into an overall quality score; rather, we considered and reported them individually. We favored this approach for several reasons:

• Previous work has shown that numeric grading systems may not discriminate well between “high” and “low” quality studies, even for randomized trials (Jüni, Witschi, Bloch, et al., 1999; Moher, Cook, Jadad, et al., 1996).

• Development and use of a new quality score would require additional work for validation, for which there is no time or budget allocation in the task order.

• Identification of specific weaknesses in each study will be helpful in identifying trends, which in turn will assist with our recommendations for future research.

• Describing key design components, rather than assigning a single aggregate score, is also consistent with recent recommendations from an expert panel on meta-analysis of observational studies (Stroup, Berlin, Morton, et al., 2000).

Summaries of each quality evaluation are provided in the far right column of the evidence tables. Grades were assigned by the primary abstractor and confirmed by the over-reader. When required, additional notes were made in the same column of the evidence table.

Quality Control Procedures

We employed quality-monitoring checks at every phase of the literature search, review, and data abstraction process to reduce bias, enhance consistency, and check the accuracy of screening. The quality checks included:

• Medical librarian review of the literature search strategy;

• Review of literature search strategies by advisory panel of technical experts;

• Check on completeness of the literature search results through reference list checks by the screener of each article;

• Reconciliation of all differences of opinion by reviewers on all full-text articles;

• Agreement of two reviewers for all eligible studies;

• Data abstractions completed by one investigator and reviewed (over-read) by another;

• Additional checks of evidence table entries for completeness and accuracy by a non-physician abstractor;

• Solicitation of advice at key decision points from the advisory panel of technical experts;

• Expert peer review of complete draft evidence report.

References

Costs and Work Performance

Introduction

This section addresses key research questions 1 and 2:

1. How do currently clinically available treatments for allergic rhinitis affect costs and work performance?

2. What is the relationship between symptom outcomes or disease-specific quality-of-life measures and work performance among adults with allergic rhinitis? Can data on symptomatic outcomes or quality of life be reliably translated into work performance measures?

To address the first question, we considered burden-of-illness studies of allergic rhinitis, as well as cost-comparison and cost-effectiveness studies. For the second question, we sought data correlating work performance either with symptoms of allergic rhinitis or with disease-specific quality of life. A strong association would permit the use of symptom or quality-of-life data, which are much more commonly reported than work-performance data, in economic analyses comparing treatment approaches.

After consulting with the project's advisory panel of experts, we elected to include data on school performance in children as a proxy for work performance in adults, because of the limited data on adults.

Table 13. Summary of types of data reported in studies (more...)

Table 13. Summary of types of data reported in studies abstracted in Evidence Table 1

Study	Data source	Per-patient burden of illness for selected populations	Total burden-of-illness estimates for US population	Cost- effectiveness	Work performance	Symptoms	Health-related quality of life
Blanc, Trupin, Eisner, et al., 2001	Telephone survey	X1			X	X	X
Burton, Conti, Chen, et al., 2001	Survey, work productivity data				X
Cockburn, Bailit, Berndt, et al., 1999a; Cockburn, Bailit, Berndt, et al., 1999b	Prescription claims data, work productivity data				X
Crystal-Peters, Crown, Goetzel, et al., 2000	1995 National Health Interview Survey and Bureau of Labor Statistics		X2
Cuffel, Wamboldt, Borish, et al., 1999	Health care claims
Donahue, Greineder, Connor-Lacke, et al., 1999	Health care claims	X
Fell, Mabry, and Mabry, 1997	Survey	X1			X	X
Gilmore, Alexander, Mueller, et al., 1996	Health care claims
Keith, Haddon, and Birch, 2000	Randomized controlled trial	X		X3
Kessler, Almeida, Berglund, et al., 2001	Survey	X			X
Kozma, Schulz, Sclar, et al., 1996	Randomized controlled trial			X		X
Lee, Cummins, and Okamoto, 2001	Health care claims	X
Leickly, Sears-Ewald, and Ownby, 1989	Randomized controlled trial					X
Liao, Leahy, and Cummins, 2001	Health care claims	X
Malone, Lawson, Smith, et al., 1997	1987 National Medical Expenditure Survey		X
Manor, Matthews, and Power, 2001	Survey					X
McMenamin, 1994	Multiple national surveys, government statistics		X
Meltzer, Casale, Nathan, et al., 1999	Randomized controlled trial				X		X
Ray, Baraniuk, Thamer, et al., 1999	Multiple national surveys, expert opinion		X
Reilly, Tanner, and Meltzer, 1996	Randomized controlled trial				X	X	X
Revicki, Leidy, Brennan-Diemer, et al., 1998	Survey				X	X	X
Ross, 1996	Multple national surveys, government statistics				X
Santilli, Nathan, Glassheim, et al., 2001	Survey	X1				X
Santos, Cifaldi, Gregory, et al., 1999 (Study 1)	Health care claims	X
Santos, Cifaldi, Gregory, et al., 1999 (Study 2)	Randomized controlled trials	X			X	X	X
Schädlich and Brecht, 2000	Multiple published estimates			X
Stahl, van Rompay, Wang, et al., 2000	Randomized controlled trials			X
Storms, Meltzer, Nathan, et al., 1997	Survey		X		X
Sussman, Mason, Compton, et al., 1999	Randomized controlled trials				X	X
Tanner, Reilly, Meltzer, et al., 1999	Randomized controlled trials				X		X
Trotter, 2000	Prescription claims	X
Yawn, Yunginger, Wollan, et al., 1999	Patient registry	X

1

Costs not assigned, but estimates of resource utilization reported.

2

Indirect costs only.

3

Cost-benefit analysis in which benefits were measured with a willingness-to-pay survey.

Thirty-two studies were included in the analysis of these questions (Table 13 and Evidence Table 1). Studies of costs included burden-of-illness studies (per-patient burden-of-illness studies of selected populations and total burden-of-illness studies for the US population) and cost-effectiveness studies (including cost-benefit and cost-minimization studies). Table 13 indicates which studies reported work-performance outcomes, and which of these also reported data on symptoms and/or health-related quality of life.

Environmental Measures

Immunotherapy

Results

Studies Identified

Sixty trials were included (see Evidence Table 3). All were required to report a clinical outcome measure based on patient assessment of symptoms and/or medication use for symptom relief. For the purposes of this discussion, trials have been separated into studies of immunotherapy for seasonal and perennial allergic rhinitis. The rationale for this division is based upon differing patterns of allergen exposure, which often correspond to differing immunotherapy protocols. Patients with seasonal allergic rhinitis symptoms may experience short periods of allergen exposure with relatively asymptomatic periods between exposures, whereas patients with perennial allergic rhinitis may have allergic responses to year-round allergens such as dust mite and cat dander. Alternatively, a significant percentage of patients experience year-round symptoms, but have multiple sensitivities to pollen, mold, and environmental allergens. Regarding immunotherapy protocols, seasonal rhinitis IT may be given continuously year-round or pre-seasonally only. The vast majority of trials considered in this report relate to seasonal allergic rhinitis caused by pollen. Only a small number of placebo-controlled trials have been performed to assess the effectiveness of IT to house dust mite or pet allergens.

Seasonal Allergic Rhinitis

Table 14. Placebo-controlled randomized controlled (more...)

Table 14. Placebo-controlled randomized controlled trials of injection immunotherapy (IT) for seasonal allergic rhinitis, by type of allergen

Allergen	Number of trials	Number of subjects	Number of trials favoring IT	Number of trials with negative or equivocal results
Ragweed	18	990	14	4
Grass (any)	13	604	12	1
Tree (any)	7	168	7	0
Parietaria	4	170	4	0

General literature review. Forty-eight trials of IT in the treatment of seasonal allergic rhinitis, with a total enrollment of 2,827 subjects, are summarized in Evidence Table 3. Ragweed pollen was the most commonly studied allergen, followed by grass pollen, tree pollen, and the weed pollen, Parietaria. All but 14 of the studies employed a seasonal treatment protocol in which subjects were given IT for 4 to 40 weeks prior to the expected pollen-exposure period. Most subjects were recruited into seasonal allergic rhinitis trials based upon symptoms occurring during the period of known exposure to the study allergen. The majority of studies employed a combined symptom-medication scale to collect patient response data. The method and grading system used to collect these data varied from study to study. None of the published studies gave detailed descriptions of measures used to ensure compliance with symptom/medication diary recording, and none provided detailed information on the percentage of expected data points that were actually collected. With one exception, all trials employed a single allergen or class of allergens (e.g., ragweed allergen or mixed grass allergen) in the treatment protocol. This is in contrast to the common clinical practice of formulating vaccines that include most or all of the allergens to which a patient is sensitive. A summary of the results of placebo-controlled trials of IT for seasonal allergic rhinitis is provided in Table 14.

Among the 48 included trials were several unique trial designs. Two trials compared a method of low-dose immunotherapy, designated the Rinkel method, with standard IT or placebo (Hirsch, Kalbfleisch, Golbert, et al., 1981; Van Metre, Adkinson, Amodio, et al., 1980). In both trials, the Rinkel method was found to be no more effective than placebo. As a result, expert panels have recommended against using the Rinkel method of immunotherapy (Bousquet, Lockey, and Malling, 1998). Two trials employed a withdrawal of therapy strategy in which subjects receiving maintenance doses of IT were randomized to receive continued immunotherapy or placebo for from 1 to 3 years (Durham, Walker, Varga, et al., 1999; Naclerio, Proud, Moylan, et al., 1997). The intent of these studies was to determine the durability of clinical and immunological responses to standard immunotherapy. At the end of the observation periods, the placebo group in each trial maintained clinical response levels similar to those measured in the group receiving continued treatment, indicating that clinical responses related to IT were durable beyond the actual treatment period.

Three trials compared immunotherapy with active medical treatment. In a 3-year trial comparing grass pollen immunotherapy with ketotifen (a drug approved in several European countries), the results favored immunotherapy (Dolz, Martinez-Cocera, Bartolome, et al., 1996). Two short-term trials compared birch or ragweed IT with nasal corticosteroids (Juniper, Kline, Ramsdale, et al., 1990; Rak, Heinrich, Jacobsen, et al., 2001). The results favored medical therapy over IT. However, it should be noted that the duration of immunotherapy was 6 weeks in each of these studies, which may not have been long enough to allow optimal immunologic response to IT, whereas nasal corticosteroids are known to be effective within this short time frame.

Safety data were reported in 38 of the 48 trials reviewed. The most common adverse events described were local reactions (either immediate or late) at the IT injection site. Systemic reactions characterized by generalized urticaria, increased rhinitis symptoms, increased asthma symptoms, or mild anaphylaxis were less common than local reactions and were apparently easily controlled. The percentage of subjects with systemic reactions varied from zero to approximately 25 percent. There were no reports of hospitalizations or deaths related to IT. No standardized methods for describing the characteristics or severity of allergic reactions to immunotherapy have been devised, making the interpretation of the adverse event data difficult.

Table 15. Data abstracted for meta-analysis of placebo-controlled (more...)

Table 15. Data abstracted for meta-analysis of placebo-controlled trials of immunotherapy (IT) for seasonal allergic rhinitis

Study	Allergen	Symptom measurement period	Outcome	IT mean	IT SD	IT n	Placebo mean	Placebo SD	Placebo n	Statistical test	P-value	IPD?
Ariano, Kroon, Augeri, et al., 1999	Tree	7 mo	Combined Sx/Rx	550 (median)	NR	11	1250 (median)	NR	11	Non-parametric	p = 0.02	No
Arvidsson, Löwhagen, and Rak, 2002	Tree	6 wk	Sx severity	1.3 (median)	0–5.2 (range)	22	2.1 (median)	0.6–5.6 (range)	24	Non-parametric	p = 0.05	No
Arvidsson, Löwhagen, and Rak, 2002	Tree	6 wk	Rx use	NR	NR	22	NR	NR	24	Non-parametric	p = 0.004	No
Bernstein, Tennenbaum, Georgakis, et al., 1976	Ragweed	4 wk	Sx severity	1.097 (mean daily score)	NR	58 (est.)	1.378 (mean daily score)	NR	54 (est.)	Not specified	p < 0.05	No
Bernstein, Tennenbaum, Georgakis, et al., 1976	Ragweed	4 wk	Rx use	0.411 (measured score)	NR	58 (est.)	0.584 (measured score)	NR	54 (est.)	Not specified	p < 0.01	No
Bødtger, Poulsen, Jacobi, et al., 2002	Tree	2 wk	Rx use	32.5 (median)	6.0–71.0 (range)	17	51.0 (median)	14.0–76.0 (range)	17	Non-parametric	p < 0.04	No
Bødtger, Poulsen, Jacobi, et al., 2002	Tree	2 wk	Rx use	52.0 (median)	2.0–114.0 (range)	17	102.0 (median)	2.0–186.0 (range)	17	Non-parametric	p < 0.02	No
Bousquet, Frank, Soussana, et al., 1987	Grass	6 wk	Sx severity	61.0	35.0	35	109	33	16	Non-parametric	p < 0.01	No
Bousquet, Hejjaoui, Skassa-Brociek, et al., 1987	Grass	4 wk	Sx severity	9.5 (median)	10.0	15	20.5 (median)	7	11	Non-parametric	p < 0.005	Graph
Bousquet, Hejjaoui, Skassa-Brociek, et al., 1987	Grass	4 wk	Rx use	0.84	2.25	15	2.67	1.54	11	Non-parametric	p < 0.01	Graph
Bousquet, Hejjaoui, Soussana, et al., 1990	Grass	6 wk	Sx severity	63.6	32.5	20	108.6	33.2	15	Non-parametric	p < 0.005	Graph
Bousquet, Hejjaoui, Soussana, et al., 1990	Grass	6 wk	Rx use	38.6	37.6	20	66.4	51.7	15	Non-parametric	p < 0.05	No
Bousquet, Hejjaoui, Soussana, et al., 1990	Grass	6 wk	Sx days	22.9	11.4	20	40.2	7.1	15	Non-parametric	p < 0.01	No
Bousquet, Maasch, Hejjaoui, et al., 1989	Grass	4 wk	Sx severity	14.8	22.9	18	63.5	54.6	14	Non-parametric	p < 0.001	No
Bousquet, Maasch, Hejjaoui, et al., 1989	Grass	4 wk	Rx use	22.9	39.1	18	53.7	54.1	14	Non-parametric	p < 0.001	No
Bousquet, Maasch, Hejjaoui, et al., 1989	Grass	4 wk	Sx days	9.0	10.7	18	26.5	8.6	14	Non-parametric	p < 0.01	Graph
Brunet, Bedard, Lavoie, et al., 1992	Ragweed	4 wk	Sx severity	4.7	0.7 (SEM)	13	7.5	1.2 (SEM)	14	Non-parametric	p < 0.05	Graph
Brunet, Bedard, Lavoie, et al., 1992	Ragweed	4 wk	Rx use	0.9	0.2 (SEM)	13	0.7	0.2 (SEM)	14	Non-parametric	p < 0.6	No
Cockcroft, Cuff, Tarlo, et al., 1977	Ragweed	Not specified	Sx severity	4.95	NR	21	5.75	NR	21	Parametric	p = NS (0.05 < p < 0.10)	No
Cockcroft, Cuff, Tarlo, et al., 1977	Ragweed	Not specified	Sx severity	2.29	NR	21	4.37	NR	21	Parametric	p < 0.05	No
Creticos, Reed, Norman, et al., 1996	Ragweed	4 mo pretrial observation; year-1 data	Sx severity	3.5 (year 1)	0.5	29	4.3 (year 1)	0.5	24	Parametric	p < 0.1	No
Grammer, Shaughnessy, Bernhard, et al., 1987	Ragweed	5 wk	Combined Sx/Rx	7.76	NR	30	17.4	NR	30	Parametric	p = 0.02	No
Grammer, Shaughnessy, Suszko, et al., 1983	Grass	9 wk	Combined Sx/Rx	210	75 (SEM)	10	500	115 (SEM)	13	Non-parametric	p = 0.02	No
Grammer, Zeiss, Suszko, et al., 1982	Ragweed	7 wk	Combined Sx/Rx	332.	64 (SEM)	21	530	83 (SEM)	19	Parametric	p = 0.022	No
Hirsch, Kalbfleisch, and Cohen, 1982	Ragweed	6 wk	Sx severity	24.8	15.1	20	45.9	18.6	14	Parametric	p < 0.004	No
Hirsch, Kalbfleisch, and Cohen, 1982	Ragweed	6 wk	Rx use	4.0	7.4	20	8.3	2.3	14	Parametric	p < 0.025	No
Iliopoulos, Proud, Adkinson, et al., 1991	Ragweed	Not specified	Combined Sx/Rx	NR	NR	21	NR	NR	20	Non-parametric	p < 0.04	No
Leynadier, Banoun, Dollois, et al., 2001	Grass	12 wk	Sx severity	49.5	NR	16	56	NR	13	Non-parametric	p = NS	No
Leynadier, Banoun, Dollois, et al., 2001	Grass	12 wk	Rx use	11.1	NR	16	40.8	NR	13	Non-parametric	p = 0.005	No
Lichtenstein, Norman, and Winken-werder, 1971	Ragweed	8 wk	Sx severity	7.25	NR	18	11.125	NR	21	Non-parametric	p < 0.01	Graph
McAllen, 1969	Grass	7 wk	Sx severity	54	NR	40	72	NR	20	Non-parametric	p = 0.074	No
McAllen, 1969	Grass	7 wk	Sx days	35	NR	40	28.5	NR	20	Non-parametric	p = 0.087	No
Norman, Lichtenstein, Kagy-Sobotka, et al., 1982	Ragweed	NR	Combined Sx/Rx	5.3	NR	16	8.8	NR	17	Non-parametric	p < 0.01	Graph
Ortolani, Pastorello, Incorvaia, et al., 1994	Tree	4 wk	Combined Sx/Rx	NR	NR	17	NR	NR	14	Non-parametric	p < 0.05	No
Parker, Whisman, Apaliski, et al., 1989	Tree	10 days	Combined Sx/Rx	57.0	NR	26	129.9	NR	25	Non-parametric	p = 0.0001	Yes
Pastorello, Pravettoni, Incorvaia, et al., 1992	Grass	4 wk	Combined Sx/Rx	NR	NR	10	NR	NR	9	Non-parametric	p < 0.01	No
Pence, Mitchell, Greely, et al., 1976	Tree	12 wk	Combined Sx/Rx	5.46	3.22	17	8.83	3.15	15	Parametric	p < 0.01	Yes
Van Metre, Adkinson, Amodio, et al., 1980	Ragweed	8 wk	Combined Sx/Rx	3.0	NR	15	5.0	NR	14	Non-parametric	p < 0.01	Graph
Van Metre, Adkinson, Amodio, et al., 1982	Ragweed	8 wk	Combined Sx/Rx	3.79	NR	15	11.14	NR	11	Non-parametric	p < 0.01	Graph
Varney, Gaga, Frew, et al., 1991	Grass	11 wk	Sx severity	360	NR	21	928	NR	16	Non-parametric	p = 0.001	No
Varney, Gaga, Frew, et al., 1991	Grass	11 wk	Rx use	129	NR	21	627	NR	16	Non-parametric	p = 0.002	No
Walker, Pajno, Limo, et al., 2001	Grass	11 wk (2 seasons: 1996 & 1998)	Grass	Difference between IT and placebo = 1186.5	241.5 to 1928.6	22	See IT mean	See IT SD	22	Non-parametric	p = 0.01	No
Walker, Pajno, Limo, et al., 2001	Grass	11 wk (2 seasons: 1996 & 1998)	Grass	Difference between IT and placebo = 1043.0	332.0 to 2667.1	22	See IT mean	See IT SD	22	Non-parametric	p = 0.007	No
Weyer, Donat, L'Heritier, et al., 1981	Grass	6 wk	Sx severity	16	10	17	24	8	16	Non-parametric	p < 0.09	No
Weyer, Donat, L'Heritier, et al., 1981	Grass	6 wk	Rx use	3	5	17	11	13	16	Non-parametric	p < 0.07	No
Weyer, Donat, L'Heritier, et al., 1981	Grass	6 wk	Combined Sx/Rx	10	7	17	18	15	16	Non-parametric	p < 0.03	No
Zenner, Baumgarten, Rasp, et al., 1997	Grass	10 wk	Sx severity	82.2	10.1	45	116	13.2	41	Non-parametric	p < 0.025	Graph
Zenner, Baumgarten, Rasp, et al., 1997	Grass	10 wk	Rx use	26% of 70 days	NR	45	33% of 70 days	NR	41	Non-parametric	p < 0.296	No

Abbreviations: IPD = individual patient data; IT = immunotherapy; mo = month(s); n = number of patients; NR = not reported; NS = not significant; Rx = medication; SD = standard deviation; SEM = standard error of the mean; Sx = symptom; wk = weeks

Meta-analysis of placebo-controlled trials. We performed a meta-analysis of the placebo-controlled trials of allergen immunotherapy conducted among patients with seasonal allergic rhinitis. Outcome data on total symptoms, medication use, or a combination of these measures was abstracted by one of the investigators (DM or JS) and confirmed from original reports by the other. We attempted to abstract data on the mean, variance, and numbers of subjects per treatment arm in order to estimate an effect size. However, many studies reported medians rather than means and used non-parametric statistical analyses; in such cases, it was not possible to estimate an effect size. Some studies used parametric statistical analysis on original or log-transformed data (Creticos, Reed, Norman, et al., 1996). When data on variance were not reported, we estimated individual patient data from published graphs and figures when reasonably accurate estimates were possible. We analyzed individual patient data using SAS (The SAS Institute, 2001) to estimate means and variance, using log-transformation if necessary to normalize the data. A description of the data abstracted for the analysis is provided in Table 15.

We calculated and combined effect sizes (Cohen, 1988) and tested for statistical heterogeneity using Comprehensive Meta-analysis statistical software (Biostat, 1999). Studies that did not report sufficient data to estimate effect sizes, including those that used only non-parametric statistical analysis, were omitted from the meta-analysis.

Planned subgroup analyses included the type of outcome measure (total symptom score versus medication use versus combined symptom-medication scores), type of allergen (tree, grass, or weed), type of placebo (inert, fixed histamine concentration, variable histamine concentration), and elements of the quality assessment for which sufficient variability was observed.

Figure 2. Meta-analysis of placebo-controlled trials (more...)

   Figure 2. Meta-analysis of placebo-controlled trials of immunotherapy for seasonal allergic rhinitis

Fifteen trials were included in the meta-analysis. The number of subjects in each trial ranged from 23 to 73. Seven trials reported data on total symptom severity, two reported data on medication use, and eight reported data on combined symptom severity and medication use. There was no overlap between the trials reporting total symptom severity and those reporting medication use (although both trials reporting medication use, also reported symptom severity). Our primary analysis of all 15 trials was stratified by outcome (symptom severity versus combined symptom severity and medication use). The effect sizes for individual studies showed no significant heterogeneity among either subgroup (p = 0.13 and 0.7, respectively) or the entire collection of studies (p = 0.76). Effect size estimates ranged from 0.43 to 1.3 for symptom severity, and from 0.61 to 1.4 for studies reporting combined symptom-medication scores (Figure 2). Summary effect sizes were 0.77 (95% CI, 0.53 to 1.02) for symptom severity and 0.97 (0.72 to 1.21) for combined symptom-medication scores, with an overall summary effect size of 0.87 (0.70 to 1.04).

Further subgroup analyses were performed based on allergen used, type of placebo, and selected quality measures. The effect size was estimated for four grass pollen, eight ragweed pollen and three tree pollen studies, with no significant difference (p = 0.25). Similarly, no significant difference was observed when studies were stratified by type of placebo (fixed histamine dose, variable histamine dose, and no histamine; p = 0.60). We analyzed for differences in effect size associated with quality assessment variables for which there was sufficient variability among trials, namely, double-blinding and description of dropouts. There was no statistically significant difference, but there was a trend (p = 0.07) toward a higher effect size among single-blinded compared to double-blinded studies (1.2 [0.8 to 1.5] versus 0.78 [0.58 to 0.98]). There was no difference between those trials that reported dropouts and those that did not (0.86 [0.64 to 1.1] versus 0.89 [0.61 to 1.2]; p = 0.85).

Perennial Allergic Rhinitis

Table 16. Placebo-controlled randomized controlled (more...)

Table 16. Placebo-controlled randomized controlled trials of injection immunotherapy (IT) for perennial allergic rhinitis, by type of allergen

Allergen	Number of trials	Number of subjects	Number of trials favoring IT	Number of trials with negative or equivocal results
Dust mite	7	357	5	2
Dust mite and pollen	1	10	0	1
Cat	1	28	1	0
Mold (Alternaria)	1	22	1	0
Latex	1	14	1	0
Multiple antigens	1	36	1	0

The number of clinical trials of IT in perennial allergic rhinitis is small. We identified 12 randomized controlled trials (540 subjects enrolled) that met our inclusion criteria. Seven trials assessed IT with dust mite allergen, and the others studied a combination of dust mite and pollen allergen, cat allergen, latex allergen, mold (Alternaria), or multiple antigens. Most studies (9 of 12) reported results favoring IT (Table 16).

There are important methodological concerns about some of the included trials. Most trials used an IT treatment program of 52 weeks. However, two trials (D'Souza, Pepys, Wells, et al., 1973; Ewan, Alexander, Snape, et al., 1988) had a short treatment program of 12 weeks. One trial used a Rinkel-type protocol and employed a 2-week treatment program of active IT or placebo, after which patients completed a 2-week washout period and crossed over to the opposite therapy (Radcliffe, Lampe, and Brostoff, 1996). It is unlikely that optimal clinical benefits of immunotherapy could be achieved within these short time frames. One trial reported a 41 percent dropout rate and did not collect adequate symptom and medication data to report results (Blainey, Phillips, Ollier, et al., 1984). Another trial did not collect daily symptom scores, had a high dropout rate (8/18; 44 percent), and did not collect data on concomitant allergy medication use (Krouse and Krouse, 2000).

After studies with significant methodological flaws were excluded, the remaining trials included four studies of dust mite immunotherapy in 241 patients, and three small trials (1 each) of immunotherapy using cat, mold, or latex allergen. The small number of trials and the limited number of patients enrolled in these studies underscore the need for additional clinical trials to assess the effectiveness of IT for the treatment of perennial allergic rhinitis.

Adverse event data were described for nine of 12 studies of IT in perennial allergic rhinitis. As observed in IT for seasonal allergic rhinitis, local injection site reactions were common. Systemic allergic reactions were reported in various studies to occur in from zero to 100 percent of subjects. Most of these reactions were mild. There were no reports of treatment-related hospitalizations or deaths.

Quality Assessment

Most of the immunotherapy trials abstracted in this analysis (48 of 60) enrolled patient populations that were similar to the adult US working population. None of the trials described the racial characteristics of the subjects enrolled. Sex- and age-related differences in clinical responses to IT were not reported in any of the trials. Virtually all of the studies used a single allergen or class of allergen in the treatment group. However, the external validity of this approach is questionable, given that most atopic patients are polysensitized. In contrast, most patients receiving IT in non-research settings have vaccines formulated with most or all of the allergens to which they are sensitive.

A primary clinical outcome measure used in most of the studies was a symptom or symptom-medication score compiled from a patient diary. Usually subjects were asked to score a symptom, such as sneezing, on a scale of 0 to 3. Unfortunately, this outcome measure had not been standardized. The degree to which this scale is responsive to change, and whether ceiling or floor effects occur when it is used, have not been determined. Finally, the degree of change in symptom score necessary to be clinically relevant is not known.

Other quality concerns identified in this review include the virtual absence of meaningful sample size determinations; inadequate description of procedures for generating randomization sequences and concealing them from investigators; incomplete patient follow-up; and failure to perform efficacy analyses according to the intention-to-treat principle.

Conclusions

We analyzed 60 controlled trials of immunotherapy in allergic rhinitis. No serious adverse events were reported, and immunotherapy was generally well tolerated. Our data show that immunotherapy for seasonal allergic rhinitis consistently demonstrates evidence of clinical benefit (effect size, 0.87 [95% CI, 0.70 to 1.04]). The magnitude of this effect equates to a 35 to 40 percent reduction in symptom or symptom-medication scores when individual trials with similar effect sizes are analyzed (Lichtenstein, Norman, and Winkenwerder, 1971; Van Metre, Adkinson, Amodio, et al., 1980). This effect is similar to or slightly better than that observed in clinical trials of antihistamines for seasonal allergic rhinitis (European Agency for Evaluation of Medicinal Products, 2001).

Important flaws in study quality were identified, which may affect the internal validity of the results of this analysis. Most trials enrolled a small number of patients and employed clinical outcome measures that have not been validated. Other concerns include inadequate or poorly described methods for allocation concealment and failure to employ an intention-to-treat analysis. Nevertheless, we could not identify significant differences in effect sizes among trials stratified by the presence or absence of these quality criteria. Since most trials were small, we could not accurately assess the presence of publication bias. Although only 15 of 42 placebo-controlled trials provided appropriate data to estimate effect size, the proportion of trials with statistically significant positive findings was similar for all studies (Table 14) and the subset included in the meta-analysis (Figure 2). Findings among the few studies of perennial rhinitis were consistent with a clinically important effectiveness, although the limited number of studies and important methodological problems preclude a firm conclusion or a quantitative estimate of the magnitude of any effect.

Our analysis also highlights several research needs related to immunotherapy and the treatment of allergic rhinitis. Standardized instruments for assessing clinical symptoms need to be developed. Using these tools, it should be possible to define response criteria that will allow investigators to classify patients as responders or non-responders. Large-scale clinical trials employing vaccines with most or all relevant allergens for each individual should be designed to assess IT as it is administered in most community settings. Additional future research objectives should be focused upon the following: methods to identify patients likely to benefit from IT; cost-effectiveness and quality-of-life analyses of IT; determination of whether IT alters the natural history of allergic rhinitis and reduces possible sequelae such as bacterial sinusitis and asthma; and studies clarifying the optimal duration of IT.

Combined Treatments

Results

Studies Identified

Thirty-one publications describing 32 separate randomized controlled trials met the inclusion and exclusion criteria for this topic (see Evidence Table 4); there were 49 relevant comparisons (some trials had multiple treatment arms). We did not identify any systematic reviews addressing this question.

Table 17. Randomized controlled trials comparing combination (more...)

Table 17. Randomized controlled trials comparing combination pharmacotherapy to monotherapy for allergic rhinitis

Treatment 1	Treatment 2	No. of comparisons	Results
Antihistamine + oral decongestant	Antihistamine	13	7 combination superior, 3 no significant difference, 3 no difference, no statistical test reported
Antihistamine + oral decongestant	Decongestant	10	8 combination superior, 2 possibly superior
Antihistamine + oral decongestant	Nasal glucocorticoid	1	No significant difference
Antihistamine + nasal glucocorticoid	Nasal glucocorticoid	7	3 combination superior, 4 no significant difference
Antihistamine + nasal glucocorticoid	Antihistamine	7	5 combination superior, 2 possibly superior
Antihistamine + mast cell stabilizer	Antihistamine	1	Combination superior
Antihistamine + NSAID	Antihistamine	2	Combination superior (1 study)
Antihistamine + ophthalmic antihistamine	Antihistamine	1	Combination reduced eye itching
Antihistamine + ipratropium	Antihistamine	1	Combination reduced rhinorrhea
Ipratropium + nasal glucocorticoid	Nasal glucocorticoid	1	Combination reduced rhinorrhea
Ipratropium + nasal glucocorticoid	Ipratropium	1	Combination reduced rhinorrhea
Nasal glucocorticoid + 3 days nasal decongestant	Nasal glucocorticoid	1	No significant difference
Nasal glucocorticoid + 3 days nasal decongestant	Antihistamine	1	Combination superior
Nasal antihistamine + nasal decongestant	Nasal antihistamine	1	No significant difference
Nasal antihistamine + nasal decongestant	Nasal decongestant	1	Combination superior

Most studies evaluated patients with seasonal allergic rhinitis (n = 26, 81 percent), recruited from specialty settings, and included primarily adults, with some adolescents (> 12 years of age). Study durations were ≤ 2 weeks (n = 18, 56 percent), 15 days to 6 weeks (n = 12, 38 percent), and > 6 weeks (n = 2, six percent). The majority of studies were small to moderate in size; sample sizes were < 100 (n = 13, 41 percent), 100 to 200 (n = 5, 15 percent), and > 200 (n = 14, 44 percent). A majority of studies were performed outside the US (53 percent). The most common outcomes reported were individual symptoms, symptom scales, and global symptom ratings. Health-related quality of life using the Medical Outcome Study Short-Form Health Survey (SF-36) or the Rhinoconjunctivitis Quality of Life Questionnaire (RQLQ) was reported in three studies. One study reported economic effects on work performance. Trials involving terfenadine (Seldane^®) and astemizole (Hismanal^®) were included even though these medications have been withdrawn from the market due to safety concerns. A description of treatment comparisons and an overview of results are provided in Table 17.

Table 18. Data abstracted for meta-analysis of combination (more...)

Table 18. Data abstracted for meta-analysis of combination treatment articles

Study	Combination	Monotherapy	Outcome	Combomean	Combo SD	Combo n	Monomean	Mono SD	Mono n	Statistical test	P-value	Possible to calculate Es?
A. Antihistamine + decongestant combinations versus antihistamine alone, total symptom severity (see also Figure 3)
Bronsky, Boggs, Findlay, et al., 1995	Loratadine+pseudoephedrine	Loratadine	TSS	6.72	NR	212	5.6	NR	212	ANOVA	P < 0.05	Yes
Dockhorn, Williams, and Sanders, 1996	Acrivastine+pseudoephedrine	Acrivastine	TSS	10.3	NR	176	12.3	NR	175	ANCOVA (1-sided)	P < 0.001	Yes
Falliers and Redding, 1980 (study 1)	Azatadine+pseudoephedrine	Azatadine	TSS	70% reduction	NR	30	52% reduction	NR	30	ANOVA	NR	No
Falliers and Redding, 1980 (study 2)	Azatadine+pseudoephedrine	Azatadine	TSS	82% reduction	NR	10	58% reduction	NR	10	ANOVA	NR	No
Grosclaude, Mees, Pinelli, et al., 1997	Cetirizine+pseudoephedrine	Cetirizine	TSS	0.85	NR	230	1.03	NR	226	ANOVA	P < 0.001	Yes
Henauer, Seppey, Hugenot, et al., 1991	Terfenadine+pseudoephedrine	Terfenadine	TSS	NR	NR	25	NR	NR	25	ANOVA	P = 0.69	Yes
Meran, Morse, and Gibbs, 1990	Acrivastine+pseudoephedrine	Acrivastine	TSS	1.66	2.25	40	2.04	2.25	40	ANOVA (log-transformed scores)	P = 0.45	Yes
Sussman, Mason, Compton, et al., 1999	Fexofenadine+pseudoephedrine	Fexofenadine	TSS	2.32	NR	215	2.05	NR	218	ANCOVA	P ~ 0.16	Yes
Williams, Hull, McSorley, et al., 1996	Acrivastine+pseudoephedrine	Acrivastine	TSS	8.5	NR	202	9.8	NR	202	ANCOVA	P < 0.001 (1-sided)	Yes
B. Antihistamine + decongestant versus antihistamine alone, nasal symptom severity (see also Figure 4)
Bertrand, Jamart, Marchal, et al., 1996	Cetirizine+pseudoephedrine	Cetirizine	Nasal obstruction	Graph	NR	70	Graph	NR	70	CMH (categorical)	P = 0.005	No
Bronsky, Boggs, Findlay, et al., 1995	Loratadine+pseudoephedrine	Loratadine	NSS	NR	NR	212	NR	NR	212	ANOVA	P < 0.01	Yes
Dockhorn, Williams, and Sanders, 1996	Acrivastine+pseudoephedrine	Acrivastine	NSS	3.8	NR	176	4.7	NR	175	ANCOVA (1-sided)	P < 0.001	Yes
Falliers and Redding, 1980 (study 1)	Azatadine+pseudoephedrine	Azatadine	NSS	68% reduction	NR	30	35% reduction	NR	30	ANOVA	P < 0.05	Yes
Falliers and Redding, 1980 (study 2)	Azatadine+pseudoephedrine	Azatadine	NSS	73% reduction	NR	10	27% reduction	NR	10	ANOVA	P < 0.05	Yes
Grosclaude, Mees, Pinelli, et al., 1997	Cetirizine+pseudoephedrine	Cetirizine	NSS	1.19	NR	230	1.43	NR	226	ANOVA	P < 0.001	Yes
Meran, Morse, and Gibbs, 1990	Acrivastine+pseudoephedrine	Acrivastine	NSS	1.89	NR	40	2.41	NR	40	ANOVA (log-transformed scores)	P < 0.01	Yes
Sussman, Mason, Compton, et al., 1999	Fexofenadine+pseudoephedrine	Fexofenadine	NSS	0.56	NR	215	0.36	NR	218	ANCOVA	P < 0.0005	Yes
Williams, Hull, McSorley, et al., 1996	Acrivastine+pseudoephedrine	Acrivastine	NSS	2.3	NR	202	2.7	NR	202	ANCOVA	P < 0.001 (1-sided)	Yes
C. Antihistamine + decongestant combination versus decongestant alone, total symptom severity (see also Figure 5)
Bronsky, Boggs, Findlay, et al., 1995	Loratadine+pseudoephedrine	Pseudoephedrine	TSS	6.72	NR	212	5.32	NR	212	ANOVA	P < 0.05	Yes
Dockhorn, Williams, and Sanders, 1996	Acrivastine+pseudoephedrine	Pseudoephedrine	TSS	10.3	NR	176	11.8	NR	177	ANCOVA (1-sided)	P = 0.002	Yes
Falliers and Redding, 1980 (study 1)	Azatadine+pseudoephedrine	Pseudoephedrine	TSS	70% reduction	NR	30	43% reduction	NR	30	ANOVA	NR	No
Falliers and Redding, 1980 (study 2)	Azatadine+pseudoephedrine	Pseudoephedrine	TSS	82% reduction	NR	10	55% reduction	NR	10	ANOVA	NR	No
Grosclaude, Mees, Pinelli, et al., 1997	Cetirizine+pseudoephedrine	Pseudoephedrine	TSS	0.85	NR	230	1.14	NR	231	ANOVA	P < 0.001	Yes
Meran, Morse, and Gibbs, 1990	Acrivastine+pseudoephedrine	Pseudoephedrine	TSS	1.66	2.25	40	2.92	2.25	40	ANOVA (log-transformed scores)	P = 0.014	Yes
Sussman, Mason, Compton, et al., 1999	Fexofenadine+pseudoephedrine	Pseudoephedrine	TSS	2.32	NR	215	1.42	NR	218	ANCOVA	P < 0.0001	Yes
Williams, Hull, McSorley, et al., 1996	Acrivastine+pseudoephedrine	Pseudoephedrine	TSS	8.5	NR	202	10.8	NR	202	ANCOVA	P < 0.001 (1-sided)	Yes
D. Antihistamine + decongestant combination versus decongestant alone, nasal symptom severity (see also Figure 6)
Bertrand, Jamart, Marchal, et al., 1996	Cetirizine+pseudoephedrine	Pseudoephedrine	Nasal obstruction	Graph	NR	70	Graph	NR	70	CMH (categorical)	P = 0.025	No
Bronsky, Boggs, Findlay, et al., 1995	Loratadine+pseudoephedrine	Pseudoephedrine	NSS	NR	NR	212	NR	NR	212	ANOVA	P = NS	No
Dockhorn, Williams, and Sanders, 1996	Acrivastine+pseudoephedrine	Pseudoephedrine	NSS	3.8	NR	176	4.1	NR	177	ANCOVA (1-sided)	P ~ 0.29	Yes
Falliers and Redding, 1980 (study 1)	Azatadine+pseudoephedrine	Pseudoephedrine	NSS	68% reduction	NR	30	62% reduction	NR	30	ANOVA	P ~ 0.72	Yes
Falliers and Redding, 1980 (study 2)	Azatadine+pseudoephedrine	Pseudoephedrine	NSS	73% reduction	NR	10	63% reduction	NR	10	ANOVA	P ~ 0.65	Yes
Grosclaude, Mees, Pinelli, et al., 1997	Cetirizine+pseudoephedrine	Pseudoephedrine	NSS	1.19	NR	230	1.22	NR	231	ANOVA	P ~ 0.68	Yes
Meran, Morse, and Gibbs, 1990	Acrivastine+pseudoephedrine	Pseudoephedrine	NSS	1.89	NR	40	2.88	NR	40	ANOVA (log-transformed scores)	P < 0.01	Yes
Sussman, Mason, Compton, et al., 1999	Fexofenadine+pseudoephedrine	Pseudoephedrine	NSS	0.56	NR	215	0.45	NR	218	ANCOVA	P ~ 0.059	Yes
Williams, Hull, McSorley, et al., 1996	Acrivastine+pseudoephedrine	Pseudoephedrine	NSS	2.3	NR	202	2.6	NR	202	ANCOVA	P ~ 0.01 (1-sided)	Yes
E. Antihistamine + nasal glucocorticoid versus antihistamine alone, nasal symptom severity (see also Figure 7)
Backhouse, Finnamore, and Gosden, 1986	Terfenadine+flunisolide nasal spray	Terfenadine	Nasal congestion	1.4	0.7	49	1.8	0.9	50	t-test	P ~ 0.03	Yes
Brooks, Francom, Peel, et al., 1996	Loratadine+beclomethasone nasal spray	Loratadine	Nasal congestion	NR	NR	20	NR	NR	20	ANOVA	P < 0.001	Yes
Juniper, Kline, Hargreave, et al., 1989	Astemizole+beclomethasone nasal spray	Astemizole	Nasal congestion	0.322	NR	30	0.594	NR	30	ANOVA	P < 0.05	Yes
Ratner, van Bavel, Martin, et al., 1998	Loratadine+fluticasone nasal spray	Loratadine	NSS	160	NR	150	232	NR	150	ANOVA	P < 0.01	Yes
Simpson, 1994	Terfenadine+budesonide nasal spray	Terfenadine	Blocked nose	7	NR	32	14	NR	23	ANOVA	P < 0.05	Yes
Wilson, Dempsey, Sims, et al., 2000	Cetirizine+mometasone nasal spray	Cetirizine	NSS	1.8	0.6 (SEM)	14	3.5	0.7 (SEM)	13	MANOVA with pairwise comparison	P ~ 0.07	Yes
F. Antihistamine + nasal glucocorticoid versus nasal glucocorticoid alone, nasal symptom severity (see also Figure 8)
Benincasa and Lloyd, 1994	Cetirizine+fluticasone nasal spray	Fluticasone nasal spray	NSS	1.5	1.6	227	1.5	1.4	227	t-test	P = 1.0	Yes
Brooks, Francom, Peel, et al., 1996	Loratadine+beclomethasone nasal spray	Beclomethasone nasal spray	Nasal congestion	NR	NR	20	NR	NR	20	ANOVA	P = 0.66	Yes
Drouin, Yang, Horak, et al., 1995	Loratadine+beclomethasone nasal spray	Beclomethasone nasal spray	NSS	66% improved	NR	76	59% improved	NR	78	ANOVA	P = NS	No
Juniper, Kline, Hargreave, et al., 1989	Astemizole+beclomethasone nasal spray	Beclomethasone nasal spray	Nasal congestion	0.322	NR	30	0.319	NR	30	ANOVA	P ~ 0.98	Yes
Purello-D'Ambrosio, Isola, Ricciardi, et al., 1999	Loratadine+flunisolide nasal spray	Flunisolide nasal spray	Nasal blockage	19.9%	NR	15	20%	NR	15	ANOVA	P ~ 1.0	Yes
Ratner, van Bavel, Martin, et al., 1998	Loratadine+fluticasone nasal spray	Fluticasone nasal spray	NSS	160	NR	150	192	NR	150	ANOVA	P < 0.05	Yes
Simpson, 1994	Terfenadine+budesonide nasal spray	Budesonide nasal spray	Blocked nose	7	NR	32	5.5	NR	30	ANOVA	P ~ 0.58	Yes

Abbreviations: ANCOVA = analysis of covariance; ANOVA = analysis of variance; CMH = Cochran-Mantel-Haenszel; ES = effect size; MANOVA = multivariate analysis of variance; n = number of patients; NR = not reported; NS = not significant; NSS = nasal symptom severity; SD = standard deviation; TSS = total symptom score

For the comparisons for which there were more than two trials, we attempted a quantitative meta-analysis. We extracted outcome data at 2 weeks for (a) total symptom relief scores and (b) nasal symptom scores and/or nasal congestion scores (Table 18). The 2-week time point was chosen to maximize comparability between trials despite differences in duration of treatment and followup. We used data on continuous measures to calculate effect sizes or standardized mean differences (Cohen, 1988) based on reported means and standard deviations or p-values from parametric statistical analyses using Comprehensive Meta-analysis statistical software (Biostat, 1999). Studies that did not report sufficient data to estimate effect size, including those that used only non-parametric statistical analysis, were omitted from the analysis. Where similar trials provided data, we tested the individual study effect size estimates for homogeneity, and, if homogeneous, used a fixed-effects model meta-analysis to combine the estimates. We planned a priori to compare the effect among subgroups of studies using sedating versus non-sedating antihistamines.

Table 19. Summary of meta-analysis of randomized controlled (more...)

Table 19. Summary of meta-analysis of randomized controlled trials comparing combination pharmacotherapy to monotherapy for allergic rhinitis

Combination	Comparator drug	Number of studies	Total number of patients	Outcome evaluated	Summary effect size (95% confidence interval
Antihistamine-decongestant	Antihistamine	7	2298	Total symptom score	0.23 (0.15 to 0.32)
Antihistamine-decongestant	Decongestant	6	2154	Total symptom score	0.31 (0.22 to 0.39)
Antihistamine-decongestant	Antihistamine	8	2233	Nasal symptom score	0.33 (0.24 to 0.41)
Antihistamine-decongestant	Decongestant	7	1806	Nasal symptom score	0.16 (0.07 to 0.25)
Antihistamine-nasal glucocorticoid	Antihistamine	6	559	Nasal sympttom score	0.44 (0.27 to 0.61)
Antihistamine-nasal glucocorticoid	Nasal glucocorticoid	6	946	Nasal symptom score	0.9 (-0.4 to 0.22)

A summary of the results of the meta-analysis is provided in Table 19.

Antihistamines with or without a Decongestant

Thirteen studies, conducted in North America (n = 7), Europe (n = 5), and India (n = 1) compared antihistamines to the combination of an antihistamine with pseudoephedrine. The antihistamines assessed included acrivastine (n = 4), cetirizine (n = 2), azatadine (n = 2), terfenadine (n = 2), and one trial each for loratadine, triprolidine, and fexofenadine. Overall, seven studies showed that the antihistamine-decongestant combination was superior to antihistamine alone for reducing symptoms (Bertrand, Jamart, Marchal, et al., 1996; Dockhorn, Williams, and Sanders, 1996; Falliers and Redding, 1980 [two studies]; Grosclaude, Mees, Pinelli, et al., 1997; Panda and Mann, 1998; Williams, Hull, McSorley, et al., 1996). Three trials found no statistically significant difference (Henauer, Seppey, Huguenot, et al., 1991; Meran, Morse, and Gibbs, 1990; Sussman, Mason, Compton, et al., 1999). Finally, three other studies showed essentially similar symptom scores (Bronsky, Boggs, Findlay, et al., 1995; Diamond, Gerson, Cato, et al., 1981; Vuurman, van Veggel, Sanders, et al., 1996); no formal statistical tests were reported, so these were interpreted as negative. Interestingly, the studies comparing the combination of a sedating antihistamine and decongestant were more often positive compared to antihistamine alone than similarly designed studies using a non-sedating antihistamine.

To quantitatively examine the variability in findings and to calculate a summary estimate of the effect size, we performed a meta-analysis of these studies for two outcomes, total symptom relief and nasal symptom relief. Eleven of the 13 studies reported a total symptom score. Six studies were excluded from the analysis, two because of study duration less than 2 weeks (Diamond, Gerson, Cato, et al., 1981; Vuurman, van Veggel, Sanders, et al., 1996), and four because an effect size could not be calculated (Bertrand, Jamart, Marchal, et al., 1996; Falliers and Redding, 1980 [two studies]; Panda and Mann, 1998). Effect size estimates for total symptom relief from treatment with combination antihistamine-pseudoephedrine versus antihistamine alone are shown in Figure 3. A test of homogeneity was insignificant (p = 0.84). The summary effect size was 0.23 (95 percent confidence interval [95% CI], 0.15 to 0.32), showing that total symptom scores were better, that is, there was a greater reduction in symptoms, in the patients receiving combination therapy.

Studies of non-sedating antihistamines (Bronsky, Boggs, Findlay, et al., 1995; Henauer, Seppey, Huguenot, et al., 1991; Sussman, Mason, Compton, et al., 1999) had a combined effect size of 0.16 (95 % CI, 0.03 to 0.29), while studies employing a sedating antihistamine (Dockhorn, Aaronson, Bronsky, et al., 1999; Grosclaude, Mees, Pinelli, et al., 1997; Meran, Morse, and Gibbs, 1990; Williams, Hull, McSorley, et al., 1996) had a summary effect size of 0.29 (95% CI, 0.18 to 0.39). The difference between the two was not statistically significant (p = 0.15).

A meta-analysis of the same studies using a nasal symptom score or nasal congestion score (if the total nasal symptom score was not reported) was also performed. Estimates of the effect of combination antihistamine-pseudoephedrine compared to antihistamine alone, based on the nasal symptom/nasal congestion score, are shown in Figure 4. A test of homogeneity was insignificant (p = 0.71). The summary effect size was 0.33 (95% CI, 0.24 to 0.41), showing that relief of nasal congestion was greater in patients receiving combination therapy. There was no significant difference in effect sizes between studies using a sedating (n = 2) versus a non-sedating antihistamine (n = 6; p = 0.55).

A third treatment arm, comparing an antihistamine-decongestant combination with pseudoephedrine alone, was evaluated in 10 of the 13 studies described above. The majority of these studies (eight of 10) showed that the antihistamine-decongestant combination was superior to decongestant alone for the treatment of rhinitis symptoms (Bertrand, Jamart, Marchal, et al., 1996; Dockhorn, Williams, and Sanders, 1996; Falliers and Redding, 1980 [two studies]; Grosclaude, Mees, Pinelli, et al., 1997; Meran, Morse, and Gibbs, 1990; Sussman, Mason, Compton, et al., 1999; Williams, Hull, McSorley, et al., 1996); two of these trials showed there was no statistical difference only in one symptom, namely, nasal congestion. Diamond and colleagues (1981) and Bronsky and colleagues (1995) failed to report any statistical comparison for the symptom scores, but the mean scores for the combination treatment were better than those for the decongestant. The treatment of allergic rhinitis with pseudoephedrine alone failed to alleviate symptoms such as sneezing, itching, and rhinorrhea, but was beneficial in reducing nasal congestion.

All 10 studies comparing pseudoephedrine alone to antihistamine-pseudoephedrine combination reported a total symptom score. Four studies were excluded from the meta-anlaysis, one because the study duration was less than 2 weeks (Diamond, Gerson, Cato, et al., 1981), and three because an effect size could not be calculated (Bertrand, Jamart, Marchal, et al., 1996; Falliers and Redding, 1980 [two studies]). Estimates of the effect of the combination of antihistamine and pseudoephedrine to decongestant alone are shown in Figure 5. A test of homogeneity was insignificant (p = 0.67). The summary effect size was 0.31 (95% CI, 0.22 to 0.39), showing that total symptom scores were better, that is, there was a greater reduction in symptoms, in the patients receiving combination therapy. There was no significant difference in effect sizes between studies using a sedating (n = 2) versus a non-sedating antihistamine (n = 4; p = 0.66).

A meta-analysis using a nasal symptom score/nasal congestion score was performed as well. Estimates of the effect of the combination of an antihistamine and pseudophedrine to decongestant alone, based on the nasal scores, are shown in Figure 6. A test of homogeneity was insignificant (p = 0.39). The summary effect size, 0.16 (95% CI, 0.07 to 0.25), shows that relief of nasal congestion was greater for patients receiving combination therapy. There was no significant difference in effect sizes between the single study using a sedating antihistamine compared to the six studies using a non-sedating antihistamine (p = 0.77).

Thus, the combination of an antihistamine and a decongestant (pseudoephedrine) provides greater relief of total and nasal symptoms than either an antihistamine alone or pseudoephedrine alone. Furthermore, studies using a sedating versus non-sedating antihistamine found similar results when combined with a decongestant.

Antihistamine With or Without Nasal Glucocorticoid

Ten studies conducted in Europe (n = 6) and North America (n = 4) compared the combination of an antihistamine with a nasal glucocorticoid with either antihistamine alone (n = 7 trials) or nasal glucocorticoid alone (n = 7 trials). The combinations studied included terfenadine-flunisolide, terfenadine-budesonide, astemizole-beclomethasone, loratadine-beclomethasone, loratadine-fluticasone, loratadine-flunisolide, cetirizine-mometasone, and cetirizine-fluticasone.

Of the seven studies comparing the combination of antihistamine-nasal glucocorticoid to antihistamine alone, five showed statistically significant differences favoring the combination (Backhouse, Finnamore, and Gosden, 1986; Brooks, Francom, Peel, et al., 1996; Juniper, Kline, Hargreave, et al., 1989; Ratner, van Bavel, Martin, et al., 1998; Simpson, 1994). Two studies did not formally test the significance of the mean symptom scores between the two treatment groups, but the mean symptom scores were better with the antihistamine-nasal glucocorticoid combination than with antihistamine alone (Berger, Fineman, Lieberman, et al., 1999; Wilson, Dempsey, Sims, et al., 2000); we interpreted these two studies as possibly showing superiority of the combination.

Only two of the seven studies reported a total symptom score; therefore we used either the total nasal symptom score or nasal congestion score in a meta-analysis to assess treatment efficacy. One study was excluded because it compared a nasal antihistamine to the combination of an oral antihistamine (a different antihistamine) with a nasal steroid (Berger, Fineman, Lieberman, et al., 1999). Estimates of the effect of combination antihistamine-nasal steroid to antihistamine alone are shown in Figure 7. A test of homogeneity was insignificant (p = 0.22). The summary effect size was 0.44 (95% CI, 0.27 to 0.61), showing that nasal symptom scores were better, that is, there was a larger reduction in symptoms, in the patients receiving combination therapy. No subgroup analysis of non-sedating and sedating antihistamines was performed since only one study used a sedating antihistamine.

Of the seven studies that compared antihistamine-nasal glucocorticoid to nasal glucocorticoid, three found the combination superior for reducing allergic rhinitis symptoms (Drouin, Yang, Horak, et al., 1995; Purello-D'Ambrosio, Isola, Ricciardi, et al., 1999; Ratner, van Bavel, Martin, et al., 1998). The three combinations studies were loratadine-fluticasone, loratadine-flunisolide, and loratadine-beclomethasone. Four studies found no significant difference between the two treatments (Benincasa and Lloyd, 1994; Brooks, Francom, Peel, et al., 1996; Juniper, Kline, Hargreave, et al., 1989; Simpson, 1994).

For the meta-analysis, one study was excluded because data were not available to calculate an effect size for the total nasal score or nasal congestion score (Drouin, Yang, Horak, et al., 1995). Estimates comparing the effect of a combination nasal steroid-antihistamine to nasal steroid alone are shown in Figure 8. A test of homogeneity was insignificant (p = 0.76). The summary effect size was 0.09 (95% CI, -0.04 to 0.22), showing that nasal symptom scores were not significantly different between combination therapy and nasal steroid monotherapy. No subgroup analysis of non-sedating and sedating antihistamines was performed since only one study used a sedating antihistamine.

Thus, the addition of a nasal glucocorticoid to antihistamine relieves allergic rhinitis symptoms better than antihistamine alone; however, the combination of antihistamine-nasal glucocorticoid has not been shown to be better than nasal glucocorticoid alone, and confidence intervals suggest that the effect cannot be large.

Antihistamine-Decongestant versus Nasal Glucocorticoid

Only one study assessed the combination of antihistamine-decongestant (astemizole-D) compared to intranasal steroid (beclomethasone) for the treatment of seasonal allergic rhinitis over a 4-week period (Negrini, Troise, Voltolini, et al., 1995). There was no difference in the mean area under the curve for symptom severity in nasal congestion, sneezing, rhinorrhea, nasal itching, or total symptom scores. There was less use of ophthalmic rescue medication in the astemizole-D group compared to beclomethasone.

Antihistamines Combined With Other Therapies

Antihistamines in combination with a non-steroidal anti-inflammatory, ophthalmic antihistamine, ipratropium bromide, or mast cell stabilizer, have been compared to antihistamine alone for the treatment of allergic rhinitis.

A comparison of a nasal antihistamine (levocabastine) with or without a nasal decongestant (oxymetazoline) for 1 week in 977 seasonal allergy patients from the US and Canada found no statistically significant difference between the combination and the nasal antihistamine alone, but found the combination superior to the nasal decongestant alone for the relief of symptoms (Busse, Janssens, and Eisen, 1996). Most frequent side effects were headache or application site reactions (no significant difference, but higher in oxymetazoline and combination groups). The global assessment of efficacy was higher in the levocabastine and levocabastine-oxymetazoline groups.

A study comparing terfenadine plus ipratropium bromide nasal spray with terfenadine alone for 2 weeks in 305 patients with perennial allergic and non-allergic rhinitis showed reduction in rhinorrhea severity and duration with the combined therapy, but no statistical difference in congestion or sneezing. Compared to terfenadine alone, the patient global assessment favored combined therapy (69 vs. 53 percent, p = 0.0008) (Finn, Aaronson, Korenblat, et al., 1998).

A comparison of terfenadine with or without nimesulide (a non-steroidal anti-inflammatory) showed a reduction in symptom severity scores (p = 0.005; 30-day treatment, seasonal allergic rhinitis) (Andri, Senna, Betteli, et al., 1992). A 7-day study evaluating terfenadine with or without flurbiprofen for seasonal allergic rhinitis showed differences in mean daily symptom scores for congestion and sneezing on day 3, and for running/blowing nose on day 4. The differences pre- and post-treatment were not compared; the treatment period may have been too short to adequately compare the treatments (Brooks and Karl, 1988).

A study evaluating astemizole with or without nedocromil sodium (1%) nasal spray and placebo control (mast cell stabilizer) showed lower mean symptom summary scores at the end of 4 weeks of treatment for ragweed seasonal allergies (combination > astemizole alone > placebo) (Bukstein, Biondi, Blumenthal, et al., 1996). Likewise, a comparison of loratadine with or without olopatadine ophthalmic solution for seasonal allergic conjunctivitis showed significantly lower itching with combination therapy after 1 week of treatment. RQLQ scores were significantly lower on combination therapy (Lanier, Gross, Marks, et al., 2001).

Nasal Glucocorticoids Combined With Other Therapies

Nasal glucocorticoids in combination with ipratropium bromide or a nasal decongestant have been studied in two trials. A comparison of a nasal steroid (budesonide) plus nasal decongestant (oxymetazoline for the 1^st 3 days) versus nasal steroid alone or antihistamine alone showed that the two nasal steroid groups (combination and alone) were better than antihistamine alone for improving all nasal symptoms (p < 0.05; 3-week treatment, perennial rhinitis) (Lau, Wei, Van Hasselt, et al., 1990). The addition of oxymetazoline led to faster relief compared to budesonide alone, 1 day versus 7 days (P < 0.05). Interestingly, the patient global assessment of efficacy was not significantly different among the three groups.

One study compared ipratropium plus beclomethasone dipropionate nasal spray with ipratropium alone, beclomethasone alone, and placebo (2-week treatment, seasonal allergic rhinitis and non-allergic rhinitis) (Dockhorn, Aaronson, Bronsky, et al., 1999). All three active treatment groups were significantly better than placebo in reducing rhinorrhea severity and duration. Patients treated with the combination of ipratropium plus beclomethasone had greater percentage in the reduction of rhinorrhea severity and duration than ipratropium alone, which was better than beclomethasone alone. Patient global assessment of efficacy (good or excellent control of rhinorrhea) was combination > ipratropium > beclomethasone > placebo. RQLQ scores improved from baseline for all four groups (combined > ipratropium or placebo, p < 0.05). Rates of minor adverse events (headache, nasal dryness, epistaxis) were similar among all groups.

Clinician Specialty Differences

Racial and Ethnic Variation

Costs and Work Performance

• Allergic rhinitis is associated with enormous direct and indirect costs in the US, with estimates as high as $4.5 billion and $7.7 billion annually, respectively; an updated comprehensive burden-of-illness study is necessary to more precisely estimate direct and indirect costs, for which currently available estimates vary four- to six-fold.

• There are few well-conducted, generalizable studies of direct and indirect costs for currently available clinical treatments.

• Economic evaluations of allergic rhinitis treatments often do not adequately consider uncertainty about estimates of the efficacy of treatments, often inappropriately using cost-minimization analyses rather than cost-effectiveness analyses.

• There is a lack of consensus on an appropriate and clinically meaningful measure of “effectiveness” to be used in the denominator of a cost-effectiveness ratio.

• The few available standardized instruments that assess allergic rhinitis symptoms are not yet widely used.

• Additional studies are needed to better understand how the severity of allergic rhinitis symptoms and the various medications used to treat those symptoms affect productivity.

• In order to better estimate indirect costs of allergic rhinitis treatments, objective measures of work performance are needed to determine the relationship between symptom outcomes and work performance.

Environmental Measures

Based on the pathophysiology of allergic rhinitis, treatments that decrease allergen exposure sufficiently through environmental control measures can be expected to control symptoms. Our systematic review showed that:

• Allergen avoidance measures have been studied more often in children than in adults with allergic rhinitis.

• Studies of air filtration systems do not show strong evidence for decreasing rhinitis symptoms; however, studies were likely underpowered to detect clinically relevant differences.

• A few trials in highly selected patients suggest that dust-mite control measures such as an acaricide, impervious covers, and extra house cleaning may decrease rhinitis symptoms.

• Studies of mite-sensitive asthmatics do not demonstrate any overall clinical benefit of a variety of measures designed to reduce mite exposure. Although the small number of studies evaluating this question did not yield a definitive answer, the data for house dust mite controls are encouraging.

Immunotherapy

• Nearly all of 60 clinical trials of immunotherapy in allergic rhinitis reporting symptom outcomes favored injection immunotherapy over placebo.

• No serious adverse events were reported, and immunotherapy was generally well tolerated.

• A quantitative meta-analysis showed a consistent effect for immunotherapy for seasonal allergic rhinitis, but the conclusion about the effectiveness of immunotherapy for perennial allergic rhinitis was less certain.

• Primary quality concerns in this literature were related to small trial size, lack of standardized clinical outcome assessments, and trial design issues related to blinding.

Combined Treatments

• Combination symptomatic pharmacotherapy with antihistamines plus decongestants has been well studied and overall shows improved total and nasal symptom relief compared to monotherapy with either antihistamines or decongestants alone.

• Combination treatment with antihistamines plus nasal glucocorticoids improves nasal symptoms more than antihistamine alone, but not significantly more than monotherapy with nasal glucocorticoids.

• Other combinations have been studied in a small number of trials and overall show that compared with antihistamines alone: (a) the addition of ipratropium is beneficial for rhinorrhea symptoms; (b) ophthalmic antihistamine reduces eye itching; and (c) the mast cell stabilizer nedocromil sodium and non-steroidal anti-inflammatory drugs improve overall rhinitis symptoms.

Clinician Specialty Differences

Although differences in care and outcomes have been demonstrated between generalist and specialist care in other conditions, including asthma, few data are available in allergic rhinitis.

• Clinician-delivered patient education interventions coupled with medical treatment may improve allergic rhinitis symptoms more than medical treatment alone, as suggested in two studies.

• Several studies point to less-than-adequate knowledge regarding allergy treatment among patients in general medical practice.

• Few objective data are available to describe case mix and practice patterns in generalist and specialist care.

• Although survey data suggest that many patients are referred from generalist practices to specialist providers based on the severity of symptoms, there are no empirical published data to support that specialist practice has more severely affected patients.

Racial and Ethnic Variation

• There are few studies addressing any aspect of racial variation in relation to prevalence, treatment patterns, or response to treatment for patients with allergic rhinitis.

• The largest and most representative study, The National Health and Nutrition Examination Survey, 1976-80, does not show a consistent relationship between allergic rhinitis prevalence and race.

• Among the randomized trials reviewed for other questions addressed in this literature synthesis, only 13 of 116 described the racial characteristics of the study population.

• The only data on variation in treatment patterns with respect to race or ethnicity suggested that in a pediatric population, whites were more likely to continue injection immunotherapy treatment than non-whites.

• No data exist that describe variation in treatment outcomes by race.

Costs and Work Performance

Although several studies have estimated the burden of illness due to allergic rhinitis, cost estimates vary widely, and both methodological issues and changes in current practice limit the applicability of these studies. Methodological challenges include: the definition of allergic rhinitis (particularly when using administrative datasets); valid cost estimates that include over-the-counter medications; and valid, objective measures of productivity changes. Additional data are needed regarding how allergic rhinitis in children affects working parents' productivity. In addition, existing analyses antedate the increased use of non-sedating antihistamines and nasal glucocorticoids. An updated study that adequately addressed these issues would give a more valid estimate of the direct costs associated with allergic rhinitis.

Ideally, the effects of treatment on work performance would be determined from randomized trials that include objective measures of work performance. Alternatively, one could model the impact of treatments on work performance if valid links existed between symptom outcomes or health-related quality of life (HRQOL) measures and work performance. Unfortunately, we did not identify any studies that establish these links. Since symptom outcomes and HRQOL are typically easier to measure than productivity, studies that would allow one to associate a given change in symptom or HRQOL score with a corresponding change in work productivity across a variety of jobs would be a particularly valuable contribution.

Environmental Measures

Based on the pathophysiology of allergic rhinitis, interventions that decrease allergen exposure through environmental control measures are conceptually appealing. The small number of studies evaluating such interventions did not yield definitive results, but the data for house dust mite controls are encouraging. Future studies will need to overcome a number of conceptual and methodological challenges. Since individuals are often allergic to more than one allergen, allergen avoidance measures may be needed for each significant allergen. Most studies to date have focused environmental controls on house dust mites or indoor aeroallergens. More comprehensive measures, such as those recommended in the National Heart, Lung, and Blood Institute's “Practical Guide for the Diagnosis and Management of Asthma” (National Heart, Lung, and Blood Institute, 1997), should be tested in patients with allergic rhinitis and significant functional impairment. If comprehensive measures are effective, future studies should identify the most critical components, since lifestyle changes are often difficult for patients to adopt. Another practical issue is whether allergen avoidance measures are more effective when tailored to an individual patient's specific allergic sensitivities, or whether more general recommendations without specific allergy testing are adequate.

Immunotherapy

Immunotherapy (IT) is a potentially important treatment for allergic rhinitis. However, it requires special expertise, a committed patient, and is relatively expensive. Immunotherapy may be administered by injection, nasally, or sublingually, but there are few studies using the latter two routes of administration. Most studies have focused on patients with grass-pollen- or ragweed-induced seasonal allergic rhinitis. To better understand the role of IT in the treatment of allergic rhinitis, we need clinical trials employing vaccines containing most or all of the relevant allergens for each individual, which would allow us to assess IT as it is administered in most community settings. Such polyantigen studies would require new approaches to outcome measurement; currently, studies on seasonal allergens rely on timing symptom assessment to peak allergen levels. Additional future research objectives should be focused on the following: methods to identify patients likely to benefit from IT; cost-effectiveness and quality-of-life analyses of IT; determination of whether IT alters the natural history of allergic rhinitis and reduces possible sequelae such as bacterial sinusitis and asthma; comparisons of immunotherapy and the best available medical management and/or allergen avoidance; and studies clarifying the optimal duration of IT. Studies should be of sufficient duration to evaluate the short- and long-term effects of treatment, and adverse effects should be collected and reported systematically. An important subgroup to study is patients with co-occurring asthma, since effective treatment for allergic rhinitis has the potential to improve asthma symptoms.

Combined Treatments

To develop the most cost-effective management strategies, it is important to determine the relative efficacy of combinations of treatments compared to monotherapy. Compared to monotherapy, combined treatments are significantly more costly, and the potential effects range from no additional benefit to synergistic increases in efficacy.

The combination of an antihistamine plus a decongestant compared to either medication alone has been well studied in a large number of relatively short-term trials. Similarly, antihistamines plus nasal glucocorticoids have been compared adequately evaluated compared to either medication alone. Over 80 percent of these studies were done in patients with seasonal allergic rhinitis; longer duration studies in patients with perennial allergic rhinitis would provide useful efficacy data. In addition, longer duration “effectiveness trials” that included outcomes such as health-related quality of life and cost-effectiveness in primary care populations with clinically diagnosed seasonal or allergic rhinitis could guide policy. Other combinations (antihistamine, mast cell stabilizer, nonsteroidal anti-inflammatory drugs, ophthalmic antihistamine, and ipratropium) have been evaluated in single trials and more data are needed to better understand the efficacy of these combinations.

Clinician Specialty Differences

To understand the quality of current care for patients with allergic rhinitis, we need studies describing current practice patterns. Theoretically, earlier and more aggressive treatments that include allergy avoidance measures, immunotherapy, and medications may lead to better functional status, better work productivity, and fewer disease-related complications. Observational studies that compare treatment patterns and outcomes across specialties will need to pay careful attention to case-mix adjustment. A standardized and validated severity-of-illness scale would facilitate this research. In addition, prospective studies that compare symptomatic treatment to allergen identification with specific treatment would directly address two approaches commonly used in generalist and specialty practices. The development, implementation, and testing of clinical practice guidelines may provide the impetus for studying clinician practice patterns and outcomes as well as a framework for improving practice and evaluating outcomes. Finally, studying patient preferences and expectations for treatment and consulting behavior may provide important insights into clinician specialty case mix, practice patterns, and outcomes.

Racial and Ethnic Variation

Racial variability in disease prevalence, treatment patterns, or response to treatment can serve as cues to underlying differences in genetic susceptibility, environmental exposures, access to care, quality of care, or differing patient preferences for care. The few studies of disease prevalence did not show important differences by race. We did not identify any studies that described differences in treatment patterns or treatment response, in part because study populations were often incompletely described. We recommend that future studies give more complete descriptions of patient populations, including racial descriptors that might permit important subgroup analyses.

Need for Improved and More Uniform Trial Reporting

This evidence report highlights the need to improve the quality and homogeneity of trial reporting. Better reporting would aid interpretation and application of research findings and facilitate future literature syntheses. For clinical trials, the process for recruiting the study population and the population's clinical and demographic characteristics were often inadequately described. Thus the generalizability of study findings was often unclear. Design characteristics that help clinicians assess the validity of trial results were often incomplete, particularly information on randomization, allocation concealment, and, in some instances, blinding. Following the recommendations of the Consolidated Standards of Reporting Trials (CONSORT) statement for reporting trials would improve assessments of generalizabilty and validity (Moher, Schulz, Altman, et al., 2001).