Prevalence of Sinusitis

"Sinusitis" is often loosely used to denote a broad group of clinical syndromes with different causes and presentations. Estimates of disease prevalence vary, in part from differences in definitions of the disease and in part from differences in the populations in which the prevalence is assessed Sinusitis, as broadly defined and estimated through insurance reimbursement claims and population statistics (e.g., National Ambulatory Medical Care Survey [NAMCS]), is one of the most common health complaints in the United States (Kennedy, 1990). Gwaltney (1996), using estimates of the average number of acute respiratory illnesses per year (six to eight for children; two to three for adults) and data suggesting that 90 percent of patients with colds have sinusitis (viral or otherwise), estimated that there are over a billion cases of sinusitis annually.

Recent outpatient data from the 1995 National Ambulatory Medical Care Survey (NAMCS) estimated that the roughly 700 million visits to nonfederally employed physicians in office-based practices included about 3 million cases of acute sinusitis (International Classification of Diseases, 9^th edition, clinically modified [ICD-9-CM]-461). The NAMCS data from 1990 to 1995 (Table 1; National Center for Health Statistics, 1990-1995) show an increasing trend in the prevalence of this diagnosis (test for trend using linear regression; r²=0.63; p=0.06). It is not clear whether this trend represents an actual increase in disease prevalence, potential changes in reporting (e.g., disease definition, billing practices), or changes in patients seeking or accessing care (Kaliner, Osguthorpe, Fireman, et al., 1997; McCaig and Hughes, 1995; Stoller, Forster, and Portugal, 1993). The National Hospital Discharge Survey documented 61,000 hospital discharges for patients with diagnoses (primary or other) of acute sinusitis (ICD-9-CM-461) for 1993; 66,000 for 1994; and 84,000 for 1995 (National Center for Health Statistics, 1993, 1994, 1995).

Although population studies agree that sinusitis is common, they use broad and different definitions of the condition that include several pathophysiologic conditions. Estimating the prevalence of various sinusitis subgroups (e.g., acute bacterial rhinosinusitis) requires the use of more specific diagnostic criteria in research studies. Even if the same definition is used, however, selection biases can lead to differing estimates for the general population, for patients in specific clinical settings, and for patients enrolled in studies where other initial selection criteria are applied. The relationship between the prevalence estimates obtained from population-based prevalence data (e.g., from insurance claims and survey data) and those from various subgroups of rhinosinusitis defined by clinical or diagnostic criteria from research studies remains unclear.

Estimated Costs of Health Care: Individual and Societal

The high prevalence of sinusitis translates into high costs for individual health, work time lost, and medical expenditures. McCaig and Hughes (1995) analyzed the 1985, 1989, and 1992 NAMCS data and found an increasing trend of office visits for sinusitis. They reported sinusitis (ICD-9-CM-461 and -473, including both acute and chronic sinusitis) as the fifth most common diagnosis for antibiotic prescriptions, representing 7 percent, 9 percent, and 12 percent of all recorded prescriptions in 1985, 1989, and 1992, respectively. The use of more expensive, broad-spectrum drugs (e.g., cephalosporins) increased, and the use of less expensive, narrow-spectrum antibiotics (e.g., the penicillins) decreased during this period. These trends have important ramifications in terms of health care costs, as well as the development of resistance to antimicrobials.

In 1992, in the United States, approximately $200 million was spent on prescription cold medications for sinusitis, a $50 million increase over 1989 (Gwaltney, Jones, and Kennedy, 1995). In addition, the U.S. population spends more than $2 billion annually on over-the-counter medications for nasal and sinus disorders (Williams, Aguilar, Makela, et al., 1997). Increasing rates of antibiotic resistance have led to a recognition of the need for more prudent use of antibiotics and to limiting their use to the treatment of bacterial infections (Levy, 1998). The costs of antibiotic use need to be balanced against the limitations of diagnostic certainty and risks of nontreatment. The high prevalence of sinusitis and its associated costs highlight the need for optimizing effective therapy.

Biology of the Disease

The paranasal sinuses consist of four pairs of air-filled cavities in the skull (the frontal, maxillary, ethmoid, and sphenoid sinuses), which are lined by mucosa and connect to the nasal passages (Gwaltney, 1996; MacLeod, 1991). Normal mucous secretions contain antibodies, help collect soluble pollutants, and, together with ciliary action, work to clear particulate matter, including bacteria, from the sinuses (Gwaltney, Jones, and Kennedy, 1995). Maintaining the mucociliary flow and an intact local mucosal surface are key host defenses against infection and for maintaining sinus health (Gwaltney, 1996; Gwaltney, Jones, and Kennedy, 1995). It is widely believed that the paranasal sinuses are normally sterile, although there have been conflicting reports (Bjorkwall, 1950; Brook, 1981; Gwaltney, Scheld, Sande, et al., 1992).

Sinusitis vs. Rhinosinusitis: Symptomatic vs. Asymptomatic

In addition to inflammation of the paranasal sinuses, most cases of sinusitis are accompanied by inflammation of the nasal passages (Gwaltney, Phillips, Miller, et al., 1994; Lund and Kennedy, 1995). As such, the clinical condition often referred to as "sinusitis" is, in fact, rhinosinusitis: inflammation of the sinuses with concomitant inflammation of the nasal passages. In some circumstances, inflammation may also include portions of the surrounding bone (Lanza and Kennedy, 1997). For the purposes of this report, this grouping of conditions will be referred to as rhinosinusitis.

Studies using computed tomography or magnetic resonance imaging (MRI) have reported sinus mucosal abnormalities in 15 percent to 49 percent of patients who have no symptoms suggesting rhinosinusitis; the clinical correlations of these findings remain unclear (Axelsson and Chidekel, 1972; Calhoun, Waggenspack, Simpson, et al., 1991; Gordts, Clement, and Destryker, 1997; Kaliner, Osguthorpe, Fireman, et al., 1997; Lloyd, Lund, and Scadding, 1991; Patel, Chavda, Violaris, et al., 1996). Although all cases of rhinosinusitis involve inflammation of the mucosal linings, in the practice setting, the focus is on those patients in whom this inflammation leads to symptoms. Table 2 lists several possible clinical presentations of rhinosinusitis, although these symptoms are not present in all patients and are not very specific (Hadley and Schaefer, 1997; Shapiro and Rachelefsky, 1992; Williams and Simel, 1993; Williams, Simel, Roberts, et al., 1992).

Causes of Rhinosinusitis as a Basis for Therapy

Clinical diagnosis aims to identify cases that have a similar pathophysiologic cause and presentation and that therefore would presumably benefit from similar treatment. Factors affecting inflammation of the sinuses include infectious agents, allergic conditions, anatomic abnormalities, systemic diseases (endocrine, metabolic, genetic), trauma, and noxious chemicals (Gwaltney, 1996; Gwaltney, Jones, and Kennedy, 1995; Lanza and Kennedy, 1997). In some cases, these factors cause inflammation of the mucosal linings directly (e.g., allergic, infectious); in others, host conditions predispose the mucosal linings to inflammation, infection, or both (e.g., anatomic abnormalities, neoplasms, ciliary function abnormalities). Although infectious agents can be the primary cause of sinus inflammation, they also may represent a secondary infection. In these cases, the initial inflammation predisposes the sinuses to infection, as for example, when bacterial rhinosinusitis follows viral rhinosinusitis (Berg, Carenfelt, Rystedt, et al., 1986; Gable, Jones, Floor, et al., 1994).

In many instances, a patient's symptoms are the result of several environmental and host factors working together. In clinical practice, diagnosis is undertaken with an eye toward therapeutic or preventive interventions. As such, diagnoses are developed and refined to identify specific clinical conditions for which a therapy may be effective. For treating rhinosinusitis, these therapies may be directed toward current symptoms, the underlying cause, or both (Table 3).

Bacterial Rhinosinusitis

Microbiologic classification of infectious causes of rhinosinusitis includes bacteria, viruses, and fungi. Data on the prevalence of cases resulting from each of these microbiologic agents are lacking. However, viruses, as the leading cause of upper respiratory infections, are among the most common (Wald, 1996). Approximately 0.5 percent to 2 percent of adult and up to 10 percent of pediatric cases of viral rhinosinusitis develop into bacterial infections (Berg, Carenfelt, Rystedt, et al., 1986; Dingle, Badger, and Jordan, 1964; Gable, Jones, Floor, et al., 1994; Gwaltney, 1996).

The rationale for identifying patients with increased likelihood of infectious rhinosinusitis stems from the potential use of anti-infective agents. Antibiotics are widely available and are effective at eliminating specific bacteria. The ability to identify cases of bacterial rhinosinusitis (either as a primary or secondary infection) would thus identify potential candidates for antibacterial therapies (Figure 1). Adequate levels of antibiotics can eradicate bacteria from maxillary sinus fluid aspirates (Eneroth and Lundberg, 1976; Hamory, Sande, Sydnor, et al., 1979).

In this report, we focus on acute bacterial rhinosinusitis because antibiotics are widely available and should be used only for bacterial infection. Bacterial infection of the sinuses can result in chronic sinusitis, as well as in other serious complications (e.g., meningitis, brain abscess). Antibiotics can be used to prevent these developments. At the same time, concerns have increased about the inappropriate use of antibiotics, both for the individual and society. For the individual, the concerns are for potential side effects and out-of-pocket expenses. For society, the concerns are for overall costs and the development of antibiotic resistance leading to loss of the therapeutic efficacy of antibiotics (Levy, 1998).

Bacterial rhinosinusitis can refer to several physiologic conditions, all of which include the presence of bacteria concomitant with sinus inflammation. The inflammation may be a direct response to bacterial infection or to nonbacterial causes that provide a setting for a secondary bacterial infection. Given, however, that in all instances bacterial rhinosinusitis entails bacterial infection, a microbiologic definition remains the current accepted diagnostic reference standard: more than 10⁴ colony-forming unit (CFU)/ml in sinus aspirate (Turner, Cail, Hendley, et al., 1992; Winther and Gwaltney, 1990). Lower colony concentrations could potentially represent early infection. However, studies of maxillary sinus aspirates generally yield titers above 10⁵ CFU/ml (Winther and Gwaltney, 1990), and the cutoff choice of 10⁴ CFU/ml of sinus aspirate (rather than complete absence of bacteria) may in part compensate for the potential contamination of the sinus aspirate during collection (Wald, 1991). Whether the sinuses are sterile under normal circumstances or whether they are routinely colonized with anaerobic and aerobic bacteria is still debated (Bjorkwall, 1950; Brook, 1981; Gwaltney, Scheld, Sande, et al., 1992).

Several studies have reported bacterial species profiles isolated from maxillary sinus aspirates and have looked at changes in the predominant species over time (Berg, Carenfelt, and Kronvall, 1988; Bjorkwall, 1950; Gwaltney, Scheld, Sande, et al., 1992; Jousimies-Somer, Savolainen, and Ylikoski, 1988; Suzuki, Nishiyama, Sugiyama, et al., 1996; Urdal and Berdal, 1949). Whereas -hemolytic streptococci and Streptococcus pneumoniae were the most frequent isolates in studies circa 1950 (Bjorkwall, 1950; Urdal and Berdal, 1949), studies in the 1970s and 1980s noted that S. pneumoniae and Hemophilus influenzae predominate (Berg, Carenfelt, and Kronvall, 1988; Gwaltney, Sydnor, and Sande, 1981; Jousimies-Somer, Savolainen, and Ylikoski, 1988; Van Cauwenberge, Verschraegen, and Van Renterghem, 1976). Followup studies in the 1990s reported that these two organisms remained the major species in adults, whereas Moraxella catarrhalis was an additional, high-prevalence species isolated from children (Benninger, Anon, and Mabry, 1997; Gwaltney, Scheld, Sande, et al., 1992; Suzuki, Nishiyama, Sugiyama, et al., 1996; Wald, Reilly, Casselbrant, et al., 1984). Other species were also cultured in many of these studies (including -hemolytic streptococci, Staphylococcus aureus, and anaerobes), but their prevalence was much lower in cases of acute bacterial rhinosinusitis (Benninger, Anon, and Mabry, 1997; Berg, Carenfelt, and Kronvall, 1988; Brook, 1996; Gwaltney, Sydnor, and Sande, 1981; Gwaltney, Scheld, Sande, et al., 1992; Jousimies-Somer, Savolainen, and Ylikoski, 1988; Wald, 1991).

Even though direct microbiologic evaluation of sinus aspirates has been the diagnostic reference standard, the puncture technique has been largely limited to sampling the maxillary sinuses. One study reported good agreement between the bacterial species obtained from the frontal sinus trephination and those obtained from the same patients' maxillary sinus (Antila, Suonpaa, and Lehtonen, 1997). However, the prevalence of bacterial species in sinuses other than the maxillary sinuses and the potential clinical significance of different patterns of the sinuses affected require further study.

Although sinus puncture and culture is the diagnostic standard for bacterial rhinosinusitis, its routine use in primary care practices is not feasible. As such, the practical identification of patients with acute bacterial rhinosinusitis remains a clinical diagnosis that relies on alternate diagnostic methods. Less invasive methods for direct sampling and microbiologic testing (i.e., nasopharyngeal swabs and middle meatal sampling by endoscopy) have been compared with sinus puncture aspirates, but the degree of agreement was weak (Axelsson and Brorson, 1972; Evans, Sydnor, Moore, et al., 1975; Gwaltney, Sydnor, and Sande, 1981; Williams, Holleman, Samsa, et al., 1995). However, modifications in the techniques for middle meatal sampling by endoscopy may improve their accuracy (Druce, 1992; Ferguson and Mabry, 1997; Klossek, Dubreuil, Richet, et al., 1996).

Further Subgrouping Based on Patterns of Disease Presentation

In patients with bacterial rhinosinusitis, further distinct subgroups may be defined that differ in pathophysiology and whose identification and distinction could further direct specific therapy. Subgrouping criteria have included patterns of disease presentation (e.g., temporal patterns) and patient characteristics (e.g., age, comorbidities).

Choice of Therapies

This report focuses on acute bacterial rhinosinusitis, given the question of the effectiveness of antibiotics for treatment. Although antibiotic use was the a priori basis for focusing on this subgroup, we also explored evidence for other therapies for this diagnosis (Figure 3).

Outcome Measures for Assessing Efficacy

The eventual outcomes of acute rhinosinusitis, including acute bacterial rhinosinusitis, are generally excellent even without specific treatment. Short-term, placebo-controlled studies show improvement or resolution of symptoms without antibiotic treatment in approximately two-thirds of patients between 8 and 12 days after presentation to the physician. Even in the preantibiotic era, serious complications (meningitis, brain abscess, and osteomyelitis) were rare. Efficacy estimates for the treatment of acute bacterial rhinosinusitis depend on the outcomes measured and may include relief of symptoms and reduced incidence of serious complications, rates of relapse and reinfection, progression to chronic rhinosinusitis, and adverse side effects of the treatments.

Prevalence

Because the unequivocal diagnosis of acute bacterial rhinosinusitis requires a bacteriologic evaluation (routinely obtained by maxillary sinus puncture), the true prevalence of acute bacterial rhinosinusitis is difficult to obtain reliably. There are no epidemiologic studies that used sinus puncture to estimate the prevalence of this condition in a given population. In this evidence report, the prevalence of sinusitis is estimated from diagnostic test studies and from treatment studies that used sinus puncture as the reference standard. In addition, we review data from several observational studies.

Diagnostic Tests

Studies identified in the literature search described in the Search Strategy and Study Selection sections were included if they presented data prospectively comparing the performance of two or more tests in the diagnosis of acute bacterial rhinosinusitis. These diagnostic tests included clinical criteria (studies had to evaluate a composite measure such as overall clinical impression or a decision aid such as a risk score), radiographs, ultrasonography, or sinus puncture/aspiration. Although studies evaluating computed tomography or magnetic resonance imaging would be eligible, no comparative studies of these tests meeting inclusion criteria were identified (there were several studies of chronic sinusitis). Studies were excluded if diagnostic tests were evaluated on individuals not presenting with symptoms of acute bacterial rhinosinusitis. However, the definition of acute bacterial rhinosinusitis varied among studies; for example, two studies evaluated some subjects with prolonged symptoms (Berg and Carenfelt, 1988; Williams, Simel, Roberts, et al., 1992), whereas others did not provide a definition of sinusitis. To avoid the problem of verification bias (Irwig, Tostetson, Gatsonis, et al., 1994), studies were excluded if some subjects did not undergo all tests being compared.

A comparison matrix (Evidence Table 1) using studies that met the inclusion criteria was constructed to help visualize the number of studies available for analysis. Meta-analyses were performed on comparisons where there were at least three studies or subgroups. These meta-analyses provided results to answer question 2.

Treatment Trials

Evaluation of antibiotics and ancillary treatment trials were considered separately. A matrix of antibiotic comparisons (Evidence Table 9) was constructed using randomized controlled trials meeting the minimum inclusion criterion of including patients with suspected acute bacterial rhinosinusitis of 4 weeks or less in duration. The intersection of a row and a column in this matrix denotes a comparison between two different antibiotics or a dosage or duration comparison of the same antibiotic. The number within a cell denotes the number of comparisons. An empty cell has no comparison available. For example, four comparisons are listed in the cell intersecting the row heading of amoxicillin and column heading of cefixime. The references for the comparisons are provided in Evidence Table 10. A total of 74 unique comparisons from 72 articles were identified.

The comparison matrix is used to help assess whether a relevant clinical question is addressable by determining whether a sufficient number of potentially useful studies is available to conduct a meta-analysis. In consultation with the technical experts, three meta-analyses of antibiotic trials were identified. The meta-analyses are based on an article published by several EPC members in the British Medical Journal (deFerranti, Ioannidis, Lau, et al., 1998). This meta-analysis was updated and provided the basis for the supplemental analyses of antibiotic treatment in this evidence report. Parts of this published article are used in this evidence report with permission from the journal.

Key questions 3 and 4 formulated by the technical expert panel concern the efficacy of antibiotics compared with that of placebo and the comparative efficacy of different antibiotics. Examination of the matrix (Evidence Table 9) identified 10 comparisons of antibiotics with placebo to address question 3. Some of these studies did not meet the inclusion criteria and were excluded from the meta-analysis. Similarly, two additional meta-analyses were identified for question 4.

Trials were eligible for inclusion in a meta-analysis if three criteria were met: (1) the trial compared amoxicillin or a folate inhibitor agent (e.g., trimethoprim/sulfamethoxazole) to another agent, generally one with a broad spectrum of activity including cephalosporins, penicillins with β-lactamase inhibitors, tetracyclines, quinolones, and macrolides; (2) patients were assigned randomly; and (3) the trial evaluated acute sinusitis or an acute exacerbation of chronic sinusitis ("acute-on-chronic"). Both adult and pediatric studies qualified. Trials of subacute or chronic sinusitis (greater than 4 weeks mean symptom duration) were excluded. Although dosage or duration comparison studies would provide useful information, a quick scan of the matrix revealed few such studies, and a meta-analysis was not possible. Placebo-controlled studies were examined to assess the effect of antibiotics on the natural history of acute bacterial rhinosinusitis.

Because the focus of this evidence report is on uncomplicated, community-acquired, acute bacterial rhinosinusitis, we excluded studies with immunocompromised patients such as patients with malignancy receiving chemotherapy, human immunodeficiency virus (HIV) infection, cystic fibrosis, asthma, Kartagener's syndrome, IgA and IgG deficiencies, and trauma or surgery-related sinus infections. Even though HIV infection and asthma are part of the exclusion criteria, we screened the MEDLINE search results for randomized controlled trials specific for these populations with the intention of performing separate subgroup analyses. We found none.

Since we found only 10 randomized controlled trials of ancillary treatments, a comparison matrix is not useful. Because of the heterogeneous mix of treatments, diagnostic definitions, and protocols, a meaningful meta-analysis was not possible for ancillary treatments. The data from these studies were abstracted and summarized in the evidence tables.

Diagnostic Test Studies

For each included study, we extracted test data cross-classifying individuals as having or not having bacterial rhinosinusitis. To calculate sensitivity and specificity in a comparison of two diagnostic tests, it is necessary first to decide which test is the "test of interest" and which is the "reference test." For included studies, comparisons of diagnostic tests were derived in a manner consistent with a "hierarchy" of accuracy: most accurate was sinus puncture, followed by radiography, ultrasonography, and then clinical criteria. Therefore, for example, estimates of sensitivity and specificity of ultrasound were derived with respect to radiography (and not vice versa). Many studies reported data only for "sinuses" and not for "patients"; these sensitivity and specificity data were used in the analyses.

When studies presented test performance data for more than one threshold or cut-point for tests of interest, we extracted data for each cut-point separately. For instance, for the clinical examination compared with radiography, data for overall clinical impressions of intermediate probability and high probability were included separately.

The following data were also extracted from each study: country where the study was performed and publication year of the study; age of study participants and duration of their symptoms; location of the study (hospital, office practice, or emergency department); and type of physicians who evaluate patients (primary care physicians or otolaryngologists). We also noted whether each diagnostic test was evaluated in a manner blinded to the results of the other evaluated tests.

Treatment Trials

Outcomes of interest were clinical "cure," "improvement," and "failure" as assessed within 48 hours of the end of treatment. Although we would have been interested in analyzing other outcome measures such as rates of improvement in the treatment and control arms and relapses, most studies do not report them. Cures and failures were recorded as defined by the individual study; "cure" generally meant resolution of all signs and symptoms, and "failure" generally signified no change or worsening of signs and symptoms. Data on radiographic "cure," "improvement," or "failure" and bacteriologic "cure" or "failure" were also extracted as defined by each study. The main analyses used clinical outcomes as the endpoint most relevant to clinicians because primary care practitioners do not routinely obtain sinus films for uncomplicated acute bacterial rhinosinusitis and almost never perform cultures of sinus aspirates. Furthermore, there is only limited evidence suggesting a correlation between radiographic or bacteriologic failure and clinical outcomes. Separate analyses assessed bacteriologic failures, radiographic failures, and patient withdrawals due to adverse drug effects.

In addition to clinical outcomes, data were also extracted on study design characteristics such as blinding, disease definition, publication year, and age group. These factors provided the basis for sensitivity analyses.

Meta-Analysis of Diagnostic Tests

Evidence Tables 3 through 8 list the items frequently considered in various quality assessments of diagnostic test studies. These items include specification and diagnostic criteria of the reference standard and the test and blinding of the interpreter of a test to the clinical information and the results of the other test. There were too few studies in each of the categories of the diagnostic test comparisons to allow a meaningful sensitivity analysis. Therefore, these items are included for descriptive purposes only. To avoid the problem of verification bias (Irwig, Tostetson, Gatsonis, et al., 1994), studies were excluded if some subjects did not undergo all tests being compared.

Meta-Analysis of Treatment Trials

In each trial, the following characteristics pertaining to the quality of study design, conduct, and reporting were assessed by two investigators with subsequent consensus: blinded vs. unblinded design, specification of criteria for the diagnosis of sinusitis, detailed reporting as to the use of decongestants, and robustness of the assessment of clinical outcomes and completeness of the information on outcomes (losses to followup). The diagnosis of sinusitis was categorized as "firm" if a trial performed sinus aspirations and culture or radiographic evaluations (assessing the presence of air-fluid levels, mucosal thickening greater than 6 mm, or opacification of sinuses). Any other diagnostic criteria, including clinical judgment or nasal swabs, were categorized as "subjective." Outcome criteria were judged to be well-specified when a study scaled symptoms or signs as assessed by patients and/or physicians in a way that could be replicated. Trials, which specified criteria to some extent, noted the signs or symptoms used to evaluate cure, improvement, or failure but were not specific about how these data were evaluated. Trials with unclear criteria made no mention of how clinical outcomes were determined.

In addition to this subject-specific assessment of quality components, we also used a previously developed, validated scale of assessing the methodologic quality of the trials on a scale of 0 to 5 with a value of 5 being the perfect score of a specific study (Jadad, Moore, Carroll, et al., 1996). This scale focuses on randomization, double-blinding, and description of withdrawals and has been widely implemented.

We further explored quality issues by conducting subgroup analyses by dichotomizing studies into two groups using factors thought to be associated with higher quality. Cumulative meta-analyses ordered by methodologic quality of the studies were used to explore possible treatment effect trends as studies with lower quality scores were added to studies with higher quality scores.

Meta-Analysis of Diagnostic Test Studies

We followed the general principles of conducting a meta-analysis of diagnostic test studies described several years ago in the Annals of Internal Medicine (Irwig, Tosteston, Gatsonis, et al., 1994). For each combination of the test of interest and the reference test, a summary receiver operator characteristic (SROC) curve was constructed based on the method described by Moses, Shapiro, and Littenberg (1993). Multiple data points from studies that provided data at different cut-points were used to derive these curves; because these observations were not independent, confidence intervals around SROC curves could not readily be calculated. SROC curves were derived for data points weighted by the inverse of the variance. When studies provided estimates of specificity over a wide range (approximating the total possible range from 0 to 1), the area under the SROC curve was calculated by extrapolation. In addition to the SROC method, random-effects weighted average was used to calculate the average sensitivity and specificity for each comparison. Although pooling these values separately as we did tends to underestimate the true test sensitivity and specificity, they are nonetheless useful estimates of the average test performance. We assessed the appropriateness of this method by noting the distance of the estimates from the SROC curve. Statistical analyses using the SROC curve method was performed using "Meta-Test" version 0.6, a computer program developed by the EPC director (Dr. Lau).

Alternative Method for Summarizing Diagnostic Test Data

In addition to the SROC curve method, we explored a method described by de Bock, Dekker, Stolk, et al. (1997) for combining studies to estimate the sensitivity and specificity of diagnostic tests either when there is no reference standard or when a reference standard is not available for all comparisons. We applied this method to the diagnostic studies for acute rhinosinusitis.

Estimates are obtained by maximizing the likelihood through the Expectation Maximization (EM) algorithm (Dempster, Laird, and Rubin, 1977). Through an iterative procedure, the method estimates the true sensitivity and specificity of each test without assuming that any of the tests is a reference standard. First, starting values are assigned for the sensitivity and specificity of each test and for prevalence of the disease associated with each study. Second, based on these parameter values, the number of diseased and not diseased are estimated for each study. This is called the E-step, for estimation. Third, the maximum likelihood estimates are computed using the estimates from the E-step. This is called the M-step, for maximization. The estimates from the M-step are then used in the E-step, and the process iterates until convergence. The process should be repeated with different starting values to make sure the initial choice of parameters does not affect the final estimates.

The validity of the method relies on two key assumptions. One assumption is that an individual's results on two tests are independent, conditioned on whether the person has the disease. The other assumption is that the sensitivity and specificity of each test do not vary from study to study. Both of these assumptions are violated by the data available in the literature. Most studies did not blind the interpreter of the reference test to the results of the test of interest, so conditional independence is unlikely. Another way conditional independence can be violated is if the disease is not dichotomous, and different tests pick up the disease at different severity levels, as pointed out by the authors of this method. In our meta-analysis of diagnostic tests, the SROC curve analysis is evidence that the sensitivity and specificity for a given modality vary. Indeed, there are several criteria for defining a positive sinus radiograph with sensitivity estimates varying from 0.41 to 0.90.

We abandoned this method because of the violations of its assumptions and based our conclusions on the SROC curve method.

Meta-Analysis of Antibiotics Randomized Controlled Trials

First, treatment outcomes from studies that had placebo arms were pooled to determine the effect of treatment with any antimicrobial on the natural history of acute bacterial rhinosinusitis. Second, two main comparisons between antibiotic groups were made: newer and/or expensive antibiotics including cephalosporins, macrolides, quinolones, tetracyclines, and penicillins with a β-lactamase inhibitor vs. amoxicillin, and newer and/or expensive antibiotics (as above) vs. folate inhibitors.

The general method of quantitative synthesis was followed (Laird and Mosteller, 1990; Lau, Ioannidis, and Schmid, 1997). Pooling of risk ratios, risk differences, and control group event rates were performed using both the Mantel-Haenszel fixed effects model (Mantel and Haenszel, 1959) and the DerSimonian and Laird random effects model (Fleiss, 1993; Ioannidis, Cappelleri, Lau, et al., 1995). The random effects model takes into account the variability of the true treatment effect between studies and provides a wider confidence interval (compared with the fixed effect model) when heterogeneity is present. Heterogeneity between studies was assessed with a chi-square statistic. This test is not very sensitive; therefore, heterogeneity was considered statistically significant if p<0.10 (Lau, Ioannidis, and Schmid, 1997). Weighted rates are also reported; rates were weighted by the inverse of their variance with random effects.

For the main outcome of clinical failure, we performed sensitivity analyses excluding: (1) pediatric studies; (2) studies in which amoxicillin or folate inhibitors were compared against tetracyclines which have been available almost as long, but continue to be more expensive; (3) studies in which patients with resistant organisms had been excluded; (4) studies in which diagnosis of sinusitis was not made on firm criteria; (5) studies with unclear assessment of outcomes; (6) studies that were not double-blind; (7) studies published before 1993; and (8) studies with a "Jadad" quality score of lower than 3. In each case, heterogeneity between excluded and remaining studies was assessed by the chi-square test and deemed significant for p<0.1. Statistical analyses of pooling treatment effects from randomized controlled trials were performed using "Meta-Analyst" version 0.991, another computer program developed by the EPC director (Dr. Lau).

Use of Sinus Puncture as the Research Diagnostic Reference Standard

The research reference standard for verification of bacterial infection of the sinuses relies on microbiologic culture of sinus aspirates. Of the studies reviewed for inclusion in this evidence report, 7 of 14 studies of diagnostic test comparisons, 5 of 30 antibiotic treatment trials, and none of the 10 ancillary treatment trials used this technique for evaluation. The procedure requires technical expertise and is invasive, which may account in part for its low rate of utilization, particularly in treatment evaluations (Laine, Maättä, Varonen, et al., 1998; van Buchem, Peeters, Beaumont, et al., 1995). Anatomy of the sinus cavities limits current sinus puncture sampling to aspirates of the maxillary sinuses, and as such, use of this reference standard in research leads to direct conclusions regarding maxillary sinus infections and not infections of other sinuses (Gwaltney, Scheld, Sande, et al., 1992). Data comparing microbial flora from infected maxillaryand frontal sinuses in individual patients correlated well but are very limited (Antila, Suonpaa, and Lehtonen, 1997).

Duration of Symptoms

One of the criteria for paper inclusion was study of patients with acute bacterial rhinosinusitis defined as symptoms of 4 weeks duration or less. Review of the literature revealed that many of the papers only specified "acute" without providing specific time periods. This was true for studies of the diagnostic tests (Evidence Tables 3-6) as well as studies of both antibiotic (Evidence Tables 15-17) and ancillary (Evidence Tables 22-24) treatments. In some instances, the duration of symptoms was reported as a mean value. In these cases, studies reporting mean estimates of 4 weeks or less were included. The lack of clearer definition of the populations used in these research studies raises questions regarding the validity of directly comparing the results of different studies and applying these results to specific patient populations.

Patient Age

Most studies of diagnostic tests focused on adult populations. In many studies, patients under 18 years of age were also included (usually teenagers), but the reported results did not distinguish these distinct age groups. Two studies comparing radiography with puncture did include patients as young as 7 and 10 years old (McNeill, 1962, and Revonta, 1980, respectively) but did not separately report the pediatric population findings. One study from Sweden (Jannert, Andreasson, Helin, et al., 1982) compared clinical examination and radiography in a solely pediatric population. Other pediatric population studies (Wald, Chiponis, and Ledesma-Medina, 1986; Wald, Reilly, Casselbrant, et al., 1984) reported on the results of diagnostic studies in their patients but did not perform all tests on all patients, precluding use of these data for diagnostic test comparisons.

Most studies on antibiotic treatment of rhinosinusitis used adult populations. Two studies focused specifically on pediatric populations (Wald, Chiponis, and Ledesma-Medina, 1986; Wald, Reilly, Casselbrant, et al., 1984). Although adult medicine in the United States usually includes patients over 18, many of the trials also included some younger patients, from as young as 10 years old (Wallace, Marsh, and Talbot, 1985), within their study populations. In these studies, however, reported results did not distinguish between the adult and pediatric patients.

Among studies of ancillary treatments, two studies focused specifically on pediatric populations, each testing different ancillary medications (Barlan, Erkan, Bakir et al., 1997; McCormick, John, Swischuk, et al., 1996). Most of the other ancillary treatment trials included some pediatric patients (most over 9 years of age) but did not separately analyze the outcomes for these patients as compared with the adults.

Sinus Radiograph Compared with Sinus Puncture

Six studies compared sinus radiographs with sinus puncture (Kuusela, Kurri, and Sirola, 1982; Laine, Maättä, Varonen, et al., 1998; McNeill, 1962; Revonta, 1980; Savolainen, Pietola, Kiukaanniemi, et al., 1997; van Buchem, Peeters, Beaumont, et al., 1995). Each of the studies used a series of three or four radiographs; all included the occipitomental (Water's) view. Because three studies provided more than one comparison of test performance, 10 values of sensitivity and specificity were available for analysis.

Figure 5 displays these 10 values of sensitivity and specificity and the SROC curve derived from them. The 10 data points appear to be well described by the curve. Shown as dark gray ellipses are the five observations of sensitivity and specificity using the criterion "sinus fluid or opacity" to define positive radiographs. Shown as light gray ellipses are the three observations based on the criterion "sinus fluid or opacity or mucous membrane thickening" to define positive radiographs. The remaining two estimates are shown as white ellipses: One of these studies used the criterion "opacification of sinus," and the explicit criteria defining a positive radiograph were not available for one estimate.

Adding "mucous membrane thickening" as one of the criteria for a positive radiograph increased the sensitivity of radiographs and decreased their specificity. The random effects estimates for sensitivity and specificity, using "fluid or opacity" as the definition of a positive radiograph, were 0.73 (95 percent confidence interval [CI], 0.60-0.83) and 0.80 (0.71-0.87), respectively. With the definition of positive radiograph "sinus fluid or opacity or mucous membrane thickening," the estimates for sensitivity and specificity were 0.90 (0.68-0.97) and 0.61 (0.20-0.91), respectively. These random-effects estimates and confidence intervals are displayed as "X's " and rectangles in Figure 6. With positive radiographs restricted to "opacification of sinus," specificity increased only slightly to 0.85 (0.76-0.91), but sensitivity decreased dramatically to 0.41 (0.33-0.49). The area under the weighted SROC curve was 0.83.

Clinical Criteria Compared with Sinus Puncture

A single study compared clinical criteria with sinus puncture (Berg and Carenfelt, 1988). This study reported data on clinicians' overall impression and also provided a risk score derived from the number of findings present from the following four-item list: purulent rhinorrhea with unilateral predominance, local pain with unilateral predominance, bilateral purulent rhinorrhea, and presence of pus in the nasal cavity.

Figure 7 plots the five values of sensitivity and specificity derived from this report. The four-item risk score (shown with dark gray ellipses) appears to have better discrimination than the overall clinical impression (white ellipse); the SROC curve, fit only to the risk score thresholds, has an area-under-the-curve of 0.91.

Unfortunately, characteristics of this study throw into question its internal validity. First, the reference test, sinus puncture and aspiration, is poorly described because it is not clear whether radiography was used in conjunction with aspiration in identifying those with sinusitis. Second, it is unclear how to use both "purulent rhinorrhea with unilateral predominance" and "bilateral purulent rhinorrhea" as independent risk-score predictors of sinusitis.

Clinical Examination Compared with Sinus Radiography

Three studies evaluated clinical examination in comparison to sinus radiographs. One study compared an otolaryngologist's overall clinical impression with sinus radiography (Axelsson and Runze, 1976). The study by Jannert, Andreasson, Helin, et al. (1982) evaluated a clinical risk score for children in which individuals could have 0 to 3 of the following findings: purulent nasal secretions on examination, history of upper respiratory infection during the 2 weeks prior to presenting symptoms, and sinus pain or tenderness. The study by Williams, Simel, Roberts, et al. (1992) evaluated a clinical risk score for adults in which individuals could have 0 to 5 of the following findings: maxillary toothache, abnormal transillumination, poor response to decongestants, purulent secretions on examination, and colored nasal discharge by history. The Williams, et al., (1992) study also included data for overall clinical impressions of "intermediate" or "high" probability of bacterial rhinosinusitis. There were therefore 11 values of sensitivity and specificity available for analysis.

Figure 8 displays these 11 values of sensitivity and specificity and the SROC curve. The data points are well described by the SROC curve. The risk score results, shown as gray ellipses, appear to have similar discrimination to the overall impressions of clinicians, shown as white ellipses. The area under the weighted SROC curve was 0.74.

Ultrasound Compared with Sinus Puncture or Radiography

Five reports provided data comparing ultrasound of the sinuses with sinus aspiration (Kuusela, Kurri, and Sirola, 1982; Laine, Maättä, Varonen, et al., 1998; Revonta, 1992; Savolainen, Pietola, Kiukaanniemi, et al., 1997; van Buchem, Peeters, Beaumont, et al., 1995). Reports provided data on more than one set of patients or for more than one diagnostic cut-point, so there were 10 values of sensitivity and specificity available for analysis (Figure 9). Of note, these points do not appear well described by the SROC curve, implying that variability in test performance is present. Ultrasound performance appeared to be poorest in the study by Laine, Maättä, Varonen, et al. (1998) (shown as a white ellipse in Figure 9); this was the only study in which untrained primary care physicians performed and interpreted the ultrasounds.

Three reports provided data for five comparisons of ultrasound to sinus radiograph (Figure 10) (Berg and Carenfelt, 1985; Jensen and von Sydow, 1987; Rohr, Spector, Siegel, et al., 1986). It is difficult to interpret these comparisons because the five data points fall close together in the SROC plot, and it is unclear how well the SROC curve describes them or how the SROC curve can be extrapolated.

Clinical Examination Compared with Ultrasound

A single study compared findings on clinical examination with ultrasound (van Duijn, Brouwer, and Lamberts, 1992). Although this study provided data on sensitivity and specificity of individual clinical symptoms and signs, it provided no data for an overall clinical impression or risk score. However, summarizing data from this study, a published meta-analysis reported a sensitivity of 0.36 (95 percent CI, interval 0.29-0.43) and specificity of 0.90 (0.85-0.93) for clinical examination compared with ultrasound (de Bock, Houwing-Duistermaat, Springer, et al., 1994).

3. Given a (clinical) diagnosis of acute bacterial rhinosinusitis, are antibiotics effective in resolving symptoms and in preventing complications or recurrence?

The evidence to answer this question and question 4 is based on a meta-analysis of randomized controlled trials of antibiotic treatment for acute bacterial rhinosinusitis recently published by EPC members (de Ferranti, Ioannidis, Lau, et al., 1998). Attachment C presents a detailed description of the meta-analysis of antibiotic treatment trials and the results. The key results are reproduced here.

Six studies compared any antibiotics against placebo (Axelsson, Chidekel, Greblius, et al., 1970; Ganaça and Trabulsi, 1973; Lindbaek, Hjortdahl, and Johnson, 1996a; Stalman, van Essen, van der Graaf, et al., 1997; van Buchem, Knottnerus, Schrijnemackers, et al., 1997; Wald, Chiponis, and Ledesma-Medina, 1986) (Evidence Table 13 and Figure 11). Antibiotics were significantly more effective than placebo, reducing treatment failures by almost one-half. However, symptoms improved or were cured in 69 percent of patients without any antibiotic treatment at all (95 percent CI, 57 to 79 percent). Although the observed heterogeneity did not reach statistical significance, there was a suggestion that one trial (Stalman, van Essen, van der Graaf, et al., 1997) that included patients simply on the basis of sinusitis-like symptoms without further diagnostic documentation had the highest cure or improvement rates in the placebo group (85 percent at 10 days) and showed absolutely no benefit from antibiotics, whereas trials with more tightly defined patient populations and lower spontaneous improvement rates showed a more clear benefit from antibiotics.

4a. In treatment of acute bacterial rhinosinusitis, what is the efficacy of antibiotics compared with that of placebo, and among the various antibiotics, what is their comparative efficacy?

Fourteen trials compared newer antibiotics, most of which have an expanded spectrum of activity, to amoxicillin (Evidence Table 14 and Figure 12). First, the agents of our comparative arms were heterogeneous. By and large, they are all substantially more expensive than amoxicillin or folate inhibitors. However, their antibacterial profile is not identical. Some of the newer agents have a very broad spectrum that should readily encompass β-lactamase producing strains of H. influenzae and M. catarrhalis (e.g., amoxicillin-clavulanic acid, cefuroxime axetil, and cefixime). Others, such as the macrolides (azithromycin, clarithromycin, and roxithromycin) are not exceptional in this respect. The pooled failure rate in patients treated with amoxicillin was low (11 percent; 95 percent CI, 8-14), and the further decrease in clinical failures with broad-spectrum antibiotics was not clinically important. Treating 100 patients with amoxicillin would lead to only 0.85 more failures (95 percent CI, 3.1 more to 1.4 fewer failures).

Similar results were obtained through analysis of the failure data of newer antibiotic vs. folate inhibitors (Evidence Table 15 and Figure 13); treatment failures occurred at an essentially identical rate with other antibiotics as with folate inhibitors. However, data on folate inhibitors were limited and of poor quality. An analysis with respect to cures gave comparable results with risk ratios close to one, both for any other antibiotic vs. amoxicillin and any other antibiotic vs. folate inhibitors. The risk differences of clinical cure using amoxicillin or folate inhibitors compared with other antibiotics were not clinically important (3.2 percent [95 percent CI, -1.5 to 7.8 percent], and 1.2 percent [95 percent CI, -10 to 12.4 percent], respectively). The results were comparable when a trial comparing penicillin with azithromycin was added to the amoxicillin comparisons: risk ratio for clinical cures is 1.05 [95 percent CI, 0.99 to 1.11], and risk ratio for clinical failures is 0.82 [95 percent CI, 0.62 to 1.11].

There was no statistically significant heterogeneity of treatment effects in the amoxicillin comparisons; however, there was some evidence of heterogeneity between the studies that compared folate inhibitors with other antibiotics (p=0.09 for clinical cures; p=0.18 for clinical failures), possibly because trimethoprim-sulfamethoxazole seemed less effective than pivampicillin/pivmecillinam in one study (Osman and Menday, 1983).

A number of sensitivity analyses showed that the results were similar when the selection of studies was restricted according to various criteria (see Attachment B). In all sensitivity analyses, there was an estimated 11 to 20 percent risk reduction in clinical failures with other antibiotics over amoxicillin, which was not statistically significant, possibly because of the small numbers of patients. Still, this reduction was clinically negligible (less than 1 failure averted per 100 patients). Sensitivity analyses were less informative for folate inhibitors because the data were sparse.

Radiographic and bacteriologic data were not available for many trials (Evidence Tables 14 and 15). Rates of radiographic failures within 48 hours of the end of treatment were not significantly different between patients treated with other antibiotics and patients treated with amoxicillin, penicillin, or folate inhibitors. Likewise, rates of bacteriologic failure were not significantly different between patients treated with newer antibiotics and those treated with amoxicillin or folate inhibitors, although the majority of samples were obtained using nasal swabs, and the data therefore are not reliable.

4b. What evidence do these comparative studies provide regarding side effects?

The common side effects of amoxicillin and folate inhibitors are well known: skin rash, diarrhea, and gastrointestinal disturbances. There is no evidence suggesting that the side effects of these antibiotics are different for patients with acute bacterial rhinosinusitis.

There was no significant difference between regimens in the rate of withdrawal from treatment, either between other antibiotics and amoxicillin or between other antibiotics and folate inhibitors. In some instances, rash or gastrointestinal side effects led to discontinuation of the medications, whereas in other cases they were reported as minor adverse events (Evidence Tables 19 and 20). The reported adverse reactions numbered under 20 percent in the majority of trials. However, in a few studies, the percentages were higher (up to 50 percent in one study, Lindbaek, Hjortdahl, Johnsen, 1996a), mostly reported as gastrointestinal related. With some agents (e.g., macrolides, quinolones, tetracyclines), drug interactions may be a problem. Adverse reactions from interactions generally were not specifically identified, perhaps in part because most studies did not standardize the use of ancillary medications.

5a. Are there data to support the use of other types of treatments for acute rhinosinusitis and acute bacterial sinusitis, specifically: decongestants, steroids, antihistamines, drainage, flushing, others?

Search of the literature revealed few reliable studies regarding ancillary treatment of acute bacterial sinusitis. Ten randomized controlled trials were found (Evidence Table 22). Several different medication types used in these studies included proteolytic enzyme (bromelain), -adrenergic agonists (xylometazoline, oxymetazoline), mucolytic agents (bromhexine), glucocorticoids (budesonide, flunisolide, and antihistamine (loratadine). In most of the studies, these medications were used in conjunction with antibiotics. Although some of the studies tested different mechanisms for administering the treatments (e.g., nasal bellow vs. nasal spray; Wiklund, Stierna, Berglund, et al., 1994), no randomized trials comparing other nondrug treatments (e.g., sinus irrigation) were identified. The studies were carried out in both general medical and specialty clinics. Several studies combined patients with both acute and chronic sinusitis in the study population. Two of the studies were carried out in pediatric populations (Barlan, Erkan1, Bakir, et al., 1997; McCormick, John, Swischuk, et al., 1996).

Results of the randomized controlled studies of ancillary treatments are listed in Evidence Table 24. Three studies looked at bromelain (a proteolytic enzyme) as compared with placebo for patients who were also given antibiotics. These studies reported the same commercial source for bromelain, although the measured outcomes differed somewhat, and results therefore were not combined. All three studies reported higher positive outcomes in the patients receiving bromelain, although it was statistically significant (95 percent confidence level) only in one study (Seltzer, 1967). The remainder of the ancillary therapy studies looked at different treatments and reported varying efficacy.

5b. What is the efficacy of antibiotics compared with that of other types of treatment?

None of the randomized trials of ancillary treatments compared antibiotic to nonantibiotic therapy. In most studies, patients were treated with 7 to 21 days of antibiotics in conjunction with the assessed ancillary treatments. One study did not give antibiotics to all patients (Lewison, 1970) but reported that antibiotics were given to patients in either trial arm when bacterial infection was "obvious."

5c. What evidence do any comparative studies provide regarding side effects?

Three of the ten ancillary treatment trials reported patients discontinuing treatment because of major adverse reactions. Two of the studies assessed steroid preparations (Barlan, Erkan, Bakir, et al., 1997; Melzer, Orgel, Backhaus, et al., 1993); the third assessed a combination of drugs (McCormick, John, Swischuk, et al., 1996). In all studies, however, discontinuation was the result of adverse reactions to the antibiotics, not to the ancillary medications. Among the studies that provided information, two reported patients with minor side effects (Evidence Table 24).