Evidence Report/Technology Assessment

Number 9

Prepared for:
Agency for Health Care Policy and Research

U.S. Department of Health and Human Services
2101 East Jefferson Street
Rockville, MD 20852
http://www.ahrq.gov

Contract No. 290-97-0019

Prepared by:
New England Medical Center, Boston, MA
Joseph Lau, M.D.
Principal Investigator

Deborah Zucker, M.D., Ph.D.
Eric A. Engels, M.D., M.P.H.
Ethan Balk, M.D.
Michael Barza, M.D.
Norma Terrin, Ph.D.
Deirdre Devine, M.Lit.
Priscilla Chew, M.P.H.
Thomas A. Lang, M.A.
David Liu, B.A.

AHCPR Publication No. 99-E016

March 1999

The Agency for Health Care Policy and Research (AHCPR), through its Evidence-based Practice Centers (EPCs), sponsors the development of evidence reports and technology assessments to assist public- and private-sector organizations in their efforts to improve the quality of health care in the United States. The reports and assessments provide organizations with comprehensive, science-based information on common, costly medical conditions and new health care technologies. The EPCs systematically review the relevant scientific literature on topics assigned to them by AHCPR and conduct additional analyses when appropriate prior to developing their reports and assessments.

To bring the broadest range of experts into the development of evidence reports and health technology assessments, AHCPR encourages the EPCs to form partnerships and enter into collaborations with other medical and research organizations. The EPCs work with these partner organizations to ensure that the evidence reports and technology assessments they produce will become building blocks for health care quality improvement projects throughout the nation. The reports undergo peer review prior to their release.

AHCPR expects that the EPC evidence reports and technology assessments will inform individual health plans, providers, and purchasers as well as the health care system as a whole by providing important information to help improve health care quality.

We welcome written comments on this evidence report. They may be sent to: Director, Center for Practice and Technology Assessment, Agency for Health Care Policy and Research, 6010 Executive Blvd., Suite 300, Rockville, MD 20852.

John M. Eisenberg, M.D.	Douglas B. Kamerow, M.D.
Administrator	Director, Center for Practice and Technology Assessment
Agency for Health Care Policy and Research	Agency for Health Care Policy and Research

The authors of this report are responsible for its content. Statements in the report should not be construed as endorsement by the Agency for Health Care Policy and Research or the U.S. Department of Health and Human Services of a particular drug, device, test, treatment, or other clinical service.

We thank our colleagues, Helena Varonen, Marjukka Mäkelä, Seppo Savolainen, and Esa Läärä, from Finland, and Jörgen Hilden from Denmark, for generously providing us with a copy of their manuscript on a meta-analysis of the diagnostic tests for acute sinusitis. We would like to give special thanks to Dr. Varonen for her assistance in extracting the necessary data from a Finnish language article.

We thank the technical experts representing the partner organizations for their most valuable contributions. They provided critical insights on this topic, reviewed an earlier draft of this report, and provided prompt assistance when requested. We also thank our patient representatives for taking time out to provide important insights from the patient's perspective in the clinical management of this condition.

We are very grateful to the British Medical Journal for permission to use in the evidence report substantial portions of the material, including text, tables, and figures, from a meta-analysis authored by several EPC members.

This report summarizes the published evidence on the diagnosis and treatment of community-acquired acute bacterial rhinosinusitis in children and adults. The performance of diagnostic tests and the efficacy of antibiotics and ancillary treatments were assessed.

Human studies of sinusitis were identified from MEDLINE (1966 to May 1998), technical experts, and bibliographies.

Acute bacterial rhinosinusitis was defined as bacterial infection of the paranasal sinuses and symptoms lasting up to 4 weeks. Only diagnostic studies that compared two or more tests were used. Only randomized controlled trials were used to assess treatment efficacy. Studies that met methodologic quality criteria were used in meta-analyses.

Data for meta-analyses were extracted in duplicate. Summary receiver operating characteristics curves were created from the meta-analyses of diagnostic tests. The random-effects model was used to estimate antibiotic efficacy. Decision and cost-effectiveness analyses were conducted to evaluate clinical management strategies.

Compared with sinus puncture, the reference standard for diagnosing acute bacterial rhinosinusitis, sinus radiography has moderate sensitivity (76 percent) and specificity (79 percent). Sinus ultrasonography has similar test characteristics, but the results are more variable and the procedure is not commonly used in the United States. Limited evidence suggests that diagnoses based on clinical criteria may be as accurate as those using sinus radiography. In a meta-analysis of six placebo controlled trials, antibiotics reduced the incidence of clinical failures by one-half (risk ratio [RR], 0.54; 95 percent confidence interval [CI], 0.37 to 0.79), although about two-thirds of the patients improved by 14 days without antibiotics. The risk of clinical failure did not differ significantly between amoxicillin (14 trials, RR, 0.85; 95 percent CI, 0.62 to 1.17) or folate inhibitors (e.g., trimethoprim/sulfamethoxazole) (nine trials, RR, 1.01; 95 percent CI, 0.52 to 1.97) and newer, more expensive antibiotics. No serious complications from lack of treatment were reported.

Some of the 10 trials assessing nonantibiotic, ancillary treatments reported statistically significant results. However, these trials included different treatments, often in conjunction with antibiotics, so further analyses were limited.

To minimize symptom duration, initial symptomatic treatment is the most cost-effective strategy up to a prevalence of 25 percent, the use of clinical criteria to guide treatment is most cost-effective for a prevalence between 25 percent and 83 percent, and empirical antibiotic treatment with amoxicillin or a folate inhibitor is cost effective only above the prevalence of 83 percent. Sinus radiography-guided treatment, as an initial management strategy for uncomplicated patients, is not cost effective at any prevalence.

The quality of the diagnostic and treatment trials reviewed was suboptimal. However, the results were generally consistent, sensitivity analyses showed the results to be robust, and the conclusions are consistent with the majority of expert opinions.

In primary care, where the prevalence of bacterial infection in patients presenting with symptoms suggesting acute rhinosinusitis is low to moderate, initial symptomatic treatment or the use of clinical criteria to guide treatment is the most cost effective approach for patients with uncomplicated infections. If antibiotics are given, amoxicillin or a folate inhibitor should be considered initially, as should the severity of the symptoms.

Future studies should use more rigorous diagnostic standards and clinical outcome measures. The optimal duration of antibiotic treatment, the role of patient preferences in clinical decisionmaking, and the issue of emerging antibiotic resistance also need to be addressed.

This document is in the public domain and may be used and reprinted without permission.

Suggested citation:

Lau J, Zucker D, Engels EA, Balk E, et al. Diagnosis and treatment of acute bacterial rhinosinusitis. Evidence Report/Technology Assessment No. 9 (Contract 290-97-0019 to the New England Medical Center). Rockville, MD: Agency for Health Care Policy and Research. March 1999.

Acute rhinosinusitis, viral or otherwise, is one of the most common infections in the United States. Millions of cases occur each year, affecting all age groups and all segments of the population. Although only a small percentage of these cases come to the attention of a physician, this high prevalence translates into high costs for individual health, work time lost, and medical expenditures. In 1992, in the United States, $200 million was spent on prescription cold medications for rhinosinusitis and more than $2 billion for over-the-counter medications.

In the majority of cases, inflammation of the paranasal sinuses (sinusitis) is accompanied by inflammation of the nasal passages (rhinitis); thus, the clinical condition often referred to as "sinusitis" is, in fact, rhinosinusitis: inflammation of the sinuses with concomitant inflammation of the nasal passages. In clinical practice, the focus is on patients in whom this rhinosinusitis results in clinical symptoms. Conditions that cause or predispose individuals to rhinosinusitis include infectious agents (bacteria, viruses, and fungi), allergic conditions (allergic rhinitis), anatomic abnormalities, systemic diseases (endocrine, metabolic, genetic), trauma, and noxious chemicals. The prevalence of rhinosinusitis resulting from each cause is unknown, although certain causes, such as viral infection, are more common. In some cases, the cause may be multifactorial (e.g., viral infection with bacterial superinfection).

Despite the common nature of rhinosinusitis, its management is controversial. Therapies are usually directed to alleviating or reducing symptoms, eradicating the underlying cause, or both. A major question is whether antibiotics should be used, and if so, which one? Because the premise of treatment with antibiotics is that bacterial infection will be eliminated, patients with bacterial rhinosinusitis need to be identified. In addition, other disease and patient characteristics, such as age and duration and pattern of illness, may help in distinguishing patient subgroups for more specific types of treatment (e.g., antibiotics to eradicate specific bacterial species). Because bacterial infection of the sinuses is potentially serious, the use of antimicrobials to prevent these complications is of interest. However, concern is increasing about the overuse or abuse of antibiotics, both for the individual, in terms of potential side effects and financial costs, as well as for society, in terms of cost and the development of antibiotic-resistant bacteria.

In this report, we summarize the evidence for the diagnosis and treatment of uncomplicated, community-acquired, acute bacterial rhinosinusitis in children and adults. We present evidence regarding the prevalence of this illness in both general primary care and subspecialty clinic settings. We analyzed the data from clinical studies that compared the performance of various diagnostic tests (including clinical examination criteria) for identifying patients with acute bacterial rhinosinusitis. We assessed randomized controlled trials that compared the treatment effects of antibiotics with those of placebo and the effects of inexpensive antibiotics, such as amoxicillin and folate inhibitors (e.g., trimethoprim/sulfamethoxazole), with those of newer, more expensive antibiotics (e.g., cephalosporins). We also collected evidence on ancillary therapies, such as decongestants, steroids, and sinus irrigation. Finally, we combined the evidence in a decision analysis and a cost-effectiveness analysis to compare clinical strategies in managing patients with acute bacterial rhinosinusitis to help translate the evidence into practice. Although sinusitis can include acute, recurrent, and chronic forms, this report focuses on the Agency for Health Care Policy and Research (AHCPR)-designated topic of acute sinusitis and more specifically, on community-acquired, acute bacterial rhinosinusitis.

The Evidence-based Practice Center (EPC) staff, along with a panel of technical experts, including representatives from four professional organizations (the American Academy of Family Physicians, American Academy of Otolaryngology-Head and Neck Surgery, American Academy of Pediatrics, and American College of Physicians), formulated the following questions to be addressed in this evidence report for acute sinusitis in children and adults.

1. What is the prevalence of bacterial infection in patients presenting with acute rhinosinusitis in primary care and specialty settings?

2. What is the diagnostic value of clinical features and imaging technologies for identifying acute rhinosinusitis and acute bacterial rhinosinusitis?

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

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

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

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

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

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

We systematically reviewed the literature for evidence addressing these questions. Prospective studies that compared two or more diagnostic tests were used to assess diagnostic test performance, and randomized controlled trials were used to assess treatment efficacy. We searched for English-language articles indexed in the MEDLINE database between 1966 to May 1998 using several sensitive search strategies for human studies on sinusitis. The titles, MeSH headings, and abstracts of the retrieved citations were manually screened to identify articles for retrieval. Technical experts were consulted, and bibliographies of retrieved primary studies, review articles, and published and unpublished meta-analyses on the diagnosis or treatment of acute rhinosinusitis were examined for additional references. A separate MEDLINE search for potentially useful foreign-language articles was also conducted to assess the magnitude of the bias caused by excluding foreign-language articles from the primary search strategy. Several studies published in other languages were included in our analyses.

Data from primary clinical studies that met inclusion criteria were extracted to develop evidence tables pertaining to the specified questions. A summary receiver operating characteristic (SROC) curve was constructed from a meta-analysis to assess the performance of clinical criteria and various imaging technologies commonly used to diagnose acute bacterial rhinosinusitis. Meta-analyses were also performed to pool the clinical outcomes of patients treated with and without antibiotics and to compare different individual and classes of antibiotics. Several subgroup analyses were performed to identify factors that may be related to treatment variations.

A decision analysis was performed from the patient's perspective to evaluate several diagnostic tests and treatment strategies for managing a patient presenting with the symptoms of acute bacterial rhinosinusitis. We also performed a cost-effectiveness analysis from the payer's perspective to estimate the cost-effectiveness of several common treatment strategies. We used both a single-time-point decision tree and a Markov process to model the clinical decisions, possible events, and clinical outcomes. The models used estimates from the evidence report's meta-analyses, primary studies, review articles, expert opinions, and consensus.

The overall methodologic quality and reporting of both diagnostic and treatment studies on this topic are poor. Few studies were conducted in North America. So few studies met strict diagnostic criteria (sinus puncture with bacterial culture) that we had to relax the criteria to have enough studies for meta-analyses (we accepted the investigators' diagnoses). Still, only 14 of 48 diagnostic test comparison studies and only 30 of the 74 randomized controlled trials on antibiotics met the revised criteria. For studies of children, only one diagnostic test study and two antibiotic treatment studies met the revised criteria for their respective meta-analysis. Although there is a pathophysiologic basis for differentially treating children and adults, the lack of evidence for children precluded making distinctions in diagnosis and treatment in these populations beyond inspection of the individual studies. Data on prevalence were obtained from the studies reviewed, although estimates from additional observational studies are also described.

1. What is the prevalence of bacterial infection in patients presenting with acute rhinosinusitis in primary care and specialty settings?

• Prevalence data for acute bacterial rhinosinusitis in the general population are sparse. The 1994 National Health Interview Survey report on chronic sinusitis estimated 35 million cases.
• The prevalence of acute sinusitis appears to be increasing, according to data from the National Ambulatory Medical Care Survey (from 0.2 percent of diagnoses at office visits in 1990 to 0.4 percent of diagnoses at office visits in 1995).
• Up to 38 percent of patients with symptoms of sinusitis in adult general medicine clinics may have acute bacterial rhinosinusitis. In otolaryngology practices, the prevalence was higher (50 to 80 percent). Between 6 and 18 percent of the children in the primary care setting presenting with upper respiratory infections may have acute bacterial sinusitis.

2. What is the diagnostic value of clinical features and imaging technologies for identifying acute rhinosinusitis and acute bacterial rhinosinusitis?

• Bacterial rhinosinusitis has been diagnosed from clinical criteria, sinus puncture with culture of the aspirate, sinus radiography, ultrasonography, and computed tomography.
• Although sinus puncture with culture is the diagnostic reference standard, it is rarely used because it is invasive and costly; it is not a practical routine procedure.
• A meta-analysis of six studies shows that sinus radiography has moderate sensitivity (76 percent) and specificity (79 percent) compared with the sensitivity and specificity of sinus puncture in the diagnosis of acute bacterial rhinosinusitis.
• Studies comparing sinus ultrasonography with puncture or sinus radiography were inconclusive in determining how well ultrasonography identifies patients with acute bacterial rhinosinusitis. The results of ultrasonography varied substantially, possibly because of differences in patient populations, ultrasonography techniques, or medical personnel involved in diagnostic testing.
• Limited evidence suggests that clinical criteria (i.e., the presence of three or four of the following symptoms: purulent rhinorrhea with unilateral predominance, local pain with unilateral predominance, bilateral purulent rhinorrhea, and the presence of pus in the nasal cavity) may have a diagnostic accuracy similar to that of sinus radiography.
• We found no studies comparing magnetic resonance imaging or endoscopy. The one randomized trial comparing computed tomography with sinus radiography was inadequately reported.

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

• More patients were cured, and cured earlier, when treated with antibiotics rather than placebo.
• About two-thirds of the patients receiving placebos recovered without antibiotics.
• Serious complications of rhinosinusitis, such as meningitis, brain abscess, and periorbital cellulitis, are rare, and none was reported in the clinical trials we examined.
• Most clinical trials have only short-term followup and report no data on relapse.

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

• Antibiotics are significantly more effective than placebo for treating acute bacterial rhinosinusitis, reducing the clinical failure rate by one-half (risk ratio [RR], 0.54; 95 percent confidence interval [CI], 0.37 to 0.79). Patients are cured more quickly and more often when treated with antibiotics compared with no treatment.
• Amoxicillin or folate inhibitors were as efficacious as the newer and more expensive antibiotics. The current evidence does not justify the use of the newer antibiotics for treating uncomplicated, community-acquired acute bacterial rhinosinusitis.

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

• About 4 percent of the patients in the amoxicillin arms of the clinical trials withdrew as a result of side effects, but this percent did not differ statistically from that in patients treated with other antibiotics. The data for folate inhibitors are more limited but similar.

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

• Ten randomized controlled trials evaluated ancillary treatment for rhinosinusitis. Meta-analysis was not possible because of the differences in treatments, diagnostic criteria, and outcomes measures among the studies, and because the concurrent and inconsistent use of antibiotics in many studies confounded statements about the efficacy of treatments.

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

• Many studies of antibiotic treatment also included ancillary therapies. However, these therapies were seldom standardized, which prevented an analysis of their benefits.

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

• Data from randomized controlled trials are insufficient to answer this question.

We conducted a cost-effectiveness analysis to compare four treatment strategies: (1) a sinus radiography-directed strategy, (2) the use of clinical criteria to guide treatment, (3) initial symptomatic (ancillary) treatment, and (4) routine empirical use of antibiotics, with either amoxicillin or a folate inhibitor. The result is essentially a "toss-up" in terms of symptom days for empirical, radiography-guided, and clinical-criteria-guided treatments. Symptomatic treatment alone provided fewer symptom-free days at all but the very lowest prevalence of acute bacterial rhinosinusitis. In terms of cost, the use of clinical criteria to guide treatment and initial symptomatic treatment is a toss-up at any prevalence. Empirical treatment is more costly at all but the highest range of prevalence. Radiography is considerably more costly at any prevalence. Initial symptomatic treatment is the most cost effective strategy, at prevalence of up to 25 percent, the use of clinical criteria to guide treatment is most cost effective for a prevalence between 25 and 83 percent, and empirical antibiotic treatment with amoxicillin or a folate inhibitor is cost effective only at prevalence greater than 83 percent. Sinus radiography is never a cost-effective strategy at any prevalence. The prevalence thresholds for various strategies are moderately sensitive to the severity of sinus symptoms as reflected in the utilities for computing quality-of-life adjustments.

At the prevalence of acute bacterial rhinosinusitis likely to be encountered in most primary care settings, a strategy of either initial symptomatic treatment or the use of clinical criteria to guide treatment is an effective and cost-effective approach for patients with uncomplicated cases. Given our finding that most patients' symptoms resolve without antibiotic treatment and that serious complications are rare, watchful waiting (before giving antibiotics) for 7 to 10 days after onset of "sinus" symptoms is a reasonable strategy. If antibiotics are to be given, amoxicillin or a folate inhibitor should be the initial choice. The severity of the patient's symptoms affected the utilities used in the decision models and thus may also need to be considered in the management decision.

• Many patients with acute rhinosinusitis are not seen by health care providers. The prevalence of this condition needs to be known to help distinguish those people requiring treatment with antibiotics from those not requiring antibiotics or further evaluation.
• Because of the developmental anatomical differences in children and adults, diagnostic and treatment studies should be conducted on pediatric populations.
• Future studies should also be dedicated to studies of patients with comorbidities (e.g., allergies, asthma, and human immunodeficiency [HIV] infection) that may influence the development, progression, and response to treatment of acute bacterial rhinosinusitis.
• Involvement of other sinuses other than maxillary sinuses needs to be studied.
• The diagnostic reference standard of sinus puncture with culture of aspirate is infeasible in routine practice, and most trials based diagnosis on other criteria. Alternative less invasive reference standard methods for diagnosing acute rhinosinusitis are needed.
• Future studies of clinical criteria (including risk scores), ultrasonography, and endoscopy with middle meatal sampling, ideally comparing them with sinus puncture in a variety of research and clinical settings, are needed to establish their diagnostic utility.
• The designs of future studies need to be improved. In particular, definitions of the populations to be treated, the test methods, and the criteria for diagnosis need to be more precise and investigators need to be masked.
• The role of antibiotic resistance in individual clinical decisionmaking needs to be clarified. More data are needed on patients with resistant organisms and their responses to therapies and on the association between laboratory and clinical resistance.
• Outcome measures need to be reassessed. In particular, assessing outcomes at different time points may better represent the differential effect of therapies. In addition to a better understanding of the connection between treatment and time to resolution of symptoms, increased knowledge regarding treatments and relapse rates or the potential development of recurrent sinusitis is also needed.
• In addition to better understanding of the connection between treatment and time to resolution of symptoms, there is a need for increased knowledge regarding treatments and relapse rates or the potential development of recurrent sinusitis.
• Standardization and focused evaluations of ancillary treatments are needed.
• The influences of several factors on patient-assigned utilities (patient-physician interactions, availability of "time for sickness," and variability of severity of episodes) need to be better understood when evidence is applied to clinical practice.