Acute Sinusitis—To Treat or Not to Treat?

Acute sinusitis is a common clinical problem in general practice. Approximately 0.5% to 2% of all episodes of the common cold develop into a purulent sinusitis; and in European general practice, 1% to 2% of all patient visits to physicians are for suspected acute sinusitis.¹ In the United States, there is a clear discrepancy between the patient's perception and the physician's diagnosis. “I have a sinus problem” is one of the most common reasons for clinical encounters in the United States and accounts for 25 million office visits per year.²

Despite randomized controlled trials showing lack of benefit of antibiotics for clinically diagnosed acute sinusitis and a number of rather strict guidelines for the use of antibiotics, acute sinusitis is still one of the most common reasons that clinicians prescribe antibiotics to adults. In the United States and the United Kingdom, more than 90% of adults seen for acute sinusitis in general practice settings receive antibiotics,³ whereas in Holland and Scandinavia, the proportion is lower, 68% to 80%.⁴ International guidelines⁵ for the treatment of acute sinusitis support the lack of effectiveness of antibiotics for clinically diagnosed sinusitis, including one guideline for patients with sinusitis confirmed with a reference standard (computed tomography, x-ray, or culture) that recommends antibiotics in cases with a duration longer than 7 days and another guideline based on patients with a clinical diagnosis that recommends symptomatic treatment only.

Antibiotic treatment of acute sinusitis corresponds with considerations regarding the limited benefit of antibiotic treatment of other common acute respiratory tract infections such as acute otitis media (especially in children), sore throat, and acute bronchitis.^{1 - 2 ,6 - 8} A number of studies have demonstrated a limited effect of antibiotics for these conditions, and guidelines have developed with more conservative treatment recommendations. This has been reinforced by the development of resistant strains of the most common respiratory pathogens, such as pneumococci, Haemophilus influenzae, group A streptococci, and Moraxella. One study reported a 10-fold increase in resistant pneumococci among various European countries and demonstrated a clear association with increased use of antibiotics.⁹

In this issue of JAMA, Williamson and colleagues¹⁰ report their findings from an important, large randomized controlled trial involving patients with a clinical diagnosis of acute sinusitis. The authors used a factorial design including either amoxicillin or budesonide (a topical steroid spray), or both, and concluded that neither regimen was better than placebo for resolving symptoms. This study is a useful contribution to the evidence on the effect of antibiotics and topical steroids in treating sinusitis but has some limitations. First, an objective reference standard was not used. Patients were included who had at least 2 of 4 Berg and Carenfeldt criteria,¹¹ which have not been validated in a primary care setting. Second, the positive likelihood ratio of 6.75 is fairly high, but depends on a high pretest probability of patients with purulent sinusitis. With a possible pretest probability of 20%, the proportion of purulent sinusitis would be 65%, leaving 35% with a possible viral infection. Therefore, it cannot be ruled out that a number of viral cases were included in the study. Third, the recruitment of patients for each clinician was low, around 1 case per primary care physician during 1 year, whereas most primary care physicians might be expected to see as many as 50 cases of sinusitis per year. This may imply a selection bias and limits the external validity of the study. On the other hand, the use of only clinical features as inclusion criteria increases the external validity.

The study by Williamson et al¹⁰ has implications for clinicians who treat patients with acute sinusitis–like symptoms. This study reinforces the lack of benefit from antibiotics shown in a number of other studies that recruited patients based on clinical symptoms and findings.^{12 - 14} Most patients with acute purulent sinusitis recover without antibiotic treatment, as was also observed in a study that used sinus computed tomography scans as the diagnostic standard.¹⁵ But some patients with sinusitis are more ill than others with fever, malaise, and deteriorated general condition. These patients still are in need of antibiotics, although they are relatively uncommon in general practice. So far there is no reliable way to distinguish viral sinusitis from bacterial sinusitis in the general practice setting, and a point of care test that could single out patients who could benefit from antibiotic treatment is not available. In a study using C-reactive protein as a point of care test in patients with a high degree of sinus pain,¹⁶ patients with an increased level of C-reactive protein who were treated with penicillin V had more rapid resolution of pain than those treated with placebo. Another study demonstrated that use of C-reactive protein reduced antibiotic prescribing for acute sinusitis from 78% to 59% of cases.¹⁷

The evidence on use of topical steroids in acute sinusitis is conflicting.¹⁸ A study involving patients with recurrent sinusitis concluded that the combination of antibiotics and steroids may provide benefit, although some patients included in this study may have had concomitant allergic rhinitis.¹⁹ In a large study in patients with sinusitis-like symptoms, steroids were more effective than placebo; however, the clinical significance of the size of the treatment effect was modest, at best.²⁰

Delayed prescribing of antibiotics for respiratory tract infections has been effective in reducing unnecessary prescription of antibiotics in some studies. In an English study,²¹ only 30% of patients with sore throat actually picked up their prescription for an antibiotic. For many patients with suspected acute sinusitis who had symptom duration of less than 1 week and mild symptoms, a delayed prescription may be useful to reduce unnecessary antibiotic use.

Although the study by Williamson et al¹⁰ has demonstrated that patients with a clinical diagnosis of sinusitis do not benefit from treatment with an antibiotic or a topical steroid, there may be subgroups that might benefit from either. This issue may be better studied in a large meta-analysis with individual patient data, such as the meta-analysis by Rovers et al,²² which included children with acute otitis media. This meta-analysis demonstrated that children younger than 2 years with bilateral otitis media or with otorrhea benefited from antibiotic treatment. Such studies may be informative for adults with acute sinusitis based on clinical diagnosis if subgroups of patients who may benefit could be identified. In the meantime, cautious use of antibiotics in the general practice setting for patients with sinusitis is warranted.