Amoxicillin-clavulanate ineffective for suspected acute sinusitis

Journal of Family Practice, Dec, 2003 by Eric A. Jackson

Bucher HC, Tschudi P, Young J, et al. Effect of amoxicillin-clavulanate in clinically diagnosed acute rhinosinusitis: a placebo-controlled, double-blind, randomized trial in general practice. Arch Intern Med 2003; 163:1793-1798.

* PRACTICE RECOMMENDATIONS

Amoxicillin-clavulanate was no more effective than placebo in quickly relieving symptoms in patients diagnosed clinically with acute sinusitis in a general practice setting. It was, however, much more likely to cause diarrhea. Because most patients will improve spontaneously, antibiotics should be reserved for patients with prolonged symptoms. An inexpensive, narrow-spectrum drug such as amoxicillin is a good initial choice.

* BACKGROUND

Fifty percent of patients with acute sinusitis will have a viral infection and will not benefit from antibiotic therapy. This study evaluated the effect of a broad-spectrum antibiotic on adults with acute sinusitis diagnosed clinically in general practice.

* POPULATION STUDIED

The investigators enrolled 252 adult patients over 4 winter seasons, recruited from general practices and the internal medicine and otolaryngology outpatient clinics of a university hospital in Basel, Switzerland. Clinical diagnosis was based on a history of repeated purulent nasal discharge and maxillary or frontal sinus pain for at least 48 hours but less than 1 month. Presence of pus under rhinoscopy was an initial inclusion criterion, but was dropped after the first winter season. After randomization, treatment groups were relatively well-matched for characteristics that would affect the outcomes of interest.

Patients were excluded if they were aged <18 years, pregnant, breastfeeding, immuno-suppressed, allergic to amoxicillin-clavulanate, had an upper respiratory tract infection or used antibiotics for any reason within 4 weeks, or had malformation of the nasal cavities or pharynx.

* STUDY DESIGN AND VALIDITY

Eligible patients were consecutively enrolled and randomly assigned (with concealed allocation) to receive amoxicillin 875 mg/clavulanic acid 125 mg or matching placebo twice daily for 6 days. Patients also received a topical nasal decongestant and acetaminophen, and were allowed to use steam inhalation. Central randomization was stratified by treatment site in blocks of 6.

At baseline and day 7, physicians performed a focused clinical exam and recorded sinusitis-related symptoms and the number of days during which sinusitis restricted activities at home or work. At these times, patients also completed a questionnaire on their symptoms and any adverse effects from antibiotics. Patients also rated severity of symptoms on a 10-point Likert scale. A study nurse interviewed patients by telephone on days 14 and 28, asking about sinusitis-related symptoms and adverse effects as well as visits to other physicians.

Overall the methodology was good: it included triple-blinding (physician, study nurse, and patient) with concealed allocation (determined centrally by a statistician not involved with the analysis) and intention-to-treat analysis. The study was adequately powered to detect a 50% increase in cure rate at day 7 after making assumptions about spontaneous cure rates for patients with and without bacterial sinusitis. Weaknesses included dropping "pus under rhinoscopy" as an inclusion criterion after the first year and the need for 2 definitions of the primary outcome.

* OUTCOMES MEASURED

The primary outcome was time to cure, defined as no days since the previous visit or interview during which symptoms restricted activity at home or work. Because 10% of recruited patients reported no restriction on activities, analysis of time to cure was repeated using a 10-point visual analog scale to describe degree of restriction on activity. Secondary outcomes included number of days during which symptoms restricted activity and frequency of adverse effects of antibiotics.

* RESULTS

Using either definition of cure, there was no difference in the primary outcome (time to cure) between treatment groups. Cure rates for amoxicillin-clavulanate and placebo at 1 week (29.8% and 30.7%) and 2 weeks (76.6% and 74.0%) were similar for the 2 groups. The time to cure was also similar between antibiotic-treated and placebo-treated patients. There was no significant difference between the 2 treatment groups in the secondary outcome of number of days of restricted activity. However, at 1 week patients who took amoxicillin-clavulanate were almost 4 times more likely to have diarrhea (odds ratio=3.89; 95% confidence interval, 2.09-7.25).

Eric A. Jackson, PharmD, Department of Family Medicine, University of Connecticut School of Medicine and Saint Francis Hospital and Medical CentER, Hartford. E-mail: ejackson2@stfranciscare.org.