Broad Versus Narrow Spectrum Antibiotics For Sinusitis

Journal of Family Practice, Dec, 1998 by Lynda Montgomery, Vance Brown

DeFerranti SD, Ioannidis JP, Lau J, Anninger WV, Barza M. Are amoxicillin and folate inhibitors as effective as other antibiotics for acute sinusitis? A meta-analysis. BMJ 1998; 317:632-7

Clinical question Are antibiotics indicated for acute, uncomplicated sinusitis, and if so, are broad-spectrum agents more effective than amoxicillin or a folate inhibitor?

Background Acute sinusitis is a common infection often treated with antibiotics and decongestants. While the most common pathogens are usually sensitive to narrow-spectrum antibiotics like amoxicillin and folate inhibitors (the most widely used is trimethoprim-sulfamethoxizole [TMP-SMX]), the use of more broad-spectrum antimicrobials has been advocated. This study reviewed randomized trials comparing narrow-spectrum antibiotics with placebo or more broad-spectrum agents.

Population studied The authors performed a meta-analysis of 27 trials with a total of 2717 adult and pediatric patients. Trials compared amoxicillin or a folate inhibitor with another antibiotic or placebo; patients were randomly assigned to treatment arms; and patients had acute sinusitis or an acute exacerbation of chronic sinusitis.

Study design and validity Trials were identified through a MEDLINE search, manual review of the Excerpta Medica, abstracts from recent conferences on antimicrobial therapy, and the references of published studies identified. No language restriction was applied. The quality of the trials was assessed by subject-specific criteria (ie, how clearly the clinical outcomes were specified) and general criteria using a standard scale for clinical trials (the Jahad scale). Pooled estimates of treatment effect were calculated using both fixed effects and random effects models, and appropriate sensitivity analyses were performed.

Outcomes measured The primary outcomes were clinical cure, improvement, or failure assessed at 48 hours after cessation of therapy. Additionally, data on radiographic cure and bacteriologic cure (by nasal swab or antral puncture) were abstracted from trials that included them. Both radiographic and bacteriologic cure were appropriately deemed secondary outcomes because their use is uncommon in primary care practice.

Results Of 80 trials of antibiotic therapy for acute sinusitis, only 27 met eligibility criteria for the meta-analysis. Six studies were placebo-controlled, 11 were double-blinded, and 6 were single-blinded (5 were blind to the investigator). The sample sizes ranged from 14 to 323 patients; only 14 studies had a Jahad scale score of 3 or higher out of a possible 5. Using pooled data, the authors determined that using any antibiotic was more effective than placebo (the risk ratio for clinical failure when taking antibiotics was .54; 95% CI, .37 - .79). However, 69% of placebo-treated patients improved with time. Though not statistically significant, trials that reported the most patient improvement regardless of treatment with antibiotics or placebo were also those that tended to use broader inclusion criteria. In trials comparing amoxicillin with other antibiotics, there were no differences in cure or failure rate (absolute risk difference for clinical cure 3.2%; 95% CI, -1.5% to 7.8%). Similarly, among trials that compared folate inhibitor drugs with broad-spectrum agents, no statistically significant differences were seen (absolute risk difference for clinical cure 1.2%; 95% CI, -10% to 12.4%). The authors noted that the pooled sample size was small for comparisons of narrow and broad-spectrum antibiotics, limiting the ability of the trials to demonstrate treatment differences. Since treatment outcomes were pooled across different antibiotics, a superior treatment effect of a single agent compared with amoxicillin or TMP-SMX could be hidden. However, publication bias, lack of consistent blinding, and other biases related to the poor quality of the studies would tend to cause overestimation of the advantages of newer antibiotics.

Recommendations for clinical practice The lack of large, well-designed, double-blinded trials available for this meta. analysis highlights the need for additional research into antibiotic therapy for acute sinusitis. Nevertheless, a few conclusions that are helpful in day-to-day practice can be drawn from this study. It is common for patients with symptoms of acute sinusitis to improve spontaneously without antibiotics. If the physician's goal is to minimize the chance for failure and maximize the chance for cure, antibiotic therapy averts I clinical failure for every 7 or so patients treated. When choosing an antibiotic, the published evidence supports the use of inexpensive, narrow-spectrum agents like amoxicillin or TMP-SMX as a first choice.

Lynda Montgomery, MD
Vance Brown, MD
University of Massachusetts Medical School
Worcester
E-mail: vance.brown@banyan.ummed.edu