Can point-of-care tests reduce inappropriate antibiotic use?

Why do physicians prescribe antibiotics for viral infections when they ought to know better? An editorial from the Centers for Disease Control and Prevention in the August 1st issue of American Family Physician suggests several talking points that may persuade the patient who wants "Vitamin Z" (azithromycin) for his (or his child's) common cold and reviews evidence that delayed antibiotic prescriptions, effective communication strategies, and public commitments to use antibiotics appropriately reduce unnecessary antibiotic use. But what if these interventions aren't enough?

The rapid strep test helps to distinguish a viral sore throat from streptococcal pharyngitis. In patients with an intermediate pre-test probability, a negative rapid strep test lowers the post-test probability of strep enough for doctors (and most patients) to feel comfortable withholding antibiotics. Wouldn't it be nice if other point-of-care tests could effectively rule out bacterial infections and curb antibiotic prescribing in patients with acute respiratory symptoms?

Procalcitonin seems ready to go. A FPIN Clinical Inquiry in the July 1st issue of AFP evaluated the effects of a procalcitonin-guided antibiotic therapy algorithm on antibiotic use and clinical outcomes. A Cochrane review and meta-analysis of 14 randomized, controlled trials (RCTs) comparing procalcitonin-guided to standard care in European adults with acute respiratory infections found that patients in the procalcitonin group received 3.47 fewer days of antibiotic therapy with no differences in 30-day mortality or treatment failure. In a single RCT of 337 children presenting to pediatric emergency departments in Switzerland, patients in the procalcitonin group were as likely as the standard care group to receive antibiotic prescriptions, but received nearly 2 fewer days of therapy.

C-reactive protein (CRP): the jury is still out. A Cochrane for Clinicians examined the performance of point-of-care measurement of CRP on similar outcomes. Although treatment thresholds varied, most studies considered a CRP level of less than 20 mg per L to suggest a viral infection and no need for antibiotics. A Cochrane review of 6 RCTs conducted in primary care settings (mostly in adults) in Europe and Russia found that groups assigned to CRP-assisted evaluation were 22 percent less likely to receive antibiotic prescriptions for acute respiratory infections, with no differences in clinical improvement at day 7, complications, or mortality. However, Dr. Irbert Vega observed in the Practice Pointers that "the meta-analysis did not identify an optimal algorithm and therefore should be considered proof of concept until further research can be performed, including research in the U.S. population."

**

This post first appeared in a different form on the AFP Community Blog.