Clinical Question

Does point-of-care measurement of C-reactive protein (CRP) reduce inappropriate antibiotic prescribing for patients with acute respiratory infections?

Evidence-Based Answer

Point-of-care CRP testing used as an adjunct to a physician's clinical examination can modestly reduce antibiotic use. Measurement of CRP to guide antibiotic prescription does not appear to affect the duration of illness or recovery, although one study suggests that it increases the risk of hospitalization. The best algorithm is not known, although most state that a CRP level of less than 20 mg per L (190.5 nmol per L) suggests a viral infection. (Strength of Recommendation: B, based on inconsistent or limited-quality patient-oriented evidence.)

Practice Pointers

Acute respiratory infections are among the most common symptomatic reasons for visits to family physicians.^1,2 These predominantly viral infections are the most common indication for an antibiotic prescription, despite a lack of benefit for most patients.^2–5 An estimated 41 million unnecessary antibiotic prescriptions are written at a cost of $1.1 billion per year for noninfluenza viral respiratory infections.⁶ Guidelines already advocate the use of CRP to help determine the appropriateness of antibiotics in patients with lower respiratory infection.⁷

The authors of this Cochrane review examined the evidence for point-of-care biomarkers to guide antibiotic prescribing in primary care settings and found only studies of CRP. They identified six randomized controlled trials with 6,183 participants from primary care settings for this systematic review; the mean age of participants was 46 years, and 139 were children. CRP was generally not used if the clinician was confident about the decision to initiate or withhold antibiotic treatment. A variety of algorithms were used, with a CRP level of less than 20 mg per L suggesting a viral infection and no need for antibiotics. The studies were conducted in Europe and Russia between 1995 and 2013; two of the studies were directly supported by manufacturers of QuikRead CRP analyzers (Orion Diagnostica) and NycoCard Reader II (Nycomed Pharma). Overall the studies had a low to moderate risk of bias.

The primary outcome was the number of patients given an antibiotic prescription at the index consultation and at follow-up 28 days later. All studies showed a statistically significant reduction in the number of antibiotic prescriptions issued for acute respiratory infections when CRP was used to guide therapy (relative risk [RR] = 0.78; 95% confidence interval [CI], 0.66 to 0.92). Studies in which practices were randomized had a greater effect (number needed to treat = 6) than those in which individual patients were randomized (number needed to treat = 20), although there was significant variability between studies. The effect was maintained at day 28. No difference was found between groups for the number of patients with substantial improvement at day 7, and no deaths or serious complications were reported.

The number of patients in need of hospital admission at 28 days was based on a single study. Out of 30 hospitalizations in 4,264 patients, 22 hospitalizations occurred in the CRP groups vs. eight in the control group. The effect was no longer statistically significant after adjusting for whether patients or practices were randomized (RR = 2.45; 95% CI, 0.65 to 9.19). No data were available on which hospitalized patients did not initially receive antibiotic treatment or on their initial CRP levels. There were no differences in the number of patients requiring reconsultation at 28 days, the duration of acute respiratory infections, the number of satisfied patients, or the number of patients with substantial improvement at 28 days.

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 intervention promotes improved antimicrobial use by influencing prescribing practices consistent with the goal of antimicrobial stewardship. Current guidelines recommend a no-antibiotic prescribing policy with deference to case-by-case evaluation, and appropriate patient education for simple acute otitis media, sore throat, pharyngitis, tonsillitis, common cold, rhinosinusitis, and bronchitis.^3–5