In response to concerns about the unnecessary use of antibiotics, researchers have sought to identify simple, non-specific tests to help clinicians determine which patients require antibiotics and which do not. Indeed, such tests could facilitate challenging management decisions, which currently rely predominately upon a thorough clinical assessment. To compound the challenge, patients often have pre-conceived ideas about whether or not they require antibiotics – which may inappropriately influence clinicians’ decisions. Adding greater objectivity to the assessment by incorporating more laboratory testing could help make these decisions easier for clinicians and could provide clearer justification to patients.
At first blush, a study evaluating the role of C-reactive protein (CRP) testing for guiding antibiotic decisions published earlier this summer in the New England Journal of Medicine might seem to represent a breakthrough. In the study, 653 British patients presenting to their primary care clinician with an acute chronic obstructive pulmonary disease (COPD) exacerbation were randomized to receive either a usual assessment or CRP-guided care. In the CRP group, clinicians used the result of a point-of-care CRP test to help determine whether or not to prescribe an antibiotic (clinicians were informed that antibiotics are unlikely to be beneficial for patients with a CRP level <20 mg per liter; for those with levels 20-40 antibiotics may be helpful; and for patients with CRP levels >40, antibiotics are likely to be beneficial). In the control group, clinicians determined the need for antibiotics based on a clinical assessment alone.
In the four weeks following randomization, 59.1% of patients assigned to the CRP group received antibiotics vs 79.7% in the usual care group. In addition, after two weeks, patients in the CRP group had significantly better scores on a questionnaire assessing COPD symptoms. Based on these findings, the authors of an accompanying editorial concluded, “[T]he findings from this study are compelling enough to support CRP testing as an adjunctive measure to guide antibiotic use in patients with acute exacerbations of COPD.”
While the observed reduction in antibiotic use, accompanied by improved clinical outcomes, might seem encouraging, we remain unconvinced that CRP-guided management of COPD exacerbations is a promising approach. Despite the fact that many patients with COPD exacerbations like the ones in this study do receive antibiotics in real-world situations, the data generally do not support this practice, except among the sickest patients. A 2012 Cochrane review, for example, concluded “Antibiotics for COPD exacerbations showed large and consistent beneficial effects across outcomes of patients admitted to an ICU. However, for outpatients and inpatients the results were inconsistent.” Given that the recent New England Journal of Medicine study focused on outpatients, two-thirds of whom were classified as having mild or moderate exacerbations, most of those in this study likely should not have been considered for antibiotics in the first place. In fact, it is likely that any intervention that led to a reduction in the use of antibiotics in this low-risk population would have been associated with excellent outcomes.
Furthermore, we worry that the use of a test like CRP may, paradoxically, justify antibiotics among some patients unlikely to benefit. Anecdotally, we have observed precisely this since the recent introduction of procalcitonin – another non-specific test used to guide antibiotic prescribing. In support of this contention, a 2018 New England Journal of Medicine study involving more than 1,600 patients presenting to the emergency department with respiratory symptoms found that the use of procalcitonin to guide antibiotic therapy did not lower antibiotic usage rates, nor did its use favorably impact outcomes. In both the CRP and procalcitonin studies, it is likely that these tests led clinicians to prescribe antibiotics to at least some patients who were unlikely to benefit.
For these reasons, we do not yet share the enthusiasm some experts have for CRP and procalcitonin to guide antibiotic management for respiratory illness. In our view, careful clinical assessment accompanied by microbiologic testing will lead to better antibiotic decisions. Perhaps in the future, research will demonstrate a role for these tests in truly borderline clinical situations. But for now, the best guides for antibiotic prescribing remain our clinical assessments, along with a clear understanding of the evidence base, which shows that antibiotics benefit only the sickest patients experiencing respiratory illness.
Michael Hochman, MD, MPH, directs the Gehr Center for Health Systems Science at the University of Southern California’s Keck School of Medicine. Pieter Cohen, MD, is a general internist at Cambridge Health Alliance in Somerville, Massachusetts, and associate professor of medicine at Harvard Medical School.
last updated 09.30.2019