Mrs. Mayer is a 73-year-old woman who comes to see you in outpatient clinic after 5 days of cough and sputum production. She’s been feeling short of breath and more fatigued than usual. Her vital signs are normal, but you hear wheezing and rhonchi in her lungs. She asks whether antibiotics would help her symptoms; how would you respond?
Every day across the United States, thousands of patients like Mrs. Mayer visit their doctors with acute lower respiratory tract infections, and many of them receive antibiotics for what are most likely self-limited viral infections. Making the right diagnosis still relies on clinical judgment, but more recently, procalcitonin has entered the scene as a way to identify bacterial infections from viral and other non-infectious etiologies.
Procalcitonin is a prohormone for calcitonin and is primarily produced in neuroendocrine tissue. In healthy individuals, procalcitonin levels are quite low, but in people with systemic inflammation, procalcitonin levels rise significantly. Previous studies have shown that higher levels of procalcitonin are associated with bacterial infections and that its use may reduce antibiotic prescribing for lower respiratory tract infections without an increase in adverse outcomes. However, these initial trials were relatively small and primarily conducted in Europe under strict protocols and researchers questioned whether the results were applicable to routine practice in the United States.
In the ProACT trial, investigators randomized 1664 adults who presented to 14 emergency departments in the United States with suspected acute lower respiratory tract infections and uncertain need for antibiotics to receive usual care or to undergo additional procalcitonin testing. In the procalcitonin group, clinicians received the results of initial and serial (if subsequently hospitalized) procalcitonin testing and guidelines on how to prescribe antibiotics based on procalcitonin levels (e.g., discouraged for ≤0.25 μg/L and recommended for >0.25 μg/L). Clinicians had autonomy with treatment decisions, and investigators implemented the approach primarily through educational outreach to simulate standard quality-improvement methods. The primary outcome — the number of antibiotic-days within 30 days after enrollment — did not differ significantly between the procalcitonin and control groups (mean, 4.2 vs. 4.3 days, respectively; P=0.87). The use of procalcitonin levels was also noninferior to usual care in terms of adverse outcomes (11.7% vs. 13.1%, P<0.001 for noninferiority) within 30 days.
Because the study was designed as a pragmatic evaluation of how procalcitonin would be used in the emergency department or inpatient settings, its strengths and limitations are evident in some other observations. First, to arrive at the study cohort, the investigators had to screen more than 8300 patients, and clinicians of approximately 1000 patients did not want to use procalcitonin in antibiotics decision making. Second, adherence to the procalcitonin protocol was only 65%; however, sensitivity analysis in which the intervention group was restricted to patients whose clinicians followed all guideline recommendations did not show a benefit from procalcitonin either. Third, overall antibiotics exposure was lower on average in this study than in previous studies.
Although it may seem disheartening to know that procalcitonin isn’t the magic bullet that will ease the dilemma of prescribing antibiotics for patients like Mrs. Mayer, the results of the ProACT trial suggest that good clinical judgment and judicious use of antibiotics is still the best way to care for most patients. It remains to be seen whether there are subgroups of patients with suspected lower respiratory tract infections who would truly benefit from the additional information of procalcitonin testing.
Browse more From Pages to Practice
Mike was a 2017-2018 NEJM Editorial Fellow and is a hospitalist at Beth Israel Deaconess Medical Center. He graduated from Harvard Medical School and completed his internal medicine training at BIDMC.