Low-value use of procalcitonin testing after introduction at an academic medical center

Background: Procalcitonin (PCT) is a biomarker that correlates with the presence of bacterial infection in certain clinical scenarios; it has increasingly accepted indications for antibiotic stewardship. Most sources suggest that antibiotics should be considered for patients with suspected infection and PCT values ≥ 0.25-0.5 ng/ml; conversely, clinicians should consider withholding antibiotics for patients with PCT values < 0.25. Shortly after introduction of the PCT assay at our large academic medical center, we sought to examine prevalence and appropriateness of use.

Methods: Rapid-turnaround PCT assay was introduced at our 808-bed tertiary care academic medical center in April 2017. Three months after the test became available, we performed a retrospective chart review of all inpatients with at least 1 PCT assay performed during their hospital stay. Data on test indication, antibiotic use, and PCT value was abstracted. Indications for testing were classified as use for either diagnosis or de-escalation.

Antibiotic de-escalation was defined as discontinuation of antibiotics when PCT values dropped below 0.25. Antibiotic use was defined as concordant with PCT if antibiotics were started for values ≥ 0.25 or withheld for values < 0.25; otherwise, antibiotic use was considered discordant. In addition, reviewers made a clinical judgment as to whether the application of the PCT test added value to the patient care. To be deemed high-value, the reviewers sought chart evidence that the treating team applied the test result to the care of the patient in a manner consistent with current guidelines and evidence.

Results: Data from 205 unique patient encounters between July and September 2017 was abstracted, encompassing 374 separate PCT tests. 50.0% of initial PCT values were ≥ 0.25, with 32.2% > 0.5. Mean (SD) initial PCT level was 3.76 (21.43), with a range of 0.02-247; median initial PCT was 0.23. PCT was overwhelmingly used for diagnosis (76.2%) rather than de-escalation (8.4%); the remainder of tests were for unclear indications. Respiratory symptoms (38.1%) and sepsis (33.5%) were the dominant clinical indications for testing. In 36 cases (17.5% of the sample), PCT was used for indications lacking established evidence, such as cellulitis, UTI, leukocytosis, gastrointestinal infections, and nonspecific infectious syndromes.

In our sample, 31.7% of patients had antibiotic use discordant with the PCT value; 86.2% of the discordant use involved patients continued on antibiotics despite a low PCT. Based on reviewers’ clinical judgment, 62.2% of the PCT use was felt to be low-value.

Conclusions: Low-value and non-evidence based use of serum PCT testing was frequent shortly after its introduction at a large tertiary-care academic medical center. As the availability of this biomarker grows more prevalent, attention to appropriate implementation strategies is important to ensure high-value application.