Clinical Utility of Procalcitonin on Antibiotic Stewardship: A Narrative Review

A machine learning strategy to mitigate the inappropriateness of procalcitonin request in clinical practice

Aim

The aim of this study was to develop machine learning (ML) models to mitigate the inappropriate request of Procalcitonin (PCT) in clinical wards.

Material and methods

We built six different ML models based on both demographical data, i.e., sex and age, and laboratory parameters, i.e., cell blood count (CBC) parameters, inclusive of monocyte distribution width (MDW), and C-reactive protein (CRP). The dataset included 1667 PCT measurements of different patients. Based on a PCT cut-off of 0.50 ng/mL, we found 1090 negative (65.4%) and 577 positive (34.6%) results. We performed a 70:15:15 train:validation:test splitting based on the outcome.

Results

Random Forest, Support Vector Machine and eXtreme Gradient Boosting showed optimal performances for predicting PCT positivity, with an area under the curve ranging from 0.88 to 0.89.

Conclusions

The ML models developed could represent a useful tool to predict PCT positivity, avoiding unusefulness PCT requests. ML models are based on laboratory tests commonly ordered together with PCT but have the great advantage to be easy to measure and low-cost.

Reduced Reliability of Procalcitonin (PCT) as a Biomarker of Bacterial Superinfection: Concerns about PCT-Driven Antibiotic Stewardship in Critically Ill COVID-19 Patients-Results from a Retrospective Observational Study in Intensive Care Units

BACKGROUND

The aim of this study was to assess whether procalcitonin levels is a diagnostic tool capable of accurately identifying sepsis and ventilator-associated pneumonia (VAP) even in critically ill COVID-19 patients.

METHODS

In this retrospective, observational study, all critically ill COVID-19 patients who survived for ≥2 days in a single university hospital and had at least one serum procalcitonin (PCT) value and associated blood culture and/or culture from a lower respiratory tract specimen available were eligible for the study.

RESULTS

Over the research period, 184 patients were recruited; 67 VAP/BSI occurred, with an incidence rate of 21.82 episodes of VAP/BSI (95% CI: 17.18-27.73) per 1000 patient-days among patients who were included. At the time of a positive microbiological culture, an average PCT level of 1.25-3.2 ng/mL was found. Moreover, also in subjects without positive cultures, PCT was altered in 21.7% of determinations, with an average value of 1.04-5.5 ng/mL. Both PCT and PCT-72 h were not linked to a diagnosis of VAP/BSI in COVID-19 patients, according to the multivariable GEE models (aOR 1.13, 95% CI 0.51-2.52 for PCT; aOR 1.32, 95% CI 0.66-2.64 for PCT-72 h).

CONCLUSION

Elevated PCT levels might not always indicate bacterial superinfections or coinfections in a severe COVID-19 setting.

Misleading Procalcitonin in Patients With Staphylococcus aureus Bacteremia: A Report of Two Cases

Procalcitonin (PCT) is an important biomarker for bacterial infection with a high negative predictive value. It is almost always positive in patients who are bacteremic with pathogenic bacteria. Here, we report two cases of Staphylococcus aureus bacteremia, where PCT levels were unexpectedly negative. This uncommon occurrence challenges the assumption of PCT's infallibility as a diagnostic marker in patients with true bacteremia. The first case is a 55-year-old woman with no past medical issues who presented with one week of generalized weakness and two days of fever and chills. Though her white blood cell (WBC) count and c-reactive protein (CRP) were elevated, PCT was normal, with no apparent source of infection, and hence antibiotic differed. However, her blood cultures returned positive for methicillin-resistant Staphylococcus aureus (MRSA). The patient was started on vancomycin and discharged on daptomycin, she responded appropriately and improved. The second case is an intravenous (IV) drug user, a 40-year-old woman, who presented with septic arthritis and osteomyelitis involving the right hip. She had blood cultures positive for methicillin-susceptible Staphylococcus aureus (MSSA); however, a PCT check on the day of positive blood cultures and various occasions subsequently was normal. These two cases remind us that we cannot over-rely on one test to rule out bacterial infection and should consider the whole clinical picture. They highlight the need for vigilance among clinicians that PCT can rarely be negative in cases of true bacteremia in spite of its high negative predictive value. Physicians and antibiotic stewardship programs should be cautious and aware of this potential pitfall when utilizing PCT as a diagnostic tool.

Plasma Proprotein Convertase Subtilisin/Kexin Type 9 (PCSK9) as a Possible Biomarker for Severe COVID-19

Proprotein convertase subtilisin/kexin type 9 (PCSK9) reduces low density lipoprotein (LDL) uptake, leading to increased plasma levels of LDL. In addition, PCSK9 has been implicated in inflammation independently of the effects on cholesterol metabolism. The current analysis showed that our 156 patients with systemic inflammatory response syndrome (SIRS) or sepsis had higher plasma PCSK9 levels in contrast with the 68 healthy controls. COVID-19 sepsis patients had increased plasma PCSK9 levels in comparison to sepsis patients not infected by SARS-CoV-2. For further analysis, patients were divided in two groups based on COVID-19. In both sub-cohorts, plasma PCSK9 levels did not correlate with C-reactive protein, leukocyte count, and procalcitonin. Plasma PCSK9 levels of both patient groups did not significantly differ among SIRS/sepsis patients with and without dialysis and patients with and without ventilation. Furthermore, vasopressor therapy was not significantly associated with altered plasma PCSK9 levels. In the non-COVID-19 SIRS/sepsis group, patients with Gram-negative and Gram-positive infections had similar plasma PCSK9 levels as patients without a detectable pathogen in their blood. In conclusion, the current study suggests PCSK9 as a possible biomarker for COVID-19, but this needs to be validated in larger cohorts.

Plasma Chemerin Is Induced in Critically Ill Patients with Gram-Positive Infections

Chemerin is a chemoattractant protein abundantly expressed in hepatocytes. Chemerin exerts pro- and anti-inflammatory effects and acts as a pro-resolving protein. Chemerin levels are low in patients with liver cirrhosis and are increased in sepsis. The aim of this study was to identify associations between plasma chemerin levels and underlying diseases as well as causes of severe illness. The cohort included 32 patients with liver cirrhosis who had low systemic chemerin, and who were not considered for further evaluation. Plasma chemerin levels were similar between the 27 patients with systemic inflammatory response syndrome (SIRS), the 34 patients with sepsis and the 63 patients with septic shock. Chemerin in plasma correlated with C-reactive protein and leukocyte count but not with procalcitonin, a clinical marker of bacterial infection. Plasma chemerin did not differ among patients with and without ventilation and patients with and without dialysis. Vasopressor therapy was not associated with altered plasma chemerin levels. Infection with severe acute respiratory syndrome coronavirus 2 had no effect on plasma chemerin levels. Baseline levels of plasma chemerin could not discriminate between survivors and non-survivors. Notably, Gram-positive infection was associated with higher chemerin levels. In summary, the current study suggests that plasma chemerin might serve as an early biomarker for the diagnosis of Gram-positive infections in patients with sepsis.