Procalcitonin (PCT) can discriminate bacterial from viral systemic infections and true bacteremia from contaminated blood cultures. The aim of this study was to evaluate PCT diagnostic accuracy in discriminating Gram-positive, Gram-negative, and fungal bloodstream infections. A total of 1,949 samples from patients with suspected bloodstream infections were included in the study. Median PCT value in Gram-negative (13.8 ng/mL, interquartile range (IQR) 3.4–44.1) bacteremias was significantly higher than in Gram-positive (2.1 ng/mL, IQR 0.6–7.6) or fungal (0.5 ng/mL, IQR 0.4–1) infections (
Procalcitonin (PCT) is a 116-amino-acid peptide whose high levels are strongly associated with systemic bacterial infections [
Few studies are available in the literature on PCT utility in differentiating among Gram-negative, Gram-positive, or fungal bacteremias [
This prospective observational study was conducted using clinical and routine laboratory data collected from the Clinical Microbiology Unit of the General Hospital of Perugia, Italy, from March to September 2014.
Inclusion criteria were unselected consecutive blood samples for blood culture (BC) and PCT that, according to our hospital standard protocol, were collected simultaneously from each patient with suspected sepsis, defined according to Bone et al. [
PCT levels were measured in sera via the automatic analyser VIDAS B.R.A.H.M.S. PCT assay (bioMérieux, Marcy l’Etoile, France), according to the manufacturer’s instructions. The lower limit of detection of the assay was 0.05 ng/mL and the functional assay sensitivity was 0.09 ng/mL (VIDAS B.R.A.H.M.S. PCT package insert; bioMérieux).
For each sample, an aliquot of 5 to 10 mL whole blood was inoculated into BACTEC aerobic and anaerobic bottles (Becton Dickinson, Sparks, MD). BACTEC Plus bottles were used for patients under antibiotic therapy and standard bottles for untreated patients. Two sets from two different sites were collected at the same time. The bottles were incubated in a BACTEC FX automated blood culture system (Becton Dickinson). All bottles flagged positive were removed from the instrument and an aliquot was taken for Gram-stain and culture on solid media for subsequent analysis. Identification of microorganisms was performed with conventional methods and with the matrix-assisted laser desorption/ionization-time-of-flight mass spectrometry (Bruker Daltonics, Bremen, Germany).
Microorganisms detected by BCs were considered to be clinically relevant pathogens rather than contaminants according to the following conditions: (i) microorganisms identified by two or more BCs, reported by the clinician as the cause of the episode of sepsis; (ii) microorganisms detected by only one set of BCs if coincided with the results of culture from samples from the suspected infectious foci, collected from the same patient during the same infectious episode; (iii) microorganisms detected only in one set of BCs, belonging to a species included among the etiopathogenic agents of the patient infectious disease (e.g.,
Values were expressed as count and percentages or median and interquartile range (IQR). Statistical significance was assumed if a null hypothesis could be rejected at a
Samples were collected as part of standard care and those included in the database were deidentified before access. No personal information was stored in the study database. No patient intervention occurred with the obtained results. For these reasons, the study was exempt from the institutional review board.
During the entire study period, a total of 8,752 BCs were collected from 3,651 patients. PCT was not drawn concomitantly with the first BCs in 1,702 patients that were excluded from the study. A total of 1,949 patients fulfilled the inclusion criteria and were enrolled in the study. Demographic characteristics of the patients and results from BCs are described in Table
Demographic and clinical characteristics of the 1,949 patients included in the study.
Variable | Values |
---|---|
Males | 1,150 (59%) |
Females | 799 (41%) |
Age (years) | 74 (IQR 62–83)* |
Ward of hospitalization | |
Medical | 1,735 (89%) |
Surgical | 179 (9.2%) |
Intensive Care Unit | 35 (1.8%) |
Antimicrobial therapy before sampling | 1,553 (79.7%) |
Blood culture | |
Negative | 1,286 (66%) |
Contaminated | 72 (3.7%) |
Monomicrobial | 586 (30.6%) |
Polymicrobial | 5 (0.3%) |
Antimicrobial therapy was already administered in 79.5% of patients with negative BCs, 74.6% with contaminated BCs, 82.6% with Gram-negative pathogens, 77.5% with Gram-positive pathogens, 87.5% with fungal pathogens, and 60% with polymicrobial sepsis. The rates of patients assuming antimicrobial therapy according to BC results showed no significant differences (
PCT median value of positive BCs (6.72 ng/mL, IQR 1.5–23.3) was significantly higher than those observed in negative BCs (0.3 ng/mL, IQR 0.1–0.9, and
PCT median values according to BC results are shown in Figure
Comparison of PCT median values according to BC result.
To evaluate the PCT diagnostic accuracy in predicting causative organisms of bloodstream infections, ROC analysis was performed in monomicrobial BCs (Figure
Receiver operating characteristic (ROC) curves of different cut-offs of PCT in differentiating: (a) Gram-negative bacteria from Gram-positive bacteria (AUC 0.765, 95% CI 0.725–0.805;
Table
PCT median values corresponding to pathogens that were isolated from two or more patients with monomicrobial bloodstream infections.
Pathogen | Number of patients | Median PCT values (interquartile range) (ng/mL) |
---|---|---|
Gram-positives | ||
|
103 | 3.6 (1.3–9.3) |
|
43 | 0.5 (0.3–2.2) |
|
18 | 1.6 (0.9–2) |
|
18 | 6.9 (3.3–23.9) |
|
5 | 2.1 (0.2–2.2) |
|
4 | 14 (1.6–26.1) |
|
3 | 1.1 (0.6–1.2) |
|
3 | 0.3 (0.2–2.7) |
|
3 | 1.8 (1–7.9) |
|
2 | 6.3 (0.6–12.1) |
|
2 | 1.1 (0.7–1.4) |
|
2 | 5.4 (0.1–10.7) |
|
2 | 0.07 (0.05–0.1) |
Gram-negatives | ||
|
||
|
183 | 18.5 (6.5–56.4) |
|
56 | 22.3 (9.6–52.4) |
|
14 | 5.5 (3.5–7.5) |
|
12 | 11.3 (8.3–16.5) |
|
5 | 14.9 (3.8–15.5) |
|
5 | 2.3 (1.1–3.7) |
|
3 | 23.2 (18–24.7) |
|
3 | 21.5 (12.5–110.7) |
|
2 | 71.8 (68.2–75.4) |
|
2 | 21 (14.4–27.7) |
Nonfermentative obligate aerobic | ||
|
21 | 6.8 (1.3–11.9) |
|
16 | 2.2 (0.6–7.4) |
|
4 | 20.5 (9.4–166.6) |
Obligate anaerobic | ||
|
8 | 2.8 (0.5–8.5) |
Fungi | ||
|
12 | 0.5 (0.3–1.2) |
|
3 | 0.6 (0.5–0.9) |
|
5 | 0.6 (0.4–0.7) |
PCT values corresponding to pathogens isolated from single patients with monomicrobial or polymicrobial bloodstream infections.
Bloodstream infection from | Pathogen | PCT values (ng/mL) |
---|---|---|
Gram-positives |
|
1.31 |
|
0.87 | |
|
6.63 | |
|
4.52 | |
|
3.54 | |
|
0.40 | |
|
0.25 | |
|
0.21 | |
|
0.45 | |
|
||
Gram-negatives |
|
1.70 |
|
8.49 | |
|
0.42 | |
|
2.24 | |
|
0.38 | |
|
0.48 | |
|
0.44 | |
|
0.98 | |
|
24.17 | |
|
0.90 | |
|
2.49 | |
|
||
Fungi |
|
0.66 |
|
0.72 | |
|
1.06 | |
|
0.13 | |
|
||
Polymicrobial |
|
15.5 |
|
12.3 | |
|
19.5 | |
|
5.61 | |
|
170.9 |
Comparison of PCT median values in bloodstream infections by
Receiver operating characteristic (ROC) curve of different cut-offs of PCT in differentiating
The main findings of this study are that, in patients with suspected sepsis, the PCT cut-off value of 10.8 ng/mL could be of help in predicting an infection caused by Gram-negatives, with a specificity of 82.5%. A cut-off of 3.1 ng/mL could be of help in excluding an infection caused by
The ability of PCT to discriminate infections by Gram-positive or Gram-negative organisms has been recently described. Charles et al. [
Although the mechanism underlying different PCT production in response to different bacterial pathogens is not completely clear, it could possibly be explained by the different interaction of Gram-positive or Gram-negative bacteria with host’s cells, involving lipoteichoic acids or LPS, respectively, and different pathogen-associated molecular patterns (PAMPs), engaging different TLRs, expressed on human cells [
To the best of our knowledge, this is the first study showing a significant difference in the PCT values between bloodstream infections sustained by
It is conceivable that the high median PCT values found in polymicrobial bloodstream infections could be attributable to the presence of Gram-negative bacteria in all of them, specifically,
We found that PCT optimally discriminated between Gram-negative and fungal infections at the best cut-off of 1.6 ng/mL and, though with less accuracy, between Gram-positive and fungal infections. Similarly, Martini et al. [
This study has some limitations. First, the discriminatory power found for PCT could have been confounded by the lack of patients’ baseline characteristics and comorbidities. Indeed, information about factors that can influence PCT levels, such as recent transplantation, severe and prolonged cardiogenic shock, heat shock, severe pancreatitis, rhabdomyolysis, autoimmune disorders, and others [
In conclusion, PCT may be of value to distinguish Gram-negative from Gram-positive and fungal infections; nevertheless, its utility to predict different microorganisms needs to be assessed in further studies including detailed patient information. The findings of the present study show that PCT cut-off of ≥10.8 ng/mL could suggest an infection by Gram-negatives, and the cut-off ≤3.1 ng/mL could suggest exclusion of infection by Enterobacteriaceae. A PCT cut-off >1.3 ng/mL could be of help in ruling out a fungal bloodstream infection.
The authors declare that there is no conflict of interests regarding the publication of this paper.