Heparin-binding protein and procalcitonin in the diagnosis of pathogens causing community-acquired pneumonia in adult patients: a retrospective study

Study population and data collection

The patients’ data were collected in Nanjing First Hospital from September 1, 2018 to May 31, 2019. The inclusion criteria included: (1) age >18 years; (2) onset in the community; (3) new patchy infiltrates, lobar or segmental consolidation, ground-glass opacities or interstitial changes with or without pleural effusions; (4) new onset of cough or expectoration, or aggravation of existing symptoms of respiratory diseases, with or without purulent sputum, chest pain, dyspnea or hemoptysis or signs of pulmonary consolidation and/or moist rales (Cao et al., 2018); and (5) all cases were confirmed by viral nucleic acid assays with real-time polymerase chain reaction (PCR) including influenza A virus, influenza B virus, human parainfluenza virus, adenovirus and respiratory syncytial virus, using nasopharyngeal swab and sputum smear and culture. Exclusion criteria: patients with human immunodeficiency virus (HIV), hepatitis B virus (HBV), hepatitis C virus (HCV) and tuberculosis infection; autoimmune diseases; long-term corticosteroid use for >6 months. We enrolled 162 patients. Demographic patient data, including age, sex and infection sites, were obtained from medical records. The computed tomography (CT) score was used to quantitatively estimate the pulmonary involvement in all these abnormalities on the basis of the area involved (Chang et al., 2005; Pan et al., 2020). Each of the five lung lobes were visually scored from 0 to 5 as: 0, no involvement; 1, <5% involvement; 2, 25% involvement; 3, 26%–49% involvement; 4, 50%–75% involvement; 5, >75% involvement. The total CT score was the sum of the individual lobar scores and ranged from 0 (no involvement) to 25 (maximum involvement). Confusion, uremia, elevated respiratory rate, hypotension, and aged 65 years or older (CURB-65) score was enrolled.

According to the clinical manifestations, sputum culture and PCR results, the patients were divided into three groups: bacterial, fungal and viral infection groups (Tables 1 and 2).

Ethical approval and consent to participate

The study protocols were approved by the Research Ethics Committee of Nanjing First Hospital. The need for written informed consent from the participants was waived by the committee because our study was retrospective, anonymous and only used currently existing data.

Investigations

We used EDTA, heparin and citrate for assays of CRP, PCT and HBP, respectively. Blood samples were taken within 2 h after admission of patients. Complete leukocyte counts, liver function tests including serum concentration of alanine aminotransferase (ALT) and total bilirubin (TB), as well as renal function tests including blood urea nitrogen (BUN) and creatinine (Cr) were measured as described in our previous research (Cai et al., 2017). The plasma concentration of CRP was tested within 2 h by turbidimetric inhibition immunoassay, normal range: 0–8 mg/L (PA900, Lifotronic, China). The plasma concentration of PCT was measured using an automatic chemiluminescence immunoassay analyzer within 2 h, normal range: 0–0.1 ng/mL (cobas e 601, Roche, Switzerland). The detection of HBP concentration in plasma was detected within 2 h using the Axis-Shield HBP microtiter plate ELISA kits using sodium citrate tubes with the automatic rapid immunoassay system, normal range: 0–11.4 ng/mL (Jet-iStar 3000, Joinstar, China).

Statistical methods

The variance homogeneity of the data was evaluated by Levene’s test. The Shapiro–Wilk test was used to assess data normality. Normally and non-normally distributed data were shown as mean ± standard deviation and median and interquartile range (IQR), respectively. The differences in measurement data were compared with Student’s t-test if the data met normal distribution and homogeneity of variance, or with a non-parametric test. The Mann–Whitney U test was used to compare the statistical significance between the groups with non-normally distributed data. The enumeration data were compared with the chi-squared test. The receiver operating characteristic (ROC) curve was used to evaluate the significant differences of the sensitivity and specificity of indicators. The results were considered significant when the p-values were ≤0.05. All statistical analyses were conducted using GraphPad Prism 6.0 (San Diego, CA, USA) and IBM SPSS version 22 (IBM SPSS, Armonk, NY, USA).

Characteristics of the study population

The clinical characteristics of the 162 patients enrolled are shown in Table 1. There were no significant differences in the age and male proportion among the three groups. The proportion of tumors in the fungal group was significantly higher (5/21, 23.81%) than that in the viral group (0%), and there were no significant differences in the proportion of complications among the three groups. The CURB-65 score, CT score, PaO₂/FiO₂, liver function and kidney function showed no significant differences among the three groups (Table 1). The different pathogens are shown in Table 2.

Accessible blood index

HBP

There were significant differences in the concentration of HBP among the three groups (p < 0.001) (Table 1). The concentration of HBP [11.05 (8.01–16.12) ng/mL] in the viral group was significantly lower than that in the bacterial group [48.85 (33.57–78.54) ng/mL] (U = 258.5, p < 0.001) and fungal group [52.90 (8.51–126.00) ng/mL] (U = 103.0, p < 0.001). However, there was no significant difference in the concentration of HBP between the bacterial and fungal groups (Fig. 1A). Because HBP is released by neutrophils, we tested the correlation between concentration of HBP and numbers of neutrophils, and found a high correlation between the two biomarkers (r_s = 0.421, p < 0.001).

Leukocyte and neutrophils

There were significant differences in the leukocyte and neutrophils counts among the three groups (p < 0.001) (Table 1). The leukocyte counts [3.89 (2.13–7.64) × 10⁹/L] in the viral group were significantly lower than those in the bacterial group [8.99 (6.45–13.24) × 10⁹/L] (U = 598.5, p < 0.001) and fungal group [9.38 (5.78–14.62) × 10⁹/L] (U = 114, p < 0.001). The leukocyte count in the bacterial group did not differ significantly from that in the fungal group. The neutrophil count in the viral group was significantly lower than that in the bacterial and fungal groups, but the count in the bacterial group did not differ significantly from that in the fungal group (Figs. 1D and 1E).

HBP and PCT concentrations distinguished bacterial and fungal from viral infections

We found significant differences in the concentrations of HBP and PCT and leukocyte and neutrophils counts between the bacterial and viral groups. The areas under the ROC curve (AUCs) for concentration of HBP and PCT, and leukocyte as well as neutrophils counts for distinguishing bacterial infection from viral infection were 0.927, 0.892, 0.832 and 0.806 , respectively (Fig. 2A). The best cut-off value was 20.050 ng/mL for HBP, with a sensitivity of 0.861 and a specificity of 0.939. The positive predictive value (PPV) was 0.979 and negative predictive value (NPV) was 0.674, and the positive likelihood ratio (PLR) was 14.208 and negative likelihood ratio (NLR) was 0.148 for discrimination of bacterial from viral infection (Table 3). The sensitivity and specificity of PCT for distinguishing bacterial from viral infection were similar to those with HBP.

The AUCs of concentration of HBP and PCT, and leukocyte and neutrophils counts for distinguishing fungal infection from viral infection were 0.851, 0.883, 0.835 and 0.830, respectively (Fig. 2B). The best cut-off value was 29.950 ng/mL for HBP, with a sensitivity of 0.667 and a specificity of 0.970, PPV 0.933 and NPV 0.821, and PLR 22.0 and NLR 0.344 for discrimination of fungal from viral infection. The best cut-off value was 0.560 ng/mL for PCT, with a sensitivity of 0.714 and specificity of 0.879, PPV 0.789 and NPV 0.829, and PLR 5.893 and NLR 0.325 for discrimination of fungal from viral infection (Table 3).