The diagnostic value of (1,3)-β-D-glucan alone or combined with traditional inflammatory markers in neonatal invasive candidiasis

This is a retrospective study. The studied subjects were the neonates admitted to the NICU department of Guangdong Women and Children’s Hospital from January 2013 to January 2018, with available biological and clinical data. The demography data such as the birth body weight, gestational age (GA) and delivery mode were collected. The laboratory results of WBC and PLT count, PCT and hsCRP level, blood BDG level were collected from our laboratory system. The biological data of blood culture were also collected. The subjects were separated into three groups according to the results of blood culture: (a) the isolated pathogens were candidiasis, defined as invasive candidiasis group (IC); (b) the isolated pathogens were bacteria, defined as bacterial infection group (BI); (C) none pathogens were isolated, served as control group (CTRL). Those subjects without the data of BDG or blood culture data were excluded. The study protocol was approved by the Ethics Committee of Guangdong Women and Children’s Hospital. As this is a retrospective study and the patients scattered across Guangdong province (even southern of China), written consent to participate was difficult to obtain for some cases, but verbal consent to participate was collected from the parents/guardians of all the neonates via phone contact. The Ethics Committee of our institute approved the method of obtaining consent.

Blood cultures were conducted with BACT/ALERT® PF vials incubated in Bact/ALERT 3D system at 37°C for 5 days. Positive blood culture was transferred to sheep blood agar plate for isolation. The isolated pathogens were identified by mass spectrometry method with Bruker MALDI-TOF.

The serum level of (1,3)-beta-D-glucan was detected with GKT assay kit (Gold Mountain River Tech Development Co., Ltd., Beijing,China) according to the manufacturer’s introduction. Briefly, 100ul serum sample was added to the dilution tube of the kit, the tube was incubated at 70°C for 10 mins followed by incubated at 10°C for 5mins; 200ul treated sample were transferred to reaction tube, and the reaction was read at 405 nm for 60mins. The concentration of BDG of each sample was calculated automatically using the calibration curve provided by manufacturer.

Quantitative variables are presented as median and interquartile range (IQR: 25th percentile and 75th percentile) or mean and standard deviation. Categorical variables are described as percentages. The differences between groups were assessed by χ² test for categorical variables and Mann-Whitney test for quantitative variables. ROC curve of BDG, CRP, PCT and PC alone or combined were generated with SPSS to determine the ability of these indicators, and cut-off values of these indicators were also calculated.

Eighty neonates (45 male and 35 female) were included into this study. The subjects were separated as IC (30 subjects), BI (25 subjects) and CTRL (25 subjects) group according to the result of blood culture. The obstetric history, clinical characteristics of the different groups are shown in Table 1. The gender distribution and delivery model were similar among different groups. Though the premature rupture of membranes rate was similar among different groups, the preterm labor rate was higher in IC and BI groups compared with CTRL group (Table 1). The median gestational age was 32 (IQR:30–36), 33 (IQR:28–37) and 36 (IQR:36–40) for IC, BI and CTRL group, respectively. The mean birth weight was 1840 ± 747 g, 1975 ± 654 g and 2215 ± 573 g for IC, BI and CTRL group, respectively.

In IC group, C. albicans was the most prominent isolated pathogen, followed by C. parapsilosis, and they accounted for 16 and 10 isolates, respectively. C. glabrata and C.tropicalis accounted for two isolates, respectively. The data of colonization of yeast in body sites such as digestive tract, respiratory tract or skin were not available. Coagulase negative staphylococcus (CNS) was the main isolates in BI group, followed by Klebsiella pneumonia.

The data of inflammatory markers such as hsCRP, PCT, WBC and PC were collected. None of the involved neonates encountered leukopenia (WBC < 1500/μl). Though the WBC count was increased in IC and BI groups compared with CTRL group, no statistical difference was found (Fig. 1a). The PC was significantly decreased in IC and BI groups compared with CTRL group (Fig. 1b). Eighteen of neonates in IC group encountered thrombocytopenia (PC < 150,000/ul), and 11 of them were severe thrombocytopenia (PC < 50,000/ul). Nine neonates in BI group encountered thrombocytopenia and four of them were severe thrombocytopenia. Only four neonates in CTRL group encountered thrombocytopenia, and two of them were severe thrombocytopenia. The PCT level of BI group was the highest followed by IC group, but there was no statistical differences between groups (Fig. 1c). The hsCRP level was significantly increased in IC groups compared with BI and CTRL group (Fig. 1d).

The BDG level of IC group, BI group and CTRL group was 369.4 ± 91.88 mg/ml, 146.0 ± 60.80 mg/ml and 17.42 ± 2.932 mg/m, respectively. The difference between IC, BI and CTRL group was significant (Fig. 1e).

We subdivided the IC group into C. albicans, C. parapsilosis, C. glabrata and C. tropicalis group according to the isolated pathogens, and compared the level of inflammatory markers and BDG between groups (Table 2). Compared with other candidas species, C. albicans related candidaemia had a higher level of BDG. The level BDG was 276.9 mg/ml (IQR:10–1000 mg/ml), 26.3 mg/ml (IQR:10–1000 mg/ml) and 21.7 mg/ml for C. albicans, C. parapsilosis and C. glabrata, respectively. The BDG level in C. tropicalis was unavailable. The serum level of hsCRP and PCT was no statistical differences between groups. Though no statistical difference, the PLT and WBC count was much lower in C. parapsilosis group compared with C. albicans group.

The ROC curve of BDG, hsCRP, PCT or PC alone was generated (Fig. 1f). The cut-off value of BDG was 13.69 mg/ml, with sensitivity of 76% [95% CI: 55–91], and specificity of 71.4% [95%CI: 49–89]. The positive likehood ratio of BDG was 2.66, and the area under the ROC (AUR) curve was 0.79 [95%CI: 0.65–0.93]. The cut-off value of hsCRP was 4.7 pg/ml, with sensitivity of 84% [95% CI: 64–95] and specificity of 81% [95%CI: 58–94]. The positive likehood ratio of hsCRP was 4.41, and AUR was 0.8 [95%CI: 0.71–0.97]. The cut-off value of PCT was 0.378 pg/ml, with sensitivity of 72% [95% CI: 51–88] and specificity of 71.4% [95% CI: 48–89]. The positive likehood ratio of PCT was 2.52 and AUR was 0.75 [95% CI: 0.61–0.90]. The cut-off value of PC was 169,000/μl, with sensitivity of 76% [95% CI: 55–91] and specificity of 71.4% [95% CI: 49–89]. The positive likehood ratio of PC was 2.66 and AUR was 0.78 [95%CI: 0.64–0.91]. Combined ROC of BDG with hsCRP, PCT or PC increased the diagnostic value (Fig. 1g). The AUR was 0.91[95% CI: 0.81–0.99], 0.83[95%CI: 0.71–0.95] and 0.85[95% CI, 0.75–0.96] for combined ROC of BDG with hsCRP, PCT or PC; Further combine of the markers did not significantly increase the diagnostic value (Fig. 1h and i).

We evaluated the diagnostic value of BDG in differentiating candidaemia and bacteremia. Serum BDG level at a cut-off value of 42 mg/ml could distinguish candidaemia from bacteremia, with sensitivity of 72% [95% CI: 51–88], specificity of 62% [95% CI: 39–82] and a positive likehood ratio of 1.89. The AUR was 0.70 [95% CI: 0.54–0.85] (Fig. 2a). Combined of hsCRP (Fig. 2b) or PCT (Fig. 2c) with BDG increased the AUR to 0.71 [95% CI: 0.56–0.86] or 0.715[95%CI: 0.56–0.86], respectively. Combined of CRP, PCT and BDG further increase the AUR to 0.748 [95%CI: 0.61–0.89] (Fig. 2d). Further combined of PC with CRP, PCT and BDG did not increase the AUR (data not showed).

To assess the value of BDG in monitoring the effect of antifungal therapy, we collected series data of BDG of the same patient. The data of 11 patients was available. Antifungal therapy was began at 0* week when the blood culture was positive. As showed in Table 3, in general the serum of level of BDG was drop along with the antifungal therapy. We also found that the level of BDG increased at the beginning of antifungal therapy (patient 2, 5, 6, 8), this might reflect the transient release of BDG as the antifungal drug destroys the cell-wall of the Yeast. As some of the patients requested discharge before the infection was cured, the outcome of total patients was not available. None of the patient died during the hospitalization in our institute.