SOLUBLE TRIGGERING RECEPTOR EXPRESSED ON MYELOID CELLS -1 (sTREM-1) AS A DIAGNOSTIC AND PROGNOSTIC MARKER FOR LATE-ONSET SEPSIS IN PRETERM NEONATES

1. Medical Microbiology & Immunology Department, Faculty of Medicine, Mansoura University, Mansoura, Egypt. 2. Neonatal Intensive Care Unit, Mansoura University Children`s Hospital, Mansoura, Egypt. ...................................................................................................................... Manuscript Info Abstract ......................... ........................................................................ Manuscript History


ISSN: 2320-5407
Int. J. Adv. Res. 4(8), 123-131 124 Neonatal sepsis is a systemic inflammatory response to infection that is characterized by presence of signs and symptoms of infection with or without concurrent bacteremia in the first month of life (Barman and Das, 2016;Behmadi et al., 2016). It is classified into early-onset sepsis (EOS) which occurs before 72 hours of life, and LOS after 72 hours (Martin et al., 2014).
The most challenging aspect in neonatal sepsis is the difficulty of its early and accurate diagnosis as it usually presents with subtle and non-specific signs that can be confused with other non-infective conditions with rapid deterioration of clinical condition (Ayazi et al., 2014;Hedegaard et al., 2015).
Although blood culture is regarded as the gold standard for confirmation of neonatal sepsis diagnosis (Delanghe and Speeckaert, 2015), it has many limitations such as unavailability of the results until 24-72 hours after starting the culture and its low sensitivity as 57% of septic infants may have false negative results (Haque, 2010). Many diagnostic tests such as white blood cell count (WBC), absolute neutrophilic count (ANC), and C-reactive protein (CRP) have been investigated for early diagnosis of septic neonates. However, they do not possess adequate sensitivity or specificity (Benitz et al., 2015;Coggins et al., 2016).
The triggering receptor expressed on the myeloid cells-1 (TREM-1) is a member of the immunoglobulin superfamily. It is a 30-kDa transmembrane glycoprotein expressed on polymorph nuclear (PMN) granulocytes and monocytes which amplifies the inflammatory response initiated by toll like receptors (TLRs) by triggering the release of pro inflammatory cytokines, activation of neutrophil degranulation and oxidative burst (Gomez-Pina et al., 2012; Sandquist and Wong, 2014). Soluble form of TREM-1 (sTREM-1) is a 27kDa protein produced through proteolytic cleavage of membrane-anchored TREM-1 by matrix metallo-proteinases (MMPs) (Gómez-Piña et al., 2007). It peaks 2 hours after infectious exposure where it can be measured in biological fluids providing an early marker for sepsis (Dupuy et al., 2013). Studies have reported increased levels of TREM-1 and sTREM-1 in the presence of infection while it is not up-regulated in patients with inflammatory conditions without infection so, it is useful for distinction between infectious and non-infectious diseases (Paolucci et al., 2012;Alqahtani et al., 2014).
Although several studies documented the role of serum level of sTREM-1 in diagnosis of neonatal sepsis (Su et al., 2012;Saldir et al., 2015), there is insufficient data regarding its prognostic value in neonatal sepsis. Hence, our study aimed at investigating the role of serum levels of sTREM-1 in early diagnosis and prognosis of LOS.

Subjects & methods:-Study Design and Population:-
This prospective study was carried out on preterm neonates (<37 weeks of gestation) admitted to the NICU of MUCH from November 2014 to November 2015. Infants with chromosomal abnormalities or major congenital malformations were excluded. Informed consent was obtained from the infants` parents or legal guardian. The study was accepted by Mansoura Faculty of Medicine Institutional Review Board. A total of 59 preterm neonates were prospectively recruited in the study. Thirty of them had three or more clinical signs of LOS as recommended by Zaki and Elsayed(2009)constituting the sepsis group. Twenty-nine gestational age and sex matched newborns without clinical findings of sepsis served as the control group. Two blood samples were obtained from all enrolled neonates; one for blood culture and the other one for sTREM-1 assay by ELISA kits. Routine sepsis screening as CBC, TLC, ANC and CRP were recorded. The septic neonates were initially evaluated at clinical suspicion of sepsis and 48-72 hours after starting antibiotic therapy. All blood culture samples were sub cultured on blood, MacConkey and nutrient agar media (Oxoid) followed by microbiological identification by colonial morphology, Gram stained films and standard biochemical reactions (Koneman et al., 2006). Antimicrobial susceptibility testing was performed by the Kirby Bauer disc diffusion method according to CLSI, 2014 guidelines. Septic neonates were further classified into two subgroups: culture positive sepsis and culture negative sepsis according to the blood culture results.

STREM-1 Assay:-
Blood sample were allowed to coagulate for 10-20 min followed by centrifugation at 2000-3000 r.p.m for 20 min. The separated sera were kept at -20 °C until assayed by sTREM-1 ELISA kits (Sun Red Biotechnology,Shanghai, China) according to the manufacturer`s instructions.

125
Statistical Analysis:-Data were analyzed by SPSS version 21 (SPSS Inc., Chicago, IL, USA). Qualitative data were described as numbers and percentages. Quantitative data were tested for normality by Kolmogorov-Smirnov test. Parametric results were expressed as means ± standard deviations, while non-parametric data were expressed as median with inter-quartile range. The Chi-square test or Fisher 's exact test were used for comparing categorical variables between groups. Independent sample t-test was used for comparing parametric variables between groups. In the non-normally distributed variables, Mann-Whitney-U test was performed to compare between groups, while Wilcoxon signed ranks test was used to compare pretreatment and follow-up. Receiver operating characteristic (ROC) curves were plotted to determine the cutoff point and area under the ROC (AUC) curve was calculated. P value < 0.05 was considered to be statistically significant, P value <0.001 was highly significant.

Results:-
Fifty-nine preterm neonates were included in the study; 30 in the sepsis group and 29 in the control group. There were no statistically significant differences between the two groups in terms of demographic characteristics except for the postnatal age which was significantly higher in sepsis group (Table 1). Table 1:-Demographic characteristics of sepsis and control groups.
In the sepsis group, 11 (36.7%) patients were diagnosed with culture-proven sepsis while 19 yielded no growth (63.3%). Among culture-proven sepsis, 3 patients (27.27%) had polymicrobial sepsis. In our study, 14 organisms were isolated from blood culture; 9 (64.29%) Gram-negative, 4 (28.57%) Gram-positive and 1 (7.14%) candida albicans ( Table 2). The antibiotic susceptibility pattern of the isolated bacteria was determined (Table 3). Table 2:-Blood Culture results of sepsis group.   (Figure 1). For prognosis of sepsis, sTREM-1 was statistically significantly elevated in survivors than non-survivors (p 0.006) being the highest discriminative marker for prognosis of sepsis. Regarding other laboratory parameters (WBC, hemoglobin, platelet and CRP), there was no statistical significant difference between survivors and non-survivors. However, ANC was statistically significantly elevated in non-survivors (Table 5).  At cutoff 77.5 pg. /ml using ROC curve, sTREM-1 had 90% sensitivity, 51.7% specificity, 65.9% PPV, 83.3% NPV for diagnosis of LOS. This high sensitivity allows recognition of almost all septic neonates while ruling out LOS in clinically suspected neonates, while this moderate specificity would be acceptable in the setting of neonatal sepsis allowing for substantial reduction in antibiotic prescriptions. Moreover, PPV is acceptable meaning that only one out of three neonates will receive antibiotic therapy, but that is harmless than missing a truly infected neonate. The high NPV would help in ruling out LOS sepsis with confidence. 128 respectively. The moderate sensitivity and specificity values reported in our study show that CRP alone is insufficient for diagnosis of neonatal sepsis.
Since there is no single marker with ideal sensitivity and specificity that was identified as the magic bullet for diagnosis of neonatal sepsis, a combination of markers is highly recommended specially one that can have high sensitivity which is more important that specificity in life threatening conditions with high morbidity and mortality as neonatal sepsis. The combination should also support the neonatologist decision to stop or continue empirical antibiotics without posing a threat to the infant (Bohnhorst et al., 2012;Walley, 2013).
So, we evaluated a combination of CRP and sTREM-1. This combination had 100% sensitivity, 44.8% specificity,65.2% PPV and 100% NPV. Accordingly, positive tests would identify all septic neonates helping to reduce the complication and mortality resulting from delay of treatment, while negative tests would rule out sepsis with absolute confidence. Also, PPV value of 65.2 % is acceptable given that only one out of three infants will be over treated with antibiotic which is considered to be harmless in respect to the fatal consequences of withholding therapy according to a false-negative result.
Regarding the prognostic value of sTREM-1, its levels were statistically significantly higher in survivor than non survivors at diagnosis stating that it is a perfect marker to stratify and recognize neonates for whom mortality or complication are expected, so that clinicians could monitor them closely. Similar results were documented by Wang and Chen (2011) In our study, we found that sTREM-1 levels showed highly significant statistical decrease 48-72 hrs after administration of antibiotics in relation to initial sTREM-1 levels indicating that sTREM-1 is a useful monitoring marker for follow-up of septic neonates to observe the effect of treatment which is consistent with the results of Sarafidis et al. (2010) and Saldir et al. (2015). The main limitation in our study was small sized sample number which can be validated by further large scale studies before application to general population.

Conclusion:-
In conclusion, sTREM-1 is considered a perfect marker for early and accurate diagnosis of LOS with an almost perfect sensitivity approaching 100%, which is reached when combined with CRP,This combination can identify allneonates who are at high risk of sepsis and in need of antibiotics, and those in which antibiotics can be withheld. This will ultimately lead to prevention the high mortality and the morbid complications associated with late diagnosis and administration of unnecessary therapy in neonatal sepsis. Furthermore, sTREM-1 is a reliable biomarker for prognosis of LOS, and its serial levels can determine the efficiency of treatment, helping the clinicians to decide whether to stop or continue antibiotics.