In parallel with developing technologies and antimicrobial treatments, premature and low birth weight babies have an increased chance of survival. In this case, supportive therapies (total parenteral nutrition, etc.), interventional procedures (intubation, central catheter, etc.) and developing complications contribute to the increase in neonatal sepsis, making it a significant cause of mortality and morbidity. Many cytokines have been shown to play a role in the pathogenesis of sepsis, and these cytokines can help predict prognosis. There are many studies investigating the role and prognostic value of tumor necrosis factor-α (TNF-α), interleukin-6 (IL-6), interleukin-10 (IL-10) and procalcitonin (PCT) levels in the pathophysiology of sepsis [17–20]. However, there is not yet a test that can fully predict the severity and prognosis of the disease in septic newborns. Therefore, our study was conducted to investigate whether the levels of cTn-I and CK-MB, which are cardiac biomarkers, can give an idea of the prognosis of sepsis.
The definitive diagnosis of sepsis is made by culturing the causative microorganisms. Blood culture positivity rate was found 58.5% by Hacimustafaoglu et al [21] in a study conducted in Turkey. In international studies, blood culture positivity rates range from 23–50% [22]. In our study, 30% blood culture positivity was determined in accordance with the literature. The sensitivity of blood culture in newborns for the diagnosis of sepsis is 50–80% under the best conditions. It should be kept in mind that there may be false positive or negative blood culture results due to reasons such as sampling under inappropriate conditions, intrapartum antibiotic treatment to the mother, insufficient blood sample taken for culture, and therefore different rates of culture positivity may be detected [23].
Neonatal sepsis is divided into early and late onset according to the postnatal age in which it occurs. Although EOS is a community-acquired infection mainly transmitted from the mother, LOS mostly defines healthcare-associated infections. In the study by Motara et al [24], it was reported that the EOS was 5.3% and the LOS was 94.7%, while the Shehab El-Din et al [25] study reported that the EOS was 44.2% and the LOS was 55.8%. In a long-term study investigating the epidemiology of neonatal sepsis in the UK [26] the frequencies of EOS and LOS were found to be 24% and 76%, respectively. In our study, 13 patients (14.9%) had EOS, 73 (83.9%) had LOS, and one (1.1%) had very LOS. The higher diagnosis of late neonatal sepsis was attributed to the higher incidence of healthcare-associated infections.
Mortality due to neonatal sepsis takes a serious place among the causes of neonatal death. As a result of the 6-year analysis by Wu et al [27], the mortality rates due to EOS and LOS were found 10.7% and 7.4%, respectively, and Martius et al [28], found the overall mortality rate due to neonatal sepsis as 10.5%. In our study, the overall mortality rate of patients with neonatal sepsis was 21.9%. Among them, the mortality rate was 38.5% in EOS and 19.2% in LOS. We believe that the fact that the mortality rate was found to be partially higher in our patients than in other studies in the literature may be due to the admission of severe cases in our unit, since our hospital is a reference center.
The frequency of microbiological agents that cause sepsis may differ between regions, countries and hospitals, as well as these rates may change over time, even in the same NICU, depending on the changes in health policies [29]. The most common microorganism associated with EOS is Group B streptococcus (GBS) in term newborns and E. coli in preterm newborns. Other rare causes of both EOS and LOS include S. pyogenes, N. gonorrhoeae, E. faecalis, S. pneumoniae, N. meningitidis, Ureaplasma spp., M. Hominis, and Candida spp. Herpes simplex virus and enterovirus infections take the first place among viral agents, and especially herpes sepsis is associated with serious mortality and morbidities.1 In Leal et al’s [30] study, 10% of the patients with culture positivity were found to be EOS (the most common cause of CoNS), while 90% of them were LOS (CoNS, S. aureus and K. pneumoniae in order of frequency). In the study of Lim et al [31], CoNS was found to be the most common among Gram-positive agents and K. pneumoniae was the most common among Gram-negative agents. In our study, culture positivity was found to be 7.7% in the EOS group and 92.3% in the LOS group. Overall, CoNS and K. pneumoniae were the most frequently isolated microorganisms in the blood culture of patients with sepsis, respectively.
Thrombocytopenia is associated with gram negative sepsis, especially in VLBW infants. Among all pathogens, this risk has been reported to be higher, especially in infections due to gram-negative bacteria and fungi. The basic mechanism of thrombocytopenia is the inability of platelet production to compensate for destruction. Some studies have shown that thrombocytopenia alone can predict sepsis-related mortality [32, 33]. Many studies have found an association between thrombocytopenia and multiple organ failure, poor prognosis, and death in intensive care unit patients [34]. In our study, a statistically significant difference was found between the groups I + II and group III, and between the group I and the group II in terms of platelet counts. The platelet count was significantly lower especially in the culture positive sepsis group, and the mortality rate was statistically significant higher in these patients with more severe thrombocytopenia. There was also a statistically significant difference in platelet count between patients discharged with recovery and died.
Changes in WBCs count can also be detected in neonatal sepsis. It is not recommended as a sepsis biomarker since it has a poor positive and negative predictive value in neonatal sepsis [35]. Leukopenia has a higher specificity for the diagnosis of sepsis compared to the total WBCs count and the presence of neutropenia is also thought to increase the risk of mortality [36]. Hanaganahalli et al [37], who classified patients as “culture-proven sepsis”, “clinical sepsis” and “control group” similar to the grouping in our study, did not detect a statistically significant difference in WBCs count between the three groups (p = 0.336). In our study, WBCs count was similar in patients in culture positive sepsis, clinical sepsis and control groups, and no statistically significant difference was found between them.
Due to some disadvantages of blood culture, which is the gold standard test in the diagnosis of neonatal sepsis, CRP has been another test that is frequently used in the differential diagnosis of infected babies. The serum CRP value begins to rise 12 hours after the onset of clinical symptoms and plateaus after 20–72 hours. Therefore, it may not be useful as an infection marker at an early stage. These physiological properties make CRP a more valuable test in the follow-up rather than the diagnosis of sepsis [38–40]. Koksal et al [41] found statistically significant higher mean CRP values in septic newborns. Liu et al [42] divided newborns with neonatal sepsis into two groups as bacterial and non-bacterial infections and compared the data of these groups with the healthy control group. When the CRP values of these patients were compared, they found a statistically significant difference both within the sepsis group (bacterial/non-bacterial) and between the sepsis group and the control group. In the study of Bartolovic et al [43], the median CRP values were found to be statistically significant higher in septic newborns than healthy controls (p < 0.001). In our study, similar to the literature, a statistically significant difference was found between the sepsis groups and the control group in terms of CRP levels.
Lactate is mainly a product of anaerobic metabolism. Increased lactate levels are thought to be due to sepsis-associated hypotension and hypoperfusion, and disruption of the mitochondrial oxidative phosphorylation process. Lactic acid is a metabolic product that occurs early in septic preterm infants [44–46]. Sarafidis et al [45] found statistically significant higher blood lactate levels in septic newborns (p = 0.007). Rodriguez et al [46] evaluated lactate levels at the 3rd and 7th days of life in preterm infants needing mechanical ventilation to predict mortality. They found that lactate > 1.5 mmol/L had a sensitivity of 95% at day 3 and 91% at day 7 to determine the risk of mortality, while a pH value of > 7.25 had a specificity of 96% at day 3 and 91% at day 7 to determine the probability of survival. Prieto et al [47] found that the pH values of VLBW infants who died of nosocomial sepsis were significantly lower than those of surviving patients. In our study, the pH values of cases were also statistically significantly different between culture positive sepsis, clinical sepsis and control groups. The patients in the sepsis group (culture positive sepsis or clinical sepsis) had significantly higher lactate levels than the control group. However, although there was a significant difference in pH value between the culture-positive sepsis and clinical sepsis groups, this difference was not found for lactate level (p = 0.312), which has a statistically significant effect on mortality. It was found that 1 unit increase in lactate value increased the risk of death by 2.6 times.
Although the uses and reference ranges of cardiac biomarkers in adults are well defined, this remains unclear in newborns. Markers of myocardial damage can be detected in the blood, which can develop not only in the case of disease, but also due to prenatal and postnatal causes such as exposure to acute or long-term tocolytic therapy, physiological circulatory changes after birth. Studies have shown that cardiac troponin limits are much higher in healthy newborns than in adults due to these factors [48–50]. CK-MB, which is mostly of muscular origin and reflects birth stress or tissue injury, does not give clear results in the newborn, especially as a measure of myocardial damage in the first days of life. However, increases in cardiac troponin above a certain physiological value may indicate myocardial involvement. Cardiac markers have been studied in many diseases such as perinatal asphyxia, patent ductus arteriosus, persistent pulmonary hypertension, sepsis, and bronchopulmonary dysplasia, both to determine the severity of the disease and to predict the prognosis [51].
Myocardial dysfunction due to sepsis is the result of hemodynamic, molecular, metabolic and structural changes and is closely related to sepsis-related mortality and morbidities [52]. An increase in cTn levels is a common finding in patients with sepsis. Coronary angiography and autopsy results performed in these patients showed that elevated cTn may occur without significant findings in favor of coronary artery disease [53]. Peek et al [54] studied the levels of cTnI, cTnT, CK, and CK-MB in healthy newborn calves with experimental sepsis with E. coli endotoxin. They found a significant difference in cTnI and CK levels measured after endotoxemia compared to control and baseline values. In an adult study of patients followed up with a diagnosis of cancer and admitted to the emergency department due to septic shock, a significant difference was found between survivors and deceased patients in terms of CK-MB and troponin levels [55]. They concluded that cTnI level is an important determinant of early (≤ 7 days) mortality. In a study by Oliveira et al [56] in which they examined the effect of cTnI levels on survival in pediatric patients diagnosed with sepsis and septic shock, it was reported that serum cTnI levels were the only variable associated with the severity of septic disease and death. In that study, it was found that CK-MB levels had no effect on survival.
The study by Abdel-Hady et al [8] demonstrated the diagnostic and prognostic use of cTnT and the “total ejection isovolume (Tei) index” in patients with sepsis. They found that there was a positive correlation between left and right ventricular Tei index and cTnT levels in patients with neonatal sepsis. The cTnT concentrations and left ventricular Tei indices of the deceased septic neonates were found to be significantly higher compared to the survivors. In the study of Frencken et al [57], in which they evaluated the relationship of cTnI levels with sepsis-related mortality and long-term (1 year) outcomes in patients with sepsis, it was found that increased cTnI levels were associated with increased mortality rates in the first 14 days of sepsis, but not associated with mortality in the following days. It has been reported that there is an association between increased cTnI levels during sepsis and long-term cardiovascular diseases in patients who have survived an episode of sepsis, which may predict cardiovascular morbidity after discharge.
In our study, cTnI levels, a sensitive marker of cardiac damage, were found to be significantly higher in the sepsis group, and the mortality rate was similarly higher in patients with high cTnI levels. Studies in the literature have shown that cTnI levels can provide information about long-term prognosis. However, there is conflicting results associated with the fact that CK-MB levels can predict prognosis in neonatal diseases. In our study, no statistically significant difference was found between the sepsis group and the control group in terms of CK-MB levels. Similarly, no significant difference was found in terms of mortality in patients with high CK-MB levels. In our study, the cTnI cut-off value was found to be 0.13 and it was seen that an increase of 1 unit increased the risk of death 1,4 times. The relationship between cTnI levels and morbidity could not be evaluated because sufficient data could not be obtained about the short-term complications and long-term sequelae of the patients.
Limitations of our study were i) since micro-CRP, an infection marker, was used in our unit, patients were not routinely requested for PCT examination; ii) because BNP and IL-6 tests are not routinely studied in our hospital, cytokines or biomarkers such as IL-6 and BNP, which are often used in literature studies, are not included in our data; iii) although echocardiographic examination was performed in almost all of the patients included in the study, concurrent cardiac functions were not evaluated with the study of cTnI levels. Therefore, it could not be investigated whether there is a correlation between cTnI levels and the severity of cardiac dysfunction in patients with sepsis.
In conclusion, neonatal sepsis is an important health problem of newborns with high mortality. It is not always possible to recover a microbiological agent in the culture of patients with sepsis in the NICU and therefore additional tests are needed for diagnosis. The higher incidence of LOS in the NICU is due to healthcare-associated infections, and great attention to infection control measures and avoidance of unnecessary invasive procedures will reduce its frequency. CoNS and K. pneumoniae are the main pathogens causing neonatal sepsis. Although thrombocytopenia is a late finding of sepsis, it should be followed carefully in patients with sepsis because it indicates a poor prognosis. Lactate levels can be used to predict prognosis in patients with sepsis. Further studies are needed to prove the effectiveness of CK-MB levels in evaluating the prognosis of patients with sepsis. cTnI is a marker that can provide insight into the risk of short-term mortality in patients with sepsis, and more comprehensive studies are needed to be able to predict long-term outcomes.