Neonatal infection (NCI) remains a major public health problem due to its high mortality rate. Indeed, the World Health Organization (WHO) estimates that 2.8 million neonatal deaths occurred globally in 2015, 47.6% of which were due to NIN [1] . Half of these deaths occurred in sub-Saharan Africa (SSA) in 2019 [2] , while the early neonatal mortality rate is estimated at 28 per 1000 live births [3] . In the Democratic Republic of Congo (DRC), 15.6% of neonatal deaths are caused by neonatal septicaemia [4] . The incidence of neonatal infections, their characteristics, and their consequences depend on the time of onset and the context in which they occur, particularly at a gestational age [5] . The incidence of NIN has fallen significantly in industrialised countries but remains high in countries with limited resources. An incidence of 0.8 per 1000 births has been found in the United States and Europe [6] . Studies carried out in Africa have shown a frequency of 96.8% in Cameroon (1) and 6.2% in Morocco [7] . Early onset neonatal infection (EONI) refers to an infection occurring within the first 72 hours after birth [6] [8] . Neonatal infections are divided into 2 categories: early neonatal infection when signs occur within 72 hours of life and late neonatal infection when signs occur after the 3rd day of life [9] . EBNIs are the consequence of vertical bacterial transmission from mother to newborn [5] . Clinical manifestations are polymorphous and non-specific [10] in response to sepsis [11] . Because of the difficulty of diagnosis, especially during the first 72 hours after birth, and the fact that the symptoms of neonatal sepsis are non-specific, empirical antibiotic therapy can lead to the treatment of up to 30 uninfected newborns for just one who is probably diagnosed as infected [12] . Blood culture, which is the gold standard, is often not available and cannot produce immediate results [13] . Biomarkers are, therefore, becoming an alternative for the early identification of sick newborns [5] . An evidence-based approach combining anamnestic, clinical and biological data is a better strategy, especially in resource-limited countries [14] . The aim of this study was to determine the clinical and biological characteristics and outcomes of newborns suspected of having EONI in a hospital setting in Kisangani.

Of the 871 newborns admitted, 190 were selected for INBP, a frequency of 21.8%. The average age of the newborns was 48.6 hours, and the sex ratio was 1.2 (165 boys and 132 girls). Low birth weight was found in 15.5% of newborns and neonatal asphyxia in 20.2% of cases ( Table 1 ). The average age of the mothers was 26, and most of them (80%) were housewives and well-educated. The most important antenatal risk factors were meconium amniotic fluid (12.1%), duration of water sac opening ≥18 h (10.8%) and maternal temperature ≥ 38˚C (8.1%), which was statistically significant (p < 0.039). The most frequent clinical manifestations were disorders of thermal regulation including fever and hypothermia (70.4%) followed by respiratory disorders (44.8%) significant (p < 0.018) and neurological signs (26.3%) ( Table 2 ). Of the 297 newborns aged < 72 hours, blood cultures were taken in 123 and came back positive in 5 newborns, giving a confirmed hospital incidence of INBP of 4.1%. The germs isolated were as follows: two blood cultures were positive for Enterobacter spp, two for Staphylococcus spp and one for Serratia spp. PCT was positive in 86 newborns, including the 5 with a positive blood culture. The median PCT of blood culture-positive neonates was higher than that of blood culture-negative neonates ( Table 3 ). Regarding the clinical outcome of neonates, the median PCT was higher in neonates who died at 22.4 [3.7 - 50.0] compared with neonates with improved clinical status 0.8 [0.3 - 3.3] with a significant difference ( Table 4 ). The outcome was favourable in most cases (96%) and the death rate was 4%.

¹n (%); Mean (SD).

¹n (%); ²Chi-square test of independence; ³Fisher’s exact test.

¹Median [EI]; n (%); ²Wilcoxon-Mann-Whitney test; ³Fisher’s exact test.

¹Median [EI]; n (%); ²Wilcoxon-Mann-Whitney test; ³Fisher’s exact test.

INBP remains a major cause of neonatal morbidity and mortality in both developed and developing countries [16] and in the DRC. It remains a concern for clinicians because of its non-specific symptoms and diagnostic difficulties. In our study, the frequency of infection was 21.8%, higher than that found by Chemsi et al. [7] (6.2% in Casablanca, Morocco), by Kamaye et al. [10] (5.4% in Niger) and by Konaté et al. [17] (11% in Mali). This frequency was lower than that found by Kemeze et al. [1] of 96.8% in Cameroon.

The average age of the newborns was 48.6 hours, and only 11.6% were symptomatic within 24 hours of birth. The youngest was 7 hours old. The sex ratio was 1.2, the Apgar score at 5 minutes was <7 in 20.2% and 15.5% of newborns had a low birth weight. The lack of unified criteria for diagnosing neonatal sepsis means that clinicians in resource-limited settings with no access to blood cultures must rely on a range of clinical arguments. [18] The risk factors found in our series were meconium amniotic fluid, opening of the water sac ≥ 18 h and maternal temperature ≥ 38˚C, which was significant. This result is different from the premature rupture of membranes significantly associated with the risk of neonatal sepsis found by Adarate et al. in Ghana [19] . The most frequent clinical manifestations were fever, respiratory disorders (with a significant difference) and neurological disorders. The same result was found by Nyenga et al. [20] in Lubumbashi in the DRC and Berhane et al. in Ethiopia. However, the symptomatology found by Tran et al. in Vietnam [21] and Maamouri et al. in Iran [22] is different from our series represented by respiratory distress and difficulty in eating, respectively. Blood culture, which is the gold standard, diagnosed only 5 positive cases. A negative blood culture does not always rule out neonatal infection, as low blood volume and low bacteremia may be the cause. This result is similar to that found by Oeser et al. [23] in England, who also found a low rate of blood culture in EONI. With regard to the clinical outcome of neonates, the median PCT was higher in neonates who died above the 50 ng/ml threshold than in neonates with improved clinical status, with a significant difference of p < 0.0001 making PCT a prognostic marker for neonatal sepsis. The outcome was favourable in most cases (96%) and the death rate was 4%.

Our study shows that INBP is still common in hospitals in our context. The clinical picture is dominated by fever, as well as respiratory and neurological disorders. Progression is favourable in most cases, but the median PCT is very high in neonates who die, making it a prognostic marker in resource-limited settings.

The limitation of this study is the low positivity of the blood culture. This did not allow us to establish the bacterial ecology of INBP in our environment. Studies on larger cohorts of newborns should be considered in the future.

The clinical examination of newborns suspected of having INBP and the collection of data were carried out by the following authors: H. M. V., G. B. M., V. M. K., B. B. L. and D. G. M.: H. M. V., G. B. M., V. M. K., B. B. L. and D. G. M. Correction of the study protocol and manuscript were provided by B. M. M., D. F. S., E. T. K. and J. P. A. O.