INTRODUCTION

In Brazil, despite the downward trend in infant mortality in recent years, a slow reduction in the early neonatal component has been observed, with 1/5 of deaths occurring on the first day of life and the majority of causes considered preventable when adequate attention is provided to women’s and newborns’ health.¹ Investigation of this component highlights neonatal sepsis as one of the main causes of these deaths, especially in premature and very low birth weight newborns. ^2,3

Early-onset neonatal sepsis is characterized as a clinical syndrome with systemic signs of infection that occurs in the first 72 hours of life, originating from bacterial pathogens transmitted vertically from mother to newborn before or during delivery.^4,5 The microorganisms most involved in its pathogenesis are group B Streptococcus (GBS), Escherichia coli and Listeria monocytogenes, which together account for approximately 65% ​​to 70%, respectively, of all systemic neonatal bacterial diseases.⁶

The incidence of culture-proven early-onset sepsis in the United States from 2005 to 2008 ranged from 0.75 to 0.77 cases/1,000 live births, and mortality was 10.9%. Black preterm infants had higher rates (5.14 cases/1,000 live births), and 24.4% died.⁷Research at the Vermont Oxford Network, from 2007 to 2016, identified 3.7% of early sepsis in extremely premature infants.⁸ In the Rede Brasileira de Pesquisas Neonatais (RBPN, Brazilian Neonatal Research Network), from 2006 to 2017, the prevalence of this sepsis was 15.5 cases/1,000 very low birth weight newborns, and 52.9% evolved to death.²

“Suspected” early neonatal sepsis is one of the most common and challenging diagnoses in Neonatal Intensive Care Units (NICUs), given that the signs and symptoms may be minimal or nonspecific and confused with clinical conditions typical of birth and adaptation to the extrauterine environment, especially in premature infants. ^2,9 Therefore, there is a great challenge for healthcare professionals in identifying newborns with a high probability of early sepsis and initiating antimicrobial therapy as well as discontinuing this therapy when infection is considered unlikely.¹⁰

Several risk factors are involved in the genesis of early sepsis, being grouped into maternal or neonatal factors, highlighting premature labor, rupture of amniotic membranes 18 hours or more before delivery, chorioamnionitis, maternal colonization by EGB, maternal fever during or immediately after delivery,¹¹ premature newborns with low 5-minute Apgar scores and need for resuscitation at birth.^12,13

Most studies on this topic are concentrated in large centers, mainly in university hospitals, where research and prophylaxis for GBS in pregnant women are routinely carried out, and neonatal units have strict management for carrying out cultures and using antimicrobial agents, which does not occur in most maternal and child hospitals in the countryside of Brazil. Little is known about the true incidence of early neonatal sepsis in low- and middle-income countries.¹⁴ Therefore, this research aims to assess the incidence of early neonatal sepsis, the associated maternal and neonatal risk factors and the evolution of premature infants admitted to three NICUs in a city in the countryside of Bahia.

METHODS

This is a non-concurrent, hospital-based cohort study, including premature infants admitted to the three NICUs on the first day of life, from January 1, 2016 to December 31, 2017. The research was carried out in the city of Vitória da Conquista, the third largest city in the state of Bahia, headquarters of the Southwest Regional Health Center.

The study population was monitored until 27 days of life. Since this was a larger study, the exclusion criterion was being premature with a congenital anomaly (complex congenital heart disease, gastrointestinal tract atresia, abdominal wall defects, hydrocephalus, encephalocele and diaphragmatic hernia).

The data were obtained by analyzing medical records stored in the medical and statistical archive service of the three hospitals. NICUs have ten beds each, two of which are located in public hospitals and the other in a private hospital.

The sample was obtained by convenience (n=268). However, the smallest sample size necessary to represent the population of premature infants in the region was estimated at 120, considering the following parameters: infinite population size (given that it is not possible to estimate the total number of premature infants who would require neonatal intensive care); expected frequency of early neonatal sepsis of 8.5%, according to Barbosa et al. (2014), in Uberlândia;¹⁵ 5% accuracy; and 95% Confidence Interval.

Data were obtained through a specific questionnaire based on the Birth Survey for Brazil³ instrument by volunteer health researchers, after training and under the supervision of neonatologists. The main fieldwork took place from June 2018 to April 2019, using a digital questionnaire through the Kobo Toolbox 1.4.8^® software.

The dependent variable was early neonatal sepsis. The diagnostic description in medical records and the use of an empirical therapeutic regimen of ampicillin or crystalline penicillin G associated with gentamicin (protocol in the units) in the first 72 hours of life were considered.

Sepsis was categorized as presumed early-onset neonatal sepsis (PES) when clinical signs and symptoms compatible with the disease occurred, and an antimicrobial regimen was initiated. Confirmed early-onset neonatal sepsis (CNS) was considered if a positive blood culture was obtained. PES was divided into rule out neonatal sepsis (RNS), whose clinical and laboratory evolution allowed the suspension of antibiotics within four days, and treated neonatal sepsis (TNS), which were those who received the antimicrobial regimen between five days and more.¹⁶

Within the TNS group, a subdivision was made into clinically treated neonatal sepsis (CTNS) and clinically treated neonatal sepsis with at least one laboratory alteration (CTNSL). According to Procianoy et al. (2020), the number of leukocytes above 25,000 or below 5,000 and the immature neutrophil to total ratio (I/T) above 0.2 or C-reactive protein above 10.¹⁷The death certificates of premature infants with PES who died in the first four days of life were assessed. When the declared cause was early neonatal sepsis or septic shock, the TNS category was assigned to those who had been treated for less than four days. For these premature infants, a sensitivity analysis was performed by comparing the results of the analyses, including these premature infants in the group described above and excluding them from the analyses for subsequent decision-making.

For these premature infants, a sensitivity analysis was performed to compare the results obtained from the decision to include deaths in the category described above and exclude them from the analysis.

In cases of suspected early sepsis, it was routine in the three units to collect a blood culture sample before starting antimicrobial agents. Microorganism isolation was performed using the manual or automated blood culture method in the three units by an outsourced laboratory.

The independent variables analyzed were divided into two chunks. Chunk I contained maternal demographic characteristics, maternal morbidities, and prenatal and delivery care. The variables used were maternal age (<20 years, ≥20 years), marital status (with partner and without partner), preterm labor (no or yes), hypertensive syndrome (no or yes), number of prenatal consultations (up to five consultations or six and more consultations), and type of delivery (cesarean or vaginal).

Chunk II included the characteristics of premature infants, neonatal care received and clinical evolution, such as the sex of premature infants (male or female), estimated gestational age in weeks, categorized as extremely premature (less than 28 weeks), very premature (28 to less than 32 weeks) and moderate/late premature (32 to less than 37 weeks). Birth weight was measured in grams and divided into ≥ 1,500 g or < 1,500 g, 5^th-minute Apgar (< 7 or ≥ 7), hypothermia on admission to the NICU (no or yes). Clinical outcomes were assessed through the diagnosis of late neonatal sepsis and neonatal death, both categorized as no or yes.

In order to obtain information regarding the missing medical records, the nursing admission books of the NICUs were analyzed. The assessment of losses was made by comparing the sample obtained with the total population using Pearson’s chi-square test or linear trend.

Initially, a descriptive analysis of the variables studied was performed, presenting absolute and relative frequencies. All variables were also described according to the categories of early neonatal sepsis, such as absent, ruled out and treated (CTNS and CTNSL), compared using Pearson’s chi-square test or Fisher’s exact test. Sensitivity analysis was performed using Pearson’s chi-square test.

For the bivariate analysis, the early neonatal sepsis variable was recategorized as absent and treated (CTNS and CTNSL). The RNS category was removed from the analyses due to its behavior being different from the other sepsis categories (absent or TNS) and the insufficient number of observations for an analysis as an isolated category. The assessment of the progression to late neonatal sepsis and death was performed according to the groups of absent and treated early neonatal sepsis. Bivariate analysis between independent variables and treated early neonatal sepsis was performed using Poisson Regression with robust variance, obtaining estimates of the unadjusted Relative Risk (_uRR) and their respective 95% Confidence Intervals (95%CI).

For multivariate analysis, the independent variables from chunks I and II were used, which met the following criteria: p-value ≤ 20% (by the Wald test); loss of less than 10%; and the assumption of independence between variables. The models were compared by the Akaike criterion, and adequacy was assessed by the chi-square test. For all tests and for the permanence of the variables in the final model, a p-value ≤ 0.05 was used. The Stata, version 15.1 (Stata Corporation, College Station, USA), was used for data analysis.

This research was carried out taking into account the guarantee of ethical and legal principles that govern research on human beings, recommended in Resolutions 466/2012, 510/2016 and 580/2018 of the Ministry of Health, approved by the Research Ethics Committee, with Certificado de Apresentação para Apreciação Ética (CAAE, Certificate of Presentation for Ethical Consideration) 42401920.0.0000.5049, on February 5, 2018.

RESULTS

During the two years of study, 592 premature infants were admitted to the NICUs, 37 of whom were excluded due to congenital malformations, and 155 medical records were not located, leaving a sample of 400 premature infants. Medical records that were not located were assessed as possible losses. No differential loss was observed, and the use of calibration factors was not necessary to conduct the other analyses. The following variables were analyzed: hospital of origin (p=0.261); birth weight (p=0.917); gestational age (p=0.948); death (p=0.939); and according to 2016 and 2017 (p=0.827).

For the present study, 268 premature infants were included. Concerning maternal characteristics, most mothers were over 20 years old, lived without a partner and evolved to cesarean delivery. In relation to the characteristics of premature infants and neonatal care, the majority were male, classified as moderate/late premature, weighed over 1,500 g, had a 5-minute Apgar score greater than or equal to 7 and were hypothermic upon admission to the NICU (Table 1).

Fourteen premature deaths were recorded before 4 days of life, of which nine had early sepsis or septic shock as the cause of death (death certificate). These patients were removed from the RNS group and included in the CTNS or CTNSL groups. To perform a sensitivity assessment, the analyses were repeated excluding these nine newborns, and no differences were observed between the results. It was therefore decided to keep them in the analyses.

A total of 170 (63.4%) (95%CI: 57.4-69.0) premature infants had PES, of which 68 (25.4%) (95%CI: 20.4-30.9) had RNS and 102 (38.0%) (95%CI: 32.4-44.0) had TNS. Of the total number of premature infants in the study (268), 33 (12.3%) (95%CI: 8.8-16.8) had CTNS, and 65 (24.5%) (95% CI: 20.8-31.3) had CTNSL (Figure 1).

When comparing the groups and the variables of chunks I and II, it was found that there is an association between maternal age (p=0.008), hypertensive syndrome (p=0.014), premature labor (p=0.013), type of delivery (p<0.001), gestational age at birth (p<0.001), birth weight (p<0.001) and 5-minute Apgar (p=0.001) (Table 2).

The bivariate analysis of the variables in chunk I demonstrated a higher risk for TNS in mothers under 20 years of age, who lived without a partner, who developed premature labor and who had a vaginal delivery. And a protective factor for the outcome included mothers with some hypertensive syndrome during pregnancy and who had six or more prenatal consultations. Regarding the variables in chunk II, a positive association, gestational age less than 32 weeks, birth weight less than 1,500 g and 5-minute Apgar score less than 7 were evidenced.

The variables maternal age, hypertensive syndrome during pregnancy, type of delivery, gestational age, birth weight and 5-minute Apgar met the adopted criteria and were included as a chunk in the multivariate analysis. After adjustment, being born by vaginal delivery, with a gestational age of less than 32 weeks and a 5-minute Apgar score of less than 7 remained significantly associated with the outcome (Table 3).

Late-onset neonatal sepsis and neonatal death occurred in 41 (40.2%) and 25 (24.5%) premature infants, respectively, who had TNS. The two conditions analyzed were significantly associated with the outcome (Table 4).

DISCUSSION

This study demonstrates the difficulties faced in the management of early neonatal sepsis in premature infants, in which it was found that most of them had clinical signs and symptoms that motivated the introduction of antimicrobial agents. As a result of rigorous clinical and laboratory monitoring, it was possible to suspend antibiotics in 25.4% of cases. Different results were verified by the RBPN, in which 74.3% of premature infants did not require antibiotics in the first 48 hours of life, and it was possible to interrupt them in 44% of cases within 48-72 hours. This demonstrates the impact of the strategies used in the institution for the rational use of antimicrobial agents.²

Despite the routine collection of blood cultures in units before starting antibiotics, no positive culture results were detected, making diagnostic accuracy more difficult. Obtaining insufficient blood volume (due to prematurity) and the absence of automated systems (due to a lack of vials in public services) in all cultures may have influenced the results. No similar results were identified in this cohort. International and national studies demonstrate low culture positivity in early neonatal sepsis, being 5.93% in India (2017),¹⁸ 2.9% in Uberlândia (2011)¹⁵ and 1.3% in Campinas (2006-2017).²

It was found that 12.3% of premature infants treated did not have any laboratory abnormalities identified and were treated based on clinical manifestations. This is the difficulty in achieving diagnostic accuracy, since the signs and symptoms may be minimal or nonspecific and confused with other non-infectious inflammatory syndromes, and diagnostic tests have a low positive predictive value. Therefore, the decision to treat a newborn also depends on other factors, such as the presence of maternal risk factors, the frequency of observations and the degree of prematurity of newborns.^2,9,10

The incidence of early sepsis in this cohort was similar to that observed in public hospitals in Ethiopia in 2019 (38%),¹³ but high when compared with other studies. In England, the incidence in extremely premature infants (2007-1016) was 3.7%.⁸ In Brazil, in 2011, 8.5% of newborns presented the aforementioned outcome.¹⁵

It is worth highlighting the methodological differences found in several studies on this topic. Some studies only include CNS through culture, others include all neonates, which made the comparative aspect difficult. Also regarding the calculation of the incidence of sepsis, most sources demonstrate the rates for every 1,000 live births.

In this cohort, being born vaginally remained an independent factor for early sepsis. This finding was also evidenced in other studies.^12,19 This type of delivery may be associated with prolonged rupture of the amniotic membranes (> 18 hours), chorioamnionitis or exposure to GBS^11,13. In this cohort, being born at a gestational age of less than 32 weeks was associated with an increased risk of early sepsis, which can be supported by several international and national studies.^{2,10,13,20,21} It is known that premature infants present immunological dysfunction due to the absence of maternal transplacental transfer of IgG, immature cellular responses, and deficiencies of soluble proteins and peptides.²¹

Being born with a 5-minute Apgar score of less than 7 led to a 45% higher risk of early-onset sepsis. Similar results were found in public hospitals in southern Ethiopia.¹² This association can be explained by the fact that perinatal asphyxia causes an immunological insult, contributing to a worsening of the response to combat infections in premature infants who already have an impaired innate immune status.^{11, 22}

It was found that premature infants with early neonatal sepsis had a 6.5 and 12.0 times greater risk of developing late-onset sepsis and neonatal death, respectively. These findings support the research conducted by the Eunice Kennedy Shriver National Institute of Child Health and Human Development Neonatal Research Network, which followed premature infants treated for early-onset sepsis. The presence of this diagnosis was associated with late-onset sepsis, necrotizing enterocolitis and death.²³

Following the Ministry of Health recommendations, there is no routine screening for GBS in prenatal care provided by the Unified Health System in this region.²⁴ This contributes to the lack of knowledge about the prevalence of colonization of pregnant women by this agent combined with the negativity of all blood cultures, which led to difficulties in identifying the microorganisms responsible for sepsis. In order to clarify this gap, a pioneering study carried out in this city screened GBS in pregnant women and identified a prevalence of 18.1%, highlighting the importance of implementing this routine in prenatal care practices in the public health system.²⁵

Among the limitations of this study, retrospective design with possible information bias, in addition to not obtaining some important maternal variables for the outcome studied, such as time of rupture of amniotic membranes, status of colonization by EGB, use of antimicrobial agents by the mother and diagnosis of chorioamnionitis, stood out.

Early-onset neonatal sepsis is one of the main diagnoses in NICUs. Its management is also one of the greatest challenges in neonatology, given that the clinical signs and symptoms are nonspecific, especially in premature infants, and can be confused with non-infectious conditions, combined with the low sensitivity of laboratory tests. In this cohort, a high incidence of TNS was identified, with most premature infants requiring antibiotics after birth and no microorganisms being isolated in blood cultures. The independent risk factors for early-onset neonatal sepsis include being born by vaginal delivery, with a gestational age of less than 32 weeks and a 5-minute Apgar score below 7. Furthermore, early-onset sepsis behaved as a risk factor for late-onset neonatal sepsis and death. These findings demonstrate the need to improve the quality of prenatal care, strategies to prevent prematurity, and management during birth to avoid perinatal asphyxia.

REFERENCES

1. Teixeira JAM, Araujo WRM, Maranhão AGK, et al. Mortalidade no primeiro dia de vida: tendências, causas de óbito e evitabilidade em oito Unidades da Federação brasileira, entre 2010 e 2015. Epidemiologia e Serviços de Saúde 2019; 28, e2018132. https://doi.org/10.5123/S1679-49742019000100006.

2. Caldas JPDS, Montera LC, Calil R, et al. Temporal trend in early sepsis in a very low birth weight infants' cohort: an opportunity for a rational antimicrobial use. Jornal de Pediatria 2021; 97: 414-419. https://doi.org/10.1016/j.jped.2020.07.006.

3. Lansky S, Friche AAL, Silva AAM, et al. Pesquisa Nascer no Brasil: perfil da mortalidade neonatal e avaliação da assistência à gestante e ao recém-nascido. Cad Saude Publica 2014; 30:S192-S207. https://doi.org/10.1590/0102-311X00133213.

4. Carlo WA, Travers CP. Maternal and neonatal mortality: time to act. J Pediatr 2016; 92(6):543-5. https://doi.org/10.1016/j.jped.2016.08.001.

5. Klein JO. Bacteriology of neonatal sepsis. Pediatr Infect Dis J [Internet] 1990 [citado 2023 mai 6]; 9: 777-778. Disponível em: https://journals.lww.com/pidj/Citation/1990/10000/Bacteriology_of_neonatal_sepsis.39.aspx

6. Simonsen KA, Anderson-Berry AL, Delair SF, et al. Early-onset neonatal sepsis. Clinical Microbiology Reviews 2014; 27(1): 21-47. https://doi.org/10.1128/cmr.00031-13.

7. Weston EJ, Pondo T, Lewis MM, et al. The burden of invasive early-onset neonatal sepsis in the United States, 2005–2008. The Pediatric Infectious Disease Journal 2011; 30(11): 937. https://doi.org/10.1097%2FINF.0b013e318223bad2

8. Boel L, Banerjee S, Clark M, et al. Temporal trends of care practices, morbidity, and mortality of extremely preterm infants over 10-years in South Wales, UK. Scientific Reports 2020; 10(1):1-9. https://doi.org/10.1038/s41598-020-75749-4.

9. Odabasi IO, Bulbul A. Neonatal Sepsis. Med Bull Sisli Etfal Hosp 2020; 54(2): 142-158. https://doi.org I: 10.14744/SEMB.2020.00236

10. Flannery DD, Mukhopadhyay S, Morales KH, et al. Delivery characteristics and the risk of early-onset neonatal sepsis. Pediatrics 2022; 149(2): e2021052900. https://doi.org/10.1542/peds.2021-052900.

11. ANVISA. Agência Nacional de Vigilância Sanitária. Ministério da Saúde. Gerência de Vigilância e Monitoramento em Serviços de Saúde (GVIMS). Gerência Geral e Tecnologia em Serviços de Saúde (GGTES). Critérios Diagnósticos de Infecção Associada à Assistência à Saúde-Neonatologia, volume 3. Brasília: Ministério da Saúde [Internet] 2017 [citado 2023 fev 5]. Disponível em: https://bvsms.saude.gov.br/bvs/publicacoes/criterios_diagnosticos_infeccoes_assistencia_saude_neonatologia.pdf

12. Teshome G, Kabthymer RH, Abebe M, et al. Factors associated with early onset neonatal sepsis among neonates in public hospitals of Sidama region, Southern Ethiopia, 2021: Unmatched case control study. Annals of Medicine and Surgery 2022; 81:104559. https://doi.org/10.1016/j.amsu.2022.104559

13. Akalu TY, Aynalem YA, Shiferaw WS, et al. Prevalence and determinants of early onset neonatal sepsis at two selected public referral hospitals in the Northwest Ethiopia: a cross-sectional study. BMC Pediatrics 2023; 23(1): 1-9. https://doi.org/10.1186/s12887-022-03824-y.

14. Sands K, Spiller OB, Thomson K, et al. Early-onset neonatal sepsis in low-and middle-income countries: Current challenges and future opportunities. Infection and Drug Resistance 2022; 15: 933-946. https://doi.org/10.2147/IDR.S294156.

15. Barbosa NG, Reis H, Resende DS, et al. Sepse neonatal precoce em unidade de terapia intensiva neonatal de um hospital universitário terciário. Pediatr. Mod [Internet] 2014 [citado 2023 fev 10]; 50(4). Disponível em: https://pesquisa.bvsalud.org/portal/resource/pt/lil-712046

16. Wynn JL, Wong HR, Shanley TP, et al. Time for a neonatal–specific consensus definition for sepsis. Pediatric Critical Care Medicine 2014; 15(6): 523. https://doi.org/10.1097/pcc.0000000000000157.

17. Procianoy RS, Silveira RC. The challenges of neonatal sepsis management. Jornal de Pediatria 2020; 96: 80-86. https://doi.org/10.1016/j.jped.2019.10.004.

18. Meshram RM, Gajimwar VS, Bhongade SD. Predictors of mortality in outborns with neonatal sepsis: A prospective observational study. Nigerian Postgraduate Medical Journal 2019; 26(4): 216. https://doi.org/10.4103/npmj.npmj_91_19.

19. Gómez JL, González SC. Asociación de factores obstétricos y neonatales con casos de sepsis neonatal temprana. Cartagena, Colombia. Revista Habanera de Ciencias Médicas [Internet] 2018 [citado 2022 dez 10]; 17(5): 750-763. Disponível em: https://www.medigraphic.com/pdfs/revhabciemed/hcm-2018/hcm185j.pdf

20. Melville JM, Moss TJ. The immune consequences of preterm birth. Frontiers in Neuroscience 2013; 7: 79. https://doi.org/10.3389/fnins.2013.00079.

21. Palatnik A, Liu LY, Lee A, et al. Predictors of early-onset neonatal sepsis or death among newborns born at < 32 weeks of gestation. Journal of Perinatology 2019; 39(7): 949-955. https://doi.org/10.1038/s41372-019-0395-9.

22. Yismaw AE, Abebil TY, Biweta MA, et al. Proportion of neonatal sepsis and determinant factors among neonates admitted in University of Gondar comprehensive specialized hospital neonatal Intensive care unit Northwest Ethiopia. BMC Research Notes 2019; 12(1): 1-5. https://doi.org/10.1186/s13104-019-4587-3

23. Kuppala VS, Meinzen-Derr J, Morrow AL, et al. Prolonged initial empirical antibiotic treatment is associated with adverse outcomes in premature infants. The Journal of Pediatrics 2011; 159(5): 720-725. https://doi.org/10.1016/j.jpeds.2011.05.033.

24. Brasil. Ministério da Saúde. Cadernos de Atenção Básica. Atenção ao Pré-Natal de Baixo Risco. Secretaria de Atenção à Saúde. Volume 32. [Internet]. Brasília: Ministério da Saúde; 2012. [citado 2023 jan 15]. 311.p. Disponível em: https://bvsms.saude.gov.br/bvs/publicacoes/cadernos_atencao_basica_32_prenatal.pdf

25. Oliveira TVLD, Santana FAF, Souza CL, et al. Prevalência e fatores associados a colonização por estreptococo do grupo B em gestantes. Revista Brasileira de Saúde Materno Infantil 2021; 20: 1165-1172. https://doi.org/10.1590/1806-93042020000400013.

IR