Introduction

Bacterial meningitis (BM) is the leading cause of morbidity and mortality in infants [1]. There are many bacterial etiologic agents of BM in this age group [1, 2]. Among these, GBS (also known as Streptococcus agalactiae) and Escherichia coli are the predominant causes of illness. Although they are the major causes of death in infants, there is no any approved vaccine available for these pathogens yet [3]. About 20–30% of healthy women carry GBS on their vaginal or rectal mucosa [4]. This maternal colonization is the primary means for newborn GBS acquisition either via ascending infection or during birth through the infected birth canal as neonates aspirate contaminated amniotic or vaginal fluids [5]. The invasive GBS infections can be categorized into early-onset disease (EOD), late-onset disease (LOD), and rarely as ultra-late onset disease (ULOD). The EOD occurs within the first six days of life while the LOD and the ULOD occurs from 7 up to 89 days, and after 3 months of age respectively [6, 7]. Recently, GBS has emerged as a leading cause of meningitis in neonates and young infants, resulting in high mortality [8]. However, GBS is not still considered as the major bacterial etiological agent of infant sepsis and meningitis in most of the developing countries including Ethiopia [9]. Thus, this study was aimed to determine the proportion of GBS associated meningitis among Ethiopian infants using PCR. The data generated from this study will used by policymakers to better understand the epidemiology of the disease in the local setting and set disease prevention and control strategies.

Results

The majority of the study participants, 61.1% (n = 44), were neonates (aged ≤ 28 days) and the remaining 38.9% (n = 28) of them were infants who are aged >28 days. The mean age at the onset of meningitis was 49.8 ± 83.6 days (1–365 days). The male to female proportion was nearly 1.7:1 i.e. 62.5% being male infants. Fever was one of the early clinical events expected to occur in most cases of meningitis. In this study, 43% (n = 31) of the total infants were observed to have a fever. Overall, vomiting or reduced ability to suck breastfeeding (n = 45), fever (n = 31), and altered consciousness (n = 27) were the most frequent clinical presentations noted in this study. Of the clinical presentations examined in infants; 61.3% (n = 19) were with fever, 48.1% (n = 13) with impaired consciousness, and 55.6% (n = 25) who have had vomiting or reduced ability to suck for breastfeeding were positive for GBS (Table 1). Most of the infants; 54/72 (75%), had a history of antibiotics use before the spinal tap procedure, which could decrease GBS culture recoverability. According to the clinical disease onset by age, most of the infants were classified under the LOD as observed by the proportion of 76.4% (55/72). While the proportion of EOD and the ULOD was 12.5% (9/72) and 11.1% (8/72) respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Bivariate analysis of demographic and clinical features of infants suspected with meningitis at TASH from June 2018 to Oct. 2019, Addis Ababa, Ethiopia.

https://doi.org/10.1371/journal.pone.0242628.t001

The CSF white cells count, protein, and glucose concentrations were obtained for more than half of the infants. Unlike most criteria’s used to exclude bacterial meningitis, GBS was detected in more than 67.4% (n = 29) of infants without pleocytosis (as defined by CSF cells count of ≥ 20 cells /μl) in our study. Of the infants with available glucose concentration data, 8 (out of 35) had below 40 mg/dl, while infants with protein concentration data 16 (out of 46) had above 100mg/dl, as expected in bacterial meningitis cases.

Similarly, the appearance of CSF was noted and nearly 78% of all the samples were found to be clear. Most of the confirmed GBS cases in our study have presented with normal CSF parameters in terms of cell count, protein, and glucose concentrations (see Table 1). This could question the reliability of CSF parameters in the prognosis of BM. The overall magnitude of GBS in infants with suspected meningitis was 63.9% (46/72). Proportionally higher prevalence of GBS was noted in young infants aged > 7 days (Table 1). Having impaired consciousness and Outcome of infants (death) was statistically associated with GBS positivity.

Death was observed in 6 of the 72 enrolled study infants. Death might be attributed by GBS in 10.9% (5out of 46 confirmed GBS cases) meningitis cases. Four of the dead infants were in the age group less than or equal to 28 days, while one was 44 days old and the other one was 4 months old. Considering the rate of GBS positivity (GBS vs. no GBS), the probable death was assumed (10.9% (5/46) vs. 3.8% (1/26)).

Discussion

In the present study, GBS was found high in infants with suspected meningitis cases, 63.9% (46/72), which is quite comparable with the study done in Korea 69% [15] and lower the USA 86.1% [16]. The finding was higher relative to the study carried out in China (46.5%) [17], in USA (53.6%) [18], in England (52%) [19], in South Africa (44%) [20] and in Malawi (26.3%) [8]. The difference observed may be due to the detection method utilized, maternal use of intra partum antibiotic prophylaxis (IAP), study setting differences, or population variations. Possibly inclusion of infants’ age beyond 3 months and the use of a more sensitive detection tool in our study might have contributed to the increment of the proportion of GBS detection.

Detection through cfb gene-targeted PCR was found to be too enhanced from all culture-negative results of GBS to 63.9%. Similar studies have also demonstrated the positivity of culture-negative CSF samples when tested with more sensitive techniques like PCR. Culture finding could be negative, possibly due to previous antibiotic exposure before CSF collection, the low number of the bacterium, or lower inoculum size. A comparative study in Ireland has shown the reputation of CSF PCR over culture. In contrast to culture which is a considered as a gold standard tool, 77.3% more detection capacity was offered by PCR [21]. In agreement to our study, other studies in Brazil (3.8% vs. 29.2%) [22], Iran (27.7% vs. 43.8%) [23], China (0% vs. 50.7%) [24] and Jimma, Ethiopia (3% vs. 33%) were also indicated the sensitivity of PCR over culture [25].

Late-onset GBS disease (65.5%) was observed high relative to EOGBS (44.4%) (Table 1). This study finding was relatively lower compared to results in China (74%) [26] and in the US (86.5%) [16]. High community or nosocomial acquisition, exposures to colonized parents and siblings, breastfeeding during maternal mastitis, and or prematurity could be the possible reasons [27, 28]. Although our sample size was too small, infants presented with LOGBS were more likely to have death outcomes, i.e. 3 out of 5 GBS positive deaths were due to LOGBS. Similar studies have shown that the LOGBS causes higher morbidity and mortality than EOGBS and ULOGBS at the time of discharge [27, 28]. A finding from Australia has verified that infants with LOD cases were 3 times more likely to develop neurodevelopmental sequelae or die in the course of admittance than ULOD, this reflects the vulnerability of infants in the first 90 days of life [28].

Disease presentations like fever was anticipated to occur in most cases of meningitis [29]. The current study noted vomiting or reduced ability to suck breastfeeding (n = 45), followed by fever (n = 31) and altered consciousness (n = 27) as the most frequent clinical presentations. Cerebrospinal fluid parameters such as protein, glucose, and white cell count (WCC) were used to assess the likely etiological agents, provide a protein concentration <60 mg/dl and WCC <90 cells/ μl were found to be optimal cut-offs for excluding bacterial meningitis [30]. Opposing this conclusion, the present study found more than 67% of infants positive for GBS had WCC < 20 cells/ μl. A similar study in India has also described that 16% of infants with GBS had 0–20 cells/ul [31], and in the USA nearly 11% of cases lacked CSF pleocytosis [18].

The probable mortality of GBS meningitis in the current study was 10.9% (5/46), lower compared with the study carried out in South Africa (18%) [20]. This might be due to the inclusion of all the invasive GBS infections (including sepsis and meningitis) in South Africa. The use of limited sample size, unavailability of data of some infants who are transferred out, lost with their appointment, and withdraw against medical advice, may have underestimated the real burden of mortality in the current study.

This study tried to address the magnitude of GBS meningitis in infants and gave an insight into how more sensitive tools would implicate for culture-negative suspected meningitis cases. However, the main constraint to this study was the use of limited sample size and the inability of our study to address potential risk factors such as mode of delivery, maternal colonization, maternal mastitis, birth gestational age, and birth weight as this study used secondary data regarded factors were not available.

Conclusions

In this study, a high proportion of GBS was detected among infants suspected of meningitis at TASH, Addis Ababa, Ethiopia. Accordingly, GBS was found out to be the major etiologic agent of BM among neonates and young infants. The current study also showed that the onset of disease matters to the outcome of patients, indicating three out of five died infants had a late-onset presentation. In addition, the role of GBS can be underestimated by using culture detection techniques in contrast to molecular assays like PCR. This was observed by a high positivity rate (63.9%) of GBS among participants who were culture negative and a history of antibiotic exposure before the spinal tap. We recommend further studies of this kind with a larger sample sizes to be conducted so as to devise the basic prevention and control strategies.

Acknowledgments

We would like to thank the nursing and laboratory staff at TASH for their contribution in this study through data collection and specimen handling as well as processing, and the archive staff of TASH for their support in cross-checking missed CSF laboratory test results. Our gratitude also goes to the molecular lab staff of AHRI for their provision of technical updates on the molecular laboratory work. We are also grateful to all parents and infants that participated in our study.