Clinical Features and Risk Factors of Group B Streptococcal Meningitis in Infants Younger than 4 Months: A 14-Year Experience at a University Hospital
Article information
Abstract
Purpose
Group B Streptococcus (GBS) is a leading cause of invasive infections in infants. However, data on GBS meningitis in young infants remain limited. This study investigated the clinical features and risk factors of GBS meningitis in infants younger than 4 months.
Methods
Medical records were retrospectively reviewed at a university hospital over a 14-year period from 2009 to 2022. Fifty cases of invasive GBS infection were identified among 32,851 young infants. Patients were divided into a meningitis group (group M; n=20) and a sepsis without meningitis group (group S; n=30).
Results
The proportion of outborn patients (i.e., those born outside the authors’ hospital) was higher in group M than in group S (75% vs. 43%). The proportion of patients with late-onset disease, defined as onset ≥7 days after birth, was also higher in group M than in group S (85% vs. 57%). Among early clinical symptoms and laboratory findings, only an elevated C-reactive protein (CRP) level was associated with meningitis (odds ratio, 14; 95% confidence interval, 1.1 to 183.8). Hospital outcomes, including 30-day mortality, did not differ between the groups. Most patients with GBS meningitis had abnormal brain imaging findings.
Conclusion
GBS meningitis was more common among outborn infants than in those born at the authors’ hospital and frequently developed as late-onset disease. An elevated CRP level in the early stage of infection was associated with the development of meningitis.
Introduction
Group B Streptococcus (GBS) is a common cause of invasive infections, including sepsis and meningitis, in neonates and young infants [1]. Since intrapartum antibiotic prophylaxis (IAP) was introduced in 1996 for mothers colonized with GBS, the incidence of invasive GBS infection has decreased [2]. However, clinical guidelines for identifying candidates for IAP differ among countries, including recommendations on microbiological screening during pregnancy and the use of perinatal risk factors [3]. Reported rates of maternal GBS colonization also vary geographically, ranging from 5% to 30% [2]. In Korea, maternal colonization rates of 8%–10% have been reported, which are considered relatively low compared with those in other countries [4]. However, a multicenter study found that GBS accounted for 48% of invasive bacterial infections in Korean infants younger than 3 months, indicating that GBS is a major cause of systemic bacterial infection in this population [5].
Invasive GBS infection is classified by time of onset as early-onset disease (EOD) or late-onset disease (LOD). EOD results from ascending infection through the birth canal and exposure to contaminated amniotic fluid during the intrapartum or peripartum period. LOD develops through several mechanisms, including postnatal exposure to the microorganism in the community or hospital, delayed manifestation of infection after prior colonization, and transmission through breast milk [6]. GBS meningitis usually develops through hematogenous spread of bacteremia to the central nervous system (CNS) and is more commonly associated with LOD than with EOD [7]. Although intensive care for invasive bacterial infections has advanced, mortality among infants with GBS meningitis remains high. Survivors also frequently show abnormal neuroimaging findings and have a high risk of long-term neurological sequelae [7,8].
The clinical findings of meningitis in infants resemble those of sepsis. Therefore, differentiating these two conditions on the basis of early clinical symptoms or laboratory findings is difficult. Meningitis is diagnosed primarily through cerebrospinal fluid (CSF) analysis after lumbar puncture; however, lumbar puncture is often delayed in clinically unstable infants [9]. Molecular diagnostic assays (MDAs), including latex agglutination (LA) and polymerase chain reaction (PCR), can diagnose meningitis more rapidly than CSF culture and are useful in infants who received empirical antibiotics before CSF analysis. However, because universal guidelines for the use of MDAs have not been established, clinical practice varies among institutions [7,10]. Therefore, this study investigated the clinical features of GBS meningitis in young infants, including CSF examination and brain imaging findings, and identified factors associated with its development.
Materials and Methods
1. Study design and population
Medical records and radiological findings were retrospectively reviewed over a 14-year period from January 2009 to December 2022 for infants younger than 4 months who were diagnosed with GBS infection and admitted to a university hospital in Korea. Patients with clinical symptoms of systemic infection and GBS isolated from CSF or blood samples were considered to have invasive GBS infection. Patients with GBS isolated only from umbilical cord blood samples, no clinical symptoms, and negative subsequent peripheral blood cultures were excluded because these cases were considered contaminated. This study was approved by the Keimyung University Dongsan Hospital Institutional Review Board (DSMC 2024-01-012). Informed consent was waived by the board.
Among patients with invasive GBS infection, meningitis was classified into three diagnostic categories: bacterial isolation from a CSF sample (type A), bacteremia with CSF pleocytosis (type B), and CSF pleocytosis with a positive MDA result in CSF (type C) [8,11,12]. Patients with invasive GBS infection were divided into two groups: patients with meningitis (group M) and patients with sepsis without meningitis (group S). Perinatal characteristics, early clinical and laboratory findings, antibiotic susceptibility, and hospital course were compared between the groups. In addition, CSF diagnostic findings and brain imaging findings were examined in group M.
2. Definitions
Infection type was classified as EOD (<7 days after birth) or LOD (≥7 days after birth) according to the time of onset [6]. Inborn and outborn status was determined according to the place of birth. Inborn infants were defined as those delivered at our hospital, whereas outborn infants were defined as those delivered at another hospital or at a location other than our hospital. Abnormal CSF laboratory findings were defined as pleocytosis (neonate, ≥16/mm3; infant, ≥10/mm3), neutrophil predominance (>50%), decreased glucose level (neonate, <25 mg/dL; infant, <27 mg/dL), elevated protein level (neonate, ≥128 mg/dL; infant, ≥100 mg/dL), bacterial detection by Gram stain or culture, or a positive MDA result, including LA or PCR [10,13]. Apnea was defined as cessation of breathing for more than 20 seconds or for a shorter duration when accompanied by cyanosis or bradycardia [14]. Only apneic episodes associated with invasive infection were included to minimize the inclusion of apnea of prematurity. Hypotension was defined as low mean blood pressure (neonate, <30 mm Hg; infant, <35 mm Hg) [15,16]. Seizure was diagnosed clinically on the basis of recurrent abnormal paroxysmal events, including focal or multifocal clonic, focal tonic, and subtle manifestations [17]. Severe infection was defined as invasive infection with any of the following: (1) hypotension requiring intravenous fluid therapy and inotropic support, (2) acute respiratory failure requiring mechanical ventilation, or (3) multiorgan dysfunction involving at least two of the following organs or systems: lungs, kidneys, CNS, blood, and liver [18]. Septic shock was defined as invasive infection with persistent hypotension despite treatment or circulatory failure indicated by at least two of the following: metabolic acidosis, oliguria, elevated serum lactate level, or skin mottling [18]. Acute kidney injury (AKI) was defined as both oliguria (<1 mL/kg/hr) and elevated serum creatinine level (>1.5 mg/dL) [19]. Fulminant infection was defined as infection resulting in death within 48 hours after onset [20]. Other terms are described in the supplementary file.
3. Statistical analysis
The collected data were analyzed using IBM SPSS Statistics version 21.0 (IBM Corp., Armonk, NY, USA). The chi-square test and Fisher exact test were used for categorical variables, and the Mann-Whitney U test and the t-test were used for continuous variables. Binomial confidence intervals were calculated using the Clopper-Pearson exact method. Multiple logistic regression analysis was performed to identify factors associated with GBS meningitis. P values <0.05 were considered statistically significant.
Results
1. Study population
Among 32,851 infants younger than 4 months, 52 (0.2%) had GBS isolated from blood or CSF samples. Two patients had GBS isolated from initial umbilical cord blood cultures but had no clinical symptoms of invasive infection and had negative subsequent peripheral blood cultures; these cases were therefore considered contaminants. Among the 50 patients with invasive GBS infection, 20 were assigned to group M and 30 to group S. Some patients with invasive GBS infection also had cellulitis, pneumonia, or urinary tract infection.
2. Clinical characteristics and outcomes of patients with meningitis
When group M was classified by time of infection onset, three patients (15%) had EOD and 17 (85%) had LOD. Bacteremia was present in 19 cases (95%), and GBS was isolated from CSF samples in 17 cases (85%). CSF MDA was performed in 14 cases, all of which were positive. Neuroimaging was performed using ultrasound (US) or computed tomography (CT) in the early stage of infection, and magnetic resonance imaging (MRI) was performed in the late stage of infection and after discharge. Imaging was performed in 18 cases (90%) in group M, and most cases showed abnormal findings. Early brain imaging with US or CT was performed in 10 patients (50%) in group M. The most common abnormal findings were cerebral edema and subdural effusion (two cases each), followed by subdural empyema and intraventricular hemorrhage (one case each). Three patients (15%) in group M died during hospitalization, and two of them showed diffuse cerebral edema in the early stage of infection. Two other patients were transferred to other hospitals during treatment (Table 1).
Clinical data of patients with group B streptococcal meningitis
3. Diagnostic types of meningitis and cerebrospinal fluid study findings
According to the diagnostic criteria for meningitis, type A was the most common diagnostic category (17 cases, 85%), followed by type B (two cases) and type C (one case) (Table 2). In the analysis of the diagnostic sensitivity of CSF parameters, MDA showed the highest sensitivity (100%), followed by culture, pleocytosis, and neutrophil predominance (85% each), elevated protein level, and decreased glucose level. The diagnostic sensitivity of MDA was significantly higher than that of Gram stain and decreased glucose level (P<0.001) (Table 3).
Types of group B streptococcal meningitis according to diagnostic criteria
Diagnostic sensitivity of cerebrospinal fluid parameters suggesting group B streptococcal meningitis
4. Comparison of clinical findings in patients with invasive group B streptococcal infections
No significant differences in perinatal characteristics, including gestational age, birth weight, mode of delivery, and sex, were observed between the two groups. However, the proportion of outborn infants was significantly higher in group M than in group S (75% vs. 43.3%, P<0.05). Similarly, LOD was significantly more frequent in group M than in group S (85% vs. 56.7%, P<0.05). Among the initial symptoms after hospitalization, fever, decreased activity, seizure, bradycardia, and apnea were more common in group M than in group S (P<0.05). The duration of fever before hospitalization was longer in group M than in group S (1.4±1.0 days vs. 0.6±0.5 days, P<0.005). Among the initial laboratory findings after hospitalization, peripheral blood leukopenia and elevated C-reactive protein (CRP) level were observed more frequently in group M than in group S (P<0.05). Initial quantitative CRP values were also significantly higher in group M than in group S (7.5±7.3 mg/dL vs. 1.2±1.6 mg/dL, P<0.005). Infection severity, evaluated according to the rates of severe infection, septic shock, AKI, and fulminant infection, did not differ significantly between the two groups. Among other comorbidities, bacteremic pneumonia occurred only in group S (six cases, 12%). Urinary tract infection (two cases, 4%) and cellulitis (six cases, 12%) occurred in both groups. The 30-day mortality rate was 10%, with no significant difference between the two groups (Table 4).
Comparison of clinical findings in patients with invasive group B streptococcal infections
5. Risk factors for group B streptococcal meningitis
Multiple logistic regression was performed using variables that were significant in the univariate analysis of clinical and laboratory findings. An elevated CRP level in the early stage of infection (odds ratio, 14; 95% confidence interval, 1.1 to 183.8) was identified as an independent factor associated with GBS meningitis (Table 5).
Multiple logistic regression analysis of risk factors for group B streptococcal meningitis
6. Antimicrobial susceptibility of group B Streptococcus in patient isolates
All GBS isolates from the 50 patients were susceptible to ampicillin and vancomycin. The susceptibility rate for cefotaxime was 97.4%, whereas those for clindamycin and erythromycin were both 48%. No significant differences in antimicrobial susceptibility were observed between the two groups (Supplementary Table 1).
7. Brain MRI findings in patients with group B streptococcal meningitis
MRI was performed after antibiotic treatment in 14 patients in group M (70%), and abnormal findings were observed in nine of these patients (64.3%). The most common abnormality was leptomeningeal enhancement (five cases), followed by subdural empyema (four cases), encephalomalacia (three cases), and brain abscess, brain atrophy, cerebral infarction, and subdural effusion (two cases each) (Supplementary Table 2).
Discussion
The incidence of invasive GBS infection, including sepsis and meningitis, in neonates and young infants has been reported to vary by country and region, ranging from 0.5 to 2 cases per 1,000 live births [6]. This study evaluated infants younger than 4 months over a 14-year period and identified 50 patients with invasive GBS infection, representing approximately 0.2% of all hospitalized infants in this age group. Although this proportion is higher than rates reported in previous studies, it should be interpreted cautiously because the study population was limited to inpatients at a tertiary hospital. In this study, meningitis occurred predominantly as LOD, consistent with previous studies [11,21]. Although one patient in the meningitis group did not have bacteremia, the patient had received oral antibiotics for 2 days at a primary clinic; therefore, the negative blood and CSF culture results were likely false negatives. In this case, meningitis was diagnosed based on CSF pleocytosis and a positive CSF PCR result [8,11]. Cerebral edema, a potential complication of bacterial meningitis, is an unfavorable prognostic factor mediated by mechanisms such as increased intracranial pressure, reduced cerebral blood flow, and brain herniation [22]. In this study, early brain imaging with US or CT was performed in 10 patients (50%) in group M, and diffuse cerebral edema was observed in two of the three fatal cases. Therefore, neuroimaging should be performed promptly in the early stage of meningitis, and appropriate treatment for cerebral edema, such as intravenous steroid therapy, should be considered [7,23].
The gold standard for diagnosing bacterial meningitis in infants is a positive CSF culture. However, other CSF parameters, including Gram stain, white blood cell count, neutrophil predominance, glucose level, and protein level, are commonly assessed to support rapid diagnosis. Fleischer et al. [13] reported diagnostic sensitivities for rapid CSF parameters in infants with culture-positive bacterial meningitis of 80.3% for pleocytosis, 90.2% for neutrophil predominance, 49.2% for low glucose level, 73.0% for elevated protein level, and 71.9% for positive Gram stain. CSF MDAs, including LA and PCR, are rapid diagnostic tests. They can also serve as complementary diagnostic tools when false-negative culture results are likely, such as when the mother or infant received antibiotics before lumbar puncture [7,10]. In a study by Min et al. [12] that included 19 cases of GBS meningitis in neonates and infants, the diagnostic sensitivity of LA (89.5%) was higher than that of CSF culture (68.4%).
Morrissey et al. [24] reported that PCR had a diagnostic sensitivity of 100% and specificity of 99.9% compared with CSF culture for the diagnosis of GBS meningitis in infants. Therefore, CSF PCR is a reliable and useful tool for diagnosing GBS meningitis. In the present study, the diagnostic sensitivities of CSF parameters among 20 patients with GBS meningitis were 85% each for culture, pleocytosis, and neutrophil predominance; 73.7% for elevated protein level; 45% for Gram stain; and 42.1% for low glucose level. These findings were similar to those of a previous study [13]. Although MDA was performed in only 14 of the 20 cases, all MDA results were positive, and its diagnostic sensitivity was significantly higher than those of Gram stain and low glucose level. Therefore, clinical guidelines for CSF evaluation in bacterial meningitis should include MDAs, such as PCR, in addition to conventional CSF parameters to support rapid and accurate diagnosis [10,24].
Distinguishing meningitis from sepsis in infants with invasive GBS infection is often difficult on the basis of early clinical signs or initial laboratory findings. Hsu et al. [25] reported that neonates with GBS meningitis more frequently presented with apnea, bradycardia, and abdominal distension as early clinical signs than neonates with GBS sepsis without meningitis. Their laboratory findings were also more likely to show leukopenia, acidosis, elevated CRP level, and coagulopathy, and they had higher frequencies of seizures, ventilator requirement, and severe sepsis during hospitalization. In the present study, several early clinical symptoms, including fever, decreased activity, and apnea, were more frequent in the meningitis group than in the non-meningitis group. In addition, leukopenia and elevated CRP level were observed more frequently in the meningitis group. These results are consistent with previous reports [25,26]. Neonatal seizures are diagnosed based on electroclinical or electrographic findings [27]. However, in the present study, electroencephalography was performed during the early stage of infection in only some patients, which limited objective comparison between the two groups. A study of risk factors for hydrocephalus in childhood bacterial meningitis found that environmental factors, including rural residence, were potential risk factors [28]. In the present study, among infants with invasive GBS infection, the non-meningitis group had a higher proportion of inborn infants, whereas the meningitis group had a higher proportion of outborn infants. We hypothesize that outborn infants may have less immediate access to tertiary hospital care than inborn infants, which could delay the diagnosis and treatment of specific infections and potentially contribute to complications; further study is needed to evaluate this possibility [28].
In infants, GBS meningitis primarily develops through bacteremia, with the pathogen crossing the blood-brain barrier and inducing an inflammatory response in CSF that results in brain injury. Thus, meningitis may represent a more advanced stage of invasive infection than sepsis [7]. Prolonged fever (>3 days) has been reported as a risk factor for late-onset sepsis complicated by neonatal purulent meningitis [29]. In this study, the duration of fever before hospitalization was longer in the meningitis group than in the non-meningitis group, although it was less than 3 days in both groups. Infants with bacterial meningitis or severe systemic infection are known to have higher serum CRP levels than those with uncomplicated sepsis [30]. Several studies have reported that CRP levels above 7–9 mg/dL may serve as predictive markers of severe infection or meningitis in infants with invasive infection [30,31]. In this study, the mean CRP level on the first day of hospitalization was higher in the meningitis group than in the non-meningitis group (7.5 mg/dL vs. 1.2 mg/dL). Furthermore, multiple logistic regression identified an elevated CRP level in the early stage of infection as a factor associated with meningitis. Lin et al. [26] reported higher mortality in infants with complicated GBS infection, including meningitis, severe infection, and septic shock, than in those with uncomplicated infection. In contrast, van Kassel et al. [32], in a 30-year nationwide surveillance study in the Netherlands, found no difference in mortality between infants with GBS meningitis and those with GBS sepsis. Similarly, in the present study, mortality did not differ between the meningitis and non-meningitis groups. Infection severity and the prevalence of comorbidities were also similar between the two groups.
Ampicillin is the treatment of choice for invasive GBS infection [7]. In meningitis, a higher dose is required because of considerations related to blood-brain barrier permeability, CSF concentration, and ampicillin pharmacokinetics. Combination therapy with ampicillin and gentamicin is widely used in the early stage of infection, followed by conversion to ampicillin monotherapy after clinical improvement and CSF sterilization are achieved [7]. In this study, all GBS isolates were susceptible to ampicillin, whereas alternative agents, including clindamycin and erythromycin, showed partial resistance, consistent with a previous report [33]. However, cases of ampicillin-resistant GBS meningitis have been reported in Korea and may be associated with unfavorable outcomes [34]. Therefore, antibiotic selection should be guided by susceptibility findings and clinical status.
Newborns and young infants with GBS meningitis are at high risk for neurological complications. Brain MRI findings, including encephalomalacia, cerebral infarction, and multifocal or extensive lesions, are predictors of poor neurodevelopmental outcomes in infancy [35,36]. In this study, several patients who underwent brain MRI after antibiotic treatment showed various brain lesions, indicating the need for careful follow-up for potential neurological sequelae, such as motor disability and developmental delay [8].
This study has several limitations. First, it was a retrospective study based on medical records and radiological findings. Second, it was conducted at a single center and included a relatively small sample, which limited objective evaluation of the clinical characteristics of GBS meningitis. Nevertheless, these findings may contribute to a better understanding of the clinical features and factors associated with GBS meningitis in neonates and young infants. In conclusion, GBS meningitis was more common in outborn infants and frequently presented as LOD. In addition, an elevated CRP level in the early stage of infection was associated with meningitis. Multicenter studies with larger sample sizes and broader geographic representation are needed to support more definitive conclusions.
Supplementary material
Supplementary materials related to this article can be found online at https://doi.org/10.26815/acn.2025.01263
Antimicrobial susceptibility of group B streptococci isolated from patients with invasive diseases
Brain magnetic resonance imaging findings after antibiotic treatment in patients with group B streptococcal meningitis
Notes
Conflicts of interest
No potential conflict of interest relevant to this article was reported.
Author contribution
Conceptualization: SS and CSK. Data curation: SS, SYS, JHP, RL, GEK, and CSK. Formal analysis: SS, NHS, and CSK. Methodology: SS and CSK. Project administration: SS and CSK. Visualization: SS and CSK. Writing-original draft: SS and CSK. Writing-review and editing: SS, SYS, JHP, HJC, JCB, and CSK.