Neonatal Meningitis Secondary to Elizabethkingia meningoseptica Infection

Abstract

Introduction: 

Elizabethkingia meningoseptica, a rare cause of sepsis and meningitis in neonates, often associated with a wide spectrum of clinical presentation. The objective of the study was to describe the clinical characteristics and outcome of neonates, who developed meningitis secondary to Elizabethkingia meningoseptica infection at a tertiary care Neonatal unit in India.

Methods: 

This retrospective study was conducted in the neonatal unit of a tertiary care hospital in New Delhi. The clinical data including demographic data, clinical presentation, management, and outcome data were collected and analyzed.

Results: 

During the study, 7 neonates with meningitis secondary to Elizabethkingia infection were identified. Majority of the neonates were preterm with a median gestational age of 31 (interquartile range: 29–33.5) weeks and a median birth weight of 1250 g (interquartile range: 1024–2065). The median age of onset of symptoms was 7 days. Lethargy (100%), apnea (85%), seizure (71%), and feeding difficulties (42%) were the common clinical presentations. Overall mortality during the period was 28.5%, and 60% of the survivor developed hydrocephalus. Isolated strains were resistant to the commonly used antibiotics (piperacillin-tazobactam, aminoglycosides, meropenem, and colistin) effective against Gram-negative organisms. The environmental screening was done but the potential source of infection could not be identified conclusively.

Conclusion: 

Meningitis in neonates caused by Elizabethkingia represents a potentially life-threatening infection and is often associated with significant neurological impairment, especially in premature neonates. A prolonged duration of antibiotic therapy, longer hospital stay, and likelihood of adverse neurologic sequelae during the hospital stay and follow-up should be anticipated in such cases of meningitis.

INTRODUCTION

Elizabethkingia meningoseptica is a Gram-negative bacillus widely distributed in soil and water. First recognized in 1959, it is responsible for causing outbreaks of neonatal meningitis and septicemia in immunocompromised children and adults.^[1] Although rare, Elizabethkingia is known to cause infections in neonates in neonatal intensive care units (NICUs), particularly septicemia and meningitis.^[2] Premature neonates are highly vulnerable to infections.^[1,2] Such infections are often associated with a high mortality or serious long-term morbidities. Various sequelae following infection such as brain abscess, developmental delay, and hydrocephalus have been reported in literature.^[2,3] Even though it is a rare cause of neonatal meningitis, accurate diagnosis is essential because the organism is resistant to the antibiotics commonly used for Gram-negative organisms.^[1,4] The organism is ubiquitous in nature and outbreaks usually occur as a result of contamination of respiratory support equipment and ventilation circuits, hospital tap water, sink drains, disinfectants, and other liquid solutions.^[1,4]

We describe an outbreak of Elizabethkingia meningoseptica infection in NICU of a tertiary care hospital in Delhi, which affected 7 neonates over a period of 3 months.

METHODS

The study was conducted at tertiary level neonatal care unit in New Delhi with around 13000–15000 births/year.

Between August and October 2017, 4132 babies were delivered and 730 babies required admission to the NICU. Of these, 7 neonates were diagnosed with Elizabethkingia meningoseptica meningitis. In all the cases with suspected sepsis, 2 sepsis screens were taken as per hospital protocol, blood culture and lumbar puncture were performed before starting antibiotics. BacT/ALERT (bioMérieux, USA) automated system was used for microbial detection. Species identification and antimicrobial susceptibility testing were performed using VITEK 2 (bioMérieux, USA) automated system. Additional susceptibility test was performed using Kirby–Bauer disc diffusion method and was reported as susceptible, intermediate, and resistant. After confirmation of cases, environmental screening was conducted and samples were collected from all potential sources which included respiratory equipment (ventilator and continuous positive airway pressure (CPAP) tubing, oxygen humidifier chambers), saline bottles, antibiotic and lipid solutions, baby bassinets and radiant warmers, tap water, suction apparatus, and sink drains. Several meetings were organized with the departments of microbiology and infection control and thereafter stringent control measures were undertaken to curtail and limit the spread of the outbreak.

Demographic data including age, sex, clinical data, paraclinical data including investigations, imaging and therapeutic data were collected retrospectively from review of medical records. Clinical data included the time of onset of symptoms, clinical presentation, and outcome. All patients were managed as per standard guidelines. Evaluations included routine blood counts, blood culture, cerebrospinal fluid (CSF) examination (cell counts, sugar, protein, culture), and cranial ultrasound. In the survivors, magnetic resonance imaging of brain and developmental evaluation using the Denver Developmental Screening test were carried out during follow-up.

Statistical analysis was done using IBM SPSS version 21.0 (IBM Corp., Armonk, N.Y., USA) software . Baseline data were expressed using descriptive statistics. Mean and standard deviations were used to express continuous variables. Data related to clinical presentations and outcomes were expressed as percentage of total cases.

RESULTS

Over the period of 3 months, 7 neonates developed meningitis due to Elizabethkingia meningoseptica. The baseline demographic and clinical details of these neonates are depicted in Table 1. Around 85% (6 of 7 babies) were born preterm. Median gestational age and birth weight were 31 (interquartile range: 29–33.5) weeks and 1250 g (interquartile range: 1024–2065) respectively with sex ratio of 1.3:1 (male: female). These neonates were admitted to the NICU either for prematurity and low birth weight or due to the presence of respiratory distress at birth (present in 6 cases). The neonates with respiratory distress were managed with respiratory support in the form of nasal CPAP. All neonates received full enteral feeds at birth and antibiotics were started in 2 neonates as per hospital policy.

Table 1: Baseline characteristics of neonates with Elizabethkingia meningoseptica infection

The clinical course was uneventful during the first 72 h. Initial respiratory support could be weaned within 48 h in 5 (83.3%) neonates. The median age of presentation of clinical symptoms was 7 days (4–22 days) suggesting nosocomial infection from NICU environment. The most common clinical symptoms were lethargy (100%) followed by apnea (85%), seizures (71%), and feed intolerance (42%).

Lumbar puncture was performed before start of antibiotics. CSF cytology invariably showed field full of pus cells and mean protein and sugar of 544.4 mg/dl (standard deviation [SD] 259.4 mg/dl) and 34.85 mg/dl (SD 33.71 mg/dl), respectively. CSF protein values were strikingly high and continued to remain high in repeat lumber puncture performed after 48–72 h. Elizabethkingia meningoseptica was isolated from CSF of all the 7 neonates. Blood culture also showed growth of the same organism in all cases. During this period, the organism was not isolated from any other unit of the hospital. Repeat CSF cultures done after 48–72 h of starting sensitive antibiotic was sterile in 4 cases, but in the remaining 3 cases, the organism continued to be isolated from CSF suggesting a delayed clearance or lack of adequate response to antibiotics. Sensitive antibiotics were continued for at least 2 weeks after the culture was reported sterile.

Isolated strains of Elizabethkingia meningoseptica were resistant to the commonly used antibiotics such as piperacillin-tazobactam, aminoglycosides, meropenem, and colistin. Overall susceptibility to piperacillin-tazobactam, amikacin, and meropenem were 28.3%, 28.3%, and 14.2%, respectively. Susceptibility to vancomycin was nearly 85% and more than 70% of isolates were susceptible to levofloxacin and cefoperazone-sulbactam. Ciprofloxacin and tigecycline had an intermediate sensitivity against the pathogen. All the neonates were initially started empirically with a combination of antibiotics (meropenem, amikacin) as per the NICU policy for meningitis. The antibiotics were modified subsequently based on the susceptibility reports and clinical response. Six out of 7 neonates received vancomycin, 5 received cefoperazone-sulbactam and 4 required oral rifampicin.

The environmental screening was done to identify the potential source infection; however, the source could not be identified. Even though the organism could not be isolated from environmental screening, factors and processes common to all 7 babies were studied and it was found that all except one baby received respiratory support before onset of symptoms. After various meetings with the infection control and microbiology department, strict infection control measures were undertaken to curtail the spread of the outbreak. After October 2017, no new case has been documented so far.

The clinical and neurological outcomes are depicted in Table 2. The overall mortality rate was 28.5%. Of the 85% (6 of 7) who survived till discharge from NICU, communicating hydrocephalus developed in 4 out of 6 (66.7%) neonates. Neonates with ventricular dilatation were serially followed with cranial ultrasound. The babies who had a progressive ventricular dilatation were started with oral acetazolamide after consultation with neurosurgeon. The neonates with increasing ventricular size despite oral decompressive therapy were considered for ventriculo-peritoneal (VP) shunts. In most of the neonates, the VP shunt was inserted after the completion of antibiotic therapy. VP shunt was placed in 3 out of 4 (75%) neonates with hydrocephalus. Ventriculosubgaleal shunt was used as a temporary measure in one neonate, which was removed later and VP shunt was inserted. None of the neonates were reported to have features of ventriculitis. The hearing screening was done in all patients before discharge. Screening for retinopathy of prematurity (ROP) was done in 4 neonates, one of them had Stage 3 ROP, requiring laser therapy. Two neonates had failed automated auditory brainstem response before discharge. One neonate had abnormal brainstem evoked response audiometry in follow-up. Postdischarge, the neonates were constantly under follow-up on an outpatient basis. All neonates were screened for developmental delay using the Denver Developmental Screening test, during follow-up visits. Two neonates (28.5%) who did not develop hydrocephalus are neurologically well and attaining milestones as per the corrected age. Among those with hydrocephalus, one neonate died at 3 months and other three (60%) had delayed milestones and were started with early intervention therapy.

Table 2: Clinical and neurologic outcome of cases with Elizabethkingia meningoseptica infection

DISCUSSION

Outbreaks of sepsis and meningitis in neonates due to Elizabethkingia meningoseptica has been reported previously.^[5–7] In the present study, majority of the neonates were preterm and very low birth weight. Mortality before discharge was 14.3%. Five neonates (71.4%) developed hydrocephalus as a complication. One neonate (14.2%) with hydrocephalus died during follow-up. Overall mortality was 28.4%. All neonates with hydrocephalus who survived had delayed developmental milestones.

Bacterial meningitis in a neonate can be ominous because of the associated high mortality or neurological sequelae. About 15%–25% of septicemic neonates may have meningitis, with nonspecific clinical features without any sign of meningeal irritation.^[8,9] The organism (Elizabethkingia meningoseptica) primarily causes nosocomial infections. It commonly colonises in the sink basins, taps, humidifiers and ventilator tubing. The organism can survive in chlorinated water. Prolonged use of respiratory support, humidifiers, and incubators is commonly encountered as the potential reservoir of infection. Indwelling central venous catheter and prolonged use of broad-spectrum antibiotics are also contributory factors.^[10–12]

Elizabethkingia meningoseptica primarily affects immunocompromised patients. Many reports in literature have described septicemia secondary to Elizabethkingia meningoseptica in adult patients with prolonged ICU stay, underling diabetes, and malignancy. Similarly, outbreaks of infection caused by this pathogen have been reported from hemodialysis units.^[13,14]Elizabethkingia meningoseptica is a rare cause of meningitis in neonates primarily affecting premature neonates. Other infections like septicemia and pneumonia have also been documented in preterm and term infants.^[15]

Thong et al. have reported a series of 7 neonates with meningitis secondary to Elizabethkingia meningoseptica with more than 50% of neonates being low birth weight (<2500 gram). All of the neonates developed symptoms in 1^st week of life, 2 babies died and 2 out of 5 survivors (40%) developed hydrocephalus.^[16] Hoque et al. have also reported a similar outbreak at a tertiary care NICU from California. Over a period of 2 years (1994–1996), Elizabethkingia meningoseptica was isolated from 8 neonates admitted to the NICU, out of which 2 neonates developed septicemia/meningitis. More than 80% of infected neonates were premature.^[11] In our series, 6 out of 7 infants (85.7%) were premature and 6 out of 7 infants (85.7%) required CPAP support before the onset of symptoms. The onset of symptom was within 1^st week of life in 4 out of 7 infants (57.1%) and rest developed symptoms beyond 2^nd week of life. One neonate died during the hospital stay. Five out of 7 neonates (71.4%) developed hydrocephalus. During follow-up at 6 months of age, 3 out of 5 neonates (60%) were found to have developmental delay as determined by the Denver developmental screening test. Detailed developmental assessment using definitive tests were not done due to a lack of follow-up.

In 1980, Dooley et al. have summarized outcomes of 63 previously reported cases of Elizabethkingia meningoseptica meningitis that occurred from 1944 to 1976. They have reported a mortality of 65% and 61% of survivors developed hydrocephalus.^[17] Di Pentima et al. reported four cases of Elizabethkingia meningoseptica infection in neonates between 1982 and 1996 in Texas Children’s Hospital. All these neonates were treated with Vancomycin and Rifampicin, with adequate response to treatment and no adverse effects were noted. Neonates with meningitis due to Elizabethkingia meningoseptica survived without developing hydrocephalus or neurological deficits.^[18]

Elizabethkingia meningoseptica colonization in the endotracheal tubes and respiratory secretions have been reported in four neonates in a NICU from Greece between April and October 2002.^[19] None of the neonates developed clinical infection, and none received specific treatment. This study emphasized the fact that colonization in neonates does not necessarily lead to infection and that with accentuation of the standard precaution measures, outbreaks may be prevented.

The pattern of antibiotic sensitivity/resistance is unique for Elizabethkingia meningoseptica. The organism is resistant to the conventional antibiotics used for Gram-negative infections such as aminoglycosides, b-lactam agents, and carbapenems. Different resistance has been reported in different outbreaks worldwide. The organism produces two b-lactamase: class A extended-spectrum b-lactamase agents and class B metallo-b-lactamase agents which confers resistance to b-lactams and carbapenems.^[9,15] Antibiotic susceptibility data are limited as the pathogen has rarely been isolated from pathological specimen. Available data suggest susceptibility to few antibiotics commonly used for the treatment of Gram-positive organisms such as vancomycin, quinolones, rifampin, and cotrimoxazole. Among the beta-lactams, piperacillin-tazobactam was found to be the most active agent overall.^[12] Majority of neonates in the present study responded to a combination of vancomycin with rifampicin/levofloxacin, but the clinical response might have been inadequate.

Data related to clinical and long-term outcomes are limited. Infection with Elizabethkingia meningoseptica carries a high risk of mortality with a reported mortality rate of almost 57%.^[15] At the same time, it carries the risk of severe postinfectious sequelae such as hydrocephalus, brain abscess, deafness, and developmental delay.^[20] Our study highlights the single-center experience of neonatal meningitis secondary to Elizabethkingia meningoseptica infection. The case-fatality rate has been high in neonates, and early and late complications are common among survivors. In the present study, case fatality rate was almost 28.5% and 60% of survivors developed hydrocephalus. Although mortality is lower, the incidence of neurological sequelae with hydrocephalus and developmental delay is comparable to that reported in literature_.^[16,17]

CONCLUSION

Elizabethkingia meningoseptica is an emerging pathogen for neonatal meningitis predominantly affecting premature neonates with a high case fatality and neurological sequelae among survivors. Early diagnosis, prompt institution of appropriate antibiotics, good supportive care with a careful scrutiny by the infection control team and NICU staff can be highly rewarding. Active infection control measures especially hospital water supply and care of respiratory equipment are necessary to prevent outbreaks.

Ethics approval

This research analyzed de-identified information from a retrospective chart review which is exempt from ethics committee of institute. The authors followed EQUATOR Network guidelines during the conduct of this research project. Informed consent was obtained from the parents of all subjects.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.