Elizabethkingia Meningosepticum

• 1959 - American bacteriologist Elizabeth O. King was studying unclassified bacteria associated with pediatric meningitis at the Centers for Disease Control and Prevention (CDC). She isolated an organism (CDC group IIa) that she named Flavobacterium meningosepticum (Flavobacterium means "the yellow bacillus" in Latin; meningosepticum means "associated with meningitis and sepsis").
• 1994 - Flavobacterium meningosepticum was reclassified in the genus Chryseobacterium and renamed Chryseobacterium meningosepticum (chryseos means "golden" in Greek; Chryseobacterium means a golden/yellow rod similar to Flavobacterium).
• 2005 - A 16S rRNA phylogenetic tree of Chryseobacteria showed that C. meningosepticum along with C. miricola (which was reported to have been isolated from the Russian space station Mir in 2001 and placed in the genus Chryseobacterium in 2003) were similar to each other but outside the tree of the rest of the Chryseobacteria. They were then placed in a new genus Elizabethkingia named after the original discoverer of F. meningosepticum.
• 2008 - Green et al described the first case of Elizabethkingia miricola sepsis in an immunocompromised host.
• 2011 - Elizabethkingia anophelis is a bacterium isolated from the midgut of Anopheles gambiae mosquitoes originating from MacCarthy Island, The Gambia.
• 2013 - Elizabethkingia anophelis was reported as a human pathogen in the Central African Republic where a clinical case of neonatal meningitis was reported. A nosocomial outbreak was also reported in an intensive care unit (ICU) in Singapore. In both clinical cases, multidrug resistance was reported and the isolates were resistant to a wide array of antibiotics.
• 2014 – A study showed that some Elizabethkingia infections that had been attributed to Elizabethkingia meningosepticum were instead caused by Elizabethkingia anophelis.
• 2015 - An outbreak of Elizabethkingia anophelis centered in Wisconsin began in early November 2015, with 48 individuals confirmed infected in 12 counties with at least 18 deaths by March 9, 2016. Four new cases were documented just in the week of March 2 – 9, 2016. 

Elizabethkingia are bacteria belonging to the family of Flavobacteriaceae. The genus Elizabethkingia currently includes four species (Table 1):

• E. meningosepticum (or E. meningoseptica)
  • Binomial Name: Elizabethkingia meningseptica
  • Synonyms: Chryseobacterium meningosepticum, Flavobacterium meningosepticum, Elizabethkingia meningosepticum and CDC Group IIa
  • E. meningosepticum causes neonatal sepsis and infections in immunocompromised individuals.
• E. miricola
  • Binomial Name: Elizabethkingia miricola
• E. endophytica
  • Binomial Name: Elizabethkingia endophytica
• E. anophelis
  • Binomial Name: Elizabethkingia anophelis

E. meningosepticum is the only encapsulated species and the most relevant in clinical practice. The ability to produce proteolytic enzymes may be associated with its virulence and the ability to cause severe human infections.

Serologic procedures were conducted for typing strains isolated in epidemiological studies. Six serotypes have been described, A to F, of which serotype C is most commonly responsible for cases of neonatal infections.

DNA hybridization studies on a limited number of strains of E. meningosepticum have identified two main genomic groups:

• Genomic Group I
• Genomic Group II which includes four subgroups, II: 1 to 4.
  • Eighteen of 20 strains isolated from cerebral spinal fluid (CSF) belong to subgroup II: 1, leading to the hypothesis that genetic differences may play a role in pathogenicity of the different strains of E. meningosepticum.

E. meningosepticum is a non-fermenting, nonmotile, oxidase-positive, gram-negative, rod-shaped, aerobic bacillus.

E. meningosepticum typically grows within 24 hours of incubation on blood and chocolate agar. E. meningosepticum strains either are not pigmented or produce a very pale yellow, nondiffusible pigment which may not be easily evident at 24 hours. An additional 24 to 48 hours of incubation at room temperature may increase pigmentation, allowing faster identification and aiding in the appropriate empiric therapy until final results become available.  Strains growing better at 40o C are mostly associated with invasive meningitis.  

Colonies of E. meningosepticum grow poorly or not at all on MacConkey agar and are considered glucose oxidizers. They do not grow on colistin nalidixic agar (CNA) because, although they are resistant to colistin, they are susceptible to quinolones such as nalidixic acid. E. meningosepticum may show colistin resistant and vancomycin sensitive/intermediate growth which is paradoxic for a gram negative bacterium.

Little information exists about the pathogenesis of Elizabethkingia infections. Early experimental animal studies supported the low virulence of these microorganisms but these studies lacked accurate information. Adult rabbits and newborn hamsters survived without evidence of infection after intravenous and subcutaneous inoculation, respectively. Intracerebral inoculation led to a mortality rate of 27% but microorganisms were not recovered at autopsy.

Colonization of the respiratory tract usually precedes invasive infection particularly in high risk populations such as neonates and adults in ICUs. Early and late onset neonatal meningitis is hypothesized to have been preceded by colonization of the infant at the time of delivery since this has been described for other common neonatal pathogens. However, this hypothesis has not been studied.

E. meningosepticum are widely distributed in nature e.g., freshwater, saltwater and soil. They may normally be present in fish and frogs. They are not normally present in human microflora.

Strains of E. meningosepticum, sharing the same phenotypic, chemotaxonomic and genomic characteristics of human pathogens, have been isolated from infected birds. Although community acquired infections have been described after exposure to contaminated water, in most cases the source remains unknown.

Since first recognition of E. meningosepticum as a cause of neonatal meningitis in 1958, outbreaks have occasionally been described since 1961. Nosocomial outbreaks have been an important source of morbidity and mortality in neonates and immunocompromised individuals since 1956.

Hospital infections include:

• 1958 - Brody et al. described two outbreaks of meningitis caused by E. meningosepticum affecting primarily premature infants in a hospital nursery. This outbreak was described as hospital acquired but attempts to isolate the organism from human contacts were unsuccessful.
• 1961 - Cabrera and Galen reported an outbreak of 14 cases of neonatal meningitis.
• 1961 - George et al. reported 14 cases of meningitis and 30 asymptomatic nasopharyngeal carriers during an outbreak in a newborn ward.
• 1966 - Plotkin and McKitrick described two cases of E. meningosepticum meningitis which were traced to a saline solution.
• Between March and July 1975 - Two separate E. meningosepticum type E outbreaks were reported by Hazuka et al. in the Neonatal Intensive Care Unit (NICU) at Children's Hospital of Michigan. Of the 10 infants affected, only 3 in the first outbreak exhibited disease directly related to E. meningosepticum. The remaining two infants in March - April and all five infants in July were colonized but not infected. The three ill infants had positive blood cultures. Two developed meningitis. One died within 6 days and the other survived but developed hydrocephalus requiring a ventriculoperitoneal shunt. The one infant who did not develop meningitis had only a transient bacteremia, was treated with appropriate antibiotics and eventually discharged home. The organism was resistant to most antimicrobial agents tested and developed resistance to others during treatment.
• Between 1972 and 1977 - Thong et al. reported seven infants with E. meningosepticum infection. E. meningosepticum was isolated from the CSF in all seven infants, from blood in three infants and from peritoneal fluid in one infant. Two infants who died did not receive intraventricular chemotherapy. Five infants survived with three normal neurologically. One of the survivors had hydrocephalus with severe brain damage. The infection in this infant began with umbilical sepsis and peritonitis. The infecting organism was isolated from the peritoneal fluid two days before the CSF yielded the same organism. The other case also had hydrocephalus without showing any other evidence of neurological handicap.
• 1980 - Dooley et al. summarized 63 previously reported cases of E. meningosepticum meningitis that occurred from 1944 to 1976. Out of 52 infants for whom the outcome of infection was known, 34 (65%) died. Eleven (61%) of the 18 neonates who survived developed hydrocephalus. Overall, four (50%) of the eight neonates whose length of gestation was known were full-term infants. According to the available data, the mean duration that CSF cultures remained positive was 16 days (range 8 - 39 days). Ten infants had associated bacteremia. One death was related to infection due to E. meningosepticum resulting in a case fatality rate of 8.3%. Hydrocephalus developed in 70% of the surviving infants. Five infants received intrathecal antibiotics.
• Between 1982 and 1996 - Di Pentima et al. isolated four strains of E. meningosepticum from the CSF and blood of neonates with clinical infection in Texas Children's Hospital. All of the neonates tolerated treatment with vancomycin and rifampin and had no adverse effects. All three neonates with meningitis due to E. meningosepticum survived without developing hydrocephalus or neurological deficits as determined on physical examination at the time of discharge. Two infants developed mild sensorineural hearing loss. E. meningosepticum was cultured at birth in a tracheal aspirate from a premature infant. Bacteremia due to E. meningosepticum subsequently developed in this infant at 3 weeks of age resolving with a 10-day course of vancomycin and ceftazidime. The infant died of Staphylococcus aureus septicemia at 2 months of age.
• 2000 - Chiu et al. reported 17 culture-documented systemic infections due to novel, atypical strains of E. meningosepticum in two newborns and 15 immunocompromised individuals from 1996 to 1999 in a medical center in Taiwan. All clinical isolates were resistant to a number of antimicrobial agents. The isolates were characterized as atypical strains of E meningosepticum. That was the first report of a cluster of atypically variant strains of E. meningosepticum. Two individuals were newborns and one was 7-year-old child with IgA nephropathy. Three of the 15 non-neonatal individuals died of the infection. Two newborns survived with severe neurologic sequelae despite antibiotic treatment.
• 2001 - Hoque et al. reported a strain of multi-resistant E. meningosepticum isolated from eight neonates in an NICU from September 1994 to May 1996. Two neonates were infected (one had pneumonia, the other septicemia and meningitis). The remaining six neonates were colonized in the respiratory secretions. Two cases occurred that could not be explained by cross-infection during the outbreak. The two infected neonates survived. The neonate with meningitis and septicemia did not develop hydrocephalus. Both of these infants were very low birth weight neonates (less than 1500 g) which is consistent with studies that show at least 50% of neonates with this infection weighed less than 2500 g.
• 2003 - Güngör et al. reported an outbreak of E. meningosepticum in four neonates with sepsis in the NICU of a referral teaching hospital in September 2001. The organism was isolated from the blood cultures of all four neonates. The first isolate was identified 5 days after the death of the index case. The other three neonates survived with one having hydrocephalus.
• 2009 - Maraki et al. reported four neonates with E. meningosepticum colonization in the endotracheal tubes and respiratory secretions in an NICU of a referral teaching hospital in Greece between April and October 2002. None of the neonates progressed to clinical infection and none of them received specific treatment. All survived. This study suggested that E. meningosepticum colonization in neonates does not necessarily lead to infection and that such colonization outbreaks may be controlled with emphasis on standard precautions.
• 2011 – Ceyhan et al. reported three clusters of E. meningosepticum infections in a hospital in July 2006 and January 2007 involving 8 newborns and 5 older children. Seven of the newborns were premature. The index patient was from the NICU with the older patients being from other pediatric wards. Three of them had meningitis, two had primary bacteremia, five had sepsis, one had postoperative cellulitis and fasciitis and two had respiratory distress and pneumonia. The organism was isolated from the blood of all 13 and 4 from the CSF. Nine patients improved on antimicrobial treatment but 4 premature infants died after the infection. One of the neonates who died had meningitis, one had sepsis and the other two had respiratory distress and pneumonia.
• 2015 - Hsu et al. studied a total of 118 individuals with E. meningosepticum bacteremia at a medical center in Taiwan from 1999 to 2006. Among 99 preserved isolates, 84% presented with fever, 86% had nosocomial infections and 60% had acquired the infection in ICUs. 78% of the individuals had primary bacteremia followed by pneumonia (9%) and catheter-related bacteremia (6%). 45 individuals (38%) had polymicrobial bacteremia. Only 6 individuals were under the age of 18 and none were premature, a much lower percentage than previously reported. 

E. meningosepticum can be transmitted from individual to individual. This bacilla has always affected humans but it is an opportunistic pathogen. It does not usually cause illness in individuals who are otherwise healthy.

Environmental studies have revealed that E. meningosepticum can:

• Survive in:
  • Chlorine-treated municipal water supplies
  • Colonized sink basins, sink traps, sink drains, water taps and tap water
• Be cultured from:
  • Ambu bags
  • Computer keyboards
  • Doorknobs
  • Electrical buttons
  • Hands
  • Phones
  • Powdered infant formula

Colonization of patients via contaminated medical equipment/devices include:

• Antiseptic solutions such as aqueous chlorhexidine gluconate (Hibitane) solutions used for the storage of thermometers and for routine disinfection
• Central and peripheral intravascular catheterization lines including flush solutions for arterial catheters and pressure transducers
• Lipid stock bottles
• Ice chests, ice machines, syringes, sterile saline solution for flushing eyes, vials etc.
• Respirators/ventilators, ventilator tubing and aerosols, endotracheal intubation tubes and nasotracheal tubes, mist tents, humidifiers, nebulizers, incubators for newborns
• Surgically implanted devices and prosthetic valves such as tissue-allographs as described in two individuals who had received tendon and tendon-bone allographs
• Topical disinfectants
• Tube feedings
• Vials of intravenous and aerosolized drugs

E. meningosepticum has been isolated from the female genital tract. Whether this site is a source of neonatal colonization with subsequent development of early and late onset meningitis has not been studied.

In outbreaks, respiratory tract colonization occurs more often than infection.

Among all Elizabethkingia, E. meningosepticum is the most important human pathogen.

Risk factors which should raise suspicion for E. meningosepticum infections include:

• Age
  • Premature newborns and infants in NICUs especially in underdeveloped countries
• Patients with multiple chronic co-morbid conditions including:
  • Alcohol dependence
  • Alcoholic cirrhosis
  • Chronic renal disease or end-stage renal disease on dialysis
  • Diabetes mellitus
  • Immune-compromising conditions
  • Immunosuppressive treatment
  • Malignancy

General symptoms of E. meningosepticum are very non-specific and may mimic many other infections. Generalized symptoms may include:

• Fever/chills
• Shortness of breath 
• Cellulitis, a redness and swelling from a skin infection that may feel hot and tender to the touch (most commonly, on the lower extremities)

The most severe and common clinical presentation during the neonatal period is E. menigosepticum meningitis. Onset of infection may be insidious and may resemble the common signs and symptoms of early and late onset neonatal meningitis due to other pathogens common in this age group. Infections usually affect premature infants during the first few weeks of life and often occur as outbreaks. Among the various serotypes of E. meningosepticum (A to F), type C has been the cause of most reported outbreaks.

As the causative agent of neonatal meningitis, E. menigosepticum demonstrates mortality rates up to 57% and produces severe post-infectious sequelae including brain abscesses, hydrocephalus, deafness and developmental delay. The case-fatality rate has been high in neonates and early and late complications are common among survivors.

Proposed risk factors in neonates associated with poor neurological outcomes, mainly hydrocephalus and deafness, include:

• Age of onset (less than 9 days of age)
• Length of time CSF cultures remain positive (more than 10 days)
• Prematurity - The primary risk factor for E. meningosepticum infection. Half of the infections have involved neonates weighing less than 2500 g.
• Use of intrathecal antibiotics

Early diagnosis and the institution of appropriate antibiotic therapy are associated with a better prognosis yet the overall mortality of neonates from early reports is above 60%. 

Other common manifestations of E. meningosepticum infection in the neonatal period are:

• Bacteremia
• Pneumonia

E. meningosepticum can affect immunocompromised individuals of all age groups whether hospitalized or not. Among hospitalized immunocompromised individuals, clinical presentation varies according to the source of exposure (indwelling devices, infusion of contaminated solutions, head and neck surgery, peritoneal dialysis, polypectomy etc.) and underlying risk factors. Underlying immunodeficiencies and invasive surgical procedures determine the severity and the site of infection. Signs and symptoms of E. meningosepticum in all age groups may include:

• Bacteremia
• Bronchitis
• Cellulitis
• Dialysis-associated peritonitis
• Endocarditis (post-dialysis and intravenous drug abuse) including prosthetic valves
• Endophthalmitis
• Epididymitis
• Intraabdominal abscess
• Keratitis
• Meningitis
• Peritonitis
• Pneumonia
• Postoperative bacteremia
• Prosthesis-associated septic arthritis
• Sepsis following extensive burns
• Sinusitis

Nosocomial pneumonia, mainly among intubated, immunocompromised adults, has a mortality rate of 60%. E. meningosepticum has been isolated from endotracheal and nasal secretions and is not considered the true pathogen. Bacterial load, immune status and selective antibiotic pressure may prompt invasive disease. Reported predisposing immunodeficiencies have ranged from carcinoma, tuberculosis, leukemia, aplastic anemia, asplenia, congenital immunodeficiency and corticosteroid therapy.

In older immunocompetent pediatric patients and adults, community acquired Elizabethkingia infections occur infrequently. Such infections are seen mainly in geriatric patients and the most common clinical presentation is pneumonia.

Only recently has Elizabethkingia infections been found in the immunocompetent and diabetic adult patients. Signs and symptoms may include:

• Bacteremia
• Cellulitis
• Community-acquired respiratory tract infections
• Keratitis
• Necrotizing fasciitis (in a diabetic patient)
• Ulcerative granulomatous lesions
• Septic arthritis
• Sepsis

E. meningosepticum has unusual resistance patterns and mechanisms and are resistant to most β-lactam drugs, including the carbapenems and aztreonam. Many possess two different types of β-lactamases, namely:

Class A extended-spectrum β-lactamases (ESBLs)
Class B metallo-β-lactamases (MBLs)

Class B MBLs confer resistance to carbapenems, which are widely used to treat infections caused by multidrug-resistant gram-negative bacteria. Two types of MBLs, BlaB and GOB, have been identified in isolates of E. meningosepticum. Although they have similar molecular weights, these two enzyme types show only very low molecular similarity. Sequencing and analysis of genes encoding BlaB and GOB has revealed heterogeneity, with up to 12 BlaGOB and 14 BlaB alleles identified. E. meningosepticum is intrinsically resistant to most β-lactams, including carbapenems, due to production of chromosomal MBLs.

Bellais et al. reported that all E. meningosepticum isolates harbored two types of MBLs simultaneously: BlaB belonging to subclass B1 and GOB belonging to subclass B3. However, a recent survey in China reported that only 55 of 170 E. meningosepticum isolates harbored both types of MBLs and 38 isolates harbored only one type of MBL. Furthermore, no MBL genes were detected in the remaining 77 isolates, even though many of these isolates were resistant to imipenem. PCR experiments detected both genes encoding BlaB and GOB MBLs in all E. meningosepticum isolates. DNA sequence analysis revealed that E. meningosepticum isolates possessed 7 types of BlaB gene, including 5 novel variants (BlaB-9 to BlaB-13) and 11 types of BlaGOB gene, including 10 novel variants (BlaGOB-8 to BlaGOB-17). The most common combination of MBLs was BlaB-12 plus GOB-17.

These organisms are generally resistant to the conventional chemotherapeutic agents used in the treatment of neonatal meningitis, such as ampicillin, gentamicin, kanamycin, and chloramphenicol. Different resistance patterns have been reported in different outbreaks worldwide.

The appropriate choice of antimicrobial agents effective for the treatment of E. meningosepticum infections is quite difficult to make due to the following reasons:

• Elizabethkingia are resistant to most antibiotics and the use of inactive drugs as empirical therapy may contribute to the poor outcome in many infections.
• MIC breakpoints have not been established by the National Committee for Clinical Laboratory Standards (NCCLS) for Elizabethkingia. Antimicrobial susceptibility data on Elizabethkingia remain very limited since this pathogen has been rarely isolated from clinical specimens.
• Results of susceptibility testing vary when different methods are used. Results from disk diffusion methods may not be reliable, so broth reference quality microdilution tests should be performed when possible. The E-test has also been suggested as a possible alternative for testing certain antibiotics.

Hoque et al. reported a strain of multiresistant E. meningosepticum isolated from eight neonates in a NICU from September 1994 to May 1996. The strain was resistant to ampicillin, ceftazidime, imipenem, gentamicin, ciprofloxacin and trimethoprim/ sulfamethoxazole (TMP/SMX) but was susceptible to piperacillin and amikacin and had variable susceptibility to rifampicin and vancomycin.

Güngör et al. reported that although all of their isolates in an NICU were susceptible to ciprofloxacin in vitro, three patients did not respond to ciprofloxacin therapy given for 6 or 7 days. They switched the therapy to vancomycin and rifampin and all three patients survived with one having a complication (hydrocephalus).

According to the results of the SENTRY Antimicrobial Surveillance Program, Elizabethkingia are known to exhibit resistance to:

• Aminoglycoside antibiotics such as kanamycin, amikacin, gentamicin
• Tetracycline antibiotics such as tetracycline, doxycycline
• Chloramphenicol antibiotics such as chloramphenicol
• Macrolide antibiotics such as erythromycin
• Lincosamide antibiotics such as clindamycin
• Glycopeptide antibiotics such as teicoplanin

Elizabethkingia are known to be susceptible to:

• Fluoroquinolones such as ofloxacin, levofloxacin, sparfloxacin, clinafloxacin (all four agents are somewhat more active than ciprofloxacin)
• Tetracycline antibiotics such as minocycline has also shown good in vitro activity but susceptibility to doxycycline appears variable
• Trimethoprim and Sulfamethoxazole (TMP/SMX): susceptibility appears variable
• Rifampin antibiotics such as rifampin is usually active in vitro and has been used as part of a combination therapy to clear persistent infection
• Glycopeptide antibiotics such as vancomycin alone or in combination with other agents, including rifampin, have been successful in the treatment of meningitis in infants. However, the usefulness of vancomycin against Elizabethkingia infections has more recently been questioned. It was reported that vancomycin was inactive in vitro.

Ceyhan et al. reported that although all of their isolates remained susceptible to vancomycin, rifampicin and linezolid during an outbreak, resistance to imipenem and amikacin was increased in the second and third clusters. This may be due to an inducible resistance against these antibiotics. The possibility of this type of resistance should be considered when choosing an antibiotic regimen to treat E. meningosepticum infection. Four of 13 cases (1 in the second cluster and 3 in the third cluster) died despite antibacterial treatment which had appeared to be effective in sensitivity testing. This may be due to the neonates' prematurity and severe disease and also may represent a difference of in vitro activity.

In conclusion, there is no optimal regimen for the treatment of Elizabethkingia infections, and antimicrobial therapy should be based on MIC results from properly performed susceptibility tests.

E. meningosepticum is usually resistant to cephalosporins, aminoglycosides and carbapenems.

Among the extended spectrum ß-lactam penicillins, piperacillin is the most active agent. In one series of 52 isolates, all were susceptible. More recent studies have reported an increase in isolates resistant or of intermediate susceptibility. Bolash et al. reported that the susceptibilities of E. meningosepticum to piperacillin and ticarcillin were similar. Previous reports had found none of the isolates tested susceptible to ticarcillin.

The addition of a ß-lactamase inhibitor, either tazobactam or clavulanate to piperacillin, amoxicillin or ticarcillin does not confer any significant advantage. However, in a study of 7 distinct isolates of E. meningosepticum, addition of clavulanic acid had a significant impact lowering the MICs of cephalosporins, particularly ceftazidime. Clavulanic acid lowered ceftazidime, cefotaxime and cefoperazone MICs. None of the strains appeared to harbor plasmids, but a broad-substrate, constitutive, chromosomal ß-lactamase was detected.

In general, all cephalosporins have unpredictable and usually poor activity against these pathogens. Cefoperazone, cefotaxime and ceftriaxone showed similar activity with some isolates highly susceptible and many of intermediate susceptibility. However, the MIC50 for E. meningosepticum are above the NCCLS susceptibility breakpoint for almost all the strains tested. Ceftazidime, and the first and second-generation cephalosporins are generally less active. In one study, eight of 12 isolates were reported as being susceptible to cefoperazone-sulbactam. Fourth generation cephalosporins, cefepime and cefpirome have also poor activity against E. meningosepticum. 

Aztreonam has shown poor activity against the few isolates tested. Furthermore, as a result of the production of metallo-ß-lactamases, Elizabethkingia are highly resistant to both imipenem and meropenem.

Fluoroquinolones have shown good activity against E. meningosepticum, especially ofloxacin, levofloxacin and ciprofloxacin.

Interpretation of the in vitro performance of vancomycin against E. meningosepticum can be controversial since MIC breakpoints recommended by the NCCLS are for gram-positive microorganisms. 

Other non-ß -lactam antibiotics most consistently active against E. meningosepticum are:

• Lincosamide antibiotics such as clindamycin
• Rifamycin antibiotics such as rifampin
• Trimethoprim/Sulfamethoxazole (TMP/SMX)
• Macrolide antibiotics such as clarithromycin being more active than erythromycin

Among aminoglycoside antibiotics such as amikacin, therapeutic failure has occurred. 

The optimal antimicrobial therapy for infections due to Elizabethkingia, is difficult to determine. Empiric therapy is based on the unusual susceptibility pattern, the availability of effective antimicrobials, the age of the patient, clinical presentation and reported experience.

Definitive therapy should rest upon the susceptibility results of the isolated organism. Important considerations to remember when managing these infections include:

• Disk diffusion techniques are not reliable and treatment failures can occur if treatment is based on disk results.
• Antimicrobial susceptibilities should be confirmed using a quantitative method.
• Acquired antimicrobial resistance can develop during therapy so repeat cultures and susceptibility testing is recommended if the patient remains culture positive, does not improve or deteriorates.

Meningitis is not only the most common but also the most difficult E. meningosepticum infection to treat successfully. Delay in implementing appropriate therapy and the lack of effective antimicrobials commonly permits the persistence of positive CSF cultures for much longer periods than for other gram-negative pathogens causing neonatal meningitis.

Although vancomycin MICs are high and its indication is controversial, this drug has been the most effective agent reported in the literature for the management of meningitis, especially in combination with rifampin. Vancomycin resistance acquired during therapy has been described and this possibility should be considered in the event of treatment failures. A recent report advised against using vancomycin for E. meningosepticum meningitis based on the high MICs of this drug for this pathogen. Rifampin is very active against Elizabethkingia but when used as a single agent it may rapidly induce resistance.

Susceptibility of E. meningosepticum to TMP/SMX varies significantly in recent reports, ranging from 23 to 100% resistance. Linder et al. reported a successful outcome in eight of nine infants treated with TMP/SMX but none of these infants had a positive CSF culture or confirmed meningitis. TMP/SMX has excellent penetration into the CSF and anecdotal evidence for its effectiveness in E. meningosepticum meningitis is available. Special considerations, when using TMP/SMX during the neonatal period, is it association with hematological abnormalities and kernicterus.

Piperacillin should be use cautiously in CNS infections since the achievable concentration of this drug in the CSF could be sub-therapeutic. Furthermore, of the 8 reported cases who received piperacillin, five developed hydrocephalus and one died.

The use of intraventricular antibiotics has been associated with an increased risk of hydrocephalus and should not be considered unless intravenous therapy has failed to eradicate infection. In older patients, empiric treatment with minocycline and quinolones can also be considered. Minocycline has good in vitro activity against the great majority of isolates, good penetration into the CSF and is available as a parenteral formulation. Given that tetracyclines are primarily bacteriostatic, combination with another active agent such as rifampin is advisable.

The pharmacodynamics of quinolones in bacterial meningitis have shown that in order to achieve bactericidal effectiveness, the concentration of antibiotics in the CSF needs to exceed the MBC for the entire dosing interval. Quinolones are lipophilic agents and peak CSF concentrations, up to 26% of the serum concentration are reached rapidly. Although there are discrepancies in the literature, in general the MIC50 of ciprofloxacin is below the NCCLS susceptibility breakpoint for the majority of the isolates tested. Ciprofloxacin is currently FDA approved for neonatal treatment of multidrug-resistant, gram-negative infections. Nevertheless, clinical experience with quinolones in E. meningosepticum neonatal meningitis is not widely available in the English literature. Sakuma et al, recently reported the successful treatment of a 5-day-old infant with E. meningosepticum meningitis treated with intravenous ciprofloxacin and TMP/SMX. The use of TMP-SMX should be used cautiously in this age group.

Most strains of E. meningosepticum are highly resistant to all cephalosporins. Based on in vitro susceptibilities, it has been proposed that ceftizoxime might be effective treatment, however, there is not enough clinical experience reported in the English literature to support the use of this agent in neonatal meningitis. A single case of E. meningosepticum meningitis treated with cefepime was reported in the Literature. Lu et al. described a 21 year-old women with insulin-dependent diabetes and E. meningosepticum meningitis, successfully treated with a 21-day-course of cefepime. The recovered strain was sensitive to cefepime, meropenem, imipenem/cilastatin, and moxalactam.

As with other cases of gram-negative meningitis it is recommended that therapy continue for two to three weeks after sterilization of the CSF is accomplished. Recommendations for treatment of other E. meningosepticum infections follow those for meningitis and are modified by the age of the patient. Minocycline or a quinolone, in combination with rifampin, seems appropriate empiric therapy in patients older than 8 years. Although minocycline has good activity, in severe infections piperacillin is recommended since tetracyclines are bacteriostatic agents. Management of catheter related infections should follow the general considerations indicated for other gram-negative pathogens.

In gram-negative neonatal meningitis:

• Systemic antibiotics are given for approximately 2 weeks after bacteriologic cure.
• Repeat CSF examinations and cultures are usually required 48 to 72 hours after initiation of therapy until negative cultures are achieved and at the end of therapy.
• Because sterilization of CSF is commonly delayed in E. meningosepticum meningitis, antibiotic therapy is generally prolonged for three or more weeks.
• Endpoint criteria to make this final decision are based on the clinical course and CSF findings before stopping antimicrobial therapy.

Management of catheter related infections should follow the general recommendations for other gram-negative pathogens:

• In general, if the patient remains clinically stable and bacteremia resolves within 48 to 72 hours after institution of therapy, removal of the catheter may not be mandatory.
• Blood cultures should be obtained daily from each catheter port and peripherally until all cultures remain negative for 48 to 72 hours. Length of therapy is based on the duration of bacteremia and the immunologic status of the patient.
• Intravenous therapy should be continued for 7 to 10 sterile days in immunocompetent patients and for 10 to 14 sterile days in immunocompromised patients. 

No vaccines are currently available for any Elizabethkingia species.

Clinical studies demonstrate the importance of rapid instigation of epidemiological investigation, while also ensuring that fundamental infection control procedures are in place during an outbreak and to prevent an outbreak. It is worthwhile for the infection control team to ensure that the hospital water tanks are inspected and chlorinated yearly and any necessary repair work carried out.

Patients known to be infected or colonized with E. meningosepticum should be placed in isolation following the institutional guidelines for multiple antibiotic-resistant organisms. Particular attention should be placed when handling respiratory equipment, especially nebulizers with a reservoir. Elizabethkingia can multiply to substantial concentrations in nebulizer fluid and increase the risk of spreading nosocomial pneumonia.

Outbreaks of gram-negative bacterial infections are usually due to transient carriage of the organisms on the hands of healthcare workers. Susceptible patients may become colonized after acquiring the organism from a healthcare worker but infection may or may not develop.

Gram-negative bacteria can have an inanimate reservoir such as hospital sinks. It is generally considered that this is not an important factor in endemic hospital-acquired infections. In those susceptible to infection, however, small numbers of potentially pathogenic organisms present on healthcare workers' hands after washing them in contaminated tap water may cause infection.

To detect the source of an outbreak of E. meningosepticum, it is important to obtain cultures from food and infant formulas, wet areas, dry surfaces, equipment and the hands of healthcare workers who have contact with patients. Obtaining cultures on a periodic basis is necessary.

Measures that have been used to eradicate E. meningosepticum outbreaks in pediatric wards include:

• Changing the prescribing policy for empiric antibiotics
• Restriction of further admissions
• Strict supervision of the preparation process of intravenous lipid solutions, as well as, other intravenous solutions and infant formulas
• Thorough disinfection of the ward(s)
• Toileting of babies with sterile instead of tap water

Other studies have shown successful control of an outbreak with milder measures including:

• Discarding opened creams, ointments, sterile water and hand-washing solutions
• In-service training should reemphasize handwashing, contact precautions and standard precautions to all staff
• Use of alcohol based hand rub after washing of hands
• Isolation of the patients with E. meningosepticum positive cultures
• Particularly in ICUs, frequent disinfection of electrical buttons, computer keyboards, phones, doorknobs and ambu bags
• Repairing, cleaning, super chlorination, isolation of the water tanks from all the hospital feeder tanks and changing the sink taps
• Restriction of staff exchange between wards
• Scrubbing units and wards thoroughly using two disinfectants (hypochlorite solution and isopropanol spray) with special emphasis on objects containing or in contact with water

All healthcare facilities and staff should follow standard precautions when caring for patients with E. meningosepticum infections, which include:

• CDC hand hygiene guidelines in health-care settings
• CDC environmental infection control guidelines in health-care facilities

In conclusion, surveillance for the reservoir and maintenance of rigorous infection control measures are essential to control E. meningosepticum outbreaks in the hospital setting.

Patients with any Elizabethkingia infection who have high risk factors have an increased mortality.

Several predictors of poor outcome include:

• Central venous line infection
• Hypoalbuminemia
• Increased pulse rate at the onset of infection

Healthcare facility response measures should include:

• Immediately reporting the identification of any isolate of Elizabethkingia from any sterile site specimen (blood, cerebral spinal fluid, synovial fluid, pleural fluid or other sterile site) to the respective State Health Department.
• Faxing requested medical records (including face sheet) to the respective State Health Department.
• Some clinical laboratories use bacterial detection systems with software that is not updated to report Flavobacterium meningosepticum or Chryseobacerium meningosepticum as Elizabethkingia meningosepticum. Therefore, it is advised to report the detection of any isolate from a sterile site specimen that is identified as F. meningosepticum, C. meningosepticum or E. meningosepticum.
• Submit all Elizabethkingia isolates (or F. meningosepticum or C. meningosepticum) expeditiously to the respective State laboratory for confirmatory testing via the facility clinical microbiology laboratory.
• Cases should be reported to Public Health officials – as should cases caused by similar bacteria. In particular, other Elizabethkingia species or related bacteria should be reported as they may have been misidentified.

Elizabethkingia meningosepticum has been deemed a potentially important threat to patients in critical care areas because of its multidrug-resistant phenotype and its ability to adapt to various environments. An increased incidence of E. meningosepticum bacteremia has increased over the last decade. Patients at high risk of E. meningosepticum infection include preterm infants, the immunocompromised and those exposed to antibiotics in critical care units. Vancomycin, rifampicin, newer fluoroquinolones, piperacillin–tazobactam, minocycline and possibly tigecycline are preferred empirical choices for E. meningosepticum infection according to in-vitro susceptibility data. Combination therapy has been used for infections not responding to single agents. Saline, lipid and chlorhexidine gluconate solutions, as well as, contaminated sinks are only a few of the implicated sources of infection following outbreak investigations. In addition to reinforcement of standard infection control measures, actions that have successfully terminated E. meningosepticum outbreaks include pre-emptive contact isolation, systematic investigations to identify the source of the bacterium and thorough cleaning of equipment and environmental surfaces. As the clinical complexity and incidence of E. meningosepticum infections increase, there is a need for heightened awareness of the potential for this bacterium to cause outbreaks. Thus, timely initiation of active surveillance for infected/colonized patients should occur, as well as, investigations to identify the likely source of the bacterium, which will, in turn, allow implementation of appropriate infection control measures.