Keywords
Ureaplasma urealyticum
- neonate - ventriculitis - CNS infection - macrolides
Bacteria of the genus Ureaplasma represent frequent commensals in the lower genital tract of sexually active women.
In the context of perinatal medicine, ascending Ureaplasma infections are relevant triggers of chorioamnionitis, which itself is one of the
main risk factors for preterm birth.[1] Of the two known Ureaplasma species, Ureaplasma
parvum and Ureaplasma
urealyticum (Uu), U. parvum is more frequently detected in cases of chorioamnionitis, preterm birth, and neonatal
CNS infections.[2] Neonatal colonization results in increased morbidity, particularly due to the development
of congenital pneumonia, septicemia, and meningitis, as well as increased overall
mortality.[3] Respiratory colonization with Uu in extremely low birth weight infants (ELBWI) increases
the risk of developing bronchopulmonary dysplasia (BPD).
Only single-case reports of Ureaplasma meningitis or central nervous infections in preterm are available.[4]
[5]
[6]
[7] Animal models have demonstrated mechanisms which may be indicative of an impaired
blood-brain barrier after Ureaplasma exposure and may thus promote neuroinflammation.[8]
Case Presentation
We report on an ELBWI born by caesarean section at 245/7 weeks of gestation with a birth weight of 630 g. Postnatal care was initially uncomplicated
in the context of prematurity; surfactant (porcine) application was administered after
endotracheal intubation. Invasive ventilation via endotracheal tube was necessary
until the 26th day of life (DOL), followed by noninvasive ventilation up to 51st DOL.
In view of postnatal detection after routine testing of Uu by targeted polymerase
chain reaction (PCR) in tracheal secretion (TS), macrolide therapy (azithromycin 10 mg/kg,
p.o. over 5d) was initiated in the first week of life. Serum CRP, serum IL-6, and
white blood cell count (WBC) were unremarkable. There were neither typical clinical
symptoms of a Uu infection of the respiratory tract nor any signs of BPD in the chest
X-rays that were taken repetitively.
On the third DOL, bilateral IVH III° with parenchymal involvement was diagnosed on
routine cranial ultrasound (CUS). Progressive posthemorrhagic ventricular dilatation
(PHVD) was developed and managed with three consecutive lumbar punctures in the 3rd
week of life. Considering progressive PHVD and suspected blockage of the foramen of
Monro, bilateral insertion of ventriculostomy access devices' (VAD's) placed in both
lateral horns was performed on the 23rd DOL. Following daily CSF drainage via VAD's,
ventricular width was stable. The CSF volume taken was adjusted to the ventricular
width according to sonographic findings. The CSF was examined twice weekly by conventional
microbiology (not for Uu) and CSF biochemistry marker (protein, glucose, lactate,
and WBC).
On the 65th DOL, the CSF showed granulocytic pleocytosis (344 cells per µL, 85% granulocytes)
with persistent high protein levels, which further increased on the 72nd DOL (721
cells per µL, 85% granulocytes) ([Fig. 1]). CSF culture and multiplex PCR of the CSF (BioFire FilmArray ME Panel, used 2022)
failed to detect any microorganisms. On CUS, membranes in the lateral ventricles were
detected at this time, first on the right and later on the left side ([Fig. 2]).
Fig. 1 Changes in CSF WBC, protein, lactate, and glucose during the clinical course. The
change in CSF parameters from the first CSF tap (lumbar) to the daily CSF drainage
via the VAD's as well as the antimicrobial treatment cycles are shown. The WBC was
elevated on DOL 65 and increased again on DOL 72; lactate in the CSF was elevated
as well. In addition, protein was persistently elevated above the limit of quantification
(>6,000 mg/dL), and the glucose level was below the limit of detection (<0.11 mmol/L).
After the start of the first therapy cycle, the WBC and lactate levels decreased rapidly,
while with some delay protein levels decreased and glucose increased.
Fig. 2 CUS of the brain through the anterior fontanelle before (1a–c) and at the onset (2a-c)
of central Uu infection. Transfontanellar sonographic view of the brain before (1a–c)
and at the onset (2a–c) of infection. 1a/2a: coronal section plane: lateral ventricles
dilated, 2a: ventricular septation on the right side. 1b-c/2b-c: sagittal sectional
plane through the lateral ventricles, 2b–c: clear septation on both sides.
Serum CRP and IL-6 measurement did not reveal systemic inflammation and conventional
CSF microbiology was negative. In view of the clinical picture of ventriculitis, intrathecal
administration of vancomycin (5 mg each) was performed on the 65th DOL via the right
VAD and on the 72nd and 77th DOL via the left VAD.[9] CSF pleocytosis and protein elevation persisted during therapy, and microbial cultures
remained negative. The clinical condition was always stable.
Uu was detected by specific PCR in the CSF and in the oro-pharyngeal secretion (OPS).
Microbiological CSF cultural testing for Uu performed in a reference laboratory to
assess antimicrobial susceptibility was unsuccessful. For the treatment of suspected
Uu ventriculitis macrolide therapy with clarithromycin (10 mg/kg body weight, intravenously
for 10 days, then p.o. for 5 days) was administered starting on the 81st DOL, 64 days
after the completion of the first macrolide therapy. An attempt of therapeutic drug
monitoring (TDM) was made but yielded undetectable levels of clarithromycin in the
CSF.
Nonetheless, a decrease in WBC and protein levels in the CSF (from > 6,000 to 4,500 mg/L
on day 15 of therapy) was eventually observed under the targeted therapy. Molecular
biology continued to detect Ureaplasma DNA during therapy. Although DNA detection does not equal the presence of vital and
reproducing bacteria (cultures remained negative), we decided to start a second cycle
of oral therapy for 30 days, until negative Uu-PCR results were obtained from CSF,
to achieve safe eradication. The protein level in the CSF further decreased to 1,900 mg/L.
Clarithromycin treatment showed no adverse effects, even during prolonged therapy.
There were no QT prolongation, cardiac arrhythmias, or significant interactions with
other drugs due to CYP3A4 inhibition. Additionally, no hepatotoxic or nephrotoxic
effects were observed. In addition to Uu therapy, apart from the prophylactic administration
of an antibiotic for PPROM for 3 days within the first 14 days, empirical single administration
was necessary twice in the context of clinical deterioration until the infection was
ruled out. Magnetic resonance imagings of the brain performed two times (on 114th
and 195th dol) showed the findings already known from the sonography.
At the time of discharge, at the corrected age of 4 months, neurological examination
demonstrated unremarkable early childhood movement pattern, mild generalized muscular
hypotonia with trunk instability and only short phases of head control, and unremarkable
newborn reflexes; regardless of the necessity of a ventriculoperitonal shunt.
Discussion
Clinical symptoms, like a bulging fontanel as well as unexpected additional findings
on CUS, as demonstrated in this case, may be indicative of an infectious process,
notably of device-associated bacterial ventriculitis, and should prompt further investigation.
Gram-positive bacteria, such as staphylococci, are the most common cause of device-associated
ventriculitis, with an incidence of 6% reported in the literature.[10] No incidence can be given for Uu-ventriculitis due to the few cases described so
far.[4]
[5]
[6]
[7] Molecular or cultural detection of Uu in CSF in cases of suspected CNS infection
is not routinely performed. Nevertheless, in the setting of culture-negative ventriculitis
characterized by elevated WBC and persistently high CSF protein levels despite regular
CSF drainage via VAD, a targeted investigation for non-culturable pathogens should
be performed, especially in patients with a history of colonization with Uu. Even
though Uu has long been known as a pathogen associated with premature birth and CNS
infections,[11] a long time usually elapses before a diagnosis is established.[4]
[7] This is because these cell wall-less bacteria cannot be cultured with conventional
methods in routine laboratories, and appropriate detection methods (targeted PCR for
urogenital Ureaplasma and Mycoplasma spp.) are only requested based on clinical suspicion. In addition, the usual empirical
therapy regimens in suspected ventriculitis do not cover atypic bacteria like Ureaplasma. The possible therapeutic approaches for covering Ureaplasma are associated with relevant adverse drug reactions in infants; for example, doxycycline
leads to permanent tooth damage, and fluoroquinolones can lead to tendon inflammation
and rupture.[12] The treatment of CNS infections with macrolides is the subject of controversy. Erythromycin
shows poor CSF penetration and is known to be associated with hypertrophic pyloric
stenosis[13]; azithromycin, on the other hand, penetrates sufficiently but, like all macrolides,
achieves low CSF levels due to the rapid distribution and accumulation in brain tissue.[14] Evidence regarding antibiotic choice and duration of therapy is limited to case
reports or small case series.[6]
[7] Spontaneous remission has been reported even without Ureplasma-specific antibiotic administration.[11]
Regarding tracheal colonization with Uu, routines for both diagnosis and therapy have
already been established in most neonatal intensive care unit. As a general practice,
macrolides are used as an oral formulation. For eradication with azithromycin, differences
between centers exist regarding the administered dosage.[15] Consideration should be given to standard monitoring of the success of eradication
therapy when Uu is detected in the TS or OPS, as persistent colonization is a potential
source of CNS infection. Lastly, it remains unclear in this case whether it is a case
of inadequate eradication or reinfection.
To date, it is unknown how Ureaplasma spp. infiltrate the central nervous system, but a disrupted blood-brain barrier appears
to be associated with infection.[11] In this context, barrier disruption may be primarily caused by infection with Ureaplasma spp,[8] but it may also occur as a consequence of inflammatory processes following events
such as intracerebral hemorrhage, thus facilitating the translocation of bacteria.[16]
Increased permeability of the blood-brain barrier could explain the efficacy of clarithromycin
in our case, although in the literature it has been described as having little or
no CSF penetration.[17] However, attempting to quantify drug levels in CSF with TDM did not yield measurable
levels in our case. This may be due to the rapid distribution of macrolides from fluid
into tissue. In addition, no established and standardized TDM protocols are available
for clarithromycin which would specify the ideal sampling time relative to drug administration
as well as storage and transport conditions. Finally, studies correlating detected
CSF drug levels to clinical outcomes are lacking. Overall, the benefit of TDM for
clarithromycin in CSF is not yet determined.
Conclusion
Ureaplasma urealyticum (Uu) is a rare causative agent of ventriculitis in preterm infants. Unlike fulminant infections
caused by gram-positive or gram-negative pathogens, Uu infections are usually subclinical
or even asymptomatic. Therefore, CSF changes indicative or compatible with infection
and not improving under empirical antimicrobial therapy should prompt a targeted investigation
for Ureaplasma, especially in infants with previously diagnosed colonization with Ureaplasma in TS. In the case described here, prolonged therapy with clarithromycin led to therapeutic
success. Consequent eradication of Ureaplasma in the TS or OPS with subsequent monitoring of success could contribute to reducing
the risk of central nervous infection.
