Multiple brain abscesses caused by Serratia marcescens in preterm newborn

Introduction

The incidence of bacterial sepsis and meningitis is greater in the neonatal period than in any other period in life [1]. Brain abscesses are a possible but very uncommon complication of these conditions. The first report of brain abscess in a neonate was published 100 years ago [2]. The most commonly reported etiologic microorganisms are Citrobacter diversu, Proteus mirabilis, Pseudomonas aeruginosa, Serratia marcescens and other Enterobacteriaceae [3]. Very low-birth-weight (VLBW) and premature infants are especially susceptible to infections. Furthermore, modern neonatal intensive care techniques that include endotracheal intubation, parenteral nutrition, chronic blood vessel cannulation (umbilical, percutaneous intravenous and other central vessel catheterization), delay in feeding and administration of broad-spectrum antibiotics are risk factors for a number of acquired infections. We report a case of multiple brain abscesses in a preterm neonate as a complication of S. marcescens sepsis.

Case report

A 32 week-of-gestation female infant weighing 1990 g was delivered by an emergency cesarean section due to other’s severe preeclampsia at a tertiary level maternity hospital. Apgar scores at 1 and 5 min of life were 8. There were no indicators of perinatal infection. Immediatelly after delivery the baby was admitted to the Neonatal Intensive Care Unit (NICU). Due to respiratory distress she required nasal continuous positive pressure (nCPAP) with FiO₂ of 35% and a mean airway preassure (MAP) of 5.0 mm Hg. An umbilical vein chatheter was inserted and a blood sample taken for hemoculture and laboratory tests (blood count, acid-base status, blood glucose). Because of the preterm delivery and signs of respiratory distress we started the routine course of first line antibiotics (ampicillin, amikacin) until hemoculture. All laboratory tests were within the normal range and the hemoculture was negative. On the 3^rd day of life the clinical status deteriorated with hypotension, skin marmorization, prolonged capillary refill, shallow and irregular breathing pattern with the inability to maintain oxygen saturation. We intubated and started the mechanical ventilation of the baby. Fluid boluses and continous inotropes were needed. The hemoculture was repeated. The laboratory tests showed a positive sepsis screen – total leukocyte count 5.00 × 10⁹/L, absolute neutrophil count 1.9 × 10⁹/L, thrombocytes 28 × 10⁹/L, serum C-reactive protein 152.6 mg/dL. Because of a high suspision of sepsis and very difficult clinical conditions we decided to act empirically and change antibiotics prior to the laboratory isolation of the etiologic agent. We started a combination of vancomycin at 15 mg/kg/dose q 12 hourly and amikacin at 18 mg/kg/dose q 36 hourly. In the next 3 days the baby stabilized, the inotropes were stoped and she was extubated. After the isolation of S. marcescens in hemoculture, vancomycin was stopped and a single therapy with imipenem was continued. A cranial ultrasonogram performed on the 10^th day of life revealed multiple hypoechoic lesions in the parietal lobes bilaterally. A magnetic resonance imaging (MRI) of the brain performed on the same day revealed five hypodense lesions with peripheral enhancement suggestive of intra-parenchymal abscesses. One lesion was located in the right frontal lobe with a craniocaudal diameter of 28 mm another four lesions were located in the left hemisphere in the frontal, temporoparietal, paraventricular and temporal subcortical regions with diameters ranging from 11 to 28 mm (Figure 1). There was no hydrocephalus or extra-cranial collection of fluid. The cerebrospinal fluid examination was normal with a sterile culture and with no micro-organisms seen in the Gram stained slide, indicating the absence of meningitis. The neurological examination of the baby was normal. No convulsive attacks were noted. We consulted the neurosurgeon which suggested a conservative treatment with antibiotics and weekly neuroimaging follow-up. The MRI after 2 weeks showed an improvement, especially in the frontal regions (Figure 2). The MRI done after the next 2 weeks showed further regression of the lesions with the tendency to shrinking. The final MRI after a total of 8 weeks of antibiotic therapy showed the formation of small cavitations in the areas of previous abscesses (Figure 3). The clinical and neurological examination of the baby was normal as was the EEG. The baby was discharged with a recommedation of neurological follow-up.

Figure 1:

The initial MRI brain images showing multiple hypodense lesions with the peripheral enhancement located in the right frontal lobe with craniocaudal diameter of 28 mm and in the left hemisphere in the frontal, temporoparietal, paraventricular and temporal subcortical regions with diameters ranging from 11 to 28 mm.

Figure 2:

The MRI 2 weeks after the initial one showing an improvement especially in frontal regions.

Figure 3:

The final MRI after a total of 8 weeks of antibiotic therapy showing the formation of small cavitations in the regions of previous abscesses.

Discussion

Brain abscesses are a rare clinical condition and there have been no randomized controlled trials of the various treatments for brain abscesses. Brain abscesses in children should be managed by a multidisciplinary team that includes neurosurgeons and infectious disease practitioners, as well as neonatologists in the case of a neonatal patient. The management of brain abscesses may be influenced by the neurological status of the patient, the location of the abscess, the number and size of the abscesses and the stage of abscess formation [4], [5], [6]. Systemic treatment with antibiotic agents plays a critical role in the treatment. The duration of the antibiotics course is usually 6–8 weeks [7]. Stereotactic aspiration or surgical excision should be performed for abscesses larger than 2.5 cm or smaller lesions that are causing a significant mass effect [8]. A decline in the patient’s clinical status, enlargement of the abscess after 2 weeks of antibiotics, or the observation of no reduction in the size of the abscess after 4 weeks of antibiotic therapy should warrant a operation [9], [10]. Empiric antibiotic treatment with broad-spectrum agents is usually started until intraoperative cultures can be obtained, allowing tailoring of the antimicrobial agents to the identified pathogens. In patients who are treated by nonsurgical means only, specimens for culture will not be available and broad-spectrum antibiotics are needed. Serial imaging studies should be conducted on a weekly basis to assess the effectiveness of antibiotic therapy [8], [11], [12].

In our case, laboratory tests showed that a clinical deterioration was caused by S. marcescens sepsis. The clinical course implies that the brain abscesses were the consequence of hematogenic spread, althought only needle aspiration or surgical excision would provide the exact ethiology. The cerebrospinal fluid (CSF) was sterile which is not uncommon in cerebral abscesses. All of the five abscesses except the one in the right frontal lobe, measuring 2.8 cm, were smaller than the cut-off limit for surgical intervention, so we decided to treat conservatively. At first the baby was treated empirically with vancomycin and imipenem targeting Staphylococcus aureus and Gram-negative bacteria and after the isolation of S. marcescens we continued imipenem as a single therapy for a total of 8 weeks. Serial weekly MRI imiging showed rapid reduction of lesions. On the final MRI there was a total regresion of former abscesses with the formation of small cavitations.

At admission, the first-line antibiotics are routinely started and if the cultures are negative, antibiotics would be discontinued after 48–72 h. The need for a routine antibiotic therapy in all premature newborns after admission to the NICU is still questionable [13]. The majority of recent studies show that without the other risk factors this practice is unnecessary. When our patient’s condition deteriorated we started vancomycin and imipenem empirically targeting S. aureus and Gram-negative bacteria.

Our treatment had the most favorable outcome, and after the 8 weeks of systemic antibiotic therapy there was a total regression of abscesses. For the whole duration of the treatment the baby was stable and there were no neurological symptoms.

When treating central nervous system (CNS) infections the antimicrobial agents need to penetrate the blood-brain barrier and achieve concentrations in the CNS adequate for eradication of the infecting pathogen. The entry into the CNS depends on the compartment studied, molecular size, electric charge, lipophilicity, plasma protein binding, affinity to active transport systems at the blood-brain/blood-cerebrospinal fluid barrier and host factors such as meningeal inflammation and CSF flow. Carbapenems are beta lactams with the broadest antibacterial spectrum currently available. They are widely distributed in the body and penetrate well into tissues and fluids. In children aged 1–48 months of age and who are receiving meropenem in doses of 20 mg/kg or 40 mg/kg, CSF concentrations varied from 0.1 to 2.8 mg/L and from 0.3 to 6.5 mg, respectively. The large individual variations were probably related to the degree of meningeal inflammation. At a dose of 40 mg/kg the CSF concentrations were in excess of the reported MIC90 for most bacterial pathogens causing meningitis [14]. Imipenem has the potential to induce seizures in those on treatment for CNS infections and this effect is dose related. As at the time meropenem was not available in our situation, we administred imipenem and had the most favorable outcome without convulsions or other complications.

Conclusion

A multidisciplinary team approach, including neurosurgeons, neonatologists and infectious disease specialists, is needed for the desision on treatment of brain abscesses in a neonate. The treatment must be tailored based upon the clinical scenario of a patient, with the neurosurgeon playing a vital role in determining the type of operative procedures to be performed, if any. During the treatment regimen, serial imaging is important in the assessment of efficacy.

As the first hemoculture was sterile in this case it was probably a nosocomial infection. Unfortunatelly, infection is a common complication of many invasive procedures in modern NICUs but it is the price we must pay for the benefit we get.