Nosocomial bacterial meningitis commonly results from invasive procedures or complicated head trauma. The responsible etiological agent can be both Gram-negative and Gram-positive bacteria and depends on different features such as the pathogenesis and timing of the infection after a predisposing event.1 Nosocomial bacterial meningitis has an incidence rate up to 20% if catheters are placed for drainage.1

Pseudomonas aeruginosa neurosurgical meningitis is a rare entity, usually related with intraventricular catheters (IVC), and associated with high mortality rates and increased resource utilization and costs (including length of stay and antibiotic use/duration).2 The treatment of P.aeruginosa meningitis pose a severe challenge due to the scarce number of drugs available for its treatment and also for the poor diffusion of most of the antimicrobial agents into the central nervous system through the blood brain barrier.2–6 Traditionally, some drugs such as cephalosporins, quinolones, carbapenems, colistin or aminoglycosides, the two-latter used also intrathecally, have been used for the treatment of meningitis caused by P. aeruginosa.3–7 However, in recent years, the emerging and spread of multidrug-resistant (MDR) strains, especially those resistant to carbapenems, have really complicated the treatment of P. aeruginosa infections.8

The aim of the present study is to describe the clinical features and outcomes of a series of patients with nosocomial neurosurgical meningitis caused by P. aeruginosa treated with different therapeutic approaches and to identify the different prognostic factors related with these rare infections. Also, the scarce literature related to this rare entity has been thoroughly reviewed.

During the period of the study 51 CSF cultures of 51 different patients positive for P. aeruginosa were reviewed. Fifty-eight percent of patients were male and the mean age was 5018 years. All patients had undergone surgical procedures and also carried an IVC. The time elapsed between surgery and the onset of infection was 2120 days (limits 3-112 days). The mean of permanence of IVC before meningitis diagnosis were of 2214 days.

The most frequent underlying diseases were neoplasm and hemorrhage (33.3% both) followed by head trauma (27.5%) and hydrocephalus (6%). From a clinical point of view, fever was the most frequent symptom (100%), followed by headache (43%) and low level of consciousness (39%). Signs of meningeal irritation were detected in only 12 patients (23%). The CSF features were as follows: white cell count 6964 [33,569]cells/mm3, protein 321 [314]mg/dL, and glucose 49 [37]mg/dL.

Ten patients (19.6%) were diagnosed with polymicrobial meningitis. The other accompanying microorganisms recovered were Staphylococcus aureus (5 cases), Acinetobacter baumannii (3 cases) and Enterococcus faecalis (2 cases).

Regarding treatment, 33 patients received intravenous monotherapy with antipseudomonal cephalosporins (27 patients) or carbapenems (six patients). Other patients received combined therapy with ceftazidime and aminoglycoside (three cases). Fifteen patients received both intravenous and intrathecal therapy with cephalosporin or cabapenems plus intrathecal aminoglycosides (nine cases), or intravenous plus intrathecal combined therapy with colistin (six cases). The median duration of antimicrobial therapy was 19 days (range 2-29 days). In patients with combined therapy (intravenous and intrathecal) both treatments were maintained during the same time (mean 17 days; range 2-23). Furthermore, in 29 patients the antimicrobial therapy was combined with removal of the IVC. In 19.7% of the cases the empirical treatment established was considered inadequate. Treatment features are summarized in Table 1.

There were not statistically significant differences in age, sex, underlying diseases, presence of mixed culture and catheter removal between the different treatments (Table 1). Only three patients had a relapse. Seventeen patients (33.3%) died as a direct consequence of the infection. They had been treated with ceftazidime (12 cases, 70.5%), carbapenems (2 cases, 11.7%), carbapenem plus intrathecal aminoglycosides (2 cases 11.7%) or combined intravenous therapy (1 case, 5.8%).

In the univariate analysis mortality was higher in the group of patients treated with ceftazidime (12 vs. 15, p=0.068, OR: 3.040 [0.877-10.544]) and also in patients under inadequate empirical treatment (7 vs. 3 p=0.010, OR: 7.230 [1.31-44.77]. Moreover, mortality was lower in patients whom had received intrathecal therapy (2 vs. 13, p=0.050, OR: 4.64 [0.80-34.93], without differences observed between those patients treated with aminoglycosides and those treated with colistin. Finally, no patients treated with colistin died (0 vs. 6 p=0.067, OR: not defined) (Table 2). In the multivariate analysis mortality was only significantly higher in patients without catheter withdrawal (p=0.014) and statistically significant lower in patients who received intrathecal therapy (p=0.05) or adequate empirical treatment (p=0.006).

P. aeruginosa meningitis is uncommon and is frequently associated with CSF fistulas or invasive procedures derived from neurosurgical interventions. As in most of nosocomial meningitis, the presence of IVC has been shown to be an important risk factor for the onset of infection by this microorganism.1,12

The lack of clinical expression of meningitis caused by P. aeruginosa can cause it to be masked thus delaying its diagnosis. As previously described in similar studies, the most frequent clinical manifestation in our patient series was fever which was present in all patients studied, followed by other symptoms such as headache or low level of consciousness.13 Interestingly, only 23% out of the patients presented with meningeal irritation signs. The absence of clinical expression may be particularly relevant in critical patients subjected to mechanical ventilation, for whom the presence of clinical signs go unnoticed because of their sedation status. This emphasizes the importance of a prompt and an accurate diagnosis taking control microbiological cultures to each patient with fever of unknown origin carrying an IVC or recently submitted to neurosurgical interventions, in order to establish an early diagnosis.

In addition to diagnosis, an important complexity of the P. aeruginosa meningitis is its therapeutic approach. The emerging and spread of MDR strains, resistant to drugs traditionally used against P. aeruginosa infections, complicates the antimicrobial therapy and is associated with higher morbi-mortality rates.8,14 Ceftazidime has been so far the treatment of choice for infections caused by P. aeruginosa susceptible strains.15 However, mortality in our series was higher in patients treated with ceftazidime, although this did not reach statistical significance. Other antipseudomonal cephalosporins such as cefepime, has been used to treat P. aeruginosa meningitis. Huang et al., compared the clinical efficacy of cefepime administered in prolonged perfusion versus intermittent perfusion for the treatment of 68 patients with nosocomial meningitis, seven of them caused by P. aeruginosa.16 They found a higher bactericidal activity and a length of stay reduction in patients treated with prolonged perfusions. Regarding carbapenems, meropenem is the only drug of the family approved for the treatment of meningitis, due to its less common neurological adverse effects. It is known to be a secure alternative for the treatment of susceptible strains,17,18 however patients treated with this drug presented a mortality rate close to 12% in our series. The establishment of an optimal meropenem dosage for treating meningitis has been the object of numerous studies.3,17 For instance, Lu et al. studied a series of 82 patients receiving meropenem 2g every 8h, 1g every 8h, or 1g every 6h for at least 3 days, concluding that the best curing results were obtained when the 2g were administered every 8h in extended perfusion (4h).17 Other studies had displayed that prolonged 4h meropenem perfusion is associated with clinical improvements of infection, microbiological eradication and less mortality than conventional dosage.18 With respect to other carbapenems, doripenem is more active that meropenem against P. aeruginosa and is capable to cross the blood-brain barrier in the absence of inflammation.19 However, there are only a few reports of doripenem usage to treat meningitis caused by P. aeruginosa leading to a good clinical evolution, but in combination with rifampicin or colistin.20

The emerging of MDR microorganisms have lead to an increase of the intrathecally treatment of meningitis, especially using aminoglycoside plus β-lactamic or colistin intrathecal plus colistin iv combinations.5,6,13,14,21–31 The crossing of the colistin through the blood brain barrier is limited, even in the presence of meningeal inflammation, so the combination of both routes of administration seems mandatory. Nevertheless, the correct dosage is unclear with recommendations ranging from 2.61 to 20mg/day.5,6,24–28 Although the 20mg/day administered in our study is a higher dose than the 10mg/day recommended by the IDSA,15 not local adverse effects were observed in our study in patients receiving this regimen.

Taken all data together, the use of intrathecal route of administration was a factor favoring good evolution in our study according to the multivariable analysis, without differences observed between the patients treated with aminoglycosides versus colistin. To highlight, no patients treated with colistin died. These results are similar to those obtained by Tangden et al., who found the highest survival rates in patients with neurosurgical Gram-negative bacilli ventriculitis treated with aminoglycosides administered intrathecally,23 although only four meningitis by P. aeruginosa were included in their study. Contrary, Shah et al., found a 3-fold increases mortality in children suffering Gram-negative bacilli meningitis treated with intraventricular aminoglycosides compared to those treated with intravenous monotherapy.31 However, the results obtained from last study are not extrapolable to ours, in which most of patients were adults in many cases without other therapy regimens.

The attributable mortality to P. aeruginosa meningitis was high in our study (33%), especially in patients without catheter withdrawal and for which the intrathecal route of administration was not used. This agrees with other studies in which mortalities of P. aeruginosa meningitis range between 20 and 80%.2,14 However, it contrasts with the study of Pai et al., in which not deaths attributable to P. aeruginosa ventriculitis occurred during the study, although a high diversity of the treatments were administered.13 Catheter withdrawal, was an independent factor of good outcome in our series, what supports the IDSA recommendations.15

As far as we know, our work is the broader described series of nosocomial post-surgical meningitis by P. aeruginosa and it shed some light in issues such its therapeutic approach. However, it presents limitations such as it is an observational retrospective study, not randomized, and in which the antimicrobial dosage was not constant since it was prescribed according to different year guidelines. For instance, any patient was treated with loading dose of colistin, as recommended by the currently available guidelines.

While recognizing the limitations of our study, we want to highlight that nosocomial post-surgical meningitis by P. aeruginosa is an entity associated with high morbi-mortality, especially in those patients without intrathecal treatment and without catheter withdrawal. More studies are required to contribute to improve management of patients with this uncommon infection.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

The authors declare no conflict of interests.