Sir, Post-surgical meningitis caused by MDR Acinetobacter baumannii continues to be a serious problem owing to its high mortality and the limited therapeutic options, so combined treatment with IV and intrathecal colistin is currently considered as the treatment of choice. 1, 2 Perez-Alba et al. 3 describe the treatment of nine patients treated with a combination of high-dose meropenem and 20 mg every 24 h of intrathecal colistin. The study had to be suspended owing to the appearance of convulsions in six patients, worsening on the Glasgow Coma Scale in three patients and the persistence of a positive culture in another of the patients. The appearance of toxicity related to the local administration of colistin, especially as chemical meningitis, has been described on several occasions, with varying prevalence between 0% and 9%. 4–7 From our first description, 4 both our own experience and that of other authors 5, 6 have demonstrated the safety of the intrathecal administration of colistin, whether it is used for the treatment of meningitis caused by A. baumannii or by other Gram-negative bacilli. No adverse effects were found to be associated with it, although it must be noted that, as stated by Perez-Alba et al., 3 we divided the total dose into two daily administrations. On the contrary, Karaiskos et al. 7 performed a review of 83 cases, published in the literature, of patients treated with colistin by both routes, with doses varying between 1.6 and 40 mg, much higher than in Perez-Alba and colleagues’ work. Cure rates of 89% were described, with the appearance, in only nine patients (11%), of mild chemical meningitis that resolved after a decrease in the dose used or an increase in the administration interval; the suspension of treatment was only required in two patients. Chusri et al. 8 reviewed 33 patients with A. baumannii meningitis, 17 treated with colistin by both routes and 16 only with IV colistin, finding lower mortality, a shorter hospital stay in general and less time in the ICU, particularly in those treated by both routes. Only three patients developed chemical meningitis and two of them received 200 000 IU of colistin in a single dose. In all cases, the disease was mild/moderate and resolved with the suspension of treatment. The role that fractionation of the intrathecal dose plays in the occurrence of adverse effects remains to be clarified. Probably, the reasons for the discrepancy between these results and those presented by Perez-Alba et al. 3 are multifactorial. In their study, no data are provided on the degree of control of the underlying neurological diseases presented by the patients, which could explain the appearance of seizures or neurological deterioration. Furthermore, there are no data on whether patients received concomitant steroid treatment. In our series, 4 most of the patients received treatment with IV dexamethasone, following the existing protocols in our hospital, which could have contributed to better control of cerebral oedema associated with the underlying disease and a more moderate expression of chemical meningitis when this appeared. Regarding a possible role for meropenem in the appearance of seizures, we feel that this cannot be entirely ruled out, but is exceptional. Practically all literature reviews coincide in reporting its excellent safety profile in the case of underlying neurological pathology. 9 The authors also do not give information about the MICs of meropenem in the patients. Unlike in other infections, 10 in the case of meningitis the efficacy of the combination of colistin and carbapenems has not been intensively studied, which has resulted in only weak evidence for its use. It is recommended by some authors only if the MIC is …