Editor—Antimicrobial resistance in enterococci, particularly E faecium, may greatly restrict the choice of treatment. We report the case of a neonate with vancomycin resistant E faecium(VRE) meningitis successfully treated with quinupristin/dalfopristin (a semisynthetic injectable streptogramin) and chloramphenicol.

The patient underwent repair of obstructed infradiaphragmatic total anomalous pulmonary venous drainage on the first day of life. Recovery was complicated by an episode of necrotising enterocolitis for which cefotaxime and metronidazole were given. At the age of 13 days abdominal signs had resolved but the patient was febrile. Over the next three days six blood cultures yielded E faecium resistant to vancomycin, amoxycillin, erythromycin, rifampicin, and sensitive to chloramphenicol. E faecium with the same susceptibility pattern was grown from three intravascular lines removed at this time. Two doses of vancomycin and a single dose of gentamicin had been given when the lines were removed. There was only modest clinical improvement so treatment with chloramphenicol (30 mg/kg/day) was started. There was no evidence on echocardiography of intracardiac infection.

During the first three days of chloramphenicol treatment three more blood cultures yielded VRE. Teicoplanin was added—a loading dose of 16 mg/kg followed by 10 mg/kg daily. After an initial improvement the patient again became febrile and lumbar puncture was performed. Cerebrospinal fluid contained 16 million red cells/l and 380 million white cells/l. Culture yielded VRE. Minimum inhibitory concentrations of a range of antibiotics were measured by broth microdilution technique for the cerebrospinal fluid and blood isolates (table 1). The dose of chloramphenicol was increased to 75 mg/kg/day and the patient enrolled (with informed parental consent) in a compassionate use programme to receive quinupristin/dalfopristin, 7 mg/kg/day in three divided doses. Teicoplanin was discontinued. A second sample of cerebrospinal fluid obtained after seven days contained <1 million red cells/l, 15 million white cells/l, and was sterile. Chloramphenicol was continued with quinupristin/dalfopristin for 13 days. No isolates were detected in the stools of patients in the ward screened for the presence of VRE. The patient made a full recovery with no adverse events.

The baby probably developed intravascular line infection, possibly from an intestinal source, which failed to be controlled by removal of the lines and administration of chloramphenicol and teicoplanin, and this resulted in meningitis. Although enterococcal meningitis has been described in neonates with bacteraemia associated with central venous lines,1 meningitis with E faecium is rare.2 Central nervous system infections with VRE present a particularly difficult therapeutic problem as β lactams are inactive and teicoplanin penetrates cerebrospinal fluid poorly3 and in any event has little activity against some VRE phenotypes. Chloramphenicol is not bactericidal and there are few data on its use in VRE meningitis.

We were unable to measure the concentration of quinupristin/dalfopristin in the patient’s cerebrospinal fluid to assess penetration, but our clinical experience confirms that the combination has a role in the treatment of serious infection with vancomycin resistant E faecium; furthermore, it complements in vitro evidence that quinupristin/dalfopristin has an additive effect in the presence of chloramphenicol.(Messick CR, Pendland SL. Abstract EO111 presented at the 36th Interscience Conference on Antimicrobial Agents and Chemotherapy, New Orleans, 1996.)