Food allergy: response to treatment with sodium cromoglycate.

decline in chloramphenicol blood levels despite sustained IV treatment in high dosage. This decrease is probably explained by the induction of hepatic microsomal enzymes by phenobarbitone. Increased glucuronidation of chloramphenicol and a consequent decrease in therapeutic activity have been demonstrated in rats pretreated with phenobarbitone (Bella et al., 1968). The conventional chloramphenicol dose for meningitis and septicaemia is 50-100 mg/kg per 24 h in infants and children. A lower dose is usually recommended for neonates whose renal and liver immaturity delays chloramphenicol excretion, but Black et al. (1978) showed that doses of up to 95 mg/kg per 24 h may be needed to maintain therapeutic blood levels in this age group also. Our two patients show that during treatment with phenobarbitone even the maximum recommended dose of chloramphenicol, 100 mg/kg per 24 h, may be insufficient to maintain peak blood levels within the therapeutic range 15-25 mg/l. We therefore recommend that in all patients who receive enzymeinducing anticonvulsants (Conney, 1967), chloramphenicol blood levels should be measured every one or 2 days to avoid the hazards of excessive or inadequate dosage. It is possible that newer anticonvulsants with less enzyme-inducing activity, such as sodium valproate, will prove to have less effect on chloramphenicol metabolism than phenobarbitone.