The optimal dosage of ornithine α-ketoglutarate (OKG) for repleting tissue glutamine (Gln) concentrations and maintaining N homeostasis after injury is unknown. We set out to perform ‘dose-ranging’ of OKG supplementation after an endotoxaemic challenge. Sixty-one male Wistar rats were injected with 3 mg lipopolysaccharide (LPS) from Escherichia coli/kg (n 50) or saline vehicle (9 g NaCl/l; controls n 11). After a 24 h fast, survivors were fed by gavage for 48 h with a polymeric standard diet (879 kJ/kg per d and 1·18 g N/kg per d) supplemented with non-essential amino acids (control, n 11; LPS-OKG-0·0, n 9), or with 0·5 g OKG/kg per d (LPS-OKG-0·5, n 12), 1·5 OKG/kg per d (LPS-OKG-1.5, n 11) or 4·5 g OKG/kg per d (LPS-OKG-4·5, n 10). The diets for all groups were made isonitrogenous with the LPS-OKG-4·5 diet by adding an appropriate amount of non-essential amino acids. Rats were killed on day 3 for blood and tissue sampling (muscle, jejunum mucosa, liver). Urine was collected daily for 3-methylhistidine and total N assays. The OKG dose was correlated with Gln concentrations in every tissue and with cumulative N balance (Spearman test, P<0·01). 3-Methylhistidine excretion was increased in endotoxaemic groups compared with controls (ANOVA, P<0·05) except in the LPS-OKG-4·5 group. Only the LPS-OKG-4·5 group achieved a positive post-injury N balance (t test, P<0·05). In conclusion, OKG exerted a dose-dependent effect on tissue Gln concentration and N balance, but only the highest dosage counteracted myofibrillar hypercatabolism and caused a positive N balance.