No abstract

1 The biliary excretion of erythromycin has been studied following parenteral administration in 23 patients. 2 Mean bile levels of the drug were approximately ten times higher than corresponding serum concentrations 1 h after i.v. (erythromycin lactobionate) and i.m. (erythromycin ethylsuccinate) injection. 3 Thus, unlike many antimicrobial agents, these compounds are well concentrated in the bile.

SUMMARY Patients with chronic active hepatitis or alcoholic cirrhosis have serum antibodies to many more serotypes of Escherichia coli than do patients with primary biliary cirrhosis or cryptogenic cirrhosis, or normal controls. They also have antibodies against more serotypes than cirrhotic patients with a portacaval shunt. These observations suggest that factors other than shunting of blood away from the liver are responsible for the increased range of antibodies. These factors are discussed. There was no correlation between the number of serotypes to which antibodies were present and the serum immunoglobulin concentration. In three patients, each with chronic active hepatitis, the antibodies were predominantly of the IgM class, while the elevation of globulin in general was mainly due to increased IgG and IgA levels. Antibodies to Escherichia coli, therefore, probably contribute only a small part of the increased globulin levels found in patients with chronic liver disease.

The incidence of primary resistance to rifampicin in Mycobacterium tuberculosis has been analysed in countries where rifampicin is restricted to use for treating tuberculosis and in countries where its use is not restricted. There is no evidence that rifampicin-resistant M tuberculosis strains are more common where the use of the drug is unrestricted. Resistance to rifampicin is less common than is resistance to streptomycin or to isoniazid. We can thus see no danger of producing resistant strains of M tuberculosis if rifampicin therapy is used for short periods for non-tuberculous infections. The problem of resistant mutants arising in the non-tuberculous species being treated is overcome by combining rifampicin with trimethoprim.Rifampicin has a uniquely wide spectrum of antimicrobial activity, which includes Grampositive and Gram-negative organisms as well as Mycobacterium tuberculosis.' Because of this, in a relatively early phase of the clinical use of this antibiotic for non-tuberculous indications, rifampicin was employed alone in infections of the chest2 and urinary tract.3 4 However, the observation that in these circumstances resistant mutants were rapidly selected caused the use of monotherapy with rifampicin in non-tuberculous infections to fall into disrepute.As a logical means of overcoming this problem, a combination of rifampicin and trimethoprim was envisaged, based on a "double blockade" type of approach. The basic assumptions were confirmed in a series of microbiological studies,5 6 which indicated both the existence of synergism between the two compounds and the prevention of the selection of resistant mutants. The situation concerning fears for the development of resistance in M tuberculosis7 as a result of rifampicin/trimethoprim being used was summarised in a provocative editorial.8 Since then, the pharmacokinetic properties of the combination9 10 and its clinical effectiveness" have been evaluated, and confirmed the encouraging preclinical evidence. A further and most important part of the development of the combination has been, and continues to be,

A Lewis's statement that rifampicin "should be effective in infections caused by typically intracellular organisms such as. .. Brucella spp" lacks conviction (7 July, p 3). Kosmidis et al in a study carried out in Greece reported that rifampicin plus doxycycline was better than the recommended World Health Organisation regimen (tetracycline plus streptomycin),' and Buzon et al found rifampicin plus tetracycline significantly more effective than co-trimoxazole in treating human Brucella melitensis infections acquired from goats in Spain.2