Of stewardship, motherhood and apple pie.

Antibiotic stewardship is universally agreed to be desirable, but optimal models for stewardship remain uncertain. UK stewardship targets the particular antibiotic families-cephalosporins and fluoroquinolones-blamed for the selection of Clostridium-difficile-associated disease. To balance this there have been dramatic increases in the use of penicillin-β-lactamase inhibitor combinations. By channelling selection pressure in this way, we hazard destroying the utility of these antibiotic classes in turn, as happened with gonorrhoea where penicillins, fluoroquinolones and cefixime were sequentially lost as therapies. Strikingly, in context, almost all carbapenemase-producers are highly resistant to penicillin-β-lactamase inhibitor combinations, which may select for them. There is an urgent need to explore an alternative stewardship model, seeking to limit total antibiotic use but to maintain heterogeneity in what is used, avoiding concentrated selection pressure. There is also a great need to improve and accelerate diagnostics for infection and resistance, reducing or removing the need for protracted empirical treatment with broad-spectrum agents.