Prescribing antibiotics in primary care

Abstract

Studies of antibiotic resistance emphasise the importance of conserving this non-renewable resource

In this week's BMJ, Chung and colleagues report that community prescribing of a β lactam antibiotic for acute respiratory infection doubled the prevalence of antibiotic resistant bacteria in individual children.¹ This antibiotic resistance was transferred as a gene encoding β lactamase from other species of bacteria to Haemophilus. What do these results mean for the future of antibiotic prescribing in general practice?

Antibiotic resistance will probably eventually appear by natural selection for every new antibiotic developed by the drug industry, and the race to produce new drugs ahead of resistance is run ever closer. Antibiotics should be thought of like oil, a non-renewable resource to be carefully husbanded. What we use now cannot be used some time in the future.

The problem is that there is no scientific solution to convincing people not to seize for their own benefit a common resource best nurtured for the good of the community.² It is especially difficult for doctors treating a sick child not to provide (or, at least seem to provide) the best care, covering all possible outcomes, whatever the potential consequences for antibiotic resistance. What tools could clinicians have to balance the needs of the patient against protecting community resources?

It seemed for a while that draconian rationing would be needed to curb doctors' freedom to prescribe antibiotics—special administrative barriers to obstruct their use, rules of indications settled centrally, and so on.³ There are three reasons why this has not been necessary.

Firstly, one of the consequences of evidence based practice is the finding that antibiotics are minimally effective against most common childhood infections, mainly acute respiratory infections. This applies not only to infections commonly caused by viruses (colds⁴ and coughs⁵), but even those usually caused by bacteria (especially sore throats,⁶ and acute otitis media⁷). This message has been slow to infiltrate the community, although it is now appreciated by more people, and doctors now prescribe antibiotics less frequently—especially for sore throats and colds, if not for acute otitis media and bronchitis.⁸

Secondly, in some countries doctors delay the access to antibiotics as a type of educational ploy. A prescription is written (to show willing), but either left to be picked up later, or given to the parents to be filled at the pharmacy only if the child does not recover. This is effective at reducing antibiotic use.⁸ However, there is a problem. It is slightly dishonest, implying that starting antibiotics later is no less effective than using them immediately, which in at least some cases is not true.⁹ In a perfect world we would educate patients in a less indirect, if more difficult and longer, way.

The third reason comes from the personalisation of resistance from a population perspective (remote and indirect) to the individual (immediate and direct). The theoretical possibility that uncontrolled antibiotic use would increase resistance was confirmed by empirical studies. Australian children prescribed antibiotics in the previous two months were twice as likely to have β lactam penicillin resistant respiratory streptococci than children who were not prescribed these drugs, an effect that had not worn off six months later.¹⁰ The study by Chung and colleagues helps us better understand the mechanisms—how resistance is transferred between species in the child.¹

Why does this matter? These data will be seized on by those trying to curb antibiotic prescribing to provide a more cogent argument for not using antibiotics. The argument now focuses on the risks of antibiotic resistance to the individual patient—carrying bacteria with those genes might confer a greater risk of resistance if they later have a severe infection that needs antibiotic rescue. This possibility opens an agenda for future research. Can we test whether people with a serious infection died because of a prescription of antibiotics for a more trivial infection in the recent past?

Another potential research question might come from trying different approaches entirely, thinking about that vast horde of bacteria we carry around with us. To what extent is infection the consequence of an imbalance of this population, rather than our current oversimplified model of a pathogenic bacterium entering the body? We can offset diarrhoea caused by antibiotics by populating the gastrointestinal tract with “friendly” commensal lactobacilli.¹¹ Some preliminary research has shown the same can work for acute otitis media.¹²

In the meantime, doctors have new information to help convince patients (and themselves) that prescribing antibiotics for minor upper respiratory infections should be reserved for occasions when we really need them.