Effect of antibiotic prescribing on antibiotic resistance in children in primary care

A prospective cohort study has assessed the effect of antibiotic prescribing for acute respiratory infection on the prevalence of antibiotic resistance in individual children in primary care. The study involved general practices in Oxfordshire and 119 children, of whom 71 received a beta-lactam antibiotic, (amoxicillin 70, cephradine 1) at presentation and 48 received no antibiotic. The main outcome measures were antibiotic resistance as assessed by the geometric mean minimum inhibitory concentration (MIC) for ampicillin and presence of the ICEHin1056 resistance element in up to four isolates of Haemophilus species recovered from throat swabs at recruitment, 2 and 12 weeks.

The following findings were reported:

• In children who did not receive an antibiotic, the MIC for ampicillin was the same (2.7 microg/ml) at both follow-up points, with no significant change from the initial level of 4.1 microg/ml.
• In children who received an antibiotic, the MIC increased about fourfold to 9.2 microg/ml at the two week follow-up (ratio to no antibiotic group 3.5, p = 0.005); at the 12 week follow-up, it fell back to 5.7 microg/ml (ratio to no antibiotic group 2.1, p = 0.06).
• The ratio of MIC at 12 weeks vs initial visit was 2.33 in children on an antibiotic and 0.71 in those not on them (p = 0.05).
• Prescribing amoxicillin doubled the risk of isolation of the ICEHin1056 resistance element in Haemophilus isolates (67% vs 36%; relative risk 1.9, 95% CI, 1.2 to 2.9) two weeks later; this increase was transient and ampicillin resistance fell close to baseline by week 12. This element was recovered from most children from whom Haemophilus species were isolated (83%, 95% CI, 76% to 89%) at some point in the study, irrespective of whether they received an antibiotic.

Based on these findings, the authors suggest that in the few cases where it is appropriate to repeat the prescription of an antibiotic in under 3 months, it may be sensible to choose one that has activity against beta lactamase producing strains, rather than give a further course of amoxicillin. They add that from a population perspective, the issue of concern is the high (35%) equilibrium level of recovery of resistant Haemophilus species to which the children receiving antibiotics returned after 12 weeks and the endemic carriage of the ICEHin1056 resistance element by nasopharyngeal Haemophilus species. Therefore any reduction in community resistance is likely to need a substantial and sustained reduction in community prescribing; one option that deserves further investigation is to reduce the duration of each course of antibiotics prescribed in the community, another more radical option is to stop prescribing antibiotics to any child with respiratory infection in the community except in well defined and exceptional circumstances, though it does run the risk of some children with bacterial disease being treated later in the course of their illness.

The study concluded that prescribing amoxicillin to a child in general practice doubles the risk of recovering a beta lactamase encoding resistance element from that child's throat two weeks later; this risk falls back to the level seen before treatment within 12 weeks. The researchers note that though the short term effect of amoxicillin is transitory in the individual child, it is sufficient to sustain a high level of antibiotic resistance in the population. They add that “the ICEHin1056 resistance element is endemic in UK children, but reversal of this will require further substantial and sustained changes in antibiotic prescribing in the community.”

BMJ, published early online 26 July 2007; doi:10.1136/bmj.39274.647465.BE (link to abstract)