The relationship between antibiotic prescribing and resistance

Antibacterial resistance is a global threat to the effectiveness of antibacterials, challenging their value for both the prevention and treatment of infections. The concept of antimicrobial stewardship has developed as a response to the problem of resistance. The assumption underlying antimicrobial stewardship is that by controlling the use of antibacterials, the rise in resistance maybe stemmed, and possibly even reversed. However, this assumption depends on the truth of the statement, antibacterial use leads to resistance. What is the evidence for this?

There are three main strands of evidence that point to resistance emerging as a consequence of antibacterial use. Laboratory evidence, ecological studies, and individual level data. Early laboratory experiments demonstrated that resistance to certain antibiotics could be readily generated simply by exposing bacteria to antibacterials. This type of resistance is usually due to a small number of mutations in the bacterial, chromosomal DNA. Mutations occur frequently in bacterial DNA. But the probability that any mutation will lead to resistance is very small.

However, because the number of bacteria in a test tube, on an agar plate, or in an infection run into the billions, it’s actually quite probable that there will be at least one successful mutation that will generate a new strain of bacteria resistant to the test antibiotic. While not all antibiotics are readily susceptible to point mutations like this, this type of evidence provides biological plausibility to the claim that resistance emerges as a consequence of the interaction of bacteria and antibacterials. Ecological studies examine the relationship between antibiotic use and resistance rates at the population level.

These types of studies can show a correlation between the amount of antibacterials used in a group of people and the proportion of bacterial isolates in that population that are resistant to those antibacterials. But it isn’t possible to say with ecological studies that patients with resistant bacteria are those who have received the antibiotics. That sort of relationship requires studies at the level of individual patients or volunteers when it is possible to demonstrate that prior antibiotics are followed by an increased risk of carriage or infection with resistant bacteria in the same person.

The next few slides are based on a very helpful review, which is referenced at the bottom of the slide. There are various ways of looking at and describing prior antibacterial use. One way is just to ask the question, has the patient received antibiotics? Yes or no? It is usually necessary though to set a time frame for antibacterial exposure. For example, has the patient received any antibiotics in the previous three months? A further refinement to this question maybe to ask about the length of course of antibiotic treatment, or to categorise exposure by asking about specific drugs, classes of antibiotics, or spectrum of activity.

Resistance can also be described in a number of different ways. The simplest way is to ask the question, are there any bacteria present in a specimen that are resistant to a particular antibiotic? Again, the question may be refined by asking about the level of resistance? By, for example, measuring the MIC. Another approach to detecting resistance is to look for molecular evidence such as the presence of resistance genes. There may be a focus on resistance to a specific antibacterial class, or alternatively, resistance to multiple antimicrobial classes maybe investigated. Resistance can be detected in a number of different specimen types. For example, surveillance cultures, such as throat swabs, or faecal specimens.

While the presence of resistance maybe looked for in clinical specimens, either any type of clinical specimen, or particular types of specimens such as blood cultures or sputum.

There are a number of features which support the claim that resistance follows and is caused by prior antibiotic exposure. A consistent pattern of resistance following exposure in a number of different, unrelated settings helps provide evidence of causality. A dose-effect relationship, where the amount of resistance changes in line with the amount of prior antibiotic use is another useful clue. The concept of cause and effect requires a temporal relationship, such that the cause, antibiotic exposure, precedes the effect, resistance. Statistical approaches such as time series analysis can help quantify this relationship. Finally, it is always helpful if the relationship between antibiotic exposure and emergence of resistance make sense. In other words, it is biologically plausible.

The following two slides illustrate some of these ideas. The first slide is a summary of a study carried out in Finland over six years. In this study, the authors gathered information on the amounts of penicillins, cephalosporins, and macrolides consumed in different regions of Finland, and correlated this to penicillin and macrolide resistance in Streptococcus pneumoniae isolates in subsequent years. This study demonstrated that macrolide and azithromycin use were associated with a subsequent increase in macrolide resistance in pneumococci in the following years. And that beto-lactam and cephalosporin use was associated with an increased rate of low-level penicillin resistance. However, the story was slightly complicated by the observation that high-level use of penicillins was not connected to increased rates of low-level penicillin resistance.

The final slide summarises a systematic review and meta-analysis of the relationship between antibiotics prescribed for community patients, and the subsequent presence of resistance in these individuals who had been prescribed antibiotics. 24 studies were included in the systematic review. The overall findings were that community patients prescribed antibiotics for infections of the urinary tract or respiratory tract had substantially increased risks of carriage of resistant bacteria over the following 12 months. Patients prescribed longer courses, or multiple courses of antibiotics, were even more likely to carry resistant organisms.

There is an increasing body of further evidence which points in the same direction, that antibiotic use leads to carriage of resistant bacteria in those individuals who have received the antibiotic. The clinical consequences of this effect include the possibility that patients treated with multiple courses of antibiotics are at risk of infections caused by difficult to treat resistant bacteria. Furthermore, people colonised with resistant bacteria may be a hazard to others in the community if the resistant bacteria spread from person to person.