Some weeks ago we published a post written by José María López-Lozano (see here). The post addressed one of the hottest topics in antimicrobial stewardship, which is the relationship between antibiotic use and resistance. Indeed, José María introduced the concept of antibiotic thresholds -thresholds of antibiotic consumption that could boost antimicrobial resistance- and antibiotic quotas .
One to the aims of the post was to introduce a debate on this topic. Herein we reproduce the thoughts of Dr. López Lozano about the questions that were brought up by Conan MacDougall (See comments in the original post). We thank both for their constructive discussion.
” Thank you very much for your interest, Dr. Macdougall.
First of all, I must say that non-linear approach is as new for you as for me and that unfortunately we do not have satisfactory answers to all questions. We explored some resistance phenomena in Aberdeen and found some very suggesting results (See here). Then, we continued exploring the same phenomena with other microorganisms in hospitals of four different countries, and observed similar finding (in a few daysthese latest results will be published).
Are we wrong? We do not know yet. Scientific knowledge needs independent contrasting. That is the reason for publishing and thus, our interest is to stimulate other researchers to independently challenge these concepts.
Regarding the existence of different time dependent thresholds, the selected methodology (MARS), allows identifying different knots delimiting different thresholds, whereas the explaining variable acts showing different intensities in its effect over dependent series. In our new paper we will show some examples. At the individual level, maybe your question needs another approach. MARS methodology apply only, under my knowledge, to the analysis of time series.
Our approach is based on the temporal study of the evolution of resistance and its influence factors (not only antimicrobial use). Maybe other approaches are desirable, for example, comparing different places with different use and resistance levels. Nevertheless under a time perspective, the theoretical framework could be something similar to this schema:
Depending on the window through which we look, we will have different possibilities:
- Window #1. A threshold of antibiotic use has been identified for a given antibiotic and it has been surpassed. Some time later, resistance begins its ascent. Remember that the relationship between antibiotic use and resistance is concomitant but not contemporary (delayed): See here and here.
- Window #3. Antibiotic consumption returns below the threshold and (some time later) resistance falls. Interestingly, the antibiotic could be used again with low risk of selecting resistance (at a population level).
- Window #2. We are unable to detect any threshold because there are not changing levels. This is, by far, the most frequent window.
What should we do if the threshold theory works? (just a little bit of sci-fi or, simply, prospective)
- To try to identify our windows/thresholds and,
- If we are in the second scenario to try to move to the third one and when we observe the diminution of resistance, continue using the antibiotic below the threshold.