Antibiotic resistance is an enormous and challenging problem for science and medicine. It needs to be a focus of research in order to minimize the development of resistance, but right now we are losing this battle. Since bacterial resistance is essentially an evolutionary process, we may be able to exploit evolutionary principles to minimize it.
The “golden age” of antibiotics lasted only from the 1930s to the 1960s, a very brief window. Since then the pace of bacterial resistance to antibiotics has been faster than the pace of development of new antibiotics. We have therefore been slowly sliding into bacterial resistance for the last 60 years. Resistance emerges when a specific species of infectious bacteria evolve a trait that allows them to bypass the mechanism by which an individual or class of antibiotics work. This happens at a rapid pace because of the number of bacteria that reproduce even in a single individual with an infection is very high, so the opportunity for mutations is also high. Further, the presence of antibiotics provides a selective pressure, favoring those with genes for resistance.
Bacteria pose a further challenge, because bacteria can exchange genes in packages called plasmids, even between different bacterial species. This means that once a gene that provides antibiotic resistance is out there, it can spread horizontally and confer resistance to other bacteria. Plasmids can even contain many resistance genes at the same time. The catalogue of all bacterial genes that confer resistance is called the “resistome” and currently includes about 20,000 resistance genes. There are also about 1,400 microbial species (including viruses, bacteria, fungi, protozoa and helminths) that infect humans. According to the CDC, in the US there are about 2.8 million infections each year with a resistant bacteria, causing 35,000 deaths.
This is also a global problem (which we should all be aware of, especially in the middle of a pandemic). A resistance gene that emerges in a distant corner of the world can find its way around the globe. This means that the entire world needs to observe best practices. Of course, this is not happening. In some countries people can purchase antibiotics over the counter, without a prescription. As many as half of antibiotic prescriptions may also be unnecessary.
While we are still struggling to enforce basic good practices to reduce antibiotic resistance, researchers are looking to tweak those best practices. At the very least, physicians who prescribe antibiotics should be doing so not only responsibly, but optimally. Since the emergence of resistance is an evolutionary problem, we may be able to exploit evolutionary principles to help minimize it (notice how I never say “eliminate” resistance – that is not even an option with current technology).
One of the evolutionary principles that has emerged in recent years is the sequential use of heterogenous, rather than homogenous, antibiotics. This involves using multiple antibiotics, either together or in series, that use completely different mechanisms of action. This reduces the probability that a single bacterium will be able to resist all of the antibiotics, so it will still be affected. The probability of simultaneously developing multiple resistances is a lot lower than evolving a single resistance.
A recent study, however, shows that there may be a lot of complexity to this basic principle. They show that sequential use of homogenous antibiotics of the same class can be effective. They tested different heterogenous and homogenous series of antibiotics and found:
To our surprise, we found that fast switching between only β-lactam antibiotics resulted in increased extinction of bacterial populations. We demonstrate that extinction is favored by low rates of spontaneous resistance emergence and low levels of spontaneous cross-resistance among the antibiotics in sequence.
In other words, if you use antibiotics that have an inherently low rate of spontaneous resistance emergence, you can still rapidly switch between antibiotics of the same class and this will increase killing of the bacteria. You can also use antibiotics that have a low rate of cross-resistance, meaning that resistance to one does not necessarily confer resistance to the other. This is an interesting result, because it’s surprising, but it remains to be seen if this will usefully inform prescribing practices.
The more important point is that prescribing patterns optimized to minimize resistance can be highly complex. Perhaps we should not be leaving this up to individual prescribers.
The end game
In just a few decades since the introduction of antibiotics, resistant strains started to appear, and now they pose a serious health threat. Where will this all lead? Some experts fear we will enter a post-antibiotic era, with most infections being entirely resistant to antibiotics. In fact, it is reasonable to argue that this outcome is inevitable. The only real question is how long it will take, not if it will occur. The more optimistic view is that we will be able to identify new antibiotics fast enough to indefinitely stave off the post-antibiotic crisis. It’s simply true that this is not happening now. If anything, resistance is accelerating, while the emergence of new antibiotics is slowing.
For the medium-term the goal is to accelerate the research of new antibiotics, while slowing the emergence of resistance. This will take a coordinated global effort. Unfortunately, such an effort is unlikely to be robust enough to work (a reality that applies to many urgent global problems). In addition to the WHO’s list of best practices, we will likely need a fundamental rethink about how antibiotics are used.
For example, hospital prescriptions of antibiotics usually do undergo some expert review, but arguably this process could be more ubiquitous. Perhaps it will also need to extend to out-of-hospital prescriptions. This would be logistically challenging with existing infrastructure, and may require new mechanisms entirely. For example, we could theoretically leverage artificial intelligence algorithms to optimize antibiotic prescribing – take it entirely out of the hands of human prescribers who have no hope of tracking all the data necessary to make the optimal decision. There are also some common sense systematic changes we can make, such as an international ban on OTC availability of antibiotics. We may further need to take specific antibiotics out of rotation at times in order to allow resistance to fade. Vaccines, of course, can play a critical role in preventing infections in the first place.
I think it’s unlikely that anything close to what is truly required will happen (but we can always hope). Antibiotic resistance requires a centralized global response that coordinates all antibiotic use. Antibiotics need to be thought of as a precious limited resource. Every use brings us one step closer to a post-antibiotic world of extensive resistance. I know there is great resistance to the notion of any kind of “centralized global control” but we just have to recognize that without it, worsening antibiotic resistance is inevitable. We are collectively mostly ignoring this problem while it slowly advances.
There is still the hope that we will technology our way out of this problem, although we cannot rely upon quirky scientific breakthroughs that haven’t happened yet. We may develop entirely non-antibiotic mechanisms for treating bacterial infections. Engineered bacteriophage viruses, for example, may be one option. Antibacterial nanotechnology would be nice. Generations, however, may live in a post-antibiotic world before such transformational technology emerges. Even if we are techno-optimists, it is in our best interest to slow down bacterial resistance as much as possible. We need to buy time for these new technologies to emerge.