Tag Archives: selection pressure

“The Swedish Variant: Selection Pressure by Diagnosis”

When we think about selection pressure the first thing that comes to mind are antibiotics that selectively kill susceptible bacteria and thus allow more resistant bacteria to fill the ecological niche.

But fewer people realise that selection pressure can also be caused (indirectly) by laboratory diagnosis. Microbes which are diagnosed in the laboratory often end up getting treated and eradicated. However a microbe which mutates sufficiently to avoid diagnosis will have a selection advantage over its diagnosable counterpart. This concept is particularly applicable to microbes which are diagnosed by molecular techniques such as PCR where only a minor mutation or deletion can potentially create sufficient change in the base sequence to make the microbe undetectable by the original molecular test.

The most classic example of this is the “Swedish Variant”.

In 2006, a drop in Chlamydia trachomatis diagnoses was noticed on a particular molecular platform X, but not on others in use within Sweden. Further analysis revealed that a mutant strain of Chlamydia trachomatis (nvCT) containing a 377 base pair deletion was circulating. This was undetectable on platform X, but detectable on other molecular platforms.

Interestingly the nvCT strain had a much higher prevalence in geographical areas where platform X was used. In areas where other platforms were utilised, it wasn’t so successful as it didn’t have any selection advantage. But this makes perfect sense when you realise that a strain that avoids laboratory detection and consequently destruction is bound to do better than a strain that is easily diagnosed.

So what implications does all this have for laboratory practice?

Centralisation, tendering, and “packaged” contracts means that we are increasingly relying on just the one molecular assay to diagnose a particular pathogen within a large geographical area.

Laboratories or regions, or even countries which just rely on just one molecular test to diagnose a pathogen are always vulnerable to “escape mutants” such as nvCT emerging which escape detection and thus thrive in the population.

Testing a cohort of samples on alternative molecular platforms to validate the results and to look for these escape mutants is an important quality assurance measure.

The story of the Swedish variant also demonstrates the importance of using the percentage positivity rate of a molecular test over time as a Quality Control measure.

Even though the Swedish variant was diagnosed over 10 years ago, the lessons that can be learned from this episode are probably even more important in the large volume, centralised laboratory landscape that we have today.

In summary, one must be careful not to put all their eggs in one basket…


Check out this article for a more detailed overview of the Swedish variant. (about a 10 minute read)

Plasmids and Team Players

Let’s say you have a problem at your hospital with carbapenemases.

One of the obvious solutions would be to reduce the use of carbapenems in order to reduce the selection pressure.

However even if you stopped carbapenem usage altogether the carbapenemases would not necessarily disappear…

This is because carbapenemases are often plasmid borne, and there are often antibiotic resistance genes for other antibiotics, e.g. A, B & C sitting on the same plasmid.

As long as the (high) usage of antibiotics A, B & C continued then the selection pressure would favour plasmid retention in the bacterium, and thus allowing persistence of the carbapenemase.

Selection pressure by proxy.

Are we all doomed?

Not necessarily…

A gene expressing one antimicrobial resistance determinant comes at an energy cost to a bacterial cell. Plasmids expressing multiple resistance genes come at even more energy cost to the cell. You can be sure if it did not need the plasmid to ensure its survival, it would be mercilessly dumped, and probably sooner rather than later.

Therefore even a modest reduction in carbapenem usage, along with a reduction in antibiotics A, B & C may go a long way to solving your problem.

Advances in molecular methods and whole genome sequencing over the next decade will mean that it will become much easier to work out exactly which resistance genes are contained in the plasmids circulating in our local hospitals, and anti-microbial stewardship can thus be optimised accordingly.

Sounds space age?

Not really, we just need to be aware that resistant bacteria are very smart in an evolutionary sense, and we need to stay alert, and not give them the niches they are looking for…


Illustration courtesy of www.biologyfun.blogspot.co.nz

“The Antibiotic Free Period”

The patients who get the most courses of antibiotics are as expected, the ones who get the most infections, and these infections are often recurrent at the one site. A few examples are the elderly person who gets recurrent urinary tract infections, the toddler who gets recurrent otitis media, or the patient with chronic obstructive pulmonary disease (COPD), who gets recurrent bouts of bronchitis.

The antibiotic selection pressure on such patients is often intense, and one can often see by observing their microbiology results over time, that the infecting organisms become increasingly resistant, until multi-drug resistant organisms (MDROs) appear, and the clinician is forced to resort to less routine and more exotic antibiotics to treat the infection.

Whilst some of these “infections” will absolutely require antibiotics, many don’t, and many more were probably not bacterial infections in the first place.

What these patients really need is “An Antibiotic Free Period” . A period where the playing field is level. When MDROs have to compete against their susceptible counterparts in the absence of selection pressure, the increased fitness of the susceptible bacteria will win in the end. (This may take a while, and varies from patient to patient, and from organism to organism, but it will happen eventually.)

MDROs really don’t like level playing fields, they much prefer the odds slanted in their favour…

How can the microbiology laboratory assist in creating antibiotic free periods for patients?

Well we can add a comment to the result, for example “Uncomplicated otitis media does not routinely require antibiotic therapy” or “The isolation of pseudomonas from a patient with COPD does not imply acute infection.”, etc., etc.

or we can simply withold susceptibilities. For example an E. coli in a urine from an elderly Rest Home patient could have a comment along the lines of: “No clinical details have been provided with this sample. Asymptomatic bacteruria occurs in a significant proportion of elderly patients. If this patient has urinary symptoms, and they are continuing, please contact the laboratory for antimicrobial susceptibilities.”

It is my experience that the presence of an MDRO on a microbiology result report causes a reflex reaction from the requestor and increases the chance that the patient will be treated with antibiotics.

However the exact opposite should really apply. The threshold for treating an infection caused by an MDRO, as opposed to a susceptible one, should go up, not down.

One of the roles of microbiology laboratories, and clinical microbiologists, should be to facilitate antibiotic free periods where the opportunity arises…


There were seven new subscribers last week during the “Purple Cow Giveaway”. Rather than draw 5 from 7, I will ensure all seven new subscribers get a copy of my book. 🙂