The most common hospital-acquired infections in intensive care units are ventilator-associated pneumonias ('VAP'), which have mortality rates between 20-50 percent. The more quickly the appropriate antibiotic can be given to the patient, the more likely the patient is to survive. Researchers in Spain have developed a rapid technique for determining which antibiotic (or antibiotics) to use.
First, a very brief bit about VAP. These infections usually result from intubation--a tube is inserted into the the mouth, through the larynx, and into the trachea, and is hooked up to a ventilator (think of it as a mechanical bellows for your lungs). VAP, and pneumonia in general, is very hard to diagnose accurately and quickly because there are often multiple organisms that are associated with the infection. Usually, one species of bacterium is the primary cause--get rid of it and the other bacteria will soon clear the lung.
To figure out which antibiotics will be effective takes at least two to three days:
1) A sample is collected, and sent to the lab.
2) The organisms have to be cultured (18-36 hours).
3) These individual strains then have to be tested for antibiotic resistance (4-24 hours depending on the organism, antibiotics to be tested, and available technology).
What these researchers did is develop a method that ignores step two. Rather then spending time trying to 'purify' the sample (i.e., tease apart the different organisms and then grow them in isolation), the sample (basically a small amount of bacterial laden snot from the lung) is smeared on a petri dish, and then the isolates are screened for resistance using plastic strips impregnated with antibiotics:
(note: This shows how the strips work; this is not from the paper--not that anyone could tell, myself included)
The numbers are the concentration of antibiotic at which the bacterium can grow. Growth above a certain concentration means that the bacterium--or in this case, the bacterial consortium--is resistant to the antibiotic.*
Not only is this method faster, but it improves patient outcomes: patient fevers were decreased in length (4.61 vs. 7.84 days), antibiotics were used for a shorter time (15.72 vs. 18.92 days), less antibiotics were used (31.43 vs 42.72 defined daily doses), and had far less Clostridium difficile**-associated diarrhea (1.8% vs 9.6% of cases), and patients came off ventilators faster (8 vs. 12 days). In addition, even factoring in the additional tests (the slow tests are still conducted for confirmation), 318 euros per patient was saved--which, sadly, might be the best way to motivate the current U.S. healthcare system. This procedure doesn't appear to lower the mortality rate, but in the control group, the patients were dropping like flies, so a larger sample might be needed to detect an effect.
This technique could also be used in the developing world, where cost savings are always critical. Instead of the plastic strips (which aren't cheap), one could use paper disks impregnated with antibiotics. It would look something like this:
There are some problems with the paper. How they decided the cutoff values for mixed bacterial cultures is unclear. Also, there are some minor statistical issues (corrections for multiple tests weren't reported; also I always like to see 'p values' reported). Nonetheless, this is a cheap, effective medical intervention that can dramatically improve infection outcomes--and lower antibiotic use and healthcare expenditures. With larger sample sizes, we might even detect if this saves lives.
*The correspondence between the culture results and what happens in a patient is determined by several things, including theoretical studies, animal models, and observed patient outcomes.
**You might have heard of this organism referred to as "C. diff."
- Log in to post comments
Interesting, the beauty is in the simplicity. Would love to see the paper. What journal is it in? Any Pubmed abstract or links?
Also - hasn't this basic culture resistance technique been available for decades? Sometimes it seems the greatest resistance lies in the willingness to adopt new or modify established methods. This is often justified, but there are times I just don't see the pragmatism.
Pardon my snark, but it seems like a medical doctor finally got around to applying something they learned in immunology besides "Throw more antibiotics at it". My attitude harkens back to when I worked at a university lab. The cross talk between med and grad students was very low to none at all. At times they were completely oblivious of each other.
Then I remind myself, even Mendel's work sat ignored for decades. So it goes.
Link to the journal article:
http://www.journals.uchicago.edu/CID/journal/issues/v44n3/40432/brief/4…
This procedure doesn't appear to lower the mortality rate, but in the control group, the patients were dropping like flies, so a larger sample might be needed to detect an effect.
Why would the high mortality rate in the control group obscure an effect? Wouldn't that make you more likely to see an effect on mortality than if only a few of the controls were dying?
Do you know where I could find an explanation for a layperson of how the e-test is different from the slow test? It's not clear what the costs and benefits are of cutting out the culture step. Or how they eliminated the need for a standard inoculum size.
Shauna --
The biggest difference I can see from the abstract is that the slow test (standard procedure) takes 2-3 days to figure out what antibiotic to use. What they do now is take the most common bugs for VAP, compare it to the susceptibility profile of their institution (ideally, but probably rare in practice), and start anywhere from 2-5 antibiotics (IV) to cover what they guess it is and what it's susceptible to. Then they wait until the culture comes back, then the susceptibilities (which aren't always ordered in practice). By this time, the patient has gotten anywhere from 6 to 25 doses of antibiotics, maybe half of which were needed. Drugs aren't cheap, and the cost of the drug plus the pharmacist's time plus nursing time = $$$. In the ideal world, the physicians will stop the unneeded drugs. In the real world, pharmacists pester the docs until they stop the unneeded drugs.
For the rapid test described in the article, you take the gunk, plate it out, and drop antibiotic-laden strips onto it before even determining the type of bug you've got. So you find out, in a very quick and dirty way, what'll kill the bugs. Though, as Mike points out, there are a few questions left unanswered (like how to account for not having standard inoculum.) I would suspect they still performed a good C&S for confirmation, or they should.
The e-test is what is currently done, but only after they've done a 24-48-hour culture and isolated the specific bacterium/a. The difference is that this paper reports a rapid e-test without waiting for the culture.
So the cost-benefit analysis is that is saves drugs and saves money, and that it decreases (risk of) antimicrobial resistance by decreasing exposure to unnecessary antibiotics.
Okay, perhaps I'm looking for more of a scientifc explanation than a layman's. I'm writing up this story for a class and I want to understand it thoroughly.
They did perform a culture in addition to the e-test (I have the full article) in the e-test group, to see if there were errors picking what would work and what wouldn't. There were 20-something errors with the e-test but nearly all of them were for one particular bug (p. aeruginosa).
I guess it's just hard to understand why something this useful isn't widely practiced.
Shauna,
the method is novel because it cuts out the step of culturing each bacterium individually. That will save anywhere from 24-48 hours (depending on how fast the bacterium will go). Essentially, from the time the sample reaches the lab until you get a result will take 24 hours instead of 2-4 days. Regarding the E-test, it is really just a high tech version of the older disk diffusion method (the time for each is comparable). The E-test has the advantage that its results are comparable to other, liquid culture methods, whereas disk diffusion is not.
Emily,
thanks for posting the link
thanks