Sniffing stools speeds diarrhoea diagnosis

[tangent]

Nature 19 February 2004

Sniffing stools speeds diarrhoea diagnosis
HelenR. Pilcher ,
Bugs' gases give them away.

The fumes from stools could help doctors diagnose diarrhoea in an hour. A new method uses the distinctive gases to identify the infectious bugs. Current detection techniques can take days.

Many different infections, including viruses, bacteria and parasites, can trigger diarrhoea and cause severe dehydration. The condition can also be the symptom of a larger problem, such as cholera or typhoid. So the exact cause should be diagnosed and treated as quickly as possible, says Chris Probert from the University of Bristol, who helped develop the test1.

The technique works by analysing diarrhoea's odour. Samples are placed in sealed glass vial and heated gently. "Essentially, you've got a fart in a tube," says Probert. The fumes are captured on a fibrous stick, which is fed into a machine for analysis. The gas is split into its constituent chemicals, which are identified by their mass and other properties.

Each infection has its own chemical signature, the team found. Patients with rotavirus - the leading cause of childhood diarrhoea - have ethyl dodecanoate in their gas. This compound, which smells of pears, was missing in samples from patients with other diseases.

Campylobacter jejuni, which can cause food poisoning and Clostridium difficile, which can inflame the colon, also have distinctive chemical fingerprints.

Diarrhoea kills an estimated 2 million people worldwide every year, mostly in the developing world. Children, people who are malnourished and those with weakened immune systems are most vulnerable.

Follow your nose

At present, the analysis machinery is bulky and expensive. But the equipment could be shrunk from fridge- to shoebox-size, and could cost as little as £10,000 (US$19,000), says Probert. After the initial outlay, the main expense is odour detection sticks, which cost around 30 pence each and can be reused many times.

Diagnosing disease by scent is nothing new. Patients with ketoacidosis, a potentially fatal condition that sometimes accompanies diabetes, have fruity breath. Suffers of colitis, an intestinal disease, sometimes have strange-smelling stools. Physicians can spot this, says intestinal infections expert Edward Ryan of Massachusetts General Hospital, Boston. The new technique automates this skill.

The team tested the method with samples from 35 patients and verified their diagnoses against standard methods for identifying bugs in the laboratory.

They now need to do tests in more patients, says diarrhoea expert Kellogg Schwab from Johns Hopkins Bloomberg School of Public Health, Baltimore. "The method is interesting but needs a lot of evaluation before it will be a functional assay," he cautions.

References

1. Probert, C. S. J. et al. A novel method for rapidly diagnosing the causes of diarrhoea. Gut, 53, 58 - 61, (2004).

http://gut.bmjjournals.com/cgi/content/abstract/53/1/58
(full text online w/ links to 3 articles that cite this one)

[DrBaboon]

Interesting report.

In no particular order - some thoughts.

The link confirms they are are talking about using Gas Chromatography/Mass Spectroscopy. As an undergraduate, MS was beyond my college's well equipped chemistry lab. It's more available now. Gas Chromatography isn't austere in the usual sense, although I wouldn't be surprised if limited versions have been incorporated into space probes. I did an undergraduate project which included refurbishing the chemistry department's gas chromatograph, and then fabricating a hardware interface and machine language programming to get it to send data to the physics department's minicomputer. I think it took me about 6 months.

That being said, there's a lot you can dig out of a sample via GC. In a crude way, you can even get nitrogen and oxygen peaks (in the correct relative amounts) off an air sample injected into GC if the column is appropriate, and you have detector that will respond to nitrogen and oxygen. We mostly used thermal conductivity (generally applicable), but I also revitalized a flame ionization detector revitalized. In the application here (volatiles from stool), MS is basically a fancy detector.

How possible would it be for a trained person to dig a GC unit out of an abandoned school and put it to use? Probably "easier" than many other fancy instruments. But that's a real leap, and the relative comparison still makes it pretty ambitious. By fancy, I mean things like NMR, IR spectroscopy, MS. Simple UV/visible spectroscopy is a snap compared to even jumpstarting GC.

To me the more useful tidbits are: 1) it reminds me of the various times I've been involved in discussions with other physicians and nurses of whether a person's stool smells like Clostridium difficile -- so simple bedside acumen is reinforced by this report, which is always applicable for us; 2) if you have endemic bugs, or an epidemic, you have a high prior probability, and so a simple yes/no result of an assay may have a helpful ability to predict if the person has/does not have the illness you are considering. Put that into play, and a wet chemistry assay for something related to what you are looking for in that one bug is possibly within the scope of austere medicine. For example - a rapid urease test (possibly even pH change once supplied with urea) as the basis of CLO in possible Helicobacter pylorii - it doesn't have to be a *wildcard* anything-goes-tell-us-what-it-is test (like GC-MS identifying a volatile chemical in extracted material). It doesn't even have to be the various breath tests used for Helicobacter to be useful. 3) all of this takes me back to the semi-automated (or at least coded) identification of bacteria. Everyone who has taken a microbiology course has seen those tree diagrams that work you through deciding the identity of a bacteria colony. It turns out that each yes/no decision and branch in the tree can be converted into binary results of 1 or 0. If you group the different branches into 3's, you have octal numbers. Roche at one point (maybe still does - or perhaps even this is too inefficient for modern clinical labs) had nice tube that were cigar length, that you could innoculate across a bunch of different agars/reaction mixtures, and get a 1 or 0 result for each, and construct your octal number. The code book for the system gave you percent likelyhoods for different species associated with an octal result.

Again - complete octal results may not be what's needed - are more qualitative result, or one that is fitted into the context of clinical/community status, may do just fine. That context establishes high or low prior probability.

I'd tend to look at the information provided by this report as something we can use to train our clinical sensibilities, so we are better able to dig out one "wet test" to apply to a situation.

[FlightERDoc]

You may be able to refurbish a GCMS from a school lab, but the necessary gas (silane?) won't be available....when the bottle is out, it's out.

C. difficile and C. jejuni have ... unique ... scents that are more or less useable for diagnosis. British trained docs (including some eastern Canadians I work with) are familar with them.

[DrBaboon]

I risk making it appear as if I'm somehow promoting GC inclusion in austere labs, which I'm not. There might be very narrow niches to make use of it - for example - for passing on skills/teaching lab tech & the litany of modern medicine during a rebuilding of society following a cosmic event, which is not what I'd base my plans on, but felt I should be fair by mentioning it.

More in the train of depositing information - even information that simply helps to broaden understanding, or information that is used to cross something off a list, or information that explains how things work, etc....

A variety of gases can be used with GC -- they are all inert. Helium, argon, CO2, nitrogen can be used, depending on the application, the column used, and what it is that you're trying to chromatograph.

Flow rates are in the milliliter per minute range. Higher flows may speed up the appearance of chemicals at the detector, but often don't allow separation of components in the mixture being analyzed. Slower flow rates produce longer "dwell" and better separation of components, but have long waits for components to pass by the detector. Slow flow is more discrete as a rule.

It's not too different of a situation of shielding gases in welding - whether one's a proponent of calling it TIG or a newer name like GMAW.

If you're going to run samples on GC, the oven needs to be on for a decent length of time to stabilize the temps, and gas needs to flow for the duration.

I don't recall changing gas cyllinders that often back in school.

I don't do a lot of hobby welding, but I have an inverter-type TIG unit, and an argon tank. IIRC, I've had the tank about 5 years and haven't run it dry, yet. It'd run dry a lot faster using it for GC. And it'd certainly run dry a lot quicker if I were welding more often.

I think we've got our 2 votes in favor of paying attention to odors in a clinical situation, rather than insisting on high tech.

OTOH - one of my own shortcomings is that I have difficulty taking a text description of odors and recognizing something clinically from that description. Maybe others don't have as much trouble as I have over the years, but for folks who aren't good at that way of figuring things out, I could see where it would help to train one's nose by confirming what you smelled with GC-MS.

IOW - it's a high tech substitute for a "scratch and sniff card." I have never found such a card, but my intuition is that my own learning style would be more compatible with an example "scratch and sniff card" than a text description.

[tangent]

I wasn't advocating GC in austere situations eithor, however the machine under development and the CBW GC's that have recently been developed (shoebox size) will make it more possable as the price drops.

Was kind of thinking, if you've got distinct chems after the first step on that swab, wouldn't there be SOME reagent or indicator that a drop of would say yes or no? - without all the bother of a GC?

"Scratch and Sniff" is a good idea, but I don't think the WWW Consortium has figured out how to embed it in a web page yet... :)

As to odors, what else can you Dx by sniffing?

-t

[FlightERDoc]

Ketoacidosis is a classic olfactory Dx

[tangent]

Yes, sweet smelling breath, but any less classic ones out there?

thanks,

-t

[blue gecko]

I'd like to comment on the odor ID concept from a mom's point of view...how can I do this without being laughed off this thread? Here goes. When my kids were sick as babies there usually was a distinctive air to their diapers. I always knew when the other child or even myself was coming down with the 'bug' before symptoms because I could smell it.
As the daughter of a microbiologist and having a bit of lab experience myself in food science, I find it rather interesting how discerning the nose can be even when presented in tiny amounts. Consider the trained noses in the perfume business...maybe there should be a medical course on odor. How many of you can recognize the smell of an open body cavity or formaldehyde? I've only done a few autopsies and I don't even have to have a sample around to remember the smell.
Now here is another interesting (at least to me) tidbit. A few months ago I purchased some young hens. While gathering the hens I noticed a distinctive odor to their feces. I just assumed it was stress. When I got home and after I got the girls settled I washed up. The next day that same odor (distinctly the same odor) was in my own feces. I didn't get symptomatic with anything but definitely had picked up something from the chickens. At the very least it was a real eye opener. Would I recognize the smell again? Absolutely.
My point on all of this is do you really need a GC in the field if you were to have some scent specific training ahead of time? It would be along the same lines as we do in Fd Sci to train taste testers or as the perfume business does for thier noses or dogs trained to scent whatever.

[DrBaboon]

Qualitatively, I agree with you. Thank you for adding your experience.

I believe that humans have differing abilities to describe and differing abilities to apply a narrative description of something.

My own abilities don't give me the words to describe the odor I recognize with Clostridium difficile, and I also don't do well hearing or reading someone else's description of an odor and then being able to apply it when I encounter an odor.

I am speaking of the association or cognitive aspect of the process, not the olfactory component of encountering an odor. It's as if I have expressive and receptive "aphasias" that prevent me from making use of those descriptions. I am able to qualitatively recognize an odor I have previously encountered, even if I am not able to put useful words to it.

There are folks who can hear music when reading sheet music in a quiet room, and folks who must encounter actual music to hear it.

I think you and I are saying similar things. IOW - that experience encountering an odor is what helps.

For those individuals who are blessed with the cognitive abilities to be able to translate the perception of an odor into a useful description and/or folks who can take a spoken/written description and translate that into a "virtual odor" they can use to compare to an odor they are experiencing, I have to say they are blessed.

On a separate but related issue: So-called "special sense" problems are an area of healthcare where we haven't spent much time and research. I have heard of an ENT group who focuses on it (things like taste and smell disorders), but it seems to be primitive so far.

For my own geriatric interests, I have seen altered ability to smell or taste disrupt appetite and contribute to adult failure to thrive. Some of that is medication related, some seems to be dementia related (anosmia or dysosmia is somewhat sensitive for dementia, but not specific for dementia), and perhaps some is part of aging (whether abnormal or normal aging isn't clear).

On the medication angle, dysgeusias are something I encounter with some medications, most often ACE inhibitors and quinolone antibiotics.

None of which helps me with my afferent or efferent "deficit" on odor descriptions.

Yet another reason to keep coming back to promoting experience, experience, experience.

[tangent]

hmmm... I wonder what the shelf life of "scratch and sniff" cards is? ;)

It would make sence that a system could be worked out. For sounds in language we have phonems so it would make sence that someone could or has developed a system of scents...

-t

[blue gecko]

I know that in the food flavor business they have become very adept at oder synthesis. That's why McDonalds has been so successful with their burgers and fries. They actually add 'angus' to the products. I would think that if there were some real interest in the med community something could be developed. However, I doubt it will happen. Not concrete enough for records and no money in it without testing. In a remote, emergency situation though I can see its use as a basic diagnostic. Tangent maybe this is your ticket to fame and fortune! I can see the cards now: yeast, staph, acidosis, old folks lung smell (you know the one I'm talking about), bowel...
Could be quite a boon for the wilderness info community as far as classes are concerned!
Right now I'm experiencing a scent bonanza. My dogs were skunked last night and I've decided the smell is a blend of blood, dog poop and chanel #5 LOL. My plan of action is Biz an enzymatic product and Simple Green an organic cleaner....so far the spritz is neutralizing the air nicely now I'm off to scrub the dogs with it...wish me luck! B

[tangent]

Inexpensive electronic nose:

http://www.coe.berkeley.edu/labnotes/0106/chang.html

-t