Lots of stories on the wires (e.g., here) about a Nature Medicine paper describing a handheld microfluidic lab-on-a-chip to detect H5N1 inexpensively in less than 30 minutes. It was hard to understand what was involved from the news articles so I retrieved the paper (published online in advance of regular appearance in the journal hardcopy). It wasn't a particularly easy read, but here is what I was able to decipher.
This device makes use of microfluidics technologies, essentially an emerging set of techniques for manipulating very tiny volumes of material -- tiny as in millionths to billionths of a liter or even less. It is used in various so-called nanotechnologies (micropropulsion devices and other things I know nothing about) and also for the current application, a lab-on-a-chip. A lab-on-a-chip is a tiny wafer upon which several biochemical operations are performed in sequence, all in tiny volumes. Microfluidics is used to move the sample around to different locations on the wafer where various processes and analyses are carried out. The new device uses superparamagnetic particles to simultaneously isolate and then move the sample in a droplet from one tiny station to another on the chip. Paramagnetism is the property of becoming a magnet when another, external magnetic field is applied. When the external field is turned off, the material is no longer magnetic. Paramagnetism can appear when a regular (ferromagnetic) material is heated above a certain critical temperature (the Curie temperature) but superparamagnetic materials develop this property even below the Curie point. The loss of magnetism is a key element in this application because if the particles remained magnetic they would clump together. By turning off the external field, they can be separated and resuspended multiple times.
The purpose of the paramagnetism is to use another external magnetic force to move a droplet in which the particles are suspended. Thus this is a droplet-based microfluidic device, unlike many other devices where the fluid flows in tiny channels. The droplet technique allows the various processes to be changed rapidly without having to manufacture an entirely new nano device. You just move the droplet from one spot to another. This also allows the specialized moving mechanism to be separate from a disposable "bench" chip.
When operating with such tiny drops there are some unusual problems that appear. The surface area to volume ratio of tiny drops is much larger than the macro drops we are used to, and this device has several innovations or adaptations of existing technologies that get around these problems. The superparamagnetic particles can be attached to materials like silica that have biological affinities for various biological materials, including viral RNA. The particles with the RNA attached are then pulled out of the drop with a magnetic field, thus concentrating the RNA. they are then moved to another couple of drops in sequence to rinse the particle-attached RNA and moved to a droplet where thermocylcing is performed to amplify segments typical of a stretch of H5N1 hemagglutinin-coding RNA (see our explanation of PCR here). A detector is used to monitor the development of the target sequence -- if it's there. PCR at such small volumes is much faster, so the analysis time is greatly shortened. Here is the breakdown of times as given in Figure 5 of the paper: extraction and concentration of RNA on particles, 300 seconds; four washes, 40 seconds; reverse transcription, 180 seconds; hot start in thermocycler, 30 seconds; thermcylcing and melting curve analysis, 1100 seconds. This adds up to about 28 minutes. Current H5N1 tests run about four hours.
At least that's how I read the paper. If you have a subscription to Nature Medicine you will find the paper here. I will say I think this is pretty nifty. The authors claim the method is 440 per cent faster, and between 2,000 and 5,000 per cent cheaper than current methods while being just as sensitive. Could be. They haven't tested it in the field yet (it sounds like they tested it by spiking a throat swab from one of the investigators with some viral RNA because they didn't have a BSL3 lab to work with infective virus), and there is no mention of the specificity of the method, that is, a measure of its tendency to produce false positives (sensitivity measures its tendency to produce false negatives, that is, to miss cases that are really there). So we'll have to see if this device is as good as it sounds.
But it sounds good.
- Log in to post comments
More like this
I do think a more rapid, cheap accurate bedside test is required and this does sound promising, even if the science behind it is beyond me. I would very much like to see the specificity and sensitivity analysis. Diagnostics is one area of pandemic planning that is problematic. Timing is critical to control a pandemic virus. If containment is going to work it will be because of rapid diagnosis combined with a good public health response. I don't feel confident we have either in many of the places where bird flu viruses are currently spreading. If I recall average time from onset of illness to diagnosis of human cases around Jakarta was around 8 days and is part of the reason for the high mortality rate despite Tamiflu treatment (obviously not within the first 48 hours.) Even in first world countries there are many areas where diagnosis may take days if not weeks. These could be crucial in stopping the spread. I look forward to hearing more about this test and its application in the field.
From what has been described, here, I would assume that this method of analysis would be applicable (and it strikes mea as a very ingenious piece of technology; micro-propulsion coupled with RNA analysis; that is very impressive) only under circumstances where the individual being tested is already beyond the incubation period, and has reached, at least, the earliest stage of actively shedding viral particles.
If it does detect "only" H5N1's HA signature, (as you describe, Revere; and I believe this is a reasonable assumption, otherwise I doubt that the developers would be unveiling the device right now), then it seems to me that the possibility of a false positive is relatively remote. It could, I suppose, detect H5N2 and the other various Ns associated with H5 (that I don't believe we have encountered, yet), but I doubt that this really constitutes any problem.
I think that this is a genuinely impressive achievement, and not at all like so much of the other fluff and nonsense that seems to constantly be coming down the pipeline. Imagine being able -- finally -- to definitively separate an early stage H5N1 infection from all of the myriad diseases out there that so closely mimic it symptomatically. In a half hour. That is truly amazing. There is even a hint here, I believe, that the time factor might possibly be amenable to some future tweaking.
estimates, when such devices will be available ?
estimates, how likely this will work ?
estimates of the cost of one sampling ?
do they register short HA-subsequences (epitopes) ?
what area ? how long ?
could it be used for other epitopes ? for NA ?
revere, do you think it's important that we understand,
how it works ??
anon: We write here about what interests us and we like to know how things work. This was news so we looked into it. We still think the big issue is how it will or will not work in the field and what its specificity is. But the technology is nifty.
Dylan: If you use a throat swab as the sample I think you are right. The method, however, seems applicable to almost any environmental or biological sample as it cleans up using the silica adsorbent. My main issue with this is that it is a prototype not tested under field conditions with field condition samples, so we'll need to see both sensitivity and specificity (which was not reported and apparently not tested). So there's a long way to go. On the other hand, this is an example of new technologies being applied to a useful purpose (as opposed to studying crumbliness). I think/hope we'll see lots more of this. Even if it were only as good as existing technology but ten times cheaper and faster we'd be way ahead.