RMstudents Technology meets medical education

Taken without permission from The Economist's website (this was originally printed in Oct 11, 2007 issue). I couldn't link to it because you need to be a member. With the amount of reading I need should be doing in school I just don't have time anymore so I've cancelled the subscription. I'll miss reading things like this.

Medicine

Blood simple

Oct 11th 2007

Because they lack an essential component, blood transfusions may be killing some of the people they are intended to save
<!--back-->IF THERE were any sure bets in medicine, you might think that “blood transfusions save lives” would be one of them. But there aren't. Even though deaths caused in the 1980s by accidental HIV infection mean that donated blood is now screened meticulously to keep it free of infectious agents, there is still a nagging feeling that something is wrong.

In 2004, for instance, Sunil Rao of Duke University Medical Centre, in North Carolina, carried out a study of people suffering from acute coronary syndrome (a specific type of heart attack). One conclusion that could be drawn from his research was that unnecessary blood transfusions might be causing tens of thousands of deaths in America alone. Dr Rao found that patients who had had a transfusion because of a low red blood-cell count had an 8% chance of dying within 30 days. Without a transfusion, only 3% died. Those numbers need to be treated with caution. As Dr Rao points out, the patients who underwent transfusion were, on average, sicker and older than those who did not. Nevertheless, his study is not the only indication of something amiss.

In recent years, research has suggested that transfusion is not necessarily a good thing for patients suffering from serious injuries, for those who have undergone surgery and even for those who are anaemic. And a study carried out earlier this year found that critically ill children whose red blood-cell counts had dropped by half fared no better after a transfusion than those who did not receive one.

As a result of all this, questions are being asked about whether something happens to blood when it is banked that causes it to stop working properly. What that might be has remained a mystery. But it may be one no longer. A group of Dr Rao's colleagues, led by Jonathan Stamler, think the answer is a gas called nitric oxide—or, rather, a lack of it.

The main reason for giving a patient blood is that it carries oxygen. It carries lots of other things, too, such as glucose. But it is a lack of oxygen that will kill you quickest. However, as Dr Stamler points out, what determines whether transfused blood works as a treatment is not merely how much oxygen it is carrying, but whether that oxygen can reach the tissues that need it. This is where nitric oxide comes in.

Nitric oxide increases the flow of blood to tissues by dilating the arteries that penetrate those tissues. The best known example is the erectile tissue of the penis (Viagra works by sustaining the signal that the gas gives). However, it is not just penile blood vessels that nitric oxide relaxes. When a red blood cell reaches any tissue in need of oxygen it releases nitric oxide in order to dilate the capillaries. Only then can it deliver its cargo. And that is doubly true of the cells in stored blood since red blood cells become less flexible with age, and thus less able to squish into capillaries. Dr Stamler thus wondered if a lack of nitric oxide was causing the problems associated with transfusions.

What he and his colleagues discovered, and published this week in the Proceedings of the National Academy of Sciences, was that the amount of nitric oxide in stored blood does indeed decrease—and does so rapidly. Within a day of storage, blood loses 70% of its nitric oxide. After a few days, up to 90% has been lost.

A second paper in the same journal, by Dr Stamler's colleague Timothy McMahon, confirmed this result (in fact, it showed that the initial drop of around 70% happens within three hours of collection) and showed that it was not caused by the way blood is processed, but merely by the passage of time. Dr McMahon also established that stored blood does indeed lose its ability to dilate blood vessels.

Dr Stamler is in little doubt about the significance of these findings. Furthermore, he warns that putting blood lacking nitric oxide into the body does not merely dilute what gas is already present in the bloodstream. Blood that is poor in nitric oxide will scavenge the gas from other tissues, causing the vessels in those tissues to constrict. If the tissue in question is heart muscle, the result will be a heart attack.

These papers, therefore, make a strong case that a lack of nitric oxide is creating the problems with transfusions—though as Michael Strong, the president of the American Association of Blood Banks, points out, they do not settle the issue once and for all. That would require a proper, randomised clinical trial.

And therein lies the rub. Because blood transfusion is such an old practice (it dates back to 1818) it has never been subjected to modern clinical standards. Nobody is questioning whether car-crash victims, say, should have transfusions after massive blood loss. Without it they would undoubtedly die. But for those not threatened with exsanguination it is far less clear whether a transfusion is a good idea. There are no rules about when to transfuse and who to do it to. These are matters of judgment, and knowledge is typically passed from doctor to doctor.

The good news from this study is that the problem should be easy to correct. If nitric oxide is what is needed, it can be added to banked blood just before transfusion. As part of the project, Dr Stamler tried this with dogs. He found that old blood replenished with nitric oxide is as good as fresh blood at relaxing blood vessels. And that, he thinks, points to a bigger possibility than merely returning blood to normal. Blood boosted with nitric oxide might be used as a therapy for people who have had heart attacks by providing extra oxygen in the crucial minutes after an attack, before the affected heart muscle has died. At that point, blood transfusions would no longer be part of the problem: they would be part of the cure.

Came across this article on how to build a computer and install OS X on it. Thought it would be fun to compare the custom-build against pre-configured Macs, and see which is the better deal.

With rebates the build was less than $800 (without, it was roughly $850). It doesn't use the best video card or processor, but they're certainly decent enough to run Windows Vista or OS X. The exact specs are:

• 2.2GHz Intel Core 2 Duo processor - my guess is this is an e4600, high end of the low-mid end of Intel's lineup.. my current build uses an e4300 (so it is of the same series) which I have been able to overclock from 1.8Ghz to 3.0Ghz, so I think the e4600 can offer the same sort of performance if not better (with the new G0 stepping). But with no virtualization support and 2MB of cache, I'd spring for the e6xxx (or q6xxx) series. That way you can run Windows on top of OS X via Parallels and utilize the full power of the processor, and have a faster FSB AND bigger cache. Or what the hell, if it's 2008 get a Penryn-based processor.
• 4GB of RAM (four sticks at 1GB each) - RAM is so cheap now, and although I'd spring for low cas latency I guess they're out to prove a point so they just got real cheap RAM. Also OS X can use all 4 sticks, which is nice when you compare it to Vista 32-bit, which can only use 3 GBs max.
• ASUS P5W DH Deluxe motherboard - built-in Wifi and what not, I love ASUS but my latest motherboard is MSI and I really like the cost to feature ratio on it, and support has been awesome, so I think I'd get an MSI for this kind of build. Advantage of the ASUS is we KNOW it works.
• GeForce 7300GT - I think they did this to ensure compatibility with the graphics API in OS X .. but I'd prefer to use an 8800GT
• 500GB hard drive - I'd buy 2 of these.. there's a crack or something out there that'll let you use an internal hard drive with Time Machine. Usually Time Machine forces you to use an external HD (by not allowing you to set internal HDs as a backup).
• DVD burner

Wow, so that for $800 right? Let's see.. what can I get for $800 from Apple? Well certainly not an iMac since they start at $1,199. But it isn't fair to compare them because you get a big beautiful screen w/ an iMac. Fine. Let's ditch the MacBook and MacBook Pro as well (but those were much more than $800 anyway). We're left with the Mac Mini (the gimped Macs, in my opinion). Currently there is a $799 pre-configured Mac Mini available at the Apple Store online and it has the following specs:

• 2.0 Ghz Intel Core 2 Duo - This one has 4MB of cache and a 667Mhz FSB (the e4600 is 2MB Cache, 800Mhz FSB) .. I'd prefer cache over FSB, but one of the few places you really see the advantage of a bigger cache is gaming .. but w/ the e4600 sporting a faster frequency, a faster FSB, and because it is a custom build so you can overclock i, I'd say the custom-build wins here.
• 1 GB of RAM (2x512MB) - 4x the RAM on the cusom build... custom-build wins again .. and even if you put 2x2GB sticks in the Mac Mini, OS X wouldn't be able to see 4 Gigabytes because the Mac Mini doesn't use the new Santa Rosa chipset!
• 120GB HD - yeah, custom-build again .. 500GB!
• Intel GMA 950 Graphics - Ew, onboard graphics. Worse, this one is really outdated. Compare that to the 7300GT? No contest. Custom-build wins.
• DVD Burner ("Superdrive") - they're even here.

Now look at that. For the cost of a Mac Mini I can custom-build a computer and get pretty much the same processor (but with the freedom to overclock it), four times the RAM, a little more than 4x the storage (4.166x to be exact), and a dedicated GPU. The Mac Mini is overpriced, or it needs to be updated, or both.

Some would argue that you can't beat the Mac Mini's form factor. Maybe. But the custom build is only using one HD, one DVD drive, and a video card.. you can probably use a different motherboard, a microATX form factor, for example, and build a similar small form factor computer. It won't be as pretty but it will get the job done and it would get it done faster (video card) with more RAM and storage space.

This is one example of why the PC model is better for the consumer. I can pick and choose my components, find the best prices and the products which meet my specific needs, and better utilize my hardware (like overclocking the CPU or GPU).