Slashdot Mirror
Retina Blood Vessels Predict Common Fatal Diseases
An anonymous reader writes "LiveScience is reporting that Tien Wong, of the Center for Eye Research Australia at the University of Melbourne, is claiming that abnormalities in the blood vessels of the retina can be used to predict diabetes, hypertension, stroke, and heart disease. These results were the culmination of several large studies. This could go a long way towards advancing medicine in the developed world as these disorders are some of the most common causes of death, hospitalization, and disability."
. . . with the potential to prolong many lives -- except that insurance companies will use the information to deny coverage, making it impossible for those identified to afford care.
I too have felt the cold finger of injustice.
...I'm just on an eight day cocaine/speed/Red Bull bender.
When I first went to a good eye doctor in 2002, she told me that the retina can tell a lot about a person's general health. She claimed she saved several patients lives when she found cholestorol build up in the vessels of the eye.
Nevertheless this is a good nonintrusive way to diagnose someone.
Really? They can tell I eat oats?
Be relentless!
If this is true, I sure as hell don't want my employer, or the government to have pictures of my retina to implement biometric security.
AccountKiller
Microhaemorrhages (bleeds) and aneurysms (a bulging section of a vessel) may be present due to hypertension in the vessels because of diabetic changes to the retinal vessels, or systemic hypertension. Having more blood vessels than usual indicates that the existing ones are not supplying enough oxygen, as will be the case in advanced diabetes. Changes in the macular (the part of the eye that sees most detail) can be apparent if the diabetes is causing problems with vision.
Stroke, heart disease, hypertension and atheroscelerosis are all intimately linked anyway - people at risk of one are often at risk of others. And it's difficult to characterise the changes to a specific cause. But they're still an important thing to look at.
Another time a doctor will look at the retina, is in an emergency situation where the is a blow to the head, they'll look for papilloedema (a swelling behind the retina due to increased pressure inside the skull). This also happens with other causes of high intracranial pressure, such as tumours.
For another thing, much of the body's inner workings remain a mystery. There are all sorts of incredibly important things floating around in the blood, and we have no idea they even exist.
For yet another thing, defining normal levels is difficult, in some cases impossible. Many body processes operate in negative feedback loops, where chemical levels are adjusted to produce a desired end result. If a person has a slightly insensitive detector for some molecule, and a slightly overactive emitter of it, their level might read as high when nothing is wrong. Worse, many blood levels depend on the exact conditions of the moment: sleep, hormone cycles, meals, psychological stress, minor viral infection, and so forth. Interpreting the results of a complete blood analysis would not be easy.
There is the Keyser-Fleischer ring (cupper deposits) that can be seen in the outer rim of the iris in Wilsons Disease (deficit in cupper metabolism). Although often times the ring is not seen after diagnosis has been made by blood tests. The disease is fatal unless treated, and also has a high rate of neurological deficits if treatment is initiated too late.
This is text book Chinese medicine. Looking for signs of these things in the face and eyes has a history of about 1500 to 2000 years. I guess it is good that Western medicine has finally come around, but this is far from amazing.
All data is speech. All speech is Free.
You raise a good point, but the problems are severalfold.
First, you need to identify the disease you're looking for.
Next, you go to the lab and spend lots and lots of time looking to develop a test that is both sensitive and specific for that disease.
A couple of definitions:
Specific - Only picks up that disease (in reality, a test with a low false positive rate)
Sensitive - Able to detect that disease every time it's there (a low false negative rate)
Take prostate cancer for example.
In laymans terms many people think the following:
Some cells of the prostate produce PSA (prostate specific antigen). If the PSA level is raised this means cancer.
A urologist thinks:
This patient has a raised PSA. Many conditions other than prostate cancer can produce elevated PSA including infections, trauma and masturbation. Also, some prostate cancers don't produce elevated PSA. So, what's the chance this one-off screening test showing a mildly elevated PSA means this patient has cancer? Also remember that this result is distributed along a normal curve and may simply represent what is normal for this particular patient despite being high against the population as a whole.
And that's the problem.
1) Identifying a test-able target molecule(s)
2) Screening the population at large to define normal limits. Recognising 5% of patients fall outside these limits given a normal distribution curve
3) Finding what conditions other than your specific disease cause an elevation in your marker molecule(s)
4) Finding the number of patients who have your disease but don't produce your marker molecule(s)
So the actual testing of a target, once defined and fully catalogued, is actually the easiest part of the process given modern analytical techniques.
It's the thousands of man-hours of lab work, research PhDs and patients involved in defining a test to begin with that's prohibitive.
And that's why many rarer diseases don't have simple tests to uncover them. It's also why your doctor will produce a whole battery of tests before giving you a definitive diagnosis.
It's not because we're procrastinating or not willing to tell you the answer.
-Nano.
Well, checking the ears won't identify "the runs" (diarrhea), but I routinely check patients' ears if they give a history of cough, headache, pink eye, or other head/neck symptoms. This is to help identify nasal congestion, a common cause of head/neck problems. Fluid (normally) produced in the sinuses, middle ear (behind the eardrum, so it can't come out through the eardrum), and eyes (excessive amounts would be called "tears") all normally drain into nasal cavity. If your nose is plugged, then you'd get sinus fluid buildup causing sinus headache, ear fluid buildup causing plugged and sore ears, and failure to adequately lubricate the eyes causing irritation and possibly infection. Furthermore, if your nasal mucus is running down the back of your throat, the "post-nasal drip" can cause a cough that won't go away with cough medicine. But try a decongestant, and the cough medicines will magically start working.
When I look in the ears, I can see whether there is fluid behind the eardrum, and tympanic bulging, which tells me that the eustachian tube (and therefore the nose) is plugged. If it is, then I can predict that the patient will probably have the sinus/eye/cough symptoms mentioned above. So the conversation would be more like:
P: I've got a cough and a headache.
D: Let me look in your ears. Yup, you've got a cough and headache.
By the way, about half my patients don't believe me when I tell them about their nasal congestion, because there's no mucus running out the nostrils, and their nose is unplugged enough so that they can breathe ok. Doesn't mean the eustachian tube's not blocked.
404555974007725459910684486621289147856453481154 in hex is "You sank my Battleship?"
[GPG key in journal]
Very good point: the eye, being transparent (in parts), is the one place where you get to see nerves and blood vessels directly, without cutting anything open. You can tell whether there is increased pressure in the brain (the blind spot will have ill-defined borders), and you can actually see whether there is cholesterol coating the blood vessels, or damage from high blood pressure or diabetes. You can even see arterial pulsations, if you look for them. This is a pretty routine exam I do when I see patients.
All this is from a $80 ophthalmoscope that you can fit into your pocket. As other posters have pointed out, this is pretty much old news.
404555974007725459910684486621289147856453481154 in hex is "You sank my Battleship?"
[GPG key in journal]
For the past couple of years my eye doctor has used a device called "Optomap" (http://www.optomap.com/) to capture a digital image of my retina. The first time he used it he identified that I had hypertension, which I had been diagnosed with a number of years before, and can even track it's progress. Had this device been in use before I was diagnosed my hypertension might have been caught earlier.
Some of what I say is fact, some is conjecture, the rest I'm just blowing out my ass...you guess.
... and a long time lurker, now turned anonymous coward :)
Anyways, as many others have noted, this concept is really not new at all. The fact that examining the retina can clue us in to systemic disease has been around for over 100 years. The novelty here, I believe, is that the researchers have prospectively (I'm guessing from the context of the article) examined digital fundus photographs, and found that they can be used predictively in determining the risk of diabetes/hypertension etc... (essentially, diseases that preferentially affect the small vessels in the body). But even this doesn't seem that novel, I can personally recall reading an article about 5 years ago which specifically looked at examining digital photographs to screen for diabetic retinopathy.
The other interesting part of the article has to do with their work on setting up a web-based digital review center. While the idea sounds great, this type of telemedicine runs into a lot of logistical problems, a sample of which would include:
-Limited sensitivity - ocular manifestations of systemic disease are not always present, and certainly are often NOT present early in the disease course. If we give high-risk individuals a "false negative" reading, are we really enhancing their overall health?
-Medical liability - who would be reading these? A trained physician? Licensed where? I am trained in the US, and fully licensed to practice, yet I can't even cross a state border and perform an exam without breaking the law.
-How are the digital images supposed to be obtained? Digital funduscopic cameras are pretty expensive.
-Practicality -- let's see, I could either buy a digital imaging system, and submit a photo to a website for review, or I could just check this patient's blood pressure, in the office (or check a blood sugar, or cholesterol level, etc...). These tests are actually much more definitive (we check patient's blood pressures in the office, before a retinal exam!), than an eye exam.
Ideally, what would probably be more effective would be something like a software package that can automatically analyze these images (with the proper legal disclaimers, of course), bundled with a portable digital imaging system. These would be wonderful for something like large scale screenings of under-served, high-risk populations -- an indian reservation, for instance. Then, you could identify especially high risk individuals, and recommend follow-up with either an ophthalmologist or internist. Though, really, everyone should probably at least have an occasional screening physical by a real doctor (ok, ok, enough soap-boxing).
Good grief. Sorry, you've pushed a button in me; I get this request from my patients all the frick'n time, so maybe if I answer this one, it will save a lot of doctors (and patients) some trouble.
Why don't we just do a "blood test that checks for everything"? The short answer that "there is no such test", but then my patient asks for a CT scan or MRI scan, or wonders when we will be technologically advanced enough so that we can invent such a test. It will never happen. Here's why:
Each test result is imperfect and must be interpreted in light of the clinical context. In other words, you have to know what you are looking for, or else you won't be able to tell what's normal or not normal. I view each test as being a somewhat inaccurate answer to a Yes/No question, e.g. "Is it Horrible Disease X?", but there is no test that tells you, "What disease is it?"
Just casting a net far and wide, looking for "everything", will net you all sorts of useless false positives that will waste time, cause worry, and worst of all, result in further unnecessary testing that may be risky.
There are two main reasons for this: test accuracy, and test relevance. Let me illustrate:
Elderly men everywhere are demanding to have their Prostate Specific Antigen (PSA) blood levels checked, because it's not as embarrassing as having the doctor stick his finger up the bum to feel the prostate. Hey, you gotta check the PSA, because --what if you have cancer!!?? How can you be SURE that I don't have prostate cancer, doctor? (Answer: you probably do, and the prostate cancer will kill you in 135 years.) PSA testing is still controversial, and in fact the Canadian task force actually labeled PSA testing as harmful because then you need to surgically sample the prostate, leading to the risks associated with minor surgery.
How about that? A test that's actually harmful, hmm? You'd think you could just ignore the test result if that were the case, but nooo... now there's a medicolegal obligation to follow through with prostate biopsy.
On the other hand, if you carry a high risk for prostate cancer, then the PSA is a very useful test. If your father and his father and his father all had (clinically significant) prostate cancer, then I would definitely get you a PSA without you asking for it.
There's a mathematical basis behind this, so let me illustrate.
Suppose there's a Horrible Disease X that everyone is afraid of. Let's say it's HIV/AIDS, which some 1 million people have in the USA, but this applies to any other disease.
Suppose I have a test that is 99% sensitive; that is, if you really do have Disease X, then only 1% of the time will the test erroneously say that you are okay. The test is also 99% specific; that is, if you DON'T have X, then only 1% of the time will the test erroneously say that you do have it.
I need a volunteer from the USA population to try my Super Duper Get-Checked-In-Case-You-Have-X test. You there, with the Slashdot username "lifeisgreat", why don't we try it on you?
Horrors! The test is positive! Life is not great, after all! What's the chance that you have Disease X? What's the probability that my Super Duper test, with 99% accuracy, is wrong?
Think out your answer before reading on. You have just tested positive for one of those "every imaginable things" test that you were asking for. What will you do now? Call a doctor? Talk to family? Write a will? Will you ever live to have kids? Will you ever get frist psot on Slashdot again?
The answer: there is les
404555974007725459910684486621289147856453481154 in hex is "You sank my Battleship?"
[GPG key in journal]