Saturday, May 30, 2009

Analysis of Meta-analysis

There have recently been several articles published in noted medical journals which derived from meta-analysis of pooled studies. Usually the purpose of a medical study/experiment is to see if intervention (X) affects condition (Y) in a patient, which condition may be discrete (e.g. heart attack) or continuous (e.g. blood pressure). Other studies are retrospective epidemiological studies, which should be used to suggest interventional studies. Unfortunately retrospective studies (e.g. a low fat diet lessons the risk of cancer) as well as meta-analyses (e.g. beta-carotene lowers the heart attack rate) have been shown on prospective, interventional studies to be incorrect. (Of course, the fact that a 5-year low fat diet in white females failed to lower their incidence of either breast cancer or colon cancer has not prevented the American Cancer Society from recommending a low fat diet to prevent cancer, but that discussion is for another blog.)

I find that medical statistics are still poorly understood by the medical-reading public, as well as by many doctors. For instance, the first study (of male cigarette smokers in Finland) to evaluate the effect of cholesterol-lowering by fibrates on the heart attack rate used a 2-tailed t-test, in case cholesterol lowering was harmful. When a null result was found, they redid the study, saying that lowering cholesterol could not possibly be harmful, and by using a 1-tailed t-test, the positive effect of cholesterol-lowering reached statistical significance. It is difficult to explain, without the use of mathematics, why changing the analysis in this matter is incorrect from a statistical point of view. (Just look at all the arguments about the Monte Hall problem, which is really trivial Bayesian analysis.)

Definition: meta-analysis is the combining and analysis of several studies, none of which usually reaches statistical or clinical significance because of small sample size, to determine the effect on a patient of an intervention.

In any event, it is my opinion that the result of any meta-analysis should be treated as an epidemiological event, and should only be a basis to suggest further interventionl testing. The real problem arises when a meta-analysis falsely suggests a negative effect, and then doctors are fearful of doing a prospective study because of the threat of malpractice. The basic problem with meta-analysis is its lack of mathematical rigor, and its basis on the subjective views of the lead investigator. I propose there should be an on-line journal of medical meta-analytic results, rather than formal peer-reviewed publication, and the results should be interpreted as tentative conclusions to be tested, rather than hardened "fact".

We also have read articles where two different meta-analyses of the same clinical intervention (e.g. low molecular weight heparin to prevent DVT's) reach opposite conclusions, and it is impossible to reject either study.

The following is a list of some of the problems that, in my opinion, prevent meta-analysis from reaching the statistical rigor and reproducibility that we should require of any process that affects human life and health:

Interpretation of "effect size" is totally arbitrary.

At least two authors are usually required to code the individual studies for grouping, because of the acknowledged subjective interpretation that occurs in coding. "Valid" studies require a reliability above 0.8!

The underlying level of risk in every studied group is of key importance, and often cannot be weighted properly. (e.g. the higher your cholesterol, the greater is the benefit of lowering it.)

The danger of Simpson's Paradox (or the Yule-Simpson Effect). This occurs when two smaller studies point to one result of intervention, and the pooled study points in the other direction. (Its existence is easy to demonstrate in vector space, but you can ask one of your statistical buddies to explain its basis to you.)

There are two different methods of combining results: the fixed effects model or the random effects model. The choice of approach strongly determines the results of the meta-analysis, and neither is always demonstrably "right".

Every meta-analysis requires subjective decisions, and few if any readers will redo the calculations with their own decisions.

Bias in selecting which studies to use.

There are TWO methods to measure effect: standardized mean difference, and correlation, and therefore there can be two different results (but only one is published).

There are various sources of heterogenicity and selection bias that cannot be determined.

There is often post-hoc sub-group analysis in meta-analyses, and this can easily lead to false statistical conclusions. This is similar in concept to the difference between an everywhere continuous function and a piecewise continuous function.

Results are often discarded if they do not make "medical sense", but it is always tricky and suspect to discard any result.

Comparing randomized trials may not be itself a random process.

The capacity to detect bias is small if only a small number or trials is considered.

There is no statistical solution to the problem of failure to reject the null hypothesis of homogeneous events, even if significant differences exist between the studies.

And finally: assessing the quality of a study in order to determine whether or not to include it in one's meta-analysis, is purely subjective.

Saturday, May 23, 2009

Men, Enlarged Prostates, and Cataract Surgery

The purpose of this issue of my blog is to analyze and discuss a recent article in the Journal of the American Medical Association issue of the week of May 20, 2009: "Association Between Tamulosin (i.e. Flomax) and Serious Ophthalmic Adverse Events in Older Men Following Cataract Surgery", Bell, C.M. et. al., JAMA, vol. 301, pp. 1991-1996, 2009.

A very brief "summary" would suggest that all men with BPH who are on Flomax, 0.4 mg., 1 or 2 at bedtime, should ask their internist or urologist to switch them to Uroxatral, 10 mg. (and no, I do not receive any money from drug companies).

A short discussion of human anatomy and physiology is in order. The kidneys make urine continuously, which collects in the renal calyces and drains by peristaltic action through the bilateral ureters into the bladder. At the bottom of the bladder, the urine exits through the ureter. The ureter first passes through the prostate gland, which sits directly under the bladder, and then through the center of the penis to the outside world.

A few parts of the human body continue to grow throughout adult life. Some of these are the prostate gland, the tip of your nose (which is why witches are always pictures as having long, curved noses), your ears, ...Now the prostate has 3 lobes, the right, left and middle. Your physician examines and estimates the size of your prostate (and looks for prostate cancer) when he/she does a rectal exam. However, the middle lobe, which is almost impossible to feel or estimate, is the part that concerns us here. The urethra passes through the middle lobe, where it contains smooth muscle in its walls, which helps to maintain the urinary sphincter so that you don't dribble urine, or accidentally urinate when you are ejaculating sperm and seminal fluid.

As the prostate grows, the prostatic urethra slowly gets compressed, so that its diameter is functionally reduced. It then takes you longer to start urinating, and longer to complete urinating. It also makes it difficult to empty your bladder completely. Since you make approximately the same amount of urine each day, then if you can't empty your bladder as completely as you once did, then you have to empty it more often, in order to maintain the same 24 hour output to the outside world.

Often the earliest sign of this that is troubling to men is the awakening at night caused by the need to evacuate some urine. By the time a patient has to awaken two or three times, then quality of life generally enables them to agree to take medicine for this condition.

The lining of the smooth muscle walls of the urethra in the prostate contain alpha-receptors, as do the walls of blood vessels. Without going into overwhelming detail, stimulation of the alpha receptors (e.g. by nor-epinephrine, a hormone produced by the adrenal gland) causes the urethra to contract further. An alpha-blocker enables the prostatic urethra to relax. This permits a man to empty his bladder more completely, and therefore, usually, to sleep through the night or awaken no more than once. (I am not discussing the use of Proscar or Avodart here.)

Unfortunately, alpha (actually alpha-1) receptors are also present in the dilator muscle of the iris, as well as its arteriolar walls. Because of this, then alpha blocking can cause the "floppy iris syndrome" during and after cataract surgery. The conclusion of the article was that "exposure to tamulosin within 14 days of cataract surgery was significantly associated with serious post-operative ophthalmic events. There were no significant associations with exposure to other alpha-blocker medications used to treat BPH".

You will have to read the article yourself or discuss it with your ophthalmologist to help evaluate how this article applies to you. I have already switched all my Flomax patients to Uroxatral, and explained why. I have had some battles with HMO's over this, since Flomax is available in a generic form while Uroxatral is not.

Good Luck!

Tuesday, May 19, 2009

Carcinogens and Vegetables

Many people worry about the carcinogens in their daily food. Cancer can occur when a cell divides if , when the DNA is duplicated, there is an error in transcription. A cell can then occur that reproduces uncontrollably, without contact inhibition or showing the Hayflick effect. The uncontrollable reproduction is a hallmark of all cancer cells. Thus cancer cells are truly immortal. The error in transcription can occur spontaneously, as a true mutation, or be induced by an outside process, called a carcinogen. The carcinogen can be a naturally occurring chemical, a man-made chemical, external energy (e.g.X-rays), it can be a chronically irritative process that stimulates the cell to divide, and thereby increases the chance of a mutation, or it can be a virus. To be technically correct, most carcinogens are inducers, i.e. they either increase the chance of a mutation or they increase the susceptibility of the cell to outside mutating influences. A few carcinogens directly cause cancer, but none of them will do so in 100% of the cases.

The first carcinogen was discovered by Dr. Percival Pott, a London physician (after whom Pott's Disease, aka tuberculosis of the spine, is named). He observed that London chimney sweeps (who generally started working when they were young boys, because of their size) almost invariably developed cancer of the skin of the scrotum, an extremely unusual cancer, at a very young age. Since they never bathed or could afford to change their underwear, he hypothesized that chronic contact with chimney soot was the inducer. By the simple expedient of having them wash their underwear weekly to remove the soot, he virtually eradicated this disease. His epoch discovery should be as well known as the story of the water pump handle and the cholera in London.

We all know that cigarette smoke is an inducer for lung cancer, as is exposure to radon, a radioactive gas, by uranium mine workers. Uranium workers who smoke have a synergistically enormously high rate of lung cancer. What is less well known is that 100% of smokers develop COPD, or emphysema, and sophisticated pulmonary function tests always show an abnormality in the closure of small airways in all smokers.

Radiation is considered to be carcinogenic, viz. Madame Curie, but consider cosmic rays. They are extremely high energy gamma rays reaching the earth from outer space. Denver, which is a mile high, has much less atmospheric protection from these high-energy rays than does New York City, yet their cancer rates are not significantly different.

The HIV virus is a significant inducer for Kaposi's sarcoma, the infectious mononucleosis virus for Burkitt's lymphoma, Hepatitis C virus leads to cancer of the liver (as does any process that has caused cirrhosis) and no doubt there are other cancer-inducing viruses waiting to be discovered. (Yes, I know this is not a complete list.)

However, a main human concern seems to be carcinogens in our food. Now, how do we decide that a chemical is carcinogenic? We usually assume that any chemical that causes cancer in rats, no matter at what dose, is carcinogenic. The fact that rat physiology and human physiology are different does not enter into our considerations, although their are chemicals that induce cancer in rats and not in humans , and vice versa.

Just to emphasize human-animal biochemical and physiological differences: only man and the guinea pig need Vitamin C in their diet; all other animals have the enzyme to synthesize it. Only the Dalmatian dog breed can get uric acid kidney stones, because other dogs excrete a different protein breakdown product in their urine. Many animals carry viruses that cannot affect or infect man, and some viruses that co-exist in animals can be lethal for man (e.g. swine flu).

Representative Delaney, from New York, had the Delaney clause added to a bill in Congress. This clause stated that any chemical that caused cancer in lab rats, could not be added to food. Of course, as soon as saccharine, the only available sweetener was shown to cause bladder tumor in rats, Congress immediately passed a saccharine exception to the Delaney clause.

Dr. Bruce Ames, a noted biochemist also developed the Ames test, which looked for DNA damage to bacteria in a Petri dish to determine possible carcinogens.

However, Dr. Ames also pointed out that there are many rodent carcinogens that occur naturally in vegetables, because the vegetables apparently develop these chemicals as their own pesticides. In other words, almost all vegetables contain carcinogens which, if extracted from the vegetable would be illegal to add to food, but the whole vegetable can be. How many of you remember when coffee was thought to induce cancer of the pancreas, or that preparing hot tea with lemon in a styrofoam cup liberated a carcinogen?

In any event, in the interests of completeness, the following is a partial list of everyday vegetables that create and contain their own carcinogens, even if you wash all the pesticides off:


saffrole oil

sassafras oil

black pepper



white potatoes

sweet potatoes















cherry tomatoes





jasmine tea



and soy lowers men's sperm count (it contains a phytoestrogen)

Friday, May 15, 2009

Prostate Cancer: Knowns and Unknowns

Prostate cancer is one of the (too many) fields of medicine where there is more heat than light, and more data than verified theories, and very few reproducible facts. I shall begin by listing what are , in my opinion, mutually agreed-upon facts. And please remember, that if there is argument in the refereed journals about the interpretation of facts, then there is no agreed-upon interpretation. Also, this blog is for both doctors and patients, so please bear with me if some of the data is well-known to you.

1) Only men get prostate cancer.

2)Eunuchs never get prostate cancer.

3) One treatment that slows down the progression of metastatic prostate cancer is androgen reduction and/or blocking its actions.

4) Yet, giving a man with low testosterone androgen supplements does not increase his risk of getting prostate cancer.

5)Sexual activity, whether heterosexual intercourse, homosexual intercourse, or masturbation does not seem to affect the incidence of prostate cancer (although some husbands tell their wives that it is a sure preventative).

6) If your father had prostate cancer, this increases your risk of getting prostate cancer.

7) There is no known dietary preventative, but the data does get twisted. For instance, one popular theory was that lycopenes in vegetables reduces the incidence of prostate cancer. A large dietary study seemed to indicate that both tomato and strawberry ingestion reduced the risk of prostate cancer. Yet, because tomatoes have lycopenes and strawberries do not, the authors of the studies immediately stated in their conclusion that strawberries had to be a statistical error, because it contained no lycopenes. This happens all too often in medical "science", where an inconvenient piece of data is "assumed away" because it does not fit the theory. In physics, we try to explain anomalous data, or re-investigate it (the advance of the perihelion of Mercury, atomic spectral lines, the black body catastrophe, etc.). In medicine it is ignored, so we never have "a beautiful theory ruined by an ugly fact".

8) There is no convincing evidence that surgery is superior to radiation therapy, and little evidence that either is superior to "watchful waiting". Yet in the newspaper reports of Mayor Giuliani's prostate cancer, watchful waiting was never mentioned as a possibility by any of the news articles, thereby planting a false impression of the "proper" treatment in non-medical readers. It has been my experience that only single men choose watchful waiting; a wife seems to be psychologically incapable of letting an untreated cancer stay in her husband.

9)THERE IS NO INCONTROVERTIBLE EVIDENCE THAT EARLY DETECTION AND TREATMENT EXTEND LIFE EXPECTANCY. If you look at the recommendations of the USPHS, vs Canadian HS, vs. AMA, vs. ACP, vs. Am. Urological Assn., you of course get differing opinions, much as urologists overwhelmingly recomment surgery, and radiation oncologists overwhelmingly recommend radiation. We internists are supposed to lay out all the data for the patient, and let him decide.

10) Proscar, aka finasteride, which blocks adrogen action on the prostate, reduces the incidence of prostate cancer, but then the cancer that does occur is more aggressive.

11) Under "first, do no harm", my inclination is against drawing PSA levels, but if there is a familoy history and the man's father died from prostate cancer, he generally wants to know.

12) Let me throw out a question to the doctors in the audience: We know that having had cancer is a risk factor for DVT. If a patient who has been treated for prostate cancer develops a DVT, should one assume that the cancer has recurred, and search aggressively for its spread?

Friday, May 8, 2009

Cholesterol and Heart Disease (1)

There have been many, many articles published, both in refereed medical journals and in newspapers, as well as reports from medical conferences about the correlation/statistical link between elevated cholesterol and heart disease, and about who would benefit from having his/her cholesterol lowered.

All commentators agree that elevated cholesterol is a "risk factor" for heart disease, in that, all else being equal, the higher the cholesterol, the greater the risk for plaque build-up in the coronary arteries. The real question, of course, is whether or not lowering the cholesterol lowers the risk for heart disease and/or heart attacks. The answer ab initio is far from obvious. No one would expect to lower an elevator in an office building by pulling down on the indicator arrow.

An excellent demonstration of the disconnect between risk indicators and risk improvement has been shown between high blood pressure, heart attacks, and strokes. Many studies have shown that untreated high blood pressure is a risk factor (i.e. is positively correlated with) both heart attacks and strokes, but, as far as I know, every study on the benefits of lowering elevated blood pressure only shows a reduction in the risk for strokes, and NOT a reduction in the risk of heart attacks. We may believe that we also help prevent heart attacks by lowering blood pressure, but there is no scientific data to support this belief. A doctor should only act on facts demonstrated by scientific research, or, to be fair to the patient, specify when he/she is acting on his own beliefs.

We also have to discriminate between primary prevention and secondary prevention. The greatest risk factor for having a heart attack is having had one already. Preventing a second heart attack would be called "secondary prevention", and since this group is at such high risk
it is easy to show the benefits or non-benefits of intervention. Showing the benefits of primary prevention is much more difficult, since the event rate is lower.

There has been a lot of argument in journals about various methods to obtain primary prevention, or reduction in primary risk. As a general rule, if scientists have to argue about the proper interpretation of data, then the data is probably useless. (Just try to follow any of the arguments of the use of meta-analyses.) Also, many studies confuse statistical significance (chance of interpretation being false less than 5%) with clinical significance (saves a significant number of lives in an absolute sense). For instance, if I could reduce your chance of being killed by a lightning bolt by 50% (relative benefit), but this meant a risk reduction from 1:1,000,000 to 1:2,000,000, you probably wouldn't be interested.

This initial blog is just setting up the framework for an in-depth discussion of cholesterol-lowering and heart disease prevention. Two caveats:(1) no one knows why lowering cholesterol with the use of Zetia or Vytorin showed no additional benefit, (2) the medical patients with the lowest total cholesterol levels, approx. 100, are patients with ulcerative colitis, who also have one of the highest rates of colon cancer.

Friday, May 1, 2009

The Flu and the Swine Flu

First of all, let me say that the current (Influenza A) flu vaccine does NOT give any protection against the Swine Flu. (Yes, I know the CDC is calling it A/@#$%^, but everyone knows what we mean by the swine flu and Lyme Disease, and Legionnaire's Disease, and Marburg virus). However, everyone should get a booster dose of the regular flu vaccine!

We have always known that the half-life of the adult antibody level induced by the Influenza A vaccine is only 3 to 4 months, so you should get a booster 6 months after your first inoculation with the same vaccine. I and my family all take it, as does my entire staff. I offer it to my patients, but 90% of them refuse it, since HMO's and MCR do not pay for the booster shot.

In fact, there was a recent article in a refereed medical journal that demonstrated that a vaccine with 4 times the strength of the standard dose would maintain viable flu antibody levels for one year.

Insofar as the new strain of flu, called swine flu is concerned, let me point out several facts, some of which need further investigation:

1) Pigs can carry viruses without getting ill (just as monkeys can carry simian hepatitis virus, which is lethal to humans, but not to the monkeys.) A pig can carry strains of virus from other species (chicken, duck, human), and they can mix and exchange RNA strands with each other, conferring new properties. Two new properties seem to be the ability to jump to humans with the ability to then jump to other humans, and, unlike the current Influenza A for which you may have been vaccinated, it is susceptible to Tamiflu. (75 mg 2 x day for 5 days for treatment, or 1/day for 10 days for exposure)

2) Fortunately for us humans, the infection is relatively non-lethal, and confers immunity. Since your infectivity begins the day BEFORE you show any symptoms, it is rather easy to tranmit. However, ordinary flu killed over 10,000 patients in the US alone this winter, nowhere near what has happened yet.

3)The flu apparently started either in Mexico, or on the California-Mexico border. We cannot be sure, because we do not have the index aminal. A greater problem is that we have absolutely no idea why it is so terribly more lethal in Mexico that elsewhere in the world---many theories, but no proofs.

4)Sterilizing your hanbds before eating is always a good idea, and perhaps so is wearing a properly applied surgical mask in crowds, as well a covering your cough. Patients with immunosuppression (steroids, AIDS, chemotherapy, anti-organ rejection drugs, etc.) should check with their doctors about special precautions to minimize exposure.

5) Two good sources of relatively unbiased information is the CDC website (Centers for Disease Control of the National Institutes of Health), as well as the IAC website (International Anti-infection Coalition).

6) I wonder if any organization is doing routine influenza surveillance on pigs around the world, much as they survey birds in China. It might be worthwhile to include swine influenza antigens in the next flu vaccine. Rather than protect against 2 strains of A and one strain of B, why not protecgt against 3 different strains of influenza A, since B is almost never lethal.

7) Take any government warning as understating the case, and any government re-assurance as too positive. The flu is like the economy, with random variations and unpredictability, and anyone with a definite forecast is operating on unvoiced assumptions that have not been tested or proven.