Wednesday, April 22, 2009

Fat People Cause Global Warming

We all hate fat people. That's a given. But why do we hate fat people? (and by "we" I mean white, Anglo-Saxon, male researchers). The answer is simple: We're not allowed to hate black people. Don't believe me? Let's examine the stereotypes. "Black people are lazy" is now "fat people are lazy." First it was "black people like chicken." Now it's "fat people have a genetic condition that makes fried chicken taste delicious." There are also historical similarities between the two groups. There used to be a "poll tax" to stop black people from voting. Now we have a "twinkie tax"(1) to stop fat people from eating. Just like the "black codes" that stripped black people of their individual liberty, we now have "food codes"(2) that strip fat people of their dietary liberty. Then there was the "science" that proved "negro inferiority"(3). Now, we have the "science," which proves that fat people cause global warming (4). At least that's what a new article published in this April's International Journal of Epidemiology suggests (I like how the press release was published on Earth Day*).

So why are we allowed to hate fat people? Supposedly, it's unhealthy, "world-wide, over 1 billion adults are overweight and around 300 million are obese. The increasing global relevance of overweight and obesity has serious implications for health, increasing the risk of type 2 diabetes, cardiovascular disease, stroke, and some cancers."(4) As it turns out, "normal people" aren't that healthy either (5a, 5b, 5c). The other reasons that make it acceptable to hate fat people, they're unattractive and fat jokes are funny, were not examined in this study.

"There is some evidence that the entire population distribution of BMI may be shifting upwards, increasing the risks of disease for the whole population and not only for the most overweight in the upper tail-2"(4). In layman's terms, people are getting fatter. However, take a look at the title of the study cited, "Weight gain and its predictors in Chinese adults." Apparently, what happens to Chinese people can be generalized to the global population. This is another example of not appropriately supporting your statements.

So why should we direct our liberal rage at fat people? "The upward shift in the population distribution of BMI could also have important environmental consequences. Food production accounts for an estimated 20% of global greenhouse gas (GHG) emissions and food consumption is intimately linked to BMI." (I question that 20% figure; 6).

Here is the stated goal of study, "to estimate the impact on GHG emissions of increase in the population distribution of BMI." How can you do a study about fat people and global warming and not examine flatulence? According to angry liberals, cow flatulence is bad for the environment (cow). Since the fat equivalent of "jungle bunny" is "cow," it seems only natural to study this too.

Here's why it was not studied: NO people and their habits (or flatulence) were actaully studied. They assumed the characteristics of a "normal adult (30-59) population of 1 billion people with mean BMI of 24.5 kg/m2 and 3.5% obese, with a corresponding overweight population with mean BMI of 29.0 kg/m2 and 40% obese." While certain human characteristics such as height and weight are normally distributed, other examined behaviors in this study such as driving, walking, and flying are not normally distributed (or they could be, I just made that up). Other methodological problems can be seen here.

"Our normal population BMI distribution reflects the UK situation in the 1970s and our overweight population BMI distribution reflects that predicted for the UK in 2010." Here's something that has never been made clear to me; in research that includes obesity data collected before 1998, what definition is being used? The definition changed after 1998 (7, 8). Current estimates indicate that approximately 30% of Americans are obese (9). However, using the pre-1998 definition, approximately 12% of Americans are considered obese, when in 1980, 14% of Americans were considered obese (10). Has the prevalence of obesity increased or decreased?

The remainder of the paper is an exercise in statistical masturbation. They concluded that "compared with a normal population distribution of BMI, a population with 40% obese requires 19% more food energy for its total energy expenditure. Greenhouse gas emissions from food production and car travel due to increases in adiposity in a population of 1 billion are estimated to be between 0.4 Giga tonnes (GT) and 1.0 GT of carbon dioxide equivalents per year." Here's the Cliff's notes version (i.e., minus the soft bigotry of scientific jargon): fat people eat more, walk less, drive and fly more, and are affected by gravity more than "normal" people.

Although the authors acknowledge that all of their "assumptions" can "be questioned," (because there is no scientific rationale to that they're correct), they still state that "the assertion that increasing population adiposity (i.e., people are getting fatter) will result in an increase in GHG emissions is justifiable and provides further evidence of the link between human health and climate mitigation (i.e., global warming)."

Since these teabaggers are allowed to make wild assumptions in the name of science, I can do the same. Here is a picture of Al Gore when he was Vice President (11). He's thin. At 6'1" and 195lbs, his BMI was 25. Current standards classifies this as overweight. Here is Al Gore now (12). He's fat. I'm going to assume, based on these pictures, that he has gained a good 35 lbs. At 6'1" and 230 pounds, his current BMI is 30. That means that he's obese. Therefore, I can concluded that Al Gore's fat assedness (i.e., increased adiposity) is leading to an increase in GHG emissions, providing further evidence of the link between human health and global warming (i.e., climate mitigation). Perhaps I'll submit this to the International Journal of Epidemiology. If I'm lucky, they'll publish it on Earth Day next year.

Edwards, P., & Roberts, I. (2009). Population adiposity and climate change International Journal of Epidemiology DOI: 10.1093/ije/dyp172

Wednesday, April 8, 2009

They Should Have Seen This Coming

In the April 2009 issue of Mayo Clinic Proceedings (1), a research article titled, "Frequency of New-Onset Pathologic Compulsive Gambling or Hypersexuality After Drug Treatment of Idiopathic Parkinson Disease" was published. More specifically, the authors looked at two dopamine agonist drugs (i.e., pramipexole & ropinirole). There are many other dopamine agonist on the market (e.g., bromocriptine, apomorphine), but since those drugs are old (and cheaper), they are not prescribed as widely as the newer (and more expensive) drugs. Pramipexole (Mirapex) and ropinirole (Requip) are prescribed for Parkinson's disease, but are heavily advertised for the bogus condition, restless leg syndrome* (RLS; 2, 3).

Here's what they found. "Among the study patients with PD, new-onset compulsive gambling or hypersexuality was documented in 7 (18.4%) of 38 patients taking therapeutic doses of dopamine agonists but was not found among untreated patients, those taking subtherapeutic agonist doses, or those taking carbidopa/levodopa alone. Behaviors abated with discontinuation of agonist therapy or dose reduction."

Anyway, don't you think that compulsive gambling and hypersexuality are pretty weird "side effects?" (Binge eating is another common side effect).

As it turns out, not all dopamine agonist are created alike. Typically, they fall into two categories, ergot derivatives and non-ergot derivatives (apomorphine is a morphine derivative). Ergot is a substance produced by the parasitic fungus Claviceps purpurea (common in rye and wheat). LSD is synthesized from ergot (which is good stuff by the way). Pramipexole and ropinirole are non-ergot derivatives, while bromocriptine is an ergot derivative.

The dopaminergic system is comprised of 5 different dopamine receptors (D1-5). D2 agonism is the primary method of treating Parkinson's disease and RLS. Pramipexole and ropinirole have affinities (i.e., chemical attraction) for the D3 receptor, and then the D2 receptor. Bromocriptine primarily has a high affinity for the D2 receptor, and apomorphine has a high affinity for the D1 receptor, followed by the D2 receptor. As the late, great, last psychiatrist discussed (4), if a drug has multiple affinities (e.g., D3 & D2), that specific receptor system will need to be saturated before the other receptors are affected. It other words, with Pramipexole and ropinirole, the majority of the D3 receptors in the brain will become stimulated before the D2 receptors.

Here is where this story gets interesting (that's if you find this shit interesting): The D3 receptor is primarily found in the nucleus accumbens. This is the part of the limbic system that helps to modulate emotional behavior and also plays an important role in the rewarding and reinforcing effects of many abused drugs (and other behaviors such as eating and sex). Moreover, the D3 receptor is the most sensitive to stimulation, requiring less dopamine to generate action potentials.

To the best of my knowledge, it's only the dopamine agonist that stimulate D3 that have the associated "side effects" of compulsive gambling, hypersexuality, and binge eating. I have not seen any reports that bromocriptine and apomorphine have these side effects (5, 6, 7 , 8).

What we have, is a situation where people with a legit disease (i.e., Parkinson's disease) are getting treatment from a drug that, before the therapeutic receptor is stimulated, the behavioral reinforcing receptor gets stimulated and saturated first.

Bromocriptine and apomorhpine are older drugs and have been used on Parkinson's patients for a quite a while. Pramipexole and ropinirole are newer and are the only drugs in the US approved for the specter that is RLS. As is the pattern, when docs start prescribing a drug for one condition, they will prescribe it for other conditions as well (any psychiatric drug fits this model). By pushing these drugs for RLS (a vague and amorphous condition), docs will, by extension, prescribe these drugs for Parkinson's disease too, while the older drugs are prescribed less often.

A months supply ( 60 tablets) of pramipexole (1mg for twice a day) costs approximately $170. A month's supply of ropinirole (3mg twice a day) cost approximately $180 (the drug recently went generic, so it should be cheaper soon). A month's supply of bromocriptine (5mg once a day) costs $130. Apomorphine is mainly administered by injection or pump, so a price cannot be provided.

I know what you're thinking, $40 or $50 is not a huge difference. Except, when you realize that some of the patients who developed compulsive gambling lost close to $60,000. A large price to pay for a drug that's probably no more effective than what's already been on the market for years.

I have not been able to find any comparator trials for bromocriptine, apomorphine, pramipexole, and ropinirole. So, why prescribe the newer drugs when nothing seems to be wrong with the older drugs? The side effects of compulsive gambling and hypersexuality arose after the new drugs had gone to market. Docs were prescribing the drugs without knowing the true risks (as is the case with any new drug). This is a common practice; out with old, in the new. It doesn't seem to matter that no evidence exist to indicate that new drugs are more effective than old drugs.

Here is my main point: I'm an idiot (not my main point); if I can look up the affinities of these two drugs on wikipedia (9, 10), and then deduce how they can cause these "side effects," the people who developed these drugs should have seen this controversy coming (that's my main point). It appears as if marketing, not science, was the prevailing force here, once again.

Here's another crime against the sick: The FDA (The Federal Douche-bags of America) recently approved tetrabenazine-TBZ (Xenazine) for the treatment of chorea associated with Huntington's disease (Was that the drug 13 was on?). This drug has been available in other countries for decades (which means it should be cheap). "A years' supply of TBZ at a dose of 50mg/day will cost about $40,000" (11). Fuck! The pharmaceutical industry is the largest single lobbying sector in Washington DC (12). As a result, not only do Americans pay more money for their medications than other nations; we also cannot import drugs produced in other countries, or re-import drugs that were originally manufactured in this country. The cocksuckers!

* To people diagnosed with RLS, I have two things to say: Firstly, it was not my intention to diminish your suffering, as I acknowledge that it is real suffering; and secondly, fuck you!

To people diagnosed with chronic fatigue syndrome, fibromyalgia, multiple chemical sensitivities, toxic mold and sick building syndrome, and Persian gulf war syndrome; the above applies to you too.

UPDATE: Another study showing the same thing. Waste of time and money (13)

(image credit: Anthony Flores)

1. J. Michael Bostwick, MD,, 2. Kathleen A. Hecksel, MD,, 3. Susanna R. Stevens, MS,, 4. James H. Bower, MD and, & 5. J. Eric Ahlskog, MD, PhD (2009). Frequency of New-Onset Pathologic Compulsive Gambling or Hypersexuality After Drug Treatment of Idiopathic Parkinson Disease Mayo Clinic Proceedings

Click here to nominate this post for Open Lab 2009

Tuesday, April 7, 2009

Topiramate Does Not Treat Alcohol Dependence: Part 2

Remember that, in my first post (1), I pointed out that the primary outcome measure for the study was useless? In this second paper (2), the authors admit how unimportant their primary measure was, "...its extensive secondary physical and psychosocial consequences (alcohol dependence) are what make the burden of the disease so devastating. This realization is enshrined in practice in that no commonly used diagnostic scheme for alcoholism or alcohol dependence (i.e., DSM & ICD) includes a quantification of drinking behavior-only its physical and psychosocial sequelae." It's the disability caused by the behavior that is considered the illness, not the behavior itself.

Then, just like H.M., the authors forgot what they had stated 20 seconds prior, "Topiramate, a sulfamate-substituted fructopyranose derivative, (I still don't know what the hell that means) has been shown to decrease the primary aberration of excessive drinking by reducing the reinforcing effects of alcohol."

Two more points: First, the word "aberration" means straying from the norm, which does not distinguish between something being inherently "bad" from being perceived as "bad." Context is important. Second, It's not proven that topiramate reduces the reinforcing effects of alcohol. The paper they site for that (3), is a theoretical work, not scientific in nature. (Many patients that I have seen say that mixing topiramate and alcohol makes them sick).

Remember that effect size that the authors didn't calculate in the previous study? In a situation that is reminiscent to CL psych's critique of Seroquel's BOLDER study (4), the authors finally calculated one: 0.52 (that's medium). The effect size that I calculated was a Cohen d (0.20, small), which is the standard in the field. The authors didn't state the method used to make their calculation. And why are they providing the effect size in a different paper as opposed to the original paper? Transparency People!

"Topiramate appears to be a promising medication for the treatment of alcohol dependence" This is just plain fucking stupid. Binge drinkers (the actual group studied), represents a minority of people with alcohol dependence (23% in men and 9% in women). And again, drinking behavior is not apart of the definition of alcohol dependence, so you can't say the drug treats alcohol dependence.

"Topiramate's therapeutic effect on drink outcomes appears to be the largest reported for a medication in a multisite alcoholism trial." Can you guess what paper the authors reference for that statement?

"A meta-analysis comparing topiramate, disulfiram, and acomprosate?"

Ted Stevens: "No!"

It's the same study that I wrote about in my previous post, the same study they used to suck more data from to write this second paper. It's a clinical trial, not a comparator trial.

Enough dwelling on the past, in this new paper, the authors discuss how topiramate "treats" certain physical and psychosocial consequences of alcohol dependence.

"Topiramate administration was associated with improvements in physical health in 3 domains of total cholesterol level, hepatic function, and hemodynamic cardiovascular status (that's blood pressure to the non-pretentious).

This appears in the discussion section: "long-term heavy drinking has been associated with elevated lipid levels and the development of fatty liver disease." Then a study is referenced.

This needs to be said: When writing a research paper, all the references cited should be in the introduction/literature review, not in the discussion section (or at least cited in the introduction first). The introduction builds the case for why the study needs to be done and why you're looking for what you are (the introduction to this paper is mostly spent stroking the chemical cock of topiramate). Instead, the authors use the scientific literature to build their case in the discussion section, the last, and sometimes only part people read. This makes the paper more persuasive rhetorically (i.e., not based on substance). But I digress...

Anyway, topiramate beat placebo in lowering total cholesterol level. Cholesterol was lowered by 12 points. That's not close to the nearly 60 point drop you see in LDL-C with high dose statins, but hey, at $210 for one month's supply (topiramate's cost), you pay $17.50 for every point dropped as opposed to $1.30 per point with Lipitor.

I know what you're thinking, "you made that comparison because the drug actually did something you can't refute!"

Actually, I made that comparisons to refute this statement instead, topiramate, by reducing cholesterol "adds considerably to its general medical utility in treating alcohol-dependent individuals." I'd like to point out that high cholesterol is an American problem (and primarily age and diet dependent), not just an alcoholic problem. But in the context created by this study, one might believe that this effect actually has a clinical significance. It doesn't.

"Topiramate administration was associated with a significant decrease in liver enzyme levels, probably due to the reduction in heavy drinking." The liver enzymes were AST, ALT, and GGT. In the first study, they used those enzymes as an objective measure of alcohol consumption; now they are using the same data to say it's an improvement in liver function. Watch how that prestidigitation works:

Tompirate has been shown to reduce the number of drinks consumed (1) and improve liver functioning (2). I just referred to the same data, which was used in two different papers; however, it appears as if I cited two entirely different research experiments.

Here is some more magical thinking, "Taken together with the lowered lipid levels and the consequent reduced risk of fatty liver degeneration (did they test for that disease? NO!), the propensity toward progressive liver disease and eventual cirrhosis could, perhaps, be diminished if alcohol-dependence individuals were treated with topiramate."

The mean difference in AST and ALT from placebo is 4.70 and 6.74 U/L. Both mean levels before and after were within the normal range. I stress "normal range" because AST and ALT lack sensitivity in the diagnosis of chronic liver disease. 33% of the time, AST and ALT are normal, even when the liver is not (Don't these people watch House?). Even the authors attribute the drop in liver enzymes to the reduction of drinking.

"But wasn't the reduction of alcohol consumption do to the drug?"

Recall from my first post that the data are skewed, which makes any attributable effect hard to determine.

"But the data for AST and ALT changes aren't skewed you asshole!"

Here's why: The ranking of organs that carry AST and ALT 1) liver 2) heart 3)skeletal muscles 4) pancreas 5)kidney 6)brain 7)lung 8)white blood cells 9)red blood cells. The enzymes are not specific to liver function or decline.

Also, ALT levels can vary as much as 30% from one day to the next (10% for AST). ALT is lowered by 20% after exercise (something that wasn't controlled for in this study).

Right now, I can make a pretty good argument for why a cheap generic statin and group CBT are better at treating alcohol dependence than topiramate. Also, a combined statin and group therapy are cheaper than a month's supply of topiramate.

Now we're onto blood pressure. Topiramate lowered systolic blood pressure by 9% and diastolic blood pressure by 10%. Not bad. Hydrochlorothiazide could cause a 12% reduction for $26 a month (statin + diuretic + group CBT = still cheaper than topiramate). Moreover, the people in this study didn't have hypertension, although some had "prehypertension." Oh here we go again. First there was prehypertension. Then there was prediabetes. You know what I tell these people? Why don't you pre-suck my genital situation! Sorry, I was channeling George Carlin for a moment.

"Excessive alcohol consumption is generally a risk factory for obesity." They cite one study from 1997. As it turns out, there's not a lot of research on this topic. What I can say is that, of the identified risk factors for obesity, alcohol consumption barely ranks.

So why mention it? Because you can't have a study with topiramate without the obligatory shout-out for it's weight loss proprieties. They knew the answer going in, so why not include it. The mean BMI (which is a pointless measure by the way) was reduced by 1 point, which was statistically significant. Yippee!

Now we get the the so-called psychsocial factors. On the Obsessive Compulsive Drinking Scale (OCDS), at week 14, the mean score for topiramate was 7.67 (SD 6.793). Remember what I said about a large SD? The data here are skewed toward the right. Two other measures: Clinical Global Impression scales for improvement and severity (CGI & CGI-S), also have skewed data (to the right naturally). On the last measure to show a statistically significant difference, Drinker Inventory of Consequences scale, Recent Consequences (DfInC-2R), the SD (20.14) is larger than the mean (17.98). That's more skewed data.

Here's how to tell if data are skewed: 1) round the mean and SD to whole numbers 2) multiply the SD by two 3) subtract or add the result of step 2 from the mean. If the result of step 3 is lower than the lowest possible score (which is usually 0), the distribution is skewed to the right. If the result is higher than the highest possible score, the distribution is highly skewed to the left. Of course, this is only possible is the SD (or SE) are provided. If you pay attention, you'll discover that it's quite common for articles to not list the SD, which makes this little trick difficult to perform.

"In summary, our findings demonstrate that topiramet appears to be a generally effective treatment for alcohol dependence because it improves not only the 'symptom' of drinking but also its physical and psychosocial sequelae. " Morons, idiots, or imbeciles. You chose the descriptive term that best describes the authors of that last sentence.

Bankole A. Johnson, DSc, MD, PhD, MPhil, FRCPsych; Norman Rosenthal, MD; Julie A. Capece, BA; Frank Wiegand, MD; Lian Mao, PhD; Karen Beyers, MS; Amy McKay, PharmD; Nassima Ait-Daoud, MD; Giovanni Addolorato, MD; Raymond F. Anton, MD; Domenic A. Ciraulo (2009). Improvement of Physical Health and Quality of Life of Alcohol-Dependent Individuals With Topiramate Treatment Archives of Internal Medicine