Vioxx was a big story, underappreciated in my view. Vioxx (rofecoxib) was a new NSAID (non-steroidal anti-inflammatory drug) developed by Merck and approved by the FDA in 1999.

Aspirin interferes with two forms of cyclooygenase, Cox-1 (involved with protecting the stomach lining) and COX-2 (involved with pain and inflammation). So you get relief of pain and inflammation, but also stomach bleeds, in some cases serious. with Vioxx, the goal was something that only inhibited COX-2.

There was a study comparing Vioxx and naproxen – VIGOR, (Vioxx GI Outcomes Research). They found an elevated risk of heart attacks in the VIOXX users – a fourfold increase. This was bad, since heart attacks are A. not rare, especially in the painkiller population and B. very dangerous. Merck defended their new drug through creative bullshit: they argued that naproxen had a protective effect, and that heart-attack-prone Vioxx users in the study were simply suffering from a naproxen deficiency. at the time, there was no evidence supporting that. In order for their arguments to have been valid, naproxen would have had to had a such a strong protective effect against heart attacks that it would have one of the most valuable drugs on the market, far more valuable than Vioxx… As it turns out, naproxen has an anti-clotting effect similar to aspirin, and probably a mild protective effect, like aspirin: but not a strong protective effect.

The Vigor study was submitted in February 2001. The New England Journal of Medicine, a few months later, found that the authors had failed to mention some strokes and heart attacks near the end of the trial. The authors used a cutoff for cardiovascular effects (bad news) that was earlier than the cutoff for gastrointestinal effects (good news). why? Because they were weasels, of course. There was a lot of money riding on this drug’s success.

Other people began to notice the increased heart risks – looking at data from HMOs and such. Merck fought back. There was an MD at Stanford that was concerned about Vioxx: Merck called up the dean of Stanford Medical School at home and warned him about possible loss of financial support: he told them to go fuck themselves.

Vioxx was a bad deal. Although it caused fewer stomach bleeds, it doubled heart attack rates (meta-analysis by Juni et al, Lancet November 5 2004), which increased heart attacks by about 5 x 10-3 per year, about 30-40% of which are expected to be fatal. So, on the one hand it reduced one kind of risk by about 1.0 x 10-4 per year while increasing another kind of risk by about 1.5 x 10-3.
Fifteen times and what do you get?

With those risk numbers, it is hard to imagine any subpopulation that would have a net benefit from Vioxx. Old people are more prone to bleeds, but they’re also more prone to heart attacks.

Now it is true that there were studies that showed greater efficacy: 21 such were reported by Scott S. Reuben, former chief of acute pain at Baystate medical Center in Springfield Mass. But as it turns out, he made them all up. There’s is no evidence that Merck knew about this, but it does perhaps say something about the general climate in big pharma.

Merck knew the gist of this for four years before they pulled the plug on the drug. They had their their drug reps lie about cardio risks, threatened researchers and sued journals that talked about the emerging cardio risks. Few physicians were aware of these risks, even though a close reading of the journals would have suggested it – because hardly any physicians read the journals.

Epidemiologists think that Vioxx caused something like 40,000 deaths. Ron Unz has argued that it likely caused far more than than that – but his case is without any strength. He says that the crude death rate increased from 1998 to 1999 – usually it slightly decreases from year to year – and that it decreased dramatically in 2004, when they pulled Vioxx.

You don’t want to look at national figures: you want to look at a known population of users, so that any effect is not drowned by other effects swirling around in the general population. If you want to know the effects of height on income, you look at the correlation between individual height and income, not the relation between income in tall states and income in states that are on average short. You want to disaggregate. For example, the crude death rate is strongly affected by the age structure: if there was a baby boom many years ago, such that the boomers are now entering prime heart-attack years, that wouldn’t mean that age-adjusted heart disease was getting worse – it would mean that the country was, on average, getting older.

The age-adjusted death rate did go up 0.7% from 1998 to 1999: but deaths due to heart disease actually dropped, as did strokes. COPD, septicemia, and Alzheimers deaths went up. Something may have been happening, but it doesn’t look like a Vioxx effect. Anyhow, Vioxx only showed up in the second half of 1999 in any event.

At the same time, the epidemiologists have strong evidence that Vioxx really did cause something like 40,000 extra deaths over those five years – but that’s not enough to generate a clear signal in the overall mortality rate. For that you need something big, like the Holodomor or the 1918 flu.

Moreover, although the age-adjusted death rate dropped from 832.7 in 2003 to 816.5 in 2004, a bigger drop than from 2002 to 2003 (845.3 to 832.7), it dropped to 798.8 in 2005 (a bigger drop) and 776.5 in 2006 (a still bigger drop).

It’s not unlike the Maguffin in The Fugitive: Harrison Ford finds out that Devlin MacGregor is pushing a dangerous drug, However, the reaction in the movie was very different from the actual reaction to Vioxx. In the movie, people were outraged. In real life, not so much. I don’t think that movie was unrealistic about that – rather, times have changed.

I’ve seen many libertarian types defend Vioxx. Sebastian Mallaby did so in the editorial pages of the Washington Post. But then he’s apparently a marsupial, with low encephalization: he argued that Bush pardoning Merck would somehow solve all our economic troubles.

In the 70s corporations were the bad guys, whether they were or not. Today, billionaires are your friend.

Posted in Uncategorized | 31 Comments


James Miller interviewed me a couple of days ago. If you’re into that sort of thing, you can listen to it here. We talk about possible increases in human intelligence,  cousin marriage, tularemia at Stalingrad, and Iskander.

Posted in Uncategorized | 39 Comments

Throwing down the gauntlet

grip strength

Razib Khan threatens to give in to despair, after hearing some dipshit sociology professor argue for the end of sex segregation in sports – not just because this dame is nuts, but because she’s an example of the general solipsistic trend. Sex is a social construct, race is a social construct, poop is a social construct: wishing can make it so!

Well, despair is a sin, and solipsism isn’t that hard to defeat. People with a fair grounding in reality are much more effective foes.

I am volunteering to do my part. If she agrees, and people put up a reasonable purse, I am willing to arm-wrestle this young lady. And for the loser – vae victis!

Posted in Uncategorized | 125 Comments

Trust Issues

A while ago I was wondering about who you could trust to work in a modern equivalent of the Manhattan project. Thinking about it again, one problem is that people, if for example you consider the typical recent Ivy League graduate to be a human being, are bound and determined to be stupid about this question.

Imagine how we would have dealt with Japanese-Americans in 1942 if we had been informed by modern sensibilities.

Our stated and enforced policy would have been based on the notion that both Issei and Nisei were perfectly trustworthy, no more likely to aid the Empire of Japan than the Dutch in Grand Rapids

So we would have drafted them into the armed forces just like anyone else, and employed them where their skills seemed useful. We would have had them translating Japanese navy intercepts: we were short on Japanese-language translators, so why not? There would have been a bunch of them working with Hypo, down in the basement. Some would have worked in the Manhattan Project. They would have had jobs in the OSS, in the FBI. What could possibly have gone wrong?

Well, some of them were in fact disloyal: not most, but a far higher percentage than in most other ethnic groups in the US. There is nothing magical about this: it often happens. Were the Anglos that moved into Texas loyal to Mexico? Were the Sudeten Germans loyal to Czechoslovakia – was Conrad Henlein just misunderstood? Consider the Niihau incident.

If many Japanese Americans had been privy to the breaking of the Japanese fleet code – plausible, because of the large Japanese population in Hawaii and the need for people with Japanese language skills – the American Magic would have gone away. No Midway, not as we knew it. I figure that we would have lost tens of thousands more KIA in the Pacific than we did in this timeline. Also, probably hundreds of thousands of extra casualties in occupied Asia. Japan would have still have lost, though.

Detailed knowledge of the results from the Manhattan Project wouldn’t have done the Japanese any good, because they didn’t have the industrial muscle and sophistication to make anything of it.

I’m sure there wouldn’t have been any problems with Japanese Secret Service members, any more than Indira Gandhi ever had trouble with her Sikh bodyguards. Besides, in a democracy, no one is indispensable – I’m sure that Henry Wallace would have been a good President, for a gullible, superstitious pinko. While Sam Rayburn was a fine man!

The funniest part would been the many examples of people making excuses for terrorism and treason. When some young Japanese pilot talked about how he should perhaps crash his plane into the White House, his colleagues would have sedulously ignored those ravings, just as our contemporaries did with Major Nidal Hasan. At least they wouldn’t have had to constantly make excuses for his incompetence, as they did with Hasan – Japanese aren’t stupid. After the crash, the new President would have said that no one really knows what motivated the pilot, although back in those days, there really was a way of knowing what evil lay in the hearts of men.

After enough crap, one presumes that the press would have been instructed not to publish the faces of the miscreants, lest the general public get the wrong idea.

Our actual response was suboptimal: people who knew the score (J. Edgar Hoover) thought that putting the Japanese into camps was a mistake. Watching and infiltrating known pro-Nippon groups, punishing those that actually committed crimes was perfectly feasible; combined with reasonable discretion in assigning Japanese to useful but nonclassified jobs, you would have a policy that was more effective than the one we actually pursued.

Locking them up (except in Hawaii !), wasn’t the best course, but it was a million times more sensible than what we would do today. Because in 1942, Americans weren’t crazy:  today, they are.

Posted in Uncategorized | 119 Comments

The Poop Gap

There’s a new article out in Science tracing the splits in gut flora. It looks as if the gut bacteria in chimpanzees split with those in humans 5.3 million years: doesn’t quite match our genetic estimates based on Human/chimp autosomal DNA differences, but it’s in the ball park. They estimate the human-gorilla split at 15.6 million year ago, but that can’t be right: we know that gorillas split off just a bit before the human-chimp split. Perhaps gorilla diet changed drastically, and maybe they picked up new bacteria from some other species.

Different populations of modern humans apparently have pretty different microbiomes. The gut bacteria from people in Malawi appear to have diverged 1.7 million years ago from those in Europeans (people from Connecticut). That is surely too old to be a consequence of modern humans’ trek out of Africa: it looks as if AMH, after leaving Africa, picked up gut flora from archaic sapiens like Neanderthals and Denisovans and dwarves.

Microbiomes are trendy. We know that fecal transplants can cure C difficile (pseudomembranous colitis) lickety-split: they might help with Crohn’s disease and ulcerative colitis. Some researchers think the microbiome has something to do with the initiation of multiple sclerosis. Others suspect that it may play a role regulating how people think and feel – in particular, mood disorders. Autism has been mentioned.

So.. Poop matters: it certainly can affect health, and it may influence brain function. People from sub-Saharan have divergent poop, or you could say that Eurasians do. Are there differences in brain function between sub-Saharan Africans and Eurasians? Sure: Africans do poorly on IQ tests and in academic subjects. They have significant higher rates of schizophrenia, higher murder rates, etc.

Maybe it’s the poop. It’s worth checking out. Perhaps the fault lies not in our stars, or our genes, but in our stool.

Already, eager experimenters – paleos on stilts – are trying to dramatically modulate their internal flora.

“AS THE SUN set over Lake Eyasi in Tanzania, nearly thirty minutes had passed since I had inserted a turkey baster into my bum and injected the feces of a Hadza man – a member of one of the last remaining hunter-gatherers tribes in the world – into the nether regions of my distal colon. I struggled to keep my legs in the air with my toes pointing towards what I thought was the faint outline of the Southern Cross rising in the evening sky. With my hands under my hips – and butt perched against a large rock for support – I peddled an imaginary upside down bicycle in the air to pass the time as I struggled to make sure my new gut ecosystem stayed put inside me.”

The problem, he’s likely running this in the wrong direction, unless he’s just gotten a high-paying job as a professional hunter-gatherer. Generally, we want to better adapt people to key features of industrial civilization, like eating Chicago-style pizza while doing calculus homework, rather than shooting pizened arrows at gazelles.

Judging from Common Core, the standard approach with a potentially gap-closing panacea is to impose it on the entire country without running any careful, small-scale tests. Fortunately, due to Sturgeon’s Law, we don’t ever have to worry about running out of raw material.

Now, I know what you’re thinking. There’s a potential show-stopper

The question is – are blacks in the mood to take any more shit from the Man?

Posted in Uncategorized | 46 Comments

Still Italian

In the early days of the empire, Rome was big, probably around 1 million. There were some number of Jews (thousands at least, perhaps as many as 40,000 by some estimates, although that’s probably high). There were also other foreigners in Rome – probably mostly Greeks, but also some Gauls and Syrians and such. There were also many slaves in Italy, perhaps a third of the population.

Mostly those slaves were from Europe, obtained in wars of expansion: Gaul, Hispania, Germany, Britannia, Greece, etc. Wiki suggests that their European origins is why they didn’t have much affect on Italian genetics because they were European, but that’s wrong. You can certainly detect genetic differences between Gauls and Italian, Germans and Italians, etc. Wiki is correct in saying that you don’t see much sign of this ancient immigration in Italian genetics, but it’s not because they were just like Italians: it has to be because they died out.

The cities were population sinks, and collapsed with the Empire. I doubt if slaves had high birth rates: certainly those working in mines or quarries didn’t. Nor did gladiators. One way or another, the foreigners in Italy, the vast majority of them, disappeared.

Now you see some signs of other stuff in Sicily or Calabria today, but that seems to be later, from Arab or Byzantine times.

Posted in Uncategorized | 46 Comments

Even more on Ashkenazi ancestry

There’s yet another paper out on Ashkenazi ancestry. It’s clear that this problem is a bit tricky, because the ancestral groups are not as different as one would like – this makes distinguishing the origins of chromosomal segments more difficult. Drift doesn’t help. So they check out and calibrate various algorithms with simulated scenarios, which makes sense.

Here they’re looking at finer details. When they analyze the origins of the European component of Ashkenazi ancestry, they conclude that most is southern – probably Italian, but that smaller amounts originated from (probably) Western Europe and (more certainly) Eastern Europe: and in that temporal order. They conclude that the Italian admixture slightly predated a late medieval founder event. Different methods came up with somewhat different estimates for the total amount of European ancestry: the local ancestry inference (LAI) approach came up with 53% European, while the GLOBETROTTER analysis came up with an estimate of 67% European ancestry (after calibration by simulations). In their best guess, they split the difference and go for 60% European.

To sum up, their model is that a population from the Levant mixed with Italians, and shortly thereafter moved to the Rhineland (the founding bottleneck), perhaps mixing to some degree with the local Europeans there, and certainly mixing some with Slavic types when they moved to the Polish-Lithuanian Commonwealth.

How do their conclusions differ from those in the last report? Previously they were thinking that the bottleneck was around 1350, a product of the Black Death and savage persecution – now they’re talking the original settlement in the Rhineland. Previously they had a somewhat lower estimate of European ancestry (~48%, now 60%). I thought these two conclusions likely a couple of years ago.

The big new point, important if correct, is that the admixture with Italians is relatively recent – too recent to have happened back in Roman times. In their model, this main admixture event is 25-55 generations ago, while the founding bottleneck is 25-35 generations ago. It’s not impossible that the admixture happened at the same time as the founding. This I didn’t expect.

How to improve our understanding? aDNA – dig we must!

Posted in Uncategorized | 69 Comments