Running a Cholesterol Confusionist Gauntlet, Part 6 –
Fukui City and Japan
Mr Colpo likes a few epidemiological studies that I haven’t talked about in my other videos.
The first which I will address is the third reference from that blog post that earned him a place in The Primitive Nutrition Series. Let’s look at it.
It deals with residents of Fukui City in Japan. The authors investigated the relationship between serum cholesterol and all-cause mortality. They concluded that anything over 240 mg/dL is safe for cholesterol. 240 is considered high cholesterol by conservative standards. They are saying that 240 and up is just fine, without stipulating any upper limit. Wow! Either the authors believe cholesterol poses no threat no matter how high it gets – which would be a totally ignorant belief – or they forgot to mention there should be an upper limit at some point. Either way, this is a pretty serious omission telling us that this study should be approached with skepticism. For them, only the presence of a known genetic disorder should provide cause for concern.
They actually go so far as to imply that lacking such genetic disorders, cholesterol-lowering medications should not be prescribed, regardless of how high the patient’s numbers are.
Here you can see some of the other work of two of the authors of this paper, including the lead author. They tell us that in their super-sophisticated studies, they were able to determine that two eggs are more filling than two balls of white rice. That’s amazing! Good work, guys. They also tell us that in a controlled experiment, people eating at least two eggs a day lost more weight than people on the control diet. What was the control diet? They were asked to continue their usual diets. Did all the people in the control group eat the same “usual diet”? That seems unlikely. They state their conclusions from their work at the bottom here. Cholesterol is not a risk factor in Japan for all-cause mortality and eggs might be useful for body weight control. This is blue ribbon stuff.
In the study of people in Fukui City they were actually reworking the numbers from an older study, so these authors are using 18-year-old data.
Limitations of this paper stated by the authors are a selection bias of the participants and no exclusion for subjects with a history of serious diseases. This creates a strong possibility of confounding, as participants who were sick with cancer or anything else that might affect cholesterol levels might have affected the distribution of the death patterns. This is once again the phenomenon of reverse causation about which I have talked so much by now. Realize the subjects here were ages 40 to 79. Notice also their qualification that they are limiting the relevance of their conclusions to Japan. If you don’t live in Japan, they are not suggesting this study applies to you. I guess they are also suggesting that the Japanese are uniquely immune to the effects of high cholesterol for some reason. They do not attempt to explain why they think this might be so.
There really is a serious chance of confounding when you look at their numbers. This study lacked statistical power. I have highlighted the number of participants and the distribution of deaths from cardiovascular diseases. I have already done enough work to show you that low cholesterol by itself does not cause cancer or infectious diseases or shorten lifespan. All-cause mortality is just a distraction. Colpo is disputing the lipid hypothesis so lets look just at cardiovascular disease. These are the numbers of deaths over a four-year timespan ending in 1990. That isn’t a very long time. It may seem significant that women had the highest age-adjusted relative risk of cardiovascular death with cholesterol scores between 140 and 159, but if you look at the actual numbers, we’re talking about only 12 women out of more than 15,500. It’s hard to conclude much from that just because with so few deaths that ratio more likely is influenced by confounders or it’s just statistical noise.
Moreover, look at the men in particular. Do you see how the relative risk jumps around in the three categories of cholesterol over 220? It goes from 1.9 down to 1.0 and back up to 3.1 in the highest cholesterol group. That’s not much of a trend, is it? Well, you should not expect to see much of a trend based on 9, 3, and 4 deaths in each category. That’s just not enough data. It seems that the Japanese in this study were like the Japanese in other comparisons. They had very low rates of heart disease. There just weren’t a lot of deaths here so the relationship of cholesterol to cardiovascular deaths is not clear within this data. Remember, the researchers did not exclude people with cancer or other serious illnesses. There were probably other important factors at play here. Maybe we could learn more from a population that experienced a lot of heart disease. Also realize the category named here is not coronary heart disease, it is cardiovascular disease.
In Asian countries, stroke is more common than coronary heart disease. The paper did not categorize mortality this way. Instead, it resolutely focused just on all-cause mortality, so it really isn’t even attempting to address coronary heart disease. Now, high cholesterol is a key risk factor for stroke, so don’t get me wrong about that. But the problem is, Asians have other important risk factors affecting their stroke rates. They eat a lot of salt and have more people with high blood pressure. They smoke a lot. That’s really important for stroke.
Here is further support for these assertions. In Japanese men, dietary salt increases the risk of death from stroke.
And the Japanese are some of the biggest salt consumers in the world.
They also smoke a lot, as I said. The study covered a period ending in 1990. Look how high smoking rates were there back in 1993.
Lastly, no mention is made of cholesterol lowering drugs. The Japanese were quick adopters of statins. On a side note, the research breakthrough that lead to the development of statins took place in Japan. To be clear, statin use probably did not affect that many people in that time period, but it is just one more factor that might be important when we are talking about such low rates of death from cardiovascular disease. As you can tell, I don’t find much in this study to threaten the lipid hypothesis. Indeed, it doesn’t even try to address that directly. It only looked at all-cause mortality. Colpo has picked another weak study for his weak case.
Do you remember how I said Colpo picked three studies that don’t support his title here? The first was highly technically flawed. The second did not look at saturated fat consumption in general, only dairy fat. And the third lacked adequate statistical power. All three had telltale signs of bias from their authors. Yet Colpo thought they made a strong case against the diet-heart idea.
More epidemiology is ahead in my next video.