Excess leads to harm on the same metric as benefit from moderation

Moderation is the key to health. For example, sunbathing can help synthesis of vitamin D in the body, but excessive sunbathing increases the risk of skin cancer. Resistance exercise can help increase muscle mass, but excessive exercise causes muscle strain. These excesses have one thing in common: the harm they cause is not on the same metric as the benefit brought about by the moderate use. For example, too much sun exposure will not actually reduce the synthesis of vitamin D. Excessive resistance exercise itself will not lead to muscle atrophy.

So, are there behaviors or food intakes where overdoing creates the opposite result on the same metric as doing it in moderation? Here is an example:
Can Eating Mangoes Reduce Women’s Facial Wrinkles? Pilot Study Shows Promising Results
A 2020 randomized clinical trial at the University of California, Davis, found that eating 85 grams of mango per day can reduce facial wrinkles, but eating 250 grams can increase wrinkles. The experiment described in the research article is simple and was well done. As for why increasing the intake of mango several times will cause increase in wrinkles, the researchers can only guess: It may be related to the large amount of sugar contained in mango. This guess is not very convincing, but there is currently no better explanation. The paper has been cited in other articles over the past two years, but the test was not further studied by any other team.

The above test only examines one indicator: the amount of facial wrinkles. Moderate intake and excessive intake of mangoes produce opposite results on the same indicator. It is not that excessive intake causes some damage unrelated to the indicator, nor is it that large intakes reduce the amount of wrinkles more, or reach a peak (i.e. Mango intake is no longer associated with wrinkle reduction). I don’t know if there is a special term for this type of test results in epidemiology, but it is certain that such test results are not common.

Another benefit about red meat

Red meat, most typically beef, but also including some others such as pork in spite of its white color after cooking, has had a bad reputation in recent years. International Agency for Research on Cancer (IARC) of the World Health Organisation (WHO) classifies red meat as probably carcinogenic to humans, but scientists caution against misreading of the news. Consumers should not rush to conclude that eating even a moderate amount of red meat is bad. In fact, unlike certain things some of us may intentionally or unintentionally consume, e.g. tabacco, or mercury, which have no safe low limit, red meat with a moderate intake is good to us because it is very rich in nutrition, especially protein.

Now, a new research, published in Nature has just found that red meat, as well as dairy products, can boost cancer therapies. More amazingly, the research even found benefit in trans-fat, which we have for quite some time regarded as a heinous villain against our health. On the other hand, reading this research, we must keep in mind that red meat is said to boost certain cancer treatments and the specific type of trans-fat is involved, but this definitely does not suggest we should eat a significant amount of red meat or even a small amount of trans-fat, because (1) this research is still new, and (2) the cancer treatments in this research are specific, at least for now. Nevertheless, this research is something we may keep an eye on because it sheds new light on our conventional wisdom.

East Asians more likely to carry a gene variant associated with long COVID but ...

According to a preprint research article Genome-wide Association Study of Long COVID, "[t]he allele frequency of rs9367106-C at the FOXP4 locus varied greatly among the different study populations, with frequencies ranging from 1.6% in studies with non-Finnish Europeans to higher frequencies such as 7.1% in Finnish, 19% in admixed Americans, and 36% in East Asians". In other words, non-Finnish Europeans have a 1.6% chance to have this gene variant associated with an increased risk of Long COVID and East Asians have a 36% chance. This is clear enough. But some people misread these simple numbers. One journalist's article LONG COVID GENE FOUND says that "[t]he Long Covid FOXP4 mutations are 7x more common in Finnish people than in non-Finnish Europeans, and 36x more common in East Asians". The author must be interpreting the ratios in the research paper as a number (numerator) over some unit denominator. But if we do care about an inter-race comparison, we ought to say that the mutations are 7.1/1.6=4.44 times more common among the Finnish people than non-Finnish Europeans, not 7 times more. Likewise, East Asians are 36/1.6=22.5 times, not 36 times, more likely than the non-Finnish Europeans to have this mutation.

When that journalist's English article was translated into Chinese and posted to Weibo, the Chinese Twitter-like social media network, an even worse headline was added, "东亚人患新冠长期后遗症的概率是欧洲人的36倍" (East Asians are 36 times more likely than Europeans to suffer from Long COVID), naively equating the ratio of carrying a gene variant to the ratio of getting the disease, a mistake I pointed out in my Weibo posting (along with the Chinese translation). If the doctor of Angelina Jolie, the actress who opted for mastectomy to forestall a possible breast cancer, had told her she would definitely get breast cancer because she carries that gene, he could be instantly famous among the cancer researchers world wide, as a laughing stock.

So, what is the chance of getting long COVID for each race? According to a February 14, 2023 article summarizing data from the US census, Asian Americans are the least likely to get long COVID among the five race categories. However, Asian Americans with long COVID have the highest chance to experience severe symptoms. To put simply, Asian Americans don't easily get long COVID, but if they do get it, their symptoms are very serious, relative to any other race. It would be interesting to reconcile this data with the latest research about the gene variant. It's possible that Asian Americans, the relatively well-educated and well-off group of Americans, follow common hygiene practices better than other Americans, hence lower rate of COVID and therefore long COVID. But due to the FOXP4 gene variant which many East Asians carry, long COVID causes the most severe symptoms to East Asians. There are still missing links in this logic, though. For example, Asian Americans are not necessarily East Asian Americans.

Sleeping pills and Alzheimer

Two seemingly conflicting reports about sleeping pills and Alzheimer:

1. Sleeping Pills May Increase Dementia Risk: Study https://www.webmd.com/sleep-disorders/news/20230303/sleeping-pills-may-increase-dementia-risk-study

2. Alzheimer’s Disease Protein Levels Reduced by Sleeping Pill https://www.insideprecisionmedicine.com/topics/patient-care/neurological-disorders/alzheimers-disease-protein-levels-reduced-by-sleeping-pill/

How do we reconcile these two reports? The key here is to make a distinction among these drugs. The good one, in the second article, is suvorexant, which according to a Google search, is sold under the brand name Belsomra. The bad ones, in the first article, are flurazepam (brand name Dalmane), triazolam (Halcion), temazepam (Restoril), trazodone (Desyrel and Oleptro), and zolpidem (Ambien). So if we ever need a sleeping pill, choose Belsomra and avoid the ones in the latter list, in which the first three here are benzodiazepine (benzo for short). There are more research articles confirming the association of benzodiazepine with dementia than otherwise. To be safe, avoid it.

But why bother with sleeping pills? If you do need one, why not take melatonin, which is naturally secreted in your body and has no truly harmful effect?