Epistemological status: I studied bioinformatics but I'm no domain expert and lay out my current view on the issue based on the facts I found in the literature. This is not health advice.
There’s strong evolutionary selection for producing enough Vitamin D. Vitamin D works together with Vitamin K2, and as a result high Vitamin D supplementation should come with Vitamin K2 supplementation.
Personally, I decided to consume every morning 10 000 IU of Vitamin D3 and 200 µg of K2 (all-trans MK7) and believe that it’s likely beneficial for many fellow rationalists who don’t spend a lot of time exposing a lot of skin to the sun to take the same supplements.
Before reading further, I want you to look at the graphic below and assess to what extend you believe the two maps are the same or differ:
It's interesting how similar those two maps happen to be. The map is from the chapter Human Skin Pigmentation as an Adaptation to UV Radiation written by Nina G. Jablonski and George Chaplin. A is a map for UVB intensity over the world. On the other hand, B is about the skin color of the residents of the area by their sensitivity for the wavelength of 305mn.
There's a lot of human migration. Given that human likely only arrived in Australia in 65000 ago, evolution didn't have much time to evolve skin color to be able to allow different Australian natives to have a skin color that matches the the UVB radiation. This illustrates that there's huge selection pressure on skin color and this is likely selection pressure that's driven by the need to produce Vitamin D3 via UVB radiation.
Take a look at the two maps again and ponder how well they are aligned.
The importance of Vitamin D
Given that Vitamin D3 seems important enough to get the skin color changed to match UVB radiation it's reasonable to assume reducing the amount of UVB radiation humans get produces Vitamin D3 deficiency that has negative health consequences. Gwern's meta-analysis on Vitamin D suggests that the existing academic publications suggest that this costs 0.33 years of life. Later in this article I will make the case why I believe that the literature is likely biased to underrate the effect for systematic reasons.
Which Vitamin D should be supplemented?
Humans produce Vitamin D3. While Vitamin D2 that gets produced by mushrooms can also be used by humans, it makes sense to supplement Vitamin D3 as it’s the form of Vitamin D around which evolution optimized us given that it’s what’s available in the natural environment. Besides the evolutionary argument The case against ergocalciferol (vitamin D2) as a vitamin supplement lists a few other reasons as well.
How much Vitamin D3 should be taken?
For clinical care, it appears that all current methodologies are adequate if one targets a 25(OH)D value higher than current cut points; for example, a value of 40 ng/ml (100 nmol/L) is without toxicity and virtually ensures that the individual's “true” value is greater than 30 ng/ml (75 nmol/L). A clinical approach of targeting a higher 25(OH)D value seems prudent in that improving vitamin D status should reduce multiple adverse consequences of vitamin D deficiency at extremely low cost with minimal toxicity risk.
In a German study they found median 25(OH)D levels of 19.8 ng/ml. While this isn't a meta-review it illustrates that plenty of people are strongly under the recommendation of the Endocrine Society. I'm German and I used to take 5000 IU vitamin D3 per day and got tested by my doctor and got a value of 33ng/ml. That's a bit over the recommended minimum of 30 ng/ml but not enough to get over the inherent error of the essay. As a result I upped my daily D3 intake to 10000 IU which is maintenance tolerable upper limits recommended by the Endocrine Society Guideline.
How high is 10000 IU? According the Endocrine Society Guideline, when an adult wearing a bathing suit is exposed to one minimal erythemal dose of UV radiation (a slight pinkness to the skin 24 h after exposure), the amount of vitamin D produced is equivalent to ingesting between 10,000 and 25,000 IU. At the same time 10000 IU is much higher than the recommended daily intake of 600 IU by IOM (US), 600 IU by EFSA (EU) and 800 IU (20 µg Vitamin D) by the DGE in Germany.
Dimitrios T. Papadimitriou argues in The Big Vitamin D Mistake that 10 000 IU should be consumed to reach the 40 ng/ml 25(OH)D blood level and that due to statistical errors it’s assumed in the official guidelines that less Vitamin D3 supplementation is required to reach that level.
Above I spoke about my belief, that the existing studies underrate the benefits of Vitamin D3 supplementation. I believe that for two reasons. The first is about the timing of Vitamin D3 supplementation and the second is about K2.
The effect of timing of Vitamin D supplementation
The studies we have generally make the assumption that timing doesn't matter and blood 25(OH)D levels are the only interesting variable. Given that we don’t have an routine clinical essay to measure vitamin D3 or vitamin D2 serum concentration we can’t focus on their serum levels.
Another name for 25(OH)D level is Calcifediol (or 25-hydroxyvitamin D / calcidiol) while the substance we supplement or that our skin produces in response to UVB light exposure is cholecalciferol (or Vitamin D3). Calcifediol gets produced in our liver from cholecalciferol. Additionally our kidney turns calcifediol into calcitriol (1,25-dihydroxyvitamin D). Both calcifediol and calcitriol are used in many different pathways. Calcifediol has a biological half-life of 2–3 weeks while calcitriol has a half-life of 4–15 h.
The mainstream belief that timing is irrelevant is supported by the fact that calcitriol levels doesn’t get directly affected affected by Calcifediol levels. At the same time Seth Roberts gathered examples of multiple people whose sleep improved when they took Vitamin D3 in the morning and whose sleep got worse when they took it in the evening. Multiple folks in Quantified Self replicated that effect for themselves but unfortunately there aren’t studies that investigated it deeper.
Given that there’s no harm in taking it in the morning I personally take my Vitamin D3 in the morning even when there’s no high quality evidence for it.
The role of K2
The second important variable is K2. Atli Arnarson makes the case in Is Vitamin D Harmful Without Vitamin K? that Vitamin D toxicity at high doses is usually about K2 deficiency because both are needed in the same pathways and more K2 is needed when there's more Vitamin D. Vitamin D toxicity leads to hypercalcemia where there's too much Calcium in the blood. Calcifediol moves some calcium from the bone to the blood and K2 is needed to put the calcium in the bones. Hypercalcemia is bad because it lead to blood vessel calcification. Observational studies link low Vitamin K2 levels to blood vessel calcification with K2 being more important than K1.
Why might we have a K2 deficiency compared to the ancestral environment? K2 is found in animal liver and fermented foods which means food in which bacteria grew. In the ancestral environment it was hard to prevent bacteria from growing in the food that was consumed and thus the related consumption was higher. Seth Roberts makes here the point that we value the taste of umami that primarily comes from eating microbe-rich foot in the ancestral environment. Today, most westerners also don't eat animal liver while ancestral humans consumed it.
There are multiple forms of K2 (menaquinone) that can theoretically be used to supplement. The most commonly discussed are MK-4 and MK-7. According to Sato et al that MK-4 has from supplements has no direct bioavailability coming to the conclusion “MK-7 is a better supplier for MK-4 in vivo than MK-4 itself.” Schurgers et all writes however that he has unpublished data that suggest MK-4 bioavailability. It would be good to have more research to get more clear about MK-4 bioavailability. There’s also MK-5, MK-8 and MK-9 however it’s not widely used as a supplement and there’s more research needed.
Given the current research it seems sensible to me to go with pure MK-7 supplements.
MK-7 exists in a trans- and a cis-form where only the trans-form is used by humans. Given that some supplements contain a mix of both forms it’s desirable to buy a MK-7 supplement that specifies that it’s all-trans (or 99% all-trans).
On that basis I have decided for myself to consume for now 10000 IU Vitamin D3 per day and 200 µg Vitamin K2 (all-trans MK-7)*. I take it every morning. I don’t have strong reasons for 200 µg but it’s the default size of supplements and there’s no known toxicity of it. While going out into the sun would also be a way to acquire Vitamin D3, it causes wrinkles in the face that while I don’t try to minimize sun exposure I won’t maximize it when I can get my Vitamin D3 through a cheap supplement.
* I brought me current K2 supplement before doing this research and it doesn’t specify trans vs. cis but I will buy all-trans MK7 the next time.