How Much Vitamin D Should I Take?

A new study of a quarter million Copenhagen residents found that those with Vitamin D blood levels of 40-70 nmol/L [16-28 ng/ml] had the lowest death rate. People with lower and higher amounts had higher death rates, in other words. The death rate versus blood level function has a reverse-J shape, i.e., too little is worse than too much. About 1% of the sample had levels above 140 nmol/L [56 ng/ml], for practical purposes a “high” level.

Because Vitamin D3 seems to have a big time-of-day-dependent effect on sleep (Vitamin D in the morning improves sleep, Vitamin D in the evening makes sleep worse) it is plausible that people with high Vitamin D levels were more likely to take it in the evening than those with moderate levels and this is why they had higher mortality. Likewise, it is plausible that those with moderate levels were more likely to take Vitamin D in the morning than those with low levels and better sleep explains the lower mortality. Although epidemiologists adjust for smoking in studies like this, they don’t yet adjust for sleep quality. It is also plausible that people who were more sick took more Vitamin D — hoping it would improve their health.

I think I have a better way to decide how much Vitamin D3 to take: choose the minimum amount that produces the best sleep. Sleep is so strongly connected with health that I wouldn’t want to choose worse sleep over better sleep simply because of epidemiology. At the same time that I greatly improved my sleep, I stopped getting easy-to-notice colds. Apparently my immune system was doing a better job of fighting them off.

There is evidence that Vitamin D improves immune function independently of its effect on sleep. A 2009 survey found that “those with less than 10 nanograms of vitamin D per milliliter [25 nmol/L] of blood, considered low, were nearly 40 percent more likely to have had a respiratory infection [over what period of time?] than those with vitamin D levels of 30 ng/ml [75 nmol/L] or higher.”

So those three studies (epidemiology, lab, epidemiology) taken together make a good case that my Vitamin D levels should be at least 25 ng/ml. I will have my Vitamin D level measured soon and it will be interesting to see how much an approach based purely on self-measurement (find the minimum amount of Vitamin D that optimizes sleep) agrees with this.

Thanks to Chase Saunders.

More. In an earlier version of this post I confused ng/ml with nmol/L.

21 Responses to “How Much Vitamin D Should I Take?”

  1. Joe Says:

    Looking forward to your vitamin D3 test results!

    Mine is 90 ng/ml, and I generally sleep like a baby.

  2. Caleb Coop Says:

    That’s a good point Seth; based on how people take pills it’s likely many with the high levels were taking a 5000 w/ breakfast and a 5000 w/ dinner. So it’s still an open and very interesting question what the long term effects high levels obtained differently might be. Since the body tops out around 10,000 when producing from the skin, I tend to think supplementing under that would be safe.

    It’s finally warm enough to make Vitamin D naturally where I live, and I believe there are additional hormeosis benefits from moderate solar radiation damage below the sunburn level. So I’ve stopped taking my D in the morning because of worry about unnaturally high levels.

    Seth, would you still take your 6000-7000ish dose in the morning if you were going to be getting enough sun exposure later in the day to top out skin production?

    My default is to be conservative in exceeding what would have happened in nature. Though I do believe some bio-hacks could surpass the benefits of simple ancestral reenactment, and really respect those like you who try them out on themselves. We’re all richer for you sharing your results.

    Seth: I take 5000 IU, not 6000-7000 IU, in the morning. Yeah, I’d still take it even if I were going to the beach. Which is purely hypothetical!

  3. P. Winter Says:


  4. Guv Says:

    Hi Seth, looks like you have a couple of typos…..
    The two nmol references should read nmol/liter (not nmol/milliliter).

    For those people not familiar with the nmol/L measurement, divide by 2.498 (or 2.5) for ng/ml.
    eg. 40-70 nmol/L = 16-25 ng/ml
    This range is a lot lower than all the various recommendations i have seen around the web; which generally quote the optimal range anywhere between 40 ng/ml (100 nmol/L) on the bottom end to 80 ng/ml (200 nmol/L) on the top end, depending on where you look.

    I have seen higher ranges recommended to treat illnesses, ie. the Mercola web site suggests 70-100 ng/ml (175-250 nmol/L) to treat Cancer & Heart Disease.

    Seth: I have fixed the mistakes, thanks. If Vitamin D has a strong of time-of-day effect — i.e., it matters a lot what time of day you take it — then I believe all recommendations will need to be reconsidered in the light of new evidence that measures or controls time of day.

  5. Guv Says:

    thx Seth, i too will be interested to hear what your Vit D level is measured at.
    (&amp how you interpret the result).

    I actually had mine tested last week, tho i have not been back to get the results yet.

  6. Chuck Says:

    I just had mine checked – 84 nmol/L. I take 5000 IU each morning (I recently switched from capsules to liquid on Robb Wolf’s advise – capsules maybe less effective) and get plenty of sun (7 – 15 hrs/wk). So I”m not sure if it’s the sun or supplements, or both.

    I began supplementing 3 years ago and haven’t had one cold in that time – usually got one cold / flu per year.

    Seth: That’s very interesting. When you started the supplements 3 years ago, did you also increase how much sunlight you got? Or did amount of sunlight stay more or less constant?

  7. P.Bear Says:

    Please read:
    Vitamin D and host resistance to infection? Putting the cart in front of the horse
    I recommend that only a very few of my patients take vitamin D, only those whose levels are below 14ng/ml. Higher levels of 25-D above (about 40ng/ml) are associated with increased all cause mortality.

    Seth: Thanks for the link. That paper references an experiment measuring whether 2000 IU/day Vitamin D reduced colds. It found no effect. This may be because the experiment used too low a dose and did not control the time of day that the subjects took Vitamin D — that is, Vitamin D in the morning may help but not at other times of day.

  8. Sara Says:

    Seth, I wonder if you are aware of the link between early D3 deficiency and Type I diabetes? The research is pointing toward D3 having a modulating effect on the immune system (Type I diabetes is an autoimmune disorder). There is probably a factor of genetic vulnerability due to a polymorphism of the VDR gene which combined with an early (even in utero) lack of D3 is linked fairly conclusively with development of Type I diabetes. I found it veeeeery interesting that high intake of D3 seems protective against a hyperimmune response. I can link you up to some papers if you are interested. Obviously each paper has its issues (for example, you can’t really induce a deficiency or leave one untreated in order to observe the effect) but the body of evidence taken together is quite convincing.

    Seth: I was not aware of the link, thanks for pointing it out. Please post another comment with links to the papers, if that’s possible.

  9. David Says:

    What did the study that you cited mean by “lower” death rate? Lower by some material amount? After watching this TedX video, I have a renewed appreciation for asking about the significance of most scientific studies.

    Seth: The highest and lowest death rates differed by a factor of 2.

  10. Pete Says:

    As vitamin d as taken is inert the time of day is unlikely to matter. It is then converted to 25(OH)D which is inert. It is only converted to an active compound when needed, so when it is taken is very unlikely to matter. You body is used to getting it at midday but I doubt it cares. Please do not take sustained daily doses above 20,000IU a day without being told to do so by a doctor it is dangerous. If you can take those high doses without getting ill it is because your parathyroid has failed.

    Seth: Since the time of day obviously matters, at least some of the time — see the 20-odd examples I have collected — something is wrong with your theory.

  11. Pete Says:

    20 examples is not many so it could easily be a fluke. However, it is more likely one of the other ingredients. The oil could give you reflux and that could be affecting you sleep. Do you get the same effect from sun light? Do you get the same effect if the vitamin d is applied to the skin rather than taken orally?

    Seth: “Easily be a fluke”? Do you know of a case where someone collected twenty examples of something that turned out to be a fluke? I found that Vitamin D in the morning improved my sleep. I found the same effect with gelcaps (with oil) and tablets (without oil). So it’s not the oil. I have found that sunlight in the morning improves my sleep. I have not tried applying Vitamin D to my skin.

  12. Dave Says:

    In the study claiming higher blood levels were correlated with a higher death rate, did they try to correct for the possible effect of people with poor health being more likely to supplement with vitamins, or that sort of thing?

    Seth: No they didn’t. That’s what I meant by “It is also plausible that people who were more sick took more Vitamin D — hoping it would improve their health.”

  13. P.Bear Says:

    This video may interest some of you artificial seco-steroid “Vitamin D” takers:

    One might also read this interesting blog:

    This article in Townsend Letter is of interest, but I would not follow any medical advice from non health care professional MP cult members.
    for health, P.Bear

  14. Carol Says:

    For years I’ve had bad sleep, and once I was in peri-menopause, it got worse. I had trouble going to sleep, and if I had a hotflash during the night, I would wake up, and be unable to go back to sleep again. I was the walking dead some days. I worked in a windowless office 5 days a week, for 9+ hours a day for the last 13 years. I also live in the PNW, so sunshine is a fleeting thing most of the time. I had been taking Vitamin D (low dosage) in the evening, but then I switched to the morning, and was taking 10,000 a day for about 2 months. The sleeping got progressively better, and I’m now to the point that I fall asleep easily, and even a hot flash only wakens me a bit, and I can go back to sleep. This is HUGE for me. I had my levels tested, and it is 63.4 ng/ml, which is in range (30-100). I’ve switched to 5,000 iu a day now, since I get almost no sun exposure opportunites. I also took both dry and oil based capsules. I didn’t notice a difference.

  15. Antonio Pedro Says:

    Hi all,

    Have you guys heard about this?

    It is not quite the same as what has been discussed here but it is surely related.


    Antonio Pedro

  16. Chuck Says:

    A qualified yes to the increase of sunlight. I’ve always gotten some sun exposure during the summer – more some years than others – but it wasn’t until last year (~16 mos ago) that I really put an effort into getting daily sun. Living in So CA you have opportunities for sun exposure – you can have a warm sunny day any month of the year – but you still have to have ability to go outside and take your shirt off. Last year my ability became open ended.

    I started with 15 minute sessions – front and back, shirt off, pants on, no sunscreen (this is important) and I do wear a baseball cap most days (it makes it easier to read) – when the sun was at it’s highest point. You’d be surprised at what 15 minutes of sun on white skin will do. I would do this on every sunny day, extending the time a little each week or 5 straight days of sun exposure. I can now spend two hours in the sun with no sign of burning. If I start feeling any burning sensation, I cover up or get in the shade.

    Most days I’ll just sit out facing the sun with my shirt off for an hour or so. Sometimes I’ll split the time between front and back. As it gets hotter, I actually spend less time in the sun, unless it’s before noon or after 3 PM, or I’m on the beach close to the water. I’ve found, if you listen to what your body is telling you, you won’t get burnt or sun stroke. Like most other things in life, the poison is in the dose.

  17. Guv Says:

    it looks a new study on Vitamin D levels has been released since the Copenhagen study (discussed in the article above).

    This is some blurb from the Vitamin D Council.
    “While a study out of Denmark reports, in an observational study, that both high and low vitamin D levels are associated with an increase in mortality, a study published just last week out of Israel reports that only low vitamin D levels are associated with increased risk of mortality.
    Published in the Journal of Endocrinology and Metabolism, lead investigator Dr Walid Saliba of the Carmel Medical Center found that in an analysis of 182,152 subjects, vitamin D levels above 20 ng/ml were better than levels below 20 ng/ml.”

    That is all the info i have on it, to find out more a login is required, which i do not have.

  18. Leszek Says:

    Long story, but interesting. Once I get better, I’ll be brief :)

    Yesterday, I took D3 in the morning (just 800 IU) and for the first time in several months, if not years, I slept soundly all night. Too early to proclaim a victory but I have no explanation for this other than early morning D3 dosing. Need some advice. Need to tell my story.

    I experimented with D3 before, with mixed results. I have Hashimoto Thyroiditis and have had cognitive, mood, and sleep problems for about 15 years. Thyroid medication, which I have been taking last two years (an equivalent of about 120 mcg levothyroxine, which is a fairly typical dose) has not worked well, until about a year ago, when I briefly experimented with D3 supplementation. I am not sure how much I took then, I think about 8000 IU, twice a day, 4000 in the morning and 4000 with dinner. Ten days into D3 supplementation, I had my levels measured and I was low (25 ng/mL). Two weeks into supplementation, I started feeling normal, confident, clear-headed, energetic, just like what I had prayed for for years. I am not sure how well I slept at that time. This improvement did not last. One month later I felt thirsty all the time, got up at night every half an hour, my normally very slow heart rate increased uncomfortably and I was in a worse shape than ever. It had felt like an excess thyroid supplementation (except for thirst which seemed odd to my doctors), but it was not.

    I discontinued D3 and went back to my pre-D3 normal. I resumed supplementation, and that again caused restless nights, thirst, frequent urination, etc. I discontinued D3 for about six months. I thought for good.

    Last six months I nearly lost all hope: sluggish, sleepy, confused, getting lost in familiar places, avoiding people, stiff and achy all over, particularly in calves and achilles tendons, in a word: your standard low-thyroid mess.

    Last month, I spent some time in Utah and felt a touch better. I don’t have much sun here (Seattle) so I thought perhaps it is a matter of D3 after all. This time (a week ago) I took 2000 IU (half in the morning, half at night) and excessive thirst and fast heart rate, particularly at night, had returned immediately, within just 24 hours. After just 2000 IU of D3! This is really not that much for someone with just D3 at 35 ng/ML (as of two months ago).

    So I bought 400 IU tablets and yesterday decided to take two (800 IU), in the morning (I am not sure why in the morning, pure accident). The effect on sleep was so good and so dramatic that I started looking for empirical evidence that perhaps all I need to get over this nightmare is to take D3 in the morning.

    I think 800 might not be enough to improve my other problems. I want to take 2000 early in the morning for a week and see what happens. What do others think? Keep your fingers crossed for me.

  19. Leszek Says:

    Let me just add that thyroid problems are highly correlated with D3 and, in women, with iron deficiency.

    Furthermore, I was briefly on a too high dose of levothyroxine (my TSH was 0.08) and it felt very much like what I was experiencing during night-time D3 dosing, except for excessive thirst.

    I have no proof but it seems to me that D3 is making thyroid hormones more effective, perhaps via increased t4 to t3 conversion.

  20. Leszek Says:

    Relative vitamin D insufficiency in Hashimoto’s thyroiditis.

  21. Sara Lake Says:

    Here you go, Seth. Some light reading.. Each one of these takes a slightly different angle.
    Borkar, V., Devidayal, V. & Bhalla, A (2010). Low levels of vitamin D in North Indian children with newly diagnosed type 1 diabetes. Pediatric diabetes, 11, 345-350.
    Giulietti, A., Gysemans, C., Stoffels, K., van Etten, E., Decallonne, B., Overbergh, L., Bouillon, R. & Mathieu, C. (2003). Vitamin D deficiency in early life accelerates type 1 diabetes in non-obese diabetic mice. Diabetologia, 47, 451-462
    Hyppönen, E., Läärä, E., Reunanen, A., Järvelin, M. & Virtanen, S. (2001). Intake of vitamin D and risk of type 1 diabetes: a birth-cohort. The Lancet, 358, 1500-1503
    Stene, L., Joner, G. & the Norwegian Childhood Diabetes Study Group. (2003). Use of cod liver oil during the first year of life is associated with lower risk of childhood-onset type 1 diabetes: A large, population-based, case-control study, The American Journal of Clinical Nutrition, 78, 1128-1134
    The EURODIAB Substudy 2 Study Group (1998). Vitamin D supplement in early childhood and risk for type I (insulin dependant) diabetes. Diabetologia, 42, 51-54.
    Zhang, J., Li, W., Liu, J., Wu, W., Ouyang, H., Zhang, Q., Wang, Y., Liu, L., Yang R., Liu, X., Meng, Q. & Lu, J. (2011). Polymorphisms in the vitamin D receptor gene and type 1 diabetes mellitus risk: An update by meta-analysis. Molecular and Cellular Endocrinology, 355, 135-142.
    Zipitis, C. & Akobeng, A. (2008). Vitamin D supplementation in early childhood and risk of type I diabetes: a systematic review and meta-analysis. Archives of Disease in Childhood, 25 March 2008, 1-15