From November 1985 to May 1987, on cloudless days as close to the 15th of each month as the weather allowed, Ann Webb set small quartz vessels on the roof of a south-facing balcony in Boston and left them in the sun from 11:30 in the morning to 2:30 in the afternoon. She worked in Michael Holick's vitamin D laboratory at Boston University. Each vessel held a trace of 7-dehydrocholesterol, the molecule in skin that sunlight begins to turn into vitamin D, dissolved in methanol. When any were available, pieces of human skin went out too: foreskins from newborn circumcisions, each cut in two, one half set in the sun and the other kept in the dark. A co-author at the University of Alberta set the same vessels out in Edmonton, ten degrees of latitude farther north. 1
In Boston, from November through February, neither the solution nor the skin made any previtamin D3, the first product of the reaction. In Edmonton the empty months ran from October through March. No skin was available in December, January, or September, so those months rest on the solution alone. The paper's title gave the result, that winter sunlight in Boston and Edmonton "will not promote vitamin D3 synthesis in human skin", and the abstract called the empty months the "vitamin D winter". 1
Skin makes vitamin D when ultraviolet B (UVB), the short-wave part of sunlight that also causes sunburn, reaches 7-dehydrocholesterol in its outer layer. How much of it ends up in the blood depends on the season and the hour, how much skin is bare, skin color, body fat, and differences between people that no study has fully explained. When 22 light-skinned adults in Copenhagen were given identical lamp doses for nine winter weeks, the rise in their blood vitamin D ran from 3 to 139 nanomoles per liter. 2
Blood tests measure 25-hydroxyvitamin D, or 25(OH)D, the form the liver makes from vitamin D, in nanomoles per liter (nmol/L) or nanograms per milliliter (ng/mL). Fifty nmol/L is 20 ng/mL, the level that a 2011 report from the US Institute of Medicine judged enough for at least 97.5 percent of people. Of 1,406 adults tested in six regions of Iran, between 29 and 37.5 degrees north, 90.7 percent were below it. 3; 4
In VITAL, the Vitamin D and Omega-3 Trial, 25,871 American men over 50 and women over 55 took 2,000 international units (IU) of vitamin D a day or an identical dummy pill for a median of 5.3 years. The pills raised blood levels. Of about 12,900 people in each group, cancer was diagnosed in 793 taking vitamin D and 824 taking the dummy pill, and a heart attack, stroke, or cardiovascular death struck 396 and 409. The largest randomized trial of real sunlight we found, a trial in which chance decided who got the extra sun, enrolled 602 residents of care homes in Sydney. 5; 6
This page covers how sunlight came to cure rickets and how vitamin D became a product, what skin can and cannot make, how season, hour, clothing, skin color, and age change the dose, why people in sunny countries run low, who decided what counts as enough, and what happened when sun and pills were tested against each other. Each study is named as real sun, a lamp, pieces of skin or a solution under light, or a measurement in blood.
Key facts
- In Boston, human skin and a 7-dehydrocholesterol solution set out in cloudless winter sun made no previtamin D3 from November through February; in Edmonton the gap ran from October through March. 1
- Under a lamp built to imitate equatorial noon, skin samples of every type stopped making previtamin D3 at about 15 percent of their 7-dehydrocholesterol, and darker skin took longer to get there. 7
- Twenty-two light-skinned adults given identical lamp doses for nine weeks raised their blood vitamin D by anything from 3 to 139 nmol/L. 2
- Of 50 adults in Karachi whose blood was tested, 49 were deficient or insufficient, and the group that spent the most time in the sun averaged 17 ng/mL. 8
- In a winter trial with 250 British Army recruits, a lamp cabinet and a daily pill each brought 95 percent or more of the men who received them to 50 nmol/L within five weeks. 9
Sunlight, lamps, and cod liver oil each cured rickets by 1922
Rickets, a disease of growing children in which the bones stay soft, was described by Daniel Whistler in his 1645 medical thesis at Leiden as the disease "of English Children". In 1822 Jędrzej Śniadecki, a Polish physician, saw that it was commoner among infants in the sunless tenements of Warsaw than in the countryside, and blamed the missing sun. In 1890 Theobald Palm set rickets against latitude and recommended the "systematic use of sun-baths". By the early twentieth century it was epidemic in the smoky industrial cities of western Europe and the northern United States. 10
In 1919 Kurt Huldschinsky cured rickets in infants with the light of a mercury vapor lamp, and in 1921 Alfred Hess and Lester Unger in New York reported healing it with sunlight. 10; 11
In the summer of 1922 Benjamin Kramer and Frank Boone, pediatricians at Johns Hopkins in Baltimore, set out to learn whether "the pigmented skin" of Black children "interfered with the therapeutic action of the sun's rays". They took seven Black children with active rickets, aged 6 to 16 months, and fed them milk dilutions and orange juice, a diet they had found would not cure the disease on its own. The children went into direct sun for fifteen minutes a day, then half an hour, then an hour. X-rays of their arms and legs showed healing in every child during the third week, and the calcium and phosphorus in their blood reached normal within three weeks. One child had first spent three weeks on the ward on the same diet with no change on the X-ray, and healed within 14 days of going into the sun. 11
The same year Elmer McCollum heated cod liver oil until it lost its power against the eye disease of vitamin A deficiency, and found it still cured rickets. He called the surviving factor vitamin D, because it was the fourth vitamin to be found. 10
A lamp shone on food made vitamin D a patented product by 1928
In 1924 Harry Steenbock at the University of Wisconsin found that shining an ultraviolet lamp on food made it cure rickets in rats, and Alfred Hess found that the same light made cottonseed oil, linseed oil, wheat germ, and lettuce work against rickets too. Steenbock and E. M. Nelson also saw rats on a rickets diet begin to grow when irradiated rats were put into their cage, and blamed the "photochemically activated" droppings of the irradiated rats. 10
Steenbock filed for a patent on 30 June 1924. It was granted on 14 August 1928 and assigned to the Wisconsin Alumni Research Foundation, which two university deans had persuaded alumni to set up in 1925 to manage it. The patent listed what the lamp could activate, from cereals, olive oil, and lard to "dog biscuit, feed for chicks", and oleo oil. It also warned that seventeen hours under the lamp destroyed the effect in olive oil and cod liver oil. 12; 10
The foundation turned away requests from the makers of "chewing gum and tobacco, beer and whiskey, pillows and bed sheets". It licensed Quaker oatmeal, Fleischmann's yeast, and five drug companies that sold irradiated ergosterol as Viosterol, and in 1932 it licensed evaporated milk. It refused margarine, although oleo oil was on the patent's list, because the university's deans wanted to protect Wisconsin butter. In 1944 a federal appeals court called oleomargarine "one of the foods of the poor" and ordered the Steenbock patents invalidated. 13
The 1928 Nobel Prize in Chemistry went to Adolf Windaus for his work on sterols and their connection with the vitamins. The presentation speech named ergosterol, a sterol found in yeast and ergot, as the mother substance of vitamin D, and noted that "5 mg of irradiated ergosterol has the same action" against rickets "as 1 litre of good cod-liver oil". In the mid-1930s Windaus and F. Bock identified the precursor in animal skin, from the sterol of pig skin, as 7-dehydrocholesterol. Vitamin D made from ergosterol is called vitamin D2, and the kind made in skin from 7-dehydrocholesterol is vitamin D3. 14; 15; 10
Skin stops making previtamin D3 at about 15 percent of its precursor
Once ultraviolet B has turned 7-dehydrocholesterol into previtamin D3, the warmth of the skin converts it into vitamin D3 over the next few days, and a carrier protein moves it into the blood. The liver turns vitamin D into 25(OH)D, and the kidney turns that into the active hormone, which helps the gut absorb calcium. 16; 17
In 1967 W. F. Loomis proposed in Science that skin color had evolved to regulate this reaction. Pale skin in the north let in enough of the weak winter ultraviolet, and dark skin near the equator could block "up to 95 percent" of it and so prevent "toxic doses of vitamin D". 18
Holick's group in Boston tested the idea on surgically removed human skin in 1981. The Fitzpatrick scale sorts skin into six types, from type I, which always burns, to type VI, which is deeply pigmented, and the group put samples of type III, V, and VI skin under a 2.5-kilowatt xenon lamp filtered to imitate noon at the equator. Lamps of this kind are called solar simulators, and this one was strong enough that about ten minutes under it matched three hours or more of equatorial sun. In every skin type previtamin D3 rose to about 15 percent of the 7-dehydrocholesterol and stopped. Longer exposure turned the previtamin D3 into lumisterol and tachysterol, which the authors called biologically inert. 7
Type III skin reached the ceiling in the equivalent of 30 minutes of equatorial noon, type V in about an hour, and type VI in about three hours. "In order of importance," the authors wrote, the limits on making previtamin D3 in skin were "(i) photochemical regulation, (ii) pigmentation, and (iii) latitude." 7
A 1982 follow-up from the same laboratory shone narrow bands of lamp light on skin samples and found that the reaction ran best between 295 and 300 nanometers, with a peak near 297, at the short end of the ultraviolet B that reaches the ground. It was paid for in part by the National Institutes of Health and the Duro-Test Corporation, and it ended by suggesting that the results "may also be helpful in the design of radiation sources" that would raise previtamin D3 in skin. 19
In 1989 Webb and Holick exposed human skin that already held vitamin D3 to Boston sunlight and found the vitamin "exquisitely sensitive to sunlight", breaking down into other compounds. Holick later wrote that "excessive exposure to sunlight will not result in vitamin D intoxication". 20; 21
In Boston, winter sun made no previtamin D3 from November through February
In June the Boston solution showed previtamin D3 after two minutes of noon sun, and 2 percent of the 7-dehydrocholesterol had converted after five. The amount climbed to about 9 percent in the first hour and levelled off at 12.3 percent after three. In January, one hour of cloudless noon sun converted 10 percent in Puerto Rico, at 18 degrees north, and 3 percent in Los Angeles, at 34. In spring the Boston solution first made previtamin D3 on 28 February, and the skin beside it not until 17 March. 1
In 1989 Matsuoka's group in Philadelphia looked for the smallest ultraviolet dose that raises vitamin D in people, giving 32 untanned young white adults single whole-body doses from lamps in a phototherapy unit. Blood vitamin D3 rose significantly only from 18 millijoules per square centimeter upward, and the authors wrote that winter sunlight in the northern United States "does not generally reach this threshold level". 22
In 2012 a New Zealand group gave volunteers twelve winter weeks of twice-weekly sessions in two kinds of phototherapy booth, then combined the blood results with measurements of sunlight. They inferred that "the production of 25(OH)D from sunlight should be possible throughout the year, although in winter the exposures necessary to maintain optimal levels of 25(OH)D would be impractically long." That, they wrote, "is at variance with the commonly-held view that no vitamin D is produced at mid-latitudes in the winter." 23
A Portuguese study of 1,497 people aged 65 or older matched each blood sample to the average ultraviolet (UV) index, the standard scale of how strong the sun's burning ultraviolet is, in the person's district over the previous 30 days. The index began to climb in February, but blood levels did not start to rise until April, when that average passed 3. 24
Midday sun makes the most previtamin D3; early-morning sun and clothing make little or none
In Tirupati, in south India at 13 degrees north, a team set sealed glass ampoules of 7-dehydrocholesterol in the sun every hour from 8 a.m. to 4 p.m. between May 2007 and August 2008. At 9 a.m. on 18 January 2008, 1.3 percent converted to previtamin D3; at 1 p.m. the same day, 18 percent of the 7-dehydrocholesterol converted. Conversion peaked between 11 a.m. and 2 p.m. in every season, and the authors concluded that sun in those hours "will promote vitamin D production in the skin year round". 25
In Abu Dhabi, at 24 degrees north, the same kind of ampoules set out hourly on one day a month for a year made "very little if any previtamin D3" before 9 a.m. or after 4 p.m. Both studies measured a solution in glass, and neither measured anyone's blood. 26
At a subtropical site in the Southern Hemisphere, in autumn and winter, the ultraviolet that makes previtamin D3 measured about 52 percent of full sun under a shade umbrella and 55 percent in tree shade. Under a covered veranda it was about 11 percent, and inside a car with the windows closed it was zero. 27
Ordinary window glass stops ultraviolet B, and the US National Institutes of Health states that sunshine through a window "does not produce vitamin D". When a team in Al Ain, in the United Arab Emirates, built a sun-tracking mirror in 2026 to bring sunlight to people sitting indoors, they kept the window "fully open to avoid attenuation of the reflected light." 17; 28
In 1992 Matsuoka's group dressed volunteers in jogging outfits of cotton, wool, or polyester, in black or white, covered their faces, necks, and hands with a sunscreen of sun protection factor (SPF) 39, and gave them whole-body doses from a lamp in winter. Black wool blocked 98.6 percent of the ultraviolet B, and white cotton blocked the least. No fabric let blood vitamin D3 rise, even at six minimal erythema doses, six times the dose that turns each person's skin just pink, and ordinary street clothes did the same. "Even light white cotton, plain weave cloth prevents vitamin D photosynthesis in Caucasians," the authors wrote. 29
The sunscreen page covers sunscreen and vitamin D. Under a lamp a sunscreen blocked vitamin D almost completely, while over an Australian summer and on a holiday in Tenerife users' levels still rose. A 2019 review of four lamp studies, three field trials, and 69 observational studies found "little evidence that sunscreen decreases 25(OH)D concentration when used in real-life settings", and noted that "there have been no trials of the high-SPF sunscreens that are now widely recommended." 30
Darker skin needs more ultraviolet for the same vitamin D, and how much more is disputed
Melanin, the pigment of skin, absorbs ultraviolet B before it reaches the 7-dehydrocholesterol in the deeper layers of the epidermis, the outer layer of the skin. In 1982 Holick's group gave one minimal erythema dose from a lamp to two white and three Black volunteers. Blood vitamin D rose up to 60-fold in the two white volunteers and did not change significantly in the three Black volunteers, until one of them took a dose six times larger and matched the white response. 10; 31
In 1986 the same laboratory tried again with six Indian and Pakistani immigrants aged 18 to 30 from a Boston university community and four white volunteers, and gave each person 1.5 times their own burn dose from sunlamps. The Asian volunteers needed 49 to 133 millijoules per square centimeter to redden, against 31 to 48, so they received more light. Their blood vitamin D peaked at 38 ng/mL against 27, a difference the study did not find significant. Asians "may need longer exposure to sunlight than Caucasians do to give a similar response, but the capacity to produce vitamin D is no different," the authors wrote. 32
In a study published in 2007, Holick's group set skin samples out in the noon sun of a cloudless 20 June in Boston. Type II skin from a 60-year-old turned 0.67 percent of its 7-dehydrocholesterol into previtamin D3 in 5 minutes, and type V skin from a 57-year-old made none. After 30 minutes the figures were 2.78 and 0.29 percent. 33
In Manchester in 2013, 51 South Asian adults finished six weeks of simulated summer sun from a lamp, three sessions a week, in doses equal to 15 to 90 minutes of noon sun in casual clothes. All 51 started below 20 ng/mL and 35 percent below 5. Those given the equivalent of 45 minutes or more averaged 15.7 ng/mL at the end, and 6 of the 51 reached 20. 34
A 2016 Manchester study gave 10 white and 6 South Asian volunteers an identical six-week lamp course and looked at their skin as well as their blood. Blood vitamin D rose from 36.5 to 54.3 nmol/L in the white group and from 17.2 to 25.5 in the South Asian group, the same proportional gain of about half. At the end of the course, the white skin had a median of 234 cells per square millimeter with ultraviolet damage to their DNA, and the South Asian skin 12. A day later the white skin had cleared 44 percent of that damage and the South Asian skin 97 percent. 35
In 2020 Antony Young's group in London gave 102 young adults of skin types II to VI, "white to black", the same dose below the burn threshold on 85 percent of the body, five times at intervals of three or four days, from a solar simulator or from a narrowband UVB lamp, which gives out a narrow band of ultraviolet near 311 nanometers. Only type II rose faster than the others, and the palest skin made vitamin D 1.3 to 1.4 times as fast as the darkest. The authors concluded that the effect of melanin "is small, compared to erythema", the medical word for reddening, "but that this difference may be sufficient to explain the epidemiological data." 36
Identical lamp doses raised blood vitamin D by anything from 3 to 139 nmol/L
In the autumn of 2010 dermatologists at Bispebjerg Hospital in Copenhagen, among them Pameli Datta and Hans Christian Wulf, gave 22 light-skinned adults from Denmark, the Faroe Islands, and England, described in the paper as "sun worshippers", identical ultraviolet B doses on about 80 percent of the body for nine weeks. The rise in blood vitamin D ran from 3 to 139 nmol/L. Two volunteers who started at almost the same level, 103.4 and 99.7 nmol/L, ended 85 nmol/L apart. Age, height, untanned skin color, and variants of the vitamin D receptor gene together explained 15.1 percent of the difference in how fast people rose. 2
Nineteen of them then went for a week of sunbathing in Hurghada, in Egypt, wearing electronic ultraviolet meters, to check that the lamps had taken them to their ceiling. They received a mean ultraviolet B dose of 66 kilojoules per square meter, or 22 once their measured sunscreen use was taken into account, against 4.2 in the last lamp week. Their mean blood vitamin D was 132 nmol/L before the holiday and 129 after. 2
A 2019 study from the same group put 40 volunteers whose ancestry spanned 16 countries, 22 of them from the 2016 study, through the same lamp course and found rises from 2.9 to 139 nmol/L. Measured skin color explained less than 3 percent of it, and variants of pigment genes explained more. 37
In 204 Swedish twins, genes accounted for half of the person-to-person spread in summer vitamin D. In winter, 72 percent of the spread came from what each pair of twins shared, which the authors took to be the height of the sun. 38
In 1985 surgically removed skin from people aged 77 and 82 made less than half the previtamin D3 of skin from people aged 8 and 18 when both were exposed to ultraviolet in the laboratory. In 2024 a Manchester group took biopsies from 21 healthy volunteers and found as much 7-dehydrocholesterol in the skin of 10 people aged 65 to 89 as in 11 people aged 18 to 40. Among 15,148 adults in the third US National Health and Nutrition Examination Survey (NHANES III), a blood survey, those over 60 who were active outdoors every day averaged 77 nmol/L, close to the 79 of adults aged 20 to 39. 39; 40; 41
When obese and lean white adults received whole-body ultraviolet from a lamp, blood vitamin D3 rose 57 percent less in the obese group 24 hours later, although their skin held as much 7-dehydrocholesterol and converted it as well. The authors attributed the difference to vitamin D being stored in body fat. 42
The minutes in public advice come from lamp studies and models
The World Health Organization's website said in 2014 that "5 to 15 minutes of casual sun exposure of hands, face and arms two to three times a week during the summer months is sufficient", and that "for most people, vitamin D deficiency is unlikely." Its 2024 answer is face, hands, and arms bare two or three times a week for half the time it takes unprotected skin to burn, "around 12 minutes for sensitive people (skin type II)" at a UV index of 7. The same page calls vitamin D production "the only beneficial effect of UV". 43; 44
The US National Institutes of Health cites "approximately 5 to 30 minutes of sun exposure, particularly between 10 a.m. and 4 p.m." from "some expert bodies and vitamin D researchers". In 2016 the UK's Scientific Advisory Committee on Nutrition declined to give any figure for how much summer sun would carry people through the winter, "because of the number and complexity of factors that affect endogenous vitamin D production." 17; 45
In January and February of one winter, when British sun holds almost no ultraviolet B, Lesley Rhodes's group in Manchester put 120 white adults aged 20 to 60, in T-shirts and knee-length shorts, into a lamp cabinet three times a week for six weeks. Each session gave 1.3 standard erythema doses, a standard erythema dose being a fixed amount of skin-reddening ultraviolet, the same for everyone. Each session matched 13 minutes of clear June noon in Manchester on about 35 percent of the skin. Among the 109 who finished, blood vitamin D rose on average from 17.6 to 28.0 ng/mL, and the individual rises ranged from 2.1 to 31.9. 46
From those numbers the team projected that 90 percent of white adults in Greater Manchester would reach 20 ng/mL after a summer of such sun. They calculated that the same dose would take 9 minutes of midsummer noon in New Orleans or Washington, 10 in Boston, and 16 in Oslo. The study was paid for by Cancer Research UK, and Rhodes declared that she had advised SunSmart, the government-funded campaign that told people to "seek shade" between 11 a.m. and 3 p.m., the midday hours whose sun the cabinet imitated. 46
In 2018 Ann Webb, by then at the University of Manchester, and Rhodes combined those lamp courses with ten years of British weather records. By their model, white adults would need nine minutes of lunchtime sun a day from March to September, with forearms and lower legs bare in June to August, and adults with brown skin, type V, about 25 minutes. "Exposing only the hands and face throughout the summer does not meet requirements," they wrote. 47; 48
Twenty-three Finnish adults with eczema who were sent to Gran Canaria to sunbathe for two weeks in real sun received a median of 60 standard erythema doses in January and 109 in March, estimated from their diaries and a meter, and their vitamin D rose by 13.4 and 24.0 nmol/L. 49
Low vitamin D is common in some of the sunniest countries
In the winter of 2002 to 2003, Hossein Dahifar's group in Tehran tested 414 schoolgirls aged 11 to 15. They spent a mean of 10 minutes a day in the sun, and all took in less than 400 IU of vitamin D a day. Fifteen of them, 3.6 percent, had blood signs of rickets, with vitamin D averaging 7.8 ng/mL. The paper explains that from the age of nine the dress code covers the body "with long dark garments and maghnaeh", a head scarf, and that schoolgirls "can not take off their maghnaeh and long garments in school even during break and exercise hrs." 50
In Honolulu, 93 young adults who reported 28.9 hours a week in the sun averaged 31.6 ng/mL, and 51 percent were below the 30 ng/mL cutoff the authors used. 51 In Karachi, 50 adults chosen for low, moderate, or high sun habits wore ultraviolet meters for four days; the high group spent a mean of 329 minutes in the sun over those days and averaged 17 ng/mL, and 49 of the 50 were deficient or insufficient. 8 In south India, surveys had found deficiency even among "agricultural laborers exposed to sunlight for more than 4 h". 25
In Malaysia in 2013, women over 45 in a rural settlement spent a median 7.83 hours a week in the sun and had a median of 69.5 nmol/L. Women in Kuala Lumpur spent 2.92 hours and had 31.9, although they bared more skin. 52
Across six Brazilian cities, from Recife at 8 degrees south to Porto Alegre at 33, 1,933 women screened for a trial of the osteoporosis drug arzoxifene had less vitamin D the farther south they lived; in Porto Alegre 24.5 percent were below 50 nmol/L, although women below 25 had been excluded. 53 Across six regions of Iran, from 29 to 37.5 degrees north, 1,406 adults averaged 26.8 nmol/L, and latitude made no difference. 4
In Sbea, in Libya, pregnant women who covered all but face and hands averaged 13.2 ng/mL, and those who bared face, arms, and legs averaged 27.6. 54 The 16 healthy adults in the 2026 Al Ain mirror study started at a mean of 5.0 ng/mL and ended at 7.7 after eight weeks of sessions. 28
In a 1999 trial of 123 Nigerian children with rickets, who ate about 200 milligrams of calcium a day, calcium alone brought 61 percent to near-complete healing within 24 weeks, calcium with vitamin D 58 percent, and vitamin D alone 19 percent. The authors concluded that these children ate too little calcium and did better on calcium, with or without vitamin D, than on vitamin D alone. 55
Committees drew the line for deficiency at 12, 20, and 30 ng/mL
The 2011 Institute of Medicine report set the Recommended Dietary Allowance, the daily intake meant to cover nearly everyone, at 600 IU a day to age 70 and 800 IU after, matched to a blood level of at least 20 ng/mL. It derived those figures "based on conditions of minimal sun exposure due to wide variability in vitamin D synthesis from ultraviolet light and the risks of skin cancer." It concluded that "the prevalence of vitamin D inadequacy in North America has been overestimated." The US National Institutes of Health, following the same committee, puts the risk of deficiency below 12 ng/mL, or 30 nmol/L. 3; 17
In July 2011 the Endocrine Society published a guideline, with Holick as chair, that called deficiency "very common in all age groups" and set its target at 30 ng/mL. Members of the Institute of Medicine committee replied that they disagreed "that serum 25-hydroxyvitamin D ... levels of 30 ng/ml or higher compared with 20 ng/ml provide increased health benefits", and that the guideline "inflates the number of persons deemed to be vitamin D deficient." Commercial laboratories including Quest and LabCorp adopted the higher figure. Clifford Rosen, a co-author of the Institute of Medicine report, told Kaiser Health News that it labels 80 percent of Americans as having too little. 56; 57; 58
The guideline's task force "received no corporate funding or remuneration." In 2018 Kaiser Health News reported that Holick had worked as a consultant to Quest Diagnostics, which performs vitamin D tests, since 1979, and had received nearly $163,000 from drug companies from 2013 to 2017. It also reported that the UV Foundation, a nonprofit arm of the Indoor Tanning Association, gave Boston University $150,000 for his research from 2004 to 2006. A 2007 study from his laboratory, in which volunteers lay in a tanning bed dosed "based on the manufacturer recommendation", acknowledged "a grant from the UV Foundation." Holick told the reporter that industry funding "doesn't influence me in terms of talking about the health benefits of vitamin D." 56; 58; 33
Doctors ordered more than 10 million vitamin D tests for Medicare patients in 2016, 547 percent more than in 2007, at a cost of $365 million. 58 In 2004 Binkley's group sent split blood samples from 10 healthy adults to six laboratories and got means from 17.1 to 35.6 ng/mL; whether a person had low vitamin D, they wrote, "is a function of the laboratory used." 59
In 2024 the Endocrine Society issued a new guideline, with a different panel, that found "no clear evidence defining the optimal target level of 25(OH)D" and advised against routine testing, including for people with obesity or dark skin. The US Preventive Services Task Force had written in 2021 that "no one serum vitamin D level cutpoint defines deficiency". On the 30 nmol/L line, 5.0 percent of Americans were at risk of deficiency in 2011 to 2014; in a standardized European survey of 55,844 people, 13.0 percent were below 30 and 40.4 percent below 50. 60; 61; 62; 63
Vitamin D pills raised blood levels but did not prevent cancer, heart disease, or fractures in large trials
VITAL was paid for by the US National Institutes of Health. Pharmavite donated the vitamin D and the dummy pills, and Quest Diagnostics measured blood vitamin D at no cost to the study. The volunteers were not chosen for low vitamin D, and at the start they averaged 30.8 ng/mL, with 12.7 percent below 20. In a subset of 1,644 retested after a year, the vitamin D group had risen from 29.8 to 41.8 ng/mL. 5
For cancer, the true effect of the pills could be anywhere from about 12 percent fewer cancers to 6 percent more. For heart attacks, strokes, and cardiovascular deaths it could be anywhere from 15 percent fewer to 12 percent more. Fractures came to 769 of 12,927 people in the vitamin D group against 782 of 12,944 in the dummy-pill group. 5; 64
In Australia's D-Health trial, 21,315 people aged 60 or older took 60,000 IU once a month or a dummy pill for five years. Deaths came to 562 against 538, and the true effect could be anywhere from 7 percent fewer deaths to 18 percent more. The dummy-pill group averaged 77 nmol/L during the trial, and the authors wrote that the dose "might not be appropriate in people who are vitamin D-replete." 65
A 2018 meta-analysis, a study that pools earlier studies, combined 81 randomized trials with 53,537 people and found that vitamin D pills did not reduce fractures, hip fractures, or falls. In one trial in Victoria, Australia, 2,256 women aged 70 or older took a single 500,000 IU dose each autumn or winter and fell more often than those on dummy pills, 83.4 falls for every 100 women followed for a year against 72.7. 66; 67
Where real sun was tried against pills, the pills raised blood levels more
In the Tehran study, the 15 girls with signs of rickets were split in two. One group sat with hands and face in the winter sun for an extra hour a day for 20 days, and their vitamin D rose from 7.1 to 13.9 ng/mL. The other took 50,000 IU of vitamin D every day for 20 days, a million units in all, and rose from 7.4 to 27.9. 50
The Sydney trial randomized 51 care homes and their 602 residents, whose mean age was 86.4, to an extra 30 to 40 minutes of early-morning sun a day for a year, with or without calcium, or to their usual routine. The homes employed "sunlight officers" to bring residents outside. The median resident came to 26 percent of the sessions, and blood vitamin D, which started at a median of 32.9 nmol/L, "increased only slightly". Falls did not change, and the true effect on falls could be anywhere from about a quarter fewer to half as many again. The authors attributed the failure "most likely" to poor adherence. 6
When the researchers interviewed residents and sunlight officers, they found that the hoped-for health benefits did not bring people out. Socializing and being outdoors did, and the barriers were the "regimentation of daily attendance", clashes with other activities, "unsuitable timing and heat discomfort". A later analysis concluded that "natural sun exposure is not a practical intervention for vitamin D deficiency in this population." 68; 69
In Pune, India, men aged 40 to 60 with blood levels below 50 nmol/L were randomly given either at least 20 extra minutes of sun on face and forearms between 11 a.m. and 3 p.m. or 1,000 IU of vitamin D3 a day, for six months. Among those who finished, 32 of 50 in the sun group and 37 of 50 in the pill group, the sun group rose from 35.6 to 48.3 nmol/L and the pill group from 31.9 to 57.1. The trial was built to test cholesterol, and the sun group's total cholesterol fell while the pill group's rose. Sun time was reported by the men themselves, and the study reported no funding. 70
In Isfahan, Iran, 87 pregnant women with low vitamin D were advised either to sit in summer sun for 30 minutes a day with about 30 percent of the body bare and no sunscreen, or to take 4,000 IU a day, for ten weeks. The pill group ended at 31.27 ng/mL and the sun group at 19.79, and the babies' length, weight, and head size did not differ. Nobody measured how much sun the women got. 71
In the winter of their basic training at Catterick, in the north of England, 250 British Army recruits were assigned by chance to a lamp cabinet giving 1.3 standard erythema doses in T-shirt and shorts, three times a week for four weeks and then weekly, to daily vitamin D3 pills, or to a placebo, a dummy treatment. The placebo cabinet was the same cabinet with its tubes wrapped in a film that blocked 99.9 percent of the ultraviolet. The lamp sessions lasted a mean of 222 seconds, and lamp and pill each brought 95 percent or more of the men to 50 nmol/L by week 5. Against placebo, the two vitamin D groups had 36 percent fewer days with a cold, and the lamp and pill groups did not differ. The trial was paid for by the UK Ministry of Defence. 9; 72
Misconceptions
What people believe
- Morning sun, before 9 or 10 a.m., is the best sun for making vitamin D.
- Fifteen minutes of sun a day gives everyone enough vitamin D.
- People with dark skin cannot make vitamin D from sunlight.
- People who live in sunny countries do not run short of vitamin D.
- More sun always means more vitamin D, and enough of it can poison you.
- Vitamin D pills need sunlight to work.
- A vitamin D pill does what sunlight does.
What the studies show
- Sealed ampoules in India and Abu Dhabi made little previtamin D3 before 9 a.m. and the most between 11 a.m. and 2 p.m. 25; 26
- The World Health Organization's 2014 figure was 5 to 15 minutes two or three times a week in summer. The minute counts behind current advice come from lamp courses on white adults in Manchester and weather models. 43; 46; 47
- Darker skin samples reached the same ceiling as pale ones, only more slowly, and in 102 young adults under the same lamp dose the palest skin made vitamin D 1.3 to 1.4 times as fast as the darkest. 7; 36
- Low levels were common among adults in Iran, Karachi, urban Malaysia, and Honolulu. 4; 8; 52; 51
- Skin caps previtamin D3 at about 15 percent of its precursor, sunlight breaks down vitamin D3 in skin, and 19 volunteers already at their lamp ceiling gained nothing from a week of Egyptian sun. 7; 20; 2
- Pills raised blood vitamin D in soldiers training through a British winter who got 0.22 standard erythema doses of sun a day. 73
- In VITAL and D-Health, pills raised blood levels without lowering cancer, cardiovascular events, or deaths. No trial of comparable size has tested sunlight. 5; 65
Questions people ask
How long do you need to be in the sun to get vitamin D?
No study has timed real sun against blood levels across skin types and places. The Manchester lamp course, matched to 13 minutes of June noon three times a week on 35 percent of the skin, took most white adults to 20 ng/mL, and a model built on it gave nine minutes of lunchtime sun a day from March to September for white skin and about 25 for brown. 46; 47; 48
Is morning sun good for vitamin D?
In sealed ampoules in south India, 9 a.m. January sun converted 1.3 percent of the precursor and 1 p.m. sun 18 percent. In Abu Dhabi almost nothing formed before 9 a.m. or after 4 p.m. 25; 26
Can you get vitamin D through a window?
Ordinary window glass blocks the ultraviolet B that starts the reaction, and the US National Institutes of Health states that sun through a window does not produce vitamin D. The evidence comes from light measurements and lamp work, and no trial has tested it in people. 17
Can you get vitamin D in the shade or on a cloudy day?
At one subtropical site, a shade umbrella let through about half the ultraviolet that makes previtamin D3, and a closed car none. In Boston and Edmonton, cloudless winter sun made none for four to six months of the year. 27; 1
Is sunlight better than vitamin D supplements?
In trials in Tehran, Pune, and Isfahan, pills raised blood levels more than advised sun, and in the British Army a lamp and a pill did equally well, on blood levels and on colds. No large trial has compared real sun and pills on a disease. 50; 70; 71; 9
Why are people in sunny countries low in vitamin D?
The studies point to covered skin, time indoors, and diet: covered women in Libya and Tehran had lower levels, and urban Malaysian women spent less time in the sun than rural ones. No study has separated these factors in the same people. 54; 50; 52
Does darker skin make less vitamin D?
Darker skin makes vitamin D more slowly, and the estimates of how much more slowly differ. In 1982 one Black volunteer needed six times the lamp dose to match two white volunteers; in 2020, among 102 young adults, the palest skin made vitamin D 1.3 to 1.4 times as fast as the darkest. 31; 36
Can you get too much vitamin D from the sun?
Skin stops making previtamin D3 at about 15 percent of its precursor, and further sun breaks down vitamin D3 in skin. The highest level among 93 sun-exposed young adults in Honolulu was 62 ng/mL. 7; 20; 51
Do tanning beds give you vitamin D?
In a Boston tanning bed study paid for partly by the UV Foundation, 12 weeks of sessions raised blood levels by 140 to 310 percent depending on skin type. The US National Institutes of Health says moderate use of tanning beds "is also effective", and the World Health Organization says artificial tanning "should never be considered as an option to achieve sufficient vitamin D status." 33; 17; 74
What we do not know
Nobody knows how many minutes of real sun a person with brown or black skin needs in a sunny country. The minute counts in circulation were calculated from lamp courses in Manchester and weather models, mostly for white skin, and the real-sun trials in Pune, Isfahan, and Tehran gave advice rather than measured doses. 46; 48; 70
Why people in sunny countries run low is unsettled. Clothing, time indoors, skin color, body fat, and diet each went with lower levels in some study, and no study has measured them together in the same people. 50; 52; 54
Nobody knows what blood level is enough for health. The 2011 committees disagreed on 20 against 30 ng/mL, the 2024 Endocrine Society panel found no clear target, and the same blood can read 17 or 35 ng/mL depending on the laboratory. 57; 60; 59
Why identical doses raise one person's vitamin D by 3 nmol/L and another's by 139 is mostly unexplained. Age, height, skin color, and receptor gene variants accounted for 15 percent of the difference in 22 people. 2
Nobody has tested, in a trial large enough to count fractures, cancers, or deaths, whether regular sunlight does something that vitamin D pills do not. The largest randomized trial of real sunlight we found enrolled 602 residents of Sydney care homes, and the median resident came to about a quarter of the sessions. 6
Key studies
- Kramer and Boone, 1922, Baltimore. Real sun. Seven Black children with rickets healed within three weeks of daily sun. 11
- Holick, MacLaughlin, and Doppelt, 1981, Boston. Solar simulator on skin samples. Previtamin D3 stopped at about 15 percent in every skin type; darker skin took longer. 7
- MacLaughlin, Anderson, and Holick, 1982, Boston. Lamp on skin samples. Previtamin D3 formed best near 297 nanometers. 19
- Clemens and colleagues, 1982, Boston. Lamp. Blood vitamin D rose in two white volunteers and not in three Black volunteers at the same dose. 31
- Lo, Paris, and Holick, 1986, Boston. Lamp. Six South Asian and four white volunteers made similar vitamin D at doses scaled to their burn threshold. 32
- Webb, Kline, and Holick, 1988, Boston and Edmonton. Real sun on a solution and skin samples. No previtamin D3 from November through February in Boston. 1
- Matsuoka and colleagues, 1992, Philadelphia. Lamp. Clothing of every fabric tested stopped the rise in blood vitamin D3. 29
- Rhodes and colleagues, 2010, Manchester. Lamp. A course matched to 13 minutes of June noon raised 109 white adults from 17.6 to 28.0 ng/mL. 46
- Datta and colleagues, 2016, Copenhagen. Lamp, then real sun. Identical doses raised 22 people by 3 to 139 nmol/L; a week in Egypt added nothing. 2
- Felton and colleagues, 2016, Manchester. Lamp. South Asian skin made less vitamin D and far less DNA damage than white skin. 35
- Young and colleagues, 2020, London. Lamp. In 102 young adults, melanin slowed vitamin D synthesis by a factor of 1.3 to 1.4. 36
- Harinarayan and colleagues, 2013, Tirupati. Real sun on sealed ampoules. Previtamin D3 formed mostly between 11 a.m. and 2 p.m. 25
- Dahifar and colleagues, 2007, Tehran. Blood survey, then real sun against pills. Signs of rickets in 15 of 414 covered schoolgirls. 50
- Humayun and colleagues, 2012, Karachi. Real sun with meters, and blood. 49 of 50 adults were low. 8
- Nikooyeh and colleagues, 2017, Iran. Blood. 90.7 percent of 1,406 adults below 50 nmol/L, with no latitude gradient. 4
- Sambrook and colleagues, 2012, Sydney. Real sun, randomized. Early-morning sun in care homes barely moved vitamin D or falls. 6
- Patwardhan and colleagues, 2017, Pune. Real sun against pills, randomized. Pills raised vitamin D more. 70
- Harrison and colleagues, 2021, Catterick. Lamp against pills, randomized. Both brought 95 percent or more of recruits to 50 nmol/L. 9
- Manson and colleagues, 2019, United States. Pills, randomized. 2,000 IU a day did not lower cancer or cardiovascular events in 25,871 people. 5
- Neale and colleagues, 2022, Australia. Pills, randomized. 60,000 IU a month did not lower deaths in 21,315 people. 65
- Ross and colleagues, 2011, and Holick and colleagues, 2011. The Institute of Medicine and Endocrine Society thresholds of 20 and 30 ng/mL. 3; 56
Definitions
- 25-hydroxyvitamin D, or 25(OH)D. The form of vitamin D that the liver makes and that blood tests measure. Also called calcidiol.
- 7-dehydrocholesterol. A molecule in the outer layer of the skin that ultraviolet B turns into previtamin D3. It is a close chemical relative of cholesterol.
- Confidence interval. The range within which the true effect of a study probably lies, given the play of chance.
- Dosimeter. A small meter worn on the body that records how much ultraviolet the wearer receives.
- Ergosterol. A sterol found in yeast and fungi. Ultraviolet light turns it into vitamin D2.
- Fitzpatrick skin type. A six-step scale of how skin responds to sun, from type I, which always burns and never tans, to type VI, which is deeply pigmented and almost never burns.
- Heliotherapy. Treatment with real sunlight, usually for skin disease.
- International unit (IU). A measure of vitamin D in food and pills. 40 IU is one microgram.
- Lumisterol and tachysterol. Compounds that form when previtamin D3 keeps absorbing ultraviolet. They do not act as vitamin D.
- Meta-analysis. A study that pools the results of earlier studies into one estimate.
- Minimal erythema dose. The smallest dose of ultraviolet that leaves one person's skin just visibly red a day later.
- Nanomoles per liter and nanograms per milliliter. The two units for blood vitamin D. One ng/mL is 2.5 nmol/L, so 20 ng/mL is 50 nmol/L.
- Narrowband UVB. A lamp that gives out a narrow band of ultraviolet near 311 nanometers, used to treat skin disease.
- Placebo. A dummy pill or treatment that looks like the real one but does nothing.
- Previtamin D3. The first product when ultraviolet B strikes 7-dehydrocholesterol. The warmth of the skin turns it into vitamin D3 over a few days.
- Randomized trial. A study in which chance decides who gets which treatment, so that the groups are alike in everything else, apart from chance.
- Rickets. A disease of growing children in which the bones stay soft, from too little vitamin D, too little calcium, or both.
- Solar simulator. A lamp built to give off a spectrum close to that of sunlight.
- Standard erythema dose. A fixed amount of skin-reddening ultraviolet, the same for everyone, used to report a dose.
- UV index. The standard scale of how strong the sun's burning ultraviolet is at a given place and time.
- UVB. Ultraviolet light from 280 to 315 nanometers. It causes most sunburn and starts vitamin D production in the skin.
- Vitamin D2 and vitamin D3. Vitamin D2 is made by shining ultraviolet on ergosterol; vitamin D3 is the kind skin makes from 7-dehydrocholesterol. Both are sold as pills.
- Vitamin D winter. The months when the sun at a given latitude is too low to start vitamin D production in skin.
Statistics
Human skin samples and a 7-dehydrocholesterol solution set out in real sun on cloudless days in Boston (42.2°N) made no previtamin D3 from November through February. 1
In Edmonton (52°N), the months in which real sun made no previtamin D3 in the test solution ran from October through March. 1
In Boston noon sun in June, a 7-dehydrocholesterol solution showed previtamin D3 after only 2 minutes, and after 5 minutes 2 percent of the starting material had converted. 1
In June sun in Boston, previtamin D3 in the test solution rose to about 9 percent of the starting 7-dehydrocholesterol in the first hour and then levelled off at 12.3 percent after 3 hours. 1
One hour of January noon sun on a cloudless day converted 10 percent of the 7-dehydrocholesterol in the test solution to previtamin D3 in Puerto Rico (18°N) and only 3 percent in Los Angeles (34°N). 1
In spring in Boston, a 7-dehydrocholesterol solution in the sun first made previtamin D3 on February 28, but human skin samples beside it did not until March 17. 1
In Edmonton, cloudless-day sun did not start making previtamin D3 in the test solution until the beginning of April and stopped after October. 1
In surgically obtained skin samples under a solar simulator set to equatorial noon, the time needed to reach the most previtamin D3 rose from 30 minutes in type III skin to 1 hour in type V and 3 hours in type VI. 7
Whatever the skin type of the surgically obtained samples, previtamin D3 under the solar simulator stopped rising at about 15 percent of the starting 7-dehydrocholesterol. 7
In type III skin samples held under simulated equatorial sun for 8 hours, 7-dehydrocholesterol fell to 30 percent of its starting level while inert lumisterol rose to 50 percent. 7
In surgically obtained human skin under narrow-band lamp light, previtamin D3 formed best between 295 and 300 nanometers, with an apparent peak near 297 nanometers. 19
Adjacent skin samples given the equivalent of 15 to 30 minutes of simulated equatorial sun turned at most 15 to 20 percent of their 7-dehydrocholesterol into previtamin D3. 19
In six Indian and Pakistani adults given whole-body lamp UV at 1.5 times their burn dose, blood vitamin D rose from 2 to a peak of 38 ng/mL at 24 hours and fell back to baseline by day 9. 32
After the same whole-body lamp dose of 1.5 burn doses, peak blood vitamin D was 38 ng/mL in the six Asian volunteers and 27 ng/mL in four white volunteers, a difference the study did not find significant. 32
The lamp dose that reddened the skin was 49 to 133 mJ/cm2 in the six Asian volunteers, against 31 to 48 mJ/cm2 in the four white volunteers. 32
After 30 minutes of noon June sun in Boston, skin samples of type II had turned 2.78 percent of their 7-dehydrocholesterol into previtamin D3, against 0.29 percent in type V skin. 33
After 5 minutes of noon June sun in Boston, type II skin samples had made measurable previtamin D3 (0.67 percent of their 7-dehydrocholesterol) and type V samples had made none. 33
After 12 weeks of tanning-bed sessions at three quarters of a burn dose, blood 25(OH)D rose by 310 percent in skin type II volunteers, 287 percent in type III, 225 percent in type IV, and 140 percent in type V. 33
Young people with cystic fibrosis in Massachusetts had the lowest blood 25(OH)D of the study in winter, 15 ng/mL, and 3 of the 10 were at or below 10 ng/mL, although the patients took about 1,000 IU of vitamin D a day. 75
Long-stay geriatric patients in Helsinki spent the least time outdoors of four groups followed through 1980 and spent no time outdoors at all in winter, spring, and autumn. 76
Among 538 white Dutch adults aged 60 to 87, 51 percent had blood 25(OH)D below 50 nmol/L in winter, against 34 percent in summer. 77
In 100 Swedish women aged 61 to 83 at 60°N, mean blood 25(OH)D rose 38 percent from winter to late summer, from 72 to 99 nmol/L. 78
In the same 100 Swedish women, preferring sun, having a skin type that does not burn easily, and normal weight together went with summer 25(OH)D 64 nmol/L higher than preferring shade, sensitive skin, and obesity. 78
In late summer, 2 of the 100 Swedish women (2 percent) were vitamin D deficient at 50 nmol/L or less, and 20 percent were in the 50 to 75 nmol/L range. 78
In 204 Swedish twins aged 39 to 85 at 60°N, blood 25(OH)D averaged 84.8 nmol/L and ran 24.2 nmol/L lower in winter than in summer. 38
In the Swedish twins, genes explained half of the person-to-person spread in blood 25(OH)D in summer, while in winter 72 percent of the spread came from shared environment. 38
In 109 white Manchester adults who completed six weeks of simulated summer sun (1.3 standard erythema doses, three times a week, in T-shirt and shorts) during winter, mean blood 25(OH)D rose by 10.4 ng/mL, from 17.6 to 28.0 ng/mL. 46
After the same 6 weeks of simulated summer sun, the rise in blood 25(OH)D ranged from 2.1 to 31.9 ng/mL between the Manchester volunteers. 46
The Manchester team projected that after a summer of such short exposures, 90 percent of white Manchester adults would reach 20 ng/mL of 25(OH)D and 26.2 percent would reach 32 ng/mL. 46
The Manchester lamp dose matched 13 minutes of cloudless midday summer sun on 35 percent of the skin, three times a week for 6 weeks. 46
The weekly rise in mean blood 25(OH)D shrank as the Manchester lamp course went on, to only 0.8 and 0.2 ng/mL in the last two weeks. 46
From the winter baseline, the Manchester team estimated that 62.5 percent of white Manchester adults aged 20 to 60 are below 20 ng/mL of 25(OH)D in winter and 5 percent are below 5 ng/mL. 46
The Manchester team calculated that the same vitamin D dose would take 9 to 16 minutes of midday midsummer sun at North American and European locations, against 13 minutes in Manchester. 46
In 22 light-skinned adults from Denmark, the Faroe Islands, and England given identical lamp UVB doses on about 80 percent of the body for nine winter weeks, the total rise in blood 25(OH)D ran from 3 to 139 nmol/L, with a mean of 48. 2
In the same 22 adults, the highest blood 25(OH)D reached after identical lamp doses ranged from 85 to 216 nmol/L, with a mean of 134. 2
Two of the Copenhagen volunteers started at nearly the same blood 25(OH)D, 103.4 and 99.7 nmol/L, yet after identical lamp doses their total rises differed by 85 nmol/L. 2
In 19 of the Copenhagen volunteers who went on a one-week sun holiday straight after the lamp course, mean blood 25(OH)D did not change, 132 before and 129 nmol/L after. 2
Across the 22 Copenhagen volunteers, blood 25(OH)D rose on average 5.5 nmol/L a week under the lamp, with personal weekly rises from 0.32 to 15.4 nmol/L. 2
In 40 Danish residents with a wide range of skin color given identical lamp UVB doses on identical body areas for nine winter weeks, the rise in blood 25(OH)D ran from 2.9 to 139 nmol/L. 37
In the same 40 Danish residents, measured tanned skin color explained 2.6 percent and untanned skin color 1.9 percent of the person-to-person spread in the lamp-driven 25(OH)D rise. 37
Across the 40 Danish residents, mean blood 25(OH)D went from 67 nmol/L (range 10 to 120) before the nine-week lamp course to 118 nmol/L (range 73 to 216) after it. 37
In 102 young adults of skin types II to VI given the same below-burn lamp dose to 85 percent of the body five times, the palest skin (type II) made 25(OH)D3 only about 1.3 to 1.4 times as fast per dose as the darkest (type VI). 36
Under a solar-simulator lamp, the rise in 25(OH)D3 per dose was the same in skin types III, IV, V, and VI; only type II was steeper. 36
After 6 weeks of simulated UK summer sun, more than half of 6 South Asian (type V) volunteers reached 25 nmol/L of 25(OH)D, but none reached 50 nmol/L. 35
Under the same 6-week simulated summer course, 10 white (type II) volunteers raised blood 25(OH)D from 36.5 to 54.3 nmol/L and 6 South Asian (type V) volunteers from 17.2 to 25.5 nmol/L, the same relative gain of about 48 to 49 percent. 35
Three lamp sessions a week in the Manchester study, each taking about 6.5 minutes, were calculated to equal 13 to 17 minutes of UK June midday sun six times a week. 35
In Pune, India, men aged 40 to 60 with 25(OH)D below 50 nmol/L who added at least 20 minutes of midday sun on face and forearms for 6 months went from 35.6 to 48.3 nmol/L (32 finished), while those taking 1,000 IU of cholecalciferol a day went from 31.9 to 57.1 nmol/L (37 finished). 70
Of 50 men in each group of the Pune trial, 39 in the comparison group, who were already above 50 nmol/L and received nothing, 32 in the extra-sun group, and 37 in the supplement group finished the 6 months. 70
Among 87 vitamin D deficient pregnant women in Isfahan, Iran, 10 weeks of 4,000 IU a day raised 25(OH)D3 to 31.27 ng/mL, measured after delivery, against 19.79 ng/mL after 10 weeks of advice to spend 30 minutes a day in summer sun with 30 percent of the body bare. 71
In the Isfahan pregnant women who started below 10 ng/mL, 25(OH)D3 rose from 5.85 to 15.57 ng/mL on 4,000 IU a day but only from 5.56 to 7.93 ng/mL with daily sun. 71
Before the Isfahan trial began, mean 25(OH)D3 was 15.95 ng/mL in the women later given pills and 15.09 ng/mL in those advised to spend time in the sun. 71
Of 1,644 British Army recruits followed through basic training across the seasons, only 21 percent of those starting in winter had blood 25(OH)D of 50 nmol/L or more. 9
In the winter trial arm of recruits, only 27 percent were at 50 nmol/L or more before treatment, and both a solar-simulator lamp and oral vitamin D3 brought 95 percent or more to that level by week 5. 9
The lamp dose in the recruits' trial, 1.3 standard erythema doses on about 40 percent of the skin, matched about 15 minutes of midday summer sun six times a week for a casually dressed person in northern England. 9
Among 447 UK military trainees studied across the year, 81 percent of those measured in winter had blood 25(OH)D below 50 nmol/L and 32 percent were below 30 nmol/L. 73
In 119 men given a solar-simulator lamp, oral vitamin D3, or placebo from the winter low, 95 percent of those on either active treatment were at 50 nmol/L or more by week 5, with no difference between lamp and pill. 73
In the same winter trial, participants got only 0.22 standard erythema doses of real sun a day and 112 IU of vitamin D a day from food. 73
The same lamp dose that matched about 15 minutes of midday summer sun in northern England was calculated to equal about 12 minutes in Philadelphia (40°N) and about 18 minutes in Oslo (60°N). 72
Of 250 young men starting British Army training in winter, 29 percent were at 50 nmol/L of 25(OH)D or more, 37 percent were at 30 to 49.9, and 34 percent were below 30. 72
In 16 healthy adults in the United Arab Emirates who sat in real sunlight redirected indoors by mirrors for 10 to 20 minutes four times a week for eight weeks, mean total 25(OH)D rose from 5.0 to 7.7 ng/mL and 25(OH)D3 from 1.99 to 4.11 ng/mL. 28
Each indoor sunlight session in the UAE study lasted 10 to 20 minutes, set by skin tone and UV intensity, four times a week for eight weeks. 28
Among 1,933 Brazilian postmenopausal women sampled in fall and winter, 17 percent (329) had 25(OH)D below 50 nmol/L, even though women below 25 nmol/L had been excluded. 53
Low vitamin D in the Brazilian women rose from north to south, reaching 24.5 percent below 50 nmol/L in Porto Alegre (33°S), more than twice the share in Recife and Salvador. 53
Of the same Brazilian women, only 31.7 percent (612) had 25(OH)D above 75 nmol/L. 53
Across the six Brazilian cities, city mean 25(OH)D and latitude correlated at r = −0.88. 53
In Malaysian women over 45, median 25(OH)D was 69.5 nmol/L in 293 rural women and 31.9 nmol/L in 107 urban women. 52
Rural Malaysian women spent a median 7.83 hours a week in the sun against 2.92 hours for urban women, although the urban women bared more skin. 52
43.9 percent of the urban Malaysian women were below 30 nmol/L, while 88.1 percent of the rural women were above 50 nmol/L. 52
In the Malaysian women, living in the countryside added 31.74 nmol/L to 25(OH)D in a regression model, and each unit of sun index (hours of sun a week times fraction of skin bared) added 1.93 nmol/L. 52
In 100 cord blood samples from healthy full-term newborns in Hawai'i, mean 25(OH)D was 24.5 ng/mL, 28 percent were below 20 ng/mL, and 50 percent were at 20 to 30 ng/mL. 79
In the Hawai'i cord blood samples, mean 25(OH)D was 30.5 ng/mL in 19 white newborns, 25.1 in 43 Asian newborns, and 20.0 in 25 Pacific Islander newborns. 79
In 144 pregnant women in Sbea, Libya, mean 25(OH)D was 13.7 ng/mL, and 43.8 percent were below 10 ng/mL. 54
Only 9 percent of the 144 pregnant women in Sbea had 25(OH)D above 30 ng/mL. 54
Libyan pregnant women who reported 15 minutes of sun had mean 25(OH)D of 13.2 ng/mL, against 19.2 ng/mL in those who reported an hour. 54
Libyan pregnant women who covered all but face and hands had mean 25(OH)D of 13.2 ng/mL, against 27.6 ng/mL in those who bared face, arms, and legs to the sun. 54
In winter in Tehran, girls aged 11 to 15 with low vitamin D who put hands and face in the sun for an extra hour a day for 20 days went from 7.1 to 13.9 ng/mL of 25(OH)D, while girls given 50,000 IU of vitamin D a day for 20 days went from 7.4 to 27.9 ng/mL. 50
Of 414 Tehran schoolgirls aged 11 to 15 tested in winter, 15 (3.6 percent) had biochemical signs of rickets with mean 25(OH)D of 7.8 ng/mL. 50
The 414 Tehran schoolgirls spent a mean of 10 minutes a day in the sun (range 5 to 40), and all took in less than 400 IU of vitamin D a day. 50
Of 50 healthy adults in Karachi, Pakistan, whose vitamin D was measured, 49 (98 percent) were deficient or insufficient, with a mean of 12 ng/mL. 8
In the Karachi adults, mean vitamin D was 9.8 ng/mL in the low sun-exposure group, 11 in the moderate group, and 17 in the high group. 8
Over four days wearing dosimeters, the Karachi high-exposure group spent a mean 329 minutes in the sun, against 69.5 minutes for the low group. 8
In 1,406 healthy Iranian adults aged 19 to 60 from six regions between 29°N and 37.5°N, mean 25(OH)D was 26.8 nmol/L and 90.7 percent were below 50 nmol/L. 4
Across Iran, the lowest regional mean 25(OH)D (24.4 nmol/L) was at latitude 31.3°N and the highest (29.5 nmol/L) at 37.5°N, the northernmost region. 4
In the same Iranian adults, 69.3 percent were below 27.5 nmol/L and only 9.3 percent were above 50 nmol/L. 4
In 109 adults in northern Poland (54°N) tested twice in 2012, mean 25(OH)D was 13.3 ng/mL in late winter, with 81.1 percent deficient, and 22.8 ng/mL in early autumn, with 42.2 percent still deficient. 80
In 1,497 Portuguese adults aged 65 and over sampled from December to June, median blood 25(OH)D was 35.9 nmol/L. 24
Of 580 medical students aged up to 24 in Istanbul, only 51 (8.8 percent) had 25(OH)D of 30 ng/mL or more, and 54.1 percent were at 20 ng/mL or less. 81
Among women aged 19 to 39 in Daegu, Korea, the 52 measured in summer spent 23.8 minutes a day outdoors and had 25(OH)D of 17.5 ng/mL, while the 83 measured in winter spent 10.8 minutes and had 13.4 ng/mL. 82
In 52 mainly white elite athletes in the Netherlands who all started above 75 nmol/L in June, mean 25(OH)D peaked at 113 nmol/L at the end of summer and fell to 78 nmol/L by the end of winter, when nearly 20 percent were below 50 nmol/L. 83
In 603 Swiss athletes, the predicted chance of 25(OH)D below 75 nmol/L was 70 percent in winter and 17 percent in summer, and 58 percent for indoor sports against 43 percent for outdoor sports. 84
Of 29 elite swimmers aged 16 to 24 training indoors at 55 to 56°N, measured in March and April, 45 percent had 25(OH)D below 50 nmol/L, with a mean of 52.6. 85
In 19 Swedish girl soccer players aged 13 to 16 with about 32 to 36 hours of sun a month in winter, mean 25(OH)D was 50.5 nmol/L. 86
In 158 overweight Swedish adults aged about 60, mean 25(OH)D rose over the summer from 55.1 to 66.3 nmol/L, yet 66 percent stayed below 76 nmol/L. 87
In 7,560 children in southwest England with a mean age of 9.9, 29 percent had total 25(OH)D below 20 ng/mL, and winter and less time outdoors were among the main risk factors. 88
Across 15,148 US adults in NHANES III, mean 25(OH)D was 79 nmol/L at ages 20 to 39, 73 at 40 to 59, and 68 at 60 and over. 41
US adults who were active outdoors every day had 25(OH)D 13 nmol/L higher (ages 20 to 39) and 16 nmol/L higher (60 and over) than those with no outdoor activity in the past month. 41
US adults aged 60 and over with daily outdoor activity averaged 77 nmol/L of 25(OH)D, close to the 79 nmol/L of adults aged 20 to 39. 41
In a model built on 1,095 US male health professionals, being African American went with plasma 25(OH)D 12.8 nmol/L lower and being Asian 13.3 nmol/L lower than being white. 89
Of 23 Finnish adults with atopic eczema sent for 2 weeks of sun in the Canary Islands, 17 (74 percent) were below 50 nmol/L of calcidiol before the trip and 4 (17 percent) after it. 49
During 2 weeks of Canary Islands sun, the January group received a median 60 standard erythema doses and the March group 109, and their calcidiol rose by 13.4 and 24.0 nmol/L. 49
In Swedish psoriasis patients treated in winter, broadband UVB lamps raised 25(OH)D from 28.5 to 64.6 ng/mL and narrowband UVB lamps from 28.3 to 47.2 ng/mL. 90
Swedish psoriasis patients given 2 weeks of heliotherapy in real sun raised 25(OH)D from 22.9 to 41.8 ng/mL. 90
In 26 psoriasis patients who took a mean 41 minimal erythema doses of real sun over four weeks at Davos, blood vitamin D3 rose by 36.6 ng/mL, from 25.8 to 62.4 ng/mL, while 14 patients in the comparison group, who received 6 doses, barely changed. 91
In 24 patients with skin disease in the north of Scotland treated in winter with narrowband UVB lamps, mean 25(OH)D rose from 34 nmol/L to 58 after 2 weeks and 78 after 4 weeks. 92
The median narrowband UVB dose over the Aberdeen course was 39.1 standard erythema doses, about a quarter of the median summer sunlight dose people receive there. 92
In a 12-week winter trial of twice-weekly exposure in phototherapy booths, mean 25(OH)D rose by more than 30 nmol/L from a cumulative 17 standard erythema doses in a mainly UVA booth, and by about twice that from 268 doses in a mainly UVB booth. 23
In 201 New Zealand volunteers who wore UV dosimeters and then took a single whole-body dose of about 2 standard erythema doses from a lamp, 25(OH)D3 rose 1.78 nmol/L per standard erythema dose. 93
Unexposed skin of 10 adults aged 65 to 89 held as much 7-dehydrocholesterol (0.25 µg/mg) as the skin of 11 adults aged 18 to 40 (0.22 µg/mg). 40
In 21 UK adults (11 aged 18 to 40, 10 aged 65 to 89) given one below-burn solar-simulator dose of 1.3 standard erythema doses on about 35 percent of the body in winter, 25(OH)D3 rose from 44 to 48 nmol/L at 24 hours. 94
In 22 hemodialysis patients whose front legs were given UVB three times a week during dialysis for 14 weeks, 25(OH)D3 rose into a range of 70 to 78 ng/mL. 95
In 1922, seven Black children with active rickets were put in direct sun for 15 minutes a day at first, then half an hour, then an hour. 11
X-rays showed the rickets healing in every one of the seven children during the third week of sun, although some of the children had very dark skin. 11
Of 44 African American women over 70 in Waco, Texas, 86.4 percent had 25(OH)D below 32 ng/mL, and 6 to 8 weeks of spring sun brought no clinically significant rise. 96
Under cotton, wool, or polyester in black or white, a 7-dehydrocholesterol test solution made no previtamin D3 even after 40 minutes of simulated sunlight. 29