In a temperature-controlled room at the University of Edinburgh, 24 young volunteers in their underwear lay 20 centimeters from a bank of eight UVA bulbs for 22 minutes. The first time, they were covered with a metal foil "space blanket" that let the lamp's heat through and kept its ultraviolet off their skin. An hour later they lay under the same lamp without it. Their skin warmed by about a degree and a half both times. After the uncovered session the average pressure in their arteries fell by 3.5 millimeters of mercury and stayed down for 20 minutes, and the vitamin D in their blood did not change. 1
The foil session lowered the pressure too, by 2.8 millimeters, but only while the lamp was on. Donald Liu, the Edinburgh dermatologist Richard Weller, and their colleagues put the longer fall down to nitric oxide, a small molecule that relaxes the walls of blood vessels, released from a store in the skin by UVA, the longer-wave ultraviolet in sunlight. Their 2014 paper set death rates for 2008 side by side, 2.6 per 100,000 people a year from melanoma against 112.5 from cardiovascular disease. They wrote that they were "concerned that well-meaning advice to reduce the comparatively low numbers of deaths from skin cancer may inadvertently increase the risk of death" from cardiovascular disease and stroke. 1
In people, ultraviolet lamps lowered blood pressure by a few millimeters for less than an hour and turned down parts of the immune response. Five days of low lamp doses cut the redness of a tuberculin skin test, the injection that checks for immunity to tuberculosis, by about a fifth in 17 volunteers, and one dose on the lip brought back cold sores in 27 of 38 people. The one trial in people with high blood pressure treated 13 of them for two weeks and found no change in their pressure over 24 hours. 2; 3; 4
For heart disease, cancers inside the body, multiple sclerosis, bones, weight, and death, the evidence comes from latitude, season, satellites, and questionnaires, and it points in both directions. Of 29,518 Swedish women followed for 20 years, 2,545 died, and those who said they avoided the sun died at about twice the rate of the most sun-seeking. Of 346,615 older white adults in the United States followed for 12 years, 41,425 died, and those living in the sunniest quarter of census tracts had about 6 percent more deaths than those in the least sunny. Apart from lamp doses in laboratories and a few sunbathing studies, none of this work measured the light that reached anyone's skin. 5; 6
This page covers what light does to microbes, what maps and seasons say about cancer, bones, and multiple sclerosis, what sunbathing and lamps did to immunity, blood pressure, and mice on a high-fat diet, and what cohorts of sun seekers and sun avoiders found about death. Each study is named as real sun, a lamp, or a stand-in such as latitude or season, with the path its authors proposed. Vitamin D itself, its blood levels, and the vitamin D pill trials are on the Vitamin D page.
Key facts
- In the Edinburgh UVA blood pressure study, 22 minutes under a UVA lamp lowered the mean arterial pressure of 24 young volunteers by 3.5 millimeters of mercury for about 20 minutes, with no change in their vitamin D. 1
- In the Edinburgh home UVA lamp trial, 13 people with mildly high blood pressure used a UVA lamp daily for 14 days, and their pressure over 24 hours did not change. 4
- Five daily lamp doses at about seven tenths of a burn dose cut the redness of a tuberculin skin test by an average of 21.4 percent in 17 hospital staff in Sydney. 2
- Among 421,461 men and 90,141 women in US Army basic training, recruits from home areas with low ultraviolet had slightly lower odds of a stress fracture than recruits from areas with high ultraviolet. 7
- Swedish women who answered no to four questions about sun habits died at about twice the rate of the most sun-seeking women over 20 years, and at the start they were older, poorer, and more often on drugs for diabetes or the heart. 5; 8
Sunlight kept bacteria out of test tubes in 1877; in hospitals its effect has not been isolated
On 24 April 1877 Arthur Downes and Thomas Blunt filled eight thin test tubes with Pasteur's solution, a broth used to grow microbes, and cased four of them in thin sheet lead. They set all eight in a rack outside a window facing south-east, about thirty feet above the ground. By 4 May the broth in the four cased tubes had turned milky with bacteria, and the four bare tubes were still clear. When one June run began under two days of overcast sky, with "not one ray of direct sunlight," the bare tubes began to cloud as well. 9
Downes and Blunt reported to the Royal Society that a solution "may be absolutely and perfectly sterilized by sunlight," though in one tube the germs of a fungus survived an amount of sun that stopped the bacteria. In 2026 a Manchester group lit eleven species of skin bacteria in culture with a solar simulator, a lamp built to imitate sunlight. The lower dose made Staphylococcus aureus grow faster, and the study was funded in part by No.7 Beauty Company, part of the Boots Group. 9; 10
In September 1918 influenza reached the training ships in East Boston harbor, and 1,200 of 5,100 merchant sailors fell ill. The Massachusetts State Guard put up 13 tents on Corey Hill in Brookline in seven hours and called it the Camp Brooks Open Air Hospital. The medical officer, Major Thomas Harrington, had found that the worst pneumonia came from the worst-ventilated parts of the ships. In good weather the patients were carried out of their tents into "a maximum of sunshine and of fresh air day and night." Between 9 September and 12 October the camp admitted 351 sailors, and 36 of them died. 11
Anyone near a patient wore a mask of five layers of gauze on a wire frame made from a gravy strainer, replaced every two hours, and a silver ointment called Argyrol went into patients' noses twice a day. Of the 154 staff, six nurses and two orderlies caught influenza. According to one estimate that Richard Hobday and J. W. Cason cite, the regimen cut the death rate among hospital cases from 40 percent to about 13. In their 2009 retelling they wrote that "Whether the patients at Camp Brooks or other temporary hospitals were spared the worst of the influenza pandemic because they slept in the open is uncertain." 11
In 2013 Hobday and S. J. Dancer reviewed the case for sunlight and fresh air in hospitals and found "recent evidence to support historical data on the effects of natural ventilation but virtually none for sunlight." In Fayetteville, North Carolina, a respiratory therapist and a psychologist calculated how many minutes of direct sun came through the window of each room on a ventilator weaning ward, from the window's direction and the sun's path. Among 72 patients who came off the ventilator between 2008 and 2010, more direct sun went with fewer days to wean, and total daylight did not. The rooms were not assigned by chance, and nothing was measured at the bed. 12; 13
Sunnier states and summer diagnoses went with fewer cancer deaths, in studies that never measured the sun
In 1941 Frank Apperly, a pathologist at the Medical College of Virginia in Richmond, set cancer death rates in the US states and Canadian provinces against the sunshine they received. He started from a finding by a researcher named Peller from the US Navy, where the skin cancer rate was 8 times that of men the same age outside it and deaths from cancers of other organs were two-fifths of the expected number. Peller had suggested that "we should deliberately induce cancer of the skin, which grows slowly and is easily treated, by applying light rays of suitable intensity to suitable surfaces." Apperly thought the Navy's numbers "too small to warrant such conclusions." 14
Apperly worked from death records, the share of each state's people who farmed, and a solar radiation index for each state, on the reasoning that "A group of office clerks could not be expected to have as much skin cancer as a group of farmers." He found that "the general cancer rate declines with increasing solar radiation," and proposed that sunlight "may in some other way produce a relative immunity to cancer in general." In 1980 C. F. and F. C. Garland asked whether sunlight and vitamin D reduce the likelihood of colon cancer, and proposed vitamin D as the path. 14; 15; 16
In 2008 Alina Porojnicu and colleagues in Oslo followed 249,373 patients from the Cancer Registry of Norway with one of seven cancers for three years after diagnosis. Against diagnosis in winter, diagnosis in summer went with a risk of death 24 percent lower for prostate cancer, 25 percent lower for breast, 21 percent lower for colon, and 16 percent lower for Hodgkin lymphoma. Ovarian, bladder, and lung cancer showed no seasonal difference. Season stood in for sun, and the vitamin D levels in the authors' explanation were calculated from a model of skin and sunlight, never measured in a patient. 16
Lymphoma, a cancer of immune cells, was once suspected of being caused by sunlight, because it is commoner in people who have had skin cancer. When an Australian team set 31,234 non-Hodgkin and 2,418 Hodgkin lymphomas registered from 2002 to 2006 against three bands of latitude, several types ran the other way. Diffuse large B-cell lymphoma was 1.37 times as common at 37 degrees south or farther as north of 29 degrees. The gradient was "opposite that observed in Australia for cutaneous melanoma," the authors wrote, "arguing against a shared etiological role" for ultraviolet. They proposed vitamin D acting on the immune system and measured no vitamin D. 17
At Asan Medical Center in Seoul, 89.1 percent of 6,178 breast cancer patients in Seoul and 91.3 percent of 3,909 in the sunnier south survived, and once other factors were accounted for neither region nor season of diagnosis predicted survival. In Spain, 2,140 people with colorectal cancer and 3,950 without were interviewed about their sun. Frequent sun in the previous summer went with about a fifth lower odds of the cancer, a true figure anywhere from 5 to 31 percent lower, and sun ten summers earlier showed no link. The patients were interviewed after their diagnosis, and the authors noted that "there is no direct measurement of sun exposure." 18; 19
Sunbathing shifted immune cells in the blood in Sydney, the Canary Islands, and Estonia
One spring in the early 1980s, from September to November, Peter Hersey's immunology unit at Sydney Hospital asked 15 of the hospital's staff, all untanned, to sunbathe in minimal clothing for half an hour on each side of the body at midday, on 12 days over two weeks. Thirteen other staff, matched for age and sex, agreed to avoid the sun, and "With two exceptions compliance with this request was good." A meter lent by Australia's national science agency logged the ultraviolet each day. The weather kept 5 of the 15 from reaching their 12 hours. 20
In the sunbathers' blood the balance of T cells, the white blood cells that direct and carry out much of the immune response, shifted toward the suppressor type and away from the helper type. The activity of suppressor cells against antibody production rose, and in many of the sunbathers it was still raised two weeks later. The authors wrote that sunlight "may favor the induction of suppressor pathways in response to antigenic stimuli and that this may limit immune responses against tumor cells such as melanoma." 20
In a study published in 1993, the photobiology laboratory at Copenhagen University Hospital sent 22 volunteers on a week's holiday in the Canary Islands in early May and asked them "to stay in the sun as long as possible." The weather was "only moderately good," nobody burned, and afterward their lymphocytes carried more breaks in their DNA while their T-cell types did not change. 21
In Pärnu, Estonia, 16 women sunbathed through the summer of 2000 on their own schedule without sunscreen, and a spectrometer on the beach turned their diaries into doses. Each woman's summer total came to between 11 and 95 minimal erythema doses, a minimal erythema dose being the smallest dose that leaves a given person's skin just visibly red. Their natural killer cells fell from 11.1 to 8.4 percent of lymphocytes and a regulatory type of T cell rose from 9.1 to 14.7 percent, all within the normal range. Neither study had a group that stayed out of the sun, and the Estonian authors wrote that "it cannot be excluded that these were seasonal changes." 22
Ultraviolet lamps weakened the tuberculin skin test and brought back cold sores
Cold sores are caused by the herpes simplex virus, which stays in the body between outbreaks. In 1991 researchers at the US National Institutes of Health and the University of California, Los Angeles, recruited 38 people who got them again and again and shone four times each person's burn dose of mostly UVB lamp light, the short-wave ultraviolet that causes sunburn, on the part of the lip where the sores came back. Before one session the lip was painted with a sunscreen of sun protection factor 15 and before the other with a matching dummy solution, and neither patients nor doctors knew which was which. After the dummy, 27 of the 38 got a cold sore, a mean of 2.9 days later, and after the sunscreen none did, though 1 of 35 shed the virus. 3
The tuberculin test, also called the Mantoux test, checks for immunity to tuberculosis by injecting a protein from the bacterium into the skin of the forearm and reading the swelling three days later. In Sydney in 1998, Diona Damian, Gary Halliday, and their colleagues recruited 17 staff at Royal Prince Alfred Hospital, all vaccinated against tuberculosis and all positive on the test. Small patches of their lower backs were lit for five days by a fluorescent lamp filtered to imitate sunlight, and tuberculin was then injected into those patches and into skin nearby that had not been lit. At the lowest dose, about seven tenths of a burn dose a day, the redness of the reaction fell by an average of 21.4 percent, and at a full burn dose by 40.6 percent. Skin that had not been lit reacted normally. 2
Those figures came from a meter that measures redness. Clinics read the test with a ballpoint pen, drawn toward the swelling from each side until "the tip of the pen reaches the outer limit of induration" and "comes to a halt," and by the pen the lowest dose shrank the reaction by only 5.5 percent. The authors warned that "false negative Mantoux reactions might result from sun exposure prior to testing." In six volunteers who took the same dose for four or five weeks, the suppression was smaller than after five days, 17 percent against 41.6. 2
In 2008 the Sydney group reported that men were suppressed by simulated sunlight at doses "three times lower than those required to immunosuppress women." In 2003 Kuchel, Halliday, and a colleague had found that blocking the enzymes that make nitric oxide in skin protected the immune reaction of nickel-allergic volunteers under a solar simulator. That nitric oxide was made by enzymes, a different route from the store the Edinburgh group later tied to blood pressure. In a 2001 study in Utrecht, 97 of 191 volunteers took a burn dose of UVB on five days before a hepatitis B vaccine, and a skin allergy test and their natural killer cells were suppressed while antibodies to the vaccine came out the same. 23; 24; 1; 25
In 2017 a team from ISGlobal in Barcelona and King's College London stood nine fair-skinned men in underpants in a lamp cabinet for one whole-body dose of simulated sunlight, about one burn dose for their skin type. Six hours later 20 genes in their blood were turned down and one turned up, genes the authors linked to immune regulation, cancer, and blood pressure, and all were back to normal within two days. Vitamin D rose by an average of 8.8 nanomoles per liter, and a statistical test found the gene changes independent of it. The link to health outcomes, the authors wrote, "can only be hypothesized." 26
On 20 December 2016 Georgetown University Medical Center announced that sunlight "Energizes Infection Fighting T Cells," and quoted Gerard Ahern, the senior author, saying that "sunlight directly activates key immune cells by increasing their movement." The study had shone blue light on T cells in dishes, cells "isolated from mouse cell culture and from human blood," and found that they moved faster. The paper itself ended, "Although it needs to be confirmed in vivo, we propose that light may enhance the motility and migration of T cells through skin." 27; 28
Fractures peak in winter, but Army recruits from low-ultraviolet homes broke slightly fewer bones
In the 5 percent sample of US Medicare records from July 1986 to June 1990, Bischoff-Ferrari and colleagues counted 69,857 fractures of the hip, wrist, upper arm, and ankle in people over 65, and all four were commonest in winter and rarest in summer. Against weather records, each 20 inches of winter snow went with 5 percent fewer hip fractures and 6 to 12 percent more of the others. In sunny autumn weather hip fractures fell, by about 6 percent for every 14 sunny days, and wrist, arm, and ankle fractures rose by 10 to 30 percent. The authors noted that hip fractures happen mostly indoors among frail people, and the others among active people who go out. 29
From 1997 to 2007, 421,461 men and 90,141 women went through US Army basic training, and Montain and colleagues looked up the average UV index, the standard daily scale of the sun's ultraviolet strength, at each recruit's home of record. "Contrary to the hypothesized effect," recruits from low-UV homes had slightly fewer stress fractures, with 8 percent lower odds for men, a true figure anywhere from 3 to 13 percent lower, and 11 percent lower odds for women. Black recruits had fewer stress fractures than others, including those from low-UV areas. Stress fractures struck 18.8 of every 1,000 men and 79.6 of every 1,000 women. 7
In Catalonia, 90,149 hip fractures in people 65 and older from 2010 to 2019 peaked in autumn and winter, and of all the weather measures only hours of sunshine consistently tracked the monthly rate. In Rijeka, Croatia, 398 men aged 70 to 80 on hormone therapy for prostate cancer, none taking vitamin D, reported how many hours a week they spent in the sun. Of the 224 who reported under three hours, 49 had osteoporotic fractures, and of the 174 who reported over ten, 19 did. The paper reports no adjustment for general health or physical activity. 30; 31
From 2003 Sato and colleagues published randomized trials in which sunlight raised bone density and cut hip fractures, among them 129 stroke patients given regular sun, of whom 1 broke a hip, against 6 of 129 kept out of it. In 2016 Bolland and colleagues at the University of Auckland analyzed all 33 of the group's trials and found the treatment groups "improbably similar" and the hip fracture reductions very large "regardless of intervention." The stroke trial was retracted in 2016, and a trial in women with Alzheimer's disease in 2015. 32; 33; 34; 35; 36
Multiple sclerosis is commoner farther from the equator; the one ultraviolet trial enrolled 20 people
Multiple sclerosis is a disease in which the immune system attacks the insulating sheath around nerves in the brain and spinal cord. A 2011 meta-analysis, a study that pools the results of earlier studies, gathered 650 prevalence estimates and found the disease commoner at higher latitudes everywhere except in Italy and northern Scandinavia. A 2019 update with 880 estimates put the gradient at about 4.3 more cases per 100,000 people for each degree of latitude, after adjusting for age. The authors of the first named "ultraviolet radiation (UVR)/vitamin D" as "the most prominent candidates." 37; 38
From November 2003 to December 2006 the Ausimmune study, led by Robyn Lucas at the Australian National University, enrolled 216 adults at four Australian centers from 27 to 43 degrees south who had just had a first attack of the kind that often comes before multiple sclerosis, and 395 people matched to them. Each person's leisure time in the sun since age six was reconstructed by questionnaire, the sun damage on their skin was graded, and their blood vitamin D was measured. For each 1,000 kilojoules per square meter of accumulated leisure sun, the odds of a first attack were 30 percent lower, a true figure anywhere from 6 to 47 percent lower, and sun and vitamin D each held up with the other taken into account. Both, the authors wrote, "will need to be evaluated in clinical trials." 39
In a Mendelian randomization study, which uses gene variants that raise or lower vitamin D as a natural experiment, people whose genes lowered their blood vitamin D by one standard deviation had twice the odds of multiple sclerosis, a true figure anywhere from 1.7 to 2.5 times. In 2022 a study of more than 10 million young adults on active duty in the US military found that the risk of multiple sclerosis rose 32-fold after infection with the Epstein-Barr virus, and its authors called the virus "the leading cause of MS." 40; 41
From November 2014 Prue Hart's group at the Telethon Kids Institute in Perth enrolled people within four months of a first attack, and 10 were assigned by chance to 24 sessions of narrowband UVB lamp treatment over eight weeks while 10 had none. Everyone's blood vitamin D was held above 80 nanomoles per liter, with pills for the seven who started below it, so that the lamps were tested apart from vitamin D. After 12 months all 10 without the lamp and 7 of the 10 with it had gone on to multiple sclerosis, a difference too small to rule out chance. The trial stopped in March 2017 for lack of recruits. 42
In Madison, Wisconsin, Hector DeLuca's biochemistry group was treating mice with experimental autoimmune encephalomyelitis, or EAE, a laboratory model of multiple sclerosis, using a narrowband UVB lamp that earlier work had shown suppresses the disease without raising blood vitamin D. They put sunscreen on the mice to block the lamp, and "Quite unexpectedly, some sunblock preparations themselves completely prevented EAE without UV radiation." Four of six brands worked, in groups of 12 mice, and the paper traced the effect to two salicylate sunscreen chemicals, homosalate and octisalate, whose presence, it said, "exactly correlated" with effectiveness. Its own Table 1 lists Blue Lizard, one of the four brands that worked, as octinoxate, octocrylene, oxybenzone, and zinc oxide, with no salicylate. The Wisconsin Alumni Research Foundation paid for the work. 43
On 17 July 1988 a 35-year-old woman with multiple sclerosis, sunbathed alone for several hours in 95 degree Fahrenheit weather and was dead when the rescue squad arrived; three years earlier she had been disabled by heat while sunbathing and "admonished never to sunbathe alone." In 2017 a single author at Rider University in New Jersey set death rates from multiple sclerosis in the 48 contiguous states against their weather. The rates tracked temperature closely, with a correlation of −0.812, and did not track hours of sunshine or solar radiation. The author credited long-term acclimation to heat. 44; 45
UVA lamps lowered blood pressure for up to an hour in two laboratories and not in a third
In 2009 Opländer, Suschek, and colleagues reported that whole-body UVA lowered the systolic and diastolic blood pressure of healthy volunteers by 11 percent at 30 minutes, an effect that lasted up to an hour. In pieces of human skin, UVA freed nitric oxide from compounds stored in the skin. 46
In the Edinburgh study of 2014, 12 young men had a drug infused into the artery of one arm to block the enzymes that make nitric oxide, and UVA still raised the blood flow in that forearm by 23.7 percent, which the authors took to mean that the nitric oxide came from a store already in the skin. Blood nitrite rose and nitrate fell after the lamp. The authors put their dose at the equivalent of "30 minutes at noon on a sunny day in Southern Europe" and named their own limit, that "all interventions were acute and carried out in healthy normotensive young individuals." The foil session always came first, and the order was never switched. 1
In the summer of 2013, near Glasgow, nine trained male cyclists rode 16.1 kilometers against the clock after the same UVA lamp or a foil sham, and after nitrate gels, which Science in Sport supplied free, or a peach-flavored dummy gel. UVA alone gave 1,450 seconds against 1,469 in the control ride, which the authors did not count as a real change. The summer turned out unusually warm, and the cyclists' sun outside the laboratory was neither controlled nor measured. 47
In 2018 a group including Weller and Martin Feelisch of Southampton gave 10 healthy men no light, 15 minutes, or 30 minutes of UVA. The larger dose raised nitrite in the blood, and blood pressure did not change at either dose. 48
Two weeks of home UVA did not lower 24-hour blood pressure, and UK Biobank linked outdoor hours with more high blood pressure
In 2020 Weller's group took the blood pressures recorded before dialysis for 342,457 patients at 2,178 US dialysis clinics over three years and set them against the ultraviolet and temperature at each clinic, from National Oceanic and Atmospheric Administration records. Systolic pressure fell as UVA, UVB, and temperature rose, and the link with ultraviolet held after allowing for temperature. Nobody measured the sun on any patient, and three of the authors, from Fresenius Medical Care, held stock in the company. 49
For a trial, the group aimed to recruit 80 people in Scotland with untreated, mildly high blood pressure, enough to detect a fall of 2 millimeters over 24 hours, and enrolled 13, with a mean age of 61. Each lit their whole body at home with a UVA lamp for about 10 minutes a side, daily for 14 days, and for another 14 days with a sham lamp filtered to pass nothing shorter than 500 nanometers, in an order set by chance. Pressure measured over 24 hours did not change. Pressure measured in the clinic within 90 minutes of the last session fell by 8.0 over 3.8 millimeters with the UVA lamp and did not fall with the sham. 4
"We had unanticipated difficulty recruiting subjects, largely because of the problems of delivering lamps to subjects' homes," the authors wrote. The British Heart Foundation paid for the trial, and Weller declared that he "has co-founded a company developing Sunscreens (Dr Weller Ltd)." The authors suggested that more frequent UVA, or UVB, might work. 4
In a 2024 preprint, a paper not yet reviewed by other scientists, a team at Xiamen Cardiovascular Hospital in China used UK Biobank data to follow more than 300,000 people for a median of 13.4 years. Each extra hour a person said they spent in outdoor light on a typical summer day went with 1.1 percent higher risk of developing high blood pressure and 1.2 percent lower risk of low blood pressure. "Contrary to previous findings," the authors wrote, their study "revealed a positive association" between outdoor light and blood pressure, and they estimated that "curtailing light exposure within one hour could potentially avert approximately 45.76 million cases of hypertension" worldwide. The exposure was a single questionnaire answer. 50
In southern Sweden, among 23,593 women who described their sun habits in 2000, those with the fewest had 41 percent higher odds of taking blood pressure drugs in 2005 to 2007 than those with the most, in a score that counted tanning beds as sun. 51
At a residential college in South India, 20 young adults were wrapped in banana leaves in the morning sun for 30 minutes, "with a small slit on the leaves above the nostrils to facilitate breathing," and their systolic pressure rose from 105.8 to 114.0. The European Society of Hypertension calls lower blood pressure in summer "a global phenomenon" and ties it to temperature. A 2026 review of 14 studies found that most reported a fall in blood pressure with ultraviolet and called for large trials before any ultraviolet treatment is recommended. 52; 53; 54
Simulated sunlight inactivated the COVID-19 virus in minutes, and counties with more UVA had fewer COVID-19 deaths
In 2020 a US Department of Homeland Security laboratory in Maryland dried SARS-CoV-2, the virus that causes COVID-19, onto stainless steel and put it under a solar simulator set to the summer solstice at 40 degrees north. Ninety percent of the infectious virus was inactivated every 6.8 minutes in simulated saliva. In droplets suspended in air, 90 percent was lost in 8 minutes under summer light and in 286 minutes in the dark. 55; 56
In 2021 Weller's group, with geographers at Edinburgh, took 2,474 US counties where, from January to April 2020, the UVB was too weak for skin to make vitamin D, so that any link would not run through vitamin D, and set satellite estimates of UVA against COVID-19 deaths. Deaths fell by 29 percent for each 100 kilojoules per square meter of mean daily UVA, a true figure anywhere from 15 to 40 percent, and analyses of 6,724 small areas of England and of Italian municipalities gave results in the same direction. The authors wrote that "Ambient UVA is likely to be a poor measure of individual-level UVA level exposure" and proposed nitric oxide released from skin as the path. Weller declared that he was "a director of Dr Weller Ltd (sunscreen development)." 57
Ultraviolet slowed weight gain in mice on a high-fat diet; we found no trial of sunlight and weight in people
In 2014 Geldenhuys, Gorman, and colleagues at the Telethon Kids Institute in Perth, with Feelisch and Weller among the authors, fed male mice a high-fat diet and gave them repeated ultraviolet lamp doses, some below the burn threshold and some at it. The irradiated mice gained less weight and showed less glucose intolerance, insulin resistance, and fatty liver. "Many of the benefits of UVR were not reproduced by vitamin D supplementation," the authors wrote, and raising nitric oxide in the skin reproduced many of them. 58
In 2017 the same group gave overweight mice the same UVB dose from two lamps. One lamp, 65 percent UVB, slowed weight gain, and the other, 96 percent UVA, did not, though it lowered LDL cholesterol, the kind tied to heart disease. In 2025 a pilot study from the UK Health Security Agency, whose title says that ultraviolet "reduces weight gain" in mice on a high-fat diet, reported in its results that "Non-significant trends are observed for weight amelioration." 59; 60
In people, a 2019 review by Gorman, de Courten, and Lucas found 25 studies of ultraviolet and metabolism, most with 100 people or fewer, most in middle-aged patients being treated for psoriasis, and most lasting 12 weeks or less. Few measured blood sugar. In a cohort of women in southern Sweden, those with the fewest sun habits had 2.47 times the odds of type 2 diabetes at 11 years, a link the authors suggested might run "possibly via nitric oxide." 61; 62
Swedish women who avoided the sun died at about twice the rate of sun seekers; a US cohort found more deaths with more ultraviolet
From 1990 to 1992 a study of melanoma in southern Sweden sent a questionnaire to women aged 25 to 64 drawn at random from the population register, and 29,518 answered, 74 percent. Four of its questions asked about sun: did they sunbathe in summer, sunbathe in winter such as on mountain holidays, use tanning beds, or travel abroad to swim and sunbathe. Twenty years later P. G. Lindqvist, of the obstetrics and gynecology department at the Karolinska University Hospital, and colleagues counted 2,545 deaths. The 1,721 women who answered no to all four questions died at "approximately twofold" the rate of the women with the most sun habits. The authors concluded that very restrictive sun advice "might in fact be harmful to women's health." 5; 8
In 2016 they wrote that nonsmokers who avoided the sun had "a life expectancy similar to smokers in the highest sun exposure group." At the start, 58.7 percent of the avoiders were aged 55 to 64, against 13.9 percent of the most sun-seeking; 45.7 percent had a low income, against 14.8; 43.1 percent had nine years of school or less, against 10.9; and 20.4 percent took drugs for diabetes, blood clotting, or the heart, against 6.7. The sun seekers smoked more, 42.5 percent against 26.9. 8
The authors adjusted for these differences and wrote that "it is not possible to differentiate between active sun exposure habits and a healthy lifestyle, and secondly, the results are of an observational nature; therefore, a causal link cannot be proven." The questionnaire did not ask about exercise until the second round in 2000. 8
Among 346,615 white adults aged 50 to 71 in the NIH-AARP Diet and Health Study, the ultraviolet over each person's census tract was read from NASA satellite data, and over 12 years 41,425 of them died. Those in the sunniest quarter had 6 percent more deaths than those in the least sunny, a true figure anywhere from 3 to 9 percent more. The authors concluded that ultraviolet exposure "might not be beneficial for longevity." Among 8,472 white adults whose skin was examined for sun damage in the first National Health and Nutrition Examination Survey from 1971, those with severe damage had 45 percent higher death rates by 1992, a true figure anywhere from 22 to 72 percent higher. 6; 63
Among 38,472 Swedish women, those who took sunbathing holidays more than once a year for three decades had lower death rates, and those who used a solarium once a month or more for a decade had higher ones. In UK Biobank, Weller's group reported in 2024 that solarium users had lower death rates than non-users, and a correction to the paper was published in 2025. Among 83,205 members of the Adventist Health Study 2 in North America, two hours a day outdoors in the warmer months went with 10 percent fewer deaths than half an hour, and no difference in cancer deaths. 64; 65; 66; 67
In 2025 a systematic review from the University of Bristol, commissioned by the UK's National Institute for Health and Care Research, gathered 73 articles on sunlight and death. It found results "subject to high risk of bias, largely due to the likelihood of uncontrolled confounding and the use of indirect measures of sunlight exposure," and concluded that they "do not provide a strong rationale for changes to sun protection guidance." A figure often quoted the other way, that too little sun may be responsible for 340,000 deaths a year in the United States and 480,000 in Europe, comes from a 2020 review whose authors' meeting was paid for by the Sunshine Health Foundation. Its source was a 2014 meta-analysis that attributed 12.8 percent of US deaths to low blood vitamin D, and it measured no sunlight. 68; 69; 70
The World Health Organization names vitamin D as the only benefit of ultraviolet; an Australian panel called others plausible
"The only beneficial effect of UV, triggered by UV-B, is the stimulation of the production of vitamin D," the World Health Organization's page of questions and answers on ultraviolet read in July 2024. The same page gives "the recurrent eruption of cold sores" as an example of the immune suppression that ultraviolet causes. 71
In March 2021 the Australian Skin and Skin Cancer Research Centre convened a two-day Sun Exposure Summit to review the country's advice on sun, and its workshop asked, "Should the statement give advice about sun exposure for the non-vitamin D benefits? If so, what should this be?" The statement, published in 2024, says that ultraviolet "may have benefits independent of vitamin D through mechanisms such as immune system modulation and the release of nitric oxide," and that it is "plausible, although not yet well established" that blood vitamin D marks having had enough ultraviolet to get them. The panel "considered that the evidence for these benefits is limited but, following the precautionary principle, recommended that people at intermediate or low risk of skin cancer obtain sufficient controlled exposure to UV radiation to maintain adequate vitamin D status." 72
Misconceptions
What people believe
- Sunlight boosts the immune system.
- Sunlight lowers blood pressure.
- Avoiding the sun is as dangerous as smoking.
- Trials have shown that sunlight prevents hip fractures.
- Sitting in the sun helps you lose weight.
- Because sunlight kills germs, it can cure a fungal infection or the flu.
What the studies show
- Ultraviolet turned parts of the immune response down: under lamps it weakened the tuberculin skin test and brought back cold sores in 27 of 38 people. The claim that sunlight "energizes" T cells came from blue light on cells in dishes. 2; 3; 27
- UVA lamps lowered blood pressure for up to an hour in healthy volunteers, and two weeks of home UVA did not lower 24-hour pressure in 13 people with mildly high blood pressure. 1; 4
- The Swedish sun avoiders were older, poorer, and sicker at the start than the sun seekers, and a US cohort of 346,615 found slightly more deaths with more ultraviolet. 8; 6
- Two sunlight and hip fracture trials by Sato's group were retracted after an analysis of the group's trials found the treatment groups "improbably similar." 34; 35; 36
- Ultraviolet slowed weight gain in mice on a high-fat diet, and we found no trial of sunlight and weight in people. 58; 61
- Simulated sunlight inactivated the COVID-19 virus on steel and in air within minutes, and ultraviolet on the lip reactivated the cold sore virus. We found no trial of sunlight for a fungal or bacterial skin infection. 55; 56; 3
Questions people ask
Can sunlight kill fungus or bacteria?
Sunlight stopped bacteria growing in glass tubes in 1877, and simulated sunlight inactivated the COVID-19 virus on steel in minutes. We found no trial of sunlight on a fungal infection of the skin or nails, and in one culture study a low lamp dose made Staphylococcus aureus grow faster. 9; 55; 10
Does sunlight lower blood pressure?
Under UVA lamps, blood pressure fell by a few millimeters for up to an hour in healthy volunteers. In 13 people with mildly high blood pressure, two weeks of daily home UVA did not change pressure over 24 hours. 1; 46; 4
Can the sun raise your blood pressure?
In a UK Biobank preprint, each extra hour a day that people said they spent outdoors in summer went with a 1.1 percent higher risk of high blood pressure. Twenty adults wrapped in banana leaves in the morning sun had a rise in systolic pressure from 105.8 to 114.0. 50; 52
Does the sun boost your immune system?
Ultraviolet turned parts of the immune response down. Two weeks of sunbathing shifted blood T cells toward the suppressor type in 15 Sydney hospital staff, and lamp doses weakened the tuberculin skin test and reactivated cold sores. 20; 2; 3
Is sunlight good for your bones?
Fractures in older people in the United States peak in winter, US Army recruits from low-UV home areas had slightly fewer stress fractures, and the best-known sunlight and fracture trials were retracted. 29; 7; 34
Is sunlight good for multiple sclerosis?
Multiple sclerosis is commoner farther from the equator, and in the Ausimmune study more leisure sun went with lower odds of a first attack. The one ultraviolet lamp trial enrolled 20 people and could not rule out chance. 38; 39; 42
Can sitting in the sun help you lose weight?
Ultraviolet lamps slowed weight gain in male mice on a high-fat diet, and a later pilot in mice found only non-significant trends. We found no trial of sunlight and weight in people. 58; 60; 61
Is avoiding the sun as bad as smoking?
That comparison comes from one Swedish cohort, whose sun avoiders were older, poorer, and more often on heart or diabetes drugs at the start. A US cohort of 346,615 older adults found 6 percent more deaths under the most ultraviolet. 8; 6
What we do not know
Nobody knows whether regular sunlight, over months or years, does anything for health past the skin that vitamin D does not. The trials of ultraviolet itself are small and short: 13 people for two weeks for blood pressure, 20 people for eight weeks for multiple sclerosis. No trial has assigned real sunlight to people and followed them for heart disease, cancer, infection, or death. 4; 42
Whether the nitric oxide that UVA frees from skin matters over months is unknown. The lamp studies followed it for an hour, and we found no study that measured nitric oxide, vitamin D, and a health outcome in the same people under real sun. 1
Whether the immune suppression that ultraviolet causes changes the course of infections in people is unknown. The Dutch vaccine study used vaccination as a stand-in because, in its authors' words, the effects on infections "cannot be determined by experimentation on humans." 25
Whether sun seekers live longer because of the sun or because of what sun seekers are like is unsettled. The cohorts disagree in direction, and the 2025 review found their results at high risk of bias from confounding and from indirect measures of sun. 68
For fungal infections of skin and nails, wound healing, and weight in people, we found no trial of sunlight.
Key studies
- Downes and Blunt, 1877, Royal Society of London. Bacteria grew in broth in lead-cased test tubes on a window ledge and not in the bare tubes beside them. 9
- Hobday and Cason, 2009, Boston. Camp Brooks Open Air Hospital admitted 351 sailors with influenza in 1918, and 36 died. 11
- Apperly, 1941, North America. Cancer death rates fell with state solar radiation indexes, on death records alone. 14
- Hersey and colleagues, 1983, Sydney. Two weeks of real-sun sunbathing shifted blood T cells toward the suppressor type in 15 hospital staff. 20
- Rooney and colleagues, 1991, Bethesda and Los Angeles. A UVB lamp on the lip brought back cold sores in 27 of 38 people, and none after sunscreen. 3
- Damian and colleagues, 1998, Sydney. Five days of lamp doses below the burn threshold cut the tuberculin reaction by 21.4 percent in 17 volunteers. 2
- Bischoff-Ferrari and colleagues, 2007, United States. 69,857 Medicare fractures peaked in winter, and sunny weather went with fewer hip fractures and more of the others. 29
- Lucas and colleagues, 2011, Australia. More leisure sun and more vitamin D each went with lower odds of a first attack of the kind that precedes multiple sclerosis. 39
- Montain and colleagues, 2013, United States. Army recruits from low-UV homes had slightly fewer stress fractures. 7
- van Leeuwen, Vajdic, and colleagues, 2013, Australia. Several lymphomas were commoner farther from the equator. 17
- Lin and colleagues, 2013, United States. More satellite-measured ultraviolet over home went with 6 percent more deaths among 346,615 older adults. 6
- Liu, Weller, and colleagues, 2014, Edinburgh. A UVA lamp lowered mean arterial pressure by 3.5 millimeters for about 20 minutes, with no change in vitamin D. 1
- Lindqvist and colleagues, 2014 and 2016, southern Sweden. Women who avoided the sun died at about twice the rate of sun seekers, and were older, poorer, and sicker at the start. 5; 8
- Geldenhuys and colleagues, 2014, Perth. Ultraviolet slowed weight gain in male mice on a high-fat diet, and vitamin D did not reproduce most of it. 58
- Wang, DeLuca, and colleagues, 2017, Madison. Sunscreen alone blocked a mouse model of multiple sclerosis, including a brand whose listed ingredients contain no salicylate. 43
- Hart and colleagues, 2018, Perth. In the PhoCIS trial 10 of 10 untreated and 7 of 10 lamp-treated people went on to multiple sclerosis. 42
- Cherrie, Weller, and colleagues, 2021, United States, England, and Italy. More satellite UVA went with fewer COVID-19 deaths in areas too dim for vitamin D synthesis. 57
- Weller and colleagues, 2023, Scotland. Two weeks of home UVA did not change 24-hour blood pressure in 13 people. 4
- Li and colleagues, 2024, UK Biobank. More reported summer hours outdoors went with slightly more high blood pressure. 50
- Parkhouse and colleagues, 2025. A review of 73 articles on sunlight and death found the evidence mixed and at high risk of bias. 68
Definitions
- Case-control study. A study that compares people who have a disease with people who do not and asks about their past exposures.
- Cohort study. A study that records exposures in a group of people and follows them to see who develops a disease or dies.
- Confidence interval. The range of values within which the true effect probably lies, given the size of the study and the play of chance.
- Ecological study. A study that compares whole places, such as counties or states, instead of individual people.
- Experimental autoimmune encephalomyelitis. A disease induced in laboratory mice as a model of multiple sclerosis.
- Immunosuppression. A weakening of the immune response. Ultraviolet causes it in the skin it reaches, and after larger doses sometimes in skin it has not reached.
- Langerhans cell. An immune cell in the outer layer of the skin that picks up foreign material and carries it to the lymph nodes. Ultraviolet reduces their number.
- Mean arterial pressure. The average pressure in the arteries over a heartbeat, between the systolic and diastolic readings.
- Mendelian randomization. A study that uses gene variants that raise or lower a trait, such as vitamin D, as a natural experiment.
- Meta-analysis. A study that pools the results of earlier studies into one estimate.
- Minimal erythema dose. The smallest dose of ultraviolet that leaves a given person's skin just visibly red a day later.
- Multiple sclerosis. A disease in which the immune system attacks the sheath around nerves in the brain and spinal cord.
- Natural killer cell. A white blood cell that kills infected and cancerous cells.
- Nitric oxide. A small molecule that relaxes the walls of blood vessels. Skin stores forms of it that UVA can release.
- Odds ratio, relative risk, and hazard ratio. Ways of comparing how often something happens in one group against another. A value of 1 means no difference, 0.8 means about a fifth lower, and 1.5 means half again as high.
- Preprint. A paper posted publicly before other scientists have reviewed it.
- Randomized trial. A study that assigns people to groups by chance, so that the groups differ only in what the researchers gave them.
- Solar simulator. A lamp built to give off a spectrum close to that of sunlight.
- Standard erythema dose. A fixed amount of ultraviolet, weighted by how strongly each wavelength reddens skin, and the same for everyone.
- T cell. A white blood cell that directs or carries out much of the immune response. Helper T cells promote responses, and suppressor or regulatory T cells hold them back.
- Tuberculin skin test. A test, also called the Mantoux test, that injects a protein from the tuberculosis bacterium into the skin and reads the reaction three days later.
- UVA. Ultraviolet light from 315 to 400 nanometers. Most of the ultraviolet reaching the ground; it penetrates deeper into skin and burns only in large doses.
- UVB. Ultraviolet light from 280 to 315 nanometers. It causes most sunburn and drives vitamin D production in the skin.
Statistics
In 24 healthy volunteers given whole-body UVA irradiation (2 standard erythemal doses), mean arterial pressure fell by 3.50 mmHg on average during and for 20 minutes after the exposure. 1
In the same 24 volunteers, diastolic blood pressure fell by 4.90 mmHg during active UVA irradiation and stayed down for 30 minutes, against a non-significant 2.79 mmHg fall under sham light. 1
After UVA irradiation, volunteers kept on a low-nitrate diet for 2 days had blood nitrite rise from 0.50 to 0.72 micromoles per litre and stay there for 40 minutes. 1
In the same experiment, blood nitrate fell from 11.79 to 8.99 micromoles per litre after active UVA, and fell only after active exposure, not sham. 1
With the body's own nitric oxide enzymes blocked by an infusion, UVA on that forearm still raised its blood flow 23.7 percent above its blocked baseline 30 minutes later. 1
The Edinburgh team calculated, from the solar spectrum and nitrite chemistry, that UVA accounts for more than 80 percent of the nitric oxide sunlight releases from nitrite. 1
Nine trained cyclists rode a 16.1 km time trial in 1,447 seconds after nitrate gels plus UVA light, against 1,469 seconds in the control condition. 47
In the same nine cyclists, UVA light without nitrate gave a 1,450-second time trial, not significantly faster than control. 47
After UVA alone, the cyclists' blood nitrite was 282 nanomoles per litre against 215 in the control condition, a small, non-significant difference; nitrate gels plus UVA gave 456. 47
In 12 day-shift nurses measured on a ward with windowed nursing stations and on one without, mean arterial pressure was 7.65 mmHg lower on the windowed, daylit ward (86.9 against 94.6 mmHg). 73
In 20 adults wrapped in banana leaves and left in morning sun for 30 minutes, systolic blood pressure rose from 105.80 to 114.03 mmHg. 52
In the same 20 banana-leaf sunbathers, pulse fell from 74.93 to 71.08 beats per minute and diastolic pressure did not change significantly (66.84 to 69.22 mmHg). 52
Six hours after nine healthy men took 2.5 to 3 standard erythemal doses of whole-body simulated sunlight from fluorescent lamps, 20 genes in their blood were turned down and one turned up; all were back to baseline by 24 or 48 hours. 26
Six healthy adults who lay in real sun for 4 hours on 14 May 1985, their first sun of the season, in bathing suits covering about 10 percent of the body, all developed erythema. 74
In the same six sunbathers, urine noradrenaline in the 6 hours after the sunbath was higher than on the two control days (p < 0.01), while adrenaline did not change. 74
Cut scalp hair from nine volunteers, left in real sun for 40 hours in total, lost a median 54 percent of its cortisol against matched hair kept in the dark; cortisol fell in all 9 samples. 75
In the same sun-exposed hair, cortisone fell in 8 of 9 samples, by a median 32 percent. 75
The share of blood lymphocytes that were T cells went from 76 to 72 percent in the sunbathers over their two weeks of sun and back to 75 percent two weeks later; in sun-avoiding controls it went 72, 73, then 79. 20
Of the sunbathing volunteers, 5 could not total the planned 12 hours of sun because of the weather. 20
Sixteen women sunbathing on their own schedule at Pärnu beach, Estonia (58°N), from May to August 2000 gathered individual real-sun doses of 11 to 95 minimal erythema doses over the summer. 22
Across the Pärnu summer of sunbathing, natural killer cells fell from 11.1 to 8.4 percent of the 16 women's blood lymphocytes, and from 266 to 197 cells per microlitre. 22
In the same 16 Pärnu sunbathers, CD4+CD25+ T cells, a type with regulatory properties, rose from 9.1 to 14.7 percent of blood lymphocytes over the summer; the authors note all changes stayed within the normal range. 22
During the Pärnu study summer, the midday UV index averaged 3.2 and peaked at 5.2. 22
In a 160-person trial, volunteers sensitized to the chemical DNCB on skin given 3 minimal erythema doses of simulated sunlight 3 days earlier became sensitized 50 percent of the time, against 95 percent without the exposure. 76
In 17 tuberculin-positive volunteers, five daily doses of simulated sunlight from a fluorescent lamp at 95 mJ per cm2 cut the redness of a later tuberculin skin test on that skin by 21.4 percent on average. 2
At a daily dose equal to each person's burn threshold, the tuberculin test reaction was cut by 40.6 percent on average in 13 subjects, and by 47.6 percent at 150 mJ per cm2. 2
In six subjects given the same low daily lamp dose for 5 days on one side of the back and 4 to 5 weeks on the other, the short course suppressed the tuberculin reaction by 41.6 percent and the long course by only 17 percent. 2
In the tuberculin study, age and burn threshold did not predict how much a low lamp dose suppressed immunity (r = 0.01 and 0.003), and darker-skinned subjects were suppressed as much as paler ones. 2
Healthy people in four groups of about 34 took whole-body simulated sunlight at 1.2 standard erythema doses a day for 2, 10, or 30 days, or none; only the 30-day group showed a significantly weaker allergic response to a test chemical applied to exposed skin. 77
In nickel-allergic volunteers, four daily below-burn doses of simulated sunlight, each about 6 minutes of midday summer sun, suppressed the nickel allergy reaction by about 34 percent. 78
In the same experiment, UVA alone suppressed the nickel reaction most after 3 daily doses and then faded to non-significant levels by the 4th and 5th doses, while simulated sunlight stayed suppressive through 5 days. 78
A single 3-minimal-erythema-dose exposure to simulated sunlight on a small patch of skin completely blocked a new contact allergy at that patch in 12 of 12 volunteers, and at an unexposed site elsewhere on the body in 10 of 12. 79
Of 18 black-skinned volunteers given four daily UVB doses of 144 mJ/cm2 to buttock skin, only 9 showed contact allergy when challenged 30 days later. 80
In volunteers given simulated sunlight before a tuberculin skin test, the dose needed to suppress the immune reaction was at least three times lower in men than in women, and at the highest dose suppression was 40 percent greater in men. 23
In the same 30 volunteers, a week of oral nicotinamide at 500 mg three times a day reduced that sun-lamp suppression to 7, 14, and 33 percent. 81
In 33 nickel-allergic women taking a control oil, simulated sunlight at 7.6 J/cm2 cut their nickel allergy reaction by 43 percent. 82
In women who took 5 g a day of EPA-rich fish oil for 3 months, the same 7.6 J/cm2 dose cut the nickel reaction by 40 percent, slightly less than in the control group. 82
After 1.25 times their burn dose of simulated sunlight, Langerhans cells, the skin's immune sentinels, fell by an average 32 percent in 34 basal cell carcinoma patients and by 32 percent in 21 healthy volunteers. 83
At the same 1.25-burn-dose exposure, the loss of Langerhans cells ranged from 0 to 74 percent between individuals. 83
Seventy-two hours after a single exposure of skin to twice the burn dose of simulated sunlight, Langerhans cell density fell by 45 percent measured with one marker and 42 percent with another. 84
In unexposed buttock skin of these volunteers, Langerhans cell density ranged from 443 to 868 cells per square millimetre. 84
In volunteers given single doses of simulated sunlight, Langerhans cells in the exposed epidermis fell by 34 percent after 2 minimal erythema doses and by 48 percent after 3. 85
The same 3-burn-dose total of simulated sunlight cut Langerhans cells by 30 percent when given as 4 doses of 0.75 or 1 dose of 3, and by 50 percent when given as 2 doses of 1.5. 85
Even at half a burn dose of simulated sunlight, the Langerhans cells that remained in the epidermis were 1.7 times their normal size. 85
When 38 people with recurrent cold sores had four minimal erythema doses of UVB-rich lamp light shone on the lip where outbreaks recurred, 27 (71 percent) developed a cold sore after a placebo lip solution, a mean 2.9 days later. 3
With sunscreen on the lip before the same lamp exposure, none of the patients developed a cold sore, though 1 of 35 shed herpes virus at the exposed site. 3
Of 34 people with a history of recurrent cold sores given three minimal erythema doses from a solar simulator, 13 (38 percent) developed a cold sore. 86
Twelve HIV-infected patients and 12 healthy controls had nearly the same burn threshold under a solar simulator, 69.25 against 70.25 mJ/cm2. 87
Among long-term ventilator patients in one hospital, the 41 whose rooms got more than the average calculated minutes of direct sun through the window took a mean 55.56 days to come off the ventilator (median 40), against 101.06 days (median 51) for the 31 with less. 13
In the same 72 patients, minutes of direct sunlight in the room correlated with fewer days to wean (r = -0.251), while total daylight did not (r = 0.08). 13
Across 2,474 US counties where sunlight was too weak for skin to make vitamin D at any time in the study period, the COVID-19 death rate fell by 29 percent for each 100 kJ per square metre more mean daily UVA. 57
Replicating the analysis in England and Italy gave COVID-19 mortality rate ratios of 0.51 and 0.81 per 100 kJ per square metre more daily UVA, and a pooled 32 percent fall across the three countries. 57
From January to April 2020, mean daily UVA averaged 696 kJ per square metre across US counties and 412 across England's small areas. 57
In Chile, hospital admissions for inflammatory bowel disease ran at 9.43 per 100,000 people in the far south (latitudes 48.9 to 56°S) against 4.06 in the far north (17.5 to 25.9°S). 88
Chile recorded 12,869 hospital admissions for inflammatory bowel disease between 2001 and 2012, a national rate of 6.52 per 100,000 inhabitants. 88
Among 46,205 Norwegian men diagnosed with prostate cancer from 1964 to 1992, those diagnosed in summer or autumn had a relative risk of death within 36 months of 0.8 against those diagnosed in winter or spring; season stands in for sun exposure and blood vitamin D. 89
Of 249,373 Norwegian cancer patients followed for 3 years, those diagnosed in summer had relative risks of death of 0.76 for prostate cancer, 0.75 for breast, 0.79 for colon, and 0.84 for Hodgkin lymphoma, against winter diagnosis; ovarian, bladder, and lung cancer showed no seasonal difference. 16
In a Spanish case-control study of 2,140 colorectal cancer patients and 3,950 controls, people who reported frequent or very frequent sun in the previous summer had an odds ratio of 0.81 for colorectal cancer; sun reported for the summer 10 years earlier showed no such link. 19
In the same Spanish study, working outdoors as a farmer or mason went with an odds ratio of 0.85 for colorectal cancer, and frequent sunburn over 30 years with 0.87, neither quite significant. 19
Among 34,712 lymphoid cancers registered in Australia from 2002 to 2006, diffuse large B-cell lymphoma was 1.37 times as common at 37°S or further south as north of 29°S, with latitude standing in for ambient UV. 17
In the same Australian data, plasmacytoma was 1.52 times and plasma cell myeloma 1.15 times as common in the southernmost band as in the northernmost. 17
Two forms of Hodgkin lymphoma, mixed cellularity or lymphocyte-depleted and nodular sclerosis, were 1.60 and 1.57 times as common in Australia's southernmost latitude band as in its northernmost. 17
Of Korean breast cancer patients, 89.1 percent (6,178) in Seoul and 91.3 percent (3,909) in the sunnier south survived, but after adjustment neither region, a stand-in for yearly UVB, nor season of diagnosis predicted survival. 18
Across the 48 contiguous US states, death rates from multiple sclerosis tracked average temperature closely (r = -0.812) but did not correlate significantly with hours of sunshine (r = -0.245) or solar radiation (r = -0.14). 45
US multiple sclerosis death rates are more than three times higher in the coldest states than in the warmest. 45
Five severely disabled children aged 6 to 8, who spent a mean 0.48 hours a day outdoors, had hand-bone density of 0.52 to 0.55 with no change over a year, against 0.80 to 0.82 in children of the same age. 90
After the five children spent more time in the sun, bone density rose significantly at 3, 6, 9, and 12 months in all five; three of them had broken a bone before. 90
In the same five children, more time outdoors was followed by no broken bones during 2 years of observation. 90
Among 69,857 fractures in US Medicare patients, 35,007 of them hip fractures, rates at the hip, wrist, upper arm, and ankle were all highest in winter and lowest in summer. 29
In the same older Americans, hip fracture risk was 6 percent lower for every 14 more sunny days in autumn (relative risk 0.94), while wrist, upper arm, and ankle fractures ran 1.1 to 1.3 times higher with more sunny days. 29
Winter sunshine was not significantly linked to fracture risk at any of the four sites in these Medicare patients. 29
Of 90,149 hip fractures in the Catalan study, 27.2 percent came in autumn and 25.7 percent in winter, against 23.5 percent in summer and 23.6 percent in spring. 30
In the Catalan time-series model adjusted for trend and the other weather measures, hours of sunshine were the only climate variable consistently linked to the hip fracture rate. 30
Among 421,461 men and 90,141 women in US Army basic training, recruits from home areas with a low yearly UV index had slightly lower odds of stress fracture than those from high-UV areas (odds ratio 0.92 for men, 0.89 for women), against the authors' hypothesis. 7
In the same recruits, women had 4.5 times the odds of a stress fracture that men had. 7
Among 398 men aged 70 to 80 on 4 to 6 years of hormone-blocking therapy for prostate cancer, 49 of 224 (21.86 percent) who reported under 3 hours of sun a week had osteoporotic fractures, against 19 of 174 (10.92 percent) who reported over 10 hours. 31
In a dish, blue light at 17 mW per square centimetre raised calcium signalling in about 90 percent of cultured human Jurkat T cells, with half of them responding within 38 seconds. 28
White light at more than 2 J per square centimetre doubled the movement of cultured T cells toward an immune signalling chemical, a peak 100 percent increase. 28
In cultured human Jurkat T cells given UVA, the female hormone estradiol produced up to 28 percent cell killing at 20 ng/ml, while testosterone did not change UVA's effect. 91
Skin bacteria grown in the lab and given simulated sunlight equal to about 1 hour 20 minutes or 4 hours at UV index 6 lost numbers in at least 7 species, including Staphylococcus aureus and Staphylococcus epidermidis at the higher dose, while 3 species (Micrococcus luteus, Staphylococcus warneri, and Staphylococcus hominis) were unaffected. 10
The lower of the two simulated-sunlight doses increased the growth of Staphylococcus aureus in culture. 10
Both simulated-sunlight doses raised reactive oxygen species inside 10 of the 11 skin bacterial strains tested. 10
In 22 hemodialysis patients whose lower legs were given UVB three times a week during dialysis, daytime systolic blood pressure fell from 141 to 133 mmHg and night-time systolic from 145 to 135 mmHg. 92
In the same 22 dialysis patients, median parathyroid hormone fell 25 percent, from 12 to 9 pmol/l, and the bone-turnover marker osteocalcin fell 45 percent, from 16 to 11 μg/l. 92
In a cultured human skin cell line, UVB raised the activity of the Merkel cell polyomavirus small T antigen promoter 1.8-fold. 93
In 16 tuberculin-positive volunteers, a 5 percent melatonin cream applied right after each of three daily doses of simulated sunlight did not prevent suppression of the tuberculin skin reaction, and melatonin did not change the burn threshold. 94
The Edinburgh authors put their UVA dose at the equivalent of 30 minutes of noon sun on a sunny day in southern Europe. 1
The artificial UV source used on the Rotterdam hair gave an irradiance equal to a UV index of 56, beyond any UV index measured on Earth, and cut hair cortisol by a median 75 percent. 75
Over the Pärnu summer, activated T cells (CD3+HLA-DR+) rose from 4.6 to 8.1 percent of the 16 sunbathing women's blood lymphocytes. 22
T-cell lymphomas followed the same Australian latitude pattern: primary cutaneous anaplastic large cell lymphoma was 4.26 times, mycosis fungoides 1.72 times, and peripheral T-cell lymphoma 1.53 times as common in the southernmost band as in the northernmost. 17
After simulated sunlight, 8 of 11 skin bacterial species released more substances that switch on a skin-barrier receptor (AhR). 10
In nickel-allergic volunteers, adding 17.8 J/cm2 of UVA before simulated sunlight roughly doubled the suppression of the nickel reaction, and lowered the smallest suppressive simulated-sun dose from 2 to 1.5 J/cm2; the UVA alone suppressed nothing detectable. 95
In nickel-allergic women, the UVB band split off a solar simulator suppressed the nickel reaction within 24 hours, while the UVA band took 48 hours to suppress it and the effect was gone by 72 hours. 96
Lower-limb fractures in US Army recruits showed no difference between those from low-UV and high-UV home regions (odds ratio 0.95). 7
Across the range of UVA each country experienced from January to April 2020, the models predicted COVID-19 death risk falling by roughly a third to a half from the least to the most UVA. 57
In biopsies after twice the burn dose of simulated sunlight, Langerhans cells were down 60 percent at 15 hours and 88, 95, and 100 percent at 36, 48, and 72 hours, from about 15 cells per millimetre in unexposed skin. 85
In a second group of 31 volunteers, simulated sunlight at 1, 2, and 4 J/cm2 a day cut the tuberculin reaction by 14, 24, and 43 percent on placebo and by 8, 14, and 22 percent on 500 mg of nicotinamide a day. 81
In the same dialysis patients, daytime diastolic pressure fell 7 percent, from 81 to 75 mmHg, and night-time diastolic barely moved, from 71 to 70 mmHg. 92
Lymphoplasmacytic lymphoma was 1.34 times and mantle cell lymphoma 1.29 times as common in Australia's southernmost latitude band as in its northernmost. 17
A low dose of blue light, 120 mJ per square centimetre, increased the random movement of activated T cells crawling on a coated dish, and the effect lasted after the light stopped. 28