On a fine summer day in 1956, on a ward round at Rochford General Hospital in Essex, Sister J. Ward showed the consultant pediatrician, R. H. Dobbs, a premature baby she had carefully undressed. The baby was pale yellow except for one sharply edged triangle of skin, much yellower than the rest. Dobbs asked her what she had painted it with, iodine or flavine, and why. She said she thought it must have been the sun, and that the dark patch was where a corner of the sheet had covered the skin. "It's the rest of the body that seems to have faded." 1
Sister Ward ran the hospital's premature unit, a set of rooms built before the war around a paved courtyard open to the south. She had been "chosen because of her known skill in rearing puppies," Dobbs wrote. On warm days she wheeled the more delicate babies out into the courtyard, sure that fresh air and sun would do them more good than an overheated incubator, and brought them back in when consultant rounds were near. 1
The yellow was bilirubin, a pigment the body makes when it breaks down heme from red blood cells, and which a newborn's liver cannot yet clear quickly. More than 80 percent of newborns show some jaundice. In most it is plainest around the third day and gone in about two weeks, and in a few the bilirubin climbs high enough to injure the brain. 2; 3
Sunlight lowers bilirubin because the blue light in it turns the pigment into forms the body can pass without the liver's help, as hospital phototherapy lamps do. 4; 5 In the largest of four randomized trials in Nigeria, sunlight filtered through window-tinting film kept bilirubin under control on 93 percent of the treatment days that could be judged, against 90 percent for the hospital's electric lamps, in 447 babies. 6; 7 No trial has shown that plain sun at home treats jaundice. In a Northern California study, a doctor had recommended sunlight for 11 of 62 babies whose bilirubin went on to reach a dangerous level, and for 35 of 248 similar babies whose bilirubin did not. The American Academy of Pediatrics and Britain's National Institute for Health and Care Excellence both advise against treating jaundice with sunlight. 8; 2; 9
This page covers how sunlight came to treat jaundice, what the light does to bilirubin, the filtered-sunlight trials in Nigeria, why guidelines tell parents not to sun a jaundiced baby, how much sun babies get, and what the studies show about sun, melasma, and sunscreen in pregnancy. Each study is named by its exposure: real sun, sun through film or glass, a lamp, sun estimated from weather records, or a chemical measured in urine.
Key facts
- At Rochford in 1956, a blood sample left in the sun on a window sill read 9 milligrams of bilirubin per 100 milliliters, while a fresh sample from the same baby read 24. 1
- In the Lagos filtered sunlight trial of 447 babies, filtered sunlight kept bilirubin under control on 93 percent of the treatment days that could be judged and electric lamps on 90 percent; temperatures above 38 °C came in 5 percent of the sunlight group against 1 percent. 6
- In a California study, 11 of 62 babies whose bilirubin climbed to a dangerous level had been advised sunlight by a doctor, against 35 of 248 babies whose bilirubin did not climb. 8
- In 8 of 20 negligence claims against the UK's National Health Service over newborn jaundice from 2001 to 2011, most of them for cerebral palsy, the baby had been put in sunlight. 10
- In 556,376 Scottish births, pregnancies whose first trimester had the most available sunshine had about 9 percent lower odds of preterm birth than those with the least; no one measured the sun on the mothers' skin. 11
Sunlight faded jaundice in premature babies at Rochford in 1956, and a blue lamp replaced it by 1958
A few weeks after the yellow triangle, blood from another deeply jaundiced baby went to the hospital laboratory and came back hours late with a bilirubin of 13 to 14 milligrams per 100 milliliters, far too low for so yellow a child. A fresh sample went up to the laboratory with a request for an explanation. The biochemist, P. W. Perryman, apologized for the delay: "I found the tube lying on the window sill and I did it myself, so I am sure it's correct." The fresh sample read 24. The morning's tube, still lying in full sunlight on the sill, now read 9. 1
R. J. Cremer, the pediatric registrar, remembered a different sample, taken just before an exchange transfusion on a sunny July afternoon, mislaid, and found afterward on a window sill with its serum green instead of yellow. An exchange transfusion replaces a baby's blood a little at a time with donor blood, and it was then the main treatment for dangerous jaundice. From that day, Cremer wrote, samples stood on the laboratory window sill "behaving like traffic lights as they turned from red to yellow and then green." Serum kept in glass at a steady temperature lost its bilirubin just the same, so the cause was light and not heat. 1
On the ward, jaundiced babies were then "cautiously exposed to sunlight for short periods." Their bilirubin fell, and the yellow left their skin everywhere except where clothing or a shadow covered it. Taken out of the sun, they turned yellow again and their bilirubin rose, until they went back out. "Rejoicing was short lived, as at this stage the sun deserted us," Cremer wrote. 1
The laboratory filled with "every kind of electric light equipment," in Cremer's words, "including a Southend street lamp." A plain blue fluorescent tube came out best, "somewhat less effective than sunlight" but "perhaps safer." The hospital engineers built a cradle of eight 40-watt blue tubes to hang over a cot, and the baby's eyes were covered with a plastic shield that head movements tended to knock loose. Six hours on and two hours off worked better than steady light, "for reasons not fully understood." 1
The lamp could not keep up with the "torrent of bilirubin" in babies with rhesus disease, in whom the mother's antibodies destroy the baby's red blood cells. It often held the slower rise of bilirubin in premature babies, and it spared some of them an exchange transfusion. Cremer, Perryman, and D. H. Richards published the work in The Lancet in 1958. "Ten years were then to pass," Cremer wrote, "before phototherapy was virtually rediscovered in America." 1; 12; 2
Blue light, not ultraviolet, turns bilirubin into forms the body can pass
The Rochford team found early on that "it was not ultraviolet but visible blue light that was most effective," and that the light acted only on unconjugated bilirubin, the form the liver has not yet processed. They took the green in their tubes for biliverdin and supposed that light was oxidizing the bilirubin. 1 In a 2001 review, A. F. McDonagh wrote that the main effect "is not photodegradation of bilirubin, but conversion of the pigment to structural isomers that are more polar and more readily excreted." These isomers are the same molecule bent into shapes that dissolve in water and leave the body without the liver's work. 13; 5
The American Academy of Pediatrics puts the best light at blue to green, about 460 to 490 nanometers, where bilirubin absorbs most strongly and the light reaches well into the skin. The devices it rates highest are light-emitting diodes of that color, with "low infrared (IR) emission, and no ultraviolet (UV) emission." For intensive phototherapy it asks for about 30 microwatts per square centimeter per nanometer of that light, a measure called irradiance. 4
In 1968 Jerold Lucey and his colleagues in the United States supplied the rediscovery Cremer mentioned, a controlled trial of blue lamp light in 111 premature babies: 53 lay under the lights from 12 to 144 hours of age and 58 did not. Light treatment, they wrote, "has not been widely used because of doubts as to its effectiveness and concern for the possible toxicity of the photochemical decomposition products of bilirubin." The treated babies kept lower bilirubin. 14
In 2013 Kyohei Sumida and colleagues at Kitasato University in Tokyo put 3-day-old mice under a lamp for 60 minutes. The lamp gave ultraviolet B, or UVB, the shorter-wave ultraviolet that causes most sunburn. The mice carried the human gene for UGT1A1, the only enzyme that can prepare bilirubin for removal, and like human newborns they become jaundiced after birth. A day later the lamp-treated mice averaged 6.2 milligrams of bilirubin per deciliter against 9.2 in the unexposed mice, and their skin held more than three times as much of the enzyme. The authors wrote that sunlight "makes it possible to treat neonatal jaundice while allowing mothers to breast-feed neonates," and later in the paper that "clinical data will be needed." 15
Ward daylight went with less jaundice in Finland; doctor-advised sunlight did not lower the risk of severe jaundice in California
In Oulu, in northern Finland, daylight runs from 3 to 22 hours depending on the season. A 1975 study there followed 86 premature babies born one after another over a year, lying in incubators in nothing but nappies. From the fifth day of life, babies born in the light half of the year had lower bilirubin, and 4 of them went above 255 micromoles per liter, about 15 milligrams per deciliter, against 10 babies born in the dark half. The authors concluded that "natural light conditions are of importance in the design of wards for newborn infants." 16
In 1985 P. Barss and K. Comfort reported that after builders extended the roof overhangs of a tropical obstetric ward to several meters, shutting out most of the daylight, the share of newborns with high bilirubin rose from 0.5 percent to 17 percent. Their abstract names neither the hospital nor the country and gives no count of babies. 17 A 1984 study of 498 healthy full-term babies in Dunedin, New Zealand, found that "natural illumination" made no difference to jaundice. 18
In a 2008 study from Kaiser Permanente hospitals in Northern California, Kuzniewicz and colleagues took the 17,986 babies born from 1995 to 2004 whose bilirubin sat in a borderline range after two days of age. They compared the 62 whose bilirubin went on to reach 25 milligrams per deciliter with 248 similar babies whose bilirubin did not. A doctor had recommended putting the baby in sunlight for 11 of the 62 (18 percent) and 35 of the 248 (14 percent). Babies advised sun had about 1.3 times the odds of reaching the dangerous level, an estimate that left room for anything from about a third lower to nearly three times higher, and the authors wrote that "placing the infant in sunlight had no apparent efficacy." Hospital phototherapy started within 8 hours was 85 percent effective at preventing it. 8
The one randomized trial of plain sun for jaundice ran in Shenzhen, China, from 2004 to 2007. Parents of 960 healthy newborns were assigned to put their babies in direct sun morning and afternoon for 30 to 60 minutes at a time, "with as much skin exposed as possible"; to bathe them in water boiled with Chinese herbs; to do both; or to do neither. Among the 482 babies in the sun-only and neither groups, sunned babies had about 40 percent less jaundice, a true effect anywhere from about a fifth less to a little over half less, and 2.2 fewer days of it. The Cochrane review, a 2021 systematic summary of the randomized trials of sunlight for jaundice, rated the evidence very low in certainty and noted that the trial reported no data on safety. 5; 7
Severe newborn jaundice is about 180 times as common in Africa as in Europe
In 2010, by one estimate, 24 million of the 134 million babies born worldwide at 32 weeks of pregnancy or later were at risk of harm from jaundice, and about 114,100 died of extreme jaundice or rhesus disease. Three quarters of those deaths were in sub-Saharan Africa and South Asia. 19 A 2017 review put severe jaundice, meaning jaundice that led to brain injury, an exchange transfusion, or death, at 667.8 cases per 10,000 live births in Africa and 3.7 in Europe. It found only 21 population-based studies to draw on, 16 of them from rich countries. 20
Kernicterus is the lasting brain damage of high bilirubin, with a form of cerebral palsy, hearing loss, eyes that cannot look upward, and damaged baby teeth. 2 In California it was diagnosed in 0.42 of every 100,000 births from 1988 to 2014, falling to 0.04 after 2009. In Denmark the rate was 1.2 of every 100,000 from 2000 to 2015. 21; 22
At Massey Street Children's Hospital in Lagos, 1,153 of the 5,229 newborns admitted from 2012 to 2014 came in for jaundice. Of the 1,118 with complete records, 17 percent had acute bilirubin encephalopathy, the early brain injury of high bilirubin, 31.5 percent needed an exchange transfusion, and 61 died. The public power supply was so unstable that every electric device, the phototherapy lamps included, needed a second power source. Staff asked every family on admission about menthol baby products and camphor balls, which can trigger the breakdown of red cells in babies with G6PD deficiency, an inherited shortage of the enzyme glucose-6-phosphate dehydrogenase that is common there. 23
Sunlight under window-tint film controlled bilirubin about as often as electric lamps in four Nigerian trials
In 2013 Hendrik Vreman's laboratory at Stanford tested nine semi-transparent window-tinting films, under lamps and then under a cloudless sky. The films passed 24 to 83 percent of the blue light, only 0.1 to 7.1 percent of the UVA, the longer-wave ultraviolet, and cut the heat under the sun by 5 to 10 °C. 24
Tina Slusher of the University of Minnesota and Bolajoko Olusanya of the Center for Healthy Start Initiative in Lagos built a treatment around those films. Their protocol for a trial at Island Maternity Hospital in Lagos called for two canopies, one with film passing about a third of the blue light for sunny hours and one passing about four fifths for overcast ones. The babies' eyes were to be protected "with low-cost eye covers made from the elastic tops of any color socks," and mothers sat beside them on reclining chairs painted white, in white aprons, and breastfed as often as they liked. The protocol noted that "the single most important reason for replacing films is wind damage." 25
The randomized trial ran at the same hospital from November 2012 to September 2013. Of 447 babies, 224 were assigned to filtered sunlight and 223 to the hospital's electric phototherapy, and every baby who needed treatment went under electric lamps at night and on rainy days. A day counted as a success if bilirubin rose by less than 0.2 milligrams per deciliter an hour in a baby up to three days old, or fell in an older one. Filtered sunlight succeeded on 93 percent of the days that could be judged and electric light on 90 percent, and the canopies gave more than twice the blue light of the hospital's lamps, 40 microwatts against 17. Temperatures above 38 °C came in 5 percent of the sunlight group against 1 percent, but no baby had to be withdrawn for safety and none needed an exchange transfusion. 6; 5
In Ogbomoso, in what the team called "a simulated rural setting," a second trial from 2015 to 2017 treated 174 babies with higher bilirubin, half of them in a transparent polycarbonate room lined with film. On the days with at least 4 hours of treatment, filtered sunlight succeeded on 87.2 percent and intensive electric phototherapy on 88.8 percent. The true difference could have been anywhere from about 10 points worse to about 7 points better for sunlight, inside the 10-point margin the team had set in advance for calling it no worse. 26
At Massey Street, in a trial reported in 2024, Olusanya's team treated 192 newborns with severe to hazardous jaundice, some with bilirubin as high as 62 milligrams per deciliter, either in an outdoor film-lined enclosure or under intensive electric lamps. Filtered sunlight succeeded on 94.2 percent of the days that could be judged and electric light on 97.1 percent, and 7 babies died in each group. The authors wrote that "about 80% of neonates required night-time phototherapy which makes the availability of IEPT imperative in any setting," IEPT being intensive electric phototherapy. 27 A fourth trial, reported in 2026, treated 104 babies with moderate to severe jaundice in Nigerian birth hospitals; filtered sunlight averaged 27.7 microwatts, below the lamps' 36.1, and both succeeded on about 93 percent of days. 28
In a 2026 summary, Olusanya's group put the cost of an enclosure at US$44 to US$120, with room for up to six mothers and babies, used for at least 4 hours a day, "typically between 10:00 and 15:00, when there is no overcast." "In low-resource settings," they wrote, "the choice is not between conventional electric-powered phototherapy and filtered sunlight, but between filtered sunlight and direct sunlight or no treatment at all." 7
Across the two filtered-sunlight trials in the Cochrane review, 621 babies in all, temperatures above 37.5 °C were more common under sunlight: for every three babies treated with filtered sunlight instead of lamps, one more overheated. 5 Of 191 mothers whose babies had filtered sunlight in Lagos, 98.4 percent said they would use it again or recommend it, and their most common complaint was the exposure of their newborns to "hot" sun. 29
In 2026 Slusher and two colleagues reported BiliRoo, a kangaroo carrier with a filtered-sunlight panel meant to treat a baby held skin to skin against a parent. On a bench from 10 a.m. to 8 p.m., its films passed enough blue light at every angle tried and blocked more than 99 percent of the UVA. The test used a bench model and no baby. 30
Guidelines in the United States, Britain, and Australia say not to treat jaundice with sunlight
The American Academy of Pediatrics set out its position in 2004 under the heading "Sunlight Exposure." Sunlight, it said, "provides sufficient irradiance in the 425- to 475-nm band to provide phototherapy," but "the practical difficulties involved in safely exposing a naked newborn to the sun either inside or outside (and avoiding sunburn) preclude the use of sunlight as a reliable therapeutic tool, and it therefore is not recommended." The same guideline says jaundice should be looked for "in a well-lit room or, preferably, in daylight at a window." 31
Britain's National Institute for Health and Care Excellence gave its rule in 2010: "Do not use sunlight as treatment for hyperbilirubinaemia," the medical name for high bilirubin. Like the academy, it tells staff to check the naked baby "in bright and preferably natural light." 9 Cancer Council Australia's 2017 fact sheet on babies says that "exposing babies to direct sun is inappropriate to treat neonatal jaundice." 32
The academy's 2022 revision kept the advice and added a sentence about the trials in Nigeria: "Although filtered sunlight has been safely used in resource-constrained settings where phototherapy is not readily available, these guidelines were not developed for use in such settings." 2 Its 2024 technical report credited Sister Ward's observation with leading to phototherapy devices; it calls her Jean, and Dobbs gave only her initial. The report says that "natural unfiltered sunlight cannot be applied at prescribed dosages," and that undressing newborns and putting them in the sun at home "cannot be recommended given the concerns for hypothermia, hyperthermia, sunburn, and long-term risk of various skin malignancies." 4
The 2021 Cochrane review found three randomized trials with 1,103 babies, all small, unblinded, and at high risk of bias. It concluded that sunlight "may be an effective adjunct to conventional phototherapy" where lamps are short, and that "these studies have not demonstrated that sunlight alone is effective for the treatment of hyperbilirubinemia given its sporadic availability and the low or very low certainty of the evidence in these studies." 5
Parents in many countries put jaundiced babies in the sun or by a window, often on a health worker's advice
In Townsville, in tropical Queensland, Simone Harrison and her colleagues at James Cook University interviewed 114 women who had just given birth, for a 1999 paper titled "Why do mothers still sun their infants?" Thirty-six percent favored sunlight to treat jaundice, and 40 percent of those with older children had already sunned a child for it. Most of the advice had come from a midwife or nurse (41 percent) or a doctor (28 percent). 33
In eight Queensland hospitals in 1999 and 2000, Harrison's group put the question to the staff: 49.5 percent of 285 nurses and 34.9 percent of 130 doctors caring for new mothers believed sunlight treats newborn jaundice, and midwives recommended it more often than other nurses. 34 After the group ran teaching sessions for the nurses of one maternity ward, the share of new mothers who said they would sun a baby with suspected jaundice fell from 28.8 percent of 106 to 13.3 percent of 203. 35
A 2026 review led by Olusanya gathered 109 studies from 30 countries, most of them surveys of mothers, and found the share of people using or recommending sun for jaundice ranging from 1.1 percent in the United States to 100 percent in India and Mexico, with a median of 40 percent. The commonest practice was unfiltered early-morning sun, between 8 and 10 a.m., for 15 to 30 minutes a day. Some families put the baby beside a closed window "on the explicit assumption that the glass attenuated harmful spectra while permitting therapeutic ones." Harms such as sunburn, dehydration, overheating, and late arrival at a clinic "were under-reported," the reviewers wrote. 36
Janet Rennie and two colleagues read the UK negligence claims over newborn jaundice held by NHS Resolution for 2001 to 2011; 20 had enough detail to analyze, and 15 of those children had cerebral palsy. Sunlight had been used for eight of the babies, three of them born after the National Institute for Health and Care Excellence advised against it in 2010. Mothers reported comments such as "the midwife said the jaundice was nothing to worry about, and that putting the baby in the window would be sufficient." All but two of the 20 babies had an underlying condition, G6PD deficiency in six of them. 10 In 1986 the journal Burns reported a newborn sunburned during sun treatment for jaundice. 37
In Spain, where the advice to put a jaundiced baby "in the sun, by the window" is "commonly issued to parents," Sequí-Canet's group in Gandia set a spectroradiometer, an instrument that measures light wavelength by wavelength, behind seven kinds of household glazing. Direct sun on a clear day gave about 230 microwatts per square centimeter per nanometer in the blue band that treats jaundice, several times the 30 to 65 used in phototherapy, and ordinary glass let most of it through. Below 400 nanometers the glass cut the light very little, "with transmittance values close to 100% for many types of glass," so much of the UVA passed, and so did the near-infrared that carries heat. The dose, they wrote, depends on "time of day, season, weather, orientation, shadows, dirt on the glass, distance from the window, and surface area of skin exposed." Their instrument could not measure UVB, and the study included no babies. 38
Babies wore ultraviolet badges in Townsville and Perth, and Perth babies who later had eczema had received less sun
In Townsville, Aurel Moise, Harrison, and Peter Gies fixed ultraviolet-sensitive film badges to the chests and shoulders of 1-year-olds for a week in October 1995 and five days in April 1997. The badges counted standard erythema doses; erythema is the medical word for skin redness, and one standard erythema dose is a fixed amount of the ultraviolet that reddens skin. The median 1-year-old received 0.4 doses a day on the chest and on the shoulder. 39 The team put a year's dose for these babies at 84 standard erythema doses, with a winter month about 40 percent of a summer month; in Britain, they noted, a winter month gives 0.5 percent. 40
Kristina Rueter's group in Perth ran a trial from 2012 to 2017 in which 195 newborns from families with allergic disease were given 400 international units of vitamin D a day, or an identical drop of coconut and palm kernel oil, until 6 months of age. The vitamin D made no difference to eczema. A subgroup of 82 babies also wore ultraviolet badges clipped at head or shoulder height from birth to 3 months, their sensor a film of spores of Bacillus subtilis, a bacterium whose DNA responds to ultraviolet much as skin does. The babies received a median of 952 joules per square meter over the three months, "typically with faces, hands, and arms exposed." The 16 diagnosed with eczema by 6 months had received a median of 555, against 998 for the 64 without it. 41
Followed to two and a half years, children with eczema had received a median of 747 joules per square meter in their first three months, against 1,204 for those without, and their parents reported 7 minutes a day outdoors between 11 a.m. and 3 p.m. in those months, against 20. 42 A 2020 appraisal in the British Journal of Dermatology noted that the badges had been handed out "in a nonrandomized fashion," and called the finding the result of "exploratory post-hoc analyses in a nonrandomized subset." 43
In Xi'an, China, custom keeps mothers and newborns indoors for about the first two months, a practice called zuo yuezi, or "sitting the month." A team at Xi'an Jiaotong University followed the babies of 229 mothers with depression from 2015 to 2017 and grouped them by what the mothers reported, including vitamin D drops alone and drops plus 2 to 14 hours of real sun a week, 30 minutes at a time. The 50 babies who got 7 to 14 hours a week had the best motor scores and the least cortisol, a stress hormone, in their hair. The groups were not randomized, and the authors wrote that they "should have measured the serum 25OH-vitamin D concentrations," the usual blood test for vitamin D. 44
Advice on sunscreen for babies under six months rests on skin structure, not on tests in babies
The American Academy of Pediatrics tells parents to "keep babies younger than 6 months out of direct sunlight," and to use sunscreen "on small areas of the body, such as the face, if protective clothing and shade are not available." 45 Its 2011 policy statement says that "to date, no data show toxicity from absorption of sunscreen ingredients in infants and young children," and that preterm babies, "because of a thinner stratum corneum, may have a higher susceptibility to the absorption of sunscreen ingredients." The stratum corneum is the outermost layer of dead skin cells, the skin's main barrier. 46
The US Food and Drug Administration titles its advice "Should You Put Sunscreen on Infants? Not Usually," and says infants "are at greater risk than adults of sunscreen side effects, such as a rash," without citing a study. 47 Every sunscreen sold in the United States must carry the line "children under 6 months of age: Ask a doctor." 48
A 2010 study looked into the skin of 20 babies aged 3 to 24 months, and of their mothers, with a laser microscope. The babies' stratum corneum was 30 percent thinner, and their epidermis, the whole outer layer of skin, 20 percent thinner. All five authors worked for Johnson and Johnson's consumer products division, which makes baby products. 49
We found no study that tested sunscreen on babies under six months, for absorption or for skin reactions. In 1999 a Queensland group tested three brands that sold a baby or toddler sunscreen beside an adult one; each pair held the same filters at the same strengths. On epidermis from adults' tummy-tuck surgery, held for 8 hours in glass chambers that collect whatever passes through, oxybenzone, a common sunscreen chemical, was the only filter to cross, up to 10 percent of the dose, and the children's and adult versions behaved alike in two of the three pairs. 50
Parents in Perth reported never putting sunscreen on 122 of 163 babies at 3 months and always on 5. 42 In Japan, a girl born at 27 weeks weighing 542 grams, already taking zinc by mouth, reached a blood zinc of 427 micrograms per deciliter after zinc oxide ointment, the same mineral used in mineral sunscreens, went on her nappy rash. 51
In 199 Philadelphia-area babies aged 6 to 12 weeks, the median level of oxybenzone in urine was 3.6 micrograms per liter. 52 Measurements of sunscreen chemicals in breast milk, placentas, and cord blood are on the sunscreen page.
Melasma darkened the faces of 16 to 46 percent of pregnant women in surveys; two small studies tested sunscreen against it
Melasma is a brown to gray-brown patchiness of the face, called "the mask of pregnancy." 53 The British Association of Dermatologists says ultraviolet "can trigger melasma or make it worse," that high-energy visible light "may also contribute," and that in pregnancy "it may go away a few months after delivery." 54
Surveys of pregnant women found melasma in 15.8 percent of 400 in Tehran, 37 percent of 415 in Casablanca, 37.8 percent of 450 in central Anatolia, 43.4 percent of 403 in Gondar, Ethiopia, and 46.4 percent of 140 in Pakistan. 55; 56; 57; 58; 59 In Tehran, melasma showed no link with sunscreen use. 55 A figure of 50 to 70 percent opens the abstract of a 2007 sunscreen study, and the British Association of Dermatologists says "up to 50%." 60; 54
In Santiago, Chile, a 1987 double-blind trial gave 65 pregnant women a placebo cream or a sunscreen with a sun protection factor (SPF) of 10 to put on their faces daily in the second trimester, in summer. The sunscreen absorbed from 290 to 360 nanometers, leaving the longest UVA uncovered. Among the 30 women who used the cream as directed, 16 on placebo and 14 on sunscreen, melasma was significantly less common with sunscreen. 61
In Casablanca, a 2002 survey of 415 pregnant women by Benchikhi, Razoki, and Lakhdar found that 18.8 percent of those with melasma used sunscreen, and 16.6 percent of those without. Its authors wrote that "no study has confirmed that sunscreens can prevent melasma in dark phototype women with pregnancy. It would therefore be difficult to recommend them to pregnant women." 56
Five years later Lakhdar's group gave 200 pregnant women in Casablanca an SPF 50+ sunscreen with strong UVA protection for 12 months. Of the 185 who finished, 5 developed melasma (2.7 percent), "much lower than the 53% previously observed in an usual condition study (same investigators, same geographical area and same time frame)." The study had no group without sunscreen of its own. Three of its seven authors worked for La Roche-Posay Pharmaceutical Laboratories, which makes sunscreens. 60
The American College of Obstetricians and Gynecologists advises "sunscreen and a wide-brimmed hat every day when you are outside" to keep melasma from getting worse, and notes that it usually fades after birth, though "some women may have dark patches that last for years." 53
Pregnancies with more sunshine had less preeclampsia and preterm birth in record studies; none measured sun on the skin
A 2017 review found eight studies of ultraviolet and pregnancy, all observational and all at high risk of bias, in which more ultraviolet in the first trimester went with better fetal growth and with "increased hypertension in pregnancy." 62 Lauren Megaw, Sarah Stock, and their colleagues in Edinburgh then linked 556,376 Scottish births from 2000 to 2010 to the sunshine hours recorded for each mother's 5-kilometer grid square from Met Office records and Meteosat satellites. Six percent of the babies were born preterm, before 37 weeks. Mothers in the sunniest quarter for the first trimester had 0.91 times the odds of a preterm birth of mothers in the least sunny quarter, after adjustment: about 9 percent lower, somewhere between 7 and 13 percent. Second-trimester sunshine showed no link, and the pattern held when the team compared term and preterm births of the same mothers. 11
A companion study linked 522,896 Scottish deliveries to ultraviolet B measured from satellites. Preeclampsia, a dangerous rise in blood pressure in the second half of pregnancy, came in 8,689 (1.66 percent), and mothers in the top fifth of ultraviolet had 0.57 times the odds of the bottom fifth, a true effect somewhere between 28 and 56 percent lower. "No data were available on changes in UV exposure attributable to holidays and trips abroad," the authors wrote. 63
At Edinburgh in 2015 and 2016, 19 women in the second trimester were given 20 minutes of low-dose UVA from a phototherapy unit or a sham exposure. The registered title states a result: "UVA Irradiation of Skin in Pregnant Women Increases Uterine Artery Blood Flow and Reduces Blood Flow Through a Nitric Oxide Pathway." The trial was completed in March 2016 and has posted no results. 64
Across 15 studies, a raised body temperature in early pregnancy went with 1.92 times the odds of a neural tube defect such as spina bifida, somewhere between 1.6 and 2.3 times. 65 Across 70 studies from 27 countries, the odds of a preterm birth rose 5 percent for each 1 °C of air temperature and 16 percent during heatwaves. 66
In June 2017, a routine blood test at 28 weeks of pregnancy in North Devon found a woman's white cell count at 29.9 billion per liter, up from 8.61 a week before. She "had spent an afternoon on the beach 4 days before her presentation and had neglected to protect her skin with clothes or sun cream." The ultraviolet index had reached 7 in the town, unusual for Britain, and her count was back to 12.1 by August. 67
Sunscreen chemicals in pregnant women's urine rose with sunscreen use, and studies of birth size point both ways
MotherToBaby, a counseling service run by specialists in birth defects, wrote in 2015 that "to date there is no published information suggesting that sunscreens cause an effect to the developing fetus/unborn baby." 68 In 2020 it added that "many of these sunscreen chemicals have not been studied very well in pregnancy, although they are not known to have a negative effect on the pregnancy or baby." 69
In northern Puerto Rico, the 30 pregnant women who reported using sunscreen in the previous 48 hours had oxybenzone in their urine at a geometric mean, a kind of average suited to skewed data, of 187.2 nanograms per milliliter, against 39.5 in the 869 who had not, about five times as much. 70
In 3,619 mothers and babies from 11 US cohorts, each tenfold rise in a mother's average oxybenzone went with 29 percent higher odds of a baby small for its gestational age, somewhere between 5 and 58 percent higher, and with no change in preterm birth or low birth weight. 71 In 922 Puerto Rico pregnancies, higher oxybenzone went with pregnancies 1.90 days longer and lower odds of preterm birth. 72 In Boston, 476 mothers sampled three times showed no link with fetal weight in either sex, and in a later Boston study of 900, mothers of small babies had the lowest oxybenzone. 73; 74 In France, oxybenzone went with larger growth measures on ultrasound in boys but not girls, and with 26 grams more birth weight in boys in another cohort, an estimate whose range ran from 2 grams less to 54 more. 75; 76
The US National Library of Medicine's LactMed database, revised in February 2026, says most chemical sunscreens are detectable in breast milk, "but the amounts that an infant would ingest are usually lower than current limits of concern." It says products with octinoxate and enzacamene, two other sunscreen chemicals, "and especially oxybenzone should probably be avoided during breastfeeding because of their possible endocrine-disrupting effects," and that mineral sunscreens "may be better alternatives for nursing mothers." On levels in breastfed babies and effects on them, it says: "Relevant published information was not found as of the revision date." 77
Misconceptions
What people believe
- A little morning sun, or a spot by a window, is a good home treatment for newborn jaundice.
- The ultraviolet in sunlight is what clears jaundice.
- Window glass filters out the harmful part of sunlight and lets the healing part through.
- Guidelines advise against sunlight for jaundice because it does nothing.
- Babies under six months must never wear sunscreen.
- Most pregnant women get melasma.
- Sunscreen chemicals have been shown to harm the unborn baby.
What the studies show
- Doctor-advised sunlight did not lower the odds of dangerous bilirubin in a California study, and no trial has shown that plain sun at home treats jaundice. 8; 5
- Blue to blue-green light, about 460 to 490 nanometers, does the work. The Rochford team found blue light more effective than ultraviolet. 1; 4
- Household glass let through most of the UVA and near-infrared heat along with the blue light in a 2026 Spanish study. 38
- Guidelines accept that sunlight lowers bilirubin and advise against it because the dose cannot be controlled and babies can burn or overheat. Filtered sunlight worked in Nigerian trials. 2; 4; 6
- The American Academy of Pediatrics allows sunscreen on small areas such as the face when shade and clothing are not available. 45
- Surveys found melasma in 15.8 to 46.4 percent of pregnant women. 55; 59
- Studies of oxybenzone in pregnant women's urine and their babies' size point in both directions, and no published study has shown harm to a baby from a mother's sunscreen. 71; 72; 68
Questions people ask
Does sunlight help newborn jaundice?
Sunlight lowers bilirubin, which is how phototherapy was found at Rochford in 1956. Filtered through window-tinting film, it controlled bilirubin about as often as electric lamps in Nigerian trials, with hourly temperature checks and lamps at night. No trial has shown that plain sun at home treats jaundice, and US, British, and Australian guidelines advise against it. 1; 6; 5; 2; 9
Can I put my newborn by a window for jaundice?
A 2026 Spanish study found that household glass passes plenty of the blue light that treats jaundice, and most of the UVA and heat as well, with no way to control the dose. The authors concluded that sun through home windows "cannot be considered an appropriate or safe intervention to prevent jaundice in developed countries." 38
How long should a jaundiced baby be in the sun?
US, British, and Australian guidelines give no time, because they advise against it. The Nigerian trials used film canopies for at least 4 to 5 hours a day, mostly between 10 a.m. and 3 p.m., with temperatures checked every hour. The commonest home practice worldwide, 15 to 30 minutes of early-morning sun, has not been tested in a trial. 2; 7; 36
Is it blue light or UV light that treats jaundice?
Blue to blue-green light, about 460 to 490 nanometers, and the lamps rated highest give off no ultraviolet. In one mouse study, a UVB lamp raised a skin enzyme that clears bilirubin, and the authors wrote that clinical data would be needed. 4; 15
Can you put a newborn in the sun?
The American Academy of Pediatrics and the Food and Drug Administration advise keeping babies under six months out of direct sunlight. Cancer Council Australia advises keeping babies under 12 months out of direct sun when the ultraviolet index is 3 or more, and considers a few minutes below 3 safe. The studies of childhood sunburn and melanoma are on the sunburn page. 45; 47; 32
Can babies under six months wear sunscreen?
The American Academy of Pediatrics allows sunscreen on small areas such as the face when shade and clothing are not available, and US sunscreen labels say "Ask a doctor" for that age. We found no study of sunscreen absorption or skin reactions in babies under six months. 45; 48; 46
Is sun exposure safe during pregnancy?
The studies of sunlight and pregnancy used weather or satellite records, and in Scotland more sunshine went with less preterm birth and less preeclampsia. The measured risks come from heat, both a raised body temperature in early pregnancy and hot weather. Sun can also darken melasma. 11; 63; 65; 66; 54
Does sunscreen prevent melasma in pregnancy?
Two small studies found less melasma with sunscreen. The 1987 trial's main result came from 30 of its 65 women, and the 2007 study had no comparison group of its own and included authors from a sunscreen maker. A 2006 Tehran survey found no link between melasma and sunscreen use. 61; 60; 55
Is sunscreen safe during pregnancy and breastfeeding?
No published study has shown harm to a baby from a mother's sunscreen, and studies of oxybenzone and birth size point in both directions. LactMed suggests avoiding oxybenzone while breastfeeding and calls mineral sunscreens possibly better alternatives. 68; 69; 71; 72; 77
What we do not know
No study has tested whether the sun families give jaundiced babies at home, most often 15 to 30 minutes early in the morning, lowers bilirubin. The one randomized trial of plain sun used longer sessions and reported no safety data. 36; 5
No registry records how often home sunning leads to sunburn, overheating, dehydration, or a late arrival at hospital, and the Lagos group's 2026 review called for one. We found no study that followed babies treated with sunlight or filtered sunlight for later skin cancer. 36
In the sickest babies, about four in five needed electric lamps at night, whichever treatment they had by day. 27
We found no study of sunscreen absorption or reactions in babies under six months. The advice rests on measurements of skin thickness and on caution. 49; 46
The Perth link between more sun and less eczema came from a subgroup that was not randomized. 43
No study of sunlight and pregnancy outcomes measured the sun on a mother's skin, and the one registered trial of ultraviolet in pregnancy has posted no results. The evidence that sunscreen prevents melasma rests on two small studies. 11; 64; 61; 60
Key studies
- Dobbs and Cremer, 1975, Rochford. The doctors' own account of how sunlight faded jaundice in premature babies in 1956 and a blue lamp replaced it. 1
- Lucey and colleagues, 1968, United States. A controlled trial of blue lamp light in 111 premature babies. 14
- Anttolainen and colleagues, 1975, Oulu. Premature babies born in the light half of the year had lower bilirubin. 16
- Barss and Comfort, 1985. Jaundice rose from 0.5 to 17 percent of newborns after a ward's roof overhangs shut out daylight. 17
- Kuzniewicz and colleagues, 2008, Northern California. Doctor-advised sunlight did not lower the odds of dangerous bilirubin. 8
- Sumida and colleagues, 2013, Tokyo. A UVB lamp lowered bilirubin in newborn mice and raised a skin enzyme that clears it. 15
- Vreman and colleagues, 2013, Stanford. Window-tinting films passed blue light and blocked most UVA. 24
- Slusher and colleagues, 2015, Lagos. Filtered sunlight controlled bilirubin on 93 percent of judged treatment days against 90 for electric lamps, in 447 babies. 6
- Slusher and colleagues, 2018, Ogbomoso. Filtered sunlight matched intensive electric phototherapy in 174 babies with higher bilirubin. 26
- Olusanya and colleagues, 2024, Lagos. Filtered sunlight in 192 babies with severe jaundice; about four in five needed lamps at night. 27
- Horn and colleagues, 2021, Cochrane. Three trials with 1,103 babies; sunlight may help alongside lamps and has not been shown to work alone. 5
- Rennie and colleagues, 2019, UK. Sunlight had been used in 8 of 20 jaundice negligence claims. 10
- Olusanya and colleagues, 2026, 30 countries. The share of families using sun for jaundice ran from 1.1 to 100 percent, median 40. 36
- Sequí-Canet and colleagues, 2026, Gandia. Household glass passed blue light, UVA, and heat with no control of dose. 38
- Rueter and colleagues, 2019, Perth. Babies with eczema by 6 months had worn badges recording less ultraviolet. 41
- Stamatas and colleagues, 2010. Babies' outer skin layer was 30 percent thinner than their mothers'. 49
- Lakhdar and colleagues, 2007, Casablanca. Melasma in 5 of 185 pregnant women using SPF 50+, with no comparison group. 60
- Megaw and colleagues, 2021, Scotland. More first-trimester sunshine went with fewer preterm births in 556,376 births. 11
- Hastie and colleagues, 2021, Scotland. More ultraviolet B in pregnancy went with less preeclampsia in 522,896 deliveries. 63
Definitions
- Acute bilirubin encephalopathy. The early brain injury of high bilirubin in a newborn, which can pass or can lead to kernicterus.
- Bilirubin. A yellow pigment made when the body breaks down heme from red blood cells. The liver prepares it for removal.
- Cochrane review. A systematic review of the randomized trials on one question, done to the methods of the Cochrane collaboration.
- Confidence interval. The range of values the true effect could plausibly take, given the data.
- Exchange transfusion. Replacing a baby's blood, a little at a time, with donor blood to remove bilirubin.
- Filtered-sunlight phototherapy. Treating jaundice with sunlight passed through window-tinting film that blocks most ultraviolet and some heat.
- G6PD deficiency. An inherited shortage of the enzyme glucose-6-phosphate dehydrogenase, which lets red blood cells break down more easily and raises the risk of severe jaundice.
- Geometric mean. An average taken on a multiplied scale, used for measurements that a few very high values would skew.
- Irradiance. How much light of a given color falls on a surface, here in microwatts per square centimeter per nanometer.
- Jaundice. Yellowing of the skin and eyes from bilirubin in the blood.
- Kernicterus. Lasting brain damage from high bilirubin, with cerebral palsy, hearing loss, and trouble looking upward.
- Melasma. Brown to gray-brown patches on the face, common in pregnancy.
- Neonatal lupus. A rash, and sometimes heart or blood problems, in a baby whose mother's antibodies crossed the placenta.
- Neural tube defect. A birth defect of the spine or brain, such as spina bifida.
- Odds ratio. The odds of an outcome in one group divided by the odds in another. 1.0 means no difference.
- Oxybenzone. A sunscreen chemical, also called benzophenone-3, that absorbs UVB and some UVA.
- Phototherapy. Treating jaundice with light, now usually blue to blue-green light of about 460 to 490 nanometers.
- Preeclampsia. A dangerous rise in blood pressure in the second half of pregnancy.
- Preterm birth. Birth before 37 weeks of pregnancy.
- Randomized trial. A study that assigns people to groups by chance, so that the groups differ only in the treatment given.
- Rhesus disease. Destruction of a baby's red blood cells by antibodies from the mother, which can cause severe jaundice.
- Small for gestational age. A baby smaller than most babies born at the same stage of pregnancy.
- Standard erythema dose. A fixed amount of ultraviolet, weighted by how strongly each wavelength reddens skin.
- Stratum corneum. The outermost layer of dead skin cells, the skin's main barrier.
- Unconjugated bilirubin. Bilirubin that the liver has not yet processed, the form that phototherapy acts on.
- UVA. Ultraviolet light from 315 to 400 nanometers, the longer-wave part that reaches deeper into skin.
- UVB. Ultraviolet light from 280 to 315 nanometers, the shorter-wave part that causes most sunburn.
Statistics
In humanized UGT1 mice, a strain bred to get newborn jaundice, 60 minutes under a UVB lamp at 3 days old lowered total serum bilirubin from an average of 9.2 to 6.2 mg/dl one day later (n = 8). 15
In the same newborn mice, the UVB lamp raised the skin's level of UGT1A1, the only enzyme that clears bilirubin, to 3 times that of unexposed mice. 15
In humanized UGT1 mice, skin carried 2.5 times as much UGT1A1 as the liver at 3 days after birth and 4.5 times as much at 14 days, the age these mice develop severe jaundice. 15
In adult human tissue, the liver expressed UGT1A1, the only enzyme that clears bilirubin, at 20,000 times the level found in skin. 15
Skin from UVB-treated newborn mice cleared the UGT1A1 test substrate at 1.5 pmol/min per milligram of microsomal protein, against 1.0 in untreated mice (3 samples each). 15
In cultured human skin cells, tryptophan exposed to a UVB lamp raised UGT1A1 activity from 1.9 to 5.5 pmol/min per milligram of microsomal protein. 15
In Perth, 82 infants who wore UV dosimeters from birth to 3 months of age received a median of 952 J/m2 of real sun, typically on the face, hands, and arms. 41
In the Perth infants, winter-born babies received a median of 1018 J/m2 of sun in their first 3 months and summer-born babies 1203 J/m2, a difference too small to reach significance. 41
In the Perth infants, measured sun dose in the first 3 months did not correlate significantly with blood 25(OH)D at 3 months (rho 0.250, 61 infants) or 6 months (rho 0.245, 62 infants). 41
In the Perth trial, the 16 infants with doctor-diagnosed eczema by 6 months had received a median of 555 J/m2 of sun in their first 3 months, against 998 J/m2 in the 64 infants without eczema. 41
In the Perth trial of infant vitamin D drops (400 IU a day) against placebo, 35 infants had doctor-diagnosed eczema by 6 months: 19 (21.8 percent) on vitamin D and 16 (19.3 percent) on placebo, no significant difference. 41
Followed to 2.5 years, the Perth children with doctor-diagnosed eczema had received a median of 747 J/m2 of real sun in their first 3 months, against 1204 J/m2 for children without eczema (40 and 39 children with dosimeter data). 42
Eczema was the commonest allergic disease in the Perth cohort, reaching 81 of 164 children (49.1 percent) by 2.5 years of age. 42
Parents of the Perth children reported a median of 65 to 75 minutes a day outdoors at 3 months of age, rising to 150 to 160 minutes a day at 30 months. 42
In their first 3 months, the Perth children later diagnosed with eczema had spent a mean of 7 minutes a day outdoors between 11 a.m. and 3 p.m., against 20 minutes for children without eczema (79 and 84 children). 42
At 3 months of age, parents reported never putting sunscreen on 60 of 84 (71.4 percent) Perth infants who stayed free of eczema and 62 of 79 (78.5 percent) who later had eczema. 42
By 30 months, only 2 of 84 (2.4 percent) and 4 of 76 (5.3 percent) Perth children in the two eczema groups never had sunscreen, and about a third always did. 42
The Xi'an infants in the sun group got 2 to 14 hours of real sun a week, 30 minutes at a time two or three times a day, against under 2 hours a week in the control infants, with the mothers deciding how much. 44
Of the Xi'an infants of depressed mothers, the 50 who got 7 to 14 hours of sun a week had the highest motor scores and the lowest hair cortisol, against 45 at 2 to 7 hours and 103 at under 2 hours. 44
The Xi'an researchers saw no rash, dehydration, or other adverse reaction to sun in any of the newborns exposed for up to 14 hours a week. 44
In Xi'an, China, the 95 newborns of depressed mothers who got real sun plus 400 IU a day of vitamin D had better early motor development than the 103 who got vitamin D alone at 400 or 1000 IU. 44
In Tampere, Finland, at latitude 61.5, daily light ranged from 5.3 hours in the darkest winter weeks to 19.5 hours in summer, the setting for a study of 85 infants' sleep by birth season. 78
In overnight sleep recordings of one-month-old Tampere infants, the 15 born in summer spent 40.6 percent of sleep in stage R, the forerunner of REM sleep, against 33.3 percent in the 16 born in winter. 78
The summer-born Finnish infants spent 17.4 percent of sleep in transitional sleep, a less mature stage, against 24.3 percent in winter-born infants. 78
The summer-born Finnish infants woke a median 57 times in the recorded night against 74 times for the winter-born. 78
In two of five infants with neonatal lupus erythematosus, the rash did not appear until the babies had been in direct sunlight. 79
A Japanese girl born at 27 weeks weighing 542 g, already on oral zinc, reached a serum zinc of 427 µg/dL after zinc oxide ointment went on her diaper rash. 51
Of three SPF 15+ sunscreen brands sold in a toddler or baby version and an adult version, each pair held the same filters at the same strength, with oxybenzone at 20 to 60 g per litre. 50
In 199 healthy Philadelphia-area infants aged 6 to 12 weeks, median urinary oxybenzone was 3.6 µg/L. 52
In 61 Danish infants sampled twice, daily urinary excretion of benzophenones was higher or similar while exclusively breastfed compared with after mixed food began. 80
In northern Puerto Rico, the 30 pregnant women who reported sunscreen in the previous 48 hours had urinary oxybenzone at a geometric mean of 187.2 ng/mL, against 39.5 ng/mL in the 869 who did not, about 5 times higher. 70
In 3,619 mother and infant pairs from 11 US cohorts, each tenfold rise in a mother's pregnancy-average oxybenzone went with 29 percent higher odds of a baby small for gestational age. 71
In the same 3,619 US pregnancies, oxybenzone showed no link with preterm birth or low birth weight. 71
In 922 pregnant women in Puerto Rico, each interquartile rise in average urinary oxybenzone went with a pregnancy 1.90 days longer. 72
In the 922 Puerto Rico pregnancies, higher average oxybenzone went with lower odds of preterm birth, an odds ratio of 0.54 per interquartile rise. 72
In 476 Boston mothers whose urine was sampled three times in pregnancy, oxybenzone showed no link with fetal weight in either sex. 73
In 900 pregnancies in the LIFECODES fetal growth case-cohort, median urinary oxybenzone was 37.4 ng/mL in mothers of small-for-age babies, against 49.0 ng/mL for average-size and 44.0 ng/mL for large babies. 74
In the same 900 LIFECODES pregnancies, higher urinary oxybenzone went with a 0.09 larger estimated fetal weight z-score on ultrasound. 74
In 478 pregnant women in the French SEPAGES cohort, each measured from two week-long pools of 21 urine samples, oxybenzone went with larger fetal growth measures on ultrasound in boys but not girls. 75
In 191 French mothers of boys, each unit rise in log urinary oxybenzone went with 26 g more birth weight, a confidence interval of minus 2 to 54 g that crosses zero. 76