Sunscreen use and increased duration of intentional sun exposure: Still a burning issue
Sunscreen use is often proposed for sun protection because of their ability to block UV-induced sunburns (the sun protection factor – SPF). Among suntan seekers, however, risk of cutaneous melanoma may be increased because of extended sun exposure duration. We made a systematic review of the evidence linking sunscreen use to sun exposure duration. Five observational studies found that when sun exposure was associated with willingness to get a tan or to stay longer in the sun (i.e., intentional sun exposure), sunscreen use was associated with duration of sun exposure 13–39% longer. Paradoxically, sunburns tend to be more frequent among sunscreen users, probably because of greater natural sun sensitivity. When sun exposure was not intentional, sunscreen use did not increase time spent in the sun. Two European double-blind randomized trials conducted among young sun seekers found daily sun exposure duration, especially sunbathing, 19–25% longer with use of SPF 30 than with use of SPF 10 sunscreens. One randomized trial in a holiday resort in France found a 3–13% increase in sun exposure duration with use of SPF 12 versus SPF 40 sunscreen. But, the SPF 12 groups used 3.6–4.2 more sunscreen than the SPF 40 group, and thus the actual SPF in the SPF 12 group was higher than in the SPF 40 groups. In conclusion, sunscreen use leads to longer duration of sun exposure when sun exposure is intentional, but not when sun exposure is non intentional. © 2007 Wiley-Liss, Inc.
Sunscreens were primarily designed for sunburn prevention, but animal and human experiments showed their ability to reduce UV-induced skin lesions such as solar keratoses and squamous cell carcinoma (SCC).1, 2 The protection afforded increases with the sun protection factor (SPF), i.e., the ability of a sunscreen to retard UV-induced skin erythemal reaction. Consequently, high SPF sunscreens (i.e., SPF ≥ 15) have often been recommended for sun protection.
Sunscreen use for sun protection has been challenged by repeated observation that not only sunscreen use (including recent high-SPF broad-band sunscreens) did not protect against cutaneous malignant melanoma (melanoma), basal cell carcinoma (BCC), and higher nevus counts, but that it was often associated with increased risk of these tumors, whereas wearing of clothes was associated with no change or with decrease of these tumors.3, 4, 5
An alternative hypothesis was that sunscreen use could encourage sun exposures of longer duration, possibly leading to increased risk of melanoma and of BCC. In this article, we review observational and randomized studies that examined sunscreen use and sun exposure duration, according to the sun exposure type associated with sunscreen use.
Intentional and nonintentional types of sun-exposure
Substantial proportions of melanoma and BCC are associated with intermittent sun exposure rather than with lifetime accumulated sun exposure,6i.e., light-skinned subjects spending most of their daily life time indoor but enjoying intense sun exposure during holidays, and often eager to acquire a sun tan. Intermittent sun exposure is thus often intentional as subjects look for a biological effect.1, 2 During intentional sun exposure (ISE), significant portions of the trunk, shoulders, and of the upper parts of limbs are frequently uncovered. Sunbathing is the most typical ISE behavior.
Nonintentional sun exposure (NISE) represents sun exposure during daily life activities, without special willingness to acquire a tan or to being able to spend long time in the sun. During NISE, skin areas most usually sun exposed are the head and neck, the hands, and the forearms. Examples of NISE are outdoor activities such as walking, hiking, gardening, skiing, or work on building construction sites or in farming fields. Lifetime accumulated NISE is mainly associated with occurrence of solar keratoses and of squamous cell carcinoma (SCC).
Methods for literature search
We started the literature search with materials gathered for the IARC Handbook on Sunscreens1 and with bibliography gathered by authors. We then performed a systematic literature search in the MEDLINE until August 2006 without restriction on type and language of article. Use of variable combinations of MeSH terms “sunscreen agent”, “sunscreening agents”, “sunlight”, “sunburns,” and “time” until March 2006 conducted to a selection of 155 articles including words in the title or in the abstract (when available) suggesting relevance for the study. Full copies of these articles were obtained and independently revised by P.A. and M.B. We made a similar search in the ISI Web of Knowledge, Science Citation Index Expanded, covering the science literature from 1945 until August 2006. Examination of title and available abstracts of articles did not conduct to finding further articles than those found using the MEDLINE. Data from relevant articles were abstracted in a table summarizing key variables and results. Reported data had to provide or to allow the calculation of time spent in the sun during parts of day or during days during which there was effective sun exposure, with knowledge of sunscreen use before or during effective sun exposure. Relevant information of methods or on results were sometimes found in the Discussion section of articles, e.g., time spent in the sun in 1 Danish study,7 or the notion that during a randomized trial in France, an investigator was permanently present in holiday villages and had daily contacts with trial participants.8
Observational studies on sunscreen use during ISE
We identified 6 observational studies conducted during predominantly ISE situations that measured time spent in the sun according to sunscreen use, and published in 7 articles.7, 9, 10, 11, 12, 13, 14 A cross- sectional study among Norwegian adolescents13 could not be used because data on sunbathing time were not reported according to sunscreen of to sun protection factor (SPF) used, and skin lotions with SPF 0–2 were incorrectly considered as sunscreens. Table I summarizes the 5 studies with relevant data.
Table I. Observational Studies on Sunscreen Use and Duration of Sun Exposure or of Exposure to Ultraviolet (UV) Radiation
|Berwick et al., 19929||Cross-sectional survey||Participants to a skin screening campaign in Connecticut, USA, 1988||153 Adults of all ages||Questionnaire on summer of previous year||Average sun exposure duration per weekend day2||Never||2.57 hr2||Ref.||NR|
|p = 0.06|
|Stender et al., 199610; Wulf et al., 199711||Cross-sectional survey||Beaches and parks in Denmark in July 1994||805 Sunbathing Caucasians (mean age: 28) on 4 beaches and in a park||(1) Questionnaire during one sunny day||Difference in minutes between time of arrival at location and time for expected departure||No||197 min||Ref.||34%3|
|p = 0.186||NR|
|(2) Single UV-dosimeter installed on the location||UV dose (SED5) recorded between time of arrival at location and time for expected departure||No||4.5 SED||Ref.||–|
|p = 0.035|
|McCarthy etal., 199912||Cross-sectional survey||Beaches in Texas, USA, in July 1997||55 Beachgoers16–59 years old||Questionnaire on sunbathing activities of the day to subjects preparing to leave the beach||Time (hr) of sunbathing||None||2.8||Ref.||54%|
|SPF > 15||3.5||25%||73%|
|Robinson et al., 200014||Nationwide cross-sectional survey||Representative sample of families in USA, in July-August 1997||503 children <13 years||Questionnaire toparents on Mondays following a weekend||Mean time outdoor between 10 am and 4 pm||No||4.6 hr||Ref.||7%|
|p < 0.05||p < 0.05|
|Thieden etal., 20057||Prospective study of subjects||Danish citizens 4 to 68 years old during 3 summer seasons, 1999–2001|| ||(1) Daily diaries||Number of sun exposure hours per risk behaviour day4||No||4.1||Ref.||4.8%6|
| ||(2) Recordings of individual UV dosimeters during risk behaviour days in Southern Europe||Median SED per risk behaviour days4|| || || || |
|93 Subjects|| || ||No||3.5 SED||Ref.||NR|
|19 Children|| || ||No||1.5 SED||Ref.||NR|
|p < 0.05|
|22 Sun worshipper|| || ||No||1.8 SED||Ref.||NR|
| || || ||8 Gardeners|| || ||No||8.1 SED||Ref.||NR|
| || || ||Yes||10.2 SED||26%||NR|
| || || ||NR|| |
In some studies, statistical tests for some results were not reported, or data reporting precluded statistical analysis. In all 5 studies, for adults and for children, sunscreen use was associated with duration of sun exposure 13–39% longer than if no sunscreen was used. One study found that UV doses received were considerably higher when sunscreens were used.7 Four studies recorded sunburns and found higher proportions of subjects with sunburn when a sunscreen was used, especially when the SPF was high. According to reports, differences in sun sensitivity between sunscreen users and nonusers were not likely reasons for explaining results on durations and on sunburns, although these results were never statistically adjusted on sun sensitivity of study participants. Interestingly, 1 study7 examined also gardeners (a NISE behavior), and found no difference in UV dose received according to sunscreen use. This study did not report data on sunburn occurrence during NISE.
These observational studies could however not assess whether longer duration was a result of sunscreen use that sunscreen users were not aware of, or a result of the willingness of sunscreen users to spent long time in the sun without (did they believe) incurring sunburn. Qualification of the exact cause-effect chain of events could only be determined by randomized trials.
The European randomized trials
The European Organization for Research and Treatment of Cancer (EORTC) Melanoma Group conducted 2 double-blind randomized trials among students 18–24 years of age eager to engage in intentional sun exposure during their summer holidays (Table II).16, 17 These trials were representative of sun exposure behaviors of millions of light-skinned young subjects eager to acquire a tan during holidays or during leisure times. The 2 trials demonstrated that sun exposure, mainly sunbathing, was 19–25% longer duration with use of SPF 30 than with use of SPF 10 sunscreens. High SPF susncreens also allowed more hazardous sun exposure behaviors that would not be possible otherwise, like for instance sunbathing with naked breasts. Average quantity of sunscreen used was similar in randomization groups and the in-holiday sunburn experience was identical for participants regardless of the SPF of the sunscreen used. For many participants, sunburn occurrence was the factor limiting sun exposure duration, but sunburn occurred later in the high SPF group than in the low SPF group. Hence, it was sunscreen use during ISE that led to longer sun exposure sessions, without affecting sunburn occurrence, and trial participants were not aware of this increase in sun exposure duration.
Table II. Randomised Trials on Use of Sunscreens Having Different Sun Protection Factor (SPF) and Duration of Sun Exposure or of Exposure to Ultraviolet (UV) Radiation
|Autier et al., 199916||Double-blind controlled randomized trial||Sunny resorts where students spent their holiday in July–August 1997||87 French and Swiss students 18 to 24 years old taking holidays in sunny resorts in July–August||Self-administered daily diaries||Mean hours of sun bathing per day with sunbathing (average of 11 days)||SPF 10||2.6 hr||Ref.||47%||72.3 (mean)|
| || || || || || ||SPF 30||3.1 hr||19%||43%||71.6 (mean)|
| || || || || || || || ||p = 0.0013||p=0.90||p = 0.95|
|Autier et al., 200017||Double-blind controlled randomized trial||Sunny resorts where students spent their holiday in July–August 1998||48 French and Belgian students 18 to 24 years old taking holidays in sunny resorts in July–August||(1) Self-administered daily diaries||Mean hours of sun bathing per day with sunbathing (average of 9 days)||SPF 10||2.4 hr||Ref.||38%||67 (mean)|
| ||SPF 30||3.0 hr||25%||41%||77 (mean)|
|Test for difference between groups|| || ||p = 0.054||p = 0.46||p = 0.22|
|(2) Individual UVB and UVA dosimeters|| ||SPF 10||UVB: 841 J/m2; UVA: 727 KJ/m2||Ref.||–||–|
|SPF 30||UVB: 984 J/m2; UVA: 812 KJ/m2||UVB: +17%; UVA: +11%||–||–|
|UVB: p = 0.15; UVA: p = 0.70||–||–|
|Dupuy et al., 20058||Randomized trial with sunscreen tubes labelled as “basic” or “high protection”||Summer villages located on the sea in France in July–August 2001||367 adults 18 to 78 years old (mean = 39) during one week of holidays in July–August||Daily diaries and questionnaires completed with help of on-site investigator||Duration (hr) of sun exposure over one week while wearing a swimming suit or equivalent||SPF 12, “basic protection”||14.6 hr||Ref.||24%||109 (median)|
|SPF 40 “basic protection”||12.9 hr||−12%||14%||30 (median)|
|SPF 40 “high protection”||14.2 hr||−3%||16%||26 (median)|
|p = 0.06 (*)||p = 0.049*||p < 0.001*|
In 2001, a Working Party convened by the IARC concluded that “use of sunscreen can extend the duration of intentional sun exposure, such as sunbathing. Such an extension may increase the risk for cutaneous melanoma.”1, 2 The US National Cancer Institute and the US Preventive Services Task Force came to similar conclusions.18, 19
The French randomized trial
A randomized trial funded by a major sunscreen manufacturer appeared to contradicted results of European trials, finding no significant difference in the duration of time spent in the sun according to sunscreen SPF.8 However, this trial involved a population with low interest in sunbathing. Also the way this trial, was designed and conducted was likely to produce a negative result, i.e., no difference in sun exposure duration according to SPF of sunscreen used.
Notwithstanding study design issues, interpretation of results may be different than that provided by authors. The French trial reported that the SPF 12 and SPF 40 sunscreens especially made for the study had different textures, SPF 12 being easier to spread, which may partly explain the 3.6- to 4.2-fold difference in amount of sunscreen used in the SPF 12 group as compared with the 2 SPF 40 groups (Table II). Taking into account data on sunscreen use in European randomized trials20 and differences in average daily sunbathing duration in European and in the French trials, participants in the SPF 40 groups may have applied between 0.25 and 0.5 mg/cm2 of sunscreen onto their skin. With a 0.25 or 0.5 g/cm2 sunscreen application, the actual SPF is about the eighth and the fourth square root of the SPF indicated on the bottle,12i.e., 1.6 or 2.5. Assuming no difference in sun exposure duration according to SPF used and of exposed skin areas, use of 3.6- to 4.2-fold more SPF 12 sunscreen than of SPF 40 may have resulted in an actual sun protection potency 2–3 times higher in the SPF 12 group than in the SPF 40 groups. The consequence was probably the borderline statistically significant 12% increase in sunbathing duration observed in 1 of the SPF 12 groups (Table II). Hence, results from this French randomized trial were in fact quite similar to results of the European randomized trials.16, 17
Sunscreen use for protection against solar keratoses and squamous cell carcinoma
Trials with solar keratoses and squamous cell carcinoma (SCC) as endpoint were conducted among older subjects whose sun exposure was not intentional but due to normal circumstances of daily-life. Two trials in volunteers relatively aged and having a history of sun-induced skin damage showed the ability of sunscreen use to reduce new solar keratoses.21, 22 The Nembour trial showed that sunscreen use can reduce the incidence of SCC.23 This trial was performed in Queensland, Australia, in a population living in an area with high ambient sunshine all the year round, and where skin cancer incidence is the highest in the World. In these 3 trials, sun exposure was essentially nonintentional, and sunscreens (or placebo lotions) were mainly used during daily life and applied essentially on the face, ears, neck, and hands. Apparently, subjects did not experience sunburn or the number of sunburns was significantly lower in the intervention group.23, 24 None of the 3 trials reported measurements of in-trial sun exposure durations, but the Nembour trial stated there was no evidence of differences in the time spent in the sun among subjects allocated to the intervention group.23
Randomized trials on sunscreen use and numbers of acquired nevi
In 1998 until 2001, 2 randomized trials tested the ability of broadband sunscreen use on the development of nevi in schoolchildren.25, 26 The Vancouver trial in Canada found a reduction in the development of new nevi in children with dense facial freckling and found no effect in children without dense facial freckling.25 The German trial failed to change patterns of sunscreen use between the randomization groups because apparently, sunscreen use was already highly prevalent in all groups at study start.26 The Vancouver trial reported estimations of the total amounts of time spent in the sun during the 3-year trial duration but did not report duration of sun exposure per day with or without sunscreen use.
All available observational and experimental data in humans provided evidence that intentional sun exposure tends to be of longer duration when a sunscreen is used or when SPF increases. Results of the European randomized trials suggest that sunscreen users are unaware of the impact sunscreen use has on their sun exposure behaviors.
A paradoxical result of observational studies was the higher numbers of subjects reporting sunburns when a sunscreen was used, mainly when the SPF was high. It is well known that the majority of sunscreen users apply only a fifth to a third of quantities of sunscreens used in laboratory for testing their SPF.11, 20 Also, thickness of sunscreens of a same commercial brand does not change much with SPF, and 1 study showed that quantities of sunscreen applied onto the skin did not vary much with SPF.27 Observational studies did not perform adjustment of their results on natural sun sensitivity of study participants. Hence, the higher number of sunburns among (high SPF) sunscreen users in observational studies could have been due to greater sun sensitivity.
In contrast to what happens during ISE, during NISE, observational and experimental data in humans provide evidence that sunscreen use would not increase time spent in the sun, and would decrease sunburn occurrence. In this respect, impact of sunscreen use during NISE situations would meet expectations raised by laboratory experiments that showed the ability of sunscreens to decrease the incidence of UV-induced skin erythemal reactions and nonmelanocytic skin cancers.1 In ISE situations, these expectations are not met because of the influence sunscreen use has on behaviors of humans eager to get a tan or to stay long in the sun.
During the second half of the nineteenth century, the sun tanning fashion exploded among light-skinned populations and growth of sunscreen commercialization paralleled that fashion.28, 29, 30 Sunscreen are often considered as tanning aid,1, 31 and advertising sometimes persuade sun seekers that sunscreens may ensure acquisition of a “safe tan”. In spite of uncertainties about their exact role in melanoma and BCC occurrence, and in spite of recommendations that sunscreen use should just be an adjunct to other more natural forms of protection, such as use of hats, shirts, and search for shade, sunscreens remain the most frequently used sun protection method, mainly among adolescents and young adults, while in the same time, younger adults declare to be more likely to sunbathe deliberately than other people.32 In Australia sun protection no longer relies on sunscreen use.33
In conclusion, examination of studies on sun protection methods should always take into account the type of sun exposure that was addressed. Also, information on sunscreens should make a clear difference between situations of intentional or of nonintentional sun exposure. When intentional sun exposure is concerned, information to the general public should be closer to uncertainties on their efficacy and to knowledge of the possible impact they may have on sun exposure behaviors, and on melanoma risk.