Association of Sun-Protective Behaviors With Bone Mineral Density and Osteoporotic Bone Fractures in US Adults

This cross-sectional study uses data from the National Health and Nutrition Examination Survey to investigate the association of sun-protective behaviors with bone mineral density and the prevalence of osteoporotic fractures among US adults.

Key Points

Question

Does sun protection increase the risk of osteoporosis and/or osteoporotic bone fractures?

Findings

In this population-based cross-sectional study of 3418 US adults 20 years and older, sun-protective behaviors (ie, staying in the shade, wearing long sleeves, applying sunscreen) were not associated with diminished site-specific bone mineral density or increased prevalence of osteoporotic bone fractures.

Meaning

These findings support the safety and continued practice of sun-protective behaviors, with no considerable negative association with bone health or the risk of osteoporotic fractures.

Abstract

Importance

Despite the reassuring emerging evidence on the lack of a causal relationship between sun protection and vitamin D deficiency, there is scarce data on whether multimodal sun protection is associated with reduced bone mineral density (BMD) and/or increased prevalence of osteoporotic bone fractures. This lack of data may lead to worry and decreased sun-protective behaviors on the part of patients.

Objective

To investigate the association of sun-protective behaviors with BMD z scores and the prevalence of osteoporotic fractures.

Design, Setting, and Participants

This population-based cross-sectional study included data from US adults who participated in the 2017 to 2018 cycle of the National Health and Nutrition Examination Survey (NHANES). Data were analyzed between September and November 2020.

Main Outcomes and Measures

Definition of sun-protective behaviors (staying in the shade, wearing long sleeves, and sunscreen use), site-specific and total BMD, and osteoporotic fractures (hip, wrist, and spine) in the NHANES data.

Results

Data from 3418 adults 20 years and older (average age, 39.5 [95% CI, 38.6-40.4] years; 1612 [47.2%] men and 1806 [52.9%] women) who completed the NHANES dermatology questionnaire were included in this study. The prevalence of frequent staying in the shade, wearing of long sleeves, and sunscreen use were 31.6% (95% CI, 27.8%-35.7%), 11.8% (95% CI, 10.6%-13.1%), and 26.1% (95% CI, 23.5%-28.8%), respectively. The use of individual sun-protective behaviors was not associated with diminished site-specific and total BMD z scores in the multivariate models (estimate, −0.23 [95% CI, −0.47 to 0.02], P = .18; −0.08 [−0.27 to 0.12], P = .72; and −0.10 [−0.32 to 0.13], P = .15 for frequent staying in the shade, wearing of long sleeves, and sunscreen use, respectively). Moderate to frequent staying in the shade was associated with reduced prevalence of spine fractures in the multivariate model (odds ratio, 0.19 [95% CI, 0.04-0.86], P = 0.02).

Conclusion and Relevance

In this cross-sectional study, routine use of sun-protective behaviors among the US adult population was not associated with decreased BMD or increased risk of osteoporotic fracture. Sun protection may be associated with a modest decrease in the prevalence of osteoporotic fractures, possibly owing to risk-averse behaviors. These reassuring findings add to the growing body of evidence on the safety of sun protection, with no considerable negative association with bone health.

Introduction

The use of sunscreen and other sun-protective measures is clinically recommended for risk reduction of melanoma and nonmelanomatous skin cancers and for accelerated skin-aging prevention.^1,2 Given that the UV irradiation range (approximately 290-320 nm) to induce effective vitamin D photosynthesis is within the photocarcinogenic UV-B spectrum, a potential concern for the use of sunscreens and other sun-protective measures to induce vitamin D deficiency has arisen among both the medical and lay communities.³

Recent reassuring information has been gained from meticulously designed studies on the use of sunscreen and other sun-protective measures with respect to their association with vitamin D status. Libon et al⁴ demonstrated that the use of sun protection factor (SPF) 50 and higher sunscreens block cutaneous vitamin D production, with only a negligible association with circulating 25-hydroxyvitamin D levels. Optimal sunscreen use, applied by an appropriate thickness and with a very high UV-A–protection factor allowing minimal skin sunburn, resulted in improved vitamin D synthesis during a weeklong sun holiday in another study.⁵ In light of these new findings, an international panel of multidisciplinary experts concluded that daily sunscreen use for recreational sun protection does not compromise vitamin D synthesis, even when applied under optimal conditions.⁶

However, little is known regarding the potential effects of sunscreen use and other sun-protective behaviors on the clinically relevant potential complications of vitamin D deficiency, such as diminished bone mineral density (BMD) and the risk of osteoporotic fractures. One small study found no evidence of substantial bone loss after 2 years with the use of sunscreen SPF 15 in a clinic-based sample of 28 elderly patients and their healthy controls,⁷ but, to our knowledge, no prior population-based study has explored this scenario. This potential relationship may affect sun-protective behaviors in the lay population; therefore, further data regarding any potential effect is essential.

To address this gap of knowledge, in this cross-sectional study, we included and analyzed data from the 2017 to 2018 cycle of the National Health and Nutrition Examination Survey (NHANES) to determine correlates of sun protection and to investigate any potential association of multimodal sun-protective behaviors with BMD and the prevalence of osteoporotic bone fractures among US adults.

Methods

Survey Details

The NHANES captures nationally representative statistics of the US noninstitutionalized civilian population on a biennial basis based on complex survey design and population-specific sample weights. Briefly, a series of sampled household interviews and standardized physical examinations in designated mobile examination centers were arranged across the country. We extracted and analyzed information from the most recent cycle of the NHANES (2017-2018), comprising data from 5569 individuals 20 years and older. Of these individuals, 3418 (aged 20-59 years) completed the dermatology questionnaire during the interview part of the survey. The final subsamples selected to analyze the associations of sun-protective behaviors with BMD and osteoporotic bone fractures contained data from 2585 (aged 20-59 years) and 917 (aged 50-59 years) individuals, respectively. Owing to use of publicly available, deidentified data released by the Centers for Disease Control and Prevention, institutional review board approval was not necessary.

Sun-Protective Behaviors, Sun Sensitivity, and Sunburns

Sun sensitivity was defined based on sun exposure (ie, “going out in the sun”) for 30 minutes without any sun protection after several months of not going out in the sun (Table 1). Subsequently, survey respondents were categorized as severely sun sensitive (ie, “severe sunburn with blisters” and “severe sunburn for a few days with peeling”), mildly sun sensitive (ie, “mildly burn with some tanning” and “turn darker without a sunburn”), and not sun sensitive (ie, “nothing would happen”) according to their answers (Table 1). History of recent sunburn was asked using the question “In the past year did you have sunburn?” and the answers were recorded as the number of sunburns or never (Table 1). To characterize sun-protective behaviors, we used 3 practices collected from participants during the interview part of the survey: (1) staying in the shade, (2) wearing a long-sleeved shirt, and (3) using sunscreen on a very sunny day. All 3 practices were recorded in the survey as “always,” “most of the time,” “sometimes,” “rarely,” and “never.” All recorded answers within 3 major categories were identified: frequent (always or most of the time), moderate (sometimes), or rare (never or rarely) (Table 1). Number of multimodal sun-protection measures was calculated from 0 to 3, according to the number of behaviors with frequent use in survey respondents.

Table 1. National Health and Nutrition Examination Survey Sun-Related Terminologies and Questions.

Terminology and survey question	Categories of participants’ responses
Sun sensitivity If after several months of not being in the sun, you then went out in the sun without sunscreen or protective clothing for a half hour, which one of these would happen to your skin?	Severe: “I get a severe sunburn with blisters” or “I get a severe sunburn for a few days with peeling” Mild: “I get mildly burned with some tanning” or “I turn darker without a sunburn” None: “Nothing would happen in half an hour” Excluded responses: “I don't know,” “Other,” “I refuse to answer,” and missing responses
Sunburn In the past year did you have sunburn?	“Yes” (numbers of sunburns recorded) “Never” Excluded responses: “I don't know,” “I refuse to answer,” and missing responses
Staying in the shade When you go outside on a very sunny day, for more than one hour, how often do you stay in the shade?	Frequent: “Always” or “Most of the time” Moderate: “Sometimes” Rare: “Never” or “Rarely” Excluded responses: “I don't know,” “I refuse to answer,” and missing responses
Wearing long sleeves … wearing a long-sleeved shirt?
Using sunscreen … using sunscreen?

Dual X-ray Absorptiometry BMD and Osteoporotic Fractures

The BMD (measured in grams/cm²) was evaluated using a dual x-ray absorptiometry technique (QDR 4500A fan-beam densitometers [Hologic Inc]) while the participants were visiting mobile examination centers.⁸ The mean BMD for the scanned anteroposterior length of L1 to L4 was computed and used for lumbar spine BMD reporting. The left hip (or right hip, in case of left hip arthroplasty or metal object injection) was routinely scanned to report total femur, femoral neck, and trochanter BMD. Exclusion criteria for participant dual x-ray absorptiometry assessment followed those of the NHANES recommendations.⁹ To determine self-reported occurrence of osteoporotic fractures, participants were asked, “Has a doctor ever told you that you had a broken or fractured hip/wrist/spine?” and answers of “yes” were recorded as positive.

Physical Activity

Weekly metabolic equivalent task (MET)-minute aggregated scores were calculated to measure the amount of energy expenditure used for physical activity.^10,11 Serving as a metabolic equivalent, MET represents the ratio of metabolic rate during activity and at rest.¹² The NHANES recommendations were followed to quantify weekly MET-minute composite scores based on the following formula¹⁰: (8.0 MET scores × minutes of vigorous work-related activity) + (4.0 MET scores × minutes of moderate work-related activity) + (4.0 MET scores × minutes of walking or bicycling for transportation) + (8.0 MET scores × minutes of vigorous leisure-time physical activity) + (4.0 MET scores × minutes of moderate leisure-time [recreational] physical activity).

Other Covariates

Milk consumption was defined based on answer into 3 categories of never/rarely (“never” or “less than once a week”), sometimes (“once or more a week, but less than once a day”), and often (“once a day or more”) in the past 30 days. Total vitamin D supplement intake (inclusive of D2 and D3) was determined based on dietary records of participants in the past 30 days prior to study. Intake of prednisone or cortisone was based on answering “yes” to the question “Have you ever taken any prednisone or cortisone pills nearly every day for a month or longer? [Prednisone and cortisone are types of steroids.]”.

Statistical Analysis

All statistical tests were carried out using the complex survey design component of SPSS, version 25.0 for Windows (IBM). Categorical outcomes are reported as percentage frequency (95% CI) and continuous variables as mean (95% CI). Differences in the pattern of multimodal sun-protective behaviors across binary categories (ie, gender, country of birth, annual household income) were evaluated using the design-based Rao-Scott–adjusted χ² test. Baseline and multivariate survey linear regression analyses were performed to evaluate correlations of sun-protective behaviors with site-specific and total BMD z scores, and results were reported by estimates and 95% CIs. Unadjusted odds ratios or multivariate adjusted odds ratios and 95% CIs were calculated to determine the association of sun-protective behaviors with osteoporotic bone fractures of the hip, wrist, and spine in design-based logistic regression models. Multivariate adjustments were performed to control for potential confounders (ie, gender, age, race and ethnicity, education level, country of birth, annual household income, milk consumption, current smoker status, body mass index [BMI; calculated as weight in kilograms divided by height in meters squared], total serum calcium, serum alkaline phosphatase, and MET-minute physical activity scores) in all multivariate regression models. P value for trend was calculated by treating the continuous or categorical outcomes of interest as dependent variables in linear regression. In all tests, P < .05 was considered statistically significant.

Results

Demographics, Sun-Protective Behaviors, Sun Sensitivity, and Sunburns

Data from 3418 adults 20 years and older (average age, 39.5 [95% CI, 38.6-40.4] years; 1612 [47.2%] men and 1806 [52.9%] women) who completed the NHANES dermatology questionnaire were included in this study. The final sample size for analysis included data from 3403 individuals with valid answer choices on all 3 sun-protection components of the dermatology questionnaire (Tables 1 and 2). The prevalence for frequent (ie, most of the time or always) staying in the shade, wearing of long-sleeve clothing, and use of sunscreen were 31.6% (95% CI, 27.8%-35.7%), 11.8% (95% CI, 10.6%-13.1%), and 26.1% (95% CI, 23.5%-28.8%), respectively. Descriptive study characteristics according to number of sun-protective behaviors among survey participants are summarized in Table 2. US adults practicing an increasing number of sun-protective behaviors were more likely to be older and female, originate from Mexican American or non-Hispanic Asian descents, be educated with advanced degrees at college level or above, be born abroad, be nonsmokers, engage in less intense physical activity, experience sunburns, spend less minutes outdoors, and have mild to severe sun sensitivity, lower BMI scores, higher total serum calcium, higher serum alkaline phosphatase, and lower serum albumin concentrations (Table 2). The frequency of sunburns was significantly elevated among individuals with severe (72.9% [95% CI, 65.5%-79.2%]) and mild (52.3% [95% CI, 48.1%-56.4%]) sun sensitivity compared with individuals who reported no sun sensitivity (23.3% [95% CI, 19.5%-27.5%]; P < .001).

Table 2. Characteristics of Survey Participants According to the Pattern of Their Multimodal Sun-Protective Behaviors (N = 3403).

Characteristic	No. reported	No. of sun-protective behaviors, % (95% CI)a			P value
		0	1	2-3
Prevalence in NHANES, 2017-2018	3403	49.0 (45.3-52.7)	35.3 (32.7-38.0)	15.7 (13.8-17.9)	NA
Age, mean (95% CI), y	3403	39.1 (38.1-40.1)	39.0 (37.7-40.3)	41.9 (40.8-43.1)	.004b
Gender	3403
Male	1607	59.6 (55.8-63.4)	30.6 (26.4-35.1)	9.8 (7.7-12.4)	<.001
Female	1796	38.7 (34.0-43.6)	39.8 (36.7-43.1)	21.5 (17.8-25.7)
Race and ethnicity	3403
Mexican American	515	40.1 (34.1-46.4)	42.6 (37.4-48.0)	17.3 (12.4-23.7)	.01
Non-Hispanic Asian	578	39.4 (34.9-44.2)	32.8 (26.7-39.5)	27.8 (22.4-33.9)
Non-Hispanic Black	777	51.6 (47.0-56.2)	39.0 (34.9-43.2)	9.4 (7.8-11.2)
Non-Hispanic White	1021	51.7 (46.1-57.3)	33.7 (29.1-38.6)	14.6 (11.6-18.1)
Other Hispanic	319	46.7 (39.3-54.2)	29.8 (22.6-38.0)	23.6 (16.4-32.7)
Other, including multiracialc	193	45.9 (32.7-59.7)	40.2 (27.0-55.1)	13.8 (5.3-31.6)
Education level	3401
>High school	217	49.0 (39.2-58.8)	36.1 (29.1-43.6)	15.0 (10.0-21.9)	.02b
Some high school	371	54.2 (45.0-63.2)	34.4 (27.3-42.3)	11.4 (8.1-15.7)
High school diploma	792	53.2 (45.3-60.8)	33.9 (28.7-39.5)	12.9 (9.3-17.7)
College or above	2021	46.7 (43.0-50.4)	35.9 (32.6-39.3)	17.5 (15.4-19.7)
US born	3403
Yes	2256	51.7 (47.4-55.9)	34.9 (31.8-38.3)	13.4 (11.4-15.7)	<.001
No	1147	39.0 (35.0-43.1)	36.6 (32.1-41.3)	24.4 (19.0-30.8)
Annual household income	2929
<$55 000	1542	52.0 (48.0-56.1)	35.2 (32.1-38.4)	12.8 (10.4-15.6)	.21
≥$55 000	1387	48.0 (43.0-53.0)	36.2 (32.0-40.7)	15.8 (13.2-18.7)
Current smoker status	1269
Every day	564	63.7 (58.5-68.6)	26.5 (22.7-30.7)	9.8 (6.5-14.6)	.001b
Some days	154	55.5 (40.7-69.3)	33.3 (24.3-43.8)	11.2 (5.5-21.4)
Not at all	551	47.7 (40.9-54.5)	37.5 (30.8-44.6)	14.9 (10.7-20.4)
Sun sensitivity	3357
Severe	322	29.2 (24.3-34.6)	43.5 (37.7-49.5)	27.4 (22.0-33.5)	<.001b
Mild	1560	49.7 (44.3-55.1)	34.8 (30.7-39.1)	15.5 (12.8-18.6)
Nonsensitive	1465	57.5 (53.4-61.4)	32.2 (29.3-35.3)	10.3 (7.8-13.5)
Sunburn in the past year	3393
Yes	1121	49.8 (44.2-55.5)	37.1 (32.1-42.4)	13.0 (10.5-16.1)	.04
No	2272	48.3 (44.8-51.7)	33.8 (31.8-35.8)	18.0 (15.2-21.1)
No. of sunburns in the past year	3393
0	2272	48.3 (44.8-51.7)	33.8 (31.8-35.8)	18.0 (15.2-21.1)	.09b
1	587	47.8 (41.1-54.5)	38.8 (32.5-45.5)	13.4 (9.8-18.0)
2	285	50.7 (41.5-59.8)	35.3 (29.7-41.3)	14.0 (9.6-20.0)
≥3	249	53.0 (44.2-61.5)	35.8 (26.3-46.5)	11.3 (6.6-18.5)
Minutes spent outdoors, mean (95% CI)	3374	301.4 (275.2-327.6)	237.7 (218.5-257.0)	202.1 (171.1-233.2)	<.001b
Milk consumption
Never/rarely (never or less than once a week)	1524	46.7 (41.7-51.7)	38.9 (34.0-44.1)	14.4 (10.9-18.7)	.28b
Sometimes (once or more a week but less than once a day)	997	52.3 (46.9-57.7)	31.8 (28.8-35.0)	15.9 (11.8-21.0)
Often (once a day or more)	870	49.1 (44.0-54.2)	32.8 (28.4-37.6)	18.1 (14.2-22.8)
Prednisone or cortisone intake	911
Yes	75	44.5 (26.9-63.7)	35.6 (18.6-57.1)	19.9 (7.0-45.2)	.94
No	830	47.7 (41.6-53.8)	32.6 (27.9-37.8)	19.7 (15.6-24.4)
BMI, mean (95% CI)	3198	29.7 (29.1-30.3)	30.5 (29.6-31.4)	29.3 (28.2-30.3)	.02b
Dietary intake vitamin D2 and D3, mean (95% CI), µg	1043	30.5 (23.7-37.3)	50.0 (14.1-85.9)	35.2 (27.8-42.6)	.36b
Total calcium, mean (95% CI), mg/dL	3010	9.3 (9.2-9.3)	9.2 (9.2-9.3)	9.3 (9.3-9.3)	.02b
Phosphorus, mean (95% CI), mg/dL	3010	3.6 (3.5-3.6)	3.6 (3.5-3.6)	3.6 (3.6-3.7)	.27b
Albumin, mean (95% CI), g/dL	3012	4.1 (4.1-4.2)	4.1 (4.0-4.1)	4.1 (4.1-4.1)	.03b
Alkaline phosphatase, mean (95% CI), U/L	3012	75.6 (73.5-77.7)	75.1 (72.1-78.2)	71.0 (68.0-74.1)	.04b
MET minutes, mean (95% CI)	3403	1345.9 (1236.0-1455.8)	1133.2 (1017.0-1249.4)	835.7 (679.9-991.5)	<.001b

Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by height in meters squared; MET, metabolic equivalent task; NA, not applicable; NHANES, National Health and Nutrition Examination Survey.

SI conversion factors: To convert total calcium to mmol/L, multiply by 0.25; phosphorus to mmol/L, multiply by 0.323; albumin to g/L, multiply by 10; and alkaline phosphatase to µkat/L, multiply by 0.0167.

^a Sun-protective behaviors include staying in the shade, wearing long sleeves, and applying sunscreen.

^b Indicates trend test P value.

^c A detailed list of all races and ethnicities included in this category was not provided in the NHANES.

Associations With Dual X-ray Absorptiometry BMD and Osteoporotic Fractures

Baseline and multivariate design–based survey linear regression models demonstrating the association of different sun-protective techniques and BMD z scores at major anatomic sites are summarized in Table 3. While each individual sun-protective behavior was associated with statistically significant decreases in several site-specific and total BMD z scores in baseline models, these associations largely faded following the multivariate adjustments for potential confounders. Notably, association of prednisone intake with the number of sun-protective behaviors was not statistically significant (Table 2). Therefore, prednisone intake was not considered to represent a potentially confounding variable for the associations of sun protection with BMD or osteoporotic fractures, and subsequently was not included in the multivariate linear or logistic regression analyses.

Table 3. Association of Sun-Protective Behaviors With Bone Mineral Density (BMD) z Scores in NHANES, 2017-2018.

Behavior frequency	BMD z scores by sun-protective behavior
	Staying in the shade		Wearing long sleeves		Using sunscreen
	Estimate (95% CI)	P value	Estimate (95% CI)	P value	Estimate (95% CI)	P value
Total body
Unadjusted model
Rare	1 [Reference]	.003	1 [Reference]	.02	1 [Reference]	.01
Moderate	−0.12 (−0.28 to 0.04)		−0.09 (−0.20 to 0.20)		−0.15 (−0.26 to −0.04)
Frequent	−0.27 (−0.42 to −0.12)		−0.17 (−0.27 to −0.06)		−0.15 (−0.25 to −0.06)
Multivariate adjusted modela
Rare	1 [Reference]	.18	1 [Reference]	.72	1 [Reference]	.15
Moderate	−0.07 (−0.26 to 0.12)		−0.04 (−0.23 to 0.15)		−0.22 (−0.44 to −0.001)
Frequent	−0.23 (−0.47 to 0.02)		−0.08 (−0.27 to 0.12)		−0.10 (−0.32 to 0.13)
Upper body
Left arm
Unadjusted model
Rare	1 [Reference]	<.001	1 [Reference]	.06	1 [Reference]	<.001
Moderate	−0.16 (−0.30 to −0.02)		−0.06 (−0.17 to 0.05)		−0.24 (−0.35 to −0.13)
Frequent	−0.41 (−0.55 to −0.27)		−0.18 (−0.33 to −0.04)		−0.40 (−0.50 to −0.30)
Multivariate adjusted modela
Rare	1 [Reference]	.58	1 [Reference]	.78	1 [Reference]	.86
Moderate	−0.03 (−0.13 to 0.07)		0.04 (−0.12 to 0.20)		−0.04 (−0.24 to 0.15)
Frequent	−0.10 (−0.30 to 0.10)		−0.02 (−0.14 to 0.11)		−0.05 (−0.24 to 0.15)
Right arm
Unadjusted model
Rare	1 [Reference]	<.001	1 [Reference]	.19	1 [Reference]	<.001
Moderate	−0.19 (−0.34 to −0.05)		−0.03 (−0.14 to 0.09)		−0.24 (−0.35 to −0.14)
Frequent	−0.44 (−0.55 to −0.33)		−0.13 (−0.27 to 0.01)		−0.39 (−0.49 to −0.28)
Multivariate adjusted modela
Rare	1 [Reference]	.13	1 [Reference]	.97	1 [Reference]	.29
Moderate	−0.09 (−0.20 to 0.01)		0.02 (−0.14 to 0.18)		−0.12 (−0.29 to 0.05)
Frequent	−0.15 (−0.31 to 0.02)		0.002 (−0.14 to 0.15)		−0.02 (−0.22 to 0.17)
Total thoracic spine
Unadjusted model
Rare	1 [Reference]	.31	1 [Reference]	.21	1 [Reference]	.01
Moderate	−0.09 (−0.25 to 0.07)		−0.09 (−0.20 to 0.01)		−0.15 (−0.27 to −0.03)
Frequent	−0.11 (−0.26 to 0.03)		−0.08 (−0.22 to 0.07)		−0.12 (−0.22 to −0.03)
Multivariate adjusted modela
Rare	1 [Reference]	.64	1 [Reference]	.95	1 [Reference]	.14
Moderate	−0.10 (−0.32 to 0.12)		−0.02 (−0.24 to 0.20)		−0.21 (−0.43 to 0.01)
Frequent	−0.05 (−0.26 to −0.16)		0.02 (−0.26 to 0.29)		−0.01 (−0.26 to 0.23)
Lower body
Femoral neck
Unadjusted model
Rare	1 [Reference]	.21	1 [Reference]	.43	1 [Reference]	.20
Moderate	−0.03 (−0.32 to 0.26)		−0.14 (−0.44 to 0.15)		−0.13 (−0.47 to 0.20)
Frequent	−0.20 (−0.52 to 0.13)		−0.25 (−0.64 to 0.14)		−0.16 (−0.34 to 0.02)
Multivariate adjusted modela
Rare	1 [Reference]	.047	1 [Reference]	.40	1 [Reference]	.92
Moderate	0.24 (−0.01 to 0.50)		0.16 (−0.28 to 0.60)		−0.05 (−0.41 to 0.32)
Frequent	−0.09 (−0.51 to 0.33)		−0.13 (−0.46 to 0.20)		−0.11 (−0.65 to 0.44)
Total lumbar spine
Unadjusted model
Rare	1 [Reference]	.17	1 [Reference]	.03	1 [Reference]	.15
Moderate	−0.08 (−0.24 to 0.09)		−0.10 (−0.22 to 0.03)		−0.11 (−0.24 to 0.02)
Frequent	−0.15 (−0.31 to 0.10)		−0.17 (−0.28 to −0.05)		0.07 (−0.07 to 0.20)
Multivariate adjusted modela
Rare	1 [Reference]	.27	1 [Reference]	.69	1 [Reference]	.01
Moderate	−0.03 (−0.22 to 0.17)		−0.10 (−0.33 to 0.13)		−0.35 (−0.56 to −0.14)
Frequent	−0.16 (−0.35 to 0.04)		−0.03 (−0.28 to 0.22)		−0.02 (−0.33 to 0.29)
Left leg
Unadjusted model
Rare	1 [Reference]	<.001	1 [Reference]	.04	1 [Reference]	<.001
Moderate	−0.14 (−0.28 to 0.003)		−0.07 (−0.14 to 0.003)		−0.18 (−0.28 to −0.08)
Frequent	−0.33 (−0.47 to −0.20)		−0.16 (−0.29 to −0.04)		−0.26 (−0.38 to −0.13)
Multivariate adjusted modela
Rare	1 [Reference]	.23	1 [Reference]	.95	1 [Reference]	.13
Moderate	−0.04 (−0.21 to 0.13)		−0.03 (−0.21 to 0.15)		−0.23 (−0.44 to −0.02)
Frequent	−0.18 (−0.39 to 0.04)		−0.01 (−0.16 to 0.13)		−0.13 (−0.33 to 0.08)
Right leg
Unadjusted model
Rare	1 [Reference]	.001	1 [Reference]	.02	1 [Reference]	.004
Moderate	−0.13 (−0.29 to 0.04)		−0.09 (−0.17 to −0.01)		−0.15 (−0.25 to −0.05)
Frequent	−0.32 (−0.47 to −0.16)		−0.17 (−0.30 to −0.06)		−0.22 (−0.35 to −0.08)
Multivariate adjusted modela
Rare	1 [Reference]	.16	1 [Reference]	.86	1 [Reference]	.17
Moderate	−0.05 (−0.19 to 0.09)		−0.001 (−0.21 to 0.21)		−0.16 (−0.34 to 0.01)
Frequent	−0.18 (−0.36 to 0.004)		−0.05 (−0.27 to 0.17)		−0.15 (−0.37 to 0.08)

Abbreviation: NHANES, National Health and Nutrition Examination Survey.

^a Adjusted for gender, age, race and ethnicity, education level, country of birth, annual household income, milk consumption, current smoker status, body mass index, total serum calcium, serum alkaline phosphatase, and metabolic equivalent task–minute physical activity scores.

Weak association of moderate staying in the shade with increased femoral neck BMD (0.24 [95% CI, −0.01 to 0.50]; P = .047), and association of moderate sunscreen use with diminished lumbar spine BMD (−0.35 [95% CI, −0.56 to −0.14]; P = .01) remained statistically significant in the multivariate models (Table 3). Across the entire study cohort, advancing age (−0.01 [95% CI, −0.02 to −0.001]; P = .03) and lower BMI (0.02 [95% CI, 0.01-0.03]; P = 0.01) were the statistically significant covariates associated with lower lumbar spine BMD z scores. Among the subset of people with moderate sunscreen use, an older age (−0.02 [95% CI, −0.04 to −0.002]; P = .04) was the only statistically significant correlate of diminished lumbar spine BMD z scores, while BMI (0.01 [95% CI, −0.02 to 0.04]; P = .43), MET-score physical exercise (0.01 [95% CI, −0.16 to 0.18]; P = .88), gender (0.27 [95% CI, −0.11 to 0.66]; P = .15), and current smoking status (0.11 [95% CI, −0.07 to 0.28], P = .20) were not. The association of individual sun-protective measures with BMD z scores at other remaining and minor anatomic locations is presented in the eTable in the Supplement.

Table 4 summarizes associations of sun-protective behaviors, sun sensitivity, and sunburns with the prevalence of osteoporotic bone fractures among the study population. In the unadjusted analysis, frequent staying in the shade was associated with a decreased prevalence of spine fractures (0.25 [95% CI, 0.11-0.54]; P = .01). In addition, use of multimodal (ie, at least 2 behaviors) sun protection correlated with diminished odds of fractures of wrist (0.21 [95% CI, 0.07-0.66]; P = .01) and spine (0.32 [95% CI, 0.11-0.95]; P = .04). Association of moderate to frequent staying in the shade with decreased prevalence of osteoporotic spine fractures persisted with multivariate adjustment (0.19 [95% CI, 0.04-0.86]; P = .02). With the multivariate adjustment, there were statistically significant associations of mild to severe sun sensitivity with reduced odds of wrist fractures (severe sun sensitivity, 0.06 [95% CI, 0.01-0.40]; P = .02) and spine fractures (mild sun sensitivity, 0.08 [95% CI, 0.02-0.34]; P = .01). The use of sunscreen or other sun-protective behaviors were not associated with increased odds of osteoporotic fractures of any location in the univariate or multivariate analyses (Table 4).

Table 4. Associations of Sun-Protective Behaviors, Sun Sensitivity, and Sunburns With Osteoporotic Bone Fractures in NHANES, 2017-2018.

Sun-related characteristic	Hip fracture		Wrist fracture		Spine fracture
	OR (95% CI)	P value	OR (95% CI)	P value	OR (95% CI)	P value
Staying in the shade
Unadjusted model
Rare	1 [Reference]	.46	1 [Reference]	.17	1 [Reference]	.01
Moderate	0.34 (0.05-2.29)		0.59 (0.24-1.50)		0.51 (0.18-1.45)
Frequent	0.39 (0.07-2.26)		0.39 (0.15-1.03)		0.25 (0.11-0.54)
Multivariate adjusted modela
Rare	1 [Reference]	.27	1 [Reference]	.08	1 [Reference]	.02
Moderate	0.27 (0.05-1.56)		0.29 (0.07-1.11)		0.13 (0.02-0.75)
Frequent	0.23 (0.02-3.01)		0.19 (0.05-0.77)		0.19 (0.04-0.86)
Wearing long sleeves
Unadjusted model
Rare	1 [Reference]	.28	1 [Reference]	.10	1 [Reference]	.64
Moderate	0.74 (0.15-3.68)		0.93 (0.28-3.16)		0.56 (0.14-2.30)
Frequent	3.98 (0.51-31.14)		0.30 (0.09-0.96)		1.156 (0.36-3.73)
Multivariate adjusted modela
Rare	1 [Reference]	.12	1 [Reference]	.65	1 [Reference]	NA
Moderate	0.45 (0.03-6.81)		0.70 (0.13-3.70)		NA
Frequent	16.69 (1.24-232.00)		1.93 (0.22-17.29)		NA
Applying sunscreen
Unadjusted model
Rare	1 [Reference]	.24	1 [Reference]	.40	1 [Reference]	.42
Moderate	0.15 (0.01-1.83)		1.05 (0.40-2.81)		0.70 (0.21-2.38)
Frequent	0.31 (0.05-1.87)		0.47 (0.15-1.46)		0.45 (0.12-1.65)
Multivariate adjusted modela
Rare	1 [Reference]	NAb	1 [Reference]	.20	1 [Reference]	NAb
Moderate	NAb		0.23 (0.04-1.24)		NAb
Frequent	NAb		0.23 (0.02-2.53)		NAb
Multimodal sun protections
Unadjusted model
<2 behaviors	1 [Reference]	.16	1 [Reference]	.01	1 [Reference]	.04
≥2 behaviors	0.20 (0.02-2.06)		0.21 (0.07-0.66)		0.32 (0.11-0.95)
Multivariate adjusted modela
<2 behaviors	1 [Reference]	NAb	1 [Reference]	.18	1 [Reference]	.35
≥2 behaviors	NAb		0.16 (0.01-2.55)		0.35 (0.03-3.63)
Sun sensitivity
Unadjusted model
Nonsensitive	1 [Reference]	.92	1 [Reference]	.78	1 [Reference]	.16
Mild	1.30 (0.29-5.78)		1.42 (0.40-5.02)		0.43 (0.17-1.09)
Severe	1.24 (0.28-5.59)		1.40 (0.52-3.75)		1.18 (0.51-2.73)
Multivariate adjusted modela
Nonsensitive	1 [Reference]	.77	1 [Reference]	.02	1 [Reference]	.01
Mild	0.99 (0.10-10.23)		0.64 (0.19-2.21)		0.08 (0.02-0.34)
Severe	0.44 (0.03-6.69)		0.06 (0.01-0.40)		0.43 (0.14-1.30)
Sunburn in the past year
Unadjusted model
No	1 [Reference]	.72	1 [Reference]	.44	1 [Reference]	.71
Yes	0.75 (0.15-3.89)		0.74 (0.33-1.66)		1.17 (0.49-2.78)
Multivariate adjusted modela
No	1 [Reference]	.74	1 [Reference]	.07	1 [Reference]	.73
Yes	0.55 (0.01-25.03)		0.42 (0.16-1.09)		0.75 (0.14-4.10)

Abbreviations: NA, not applicable; NHANES, National Health and Nutrition Examination Survey; OR, odds ratio.

^a Adjusted for gender, age, race and ethnicity, education level, country of birth, annual household income, milk consumption, current smoker status, body mass index, total serum calcium, serum alkaline phosphatase, and metabolic equivalent task–minute physical activity scores.

^b Unreliable estimates owing to insufficient outcome observations.

Discussion

This is, to our knowledge, the first report to investigate the associations among sun protection, sun sensitivity, BMD, and osteoporotic bone fractures in a population-based setting, specifically the US adult population. Despite initially statistically significant associations of various sun-protective measures (ie, staying in the shade, wearing long sleeves, applying sunscreen) with lower site-specific or total body BMD in the univariate regressions, these associations faded in the multivariate adjusted models, except for a possible diminishing of lumbar spine BMD with moderate sunscreen use. It is important to note that this statistically significant association was more likely to be seen among the older population. In this population, it is imperative to address additional and more likely contributing risk factors to lower spinal BMD such as lower physical activity and diminished body weight, which separately or together accelerate bone mass loss^13,14,15,16 and progression to the hard outcomes, including osteoporosis¹⁷ and osteoporotic bone fractures.¹⁸ Therefore, at-risk individuals with 1 or a combination of sun-protective behaviors would likely benefit from counseling on reversing maladaptive lifestyle patterns (eg, sedentary lifestyle, malnourishment) to offset potential effects on bone health.

There was no increased risk of osteoporotic bone fractures in those practicing various sun-protective behaviors. In fact, those with moderate to frequent staying in the shade were less likely to experience osteoporotic spine fracture, possibly because of sustained patterns of self-protective behaviors to avoid unintentional falls and risky physical activities. We ran a further exploratory analysis on the associations of sun sensitivity and sunburns with the osteoporotic bone fractures (Table 4) and found that those with mild to severe sun sensitivity were less likely to experience wrist and spine fractures. Considering the strong association between sun protection and sun sensitivity, this additional analysis further supported that people with strong sun-protective behaviors (who are also more likely to be sun sensitive) are not at increased risk of adverse bone-related outcomes.

What is striking is that almost half of US adults included in this study did not use any of the sun-protective behaviors (ie, staying in the shade, wearing long sleeves, applying sunscreen) and 35% used only 1, almost identical findings to those reported by Kim et al³ based on NHANES data from 2011 to 2014. These data reflect the trend that sun-protective practices have not changed in the US in recent years and may call for revised action policies to reinforce sun-protective behavioral habits, especially among non-Hispanic White and Black adults. However, it should be noted that the recommendation for sun-protective habits in individuals with darker skin is somewhat controversial among dermatologists. One of many potential reasons for such differing opinions is that UV exposure is not considered a major etiological risk factor for melanoma in populations with darker skin. In addition, melanoma in individuals with darker skin typically arises in sun-protected areas (eg, mucosal, plantar, palmar, subungual regions) as opposed to their White peers, who typically develop melanoma on sun-exposed areas (eg, face and neck).¹⁹ While previous melanoma intervention brochures for individuals with darker skin have been largely ineffective, the same brochures for the general public have enhanced early detection of skin cancers and honed patients’ skills in identifying abnormal lesions.²⁰

Some of the underuse might further stem from the subjective fears of both patients and physicians associated with the use of sunscreens and other sun-protective measures. Despite the lack of established data to support negative associations of sun protection with vitamin D status and bone health, many patients may still be reluctant to use sunscreen because, in their own words, “I don’t want to become vitamin D deficient.”^21(p1378) There may be other potential fears concerning the use of sunscreens that may contribute to this noncompliance in certain individuals. Additional fears of chemical absorption into the body (eg, organic and inorganic filters)^22,23 and their perceived consequences (ie, carcinogenicity), as well as any potential harm of sunscreen ingredients (eg, oxybenzone) on the environment²⁴ have been expressed by some patients. Thus, the present findings are essential to continue the important sun-protection conversation with patients.

Limitations

This study has some limitations. Owing to the cross-sectional design of the NHANES 2017 to 2018 dataset, we could not establish the causality of associations or provide any long-term data on the included individuals. The question on the use of sunscreens in the NHANES does not clearly distinguish between the use of sunscreens and other sunblocks (eg, SPF moisturizers), which may have different potency across various body areas.²⁵ Currently, the NHANES also does not record information on the specific type (eg, lotion, spray), strength (eg, SPF, UV protection), or amount of sunscreen used among survey respondents. This may partially contribute to the higher prevalence of sunburns among sun-sensitive individuals, despite their higher frequency of multimodal sun protection compared with sun-tolerant individuals. Therefore, sun-sensitive US adults may particularly benefit from targeted education on their required specifics for sunscreens and other sun-protective methods, as well as supervised application of each method for enhanced results. In a cross-sectional study of 65 patients with melanoma, the majority did not apply sunscreens the correct way, with only 8% applying sunscreens more than once daily.²⁶ Many patients cited poor patient-physician communication for the lack of knowledge, and one-third stated that they did not receive adequate information about melanoma risk factors from their physicians.²⁶ The BMD scans in the NHANES 2017 to 2018 dataset were performed in adults up to age 59 years, whereas routine BMD measurements are clinically recommended around age 65 years. Such disconnect between the study population age and bone health epidemiology potentially results in a failure to capture the time of maximal risk for osteoporosis. In addition, some of the fractures in the present study with a relatively younger population (ie, age 50-59 years) are possibly attributable to trauma. Finally, by the time of this report, serum 25-hydroxyvitmain D data from the NHANES 2017 to 2018 dataset has not been made available. To address this limitation, we performed univariate and multivariate analyses taking into account history of recent milk consumption, and serum calcium and alkaline phosphatase as surrogates of serum 25-hydroxyvimain D status and bone homeostasis (Tables 3 and 4). Notably, there was no statistically significant difference in dietary intake of vitamin D among adults with different patterns of sun-protective behaviors (Table 2).

Conclusions

Results of this cross-sectional study demonstrated that the continued practice of sun protection is not associated with increased bone-related health complications such as decreased BMD and/or increased osteoporotic fractures. Sun protection may actually be associated with a modest decrease in the prevalence of osteoporotic fractures, possibly owing to vigilant behaviors, and should be encouraged in all patients, including those otherwise at risk for decreased BMD and fractures. Future studies accounting for the type, strength, and amount of sunscreen use, as well as serum vitamin D status, are needed to further explore the thus far reassuring association between sun protection and bone health.

Supplement.

eTable. Association of sun-protective behaviors with BMD z-scores in NHANES 2017-2018