Background:

Cutaneous basal cell carcinoma (BCC) has long been associated with UV radiation (UVR) exposure, but data are limited on risks by anatomic site.

Methods:

We followed 63,912 cancer-free White U.S. radiologic technologists from cohort entry (1983–1989/1994–1998) to exit (date first BCC via 2003–2005 questionnaire). We estimated associations between cumulative ambient UVR and relative/absolute risks of self-reported BCC by anatomic location via Poisson models.

Results:

For incident first primary BCC in 2,124 subjects (mean follow-up, 16.9 years) log[excess relative risks] (ERR) of BCC per unit cumulative ambient UVR = 1.27/MJ cm–2 [95% confidence interval (CI): 0.86–1.68; Ptrend < 0.001] did not vary by anatomic site (P = 0.153). However, excess absolute risks of BCC per unit cumulative ambient UVR were large for the head/neck = 5.46/MJ cm–2/104 person-year (95% CI: 2.92–7.36; Ptrend < 0.001), smaller for the trunk (2.56; 95% CI: 1.26–3.33; Ptrend = 0.003), with lesser increases elsewhere. There were lower relative risks, but higher absolute risks, for those with Gaelic ancestry (P < 0.001), also higher absolute risks among those with fair complexion, but relative and absolute risks were not generally modified by other constitutional, lifestyle or medical factors for any anatomic sites. Excess absolute and relative risk was concentrated 5–15 years before time of follow-up.

Conclusions:

BCC relative and absolute risk rose with increasing cumulative ambient UVR exposure, with absolute risk highest for the head/neck, to a lesser extent in the trunk.

Impact:

These associations should be evaluated in other White and other racial/ethnic populations along with assessment of possible modification by time outdoors, protective, and behavioral factors.

Basal cell carcinoma (BCC) of the skin is the most common malignancy in populations of European ancestry, with estimates of over 1 million people diagnosed with BCC in the United States each year (1, 2). Rates of BCC have substantially increased worldwide over the last 30 years (3). The total economic costs of diagnosis and treatment in the United States exceeded $8.1 billion in 2011 (4) and costs are quite rapidly increasing in a number of countries (5, 6). The relationship of BCC with UVR exposure is suspected to vary by anatomic site, which could have important implications for recommendations aimed at primary prevention. On the basis of clinical reports and small and limited case–control studies, the majority of BCCs are thought to occur on typically sun exposed areas such as the head and neck (3, 7) supporting the important causative role of solar UV radiation (UVR; ref. 8). Other risk factors include occupation, male sex, older age, lighter skin, blue/green eye color, and lighter or red hair color, photosensitizing agents, indoor tanning, ionizing radiation, obesity, and certain medical conditions or medications, (3, 7, 9–14) in particular some leading to immune suppression (3, 15).

There has been very little assessment of ambient UVR or other risk factors for BCC by anatomic site. Evaluations have been limited to case series (16, 17), case–control studies in western Australia (18) and Italy (19), and a relatively small skin cancer screening study in the town of Nambour in Queensland, Australia (20). The findings have been inconsistent. Although there have been comprehensive case–control studies assessing UVR and other postulated risk factors in specific states in the United States [reviewed by Gossai and colleagues (3)] and in cohorts of health professionals from (at least) 11 populous U.S. states (21), limited information is available on nationwide occurrence and risk factor relationships for BCC and none have prospectively evaluated cumulative ambient UVR in relation to risk by anatomic site in the United States, in part because information on BCC is not systematically collected by U.S. population-based cancer registries (21, 22). Compared with case–control studies, cohort studies minimize recall bias (although residential location would be expected to be recalled with similar accuracy in both types of study), and enable estimation of absolute risks, which are important measures of public health impact. Data collected nationwide in the United States from a wide range of latitudes provides a large range of cumulative UVR exposures, thereby increasing the statistical power to detect UVR-associated risks relative to a more regionally defined study.

The objective of the current study is to assess the relationship between estimated ambient lifetime UVR with incidence of BCC by anatomic site using data from the U.S. Radiologic Technologists (USRT), a large nationwide prospective cohort, which has collected detailed information on anatomic sites of skin cancers in a population exposed to a wide range of lifetime ambient UVR. Features of this investigation include a primary emphasis on absolute risks and assessment of potential effect modification by constitutional and selected lifestyle factors.

Overview of USRT study

The USRT cohort consists of 146,022 radiologic technologists living in the United States and its territories who were certified by the American Registry of Radiologic Technologists for at least 2 years between 1926 and 1982 and followed using yearly registry recertification. For cohort members not recertified, vital status has been obtained through linkage with tracing databases and the Social Security Administration's “presumed living” search. Linkage of those known or presumed deceased with the National Death Index was conducted to confirm that they were deceased. Methods of the study have been described previously (10, 23–25) and detailed information can be found online at www.radtechstudy.nci.nih.gov. Three surveys were undertaken between 1983 and 2005. For the first survey, questionnaires were sent during 1983 to 1989 to 132,298 surviving cohort members with a postal address; 90,305 (68%) technologists completed the questionnaire. A second questionnaire was sent to 126,628 living and located members during 1994 to 1998; 90,972 (72%) technologists responded. If a technologist answered both the first and second questionnaire, the earlier of the two was designated as the baseline questionnaire. A third questionnaire was sent in 2003 to 2005 to 105,694 participants who had answered at least one previous questionnaire and were still known living; of these, 73,838 (73%) responded.

Study population

In the current study, we evaluated participants who completed a baseline (first or second or both) questionnaire, the third questionnaire, did not report history of cancer (including BCC) prior to their baseline questionnaire, and who had sufficiently complete residential history that they could be linked to a database of UVR irradiance estimates by calendar time period and location in the continental United States. Those not completing the third questionnaire, in particular those dying before this point, were not included in the analysis. There were 63,912 White participants, both Hispanic and non-Hispanic, that met these criteria. Non-White respondents were excluded from detailed analyses due to the notably lower BCC incidence rates compared with Whites (3), the likely lower UVR sensitivity to skin cancer (26), and the relatively small number of these participants; however, summary results are presented in Table 1. Cohort members were followed from baseline questionnaire completion to the earlier of first primary cancer diagnosis, including BCC, or the completion of the third questionnaire (administrative censoring).

Table 1.

Distribution (%) of selected baseline characteristics of BCC cases (including cases without specified location) and non-cases among 63,912 White and 3,278 non-White participants in the USRT study.a

BCC
PheterogeneitybPerson-yearsNo BCCBCC all sitesHead and neckUpper extremitiesTrunkLower extremities
Characteristic         
Age at entry, mean (SD) 38.8 (9.4)        
Age at exit, mean (SD) 55.8 (8.4)        
Total White 1,082,740 61,761 2,151 1,411 166 413 134 
 Non-White 53,200 3,273 
Split by risk factors among Whites 
Sex 
P = 0.6421 Male 213,408 (19.7) 12,656 (20.5) 426 (20.1) 302 (21.4) 40 (24.1) 77 (18.6) 7 (5.2) 
 Female 869,332 (80.3) 49,132 (79.5) 1,698 (79.9) 1,109 (78.6) 126 (75.9) 336 (81.4) 127 (94.8) 
Birth year 
P < 0.0001 1900–1939 161,857 (14.9) 9,795 (15.9) 550 (25.9) 412 (29.2) 33 (19.9) 74 (17.9) 31 (23.1) 
 1940–1949 369,204 (34.1) 21,069 (34.1) 723 (34.0) 485 (34.4) 57 (34.3) 130 (31.5) 51 (38.1) 
 1950+ 551,678 (51.0) 30,924 (50.0) 851 (40.1) 514 (36.4) 76 (45.8) 209 (50.6) 52 (38.8) 
Highest level of education 
P < 0.0001 Grade school/high school 86,424 (8.0) 9,707 (15.7) 224 (10.5) 147 (10.4) 17 (10.2) 47 (11.4) 13 (9.7) 
P < 0.0001c Vocational/other 4,464 (0.4) 241 (0.4) 19 (0.9) 15 (1.1) 1 (0.6) 3 (0.7) 0 (0.0) 
 Two-year rad tech program 45,626 (4.2) 2,416 (3.9) 78 (3.7) 64 (4.5) 2 (1.2) 10 (2.4) 2 (1.5) 
 College or graduate school 561,315 (51.8) 29,239 (47.3) 966 (45.5) 641 (45.4) 77 (46.4) 182 (44.1) 66 (49.3) 
 Unknown 384,910 (35.5) 20,185 (32.7) 837 (39.4) 544 (38.6) 69 (41.6) 171 (41.4) 53 (39.6) 
Skin complexion 
P < 0.0001 Dark 18,987 (1.8) 1,105 (1.8) 20 (0.9) 16 (1.1) 1 (0.6) 2 (0.5) 1 (0.7) 
P < 0.0001c Medium 482,830 (44.6) 27,807 (45.0) 752 (35.4) 495 (35.1) 47 (28.3) 167 (40.4) 43 (32.1) 
 Fair 475,985 (44.0) 27,293 (44.2) 1,210 (57.0) 808 (57.3) 103 (62.0) 216 (52.3) 83 (61.9) 
 Unknown 104,937 (9.7) 5,583 (9.0) 142 (6.7) 92 (6.5) 15 (9.0) 28 (6.8) 7 (5.2) 
Eye color 
P < 0.0001 Brown/black 312,274 (28.8) 18,052 (29.2) 505 (23.8) 345 (24.5) 37 (22.3) 93 (22.5) 30 (22.4) 
P < 0.0001c Hazel 217,290 (20.1) 12,485 (20.2) 423 (19.9) 267 (18.9) 32 (19.3) 93 (22.5) 31 (23.1) 
 Blue/green/gray 442,840 (40.9) 25,357 (41.0) 1,043 (49.1) 700 (49.6) 82 (49.4) 196 (47.5) 65 (48.5) 
 Other 3,447 (0.3) 200 (0.3) 4 (0.2) 2 (0.1) 1 (0.6) 1 (0.2) 0 (0.0) 
 Unknown 106,890 (9.9) 5,694 (9.2) 149 (7.0) 97 (6.9) 14 (8.4) 30 (7.3) 8 (6.0) 
Natural hair color at age 15 
P < 0.0001 Black 29,123 (2.7) 1,756 (2.8) 36 (1.7) 27 (1.9) 2 (1.2) 6 (1.5) 1 (0.7) 
P < 0.0001c Brown 739,571 (68.3) 42,441 (68.7) 1,374 (64.7) 911 (64.6) 101 (60.8) 273 (66.1) 89 (66.4) 
 Red/blonde 204,933 (18.9) 11,763 (19.0) 565 (26.6) 378 (26.8) 47 (28.3) 104 (25.2) 36 (26.9) 
 Other 3,584 (0.3) 204 (0.3) 10 (0.5) 6 (0.4) 2 (1.2) 2 (0.5) 0 (0.0) 
 Unknown 105,528 (9.7) 5,624 (9.1) 139 (6.5) 89 (6.3) 14 (8.4) 28 (6.8) 8 (6.0) 
Gaelic ancestry 
P < 0.0001 No 546,971 (50.5) 31,206 (50.5) 957 (45.1) 619 (43.9) 75 (45.2) 189 (45.8) 74 (55.2) 
P < 0.0001c Yes 209,501 (19.3) 12,115 (19.6) 591 (27.8) 404 (28.6) 40 (24.1) 115 (27.8) 32 (23.9) 
 Unknown 326,267 (30.1) 18,467 (29.9) 576 (27.1) 388 (27.5) 51 (30.7) 109 (26.4) 28 (20.9) 
Hispanic 
P = 0.0033 No 1,065,920 (98.4) 60,650 (98.2) 2,104 (99.1) 1,395 (98.9) 165 (99.4) 411 (99.5) 133 (99.3) 
P = 0.0010c Yes 15,407 (1.4) 989 (1.6) 16 (0.8) 12 (0.9) 1 (0.6) 2 (0.5) 1 (0.7) 
 Unknown 1,413 (0.1) 149 (0.2) 4 (0.2) 4 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) 
Cancer in any sibling, and at young age 
P < 0.0001 No cancer in siblings 874,467 (80.8) 50,107 (81.1) 1,690 (79.6) 1,116 (79.1) 130 (78.3) 333 (80.6) 111 (82.8) 
P < 0.0001c Cancer in siblings, but age unknown 3,818 (0.4) 229 (0.4) 7 (0.3) 5 (0.4) 1 (0.6) 0 (0.0) 1 (0.7) 
 Cancer in siblings, all age ≥ 30 76,974 (7.1) 4,579 (7.4) 217 (10.2) 152 (10.8) 13 (7.8) 40 (9.7) 12 (9.0) 
 Cancer in siblings, at least one at age < 30 16,125 (1.5) 914 (1.5) 42 (2.0) 26 (1.8) 4 (2.4) 9 (2.2) 3 (2.2) 
 Unknown if siblings have cancer 111,356 (10.3) 5,959 (9.6) 168 (7.9) 112 (7.9) 18 (10.8) 31 (7.5) 7 (5.2) 
Baseline age at menarche (years) (females only) 
P = 0.0642 0–7 508 (0.1) 28 (0.1) 1 (0.1) 1 (0.1) 0 (0.0) 0 (0.0) 0 (0.0) 
P = 0.0508c 8–9 12,894 (1.5) 731 (1.5) 20 (1.2) 15 (1.4) 0 (0.0) 4 (1.2) 1 (0.8) 
 10–11 159,600 (18.4) 9,069 (18.5) 293 (17.3) 198 (17.9) 16 (12.7) 56 (16.7) 23 (18.1) 
 12–14 609,378 (70.1) 34,384 (70.0) 1,177 (69.3) 757 (68.3) 94 (74.6) 236 (70.2) 90 (70.9) 
 15+ 71,104 (8.2) 4,031 (8.2) 171 (10.1) 116 (10.5) 11 (8.7) 33 (9.8) 11 (8.7) 
 Unknown 15,847 (1.8) 889 (1.8) 36 (2.1) 22 (2.0) 5 (4.0) 7 (2.1) 2 (1.6) 
Baseline age at menopause (years) (females only) 
P < 0.0001 0–24 4,996 (0.6) 264 (0.5) 8 (0.5) 6 (0.5) 2 (1.6) 0 (0.0) 0 (0.0) 
P = 0.4029c 25–34 36,957 (4.3) 2,126 (4.3) 83 (4.9) 50 (4.5) 5 (4.0) 16 (4.8) 12 (9.4) 
 35–44 50,400 (5.8) 3,146 (6.4) 154 (9.1) 109 (9.8) 11 (8.7) 21 (6.3) 13 (10.2) 
 45–54 44,942 (5.2) 3,187 (6.5) 150 (8.8) 118 (10.6) 4 (3.2) 20 (6.0) 8 (6.3) 
 55+ 2,936 (0.3) 240 (0.5) 11 (0.6) 8 (0.7) 0 (0.0) 1 (0.3) 2 (1.6) 
 Unknown 729,101 (83.9) 40,169 (81.8) 1,292 (76.1) 818 (73.8) 104 (82.5) 278 (82.7) 92 (72.4) 
Use of oral contraception (females only) 
P < 0.0001 No 186,690 (21.5) 10,856 (22.1) 427 (25.1) 305 (27.5) 28 (22.2) 56 (16.7) 38 (29.9) 
P = 0.0205c Yes 604,544 (69.5) 34,133 (69.5) 1,173 (69.1) 743 (67.0) 91 (72.2) 257 (76.5) 82 (64.6) 
 Unknown 78,098 (9.0) 4,143 (8.4) 98 (5.8) 61 (5.5) 7 (5.6) 23 (6.8) 7 (5.5) 
Duration of use of hormone replacement therapy (years) (females only) 
P < 0.0001 11,324 (1.3) 830 (1.7) 27 (1.6) 22 (2.0) 0 (0.0) 3 (0.9) 2 (1.6) 
P = 0.0114c 1–2 17,491 (2.0) 1,151 (2.3) 40 (2.4) 31 (2.8) 3 (2.4) 2 (0.6) 4 (3.1) 
 3–4 10,426 (1.2) 751 (1.5) 34 (2.0) 22 (2.0) 6 (4.8) 4 (1.2) 2 (1.6) 
 5–9 13,679 (1.6) 1,001 (2.0) 57 (3.4) 46 (4.1) 3 (2.4) 7 (2.1) 1 (0.8) 
 10+ 12,832 (1.5) 974 (2.0) 59 (3.5) 40 (3.6) 4 (3.2) 10 (3.0) 5 (3.9) 
 Unknown 803,581 (92.4) 44,425 (90.4) 1,481 (87.2) 948 (85.5) 110 (87.3) 310 (92.3) 113 (89.0) 
Rheumatoid arthritis ever 
P = 0.0002 No 939,070 (86.7) 53,881 (87.2) 1,902 (89.5) 1,259 (89.2) 149 (89.8) 375 (90.8) 119 (88.8) 
P = 0.4562c Yes 22,830 (2.1) 1,349 (2.2) 53 (2.5) 41 (2.9) 4 (2.4) 4 (1.0) 4 (3.0) 
 Unknown 120,840 (11.2) 6,558 (10.6) 169 (8.0) 111 (7.9) 13 (7.8) 34 (8.2) 11 (8.2) 
Thyroiditis ever 
P < 0.0001 No 982,681 (90.8) 51,354 (83.1) 1,874 (88.2) 1,244 (88.2) 148 (89.2) 366 (88.6) 116 (86.6) 
P = 0.3930c Yes 6,694 (0.6) 353 (0.6) 16 (0.8) 13 (0.9) 0 (0.0) 1 (0.2) 2 (1.5) 
 Unknown 93,365 (8.6) 10,081 (16.3) 234 (11.0) 154 (10.9) 18 (10.8) 46 (11.1) 16 (11.9) 
Hyperthyroidism (Graves disease) 
P = 0.6284 No 1,059,090 (97.8) 60,285 (97.6) 2,067 (97.3) 1,373 (97.3) 162 (97.6) 405 (98.1) 127 (94.8) 
P = 0.4077c Yes 8,313 (0.8) 541 (0.9) 22 (1.0) 17 (1.2) 0 (0.0) 2 (0.5) 3 (2.2) 
 Unknown 15,340 (1.4) 962 (1.6) 35 (1.6) 21 (1.5) 4 (2.4) 6 (1.5) 4 (3.0) 
Use of diuretics 
P < 0.0001 No 858,538 (79.3) 49,199 (79.6) 1,732 (81.5) 1,155 (81.9) 127 (76.5) 340 (82.3) 110 (82.1) 
P < 0.0001c Yes 115,023 (10.6) 6,739 (10.9) 243 (11.4) 161 (11.4) 25 (15.1) 40 (9.7) 17 (12.7) 
 Unknown 109,179 (10.1) 5,850 (9.5) 149 (7.0) 95 (6.7) 14 (8.4) 33 (8.0) 7 (5.2) 
Baseline smoking status 
P = 0.0015 Unknown smoking status 1,834 (0.2) 141 (0.2) 4 (0.2) 2 (0.1) 2 (1.2) 0 (0.0) 0 (0.0) 
P = 0.0006c Non-smoker 539,832 (49.9) 30,623 (49.6) 1,050 (49.4) 680 (48.2) 82 (49.4) 221 (53.5) 67 (50.0) 
 Former smoker 298,451 (27.6) 17,509 (28.3) 671 (31.6) 457 (32.4) 47 (28.3) 124 (30.0) 43 (32.1) 
 Current smoker 235,766 (21.8) 13,147 (21.3) 391 (18.4) 265 (18.8) 34 (20.5) 68 (16.5) 24 (17.9) 
 Ex-smoker, unknown if current smoker 6,857 (0.6) 368 (0.6) 8 (0.4) 7 (0.5) 1 (0.6) 0 (0.0) 0 (0.0) 
Baseline alcohol consumption (drinks/week) 
P < 0.0001 None 13,124 (1.2) 1,167 (1.9) 25 (1.2) 16 (1.1) 3 (1.8) 4 (1.0) 2 (1.5) 
P = 0.0003c 1–2 625,143 (57.7) 35,315 (57.2) 1,137 (53.5) 760 (53.9) 86 (51.8) 210 (50.8) 81 (60.4) 
 3–6 175,178 (16.2) 9,883 (16.0) 337 (15.9) 218 (15.5) 30 (18.1) 73 (17.7) 16 (11.9) 
 ≥7 159,694 (14.7) 9,024 (14.6) 364 (17.1) 228 (16.2) 28 (16.9) 82 (19.9) 26 (19.4) 
 Unknown 109,601 (10.1) 6,399 (10.4) 261 (12.3) 189 (13.4) 19 (11.4) 44 (10.7) 9 (6.7) 
Body mass index (kg/m2) 
P < 0.0001 0–18.4 35,444 (3.3) 1,889 (3.1) 77 (3.6) 37 (2.6) 6 (3.6) 21 (5.1) 13 (9.7) 
P < 0.0001c 18.5–24.9 710,830 (65.7) 39,537 (64.0) 1,466 (69.0) 963 (68.2) 108 (65.1) 296 (71.7) 99 (73.9) 
 25–29.9 229,387 (21.2) 13,778 (22.3) 391 (18.4) 277 (19.6) 35 (21.1) 67 (16.2) 12 (9.0) 
 ≥30 81,491 (7.5) 5,113 (8.3) 143 (6.7) 101 (7.2) 11 (6.6) 23 (5.6) 8 (6.0) 
 Unknown 25,587 (2.4) 1,471 (2.4) 47 (2.2) 33 (2.3) 6 (3.6) 6 (1.5) 2 (1.5) 
Strenuous exercise hours per week 
P = 0.0024 <1 641,214 (59.2) 36,661 (59.3) 1,242 (58.5) 830 (58.8) 94 (56.6) 243 (58.8) 75 (56.0) 
P = 0.0081c 1–3 177,328 (16.4) 10,185 (16.5) 359 (16.9) 227 (16.1) 29 (17.5) 76 (18.4) 27 (20.1) 
 4+ 96,480 (8.9) 5,594 (9.1) 238 (11.2) 156 (11.1) 22 (13.3) 43 (10.4) 17 (12.7) 
 Unknown 167,718 (15.5) 9,348 (15.1) 285 (13.4) 198 (14.0) 21 (12.7) 51 (12.3) 15 (11.2) 
Cumulative baseline head and neck dose (Gy) 
P < 0.0001 3,077 (0.3) 183 (0.3) 4 (0.2) 3 (0.2) 0 (0.0) 0 (0.0) 1 (0.7) 
 0–0.010 68,119 (6.3) 3,651 (5.9) 102 (4.8) 68 (4.8) 10 (6.0) 18 (4.4) 6 (4.5) 
 0.010–0.025 246,821 (22.8) 13,573 (22.0) 401 (18.9) 244 (17.3) 32 (19.3) 100 (24.2) 25 (18.7) 
 0.025–0.050 322,196 (29.8) 18,260 (29.6) 590 (27.8) 382 (27.1) 55 (33.1) 116 (28.1) 37 (27.6) 
 0.050–0.075 181,210 (16.7) 10,440 (16.9) 371 (17.5) 250 (17.7) 26 (15.7) 71 (17.2) 24 (17.9) 
 0.075–0.100 98,552 (9.1) 5,787 (9.4) 198 (9.3) 137 (9.7) 13 (7.8) 37 (9.0) 11 (8.2) 
 0.100–0.200 124,767 (11.5) 7,473 (12.1) 331 (15.6) 225 (15.9) 27 (16.3) 55 (13.3) 24 (17.9) 
 0.200–0.300 24,809 (2.3) 1,541 (2.5) 87 (4.1) 66 (4.7) 3 (1.8) 13 (3.1) 5 (3.7) 
 0.300–0.500 10,226 (0.9) 680 (1.1) 33 (1.6) 30 (2.1) 0 (0.0) 2 (0.5) 1 (0.7) 
 0.500–0.750 2,278 (0.2) 153 (0.2) 6 (0.3) 5 (0.4) 0 (0.0) 1 (0.2) 0 (0.0) 
 ≥0.750 683 (0.1) 47 (0.1) 1 (0.0) 1 (0.1) 0 (0.0) 0 (0.0) 0 (0.0) 
BCC
PheterogeneitybPerson-yearsNo BCCBCC all sitesHead and neckUpper extremitiesTrunkLower extremities
Characteristic         
Age at entry, mean (SD) 38.8 (9.4)        
Age at exit, mean (SD) 55.8 (8.4)        
Total White 1,082,740 61,761 2,151 1,411 166 413 134 
 Non-White 53,200 3,273 
Split by risk factors among Whites 
Sex 
P = 0.6421 Male 213,408 (19.7) 12,656 (20.5) 426 (20.1) 302 (21.4) 40 (24.1) 77 (18.6) 7 (5.2) 
 Female 869,332 (80.3) 49,132 (79.5) 1,698 (79.9) 1,109 (78.6) 126 (75.9) 336 (81.4) 127 (94.8) 
Birth year 
P < 0.0001 1900–1939 161,857 (14.9) 9,795 (15.9) 550 (25.9) 412 (29.2) 33 (19.9) 74 (17.9) 31 (23.1) 
 1940–1949 369,204 (34.1) 21,069 (34.1) 723 (34.0) 485 (34.4) 57 (34.3) 130 (31.5) 51 (38.1) 
 1950+ 551,678 (51.0) 30,924 (50.0) 851 (40.1) 514 (36.4) 76 (45.8) 209 (50.6) 52 (38.8) 
Highest level of education 
P < 0.0001 Grade school/high school 86,424 (8.0) 9,707 (15.7) 224 (10.5) 147 (10.4) 17 (10.2) 47 (11.4) 13 (9.7) 
P < 0.0001c Vocational/other 4,464 (0.4) 241 (0.4) 19 (0.9) 15 (1.1) 1 (0.6) 3 (0.7) 0 (0.0) 
 Two-year rad tech program 45,626 (4.2) 2,416 (3.9) 78 (3.7) 64 (4.5) 2 (1.2) 10 (2.4) 2 (1.5) 
 College or graduate school 561,315 (51.8) 29,239 (47.3) 966 (45.5) 641 (45.4) 77 (46.4) 182 (44.1) 66 (49.3) 
 Unknown 384,910 (35.5) 20,185 (32.7) 837 (39.4) 544 (38.6) 69 (41.6) 171 (41.4) 53 (39.6) 
Skin complexion 
P < 0.0001 Dark 18,987 (1.8) 1,105 (1.8) 20 (0.9) 16 (1.1) 1 (0.6) 2 (0.5) 1 (0.7) 
P < 0.0001c Medium 482,830 (44.6) 27,807 (45.0) 752 (35.4) 495 (35.1) 47 (28.3) 167 (40.4) 43 (32.1) 
 Fair 475,985 (44.0) 27,293 (44.2) 1,210 (57.0) 808 (57.3) 103 (62.0) 216 (52.3) 83 (61.9) 
 Unknown 104,937 (9.7) 5,583 (9.0) 142 (6.7) 92 (6.5) 15 (9.0) 28 (6.8) 7 (5.2) 
Eye color 
P < 0.0001 Brown/black 312,274 (28.8) 18,052 (29.2) 505 (23.8) 345 (24.5) 37 (22.3) 93 (22.5) 30 (22.4) 
P < 0.0001c Hazel 217,290 (20.1) 12,485 (20.2) 423 (19.9) 267 (18.9) 32 (19.3) 93 (22.5) 31 (23.1) 
 Blue/green/gray 442,840 (40.9) 25,357 (41.0) 1,043 (49.1) 700 (49.6) 82 (49.4) 196 (47.5) 65 (48.5) 
 Other 3,447 (0.3) 200 (0.3) 4 (0.2) 2 (0.1) 1 (0.6) 1 (0.2) 0 (0.0) 
 Unknown 106,890 (9.9) 5,694 (9.2) 149 (7.0) 97 (6.9) 14 (8.4) 30 (7.3) 8 (6.0) 
Natural hair color at age 15 
P < 0.0001 Black 29,123 (2.7) 1,756 (2.8) 36 (1.7) 27 (1.9) 2 (1.2) 6 (1.5) 1 (0.7) 
P < 0.0001c Brown 739,571 (68.3) 42,441 (68.7) 1,374 (64.7) 911 (64.6) 101 (60.8) 273 (66.1) 89 (66.4) 
 Red/blonde 204,933 (18.9) 11,763 (19.0) 565 (26.6) 378 (26.8) 47 (28.3) 104 (25.2) 36 (26.9) 
 Other 3,584 (0.3) 204 (0.3) 10 (0.5) 6 (0.4) 2 (1.2) 2 (0.5) 0 (0.0) 
 Unknown 105,528 (9.7) 5,624 (9.1) 139 (6.5) 89 (6.3) 14 (8.4) 28 (6.8) 8 (6.0) 
Gaelic ancestry 
P < 0.0001 No 546,971 (50.5) 31,206 (50.5) 957 (45.1) 619 (43.9) 75 (45.2) 189 (45.8) 74 (55.2) 
P < 0.0001c Yes 209,501 (19.3) 12,115 (19.6) 591 (27.8) 404 (28.6) 40 (24.1) 115 (27.8) 32 (23.9) 
 Unknown 326,267 (30.1) 18,467 (29.9) 576 (27.1) 388 (27.5) 51 (30.7) 109 (26.4) 28 (20.9) 
Hispanic 
P = 0.0033 No 1,065,920 (98.4) 60,650 (98.2) 2,104 (99.1) 1,395 (98.9) 165 (99.4) 411 (99.5) 133 (99.3) 
P = 0.0010c Yes 15,407 (1.4) 989 (1.6) 16 (0.8) 12 (0.9) 1 (0.6) 2 (0.5) 1 (0.7) 
 Unknown 1,413 (0.1) 149 (0.2) 4 (0.2) 4 (0.3) 0 (0.0) 0 (0.0) 0 (0.0) 
Cancer in any sibling, and at young age 
P < 0.0001 No cancer in siblings 874,467 (80.8) 50,107 (81.1) 1,690 (79.6) 1,116 (79.1) 130 (78.3) 333 (80.6) 111 (82.8) 
P < 0.0001c Cancer in siblings, but age unknown 3,818 (0.4) 229 (0.4) 7 (0.3) 5 (0.4) 1 (0.6) 0 (0.0) 1 (0.7) 
 Cancer in siblings, all age ≥ 30 76,974 (7.1) 4,579 (7.4) 217 (10.2) 152 (10.8) 13 (7.8) 40 (9.7) 12 (9.0) 
 Cancer in siblings, at least one at age < 30 16,125 (1.5) 914 (1.5) 42 (2.0) 26 (1.8) 4 (2.4) 9 (2.2) 3 (2.2) 
 Unknown if siblings have cancer 111,356 (10.3) 5,959 (9.6) 168 (7.9) 112 (7.9) 18 (10.8) 31 (7.5) 7 (5.2) 
Baseline age at menarche (years) (females only) 
P = 0.0642 0–7 508 (0.1) 28 (0.1) 1 (0.1) 1 (0.1) 0 (0.0) 0 (0.0) 0 (0.0) 
P = 0.0508c 8–9 12,894 (1.5) 731 (1.5) 20 (1.2) 15 (1.4) 0 (0.0) 4 (1.2) 1 (0.8) 
 10–11 159,600 (18.4) 9,069 (18.5) 293 (17.3) 198 (17.9) 16 (12.7) 56 (16.7) 23 (18.1) 
 12–14 609,378 (70.1) 34,384 (70.0) 1,177 (69.3) 757 (68.3) 94 (74.6) 236 (70.2) 90 (70.9) 
 15+ 71,104 (8.2) 4,031 (8.2) 171 (10.1) 116 (10.5) 11 (8.7) 33 (9.8) 11 (8.7) 
 Unknown 15,847 (1.8) 889 (1.8) 36 (2.1) 22 (2.0) 5 (4.0) 7 (2.1) 2 (1.6) 
Baseline age at menopause (years) (females only) 
P < 0.0001 0–24 4,996 (0.6) 264 (0.5) 8 (0.5) 6 (0.5) 2 (1.6) 0 (0.0) 0 (0.0) 
P = 0.4029c 25–34 36,957 (4.3) 2,126 (4.3) 83 (4.9) 50 (4.5) 5 (4.0) 16 (4.8) 12 (9.4) 
 35–44 50,400 (5.8) 3,146 (6.4) 154 (9.1) 109 (9.8) 11 (8.7) 21 (6.3) 13 (10.2) 
 45–54 44,942 (5.2) 3,187 (6.5) 150 (8.8) 118 (10.6) 4 (3.2) 20 (6.0) 8 (6.3) 
 55+ 2,936 (0.3) 240 (0.5) 11 (0.6) 8 (0.7) 0 (0.0) 1 (0.3) 2 (1.6) 
 Unknown 729,101 (83.9) 40,169 (81.8) 1,292 (76.1) 818 (73.8) 104 (82.5) 278 (82.7) 92 (72.4) 
Use of oral contraception (females only) 
P < 0.0001 No 186,690 (21.5) 10,856 (22.1) 427 (25.1) 305 (27.5) 28 (22.2) 56 (16.7) 38 (29.9) 
P = 0.0205c Yes 604,544 (69.5) 34,133 (69.5) 1,173 (69.1) 743 (67.0) 91 (72.2) 257 (76.5) 82 (64.6) 
 Unknown 78,098 (9.0) 4,143 (8.4) 98 (5.8) 61 (5.5) 7 (5.6) 23 (6.8) 7 (5.5) 
Duration of use of hormone replacement therapy (years) (females only) 
P < 0.0001 11,324 (1.3) 830 (1.7) 27 (1.6) 22 (2.0) 0 (0.0) 3 (0.9) 2 (1.6) 
P = 0.0114c 1–2 17,491 (2.0) 1,151 (2.3) 40 (2.4) 31 (2.8) 3 (2.4) 2 (0.6) 4 (3.1) 
 3–4 10,426 (1.2) 751 (1.5) 34 (2.0) 22 (2.0) 6 (4.8) 4 (1.2) 2 (1.6) 
 5–9 13,679 (1.6) 1,001 (2.0) 57 (3.4) 46 (4.1) 3 (2.4) 7 (2.1) 1 (0.8) 
 10+ 12,832 (1.5) 974 (2.0) 59 (3.5) 40 (3.6) 4 (3.2) 10 (3.0) 5 (3.9) 
 Unknown 803,581 (92.4) 44,425 (90.4) 1,481 (87.2) 948 (85.5) 110 (87.3) 310 (92.3) 113 (89.0) 
Rheumatoid arthritis ever 
P = 0.0002 No 939,070 (86.7) 53,881 (87.2) 1,902 (89.5) 1,259 (89.2) 149 (89.8) 375 (90.8) 119 (88.8) 
P = 0.4562c Yes 22,830 (2.1) 1,349 (2.2) 53 (2.5) 41 (2.9) 4 (2.4) 4 (1.0) 4 (3.0) 
 Unknown 120,840 (11.2) 6,558 (10.6) 169 (8.0) 111 (7.9) 13 (7.8) 34 (8.2) 11 (8.2) 
Thyroiditis ever 
P < 0.0001 No 982,681 (90.8) 51,354 (83.1) 1,874 (88.2) 1,244 (88.2) 148 (89.2) 366 (88.6) 116 (86.6) 
P = 0.3930c Yes 6,694 (0.6) 353 (0.6) 16 (0.8) 13 (0.9) 0 (0.0) 1 (0.2) 2 (1.5) 
 Unknown 93,365 (8.6) 10,081 (16.3) 234 (11.0) 154 (10.9) 18 (10.8) 46 (11.1) 16 (11.9) 
Hyperthyroidism (Graves disease) 
P = 0.6284 No 1,059,090 (97.8) 60,285 (97.6) 2,067 (97.3) 1,373 (97.3) 162 (97.6) 405 (98.1) 127 (94.8) 
P = 0.4077c Yes 8,313 (0.8) 541 (0.9) 22 (1.0) 17 (1.2) 0 (0.0) 2 (0.5) 3 (2.2) 
 Unknown 15,340 (1.4) 962 (1.6) 35 (1.6) 21 (1.5) 4 (2.4) 6 (1.5) 4 (3.0) 
Use of diuretics 
P < 0.0001 No 858,538 (79.3) 49,199 (79.6) 1,732 (81.5) 1,155 (81.9) 127 (76.5) 340 (82.3) 110 (82.1) 
P < 0.0001c Yes 115,023 (10.6) 6,739 (10.9) 243 (11.4) 161 (11.4) 25 (15.1) 40 (9.7) 17 (12.7) 
 Unknown 109,179 (10.1) 5,850 (9.5) 149 (7.0) 95 (6.7) 14 (8.4) 33 (8.0) 7 (5.2) 
Baseline smoking status 
P = 0.0015 Unknown smoking status 1,834 (0.2) 141 (0.2) 4 (0.2) 2 (0.1) 2 (1.2) 0 (0.0) 0 (0.0) 
P = 0.0006c Non-smoker 539,832 (49.9) 30,623 (49.6) 1,050 (49.4) 680 (48.2) 82 (49.4) 221 (53.5) 67 (50.0) 
 Former smoker 298,451 (27.6) 17,509 (28.3) 671 (31.6) 457 (32.4) 47 (28.3) 124 (30.0) 43 (32.1) 
 Current smoker 235,766 (21.8) 13,147 (21.3) 391 (18.4) 265 (18.8) 34 (20.5) 68 (16.5) 24 (17.9) 
 Ex-smoker, unknown if current smoker 6,857 (0.6) 368 (0.6) 8 (0.4) 7 (0.5) 1 (0.6) 0 (0.0) 0 (0.0) 
Baseline alcohol consumption (drinks/week) 
P < 0.0001 None 13,124 (1.2) 1,167 (1.9) 25 (1.2) 16 (1.1) 3 (1.8) 4 (1.0) 2 (1.5) 
P = 0.0003c 1–2 625,143 (57.7) 35,315 (57.2) 1,137 (53.5) 760 (53.9) 86 (51.8) 210 (50.8) 81 (60.4) 
 3–6 175,178 (16.2) 9,883 (16.0) 337 (15.9) 218 (15.5) 30 (18.1) 73 (17.7) 16 (11.9) 
 ≥7 159,694 (14.7) 9,024 (14.6) 364 (17.1) 228 (16.2) 28 (16.9) 82 (19.9) 26 (19.4) 
 Unknown 109,601 (10.1) 6,399 (10.4) 261 (12.3) 189 (13.4) 19 (11.4) 44 (10.7) 9 (6.7) 
Body mass index (kg/m2) 
P < 0.0001 0–18.4 35,444 (3.3) 1,889 (3.1) 77 (3.6) 37 (2.6) 6 (3.6) 21 (5.1) 13 (9.7) 
P < 0.0001c 18.5–24.9 710,830 (65.7) 39,537 (64.0) 1,466 (69.0) 963 (68.2) 108 (65.1) 296 (71.7) 99 (73.9) 
 25–29.9 229,387 (21.2) 13,778 (22.3) 391 (18.4) 277 (19.6) 35 (21.1) 67 (16.2) 12 (9.0) 
 ≥30 81,491 (7.5) 5,113 (8.3) 143 (6.7) 101 (7.2) 11 (6.6) 23 (5.6) 8 (6.0) 
 Unknown 25,587 (2.4) 1,471 (2.4) 47 (2.2) 33 (2.3) 6 (3.6) 6 (1.5) 2 (1.5) 
Strenuous exercise hours per week 
P = 0.0024 <1 641,214 (59.2) 36,661 (59.3) 1,242 (58.5) 830 (58.8) 94 (56.6) 243 (58.8) 75 (56.0) 
P = 0.0081c 1–3 177,328 (16.4) 10,185 (16.5) 359 (16.9) 227 (16.1) 29 (17.5) 76 (18.4) 27 (20.1) 
 4+ 96,480 (8.9) 5,594 (9.1) 238 (11.2) 156 (11.1) 22 (13.3) 43 (10.4) 17 (12.7) 
 Unknown 167,718 (15.5) 9,348 (15.1) 285 (13.4) 198 (14.0) 21 (12.7) 51 (12.3) 15 (11.2) 
Cumulative baseline head and neck dose (Gy) 
P < 0.0001 3,077 (0.3) 183 (0.3) 4 (0.2) 3 (0.2) 0 (0.0) 0 (0.0) 1 (0.7) 
 0–0.010 68,119 (6.3) 3,651 (5.9) 102 (4.8) 68 (4.8) 10 (6.0) 18 (4.4) 6 (4.5) 
 0.010–0.025 246,821 (22.8) 13,573 (22.0) 401 (18.9) 244 (17.3) 32 (19.3) 100 (24.2) 25 (18.7) 
 0.025–0.050 322,196 (29.8) 18,260 (29.6) 590 (27.8) 382 (27.1) 55 (33.1) 116 (28.1) 37 (27.6) 
 0.050–0.075 181,210 (16.7) 10,440 (16.9) 371 (17.5) 250 (17.7) 26 (15.7) 71 (17.2) 24 (17.9) 
 0.075–0.100 98,552 (9.1) 5,787 (9.4) 198 (9.3) 137 (9.7) 13 (7.8) 37 (9.0) 11 (8.2) 
 0.100–0.200 124,767 (11.5) 7,473 (12.1) 331 (15.6) 225 (15.9) 27 (16.3) 55 (13.3) 24 (17.9) 
 0.200–0.300 24,809 (2.3) 1,541 (2.5) 87 (4.1) 66 (4.7) 3 (1.8) 13 (3.1) 5 (3.7) 
 0.300–0.500 10,226 (0.9) 680 (1.1) 33 (1.6) 30 (2.1) 0 (0.0) 2 (0.5) 1 (0.7) 
 0.500–0.750 2,278 (0.2) 153 (0.2) 6 (0.3) 5 (0.4) 0 (0.0) 1 (0.2) 0 (0.0) 
 ≥0.750 683 (0.1) 47 (0.1) 1 (0.0) 1 (0.1) 0 (0.0) 0 (0.0) 0 (0.0) 

aAnalysis is based on full follow-up, starting at earliest of first or second questionnaire completed. See Supplementary Fig. S1 for population exclusions. The BCC (all sites) column includes 27 BCC without anatomic location indicated.

bPheterogeneity in distribution of cases and non-cases, evaluated using Monte Carlo version of Fisher exact test with 106 simulations.

cPheterogeneity in distribution of cases and non-cases, excluding persons without information on the classifying variable, evaluated using Monte Carlo version of Fisher exact test with 106 simulations.

BCC case ascertainment and medical validation, and skin map location

Cases included self-reported incident first primary BCC (ICD10 C44.91) among subjects who responded to the third questionnaire, and who had previously completed the first or second questionnaire (or both) as described in more detail in an earlier report (10). Follow-up stopped at the first BCC reported by a technologist, and all subsequent BCCs were ignored. Efforts to validate self-reported BCC cases, conducted for 1,355 BCCs self-reported by USRT cohort members in the second questionnaire resulted in retrieval of 50% of medical records sought among which 97% (668 BCCs) of self-reported BCCs were confirmed (27). Medical records were sought for the 4,862 BCCs self-reported on the third questionnaire and obtained for 2,058 (42.3%) of these cohort members; of these, 1,762 (85.6%) self-reports were confirmed as a BCC. Because of the high medical record confirmation of cases for which medical records were available, we include all self-reported BCC cases in this analysis. Information on behavior, in particular, time spent outdoors was also captured on the third questionnaire. However, we judge that the possibility of biased recall of activity is substantial, so that we did not use this information. There is no anatomic site information captured on the fourth questionnaire, so that these data cannot be used.

A skin map anatomic diagram was sent to 5,990 USRT cohort members (5,927 of them White) reporting BCC and other skin cancers on the third questionnaire. Cohort members were asked to indicate the anatomic site of the skin cancer on the diagram, the type of skin cancer, and the year of diagnosis. We instructed cohort members to report only skin cancers diagnosed by physicians, and not to report “pre-cancerous” skin lesions or lesions that were treated but never confirmed to be skin cancer. Each mark on the anatomic diagram was coded to anatomic location and date of diagnosis. For purposes of the current article, cases include those self-reported BCCs with recorded anatomic location. Following Scotto and colleagues (28) we grouped the BCC into four main anatomic groupings, reflecting the likely degree of UVR exposure, namely face/head/neck + supraclavicular anatomic region (henceforth referred to as the head and neck), upper extremities, trunk, and lower extremities.

UVR exposure assessment

We used total UVR (= UVA+UVB) as a measure of ambient solar ultraviolet exposure. A standard UVR exposure metric is cumulative UVR radiant exposure [in units of Joule (J) per square meter (J m–2)], which is proportional to cumulative solar UVR energy deposition on a surface over a period of time. For concision we shall refer to cumulative UVR exposure rather than cumulative UVR radiant exposure from now on. This measure of cumulative ambient UVR exposure is recommended by the Commission Internationale de l'Eclairage (CIE; International Commission on Illumination; ref. 29). We used AVerage daily total GLObal (AVGLO) ground-based solar radiation measurements which are described in more detail elsewhere (30, 31) and in the Supplementary Materials and Methods. On the third questionnaire, residential location for five age periods (age <13, 13–19, 20–39, 40–64, 65+) was collected. Individual age-specific ambient UVR levels were determined by linking the residential locations with the AVGLO database (31).

Other postulated risk factors for BCC

The USRT has collected extensive baseline information on personal constitutional (skin, eye, and hair color), demographic (gender, age at diagnosis, racial/ethnic group), lifestyle [body mass index (BMI), smoking, and alcohol use], reproductive (age at menarche, age at menopause, use of oral contraceptives, menopausal hormone therapy), medical conditions and treatments, particularly disorders and medications associated with immune dysfunction or immune suppression (organ transplant, autoimmune diseases, immunosuppressive drugs), environmental exposures (UVR, occupational ionizing radiation), and genetic factors, which have been previously associated with increased risk of BCC (3). We assessed the effect of adjusting for these potentially confounding variables on cumulative UVR and excess absolute risk of BCC. We also assessed the extent to which this broad group of known and postulated BCC risk factors modified UVR-associated BCC risk.

Statistical analysis

In the analysis of Table 1, assessment of heterogeneity was via an analog of the Fisher exact test based on 106 Monte Carlo replicates. Missing values were coded as an extra level in each classification, but were otherwise treated like all the other levels for each classification. For each classifying variable, the full dataset was analyzed as well as the subset with information on the variable under investigation. To evaluate the relationship between UVR and BCC risk, we fitted Poisson excess relative risk and excess absolute risk models, in which the expected number of cases in the stratum with person years PY, after cumulative lagged UVR exposure, H(t) (in J cm–2), at age t, and with various other explanatory covariates, |Z = ({Z_i})$|⁠, was given by:

formula

for the relative risk model and

formula

for the absolute risk model. The parameter α is the log-linear excess relative risk per unit of cumulative radiant exposure, and the parameter δ is the linear excess absolute risk per unit of cumulative radiant exposure and per year of follow-up. We provide risk estimates for both relative risk and absolute risk. Relative risk was selected as the primary risk measure because this metric is often used in epidemiologic studies of UVR and most other exposures. There are also some methodologic advantages since relative risk models were less prone to instability in model fits. Absolute risk was also used because these model estimates provide a better measure of the lifetime impact for an individual or a population of UVR exposure. In addition, absolute risk provides quantitative estimates more useful for clinical and public health purposes. Fits of linear Poisson relative risk models and Cox proportional hazards models were also attempted, but both proved very unstable, and prone to problems of non-convergence. For the purposes of the current main analyses, a loglinear model of UVR cumulative radiant exposure was used for relative risk, and a linear model of cumulative radiant exposure used for absolute risk. For most purposes cumulative UVR exposures were lagged by 5 years, by analogy with the latency often assumed for solid cancers after ionizing radiation exposure (32), but we conducted sensitivity analyses using lag periods of 2 to 15 years. Because the risk factors being analyzed were only determined from the second questionnaire, for certain analyses in relation to possible modifying risk factors (as indicated by footnotes to the relevant tables) follow-up was restricted to the point of response to the second questionnaire onward, whether or not the subject also answered the first questionnaire. Further details are given in Supplementary Materials and Methods. Model fitting was performed using R (33) and Epicure (34). Confidence intervals (CI) were estimated from the profile likelihood (35), or if this did not converge, using Wald-based CIs.

Ethics approval and consent to participate

The research protocol for the USRT cohort study of cancer risks has been approved annually by the NCI Special Studies Institution Review Board (SSIRB Protocol OH97-C-N053) and the University of Minnesota Human Research Protection Program Institution Review Board (Federal Wide Assurance number 8005M02489). Completion of any questionnaire was considered implied consent for study participation; a waiver of consent was approved annually by both institutions due to the minimal risks of the study and the large size of the study population.

Among the 63,912 White USRT cohort members with information on cumulative UVR irradiance and no prior cancer diagnosis at baseline, 2,124 cases of first BCC were reported over the follow-up period (Table 1). Further details on numbers of skin map questionnaires sent and received are given in Supplementary Table S1. We present summary results for non-Whites in Table 1, which demonstrate the much lower BCC risk for these (5/53,200 = 9.4 × 10–5 per person per year) compared with Whites (2151/1,082,740 = 2.0 × 10−3 per person per year). A flow diagram detailing exclusions is given in Supplementary Fig. S1. Supplementary Table S2 gives a more detailed listing of BCC by anatomic site and sex. Most of the cohort members without BCC were female, born after 1950, had college or graduate school education, did not smoke and drank once or twice a week. Similar distributions were generally seen among cohort members with BCC overall and by anatomic site (Table 1). However, BCC cases tended to be born earlier than non-cases, were less likely to complete higher educational levels (college or graduate school), were somewhat less likely to be a current smoker, were slightly more likely to consume 3+ drinks per week, were more likely to have fair complexion and blue/green/gray eyes and have red/blonde hair, were more likely to have Gaelic ancestry, less likely to be Hispanic, not to be overweight, less likely to use oral contraception, were more likely to have hormone replacement therapy for 10 or more years, were more likely to use diuretics, were more likely to exercise strenuously, were more likely to have siblings with cancer and were slightly more likely to have higher occupational radiation dose. All these differences were mostly highly statistically significant (P < 0.005), irrespective of whether the full dataset was analyzed or just those with information on each classifying variable (Table 1).

Table 2 demonstrates that BCC risks normalized per unit skin area were much higher on the face and head, by about a factor of 4 compared with the neck and supraclavicular region. (Supplementary Table S3 details the stratifying variables used to generate the person-year tables.) These two anatomic sites have much higher risk per unit skin area than the upper extremities, trunk and lower extremities. Although the two sexes had generally similar patterns of variation of risk per unit area, lower extremity risk per unit area in females was 5-fold higher than that for males (Table 2).

Table 2.

Summary number of BCC cases at known anatomic locations by anatomic site, and rates per unit skin surface area among 63,912 White subjects.

BCC casesPercentage of BCC cases with known locationPercentage skin surface [as determined from standard Lund and Browder charts (47)]Relative BCC risk, normalized by skin surface areaa
Males 
 Face/head 282 66.2 1 (reference) 
 Neck/supraclavicular 20 4.7 0.248 (0.153–0.380) 
 Upper extremity 40 9.4 20 0.050 (0.035–0.068) 
 Trunk 77 18.1 32 0.060 (0.046–0.076) 
 Lower extremity 1.6 39 0.004 (0.002–0.009) 
 Total 426 100.0 100  
Females 
 Face/head 1,025 60.4 1 (reference) 
 Neck/supraclavicular 84 4.9 0.287 (0.228–0.356) 
 Upper extremity 126 7.4 20 0.043 (0.036–0.052) 
 Trunk 336 19.8 32 0.072 (0.063–0.081) 
 Lower extremity 127 7.5 39 0.022 (0.018–0.027) 
 Total 1,698 100.0 100  
Total 
 Face/head 1,307 61.5 1 (reference) 
 Neck/supraclavicular 104 4.9 0.279 (0.227–0.338) 
 Upper extremity 166 7.8 20 0.044 (0.038–0.052) 
 Trunk 413 19.4 32 0.069 (0.062–0.077) 
 Lower extremity 134 6.3 39 0.018 (0.015–0.022) 
 Total 2,124 100.0 100  
BCC casesPercentage of BCC cases with known locationPercentage skin surface [as determined from standard Lund and Browder charts (47)]Relative BCC risk, normalized by skin surface areaa
Males 
 Face/head 282 66.2 1 (reference) 
 Neck/supraclavicular 20 4.7 0.248 (0.153–0.380) 
 Upper extremity 40 9.4 20 0.050 (0.035–0.068) 
 Trunk 77 18.1 32 0.060 (0.046–0.076) 
 Lower extremity 1.6 39 0.004 (0.002–0.009) 
 Total 426 100.0 100  
Females 
 Face/head 1,025 60.4 1 (reference) 
 Neck/supraclavicular 84 4.9 0.287 (0.228–0.356) 
 Upper extremity 126 7.4 20 0.043 (0.036–0.052) 
 Trunk 336 19.8 32 0.072 (0.063–0.081) 
 Lower extremity 127 7.5 39 0.022 (0.018–0.027) 
 Total 1,698 100.0 100  
Total 
 Face/head 1,307 61.5 1 (reference) 
 Neck/supraclavicular 104 4.9 0.279 (0.227–0.338) 
 Upper extremity 166 7.8 20 0.044 (0.038–0.052) 
 Trunk 413 19.4 32 0.069 (0.062–0.077) 
 Lower extremity 134 6.3 39 0.018 (0.015–0.022) 
 Total 2,124 100.0 100  

aThese numbers are the normalized relative risks per unit skin area, and are given by |\exp [{\beta _i}]$| (with associated profile-likelihood CIs) corresponding to model (S1).

Table 3 and Fig. 1 demonstrate the increasing excess relative risk of BCC at all anatomic sites combined associated with increasing cumulative lifetime UVR exposure. Cumulative (lifetime) UVR exposure of 1 MJ cm–2 for both sexes combined was associated with an overall increase in the log relative risk of BCC of 1.27 (95% CI, 0.86–1.68; P < 0.001). Table 3 also shows that there was no significant difference between the relative risks by sex overall (P = 0.153) or for any specific anatomic site. Excess relative risk of BCC in relation to UVR was much the same on the head and neck as on the trunk (Table 3). The variation in excess relative risk by anatomic site was nonsignificant (P > 0.3) for males, females, or both sexes together, whether using the two-class analysis of heterogeneity (the contrast between head and neck and all other anatomic sites), or the four-class analysis of heterogeneity (the heterogeneity between head and neck vs. upper extremity vs. trunk vs. lower extremity).

Table 3.

Log excess relative risk of BCC at known anatomic locations among Whites with cumulative UVR exposure according to sex and anatomic site.a

MalesFemalesTotal
Caseslog excess relative risk per UVR MJ/cm2PCaseslog excess relative risk per UVR MJ/cm2PCaseslog excess relative risk per UVR MJ/cm2PPHeterogeneity
All BCC at known anatomic locationsb 426 1.50 (0.98–2.00) <0.001 1,698 1.19 (0.76–1.62) <0.001 2,124 1.27 (0.86–1.68) <0.001 0.153 
Four-level anatomic-location UVR-exposure classification 
Head and neckc 302 1.52 (0.91–2.13) <0.001 1,109 1.16 (0.64–1.67) <0.001 1,411 1.26 (0.77–1.76) <0.001 0.151 
Upper extremity, trunk, lower extremityd 124 1.28 (0.32–2.22) 0.009 589 1.21 (0.43–1.96) 0.002 713 1.22 (0.47–1.96) 0.001 0.845 
Upper extremityd 40 2.40 (0.60–4.13) 0.010 126 1.87 (0.28–3.41) 0.022 166 2.05 (0.53–3.50) 0.008 0.476 
Trunke 77 1.35 (0.13–2.53) 0.030 336 1.26 (0.24–2.24) 0.016 413 1.28 (0.30–2.24) 0.011 0.847 
Lower extremityd 0.22 (−3.39 to 3.46) 0.899 127 0.06 (−1.72 to 1.75) 0.944 134 0.07 (−1.70 to 1.75) 0.938 0.920 
Pheterogeneity (head and neck vs. upper extremity +trunk +lower extremity)  0.546   0.657   0.929   
Pheterogeneity (head and neck vs. upper extremity vs. trunk vs. lower extremity)  0.893   0.696   0.397   
MalesFemalesTotal
Caseslog excess relative risk per UVR MJ/cm2PCaseslog excess relative risk per UVR MJ/cm2PCaseslog excess relative risk per UVR MJ/cm2PPHeterogeneity
All BCC at known anatomic locationsb 426 1.50 (0.98–2.00) <0.001 1,698 1.19 (0.76–1.62) <0.001 2,124 1.27 (0.86–1.68) <0.001 0.153 
Four-level anatomic-location UVR-exposure classification 
Head and neckc 302 1.52 (0.91–2.13) <0.001 1,109 1.16 (0.64–1.67) <0.001 1,411 1.26 (0.77–1.76) <0.001 0.151 
Upper extremity, trunk, lower extremityd 124 1.28 (0.32–2.22) 0.009 589 1.21 (0.43–1.96) 0.002 713 1.22 (0.47–1.96) 0.001 0.845 
Upper extremityd 40 2.40 (0.60–4.13) 0.010 126 1.87 (0.28–3.41) 0.022 166 2.05 (0.53–3.50) 0.008 0.476 
Trunke 77 1.35 (0.13–2.53) 0.030 336 1.26 (0.24–2.24) 0.016 413 1.28 (0.30–2.24) 0.011 0.847 
Lower extremityd 0.22 (−3.39 to 3.46) 0.899 127 0.06 (−1.72 to 1.75) 0.944 134 0.07 (−1.70 to 1.75) 0.938 0.920 
Pheterogeneity (head and neck vs. upper extremity +trunk +lower extremity)  0.546   0.657   0.929   
Pheterogeneity (head and neck vs. upper extremity vs. trunk vs. lower extremity)  0.893   0.696   0.397   

aFor definition of four-level anatomic site classification, see Supplementary Table S2. Analysis is based on full follow-up, starting at first or second questionnaire.

bAdjustment to the baseline BCC rate for baseline questionnaire, sex, ln[age], ln[age]2, birth year, birth year2, birth year3, birth year4, birth year5.

cAdjustment to the baseline BCC rate for sex, ln[age], ln[age]2, birth year, birth year2, birth year3, birth year4, birth year5.

dAdjustment to the baseline BCC rate for sex, ln[age], ln[age]2, birth year.

eAdjustment to the baseline BCC rate for sex, ln[age], birth year, birth year2.

Figure 1.

Relative risk for BCC in relation to cumulative UVR exposure, for all BCC with known anatomic location (A), head and neck BCC (B), upper extremity (C), trunk (D), and lower extremity BCC (E). Relative risk (+95% CI) of BCC (+95% CI) in relation to baseline level (defined by cumulative UVR exposure <800 kJ m–2) in relation to cumulative UVR exposure, with intervals 0–799, 800–999, 1,000–1,199, 1,200–1,399, 1,400–1,599, 1,600+ kJ m–2. The reference group used is the lowest UVR cumulative radiant exposure group, 0–799 kJ m–2. Baseline models are those specified in footnotes to Table 3.

Figure 1.

Relative risk for BCC in relation to cumulative UVR exposure, for all BCC with known anatomic location (A), head and neck BCC (B), upper extremity (C), trunk (D), and lower extremity BCC (E). Relative risk (+95% CI) of BCC (+95% CI) in relation to baseline level (defined by cumulative UVR exposure <800 kJ m–2) in relation to cumulative UVR exposure, with intervals 0–799, 800–999, 1,000–1,199, 1,200–1,399, 1,400–1,599, 1,600+ kJ m–2. The reference group used is the lowest UVR cumulative radiant exposure group, 0–799 kJ m–2. Baseline models are those specified in footnotes to Table 3.

Close modal

Supplementary Table S4 demonstrates that adjusting for any of the potentially confounding lifestyle or constitutional variables had little effect on the risk, although there was a reduction by about 6% in risk when adjusting for BMI (adjusted log ERR per MJ/cm2 = 1.14 vs. 1.21 unadjusted). Even when adjusting for all variables simultaneously in the two separate variable groups (e.g., endogenous vs. lifestyle/reproductive/medical factors), the ERRs for BCC associated with cumulative UVR exposure remain significant.

Table 4 evaluates the potential for modification of the association between cumulative UVR and risk of BCC according to constitutional, lifestyle, and other risk factors. As shown, there was significant heterogeneity of log ERR per unit exposure for skin complexion for the upper extremity (P = 0.001) and trunk (P = 0.043). There was borderline significant (P = 0.056) heterogeneity in log ERR by eye color for upper extremity BCC; log ERR for upper extremity BCC was particularly high for those with dark or unknown complexion, with overall increases in the log ERR of BCC per MJ/cm2 of 5.63 (95% CI, −2.28 to 11.11) and 50.10 (95% CI, 9.20 to >100) respectively (Table 4). There was also significant heterogeneity of ERR per unit exposure by Gaelic ancestry for all sites (P < 0.001), and for head and neck BCC (P = 0.003; Table 4); log ERR was particularly low for those with Gaelic ancestry, with overall increase in the log ERR per MJ/cm2 of BCC of 0.64 (95% CI, 0.04–1.23; Table 4). There were significant (P = 0.014) variations in log ERR for upper extremity BCC in relation to use of diuretics, with those using diuretics at particularly high risk, having log ERR per MJ/cm2 of 3.26 (95% CI, 0.66–5.71; Table 4). There were borderline significant (P = 0.059) variations in log ERR for upper and lower extremity and trunk BCC in relation to age at menopause, with those with menopause under the age of 25 at highest risk. ERRs were borderline significant by level of occupational radiation dose overall (P = 0.088) and for head and neck BCC (P = 0.070), with log ERR per unit cumulative radiant exposure tending to be higher for higher occupational radiation doses (Table 4). Otherwise, there does not appear to be any significant modification of log ERR of BCC risk per unit UVR cumulative radiant exposure with any lifestyle, environmental or constitutional variable, either overall or for particular anatomic locations (Table 4).

Table 4.

Modification by constitutional risk factors of log excess relative risk of BCC by anatomic site in relation to cumulative UVR exposure.a

log excess relative risk per UVR MJ/cm2 (95% CI) (= β)
CasesAll BCC, known anatomic locationsbHead and neckcUpper and lower extremity, trunkdUpper extremitydTrunkeLower extremityd
Skin complexion 
 Dark 12 2.03 (−0.38 to 4.12) 1.59 (−1.27 to 4.04) 2.95 (−1.98 to 6.59) 5.63 (−2.28 to 11.11) −14.28 (−40.56 to 1.57) 4.91 (−2.08 to 11.16) 
 Medium 540 1.46 (0.91–2.00) 1.41 (0.74–2.07) 1.51 (0.52–2.47) 3.06 (0.95–5.08) 1.53 (0.22–2.79) −0.35 (−2.67 to 1.82) 
 Fair 826 1.05 (0.53–1.56) 1.15 (0.52–1.76) 0.60 (−0.38 to 1.54) 1.21 (−0.83 to 3.14) 0.48 (−0.86 to 1.78) −0.27 (−2.36 to 1.69) 
 Unknown 10 1.10 (−1.68 to 3.66) 0.03 (−3.48 to 3.16) 3.24 (−1.64 to 7.71) 50.10 (9.20–>100f−2.99 (−12.47 to 3.76)  
Pheterogeneity  0.300 0.679 0.104 0.001 0.043 0.305 
Eye color 
 Brown/black 345 1.39 (0.78–1.99) 1.38 (0.65–2.10) 1.26 (0.14–2.35) 2.54 (0.26–4.68) 0.87 (−0.71 to 2.38) 0.44 (−1.94 to 2.65) 
 Hazel 294 1.24 (0.59–1.88) 1.26 (0.46–2.04) 1.15 (−0.01 to 2.27) 1.93 (−0.77 to 4.41) 1.42 (−0.09 to 2.86) −0.73 (−3.38 to 1.71) 
 Blue/green/gray 734 1.07 (0.54–1.60) 1.13 (0.49–1.76) 0.75 (−0.24 to 1.71) 1.63 (−0.43 to 3.58) 0.74 (−0.60 to 2.03) −0.36 (−2.54 to 1.70) 
 Other −6.53 (−28.78 to 4.89) −6.59 (−28.62 to 4.77) 3.19 (−0.32 to 6.10) 10.19 (4.03–22.38) 0.91 (−4.03 to 4.70)  
 Unknown 14 2.16 (0.10–4.01) 1.74 (−0.95 to 4.09)     
Pheterogeneity  0.450 0.700 0.409 0.056 0.789 0.613 
Natural hair color at age 15 
 Black 21 1.50 (−0.25 to 3.16) 1.37 (−0.65 to 3.29) 1.58 (−2.26 to 4.92) −0.76 (−13.68 to 7.67) 2.19 (−1.88 to 5.77) 0.15 (−1.88 to 2.05) 
 Brown/black 944 1.31 (0.81–1.81) 1.32 (0.71–1.92) 1.19 (0.27–2.08) 2.00 (0.09–3.83) 1.11 (−0.14 to 2.32)  
 Red/blonde 408 0.78 (0.17–1.39) 0.87 (0.14–1.58) 0.36 (−0.79 to 1.47) 1.92 (−0.50 to 4.19) 0.25 (−1.31 to 1.73) −1.21 (−3.80 to 1.19) 
 Other 1.21 (−2.43 to 4.22) 1.04 (−3.47 to 4.71) 1.39 (−5.63 to 6.47) −4.87 (−23.99 to 5.94) 4.40 (−3.10 to 11.06) −9.38 (−29.29 to 2.60) 
 Unknown 0.75 (−2.17 to 3.24) 0.14 (−4.03 to 3.53) 1.58 (−2.89 to 5.12) 9.88 (3.50–22.37) −2.80 (−12.37 to 3.75)  
Pheterogeneity  0.322 0.620 0.523 0.110 0.359 0.157 
Gaelic ancestry 
 No 689 1.65 (1.11–2.18) 1.71 (1.06–2.35) 1.41 (0.43–2.35) 2.61 (0.60–4.51) 1.37 (0.04–2.65) 0.01 (−2.14 to 2.03) 
 Yes 392 0.64 (0.04–1.23) 0.67 (−0.04 to 1.37) 0.31 (−0.84 to 1.43) 0.86 (−1.72 to 3.24) 0.43 (−1.10 to 1.89) −0.78 (−3.37 to 1.61) 
 Unknown 307 1.05 (0.40–1.67) 0.98 (0.21–1.74) 1.04 (−0.11 to 2.17) 2.53 (0.20–4.74) 0.63 (−0.94 to 2.12) 0.75 (−1.91 to 3.23) 
Pheterogeneity  <0.001 0.003 0.103 0.275 0.307 0.532 
Hispanic 
 No 1,372 1.32 (0.82–1.80) 1.32 (0.72–1.90) 1.21 (0.32–2.08) 2.40 (0.58–4.14) NA −0.07 (−2.05 to 1.79) 
 Yes 12 1.44 (−0.77 to 3.46) 1.79 (−0.56 to 3.99) −2.19 (−10.52 to 3.82) −1.10 (−13.47 to 7.12)  −4.10 (−17.74 to 4.44) 
 Unknown 0.56 (−3.75 to 3.94) 0.31 (−3.98 to 3.70)     
Pheterogeneity  0.915 0.789 0.300 0.451 NA 0.411 
Cancer in siblings and minimum age of development of cancer in any sibling 
 No cancer in siblings 1,176 1.24 (0.75–1.73) 1.29 (0.69–1.87) 0.99 (0.09–1.86) 2.18 (0.34–3.94) 0.86 (−0.36 to 2.04) −0.15 (−2.12 to 1.71) 
 Cancer in siblings, age unknown 1.42 (−1.65 to 4.36) 3.86 (0.22–7.98) −7.31 (−20.31 to 0.75) −9.74 (−35.04 to 3.30)  −7.62 (−24.94 to 2.25) 
 Cancer in siblings, age 30+ 140 0.96 (0.16–1.74) 1.07 (0.12–2.00) 0.70 (−0.80 to 2.14) 1.52 (−1.90 to 4.63) 0.70 (−1.23 to 2.54) −0.41 (−3.96 to 2.87) 
 Cancer in siblings, age <30 35 1.46 (−0.29 to 3.11) 1.14 (−1.13 to 3.24) 1.86 (−0.99 to 4.46) −0.75 (−9.15 to 5.73) 3.31 (−0.01 to 6.33) −2.75 (−11.64 to 3.73) 
 Unknown if siblings have cancer 31 1.78 (0.35–3.13) 1.31 (−0.39 to 2.92) 2.74 (0.01–5.20) 5.68 (1.93–9.24) 0.88 (−3.63 to 4.58)  
Pheterogeneity  0.843 0.688 0.157 0.096 0.502 0.475 
Age at menarche (years) 
 0–9 14 −0.02 (−2.88 to 2.46) 0.27 (−2.70 to 2.84) −3.91 (−15.19 to 3.91) 4.55 (1.41–7.39) −10.21 (−34.38 to 3.79) −2.16 (−16.02 to 6.88) 
 10–11 203 1.80 (1.08–2.51) 1.76 (0.88–2.61) 1.81 (0.48–3.07)  1.83 (0.07–3.48) −0.66 (−3.55 to 1.97) 
 12–14 777 1.09 (0.55–1.61) 1.18 (0.54–1.82) 0.82 (−0.14 to 1.75) 1.43 (−0.61 to 3.37) 0.74 (−0.56 to 2.00) −0.05 (−2.10 to 1.88) 
 15+ 94 0.82 (−0.16 to 1.76) 0.72 (−0.46 to 1.83) 0.90 (−1.00 to 2.65) 1.83 (−2.07 to 5.15) −0.23 (−3.34 to 2.46) 1.31 (−2.24 to 4.46) 
 Unknown 300 1.33 (0.71–1.94) 1.35 (0.63–2.06) 1.05 (−0.15 to 2.22) 3.06 (0.79–5.24) 1.01 (−0.56 to 2.51) −2.93 (−7.58 to 0.99) 
Pheterogeneity  0.166 0.426 0.419 0.153 0.320 0.475 
Age at menopause (years) 
 0–24 1.88 (−1.35 to 4.36) 1.07 (−2.98 to 4.06) 3.33 (−2.69 to 7.21) 4.23 (−1.84 to 8.23) −0.96 (<−100f to 0.00) 0.15 (−4.10 to 3.93) 
 25–34 55 1.47 (0.16–2.73) 0.52 (−1.27 to 2.22) 2.76 (0.74–4.65) 2.04 (−3.47 to 6.68) 3.75 (1.21–6.10)  
 35–44 87 1.62 (0.59–2.63) 1.32 (0.05–2.55) 2.54 (0.76–4.25) 4.06 (0.71–7.24) 1.78 (−0.93 to 4.30) 1.78 (−1.85 to 5.13) 
 45–54 100 0.26 (−0.84 to 1.34) 0.72 (−0.52 to 1.92) −0.36 (−2.68 to 1.80) −2.64 (−11.94 to 4.74) −1.79 (−5.03 to 1.19) 1.05 (−2.84 to 4.55) 
 55+ 2.18 (−1.04 to 5.08) 2.13 (−1.57 to 5.41) 2.90 (−3.50 to 7.77) 2.03 (0.14–3.84) 0.54 (−10.65 to 7.87) 3.82 (−5.00 to 10.73) 
 Unknown 1,129 1.29 (0.77–1.80) 1.41 (0.78–2.02) 0.71 (−0.24 to 1.63) NA 0.80 (−0.47 to 2.03) −1.24 (−3.54 to 0.91) 
Pheterogeneity  0.406 0.790 0.059 0.428 0.082 0.398 
Ever use oral contraception 
 No 291 0.85 (0.22–1.48) 0.95 (0.19–1.69) 0.61 (−0.57 to 1.77) 1.27 (−1.32 to 3.69) 0.14 (−1.59 to 1.80) −0.04 (−2.40 to 2.18) 
 Yes 813 1.35 (0.80–1.88) 1.37 (0.71–2.02) 1.15 (0.19–2.08) 2.48 (0.50–4.36) 1.07 (−0.22 to 2.32) −0.04 (−2.17 to 1.96) 
 Unknown 284 1.29 (0.66–1.91) 1.38 (0.64–2.11) 0.86 (−0.38 to 2.06) 2.53 (0.15–4.80) 0.74 (−0.85 to 2.27) −2.43 (−7.30 to 1.65) 
Pheterogeneity  0.218 0.417 0.604 0.515 0.484 0.500 
Duration of use of hormone replacement therapy (years) 
 0 18 1.15 (−1.24 to 3.38) 0.62 (−2.13 to 3.16) 2.89 (−1.96 to 7.00) 6.51 (0.82–11.57) 1.17 (−4.51 to 5.90) 3.81 (−5.80 to 11.75) 
 1–2 25 0.38 (−1.51 to 2.19) 0.26 (−1.94 to 2.35) 1.28 (−2.64 to 4.75)   −3.22 (−9.16 to 2.02) 
 3–4 25 1.23 (−0.54 to 2.94) 2.20 (0.04–4.29) −0.12 (−3.44 to 2.89) 0.63 (−3.74 to 4.58) 0.64 (−6.22 to 6.19) −0.25 (−10.90 to 7.94) 
 5–9 33 −0.30 (−2.02 to 1.35) 0.01 (−1.89 to 1.83) −1.36 (−5.82 to 2.61) −0.21 (−8.27 to 6.37) −1.19 (−6.57 to 3.49) 2.63 (−1.27 to 6.33) 
 10+ 48 0.61 (−0.65 to 1.85) 1.08 (−0.52 to 2.65) 0.42 (−1.69 to 2.45) −1.52 (−6.57 to 3.10) −0.01 (−2.96 to 2.80)  
 No/unknown 1,239 1.28 (0.78–1.77) 1.30 (0.70–1.89) 1.02 (0.12–1.90) 2.33 (0.48–4.09) 0.99 (−0.23 to 2.17) −0.52 (−2.59 to 1.42) 
Pheterogeneity  0.386 0.562 0.727 0.174 0.870 0.314 
Rheumatoid arthritis 
 No 1,318 1.24 (0.75–1.72) 1.27 (0.69–1.85) 1.03 (0.15–1.89) 2.42 (0.63–4.12) 0.93 (−0.27 to 2.08) −0.33 (−2.31 to 1.54) 
 Yes 41 0.74 (−0.59 to 1.99) 1.12 (−0.41 to 2.56) −0.60 (−3.63 to 1.99) −2.42 (−9.45 to 2.95) 0.94 (−3.63 to 4.55) −1.68 (−7.30 to 2.77) 
 Unknown 29 1.69 (0.22–3.10) 1.61 (−0.17 to 3.31) 1.83 (−0.98 to 4.37) NA 0.55 (−3.68 to 4.05) 1.91 (−2.38 to 5.87) 
Pheterogeneity  0.583 0.903 0.381 0.071 0.980 0.444 
Thyroiditis 
 No 1,152 1.27 (0.77–1.75) 1.41 (0.82–2.00) 0.87 (−0.04 to 1.76) 2.07 (0.21–3.84) 0.87 (−0.36 to 2.06) −0.56 (−2.60 to 1.35) 
 Yes 11 2.48 (0.27–4.55) 1.80 (−0.79 to 4.13) 5.01 (0.11–10.10) 3.22 (0.75–5.56) 6.85 (0.16–16.84) 1.51 (−7.62 to 8.89) 
 Unknown 225 0.95 (0.25–1.64) 0.66 (−0.20 to 1.50) 1.48 (0.26–2.66)  0.90 (−0.77 to 2.49) 1.14 (−1.56 to 3.67) 
Pheterogeneity  0.298 0.116 0.135 0.293 0.209 0.322 
Hyperthyroidism (Graves disease) 
 No 1,345 1.21 (0.72–1.69) 1.28 (0.70–1.86) 0.88 (−0.01 to 1.74) 2.22 (0.43–3.94) 0.85 (−0.35 to 2.01) −0.69 (−2.71 to 1.20) 
 Yes 17 2.47 (0.62–4.22) 1.38 (−0.89 to 3.48) 5.45 (1.91–9.13) 3.40 (−1.15 to 7.32) 6.33 (1.45–12.17) 2.94 (−2.84 to 8.49) 
 Unknown 26 1.24 (−0.45 to 2.82) −0.02 (−2.52 to 2.24) 2.63 (0.22–4.83)  0.20 (−4.40 to 3.93) 4.50 (0.30–8.47) 
Pheterogeneity  0.388 0.526 0.015 0.573 0.080 0.024 
Ever use diuretics 
 No 1,204 1.26 (0.76–1.74) 1.29 (0.70–1.88) 1.02 (0.12–1.89) 1.82 (−0.06 to 3.62) 0.97 (−0.23 to 2.14) 0.01 (−1.96 to 1.87) 
 Yes 167 1.07 (0.32–1.81) 1.22 (0.32–2.11) 0.69 (−0.75 to 2.06) 3.26 (0.66–5.71) 0.49 (−1.59 to 2.42) −2.40 (−5.84 to 0.70) 
 Unknown 17 1.19 (−0.79 to 3.01) 0.31 (−2.44 to 2.77) 2.40 (−0.59 to 5.08) 13.17 (4.71–31.80) 0.90 (−2.75 to 4.00) NA 
Pheterogeneity  0.863 0.736 0.534 0.014 0.866 0.088 
Smoking status 
 Missing+never smoked 713 1.33 (0.81–1.85) 1.49 (0.86–2.12) 0.90 (−0.05 to 1.83) 1.93 (−0.08 to 3.85) 0.92 (−0.36 to 2.15) −0.37 (−2.51 to 1.63) 
 Former smoker 434 1.07 (0.48–1.65) 1.03 (0.32–1.72) 1.00 (−0.11 to 2.07) 2.50 (0.29–4.60) 0.60 (−0.95 to 2.08) 0.08 (−2.28 to 2.29) 
 Current smoker 234 1.14 (0.42–1.84) 1.05 (0.18–1.89) 1.18 (−0.11 to 2.42) 2.78 (0.18–5.18) 1.13 (−0.64 to 2.79) −0.55 (−3.50 to 2.15) 
 Former smoker unknown if current 1.39 (−1.87 to 4.48) 1.20 (−2.09 to 4.32)     
Pheterogeneity  0.749 0.412 0.891 0.734 0.820 0.874 
Alcohol consumption (drinks per week) 
 Missing/unknown 23 1.18 (−0.49 to 2.77) 1.62 (−0.40 to 3.55) 0.12 (−3.12 to 2.99) −3.28 (−13.42 to 3.82) 1.04 (−3.26 to 4.73) −0.13 (−6.81 to 5.44) 
 <1 755 1.00 (0.47–1.52) 1.04 (0.40–1.66) 0.75 (−0.22 to 1.70) 2.58 (0.63–4.45) 0.76 (−0.56 to 2.04) −1.55 (−3.81 to 0.57) 
 1–2 214 1.47 (0.74–2.18) 1.63 (0.76–2.47) 0.95 (−0.43 to 2.26) 1.34 (−1.74 to 4.09) 0.93 (−0.93 to 2.66) 0.49 (−2.58 to 3.28) 
 3–6 238 1.51 (0.82–2.18) 1.68 (0.83–2.50) 1.26 (0.06–2.41) 2.49 (−0.04 to 4.84) 0.53 (−1.19 to 2.14) 1.32 (−1.14 to 3.60) 
 7+ 158 1.57 (0.83–2.30) 1.37 (0.48–2.23) 2.00 (0.58–3.34) 2.51 (−0.80 to 5.44) 2.05 (0.23–3.75) 1.24 (−2.14 to 4.30) 
Pheterogeneity  0.270 0.373 0.403 0.519 0.618 0.090 
BMI (kg/m2
 Missing 32 0.81 (−0.78 to 2.31) 0.54 (−1.33 to 2.27) 1.41 (−1.86 to 4.28) 1.43 (−4.89 to 6.47) 2.05 (−3.08 to 6.24) 0.29 (−6.81 to 6.01) 
 <18.5 56 0.95 (0.34–1.55) 0.97 (0.20–1.72) 0.59 (−0.46 to 1.62) 1.22 (−1.07 to 3.37) 0.59 (−0.87 to 1.99) −0.42 (−2.67 to 1.71) 
 18.5–24.9 930 1.10 (0.60–1.59) 1.22 (0.62–1.80) 0.74 (−0.16 to 1.62) 1.75 (−0.11 to 3.53) 0.75 (−0.47 to 1.93) −0.56 (−2.55 to 1.33) 
 25.0–29.9 267 1.17 (0.67–1.67) 1.20 (0.60–1.80) 0.98 (0.07–1.88) 2.47 (0.62–4.23) 0.88 (−0.37 to 2.09) −0.64 (−2.75 to 1.34) 
 ≥30 103 1.41 (0.84–1.96) 1.29 (0.60–1.96) 1.61 (0.59–2.60) 3.01 (0.94–4.97) 1.36 (−0.04 to 2.70) 0.66 (−1.63 to 2.81) 
Pheterogeneity  0.477 0.690 0.099 0.392 0.653 0.157 
Strenuous hours of exercise per week 
 Unknown 108 0.92 (0.05–1.77) 1.03 (0.02–2.02) 0.33 (−1.48 to 2.04) 1.82 (−1.70 to 4.98) 0.33 (−2.11 to 2.57) −1.52 (−6.22 to 2.53) 
 <1 873 1.27 (0.75–1.77) 1.31 (0.69–1.92) 1.02 (0.08–1.94) 2.81 (0.91–4.63) 0.80 (−0.48 to 2.03) −0.34 (−2.45 to 1.65) 
 1–3 249 1.24 (0.56–1.91) 1.30 (0.46–2.11) 1.07 (−0.14 to 2.23) 1.47 (−1.26 to 3.94) 1.07 (−0.60 to 2.63) 0.34 (−2.20 to 2.68) 
 4+ 158 1.16 (0.34–1.95) 1.15 (0.16–2.11) 1.08 (−0.39 to 2.47) 0.76 (−2.92 to 3.94) 1.67 (−0.33 to 3.51) −0.59 (−3.59 to 2.12) 
Pheterogeneity  0.855 0.933 0.862 0.458 0.741 0.826 
Occupational head and neck dose at baseline (Gy) 
 <0.01 68 0.60 (−0.84 to 1.90) 0.69 (−0.97 to 2.18) −0.11 (−3.05 to 2.38) −1.56 (−7.71 to 3.23) 1.56 (−2.85 to 4.75) −2.46 (−8.90 to 2.40) 
 0.010–0.049 669 1.57 (0.98–2.16) 1.76 (1.04–2.46) 0.85 (−0.24 to 1.89) 1.48 (−0.86 to 3.64) 1.13 (−0.29 to 2.48) −0.68 (−3.26 to 1.68) 
 0.050–0.099 371 0.79 (0.11–1.45) 0.69 (−0.12 to 1.49) 0.85 (−0.38 to 2.04) 3.41 (0.91–5.77) 0.76 (−0.86 to 2.31) −1.97 (−4.96 to 0.78) 
 0.100–0.199 211 1.45 (0.70–2.19) 1.48 (0.56–2.38) 1.72 (0.38–3.01) 2.63 (−0.14 to 5.26) 1.11 (−0.75 to 2.90) 1.40 (−1.37 to 3.98) 
 ≥0.200 69 0.61 (−0.68 to 1.87) 0.75 (−0.70 to 2.16) 0.55 (−2.33 to 3.22) 9.52 (1.28–19.16) −3.17 (−7.11 to 0.65) 2.00 (−2.95 to 6.43) 
Pheterogeneity  0.088 0.070 0.640 0.142 0.267 0.262 
log excess relative risk per UVR MJ/cm2 (95% CI) (= β)
CasesAll BCC, known anatomic locationsbHead and neckcUpper and lower extremity, trunkdUpper extremitydTrunkeLower extremityd
Skin complexion 
 Dark 12 2.03 (−0.38 to 4.12) 1.59 (−1.27 to 4.04) 2.95 (−1.98 to 6.59) 5.63 (−2.28 to 11.11) −14.28 (−40.56 to 1.57) 4.91 (−2.08 to 11.16) 
 Medium 540 1.46 (0.91–2.00) 1.41 (0.74–2.07) 1.51 (0.52–2.47) 3.06 (0.95–5.08) 1.53 (0.22–2.79) −0.35 (−2.67 to 1.82) 
 Fair 826 1.05 (0.53–1.56) 1.15 (0.52–1.76) 0.60 (−0.38 to 1.54) 1.21 (−0.83 to 3.14) 0.48 (−0.86 to 1.78) −0.27 (−2.36 to 1.69) 
 Unknown 10 1.10 (−1.68 to 3.66) 0.03 (−3.48 to 3.16) 3.24 (−1.64 to 7.71) 50.10 (9.20–>100f−2.99 (−12.47 to 3.76)  
Pheterogeneity  0.300 0.679 0.104 0.001 0.043 0.305 
Eye color 
 Brown/black 345 1.39 (0.78–1.99) 1.38 (0.65–2.10) 1.26 (0.14–2.35) 2.54 (0.26–4.68) 0.87 (−0.71 to 2.38) 0.44 (−1.94 to 2.65) 
 Hazel 294 1.24 (0.59–1.88) 1.26 (0.46–2.04) 1.15 (−0.01 to 2.27) 1.93 (−0.77 to 4.41) 1.42 (−0.09 to 2.86) −0.73 (−3.38 to 1.71) 
 Blue/green/gray 734 1.07 (0.54–1.60) 1.13 (0.49–1.76) 0.75 (−0.24 to 1.71) 1.63 (−0.43 to 3.58) 0.74 (−0.60 to 2.03) −0.36 (−2.54 to 1.70) 
 Other −6.53 (−28.78 to 4.89) −6.59 (−28.62 to 4.77) 3.19 (−0.32 to 6.10) 10.19 (4.03–22.38) 0.91 (−4.03 to 4.70)  
 Unknown 14 2.16 (0.10–4.01) 1.74 (−0.95 to 4.09)     
Pheterogeneity  0.450 0.700 0.409 0.056 0.789 0.613 
Natural hair color at age 15 
 Black 21 1.50 (−0.25 to 3.16) 1.37 (−0.65 to 3.29) 1.58 (−2.26 to 4.92) −0.76 (−13.68 to 7.67) 2.19 (−1.88 to 5.77) 0.15 (−1.88 to 2.05) 
 Brown/black 944 1.31 (0.81–1.81) 1.32 (0.71–1.92) 1.19 (0.27–2.08) 2.00 (0.09–3.83) 1.11 (−0.14 to 2.32)  
 Red/blonde 408 0.78 (0.17–1.39) 0.87 (0.14–1.58) 0.36 (−0.79 to 1.47) 1.92 (−0.50 to 4.19) 0.25 (−1.31 to 1.73) −1.21 (−3.80 to 1.19) 
 Other 1.21 (−2.43 to 4.22) 1.04 (−3.47 to 4.71) 1.39 (−5.63 to 6.47) −4.87 (−23.99 to 5.94) 4.40 (−3.10 to 11.06) −9.38 (−29.29 to 2.60) 
 Unknown 0.75 (−2.17 to 3.24) 0.14 (−4.03 to 3.53) 1.58 (−2.89 to 5.12) 9.88 (3.50–22.37) −2.80 (−12.37 to 3.75)  
Pheterogeneity  0.322 0.620 0.523 0.110 0.359 0.157 
Gaelic ancestry 
 No 689 1.65 (1.11–2.18) 1.71 (1.06–2.35) 1.41 (0.43–2.35) 2.61 (0.60–4.51) 1.37 (0.04–2.65) 0.01 (−2.14 to 2.03) 
 Yes 392 0.64 (0.04–1.23) 0.67 (−0.04 to 1.37) 0.31 (−0.84 to 1.43) 0.86 (−1.72 to 3.24) 0.43 (−1.10 to 1.89) −0.78 (−3.37 to 1.61) 
 Unknown 307 1.05 (0.40–1.67) 0.98 (0.21–1.74) 1.04 (−0.11 to 2.17) 2.53 (0.20–4.74) 0.63 (−0.94 to 2.12) 0.75 (−1.91 to 3.23) 
Pheterogeneity  <0.001 0.003 0.103 0.275 0.307 0.532 
Hispanic 
 No 1,372 1.32 (0.82–1.80) 1.32 (0.72–1.90) 1.21 (0.32–2.08) 2.40 (0.58–4.14) NA −0.07 (−2.05 to 1.79) 
 Yes 12 1.44 (−0.77 to 3.46) 1.79 (−0.56 to 3.99) −2.19 (−10.52 to 3.82) −1.10 (−13.47 to 7.12)  −4.10 (−17.74 to 4.44) 
 Unknown 0.56 (−3.75 to 3.94) 0.31 (−3.98 to 3.70)     
Pheterogeneity  0.915 0.789 0.300 0.451 NA 0.411 
Cancer in siblings and minimum age of development of cancer in any sibling 
 No cancer in siblings 1,176 1.24 (0.75–1.73) 1.29 (0.69–1.87) 0.99 (0.09–1.86) 2.18 (0.34–3.94) 0.86 (−0.36 to 2.04) −0.15 (−2.12 to 1.71) 
 Cancer in siblings, age unknown 1.42 (−1.65 to 4.36) 3.86 (0.22–7.98) −7.31 (−20.31 to 0.75) −9.74 (−35.04 to 3.30)  −7.62 (−24.94 to 2.25) 
 Cancer in siblings, age 30+ 140 0.96 (0.16–1.74) 1.07 (0.12–2.00) 0.70 (−0.80 to 2.14) 1.52 (−1.90 to 4.63) 0.70 (−1.23 to 2.54) −0.41 (−3.96 to 2.87) 
 Cancer in siblings, age <30 35 1.46 (−0.29 to 3.11) 1.14 (−1.13 to 3.24) 1.86 (−0.99 to 4.46) −0.75 (−9.15 to 5.73) 3.31 (−0.01 to 6.33) −2.75 (−11.64 to 3.73) 
 Unknown if siblings have cancer 31 1.78 (0.35–3.13) 1.31 (−0.39 to 2.92) 2.74 (0.01–5.20) 5.68 (1.93–9.24) 0.88 (−3.63 to 4.58)  
Pheterogeneity  0.843 0.688 0.157 0.096 0.502 0.475 
Age at menarche (years) 
 0–9 14 −0.02 (−2.88 to 2.46) 0.27 (−2.70 to 2.84) −3.91 (−15.19 to 3.91) 4.55 (1.41–7.39) −10.21 (−34.38 to 3.79) −2.16 (−16.02 to 6.88) 
 10–11 203 1.80 (1.08–2.51) 1.76 (0.88–2.61) 1.81 (0.48–3.07)  1.83 (0.07–3.48) −0.66 (−3.55 to 1.97) 
 12–14 777 1.09 (0.55–1.61) 1.18 (0.54–1.82) 0.82 (−0.14 to 1.75) 1.43 (−0.61 to 3.37) 0.74 (−0.56 to 2.00) −0.05 (−2.10 to 1.88) 
 15+ 94 0.82 (−0.16 to 1.76) 0.72 (−0.46 to 1.83) 0.90 (−1.00 to 2.65) 1.83 (−2.07 to 5.15) −0.23 (−3.34 to 2.46) 1.31 (−2.24 to 4.46) 
 Unknown 300 1.33 (0.71–1.94) 1.35 (0.63–2.06) 1.05 (−0.15 to 2.22) 3.06 (0.79–5.24) 1.01 (−0.56 to 2.51) −2.93 (−7.58 to 0.99) 
Pheterogeneity  0.166 0.426 0.419 0.153 0.320 0.475 
Age at menopause (years) 
 0–24 1.88 (−1.35 to 4.36) 1.07 (−2.98 to 4.06) 3.33 (−2.69 to 7.21) 4.23 (−1.84 to 8.23) −0.96 (<−100f to 0.00) 0.15 (−4.10 to 3.93) 
 25–34 55 1.47 (0.16–2.73) 0.52 (−1.27 to 2.22) 2.76 (0.74–4.65) 2.04 (−3.47 to 6.68) 3.75 (1.21–6.10)  
 35–44 87 1.62 (0.59–2.63) 1.32 (0.05–2.55) 2.54 (0.76–4.25) 4.06 (0.71–7.24) 1.78 (−0.93 to 4.30) 1.78 (−1.85 to 5.13) 
 45–54 100 0.26 (−0.84 to 1.34) 0.72 (−0.52 to 1.92) −0.36 (−2.68 to 1.80) −2.64 (−11.94 to 4.74) −1.79 (−5.03 to 1.19) 1.05 (−2.84 to 4.55) 
 55+ 2.18 (−1.04 to 5.08) 2.13 (−1.57 to 5.41) 2.90 (−3.50 to 7.77) 2.03 (0.14–3.84) 0.54 (−10.65 to 7.87) 3.82 (−5.00 to 10.73) 
 Unknown 1,129 1.29 (0.77–1.80) 1.41 (0.78–2.02) 0.71 (−0.24 to 1.63) NA 0.80 (−0.47 to 2.03) −1.24 (−3.54 to 0.91) 
Pheterogeneity  0.406 0.790 0.059 0.428 0.082 0.398 
Ever use oral contraception 
 No 291 0.85 (0.22–1.48) 0.95 (0.19–1.69) 0.61 (−0.57 to 1.77) 1.27 (−1.32 to 3.69) 0.14 (−1.59 to 1.80) −0.04 (−2.40 to 2.18) 
 Yes 813 1.35 (0.80–1.88) 1.37 (0.71–2.02) 1.15 (0.19–2.08) 2.48 (0.50–4.36) 1.07 (−0.22 to 2.32) −0.04 (−2.17 to 1.96) 
 Unknown 284 1.29 (0.66–1.91) 1.38 (0.64–2.11) 0.86 (−0.38 to 2.06) 2.53 (0.15–4.80) 0.74 (−0.85 to 2.27) −2.43 (−7.30 to 1.65) 
Pheterogeneity  0.218 0.417 0.604 0.515 0.484 0.500 
Duration of use of hormone replacement therapy (years) 
 0 18 1.15 (−1.24 to 3.38) 0.62 (−2.13 to 3.16) 2.89 (−1.96 to 7.00) 6.51 (0.82–11.57) 1.17 (−4.51 to 5.90) 3.81 (−5.80 to 11.75) 
 1–2 25 0.38 (−1.51 to 2.19) 0.26 (−1.94 to 2.35) 1.28 (−2.64 to 4.75)   −3.22 (−9.16 to 2.02) 
 3–4 25 1.23 (−0.54 to 2.94) 2.20 (0.04–4.29) −0.12 (−3.44 to 2.89) 0.63 (−3.74 to 4.58) 0.64 (−6.22 to 6.19) −0.25 (−10.90 to 7.94) 
 5–9 33 −0.30 (−2.02 to 1.35) 0.01 (−1.89 to 1.83) −1.36 (−5.82 to 2.61) −0.21 (−8.27 to 6.37) −1.19 (−6.57 to 3.49) 2.63 (−1.27 to 6.33) 
 10+ 48 0.61 (−0.65 to 1.85) 1.08 (−0.52 to 2.65) 0.42 (−1.69 to 2.45) −1.52 (−6.57 to 3.10) −0.01 (−2.96 to 2.80)  
 No/unknown 1,239 1.28 (0.78–1.77) 1.30 (0.70–1.89) 1.02 (0.12–1.90) 2.33 (0.48–4.09) 0.99 (−0.23 to 2.17) −0.52 (−2.59 to 1.42) 
Pheterogeneity  0.386 0.562 0.727 0.174 0.870 0.314 
Rheumatoid arthritis 
 No 1,318 1.24 (0.75–1.72) 1.27 (0.69–1.85) 1.03 (0.15–1.89) 2.42 (0.63–4.12) 0.93 (−0.27 to 2.08) −0.33 (−2.31 to 1.54) 
 Yes 41 0.74 (−0.59 to 1.99) 1.12 (−0.41 to 2.56) −0.60 (−3.63 to 1.99) −2.42 (−9.45 to 2.95) 0.94 (−3.63 to 4.55) −1.68 (−7.30 to 2.77) 
 Unknown 29 1.69 (0.22–3.10) 1.61 (−0.17 to 3.31) 1.83 (−0.98 to 4.37) NA 0.55 (−3.68 to 4.05) 1.91 (−2.38 to 5.87) 
Pheterogeneity  0.583 0.903 0.381 0.071 0.980 0.444 
Thyroiditis 
 No 1,152 1.27 (0.77–1.75) 1.41 (0.82–2.00) 0.87 (−0.04 to 1.76) 2.07 (0.21–3.84) 0.87 (−0.36 to 2.06) −0.56 (−2.60 to 1.35) 
 Yes 11 2.48 (0.27–4.55) 1.80 (−0.79 to 4.13) 5.01 (0.11–10.10) 3.22 (0.75–5.56) 6.85 (0.16–16.84) 1.51 (−7.62 to 8.89) 
 Unknown 225 0.95 (0.25–1.64) 0.66 (−0.20 to 1.50) 1.48 (0.26–2.66)  0.90 (−0.77 to 2.49) 1.14 (−1.56 to 3.67) 
Pheterogeneity  0.298 0.116 0.135 0.293 0.209 0.322 
Hyperthyroidism (Graves disease) 
 No 1,345 1.21 (0.72–1.69) 1.28 (0.70–1.86) 0.88 (−0.01 to 1.74) 2.22 (0.43–3.94) 0.85 (−0.35 to 2.01) −0.69 (−2.71 to 1.20) 
 Yes 17 2.47 (0.62–4.22) 1.38 (−0.89 to 3.48) 5.45 (1.91–9.13) 3.40 (−1.15 to 7.32) 6.33 (1.45–12.17) 2.94 (−2.84 to 8.49) 
 Unknown 26 1.24 (−0.45 to 2.82) −0.02 (−2.52 to 2.24) 2.63 (0.22–4.83)  0.20 (−4.40 to 3.93) 4.50 (0.30–8.47) 
Pheterogeneity  0.388 0.526 0.015 0.573 0.080 0.024 
Ever use diuretics 
 No 1,204 1.26 (0.76–1.74) 1.29 (0.70–1.88) 1.02 (0.12–1.89) 1.82 (−0.06 to 3.62) 0.97 (−0.23 to 2.14) 0.01 (−1.96 to 1.87) 
 Yes 167 1.07 (0.32–1.81) 1.22 (0.32–2.11) 0.69 (−0.75 to 2.06) 3.26 (0.66–5.71) 0.49 (−1.59 to 2.42) −2.40 (−5.84 to 0.70) 
 Unknown 17 1.19 (−0.79 to 3.01) 0.31 (−2.44 to 2.77) 2.40 (−0.59 to 5.08) 13.17 (4.71–31.80) 0.90 (−2.75 to 4.00) NA 
Pheterogeneity  0.863 0.736 0.534 0.014 0.866 0.088 
Smoking status 
 Missing+never smoked 713 1.33 (0.81–1.85) 1.49 (0.86–2.12) 0.90 (−0.05 to 1.83) 1.93 (−0.08 to 3.85) 0.92 (−0.36 to 2.15) −0.37 (−2.51 to 1.63) 
 Former smoker 434 1.07 (0.48–1.65) 1.03 (0.32–1.72) 1.00 (−0.11 to 2.07) 2.50 (0.29–4.60) 0.60 (−0.95 to 2.08) 0.08 (−2.28 to 2.29) 
 Current smoker 234 1.14 (0.42–1.84) 1.05 (0.18–1.89) 1.18 (−0.11 to 2.42) 2.78 (0.18–5.18) 1.13 (−0.64 to 2.79) −0.55 (−3.50 to 2.15) 
 Former smoker unknown if current 1.39 (−1.87 to 4.48) 1.20 (−2.09 to 4.32)     
Pheterogeneity  0.749 0.412 0.891 0.734 0.820 0.874 
Alcohol consumption (drinks per week) 
 Missing/unknown 23 1.18 (−0.49 to 2.77) 1.62 (−0.40 to 3.55) 0.12 (−3.12 to 2.99) −3.28 (−13.42 to 3.82) 1.04 (−3.26 to 4.73) −0.13 (−6.81 to 5.44) 
 <1 755 1.00 (0.47–1.52) 1.04 (0.40–1.66) 0.75 (−0.22 to 1.70) 2.58 (0.63–4.45) 0.76 (−0.56 to 2.04) −1.55 (−3.81 to 0.57) 
 1–2 214 1.47 (0.74–2.18) 1.63 (0.76–2.47) 0.95 (−0.43 to 2.26) 1.34 (−1.74 to 4.09) 0.93 (−0.93 to 2.66) 0.49 (−2.58 to 3.28) 
 3–6 238 1.51 (0.82–2.18) 1.68 (0.83–2.50) 1.26 (0.06–2.41) 2.49 (−0.04 to 4.84) 0.53 (−1.19 to 2.14) 1.32 (−1.14 to 3.60) 
 7+ 158 1.57 (0.83–2.30) 1.37 (0.48–2.23) 2.00 (0.58–3.34) 2.51 (−0.80 to 5.44) 2.05 (0.23–3.75) 1.24 (−2.14 to 4.30) 
Pheterogeneity  0.270 0.373 0.403 0.519 0.618 0.090 
BMI (kg/m2
 Missing 32 0.81 (−0.78 to 2.31) 0.54 (−1.33 to 2.27) 1.41 (−1.86 to 4.28) 1.43 (−4.89 to 6.47) 2.05 (−3.08 to 6.24) 0.29 (−6.81 to 6.01) 
 <18.5 56 0.95 (0.34–1.55) 0.97 (0.20–1.72) 0.59 (−0.46 to 1.62) 1.22 (−1.07 to 3.37) 0.59 (−0.87 to 1.99) −0.42 (−2.67 to 1.71) 
 18.5–24.9 930 1.10 (0.60–1.59) 1.22 (0.62–1.80) 0.74 (−0.16 to 1.62) 1.75 (−0.11 to 3.53) 0.75 (−0.47 to 1.93) −0.56 (−2.55 to 1.33) 
 25.0–29.9 267 1.17 (0.67–1.67) 1.20 (0.60–1.80) 0.98 (0.07–1.88) 2.47 (0.62–4.23) 0.88 (−0.37 to 2.09) −0.64 (−2.75 to 1.34) 
 ≥30 103 1.41 (0.84–1.96) 1.29 (0.60–1.96) 1.61 (0.59–2.60) 3.01 (0.94–4.97) 1.36 (−0.04 to 2.70) 0.66 (−1.63 to 2.81) 
Pheterogeneity  0.477 0.690 0.099 0.392 0.653 0.157 
Strenuous hours of exercise per week 
 Unknown 108 0.92 (0.05–1.77) 1.03 (0.02–2.02) 0.33 (−1.48 to 2.04) 1.82 (−1.70 to 4.98) 0.33 (−2.11 to 2.57) −1.52 (−6.22 to 2.53) 
 <1 873 1.27 (0.75–1.77) 1.31 (0.69–1.92) 1.02 (0.08–1.94) 2.81 (0.91–4.63) 0.80 (−0.48 to 2.03) −0.34 (−2.45 to 1.65) 
 1–3 249 1.24 (0.56–1.91) 1.30 (0.46–2.11) 1.07 (−0.14 to 2.23) 1.47 (−1.26 to 3.94) 1.07 (−0.60 to 2.63) 0.34 (−2.20 to 2.68) 
 4+ 158 1.16 (0.34–1.95) 1.15 (0.16–2.11) 1.08 (−0.39 to 2.47) 0.76 (−2.92 to 3.94) 1.67 (−0.33 to 3.51) −0.59 (−3.59 to 2.12) 
Pheterogeneity  0.855 0.933 0.862 0.458 0.741 0.826 
Occupational head and neck dose at baseline (Gy) 
 <0.01 68 0.60 (−0.84 to 1.90) 0.69 (−0.97 to 2.18) −0.11 (−3.05 to 2.38) −1.56 (−7.71 to 3.23) 1.56 (−2.85 to 4.75) −2.46 (−8.90 to 2.40) 
 0.010–0.049 669 1.57 (0.98–2.16) 1.76 (1.04–2.46) 0.85 (−0.24 to 1.89) 1.48 (−0.86 to 3.64) 1.13 (−0.29 to 2.48) −0.68 (−3.26 to 1.68) 
 0.050–0.099 371 0.79 (0.11–1.45) 0.69 (−0.12 to 1.49) 0.85 (−0.38 to 2.04) 3.41 (0.91–5.77) 0.76 (−0.86 to 2.31) −1.97 (−4.96 to 0.78) 
 0.100–0.199 211 1.45 (0.70–2.19) 1.48 (0.56–2.38) 1.72 (0.38–3.01) 2.63 (−0.14 to 5.26) 1.11 (−0.75 to 2.90) 1.40 (−1.37 to 3.98) 
 ≥0.200 69 0.61 (−0.68 to 1.87) 0.75 (−0.70 to 2.16) 0.55 (−2.33 to 3.22) 9.52 (1.28–19.16) −3.17 (−7.11 to 0.65) 2.00 (−2.95 to 6.43) 
Pheterogeneity  0.088 0.070 0.640 0.142 0.267 0.262 

Abbreviation: NA, not available.

aFor definition of anatomic site classification, see Supplementary Table S2. Analysis is based on follow-up starting at second questionnaire.

bAdjustment to the baseline BCC rate for baseline questionnaire, sex, ln[age], ln[age]2, birth year, birth year2, birth year3, birth year4, birth year5 and the respective risk factor.

cAdjustment to the baseline BCC rate for sex, ln[age], ln[age]2, birth year, birth year2, birth year3, birth year4, birth year5 and the respective risk factor.

dAdjustment to the baseline BCC rate for sex, ln[age], ln[age]2, birth year and the respective risk factor.

eAdjustment to the baseline BCC rate for sex, ln[age], birth year, birth year2 and the respective risk factor.

fWald-based confidence interval.

Cumulative (lifetime) UVR exposure of 1 MJ cm–2 was associated with an overall increase in the absolute risk of BCC of 9.62 (95% CI, 6.94–11.73; P < 0.001) per 10,000 persons per year (Supplementary Table S5). Supplementary Table S5 also shows that there was no significant difference between the absolute risks by sex overall (P = 0.105) or for any specific anatomic site. The variation in absolute risks by anatomic site was much more pronounced than the variation of relative risk, as shown by Fig. 2 and Supplementary Table S5. Excess absolute risk of BCC in relation to UVR was notably higher on the head and neck than on the trunk (Supplementary Table S5). The variation in excess absolute risk by anatomic site was highly significant for females (P = 0.017 for two-class analysis of heterogeneity, P = 0.006 for four-class analysis of heterogeneity) and borderline significant for males (P = 0.099 for two-class analysis of heterogeneity, P = 0.071 for four-class analysis of heterogeneity) but the difference was not significant when averaged over the two sexes (P > 0.1). Supplementary Table S4 demonstrates that adjusting for any of the potentially confounding lifestyles or constitutional variables had generally little effect on the absolute risk, although there was a notable reduction by about 30% in risk when adjusting for complexion [adjusted excess absolute risk (EAR) per MJ/cm2 per 10,000 persons per year = 10.63 vs. 15.23 unadjusted]. When adjusting for all variables simultaneously in the two separate variable groups, the ERRs remain significant using all the larger group of environmental and lifestyle risk factors (P = 0.004), but significance was lost when adjusting using all of the smaller group of familial variables (P = 0.138).

Figure 2.

Excess absolute risk per 104 person-years of follow-up for BCC in relation to cumulative UVR exposure, for all BCC with known anatomic location (A), head and neck BCC (B), upper extremity (C), trunk (D), and lower extremity BCC (E). Excess absolute risk (+95% CI) of BCC (+95% CI) over baseline level (defined by cumulative UVR exposure <800 kJ m–2) in relation to cumulative UVR exposure, with intervals 0–799, 800–999, 1,000–1,199, 1,200–1,399, 1,400–1,599, 1,600+ kJ m–2. The reference group used is the lowest UVR cumulative radiant exposure group, 0–799 kJ m–2. Baseline models are those specified in footnotes to Table 3.

Figure 2.

Excess absolute risk per 104 person-years of follow-up for BCC in relation to cumulative UVR exposure, for all BCC with known anatomic location (A), head and neck BCC (B), upper extremity (C), trunk (D), and lower extremity BCC (E). Excess absolute risk (+95% CI) of BCC (+95% CI) over baseline level (defined by cumulative UVR exposure <800 kJ m–2) in relation to cumulative UVR exposure, with intervals 0–799, 800–999, 1,000–1,199, 1,200–1,399, 1,400–1,599, 1,600+ kJ m–2. The reference group used is the lowest UVR cumulative radiant exposure group, 0–799 kJ m–2. Baseline models are those specified in footnotes to Table 3.

Close modal

Supplementary Table S6 evaluates the potential for modification of the association between cumulative UVR and risk of BCC according to constitutional, lifestyle, and other risk factors. As shown, there is significant heterogeneity of absolute risk per unit exposure for Gaelic ancestry for all sites and among head and neck BCC (P < 0.001), with EAR per MJ/cm2 among those with Gaelic ancestry of 30.05 (95% CI, 20.43–37.92) per 10,000 persons per year. There is also borderline significant (P = 0.035) heterogeneity by skin complexion for all sites, with risk particularly high for those with fair complexion, with overall EAR per MJ/cm2 of 19.21 (95% CI, 8.50–29.93) per 10,000 persons per year (Supplementary Table S6). ERRs were significant or borderline significant by level of occupational radiation dose overall (P = 0.014), for head and neck BCC (P = 0.008), for upper extremity BCC (P = 0.050) and for upper+lower extremity and trunk BCC (P = 0.060), with EAR per unit cumulative radiant exposure tending to be higher among those receiving intermediate (0.05–0.1 Gy) doses (Supplementary Table S6). There were significant or borderline significant variations in EAR per unit UVR cumulative radiant exposure for upper+lower extremity and trunk BCC by smoking status (P = 0.011) or by exercise status (P = 0.065), with EAR tending to be higher for those exercising more, and not current smokers; the same pattern in relation to exercise status was also observed for upper extremity BCC (P = 0.024). A complication with most of these findings was non-convergence of the fitted models. Otherwise there was little variation of EAR either overall or for particular anatomic locations (Supplementary Table S6).

In general, increasing the lag period resulted in increasing ERR and EAR, so that with a latency of 2 years the log ERR per MJ/cm2 was 1.21 (95% CI, 0.82–1.60; P < 0.001), which increased to 1.51 (95% CI, 0.99–2.02; P < 0.001) when the latency increased to 15 years, and likewise the EAR per MJ/cm2 per 10,000 persons per year was 8.68 (95% CI, 6.19–10.63; P < 0.001), which increased to 14.70 (95% CI, 11.00–17.50; P < 0.001) when the latency increased to 15 years (Supplementary Table S7). Supplementary Table S8 shows that there was little evidence (P > 0.15) of variation of EAR per MJ/cm2 cumulative radiant exposure per person year of follow-up with time after exposure, whether using the finer (four time window) or the broader (three time window) breakdown; a complication in interpreting these findings was non-convergence of the fitted models. However, there was evidence at borderline levels of statistical significance of variation of log ERR per MJ/cm2 cumulative radiant exposure with time after exposure, whether using the finer (four time window) breakdown (P = 0.026) or the broader (three time window) (P = 0.043) breakdown; log ERR per MJ/cm2 and the EAR per MJ/cm2 appeared to be largest for the period 5 to 14 years after exposure, with smaller risks outside this time since exposure window (Supplementary Table S8).

When competing risk models were used there was little change in log ERR overall, which became 1.25 (95% CI, 0.84–1.66; P < 0.001) per MJ/cm2 from 1.27 per MJ/cm2 using the standard model, but there was considerable diminution of EAR overall, so that the EAR per MJ/cm2 per 10,000 persons per year changed to 2.58 (95% CI, −2.42 to 5.97; P = 0.182) from 9.62 per MJ/cm2 using the standard model (Table 3; Supplementary Tables S5 and S9).

In this U.S. study, one of the first large, nationwide, prospective cohort investigations to evaluate ambient cumulative UVR exposure and BCC risk by anatomic site, we observed that estimated excess relative and absolute risk of BCC at all anatomic sites combined rose with increasing estimated cumulative lifetime ambient UVR exposure. We used a high-quality lifetime estimated ambient UVR exposure assessment based on residence lived in longest during each of five age periods linked with AVGLO environmental measurements (31, 36). Relative risks varied little by anatomic site (Table 3), whereas there was more pronounced variation by site in absolute risks, with the highest excess absolute risks per unit of UVR cumulative radiant exposure observed for BCC on the head and neck while lower excess absolute risks per unit of UVR exposure were seen on the trunk. Excess absolute risks of BCC on the upper extremities rose modestly with cumulative UVR, but there was no excess absolute risk of BCC of the lower extremities with UVR (Supplementary Table S5). We have also demonstrated modification of the cumulative ambient UVR-BCC risks with lower log ERR, but higher EAR among those with Gaelic/Celtic ancestry, and higher EAR among those with fair complexion but in general no overall significant and convincing difference between the sexes or modification by other constitutional, lifestyle, reproductive or medical factors, although for particular anatomic sites there was some observed heterogeneity (Table 4; Supplementary Table S6). The fact that the excess relative risk for those with Gaelic ancestry is lower (Table 4) but the excess absolute risk is higher (Supplementary Table S6) is consistent with the fact that the underlying risk of BCC is higher in the group with Gaelic ancestry, implying a submultiplicative interaction between the effect of ancestry and UVR-associated risk. This is not infrequently observed between the effects of some carcinogen and the baseline (e.g., genetic) risk factors (37–39).

Few previous studies have evaluated UVR and BCC by anatomic site, and none have involved large prospective cohort investigations examining dose–response employing cumulative estimates of exposure based on UVR measurements among residents from a large range of latitudes with broad assessment of potential confounders and effect modifiers. In a clinical series of 202 patients identified from skin cancer clinics in Queensland, Australia, Khalesi and colleagues (16) found that risk of BCC on solar-shielded skin (compared with solar-exposed skin) increased with increasing time spent outdoors and with numbers of actinic keratoses, but the correlation disappeared after adjusting for confounders (16). However, the study did not estimate excess absolute risks, and did not report potential effect modification by constitutional or lifestyle risk factors. Among 254 patients with histologically confirmed BCC ascertained from dermatology practices in Puerto Rico nonmelanoma skin cancers in sun-exposed areas were more common in participants who received high and very high UVR doses (17). The study design limitations of these two clinical series, both conducted in geographic regions with very high solar exposure precludes comparison with our large prospective cohort study at a broad range of latitudes. A population-based case–control study of 192 BCC and 700 controls (all identified in a population-based survey) from Western Australia found that risk of BCC on the head and neck and on the limbs declined whereas risk of BCC on the trunk rose with increasing total (cumulative occupational and recreational) sun exposure based on lifetime residential history and time outdoors on workdays and non-work days during warmer and cooler months (18). A study of 528 hospital-based BCC cases and 512 hospital controls with acute conditions from 16 centers in northern and central Italy found associations of recreational sun exposure and risk of both head/neck and trunk BCC and of occupational sun exposure with head/neck but not truncal BCC, although long duration of exposure was not linked with higher risk than short duration for any of these associations (40). Comparisons of the results of the two case–control studies with our prospective study are difficult due to study design differences, the differences in measurements (questionnaires about time spent outdoors vs. ground-based measurements), the emphasis on personal (case–control studies) versus ambient exposures (our cohort study), and metrics used (relative risk only vs. relative and excess absolute risk per unit UVR) and make comparison of results difficult. A small skin cancer cohort study in the town of Nambour, Queensland, Australia (22) assessed detailed anatomic site of 177 first BCC (598 total BCC) but had only a relatively crude 3-level occupational coding of likely UVR exposure (mainly outdoors, partly indoors and outdoors, mostly indoors).

The notably higher percentage of BCCs occurring on the face and head is well documented (41), but no studies have analyzed cumulative ambient UVR and risk of BCC by anatomic location. Our findings revealed that UVR-associated BCCs are also linked with increasing estimated cumulative ambient UVR on the trunk and, to a modest extent on the upper extremities. The trunk and upper extremities likely receive less exposure than the head and neck, possibly due to behavioral factors and clothing, but data on these factors were not collected. The findings may also be the result of some sort of abscopal effect, that is to say an effect in a part of the body distant to that exposed, something that has been suggested in a Ptch1+/− mouse model in relation to ionizing radiation exposure (42), and as also suggested by the UVA- and UVB-associated mutagenic bystander effects in vitro (43, 44). There are known involvements of UVR exposure on systemic immune function (45, 46), which could conceivably modify BCC risk on solar-unexposed parts of the body. These should be priority areas for future work in this area.

The metric used here, cumulative radiant UVR exposure, although not commonly used in epidemiologic studies, incorporates two important determinants of UVR, namely, latitude and age. As outlined previously, there are strong biophysical reasons why it should be employed (31, 36). Using data taken from that employed in the article of Little and colleagues (31), very similar to that used here, it can be estimated that at ages 65 and 80 an increase of 1 degree of latitude (mean taken over the ages 0–65, 0–80, respectively) corresponds to a reduction of cumulative radiant exposure of about 0.0240 and 0.0298 MJ cm–2, respectively. These are aggregate figures over the whole dataset; the precise degree of reduction for an individual will depend on their latitude and other factors.

The current study has several strengths including a large nationwide U.S. cohort of both sexes (80% female), residing in a wide range of latitudes, with long term follow-up and lifetime residential history. Use of ground-based AVGLO measurements of ambient UVR is a significant strength, combining the advantages of fine (1 km2 grid) geographical resolution, full-day average of solar exposure taking account of cloud cover, non-erythemally weighted measured irradiances, and ongoing data collection from 1960 to the present (36). In addition, our study provides the anatomic site of nearly two-thirds of the self-reported BCC, a large proportion of the BCCs were validated against medical records, and quantitative measurements for ambient UVR. The USRT cohort also had detailed information about constitutional, lifestyle, reproductive, and medical factors, as well as occupational information on ionizing radiation exposure that enabled us to evaluate potential effect modification across a wide variety of anatomic locations, and also assess latency and variations of EAR and ERR with time since exposure (Supplementary Tables S7 and S8), somewhat paralleling previous analysis (36).

Limitations of our study include self-report of anatomic location, possible mis-report of BCC not validated by medical records, restriction of the study population to White, indoor workers, and ambient UVR being based on self-reported residential history locations for five age periods, with no assessment of intermittent UVR exposures (e.g., during vacations) or other residential location not recorded in these five age periods. We also did not collect baseline information about number of hours spent outdoors or the clothing worn when subjects were outdoors, both of which are likely to correlate with latitude, or the use of tanning sunbeds. We do not have data on certain medical conditions and treatments, particularly disorders and medications associated with immune dysfunction or immune suppression (organ transplant, immunosuppressive drugs) and genetic factors (other than Gaelic ancestry or cancer in siblings at young age) that have been associated with increased risk of BCC (3). However, there is no reason to suppose that the prevalence of these factors and conditions would correlate with UVR exposure and so would not be expected to appreciably confound any trends of BCC with UVR. As noted in the Results, there are problems with lack of convergence, which particularly affect the absolute risk models (Supplementary Tables S5, S6, and S8). Only persons who survived and answer one of the first two questionnaires and the third questionnaire are included in this analysis. A degree of selection is thereby introduced. This would not be expected to bias log ERR, as risks are estimated in relation to a nominal baseline group, which would be equally affected by the selection, although downward bias would be expected for the measure of EAR, as excess cases would be removed by the selection.

We have demonstrated that BCC varies with increasing estimated lifetime ambient UVR exposure strongly for the head and neck, and to a somewhat lesser extent for the trunk, and modestly for the upper extremities. Our findings, of higher UVR-associated absolute excess risk associated with exposure of solar-exposed parts of the body such as the head and neck, and also of sites that might be presumed less solar exposed such as the trunk (Supplementary Table S5) parallel those suggested in other studies, albeit ones without the more comprehensive lifetime ambient UVR measures that we have (22). Our findings underscore the importance of skin cancer prevention guidelines to focus on the development of long-term sun protection habits, especially for the head and neck.

No disclosures were reported.

M.P. Little: Conceptualization, data curation, software, formal analysis, supervision, investigation, visualization, methodology, writing–original draft, writing–review and editing. T. Lee: Formal analysis, methodology, writing–review and editing. M.G. Kimlin: Conceptualization, visualization, methodology, writing–review and editing. C.M. Kitahara: Investigation, methodology, project administration, writing–review and editing. R. Zhang: Formal analysis, validation. B.H. Alexander: Resources, investigation, methodology, writing–review and editing. M.S. Linet: Resources, supervision, funding acquisition, methodology, project administration, writing–review and editing. E.K. Cahoon: Conceptualization, formal analysis, supervision, validation, visualization, methodology, writing–review and editing.

The authors would like to thank the two referees and the associate editor for their detailed and helpful remarks. The authors thank Elizabeth Platz and Paul Strickland for their input in the design and review of early study results. We would like to thank the radiologic technologists who participated in the USRT Study, Jerry Reid of the ARRT for continued support, and Diane Kampa and Allison Iwan of the University of Minnesota for study management and data collection.

This work was supported by the Intramural Research Program of the NIH, NCI, Division of Cancer Epidemiology and Genetics.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1.
Rogers
HW
,
Weinstock
MA
,
Feldman
SR
,
Coldiron
BM
. 
Incidence estimate of nonmelanoma skin cancer (keratinocyte carcinomas) in the US population, 2012
.
JAMA Dermatol
2015
;
151
:
1081
6
.
2.
Rogers
HW
,
Weinstock
MA
,
Harris
AR
,
Hinckley
MR
,
Feldman
SR
,
Fleischer
AB
, et al
Incidence estimate of nonmelanoma skin cancer in the United States, 2006
.
Arch Dermatol
2010
;
146
:
283
7
.
3.
Gossai
A
,
Barton
DT
,
Rees
JR
,
Nelson
HH
,
Karagas
MR
. 
Keratinocyte cancers
. In:
Thun
MJ
,
Linet
MS
,
Cerhan
JR
,
Haiman
CA
,
Schottenfeld
D
, editors.
Cancer epidemiology and prevention
. 4th ed.
New York, NY
:
Oxford University Press
; 
2018
.
p.
1089
117
.
4.
Chen
JT
,
Kempton
SJ
,
Rao
VK
. 
The economics of skin cancer: an analysis of Medicare payment data
.
Plastic Reconstr Surg Glob Open
2016
;
4
:
e868
.
5.
Eriksson
T
,
Tinghög
G
. 
Societal cost of skin cancer in Sweden in 2011
.
Acta Derm Venereol
2015
;
95
:
347
8
.
6.
Duarte
AF
,
Sousa-Pinto
B
,
Freitas
A
,
Delgado
L
,
Costa-Pereira
A
,
Correia
O
. 
Skin cancer healthcare impact: a nation-wide assessment of an administrative database
.
Cancer Epidemiol
2018
;
56
:
154
60
.
7.
Green
AC
,
Whiteman
DC
. 
Ultraviolet radiation
. In:
Thun
MJ
,
Linet
MS
,
Cerhan
JR
,
Haiman
CA
,
Schottenfeld
D
, editors.
Cancer epidemiology and prevention
. 4th ed.
New York, NY
:
Oxford University Press
; 
2018
.
p.
249
58
.
8.
Armstrong
BK
,
Kricker
A
. 
The epidemiology of UV induced skin cancer
.
J Photochem Photobiol B
2001
;
63
:
8
18
.
9.
Cahoon
EK
,
Kitahara
CM
,
Ntowe
E
,
Bowen
EM
,
Doody
MM
,
Alexander
BH
, et al
Female estrogen-related factors and incidence of basal cell carcinoma in a nationwide US cohort
.
J Clin Oncol
2015
;
33
:
4058
65
.
10.
Lee
T
,
Sigurdson
AJ
,
Preston
DL
,
Cahoon
EK
,
Freedman
DM
,
Simon
SL
, et al
Occupational ionising radiation and risk of basal cell carcinoma in US radiologic technologists (1983–2005)
.
Occup Environ Med
2015
;
72
:
862
9
.
11.
McDonald
E
,
Freedman
DM
,
Alexander
BH
,
Doody
MM
,
Tucker
MA
,
Linet
MS
, et al
Prescription diuretic use and risk of basal cell carcinoma in the nationwide U.S. radiologic technologists cohort
.
Cancer Epidemiol Biomarkers Prev
2014
;
23
:
1539
45
.
12.
Gerstenblith
MR
,
Rajaraman
P
,
Khaykin
E
,
Doody
MM
,
Alexander
BH
,
Linet
MS
, et al
Basal cell carcinoma and anthropometric factors in the U.S. radiologic technologists cohort study
.
Int J Cancer
2012
;
131
:
E149
55
.
13.
Cahoon
EK
,
Rajaraman
P
,
Alexander
BH
,
Doody
MM
,
Linet
MS
,
Freedman
DM
. 
Use of nonsteroidal anti-inflammatory drugs and risk of basal cell carcinoma in the United States Radiologic Technologists Study
.
Int J Cancer
2012
;
130
:
2939
48
.
14.
Yanik
EL
,
Pfeiffer
RM
,
Freedman
DM
,
Weinstock
MA
,
Cahoon
EK
,
Arron
ST
, et al
Spectrum of immune-related conditions associated with risk of keratinocyte cancers among elderly adults in the United States
.
Cancer Epidemiol Biomarkers Prev
2017
;
26
:
998
1007
.
15.
Zamoiski
RD
,
Yanik
E
,
Gibson
TM
,
Cahoon
EK
,
Madeleine
MM
,
Lynch
CF
, et al
Risk of second malignancies in solid organ transplant recipients who develop keratinocyte cancers
.
Cancer Res
2017
;
77
:
4196
203
.
16.
Khalesi
M
,
Whiteman
DC
,
Rosendahl
C
,
Johns
R
,
Hackett
T
,
Cameron
A
, et al
Basal cell carcinomas on sun-protected vs. sun-exposed body sites: a comparison of phenotypic and environmental risk factors
.
Photodermatol Photoimmunol Photomed
2015
;
31
:
202
11
.
17.
Ramos
J
,
Villa
J
,
Ruiz
A
,
Armstrong
R
,
Matta
J
. 
UV dose determines key characteristics of nonmelanoma skin cancer
.
Cancer Epidemiol Biomarkers Prev
2004
;
13
:
2006
11
.
18.
Kricker
A
,
Armstrong
BK
,
English
DR
,
Heenan
PJ
. 
A dose-response curve for sun exposure and basal cell carcinoma
.
Int J Cancer
1995
;
60
:
482
8
.
19.
Pelucchi
C
,
Di Landro
A
,
Naldi
L
,
La Vecchia
C
Oncology Study Group of the Italian Group for Epidemiologic Research in Dermatology
. 
Risk factors for histological types and anatomic sites of cutaneous basal-cell carcinoma: an Italian case-control study
.
J Invest Dermatol
2007
;
127
:
935
44
.
20.
Neale
RE
,
Davis
M
,
Pandeya
N
,
Whiteman
DC
,
Green
AC
. 
Basal cell carcinoma on the trunk is associated with excessive sun exposure
.
J Am Acad Dermatol
2007
;
56
:
380
6
.
21.
Wu
S
,
Han
J
,
Li
WQ
,
Li
T
,
Qureshi
AA
. 
Basal-cell carcinoma incidence and associated risk factors in U.S. women and men
.
Am J Epidemiol
2013
;
178
:
890
7
.
22.
Richmond-Sinclair
NM
,
Pandeya
N
,
Ware
RS
,
Neale
RE
,
Williams
GM
,
van der Pols
JC
, et al
Incidence of basal cell carcinoma multiplicity and detailed anatomic distribution: longitudinal study of an Australian population
.
J Invest Dermatol
2009
;
129
:
323
8
.
23.
Boice
JD
 Jr
,
Mandel
JS
,
Doody
MM
,
Yoder
RC
,
McGowan
R
. 
A health survey of radiologic technologists
.
Cancer
1992
;
69
:
586
98
.
24.
Doody
MM
,
Mandel
JS
,
Lubin
JH
,
Boice
JD
 Jr
. 
Mortality among United States radiologic technologists, 1926–90
.
Cancer Causes Control
1998
;
9
:
67
75
.
25.
Sigurdson
AJ
,
Doody
MM
,
Rao
RS
,
Freedman
DM
,
Alexander
BH
,
Hauptmann
M
, et al
Cancer incidence in the US Radiologic Technologists Health Study, 1983–1998
.
Cancer
2003
;
97
:
3080
9
.
26.
Brenner
M
,
Hearing
VJ
. 
The protective role of melanin against UV damage in human skin
.
Photochem Photobiol
2008
;
84
:
539
49
.
27.
Yoshinaga
S
,
Hauptmann
M
,
Sigurdson
AJ
,
Doody
MM
,
Freedman
DM
,
Alexander
BH
, et al
Nonmelanoma skin cancer in relation to ionizing radiation exposure among U.S. radiologic technologists
.
Int J Cancer
2005
;
115
:
828
34
.
28.
Scotto
J
,
Fears
TR
,
Fraumeni
JF
 Jr
.
Incidence of nonmelanoma skin cancer in the United States
.
Bethesda, MD
:
NIH
; 
1983
:
83-2433. p. i-xv+1-113
.
29.
Sliney
DH
International Commission on Illumination
. 
Radiometric quantities and units used in photobiology and photochemistry: recommendations of the Commission Internationale de L'Eclairage (International Commission on Illumination)
.
Photochem Photobiol
2007
;
83
:
425
32
.
30.
Tatalovich
Z
,
Wilson
JP
,
Cockburn
M
. 
A comparison of Thiessen polygon, kriging, and spline models of potential UV exposure
.
Cartogr Geograph Inf Sci
2006
;
33
:
217
31
.
31.
Little
MP
,
Tatalovich
Z
,
Linet
MS
,
Fang
M
,
Kendall
GM
,
Kimlin
MG
. 
Improving assessment of lifetime solar ultraviolet radiation exposure in epidemiologic studies: comparison of ultraviolet exposure assessment methods in a nationwide U.S. occupational cohort
.
Photochem Photobiol
2018
;
94
:
1297
307
.
32.
United Nations Scientific Committee on the Effects of Atomic Radiation (UNSCEAR)
.
UNSCEAR 2006 Report. Annex A. Epidemiological studies of radiation and cancer
.
New York
:
United Nations
; 
2008
. p.
13
322
.
33.
R Core Team
.
R: A language and environment for statistical computing
.
R Foundation for Statistical Computing, Vienna, Austria
; 
2019
.
Available from:
https://www.r-project.org.
34.
Risk Sciences International
.
Epicure version 2.0.1.0. 55 Metcalfe, K1P 6L5
,
Canada
:
Risk Sciences International
; 
2015
.
35.
McCullagh
P
,
Nelder
JA
.
Generalized linear models
. 2nd edition.
Boca Raton, FL
:
Chapman and Hall/CRC
; 
1989
. p.
1
526
.
36.
Little
MP
,
Linet
MS
,
Kimlin
MG
,
Lee
T
,
Tatalovich
Z
,
Sigurdson
AJ
, et al
Cumulative solar ultraviolet radiation exposure and basal cell carcinoma of the skin in a nationwide US cohort using satellite and ground-based measures
.
Environ Health
2019
;
18
:
114
.
37.
Blechter
B
,
Wong
JYY
,
Agnes Hsiung
C
,
Hosgood
HD
,
Yin
Z
,
Shu
XO
, et al
Sub-multiplicative interaction between polygenic risk score and household coal use in relation to lung adenocarcinoma among never-smoking women in Asia
.
Environ Int
2021
;
147
:
105975
.
38.
Little
MP
,
Schaeffer
ML
,
Reulen
RC
,
Abramson
DH
,
Stovall
M
,
Weathers
R
, et al
Breast cancer risk after radiotherapy for heritable and non-heritable retinoblastoma: a US-UK study
.
Br J Cancer
2014
;
110
:
2623
32
.
39.
Little
MP
,
de Vathaire
F
,
Shamsaldin
A
,
Oberlin
O
,
Campbell
S
,
Grimaud
E
, et al
Risks of brain tumour following treatment for cancer in childhood: modification by genetic factors, radiotherapy and chemotherapy
.
Int J Cancer
1998
;
78
:
269
75
.
40.
Pelucchi
C
,
Naldi
L
,
Di Landro
A
,
La Vecchia
C
Oncology Study Group of Italian Group for Epidemiologic Research in Dermatology
. 
Anthropometric measures, medical history and risk of basal cell carcinoma in an Italian case-control study
.
Dermatology
2008
;
216
:
271
6
.
41.
Karagas
MR
,
Weinstock
MA
,
Nelson
HH
. 
Keratinocyte carcinomas (basal and squamous cell carcinomas of the skin)
. In:
Schottenfeld
D
,
Fraumeni
JF
 Jr
, editors.
Cancer epidemiology and prevention
. 3rd ed.
New York
:
Oxford University Press
; 
2006
. p.
1230
50
.
42.
Mancuso
M
,
Leonardi
S
,
Giardullo
P
,
Pasquali
E
,
Tanori
M
,
De Stefano
I
, et al
Oncogenic radiation abscopal effects in vivo: interrogating mouse skin
.
Int J Radiat Oncol Biol Phys
2013
;
86
:
993
9
.
43.
Dahle
J
,
Kaalhus
O
,
Stokke
T
,
Kvam
E
. 
Bystander effects may modulate ultraviolet A and B radiation-induced delayed mutagenesis
.
Radiat Res
2005
;
163
:
289
95
.
44.
Dahle
J
,
Kvam
E
,
Stokke
T
. 
Bystander effects in UV-induced genomic instability: antioxidants inhibit delayed mutagenesis induced by ultraviolet A and B radiation
.
J Carcinog
2005
;
4
:
11
.
45.
González Maglio
DH
,
Paz
ML
,
Leoni
J
. 
Sunlight effects on immune system: is there something else in addition to UV-induced immunosuppression?
Biomed Res Int
2016
;
2016
:
1934518
.
46.
Bernard
JJ
,
Gallo
RL
,
Krutmann
J
. 
Photoimmunology: how ultraviolet radiation affects the immune system
.
Nat Rev Immunol
2019
;
19
:
688
701
.
47.
Murari
A
,
Singh
KN
. 
Lund and Browder chart-modified versus original: a comparative study
.
Acute Crit Care
2019
;
34
:
276
81
.

Supplementary data