Fat Intake and Risk of Squamous Cell Carcinoma of the Skin.

Abstract: The varied effects of different classes of dietary fatty acids on carcinogenesis suggest that fatty acid composition is an important determining factor in tumor development. In the present study, we investigated the association between dietary n-3 and n-6 fatty acid intake and risk of squamous cell carcinoma of the skin (SCC). Data were taken from a population-based case-control study of skin SCC in Southeastern Arizona. Our data show a consistent tendency for a lower risk of SCC with higher intakes of n-3 fatty acids [p (for trend) = 0.055]. The adjusted odds ratios for increasing levels of n-3 fatty acids were 0.85 [95% confidence interval (CI) = 0.56-1.27] and O. 71 (95% CI = 0.49-1.00) compared with the lower level as the referent. For the ratio of n-3 to n-6 fatty acids, the odds ratios in successively higher levels were 0.88 (95% CI = 0.59-1.32) and 0.74 (95% CI = 0.51-1.05), suggesting a tendency toward decreased risk of SCC with increased intake of diets with high ratio of n-3 to n-6 fatty acid. More studies are clearly needed to elucidate the function of dietary fatty acids so that recommendations can be made to alter the human diet for cancer prevention, particularly in light of the increasing incidence of SCC of the skin.

Introduction
Nonmelanoma skin cancer, which includes squamous cell carcinoma (SCC) and basal cell carcinoma, is the most common type of cancer in the United States and results in substantial morbidity and treatment costs. An estimated 900,000-1,200,000 new cases of nonmelanoma skin cancer are diagnosed each year in the United States ( 1). Residents of Arizona experience a three to seven times greater incidence of nonmelanoma skin cancer than the general population in the United States ( 2).

Several animal studies showed that a high level of dietary fat intake markedly shortened the time between ultraviolet radiation (UV) exposure and tumor appearance and increased the number of tumors that developed. Furthermore, high levels of dietary fat affected skin cancer development at the promotional stage of UV carcinogenesis ( 3). In addition, experimental studies in humans showed that switching to a low-fat diet significantly reduced the occurrence of actinic keratosis (AK) ( 4) and nonmelanoma skin cancer ( 5, 6).

The varied effects of different classes of dietary fatty acids on carcinogenesis suggest that fatty acid composition is an important determining factor in tumor development. There are strong epidemiological, clinical, and experimental data indicating a protective effect of the principal dietary n-3 polyunsaturated fatty acids {PUFAs: eicosapentaenoic acid [EPA, 20:5(n-3)] and docosahexaenoic acid [DHA, 22:6(n-3)]} on cancer development ( 7-11). Others have suggested, however, that the critical relationship may be the ratio of n-3 to n-6 fatty acids that occurs in the diet ( 11, 12).

In the present study, we investigated the association between dietary n-3 and n-6 fatty acid intake and risk of skin SCC. Data were taken from a population-based case-control study of SCC of the skin in Southeastern Arizona.

Subjects and Methods
A population-based case-control study was conducted with cases of SCC of the skin recruited from persons identified through the Southeastern Arizona Skin Cancer Registry. Control subjects were recruited through a random-digit dialing process. All 792 subjects (cases and controls) completed an extensive interview for demographic, behavioral, and past UV exposure information. Subjects then were asked to complete four 24-hour dietary recalls. This report uses data from those individuals who completed all four dietary recalls (n = 568).

Study Population
Cases were eligible if they were >30 years of age, had a histopathologically confirmed nonmetastatic SCC of the skin diagnosed within four months before contact, and had no prior history of a skin cancer. All cases were randomly selected from cases identified by the Southeastern Arizona Skin Cancer Registry and selected to reflect the age and gender distribution of SCC within the registry. On identification of a potential case from the registry, a letter was sent to the personal physician requesting permission to contact the patient regarding study participation. Once the physician granted permission, a letter describing the study was sent to the patient; a study interviewer then contacted the patient by phone to describe the study, determine eligibility, invite participation, and schedule the interview. Over 83% of identified cases from the registry were located and interviewed for eligibility and interest. Of these interviewed cases, 404 were eligible and participated in the baseline study. Only non-Hispanic (Anglo) and Hispanic white cases were eligible. Nonwhite subjects were defined as self-reporting at least one nonwhite parent.

Population-based controls were selected using random-digit dialing. Telephone numbers were randomly generated from the cases' residential telephone numbers. These numbers were dialed until each was resolved as a nonworking number, nonresidential number, or residential telephone. Nonresidential or nonworking numbers were excluded and replaced. At least six attempts were made to interview a member of the selected residence before replacement. If no eligible persons were identified, a new telephone number was selected. One control per household was invited to participate using modified Waksberg criteria ( 13). Controls were frequency matched to the cases by age category (same 10-yr group) and gender. Control subjects were considered eligible if they had no prior history of skin cancer, lived within the Tucson region, and were within the age, gender, and ethnicity grouping.

Dietary Assessment and Quality Control
All subjects also completed four 24-hour dietary recalls, which included four randomly selected days (3 weekdays and 1 weekend) within two weeks of the clinic visit. Subjects were called by our trained interviewer who conducted the 24-hour recalls, recorded the data, and entered the dietary information. One interviewer conducted >90% of the personal interviews and 100% of the dietary recalls. All the dietary recalls were tape-recorded, and 10% of the recorded tapes were randomly selected and reviewed by staff from the Nutrition Core of the Arizona Cancer Center. A randomly selected sample (10%) of the 24-hour dietary recalls was reentered by a staff member of the Nutrition Core for quality control and quality assurance procedures. The results of quality control and quality assurance were very high, with >97% concurrence. Daily mean nutrient intakes, including different classes of fatty acids, were calculated with the use of the Minnesota Nutrition Data System (version 2.9) ( 14).

Other Variables and Quality Control
All participants completed a structured interview detailing personal, behavioral, and demographic characteristics. The interview instrument sought information on skin characteristics, sunburns and tanning history, use of suntan lotions and sunscreens, residential history, UV exposure during the past year, family history of skin cancer, past medical history, tobacco and alcohol use, physical characteristics, and demographic information. After each interview, questionnaires were reviewed for completeness and coded. Data entry was through screen-based entry programs that included range checks.

Data Analyses
Distribution of demographic characteristics and potential risk factors were compared between cases and controls using Student's t-tests for continuous variables and X2 tests for categorical variables. Tests for trend were also calculated. The various types of dietary fatty acids between cases and controls were compared using t-tests. Means of dietary intake were adjusted for age and gender.

Crude odds ratios (ORs) and 95% confidence intervals (CIs) were calculated. Adjusted ORs were then calculated using multiple logistic regressions. The various types of fatty acids were stratified on the basis of three levels of dietary intake. The cutoffs for the three levels of fatty acid intake are based on the 50th (median) and 90th percentiles of daily intake in grams among controls (<50th, >50th to <90th, >90th). The cutoffs for the three levels of energy intake from fat (<20%, >20% to <35%, >35%) are based on the levels used in human experimental studies ( 3, 4). We assessed the potential confounding or modifying effects of age, gender, education, energy intake (kcal/day), alcohol intake (mean alcohol intake/day), smoking history (never, former smoker, and current smoker), history of actinic skin damage (self-reported diagnosed AK), self-reported ability to tan after prolonged sun exposure (no suntan/possible freckling, mildly tanned/possible freckling, moderately tanned, and deeply tanned), number of current freckles on the arms, sun exposure, and use of sunscreen. The initial adjusted model included age, gender, and energy intake (kcal/day). Other multivariate models were evaluated to determine impact of other skin cancer risk factors as potential confounding factors. Inclusion of alcohol intake and smoking status did not alter any of the results, and these variables were excluded from the final model. Age, gender, energy intake, inability to tan after prolonged sun exposure, history of diagnosed and treated AK, number of current freckles on the arms, and use of sunscreen were included in the final multivariate model. All statistical analyses were done by using Stata (College Station, TX) computer software ( 15); p

Results
The study population (n = 568) is composed of older educated Southwestern US residents, with 65.8% of cases and 64.1% of controls having some college education. There was no difference in socioeconomic level between cases and controls. We compared subjects who participated in the diet study with those who did not participate. We found no statistically significant difference between the two groups in relation to case-control status (X2 = 0.46, p = 0.50), age (t =-1.37, p = 0.17), gender (X2 = 1.71, p = 0.19), smoking (X2 = 1.07, p = 0.59), average hours in the sun per day (t = 0.86, p = 0.39), tanning ability (X2 = 1.22, p = 0.27), number of current freckles on the arm (X2 = 2.06, p = 0.84), and use of sunscreen (X2 = 6.90, p = 0.14). Participants in the diet study reported more AK history than subjects who did not participate in the study (X2 = 9.7, p = 0.00). This difference between the two groups was consistent among cases and controls.

Table 1 shows the distribution of cases and controls according to gender, age, education, smoking status, alcohol intake, reported tanning ability, history of AK, use of sun screen, daily hours of sun exposure, and total number of current freckles on the arms. Only tanning ability, history of AK, number of freckles, and use of sunscreen show a significant difference between cases and controls (p < 0.001).

Table 2 presents the ORs according to intake of various types of fat, total energy, and cholesterol. There was a positive nonsignificant association between the percentage of caloric intake from fat and the risk of skin SCC [p (for trend) = 0.09]. There was a tendency for a lower risk of skin SCC with increasing intake of total monounsaturated fatty acids (OR = 0.86, 95% CI = 0.60-1.43). We examined the associations between total PUFAs and total energy, fat, and cholesterol intake in relation to skin SCC. No consistent patterns of association were observed.

The findings regarding skin SCC risk in relation to levels of intake of the various types of PUFA are presented in Table 3. There was a consistent pattern for marginally lower risks of skin SCC with increasing intakes of EPA (OR = 0.78, 95% CI = 0.54-1.10), DHA (OR = 0.80, 95% CI = 0.57-1.31), and arachidonic acid (AA; OR = 0.70, 95% CI = 0.46-1.08), although the trends were not statistically significant.

Table 4 displays the ORs and 95% CIs for skin SCC according to levels of dietary intake of long-chain PUFA and various fatty acid ratios. The adjusted ORs and 95% CIs for increasing levels of n-3 fatty acids (EPA + DHA) were 0.85 (95% CI = 0.56-1.27) and 0.71 (95% CI = 0.49-1.00) compared with the lower level as the referent [p (for trend) = 0.055]. We found no relationship between risk of skin SCC and dietary intake of n-6 fatty acids. Because n-3 fatty acids compete with fatty acids in the n-6 family in eicosanoid biosynthesis and because n-3 fatty acids can extensively alter membrane lipid composition and reduce the production of prostaglandins derived from AA, we elected to evaluate the ratio of the total level of n-3 fatty acids to the total level of n-6 fatty acids (n-3/n-6). Other ratios that were evaluated included ratios of linolenic acid to linoleic acid (LA) and the ratio of PUFA to saturated fatty acid (Table 4). For n-3/n-6, the ORs at successively higher levels were 0.88 (95% CI = 0.59-1.32) and 0.74 (95% CI = 0.51-1.05), with the lowest level as the referent group. The adjusted ORs and 95% CIs for increasing levels of linolenic acid-to-LA ratio were 0.98 (95% CI = 0.65-1.46) and 0.81 (95% CI = 0.61-1.08).

In Table 5, the associations between percent energy from fat, sum of n-3 and n-6 fatty acids, and n-3/n-6 and the risk of skin SCC are analyzed in the strata of selected covariates. Percent energy from fat was found to be positively associated with risk of skin SCC in the various strata. This higher risk was significant among men (OR = 1.58, 95% CI = 1.01-2.48). There was a consistent tendency for a lower risk of skin SCC with increasing intake of n-3 fatty acids, with the OR ranging from 0.55 to 0.93 in the various strata. This lower risk was statistically significant among men (OR = 0.68, 95% CI = 0.47-0.97). We found no consistent relationship between risk of skin SCC and dietary intake of n-6 fatty acids. A higher n-3/n-6 in the diet was associated with lower risk of skin SCC in the various strata, and this lower risk was statistically significant among men (OR = 0.68, 95% CI = 0.46-0.99). However, for 12 of 13 evaluations, the OR estimate was lower for the sum of n-3 fatty acids than for n-3/n-6.

Discussion
In animal models, high levels of dietary fat profoundly suppress specific immune responses; in particular, high dietary fat appears to exacerbate the UV carcinogenic expression effect through immunosuppression in the skin ( 16). High dietary fat has been shown to elevate levels of prostaglandin E2 (PGE2), which is known to act as an immunoregulator of T cell function and to modulate UV carcinogenesis ( 16). These observations make a strong circumstantial case that high intake of dietary fat, especially of dietary fat rich in essential fatty acids, potentiates UV carcinogenesis via regulation of prostanoid metabolism in a manner that consequently suppresses immune responses that control the outgrowth of UV-induced tumors ( 16).

Results from epidemiological and experimental studies regarding the influence of dietary fat have been in conflict. Animal studies showed that restriction of fat calories resulted in fewer skin papillomas and carcinomas ( 17, 18). Other studies have indicated that total energy intake is a more stringent determinant of tumor growth than fat intake, because energy reduction significantly inhibits tumor growth, even when the diets are relatively high in fat ( 19). Our results show a positive nonsignificant association between the percent energy from fat and the risk of skin SCC. Furthermore, percent energy from fat was found to be positively associated with risk of skin SCC in the various strata studied. This higher risk was statistically significant among men (OR = 1.58, 95% CI = 1.01-2.48)

The degree of dietary fatty acid saturation alone does not seem to be a determinant in modulation of UV carcinogenesis. Composition of dietary fat may contribute to promotion and progression of skin carcinoma ( 20-22). The promoting effects of dietary fat are highly dependent on the type of fat ( 20, 23, 24). Essential n-6 [LA, 18:2(n-6)] and n-3 [alpha-linolenic, 18:3(n-3)] fatty acids are important modulators of the carcinogenic process and prostanoid metabolism. Prostaglandins, particularly 2-series prostaglandins, are recognized as important participants in the inflammatory response to UV, and it has been shown that n-3 fatty acids markedly lower cutaneous PGE2 levels and reduce sunburn response ( 26).

The principal n--6 fatty acid in the diet is LA, which is metabolized in some host tissues and tumor cells by a system of desaturase and elongase enzymes to AA. Our data show no consistent pattern of association between dietary intake of LA and risk of skin SCC. The major prostaglandin of mammalian epidermis is PGE2 ( 27), and human skin homogenates have been shown to be capable of transforming AA via cyclooxygenase to prostaglandins ( 28). To our surprise, we found a consistent tendency for a lower risk of skin SCC with increasing intake of AA. The vast number of compounds that can be produced via the oxidation of AA has led to the appreciation of the potential complexity of this lipid. However, the relationship of dietary AA to the production of these compounds and their subsequent biological impact remains obscure ( 29). An animal study showed that although dietary LA stimulated mammary tumor growth, this was not apparent when pure AA was added to the diet ( 30). It is possible that dietary AA does not enter the same pool as AA derived from LA or that 1-series prostaglandins derived from dihomo-gamma-LA are involved, rather than 2- series prostaglandins.

The type of fat may also be important, since fatty acids of the n-3 series always have an inhibitory effect ( 21, 31). Dietary EPA and DHA are of nutritional importance because of their influence on the metabolism of LA and its derivatives. Dietary EPA and DHA readily displace n-6 fatty acids from membrane lipids and inhibit the conversion of LA to AA ( 32). They can also directly influence the extent to which AA is converted to prostaglandins and leukotrienes by acting as competitive inhibitors for cyclooxygenase and lipoxygenase ( 33, 34). We were able to find a consistent tendency for a lower risk of skin SCC with higher intakes of EPA and DHA but not with intake of alpha-linolenic acid. This may be due to the fact that EPA and DHA are 2.5-5.0 times more effective than alpha-linolenic acid in depressing AA content and altering tissue phospholipid fatty acid composition (35, 36).

EPA [20:5(n-3)] and DHA [22:6(n-3)] may exert their effects through competition with AA [20:4(n-6)] for the enzyme cyclooxygenase, thereby curtailing the formation of AA metabolites ( 37). Diets rich in n-3 PUFA have been shown to suppress the excessive production of PGE2, whereas a diet rich in n-6 PUFA causes further increases ( 38). Our data provide supporting evidence for the protective effect of diets rich in n-3 PUFA (EPA + DHA) in relation to risk of skin SCC. We found no relationship between risk of skin SCC and dietary intake of n-6 fatty acids. There was a consistent tendency for a lower risk of skin SCC with higher intake of n-3 fatty acids, with the OR ranging from 0.55 to 0.93 in the various strata. This lowered risk was statistically significant among men (OR = 0.68, 95% CI = 0.47-0.97).

PUFA of the n-6 and the n-3 families are metabolized by the same enzyme systems and, thus, compete with each other. It is possible that there may also be competition between PUFA within the same family, which would mean that the number of competitive interactions could be very high. The serum levels of PGE2 are decreased as dietary n-3/n-6 is increased, again suggesting that the antipromotional effects of n-3 fats may be mediated through inhibition of formation of 2-series prostaglandins ( 39). The structural similarities between AA and EPA most likely explain the enhanced effectiveness of EPA as an antagonist to AA and its metabolism ( 40). For n-3/n-6, our data show that the ORs in successively higher levels were 0.88 (95% CI = 0.59-1.32) and 0.74 (95% CI = 0.51-1.05), suggesting a tendency toward decreased risk of skin SCC with increased intake of diets with high n-3/n-6. Likewise, a higher n-3/n-6 in the diet was associated with lower risk of skin SCC in the various strata, and this lowered risk was statistically significant among men (OR = 0.68, 95% CI = 0.46-0.99).

Another ratio that was higher in the controls than in the cases was the ratio of alpha-linolenic acid [18:3(n-3)] to LA [18:2(n-6)]. Again we found a similar pattern of lowered risk of skin SCC with a higher ratio of 18:3(n-3) to 18:2(n-6). The same enzyme system is responsible for the synthesis of the long-chain PUFA from the dietary LA and alpha-linolenic acid. However, alpha-linolenic acid inhibits the further metabolism of LA ( 41).

In summary, this case-control study, one of the largest to date on fat intake and SCC of the skin, shows an inverse relationship between the risk of skin SCC and the intake n-3 PUFA. The association was strong for the intake of EPA, DHA, and diet with high n-3/n-6. In this study the protection conferred by n-3 PUFA and n-3/n-6 and the increased risk conferred by higher energy intakes from fat seemed slightly stronger among men. More studies are clearly needed to elucidate the function of dietary fatty acids so that recommendations can be made to alter the human diet for cancer prevention, particularly in light of the increasing incidence of SCC of the skin.

Acknowledgments and Notes
The authors thank Dr. David Alberts (Div. of Cancer Prevention and Control) for support, Steve Rodney for assistance with data management, and Mary Lurie for assistance with interviewing and data entry. This work was made possible by National Cancer Institute Grant P01 CA-27502. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the National Cancer Institute. Address reprint requests to Iman Hakim, M.D., Ph.D., M.P.H., Arizona Cancer Center, College of Medicine, 1515 N. Campbell Ave., PO Box 245024, Tucson, AZ 85724. Phone: (520) 626-5355. FAX: (520) 626-5348. E-mail: ihakim@azcc.arizona.edu.

Submitted 16 November 1999; accepted in final form 17 December 1999.

Table 1. Selected Characteristics of Skin SCC Cases and Controls Participating in the Southeastern Arizona Health Study[a]

Legend for Chart:

A - Variable
B - Controls (n = 267): No.
C - Controls (n = 267): %
D - Cases (n = 301): No.
E - Cases (n = 301): %
F - P Value

A B C
D E F

Gender

Female 113 42
118 39 --

Male 154 58
183 61 0.450

Age group

65 22 --

61-70 yr 99 37
115 38 --

>70 yr 96 36
121 40 0.300

Education

21 7 --

High school 73 27
82 27 --

Some college or higher 171 64
198 66 0.90

Smoking

Never 93 35
100 33 --

Former 129 48
162 54 --

Current 45 17
39 13 0.301

Alcohol intake[b]

No 51 19
67 22 --

Yes 216 81
234 78 0.354

Actinic keratosis

No 175 66
77 26 --

Yes 92 34
224 74 0.000

Freckles number[c]

None 16 6
5 2 --

Mild 225 84
237 78 --

Moderate 26 2
59 20 0.000

Use of sunscreen

No 83 31
49 16 --

Yes 184 69
252 84 0.000

Tanning ability[d]

No 1 --
11 4 --

Mild 47 18
83 27 --

Moderate 139 52
165 55 --

Deep 80 30
42 14 0.000

Sun exposure[ef] 1.43 +/- 0.07
1.49 +/- 0.07 0.57
a: Percentages are rounded. SCC, squamous cell carcinoma.

b: Alcohol intake based on four 24-h dietary recalls.

c: Number of current freckles on arms: mild, <50 freckles; moderate, >50 freckles.

d: Ability to tan after prolonged sun exposure.

e: Average hours per day outside during peak sun.

f: Means +/- SD.

Table 2. ORs and 95% CIs of Skin SCC According to Levels of Energy, Cholesterol, and Fat Intake[a]

Legend for Chart:

A - Level of Intake[b]: I
B - Level of Intake[b]: II
C - Level of Intake[b]: III
D - P Value

A B
C D

Energy (total),
kcal/day >2,055.4 --

No. of cases 156 112
33 --
No. of controls 133 107
27 --

OR1 (95% CI) 1.0 0.88 (0.61-1.28)
0.99 (0.73-1.34) 0.80

OR2 (95% CI) 1.0 0.89 (0.59-1.35)
0.90 (0.63-1.29) 0.67

Calories from
total fat,[c] % 20% to <35
>35 --

No. of cases 21 197
83 --

No. of controls 22 180
65 --

OR1 (95% CI) 1.0 1.14 (0.60-2.15)
1.25 (0.87-1.80) 0.27

OR2 (95% CI) 1.0 1.07 (0.51-2.22)
1.34 (0.87-2.06) 0.09

Cholesterol, mg/day >349 --

No. of cases 166 110
25 --

No. of controls 159 81
27 --

OR1 (95% CI) 1.0 1.28 (0.88-1.87)
0.98 (0.70-1.37) 0.68

OR1 (95% CI) 1.0 1.07 (0.69-1.64)
1.01 (0.68-1.49) 1.00

Total fat, g/day >78.7 --

No. of cases 148 121
32 --

No. of controls 133 107
27 --

OR1 (95% CI) 1.0 1.08 (0.70-1.64)
1.03 (0.65-1.63) 0.77

OR2 (95% CI) 1.0 1.01 (0.64-1.64)
1.03 (0.62-1.72) 0.76
Saturated fat, g/day >26.9 --

No. of cases 137 128
36 --

No. of controls 131 108
28 --

OR1 (95% CI) 1.0 1.27 (0.84-1.91)
1.13 (0.73-1.73) 0.33

OR2 (95% CI) 1.0 1.30 (0.82-2.07)
1.18 (0.72-1.92) 0.09

MUFA, g/day >31.2 --

No. of cases 152 117
32 --

No. of controls 133 106
28 --

OR1 (95% CI) 1.0 0.98 (0.64-1.51)
0.94 (0.59-1.51) 0.50

OR2 (95% CI) 1.0 0.98 (0.60-1.59)
0.86 (0.60-1.43) 0.22

PUFA, g/day >15 --

No. of cases 152 108
41 --

No. of controls 131 104
32 --

OR1 (95% CI) 1.0 0.97 (0.63-1.49)
0.87 (0.60-1.27) 0.92

OR2 (95% CI) 1.0 1.06 (0.66-1.73)
0.93 (0.60-1.73) 0.64
a: OR, odds ratio; CI, confidence interval. OR1, OR adjusted for age, gender, and total energy intake; OR2, OR further adjusted for history of diagnosed actinic keratosis, ability to tan after prolonged sun exposure, number of current freckles on arm, and use of sunscreen. MUFA and PUFA, mono- and polyunsaturated fatty acids.

b: Cutoffs for 3 levels of fatty acid intake are based on 50th (median) and 90th percentiles of daily intake in grams among controls (<50th, >50th to <90th, >90th).

c: Categorical variable: cutoff for 3 levels: 20% to 35%.

Table 3. ORs and 95% CIs of Skin SCC According to Level of Dietary Intake of Long-Chain PUFA[a]

Legend for Chart:

A - Fatty Acid
B - Level of Intake[b]: I
C - Level of Intake[b]: II
D - Level of Intake[b]: III
E - P Value

A B C
D E

n-3 fatty acids

Linolenic acid
(C18:3), g/day >1.35 --

No. of eases 143 122
36 --

No. of controls 133 102
32 --

ORI (95% CI) 1.0 1.22 (0.81-1.83)
1.11 (0.76-1.60) 0.47

OR2 (95% CI) 1.0 1.24 (0.78-1.94)
1.01 (0.65-1.54) 0.73

EPA (C20:5), g/day 0 >0 to <0.12
>0.12 --

No. of cases 119 158
24 --

No. of controls 102 137
28 --

OR1 (95% CI) 1.0 1.02 (0.71-1.45)
0.84 (0.62-1.15) 0.45

OR2 (95% CI) 1.0 0.91 (0.60-1.36)
0.78 (0.54-1.10) 0.20

DHA (C22:6), g/day 0.03 to <0.24
>0.24 --

No. of cases 131 145
25 --

No. of controls 121 118
28 --

OR1 (95% CI) 1.0 1.16 (0.81-1.65)
0.91 (0.66-1.23) 0.86

OR1 (95% CI) 1.0 0.99 (0.66-1.48)
0.80 (0.57-1.31) 0.26

n-6 fatty acids

Linoleic acid
(C18:2), g/day >13.7 --

No. of cases 145 117
39 --

No. of controls 130 108
29 --

OR1 (95% CI) 1.0 1.13 (0.75-1.72)
0.95 (0.65-1.39) 0.53

OR2 (95% CI) 1.0 1.17 (0.73-1.87)
1.05 (0.68-1.61) 0.36

Arachidonic acid
(C20:4), g/day >0.20 --

No. of cases 144 137
20 --

No. of controls 128 113
26 --

OR1 (95% CI) 1.0 1.04 (0.73-1.49)
0.83 (0.58-1.20) 0.47

OR2 (95% CI) 1.0 0.86 (0.57-1.29)
0.70 (0.46-1.08) 0.16
a: EPA, eicosapentaenoic acid; DHA, docosaxexaenoic acid; see Table 2 footnote for other abbreviations.

b: Cutoffs for 3 levels of fatty acid intake are based on 50th (median) and 90th percentiles of daily intake in grams among controls (<50th, >50th to <90th, >90th).

Table 4. ORs and 95% CIs of Skin SCC According to Level of Dietary Intake of the Sum of Long-Chain PUFA

Legend for Chart:

A - Level of Intake[a]: I
B - Level of Intake[a]: II
C - Level of Intake[a]: III
D - P Value

A B
C D

Sum of n-3 fatty
acids,[b] g/day 0.06 to <0.45
>0.45 --

No. of eases 155 124
22 --

No. of controls 132 107
28 --

ORI (95% CI) 1.0 1.01 (0.70-1.43)
0.79 (0.58-1.08) 0.29

OR2 (95% CI) 1.0 0.85 (0.56-1.27)
0.71 (0.49-1.00) 0.055

Sum of n--6 fatty
acids, g/day 8.3 to >13.9 --

No. of cases 151 112
38 --

No. of controls 133 106
28 --

OR1 (95% CI) 1.0 1.07 (0.69-1.63)
0.92 (0.61-1.37) 0.66

OR2 (95% CI) 1.0 1.03 (0.69-1.83)
0.99 (0.63-1.56) 0.51

n-3/n-6 0.01 to >0.06 --

No. of cases 141 140
20 --
No. of controls 127 112
28 --

ORI (95% CI) 1.0 1.12 (0.79-1.59)
0.79 (0.57-1.08) 0.51

OR2 (95% CI) 1.0 0.88 (0.59-1.32)
0.74 (0.51-1.05) 0.14

Ratio of linolenic
to linoleic acid 0.10 to >0.13 --

No. of cases 140 121
40 --

No. of controls 121 102
44 --

OR1 (95% CI) 1.0 1.01 (0.70-1.46)
0.86 (0.67-1.11) 0.44

OR2 (95% CI) 1.0 0.98 (0.65-1.46)
0.81 (0.61-1.08) 0.33

PUFA/SFA[c] >0.98 --

No. of cases 152 115
34 --

No. of controls 132 112
23 --

OR1 (95% CI) 1.0 0.89 (0.62-1.26)
1.15 (0.85-1.54) 0.77

OR2 (95% CI) 1.0 0.79 (0.53-1.18)
1.16 (0.82-1.62) 0.98
a: Cutoffs for 3 levels of fatty acid intake are based on 50th (median) and 90th percentiles of daily intake in grams among controls (<50th, >50th to <90th, >90th).

b: Sum of dietary intake of 20:5(n-3) and 22:6(n-3) fatty acids.

c: Ratio of PUFA to saturated fatty acid (SFA).

Table 5. ORs and 95% CIs of Skin SCC According to Dietary Intake of the Long-Chain n-3 and n-6 PUFAs in Strata of Selected Covariates (Arizona 1993-1996)[a]

Legend for Chart:

A - Stratum
B - % kcal from fat[b]
C - summation of n-3[c,d]
D - summation of n-6[c]
E - n-3/n-6[c]

A B C
D E

Gender

Female 1.06 (0.60-1.89) 0.93 (0.54-1.57)
0.77 (0.40-1.48) 1.10 (0.66-1.83)

Male 1.58 (1.01-2.48) 0.68 (0.47-0.97)
1.32 (0.83-2.06) 0.68 (0.46-0.99)

Age group

1.34 (0.59-3.01) 0.96 (0.43-2.13)

61-70 yr 1.46 (0.80-2.62) 0.70 (0.43-1.12)
0.88 (0.47-1.64) 0.73 (0.45-1.17)

>70 yr 1.34 (0.76-2.34) 0.75 (0.47-1.16)
1.24 (0.69-2.23) 0.81 (0.50-1.29)

Energy intake

(median) 1.43 (0.86-2.37) 0.71 (0.45-1.10)
1.14 (0.61-2.12) 0.68 (0.43-1.06)

>1,467 kcal/day 1.26 (0.77-2.05) 0.78 (0.52-1.16)
1.09 (0.74-1.59) 0.92 (0.61-1.37)

Actinic keratosis

No 1.31 (0.77-2.21) 0.67 (0.41-1.09)
0.86 (0.48-1.53) 0.83 (0.52-1.31)

Yes 1.35 (0.82-2.21) 0.75 (0.51-1.10)
1.40 (0.85-2.31) 0.76 (0.51-1.130)

Tanning Ability[e]

No 0.88 (0.37-2.05) 0.56 (0.27-1.16)
0.77 (0.32-1.80) 0.73 (0.36-1.45)

Yes 1.46 (0.99-2.15) 0.74 (0.53-1.02)
1.20 (0.79-1.79) 0.76 (0.54-1.06)
Use of sunscreen

No 2.03 (0.90-4.56) 0.55 (0.26-1.13)
0.91 (0.42-1.95) 0.60 (0.29-1.20)

Yes 1.22 (0.82-1.80) 0.82 (0.59-1.13)
1.21 (0.78-1.86) 0.88 (0.63-1.22)
a: Values are ORs adjusted for age, gender, total energy intake, history of diagnosed actinic keratosis, ability to tan aider prolonged sun exposure, number of current freckles on the arm, and use of sunscreen, with 95% CIs in parentheses.

b: Categorical variable: Cutoff for 3 levels: 20% to 35%.

c: Cutoffs for 3 levels of fatty acid intake are based on 50th (median) and 90th percentiles of daily intake in grams among controls (<50th, >/=50th to <90th, >/=90th).

d: Sum of dietary intake of 20:5(n-3) and 22:6(n-6) fatty acids.

e: Tanning ability after prolonged sun exposure: No, no or mild tanning; Yes, moderate to deep tanning.

References
1. Miller, DL, and Weinstock, MA: Nonmelanoma skin cancer in the United States: incidence. J Am Acad Dermatol 30, 774-778, 1994.

2. Robinson, JK: Risk of developing another basal cell carcinoma. Cancer 60, 118-120, 1987.

3. Black, HS: Influence of dietary factors on actinically induced skin cancer. Mutat Res 422, 185-190, 1998.

4. Black, HS, Herd, JA, Goldberg, LH, Wolf, JE, Jr, Thornby, JI, et al.: Effect of low-fat diet on the incidence of actinic keratosis. N Engl J Med 330, 1272-1275, 1994.

5. Black, HS, Thornby, JI, Wolf, JE, Jr, Goldberg, LH, Herd, JA, et al.: Evidence that a low-fat diet reduces the occurrence of non-melanoma skin cancer. Int J Cancer 62, 165-169, 1995.

6. Jaax, S, Scott, LW, Wolf, JE, Jr, Thomby, JI, and Black, HS: General guidelines for a low-fat diet effective in the management and prevention of nonmelanoma skin cancer. Nutr Cancer 27, 150-156, 1997.

7. Yen, A, Black, HS, and Tschen, J: Effect of dietary omega-3 and omega-6 fatty acid sources on PUVA-induced cutaneous toxicity and tumorigenesis in the hairless mouse. Arch Dermatol Res 286, 331-336, 1994. 8. Potter, JD, Slattery, ML, Bostick, RM, and Gapstur, SM: Colon cancer: a review of the epidemiology. Epidemiol Rev 15, 499-545, 1993.

9. Reddy, BS: Chemoprevention of colon cancer by dietary fatty acids. Cancer Metastasis Rev 13, 285-302, 1994.

10. Parkinson, AJ, Cruz, AL, Heyward, WL, Bulkow, LR, Hall, D, et al.: Elevated concentrations of plasma n-3 polyunsaturated fatty acids among Alaskan Eskimos. Am J Clin Nutr 59, 384-388, 1994.

11. Caygill, CPJ, Charlett, A, and Hill, MJ: Fat, fish, fish oil and cancer. Br J Cancer 74, 159-164, 1996.

12. Kamano, K, Okuyama, H, Konishi, R, and Nagasawa, H: Effect of a high-linoleate and a high-alpha-linolenate diet on spontaneous mammary tumorigenesis in mice. Anticancer Res 9, 1903-1908, 1989.

13. Waksberg, J: Sampling methods for random digit dialing. J Am Stat Assoc 56, 33-43, 1978.

14. Nutrition Data System, version 2.9. University of Minnesota: Nutrition Coordinating Center, 1997.

15. Stata Corporation: Stata Statistical Software, Intercooled Stata, release 5.0. College Station, TX: Stata, 1997.

16. Black, HS: Diet and skin cancer. In Nutritional Oncology, D Heber, G Blackburn, and V Liang (eds). New York: Academic, 1999, pp 405-419.

17. Birt, DE, Kris, ES, Choe, M, and Pelling, JC: Dietary energy and fat effects on tumor promotion. Cancer Res 52 Suppl 7, 2035s-2039s, 1992.

18. Birt, DF, Barnett, T, Pour, PM, and Copenhaver, J: High-fat diet blocks the inhibition of skin carcinogenesis and reductions in protein kinase C by moderate energy restriction. Mol Carcinog 16, 115-120, 1996.

19. Kritchevsky, D, and Klurfeld, DM: Caloric effects in experimental mammary tumorigenesis. Am J Clin Nutr 45, 236--242, 1987.

20. Black, HS, Thornby, JI, Gerguis, J, and Lenger, W: Influence of dietary omega-6, -3 fatty acid sources on the initiation and promotion stages of photocarcinogenesis. Photochem Photobiol 56, 195-199, 1992.

21. Orengo, IF, Black, HS, Kettler, AH, and Wolf, JE, Jr: Influence of dietary menhaden oil upon carcinogenesis and various cutaneous responses to ultraviolet radiation. Photochem Photobiol 49, 71-77, 1989.

22. Ley, RD, and Reeve, VE: Chemoprevention of ultraviolet radiation-induced skin cancer. Environ Health Perspect 105 Suppl 4, 981-984, 1997.

23. Reddy, BS: Dietary fat, calories, and fiber in colon cancer. Prev Med 22, 738-749, 1993.

24. Reddy, BS, and Maruyama, H: Effect of dietary fish oil on azoxy-methane-induced colon carcinogenesis in male F344 rats. Cancer Res 46, 3367-3370, 1986.

26. Rhodes, LE, Durham, BH, Fraser, WD, and Friedmann, PS: Dietary fish oil reduces basal and ultraviolet B-generated PGE2 levels in skin and increases the threshold to provocation of polymorphic light eruption. J Invest Dermatol 105, 532-535, 1995.

27. Van Dorp, G, Volker-Dieben, HJ, Van Leeuwen, A, Kok-van Alphen, CC, and Van Rood, JJ: Major histocompatibility complex or HL-A antibodies in patients with a rejection reaction (abstr). Ophthalmologica 175, 42, 1977.

28. Greaves, MW, and McDonald-Gibson, W: Prostaglandin biosynthesis by human skin and its inhibition by corticosteroids. Br J Pharmacol 46, 172-175, 1972.

29. Whelan, J, Surette, ME, Hardardottir, I, Lu, G-P, Golemboski, KA, et al.: Dietary arachidonate enhances tissue arachidonate levels and eicosanoid production in Syrian hamsters. J Nutr 123, 2174-2185, 1993.

30. Hillyard, LA, and Abraham, S: Effect of dietary polyunsaturated fatty acids on growth of mammary adenocarcinomas in mice and rats. Cancer Res 39, 4430-4437, 1979.

31. Branden, LM, and Carroll, KK: Dietary polyunsaturated fat in relation to mammary carcinogenesis in rats. Lipids 21, 285-286, 1986.

32. Brenner, RR, and Peluffo, RO: Inhibitory effect of docosa-4,7,10,13, 16,19-hexaenoic acid upon the oxidative desaturation of linoleic into gamma-linolenic acid and of gamma-linolenic acid into octadeca-6,9,12,15-tetraenoic acid. Biochim Biophys Acta 137, 184-186, 1967.

33. Lands, WEM, LeTellier, PR, Rome, LH, and Vanderhoek, JY: Inhibition of prostaglandin biosynthesis. Adv Biosci 9, 15-28, 1973.

34. Hamberg, M: Transformation of 5,8,11,14,17-eicosapentaenoic acid in human platelets. Biochim Biophys Acta 618, 389-398, 1980. 35. Hwang, DH, Boudreau, M, and Chanmugam, P: Dietary linolenic acid and longer-chain n-3 fatty acids: comparison of effects on arachidonic acid metabolism in rats. J Nutr 118, 427-437, 1988.

36. Whelan, J, Broughton, KS, and Kinsella, JE: The comparative effects of dietary alpha-linolenic acid and fish oil on 4- and 5-series leukotriene formation in vivo. Lipids 26, 119-126, 1991.

37. Needleman, P, Sprecher, H, Whitaker, HO, and Wyche, A: Mechanism underlying the inhibition of platelet aggregation by eicosapentaenoic acid and its metabolites. Adv Prostaglandin Thromboxane Res 6, 61-68, 1980.

38. Minoura, T, Takata, T, Sakaguchi, M, Takada, H, Yamamura, M, et al.: Effect of dietary eicosapentaenoic acid on azoxymethane-induced colon carcinogenesis in rats. Cancer Res 48, 4790-4794, 1988.

39. O'Connor, TP, Roebuck, BD, Peterson, F, and Campbell, TC: Effect of dietary intake of fish oil and fish protein on the development of 1-azaserine-induced preneoplastic lesions in the rat pancreas. JNC175, 959-962, 1985.

40. Whelan, J: Antagonistic effects of dietary arachidonic acid and n-3 polyunsaturated fatty acids, J Nutr 126, 1086S-1091S, 1996.

41. Mohrhauer, H, and Holman, RT: The effect of dose level of essential fatty acids upon fatty acid composition of the rat liver, J Lipid Res 4, 151-159, 1963.

~~~~~~~~

By Iman A. Hakim; Robin B. Harris and Cheryl Ritenbaugh

I. A. Hakim and R. B. Harris are and C. Ritenbaugh was affiliated with Cancer Prevention and Control, Arizona Cancer Center, College of Medicine, University of Arizona, Tucson, AZ 85724. C. Ritenbaugh is affiliated with the Kaiser Center for Health Research, Portland, OR 97227.

Share this with your friends