Fatty Fish Consumption Lowers the Risk of Endometrial Cancer: A Nationwide Case-Control Study in Sweden¹

A recent cross-sectional study within the European Prospective Investigation into Cancer and Nutrition (EPIC) cohort in 16 regions of Europe showed significantly increased (3- to 4-fold) plasma levels of EPA,³ an omega-3 fatty acid, in people from Sweden and Denmark with high consumption levels of fatty fish (1). EPA competes with arachidonic acid as substrate for cyclooxygenases, and, therefore, high concentrations of EPA can lead to important changes in relative concentrations of tumor growth-enhancing prostaglandins (2, 3). In Sweden, for example, a recent cohort study of 6272 men found that high fish consumption was associated with a significant reduction in prostate cancer incidence and mortality (4).

Although omega-3 fatty acids contained in fatty fish have been shown to retard the growth of hormone-responsive tumors in vitro and in animal experiments (2, 3), evidence from studies of hormone-responsive cancers conducted among human populations remains sparse. Regarding endometrial cancer, five case-control studies (5, 6, 7, 8, 9) and one prospective cohort study (10) of total fish consumption have found mixed results. However, none of these studies have specifically examined the consumption of fatty fish or omega-3 fatty acids in relation to risk. Null results are not surprising in populations with a low range of total fish consumption, especially if the percentage of fatty fish is also low. Therefore, we examined the association between the consumption of fatty fish, lean fish, and endometrial cancer in a nationwide case-control study of women in Sweden, a country with a relatively wide range of fatty fish consumption (11).

The methods of this population-based case-control study have been described elsewhere (12). In brief, the study was conducted among postmenopausal women ages 50–74 years, born in Sweden and residing in Sweden from 1994 through 1995. Eligible as case patients (n = 1055) were women with an incident, primary, endometrial cancer, identified through six regional cancer registries in Sweden. Control women (n = 4216) were randomly selected from a continuously updated population register. Participation rates were 75% among case patients (789 of 1055 eligible) and 79.9% among the control women (3368 of 4216 eligible). Histological specimens for case women were reviewed by the study pathologist (Anders Lindgren) who, blinded to any characteristics of the subjects, reclassified 80 cases as endometrial atypical hyperplasia, which we excluded from the analysis.

Data were obtained through mailed questionnaires. Participants were asked how often per day, per week, or per month, on average, 1 year prior to the interview, they had consumed different types of fish and 32 other foods, using nine predefined frequency categories that ranged from “never/seldom” to “three or more times per day.” The questionnaire also covered hormone replacement therapy (12), reproductive and medical histories, anthropometrical measures, and lifestyle factors such as physical activity, smoking, and multivitamin use.

Among control subjects, 480 (14.3%) of 3368 failed to return the mailed questionnaire but agreed to a telephone interview including most questionnaire items (except family history of cancer, diet, alcohol consumption, and medical history). These controls were excluded from the present analysis. An additional 36 controls (1.2%), and 9 cases (1.3%), failed to respond to questions regarding fish consumption. This left 698 cases and 2720 controls available for analysis. ORs were computed from unconditional logistic regression models fit by maximum likelihood methods. For tests of trend, median values of each category of categorized variables were fitted in risk models as successive integers (13). Multivariate models adjusted for age (as a continuous variable), body mass index (in quartiles defined among controls), smoking (ever >1 year, never), leisure time physical activity 1 year prior to the interview (never, <1 h/week, 1–2 h/week, >2 h/week), consumption of alcohol (in quartiles), multivitamin use (never, sometimes, frequent), and history of diabetes (yes or no).

Baseline characteristics of the study subjects are shown in Table 1. Cases consumed less fatty fish than controls, although consumption of other fish did not vary according to case-control status; were less likely to smoke than controls; had a lower mean number of children; and had a lower prevalence of oral contraceptive use. Cases also had a higher median body mass index, a higher prevalence of diabetes, and a higher prevalence of hormone replacement use.

Fatty fish consumption was inversely associated with endometrial cancer in models that adjusted estimates for age and for the potentially confounding factors listed in Table 2. The OR of invasive endometrial cancer for women in the highest quartile level of fatty fish (median, 2.0 servings per week) compared to women with in the lowest (median, 0.2 servings per week) was 0.6 (95% CI, 0.5–0.8; P for trend, 0.0002). The corresponding OR for women in the highest quartile level of lean fish (median, 2.5 servings per week) compared with women in the lowest (median, 0.6 servings per week) was 1.0 (95% CI, 0.8–1.3; P for trend, 0.72). Total fish consumption was only weakly associated with endometrial cancer. The corresponding OR for women in the highest quartile level of total fish compared with women in the lowest was 0.8 (95% CI, 0.6–1.0; P for trend, 0.05). Age-adjusted results were similar to those with multivariate-adjustment, which were essentially unchanged after additional adjustment for height (in quintiles defined among controls), leisure time physical activity between ages 18 and 30 (never, <1 h/week, 1–2 h/week, >2 h/week), consumption of high fiber grains, dairy products, fruit, brassica and non-brassica vegetables (in quintiles), parity (as a continuous variable), duration of hormone replacement therapy (as a continuous variable), age at menarche (<12 years, 12–14 years, >14 years), age at menopause and age at first birth (in quartiles), history of hypertension (yes, no), and quintiles of total food consumption (estimated by totaling all of the food items in the food frequency questionnaire).

We found that consumption of fatty fish was inversely associated with endometrial cancer risk. In contrast, we observed no association when we modeled the consumption of lean fish. Even in this population with a relatively high consumption of fatty fish, total fish consumption was only weakly associated with endometrial cancer risk, which underscores the need for separate analyses according to the type of fish consumed.

Total fish consumption was inversely associated with endometrial cancer in two previous case-control studies (5, 9). The specific type of fish examined in these studies was not considered, however, which makes comparisons with our findings difficult. Two previous case-control studies found no association with total fish consumption (6, 7), whereas a case-control study (8) and a prospective cohort study (10) found that high total fish consumption increased the risk. However, the prospective cohort study excluded canned tuna from their total fish measure, which might have been one of the most important sources of omega-3 fatty acids in the population of women from the Midwestern United States.

Our data were limited by the likelihood of nondifferential measurement error with respect to fish consumption. Nondifferential misclassification would tend to attenuate true associations and can stem from measurement error in the assessment of diet because of inaccurate recall of past diet (14), or because of changes in diet over time (14). Finally, although we adjusted our estimates for a wide range of potentially confounding variables, uncontrolled confounding from dietary (or other) factors cannot be excluded.

We conclude that high fatty fish consumption may reduce endometrial cancer risk. Fatty fish are a good source of omega-3 fatty acids, which, possibly acting through inhibition of arachidonic acid-derived eicosanoid biosynthesis, may reduce concentrations of tumor growth-enhancing prostaglandins (2, 3). Given the scarcity of studies of fish consumption and endometrial cancer risk, additional investigations are warranted.