Human Carcinogenesis: The Role of Age and Gender

Discussion

It is suspected that females are healthier than males because at any given chronological age, their tissues, except for breast and thyroid, are physiologically younger than male tissues. Four differences in cancer risk warrant discussion: i) between genders, ii) among cancer types, iii) among populations, and iv) across ages.

Between genders. For non-HL, myeloma, and cancer of the mouth, esophagus, stomach, colon, liver, pancreas, bronchus, kidney, and larynx, the ASR was greater for males than females in every population studied. With minor exceptions, the same conclusion applies to HL and cancer of the brain and small intestine. The minor exceptions were as follows: For China Shanghai, the ASR for cancer of the small intestine was the same for both genders. In China Shanghai and China Hangzhou, the ASR for cancer of the brain was greater for females. In Croatia and Greater Poland, the ASR for HL was greater for females, while in Poland Lower Silesia and Ukraine, it was the same in both genders.

For cancer of the breast and thyroid, the ASR was greater for females than males in every population studied. This pattern of greater risk in males than females except for breast and thyroid cancer cannot be attributed to chance or the lifestyle/environment because neither would yield such global consistency. The gender pattern of risk must be attributed to genes that are common in populations across the globe.

Melanoma is conspicuously different, displaying a more varied pattern of gender risk. While most populations exhibit greater melanoma risk for males than females, the exceptions are numerous and notable: Melanoma risk was found to be the same for both genders in China Beijing, China Hong Kong, Japan Osaka, Japan Aichi, and Korea Seoul, and greater for females in Denmark, Norway, Ireland, UK, UK England, UK Scotland, Germany Lower Saxony, Spain Basque Country, and Ukraine. These exceptions suggest that for melanoma, population-specific environmental, lifestyle or genetic factors often overwhelm the cause of the normal gender pattern of risk.

In every population studied, LE was longer, and AMRs, GNI, suicide rates, and age-standardized smoking rates were lower in females than males. Of the 24 nations reporting gender-specific IMR and U5MR, all but China reported lower rates for females. There is a suspicion that the same would be true in China if equal care were given to babies of both genders (20-21). Females eat less meat and drink less alcohol than males (22-24). Plasma testosterone is lower in females than males (25, 26). From birth onward, females are healthier than males (27). Any or all of these differences might contribute to the differences in cancer risk between the genders. But height is, arguably, the most conspicuous difference between the genders, and cancer risk is positively correlated with height (28). Unfortunately, reliable data on height in populations across the world are lacking.

The ratio for female/male height in centimeters in the United States is non-Hispanic Black=0.927, non-Hispanic Asian=0.922, non-Hispanic White=0.921, and Hispanic=0.920 (29). Height may be a surrogate for tissue size, which may be a surrogate for cancer target size.

Cancer types. Cancer of the bronchus and colon is common, while those of the mouth and small intestine are rare, in both genders in all populations studied (Tables I and II). This is an interesting pattern because essentially every carcinogen that enters the bronchus passes first through the mouth and every carcinogen that enters the colon passes first through the small intestine. What can explain the disparity in risk of cancer between these neighboring tissues? It cannot be chance because the disparity is large and consistent. It cannot be population-specific environment, lifestyle or genes because the disparity is invariant across all populations studied. The disparity must be due to genes that are common across our species but differ in expression between tissues of high and low cancer risk. It is these genes that are the principal cause of cancer.

In both genders in all populations studied, non-HL as well as cancer of the pancreas, brain, and kidney are less common than cancer of the bronchus and colon, and more common than cancer of the mouth and small intestine, again suggesting that relative cancer risk among tissues is caused by genes that are common across the world but vary in expression among tissues of tumor origin.

Melanoma, again, demonstrates an unusual pattern of relative risk. In both genders in the Asian populations and U.S. Blacks, melanoma is less common than cancer of the mouth and small intestine, and in Australia and New Zealand, more common than cancer of the bronchus and colon. Some other cancer types also demonstrate unusual patterns of relative risk. Stomach cancer is more common than cancer of the bronchus and colon in Japan and Korea, but less common than those types of cancer everywhere else. HL is less common than cancer of the mouth and small intestine in Japan and Korea, and more common than these everywhere else. Thyroid cancer is conspicuously common in Korea, as is myeloma in U.S. Blacks. These unusual patterns suggest that population-specific factors can overwhelm the common cause of relative cancer risk among tissues. But unusual patterns are unusual because in the vast majority of cases there is no evidence for population-specific carcinogenesis.

In all populations, male breast cancer is rare. In most populations, male breast cancer is rarer than mouth and small intestine cancer. In all populations, female breast cancer is more common than cancer of the bronchus and colon. Size is one obvious explanation, as larger breasts carry higher cancer risk (30). Patients with Klinefelter syndrome have larger breasts and higher breast cancer risk than normal males, but smaller breasts and lower breast cancer risk than normal females (31). Breast cancer risk among populations may correlate with breast size. Unfortunately, reliable data on breast size across populations are lacking.

Thyroid cancer, like breast cancer, was found to be more common in females than males in all populations studied. Again, size is a possible explanation, as the female thyroid grows with the menstrual cycle (32). At the age when risk of female thyroid cancer begins to increase, i.e. 15 years (33), thyroid volume is larger in females than males (34).

Of the cancer types studied, laryngeal cancer is rarest in females, and the ratio of female/male ASR smallest. Again, size is an obvious explanation as laryngeal length and volume are larger for males than females (35).

Except for cancer of the bronchus, esophagus, and breast, Var_med for the ratio of female/male ASR was substantially less than for the individual gender rates (Tables I, II and III), suggesting that population-specific influence on ASR is the same for both genders except for this group of cancer types. Correlation coefficients for ASR values between genders across the populations studied confirm this suggestion, i.e. they were high except for cancer of the bronchus, esophagus, and breast (Table IV). Ratios of female/male ASR values for breast cancer are unreliable because male breast cancer rates are so low as to be heavily affected by noise. Risks of bronchial and esophageal cancer are, undoubtedly, related to smoking, and smoking rates vary between genders and among populations (5).

The weak or moderate correlation coefficients suggest a unique cause for each cancer type. The correlation between stomach and liver cancer, and the inverse correlation between stomach cancer and HL may be coincident of Asian culture. That for female breast cancer and non-HL, and for myeloma with cancer of the small intestine are interesting. But even a correlation coefficient of 0.8 can only explain 64% of the relationship between variables. The conspicuous conclusion is that the vast majority of cancer types do not appear to share a common cause. Except for a slight indication that smoking is related to cancer of the pancreas, bronchus, and larynx, the conspicuous conclusion is that, in the vast majority of cases, ratios of female/male risk do not share a common cause because the correlation coefficients are consistently low or moderate at best.

Population. For each population, the female/male ratio of ASR values was listed in order of cancer: Mouth, esophagus, stomach, colon, liver, pancreas, breast, bronchus, melanoma, kidney, brain, thyroid, myeloma, larynx, HL, non-HL, and small intestine. Correlation coefficients were calculated for the 49 populations in all pairwise combinations. Of these 1152 correlation coefficients, the lowest was 0.950. From this it could be concluded that the reason for the differences in ratio of gender risk among the cancer types studied is the same in all populations. This reason must be intrinsic to the tissues of tumor origin and not to anything population-specific.

Age. For each cancer type, the difference in ASR between populations measures risk from causes other than aging. For each cancer, the difference in age-specific rates between ages 70 and 30 years measures risk from aging. For melanoma and cancer of the bronchus, colon, rectum, stomach, pancreas, kidney, esophagus, larynx, liver, mouth, small intestine, and breast in populations across the world, the maximum difference in ASR between populations is less than the 90th percentile difference between age-specific rates at ages 70 and 30 years (3). By this measure, aging is more important than the collection of non-aging processes.

For melanoma, and cancer of the bronchus, colon, rectum, stomach, pancreatic, kidney, esophageal, laryngeal, liver, mouth, small intestine, and breast in populations across the world, differences were calculated in age-specific incidence rates between ages 40 and 30, 50 and 40, 60 and 50, 70 and 60, and 80 and 70 years. The difference between ages 50 and 40 years correlated best (r>0.88) with ASR (3). This high correlation coefficient suggests that the transition between ages 40 and 50 from minimal to increasing risk can explain most of the ASR. The differences in age-specific rates between ages 50 and 40 years also correlated, with coefficients greater than 0.88, with the differences in ASR between genders, races, and ethnicities (3). The magnitude of the initial acceleration of risk predicts total risk. This suggests that the mechanism of initial acceleration of risk is the principal cause of cancer, and the differences in cancer risk between tissues, genders, races, and ethnicities.

Why is cancer more common in one tissue than another, or in one gender or race or ethnicity than another? Because the difference in age-specific rates between ages 50 and 40 years is greater for the tissue, gender, race, or ethnicity where cancer is more common than in the tissue, gender, race, or ethnicity where cancer is less common. What happens between ages 40 and 50 years? Perhaps the cancer target is created, and the size of this target determines the size of the risk. Height is correlated with cancer risk (30). Perhaps height is a surrogate for target size. The incidence of cancer may be higher in men than in women because, in part, they harbor larger targets. The larynx is a good example. The breast and thyroid are exceptions, because they are larger in women than men.

Theory. For all cancer types that originate after infancy, risk changes in three phases: Early phase of low risk at young age, followed by a phase of increasing risk with increasing age, followed by a phase of constant or declining risk with increasing age. The phases might be attributed to species-specific genes that control telomere dynamics in stem cells that obey a fixed strategy at rates that vary among tissues and between genders, races, and ethnicities, and that can fluctuate in response to environmental and lifestyle factors (36). In the early low-risk phase, tumor-suppressor genes and proto-oncogenes are protected from mutation, perhaps by the absence of stem cell renewal. In the subsequent phase of growing risk, they become vulnerable to mutation, perhaps due to stem cell renewal. The rate of renewal varies among tissues and between genders, races, and ethnicities, and in response to environmental and lifestyle agents. For any given tissue type, the larger the size, the greater the chance of a carcinogenic mutation. Beyond some tissue-specific age, senescence diminishes renewal and this causes risk to plateau or decline.