Carcinoembryonic antigen (CEA), an oncofetal glycoprotein, is a widely used tumor marker due to its high expression in adenocarcinomas, particularly colorectal cancer. However, it is also expressed in normal mucosal cells¹⁾ and CEA elevation is associated with several non-neoplastic conditions, including chronic renal failure, hypothyroidism, chronic obstructive pulmonary disease, obesity, aging, and cigarette smoking.^2-6)

Recently, several studies have reported that the serum level of CEA was independently associated with metabolic syndrome, atherosclerosis, and insulin resistance.^7-10) Although the function of CEA has not been fully clarified, previous studies have reported that CEA stimulates monocytes and macrophages to trigger the production of pro-inflammatory cytokines and endothelial adhesion molecules.^11,12) Also, these pro-inflammatory features of CEA suggest that CEA could be related with insulin resistance or metabolic syndrome because insulin resistance and low-grade inflammation have been considered important pathogenic components of metabolic syndrome.¹³⁾

Because serum CEA levels are often slightly elevated in healthy cigarette smokers, only some information is available on the possible association between serum CEA levels and insulin resistance or low-grade inflammation in non-smokers including males. The aim of this study was to investigate whether the insulin resistance or low-grade inflammation was associated with increased serum levels of CEA in healthy non-smoking males and females.

The mean ages of the 543 males (22%) and 1,877 females (78%) were 46.8 ± 11.0 and 47.0 ± 11.5 years, respectively. Thirty-two percent of female subjects were menopausal. There was no difference in the prevalence of metabolic syndrome according to gender. The mean CEA level was 1.18 ± 0.85 ng/dL in males and 0.93 ± 0.73 ng/dL in females (P < 0.001). The mean CEA levels of females with and without metabolic syndrome were 0.91 ± 0.73 ng/dL and 1.07 ± 0.70 ng/dL, respectively (P < 0.05). This difference was significant only in females (data not shown). Fasting glucose levels and hsCRP levels were significantly higher in females. On the other hand, there were no differences in HbA1C. General characteristics of the study subjects are presented in Table 1. Figure 1 shows the mean levels of CEA according to HbA1C in each gender. The mean levels of CEA levels were significantly higher in the highest HbA1C group (P < 0.005) (Figure 1).

After adjustment for age, CEA level was positively correlated with fasting glucose, HbA1C, HDL cholesterol, eGFR in males. In females, CEA level was positively correlated with fasting glucose, HbA1C, HDL cholesterol, aspartate aminotransferase, and hsCRP (Table 2).

Table 3 shows that HbA1C had the strongest influence on CEA levels when age, waist and all other variables were included in the regression models in each gender. Also, in males, eGFR and age were associated with CEA levels (P < 0.05). On the other hand, age, HDL cholesterol, hsCRP, and exercise were associated with CEA levels in females (P < 0.05). Waist, blood pressure, triglyceride levels, and HOMA-IR were not associated with CEA levels in either gender. All together, these variables explain up to 3.3% and 8.3% of the variation in the CEA levels in males and females, respectively (Table 3).

Although of little use in detecting early colorectal cancer, CEA is a useful indicator for surveillance of patients with diagnosed colorectal cancer. Slightly elevated CEA levels are associated with several conditions,^2-6) but very little information exists about the significance of mildly elevated CEA levels over an extended period in individuals who do not have cancer. Recently, some epidemiological studies have demonstrated a possible association between increased CEA levels and metabolic syndrome or atherosclerosis.^7-11) Lee et al.⁷⁾ reported that CEA level was independently associated with metabolic syndrome after adjustment for all covariates in Korean female non-smokers. For the same reason, Kim et al.⁸⁾ suggested that metabolic syndrome may lead to a misunderstanding of CEA levels. Ishizaka et al.⁹⁾ reported that CEA was associated with carotid atherosclerosis in Japanese males without serious health problems. The increase in serum CEA was found to be associated with an increased prevalence of carotid plaque independent of blood pressure, fasting glucose, serum lipids, and inflammatory markers. Bae et al.¹⁰⁾ reported that CEA level was associated with arterial stiffness in healthy Korean males and females after adjustment for all covariates. These data indicate the plausibility that serum CEA may be associated with increased insulin resistance and low-grade inflammation.

Previous studies have suggested that CEA may stimulate macrophages or monocytes via binding to the CEA receptor which results in the production of pro-inflammatory cytokines such as tumor necrosis factor-α, interleukin-1b and -6, reactive oxygen species, eicosanoids, nitric oxide, and carbon monoxide, and up-regulation of endothelial adhesion molecules, which may play a role in metabolic disturbances and cardiovascular disease in addition to cancer.^11,12) Low-grade inflammation is now recognized as the central player in the pathogenesis of atherosclerosis, insulin resistance, and glucose intolerance.^16,17) Also serum CEA levels were increased in patients with chronic inflammatory disorders such as inflammatory bowel disease, collagen disease, and chronic viral hepatitis.^2,18)

In this study, we investigated the factors associated with CEA levels and whether insulin resistance or low-grade inflammation was associated with high CEA levels in healthy non-smokers including non-smoking males. Although HbA1C and age were shared determinants of CEA levels in both genders, the other factors associated with CEA levels were different in each gender. In male subjects, eGFR was associated with CEA levels, but CEA levels were associated with HDL cholesterol, hsCRP, and exercise status in females. Contrary to expectations, the HOMA-IR was not associated with CEA levels in each gender. This finding suggests that the factor affecting CEA levels is glucose levels rather than insulin resistance. On the other hand, hsCRP is associated with CEA levels in females, but not in males. Shio et al.¹⁹⁾ reported that the serum CEA level was significantly associated with fasting glucose levels in diabetic patients. Another study had revealed that serum levels of CEA and carbohydrate antigen 19-9 (CA19-9), a tumor marker of pancreatic cancer, could be influenced by high glucose levels and HbA1C in diabetic patients.²⁰⁾ These results may be due to similar components such as glucose, glycoprotein, and lipid of CEA, CA19-9, and HbA1C.²¹⁾

Several variables including HbA1C were associated with CEA levels, and these variables explain up to 3.3% and 8.3% of the variation in the CEA levels in male and female gender, respectively. This finding suggests that many other factors seem to have a stronger influence on CEA levels.

This study has a number of strengths. First, this study subjects include a large number of non-smoking males, though a fewer number than males. Second, there are several factors that influence serum CEA levels and its metabolism. For this reason, we excluded all subjects that had any smoking history including ex-smokers and abnormal liver or kidney function test results, in order to eliminate the influence of these conditions. Third, we also excluded all subjects that had any diabetic medications which may influence HOMA-IR in the identification of the association between insulin resistance and serum CEA levels. For the same reason, we exclude all subjects whose levels of hsCRP > 1 mg/dL, as this may indicate active inflammatory or infectious conditions.

At the same time, this study has some limitations. First, because the study was cross-sectional, we cannot verify the causality between the each variable and the elevation of serum CEA. Second, we cannot rule out selection bias in our subjects, who had visited a health promotion center. Therefore, these results cannot be generalized to other population groups. Third, serum CEA is known to be elevated in other nonmalignant conditions, such as hypothyroidism and end-stage lung diseases. We did not include these disorders as confounding variables, because their prevalence is considered to be very low among study populations.

In conclusion, within the normal range, serum CEA levels were significantly associated with HbA1C levels but not with HOMA-IR in the non-smoking population. These findings suggest that serum CEA levels are influenced by the glucose level itself rather than insulin resistance.