Non-alcoholic fatty liver disease (NAFLD) is a spectrum of diseases ranging from simple hepatic steatosis to non-alcoholic steatohepatitis (NASH) [1, 2]. It is the most common form of chronic liver disease in clinical practice. In some patients, steatohepatitis may progress to liver cirrhosis and its complications, including hepatocellular carcinoma . In addition, emerging studies suggested that NAFLD is related to cardiometabolic diseases, including metabolic syndrome , insulin resistance , and atherosclerosis . Currently, excessive hepatic fat accumulation and increased levels of pro-inflammatory cytokines and adipokines are considered to affect these relationships [7, 8].
Carcinoembryonic antigen (CEA) is an onco-fetal glycoprotein with a molecular weight of 180–200kDa . CEA is widely used as a biomarker for tumor detection because it is overexpressed in adenocarcinomas, especially colorectal cancer . However, CEA levels also increase in non-neoplastic conditions, including aging, smoking, hepatic insufficiency, and inflammatory bowel disease . Furthermore, the association of CEA with cardiometabolic diseases, including metabolic syndrome and carotid atherosclerosis, has also been reported [12, 13]. Although the precise role of CEA has not been fully elucidated, pro-inflammatory and pro-thrombic features of CEA, which stimulate macrophages and monocytes, are considered to link CEA to cardiometabolic diseases [14, 15].
As NAFLD is an inflammatory hepatic disease and a known risk factor for cardiometabolic diseases, we hypothesized that CEA may be associated with NAFLD. However, little information is available on the possible relationship between serum CEA levels and NAFLD. Therefore, we investigated the relationship between serum CEA concentrations and the prevalence of NAFLD in healthy Korean non-smokers.
Materials and methods
The study sample consisted of 370 Korean people who visited the Health Promotion Center at Severance Hospital for a routine health check-up between January 2011 and July 2012. All subjects completed a questionnaire about lifestyle factors, including alcohol consumption, cigarette smoking, and exercise. Of the 370 participants, we excluded 56 people who were current or former smokers, as well as 28 people who reported alcohol consumption, which was defined as drinking alcohol more frequently than once per week or drinking 70 g or more alcohol per week during the previous 1 year. We also excluded people with underlying medical conditions, including a history of hypertension, diabetes mellitus, dyslipidemia, chronic liver disease, chronic renal disease, coronary artery occlusive disease, chronic inflammatory disease (e.g., pancreatitis and chronic obstructive pulmonary disease), or cancer. We also excluded people using any medications that could affect cardiometabolic function, including anti-hypertensive medicines, oral hypoglycemic agents, lipid-lowering agents, anti-obesity drugs, and antidepressants. A total of 200 people were included in the final analyses. The study complied with the Declaration of Helsinki; the Institutional Review Board of Yonsei University College of Medicine approved this study.
Anthropometric measurements were taken by a single, well-trained examiner. Blood pressure was measured in the sitting position after a 10-min resting period. Body mass index (BMI) was calculated as weight divided by height squared. Waist circumference was measured at the umbilicus while the subject was standing.
After a 12-h overnight fast, blood samples were collected. We measured fasting glucose, aspartate aminostransferase (AST), alanine aminotransferase (ALT), γ-glutamyltransferase (GGT), total cholesterol, triglyceride, and high-density lipoprotein (HDL)-cholesterol levels using an ADVIA 1650 chemistry system (Siemens Medical Solution, Tarrytown, NY, USA). Low-density lipoprotein (LDL)-cholesterol levels were calculated using the Friedewald equation. Fasting insulin levels were determined by electrochemiluminescence immunoassay using an Elecsys 2010 immunology analyzer (Roche, Indianapolis, IN, USA). Insulin resistance was estimated by the homeostasis model assessment of insulin resistance (HOMA-IR) index: [insulin (pmol/L) × fasting blood glucose (mmol/L)]/22.5. White blood cell (WBC) counts were measured using an automated blood cell counter (ADVIA 120, Bayer, NY, USA).
CEA levels were measured by chemiluminescence immunoassay using a DxI 800 Access Immunoassay System (Beckman Coulter Inc., Brea, CA, USA). Assay precision was evaluated by testing three control concentrations of commercially available quality control materials (MAS T-Marker; Medical Analysis Systems, Camarillo, CA, USA) and pooled human sera for the respective markers. Each sample was assayed twice, in duplicate, with a minimum of 2-h intervals between runs. The inter- and intra-assay coefficients of variations were 3.80±1.96% and 3.01±1.55%, respectively. The cut-off value for an abnormal CEA measurement was set as above 5 μg/L based on previous criteria .
Hepatic steatosis was measured using the Fibroscan®. Examinations were performed by one experienced operator. The patient was placed in the dorsal decubitus position, and the probe of the transducer was placed between the ribs over the right hepatic lobe. The controlled attenuation parameter (CAP) was used to measure hepatic steatosis. The CAP measured liver ultrasonic attenuation at 3.5MHz using signals acquired by the Fibroscan. The principles of measurement of the CAP have been described previously . The median of individual values was used as the final CAP and was expressed in decibels per meter (dB/m). Hepatic steatosis was graded according to the following CAP values: grade 0: <238 dB/m, grade 1: 238–260 dB/m, grade 2: 260–292 dB/m, and grade 3; ≥292 dB/m, based on the results of the previous study .
A normal ALT level was defined as <30 U/L in men and 19 U/L in women . Populations with steatosis on the Fibroscan and normal ALT levels were classified as the ‘simple steatosis group’. Populations with steatosis on the Fibroscan with increased serum ALT levels were classified as the ‘NASH group’ based on existing criteria .
Data are described as the mean±standard deviation for normally distributed data and as the median and interquartile range for non-normally distributed data. CEA, body fat (%), fasting glucose, insulin, HOMA-IR, AST, ALT, GGT, HDL-cholesterol, and triglyceride levels were logarithmically transformed to eliminate the skewness of the distribution. The basic characteristic of the participants according to the CEA quartiles were compared using linear trend analysis. Mean CEA levels between the non-NAFLD and NAFLD groups were compared using the χ2-test. Mean CEA levels were calculated by analysis of covariance (ANCOVA) according to the grade of hepatic steatosis and among the normal, ALT elevation, simple steatosis, and NASH groups. The odds ratio and 95% confidence interval (CI) for NAFLD were calculated after adjusting for confounding variables across CEA quartiles using multivariate logistic regression analysis.
We performed all statistical analyses using the Statistical Package for the Social Sciences, version 18.0 (SPSS Inc., Chicago, IL, USA). Statistical significance was defined as p<0.05.
The clinical characteristics of the study population in relation to the CEA quartiles are shown in Table 1. CEA quartiles were categorized as follows: Q1: <1.12; Q2: 1.12–1.66, Q3: 1.66–2.50, Q4: ≥2.50 μg/L. The mean age, fasting glucose, AST, ALT, and LDL-cholesterol levels were highest in the fourth CEA quartile. The prevalence of NAFLD increased significantly according to the CEA quartile. Figure 1 compares the CEA levels between the non-NAFLD and NAFLD groups. The mean log CEA level was 0.32±0.8 μg/L in the non-NAFLD group and 0.63±0.57 μg/L in the NAFLD group (p<0.01). The median (minimum–maximum) CEA level was 1.50 μg/L (0.42–6.86) in the non-NAFLD group and 1.76 μg/L (0.57–6.61) in the NAFLD group. The percentage of people with CEA values greater than the upper reference limit was 1% in the non-NAFLD group and 2.9% in the NAFLD group. The mean log CEA level increased gradually according to the grade of hepatic steatosis (p<0.01): 0.32±0.8 μg/L, 0.52±0.42, 0.66±0.66, and 0.89±0.67 μg/L for grade 0 (normal), grade 1, grade 2, and grade 3, respectively (Figure 2). The median (minimum–maximum) CEA level was 1.50 (0.42–6.86), 1.72 (0.59–4.08), 1.77 (0.57–4.82), and 2.73 (0.67–6.61) μg/L for grade 0, grade 1, grade 2, and grade 3, respectively. There were no people with CEA levels above the cut-off value in grades 1 and 2. Three individuals (14.3%) of grade 3 had CEA levels above the upper reference limit. When classifying the groups into normal, ALT elevation, simple steatosis, and NASH, the mean log CEA level was 0.28±0.56 μg/L in the normal group, 0.43±0.54 μg/L in the ALT elevation group, 0.53±0.49 μg/L in the simple steatosis group, and 0.71±0.56 μg/L in the NASH group (p<0.05) (Figure 3).
|CEA quartiles||P for trend|
|Male, %||32 (64)||31(62)||23 (46)||34 (68)||NS|
|Body fata, %||27.60 (22.20–31.50)||28.70 (23.70–31.30)||27.70 (22.40–33.80)||28.35 (22.45–34.55)||NS|
|Blood pressure, mm Hg|
|ASTa, U/L||19.00 (16.00–23.00)||20.00 (17.00–25.00)||21.00 (20.00–25.00)||22.00 (18.00–31.25)||<0.03|
|ALTa, U/L||18.50 (16.00–23.00)||19.50 (14.00–30.00)||19.00 (15.00–26.00)||24.00 (15.00–35.00)||0.04|
|GGTa, U/L||27.00 (16.00–37.00)||28.00 (15.00–38.00)||23.50 (15.00–44.00)||26.50 (19.00–55.00)||NS|
|Fasting glucosea, mmol/L||5.16 (4.94–5.44)||5.11 (4.83–5.83)||5.47 (5.11–5.99)||5.63 (5.12–7.05)||<0.01|
|Fasting insulina, pmol/L||54.27 (30.57–67.30)||56.90 (33.65–74.12)||39.68 (25.83–53.10)||50.23(36.52–76.56)||NS|
|HOMA-IRa||1.48 (0.98–2.13)||1.98 (1.06–2.51)||1.21 (0.77–1.75)||1.99 (1.22–3.20)||NS|
|Lipid profile, mmol/L|
|HDL-cholesterola||1.17 (0.98–1.35)||1.14 (0.98–1.30)||1.24 (1.10–1.48)||1.17 (0.95–1.32)||NS|
|Triglyceridea||1.06 (0.85–1.47)||1.32 (0.84–1.79)||1.15 (0.82–1.34)||1.16 (0.84–1.78)||NS|
|Regular exercise, n (%)||25 (51.0)||23 (46.9)||20 (40.8)||20 (40.8)||NS|
|NAFLD, n (%)||18 (36)||26 (52)||27 (54)||34 (68)||0.02|
Table 2 shows the risk of NAFLD according to the CEA quartile. The multivariate-adjusted odds ratio (95% CI) for the highest versus the lowest CEA quartiles was 2.98 (1.10–8.05) after adjusting for age, gender, BMI, exercise, systolic blood pressure, diastolic blood pressure, fasting glucose, fasting insulin, total cholesterol, triglycerides, HDL-cholesterol, LDL-cholesterol, WBC counts, AST, ALT, and GGT. These positive associations persisted even after excluding those who had a CEA level over the reference range.
|Model||CEA quartile, OR (95% CI)|
|Model 1a||1.00||2.02 (0.87–4.71)||2.76 (1.15–6.62)||3.37 (1.39–8.16)|
|Model 2b||1.00||1.89 (0.80–4.43)||2.62 (1.08–6.36)||3.14 (1.28–7.69)|
|Model 3c||1.00||1.78 (0.70–4.53)||2.15 (0.80–5.77)||2.98 (1.10–8.05)|
|Model 4d||1.00||1.80 (0.71–4.55)||2.09 (0.78–5.60)||3.02 (1.10–8.28)|
Our cross-sectional study showed positive associations between CEA and NAFLD by analyzing data from 200 healthy Korean non-smokers who underwent general health screening. The positive and graded associations were found between serum CEA and NAFLD after adjusting for other confounding factors.
CEA is widely used as to monitor disease recurrence and therapeutic efficacy in colon cancer . However, mild elevation of CEA levels has been reported in many non-neoplastic conditions, including hepatic insufficiencies, such as liver cirrhosis and acute hepatitis . Furthermore, the relationship between CEA and cardiometabolic diseases has also been reported [12, 13]. Although NAFLD is the most common chronic hepatic disease and is a well-known risk factor for cardiometabolic disease and malignancy, the relationship between CEA and NAFLD is not well understood. To our knowledge, this is the first study to evaluate the relationship between serum CEA levels and the prevalence of NAFLD.
The precise underlying mechanisms that explain the relationship between serum CEA concentrations and NAFLD remain unclear. However, there are several possible mechanisms.
First, pro-inflammatory cytokines may link CEA and NAFLD. NAFLD is known to induce chronic hepatic inflammation by cytokine production and lipid peroxidation of hepatocytes . This chronic inflammation is considered to mediate the association between NAFLD and cardiometabolic diseases . Furthermore, CEA is also known to be associated with chronic inflammation. Elevated CEA levels were reported in many chronic inflammatory conditions, and CEA promotes the production of pro-inflammatory cytokines by stimulating hepatic macrophages and Kupffer cells . Although it is impossible to find a causal link between CEA and NAFLD in our study, it is possible to hypothesize that chronic inflammation may mediate the relationship between CEA and NAFLD.
Second, decreased hepatic clearance of CEA is another possible mechanism. The liver is known to play an important role in the clearance of CEA. Through the portal circulation, CEA is endocytosed into hepatocytes and is degraded by lysosomal enzymes . Hence, the lower clearance of CEA due to hepatic insufficiency may cause increased CEA levels in NAFLD.
Third, the direct secretion of CEA from fat stores in hepatocytes should be considered. Although CEA is mainly secreted by epithelial cells, the types of cells associated with the secretion of CEA have not been fully elucidated. Furthermore, our previous study reported a positive relationship between CEA and visceral fat accumulation . Therefore, we hypothesize that alterations of the cellular environment in hepatic fat stores and cellular regeneration may induce the direct expression of CEA in hepatocytes. Basic experimental studies should be performed to determine these precise mechanisms.
This study has several limitations. We cannot establish a causal relationship between CEA levels and NAFLD from this cross-sectional design. In addition, we enrolled a small number of participants who visited a single hospital for a general health check-up. Therefore, our results do not allow for a generalization of the data to the population at large. Third, although we adjusted for the metabolic parameters that may influence serum CEA levels and NAFLD, we cannot exclude the possibility that the relationship between CEA and NAFLD could be mediated by unknown factors. Finally, despite liver biopsy as the gold standard for diagnosis of fatty liver, we assessed hepatic steatosis using the CAP values measured by the liver Fibroscan. However, the CAP has been well correlated with the results of liver biopsy in previous studies [24, 25], and it is also an operator-independent and sensitive test for the detection of minimal degrees of steatosis in contrast with ultrasonography .
In conclusion, serum CEA levels were independently associated with the prevalence of NAFLD in healthy Korean non-smokers. Although it is impossible to determine causality based on this study, our findings collectively suggest that CEA may function in the pathophysiology of fatty liver disease as well as play a conventional role as a biomarker. In addition, CEA could be a possible candidate biomarker for NAFLD in addition with the known biomarker for NAFLD, CK-18. Additionally, CEA may be useful to distinguish simple steatosis from NASH. Further studies are required to understand the clinical and pathophysiological significance of our findings.
This study was partially supported by a faculty research grant from Yonsei University College of Medicine for 2011 (6-2011-0137).
Conflict of interest statement
Authors’ conflict of interest disclosure: The authors stated that there are no conflicts of interest regarding the publication of this article.
Research funding: None declared.
Employment or leadership: None declared.
Honorarium: None declared.
Hamaguchi M, Kojima T, Takeda N, Nakagawa T, Taniguchi H, Fujii K, et al. The metabolic syndrome as a predictor of nonalcoholic fatty liver disease. Ann Intern Med 2005;143:722–8.
Clark JM, Brancati FL, Diehl AM. Nonalcoholic fatty liver disease. Gastroenterology 2002;122:1649–57.
Matteoni CA, Younossi ZM, Gramlich T, Boparai N, Liu YC, McCullough AJ. Nonalcoholic fatty liver disease: a spectrum of clinical and pathological severity. Gastroenterology 1999;116:1413–9.
Targher G, Day CP, Bonora E. Risk of cardiovascular disease in patients with nonalcoholic fatty liver disease. N Engl J Med 2010;363:1341–50. [Web of Science]
Wieckowska A, Papouchado BG, Li ZZ, Lopez R, Zein NN, Feldstein AE. Increased hepatic and circulating interleukin-6 levels in human nonalcoholic steatohepatitis. Am J Gastroenterol 2008;103:1372–9.
Leclercq IA, Field J, Farrell GC. Leptin-specific mechanisms for impaired liver regeneration ob/ob mice after toxic injury. Gastroenterology 2003;124:1451–64.
Gold P, Freedman SO. Demonstration of tumor-specific antigens in human colonic carcinomata by immunological tolerance and absorption techniques. J Exp Med 1965;121:439–62.
Ruibal MA. CEA serum levels in non-neoplastic disease. Int J Biol Markers 1992;7:160.
Ishizaka N, Ishizaka Y, Toda EI, Koike K, Yamakado M, Nagai R. Are serum carcinoembryonic antigen levels associated with carotid atherosclerosis in Japanese men? Arterioscler Thromb Vasc Biol 2008;28:160–5. [PubMed] [Web of Science]
Lee JW, Park KD, Im JA, Hwang HJ, Kim SH. Serum carcinoembryonic antigen is associated with metabolic syndrome in female Korean non-smokers. Clin Chim Acta 2011;412: 527–30.
Ganguly A, Yeltsin E, Robbins J. Identification of a carcinoembryonic antigen binding protein on monocytes. Biochem Biophys Res Commun 2003;311:319–23.
Aarons CB, Bajenova O, Andrews C, Heydrick S, Bushell KN, Reed KL, et al. Carcinoembryonic antigen-stimulated THP-1 macrophages activate endothelial cells and increase cell–cell adhesion of colorectal cancer cells. Clin Exp Metastasis 2007;24:201–9. [Web of Science] [CrossRef]
Compton C, Fenoglio‐Preiser CM, Pettigrew N, Fielding LP. American Joint Committee on Cancer prognostic factors consensus conference. Cancer 2000;88:1739–57.
Sasso M, Beaugrand M, De Ledinghen V, Douvin C, Marcellin P, Poupon R, et al. Controlled attenuation parameter (CAP): a novel VCTE™ guided ultrasonic attenuation measurement for the evaluation of hepatic steatosis: preliminary study and validation in a cohort of patients with chronic liver disease from various causes. Ultrasound Med Biol 2010;36:1825–35. [CrossRef] [Web of Science]
Prati D, Taioli E, Zanella A, Della Torre E, Butelli S, Del Vecchio E, et al. Updated definitions of healthy ranges for serum alanine aminotransferase levels. Ann Intern Med 2002;137:1–10.
Clark JM, Diehl AM. Defining nonalcoholic fatty liver disease: implications for epidemiologic studies. Gastroenterology 2003;124:248–50.
Jessup J, Giavazzi R, Campbell D, Cleary K, Morikawa K, Fidler I. Growth potential of human colorectal carcinomas in nude mice: association with the preoperative serum concentration of carcinoembryonic antigen in patients. Cancer Res 1988;48:1689. [PubMed]
Loewenstein M, Zamcheck N. Carcinoembryonic antigen (CEA) levels in benign gastrointestinal disease states. Cancer 1978;42:1412–8.
Park SH, Kim BI, Yun JW, Kim JW, Park DI, Cho YK, et al. Insulin resistance and C‐reactive protein as independent risk factors for non‐alcoholic fatty liver disease in non‐obese Asian men. J Gastroenterol Hepatol 2004;19:694–8. [PubMed] [CrossRef]
Lédinghen V, Vergniol J, Foucher J, Merrouche W, Bail B. Non‐invasive diagnosis of liver steatosis using controlled attenuation parameter (CAP) and transient elastography. Liver Int 2012;32:911–8. [PubMed] [Web of Science] [CrossRef]