Journal of Integrative Nephrology and Andrology

: 2015  |  Volume : 2  |  Issue : 1  |  Page : 23--28

Impact of hemodialysis and uremia on serum levels of tumor markers

Xiaoguang Zhang, Niansong Wang, Qin Xue, Guihua Jian, Xuping Gao 
 Department of Nephrology, Shanghai Jiao Tong University Affiliated Sixth People's Hospital, Shanghai, China

Correspondence Address:
Niansong Wang
Department of Nephrology, Shanghai Jiao Tong University Affiliated Sixth People«SQ»s Hospital, Shanghai - 200 233


Objective: The aim was to study the possible impact of uremia and hemodialysis on tumor markers including carcinoembryonic antigen, carbohydrate antigen (CA)19-9, neuron-specific enolase, CA242, ferritin, β-human chorionic gonadotropin, alpha 1-fetoprotein, human growth hormone, CA153, CA125, prostate-specific antigen (PSA), and free PSA (fPSA). Materials and Methods: Under a controlled study design, we assigned 143 uremia patients receiving maintenance hemodialysis into hemodialysis group, 143 uremia patients (chronic kidney disease phase 5) who did not receive hemodialysis into nonhemodialysis group, and 429 healthy people into the control group. Serum levels of tumor markers were determined using protein chip and were compared among groups. Patients in hemodialysis group were further divided into two subgroups according to the duration of hemodialysis: Group A (maintenance hemodialysis was over 3 mo), Group B (maintenance hemodialysis was <3 mo), and serum levels of tumor markers were compared between these two subgroups. Results: Ferritin levels were 22.1% and 5.8% in hemodialysis and nonhemodialysis groups, respectively, with statistically significance (P < 0.05), but any of the other tumor marker levels in these two groups were without statistical differences (P > 0.05). In hemodialysis group, 12.23% of patients were found with higher serum levels of CA125 than normal, 9.75% with higher levels of PSA than normal, and 11.3% with higher levels of fPSA than normal. In nonhemodialysis group, 14.91% of patients were found with higher levels of CA125 than normal, 8.76% with higher levels of PSA than normal, and 10.8% with higher levels of fPSA than normal. In the control group, patients with higher levels of the above markers accounted for <5%. Conclusion: Ferritin, CA19-9 and CA125 are elevated to various degrees in patients with uremia who are receiving hemodialysis or do not receive hemodialysis. Hemodialysis is not able to eliminate tumor markers.

How to cite this article:
Zhang X, Wang N, Xue Q, Jian G, Gao X. Impact of hemodialysis and uremia on serum levels of tumor markers.J Integr Nephrol Androl 2015;2:23-28

How to cite this URL:
Zhang X, Wang N, Xue Q, Jian G, Gao X. Impact of hemodialysis and uremia on serum levels of tumor markers. J Integr Nephrol Androl [serial online] 2015 [cited 2021 Apr 21 ];2:23-28
Available from:

Full Text


Tumor markers are essentially bioactive materials that are abnormally expressed by tumor tissues and cells. These materials were either observed in the human embryo other than in normal human tissues, or were at levels exceeding those in healthy human bodies. Currently, clinical detection techniques mainly determine levels of tumor-associated antigens that are not tumor specific, but rather an issue of quantity. And no single tumor marker is completely tumor-specific so far, as one type of tumor may produce various tumor markers, and different types of tumors or one type of tumor of different histological types could produce identical or different tumor markers. In recent years, detection of various tumor markers in combining was adopted clinically in order to increase the positive rate of diagnosis. Multi-tumor-marker protein chip detection system (termed C12 briefly) is a combined detection of various tumor markers using modern technology of biochips, which is applied in early screening of tumors. [1] In patients with uremia (where there is a notable percentage of senile people with a high incidence of tumor) who have a higher incidence of complicating tumors when compared with healthy people, tumor marker screening should be performed as a routine. Serum tumor markers are mostly macro-molecular glycoproteins, and the metabolism pathway of which are not fully understood. The impaired renal function and the absorption of dialysis membrane may have a certain impact on serum levels of tumor markers. Under a controlled study design, we compared serum tumor marker levels in healthy people with those in patients receiving maintenance hemodialysis in Department of Nephrology, Shanghai Jiao Tong University Affiliated Sixth People's Hospital and patients at chronic kidney disease (CKD) phase 5 without hemodialysis in order to explore the possible impact of uremia and hemodialysis on serum levels of tumor markers.



Hemodialysis group

In hemodialysis group, 143 patients (aged 66.29 ± 11.26 years, 70 males and 73 females, without any evidence of malignancy as validated after follow-up) receiving maintenance hemodialysis in Department of Nephrology, Shanghai Jiao Tong University Affiliated Sixth People's Hospital from January 2005 to January 2008, using Fresenius AV600 dialyser (1.3 m 2 of polysulfone membrane) 2-3 times/week and 4-5 h in each time were included. Of these patients, primary disease in 51 were chronic nephropathy, 43 were hypertensive renal arteriolar sclerosis, 37 were diabetic nephropathy, 4 were polycystic kidney, 3 were lupus nephritis, 2 were vasculitis, 1 was urate nephropathy, 1 was obstructive nephropathy, and 1 was chronic pyelonephritis. According to the duration of hemodialysis, patients in hemodialysis group were classified into two subgroups: Group A (duration of maintenance hemodialysis was over 3 mo) and Group B (duration of maintenance hemodialysis was <3 mo).

Nonhemodialysis group

In nonhemodialysis group, 143 patients with uremia at CKD phase 5 (aged 65.45 ± 11.65 years, 72 males and 71 females) hospitalized in Department of Nephrology in our hospital during the same period who did not receive hemodialysis or peritoneal dialysis were selected. Of these patients, primary diseases in 52 were chronic nephropathy, 44 were hypertensive renal arteriolar sclerosis, 40 were diabetic nephropathy, 4 were lupus nephritis, 2 were vasculitis, and 1 was urate nephropathy.

Control group

At the ratio of 1:3, 429 healthy people (217 males and 212 females, aged 63.4 ± 11.05 years) receiving a physical examination in our hospital during the same period were selected as a control group. There were no statistically significant differences in age or gender distribution among three groups.


Starving veno-blood samples were collected once in the morning and serum were separated. Serum levels of carcinoembryonic antigen (CEA), carbohydrate antigen (CA)19-9, neuron-specific enolase (NSE), CA242, ferritin, β-human chorionic gonadotropin (HCG), alpha 1-fetoprotein (AFP), human growth hormone (HGH), CA153, CA125, prostate-specific antigen (PSA), and free PSA (fPSA) were determined using C-12 chips. C-12 chip kit and HD2004 automatic biochip detector were purchased from Shanghai Shukang Science and Technology Co., Limited. Each chip comprises 48 micropores (10 mm × 10 mm) in which there were 25 reactive points (24 points were precoated with 12 detecting antibodies (primary Abs). Detection steps:

Chips and reagents were incubated at room temperature for a while;Standard, correcting and internal control samples, together with l00 μL of each serum sample to be detected were sampled into each micropore one by one;The chips were placed in reactive channels of HD2004 detector, shivering automatically, washed and then took luminescence solution (the substrate of which is horseradish peroxidase);After reaction, chip detector would check the 25 reaction points in all the micropores one by one;Specialized software was used to read and analyze data, and print detection reports.

Normal reference ranges: CEA 0-10.0 ng/mL; CA19-9 0-35.0 KU/L; NSE 0-13.0 ng/mL; CA153 0-35.0 KU/L; CA242 0-20.0 KU/L; ferritin (female): 0-291 ng/mL; (male): 0-322 ng/mL; β-HCG 0-3.0 ng/mL; AFP 0-20.0 ng/mL; HGH 0-10.0 ng/mL.

Statistical analysis

SPSS 11.0 statistical software was used for data procession. Data of abnormal distribution were presented by medians. Kruskal-Wallis rank sum test was used for multi-group difference analysis. Wilcoxon Mann-Whitney rank sum test was used for group-group analysis. P < 0.05 was adopted as with statistical significance.


Results of tumor markers detection in three groups

Data in hemodialysis group were as follows: Ferritin (22.1%, n = 30), CA125 (12.23%, n = 21), PSA (9.75%, n = 14), fPSA (11.3%, n = 16), and CA19-9 (15.59%, n = 22) levels were higher than normal; Serum levels of CEA, CA153, CA242, β-HCG, HGH of 1.2% (n = 2) patients were higher than normal. Data in nonhemodialysis group were as follows: Ferritin (5.8%, n = 8), CA125 (14.91%, n = 21), CA19-9 (13.61%, n = 19), PSA (8.76%, n = 13), and fPSA (10.8%, n = 15) levels were higher than normal; Serum levels of CEA of 2.5% patients were higher than normal. There were <0.05% of patients with higher serum levels in any of these parameters in the control group. In these three groups, detection levels of tumor markers were presented by medians, 5% fractile and 95% fractile [Table 1].{Table 1}

Comparison of tumor markers in three groups

The results from Kruskal-Wallis rank sum tests showed that there were statistical significant differences in serum ferritin levels, but not in any other tumor markers (P > 0.05) in hemodialysis group and nonhemodialysis group; there were significant differences of tumor markers in serum levels of Ferritin, CA19-9 and CA125 (P < 0.05) among hemodialysis, nonhemodialysis and control group; there were no significant differences in other tumor markers (P > 0.05).

Impact of the duration of hemodialysis on serum tumor marker levels

There were no statistically significant differences (P > 0.05) in any of the tumor markers among patients in subgroups A and B, as was analyzed using Wilcoxon Mann-Whitney rank sum test [Table 2].{Table 2}


Tumor markers comprise five categories of markers including CEAs (CEA, AFP, etc.), CA153, CA125, CA242, CA19-9, etc., enzymes (PSA, NSE, etc.), hormones (HGH, HCG, etc.), and other proteins (ferritin, etc.):

CEAs, which should be no longer synthesized or secreted after birth, are resynthesized and secreted via the activation of "inactivated" genes. These abnormally expressed CEAs enter blood and lymphoid flow, causing the increased serum levels of CEA, AFP, etc.;Carbohydrate antigenic markers are antigenic materials on the surface of or secreted by tumor cells. These molecules are also called glycol-antigens as they are monoclonal antibodies. These antigens are extremely low in healthy individuals, but are abnormally elevated in patients with cancer;Enzyme activity is significantly altered in patients with cancer due to that tumor cells or tissues induce other cells or tissues to produce abnormal enzyme activity, or the enzymes in vigorously metabolizing tumor cells with increased permeability are released into blood flow; or impaired organ functions due to tumor invasion causing the impaired inactivation and excretion of enzymes; or reflux of enzymes in certain cavities into blood flow due to tumor compression;Hormones secretion are increased at the transformation of cells with hormone secretion ability, or when normal nonhormone-secreting cells are transformed into hormone secreting cancer cells;Other protein markers are with physiological functions that are abnormally increased and released into serum and other body fluids in tumor settings.

Tumor markers are of low specificity, and combined detection of the above tumor markers may increase the sensitivity of diagnosis. Tumor markers entering blood flow and the body fluid are excreted by the liver and the kidney. False positive results might be detected in patients with impaired renal functions, especially in patients with terminally impaired renal functions. In these settings, abnormal laboratory examination results should be further analyzed by doctors.

The multi-tumor marker protein chip detection system (C12) is the integration of immunology and biosignal integration, in which micro biochemical analysis system is established on fixed surface in order to capture and quantify multi-tumor markers in samples via transfer of biosignals into photoelectric signals using highly sensitive tracing techniques. It has been demonstrated in experiments that this is an easy and practical technique that promotes the early diagnosis of various diseases. [1] Sun et al. [2] observed a 68.18% of sensitivity and 97.1% of specificity using C12 in 1147 patients with cancer and 793 healthy individuals.

In the current study, we observed that: Some of the tumor markers were elevated to various degrees in patients in both hemodialysis and nonhemodialysis groups.

There have been reports on studies of small sample size with controversial results: Arik et al. [3] reported that CA199 was elevated, while no significant differences were observed in CEA, AFP and PSA levels in patients with uremia as compared with a healthy population; Xiaofang et al. [4] have reported that serum CEA and CA199 were higher in patients with creatinine clearance rate (Ccr) ≤30 mL/min than those in patients with Ccr ≥50 mL/min. And serum CEA and CA199 were inversely proportional to Ccr, while there were no significant differences in AFP in patients with various degrees of renal functions; Lye et al. [5] reported that 17% of patients receiving hemodialysis had increased CEA and 57% had increased CA19-9. In patients receiving peritoneal dialysis, 37% had increased CEA and 53% had increased CA19-9, while AFP and PSA levels were not elevated in patients receiving dialysis or kidney transplantation; Polenakovic et al. [6] reported that NSE, CEA and CA19-9 were increased in patients with uremia.

Serum CA125 is influenced by various factors and is elevated in patients with chronic hepatic diseases, hepatic cirrosis, heart failure and fluidified serous cavity. [7],[8] Increased CA125 of various degrees was observed in patients receiving hemodialysis in most studies. [6],[9] Lye et al. [5] observed that 27% of patients receiving hemodialysis had increased CA125. Bastani and Chu [10] found that 8% of hemodialyzed patients had CA125 levels above normal while most of the others had normal CA125 levels; However, Xiaofang et al. [4] reported no significant differences in CA125 among various Ccr groups. The results in our study showed that 12.23% of patients in hemodialysis group and 14.91% in nonhemodialysis group had above-normal CA125 levels. Conclusively, some of the hemodialyzed patients may have increased serum CA125. In these patients who are possibly complicated with fluidified serous cavity, gynecologic examinations should be performed, and regular follow-ups should be carried out after exclusion of tumors.

Kamata and Fushimi [11] observed that patients with end-stage renal disease (ESRD) had a higher incidence of prostatic cancer than a healthy population. In their study on 724 male patients with hemodialysis, they found a 1.4% incidence of prostatic cancer. PSA has two forms in male serum: CPSA and fPSA, and total PSA (tPSA) is a routinely examined parameter, with the ratio of fPSA: tPSA as the indicator of malignancy. CPSA has a too big molecular weight to be filtered through glomeruli, and its serum half-life is 2-3 days. Sasagawa et al. [12] studied 93 male patients receiving hemodialysis and 2298 healthy people and observed that serum tPSA was not significantly different between hemodialysis and nonhemodialysis groups. Other studies have also observed similar serum PSA in patients with terminally impaired renal functions and in healthy people; [3],[4],[5] However, Wada et al. [13] reported in a large-scale controlled clinical trial (1250 patients receiving maintenance hemodialysis and 1007 healthy people as a control) that there were statistically significant differences in tPSA levels among patients receiving maintenance hemodialysis and healthy people. However, as they also observed that male patients with ESRD with PSA >10 ng/mL had a higher incidence of prostatic cancer, they concluded that PSA level of ESRD patients was of diagnostic value for prostatic cancer. fPSA has a relative small molecular weight and could be filtered through glomeruli. It has a short serum half-life of 2-3 h. Among healthy people, increased fPSA is regarded as possibly benign, while it could no longer be granted as a reflection of benign diseases when much higher fPSA is observed in ESRD patients with impaired glomerular filtration rate. Sasagawa et al. [12] observed that fPSA level was higher in hemodialysis group than that in the control group. Bruun et al. [14] also observed that significantly more patients receiving hemodialysis and peritoneal dialysis had increased fPSA; Recently, Bruun et al. [15] observed significant correlation between fPSA and renal functions in patients with medium impaired renal functions, and significantly more CKD patients with medium impaired renal functions had increased fPSA, which may lead to misdiagnosis. Thus, prostatic cancer should be excluded in ESRD patients with either PSA or fPSA elevation. We observed in our study that 9.75% of patients in hemodialysis group had above-normal PSA, 11.3% had above-normal fPSA. In nonhemodialysis group, 8.76% of patients had above-normal PSA, 10.8% had above-normal fPSA; there were no statistical differences compared with those of the control group, which was possibly due to inadequate sample size.

Tumor markers are generally macro-molecular materials. Molecular weight of AFP is 70 kD; CEA, 180-200 kD; PSA, 33-34 kD; fPSA, 28 kD; CA125, around 200 kD; CA15-3, 400 kD; CA19-9, 5000 kD; NSE, 78 kD; HGH, 22 kD; HCG, 13-15 kD; and ferritin, 450 kD. Low- and high-flux dialysis membrane could respectively filter materials of <5 kD and 50 kD. Obviously, tumor markers could theoretically not be eliminated through low-flux dialysis membrane while some tumor markers could be eliminated through high-flux dialysis membrane. All the hemodialyzed patients in our hospital were using Fresenius F6 polysulfone membrane which is a low-flux dialysis membrane. It is demonstrated that there were no statistically significant differences in tumor marker levels among patients receiving maintenance hemodialysis for over or <3 mo. There were also no significant statistical differences in tumor marker levels between hemodialyzed patients and nonhemodialyzed patients at CKD phase 5 who are gender and age comparable patients with similar renal functions. Thus, our observation demonstrated that low-flux dialysis membrane could not eliminate tumor markers.

Previously, there were controversial ideas on whether hemodialysis could eliminate tumor markers: Tzanakis et al. [16] observed that hemodialysis could not eliminate PSA. Similarly, Monath et al. [17] observed higher serum PSA levels in 26 male patients with terminal renal diseases when compared with those after dialysis, suggesting that PSA was not eliminated by hemodialysis. In 149 patients dialyzed through high- or low-flux, Djavan et al. [18] observed that neither membrane could eliminate tPSA, and that fPSA could be eliminated by both high- or low-flux membrane with better elimination through high-flux membrane. Kashiwabara et al. [19] observed that patients who were hemodialyzed using fibrous membrane and synthesized membrane had increased CEA level due to pachemia, indicating that neither membrane could eliminate CEA.


The proportion of patients with terminally impaired renal functions or receiving hemodialysis had increased serum tumor markers such as ferritin, CA19-9, and CA125 to various degrees with a certain percentage of false positive rate; low-flux dialysis membrane could not eliminate tumor markers. Consequently, examinations for exclusion of malignancy and dynamic follow-up are necessary in patients with elevated tumor markers who had uremia or were receiving hemodialysis, as both complicating tumors and false positive detection are possible.


The research was supported by National Nature Science Foundation of China (No. 81270824); Shanghai Science and Technology Project (No. 11DZ1921904; 11410708500). The funders had no role in study design, data collection and analysis, decision to publish and preparation of the manuscript.


1Weinberger SR, Dalmasso EA, Fung ET. Current achievements using ProteinChip Array technology. Curr Opin Chem Biol 2002;6:86-91.
2Sun Z, Fu X, Zhang L, Yang X, Liu F, Hu G. A protein chip system for parallel analysis of multi-tumor markers and its application in cancer detection. Anticancer Res 2004;24:1159-65.
3Arik N, Adam B, Akpolat T, Hasil K, Tabak S. Serum tumour markers in renal failure. Int Urol Nephrol 1996;28:601-4.
4Xiaofang Y, Yue Z, Xialian X, Zhibin Y. Serum tumour markers in patients with chronic kidney disease. Scand J Clin Lab Invest 2007;67:661-7.
5Lye WC, Tambyah P, Leong SO, Lee EJ. Serum tumor markers in patients on dialysis and kidney transplantation. Adv Perit Dial 1994;10:109-11.
6Polenakovic M, Sikole A, Dzikova S, Polenakovic B, Gelev S. Acquired renal cystic disease and tumor markers in chronic hemodialysis patients. Int J Artif Organs 1997;20:96-100.
7Sevinc A, Buyukberber S, Sari R, Turk HM, Ates M. Elevated serum CA-125 levels in patients with nephrotic syndrome-induced ascites. Anticancer Res 2000;20:1201-3.
8Xiao WB, Liu YL. Elevation of serum and ascites cancer antigen 125 levels in patients with liver cirrhosis. J Gastroenterol Hepatol 2003;18:1315-6.
9Odagiri E, Jibiki K, Takeda M, Sugimura H, Iwachika C, Abe Y, et al. Effect of hemodialysis on the concentration of the seven tumor markers carcinoembryonic antigen, alpha-fetoprotein, squamous cell carcinoma-related antigen, neuron-specific enolase, CA 125, CA 19-9 and CA 15-3 in uremic patients. Am J Nephrol 1991;11:363-8.
10Bastani B, Chu N. Serum CA-125 level in end-stage renal disease patients maintained on chronic peritoneal dialysis or hemodialysis: The effect of continuous presence of peritoneal fluid, peritonitis, and peritoneal catheter implantation. Am J Nephrol 1995;15:468-72.
11Kamata T, Fushimi K. Prevalence of prostate cancer in end-stage renal disease patients. Urol Int 2008;80:419-24.
12Sasagawa I, Kubota Y, Hayami S, Adachi M, Nakada T, Miura H, et al. Serum levels of total and free prostate specific antigen in men on hemodialysis. J Urol 1998;160:83-5.
13Wada Y, Nakanishi J, Takahashi W, Kai N, Nakayama Y, Yamashita Y, et al. Mass screening for prostate cancer in patients with end-stage renal disease: A comparative study. BJU Int 2006;98:794-7.
14Bruun L, Björk T, Lilja H, Becker C, Gustafsson O, Christensson A. Percent-free prostate specific antigen is elevated in men on haemodialysis or peritoneal dialysis treatment. Nephrol Dial Transplant 2003;18:598-603.
15Bruun L, Savage C, Cronin AM, Hugosson J, Lilja H, Christensson A. Increase in percent free prostate-specific antigen in men with chronic kidney disease. Nephrol Dial Transplant 2009;24:1238-41.
16Tzanakis I, Kazoulis S, Girousis N, Kagia S, Spandidakis V, Karefillakis N, et al. Prostate-specific antigen in hemodialysis patients and the influence of dialysis in its levels. Nephron 2002;90:230-3.
17Monath JR, Burkart JM, Freedman BI, Pittaway DE, Russell GB, Assimos DG. Effects of hemodialysis on prostate-specific antigen. Urology 1993;42:398-400.
18Djavan B, Shariat S, Ghawidel K, Güven-Marberger K, Remzi M, Kovarik J, et al. Impact of chronic dialysis on serum PSA, free PSA, and free/total PSA ratio: Is prostate cancer detection compromised in patients receiving long-term dialysis? Urology 1999;53:1169-74.
19Kashiwabara K, Nakamura H, Yagyu H, Kishi K, Matsuoka T, Esaki T. Changes in squamous cell carcinoma-related antigen levels before and after hemodialysis in relation to the model of dialyzer employed. Intern Med 2000;39:291-5.