You are on page 1of 6

Cancer Detection and Prevention 29 (2005) 570–575 www.elsevier.

com/locate/cdp

Validity of serum pepsinogen levels and quininium resin test combined for gastric cancer screening
Jang-Rak Kim MD, Dr PHa,*, Kyongwu Lee MPHb, Woon-Tae Jung MD, PhDc, Ok-Jae Lee MD, PhDc, Tae-Hyo Kim MDc, Hyun-Jin Kim MDc, Jong-Sil Lee MDd, Douglas J. Passaro MD, MPHb,1
Department of Preventive Medicine, College of Medicine, Institute of Health Sciences, Gyeongsang National University, 92 Chilam-Dong, Jinju 660-751, South Korea b School of Public Health and of Medicine, University of Illinois-Chicago, 1603 W. Taylor St., Rm 958, Chicago, IL 60612, USA c Department of Internal Medicine, College of Medicine, Gyeongsang National University, 92 Chilam-Dong, Jinju 660-751, South Korea d Department of Pathology, College of Medicine, Gyeongsang National University, 92 Chilam-Dong, Jinju 660-751, South Korea Accepted 25 July 2005
a

Abstract Background: To investigate the feasibility of combining several serum markers into a valid serum screening tool for gastric cancer, we performed a study evaluating the association between gastric cancer and precancerous conditions and a blood test for gastric acidity (the blood quininium resin test [QRT]) combined with serum pepsinogen levels. Methods: We performed immunoradiometric assays of serum pepsinogen I (PG I), II (PG II) levels, and QRT’s in 10 endoscopically normal subjects, in 20 patients with chronic atrophic gastritis, and in 13 patients with biopsy-confirmed gastric adenocarcinoma. Results: Serum PG I, II levels, I/II ratio were significantly different among normal, gastritis, and cancer patients. Serum PG I/II ratios were much lower in cancer patients. Serum quinine levels by QRT were correlated with PG I/II ratio (rs = 0.39, p < 0.01). Age was negatively correlated both with PG I/II ratio (rs = À0.58, p < 0.01) and serum quinine level (rs = À0.45, p < 0.01). The screening using serum PG levels was more valid (sensitivity of 69%, specificity of 77%) than that using QRT alone. The combination of serum PG levels and QRT increased specificity for detecting gastric cancer to 87% without altering sensitivity. Conclusion: Although blood QRT is a useful addition to other serum screening tests for gastric cancer, these tests alone are not sufficiently accurate as screening tools for gastric cancer. # 2005 International Society for Preventive Oncology. Published by Elsevier Ltd. All rights reserved.
Keywords: Pepsinogen I; Pepsinogen II; Pepsinogen I/II ratio; Sensitivity; Specificity; Chronic atrophic gastritis; Combined diagnostic test; Gastric cancer screening; Quinium resin test; QRT; Hypochlorhydria

1. Introduction Gastric cancer remains the second highest cause of cancer death worldwide [1] and in Korea [2]. Primary prevention to reduce the incidence by changing the dietary and lifestyle factors and to treat H. pylori infection, while important, may be less feasible than secondary prevention to detect and treat
* Corresponding author. Tel.: +82 55 751 8753; fax: +82 55 755 2445. E-mail addresses: jrkim@gsnu.ac.kr (J.-R. Kim), doug@uic.edu (D.J. Passaro). 1 Tel.: +1 312 355 0696; fax: +1 312 996 0064.

patients earlier through screening. However, few persons are able to obtain screening endoscopies. Valid, non-invasive, and inexpensive serum screening tools are needed. Among non-invasive screening methods for gastric cancer, the most common parameters in use before the introduction of H. pylori serology were pepsinogen I (PG I) and pepsinogen II (PG II) as well as their ratio (PG I/PG II). Pepsinogens are proteases that are secreted into the gastric lumen and transformed there into pepsin. The chief cells in the gastric fundus and corpus produce PG I. Chief cells and mucous neck cells throughout the entire stomach produce PG II. Serum levels of PG I and PG II increase

0361-090X/$30.00 # 2005 International Society for Preventive Oncology. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.cdp.2005.07.005

Of the 147 enrolled persons (40 endoscopically normal subjects. severe anemia). gastric adenoma. II. This system has endoscopic and histological criteria. chronic atrophic gastritis. We calculated the sensitivity and specificity of the combination of a PG I level <70 ng/ml with PG I/PG II ratio <3. I/II ratio. <15.0 and a quinine level <9 ng/ml. pylori infection. Serum PG I. We evaluated the sensitivity and specificity of the combination of a PG I level <70 ng/ml and a PG I/PG II ratio <3. which could influence the serum pepsinogen levels or gastric acidity. Results There was a significant difference in age. The mean of two readings was taken as the final value. The QRT is based upon the principle that hydrogen ions released from gastric parietal cells displace quinine from orally taken quininium resin complex at gastric pH below 3. II. The most suitable cut-off point was a PG I concentration of less than 70 ng/ml and a PG I/PG II ratio less than 3. especially PG I and the PG I/PG II ratio were significantly lower than those of normal controls and correlated well with the extent of chronic gastritis associated with the cancerous stomach [4].J. We investigated the feasibility of combining serum PG levels and blood QRT into a valid serum screening tool for gastric cancer. a fasting blood sample was drawn. After informed consent was obtained from each eligible subject. and 26 patients with gastric cancer). II levels. Inclusion criteria for participation in this study were: age between 20 and 75 years and providing oral or written informed consent. Separated sera were kept at À70 8C for measuring PG I and PG II concentrations by immunoradiometric assay [10] using PG I. Another possible non-invasive screening test for gastric atrophy and acidity is the quininium resin test (QRT). QRT may be a useful screening tool for detecting gastric cancer and precancerous lesions. receiver-operator characteristic curves were estimated. and 13 gastric cancer patients. I/II ratio. 3. serum PG I. <12. Dubuque. Each person who also agreed to get blood QRT was asked to eat a bowl of gruel (150 cc) mixed with 80 mg/kg of quininium resin (Rohm and Hass. or mild antral gastritis) or moderate-to-severe chronic atrophic pangastritis including intestinal metaplasia. regimens for eradication of H. II RIA BEAD kits from Dainabot. 2. All cases of adenocarcinoma were confirmed histologically. The free quinine is then absorbed into the bloodstream. Exclusion criteria included the following: current pregnancy. To determine the most suitable one with serum quinine. 81 patients with chronic atrophic gastritis. All the data was analyzed by SPSS for Windows software (version 10. or antibiotics) in the past month. IA. There was no significant difference in serum quinine level. Serum quinine content was measured using photomultiplier fluoromicrophotometer (Turner Instruments. USA). Serum PG I/II ratios were much lower in cancer patients. chronic renal disease. Il).0 [6]. the most suitable cut-off point for gastric cancer screening [6].0. only 43 persons (10 endoscopically normal subjects. We recruited persons who had endoscopically normal findings (including mild superficial gastritis. / Cancer Detection and Prevention 29 (2005) 570–575 571 with gastric mucosal inflammation and decrease below normal when gland loss occurs [3]. and gastric cancer patients by Kruskal–Wallis test. Endoscopic criteria take into account the appearance of gastric inflammation. and quinine levels. Materials and methods The study subjects were persons who underwent diagnostic upper gastrointestinal endoscopies in the outpatient department of internal medicine or health examinees who underwent screening upper endoscopies in the department of health promotion center at Gyeongsang National University Hospital between June 2002 and January 2003.-R.0) (SPSS Inc. (<10. The quinine content of the blood sample is then determined using a fluorometer [7]. The sensitivity and specificity of pepsinogen screening for gastric cancer were 57 and 80%. past or present history of any general diseases or medication. six biopsies of the antral and corpus mucosa were taken from each subject. When histology was deemed to be necessary.5. Chicago. serum PG I. <11. Philadelphia. or <17). Typical cases can be diagnosed by endoscopic findings only. The serum pepsinogen levels of stomach cancer patients. The study was approved by the Ethics Committee at Gyeongsang National University Hospital. respectively in Japan in a meta-analysis [5]. and 13 patients with gastric cancer) agreed to take QRT. or cancer (adenocarcinoma confirmed histologically). Kim et al. 20 chronic atrophic gastritis. and I/II ratio among 10 endoscopically normal. After 60 min a second blood sample was drawn. <14. previous gastrectomy. 20 patients with chronic atrophic gastritis. too. benign gastric or duodenal ulcer.. An interview and medical record were used to collect information on age. PA) after upper endoscopy. . Tokyo. medical illness (chronic liver disease. or use of certain medications (acid-suppressing drugs. A variety of cut-off values to detect gastric cancer were examined to use the combination of serum PG and quinine levels for gastric cancer screening tool. Spearman’s correlation coefficients (rs) were calculated to estimate the strength of associations among age. Because a higher level of pH (greater than 4) in gastric juice is associated with chronic atrophic gastritis and gastric cancer [8]. <13. sex. We identified patients with moderate-to-severe chronic atrophic pangastritis according to the updated Sydney System [9]. and quinine levels were compared among groups of endoscopically normal subjects. diabetes mellitus. <16. and all the separated sera were kept at À70 8C.

4.02 0.0 and serum quinine level <14. .6 Serum quinine level (ng/ml) 13. Serum quinine levels correlated little with serum PG I and PG II levels. quininium resin test (QRT).01) with PG I/PG II ratio.7 Æ 4.0 23.9 Æ 4.0%) Subtotal of carcinoma absent group 7 (23.01) with advancing age.38y 43 Pepsinogen I/II ratio – – 0.0 Æ 39. / Cancer Detection and Prevention 29 (2005) 570–575 Table 1 Serum pepsinogen. Fig.9 Æ 1.0 were used as screening criteria were 69. Fig.8%) 13 (100.0 (B).4 47.68* À0.0 ng/ml (or <15.7% for detecting gastric carcinoma (Table 5.8 and 80.39* À0.05.2 Æ 1.0 among 43 study subjects Screening with serum pepsinogen Gastric carcinoma Absent Normal Positive Negative Total a b Total Present Atrophic Gastritis 7 (35.0 13.58. Table 3 Sensitivity and specificity for detecting gastric carcinoma when screening criteria is serum pepsinogen I < 70 ng/ml and I/II ratio <3.572 J. respectively (Table 4.0%) 20 (100.0%) Sensitivity.45.01 among three endoscopically defined groups by Kruskal–Wallis test.6 20. Sensitivity and specificity for detecting gastric carcinoma when serum pepsinogen I levels <70 ng/ml and I/II ratio <3. I/II ratio. and decreased significantly (rs = À0.2 and 76. II level. Fig.69* À0. Including subjects with moderate-to-severe chronic atrophic pangastritis endoscopically.0%) 9 (69. Sensitivity and specificity for detecting gastric carcinoma when serum quinine levels <9. p < 0. However.0%.0 ng/ml alone (C).0%) 10 (100.2 2.0%) 0 (0. and age in 43 study subjects Pepsinogen I Pepsinogen II Pepsinogen I/II ratio Quinine Age Number * y Pepsinogen II – À0. Serum PG II levels increased significantly (rs = 0. Kim et al.05 among three endoscopically defined groups by Kruskal–Wallis test.2%a) 4 (30. p < 0.01 43 Among a variety of combinations.8 57. p < 0.0%) 13 (65.0 ng/ml were used as the screening criterion were 53. Discussion Screening plays an important role in detecting gastric cancer in an early stage and in potentially reducing death p < 0.38.3 p < 0. serum quinine levels increased significantly (rs = 0.8 10 20 13 33.0 ng/ml (or <15. p < 0. or <16) (A). The following cut-off points are to be suitable: a serum PG I < 70 ng/ml with I/II ratio < 3.1 63. or <16) giving a sensitivity of 69.0%) 10 (100. 1). and their combination.7%. 1). Fig.4 Æ 1. Serum PG I levels showed almost no variation with age (rs = 0.1 Æ 12. 1.12 0. and a serum quinine <9. Table 2 Spearman correlation coefficients among serum pepsinogen I.01).01) with advancing age (Table 2).39. and serum quinine levels (mean Æ standard deviation) in 43 study subjects Gross endoscopic findings Normal c Atrophic gastritisd Cancer a b c d Number Mean agea Serum pepsinogen level (ng/ml) Pepsinogen I b I/II ratioa a Pepsinogen II 8. Specificity. Receiver operator characteristics (ROC) curves generated with serum pepsinogen (PG) I level with I/II ratio.3%) 23 (76.2% and a specificity of 86.2 Æ 14. and the PG I/PG II ratio decreased significantly (rs = À0.0%) 16 (37.9 Æ 12. p < 0. although it was lowest in cancer patients (Table 1).2 Æ 4.01. quinine level.-R. the most suitable cutoff point was both serum pepsinogen I <70 ng/ml with I/II ratio <3.58* 43 Quinine – – – À0.0 3. Including subjects with endoscopically normal finding.8%) 43 (100. p < 0. a serum PG I < 70 ng/ml with I/II ratio < 3.0 37.45* 43 0.1 Æ 6.5 4.7%)b 30 (100.18 0.3 9.0 and serum quinine < 14.2%) 27 (62.3 50.1 Æ 32. 1).05). respectively (Table 3.

0%b) 30 (100. the pepsinogen test method also has utility as a screening test for high-risk subjects with atrophic gastritis.2%) 20 (69. pylori serology were pepsinogen I and pepsinogen II as well as their ratio (PG I/PG II) [3]. QRT identifies hypochlorhydria accurately and non-invasively [7]. Kim et al.0 ng/ml among 43 study subjects Screening with serum quinine Gastric carcinoma Absent Normal Positive Negative Total a b 573 Total Present Atrophic Gastritis 5 (25. Endoscopy is a perfect technique for an individual patient but does not meet any of the requirements for a screening test. / Cancer Detection and Prevention 29 (2005) 570–575 Table 4 Sensitivity and specificity for detecting gastric carcinoma when screening criteria is serum quinine <9. The results are similar to 57% of sensitivity and 80% of specificity in Japan in a meta-analysis using 42 data sets [5]. <16.0%) 15 (75.0%) 16 (80. Because a higher level of pH (greater than 4) in gastric juice is associated with chronic atrophic gastritis and gastric cancer [8]. such as high convenience and efficiency. or <17). No valid.0%) 0 (0.0%) Sensitivity. Among noninvasive screening methods tested for a long time. <12.0%) Subtotal of carcinoma absent group 6 (20.2 and 76. Specificity.-R.0%) 24 (80.0 and serum quinine level <14. However.8%) 43 (100.7%b) 30 (100. Another possible non-invasive screening test for gastric atrophy and acidity is the quininium resin test.0 [6].7%.0%) Subtotal of carcinoma absent group 4 (13. We calculated the sensitivity and specificity of the combination of a PG I level <70 ng/ml with PG I/PG II ratio <3. and of blood QRT used alone for detecting gastric cancer.0%) 1 (10.2%) 20 (69.0%) 13 (30.0%) 7 (53.0%) 10 (100. .0 and a quinine level <9 ng/ml (<10. <11.3%) 26 (86. However.0%) 9 (69.0 were 69. The original versions of the QRT were 92–95% sensitive but were considered insufficiently specific for diagnostic use for gastric acid secretion [13].2% and a specificity of 86.8%a) 6 (46. The most suitable cut-off point was a PG I concentration of less than 70 ng/ml and a PG I/PG II ratio less than 3. the most common parameters in use before the introduction of H. <15. respectively in Japan in a meta-analysis [5].0%) 13 (30. This study evaluated the sensitivity and specificity of serum PG levels used alone.8%) 13 (100. where endoscopy is unavailable. However. or <16) giving a sensitivity of 69.0%) 20 (100.8%) 43 (100. rather than as a tool for cancer itself [12].0%) 20 (100. together with low burden and costs. or <16) among 43 study subjects Screening with serum pepsinogen and quinine Gastric carcinoma Absent Normal Positive Negative Total a b Total Present Atrophic Gastritis 4 (20.0%) 9 (90. In this study.0 ng/ml (or <15. QRT may be a useful screening tool for detecting gastric cancer. attributable to gastric cancer. and inexpensive screening tool for gastric cancer have yet been developed.0%) 10 (100. A variety of cut-off values to detect gastric cancer were examined to use the combination of serum PG and quinine levels for gastric cancer screening tool.0 ng/ml (or <15. The sensitivity and specificity of pepsinogen screening for gastric cancer were 57 and 80%.2%) 13 (100. <14.J. Specificity.0%) Sensitivity.7% for detecting gastric Table 5 Sensitivity and specificity for detecting gastric carcinoma when screening criteria is both serum pepsinogen I < 70 ng/ml with I/II ratio <3. the QRT performed considerably worse with a measured sensitivity of detecting gastric cancer of 54% with a specificity of 80%. <13. The most suitable cut-off point was both serum pepsinogen I < 70 ng/ml with I/II ratio <3. This study also investigated if the combination of serum PG levels and blood QRT is a more valid screening tool than serum PG levels used alone. noninvasive tests may be used as an adjunct to endoscopy for population-based screening.0 and serum quinine <14. Sensitivity and specificity for detecting gastric carcinoma when the screening criteria are serum pepsinogen I < 70 ng/ml and I/II ratio <3.0%) 10 (100. Endoscopic screening is recommended as the best screening method [11]. respectively. Studies investigating the relationship between hypochlorhydria and gastric malignancy have been limited by difficulties in the non-invasive measurement of gastric acidity. non-invasive.2%a) 4 (30.

[2] Korea National Statistical Office. Yamaki G.15:877–9. References [1] Murray CJ.-R. Meta-analysis on the validity of pepsinogen test for gastric carcinoma. another study demonstrated that there was no age-related increase in achlorhydria by QRT or in atrophic gastritis by ratio of PG I/II ratio in subjects older than 65 years. [3] Kuipers EJ. aspirin. Van Helden PD. Kobayashi K.0 ng/ml yielded sensitivity of only 25% for the detection of chronic atrophic gastritis. [5] Dinis-Ribeiro M. Gastric cancer risk is higher in corpus atrophic gastritis than in antral atrophic gastritis [15]. Kim et al. a quinine level of <9. [7] Passaro DJ. serum pepsinogen I<70 ng/ml with I/II ratio <3. [8] Jaskiewicz K.0 ng/ml (or <15. This low sensitivity suggests that serum pepsinogen and quinine levels are not suitable alone as screening tools for chronic atrophic gastritis. (R05-2002-00000273-0) from Korea Science and Engineering Foundation. In conclusion. J Med Screen 2004.44: 693–7.128:41–7. Annual Report on the Cause of Death Statistics (2002).80: 111–4. When the QRT was used in isolation. [10] Ichinose M. Houston. Jpn J Cancer Res 1989. or NSAIDs) that could influence the serum pepsinogen levels or gastric acidity despite of our exclusion efforts. the distribution of the gastric diseases were arbitrary and does not represent that of the population. Furihata C.10:833–6. Chronic atrophic gastritis. International Workshop on the Histopathology of Gastritis. 2003. It is arguable whether the 10% improvement in specificity that results from the addition of the QRT warrants its addition to the relatively insensitive PG screening approach. [6] Kitahara F. Acknowledgement This work was supported by a grant no. steroid. et al. The updated Sydney System. Kotoh T. The sensitivity of 69% provided by PG screening or by the combined tests limits the applicability of these screening approaches. nitrites and micronutrient levels in a population at risk for gastric carcinoma. The results are similar to 13. whereas. et al.126:183–91. Correa P. Anticancer Res 1990. Radioimmunoassay of serum group I and group II pepsinogens in normal controls and patients with various disease. pylori virulence factors) improve sensitivity and therefore increase the utility of serum screening tests. specificity of 77%) than that using QRT alone. The most suitable cut-off point for detecting gastric cancer. Wiid IJ. Lopez AD. However. Lancet 1997. Masunaga R. Am J Surg Pathol 1996.0 and serum quinine <14. gave a sensitivity of 20. specificity of 87%) for detecting gastric cancer. . regimens for eradication of H. The combination of serum PG levels and QRT gave a higher specificity (sensitivity of 69%. Gut 1999. The significance of low serum pepsinogen levels to detect stomach cancer associated with extensive chronic gastritis in Japanese subjects. By adding QRT to serum PG levels in this study. Yardley JH. et al.7% of specificity when screening criteria is serum I/II ratio <2. dysplasia or chronic atrophic gastritis screening. Surgery 2000. et al. A modification of the quininium resin test for assessing gastric acidity. Miki K. Sato T. only 14 patients (13 cancer and 1 chronic atrophic gastritis) in this study have been evaluated both endoscopically and histologically. Ichinose M. Eur J Gastroenterol Hepatol 2003. another study which evaluated the validity of the combination of a PG level and a gastrin level for detecting gastric cancer showed very high sensitivity (76%) and specificity (96%) [14]. Accuracy of screening for gastric cancer using serum pepsinogen concentrations. Daejeon. The use of a design that employed frequency-matching of cases and controls on the basis of age and sex would have given a sensitivity and a specificity without any age and sex effects.0%. Miki K. which is affected by the background prevalence of the condition of interest. Impact of screening survey of gastric cancer on clinicopathological features and survival: retrospective study at a single institution. but at the cost of a further decrease in sensitivity [3]. respectively. Clin Chim Acta 1982. [11] Kubota H. [9] Dixon MF. Further research could concentrate on whether additional markers (such as gastric levels or identification of specific H.7% of sensitivity and 95.0 were 35.0% and a specificity of 100.574 J. It is therefore conceivable that serum quinine levels may correlate with corpus rather than antrum pathology. et al. This slight augmentation in specificity may be counterbalanced by the inconvenience of the QRT assay. Classification and grading of gastritis.0 and 100. antibiotics.349:1269–76. Mortality by cause for eight regions of the world: global burden of disease study. Genta RM. [4] Miki K.0% for detecting atrophic gastritis. a combination of tests usually increased the specificity for the diagnosis of atrophic gastritis. pylori infection. South Korea: Korea National Statistical Office. Aliment Pharmacol Ther 2002. However. the screening using serum PG levels was more valid (sensitivity of 69%.20:1161–81. / Cancer Detection and Prevention 29 (2005) 570–575 carcinoma. Kawamura N. Triadafilopoulos G. gastric pH. et al. In through the out door: serology for atrophic gastritis.11:141–7. A second limitation of this study is that the study subjects who underwent diagnostic upper gastrointestinal endoscopies may have been receiving certain medications (acid-suppressing drugs. or <16).16:875–80. Sensitivity and specificity for detecting chronic atrophic gastritis in this study when screening criteria is serum pepsinogen I < 70 ng/ml and I/II ratio <3. which means that the age effect in this study is unlikely to be substantial [17]. Further study of a large sample is needed to confirm the associations between serum quinine levels and pathological findings. which involves a test meal followed by blood sampling 60 min later. specificity of the screening tool improved by 10% without compromising sensitivity. et al. As this study was a case-control design. Another limitation is that we could not control for age and sex because of the small sample size. We could not calculate predictive value. However.5 in another study [16]. Hurwitz A.

278:659–62. The Azur-A method as a screening test of gastric acid secretion.23:194–6.-R. Sasajima M. et al. Mohar A. [17] Hurwitz A.12:117–24. Di Mario F. . JAMA 1997.98:735–9. et al. Boeing H. Plebani M. et al. Cancer Epidemiol Biomarkers Prev 2001. Morita M. / Cancer Detection and Prevention 29 (2005) 570–575 [12] Miki K. Schaal SE. Scand J Gastroent 1966. Mexico. Gastric acidity in older adults. et al. Usefulness of gastric cancer screening using the serum pepsinogen test method. Rev Environ Health 1997. Popiela T. Am J Gastroenterol 2003. [16] Ley C. et al. A clinico-epidemiological study on gastritis in gastric carcinoma and in non-cancerous gastric pathology in Poland. Screening markers for chronic atrophic gastritis in Chiapas. [13] Christiansen PM. Kim et al.J.10:107–12. [14] Farinati F. 575 [15] Jedrychowski W. Pepsinogen A/pepsinogen C or pepsinogen A multiplied by gastrin in the diagnosis of gastric cancer? Ital J Gastroenterol 1991.1:9–20. Guarner J. Brady DA.