- Email: [email protected]
Quantification of Serum Levels of Pepsinogens and Gastrin to Assess Eradication of Helicobacter Pylori
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY 2011;9:440 – 442
BRIEF COMMUNICATIONS Quantification of Serum Levels of Pepsinogens and Gastrin to Assess Eradication of Helicobacter Pylori LUIGI GATTA,*,‡ FRANCESCO DI MARIO,‡ DINO VAIRA,§ MASSIMO RUGGE,储 ANGELO FRANZÈ,¶ MARIO PLEBANI,# GIULIA MARTINA CAVESTRO,‡ PAOLO LUCARINI,* MAURIZIO LERA,* and CARMELO SCARPIGNATO‡ *Gastroenterology and Endoscopy Unit, Versilia Hospital, Lido di Camaiore; ‡Section of Gastroenterology, Department of Clinical Sciences, University of Parma, Parma; § Department of Internal Medicine and Gastroenterology, University of Bologna, Bologna; 储Department of Medical Diagnostic Sciences and Special Therapies (Pathology Unit), University of Padova, Padova; ¶Gastroenterology Unit, Hospital of Parma, Parma; #Department of Laboratory Medicine, University Hospital of Padova, Padova, Italy
See editorial on page 373. BACKGROUND & AIMS: We investigated whether serum levels of pepsinogen (sPG)I and sPGII, the ratio of sPGI to sPGII, or serum levels of gastrin-17 (sG17), can be used to assess eradication of Helicobacter Pylori 8 weeks after treatment. METHODS: We performed a prospective study of 228 consecutive patients with H pylori infections. At the start of the trial (baseline), patients were assessed using the 13C-urea breath test (13C-UBT) and endoscopy, and serum levels of pepsinogens and gastrin levels were measured. Patients were offered a 7-day triple therapy and asked to return 8 weeks after treatment for another 13C-UBT and measurements of serum levels of sG17, sPGI, and sPGII (175 patients completed the study). RESULTS: The eradication rate of H pylori was 67%. Percentage variation in levels of sPGI and sPGII, the ratio of sPGI to sPGII, and in levels of sG17 resulted in area under the curve values of 0.858, 0.973, 0.940, and 0.810, respectively, for H pylori eradication. A decrease of 22.7% or greater in the level of sPGII detected H pylori eradication with 100% sensitivity and 96.6% specificity. Spectrum analysis did not identify differences in accuracy. CONCLUSIONS: Percentage variation of sPGII levels 8 weeks after therapy for H pylori infection correlates with eradication. Additional studies are needed to confirm these results. Keywords: Helicobacter pylori; Dyspepsia; Treatment; Follow-up.
H
elicobacter pylori is the principal cause of gastritis and peptic ulcer disease, and has been associated with gastric cancer.1 H pylori infection is associated with reversible modifications in gastric physiology such as an increase in serum pepsinogen levels and gastrin secretion.2,3 Human pepsinogens are subgrouped into 2 types: pepsinogen I exclusively produced by the chief and mucous neck cells of the corpus, and pepsinogen II produced not only in those cells but also in the cardiac, pyloric, and Brunner gland cell.3 Gastrin is a hormone produced predominantly by G cells located in the gastric antrum and to a lesser extent in proximal duodenal mucosa. Two main forms of gastrin are present in blood: gastrin-17 and gastrin-34. Antral mucosa contains predominantly gastrin-17, whereas gastrin-34 is the major molecular form in the duodenum.4 Several studies have tried to investigate whether serum pepsinogens or gastrin can be used to follow-up H pylori infection after treatment, but results are not unequivocal.5–11
The aim of the present study was therefore to assess the use of serum pepsinogen I (sPGI), serum pepsinogen II (sPGII), their ratio (sPGI/sPGII), and serum gastrin-17 (sG17) to diagnose H pylori eradication 8 weeks after the end of treatment.
Materials and Methods This was a prospective cohort study in which different diagnostic tests were compared in a blinded fashion, designed to fulfill the Standards for Reporting of Diagnostic Accuracy (STARD) recommendations.12 Between January 2007 and June 2008, there were 228 consecutive H pylori–infected dyspeptic outpatients,13 aged 18 and older, who underwent an upper endoscopy, and were asked to participate. Patients were excluded from the study if they had been treated previously for H pylori infection or were taking proton pump inhibitors, H2-receptor antagonists, antibiotics, bismuth compounds in the 4 weeks preceding the initial visit, if they were bleeding, or if blood samples were not available for analysis. The study was performed according to good clinical practice and the Declaration of Helsinki and informed consent was obtained from each patient. H pylori status was assessed using histology (specimens collected were stained with H&E and Giemsa) and the 13C-urea breath test (13C-UBT). Patients were classified as infected if both tests were positive. The determination of sPGs and sG17 was made by enzyme-linked immunosorbent assay (Biohit, Helsinki, Finland) in Parma. Biopsies, breath samples, and blood samples were coded differently so that pathologists and technicians were unaware who was the patient tested. Serum analysis was not paired but was performed at 2 separate times (before and after therapy). All recruited patients were given 7-day triple therapy. Patients with peptic ulcer disease also were given a further 4 weeks of proton pump inhibitors. Patients were asked to return 8 weeks after stopping treatment to evaluate their H pylori status using a 13C-UBT (reference standard for the assessment of H pylori eradication) and to collect a blood sample.
Statistical Analysis Because data were not distributed normally, values of sPGs and sG17 are expressed as a median with interquartile Abbreviations used in this paper: CI, confidence interval; 13C-UBT, breath test; IQR, interquartile range; LR, likelihood ratio; PUD, peptic ulcer disease; sG17, serum gastrin-17; sPG, serum pepsinogen; STARD, Standards for Reporting of Diagnostic Accuracy. © 2011 by the AGA Institute 1542-3565/$36.00 doi:10.1016/j.cgh.2010.12.009 13C-urea
May 2011
SERUM PEPSINOGENS AND GASTRIN
Table 1. Characteristics of the Patients Enrolled in the Study Median age, y (IQR) Sex, male/female Alcohol consumption Smokers Duodenal ulcer Gastric ulcer sPGI, g/L (IQR) sPGII, g/L (IQR) Ratio (IQR) sG17 (IQR)
47 (36–56) 102/113 68.84% 53.02% 19.53% 23.26% 75 (58–93) 17 (13–23) 4.3 (2.5–6.5) 13 pg/L (9–17)
ranges (IQRs). The Wilcoxon matched pair test was used when necessary. By using the percentage change in values before and after treatment, a receiver operating characteristic analysis was performed applying nonparametric methods against the predefined reference standard (13C-UBT). Sensitivity, specificity, and the likelihood ratio for a positive (LR-positive) and negative (LR-negative) test were calculated. The spectrum effect14 was evaluated for the following factors: (1) age (dichotomized as ⱕ50 or ⬎50 y), (2) sex, (3) peptic ulcer disease (dichotomized as present or absent), (4) smoking (dichotomized as smokers or nonsmokers), and (5) consumption of alcohol (dichotomized as ⱖ1 glass/day or no consumption). These criteria were determined before commencement of the study. A 2-sided P value of less than .05 for all other tests indicated statistical significance.
Results A total of 228 consecutive H pylori–infected patients were offered entry into the study. Thirteen patients were not included because of exclusion criteria, and 215 patients entered the study. Demographic features and endoscopic findings of the population enrolled are shown in Table 1. Forty patients were lost to follow-up evaluation and 175 patients (men/women, 82/93; median age, 46 y; IQR, 32–53 y) completed the study, with an eradication rate of 67% (95% confidence interval [CI], 60–74). The sPGI level decreased significantly in eradicated patients (from 76 g/L [IQR, 55– 88 g/L] to 62 g/L [IQR, 48.0–77.2 g/L]; P ⬍ .0001). In noneradicated patients, values remained
441
virtually the same (from 80 g/L [IQR, 67–95 g/L] to 86 g/L [IQR, 75–99 g/L]; P ⬎ .1). The sPGII level also significantly decreased in eradicated patients (from 17 g/L [IQR, 13–23 g/L] to 8 g/L [IQR, 7–9 g/L]; P ⬍ .0001). In those not eradicated, values remained stable (from 16 g/L [IQR, 12–24 g/L] to 16 g/L [IQR, 14 –21 g/L]; P ⫽ .705). As a consequence, the sPGI/sPGII level increased in eradicated patients (from 4.2 g/L [IQR, 2.3– 6.2 g/L] to 8.2 g/L [IQR, 5.6 –10.7 g/L]; P ⫽ .0001). In noneradicated patients, values were almost the same (from 4.7 g/L [IQR, 2.8 – 6.8 g/L] to 5 g/L [IQR, 3.5– 6.8 g/L]; P ⫽ .921). Finally, the sG17 level decreased significantly in those eradicated (from 13 pg/L [IQR, 8.5–16 pg/L] to 10 pg/L [IQR, 8 –14 pg/L]; P ⬍ .0001). In those not eradicated, values were not significantly different (from 13 pg/L [IQR, 8.5–16 pg/L] to 13 pg/L [IQR, 10 –18 pg/L]; P ⫽ .1). Receiver operating characteristic analysis established the following regarding a decrease in values: (1) sPGI greater than 6.8% as the cut-off point would result in an AUC of 0.858 (95% CI, 0.798 – 0.906); (2) sPGII greater than 22.7% as the cut-off point would result in an AUC of 0.973 (95% CI, 0.936 – 0.991); (3) an increment in values of sPGI/sPGI of 33.8% or greater as the cut-off point would result in an AUC of 0.940 (95% CI, 0.894 – 0.970); (4) a decrease in values of sG17 greater than 17.6% as the cut-off point would result in an AUC of 0.810 (95% CI, 0.744 – 0.866). The value of AUC achieved with sPGII was significantly higher than that obtained with sPGI (P ⬍ .001), ratio (P ⫽ .042), and sG17 (P ⬍ .001). Sensitivity, specificity, LR-positive, and LR-negative of sPGII were 100% (95% CI, 96.8 –100), 96.6% (95% CI, 91.5–99.1), 29.5% (95% CI, 8.6 –107), and 0% (95% CI, 0 – 0.03), respectively. Spectrum effect analysis performed for sPGII shown that there was no significant difference in terms of sensitivity, specificity, LR-positive and LR-negative with regard to age, sex, peptic ulcer disease (PUD), smoking, and alcohol consumption (Table 2).
Discussion We found that percentage change in values before and after treatment of sPGII could provide a good diagnostic accuracy in evaluating the effect of H pylori treatment 8 weeks after the end of treatment when compared with the 13C-UBT. There have been previous attempts to establish whether changes in
Table 2. Spectrum Effect for Percentage Variation in sPGII Considering Sex, Age, PUD, Smoking, and Alcohol to Assess H pylori Eradication Parameters
Sensitivity (95% CI)
Male Female Age, ⬍50 y Age, ⱖ50 y PUD No PUD Smokers Nonsmokers Alcohol No alcohol
100% (93.6–100) 100% (94–100) 100% (94.3–100) 100% (93.2–100) 100% (91.4–100) 100% (95.1–100) 100% (93.8–100) 100% (93.8–100) 100% (95.1–100) 100% (91.4–100)
aP bP
P valuea 1 1 1 1 1
Specificity (95% CI) 96.2% (81.1–99.3) 97% (84.7–99.5) 97.6% (87.4–99.6) 94.4% (74.2–99) 96.8% (83.8–99.4) 96.4% (82.3–99.4) 94.3% (81.4–98.4) 100% (86.2–100) 97.8% (88.7–99.6) 92.3% (66.7–98.6)
value for comparison of sensitivities or specificities in the 2 strata. value for homogeneity test in the 2 strata.
P valuea 1 .5 1 .5 .3
LR-positive (95% CI) 26 (5.2–146.5) 33 (⬁ to ⬁) 41 (7.9–231.5) 18 (3.8–101.2) 31 (⬁ to ⬁) 28 (5.6–157.9) 17.5 (5.3–63.2) — 46 (⬁ to ⬁) 13 (3–72.9)
P valueb .837 .475 .934 — .262
LR-negative (95% CI) 0 (0–0.06) 0 (0–⬁) 0 (0–0.05) 0 (0–0.07) 0 (0–⬁) 0 (0–0.05) 0 (0–0.06) — 0 (0–⬁) 0 (0–0.09)
P valueb .968 .911 .767 — .719
442
GATTA ET AL
serum pepsinogens or gastrin levels could be used to assess the effects of antibiotic therapy with contrasting results.5–11 Several pathophysiological factors may explain why sPGII, among others, could be a reliable indicator to perform a follow-up evaluation of patients after treatment. sPGII increases more than sPGI in H pylori–infected patients,15,16 and sPGII correlates proportionally according to the severity of H pylori gastric inflammation (including both neutrophilic and lymphomononuclear infiltration) in antrum and corpus better than sPGI.3 H pylori also might influence sPGII by interfering with the release of gastrin and somatostatin.17,18 Finally, H pylori–associated gastritis is localized mainly in the antrum, and this could affect sPGII more than sPGI.3 Our analysis presents several strengths compared with the previous studies: it was designed to fulfill the STARD recommendations,12 work-up bias and expectation bias have been avoided because all patients who had pepsinogens and gastrin blood tests also had histology and the UBT, and those who interpreted the results were not aware of the results of all other tests performed. Finally, a spectrum analysis including 5 variables (sex, age, PUD, smoking, and alcohol consumption) was performed. However, our study presents some drawbacks. First, the identified cut-off points need to be tested in future prospective studies to assess the validity, because the results were dataderived. Second, it was possible to perform the spectrum effect analysis only evaluating 5 variables; however, it could not be excluded that other factors (eg, body mass index, previous or concomitant use of nonsteroidal anti-inflammatory drugs, antisecretory drugs, food, and so forth) could affect the performance of the test. Third, the eradication status was evaluated 8 weeks after the end of treatment; data also are needed for a shorter follow-up period (eg, 4 weeks). We also could not perform a measurement of the economic impact of having this type of noninvasive diagnosis. We also did not perform a matched comparison with the radioimmunoassay. Finally, it is useful to remember that different enzyme-linked immunosorbent assay kits can have different calibrations, and cautious comparisons should be made interpreting measurements obtained using different kits.19 We found a poor eradication rate with the triple therapy. However, this finding was not so different from that recently reported in a meta-analysis in which triple therapies lasting 7 days had an eradication rate of 72% (95% CI, 70.5%–75%).20 Indeed, the 95% CIs of our eradication rate overlap with those reported in the meta-analysis (P ⫽ .10 for the difference). In conclusion, a decrease in sPGII levels of 22.7% or greater 8 weeks after H pylori therapy reliably correlated with H pylori eradication in this study. Subsequent studies are needed to confirm these results. References 1. Vakil N, Megraud F. Eradication therapy for Helicobacter pylori. Gastroenterology 2007;133:985–1001. 2. Graham DY, Dore MP. Perturbations in gastric physiology in Helicobacter pylori duodenal ulcer: are they all epiphenomena? Helicobacter 1997;2(Suppl 1):S44 –S49. 3. Di Mario F, Cavallaro LG, Moussa AM, et al. Usefulness of serum pepsinogens in Helicobacter pylori chronic gastritis: relationship with inflammation, activity, and density of the bacterium. Dig Dis Sci 2006;51:1791–1795. 4. Daugherty D, Yamada T. Posttranslational processing of gastrin. Physiol Rev 1989;69:482–502.
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 5
5. Hunter FM, Correa P, Fontham E, et al. Serum pepsinogens as markers of response to therapy for Helicobacter pylori gastritis. Dig Dis Sci 1993;38:2081–2086. 6. Plebani M, Basso D, Scrigner M, et al. Serum pepsinogen C: a useful marker of Helicobacter pylori eradication? J Clin Lab Anal 1996;10:1–5. 7. Pilotto A, Franceschi M, Leandro G, et al. The clinical usefulness of serum pepsinogens, specific IgG anti-HP antibodies and gastrin for monitoring Helicobacter pylori treatment in older people. J Am Geriatr Soc 1996;44:665– 670. 8. Furuta T, Kaneko E, Baba S, et al. Percentage changes in serum pepsinogens are useful as indices of eradication of Helicobacter pylori. Am J Gastroenterol 1997;92:84 – 88. 9. Al-Assi MT, Miki K, Walsh JH, et al. Noninvasive evaluation of Helicobacter pylori therapy: role of fasting or postprandial gastrin, pepsinogen I, pepsinogen II, or serum IgG antibodies. Am J Gastroenterol 1999;94:2367–2372. 10. Perez-Paramo M, Albillos A, Calleja JL, et al. Changes in gastrin and serum pepsinogens in monitoring of Helicobacter pylori response to therapy. Dig Dis Sci 1997;42:1734 –1740. 11. Gisbert JP, Boixeda D, Vila T, et al. Verification of decreased basal and stimulated serum pepsinogen-I levels is a useful non-invasive method for determining the success of eradication therapy for Helicobacter pylori. Scand J Gastroenterol 1996;31:103–110. 12. Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Ann Intern Med 2003;138:40 – 44. 13. Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disorders. Gastroenterology 2006;130:1466 –1479. 14. Mulherin SA, Miller WC. Spectrum bias or spectrum effect? Subgroup variation in diagnostic test evaluation. Ann Intern Med 2002;137:598 – 602. 15. Asaka M, Kimura T, Kudo M, et al. Relationship of Helicobacter pylori to serum pepsinogens in an asymptomatic Japanese population. Gastroenterology 1992;102:760 –766. 16. Plebani M, Basso D, Cassaro M, et al. Helicobacter pylori serology in patients with chronic gastritis. Am J Gastroenterol 1996; 91:954 –958. 17. Moss SF, Legon S, Bishop AE, et al. Effect of Helicobacter pylori on gastric somatostatin in duodenal ulcer disease. Lancet 1992; 340:930 –932. 18. Graham DY, Lew GM, Lechago J, et al. G-cell and D-cell numbers in Helicobacter pylori infection: effect of H. pylori eradication. Gastroenterology 1993;104:1655–1660. 19. Kazumasa M, Mitsuhiro F. Cautious comparison between East and West is necessary in terms of the serum pepsinogen test. Dig Endosc 2009;21:134 –135. 20. Fuccio L, Minardi ME, Zagari RM, et al. Meta-analysis: duration of first-line proton-pump inhibitor based triple therapy for Helicobacter pylori eradication. Ann Intern Med 2007;147: 553–562.
Reprint requests Address requests for reprints to: Luigi Gatta, MD, Section of Gastroenterology, Department of Clinical Science, University of Parma, Via Gramsci 14, 43100, Parma, Italy. e-mail: [email protected]. Conflicts of interest These authors disclose the following: Francesco Di Mario received consulting fees from and was a board member of AstraZeneca; Biohit, Inc; Giuliani SpA; Janssen-Cilag; Malesci Spa, Italy; Novartis Consumer Health, Inc; and Sofar Spa, Italy. Dino Vaira received stock options (other than mutual funds) in Meridian Bioscience. The remaining authors disclose no conflicts.
BRIEF COMMUNICATIONS Quantification of Serum Levels of Pepsinogens and Gastrin to Assess Eradication of Helicobacter Pylori LUIGI GATTA,*,‡ FRANCESCO DI MARIO,‡ DINO VAIRA,§ MASSIMO RUGGE,储 ANGELO FRANZÈ,¶ MARIO PLEBANI,# GIULIA MARTINA CAVESTRO,‡ PAOLO LUCARINI,* MAURIZIO LERA,* and CARMELO SCARPIGNATO‡ *Gastroenterology and Endoscopy Unit, Versilia Hospital, Lido di Camaiore; ‡Section of Gastroenterology, Department of Clinical Sciences, University of Parma, Parma; § Department of Internal Medicine and Gastroenterology, University of Bologna, Bologna; 储Department of Medical Diagnostic Sciences and Special Therapies (Pathology Unit), University of Padova, Padova; ¶Gastroenterology Unit, Hospital of Parma, Parma; #Department of Laboratory Medicine, University Hospital of Padova, Padova, Italy
See editorial on page 373. BACKGROUND & AIMS: We investigated whether serum levels of pepsinogen (sPG)I and sPGII, the ratio of sPGI to sPGII, or serum levels of gastrin-17 (sG17), can be used to assess eradication of Helicobacter Pylori 8 weeks after treatment. METHODS: We performed a prospective study of 228 consecutive patients with H pylori infections. At the start of the trial (baseline), patients were assessed using the 13C-urea breath test (13C-UBT) and endoscopy, and serum levels of pepsinogens and gastrin levels were measured. Patients were offered a 7-day triple therapy and asked to return 8 weeks after treatment for another 13C-UBT and measurements of serum levels of sG17, sPGI, and sPGII (175 patients completed the study). RESULTS: The eradication rate of H pylori was 67%. Percentage variation in levels of sPGI and sPGII, the ratio of sPGI to sPGII, and in levels of sG17 resulted in area under the curve values of 0.858, 0.973, 0.940, and 0.810, respectively, for H pylori eradication. A decrease of 22.7% or greater in the level of sPGII detected H pylori eradication with 100% sensitivity and 96.6% specificity. Spectrum analysis did not identify differences in accuracy. CONCLUSIONS: Percentage variation of sPGII levels 8 weeks after therapy for H pylori infection correlates with eradication. Additional studies are needed to confirm these results. Keywords: Helicobacter pylori; Dyspepsia; Treatment; Follow-up.
H
elicobacter pylori is the principal cause of gastritis and peptic ulcer disease, and has been associated with gastric cancer.1 H pylori infection is associated with reversible modifications in gastric physiology such as an increase in serum pepsinogen levels and gastrin secretion.2,3 Human pepsinogens are subgrouped into 2 types: pepsinogen I exclusively produced by the chief and mucous neck cells of the corpus, and pepsinogen II produced not only in those cells but also in the cardiac, pyloric, and Brunner gland cell.3 Gastrin is a hormone produced predominantly by G cells located in the gastric antrum and to a lesser extent in proximal duodenal mucosa. Two main forms of gastrin are present in blood: gastrin-17 and gastrin-34. Antral mucosa contains predominantly gastrin-17, whereas gastrin-34 is the major molecular form in the duodenum.4 Several studies have tried to investigate whether serum pepsinogens or gastrin can be used to follow-up H pylori infection after treatment, but results are not unequivocal.5–11
The aim of the present study was therefore to assess the use of serum pepsinogen I (sPGI), serum pepsinogen II (sPGII), their ratio (sPGI/sPGII), and serum gastrin-17 (sG17) to diagnose H pylori eradication 8 weeks after the end of treatment.
Materials and Methods This was a prospective cohort study in which different diagnostic tests were compared in a blinded fashion, designed to fulfill the Standards for Reporting of Diagnostic Accuracy (STARD) recommendations.12 Between January 2007 and June 2008, there were 228 consecutive H pylori–infected dyspeptic outpatients,13 aged 18 and older, who underwent an upper endoscopy, and were asked to participate. Patients were excluded from the study if they had been treated previously for H pylori infection or were taking proton pump inhibitors, H2-receptor antagonists, antibiotics, bismuth compounds in the 4 weeks preceding the initial visit, if they were bleeding, or if blood samples were not available for analysis. The study was performed according to good clinical practice and the Declaration of Helsinki and informed consent was obtained from each patient. H pylori status was assessed using histology (specimens collected were stained with H&E and Giemsa) and the 13C-urea breath test (13C-UBT). Patients were classified as infected if both tests were positive. The determination of sPGs and sG17 was made by enzyme-linked immunosorbent assay (Biohit, Helsinki, Finland) in Parma. Biopsies, breath samples, and blood samples were coded differently so that pathologists and technicians were unaware who was the patient tested. Serum analysis was not paired but was performed at 2 separate times (before and after therapy). All recruited patients were given 7-day triple therapy. Patients with peptic ulcer disease also were given a further 4 weeks of proton pump inhibitors. Patients were asked to return 8 weeks after stopping treatment to evaluate their H pylori status using a 13C-UBT (reference standard for the assessment of H pylori eradication) and to collect a blood sample.
Statistical Analysis Because data were not distributed normally, values of sPGs and sG17 are expressed as a median with interquartile Abbreviations used in this paper: CI, confidence interval; 13C-UBT, breath test; IQR, interquartile range; LR, likelihood ratio; PUD, peptic ulcer disease; sG17, serum gastrin-17; sPG, serum pepsinogen; STARD, Standards for Reporting of Diagnostic Accuracy. © 2011 by the AGA Institute 1542-3565/$36.00 doi:10.1016/j.cgh.2010.12.009 13C-urea
May 2011
SERUM PEPSINOGENS AND GASTRIN
Table 1. Characteristics of the Patients Enrolled in the Study Median age, y (IQR) Sex, male/female Alcohol consumption Smokers Duodenal ulcer Gastric ulcer sPGI, g/L (IQR) sPGII, g/L (IQR) Ratio (IQR) sG17 (IQR)
47 (36–56) 102/113 68.84% 53.02% 19.53% 23.26% 75 (58–93) 17 (13–23) 4.3 (2.5–6.5) 13 pg/L (9–17)
ranges (IQRs). The Wilcoxon matched pair test was used when necessary. By using the percentage change in values before and after treatment, a receiver operating characteristic analysis was performed applying nonparametric methods against the predefined reference standard (13C-UBT). Sensitivity, specificity, and the likelihood ratio for a positive (LR-positive) and negative (LR-negative) test were calculated. The spectrum effect14 was evaluated for the following factors: (1) age (dichotomized as ⱕ50 or ⬎50 y), (2) sex, (3) peptic ulcer disease (dichotomized as present or absent), (4) smoking (dichotomized as smokers or nonsmokers), and (5) consumption of alcohol (dichotomized as ⱖ1 glass/day or no consumption). These criteria were determined before commencement of the study. A 2-sided P value of less than .05 for all other tests indicated statistical significance.
Results A total of 228 consecutive H pylori–infected patients were offered entry into the study. Thirteen patients were not included because of exclusion criteria, and 215 patients entered the study. Demographic features and endoscopic findings of the population enrolled are shown in Table 1. Forty patients were lost to follow-up evaluation and 175 patients (men/women, 82/93; median age, 46 y; IQR, 32–53 y) completed the study, with an eradication rate of 67% (95% confidence interval [CI], 60–74). The sPGI level decreased significantly in eradicated patients (from 76 g/L [IQR, 55– 88 g/L] to 62 g/L [IQR, 48.0–77.2 g/L]; P ⬍ .0001). In noneradicated patients, values remained
441
virtually the same (from 80 g/L [IQR, 67–95 g/L] to 86 g/L [IQR, 75–99 g/L]; P ⬎ .1). The sPGII level also significantly decreased in eradicated patients (from 17 g/L [IQR, 13–23 g/L] to 8 g/L [IQR, 7–9 g/L]; P ⬍ .0001). In those not eradicated, values remained stable (from 16 g/L [IQR, 12–24 g/L] to 16 g/L [IQR, 14 –21 g/L]; P ⫽ .705). As a consequence, the sPGI/sPGII level increased in eradicated patients (from 4.2 g/L [IQR, 2.3– 6.2 g/L] to 8.2 g/L [IQR, 5.6 –10.7 g/L]; P ⫽ .0001). In noneradicated patients, values were almost the same (from 4.7 g/L [IQR, 2.8 – 6.8 g/L] to 5 g/L [IQR, 3.5– 6.8 g/L]; P ⫽ .921). Finally, the sG17 level decreased significantly in those eradicated (from 13 pg/L [IQR, 8.5–16 pg/L] to 10 pg/L [IQR, 8 –14 pg/L]; P ⬍ .0001). In those not eradicated, values were not significantly different (from 13 pg/L [IQR, 8.5–16 pg/L] to 13 pg/L [IQR, 10 –18 pg/L]; P ⫽ .1). Receiver operating characteristic analysis established the following regarding a decrease in values: (1) sPGI greater than 6.8% as the cut-off point would result in an AUC of 0.858 (95% CI, 0.798 – 0.906); (2) sPGII greater than 22.7% as the cut-off point would result in an AUC of 0.973 (95% CI, 0.936 – 0.991); (3) an increment in values of sPGI/sPGI of 33.8% or greater as the cut-off point would result in an AUC of 0.940 (95% CI, 0.894 – 0.970); (4) a decrease in values of sG17 greater than 17.6% as the cut-off point would result in an AUC of 0.810 (95% CI, 0.744 – 0.866). The value of AUC achieved with sPGII was significantly higher than that obtained with sPGI (P ⬍ .001), ratio (P ⫽ .042), and sG17 (P ⬍ .001). Sensitivity, specificity, LR-positive, and LR-negative of sPGII were 100% (95% CI, 96.8 –100), 96.6% (95% CI, 91.5–99.1), 29.5% (95% CI, 8.6 –107), and 0% (95% CI, 0 – 0.03), respectively. Spectrum effect analysis performed for sPGII shown that there was no significant difference in terms of sensitivity, specificity, LR-positive and LR-negative with regard to age, sex, peptic ulcer disease (PUD), smoking, and alcohol consumption (Table 2).
Discussion We found that percentage change in values before and after treatment of sPGII could provide a good diagnostic accuracy in evaluating the effect of H pylori treatment 8 weeks after the end of treatment when compared with the 13C-UBT. There have been previous attempts to establish whether changes in
Table 2. Spectrum Effect for Percentage Variation in sPGII Considering Sex, Age, PUD, Smoking, and Alcohol to Assess H pylori Eradication Parameters
Sensitivity (95% CI)
Male Female Age, ⬍50 y Age, ⱖ50 y PUD No PUD Smokers Nonsmokers Alcohol No alcohol
100% (93.6–100) 100% (94–100) 100% (94.3–100) 100% (93.2–100) 100% (91.4–100) 100% (95.1–100) 100% (93.8–100) 100% (93.8–100) 100% (95.1–100) 100% (91.4–100)
aP bP
P valuea 1 1 1 1 1
Specificity (95% CI) 96.2% (81.1–99.3) 97% (84.7–99.5) 97.6% (87.4–99.6) 94.4% (74.2–99) 96.8% (83.8–99.4) 96.4% (82.3–99.4) 94.3% (81.4–98.4) 100% (86.2–100) 97.8% (88.7–99.6) 92.3% (66.7–98.6)
value for comparison of sensitivities or specificities in the 2 strata. value for homogeneity test in the 2 strata.
P valuea 1 .5 1 .5 .3
LR-positive (95% CI) 26 (5.2–146.5) 33 (⬁ to ⬁) 41 (7.9–231.5) 18 (3.8–101.2) 31 (⬁ to ⬁) 28 (5.6–157.9) 17.5 (5.3–63.2) — 46 (⬁ to ⬁) 13 (3–72.9)
P valueb .837 .475 .934 — .262
LR-negative (95% CI) 0 (0–0.06) 0 (0–⬁) 0 (0–0.05) 0 (0–0.07) 0 (0–⬁) 0 (0–0.05) 0 (0–0.06) — 0 (0–⬁) 0 (0–0.09)
P valueb .968 .911 .767 — .719
442
GATTA ET AL
serum pepsinogens or gastrin levels could be used to assess the effects of antibiotic therapy with contrasting results.5–11 Several pathophysiological factors may explain why sPGII, among others, could be a reliable indicator to perform a follow-up evaluation of patients after treatment. sPGII increases more than sPGI in H pylori–infected patients,15,16 and sPGII correlates proportionally according to the severity of H pylori gastric inflammation (including both neutrophilic and lymphomononuclear infiltration) in antrum and corpus better than sPGI.3 H pylori also might influence sPGII by interfering with the release of gastrin and somatostatin.17,18 Finally, H pylori–associated gastritis is localized mainly in the antrum, and this could affect sPGII more than sPGI.3 Our analysis presents several strengths compared with the previous studies: it was designed to fulfill the STARD recommendations,12 work-up bias and expectation bias have been avoided because all patients who had pepsinogens and gastrin blood tests also had histology and the UBT, and those who interpreted the results were not aware of the results of all other tests performed. Finally, a spectrum analysis including 5 variables (sex, age, PUD, smoking, and alcohol consumption) was performed. However, our study presents some drawbacks. First, the identified cut-off points need to be tested in future prospective studies to assess the validity, because the results were dataderived. Second, it was possible to perform the spectrum effect analysis only evaluating 5 variables; however, it could not be excluded that other factors (eg, body mass index, previous or concomitant use of nonsteroidal anti-inflammatory drugs, antisecretory drugs, food, and so forth) could affect the performance of the test. Third, the eradication status was evaluated 8 weeks after the end of treatment; data also are needed for a shorter follow-up period (eg, 4 weeks). We also could not perform a measurement of the economic impact of having this type of noninvasive diagnosis. We also did not perform a matched comparison with the radioimmunoassay. Finally, it is useful to remember that different enzyme-linked immunosorbent assay kits can have different calibrations, and cautious comparisons should be made interpreting measurements obtained using different kits.19 We found a poor eradication rate with the triple therapy. However, this finding was not so different from that recently reported in a meta-analysis in which triple therapies lasting 7 days had an eradication rate of 72% (95% CI, 70.5%–75%).20 Indeed, the 95% CIs of our eradication rate overlap with those reported in the meta-analysis (P ⫽ .10 for the difference). In conclusion, a decrease in sPGII levels of 22.7% or greater 8 weeks after H pylori therapy reliably correlated with H pylori eradication in this study. Subsequent studies are needed to confirm these results. References 1. Vakil N, Megraud F. Eradication therapy for Helicobacter pylori. Gastroenterology 2007;133:985–1001. 2. Graham DY, Dore MP. Perturbations in gastric physiology in Helicobacter pylori duodenal ulcer: are they all epiphenomena? Helicobacter 1997;2(Suppl 1):S44 –S49. 3. Di Mario F, Cavallaro LG, Moussa AM, et al. Usefulness of serum pepsinogens in Helicobacter pylori chronic gastritis: relationship with inflammation, activity, and density of the bacterium. Dig Dis Sci 2006;51:1791–1795. 4. Daugherty D, Yamada T. Posttranslational processing of gastrin. Physiol Rev 1989;69:482–502.
CLINICAL GASTROENTEROLOGY AND HEPATOLOGY Vol. 9, No. 5
5. Hunter FM, Correa P, Fontham E, et al. Serum pepsinogens as markers of response to therapy for Helicobacter pylori gastritis. Dig Dis Sci 1993;38:2081–2086. 6. Plebani M, Basso D, Scrigner M, et al. Serum pepsinogen C: a useful marker of Helicobacter pylori eradication? J Clin Lab Anal 1996;10:1–5. 7. Pilotto A, Franceschi M, Leandro G, et al. The clinical usefulness of serum pepsinogens, specific IgG anti-HP antibodies and gastrin for monitoring Helicobacter pylori treatment in older people. J Am Geriatr Soc 1996;44:665– 670. 8. Furuta T, Kaneko E, Baba S, et al. Percentage changes in serum pepsinogens are useful as indices of eradication of Helicobacter pylori. Am J Gastroenterol 1997;92:84 – 88. 9. Al-Assi MT, Miki K, Walsh JH, et al. Noninvasive evaluation of Helicobacter pylori therapy: role of fasting or postprandial gastrin, pepsinogen I, pepsinogen II, or serum IgG antibodies. Am J Gastroenterol 1999;94:2367–2372. 10. Perez-Paramo M, Albillos A, Calleja JL, et al. Changes in gastrin and serum pepsinogens in monitoring of Helicobacter pylori response to therapy. Dig Dis Sci 1997;42:1734 –1740. 11. Gisbert JP, Boixeda D, Vila T, et al. Verification of decreased basal and stimulated serum pepsinogen-I levels is a useful non-invasive method for determining the success of eradication therapy for Helicobacter pylori. Scand J Gastroenterol 1996;31:103–110. 12. Bossuyt PM, Reitsma JB, Bruns DE, et al. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD initiative. Ann Intern Med 2003;138:40 – 44. 13. Tack J, Talley NJ, Camilleri M, et al. Functional gastroduodenal disorders. Gastroenterology 2006;130:1466 –1479. 14. Mulherin SA, Miller WC. Spectrum bias or spectrum effect? Subgroup variation in diagnostic test evaluation. Ann Intern Med 2002;137:598 – 602. 15. Asaka M, Kimura T, Kudo M, et al. Relationship of Helicobacter pylori to serum pepsinogens in an asymptomatic Japanese population. Gastroenterology 1992;102:760 –766. 16. Plebani M, Basso D, Cassaro M, et al. Helicobacter pylori serology in patients with chronic gastritis. Am J Gastroenterol 1996; 91:954 –958. 17. Moss SF, Legon S, Bishop AE, et al. Effect of Helicobacter pylori on gastric somatostatin in duodenal ulcer disease. Lancet 1992; 340:930 –932. 18. Graham DY, Lew GM, Lechago J, et al. G-cell and D-cell numbers in Helicobacter pylori infection: effect of H. pylori eradication. Gastroenterology 1993;104:1655–1660. 19. Kazumasa M, Mitsuhiro F. Cautious comparison between East and West is necessary in terms of the serum pepsinogen test. Dig Endosc 2009;21:134 –135. 20. Fuccio L, Minardi ME, Zagari RM, et al. Meta-analysis: duration of first-line proton-pump inhibitor based triple therapy for Helicobacter pylori eradication. Ann Intern Med 2007;147: 553–562.
Reprint requests Address requests for reprints to: Luigi Gatta, MD, Section of Gastroenterology, Department of Clinical Science, University of Parma, Via Gramsci 14, 43100, Parma, Italy. e-mail: [email protected]. Conflicts of interest These authors disclose the following: Francesco Di Mario received consulting fees from and was a board member of AstraZeneca; Biohit, Inc; Giuliani SpA; Janssen-Cilag; Malesci Spa, Italy; Novartis Consumer Health, Inc; and Sofar Spa, Italy. Dino Vaira received stock options (other than mutual funds) in Meridian Bioscience. The remaining authors disclose no conflicts.