Evaluation of Commercial Immunoassays for Detection of Antibody against Helicobacter pylori in Thai Dyspeptic Patients

Abstract

The performance of five immunoassays for detection of immunoglobulin G antibody against Helicobacter pylori in 191 dyspeptic patients was evaluated. The sensitivities, specificities, accuracies, positive predictive values, and negative predictive values ranged from 86.32 to 97.89%, 57.95 to 72.22%, 77.02 to 83.76%, 71.54 to 77.42%, and 83.33 to 96.23%, respectively. The immunoglobulin A test kit also gave a high sensitivity and negative predictive value (95.79 and 91.40%, respectively), while the specificity was relatively low (51.14%).

Serological assay for Helicobacter pylori antibody is a noninvasive method to detect H. pylori infection. It has been reported to have sensitivity and specificity in predicting the status of H. pylori infection in untreated patients as accurately as invasive tests (11, 12). However, it has been suggested that serological tests for H. pylori should be locally validated (7), because assays validated in one region may yield variable diagnostic performances in others. These variations may be attributed to many factors, including the source of antigen used, the prevalence of infection in each population studied, and the reference method used to determine true H. pylori infection status. Therefore, reevaluation is needed before implementing a test in different populations. In Thailand, the seroprevalence of H. pylori infection has been reported to be higher than that in industrialized countries (10), and commercially available enzyme immunoassay (EIA) test kits have been reported to have lower sensitivities and specificities compared to in-house EIAs in Thai dyspeptic patients (1). We therefore evaluated the performance of five commercial test kits for detecting of immunoglobulin G (IgG) antibody to H. pylori. Three of them use a standard enzyme-linked immunosorbent assay (Cobas Core anti-H. pylori EIA [Roche, Mannheim, Germany]; Pyloriset EIA-GIII [Orion, Espoo, Finland]; and Enzygnost anti-H. pylori II/IgG [Dade Behring, Marburg, Germany]), and two are rapid assay test kits (Pyloriset Dry [Orion] and anti-H. pylori IgG Immunocomb [Orgenics, Yavne, Israel]). One H. pylori IgA antibody test kit (Pyloriset EIA-AIII) was also evaluated.

A total of 191 patients (57 males and 134 females; age range, 16 to 83 years [mean, 39 years]) were studied. Endoscopy was performed in all patients, and 183 (95.81%) of them were diagnosed as having nonulcer dyspepsia while the remaining 8 patients (4.19%) had a duodenal ulcer. Patients who received antibiotic therapy, bismuth treatment, or a proton pump inhibitor or H₂ blocker within 1 month prior to the study were excluded. Written informed consent was obtained from all patients before the study. Five milliliters of clotted blood was obtained on the day of endoscopy. Sera were kept at −20°C until analyzed. The biopsy specimens from the antrum and stomach body were obtained for rapid urease (CLO) test and histological and cultural examination. All the tests were performed according to the manufacturer's instructions and without the knowledge of the status of the patient's infection. The results of these examinations were described previously (4).

In this study, a patient was considered infected with H. pylori when either culture was positive or both rapid urease (CLO) test and histological analysis were positive.

Statistic analyses for sensitivity, specificity, accuracy, positive predictive value, and negative predictive value were calculated against the status of H. pylori infection. As shown in Table 1, the three standard EIAs for IgG antibody gave a similar higher sensitivity (95.95 to 97.89%) and negative predictive value (92.06 to 96.23%) when compared to the rapid immunoassays. The specificity of all tests was considered low (57.95 to 69.57%), while the accuracy was similar, with the highest at 83.76% by Cobas Core anti-H. pylori EIA. The agreement between each test as analyzed by kappa statistic was relatively high among the standard immunoassays. Pyloriset Dry gave the lowest agreement with other tests, especially with Immunocomb (Table 2).

TABLE 1.

Performance of immunoassays for the detection of antibody to H. pylori

Kit	No. tested	No. of H. pylori-positive (No. of true positive)a	No. of H. pylori-negative (no. of true negative)a	Sensitivity (%)	Specificity (%)	Accuracy (%)	Positive predictive value (%)	Negative predictive value (%)
Cobas Core anti-H. pylori EIA	191	124 (99)	67 (92)	96.97	69.57	83.76	77.42	95.52
Enzygnost anti-H. pylori II/IgG	191	129 (99)	62 (92)	95.95	63.64	80.64	73.44	92.06
Pyloriset EIA-G III	183	130 (95)	53 (88)	97.89	57.95	78.69	71.54	96.23
Anti-H. pylori IgG Immunocomb	161	111 (86)	50 (75)	93.02	58.67	77.02	72.07	88.00
Pyloriset Dry	185	107 (95)	78 (90)	86.32	72.22	79.46	76.64	83.33
Pyloriset EIA-A III	183	134 (95)	49 (88)	95.79	51.14	74.32	67.91	91.40

^aNumbers of true positive and true negative were determined by the results of culture or CLO test and histology.

TABLE 2.

Agreement between test kits by kappa statistic

Kit	Agreementa (kappa) with:
	Enzygnost	Pyloriset EIA-G III	Immunocomb	Pyloriset Dry
Cobas Core anti-H. pylori EIA	0.79	0.78	0.79	0.67
Enzygnost anti-H. pylori II/IgG		0.76	0.71	0.64
Pyloriset EIA-G III			0.77	0.59
Immunocomb				0.55

^aThe degree of agreement between each test kit was significant (P < 0.001) by kappa statistic analysis.

As reported by other investigators who have found that Western antigen-based serology has relatively poor performance with samples from Asian groups (5, 6, 8), we also found a low specificity of these tests in our study. The possible reasons may be due to the high prevalence rate of H. pylori infection in the Thai population (10). Therefore, the presence of antibody in some sera may reflect past infection. Furthermore, the results of validation are highly dependent on the reliability of the reference method used, and it is generally accepted that all the tests for H. pylori have their pitfalls and limitations that may affect the status of infection. In this study, the status of infection depended on the results of culture or histology and rapid urease (CLO) test. We observed that 10 out of 35 seropositive (as demonstrated by at lease two serological tests used in this study) patients in the 92 noninfected groups were concomitantly positive by histology or urease test. Most of them had higher antibody levels than those who were positive by serology alone, as shown in Fig. 1. Therefore, it is possible that some patients of the group may have had H. pylori infection during the study period. Figure 1 also demonstrates the distribution of H. pylori IgG antibody levels in the H. pylori-infected and noninfected patients.

FIG. 1.

Distribution of anti-H. pylori IgG antibody levels in H. pylori-infected and uninfected patients (n = 191). Results of the Cobas Core anti-H. pylori EIA test, with serum antibody levels of <6 U/ml, were considered negative.

The performance of the rapid tests was similar to that reported previously (2) in that they were slightly inferior to the standard EIA tests. Although better results have also been reported (9, 14), the use of rapid tests has not been recommended (7). However, these tests are easy to perform and can be finished within a few minutes without the need of sophisticated equipment. By using these tests and with careful interpretation, the physician can determine H. pylori infection of a patient at the first consultation.

Anti-H. pylori IgA antibody was found in 91 out of 95 (95.79%) H. pylori-infected patients (Table 1). Almost all, except one, were also found to have IgG antibody (by Pyloriset EIA-GIII) in their serum. Combining the results of this Pyloriset EIA-AIII and Pyloriset EIA-GIII slightly increased the sensitivity and the negative predictive value, but the specificity was markedly decreased (data not shown). Although anti-H. pylori IgA antibody has been reported to have diagnostic values when used in conjunction with IgG (3, 13), we agree with a previous report (12) that IgA had no additional diagnostic value in our clinical settings.

With regards to the high sensitivity and negative predictive values of the commercial H. pylori IgG antibody kits used in this study, we concluded that, with careful interpretation, these tests may be used as an alternative test for determining H. pylori infection, especially in those for whom gastroscopy cannot be performed.