Diagnosis of Helicobacter pylori infection in the elderly using an immunochromatographic assay‐based stool antigen test

Abstract The diagnostic value of Helicobacter pylori stool antigen (HpSA) tests in elderly subjects remains unclear. The objective of this study was to assess the diagnostic accuracy of the immunochromatographic assay‐based HpSA test in a male elderly cohort and identify factors affecting the accuracy. Data for asymptomatic elderly male citizens (≥65 years old) who received health checkups at the Chinese PLA General Hospital between July 2007 and November 2018 were collected. The diagnostic accuracy of the HpSA test was determined using the 13C‐urea breath test as a reference standard. Associations between baseline comorbidities and the accuracy of the HpSA test were analyzed. In total, 316 participants were enrolled, including 193 in the pre‐treatment group (77.2 ± 7.8 years old) and 123 in the post‐treatment group (78.7 ± 8.3 years old). The accuracy (91.5%, 91.2%, and 91.9%) and specificity (97.6%, 98.7%, and 96.0%) were high in all participants, pre‐ and post‐treatment groups, respectively. However, sensitivities were only 68.7%, 65.1%, and 75.0%, respectively. In the pre‐treatment group, constipation was associated with decreased sensitivity (p = 0.039), while colorectal polyps were associated with increased sensitivity (p = 0.010). Multivariate analysis indicated that constipation and colorectal polyps are independent factors for the sensitivity of HpSA in the pre‐treatment group. The immunochromatographic assay‐based HpSA test achieved high accuracy with high specificity but suboptimal sensitivity in the elderly male cohort. Constipation and colorectal polyps were negatively and positively associated with HpSA sensitivity, respectively, in the pre‐treatment group.


| INTRODUC TI ON
Helicobacter pylori infection is considered an infectious disease, regardless of symptoms and the stage of the disease (Sugano, Tack, & Kuipers, 2015). Along with increasing age, the prevalence of H. pylori infection is elevated in developing countries (Bardhan, 1997). The reliable diagnosis of H. pylori infection is of utmost importance for identifying the source of infection, preventing complications related to chronic H. pylori infection, and monitoring the treatment response after H. pylori eradication.
Invasive tests, such as histopathology, H. pylori culture, rapid urease tests, and modern molecular tests (e.g., real-time quantitative PCR techniques), require gastroscopy with gastric mucosa biopsies, may need specialized laboratory facilities, and are time-consuming. Thus, researches have focused on noninvasive methods, such as the urea breath test (UBT), H. pylori stool antigen (HpSA) test, and serological assays. UBT is capable of identifying active H. pylori infections and is the most widely studied and preferentially recommended a noninvasive approach for the "test-and-treat strategy" (Malfertheiner et al., 2017). The 13 C-UBT is the best approach for the detection of H. pylori infection, with outstanding sensitivity, specificity, and performance (Gisbert & Calvet, 2013;Gisbert & Pajares, 2004a).
However, the high price and the need for skilled technical staff and complicated instruments limit the application of UBT in clinical practice. As H. pylori antibodies may remain positive for several months or longer after the eradication of bacteria, it is difficult to distinguish between current and past infections using serologic tests (Bergey, Marchildon, Peacock, & Megraud, 2003).
The HpSA test detects bacterial antigens and thus can diagnose active infections. It is easy to perform, especially for pediatric and geriatric patients, those with asthma, after gastrectomy, or in the case of achlorhydria, those in which breath test results are unreliable (Yang & Seo, 2008). It is a noninvasive alternative to UBT (Korkmaz, Kesli, & Karabagli, 2013). Previous HpSA tests with poly-/monoclonal antibodies have shown a sensitivity of 0.83 at a fixed specificity of 0.9 and a ratio of diagnostic odds ratios of 0.88 for the 13 C-UBT versus the stool antigen test (Best et al., 2018). The HpSA test can be organized into three groups: immunochromatographic assays (ICA), enzymatic immunoassays (EIA), and immunodot blot assays.
H. pylori stool antigens can be easily and rapidly detected using the ICA-based HpSA test, with reported sensitivity and specificity values exceeding 90% both before and after H. pylori treatment (Gatta et al., 2004). There is no significant difference in diagnostic accuracy between ICA-based tests and EIA-based tests in children (Yang & Seo, 2008).
The diagnostic value of the HpSA test in elderly patients remains unclear. Only a few reports involving small sample sizes have evaluated HpSA tests in these patients (Inelmen et al., 2004;Kamel et al., 2011;Salles-Montaudon, Dertheil, & Broutet, 2001. The objective of this study was to evaluate the sensitivity, specificity, positive (PPV) and negative predictive values (NPV), and diagnostic accuracy of the ICA-based HpSA test in an elderly male cohort using the 13 C-UBT as a reference standard. As elderly individuals often have concurrent chronic diseases, we adjusted their baseline comorbidities to investigate the factors related to the accuracy of ICAbased HpSA tests in the study population.

| Participants
Clinical data for elderly male citizens (age ≥65 years) who underwent health checks at the Chinese PLA General Hospital between July 2007 and November 2018 were collected. All participants received the 13 C-UBT examination and ICA-based HpSA test. Stool samples were obtained for the HpSA test, which was performed on the same day or no longer than 1 week before or after the 13 C-UBT. Subjects who took antibiotics, proton-pump inhibitors, H 2 receptor antagonists, or bismuth within recent 4 weeks of the tests were excluded.
Clinical data for concurrent drug use and chronic diseases that may affect the accuracy of tests, such as atrophic gastritis, constipation, colon diverticulum, and diabetes mellitus, were recorded. The history of anti-H. pylori treatment (triple or quadruple regimens) was also collected. Subjects with no history of anti-H. pylori treatment before 13 C-UBT and HpSA tests were regarded as the pre-treatment group. Those who were tested after anti-H. pylori treatment were

| 13 C-UBT detection
Helicobacter pylori infection was determined via the 13 C-UBT. After fasting for over 8 hr, each subject drank a solution containing 75 mg of 13 C-urea in 70 ml of water. Breath samples were collected before and 30 min after the ingurgitation of the water. Then, 13 C-enrichment was detected using a 13 C-breath test instrument (Fischer Analysen Instrumente GmbH). The results were defined as positive when the delta over baseline (DOB) was >4‰, calculated as the surplus of the isotopic ratio over the baseline isotopic ratio.

| Statistical analyses
Statistical analyses were performed using Statistical Package for Social Sciences version 25.0 (SPSS, Chicago, IL, USA). Sensitivity, specificity, PPV, and NPV with 95% confidence intervals (CI) were calculated by standard methods using 13 C-UBT as the reference standard. Continuous variables are expressed as means ± standard deviation (SD). Categorical variables are expressed as n (%). The chisquare test was used to detect differences within categorical variables. A univariate analysis was performed for all variables, including age, comorbidities, and medications, via chi-square tests. A multivariate analysis was performed to determine independent factors for diagnostic efficiency using a binary logistic regression model. A p-value of less than 0.05 (two-sided) was regarded as statistically significant.

| Demographic characteristics of participants
A total of 316 participants who underwent both 13 C-UBT and HpSA tests were enrolled. Among them, 193 subjects were assigned to the pre-treatment group and 123 subjects were assigned to the posttreatment group. The mean ages of all participants and those in the pre-treatment and post-treatment groups were 77.8 ± 8.0 years, 77.2 ± 7.8 years, and 78.7 ± 8.3 years, respectively. The positive rate for 13 C-UBT was 21.1% and that for HpSA was 16.5% (Table 1 and   Table A1 in the Appendix A). Comorbidities in each group, such as atrophic gastritis, constipation, colorectal polyps, diabetes, hyperlipidemia, hypertension, and dementia, are listed in Table 1 Table A2 in the Appendix A.

| Diagnostic efficacy of HpSA test
The performance of the HpSA test using 13 C-UBT as a reference standard was analyzed. The median time between 13 C-UBT and

TA B L E 1 Demographic characteristics of all participants
HpSA tests was 2 days (interquartile range 1-4 days). As shown in Figure 1 and

| Factors associated with HpSA sensitivity
We further investigated factors (comorbidities) that affect the sensitivity of the HpSA test. Univariate analysis indicated that the sensitivity of the HpSA test was significantly higher for participants over 78 years old than for younger participants (54.8% vs. 80.6%, p = 0.024), significantly lower in those with constipation than in those without (76.5% vs. 43.8%, p = 0.014), and higher in those with colorectal polyps than in those without (56.8% vs. 83.3%, p = 0.020).
In the pre-treatment group, the sensitivity of the HpSA test was significantly lower in participants with constipation than in those without constipation (76.7% vs. 38.5%, p = 0.039). The sensitivity was significantly higher in participants with colorectal polyps than those without (45.0% vs. 82.6%, p = 0.010). Differences in these parameters were not observed in the post-treatment group (Table 2).
Other comorbidities, such as colon diverticulum, a history of bowel surgery, diabetes, and hyperlipidemia, were not significantly correlated with HpSA sensitivity in this cohort. No medications were significantly associated with HpSA sensitivity in all groups (Table A6 in Appendix A).
To identify the most important covariate for HpSA sensitivity, various factors including age, constipation, and colorectal polyps were subjected to multivariate regression analysis. Both constipation and colorectal polyps were independent factors for the sensitivity of the HpSA test in all participants and the pre-treatment group.
All summary statistics are summarized in Table 3.
A subgroup analysis showed that the accuracy of the HpSA test in patients with constipation was lower than that in patients with-  . We did not find an association between polyps and constipation in the pre-treatment group (Table A10 in Appendix A). Large prospective studies are needed for further investigation of the association between colorectal polyps and HpSA sensitivity in the elderly.

| D ISCUSS I ON
Low sensitivity of the polyclonal HpSA test in the post-treatment setting has been reported (Gisbert & Pajares, 2004b).
However, there is evidence that the HpSA test using monoclonal antibodies shows superior sensitivity to those of tests using polyclonal antibodies, particularly in the post-treatment setting (Gisbert et al., 2006). We achieved an accuracy rate exceeding 91% in both pre-and post-treatment groups by using ICA, consistent with previous reports (Vaira et al., 2000). Our stringent inclusion criteria may explain the high accuracy obtained in both the preand post-treatment groups. In this study, factors with the potential to affect the accuracy of 13 C-UBT or HpSA, such as proton-pump inhibitors, antibiotics, bismuth therapy (Calvet et al., 2002;Gisbert & Pajares, 2001;Grino et al., 2003;Inelmen et al., 2005), a history of gastrectomy, and overt gastrointestinal bleeding, were excluded. Cases with "indeterminate results" were also excluded.
Furthermore, the 13 C-UBT and HpSA tests were performed on the same day or with an interval of no longer than 7 days to minimize the error caused by variation in detection times. Finally, we recruited only asymptomatic elderly subjects who underwent the tests for health checkup purposes to minimize the influence of active and/or severe diseases.
There were several limitations to this study. First, 13 C-UBT, believed to be an ideal noninvasive assay, was chosen as the only reference standard (Best et al., 2018). According to the literature, the false-negative rate of 13 C-UBT could be elevated in elderly individuals (Salles-Montaudon et al., 2001). A combination of invasive tests, such as histological or culture data, would be a more effective reference standard. Also, the sample size was relatively small and was limited to male subjects. Sample size calculation was performed based on the following settings: the prevalence of H. pylori infection (around 25%) in the cohort, previously reported sensitivity (76%) and specificity (93%) values for the HpSA test in the elderly (Inelmen et al., 2005), and a two-sided α level of 0.05. We noticed that only the whole study cohort matched the sample size requirement. Although no study has reported a gender difference in noninvasive detection efficiency, further studies of both male and female subjects will broaden our knowledge in this field. Moreover, participants were in a relatively higher-than-average socioeconomic status, as evidenced by their utilization of a regular health check with a low prevalence (22.3%) of H. pylori infection, compared with a reported prevalence in Beijing, China, in the general population of as high as 47.0% (Hooi et al., 2017). Hence, the generalizability of the results of this study to the whole elderly population should be performed with caution. Lastly, we cannot confirm the causality between the parameters identified, such as colorectal polyps, and HpSA sensitivity, as this is an observational study.

| CON CLUS IONS
In an observational study of an elderly male cohort, we revealed that HpSA achieves high accuracy and specificity but suboptimal sensitivity in both pre-and post-treatment groups when using 13 C-UBT as a reference

CO N FLI C T O F I NTE R E S T
None declared.