Background

Epidemiological studies have revealed a correlation between Helicobacter pylori (H. pylori) infection and some non-gastrointestinal diseases, especially those characterized by persistent and low-grade systemic inflammation [1,2,3]. Idiopathic iron deficiency anemia (IDA) is a well-recognized extragastric manifestation of H. pylori infection and has already been fully accepted and included in the current guidelines for these conditions [1, 2, 4, 5]. After the first report of a young adult whose long-standing IDA was reversed after H. pylori eradication [6], some additional trials indicated that H. pylori infection is associated with an increased likelihood of depleted iron storage, and H. pylori eradication therapy might be beneficial, in terms of increasing ferritin levels [7,8,9,10]. Reported data have supported the effectiveness of H. pylori treatment in patients with moderate to severe anemia when compared to those with mild anemia [11]. However, it should be noted that the most convincing data for the association between H. pylori infection and IDA come from meta-analyses. Meanwhile, there are also reports of conflicting findings on the relationship between H. pylori and iron storage [12,13,14,15]. Shak JR et al. even found that H. pylori sero-positivity was negatively associated with anemia [16].

It has recently been recommended that H. pylori infection be tested and treated in persons with unexplained IDA [5, 17]. The consensus was reached based on studies in which most of the subjects were children, adolescents and middle-age adults. There is no guideline or consensus on H. pylori management in the elderly population until now. Few data on the effects of H. pylori infection on anemia and iron storage in the elderly population were presented. Kaffes et al. observed an association of H. pylori infection with significantly lower ferritin levels only in elderly female aspirin users, which indicated that H. pylori infection may play a role in iron storage in this population [18]. A small survey performed in asymptomatic elderly subjects found that nutritional indices, including hemoglobin, iron, ferritin and transferrin, were not influenced by the presence of anti-HP antibodies [14]. The relationship between anemia and H. pylori infection in the elderly has so far not been established. It is evident that the prevalence of H. pylori infection increases with age in developing countries [19,20,21]. Since medical decision making for the elderly is difficult due to their declining physical and cognitive functions and social support, anemia is associated with an increased mortality risk in persons aged 85 years and older [22], it is meaningful to study if H. pylori infection is associated with anemia in this population.

In this study of an asymptomatic elderly cohort presenting active H. pylori infection, which was identified with the 13C-urea breath test (13C-UBT), we examined if they have a higher prevalence of anemia when compared to their counterparts with no evidence of active H. pylori infection. Elderly people are often presented with chronic diseases, which are major confounders in longitudinal non-randomized studies. When adjusted for comorbidities, researchers may consider comorbidities individually or through the use of summary measures, such as the Charlson Comorbidity Index (CCI). We applied the CCI scores to adjust the baseline comorbidities, in order to remove the possible confounding factors which may cause anemia in this study. We also assessed the effects of H. pylori infection on the level of serum iron, ferritin and C-reactive protein (CRP).

Methods

Study subjects

This study is conducted with a cohort of male senior citizens (age ≥ 65 years) who received an assessment of their health status at the Chinese PLA General Hospital from January 2015 to December 2015. All participants received the 13C-UBT examination. Exclusion criteria included use of antibiotics, bismuth, or proton pump inhibitor (PPI) and H2-receptor antagonist within 4 weeks. On the same day of the 13C-UBT, blood samples were collected for complete blood count, and serum iron and ferritin measurements. Serum CRP level was also tested in individuals who have available serum samples. Stool samples were obtained for occult assessment. A detailed clinical history was taken to record any significant acute or chronic illnesses that may cause anemia, such as neoplasia at advanced stages, hematologic disorders, chronic kidney diseases, gastrectomy, malnutrition, acquired immunosuppression, evident gastrointestinal bleeding, and ongoing inflammatory diseases, etc.. In addition, comorbid diseases of each participant were collected and CCI were applied to establish baseline comorbidities. The CCI is based on a number of conditions that are each assigned an integer weight from one to six, with a weight of six representing the most severe morbidity. The summation of the weighted comorbidity scores results in a summary score [23], which can give clinicians and researchers a single number that captures the individual’s health status.

Patients with evident gastrointestinal cancer, hematologic disorders, who received iron supplementation for at least 30 days, received erythropoietin injection, those with overt/occult gastrointestinal bleeding, with malnutrition, a history of gastrectomy, and those with recent hospitalization due to acute illnesses were excluded. This study was approved by the Ethics Committee of the General Hospital of Chinese PLA.

H. pylori status

H. pylori status was evaluated by the 13C-UBT, which was performed in the morning after fasting for at least 8 h. Breath samples were collected from each subject at baseline and 30 min after drinking 70 mL of room-temperature water containing 75 mg of 13C-urea. An additional breath sample was collected 30 min after the ingestion of the tracer. The test was performed with a 13C-breath test instrument (Fischer Analysen Instrumente GmbH, Leipzig, Germany). The results were expressed as a surplus of isotopic ratio over the baseline isotopic ratio (delta over baseline [DOB]). A DOB value of > 4 ‰ was considered positive, according to the instructions from the test manufacturer.

Clinical tests

Blood counts were determined using an automated electronic counter (Sysmex XN3000, Sysmex Corporation, Kobe, Japan). The serum concentration of ferritin was determined by an electrochemiluminescence immunoassay (Cobas e601, Roche Diagnostics Ltd., Switzerland). The serum levels of iron and CRP were also determined by electrochemiluminescence immunoassay (Cobas c501, Roche Diagnostics Ltd., Switzerland). The reference ranges for serum iron and ferritin were 7 - 32 μmol/L and 30–400 ng/mL, respectively. The reference range for mean corpuscular volume (MCV) was 80–100 fl, and that for CRP was 0–0.8 mg/dL. The World Health Organization defines anemia in the adult or elderly populations as hemoglobin concentrations below the level of 120 g/L for non-pregnant women, and 130 g/L for men [24]. The use of this criterion in the elderly population has been questioned because these cutoff levels were derived from statistical distributions in a study in which the participants were healthy and young. We use a hemoglobin concentration of lower than 120 g/L as the cutoff value for anemia in this study, based on both the data described in the literatures and the normal range (110–176 g/L) of the test system we applied. A cutoff value of 130 g/L was used in sensitivity analyses.

Statistical analysis

Statistical analyses were performed using Statistical Package for SPSS (Windows version 19.0). Continuous variables were expressed as −x ± s for variables of normal distribution and median (interquartile range) for variables of skewness distribution. Categorical variables were expressed as n (%). T test, χ2 test and nonparametric test were used to examine the differences in the corresponding continuous and categorical variables. Multivariable logistic regression was used to calculate the odds ratios (ORs) and 95% confidence intervals (CIs). Two-sided P value < .05 was considered as statistically significant.

Results

General information and baseline diseases of participants

Results from 646 subjects were available for analysis. The mean age was 79.4 ± 8.9 years, with a range of 65–100 years. Baseline diseases which may cause a decrease in hemoglobin concentration were documented (Table 5 in Appendix). The comorbidities were measured by CCI, with an average CCI score of 1.93 in the 646 subjects. The prevalence of H. pylori infection was 35.3%. The subjects were further divided into three subgroups based on their age [65–74 years group (n = 239), 75–84 years group (n = 162), ≥ 85 years group (n = 245)]. As summarized in Table 1, the CCI score and prevalence of comorbidities increased significantly with age (P < .001). There is no difference in the levels of hemoglobin, serum iron and ferritin, as well as H. pylori prevalence, among the three subgroups (P > .05) (Table 1).

Table 1 General Characteristics of participants according to different age groups
Full size table

H. pylori status and hematological parameters

Twenty-one cases with hemoglobin less than 120 g/L were observed in this cohort (Table 6 in Appendix). In general, the prevalence of anemia in the H. pylori positive group was higher than that in the negative group (5.3% vs 2.2%, P = .033). The hemoglobin concentration was slightly lower in the H. pylori positive group than in the negative group, but with no statistical significance (mean ± SD, 144.83 ± 13.77 g/L vs. 146.57 ± 13.11 g/L, P = .115). Either age or CCI score alone had no effects on anemia prevalence or hemoglobin levels (Table 2). When using hemoglobin 130 g/L as a cutoff value for anemia, the odds ratio (OR) of H. pylori positive on anemia prevalence was 1.27 (95% CI: 0.78–2.07; P = .332), compared with the H. pylori negative group (Table 7 in Appendix).

Table 2 Anemia prevalence and hemoglobin levels according to age, CCI score and H. pylori status
Full size table

The relationship between anemia and H. pylori status according to CCI scores and age were further analyzed. The difference of anemia prevalence in each age subgroups remained to be around 3%, but was of no statistical significance. The average CCI scores of H. pylori positive and negative group was 1.85 and 1.97, respectively (P = .643). No relation between CCI scores and anemia was found in this study population. The prevalence of anemia in high (>2) and low (≤ 2) CCI score groups was 4.1 and 2.9%, respectively (P = .442). Among the patients with low CCI score (≤ 2), there was no significant difference of the anemia prevalence between the H. pylori positive and negative groups (3.5% vs. 2.5%, P = .569). However, among the patients who had higher CCI scores (>2), the prevalence of anemia was higher in the H. pylori positive group than those in the negative group (10.3% vs.1.4%, P = .009) (Table 3).

Table 3 Relationship between anemia and H. pylori status according to age and CCI score
Full size table

Table 4 shows the OR of H. pylori status on anemia prevalence. Compared with the H. pylori negative individuals, the crude OR of H. pylori status on anemia prevalence was 2.53 (95% CI: 1.05–6.09; P = .033). We also adjusted for age and CCI scores in the logistic regression. The results were similar. In the sensitivity analysis, we calculated the ORs of H. pylori status on anemia among those without malignant tumors (n = 549), the age–adjusted OR was 2.30 (95% CI: 0.95–5.55).

Table 4 ORs and 95% CI of H. pylori positive on anemia prevalence
Full size table

Erythroid-related indices for both the H. pylori positive and negative groups were analyzed. Hematocrit (HCT) had a marginal difference between the H. pylori positive and negative groups (P = .058). The MCV of the H. pylori positive and negative group was 91.17 ± 3.94 fl and 91.18 ± 4.09 fl (mean ± SD, P = .986), respectively. Other erythroid-related indices showed no significant difference among either H. pylori status (P > .05) (Table 8 in Appendix).

Effect of H. pylori status on serum iron, ferritin and CRP levels

Similar serum iron levels were found in the H. pylori positive and negative groups (mean ± SD, 19.90 ± 6.24 μmol/L vs. 19.87 ± 5.78 μmol/L, P = .961). The results were found to be similar for serum ferritin. The median serum ferritin was 174.00 ng/mL (IQR 108.78–267.88) in the H. pylori positive group, and 181.65 ng/mL (IQR 110.68–270.60) in the H. pylori negative group (P = .613).

Among the subjects recruited in this study, CRP data were available in 262 cases, which included 66 H. pylori positive and 196 H. pylori negative ones. Although all the subjects represent CRP within the normal range, the CRP level in the H. pylori positive group (Median: 0.17 mg/dL, IQR: 0.10–0.69) was higher than that in the H. pylori negative group (Median: 0.10 mg/dL; IQR: 0.07–0.23) (P < .001).

Discussion

The prevalence of H. pylori infection increases with age, with about 50% of the population infected at ages above 60, and around 10% between 18 and 30 [19]. H. pylori infections are usually acquired in early childhood in all countries [25]. Therefore, it is reasonable to assume that chronic H. pylori infections in the elderly can cause anemia due to predisposition to gastrointestinal mucosal lesions [26].

Among the 646 elderly male subjects, we found a significant association between H. pylori infection and anemia. The prevalence of anemia in the positive group was higher than that in the negative group, with a difference of about 3%. This is consistent with the mainstream opinion that anemia is considered as a complication of H. pylori infection [9, 27,28,29]. A strong association between H. pylori infection and anemia (OR 2.53, 95% CI 1.05–6.09) was observed in this study. A similar result was reported in a recent study. Xu et al. found that when compared to H. pylori negative groups, the OR of the H. pylori positive group was 2.01 (95% CI: 0.92–4.40) for mild anemia and 2.41 (95% CI: 0.91–6.34) for moderate-to-severe anemia in male subjects after adjusting for potential confounders [27]. Because of the low prevalence of anemia in this special cohort, no significant difference of average hemoglobin levels between the H. pylori positive and negative groups (mean ± SD, 144.83 ± 13.77 g/L vs. 146.57 ± 13.11 g/L, P = .115) was found, although the positive group did show a slightly lower hemoglobin level than the negative group.

In elderly subjects, the “accepted” cutoff values for the diagnosis of anemia vary between countries, regions and laboratories [30]. The hemoglobin concentration is lower on average in older people [31]. A significant association between H. pylori infection and anemia was found by using hemoglobin concentration of lower than 120 g/L as the cutoff value for anemia in this study. This association was corroborated by a further sensitivity analysis when using the cutoff value of 130 g/L, in which the OR of H. pylori positive on anemia prevalence was 1.27. The anemia prevalence (hemoglobin cutoff value of 120 g/L) in this cohort was 3.3%, which was lower than that reported in the literatures. The anemia prevalence of Chinese city population was recently reported as 9.7% (95%CI: 9.4–10.1%), 6.8% (95%CI: 6.4–7.3%) for male and 12.8% (95%CI: 12.2–13.4%) for female, while 11.4% (95%CI: 9.8–11.3%) for male metropolis more than 60 years-old [32]. In this study, 77 cases of anemia could be identified if the cutoff value was set to 130 g/L. This prevalence (11.9%) is similar to that reported in the literature. Furthermore, we excluded those with overt/occult gastrointestinal bleeding, with malnutrition and those received iron supplementation, etc., that some anemia patients related to nutritional deficiency or iron deficiency were eliminated from the cohort. Thus, the anemia prevalence in this study is relatively low.

We found no significant association between H. pylori infection and serum iron and ferritin levels. The question as to whether H. pylori infection plays a role in IDA has been widely studied. One updated systematic review and meta-analysis study has indicated an association between H. pylori infection and an increased likelihood of depleted iron storage [8]. On the contrary, some conflicting findings have been reported. A study in Australia of individuals aged ≥65 years found no significant association, except among female aspirin users, which supports a negative effect of H. pylori on iron status [18]. A recent large cohort retrospective study also reported that after adjusting for potential confounders, subjects with H. pylori infection had a higher prevalence of normocytic anaemia [27]. Patients with gastrointestinal bleeding or those received iron supplementation were excluded in this study, the level of iron deficiency could be underestimated. H. pylori infection could induce mixed deficiencies such as concomitant iron and vitamin B12 or folate deficiency, that would result in normocytic anemia. Owing to the normal iron and ferritin levels, as well as the normal erythroid-related parameters, we could exclude iron deficiency as a direct cause of normocytic and normochromic anemia in this study group. The absence of macrocytosis can help to infer that there might not be vitamin B12 deficiency in this cohort.

Apart from hematologic malignancies which accounted for 4% of anemia cases, some common causes of normocytic anemia encountered in the elderly included chronic inflammation, mixed nutrition deficiencies, chronic diseases and hemoglobinpathies [31]. Chronic inflammation is one of the most common causes of anemia in the elderly [33]. This type of anemia is typically normocytic and normochromic with mild to moderate reductions in hemoglobin level. Anemia of chronic inflammation is believed to be driven by immune responses [31, 34]. It is reported that H. pylori infection can cause systemic immune responses and chronic inflammation, that would subsequently induce the release of cytokines and inflammatory mediators such as interleukins 1 and 17, as well as tumor necrosis factor–α and interferon-γ [35]. The work-up and diagnosis of unexplained anemia in the elderly may include the measurement of inflammatory markers such as CRP and erythrocyte sedimentation rate, interleukin-6, and hepcidin levels [36]. It is reported that H. pylori infection may increase the level of serum CRP [37]. In this study, although all the subjects presented CRP levels within the normal range, we revealed a higher serum CRP level in the H. pylori positive group than that in the negative group. The possible attribution of H. pylori infection to changes in serum CRP level could be through a complex route of inflammation that has yet to be elucidated.

It is noticeable in this study that in the H. pylori positive subjects of higher CCI scores (>2), the prevalence of anemia was significantly higher than the negative ones with higher CCI scores (10.3% vs. 1.4%). This indicates that H. pylori infection could have a strong association with anemia in people with more comorbidities. No independent effect of CCI scores on anemia or hemoglobin level was found in our study, which was different from what the literature reported [38]. This may be because the average CCI score (mean = 1.93) was low in this cohort, which means that the subjects in this study are in relatively good health, their comorbidities are not severe enough to affect their anemia prevalence. These subjects are from well-developed areas in China, who are also at a social-economic level much better than average as evident by their ability to take part in a routine annual health checkup. This can also partially explain the findings in this study that the prevalence of H. pylori infection was 35.2%, which were lower than those cited by other studies conducted on the Chinese adult population [39]. The two similar CCI scores (1.85 vs. 1.97) in H. pylori positive and negative groups also helped to verify the effects of H. pylori infection, rather than comorbidities, on anemia in this special cohort.

Our study has several limitations. First, we used only 13C-UBT to determine current active H. pylori infection, which is believed to have better sensitivity and specificity than the serological assays [40]. 13C-UBT would present more false negative results in elderly population with more prevalence of atrophic gastritis [41], which may have biased the detection of H. pylori infection. It would be more meaningful if we have serological data and the patients could be classified into H. pylori negative, past H. pylori infection and current H. pylori infection subgroups. Second, we did not obtain the data on serum folate, vitamin B12 levels, which can potentially confound the causal relationship, although the absence of macrocytosis can help to address the issue. Besides, some residual confounders such as smoking and drinking are lacking, which may bias the analysis results. H. pylori eradication can increase the hemoglobin level in IDA patients [42], it would be meaningful if we can improve the normocytic and normochromic anemia after eradication in our cohort. This will be performed in our future study. Third, due to the nature of our institution, we could only study a relatively small sample, and limited to male subjects with a higher-than-average socio-economic status. Therefore, the results and conclusion are not representative of the population, even for the elderly, as a whole. Finally, we cannot confirm the actual causality between H. pylori infection and anemia as this is an observational study. Our hope is that other institutions will perform similar studies so that we can put together a more complete picture of how H. pylori infection affects the other clinical aspects of the elderly population.

Conclusion

In summary, our data seem to suggest a positive association between active H. pylori infection, as determined by 13C-UBT, and anemia in an elderly asymptomatic male cohort, especially in those with more comorbidities. No difference was observed between 13C-UBT positive and negative subjects with respect to erythroid-related indices, serum iron or ferritin. More detailed studies are needed and could help to delineate the prevalence, and the mechanism of action that may lead to anemia after chronic H. pylori infection in the elderly.