Association of anaemia with Helicobacter pylori infection: a retrospective study

The role of Helicobacter pylori (H. pylori) infection in haematological system diseases is not well understood. We conducted this retrospective study to explore the association between H. pylori infection and anaemia in the Chinese population. This retrospective study was performed in Aerospace Center Hospital in Beijing. We derived the data from the registration system of the physical population between 2012–2016. Logistic regression models were used to explore the association between H. pylori infection and anaemia. Among 17,791 subjects, there were 7,804 (43.9%) subjects with H. pylori infection and 950 (5.3%) with anaemia. The prevalence of anaemia in the H. pylori (+) group was significantly higher than in the H. pylori (−) group after adjusting for age, sex, marriage, underlying diseases and body mass index. Compared to H. pylori (−), the OR of H. pylori (+) was 1.39 for moderate-to-severe anaemia and 1.05 for mild anaemia. The level of haemoglobin was lower in the H. pylori (+) group than in the H. pylori (−) group. This study indicates that H. pylori infection may be related to anaemia and haemoglobin level in the Chinese population.

Anaemia is a serious public health problem, and its prevalence remains unacceptably high in many regions [1][2][3] , affecting roughly a third of the world's population 4 . Anaemia is usually associated with a variety of other diseases, such as malaria, schistosomiasis, and chronic kidney disease 5 .
Helicobacter pylori (H. pylori) is a spiral shaped pathogenic bacterium found on the human gastric mucosa 6 , and its prevalence remains high in many regions of the world 7 . The prevalence of H. pylori infection was reported to range from 8.7% to 85.5% 8 , which increased with age and varies widely by geographic area, race, ethnicity and socio-economic status 9 . H. pylori infection plays a major role in the incidence of chronic gastritis 10 , but the involvement of its infection on haematological system diseases is not well understood.
H. pylori gastric colonization generally persists for decades and requires the continuous supplementation of nutrients essential for bacterial growth 2 . Rostami-Nejad et al. indicated that H. pylori was associated with iron deficiency anaemia even in celiac disease patients, which was strongly evidence based but weakly reflected in practice 11 . A range of evidence from epidemiological and clinical studies supports an association between anaemia and H. pylori infection. However, reports from different areas and countries are not consistent regarding this association between anaemia and H. pylori infection 1 and the underlying mechanisms remain unclear.
As observed in one meta-analysis 1 , which included most of the previous studies on anaemia and H. pylori, the past studies were imperfect with notable shortcomings. First, most previous international researches on H. pylori infection and anaemia was concentrated on women and children, especially pregnant women, rather than the general population, with relatively small sample sizes. Most also did not consider the severity and type of the anaemia. For studies with large sample studies, the information was often inadequate to account for potential confounding factors. To our knowledge, no published studies with large sample sizes have explored the association between H. pylori and anaemia in the Chinese adult population. Therefore, we conducted this retrospective study using data from the registration system of the physical population to explore the association between H. pylori infection and anaemia in the Chinese population.

Outcomes.
The key outcome (anaemia) was measured objectively using WHO haemoglobin cutoffs 12 . We defined anaemia as a haemoglobin level no more than 130 g/L for men and no more than 120 g/L for women according to WHO sex-based criteria. Patients with anaemia were further divided into two groups according to the severity of anaemia: mild anaemia (110 g/L ≤ hemoglobin < 119 g/L for women and 110 ≤ hemoglobin < 129 g/L for men) and moderate-severe anaemia (hemoglobin < 110 g/L for men and women). Patients with anaemia were also divided into four groups according to the type of anaemia: normocytic anaemia (NA, 80 ≤ MCV ≤ 100 pg), simple small cell anaemia (SCA, MCV < 100 pg and 320 ≤ MCHC ≤ 360 fl), large cell anaemia (LCA, MCV > 100 pg), and microcytic hypochromic anaemia (MHA, MCV < 100 pg and MCHC < 320 fl) 13 .
Statistical analysis. Descriptive statistics were performed, with continuous variables summarized as the mean and standard deviation (SD), and categorical variables summarized as frequencies and proportions. Statistical significance between various groups was tested using the χ 2 test for categorical variables and independent t-test for continuous variables. Generalized linear models were used to explore the association between H. pylori infection and haemoglobin values. Logistic regression models were used to explore the association between H. pylori infection and anaemia. Ordinal logit models were used to explore the association between H. pylori infection and differences in severity or type of anaemia. The variables of age, sex, diabetes, hypertension, coronary heart disease and BMI were adjusted for the above models. As anaemia prevalence is quite different for males or females, we also analysed the effect of sex using all the above models. Odds ratios (ORs) and their 95% confidence intervals (CIs) were estimated. A two-sided P < 0.05 was deemed statistically significant. All analyses were performed using Stata 12.0 (Stata Corp LP, College Station, TX, USA).

Description of subjects & H. pylori and anaemia distribution.
A total of 17,791 subjects were recruited from the hospital for this study. Subjects had a mean age (SD) of 46 (18) years, and 5,844 (32.5%) were female. The overall prevalence of underlying diseases was 30.2% (5,374/12,417), including 11.7% with diabetes, 25.1% with hypertension and 4.6% with coronary heart disease. As for BMI, 27.3% of the subjects were overweight and 12.6% were obese. All basic characteristics of subjects included in the study are shown in Table 1.
Among all subjects, there were 7,804 (43.9%) subjects with H. pylori infection and 950 (5.3%) with anaemia. The subjects with older age, hypertension and higher BMI value were demonstrated to have higher prevalence of H. pylori infection (all P < 0.05). There were no significant differences in the prevalence of H. pylori for diabetes, coronary heart disease or BMI groups (all P > 0.05). Significant differences in the rates of anaemia were found for the variables of age, sex, underlying diseases and BMI (all P < 0.05).
Association between H. pylori and the severity of anaemia. The prevalence of anaemia was 5.5% (428/7,804) in the subjects with H. pylori infection compared to 5.2% (522/9,987) in uninfected individuals. The prevalence of anaemia in the H. pylori (+) group was significantly higher than in the H. pylori (−) group after adjusting for age, sex, marriage, underlying diseases and BMI (OR = 1.19; 95% CI: 1.03, 1.39; P = 0.021), as shown in Table 2. When classified by sex, the anaemia rate in the H. pylori (+) group was significantly higher than in the H. pylori (−) group for the female subjects (OR = 1.25; 95% CI: 1.04, 1.51; P = 0.019), but no significant association was detected for the male subjects (OR = 1.09; 95% CI: 0.84, 1.40; P = 0.530).
The ordinal logit model was used to analyse the association between H. pylori infection and the severity of anaemia. Compared to H. pylori (−), the OR of H. pylori (+) was 1.39 (95% CI: 1.06, 1.54; P = 0.019) for moderate-to-severe anaemia and 1.05 (95% CI: 0.90, 1.22; P = 0.520) for mild anaemia after adjusting for potential confounders ( Table 2). Significant associations between H. pylori infection and the severity of anaemia were also not found in the male and female models.
Associations between H. pylori infection and type of anaemia. The association between H. pylori infection and the type of anaemia was analysed in this study. Haematological parameters for both the H. pylori (+) group and the H. pylori (−) group are summarized in Table 3. The levels of RBC, MCV, MCH, MCHC had significant differences between H. pylori (+) and H. pylori (−) groups (all P < 0.05). However, these differences were small and likely lack biological significance. The association between H. pylori infection and the type of anaemia were analysed using a multinomial logistic regression model (Table 4). Significant differences were not detected between H. pylori infection and different anaemia types in the all subjects model, female model and male model. When adjusted for potential confounders, including age, sex, underlying diseases and BMI, a significant association was found between H. pylori infection and NA (OR = 1.25; 95% CI: 1.04, 1.49; P = 0.017). This association was also found in the female model (OR = 1.28; 95% CI: 1.02, 1.60; P = 0.035) after adjusting for the above factors.

Association between H. pylori infection and the level of haemoglobin. The association between H.
pylori infection and the level of haemoglobin were analysed using the generalized linear model in Table 5. When including all subjects, the haemoglobin level was lower in H. pylori (+) subjects than in H. pylori (−) subjects    Table 5. Association between H. pylori and the level of hemoglobin. a Adjusted β and 95%CI were calculated by adjusting for the potential confounders, including age, sex, underlying diseases and BMI.

Discussion
In this retrospective study with a large sample size, we evaluated the relationship between H. pylori infection and anaemia using 17,791 subjects from the health examination population in Beijing. Subjects with H. pylori infection may have a higher possibility of anaemia and a lower level of haemoglobin. The prevalence of H. pylori infection and anaemia was 43.9% and 5.5%, respectively, which were lower than those cited by other studies conducted on the Chinese adult population 14,15 . The prevalence of anaemia was 9.7% in the Chinese rural population 16 and 9.7% in the urban population on in the 2010-2012 National Nutrition and Health Survey 15 . Another study showed that the prevalence of anaemia in urban community dwelling elderly population was 16.3% in men and 13.7% in women, which increased significantly with age 17 . This may be due to study populations from relatively developed areas in China. The prevalence of iron deficiency anaemia in Chinese population was 0.37% in the year of 2008 and 0.17% for male and 0.20% for female 18 . As we did not obtain the data on serum iron, serum ferritin and serum transferrin saturation, etc, we cannot calculate the prevalence of iron deficiency anaemia in our study. However, MHA may represent the prevalence of iron deficiency anaemia, which was higher in our study than the above.
The association between H. pylori infection and anaemia has been explored by previous epidemiological studies in different settings 9,15,[19][20][21][22][23] . Anaemia is considered as a complication of H. pylori infection 24 . We found that the subjects with H. pylori infection had higher prevalence of anaemia, especially in female subjects. A meta-analysis found a borderline significant and weak positive association between H. pylori infection and anaemia with a pooled OR of 1.15 (95% CI: 1.00, 1.32) 1 . Moreover, that study also indicated that the magnitude of this association was stronger when the analysis was limited to studies that adjusted for confounders, which is consistent with our study.
There are several candidate anaemiagenic mechanisms that might explain the association between H. pylori infection and anaemia. The most plausible mechanism is gastrointestinal blood loss due to H. pylori-induced gastritis or duodenitis 25 . Another possibility is that the H. pylori bacterial sequestration of free iron affects iron transporter molecules, thereby inhibiting free iron absorption and causing food cobalamin malabsorption 26,27 . In addition, H. pylori gastric colonization requires continuous supplementation of nutrients essential for bacterial growth and can use the host's own iron stores 28 . Severe anaemia was related to H. pylori infection in one case report on school age children, which recommends screening for H. pylori infection and appropriate treatment for severe iron-deficiency anaemia 29 . However, we did not find a significant association between H. pylori infection and severity of anaemia in this study. This may be due to the characteristics of our subjects, which had a selection bias.
Our analysis also revealed significant differences in RBC, MCV, MCH, and MCHC between H. pylori (+) and H. pylori (−) groups, but these changes are unlikely to be biologically significant. A health examination of 2,398 healthy subjects and a clinical study also indicated similar results concerning MCV, MCH, and MCHC 30,31 . Moreover, a large household controlled therapeutic trial among children with iron deficiency suggested that H. pylori infection might play a causal role in haematological outcomes 32 . In this study, we also analysed the associations between H. pylori infection and different types of anaemia. To our knowledge, this was the first to explore an association between H. pylori infection and anaemia type. After adjusting for potential confounders, subjects with H. pylori infection had higher prevalence of NA, and this association was also found in the female subjects. This finding requires further clinical or molecular studies to elucidate the mechanisms underlying the association between H. pylori infection and different types of anaemia. We observed a lower haemoglobin level in the positive H. pylori infection group for male subjects. However, this association was not found in all subjects, which is similar with previous research from Taiwan 33 . Another study indicated that haemoglobin levels had statistical significance among no infection, recent infection, long-term infection and past infection subjects. That study concluded that a lower haemoglobin level was related to the presence of gastric atrophy and older age rather than the presence of H. pylori infection itself 30 . A community-based study of Arabs found a significantly lower haemoglobin level in children aged 6-9 years who were infected with H. pylori compared with their uninfected peers 34 . A meta-analysis of randomized control trials of H. pylori eradication has indicated that eradication can increase haemoglobin levels 35 . In addition, it was reported that levels of haemoglobin, serum iron and ferritin are improved following the treatment of H. pylori infection. Based on the above results, we believe that an association between H. pylori infection and haemoglobin level exists. However, the underlying reason for this still requires further research, especially for male subjects.
Our findings should lead to careful consideration of appropriate interventions aimed at eradication of H. pylori, not only for the possibility to improve anaemia status but also to provide protection against underlying diseases (e.g., hypertension), although this will require much effort.
Our study has several limitations. First, we cannot confirm the actual causality between H. pylori infection and anaemia as this is an observational study. Second, all subjects were from the health examination population at Aerospace Center Hospital in Beijing, which leads to selection bias. The third limitation is that serological testing for the presence of anti-H. pylori IgG and IgM does not indicate a current infection and only shows exposure to these bacteria, which may have biased the detection of H. pylori infection. Moreover, we did not obtain iron status laboratory test results, so we cannot give the detailed information on the subjects with iron deficiency anaemia. Finally, the ratio of males to females was high due to the characteristics of the unit from which the subjects came. The above limitations might have biased our results.