Helicobacter pylori infection is associated with diabetes among Chinese adults

Abstract Aims/Introduction Several epidemiological studies investigated the effect of Helicobacter pylori infection on diabetes, but the conclusions remained inconsistent. We aimed to explore the relationship between H. pylori infection and diabetes, as well as glycemic metabolism profiles. Materials and Methods A cross‐sectional study including 58,482 Chinese adults was carried out between January 2016 and December 2017. H. pylori infection was diagnosed by the 13C‐urea breath test. Multivariate regression analyses were carried out to evaluate the association of H. pylori infection with diabetes. Results Of the 58,482 participants, 3,449 (5.9%) had diabetes. The H. pylori‐positive participants had a higher rate of diabetes (7.3% vs 5.2%, P < 0.001), and higher levels of fasting plasma glucose (5.36 ± 1.12 mmol/L vs 5.28 ± 0.95 mmol/L, P < 0.001) and glycated hemoglobin A1c (5.63 ± 0.68% vs 5.57 ± 0.60%, P < 0.001) than the H. pylori negative group. Multivariate regression analyses showed that H. pylori infection was positively related to diabetes (odds ratio 1.25, 95% confidence interval 1.15–1.35). Among the H. pylori‐positive participants, the elevated levels of fasting plasma glucose and glycated hemoglobin A1c were 0.033 mmol/L (95% confidence interval 0.016–0.049 mmol/L) and 0.024% (95% confidence interval 0.008–0.041%), respectively. Additionally, H. pylori infection was significantly related to diabetes in participants aged ≥44 years, but not in participants aged <44 years. Conclusions The present study showed that H. pylori infection is associated with diabetes among Chinese adults. More attention should be paid to adults with H. pylori infection for effective prevention of diabetes.


INTRODUCTION
The number of diabetes patients worldwide has increased dramatically during the past few decades, especially in developing countries 1 . In 2017, China had the world's largest diabetic population, with a total of 114.4 million adult patients 2 . A nationally representatively survey in mainland China showed that the estimated prevalence of diabetes was 10.9% among Chinese adults 3 . Diabetes seriously affects the patients' quality of life, and significantly increases the risk of all-cause mortality among adults in both rural and urban areas of China 4,5 . Risk factors for diabetes, such as genetics, adiposity and smoking, have been well established in past decades. However, whether infectious pathogens are involved in the pathogenesis of diabetes remains inconclusive. In the present study, we focused on the effect of Helicobacter pylori infection on diabetes.
H. pylori, a kind of Gram-negative bacteria that colonizes the stomach of humans, infects 44.3% of the world's population 6 . H. pylori generally causes gastrointestinal diseases, such as chronic gastritis, peptic ulcer disease and gastric malignancy. Epidemiological evidence showed that H. pylori infection also increased the risk of chronic metabolic diseases, such as obesity, dyslipidemia and hypertension [7][8][9] . A meta-analysis including 41 studies documented that diabetes was significantly associated with an elevated rate of H. pylori infection 10 , indicating that diabetes could promote susceptibility to H. pylori infection. However, does H. pylori infection inversely have an impact on diabetes development? Several studies explored the effect of H. pylori infection on diabetes, but the conclusions remained inconsistent [11][12][13][14][15][16] . A prospective study of 782 Latinos aged >60 years reported a positive association of H. pylori infection with diabetes 11 , whereas a cross-sectional study of 1,000 multiethnic Americans aged 45-84 years showed no significant association between them 12 ; both of the two studies had a relatively small sample size and used a serological test to diagnose H. pylori infection. Serological antibody for H. pylori might last for several years after successful eradication of H. pylori 17 , which means that the serological test is not able to adequately represent the current H. pylori infection status.
We carried out the present large cross-sectional study to examine the relationship between H. pylori infection, as diagnosed by the 13 C-urea breath test, and diabetes as well as glycemic metabolism profiles, including fasting plasma glucose (FPG) and glycated hemoglobin A1c (HbA1c), in a Chinese population.

Study population
We used data from a cross-sectional study, which was carried out in the physical examination center of Tongji hospital in Wuhan city, Hubei province, China. Participants aged ≥18 years who had a health checkup in the study hospital were recruited between January 2016 and December 2017. Most of the participants were urban residents employed by government or enterprises. The health checkup data included demographic characteristics, serum biochemical index, anthropometric parameters, the 13 C-urea breath test and information on medical histories. Finally, a total of 58,482 participants who underwent both the 13 C-urea breath test and the monitoring of glycemic metabolism profiles were included in the current study.
The ethics committee of Tongji Hospital, Tongji Medical College, Huazhong University of Science and Technology approved the present study (institutional review board approval number: TJ-C20160115). The study conformed to the ethical guidelines of the Declaration of Helsinki. Written informed consent was collected from each participant.

Assessment of diabetes
Fasting plasma glucose was measured by the hexokinase method using the Cobas 8000 modular analyzer (Roche Diagnostics, Basel, Switzerland). HbA1c was measured by the ionexchange high-performance liquid chromatography method using an automatic glycohemoglobin analyzer (ADAMS A1c, HA-8180; ARKRAY, Kyoto, Japan). In the current study, diabetes was assessed according to the American Diabetes Association criteria 18 as meeting any of the following criteria: (i) FPG ≥7.0 mmol/L; (ii) HbA1c ≥6.5%; and (iii) as the participants who self-reported physician diagnosis of diabetes with use of antidiabetic medication (hypoglycemic agent or insulin) might have normal levels of FPG and HbA1c, they were also considered as diabetics in the present study.

Diagnosis of H. pylori infection
The 13 C-urea breath test was applied to diagnose H. pylori infection. Participants who had the 13 C-urea breath test were overnight fasting for at least 8 h. Participants were required to take 75 mg of 13 C-urea (Urea-13 C Capsule Breath Test Kit; HEADWAY, Shenzhen, China) after providing the initial baseline breath sample. We collected the second breath sample after half an hour. The samples were detected using a 13 C-urea breath test analyzer (HCBT-01; HEADWAY). H. pylori infection was defined as positive if the difference between the two samples exceeded 4.0 parts per 1,000 of 13 CO 2 . Otherwise, it was considered negative.

Assessment of covariates
Demographic characteristics (age, sex) and medical histories were collected by a questionnaire. Standard laboratory methods were applied to obtain data of triglyceride, total cholesterol, low-density lipoprotein, high-density lipoprotein, uric acid and creatinine. Hypertension was defined as follows: systolic blood pressure ≥140 mmHg, or diastolic blood pressure ≥90 mmHg, or use of antihypertensive medication, or self-reported physician diagnosis of hypertension. Body mass index (BMI) was calculated as weight in kilograms divided by the square of height in meters.

Statistics analysis
In the current study, SPSS version 17.0 (SPSS Inc., Chicago, IL, USA) was applied to carry out all statistical analyses. Continuous and categorical variables were presented as meansstandard deviation and percentages, respectively. Differences in the data were evaluated by Student's t-tests and v 2 -tests accordingly. Regression coefficients (b) and their 95% confidence intervals (CI) from multivariate linear regression models were used to evaluate the association of H. pylori infection with FPG and HbA1c. Of the 58,482 participants, 54,404 (93%) were involved as no antidiabetic medication use. We did not include participants who took antidiabetic medication when calculating the average glucose metabolism levels and carrying out the multivariate linear regression analyses, as they might have normal levels of FPG and HbA1c. Logistic regression models were used to assess the association between H. pylori infection and diabetes. The results were presented as odds ratios (ORs) and 95% CIs. Model 1 adjusted for age and sex. In model 2, BMI was added. Model 3 further included blood lipids (triglyceride, total cholesterol, low-density lipoprotein, high-density lipoprotein), hypertension, uric acid and creatinine. We also carried out stratified analyses according to several major confounders. Age was stratified by the median value (44 years). BMI was stratified by the cut-off value of overweight for Asian populations (24 kg/m 2 ). We then carried out the Wald test to explore the interaction across the subgroup. The P for interaction was tested for multiplicative interactions. A two-sided P-value of <0.05 was considered statistically significant. 5.9%. The percentage of H. pylori infection in diabetes patients was significantly higher than that in non-diabetic participants (39.8% vs 31.7%, P < 0.05). There were 18,798 (32.1%) participants infected with H. pylori. The participants' characteristics according to the status of H. pylori infection are presented in Table 1. The H. pylori-positive participants had a higher proportion of diabetes and hypertension (all P < 0.001). The levels of age, BMI, blood pressure, triglyceride, total cholesterol, lowdensity lipoprotein, FPG and HbA1c were higher in participants infected with H. pylori (all P < 0.001). In addition, the level of high-density lipoprotein was lower in participants with H. pylori infection (P < 0.001). Table 2 shows the association of H. pylori infection with FPG and HbA1c using multiple linear regression models. In model 1, which adjusted for age and sex, H. pylori infection was related to higher levels of FPG and HbA1c (all P < 0.001). Additional adjustment for BMI in model 2 did not alter the positive association (all P < 0.001). In model 3, which further included hypertension, blood lipids, uric acid and creatinine, the positive association of H. pylori infection with FPG and HbA1c remained significant; among the H. pylori-positive participants, the elevated levels of FPG and HbA1c were 0.033 mmol/L (95% CI 0.016-0.049) and 0.024% (95% CI 0.008-0.041), respectively.
The association of H. pylori infection with diabetes is shown in Table 3. H. pylori infection was associated with an increased risk of diabetes after controlling for age and sex (model 1), with an OR of 1.34 (95% CI 1.25-1.44). After additional adjustment for BMI (model 2), the positive association between them was similarly observed (OR 1.28, 95% CI 1.19-1.38). The fully adjusted analysis (model 3) suggested that the positive association between H. pylori infection and the risk of diabetes did not materially change (OR 1.25, 95% CI 1.15-1.35).
Age, sex and BMI were the major risk factors. Subgroup analyses were carried out according to these three risk factors (Table 4). H. pylori infection was significantly related to diabetes, which was found to be similar across subgroups stratified by sex and BMI, and more evident among participants aged ≥44 years. No evidence of significant interactions between these three confounders and H. pylori infection on diabetes were found in the analyses (all P for interaction > 0.05).

DISCUSSION
In the present study, it was observed that H. pylori infection was related to an increased risk of diabetes after adjustment for  age, sex, BMI, blood lipids, uric acid, creatinine and hypertension. The present study also showed that the H. pylori-positive participants were more likely to have higher levels of FPG and HbA1c after controlling for potential confounders. In addition, the present study showed that H. pylori infection was significantly related to diabetes in participants aged ≥44 years, but not in participants aged <44 years, although there was no evidence of effect modification. The prevalence of diabetes (5.9% vs 10.9%) and H. pylori infection (32.1% vs 55.8%) among Chinese adults in the present study were lower than those reported in previous investigations 3,19 . As most of the participants in the present study were urban residents employed by government or enterprises, they might have relatively higher levels of education and socioeconomic status, and better sanitary conditions, which could help explain the lower prevalence of H. pylori infection and diabetes in our study population. In addition, the present study did not carry out the oral glucose tolerance test to diagnose diabetes, which might lead to the underestimated prevalence of diabetes. A cross-sectional study of 30,810 middle-aged and older participants from China showed that individuals carrying H. pylori presented an increased risk of diabetes, independent of potential confounders 13 . This finding was consistent with the present results, but the study was restricted to a middle-aged and older population. Evidence shows that H. pylori infection inactively occurs in young adults 20 , so it would be interesting to investigate the effect of H. pylori infection on diabetes among young adults. The subgroup analyses showed that H. pylori infection was not related to diabetes in participants aged <44 years, as Hsieh et al. 14 showed in participants aged <45 years. The impact of H. pylori infection on diabetes in the younger population remained to be elucidated. Two cross-sectional studies from Taiwan reported that participants carrying H. pylori had a higher risk of diabetes 14,15 , whereas no positive association between them was observed in a Turkish study 16 . These three studies were hospital-based, and recruited participants who underwent gastroendoscopy and gastric biopsy. We know that participants who underwent gastric biopsy were very likely patients suffering gastrointestinal diseases, potentially biasing the results. In contrast, the participants included in the present study were from the general population.
As observed in previous studies, H. pylori infection was closely correlated with higher levels of FPG and HbA1c, independent of potential confounders 13,14,21 , which was in accordance with the present findings. FPG, a frequently-used biomarker that reflects instant plasma glucose level, could be substantially influenced by diet content and the amount of exercise before the physical examination. However, participants in the present study were asked to fast overnight and not to carry out moderate-to-vigorous exercise before having health checkups. Thus, diet and exercise might have little effect on the association between H. pylori infection and FPG in the present study. HbA1c, however, a more reliable index that reflects the longterm average level of glycaemia 22,23 , is considered more valid to evaluate the association between H. pylori infection and glycemic metabolism profile.
It remains to be elucidated how H. pylori infection contributes to diabetes, but inflammation and appetite-related hormones are considered to be of importance. It was reported that individuals suffering H. pylori infection were more likely to have elevated levels of several inflammation cytokines, such as  tumor necrosis factor-a, interleukin-6 and C-reactive protein 24,25 . These cytokines were related to insulin resistance and diabetes development 26 . In addition, Gram-negative bacteria, such as H. pylori, that colonize the stomach can accelerate the secretion of lipopolysaccharide, a substance that stimulates innate inflammation processes 27 . Preceding studies documented that H. pylori infection affected the regulation of appetite-related hormones, such as ghrelin and leptin 28,29 . Ghrelin and leptin play a key role in energy homeostasis. Furthermore, these two hormones were reported to be closely correlated with the pathogenesis of obesity and glycometabolism 30,31 , leading to the occurrence and development of diabetes consequently.
In the current study, H. pylori infection was significantly related to diabetes in participants aged ≥44 years, but not in participants aged <44 years, although there was no evidence showing that the interaction between H. pylori infection and age had an effect on diabetes development. Several studies investigated the effect of H. pylori infection on diabetes among middle-aged or older individuals, and found a significant relationship between them 11,13,14 , which was in accordance with the present findings. Most new H. pylori infection occurs during childhood 32,33 . Additionally, the H. pylori infection could perpetuate throughout a lifetime unless eradication therapy is administered 33 . Thus, younger adults usually have a shorter history of H. pylori infection. The long-term cumulative effect and long history of H. pylori infection might help explain why H. pylori infection was found to be related to diabetes among participants aged ≥44 years.
In the present study, the average age of the included participants was 43.81 years. We know that most patients with type 2 diabetes mellitus are diagnosed at middle age and older. Thus, there might be a very limited number of type 1 diabetes mellitus patients in the current study. In addition, two meta-analyses showed that H. pylori infection was associated with type 2 diabetes mellitus, but not type 1 diabetes mellitus 10,34 . Therefore, the effect of type 1 diabetes mellitus was not considered to bias the finding of the present study that participants with a lower BMI were at higher risk of H. pylori-related diabetes. As we know, overweight or obesity is an important risk factor for type 2 diabetes mellitus, and contributes greatly to its occurrence and development. Hence, overweight or obesity might mask the effect of H. pylori infection on diabetes in participants with BMI ≥24 kg/m 2 . Inversely, the effect of H. pylori could be highlighted in participants with BMI <24 kg/m 2 , leading to a higher risk of H. pylori-related diabetes in participants with a lower BMI.
We hypothesized that H. pylori infection was early-onset in the present study, as people generally become infected with H. pylori during childhood 32,33 . However, if the diabetes was early-onset, we could not exclude the possibility that diabetes increased the risk of H. pylori infection. One explanation was that diabetes could impair the body's cellular and humoral immunity 35 , resulting in increased susceptibility to bacterial infection. Another explanation suggested that diabetes could reduce gastric motility and acid secretion 36,37 , consequently leading to vulnerability to H. pylori infection. Therefore, it would be interesting to validate the bidirectional association between H. pylori infection and diabetes in the future.
The present study had several strengthens. First of all, our study investigated the effect of H. pylori infection on diabetes among young adults in mainland China. Second, we included a large sample from the general population and adjusted for important risk factors, which make the results more reliable and convincing. Third, we used the 13 C-urea breath test to determine H. pylori status. This is a non-invasive and effective diagnostic method for H. pylori, with high sensitivity and specificity 38,39 . As we know, this is the first study to examine the association of diabetes with H. pylori infection using the 13 Curea breath test. Fourth, FPG and HbA1c were used to describe the glycemic metabolism profiles, indicating that we evaluated the association of H. pylori infection with both instant and long-term average levels of glycemia.
There were also some limitations. First, the causal relationship of H. pylori infection and diabetes could not be determined due to the cross-sectional design. However, people generally become infected with H. pylori during childhood 32,33 . Therefore, participants with both type 2 diabetes mellitus and H. pylori infection might suffer H. pylori infection at an earlier age. Second, although we adjusted for several major risk factors, other potential confounders that we did not measure, such as the use of antibiotics, family history of diabetes, waist circumference, alcohol consumption and smoking, might bias the results. Third, the participants were recruited from a physical examination center and might not represent all the Chinese adults, which limited the extrapolation of the present findings. Fourth, we did not carry out the validation study of the 13 Curea breath test by gold standard diagnostic tools, such as gastroscopy and subsequent biopsy, bacterial culture, and anti-H. pylori. However, previous studies have shown excellent agreement between the 13 C-urea breath test and gastroscopy with biopsy 40,41 , indicating that the 13 C-urea breath test has high diagnostic accuracy and good performance.
In summary, the present findings suggest that H. pylori infection is independently related to a higher risk of diabetes among Chinese adults. More attention should be paid to adults with H. pylori infection for effective prevention of their diabetes. The prevention and treatment of H. pylori infection might not only decrease the burden of gastrointestinal diseases, but also delay the development of diabetes, therefore reducing the heavy burden of diabetes. Well-designed prospective studies are warranted to confirm the effect of H. pylori infection on the incident risk of diabetes and corresponding confounders.

ACKNOWLEDGMENTS
This study did not obtain any funding. The authors thank the study participants in the Physical Examination Center of Tongji Hospital for participating in the current study.