Helicobacter pylori infection as a risk factor for serum bilirubin change and less favourable lipid profiles: a hospital-based health examination survey

We performed a case-control study by selecting subjects who were overtly positive for H. pylori as cases and overtly negative controls matched by sex and age. We screened subjects aged 18–79 years who were receiving annual health examinations including ¹³C –urea breath test (UBT) in Beijing Tongren Hospital from March 2016 to May 2017. Subjects with ¹³C-UBT values≥10‰ or ≤ 1‰ were defined as overtly positive (cases) or overtly negative (controls) for H. pylori, respectively. Subjects with ¹³C-UBT results between 1‰ and 10‰ were excluded from the study. Subjects in the H. pylori positive group were matched 1:1 with age and sex to H. pylori negative individuals. After the primary assessment of the baseline characters for all subjects, we further excluded participants with liver and gall bladder diseases (hepatitis, jaundice, cholecystitis, biliary calculus), abnormal liver function (alanine aminotransferase (ALT) or aspartate aminotransferase (AST) > 1.5 times upper normal limit, or bilirubin > twice upper normal limit), abnormal kidney function (Cr > upper normal limit) to better eliminate the potential biases caused by diseases.

Each subject had anthropometric measurements. Presence of systematic or previous diseases, such as diabetes mellitus (DM), hypertension, hepatitis, jaundice, cholecystitis or biliary calculus were noted. Body mass index (BMI) was measured as weight (kg) divided by height (meters) squared (kg/m²). Waist circumference (WC) was measured at the level of the umbilicus in cm. Blood pressure (BP) was measured three times when participants were seated, and the average of the last two measurements was recorded. Blood samples were collected after an overnight fasting for the determination of plasma glucose, glycosylated hemoglobin A1c (HbA1c), total cholesterol (TC), triglycerides (TG), high density lipoprotein cholesterol (HDL-C), low density lipoprotein cholesterol (LDL-C), direct bilirubin, total bilirubin, total bile acids, alkaline phosphatase (ALP), ALT, AST [9], γ-glutamyl transpeptidase (γ-GT), blood urea nitrogen [10], serum creatinine (SCr) and uric acids (UA) concentrations.

Data are presented as the mean ± SD. The baseline characteristics of subjects were compared using the chi-squared test for categorical variables and the Student t-test for continuous variables. Distribution of discrete/qualitative variables was compared by trend chi-square test. Binary logistic regression analysis was used to estimate crude and adjusted odds ratios (ORs) (95% CIs) to assess the risk of bilirubin change associated with H. pylori infection. When data were not normally distributed, the correlations of bilirubin and cholesterol were determined by the Spearman correlation coefficient analysis. Calculations were performed using SPSS 24.0 statistical software (SPSS Inc., Chicago, IL, USA), and significance was established at a two-tailed P < 0.05.

A total of 1982 subjects participated in the health examination including ¹³C -UBT, physical examination and blood tests. Based on selection criteria, 623 subjects were initially defined as cases, 1010 were defined as controls, and 349 were excluded from the study. The primary baseline characters of the two groups are shown in Additional file 1: Table S1. Unexpectedly, our results unveiled an intriguing association between bilirubin decrease and H. pylori infection (direct bilirubin 2.50 ± 0.99 vs. 2.37 ± 0.90 μmol/L, P = 0.015, total bilirubin 14.49 ± 5.66 vs. 13.87 ± 5.32 μmol/L, P = 0.049). We doubted if other disorders may influence the bilirubin levels and caused the inauthentic results, hence, we excluded participants with diseases which could affect bilirubin levels pathologically. According to the specified criteria, 29 patients were excluded from the study. Of the remaining 1953 subjects, 617 (31.6%) overtly positive for H. pylori were matched 1:1 with age and sex to 988 (50.6%) overtly negative for H. pylori, 348 subjects with ¹³C-UBT results between 1‰ and 10‰ were excluded from the study. Thus, 617 overtly positive H. pylori individuals (case group) and 617 sex and age- matched overtly negative H. pylori subjects (control group) comprised the case-control study (Fig. 1).

The demographic and biochemical parameters of the cases and the controls are shown in Table 1. The case group had significantly less favourable lipid profiles than controls: LDL-C 2.98 ± 0.76 vs. 2.89 ± 0.75 mmol/L (P = 0.033), HDL-C 1.39 ± 0.37 vs. 1.44 ± 0.41 mmol/L (P = 0.044). Unsurprisingly, the case group also had lower bilirubin levels compared with the control group: direct bilirubin 2.34 ± 0.38 vs. 2.47 ± 0.90 μmol/L (P = 0.008), total bilirubin 13.65 ± 4.84 vs. 14.28 ± 5.01 μmol/L (P = 0.026). There was no significant difference in other demographic and clinical characteristics.

To further investigate the correlation between H. pylori infection and bilirubin levels, subjects from both groups were assigned to 4 grades based on quartiles of serum bilirubin concentrations and the proportions of each grade in case and control groups are shown in Fig. 2. In the H. pylori positive group, the proportion of subjects with direct bilirubin levels in the highest quart (> 2.8 μmol/L) was smaller than that of the H. pylori negative group (21.1% vs 28.4%) (Trend χ² = 4.119, P = 0.042). The same trend was also observed in subjects assigned by total bilirubin level (Trend χ² = 6.256, P = 0.012). This suggests that H. pylori infection is associated with both direct and total bilirubin reduction.

Similar trend was observed when subjects were grouped according to quartiles of LDL-C level (Trend χ² = 4.577, P = 0.032, data not shown), but was not seen when subjects grouped according to quartiles of HDL-C level (Trend χ² = 1.185, P = 0.288, data not shown) although the mean of HDL-C was significant different in H. pylori overtly positive and negative groups (Table 1).

Given the association between increased serum bilirubin levels within the reference range and better health outcomes (or conversely lower bilirubin concentrations with higher morbidities) [11‐16], we assigned subjects into two groups with direct bilirubin (2.8) and total bilirubin (16.4) in the upper distribution quartile. Bilirubin levels below the upper quartile were defined as “bilirubin not increase”. Logistic regression analysis was performed to determine the independence of the association between H. pylori infection and risk for “bilirubin not increase”. As shown in Table 2, H. pylori infection was associated with higher risk of both direct bilirubin (OR = 1.483, 95% CI =1.143–1.925, P = 0.003) and total bilirubin (OR = 1.336, 95% CI =1.030–1.733, P = 0.029) “not increase” in univariate analysis. After adjustment for age and sex (model 2), further adjustment for BMI, ALT and AST (model 3), and then additional adjustment for HDL-C, LDL-C, TC and TG (model 4), the ORs remained significant.

We found that lipid profiles improved with increasing direct bilirubin levels regardless of H. pylori status. LDL-C was negatively correlated with direct bilirubin concentration (R = − 0.260, P < 0.0001) (Fig. 3a), while HDL-C was positively correlated with direct bilirubin level. (R = 0.063, P = 0.028) (Fig. 3b).