Patients admitted to the second affiliated hospital of Xi’an Jiaotong University from September 2016 to March 2021 who underwent ¹⁴C-UBT and thyroid ultrasonography were included in the study (Shaanxi and Gansu provinces, and Ningxia autonomous region). The individuals were divided into 2 groups: a cross-sectional study of single follow-up and a retrospective cohort study of multiple follow-ups ( Figure 1A ). Patients were excluded on the following criteria: (1) The use of antibiotics, proton pump inhibitors, H₂ receptor antagonists, and bismuth within 4 weeks before the tests; (2) History of gastric surgery; (3) Allergic to ¹⁴C-urea; (4) Missing information during the follow-up. The study was performed according to the principles of the Declaration of Helsinki and approved by the ethics committee of the second affiliated hospital of Xi’an Jiaotong University, Xi’an, Shaanxi, China.

All participant information was obtained from the clinical record database of the second affiliated hospital of Xi’an Jiaotong University. H. pylori infection was diagnosed by ¹⁴C-UBT, which had been considered the “gold standard” in clinical practice (Aung et al., 2021). In ¹⁴C-UBT, all participants were required to fast for more than 6 hours before taking ¹⁴C urea capsule. There was a prohibition against eating and drinking before taking the breath sample after waiting for 30 minutes. The positive result was defined as a standard of over 100 disintegrations per minute per millimole (dpm/mmol), while a range of 0~100 was considered negative.

The types of TNs were classified by ultrasonographers with extensive experience in the second affiliated hospital of Xi’an Jiaotong University. TNs were assessed by the American College of Radiology Thyroid Imaging Reporting and Data System (ACR TI-RADS) according to the characteristics of composition, echogenicity, shape, margin, and echogenic foci, which were divided into 5 categories from TR1 (benign) to TR5 (high suspicion of malignancy) (Tessler et al., 2017). To facilitate analysis, TNs had been separately grouped as TR ≤ 2, TR≥4, and TR=3.

Baseline characteristics included age, gender, body mass index (BMI), waist-hip ratio (WHR), systolic blood pressure (SBP), diastolic blood pressure (DBP), smoking, alcohol, seafood consumption habit, and family history of thyroid disease. BMI was calculated by dividing weight (in kilograms) by the square of height (in meters). WHR was defined as the minimum circumference between the iliac crest and the rib cage. Blood pressure was detected with a mercury sphygmomanometer on the arm after individuals rested for 5 minutes. Hypertension was considered systolic blood pressure (SBP)≥140 mmHg or diastolic blood pressure (DBP)≥90 mmHg or anti-hypertensive treatment.

Relevant covariates involved laboratory indicators related to hepatic-renal functions and thyroid hormones. The fresh blood sample after an overnight fast was obtained from a venipuncture for the following measurements: (1) Hepatic-renal functions: alanine aminotransferase (ALT), aspartate aminotransferase (AST), uric acid (UA), serum creatinine (SCr), total cholesterol (TCH), triglyceride (TG), low-density lipoprotein-cholesterol (LDL-C), high-density lipoprotein-cholesterol (HDL-C), and fasting blood glucose (FBG); (2) Thyroid hormones: thyroid stimulating hormone (TSH), free T3 (FT3), free T4 (FT4), anti-thyroid peroxidase (anti-TPO), anti-thyroglobulin (anti-Tg), thyroglobulin (Tg), thyroid stimulating hormone receptor antibody (TRAb), reverse T3 (rT3), intact parathyroid hormone (iPTH), and thyroid volume. The normal ranges of blood indicators were shown in the supplementary material in details ( Table S1 ). The thyroid volume was calculated by measuring left and right lateral lobes with the isthmus by ultrasonography.

Statistical analysis was performed by SPSS 22.0 software. The quantitative variables were shown as mean ± standard deviation ( X ¯ ± SD ) and analyzed by Student’s t-test and one-way ANOVA. The qualitative variables were expressed as a percentage (%) with a number of cases and analyzed by the chi-square (χ²) test. The crude odds ratios (ORs) were estimated with the binary logistic regression. The assignments of potential risk factors in the logistic regression were presented with numbers 0, 1, 2, and 3 ( Table S2 ). The cross-sectional study was performed by a single measurement to explore the association between H. pylori and TNs, while multiple assessments were applied to investigate TNs development with H. pylori infection in retrospective cohort study. A p-value less than 0.05 was considered to be statistically different.

A total of 9,042 individuals were enrolled with ¹⁴C-UBT and thyroid ultrasonography. After exclusion from the criteria, 8,839 patients were included in the study. A cross-sectional study of single follow-up (n=8,711) and a retrospective cohort study of multiple follow-ups (n=139) were performed, respectively. In the cross-sectional study, the patients were divided into 2 groups according to the presence (n=4,176) or absence (n=4,535) of TNs. In the retrospective cohort study, the individuals were also divided into 2 groups: 61 patients with persistent H. pylori infection and 78 individuals without infection ( Figure 1A ).

The baseline data showed the characteristics of all enrolled individuals according to the status of TNs. The average prevalence of H. pylori infection and TNs was 39.58% and 47.94% in the adults of Northwest China, respectively. The prevalence of TNs was significantly higher in patients with H. pylori infection than in those without (52.55% vs. 44.92%, p<0.01). The indicators of baseline characteristics were also higher in TNs patients than those without TNs, for example, age, female gender, smoking, alcohol, and seafood consumption habit, as well as family history of thyroid disease. For relevant covariates, the laboratory indicators of hepatic-renal functions and thyroid hormones, also had higher levels in TNs individuals compared to those without TNs, such as ALT, AST, UA, TCH, TSH, FT4, anti-TPO, anti-Tg, Tg, TRAb, and urinary iodine, apart from HDL-C, iPTH, and rT3. Although the p-values for indicators BMI, WHR, blood pressure (SBP/DBP), TG, LDL-C, and FT3 were less than 0.01, we did not consider these covariates to be different between the individuals with and without TNs, due to the small effect sizes (χ²/t). However, the levels of SCr and FBG did not have significant differences between patients with and without TNs ( Table 1 ).

The prevalence of TNs between positive and negative H. pylori patients was calculated, and the relevant covariates were also analyzed ( Table 1 ). The average prevalence of TNs was 47.94% in the northwestern Chinese adults detected by ultrasonography. The prevalence of TNs was significantly higher in H. pylori-positive individuals than those without infection (52.55% vs. 44.92%, p<0.01). In addition, significant prevalence of TNs was also observed in basic indicators, for example, advanced age, female gender, smoking, alcohol, seafood consumption habit, and family history of thyroid disease (all p<0.01). Moreover, we also found an increased prevalence of TNs in laboratory indicators, such as ALT, AST, UA, TCH, TSH, FT4, anti-TPO, anti-Tg, Tg, TRAb, urinary iodine, and thyroid volume (all p<0.01). However, there was no significant difference in SCr and FBG between patients with and without TNs (p=0.15 and p=0.59). Also, it was noted that rT3 and iPTH were all decreased in subjects with TNs than those without TNs (all p<0.01).

TNs were divided into 5 types according to ACR TI-RADS (TR1~5). We combined these types into 3 clusters, namely TR ≤ 2, TR=3, and TR≥4. There were significant differences in H. pylori infection among 3 clusters of TNs (χ^{2 =} 152.47, p<0.01, Table 2 ). The proportion of H. pylori infection in the TR≥4 group was considerably higher than that in the TR=3 group (Bonferroni, χ^{2 =} 76.46, p<0.01), but slightly lower compared with the TR ≤ 2 group (Bonferroni, χ^{2 =} 124.33, p<0.01). A binary logistic regression was performed, which obtained 4 models with odds ratios (ORs, Figure 1B ). As shown in Model 1, the unadjusted OR for the association between H. pylori infection and TNs was 1.624 (95% CI 1.242~2.123, p<0.01). Model 2 indicated that H. pylori infection was related to an increased risk of TNs with an OR of 1.731 (95% CI 1.294~2.316, p<0.01) after an adjustment for age, gender, alcohol, seafood consumption habit, WHR, and SBP. Model 3 revealed a significant positive correlation (OR=2.287, 95% CI 1.633~3.205, p<0.01) adjusted for additional relevant covariates, including smoking, family history of thyroid disease, ALT, AST, HDL-C, and TCH. On the basis of Model 3, Model 4 indicated that H. pylori was associated with TNs but with a slightly lower risk (OR=2.016, 95% CI 1.390~2.922, p<0.01) than that in Model 3 after an adjustment for BMI, DBP, FT3, FT4, anti-TPO, Tg, and rT3 ( Table 2 ).

A total of 139 patients had been followed-up by ¹⁴C-UBT and thyroid ultrasonography for 5 years. Individuals with annually positive ¹⁴C-UBT were considered persistent H. pylori infection and TN types were classified by ultrasonography according to ACR TI-RAD. The annual incidences of TNs were all significantly higher in patients with persistent H. pylori infection (all p<0.05) than those without infection at 5-year follow-ups ( Figure 1C ). There were 14 and 8 new cases with TNs existing in individuals with persistent infection of H. pylori and without infection in the 1^st year of the follow-up, respectively. After 5 years, a total of 80 new cases developed TNs during the follow-up, including 57 cases with persistent H. pylori infection and 23 cases without infection ( Table 3 ).

The association between persistent H. pylori infection and the development of TN types (TR ≤ 2, TR=3, and TR≥4) had been investigated in the 1^st year of 5-year follow-up ( Table 4 ). For TNs individuals with H. pylori infection, the indicators including WHR, alcohol, TG, rT3, and iPTH, were all significantly higher than those without infection (all p<0.05). It was interesting to note that WHR, TG, and rT3 were significantly greater in positive patients than their counterparts in the TR≥4 group (all p<0.01). However, we only observed a slight difference of UA in the TR≥4 group (p=0.047), and this difference did not occur in the TR ≤ 2 and TR=3 groups. For other remaining indicators, no differences were observed, regardless of TN types.