Long-term Risk of Upper Gastrointestinal Bleeding After Helicobacter pylori Eradication

A Population-based Cohort Study

Fang Jiang; Chuan-Guo Guo; Ka Shing Cheung; Wai K. Leung


Aliment Pharmacol Ther. 2021;54(9):1162-1169. 

In This Article


Patient Characteristics

Before PS matching, there was a total of 62 738 patients in the H. pylori eradicated cohort and 251 611 patients in the H. pylori-negative cohort. After PS matching, 62 738 H. pylori eradicated patients and 62 738 H. pylori-negative patients were analysed (Figure S1). The baseline characteristics of all study patients, before and after PS matching, are presented in Table 1.

After PS matching, the median age of receiving clarithromycin-based triple therapy or endoscopy in the two cohorts was both 54 years. The median follow-up duration of the H. pylori eradicated and H. pylori-negative subjects was 7.8 years (IQR 5.3–10.4), and 8.4 years (IQR 5.7–10.8) respectively.

Risks of UGIB

Overall, 1,017 (1.62%) H. pylori-eradicated and 709 (1.13%) H. pylori-negative patients had hospitalisation for UGIB during the follow up period with an incidence rate of 20.8 (95% CI 19.5–22.1) and 13.6 (95% CI 12.7–14.7) per 10,000 person-years respectively. Compared to H. pylori-negative patients, H. pylori eradicated patients had a significantly higher risk of UGIB (multivariable Cox regression HR: 1.65, 95% CI 1.49–1.83). The results were consistent in univariable Cox regression model after PS matching (HR: 1.50, 95% CI, 1.37–1.66), and PS matching adjusting for time-varying concomitant medications (HR: 1.44, 95% CI, 1.31–1.59; Table 2).

Kaplan-Meier curve showed a higher rate of UGIB in the H. pylori eradicated group as compared to H. pylori-negative group, particularly after 2-year of follow up (P < .001; Figure 1A). When the follow-up period was broken down into the first 2-year (Figure 1B) and after 2-year (Figure 1C), a significant difference in the rates of UGIB were only observed in the subsequent years (P < .001) but not in the first 2-year (P = .9). Analyses limited to the first 2-year follow-up showed no significant difference in risk of UGIB between the two groups (multivariable Cox regression HR 1.02, 95% CI 0.87–1.20; Table 2). Univariable Cox regression and PS matching adjusting for time-varying concomitant medications yielded consistent results. However, the risk of UGIB in the H. pylori eradicated patients was significantly higher after 2-year follow-up when compared to H. pylori-negative patients (multivariable adjusted HR: 2.18, 95% CI, 1.91–2.49). The results were also consistent in univariable Cox regression model after PS matching (HR: 1.93, 95% CI, 1.70–2.18), and PS matching adjusting for time-varying concomitant medications (HR: 1.84, 95% CI, 1.62–2.08). Sensitivity analysis limited to peptic ulcer bleeding, rather than all UGIB, as outcome showed consistent results with an overall HR of 1.81 (95% CI, 1.57–2.08; Table 2).

Figure 1.

Kaplan-Meier curves for the proportion of patients with UGIB after PS matching. (A) all follow up. (B) follow-up time <2 years. (C) follow-up time after 2 years

Moreover, sensitivity analyses limited to patients with no prior history of UGIB showed an even higher risk of UGIB (HR: 1.93, 95% CI, 1.70–2.19) in H. pylori eradicated than H. pylori-negative patients; as well as in patients with no prior UGIB or peptic ulcer (HR: 2.14, 95% CI, 1.84–2.48; Table S2 https://onlinelibrary.wiley.com/action/downloadSupplement?doi=10.1111%2Fapt.16604&file=apt16604-sup-0002-Supinfo.docx). Alternatively, when analyses were limited to patients with prior history of UGIB, and those with prior UGIB or peptic ulcer, there were also significant increase in UGIB risks in H. pylori eradicated group than in H. pylori-negative group (multivariable adjusted HR: 1.25, 95% CI, 1.05–1.49 and HR: 1.34 95% CI, 1.16–1.54 respectively).

Age-stratified Risks of UGIB

Figure 2 showed the risks of UGIB after age-stratification. The corresponding crude incidence rate of UGIB in the H. pylori eradicated group was 6.2 (95% CI, 4.9–7.7), 9.9 (95% CI, 8.6–11.4), 33.1 (95% CI, 29.9–36.6), and 91.3 (95% CI, 82.0–101.4) per 10 000 person-years in the 18–45 years, 45–60 years, 60–75 years, 75 years and above groups respectively. In the H. pylori-negative group, the corresponding crude incidence rate was 4.4 (95% CI, 3.4–5.7), 5.8 (95% CI, 4.8–6.9), 17.8 (95% CI, 15.5–20.3), and 50.5 (95% CI, 45.1–56.4) per 10 000 person-years respectively.

Figure 2.

Proportion of patients with UGIB after stratified by age

After stratified by age, there were significantly higher overall risks of UGIB in H. pylori eradicated groups than in H. pylori-negative groups, except in the youngest (18–45 years) age group (Figure 2 and Table 3). The corresponding HR of UGIB in different age groups was: 18–45 years (1.34, 95% CI, 0.94–1.90); 45–60 years (1.50, 95% CI, 1.18–1.91); 60–75 years (1.76, 95% CI, 1.48–2.10); and 75 years and older (1.65, 95% CI, 1.40–1.94). The bleeding risks increased further after 2-year of follow up, but were not significant during the initial 2-year (Table 3). Notably, the UGIB risk was also significantly in the 18–45 year group after the 2-year follow-up (HR 1.84; 95% CI, 1.17–2.90).