Helicobacter pylori Screening After Acute Myocardial Infarction: The Cluster Randomized Crossover HELP-MI SWEDEHEART Trial

Robin Hofmann; Stefan James; Martin O Sundqvist; Jonatan Wärme; Oskar Angerås; Joakim Alfredsson; David Erlinge; Gabriel Arefalk; Göran Arstad; Simon Blomberg; Ole Fröbert; Kristina Hambraeus; Per M Hellström; Jörg Lauermann; Matthias Lidin; Lars Lindhagen; Georgios Mourtzinis; Carolina Schoede; Erik Thunström; Birgitta Voldberg; Henrik Wagner; Ollie Östlund; Tomas Jernberg; Magnus Bäck

doi:10.1001/jama.2025.15047

. 2025 Sep 1;334(13):1160–1169. doi: 10.1001/jama.2025.15047

Helicobacter pylori Screening After Acute Myocardial Infarction

The Cluster Randomized Crossover HELP-MI SWEDEHEART Trial

Robin Hofmann ^1,^✉, Stefan James ^2,³, Martin O Sundqvist ¹, Jonatan Wärme ¹, Oskar Angerås ^4,⁵, Joakim Alfredsson ^6,⁷, David Erlinge ⁸, Gabriel Arefalk ^2,⁹, Göran Arstad ¹⁰, Simon Blomberg ¹¹, Ole Fröbert ^12,¹³, Kristina Hambraeus ¹⁴, Per M Hellström ¹⁵, Jörg Lauermann ^6,¹⁶, Matthias Lidin ¹⁷, Lars Lindhagen ³, Georgios Mourtzinis ^4,¹⁸, Carolina Schoede ^6,¹⁹, Erik Thunström ^4,²⁰, Birgitta Voldberg ²¹, Henrik Wagner ⁸, Ollie Östlund ³, Tomas Jernberg ²², Magnus Bäck ¹⁷

¹Department of Clinical Science and Education, Division of Cardiology, Karolinska Institutet, Södersjukhuset, Stockholm, Sweden

²Department of Medical Sciences, Cardiology, Uppsala University, Uppsala, Sweden

³Uppsala Clinical Research Center, Uppsala University, Uppsala, Sweden

⁴Department of Molecular and Clinical Medicine, Institute of Medicine, Sahlgrenska Academy, University of Gothenburg, Gothenburg, Sweden.

⁵Department of Thoracic surgery and Cardiology, Sahlgrenska University Hospital, Gothenburg, Sweden

⁶Department of Health, Medicine and Caring Sciences, Linköping University, Linköping, Sweden

⁷Department of Cardiology, Linköping University, Linköping, Sweden

⁸Department of Cardiology, Clinical Sciences, Lund University, Lund, Sweden

⁹Department of Cardiology, Karlskrona Hospital, Karlskrona, Sweden

¹⁰Department of Cardiology, St Görans Hospital, Stockholm, Sweden

¹¹Department of Internal Medicine, Cardiology, Lidköping Hospital, Lidköping, Sweden

¹²Department of Cardiology, Faculty of Health, Örebro University, Örebro, Sweden

¹³Department of Clinical Medicine, Faculty of Health, Aarhus University, Aarhus, Denmark

¹⁴Department of Cardiology, Falun Hospital, Falun, Sweden

¹⁵Department of Medical Sciences, Gastroenterology, Uppsala University, Uppsala, Sweden

¹⁶Department of Cardiology, Ryhov Hospital Jönköping, Jönköping, Sweden

¹⁷Department of Medicine Solna, Karolinska Institutet, and Department of Cardiology, Heart and Vascular Center, Karolinska University Hospital, Stockholm, Sweden.

¹⁸Department of Medicine and Emergency Care Mölndal, Sahlgrenska University Hospital, Gothenburg, Sweden

¹⁹Department of Cardiology, Norrkoping Hospital, Norrkoping, Sweden

²⁰Department of Medicine, Geriatrics, and Emergency Medicine, Sahlgrenska University Hospital Östra, Gothenburg, Sweden

²¹Department of Cardiology, Västerås Hospital, Västerås, Sweden

²²Department of Clinical Sciences, Danderyd Hospital, Karolinska Institutet, Stockholm, Sweden

Accepted for Publication: August 5, 2025.

Published Online: September 1, 2025. doi:10.1001/jama.2025.15047

^✉

Corresponding Author: Robin Hofmann, MD, PhD, Department of Clinical Science and Education, Division of Cardiology, Karolinska Institutet, Södersjukhuset, 11883 Stockholm, Sweden (robin.hofmann@ki.se).

Author Contributions: Drs Hofmann and Bäck had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Hofmann, James, Sundqvist, Wärme, Blomberg, Hambraeus, Hellström, Lindhagen, Östlund, Jernberg, Bäck.

Acquisition, analysis, or interpretation of data: Hofmann, James, Sundqvist, Wärme, Angerås, Alfredsson, Erlinge, Arefalk, Arstad, Fröbert, Hambraeus, Hellström, Lauermann, Lidin, Lindhagen, Mourtzinis, Schoede, Thunström, Voldberg, Wagner, Östlund, Jernberg, Bäck.

Drafting of the manuscript: Hofmann, Sundqvist, Lindhagen, Bäck.

Critical review of the manuscript for important intellectual content: All authors.

Statistical analysis: Hofmann, Wärme, Lauermann, Lindhagen, Östlund.

Obtained funding: Hofmann, Alfredsson.

Administrative, technical, or material support: Hofmann, Sundqvist, Wärme, Erlinge, Arefalk, Arstad, Hambraeus, Lauermann, Lidin, Thunström, Voldberg, Wagner, Bäck.

Supervision: Hofmann, Wärme, Angerås, Hambraeus, Thunström, Bäck.

Other - Study site PI: Schoede.

Other - Critical assessment of data and writing: Hellström.

Conflict of Interest Disclosures: Dr Hofmann reported receiving lecture and advisory board fees from AstraZeneca and MSD/Pfizer outside the submitted work. Dr James reported receiving grants from Amgen, Novo Nordisk, AstraZeneca, and Jansen outside the submitted work. Dr Angerås reported receiving grants from AstraZeneca and Novo Nordisk outside the submitted work. Dr Alfredsson reported receiving lecture fees from Novartis, Orion Pharma, and Boehringer Ingelheim outside this work outside the submitted work. Dr Arefalk reported receiving speaker fees from AstraZeneca, Boehringer-Ingelheim, and Novo Novartis outside the submitted work. Dr Fröbert reported receiving consulting fees from GSK, speaker fees Sanofi and Pfizer outside the submitted work. Dr Östlund reported receiving grants from the Karolinska Institutet Uppsala clinical research center, which was paid by the sponsor of the trial during the conduct of the study. Dr Jernberg reported receiving institutional grants from MSD outside the submitted work. Dr Bäck reported receiving speaker and consultant fees to his institution from Amarin, Amgen Novartis, and FreseniusKabi outside the submitted work. No other disclosures were reported.

Funding/Support: The study was funded by grants 2019-00414 from the Swedish Research Council, 2021-0275, 2024-0419 from the Swedish Heart Lung Foundation, and RS 2020-0314 and RS 2022-0674 from the Region Stockholm. Dr Hofmann was supported by grants 2021-0273 from the Swedish Heart Lung Foundation and RS 2021-0933 from the Region Stockholm. Dr Bäck was supported by grants 2023-02652 from the Swedish Research Council, 2024-0697 from the Swedish Heart-Lung Foundation, and RS 2023-0859 from the Region Stockholm.

Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Meeting Presentation: This paper was presented at the European Society of Cardiology (ESC) Congress 2025; September 1, 2025; Madrid, Spain.

Data Sharing Statement: See Supplement 3.

Additional Contributions: We thank the patients who participated in the HELP-MI SWEDEHEART trial; the clinical and research teams of the participating hospitals for their collaboration and commitment to SWEDEHEART and this trial; and the staff at the Uppsala Clinical Research Center at Uppsala University and Karolinska Institutet, Södersjukhuset, for administrative assistance. Nonauthor Collaborators: A list of participating centers, investigators and study personnel is provided in Supplement 2.

^✉

Corresponding author.

PMCID: PMC12403021 PMID: 40887995

Key Points

Question

Does routine Helicobacter pylori screening with urea breath test of patients with myocardial infarction reduce upper gastrointestinal bleeding events?

Findings

In this nationwide, cluster randomized, crossover clinical trial involving 18 466 patients with myocardial infarction from 35 Swedish hospitals, routine H pylori screening did not significantly reduce the risk of upper gastrointestinal bleeding (incidence rate, 16.8 vs 19.2 events per 1000-person-years; rate ratio, 0.90), although a benefit among patients at higher risk of bleeding may exist.

Meaning

Routine H pylori screening with urea breath test in patients with myocardial infarction cannot be recommended.

Abstract

Importance

Upper gastrointestinal bleeding is common after myocardial infarction.

Objective

To determine whether routine screening for Helicobacter pylori infection during hospitalization for myocardial infarction reduces bleeding events and improves clinical outcomes.

Design, Setting, and Participants

A nationwide, open-label, 2-period, 2-sequence, cluster randomized, crossover clinical trial using a clinical registry for study population definition and data collection merged with national Swedish health data registries. From November 17, 2021, through January 17, 2024, thirty-five Swedish hospitals grouped into 18 clusters were randomized to a sequence of 1 year with routine H pylori screening of all patients with acute myocardial infarction followed by a washout period of 2 months before crossing over to 1 year with usual care or vice versa. Patients were followed up until January 17, 2025.

Intervention

Routine addition of H pylori screening by urea breath test to standard care in all patients hospitalized for myocardial infarction during the screening periods.

Main Outcome and Measure

Upper gastrointestinal bleeding, analyzed by a negative binomial model in the intention-to-treat population.

Results

A total of 18 466 patients (median age, 71 years [IQR, 61-79], 13 138 males [71%]) with myocardial infarction were followed up: 9245 during the screening periods and 9221 during the nonscreening periods. At admission, 2284 during the screening periods and 2275 during the nonscreening periods (both 24.7%) reported proton pump inhibitor use. During screening periods, 6480 patients (70%) had undergone testing, of those 1532 (23.6%) tested positive for H pylori. After a median follow-up of 1.9 years, 299 patients in the screening group (incidence rate, 16.8 events per 1000 person-years; cumulative hazard at 3 years, 4.1%) and 336 in the usual care group (incidence rate, 19.2 events per 1000 person-years; cumulative hazard at 3 years, 4.6%) experienced the primary end point of upper gastrointestinal bleeding (rate ratio [RR], 0.90; 95% CI, 0.77-1.05; P = .18). Predefined nonmultiplicity adjusted subgroup analyses showed a heterogeneous intervention effect; for no anemia (RR, 0.98; 95% CI, 0.80-1.21), mild anemia (RR, 0.64; 95% CI, 0.42-0.98), and moderate to severe anemia (RR, 0.44; 95% CI, 0.23-0.87; P for interaction = .03).

Conclusions and Relevance

Among unselected patients with acute myocardial infarction, routine H pylori screening did not significantly reduce the risk of upper gastrointestinal bleeding.

Trial Registration

ClinicalTrials.gov Identifier: NCT05024864

This randomized cluster trial investigates whether routine screening for Helicobacter pylori using urea breath tests among patients after acute myocardial infarction reduces upper gastrointestinal bleeding events.

Introduction

The implementation of evidence-based therapies,^1,2 most notably early revascularization combined with potent antiplatelet agents, has markedly reduced recurrence of ischemic events after acute myocardial infarction.³ However, these therapeutic gains have been accompanied by a concomitant rise in bleeding complications,⁴ particularly those originating from the gastrointestinal tract.^5,6 Upper gastrointestinal bleeding is the most common bleeding source after myocardial infarction⁷ and is associated with increased mortality and recurrent cardiovascular events.⁶

The use of a proton pump inhibitor (PPI) may reduce the upper gastrointestinal bleeding risk and enhance adherence to antiplatelet therapy, thereby improving post–myocardial infarction outcomes.⁸ However, the net benefit of long-term PPI treatment remains controversial and current guidelines recommend PPI use after myocardial infarction only for patients at elevated upper gastrointestinal bleeding risk.^1,2

In this context, chronic active gastritis due to Helicobacter pylori may represent another accessible target for upper gastrointestinal bleeding prevention after myocardial infarction^9,10 because it is highly prevalent among patients with cardiovascular disease³ and associated with a markedly increased upper gastrointestinal bleeding risk when H pylori infection coincides with the use of potent antiplatelet agents.¹¹ However, data derived from randomized clinical trials evaluating the utility of H pylori screening and eradication remain scarce. Nevertheless, expert consensus in gastroenterology guidelines recommended H pylori screening and eradication for long-standing use of aspirin,^12,13 but such recommendations are not included in current cardiology guidelines^1,2 or suggested as a bleeding avoidance strategy in practice.¹⁴

To date, only one randomized study involving unselected, long-term aspirin users showed that eradication therapy of diagnosed H pylori infection reduces upper gastrointestinal bleeding but did not offer lasting protection.¹⁵ Since the use of potent antithrombotic therapy is associated with a higher risk of upper gastrointestinal bleeding, the primary objective of this study was to investigate in a real-world setting the effect of routine H pylori screening on upper gastrointestinal bleeding in addition to standard care after acute myocardial infarction.

Methods

Study Design and Setting

HELP-MI SWEDEHEART (Helicobacter pylori screening to prevent gastrointestinal bleeding in patients with myocardial infarction, Swedish Web system for enhancement and development of evidence-based care in heart disease¹⁶) was an open-label, 2-period (each lasting 1 year), 2-sequence cluster randomized crossover, registry-based, clinical trial involving 35 Swedish hospitals grouped into 18 clusters based on percutaneous coronary intervention networks. An independent statistician performed randomization before the study started. The clusters were allocated 1 to 1 to implementation of routine H pylori screening, followed by no H pylori screening, or no H pylori screening followed by H pylori screening, using a single computer-generated randomly permuted block, ensuring no more than 20 hospitals were allocated to the same intervention sequence. The latter restriction was motivated by feasibility because the number of available H pylori analyzer units in each study period was limited. H pylori analyzer units were available in hospitals only during the H pylori screening period and were subsequently relocated during a 2-month wash-out period to limit a potential carryover effect. The design and rationale of the trial have been published previously.¹⁷ This trial was investigator-initiated and conducted under the Declaration of Helsinki.¹⁸ Because allocation to H pylori screening was performed in clusters, and not at an individual level, informed consent for randomization was waived. The trial protocol was approved by the Swedish Ethical Review Authority (2020-01885, 2024-03095-02) and is available with the statistical analysis plan in Supplement 1. Details about the trial organization, as well as supplemental material are provided in Supplement 2, and the Data Sharing Statement in Supplement 3. The study is reported according to the Consolidated Standards of Reporting Trials (CONSORT) reporting guidelines.¹⁹

Patients

All adult patients admitted at a participating hospital during the study period and discharged alive with a diagnosis of type 1 myocardial infarction and registered in SWEDEHEART¹⁶ were included. To ensure completeness of follow-up, nonresidents of Sweden were excluded. A list of the inclusion and exclusion criteria is provided in eTable 1 in Supplement 2.

Study Intervention and Procedures

The intervention was the implementation of routine H pylori screening of all patients with type 1 myocardial infarction. H pylori infection was diagnosed using a bedside ₁₃C-urea breath test as part of routine care during the hospitalization period for myocardial infarction. All patients received information about the test result and decisions on eradication treatment based on those findings were at the discretion of the caring physician. A summary of recommended eradication therapy and follow-up were available to all participating units (Supplement 2).

The study organizers supplied all H pylori analysis equipment (Kibion, Dynamic, Diabact, Mayoly Spindler). The individual implementation of H pylori screening, test result, and eradication therapy prescription were recorded in electronic health records and in SWEDEHEART. Control of successful H pylori eradication was recommended to the treating physician (details in Supplement 2) but due to the pragmatic design of the study, no central follow-up was performed. In clusters randomized to no screening, all patients with myocardial infarction received usual care without routine H pylori screening.

Clinical End Points

The primary end point was upper gastrointestinal bleeding. Secondary end points were in a hierarchical order; (1) net adverse clinical events (including upper gastrointestinal bleeding, all-cause death, rehospitalization with myocardial infarction, or hospitalization for ischemic stroke); (2) major adverse cardiac or cerebrovascular events (including cardiovascular death, rehospitalization with myocardial infarction, or hospitalization for ischemic stroke); (3) all-cause death; (4) cardiovascular death; (5) rehospitalization with myocardial infarction; (6) hospitalization for ischemic stroke; and (7) hospitalization for heart failure.

All International Classification of Diseases, 10th Revision (ICD-10) codes for the above-mentioned diagnoses are summarized in eTable 2 in Supplement 2.

Data Sources

Baseline data were collected from SWEDEHEART as previously described,^16,20 with key aspects summarized in Supplement 2.

Upper gastrointestinal bleeding was obtained from the mandatory National Patient Register, including ICD codes from all inpatient and specialized outpatient visits in Sweden. No central adjudication of events was performed.

All-cause death was obtained from the Swedish Population Register,²¹ including the vital and emigration status, and the National Cause of Death Register,²² for cardiovascular death.

Data on readmission for myocardial infarction events during the first month after the index event were collected from SWEDEHEART to the study database and thereafter captured in the National Patient Register.²³

Hospitalizations for heart failure and stroke were obtained from the National Patient Register.

The linkage with the National Patient Register, the National Cause of Death Register, and the National Prescribed Drug Register (prescribed and dispensed treatments) was performed 3 months after the end of the follow-up period, which is the officially recommended time frame to allow for a delay in register reporting.

Sample Size Estimation

Based on a nationwide observational study⁶ and a pilot study involving a contemporary myocardial infarction population,²⁴ we anticipated the prevalence of active H pylori infection to be 20%²⁴ and that the event rate of upper gastrointestinal bleeding would be 2.5% over an average follow-up of 2 years.⁶ Based on consensus in the steering committee, we assumed a 30% relative risk reduction in the group screened for H pylori. Based on these assumptions, a simulation based on a total of 11 544 screened and 11 544 usual care patients using a cluster-summary analysis similar to the present trial yielded an 87% power to detect a hazard ratio of 0.70 when comparing screening with no screening, at a 2-sided α = .05.¹⁷

Statistical Analyses

The primary intention-to-treat analysis was a cluster-summary analysis. Patients were censored at emigration or death, unless part of the end point, in line with the recommendation to use cause-specific hazards for efficacy studies.²⁵ A negative binomial model was fitted with the number of events in each cluster period as the dependent variable, and intervention (screening or no screening), period (first or second), and cluster as independent variables. The results are presented as a rate ratio (RR) for screening with a 95% CI and a 2-sided P value for the null hypothesis of no screening effect. A P value of <.05 was considered statistically significant. The primary analysis using all follow-up time of the primary and secondary end points was adjusted for multiplicity using a hierarchical strategy. Sensitivity analyses included randomization inference based on the admissible randomization sequences, Cox proportional hazards and Poisson models, and modified follow-up (from admission, at 1 year, and 1- to 2-year follow-up, a landmark analysis starting 30 days after discharge).

To explore possible heterogeneity in the intervention effect, we conducted prespecified subgroup analyses of the primary end point without adjustment for multiplicity.

To assess the efficacy of H pylori screening after myocardial infarction under optimal conditions, the per-protocol populations encompass patients with a conclusive H pylori screening result, positive screening result, and positive screening with eradication therapy. These patients were identified in the screening group and compared with the corresponding, but nonidentified, patient group in the cluster periods randomized to no screening. These analyses were carried out using instrumental variable methods,²⁶ based on a flexible proportional hazards model.²⁷

Details regarding the statistical analyses are provided in Supplement 2 and the statistical analysis plan in Supplement 1. All data analyses were performed using R, version 4.4.3, 2025 (R Foundation).

Results

Patients

From November 17, 2021, to November 17, 2022, and January 17, 2023, to January 17, 2024, a total of 18 466 patients with myocardial infarction across 35 hospitals grouped in 18 clusters were discharged alive after randomization, of which 9245 and 9221 patients were followed up after routine H pylori screening and nonscreening periods, respectively (study flowchart, Figure 1; eTable 3 and eFigure 1 in Supplement 2). The baseline characteristics of the patients were well balanced between the 2 groups (Table 1). The median age was 71 (IQR, 61-79) years, (13 138 male [71%]). Overall, 6136 patients (33%) have had a previous myocardial infarction. At admission, 2284 in the screening group and 2275 in the usual care group (both 24.7%) took PPIs. Data completeness was high (eTable 4 in Supplement 2), and no imputation was performed. Cluster-level, trial group, and period baseline characteristics¹⁹ revealed no evidence of heterogeneity (eTable 3, 5, and 6 in Supplement 2).

Figure 1. — ^aA map of participating hospitals is provided in eFigure 1 in Supplement 2, and clusters with cluster periods by participating hospital and number of patients in eTable 3 in Supplement 2. *H pylori* indicates *Helicobacter pylori*.

Table 1. Characteristics of the Patients.

Characteristic	No. (%) of patients
Characteristic	Helicobacter pylori screening (n = 9245)	Usual care (n = 9221)
Demographics
Age, median (IQR), y	71 (61-79)	71 (62-79)
Sex
Female	2636 (28.5)	2692 (29.2)
Male	6609 (71.5)	6529 (70.8)
Risk factors
Hypertension	6384 (69.1)	6401 (69.4)
Diabetes	2472 (26.7)	2416 (26.2)
Previous bleeding	2018 (21.8)	2064 (22.4)
Current smoking	1718 (19.1)	1765 (19.7)
Previous gastroduodenal pathology	1566 (16.9)	1529 (16.6)
Previous anemia	801 (8.7)	790 (8.6)
Chronic kidney disease	722 (7.8)	742 (8.0)
Previous alcohol use disorder	358 (3.9)	340 (3.7)
Previous cardiovascular disease
Myocardial infarction	3124 (33.8)	3012 (32.7)
Percutaneous coronary intervention	2529 (27.4)	2467 (26.8)
Ischemic stroke	561 (6.1)	557 (6.0)
Coronary arterial bypass graft surgery	541 (5.9)	546 (5.9)
Medication at admission
ACE inhibitor or ARB	4078 (45.3)	4086 (45.4)
Statin	3148 (35.0)	3163 (35.2)
β-Blocker	2911 (32.4)	2944 (32.7)
Aspirin	2360 (26.2)	2340 (26.0)
Proton pump inhibitor	2284 (24.7)	2275 (24.7)
Direct oral anticoagulation	862 (9.6)	909 (10.1)
P2Y12 receptor blocker	491 (5.5)	476 (5.3)
Warfarin	130 (1.4)	157 (1.7)
Characteristic at presentation
Systolic blood pressure, median (IQR), mm Hg	147 (130-165)	147 (130-165)
Heart rate, median (IQR), beats/min	78 (67-92)	78 (66-91)
eGFR, CKD-EPI, mean (SD)	74.7 (22.0)	75.0 (21.9)
Anemia at admission, No. (%)
No anemia (hemoglobin ≥120 g/L)	7532 (85.6)	7535 (86.4)
Mild anemia (hemoglobin 100-120 g/L)	1052 (12.0)	1006 (11.5)
Moderate to severe anemia (hemoglobin <100 g/L)	213 (2.4)	181 (2.1)
Kidney failure at admission (eGFR<60 mL/min/1.73m²)	2222 (24.1)	2181 (23.7)
In-hospital course
Helicobacter pylori
Screened	6480 (70.1)
Tested positive	1532 (16.6)
Eradication prescribed	1481 (16.0)
Revascularization	7748 (83.8)	7732 (83.9)
ST–segment elevation myocardial infarction	3400 (36.8)	3488 (37.8)
Median duration of hospital stay, d	4 (3-6)	4 (3-6)
Medication at discharge
Statin	8725 (94.4)	8722 (94.6)
Aspirin	7817 (84.6)	7852 (85.2)
ACE inhibitor or ARB	7718 (83.5)	7632 (82.8)
β-Blocker	7002 (75.8)	6942 (75.3)
Proton pump inhibitor	5207 (56.3)	4550 (49.3)
Ticagrelor	4732 (51.2)	4927 (53.4)
Clopidogrel	2215 (24.0)	2280 (24.7)
Direct oral anticoagulation	1471 (15.9)	1514 (16.4)
Prasugrel	1108 (12.0)	743 (8.1)
Warfarin	156 (1.7)	178 (1.9)
Antithrombotic combination therapy, No. (%)
None	78 (0.8)	84 (0.9)
Single antiplatelet therapy	1011 (10.9)	1013 (11.0)
Dual antiplatelet therapy	6522 (70.6)	6429 (69.7)
Oral anticoagulation monotherapy	129 (1.4)	124 (1.3)
Oral anticoagulation with antiplatelet therapy	1498 (16.2)	1568 (17.0)

Open in a new tab

Abbreviations: ACE, angiotensin-converting enzyme; ARB, angiotensin-receptor blocker; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration equation; eGFR, estimated glomerular filtration rate.

Intervention Adherence and Positive H pylori Testing Rate

Of the 9245 patients with myocardial infarction included during screening periods, 6480 (70%) had undergone testing, 23.6% of whom tested positive for H pylori. Among the 1532 patients who tested positive, 1481 (96.6%) received eradication therapy at the treating physician’s discretion.

Primary and Secondary End Points

Patients were followed up until January 17, 2025, which corresponds to a minimum follow-up of 1 year and a maximum follow-up time of 3 years and 2 months. After a median follow-up of 1.9 years, a total of 635 patients experienced the primary outcome of upper gastrointestinal bleeding, of which 299 (incidence rate, 16.8 events per 1000 person-years; cumulative hazard at 3 years, 4.1%) occurred in the active screening group, and 336 (incidence rate, 19.2 events per 1000 person-years; cumulative hazard at 3 years, 4.6%) in the usual care group.

There was no significant difference between screening and nonscreening clusters in the primary outcome of upper gastrointestinal bleeding at the end of follow-up (RR, 0.90; 95% CI, 0.77-1.05; P = .18; Figure 2 and Table 2).

Figure 2. — The primary end point was upper gastrointestinal bleeding. The median follow-up time in the *H pylori* screening group was 1.91 years, in the usual care group, 1.86 years.

Table 2. Primary and Secondary End Points.

Period	Events, No.	Total follow-up (per 1000 person-years)^a	Incidence rate (per 1000 person-years)	Negative binomial
Period	Events, No.	Total follow-up (per 1000 person-years)^a	Incidence rate (per 1000 person-years)	Rate ratio (95% CI)^b	P value crude^c	P value adjusted^c
Primary end point
Upper gastrointestinal bleeding
Usual care	336	17.52	19.18	0.90 (0.77-1.05)	.18	.18
Screening	299	17.82	16.78	0.90 (0.77-1.05)	.18	.18
Secondary end points
Net adverse clinical event
Usual care	1809	16.62	108.85	1.00 (0.93-1.07)	.97	.97
Screening	1818	16.92	107.42	1.00 (0.93-1.07)	.97	.97
MACCE
Usual care	1236	16.96	72.87	0.97 (0.89-1.05)	.41	.97
Screening	1213	17.27	70.22	0.97 (0.89-1.05)	.41	.97
All-cause death
Usual care	965	17.93	53.82	1.01 (0.92-1.10)	.88	.97
Screening	968	18.21	53.17	1.01 (0.92-1.10)	.88	.97
Cardiovascular death
Usual care	512	17.93	28.56	1.00 (0.88-1.13)	.97	.97
Screening	511	18.21	28.07	1.00 (0.88-1.13)	.97	.97
Myocardial infarction
Usual care	696	17.11	40.67	0.95 (0.85-1.06)	.38	.97
Screening	678	17.44	38.88	0.95 (0.85-1.06)	.38	.97
Ischemic stroke
Usual care	174	17.75	9.80	0.97 (0.78-1.21)	.80	.97
Screening	177	18.03	9.82	0.97 (0.78-1.21)	.80	.97
Heart failure
Usual care	490	17.43	28.11	0.97 (0.86-1.11)	.68	.97
Screening	485	17.69	27.42	0.97 (0.86-1.11)	.68	.97

Open in a new tab

Abbreviation: MACCE, major adverse cardiac or cerebrovascular event.

^{^a}

Cluster-summary analysis of the primary and secondary end points using a negative binomial analysis.

^{^b}

Median follow-up time of 1.9 years (minimum-maximum, 1-3.2 years). The primary analyses of the primary and secondary end points were adjusted for multiplicity using a hierarchical strategy, ordering the end points as above.

^{^c}

The hierarchical strategy means that the adjusted P value is the largest of the current crude P value and all the crude P values above.

No significant between-group differences were observed in any composite or individual secondary end points at the end of follow-up (Table 2).

Sensitivity Analyses

Baseline characteristics of the 19 407 patients with myocardial infarction randomized (including the 941 patients deceased during hospitalization) were well-balanced between the groups. As expected, patients were slightly older with more baseline comorbidities (eTable 7 in Supplement 2) compared with the primary study population. Sensitivity analysis using randomization inference, Cox proportional hazard, or Poisson models generated similar primary end point measures (eTable 8 in Supplement 2) as did predefined analyses using a modified follow-up (from admission, at 30 days, at 1 year, and 1- to 2-year follow-up, a landmark analysis starting 30 days after discharge; eTable 9 in Supplement 2) corroborating robustness of the primary analysis.

Subgroup Analyses

To explore potential heterogeneity in the effect of the H pylori screening, predefined subgroup analyses identified a higher risk of upper gastrointestinal bleeding and a lower upper gastrointestinal bleeding RR after H pylori screening among patients with acute myocardial infarction with concomitant anemia (mild anemia, defined as hemoglobin 100-120 g/L; RR, 0.64; 95% CI, 0.42-0.98; moderate to severe anemia, defined as hemoglobin <100 g/L; RR, 0.45; 95% CI, 0.23-0.87; P for interaction = .03; Figure 2; eTable 10 in Supplement 2). In addition, a lower RR of upper gastrointestinal bleeding was found in patients with concomitant kidney failure (defined as estimated glomerular filtration rate [eGFR] <60 mL/min/1.73 m²; RR, 0.75; 95% CI, 0.58-0.97), whereas no difference between screening and nonscreening periods was observed in patients with normal kidney function (RR, 1.00; 95% CI, 0.82-1.23; P for interaction = .08; Figure 3; eTable 10 in Supplement 2). Whereas PPI use at admission did not affect the subsequent upper gastrointestinal bleeding risk significantly (Figure 3; P for interaction = .51, eTable 10 in Supplement 2), a post hoc subgroup analysis stratified according to PPI use at discharge showed a significant risk reduction for upper gastrointestinal bleeding in non-PPI users (RR, 0.74; 95% CI, 0.58-0.94; P for interaction = .04) in the screening group (eFigure 2, eTable 10 in Supplement 2).

Figure 3. — Mild anemia was defined as hemoglobin 100-120 g/L; moderate to severe anemia was defined as hemoglobin <100 g/L; kidney failure was defined as estimated glomerular filtration rate of less than 60 mL/min/1.73 m². STEMI indicates ST segment myocardial infarction.

Per-Protocol Populations

Baseline characteristics for the per-protocol populations were well-balanced between the groups and were similar to the primary study population (eTable 11 in Supplement 2). Lower, but statistically nonsignificant different point estimates for upper gastrointestinal bleeding were found in per-protocol analyses in patients who had been tested for H pylori (hazard ratio [HR], 0.84; 95% CI, 0.65-1.08; P = .16), tested positive for H pylori (HR, 0.47; 95% CI, 0.22-1.01; P = .05), and tested positive for H pylori positive and had received eradication (HR, 0.49; 95% CI, 0.23-1.06; P = .07) compared with a corresponding population in the usual care group. Even point estimates for secondary end points were slightly lower (eTable 12 in Supplement 2) compared with results from the primary intention-to-treat analysis.

Discussion

In this nationwide cluster randomized, crossover, registry-based clinical trial, routine H pylori screening with urea breath test during hospitalization for acute myocardial infarction did not result in a statistically significant reduction of upper gastrointestinal bleeding over a median follow-up of almost 2 years. In this real-world setting, among clusters assigned to H pylori screening, 70% of patients underwent testing, with an active infection prevalence of 23.6%. Results from subgroup analyses and analyses in secondary per-protocol populations of patients tested for H pylori may indicate a screening effect in patients at higher risk of upper gastrointestinal bleeding.

Several factors specific to the trial design and study setting should be considered when interpreting the results. First, the H pylori prevalence, while close to anticipated among tested patients, may not have been sufficient to drive changes in upper gastrointestinal bleeding incidence in an unselected myocardial infarction population. Geographical differences have been reported with a contemporary worldwide H pylori prevalence of 56.7% in cardiovascular disease populations³—more than twice the prevalence observed in the present Swedish myocardial infarction population. A potential benefit of routine screening during myocardial infarction hospitalization in regions with higher H pylori prevalence cannot be excluded. Second, adherence to screening was 70%, and the intention-to-treat analysis may, therefore, underestimate the actual effect of screening in the overall population. Importantly, higher adherence to routine screening in a real-world setting is unlikely. The per-protocol analysis, limited to patients who were tested for H pylori, revealed lower point estimates for the primary end point but with the 95% CI still including one. Notwithstanding, the more positive effect observed in the per-protocol analysis may suggest a potential benefit of the intervention. Third, the relatively high use of PPI at admission in one-fourth of the patients may have interfered with the urea breath test accuracy,²⁸ leading to false-negative test results. Additionally, preventive PPI use at discharge in nearly half of the patients with myocardial infarction may have reduced the additional benefit of routine H pylori screening. However, H pylori screening significantly reduced upper gastrointestinal bleeding in patients with an acute myocardial infarction that had an unknown upper gastrointestinal bleeding risk not motivating PPI prescription in the absence of H pylori screening results.

Subgroup analyses suggested clinically meaningful heterogeneity in line with risk groups identified in a previous study.²⁹ Among patients with anemia, particularly those with moderate to severe anemia, H pylori screening was associated with a lower risk of upper gastrointestinal bleeding. A similar benefit of screening was observed in patients with concurrent kidney failure at admission, whereas those with normal kidney function appeared to benefit less from H pylori screening. These results may suggest that selected high-risk subgroups derive greater benefit from targeted H pylori testing and support a risk-stratified approach to H pylori screening after acute myocardial infarction.

This study has several strengths, including its cluster randomized design, broad inclusion criteria of a nationwide, unselected myocardial infarction population, and a large sample size, which enhances the generalizability of the findings.

Limitations

This trial has several limitations. First, despite the large overall sample size, the number of myocardial infarction hospitalizations and the effect size were somewhat lower than anticipated in the power calculations. Second, due to the pragmatic design of this trial relying on ICD codes from National registers²³ without central adjudication, misclassification cannot be ruled out. However, diagnosed hemorrhages from inpatient and specialized outpatient care²³ have a high validity for gastrointestinal bleeding events.³⁰ Third, although treatment was not part of the intervention and that its prescription, and treatment control was at the attending physician’s discretion and not centrally followed up, a summary of current eradication therapy and follow-up recommendations were provided and almost all patients who tested positive for H pylori received a prescription of eradication treatment (96%). Likewise, H pylori testing and treatment during follow-up by general practitioners was not captured, which may have altered the baseline risk profile of the usual care period. Fourth, the crossover design could introduce period by intervention effects, although no apparent changes were observed. In addition, possible carry-over effects were minimized by a washout period before crossover and the removal of screening equipment during nonscreening periods. Fifth, subgroup analyses, though prespecified, should be interpreted with caution given the risk of false-positive findings in multiple comparisons. Specifically, the source of baseline anemia was not determined and could also reflect lower gastrointestinal blood loss or nonbleeding causes.

Conclusions

In this nationwide cluster-randomized, crossover, registry-based clinical trial among unselected patients with acute myocardial infarction, routine H pylori screening did not significantly reduce the risk of the primary end point of upper gastrointestinal bleeding.

Supplement 1.

Study Protocol and Statistical Analysis Plan

jama-e2515047-s001.pdf^{(4MB, pdf)}

Supplement 2.

Trial Organization

eTable 1. Inclusion and exclusion criteria of primary intention to treat analysis population

eTable 2. End points and corresponding ICD codes

eTable 3. Clusters with cluster periods by participating hospital

eTable 4. Missing or unknown baseline data

eTable 5. Characteristics of the patients on cluster level

eTable 6. Characteristics of the patients by trial arm and period

eTable 7. Characteristics of all enrolled patients including those diseased during hospitalization period

eTable 8. Sensitivity analyses including randomization inference, Cox and Poisson models on primary and secondary end points

eTable 9. Sensitivity analyses of the primary outcomes upper gastrointestinal bleeding with modified follow-up

eTable 10. Subgroup interaction analyses for primary endpoint censoring at 1–3 years

eTable 11. Characteristics of the patients in per-protocol populations

eTable 12. Primary and secondary end points in per-protocol populations

eFigure 1. Map of Sweden with participating hospitals by cluster period

eFigure 2. Upper gastrointestinal bleeding in predefined subgroup analyses including post hoc analysis of PPI use at discharge

eReferences

jama-e2515047-s002.pdf^{(3.2MB, pdf)}

Supplement 3.

Data Sharing Statement

jama-e2515047-s003.pdf^{(15.9KB, pdf)}

References

1.Rao SV, O’Donoghue ML, Ruel M, et al. 2025 ACC/AHA/ACEP/NAEMSP/SCAI Guideline for the management of patients with acute coronary syndromes: a report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. Circulation. 2025;151(13):e771-e862. [DOI] [PubMed] [Google Scholar]
2.Byrne RA, Rossello X, Coughlan JJ, et al. ; ESC Scientific Document Group . 2023 ESC Guidelines for the management of acute coronary syndromes. Eur Heart J. 2023;44(38):3720-3826. doi: 10.1093/eurheartj/ehad191 [DOI] [PubMed] [Google Scholar]
3.Yaslianifard S, Sameni F, Kazemi K, et al. Beyond the gut: a comprehensive meta-analysis on Helicobacter pylori infection and cardiovascular complications. Ann Clin Microbiol Antimicrob. 2025;24(1):18. doi: 10.1186/s12941-025-00788-6 [DOI] [PMC free article] [PubMed] [Google Scholar]
4.Simonsson M, Wallentin L, Alfredsson J, et al. Temporal trends in bleeding events in acute myocardial infarction: insights from the SWEDEHEART registry. Eur Heart J. 2020;41(7):833-843. doi: 10.1093/eurheartj/ehz593 [DOI] [PubMed] [Google Scholar]
5.Généreux P, Giustino G, Witzenbichler B, et al. Incidence, predictors, and impact of post-discharge bleeding after percutaneous coronary intervention. J Am Coll Cardiol. 2015;66(9):1036-1045. doi: 10.1016/j.jacc.2015.06.1323 [DOI] [PubMed] [Google Scholar]
6.Sarajlic P, Simonsson M, Jernberg T, Bäck M, Hofmann R. Incidence, associated outcomes, and predictors of upper gastrointestinal bleeding following acute myocardial infarction: a SWEDEHEART-based nationwide cohort study. Eur Heart J Cardiovasc Pharmacother. 2022;8(5):483-491. doi: 10.1093/ehjcvp/pvab059 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Bonaca MP, Bhatt DL, Cohen M, et al. ; PEGASUS-TIMI 54 Steering Committee and Investigators . Long-term use of ticagrelor in patients with prior myocardial infarction. N Engl J Med. 2015;372(19):1791-1800. doi: 10.1056/NEJMoa1500857 [DOI] [PubMed] [Google Scholar]
8.Sehested TSG, Carlson N, Hansen PW, et al. Reduced risk of gastrointestinal bleeding associated with proton pump inhibitor therapy in patients treated with dual antiplatelet therapy after myocardial infarction. Eur Heart J. 2019;40(24):1963-1970. doi: 10.1093/eurheartj/ehz104 [DOI] [PubMed] [Google Scholar]
9.Hofmann R, Bäck M. Time for routine helicobacter pylori screening in coronary artery disease? Circulation. 2023;147(23):1731-1733. doi: 10.1161/CIRCULATIONAHA.123.064944 [DOI] [PMC free article] [PubMed] [Google Scholar]
10.Saltzman JR. The challenge of nonvariceal upper GI bleeding management in patients with acute coronary syndrome receiving dual-antiplatelet therapy. Gastrointest Endosc. 2020;92(1):75-77. doi: 10.1016/j.gie.2020.03.012 [DOI] [PubMed] [Google Scholar]
11.Venerito M, Schneider C, Costanzo R, Breja R, Röhl FW, Malfertheiner P. Contribution of Helicobacter pylori infection to the risk of peptic ulcer bleeding in patients on nonsteroidal anti-inflammatory drugs, antiplatelet agents, anticoagulants, corticosteroids and selective serotonin reuptake inhibitors. Aliment Pharmacol Ther. 2018;47(11):1464-1471. doi: 10.1111/apt.14652 [DOI] [PubMed] [Google Scholar]
12.Malfertheiner P, Megraud F, Rokkas T, et al. ; European Helicobacter and Microbiota Study group . Management of Helicobacter pylori infection: the Maastricht VI/Florence consensus report. Gut. 2022;gutjnl-2022-327745. doi: 10.1136/gutjnl-2022-327745 [DOI] [PubMed] [Google Scholar]
13.Chey WD, Howden CW, Moss SF, et al. ACG clinical guideline: treatment of Helicobacter pylori infection. Am J Gastroenterol. 2024;119(9):1730-1753. doi: 10.14309/ajg.0000000000002968 [DOI] [PubMed] [Google Scholar]
14.Capodanno D, Bhatt DL, Gibson CM, et al. Bleeding avoidance strategies in percutaneous coronary intervention. Nat Rev Cardiol. 2022;19(2):117-132. doi: 10.1038/s41569-021-00598-1 [DOI] [PubMed] [Google Scholar]
15.Hawkey C, Avery A, Coupland CAC, et al. ; HEAT Trialists . Helicobacter pylori eradication for primary prevention of peptic ulcer bleeding in older patients prescribed aspirin in primary care (HEAT): a randomised, double-blind, placebo-controlled trial. Lancet. 2022;400(10363):1597-1606. doi: 10.1016/S0140-6736(22)01843-8 [DOI] [PubMed] [Google Scholar]
16.Jernberg T, Attebring MF, Hambraeus K, et al. The Swedish Web-system for Enhancement and Development of Evidence-Based Care in Heart Disease Evaluated According to Recommended Therapies (SWEDEHEART). Heart. 2010;96(20):1617-1621. doi: 10.1136/hrt.2010.198804 [DOI] [PubMed] [Google Scholar]
17.Hofmann R, James S, Sundqvist MO, et al. Helicobacter pylori screening to prevent gastrointestinal bleeding in patients with acute Myocardial Infarction (HELP-MI SWEDEHEART)—design and rationale of a cluster randomized, crossover, registry-based clinical trial. Am Heart J. 2025;286:66-74. doi: 10.1016/j.ahj.2025.03.014 [DOI] [PubMed] [Google Scholar]
18.World Medical Association . World Medical Association Declaration of Helsinki: ethical principles for medical research involving human participants. JAMA. 2025;333(1):71-74. doi: 10.1001/jama.2024.21972 [DOI] [PubMed] [Google Scholar]
19.McKenzie JE, Taljaard M, Hemming K, et al. Reporting of cluster randomised crossover trials: extension of the CONSORT 2010 statement with explanation and elaboration. BMJ. 2025;388:e080472. doi: 10.1136/bmj-2024-080472 [DOI] [PMC free article] [PubMed] [Google Scholar]
20.Leosdottir M, Dahlbom L, Bäck M, et al. Data quality assessment in the SWEDEHEART registry: Insights from serial audits on completeness and accuracy. Am Heart J. 2025;288:149-158. doi: 10.1016/j.ahj.2025.04.028 [DOI] [PubMed] [Google Scholar]
21.Ludvigsson JF, Almqvist C, Bonamy AK, et al. Registers of the Swedish total population and their use in medical research. Eur J Epidemiol. 2016;31(2):125-136. doi: 10.1007/s10654-016-0117-y [DOI] [PubMed] [Google Scholar]
22.Brooke HL, Talbäck M, Hörnblad J, et al. The Swedish cause of death register. Eur J Epidemiol. 2017;32(9):765-773. doi: 10.1007/s10654-017-0316-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
23.Ludvigsson JF, Andersson E, Ekbom A, et al. External review and validation of the Swedish national inpatient register. BMC Public Health. 2011;11:450. doi: 10.1186/1471-2458-11-450 [DOI] [PMC free article] [PubMed] [Google Scholar]
24.Wärme J, Sundqvist M, Mars K, et al. Helicobacter pylori screening in clinical routine during hospitalization for acute myocardial infarction. Am Heart J. 2021;231:105-109. doi: 10.1016/j.ahj.2020.10.072 [DOI] [PubMed] [Google Scholar]
25.Troendle JF, Leifer ES, Kunz L. Dealing with competing risks in clinical trials: How to choose the primary efficacy analysis? Stat Med. 2018. doi: 10.1002/sim.7800 [DOI] [PMC free article] [PubMed] [Google Scholar]
26.Altstein LL, Li G, Elashoff RM. A method to estimate treatment efficacy among latent subgroups of a randomized clinical trial. Stat Med. 2011;30(7):709-717. doi: 10.1002/sim.4131 [DOI] [PMC free article] [PubMed] [Google Scholar]
27.Royston P, Parmar MK. Flexible parametric proportional-hazards and proportional-odds models for censored survival data, with application to prognostic modelling and estimation of treatment effects. Stat Med. 2002;21(15):2175-2197. doi: 10.1002/sim.1203 [DOI] [PubMed] [Google Scholar]
28.Gisbert JP, Pajares JM. Review article: 13C-urea breath test in the diagnosis of Helicobacter pylori infection–a critical review. Aliment Pharmacol Ther. 2004;20(10):1001-1017. doi: 10.1111/j.1365-2036.2004.02203.x [DOI] [PubMed] [Google Scholar]
29.Moukarbel GV, Signorovitch JE, Pfeffer MA, et al. Gastrointestinal bleeding in high risk survivors of myocardial infarction: the VALIANT trial. Eur Heart J. 2009;30(18):2226-2232. doi: 10.1093/eurheartj/ehp256 [DOI] [PubMed] [Google Scholar]
30.Friberg L, Skeppholm M. Usefulness of health registers for detection of bleeding events in outcome studies. Thromb Haemost. 2016;116(6):1131-1139. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials