Helicobacter pylori infection treatment in the United States: clinical consequences and costs of eradication treatment failure

ABSTRACT

Introduction

Helicobacter pylori (Hp) is causal in benign and malignant gastrointestinal diseases. Accordingly, current guidelines recommend Hp eradication in patients with active infection. Unfortunately, treatment failure is common, exposing patients to complications associated with persistent Hp infection and consequences of repeated treatment, including promotion of antibiotic resistance. In the United States (US), data regarding eradication rates with available therapies are limited. Moreover, the clinical and economic burden of eradication treatment failure have not been thoroughly described.

Areas covered

We aimed to characterize Hp eradication rates and the clinical consequences and associated costs of persistent Hp infection among US adults. We conducted focused literature reviews using initial searches in Embase, MEDLINE, and Cochrane Database of Systematic Reviews via Ovid followed by manual searches to identify relevant publications.

Expert opinion

Hp eradication rates were suboptimal, with most studies reporting rates ≤80% with clarithromycin-based triple therapy and bismuth quadruple therapy. There was direct evidence supporting numerous benefits of successful Hp eradication, including decreased risk of recurrent or complicated peptic disease and non-cardia gastric cancer. Cost benefits of eradication were related to mitigation of conditions associated with persistent Hp infection, (e.g. complicated peptic ulcer disease, and gastric cancer) which altogether exceed US$5.3 billion.

KEYWORDS:

• Bismuth quadruple therapy
• clarithromycin triple therapy
• eradication
• gastric cancer
• Helicobacter pylori
• peptic ulcer disease
• proton pump inhibitor

1.0 Introduction

Helicobacter pylori (Hp) is a gram-negative bacterium estimated to infect more than 50% of the world’s population (4.4 billion people in 2015) and is the most common chronic bacterial infection. Hp is a human carcinogen that is responsible for 5.5% of all malignancies and 80–90% of the global burden of gastric cancer (GC) – the third leading cause of cancer-related death [1]. Persistent Hp infection is therefore considered a major threat to public health, including by the World Health Organization (WHO) [2,3]. Hp prevalence varies geographically and on the basis of industrialization, with the highest rates observed in Asian Pacific regions, Eastern Europe, and Central/Latin America, while the lowest rates overall are reported in North America [4]. However, there is also marked within-country variation, for example, based on race/ethnicity, immigrant status, geography, rural vs urban environment, and socioeconomic factors [5–7]. While the United States (US) is considered a low-prevalence country overall, non-White races and ethnic groups may have Hp prevalence up to, and in certain groups in excess of, 50% [8], compared to the non-Hispanic White population, among whom prevalence is estimated to be <15% [4].

Based on data from endemic countries, Hp infection is most often acquired in childhood and, unless eradicated, generally persists for the lifetime of the host [9,10]. Hp infection invariably causes chronic gastric inflammation; most often, this does not result in major clinical complications although it can result in bothersome nonspecific GI symptoms, such as dyspepsia [11]. Importantly, though, a minority of patients will develop serious and potentially life-threatening conditions, including complicated peptic ulcer disease (PUD), non-cardia GC (NCGC), and mucosa-associated lymphoid tissue (MALT) lymphoma [12–14]. The potential economic cost of Hp-associated GI disease morbidity and mortality is enormous. In 2016, nearly $2 billion in health-care expenditures were for PUD, gastritis, and duodenitis alone [15].

Because of the established role of Hp in malignant and benign disease, international guidelines, including the American College of Gastroenterology (ACG), recommend Hp eradication when active infection is diagnosed [14,16,17]. Eradication treatment consists of one or more antibiotics in combination with a potent acid-suppressing regimen [18–20]. In the US, common regimens are clarithromycin-based triple therapy consisting of a proton pump inhibitor (PPI) along with clarithromycin and amoxicillin (PCA) or metronidazole (PCM) or bismuth-quadruple therapy (i.e. PPI, bismuth, a nitroimidazole such as metronidazole, and tetracycline [PBMT]), though rifabutin-triple therapy (PPI plus high-dose amoxicillin and rifabutin), amoxicillin dual therapy (PPI plus high-dose amoxicillin), and other regimens are also used [16,17]. Recommended regimens for Hp eradication vary globally with the main differences being duration (i.e. in Asia Pacific many regimens are 7 days whereas most are 10–14 days in the US) and choice of acid suppressant (both PPIs and potassium-competitive acid blockers [PCABs] and are used in Asia and some other non-US geographies) [14,16,17,21].

Successful eradication of Hp is proven to decrease GC-related mortality and recurrent PUD complications. However, eradication failure is increasingly common and multifactorial. Rising rates of Hp antibiotic resistance are a leading reason, but other patient-related, genetic, and environmental factors certainly contribute [17,22]. Patients with Hp eradication failure should be prescribed an alternative first- or second-line treatment, although repeated attempts generally have a lower likelihood of success [14,16,17]. Furthermore, Hp eradication failure subjects patients to ongoing risk of Hp-associated complications, not to mention the inconvenience, cost, and complications associated with repeated treatment.

Outside the US, monitoring programs such as the European Registry on Helicobacter pylori Management provide surveillance of Hp resistance and management patterns [23,24]. The US had a country-wide surveillance program known as H. pylori Antimicrobial Resistance Monitoring Program (HARP); however, surveillance data have not been available post-2002 [25]. While global data demonstrate Hp eradication rates are declining [14,16,26–28], there is a dearth of epidemiological data from the US, particularly in recent years. A network meta-analysis of eradication therapies by Rokkas et al. demonstrated a pooled overall cure rate of 81.3% (79.5–83.1) with bismuth quadruple therapy and 75.7% (74.9–76.4) with clarithromycin-based triple therapy. Though analyses of grouped Western, Eastern Asian, and West Asian countries were conducted, evidence was insufficient to conduct a US analysis [29].

To address these clear knowledge gaps in the US and inform future research directions, we conducted a series of focused literature reviews to identify and synthesize evidence on Hp eradication success rates using US-guideline recommended Hp eradication regimens. We secondarily aimed to describe the clinical and economic impact of failed Hp eradication in the US; for example, the resulting rates, risk, resource utilization, and costs of PUD complications, gastric precancerous conditions (e.g. gastric intestinal metaplasia), and NCGC.

1. Methods

We developed three key research questions based on an initial landscape assessment of the literature in Hp to understand the existing evidence and identify gaps. The research questions were vetted by clinical experts in gastroenterology and Hp treatment and outcomes. Once the initial questions were finalized, we selected outcomes of interest that would provide relevant data to help inform our search strategy and aid in study selection (Table 1).

Table 1. Search question and outcomes of interest

Question	Relevant outcomes of interest
What are overall and regimen-specific Hp eradication rates in the US? How has this varied over time and geography?	Regimen-specific eradication rates, study location, study timeframe
What is the evidence describing the clinical consequences of persistent Hp infection?	Prevalence of conditions for which Hp is known to be a causal factor, including PUD, both complicated and uncomplicated, precancerous conditions, progression, and NCGC.^24 Association between eradication status (e.g. success vs failure) or Hp status (positive vs negative) and clinical conditions
What are the costs associated with Hp eradication treatment failure and/or Hp-associated diseases?	Healthcare resource utilization, costs, aggregate costs, and cost drivers associated with either Hp eradication treatment failure or known Hp-associated conditions

NCGC, non-cardia gastric cancer; PUD, peptic ulcer disease.

1.1. Search strategy and study selection

For each research question, we performed a comprehensive, but targeted literature search using an iterative hybrid of ‘pearl growing’ and ‘snowball’ search methods [30–32]. Briefly, consistent with this methodologic approach, we conducted an initial broad, structured search to identify publications providing data relevant to the specific outcomes of interest for each research question, followed by ‘pearl-growing’ and ‘snowball’ citation chasing methods as described below. The initial search strategies included a combination of indexed and free-text terms for H. pylori, eradication therapy, and outcomes of interest (see Supplementary Tables 1–3 for initial search strategies). We conducted these initial searches in Embase, MEDLINE, MEDLINE In-Process, Cochrane Central Register of Controlled Trials, and Cochrane Database of Systematic Reviews via Ovid.

Studies were selected based on their relevance to the specific research question. We focused on studies reporting data from adults in the US. Initial searches were limited to articles in English published in the last 10 years (January 2011–April 2021); however, in instances where there was a paucity of high-quality data from the US in the last 10 years (i.e. Question 1), we expanded our search to include publications from prior years (2000 onward). While we were most interested in real-world evidence (RWE), we also included other sources of high-quality data including published randomized controlled trials (RCTs) and meta-analyses when available evidence was limited. In addition, free-text terms were used to narrow the focus of the searches to studies conducted in the US.

We next applied ‘pearl-growing’ and ‘snowball’ citation chasing methods to the studies that were identified in the initial search. Such methods involve performing prospective searches (e.g. identify articles that cite the selected study using the ‘cited by’ search listing in PubMed) as well as retrospective bibliographic searches of key publications (e.g. identified clinical guidelines, meta-analyses) to ensure that all key studies from the US were captured. Additionally, we also identified new index terms and used them to conduct targeted manual searches on PubMed. This multi-step process ensures maximal capture of relevant studies [30–32].

Each search and study selection was conducted by a single reviewer (EH or LV) and relevant outcomes for each research question were captured. Study population size, in addition to other relevant study details, was captured, but no exclusions were made based on study sample size.

1.2. Data analysis

To evaluate whether eradication rates changed significantly over time, we conducted a linear regression model using reported eradication rates weighted by sample size with treatment timeframe as the predictor variable. We assessed for fit via ANOVA using JMP®, Version 16.1. SAS Institute Inc., Cary, NC, 1989–2021. All graphs were created in Microsoft Excel®.

2. Results

2.1. Research Question 1: What are overall and regimen-specific Hp eradication rates in the US? How have they varied over time and geography?

A total of 20 publications for 18 distinct study cohorts that reported eradication rates in the US were identified (see Figure 1) [33–50]. Overall, these studies included 7075 people; 4414 (62.4%) were treated with clarithromycin-based triple therapy, 1450 (20.5%) bismuth quadruple therapy, 236 (3.3%) received rifabutin triple therapy, 227 (3.2%) received high-dose amoxicillin dual therapy, 309 (4.4%) received another treatment regimen. Treatment cohorts in these studies were generally small, with a median of 122 patients (range: 2–1228).

Figure 1. Eradication rates in identified US clinical trials and real-world studies (a) Eradication rates (crude) from identified clinical and observational studies. (b) Eradication rates reported in identified clinical trials (weighted by sample size). (c) Eradication rates reported in observational ‘real-world’ studies (weighted by sample size). (d) US map of reported eradication rates from 2000 to 2021CT, clinical trial; RWE, real-world evidence; US, United States.

Among identified US observational studies and clinical trials of clarithromycin-based triple therapy, bismuth quadruple therapy, rifabutin triple therapy, and levofloxacin-based therapy none reported eradication rates above 90%. In the US, over the past two decades, 10 out of 14 study arms reported real-world eradication rates with clarithromycin-based triple therapy below 80% see (Table 2). According to a retrospective analysis of patients with treatment-naïve Hp infection treated with PCA or PCM for 10–14 days between 2001 and 2019 at a single center in Michigan (N = 1058), the overall eradication rate was 78.1% (range: 81.5% [2001–2005] to 79.1% [2016–2019]) [41]. In a large, multicenter retrospective analysis (N = 1966) of patients diagnosed and treated for Hp infection at US military medical facilities nationwide between 2015 and 2018, the eradication rate with clarithromycin-based triple therapy was 76% [34,45]. Across 14 total studies of 10-day or 14-day clarithromycin-based triple therapies, eradication rates ranged from 72.0% to 88.0% and 58.1% to 88.6%, respectively. Notably, the proportion of patients treated with clarithromycin-based triple therapy varied, ranging from 25% (n/N: 262/1101) to 84.6% (88/104) between 2010 and 2018 [33,44,45,49]. In 2017, the ACG guideline was updated to recommend that clarithromycin-based triple therapy only be used when local resistance to clarithromycin is <15% or the patient is confirmed to be colonized with clarithromycin-sensitive Hp. Only one study included in the present analysis was conducted after the 2017 ACG guidelines publication; in this study (2018–2019), 42.8% (80/187) were prescribed clarithromycin-based triple therapy [34].

Table 2. Observational studies describing Hp regimen-specific eradication rates in the US

						Clarithromycin triple therapy		Bismuth quadruple regimen
Study	Study type	Geography/ setting	Patient demographics	Time period	Treatment length (days)	Eradication rate (%)	n	Eradication rate (%)	n	Eradication testing	Factors associated with eradication outcome
Ginnebaugh 2020 [41]	Retrospective single center	Michigan Large academic medical center	Mean age: 51.1 years (range 18–90) Female: 61.8% white: 49.5%	2001–2005	10 or 14	81.5	189	–		Testing by UBT posttreatment	Adherence did not significantly affect eradication rates
				2006–2010	14	78.5	376	–	–
				2011–2015	10 or 14	77	239	–	–
				2016–2019	10 or 14	79.1	254	–	–
Rubin 2018 [49]	Retrospective single center	California Academic	Median age in those with eradication testing: 60 years Female: 50% White: 36.6% African American: 16.1% Hispanic: 22.8% Asian: 16.1% Gastric ulcer: 38.4% Duodenal ulcer: 21.0%	2005–2015	NR	77	115	78	9	Gastric biopsy, stool antigen test, and/or UBT ≥2 months after treatment	Similar eradication rates regardless of sex, race, and ethnicity (all p-values >0.19)
Liu 2018 [33]	Retrospective single center	Texas Academic	Mean age: 53 years Female: 76% Hispanic: 85% Mean BMI: 30.5 kg/m^2	2010–2015	14	88.6	88	75^a	16	Testing by UBT and posttreatment UBTs	Significant association between female sex and higher eradication rates (p = 0.002) Numeric association with higher BMI and treatment failure (p = 0.08)
Nayar 2018 [47]	Retrospective single-center urban	New Jersey Urban	Age range: 21–76 years Female: 62% White: 44% Hispanic or Latino: 35% Asian: 34% Black: 20% Alaskan Native: 1% Clarithromycin exposure in previous 2 years: 6% PPI exposure in previous 2 months: 34%	2011–2017	10	84	156	–	–	Follow-up UBT within 1–3 months	Negative trends not observed with prior clarithromycin exposure
Alsamman 2019 [44]	Retrospective multi-center	Rhode Island Academic	Mean age: 43–49 years Female: 64% Hispanic: 35–42% White: 13–29% Black: 4–21% Asian/Native American: 0–2%	2015–2017	10	67	101	77^a	135	Endoscopic biopsy, UBT, or stool antigen test	No significant differences in age, sex, or ethnicity were observed between eradicated vs non-eradicated patients
					14	79	161	87^a	585
Barrett-Englert 2019 [48]	Retrospective	Illinois Academic health system	Median age: 55 Female: 62%	2015–2017	NR	72	194	71	37	Not reported	–
Mertz 2020 [45]	Retrospective multicenter	Nationwide US and OCONUS; military medical facilities	Female: 60.4%	2015–2018	NR	76	1228	80^a	215	UBT or stool antigen test	Active-duty patients less likely to achieve eradication than non-active-duty patients (70.6% vs 75.5% vs p = 0.045)
Camero 2019 [46]	Retrospective single center	Texas Academic	Mean age: 53.4 years Hispanic: 63.3% No medical insurance: 44.1%	2016–2016	NR	81	21	NR	3	Not reported	–
Tariq 2020 [51]	Retrospective single center	New York Urban	Mean age: 50.41–56.82 years Female: 65.1–76.5% Hispanic: 50.5–58.7% African American: 23.5–31.7% White: 4.76–23.5% Asian: 0–4.76% BMI >30 kg/m^2: 36.5–48.5%	2017–2018	14	78.3	97	–	–	Fecal antigen testing (preferred), 4 weeks post-therapy	Sex, comorbidities (hypertension, diabetes mellitus), obesity, PUD, and smoking were not associated with treatment failure (all p-values >0.1)
Argueta 2021 [34]	Retrospective multicenter	Rhode Island	Mean age: 45.2 years (range 3–92) Female: 31.7% Hispanic: 43.9% White: 29.1% Black: 23.8% Asian: 2.1%	2018–2019	14	60.5	84	88.1 89.9^a	101	UBT, stool antigen, or biopsy	Clarithromycin-resistant strain was associated with treatment failure (p < 0.001)

^aEradication rate for PBMT regimen.

BMI, body mass index; n, number; NR, not reported; OCONUS, outside the continental US; PUD, peptic ulcer disease; UBT, urea breath test; US, United States.

Studies of other regimens, including bismuth quadruple therapy (doxycycline substituted for tetracycline), rifabutin triple therapy, levofloxacin-based triple therapy, and bismuth-based triple therapies with tetracycline or doxycycline were also identified (Table 3) [34,36–38,42,43,45]. In observational studies, bismuth quadruple therapy was associated with eradication rates ranging from 71% to 88% (see Table 2) [33,34,44,45,48,49]. Among identified clinical trials employing PBMT, eradication rates ranged from 71% to 89% [36,37,39,52]. Rifabutin triple therapy (amoxicillin 3 g, omeprazole 120 mg, and rifabutin delayed-release 150 mg, every 8 h for 14 days), which, prior to its approval as first-line treatment in November 2019, was used predominantly as salvage, was associated with an 83.8% eradication rate vs 57.7% in the comparison arm (amoxicillin 3 g and omeprazole 120 mg, every 8 h for 14 days) in the phase 3 ERADICATE Hp2 trial of treatment-naïve patients [43]. Rifabutin-triple therapy use was reported in two retrospective observational studies: one of two hospitals in Rhode Island (2 out of 187 patients) and at various military medical facilities throughout the US and OCONUS (6 patients out of 1966). Both studies, which recruited patients between 2015 and 2019, reported eradication rates of 50% with rifabutin triple therapy (Table 3) [34,45]. Levofloxacin-based regimens were reported in two studies, with Hp eradication success rates ranging from a high of 90% in a randomized single-center, open-label trial in 2011, to a low of 64% in a retrospective observational study of patients treated between 2015 and 2018 [42,45].

Table 3. Other Hp eradication regimens

Study	Study type	Time period	Geography/ setting	n	Treatment type	Duration (days)	Eradication rate (%)
Magaret 2001 [36]	Single-center prospective	2001	Michigan	28	Bismuth metronidazole–tetracycline–lansoprazole	14	71
Stevens 2002 [37]	RCT	2002	Oregon	154	Omeprazole bismuth metronidazole tetracycline	7	80.5
Sullivan 2002 [38]	Prospective, randomized, blinded comparative	2002	Virginia	27	Bismuth–clarithromycin–amoxicillin-lansoprazole	10	81
				29	Bismuth–azithromycin–amoxicillin–lansoprazole		71
Barrett-Englert 2019 [48]	RCT, single-center, open label trial	2011	New York	90	Levofloxacin–omeprazole–nitazoxanide–doxycycline	7	90
						10	88.9
Mertz 2020 [45]	Retrospective observational multicenter	2015–2018	Nationwide US and OCONUS; military medical facilities	135	Levofloxacin regimens (combined)	NR	64
				89	Levofloxacin triple		65
				118	Bismuth quad, doxycycline based		73
				6	Rifabutin triple		50
Argueta 2021 [34]	Retrospective, observational multicenter	2018–2019	Rhode Island	2	Rifabutin-PPI-amoxicillin	NR	50
				18	PPI–doxycycline–metronidazole–bismuth		75
Graham 2020 [43]	RCT, multicenter	2020	Nationwide US	228	Rifabutin–omeprazole–amoxicillin	14	83.8
				227	High-dose amoxicillin–omeprazole		57.7

N, number; NR, not reported; OCONUS, outside the continental US; PPI, proton pump inhibitor; RCT, randomized controlled trial; US, United States.

Few studies included patients who had refractory Hp infection after first-line or subsequent eradication treatment. In a recent US-based study by Argueta et al., 65.6% of Hp isolates were resistant to at least one antimicrobial used for treatment, including metronidazole (33.3%), clarithromycin (30.0%), and levofloxacin (29.6%) [34]. This contrasts a 2004 report citing a resistance rate of 29.1% to one antibiotic used for eradication therapy, either metronidazole, clarithromycin, or metronidazole [53]. Resistance rates are even higher among patients with at least one prior episode of Hp eradication treatment failure. In one single-center study in Pennsylvania of patients with refractory Hp infection, 89% of patients had confirmed resistance to at least one antibiotic commonly used for Hp eradication, including levofloxacin (69.2%), metronidazole (41.5%), clarithromycin (43.3%); not surprisingly, only 38% of refractory patients had confirmed eradication success with subsequent treatment [54].

2.1.1. Factors influencing eradication treatment outcomes

Details of included studies that analyzed or described factors associated with eradication treatment outcomes (i.e. success or failure) are reported in Table 2. In identified studies, patients lost to follow-up or not tested for eradication were common. Based on the few studies that actually reported the proportion of patients who underwent confirmatory testing, frequencies of Hp eradication confirmation testing ranged from 33.9% to 85.6% [44,46,48,49,55].

2.1.2. Temporal trends

Identified eradication rates (crude data) in the US are illustrated in Figure 1. Studies conducted prior to 2015 were predominantly clinical trials, whereas more recent evidence tended to be observational. No clear trends were identified when visualizing raw eradication rates for individual studies over time by study type. According to a linear regression model using reported eradication rates weighted by sample size with treatment timeframe as the predictor, there was no statistically significant temporal trend (p = 0.23).

2.2. Research Question 2: Hp disease-related outcomes of eradication treatment failure

2.2.1. PUD and complications

We identified no US-based studies that reported on the association between Hp eradication and risk of PUD or PUD complications. As such, to inform this research question, we extended the search to include international data, and prioritized systematic literature reviews (SLRs) with meta-analyses. We identified three meta-analyses, all of which suggested that Hp eradication therapy is associated with a higher likelihood of duodenal ulcer healing and reduced likelihood of duodenal and gastric ulcer recurrence vs no treatment (Table 4) [56]. Compared to no Hp eradication treatment, confirmed Hp eradication was associated with a significantly lower likelihood of rebleeding with or without long-term acid suppression maintenance therapy [57], as well as reduced risk of recurrent perforation following surgical repair at 1 year [58]. Among a cohort of patients in the US, 16% and 25% of the patients with gastric and duodenal ulcers, respectively, were Hp positive. The proportion of Hp-positive ulcers East Asians and Hispanics was 37% and 35%, respectively [59].

Table 4. Hp-associated PUD complications, precancerous gastric mucosal changes, and risk of GC

PUD (International)^a
Study	Study characteristics and time frame	Hp-associated results
Ford 2016 [56]	SLR and meta-analysis of RCTs on Hp-associated PUD treatment 1950–2016 55 studies	Eradication treatment was associated with: Higher likelihood of duodenal ulcer healing vs anti-acid monotherapy (RR 0.66; 95% CI, 0.58–0.76) or no treatment (RR, 0.37; 95% CI, 0.26–0.53) Reduced risk of duodenal (RR 0.20; 95% CI 0.15–0.26) and gastric (RR 0.31; 95% CI 0.22–0.45) ulcer recurrence vs no treatment
Gisbert 2004 [57]	SLR and meta-analysis of RCTs on Hp eradication therapy vs antisecretory non-eradication therapy 1966 to 2009 Seven studies, N = 578 (eradication vs non-eradication without maintenance Three studies, N = 470 (eradication vs non-eradication with maintenance	Eradication therapy was associated with lower mean percentage of rebleeding than: Non-eradication without long term maintenance antisecretory therapy (2.9% vs 20%; OR 0.17; 95% CI 0.10–0.32) Non-eradication with long term maintenance anti-secretory therapy (1.6% vs 5.6%; OR 0.24; 95% CI 0.09–0.67)
Tomtitchong 2012 [58]	SLR and meta-analysis of RCTs on ulcer recurrence in duodenal ulcer perforation in patients treated with simple closure plus post-operative Hp eradication therapy vs simple closure plus antisecretory non-eradiation therapy Database inception through to 2010 Three studies	Pooled 1-year ulcer recurrence: 5.2% (95% CI 0.7–9.7) for eradication therapy 35.2% (95% CI 0.25–0.45) for antisecretory therapy alone Pooled RR: 0.15 (95% CI 0.06–0.37)
Sonnenberg 2020 [59]	Cross-sectional study 1,289,641 patients who underwent esophagogastro-duodenoscopy patients at private outpatient endoscopy centers across the US 2009–2018	Ulcer risk with Hp Increased risk of gastric ulcer and duodenal ulcer (OR 1.68; 95% CI 1.64–1.72 and 2.97; 2.85–3.09, respectively) Hispanic gastric ulcer OR 3.20 (95% CI 2.85–3.61); duodenal ulcer OR 3.19 (95% CI 2.95–3.45) East Asian gastric ulcer OR 6.77 (95% CI 5.54 = 8.28); duodenal ulcer OR 4.57 (95% CI 3.99–5.24)
Gastric precancerous mucosal changes (US)
Study	Study characteristics	Hp-associated results	Other risk factors
Parbhu 2019 [61]	Retrospective review of upper GI endoscopies in patients from an academic database 2011–2016, N = 434	2.6% of patients diagnosed with GIM 86.5% of patients with GIM tested for Hp 26.1% positive for H. pylori	–
Huang 2020 [63]	Multicenter review of upper endoscopic procedures Two medical centers (academic and community safety net) in Washington state 1999–2014, N = 2194	Hp infection was associated with: Prevalence of GIM Higher-grade precancerous pathology Risk of GIM Multivariate OR, 2.78(95% CI 2.48–3.11; p < 0.05) Presence of H pylori at the time of endoscopy was associated with GIM (OR, 3.46; 95% CI 3.13–3.83)	Race/ethnicity was significantly associated with GIM or higher-grade lesions Prevalence (OR [95% CI]) Asian/Pacific Islander 34.7% (OR 5.21 [4.60–5.91]) Hispanic 33.8% (OR 4.94 [4.03–6.05]) African American 20.7% (OR 2.49 [2.15–2.89])
Kligman 2019 [62]	Multicenter retrospective cross-sectional study of patients undergoing elective endoscopy with gastric biopsies Six academic and community medical centers in Texas 2017, N = 1140	History of Hp infection was associated with higher prevalence of GIM: 17.5% vs 7.5% in controls (p = 0.002)	GIM prevalence: Asian/Pacific Islander 18.5% Hispanic 12.2% African American 9.7% Non-Hispanic White: 6.7% Older age and non-White race/ethnicity were associated with higher GIM prevalence
Gastric cancer (US)
Study	Study type and time frame	N	H. pylori	Other risk factors
Kumar 2020 [64]	Retrospective VA study of Hp-positive patients 1994–2018	38,535	Lower risk of GC among those with confirmed Hp eradication: SHR 0.24 (95% CI 0.15–0.41; p < 0.001) For each 5-year increase in age at the time of Hp diagnosis, the risk of GC increased: SHR 1.13 (95% CI 1.11–1.15; p < 0.001)	Increased GC risk SHRs (all p < 0.001): Black/African American race: 2.00 (95% CI 1.80–2.22) Asian race: 2.52 (95% CI 1.64–3.89) Hispanic or Latino ethnicity: 1.59 (95% CI 1.34–1.87) p < 0.001) History of smoking: 1.38 ((%% CI 1.25–1.52) Decreased GC risk SHRs: Female gender: 0.52 (95% CI 0.40–0.68; p < 0.001)
Fansiwala and Liang 2020 [65]	Single-center New York Harbor VA (GC: 80% NCGC) 1997–2018	504	Hp eradication was associated with a decreased risk of GC (OR 0.15; 95% CI 0.06–0.38; p < 0.05)	Increased GC risk in those of Black race vs White race: OR 2.73; (95% CI 1.12–6.68; p < 0.05) Decreased GC risk in former alcohol users vs never users: OR 0.26, (95% CI 0.07–0.97; p < 0.05)
Dhingra 2020 [67]	Retrospective single-center, single-cohort study at a university medical center in Massachusetts 1993–2013	358	No association between H. pylori, GIM, and progression to GAC was identified^a	Patients with GIM had 10× IR of non-GIM population: 0.8 vs 0.07 per 1000 person-years GIM IR rate by race/ethnicity (per 1000 person-years): African American: 18.4 Asian: 16.3 Hispanic: 11.3 White: 6.9
Li 2020 [66]	Retrospective study of patients with NCGC and Hp testing in a single healthcare system (Kaiser Permanente) in California 1997–2015	173	Risk of NCGC in patients with untreated Hp infection increased significantly over time from testing vs those who tested negative: <5 years: AHR 1.93 (95% CI 0.88–4.24; p = 1.03) 5–10 years: AHR 4.35 (95% CI 1.48–12.78; p = 0.006) >10 years: AHR 5.67 (95% CI 1.65–19.50; p = 0.006) Eradication therapy reduced 10-year NCGC risk by 67%	Non-White race/ethnicity Family history of GC Older age (>50 years) Male sex
Nguyen 2020 [68]	Retrospective single-center Houston, Texas VA study of patients with NCGC 2007–2018	91	Hp-positive NCGC was associated with: Active/acute gastritis: adjusted OR 3.74 Atrophic gastritis: adjusted OR 15.30 GIM: adjusted OR 3.65	Those diagnosed with Hp-positive NCGC were more likely to be Black than non-Hispanic White (60.0% vs 22.8%, p = 0.065)
Li 2016 [69]	Retrospective cohort study of patients diagnosed with GIM or dysplasia in California 1997–2006	4146	HR of progression from precancerous lesion to GAC: Hp negativity: 0.77 (0.30–1.94) vs positive (reference)	Compared to the plan member population, SIRs were highest for: Hispanic persons: 52.70 Asians: 48.73 Whites: 15.78 No African American patients with dysplasia had NCGC
Reddy 2016 [70]	Retrospective cohort study of patients diagnosed with GIM in a single healthcare system (Kaiser Permanente) in California 2000–2011	934	No significantly increased risk with Hp; none of eight cases of GC were positive for Hp infection.	Increased GC risk: Family history (HR 3.8; p = 0.012 Extent of GIM (OR 9.4; 95% CI 1.8–50.4)

^aParameters extended to include systematic reviews and meta-analyses from 2000–2021 given the low number of studies.

^bAuthors attributed lack of association to small sample size of study.

AHR, adjusted hazard ratio; CI, confidence interval; GAC, gastric adenocarcinoma; GC, gastric cancer; GIM, gastric intestinal metaplasia; HR, hazard ratio; IR, incidence rate; OR odds ratio; NCGC, non-cardia gastric cancer; PUD, peptic ulcer disease; SHR, subhazard ratio, SIR, standardized incidence ratio; SLR, systematic literature review; RCT, randomized controlled trial; RR, relative risk; US, United States; VA, Veterans Administration.

2.2.2. Precancerous gastric mucosal changes

Persistent Hp gastritis may progress to precancerous gastric mucosal changes, including atrophic gastritis and gastrointestinal metaplasia (GIM), with a minority having neoplastic transformation to dysplasia and ultimately noncardia gastric adenocarcinoma. Based on a multivariable analysis of a single VA medical center in Texas from 2009 to 2014, patients with Hp-positive gastritis were significantly more likely to be Black than non-Hispanic White [60]. According to a large cohort study in Utah from 2011 to 2016, the prevalence of GIM was 2.6% and approximately 26% of cases of GIM had active Hp infection based on histology [61]. Two retrospective studies of patients (N = 1140; 17,710) undergoing endoscopy with gastric biopsy reported significant associations between Hp infection and GIM prevalence [62,63]. In addition, GIM or higher-grade pathology was more prevalent in non-White populations in both studies with Asian/Pacific Islanders having the highest prevalence, followed by Hispanic and African American persons [62,63]. Prevalence of precancerous changes increased with age; moreover, this observation was more marked in non-White populations [63].

2.2.3. NCGC

We identified six studies that analyzed the association of confirmed Hp eradication and risk of NCGC, representing a total of 43,807 patients (range: 91 to 38,535) from 1993 to 2018 [64–69]. Five reported reduced risk of NCGC with Hp eradication [64–66,68,69], whereas one study by Dhingra et al. did not demonstrate a significant relationship over a median of 3.5 years of follow-up (interquartile range of 1.8–6.6 years) up to 18 years (Table 4) [67]. According to a retrospective cohort study of 934 patients with diverse racial/ethnic backgrounds (68.2% non-White race or ethnicity) and confirmed GIM with a median duration of follow-up of 4.6 years (interquartile range, 3.0–6.7 years), 2.7% were diagnosed with GC on follow-up between 2000 and 2011 [70]. While extensive vs limited GIM (OR 9.4 [95% CI 1.8–50.4]) and positive family history of GC were significant independent risk factors of GC, a history of Hp was not independently associated with GC risk [70].

We did not identify any US-based studies of Hp eradication and risk of metachronous GC following GC resection, nor of Hp eradication in patients who are high risk for GC based on a positive family history.

2.2.3.1. NCGC risk – ethnicity and race

All studies reporting on Hp eradication status and NCGC also identified race and/or ethnicity as risk factors for NCGC [64–68]. Therefore, we examined additional studies reporting on the association of patient characteristics, including race and ethnicity, and incidence of NCGC.

Vulnerable populations demonstrate increased rates of NCGC and Hp positivity, with studies suggesting twofold higher prevalence. Reported seroprevalence among non-Hispanic Blacks and Hispanics ranged from 53% to over 60% compared to 26% for non-Hispanic Whites in samples from the National Health and Nutritional Examination Survey (1988–1991) [71]. Prevalence in a contemporary study (2009–2018) of over one million patients undergoing esophagogastroduodenoscopy was approximately 9% to 11% for the overall population, 22.0% to 23.6% among Hispanics, and 15.1% to 21.4% among East Asians; rates declined with time and prevalence in non-Hispanic Blacks was not assessed [59]. Compared to non-Hispanic Whites, rates of NCGC for non-Hispanic Blacks and Hispanics are 3- to 4-fold higher and up to 13.3-fold higher for Korean individuals in the US [72].

According to an analysis of California Cancer Registry (CCR) data between 2011 and 2015, non-White populations ≥50 years had higher age- and sex-adjusted incidence rates of NCGC adenocarcinoma [71]. Incidence rate ratios (IRRs) were highest for Asian individuals, particularly Korean (13.3), Vietnamese (6.46), Southeast Asian (5.71), Japanese (5.18), and Chinese (4.77) individuals. Hispanics and non-Hispanic Black persons had 3.79 and 3.03 higher rates than non-Hispanic White persons [72].

An analysis of NCGC in the North American Association of Central Cancer Registries database from 1995 to 2013 yielded age-standardized rates (ASRs) of 2.16 for non-Hispanic Whites, 6.43 for non-Hispanic Blacks, and 6.15 for Hispanics [73]. The ASR for high vs low county-level poverty percentage was 4.38 vs 2.67, respectively [73]. According to an analysis of the National Cancer Institute’s Surveillance, Epidemiology, and End Results (SEER) Program data from from 2000 to 2014luding 25,825 cases of NCGC, incidence rates were higher among non-Hispanic Blacks and Hispanic persons (2.8-fold), Asians and Pacific Islanders (3.9-fold), and indigenous US populations (1.7-fold) compared to non-Hispanic Whites [74]. Increased risk of NCGC was also noted with decreasing neighborhood socioeconomic status (nSES); the effect of nSES on age-adjusted incidence rates was more marked for Black and Hispanic populations compared to non-Hispanic Whites [74]. Of note, incidence patterns observed in SEER and CCR were unique to NCGC and were essentially the opposite of those seen with cardia GC, a cancer that is not attributed to H. pylori [72,74].

Overall, NCGC rates have declined and remained stable after 2007; however, annual percent change in incidence rates were less favorable for non-White populations [74]. Of note, Asian/Pacific Islanders are more likely to be diagnosed with localized NCGC, whereas Hispanic patients’ distant-stage NCGC rates are increasing in Hispanic persons under age 50 years (annual percent change 1.78; 95% CI 0.66, 2.91) [75–77].

2.2.3.2. NCGC mortality

The search yielded no US-based studies identified that reported on Hp eradication vs no eradication and GC-related mortality. A number of studies suggested worse NCGC survival outcomes for non-White populations [75,76,78–81]. According to an analysis of the SEER database from 2004 to 2016 conducted by Laszkowska (2020), NCGC mortality rates were two to three times higher for non-Hispanic Blacks, Hispanics, and Asians/Pacific Islanders compared to non-Hispanic Whites regardless of age or disease stage [78]. However, an analysis of the CCR from 2006 to 2015 reported similar mortality risks for non-Hispanic Blacks, Hispanics, and non-Hispanic Whites with NCGC, and 20% reduced risk for Asians/Pacific Islanders [79].

In a single-center retrospective analysis of 638 patients with GC, ethnicity was an independent predictor of survival (multivariable OR: 1.69; 95% CI 1.07–2.55, p < 0.02). Non-Hispanic Whites had a median overall survival (OS) of 99 months vs 51 months for Hispanics who tended to be younger at diagnosis and present with more advanced disease [76]. Similarly, in a study of young-onset NCGC using SEER program data from 2007 to 2015, Hispanics had increased risk of diagnosis at higher grade (III/IV) compared with non-Hispanic Whites (OR 1.93; 95% CI 1.32–2.84, p = 0.001), but were 38% less likely to undergo surgery [75].

A retrospective cohort study of patients with resectable gastric adenocarcinoma identified from SEER data (N = 15,991; 2007 to 2015) found that Black patients were more likely to receive a recommendation against surgical resection than non-Hispanic White, Asian, or Native American patients (adjusted ORs for surgery recommendation were 0.86, 0.55, and 0.50, respectively) [80]. Non-Hispanic Whites were more likely to receive preoperative chemotherapy than other racial or ethnic groups, according to a study by Ikoma et al.. Use of preoperative chemotherapy mitigated racial/ethnic disparities in OS [81].

2.3. Research Question 3: economic impact of failed eradication

We did not identify direct evidence of the costs of Hp eradication treatment failure. Herein we present indirect evidence through estimated costs of Hp-associated diseases. Note that these costs also do not account for the cost of repeated Hp eradication treatment courses and downstream multifaceted consequences.

2.3.1. PUD and acute complications

PUD was associated with a total expense of US$777 million in 2015 (see Table 5) [15]. While PUD has its own cost of management, according to the studies identified in our search, complications are key drivers of the economic impact. Cost is primarily driven by hospitalizations (43.5% of total expense), medications (24.4%), and outpatient (19.2%) and emergency department (ED) visits (11.4%) [15]. Complicated PUD, including bleeding and perforation, is associated with high costs and healthcare resource utilization, including ED visits, admissions, and surgery [82]. Identified costs of complicated PUD ranged from $1,883 to $25,444 per patient prior to 2010 [82]; adjusted for inflation, this amounts to at least roughly $2,201 and $29,745 in 2021.

Table 5. Costs associated with PUD, PUD-associated complications, and GC in the US

PUD and complication costs
	Per patient/per admission cost		Aggregate cost
Overall^a PUD	$163–866 [82]		$777 million [15]^b
Complicated^c PUD	$1,883–25,444 [82]		$777 million [15]^b
GI hemorrhage	$7413/$26,019 [15]		$2.79 billion [15]
Perforation	$27,908–93,604 [90]		–
Sepsis^d	$192,525 [94]		$15 billion [92,93]^e
GC costs
	Mean per patient cost ± SD		Aggregate cost
Treatment-related costs
First line^f	$9,293 ± $9360 [99]		–
Second line^f	$6,924 ± $10,279 [99]		–
Total GC treatment^f	$14,668 ± $17,501 [99]		–
Total supportive care^f	$12,236 ± $18,251 [99]		–
Hospitalization costs
Hospitalizations during treatment^g,h	$16,880–24,706 [96,101]		$575 million [96]
Hospitalizations post-chemotherapy^g	$85,769.49 ± $178,096.47		–
Total all-cause costs
First line^g	$40,811 ± $49,916 [101]
Second line^g	$26,588 ± $33,301 [101]
Total GC^g	$70,808 ± $56,6206 [101]		$2.3 billion [101]^g
Post-chemotherapy/end of life all cause-total healthcare costs^b	$80,148.07 ± $161,421.80 [101]		–

GC, gastric cancer; GI, gastrointestinal; PUD, peptic ulcer disease; SD, standard deviation; US, United States.

^aOverall PUD includes uncomplicated PUD and PUD that is associated with complications.

^bIncluded all costs for standard care and emergency treatment of complications.

^cComplicated PUD encompassed hemorrhage, perforation, or any PUD complication requiring surgical intervention.

^dSepsis in patients with any GI bleeding (does not encompass costs related to sepsis associated with PUD perforation).

^eTotal aggregate sepsis costs, not classified by cause.

^fBased on SEER-Medicare claims (2000–2009).

^gBased on Truven Health MarketScan Research Databases (2004–2012).

^hNational Inpatient Sample (2001–2011).

Bleeding is the most common complication of PUD, observed in 15–20% of patients and accounting for roughly 40–60% of acute GI bleeding (Table 5) [83]. Among 119,684 admissions between 2003 and 2012 in the Healthcare Cost and Utilization Project’s National Inpatient Sample (NIS), 6,571 (5.4%) had GI bleeding [84]. According to the 2014 National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey, GI bleeding led to 2,147,949 office visits and 606,970 emergency department visits [15]. Three studies reported Hp status and GI bleeding rates; among them, Hp was observed in 42–69% of patients with GI bleeds [85–87].

Three studies reported PUD complication rates between 2000 and 2015 (Table 5) [84,88,89]. GI hemorrhage accounted for 84–87% of complications, whereas perforation was less common (10.6–12%) [84,88,89]. GI hemorrhage was responsible for 231,567 ED visits in 2015, with a 78.4% admission rate [15]. Median inpatient stay was 3 days, leading to a total of 896,224 hospital days, and generating total expenses of $2.14 to $2.74 billion; moreover, the observed 30-day all-cause readmission rate for patients with an index hospitalization for GI hemorrhage was 14.0%, with a median charge of $26,019 and $27,717 for the index and readmission stay, respectively [15].

PUD-related perforations were associated with higher costs than GI hemorrhage. According to a study using data from the NIS from 2007 to 2010, 5,361 patients with PUD-related perforations were identified [90]. The majority of PUD perforations required surgical intervention and were associated with longer inpatient stays (median 7 days for laparoscopic and 8 days for open repair), with median total charges ranging from $27,908 to $93,604 [90].

According to a retrospective cohort study of the NIS between 2012 and 2013 that included all patients with an admission due to PUD (N = 52,396), 76% of patients underwent surgery or endoscopy [88]. An analysis of the NIS from 2005 to 2014 found surgery reported in 1.9%, 67.4%, 20.2% of PUD-associated hemorrhage, perforation, and obstruction cases, respectively [89]. Among Medicare recipients who underwent emergency general surgery between 2008 and 2014 (N = 481,417), PUD surgery was the second most common type, accounting for 28.2% of admissions [92].

In complicated PUD, sepsis was a substantial cost driver [91]. Reported aggregate treatment costs for hospitalized patients with GI bleeding and septicemia in the US in 2008 were $15 billion in 2008 with an average annual increase of 12% [92,93]. Patients with GI bleeding and septic shock required longer stays (20.56 days vs 15.76; p < 0.001) and had higher mean hospital admission costs ($192,524.89 vs $142,688.55; p < 0.001) [94]. Based on an estimated rate of 17% Hp positivity among patients with PUD, Hp associated costs are still substantial [59].

2.3.2. Gastric cancer

In the US, there were an estimated 27,600 new cases of GC in 2020, amounting to 1.5% of all new cancer diagnoses [95]. While mortality decreased over time, hospitalizations remained stable and costs rose significantly from 2000 to 2013 [96–98]. Overall, the projected cost of GC in the US in 2020 was $2.31 billion [99], largely driven by costs of treatment and supportive care as well as hospitalizations (Table 5).

In the absence of guidelines for GC screening in the US, the majority of patients are diagnosed with late-stage GC, which is typically when symptoms are present and prompt diagnostic workup. Based on SEER data, an estimated 36% of GC cases are expected to be diagnosed at stage 4 in 2021 [95]. In a retrospective analysis of the National Cancer Database (2004–2013; N = 93,734), approximately 49% of all patients with GC received chemotherapy [100]. According to a retrospective SEER-Medicare claims analysis (2000 to 2009), total GC-related cost per patient was $70,808 ± $56,6206; 55% received additional GC treatment after first-line therapy ($14,668 ± $17,501) and 45% received supportive care only ($12,236 ± $18,251) [99]. Costs drivers include chemotherapy-related costs and inpatient care [99,101]. First-line treatment costs ranged from $15,066 to $40,811, second-line ranged from $12,699 to $26,588, and third-line was $7,199 [101,102]. In a retrospective cohort study of GC and non-GC patients between 2007 and 2012, healthcare costs for patients with GC were over 10 times that of matched non-GC [103]. While patients with GC have high healthcare resource utilization and costs overall, advanced disease is associated with the highest costs, primarily driven by hospitalizations.

Despite higher rates of diagnosis during more expensive stages of GC, the mean cost of care for Hispanic patients was less than a quarter of the average for White patients in the NIS from 2001 to 2011 [96].

3. Discussion

Untreated Hp has major clinical and economic consequences. The success rates of Hp eradication treatment are decreasing globally as Hp antibiotic resistance rises; however, the limited published evidence in the US hinders our ability to completely evaluate trends in the US. The data for regimen-specific eradication rates in US cohorts were limited and heterogeneous. The overall quality of data across studies was also inconsistent; of the seven studies that reported eradication rates with regimens other than clarithromycin-based triple or bismuth-based quadruple therapies, three were RCTs and one was a randomized prospective study. Most studies reported predominantly on clarithromycin-based triple therapy and/or bismuth quadruple therapy, with rifabutin triple therapy and levofloxacin-based therapy reported in two non-randomized studies each. We did not identify recent evidence of eradication rates with high-dose amoxicillin dual therapy apart from RCTs in the US, although this regimen is commonly used outside of the US. Studies from non-US regions appear conflicting, especially for high-dose amoxicillin dual therapy, which requires appropriate acid suppression for treatment success. This treatment is enticing though, given the consistently low rates of amoxicillin-resistant Hp strains, increased patient tolerability, relative simplicity, and parsimony of the regimen compared to other Hp treatment regimens, and the alignment with antibiotic stewardship goals. Of note, two studies of non-amoxicillin-based dual regimens were identified but excluded because they are not FDA approved or recommended by current clinical guidelines; one reported a 46% eradication success rate with a 7-day pantoprazole-clarithromycin dual regimen and the other reported 46% eradication success rate with a 10-day esomeprazole-clarithromycin dual regimen [35,40].

Eradication rates are inherently difficult to compare or generalize because of the heterogenous nature of the studies, as regimens vary by PPI type and dose used, antibiotics studied and formulations, treatment duration, patient demographics, comorbidities, and other characteristics (e.g. known PUD). Study findings may also differ from true rates given the possibility of publication biases. Moreover, eradication rates reported in clinical trials may be higher than ‘real-world’ practice; one reason perhaps being higher adherence rates in trial participants and participant bias. There is also a possibility of bias for patients who are tested vs not tested for successful Hp eradication in clinical practice. For example, patients who complete post-treatment testing may disproportionately represent patients who are at higher risk for complications related to persistent Hp infection (e.g. history of complicated PUD or family history of GC) or perhaps may represent a group of higher-adherence patients or patients with greater resources vs those who do not complete Hp eradication confirmatory testing. As such, our analysis was only descriptive and we were not able to conduct pooled analyses of the data from real-world evidence and RCTs.

Across studies, Hp antibiotic sensitivity profile was variably associated with eradication success vs failure [34,53,54]. Clarithromycin-based triple therapy was used for a substantial proportion of patients, even though its use is not recommended in regions with known or suspected clarithromycin resistance rates >15% or in patients with a history of macrolide use [16,17]. While identified studies did not provide insight as to why physicians may be selecting a potentially suboptimal regimen, limited information regarding patterns of antibiotic resistance in the US and lack of routine susceptibility testing to inform treatment decisions may be contributing factors. According to a survey by Murakami et al., physicians did not demonstrate knowledge of resistance rates and tended to over- and underestimate resistance rates [104].

Because eradication therapy typically consists of complex multidrug regimens that are associated with side effects, patients often miss doses or discontinue treatment, which may contribute to treatment failure [16,17]. While there is concern that adherence rates are generally higher in trials than in the real world, creating the potential for lower rates in practice, the scarcity of adherence data in the RWE we identified limited our ability to draw conclusions. The vast majority of studies we identified utilized 10 or 14-day regimens. A previous study demonstrated that adherence above 60% was not associated with treatment failure [105]; however, the recent ERADICATE Hp2 trial of rifabutin reported a nearly 7-percentage point difference among the whole population and the confirmed adherent population [43]. That said, even in studies with high adherence, eradication treatment failure still occurs. This suggests that the underlying causes of failure are multifactorial and include host, microbial, and system factors.

With respect to patient characteristics, in the studies identified for the primary research questions, age, sex, race, and ethnicity were not consistently associated with eradication success vs failure. Genetic polymorphisms that affect host gastric acid suppression (e.g. CYP2C19, IL-1B polymorphisms may influence the success of Hp eradication treatment. This is because effective, durable acid suppression is needed to achieve the intragastric pH range necessary to ensure Hp are replicating and therefore susceptible to antibiotics [20] and to improve bioavailability and activity of certain growth-dependent antibiotics, such as amoxicillin and clarithromycin [19]. While higher rates of Hp eradication have been noted among patients with poor vs rapid metabolizer phenotypes for first-generation PPIs in Asian studies [106], the impact of racial and ethnic phenotypic differences on eradication outcomes in the US has not been extensively explored [107].

Overall, the reported costs of conditions attributable to Hp infection in identified studies were substantial. However, we did not identify direct evidence associating failed Hp eradication with economic impact of conditions related to persistent Hp infection (e.g. dyspepsia, PUD, PUD complications, NCGC) in the US. Moreover, it was not possible to determine with certainty the proportion of the clinical conditions of interest that were attributable to failed Hp eradication and further investigations, such as modeling studies, would be of value. We were also not able to quantify indirect downstream costs related to retreatment of Hp after failed a failed episode of treatment, such as promotion of antibiotic resistance. Notably, though, such costs have been suggested to be as high as $33 billion in the US alone [108].

Strengths of this study include the comprehensive literature search and the focus on contemporary US studies. There are limitations that merit mentioning, in addition to those alluded to above such as publication and participant biases. We did not focus on costs related to adverse outcomes related to the treatment itself, including GI side effects. We will note, however, that two large US cohort studies have at least reported the lack of serious adverse effects associated with Hp treatment, including cardiovascular mortality and Clostridium difficile infection [109,110]. We also did not evaluate rarer Hp associated conditions, including MALT lymphoma, and extragastric manifestations such as immune thrombocytopenia purpura.

4. Conclusion

Hp poses an inarguable threat to human health with the potential for severe clinical consequences and high costs; our findings suggest a substantial proportion of people in the US with chronic Hp infection may not experience the benefits of eradication due to suboptimal rates of eradication and rising antibiotic resistance.

More effective eradication regimens would offer the potential for reducing both acute and long-term clinical consequences of Hp as well as reductions in costs of treatment and healthcare resource utilization for Hp-associated complications

5. Expert opinion

Hp eradication treatment failure is common and exposes patients to complications associated with persistent Hp infection, not to mention downstream consequences of repeated treatment, including promotion of antibiotic resistance. In the US, the impact of eradication treatment failure on clinical outcomes, including PUD and progression to GC, and the concomitant economic burden, have not been thoroughly described nor quantified.

According to the findings of this review, eradication rates in the US are poor, with most studies reporting rates ≤80% with current guideline-approved first-line regimens. Clarithromycin-based triple therapy was used for a substantial proportion of patients, despite its limited utility in patients with prior macrolide use or in regions with clarithromycin resistance rates >15% [16,17]. The absence of systematic regional and national registries for monitoring and recording Hp resistance patterns along with a lack of routine use of commercially available susceptibility testing to inform treatment decisions may contribute to the paucity of US data and impedes detailed understanding of Hp antibiotic resistance profiles. This is in stark contrast to the European consortium [23] where the data generated have already translated to updated clinical practice recommendations regarding selection of Hp eradication regimens. Another contributing factor is that in the US, despite current guideline recommendations [16], testing to confirm successful Hp eradication is not regularly performed in clinical practice following Hp treatment, which reduces both the volume and generalizability of RWE.

Future studies to establish the prevalence of specific strains of Hp and rates of associated clinical conditions such as PUD, precancerous lesions, and GC among vulnerable populations would be important to support improvements in monitoring, treatment, and outcomes. Potential initiatives include addressing the need for better antibiotic resistance surveillance and Hp registries in the US, streamlining culture and sensitivity testing, as well as encouraging widespread adoption of best practices for eradication treatment selection guided by clinical guidelines and publicly available local aggregate data [17].

Efforts to individualize eradication therapy based on genetic factors (IL-1B, CYP2C19, and non-genetic clinical factors [e.g. smoking, diet, obesity]) may improve treatment decision-making by identifying patients most at-risk for negative sequelae vs those who may not benefit from eradication [111]. In addition, the adoption of programs to support maximal adherence and studies investigating alternative regimens that consistently achieve eradication rates >90% with initial therapy will presumably not only improve patient outcomes and experience but would also reduce waste in the healthcare system.

For vulnerable populations who experience increased risk of Hp and GC, screening may be beneficial and cost-effective. Interestingly, a recent study in a Korean population demonstrated that for individuals with a family history of GC (first-degree relative) and Hp infection, the risk of GC was reduced by eradication treatment [112]. While early eradication seems to be important to provide the greatest protective benefit, Korean patients with endoscopically resected early GC who received Hp eradication therapy exhibited reduced rates of metachronous GC compared to patients in the placebo arm [113.

Investigation of both new and repurposed antibiotic agents as well as potentially promising adjunctive therapies is sorely needed. Excitingly, the increase in availability of molecular modalities to test for Hp antibiotic susceptibility (either on biopsies or noninvasively) has very strong potential to optimize our treatment approach. Moreover, individualized Hp eradication therapy that is further optimized for both host and Hp-specific factors, including the most appropriate antibiotics combined with potent acid suppression, is critical to achieve high rates of successful eradication and ensure that the protective benefits are experienced by patients in need.

Article highlights

• Overall, our search results demonstrated that current guideline-approved first-line regimens achieve poor eradication rates in the US, with most studies reporting rates ≤80% with current guideline-approved first-line regimens (10/14 studies of clarithromycin triple therapy and 5/7 studies of bismuth quadruple therapy). The most common reasons for Hp eradication failure are antibiotic resistance and patient nonadherence. There are limited data on clinical factors associated with Hp eradication treatment failure. Higher BMI and clarithromycin resistance were identified factors based on limited literature.

• No recent US studies have evaluated Hp eradication outcomes and risk of peptic ulcer disease (PUD) and associated complications. Based on non-US studies, successful eradication of Hp was associated with increased likelihood of ulcer healing and decreased risk of ulcer recurrence, rebleeding, and reperforation.

• Eradication of Hp is associated with reduced likelihood of precancerous mucosal changes compared to no Hp treatment. Hp infection was positively associated with incidence of precancerous lesions, which were more prevalent in non-White groups .

• Based on a handful of US studies, Hp eradication is associated with reduced risk of non-cardia gastric cancer (NCGC).

• Based on indirect evidence, the cost of Hp eradication treatment failure in the US is enormous. Reported costs associated with complicated PUD and NCGC were high, with aggregate costs over $5.3 billion driven primarily by hospitalizations, medication, and outpatient visits.

Declaration of interests

R Yadlapati is supported by NIH K23 DK125266 (PI: Yadlapati); SC Shah is supported by a 2019 American Gastroenterological Association Research Scholar Award (PI: Shah), and Veterans Affairs Career Development Award under award number ICX002027A01 (PI: Shah) The authors disclose the following relationships: R Yadlap: Institutional Consulting Agreement: Medtronic, Ironwood Pharmaceuticals; StatLinkMD; Consultant: Phathom Pharmaceuticals; Research support: Ironwood Pharmaceuticals; Advisory Board with Stock Options: RJS Mediagnostix; S Shah: Consultant, Phathom Pharmaceuticals; C Pelletier and R Jacob: Employment: Phathom Pharmaceuticals; E Hubscher and L Vinals: Employment: Cytel, Inc, which was a paid consultant on the project. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or conflict with the subject matter or materials discussed in this manuscript apart from those disclosed.

Reviewer disclosures

Peer reviewers on this manuscript have no other relevant financial relationships or otherwise to disclose.

Additional information

Funding

This review was supported by Phathom Pharmaceuticals.