Introduction

Helicobacter pylori is a Gram-negative, microaerophilic bacterium that infects more than half of the human population worldwide.1 H. pylori infection causes gastrointestinal diseases, including gastritis, gastroduodenal ulcer disease and gastric cancer as well as iron-deficiency anaemia, vitamin B12 deficiency and immune thrombocytopenic purpura.2 In all these conditions, eradication of the bacterium is considered the best course of action.3 Treatment of H. pylori is challenging, and the treatment schemes currently applied do not achieve 100% cure rates. Furthermore, the success of certain therapies may decline over time due to the increase in antibiotic resistance.4–6 In this context, expert consensuses have established that any acceptable therapy should achieve a minimal cure rate of 90% for this microbial infection.7 8

H. pylori therapies combine antibiotics and adjuvant drugs. The most frequently used schemes comprise triple therapies involving a proton pump inhibitor (PPI) plus two antibiotics or quadruple therapies with an additional antibacterial agent. Triple therapies combining a PPI with clarithromycin and either amoxicillin or metronidazole have classically been the standard of care for H. pylori eradication. However, bacterial resistance to clarithromycin markedly reduces H. pylori eradication when triple therapy is prescribed. As resistance to clarithromycin has increased steadily in recent years, triple therapy currently fails in more than 20% of cases in most settings.4 The figures are even more alarming in the case of rescue regimens, due to the high rate of secondary bacterial resistance following accumulation of previous failed attempts.9–11 For this reason, triple therapies are no longer regarded as acceptable in most settings and quadruple therapies are now considered the new standard of care by consensus conferences.3 12 13 Quadruple therapies may or may not include bismuth; those that do not (namely sequential, concomitant and hybrid regimens) include three antibiotics plus a PPI, whereas bismuth-based regimens combine a bismuth salt with a PPI and two antibiotics.14

Bismuth has been extensively used to treat different gastrointestinal diseases.15 Its use offers the following advantages: (a) a strong bacteriostatic effect that is not altered by resistances; (b) beneficial synergy when combined with several antibiotics, making it possible to overcome bacterial resistance; (c) a good tolerability and safety profile and (d) a reduction in the antibiotic load and duration of H. pylori therapies.3 16 Bismuth is mainly used in quadruple therapies (bismuth quadruple therapies, BQT). Classical BQT combines bismuth with a PPI, metronidazole and tetracycline. Clinical trials have shown that BQT eradicates H. pylori better than standard triple therapies and that its effectiveness is largely unaltered by antibiotic resistances; indeed, BQT is particularly recommended in areas with high rates of antibiotic resistance.4

The European Registry on Helicobacter pylori Management (Hp‐EuReg) was set up in 2013 to collate data regarding the diagnosis and eradication treatments, making it possible to perform time trend evaluations and thus enhance the clinical management of adult infected patients. The Hp‐EuReg currently includes more than 70 000 cases from over 300 centres in 38 countries.17

The objective of the current study was to analyse the evolution of the use, effectiveness and safety of BQT in the clinical management of H. pylori infection in Europe.

Results

During the study period, 15 582 (31%) patients out of a total of 49 690 received a bismuth-based treatment. The mean age of patients was 50 years (±17.8), 8897 (62%) were women and the most frequent indications were non-investigated dyspepsia (21%), functional dyspepsia (33%) and peptic ulcer (17%). The most frequently prescribed scheme was ScBQT concomitantly with a PPI, in 6668 cases (43% of all bismuth treatments); the remaining 8914 cases were non-scBQT (NScBQT) using different antibiotic combinations. These NScBQT encompassed 45 different treatment schemes. Of these, 26 were used in fewer than 10 patients and a further 7 were prescribed in more than 100 patients; the latter schemes were selected for further analysis, as described in the Methods section. The most frequently used NScBQTs were PPI+CAB, PPI+LAB and PPI+MTB (online supplemental file 4).

Trends in the use of bismuth-based therapy in Europe

After the commercialisation of ScBQT in Europe in 2013, there was a progressive increase in BQT prescription, especially due to the use of the single-capsule scheme. The use of BQT peaked in 2018 when 50% of the reported treatments were bismuth therapies; thereafter, the rate of bismuth therapies fell slightly with respect to non-bismuth options and remained stable at around 40% in the 2019–2021 period (figure 1).

Figure 1

Evolution of all bismuth versus non-bismuth-based prescriptions between 2013 and 2021 in Europe. White numbers represent ScBQT percentage over total treatments. Black numbers represent NScBQT percentage over total treatments; non-bismuth, all regimens without bismuth. NScBQT, non-single-capsule bismuth quadruple therapy; ScBQT, single-capsule BQT.

In 2018, 19% of the treatments were ScBQT and 31% were NScBQT; however, by the end of current study period, the prescription rates of ScBQT and NScBQT were practically the same, representing 17.8% and 21% of all treatments, respectively (figure 1).

Regarding geographical distribution, the greatest increase in BQT prescriptions occurred in Southwestern Europe, accounting for 75% of all treatments in 2018 (online supplemental file 5). ScBQT was the main treatment in this region (60%) (online supplemental file 6), while NScBQT was more frequent in Eastern Europe (60% in 2018 and 40% in 2021).

Both the rate and type of BQT prescriptions ranged widely between European countries (figures 2 and 3, online supplemental file 4).

Figure 2

Proportion (%) of bismuth quadruple therapy (over all Helicobacter pylori treatments) by European country. Countries in grey did not participate in the registry.

Figure 3

Most frequently used bismuth quadruple therapies by European country. Countries in grey did not participate in the registry. ALB, amoxicillin, levofloxacin, bismuth; ATB, amoxicillin, tetracycline, bismuth; CAB, clarithromycin, amoxicillin, bismuth; MAB, metronidazole, amoxicillin, bismuth; MTB, metronidazole, tetracycline and bismuth; PPI, proton pump inhibitor; ScBQT, single-capsule bismuth quadruple therapy.

Use and effectiveness of BQT according to treatment line, length of treatment, PPI dose, European region and year

BQT was used as first-line therapy in 9955 cases (72.3%), as second-line in 2550 (18.5%) and as rescue therapy in 1262 (9.2%). ScBQT was used in 71.7% of cases as first-line therapy, in 18.4% as second-line and in 10% as rescue therapy.

Few schemes achieved an optimal overall effectiveness (above 90%). In first-line therapy, ScBQT, PPI+CAB, PPI+MTB and PPI+MAB achieved mITT cure rates of 93%, 91%, 91% and 90%, respectively (table 1).

Table 1

Modified intention-to-treat effectiveness of bismuth-based quadruple therapy in treatment-naïve and rescue patients in Europe

According to the length of treatment, only ScBQT lasting 10 days, PPI+CAB for 14 days and PPI+MAB for 14 days achieved overall optimal effectiveness (including all lines of treatment) with rates of 92%, 92% and 91%, respectively (table 2).

Table 2

Global (all lines of treatment) modified intention-to-treat effectiveness of NScBQT schemes by treatment length

Regarding PPI dose, ScBQT using standard or high-dose PPIs (93.5%), PPI+CAB with standard or high-dose PPIs (92.6% and 91.2%, respectively), PPI+MAB with standard and high-dose PPIs (91% and 92.6%, respectively) and PPI+MTB with high-dose PPIs (95.4%) achieved a rate of effectiveness above 90%. No treatment achieved an effectiveness above 90% when combined with a low-dose PPI (table 3).

Table 3

Modified intention-to-treat effectiveness according to the proton pump inhibitor dose

Effectiveness according to European region is shown in online supplemental file 7. Only ScBQT achieved an effectiveness above 90% in all European regions.

Online supplemental file 8 shows the effectiveness of the treatments by year. The effectiveness of the different BQT remained stable.

Multivariate analysis

Among the variables studied, adherence with treatment (OR: 8.447; 95% CI: 6.46 to 11.038, p<0.000), the use of ScBQT (OR: 1.941; 95% CI: 1.634 to 2.307; p<0.000), the use of 14-day prescriptions (vs 10 days) (OR: 1.396; 95% CI: 1.167 to 1.670 p<0.000) and the combination the BQT with either standard or high-dose PPIs (vs low dose) (OR: 1.696; 95% CI: 1.48 to 1.934, p<0.000) were significantly associated with higher mITT effectiveness. The use of PPI+MDB scheme (OR: 0.435; 95% CI: 0.313 to 0.604; p<0.000) or prescription of BQT as rescue treatment (vs first line) (OR: 0.547; 95% CI: 0.481 to 0.622, p<0.000) were significantly associated with lower effectiveness (table 4).

Table 4

Predictive factors of treatment modified intention-to-treat effectiveness in the multivariate analysis

Adherence and safety

Adherence was above 95% in all BQT (online supplemental file 9).

At least one AE was recorded in 40% of cases. The most frequently reported AEs were taste disturbance, diarrhoea, nausea and abdominal pain (online supplemental file 10). Around 30% of AEs were mild and roughly 6.5% were intense/severe; fewer than 1% were serious. No significant differences were observed in the intensity of AEs between the different BQT schemes.

Treatment was interrupted due to AEs in 10% of cases.

Discussion

Our study shows that one in three treatments prescribed in Europe by the gastroenterologists participating in the Hp-EuReg between 2013 and 2021 was a BQT. BQT has progressively gained popularity; its use rose steadily from 2013 to 2018 and has remained stable since then. A second noteworthy finding is that there is an extreme heterogeneity in the different BQT schemes prescribed across the different European regions. There are many possible explanations for these findings. The first is the presence of lower bacterial resistance rates in some specific settings (eg, Northern Europe) which allowed triple therapies to achieve good cure rates.23 Second, the unavailability of ScBQT, tetracycline and/or bismuth salts in many European countries limits the use of BQT. Finally, at the time of the study, European and local consensus reports often recommended other schedules as preferred treatments.3 The use of BQT has, however, increased steadily in countries where one-in-three ScBQT is available; in most of these countries, ScBQT has become the treatment of choice given its good safety profile and the excellent effectiveness it provides; furthermore, ScBQT is the most widely used treatment in south-western Europe. By contrast, classical BQT, comprising MTB administered separately, might be more laborious and inconvenient for the patient, and these characteristics may have limited its use.

Regarding effectiveness, many of the most frequently prescribed BQT therapies achieved cure rates above 90%. Notably, ScBQT repeatedly achieved cure rates above this figure, regardless of the geographical region or the PPI dose. NScBQT schemes such as PPI+CAB, PPI+MAB and PPI+MTB also achieved optimal results, provided that they were prescribed for 14 days and combined with standard or high-dose PPIs.

Accordingly, the main variables predicting the cure of H. pylori infection in the multivariate analysis were the use of ScBQT, administration of NScBQT for 14 days and the use of standard-dose or high-dose PPIs. When PPIs were standardised according to the Hp-EuReg analysis recommendations, the one most frequently prescribed in the low-dose group was omeprazole 20 mg two times per day. The most frequent prescriptions in the standard-dose and high-dose groups were omeprazole 40 mg two times per day and esomeprazole 40 mg two times per day, respectively. Therefore, our study suggests that the PPI dose required for achieving optimal results with any of the BQT should be omeprazole 40 mg or higher two times per day.

Data on the effectiveness of ScBQT were consistent with those of previous studies24 25 and meta-analyses26 which also revealed excellent cure rates. In fact, as previously stated, ScBQT was the only treatment to consistently achieve cure rates above 90% in all the European regions. These excellent ScBQT results in spite of the increasing bacterial antibiotic resistance in Europe endorse the Maastricht VI/Florence consensus report’s recommendation of BQT as the therapy of choice in areas of high bacterial antibiotic resistance.13 27 By contrast, since our study showed ScBQT to be highly effective independently of regional antibiotic resistances, our results argue against the recommendation of performing routine antibiotic resistance susceptibility testing for treatment selection. There is no evidence proving that susceptibility-guided treatment may be superior to an adequate empirical treatment28 and BQT has always been listed as a therapy that can be given empirically.29 Therefore, when using empirical ScBQT, the determination of antibiotic resistances would definitively not provide any advantage that might increase effectiveness.

The multivariate analysis results were in line with previous reports: that is, the cure of the infection was related to the use of 10-day ScBQT, prescription of 14 days of therapy in all other NScBQT and administration of standard or high doses of PPIs. Of all BQT, ScBQT was the one most significantly related to greater treatment success and the PPI+MDB combination was the one with the lowest effectiveness.

Finally, we found a wide variety of treatment schemes in use in clinical practice. Although the local patterns of antibiotic resistance may still allow the use of triple therapies in a few privileged areas and may thus explain some of the heterogeneity, the variability in the infection treatments administered remains largely unexplained. As situations of this kind have generally been associated with suboptimal quality of care,30 the study findings suggest that there may be considerable room for improvement in H. pylori treatment.

Adherence to BQT was good overall, and in fact, was better than expected. This may be due to the nature of the registry design, carried out by dedicated gastroenterologists or alternatively may reflect a relatively low sensitivity of the questionnaires used in the registry to detect non-adherence.

The overall incidence of AEs was comparable to that of previous reports, with around 40% of patients presenting at least one AE. The rate of serious AEs, however, was below 1%.31

The limitations and strengths related to the use of registry data have been extensively discussed elsewhere.23 32 33 The limitations include the risk of selection bias, the possible underreporting of AEs and the uncertainty regarding the outcome in patients who do not complete follow-up. To minimise this bias, previous studies have proposed the use of an mITT effectiveness analysis as the most reliable approach.23 Likewise, the categorisation of the main variables studied, such as the dose of PPIs and the regional analysis aided the interpretation of the great diversity of data and helped to provide more robust conclusions.

The main strength of the study is the large sample size, around 50 000 patients, which gave considerable power to the statistical analysis. In addition, the Hp-EuReg offers a faithful reflection of routine clinical practice and provides data on situations in which randomised trials will probably never be performed.23

Our study has important consequences for clinical practice. The results strongly suggest that BQT (and, specifically, ScBQT) should be made available in regions where these therapies are not currently marketed. BQT may be useful either as first-line treatment in areas of high bacterial antibiotic resistance or as rescue therapy in the few regions where standard triple therapy still achieves good results.24 In a plausible scenario of increasing resistances, BQT (and especially ScBQT) may become the treatment of choice. A second important finding was that none of the treatments achieved an effectiveness above 90% when combined with a low-dose PPI. In fact, the study corroborates the notion that the minimal PPI dose for achieving optimal BQT results should be set to 40 mg omeprazole or an equivalent PPI dose prescribed two times per day. This finding corroborates those of previous studies34 and suggests that the dose of PPI recommended in the technical specifications of ScBQT—omeprazole 20 mg two times per day—should be updated to 40 mg two times per day. An important related aim for future research will be to determine whether the marked acid suppression obtained by the potassium-competitive acid blockers (P-CABs), in comparison to PPIs, would also increase BQT efficacy.35 Although P-CABs are not currently available in most European countries, the Hp-EuReg may be key to determining the role of these promising new drugs in increasing the BQT cure rates. A second area that might need further research is whether ScBQT results might be improved by using a 14-day scheme. As this extended ScBQT scheme is currently rarely used in Europe, our study could not provide data regarding this point. Finally, H. pylori treatment changed considerably over the course of the study period, a circumstance that highlights the importance of monitoring the trends of both the use and the effectiveness of H. pylori eradication therapies.

In conclusion, BQT is one of the most effective H. pylori treatments currently available. Specifically, the use of ScBQT homogeneously obtains excellent eradication rates in all areas and its use has increased steadily in the countries in which it is available. This study suggests that expanding the availability of this therapy may reduce the heterogeneity that characterises clinical practice at present and may increase the overall effectiveness of H. pylori treatment.