Introduction

Immune thrombocytopenic purpura (ITP) is an autoimmune disease [1, 2] with low platelet count, destroyed by antiplatelet antibodies, via autoreactivity expression of T and B cells, suppressing platelet production of megakaryocytes in the bone marrow and reducing platelets in peripheral blood. For primary ITP treatment, immunosuppressants such as steroids, intravenous immunoglobulin (IVIG), and anti-D immunoglobulin are used. Adult ITP treatment guidelines recommend starting treatment when bleeding symptoms are noticed or when platelet count drops to 30 × 103/μL or less, taking into account the effectiveness of the treatment, toxicity, risk of bleeding without treatment, and cost-effectiveness [3, 4].

The association between ITP and Helicobacter pylori (H. pylori) has been reported. In 1998, Gasbarrini et al. [5] reported that H. pylori eradication as a treatment could improve thrombocytopenia in ITP patients. In 2009, a meta-analysis study reported that the platelet response rate was more than 50% by H. pylori eradication as a treatment [6,7,8]. Besides, platelet count improvement after H. pylori eradication as a treatment was 8.8% for H. pylori-negative ITP patients and 51.2% for H. pylori-positive ITP patients. The odds ratio for this improvement was 14.5 (95% confidence interval: CI 4.2 ~ 83.0) [9]. However, most studies reporting the efficacy of H. pylori eradication as a treatment for ITP were retrospective, observational studies. There were only a few randomized studies or phase 2 and 3 studies [10,11,12,13,14,15,16].

Domestic and international guidelines for H. pylori eradication suggest standard 3-drug treatment including proton pump inhibitor (PPI) as the first-line treatment. The eradication failure rate is known to be about 20% [17]. The eradication success rate of the standard 3-drug treatment is decreasing due to an increase in antibiotic resistance [18, 19]. As an alternative to the primary treatment, sequential and concomitant treatment has been proposed [20, 21]. In the sequential treatment, PPI and amoxicillin are administered in the first half followed by administration of PPI, clarithromycin, and metronidazole in the second half. This method is superior to the standard 3-drug treatment and its effect persists even in a long-term observation [22,23,24]. A meta-analysis has reported that the sequential treatment has a higher elimination rate (79.4% vs. 68.2%) with similar compliance and side effects than the standard 3-drug treatment [25].

According to ITP treatment guidelines, treatment is not recommended for patients with moderate ITP with platelets greater than 30 × 103/ μL. Although such patients do not have severe bleeding symptoms, their physical and mental quality of life could be affected by procedures or surgery, vigorous physical activity, and restrictions on career options. Therefore, if long-term effective and safe treatment is developed, active treatment could be considered for this patient group instead of simple observation. Considering the cost, effectiveness, and side effects of H. pylori eradication therapy based on available data, it can be said that this therapy is close to an ideal treatment for H. pylori-positive ITP patients. However, existing studies have limited data on only moderate thrombocytopenia patients with ITP. Besides, there is currently no published study on whether platelet response is improved in H. pylori-positive ITP patients when sequential treatment as a new H. pylori eradication therapy is used. Therefore, we conducted a phase 3 study to confirm the efficacy and safety of H. pylori-positive ITP treatment in H. pylori-positive ITP patients with moderate thrombocytopenia using only sequential eradication treatment.

Methods

This was a prospective, multicenter, non-blind, randomized phase 3 study conducted in eight Korean institutions. The study protocol was approved by institutional review boards and registered with ClinicalTrials.gov (NCT03177629). All patients voluntarily decided to participate in this study, and written consents was obtained.

Eligible patients were adults over 19 years of age with persistent (lasting 3 to 12 months after diagnosis) or chronic (lasting more than 12 months after diagnosis) ITP as defined by the International Working Group and having platelet counts of 30 × 103/μL ~ 80 × 103/μL. Eligible patients were enrolled in this study when H. pylori was confirmed to be positive by urea breath test (UBT), rapid urease test, stool antigen test, or biopsy through gastroscopy. Exclusion criteria were previous H. pylori eradication treatment, any treatment for ITP within the last 3 months, an underlying disease or drug history that might cause secondary thrombocytopenia, a history of bleeding disease or at risk of bleeding from taking antithrombotic or aspirin.

The failure rate of H. pylori eradication therapy was assumed 20% for the treatment group. The randomization ratio between the treatment group and the control group was set to be 5:4. Patients were assigned to the treatment group or the control group according to a stratified block randomization table. Stratification factors were (1) the presence or absence of previous ITP treatment history and (2) platelets greater than 30 × 103/μL but less than 50 × 103/μL versus platelets greater than 50 × 103/μL but less than 80 × 103/μL.

The treatment group was sequentially administered four drugs as follows: pantoprazole 40 mg per os (PO) b.i.d., amoxicillin 1,000 mg PO b.i.d. for days 1–5, followed by pantoprazole 40 mg PO b.i.d., clarithromycin 500 mg PO b.i.d., and metronidazole 500 mg PO t.i.d. until days 6–10. At 3 months, success in both eradication and platelet response was confirmed in the treatment group. The control group was followed for the first 3 months. At 3 months, the same eradication treatment was performed for the control group as for the treatment group. At 6 months, eradication success and platelet response were checked.

The success of H. pylori eradication was confirmed as a negative result of UBT at 3 months of treatment. Platelet counts were checked at 1, 2, 3, and 6 months. Response evaluation followed the definition of the ITP treatment response evaluation criteria of the International Working Group. If the value of platelet count was more than 100 × 103/μL, it was called a complete response (CR). If it was more than 30 × 103/μL and at least two times higher than the baseline value, it was called response (R). If it did not meet the criteria of R, it was called no response (NR) [3]. Safety evaluation was performed for all patients who took the antibacterial treatment drug for more than one day. The severity of adverse events was graded according to the National Cancer Institute Common Toxicity Criteria (NCI-CTCAE) (Version 4.0). Drug compliance was defined as taking 85% or more of the drug.

The primary objective was the overall response rate (CR + R) based on platelets at 3 months with successful eradication and confirmed UBT negative in the treatment group and control group. Secondary objectives were the time to platelet response, success rate of H. pylori eradication, clinical characteristics between platelet responders, and non-responders after treatment, safety, and drug compliance. Among the clinical features of enrolled patients, the hemorrhagic symptom score was measured before and after treatment using the clinical hemorrhage scoring system [26].

The platelet response rate at 3 months was assumed 50% for the treatment group with successful eradication and 15% for the control group. The sample size of each group required to test this hypothesis at a one-sided significance level of 5% and power of 80% was 21 (actually, 27 in consideration of eradication failure rate in the treatment group). A total of 54 patients (30 in the treatment group and 24 in the control group) were enrolled in consideration of a 10% dropout rate.

Results of all patients included in the randomization were analyzed according to the intention-to-treat principle. Platelet responses in the treatment group of patients with H. pylori successfully eradicated, and the control group was analyzed by chi-square test or Fisher’s exact test. Clinical difference between platelet responders and non-responders was analyzed by the Kruskal–Wallis test. The period until platelet response and the cumulative platelet response were analyzed by Gray’s test. All data were analyzed using SAS version 9.4 (SAS Institute, Cary, NC, USA) (http://www.sas.com).

Results

A total of 28 participants (10 males and 18 females) were registered from May 2017 to May 2020. At the time of randomization (17 to the treatment group and 11 to the control group), the Korean national insurance and treatment guidelines for H. pylori-positive ITP patients were changed. Thus, subject registration was terminated early. Out of 17 subjects for the treatment group who received the eradication treatment, 15 were found to be UBT negative at 3 months, showing an H. pylori eradication rate of 88.2%. One patient had eradication failure and one patient dropped out due to worsening thrombocytopenia (Fig. 1). Four out of 11 control subjects withdrew their consent after randomization. Baseline disease characteristics such as age, previous ITP treatment history, baseline platelet counts, and hemorrhagic symptom score did not show statistically significant differences between the treatment group and the control group and between the treatment group with successful eradication and the control group (Table 1). Three patients in the treatment group and two patients in the control group received treatment before the study enrollment. Two received corticosteroid and IVIG; the other three received IVIG, eltrombopag, and danazol monotherapy, respectively.

Fig. 1
figure 1

Study consort diagram. The platelet response rate at 3 months as the primary endpoint was 60% for the treatment group with successful eradication and 0% for the control group (p = 0.017). aUrea breath test (UBT) not done: At the time of treatment initiation after the screening, one patient experienced decrease of platelet count to less than 30 × 103/μL. Hence, we terminated this study process after follow-up up to 2 months where the patient did not undergo the follow-up UBT. bUBT+: failure of H. pylori eradication. cUBT-: success of H. pylori eradication. dR: platelet response. eNR: no platelet response. fDrop-out: Patient who withdrew from the study after randomization. gNo eradication: At 2 months, the platelet count was 98 × 103/μL. The study was terminated at the discretion of the investigator

Full size image
Table 1 Baseline characteristics of the study populationa
Full size table

At 3 months after the beginning of treatment, 9 out of 15 patients in the treatment group had successful eradication (UBT negative) with a CR (more than 100 × 103/μL platelets). The remaining six patients did not show a response based on platelet counts. All seven control subjects showed no response at 3 months based on platelet counts. Therefore, the platelet response rate at 3 months as the primary endpoint was 60% for the treatment group with successful eradication and 0% for the control group, showing a statistically significant difference (p = 0.017) (Fig. 1). The median difference (interquartile range, IQR) of the hemorrhagic symptom score between before and after treatment was 0.00 (0.00, 0.00) for the treatment group with successful eradication and 0.00 (–0.75, 0.00) (p = 0.031) for the control group. H. pylori eradication therapy, which had been planned with 3 months of follow-up, was performed for six out of seven control patients. Of these, three had UBT negative results at 6 months of follow-up. Platelet response was confirmed as CR in one patient.

Of all 23 patients who received the sequential treatment in both of the treatment group and the control group, 18 achieved successful eradication at 3 months after eradication, showing an H. pylori eradication rate of 78.3%. The platelet response rate 3 months after H. pylori eradication was 55.6% in both groups with successful eradication.

The median time to platelet response in patients with successful eradication was 2 months (95% CI: 0.49–3.52 months) (Fig. 2). Platelet response rates at 1 month, 2 months, 3 months, and 6 months were 33.3%, 53.3%, 60.0%, and 64.3%, respectively. Continuous platelet reactions were observed.

Fig. 2
figure 2

Time to platelet response after H. pylori eradication pylori eradication. The median time to platelet response was 2 months (95% CI, 0.49–3.52 months) in patients with successful eradication among treatment groups. One subject in the control group showed platelet response at 1 month that was lost at 3 months and regained it at 6 months (the time of 3 months after eradication). aH. pylori eradication treatment was performed for the control group at 3 months. Gray’s test for was used

Full size image

The median platelet counts (IQR) at 3 months of follow-up were 108.0 × 103/μL (46.5, 160.3) and 56.5 × 103/μL (50.3, 74.8) in the treatment group with successful eradication and the control group, respectively (p = 0.070), and those at 6 months were 139.5 × 103/μL (59.3, 152.5) and 66.0 × 103/μL (50.5, 106.8) (p = 0.193), with the treatment group showing a tendency to have higher platelet counts than that of control group.

The drug compliance was 95.7% for 17 patients in the treatment group and six patients in the control group who underwent eradication at 3 months. Of all patients, 78.3% completed the full dose according to the planned regimen. During treatment, adverse events caused by drugs were mainly gastrointestinal symptoms, with nausea being the most frequent. All adverse events were tolerable with grades 1 and 2 (Table 2).

Table 2 Toxicities observed during the treatment period
Full size table

To determine which factors influenced platelet response, post hoc analysis of clinical and laboratory factors was performed for patients with platelet response at 3 months of follow-up [platelet responder group, 9] and those in the platelet non-responder group [14] regardless of the presence or absence of H. pylori eradication treatment or the randomization group (Table 3). All nine patients in the platelet responder group were treated for H. pylori. The platelet non-responder group included seven participants in the treatment group and seven in the control group. There were no significant differences in age, baseline platelet count, prevalence period after ITP diagnosis, or previous ITP treatment history between the two groups. However, the proportion of women in the platelet non-responder group was significantly higher at 85.7%. In 100% of subjects of the platelet responder group, UBT was negative at 3 months of follow-up, confirming the success of H. pylori eradication. In the platelet non-responder group, only 42.9% of subjects were found to be UBT negative, significantly (p = 0.005) lower than in the platelet responder group. In the multivariate analysis, success or failure of H. pylori eradication was the only independent predictor of platelet response at 3 months after eradication (HR 0.4, 95% CI 0.006–0.794, p = 0.047). We conducted similar analysis in 15 patients who achieved successful eradication where no statistically significant differences in clinical characteristics between the platelet responders and non-responders were seen.

Table 3 Comparison of clinical characteristics between treatment responders and non-responders at 3 months after H. pylori eradication
Full size table

Discussion

According to existing treatment guidelines, when there is no bleeding and the platelet count is 30 × 103/μL or more, patients with ITP can only be examined regularly without any specific management. This is due to treatment-related adverse effects, the possibility of long-term treatment, and the financial burden of treatment such as corticosteroid, IVIG, splenectomy, thrombopoietin agonist, and rituximab. As a result, ITP patients with moderate thrombocytopenia who were not indicated for the treatment are continuously exposed to bleeding risks, even if they are not at high risk.

For ITP patients with confirmed H. pylori infection, eradication may be considered as a treatment option. Based on clinical experiences, several studies have attempted to elucidate the mechanism of H. pylori-related ITP. Hypotheses presented so far include the molecular mimicry hypothesis stating that anti-CagA antibodies against H. pylori can cross-react with platelet surface antigen [27], autoreactivity expression of T cells and B cells as a response of the host immune system after H. pylori infection [28], and the pathogenic hypothesis stating that the platelet autoantibody continues to respond by modulating monocyte/macrophage function as a result of this autoimmune response [29]. However, the mechanism cannot be fully explained by one factor. A multi-factorial mechanism might be involved [30].

In particular, H. pylori has been pointed out as a cause of benign diseases such as atrophic gastritis, intestinal metaplasia, peptic ulcer disease, malignant tumors such as gastric cancer and mucosa-associated lymphoid tissue lymphoma, and ITP. For this reason, guidelines in Korea and Japan known to have a high prevalence of H. pylori recommend eradication treatment when H. pylori infection is diagnosed in ITP patients with mild or moderate severity [31, 32].

The evidence of eradication as a treatment of ITP has been proven based on the previous studies showing platelet responses after H. pylori eradication in ITP patients. However, in most of these studies, the target patient group was not homogeneous, and H. pylori eradication was commonly combined with other ITP treatments such as steroids [10,11,12,13,14,15, 33]. Therefore, we conducted a phase 2 study with standard 3-drug treatment enrolling only a group of ITP patients with moderate thrombocytopenia who had not been treated in the past [16]. In this study, the overall response rate was 57.7% at 3 months of treatment and 30.8% at 12 months of treatment. Most (87.5%) responders showed complete responses. This confirms an effective and durable response with relatively few side effects. However, recent studies on H. pylori eradication have reported that the eradication rate of the existing standard 3-drug treatment is less than 70% due to drug resistance [18, 19]. Sequential eradication treatment to overcome this problem has been proposed, and an excellent eradication rate of about 80% has been reported [21,22,23,24]. Considering drug resistance, this study attempted a sequential treatment rather than the standard 3-drug treatment and determined whether appropriate platelet response could be obtained with this new treatment. To the best of our knowledge, this is the first study to conduct a sequential H. pylori eradication monotherapy, multicenter, phase 3, randomized study in patients with moderate thrombocytopenia only.

In this study, the H. pylori eradication rate at 3 months was 88.2% for patients who received the sequential eradication treatment. This percentage was higher than those reported in recent studies using a standard 3-drug treatment. This sequential eradication treatment showed excellent drug compliance without causing significant toxicity. Besides, the final platelet response rate of 60% was similar to those reported after the conventional 3-drug treatment [10,11,12,13,14,15]. Therefore, the use of sequential eradication treatment might be expected to have an increasing effect on platelet responders for ITP patients compared to the existing 3-drug treatment. Besides, considering the association between H. pylori and other diseases, choosing a treatment regimen with a high eradication rate for H. pylori-positive ITP patients not only can improve platelet response, but also can prevent other H. pylori-related diseases including lymphoma and gastric cancer. The higher eradication rate of the sequential treatment used in this study is helpful for the prevention and symptom improvement of H. pylori-induced other diseases by itself. Since the drug price is relatively inexpensive, such treatment can be said to be ideal.

This study did not evaluate long-term response for more than 12 months as per the authors’ previous phase 2 study. However, at 3 months and 6 months, it was confirmed that the total platelet response rate was steadily maintained at 60% or more. Besides, six patients in the control group who did not have a platelet response at 3 months were subjected to the same eradication afterwards. Their median platelet counts also tended to increase at 6 months. To confirm the effect of bleeding symptom improvement in these patients, the median difference of the hemorrhagic symptom score with IQR between before and after treatment was also analyzed and observed to be improved in patients in the successfully eradicated treatment group compared to those in the control group. However, this finding was thought to be due to a minor bleeding event in one patient in the control group, and clinically significant bleeding did not occur in both groups of patients. After the diagnosis of ITP in patients participating in this study, the median prevalence period until they were enrolled in this study was 17.9 months (IQR, 3.1–52.7 months). About 20% of these patients had a history of using treatments including steroids before. Therefore, the results of the present study suggest that eradication treatment alone can produce an excellent response, regardless of the prevalence duration or past ITP treatment history.

A previous study suggested that some platelet improvement after eradication treatment in H. pylori-positive ITP patients might be due to other reasons rather than by the eradication, including an immune modulating effect by antibiotic therapy itself or by eradication of other commensal bacteria [34]. However, the fact that platelet response was observed in patients who achieved successful H. pylori eradication even when the eradication regimen was changed as shown in this study suggested that the eradication success itself was related to the improvement of platelet count, and not because of specific antibiotics. This has also been reported in a retrospective study analysis of more than 100 patients [35].

In this study, factors predicting platelet response were evaluated. Univariate analysis revealed that the platelet response was decreased in women (Table 3). However, the multivariate analysis did not show a statistically significant difference in platelet response between males and females. This might be due to a high proportion of women in the total patient group and the control group. The result that there was no difference in treatment effect according to sex was consistent with several previous studies that observed platelet response after ITP treatment including H. pylori eradication [36,37,38,39,40,41,42,43,44].

In contrast to results of study in Asian countries such as Korea and Japan, several previous studies conducted in other regions were not significantly superior, with platelet responses of 7–30%, after H. pylori eradication treatment [40, 45,46,47,48,49], and a meta-analysis has shown similar results [6]. The influence of ethnic differences and heterogeneity of the study participants can be considered as the cause of such difference. Ethnic differences in H. pylori species and differences in H. pylori prevalence have been reported in several studies [50, 51]. The genetic diversity of H. pylori strains may be the reason for such H. pylori prevalence difference and ethnic differences in platelet response after eradication. Besides, when geographical grouping was performed through bacterial sequencing, a pattern similar to racial distribution according to human migration that has been hypothesized and explained [52]. For this reason, H. pylori detection and eradication in areas with high prevalence and virulence are very reasonable and scientific approach for patients with ITP. In particular, when applying a treatment, it is necessary to select an appropriate treatment in consideration of H. pylori prevalence, ethnic differences of strains, and differences in resistance to antibiotics by region. International studies are needed in the future to clarify existing hypotheses about ethnicity differences of H. pylori’s prevalence, virulence, and ITP treatment response.

Several studies in which platelet response is poor after eradication mentioned above, the therapeutic effect of H. pylori eradication might have been underestimated due to heterogeneous inclusion of ITP patients with moderate and severe thrombocytopenia, a study design allowing treatment with steroids in the control group. Therefore, in H. pylori eradication studies, a study design that selects more homogenous ITP patients is needed to help determine the effectiveness of the eradication treatment.

However, in this study, indication of H. pylori eradication for ITP patients was changed due to extended insurance coverage in Korea. Therefore, the study was terminated early because an ethical problem could arise if the treatment is delayed for the control group. Due to this, the total number of enrolled patients was smaller than planned. Besides, because our study was non-blind, four patients in the control group withdrew from the study post- randomization after recognizing their assignment. This was another limitation of our study in terms of small sample size. However, results of this study still showed a positive effect of the sequential eradication treatment with a statistical significance.

Conclusion

This study confirmed that a sequential eradication treatment was effective and tolerable in H. pylori-positive ITP patients with moderate thrombocytopenia. Therefore, this might give justification for the evaluation of H. pylori infection and eradication in the H. pylori-positive ITP patients with mild to moderate thrombocytopenia, currently recommended only for follow-up without treatment. In addition, the sequential treatment could be considered as a standard eradication regimen for H. pylori-positive ITP, especially in the areas where drug resistance is increasing.