H. pylori and platelet counts


  • Potential conflict of interest: Nothing to report.

  • Disclaimer: The opinion or assertions contained herein are the private views of the authors and are not to be construed as official or reflecting the view of the Department of the Army or the Department of Defense.


Helicobacter pylori (HP) eradication therapy is a useful treatment for idiopathic thrombocytopenic purpura (ITP). Some investigators have also reported the effects of proton pump inhibitor (PPI) monotherapy on ITP. We performed a randomized study of HP eradication therapy and PPI monotherapy on ITP. Four of nine patients achieved complete remission (CR), two of nine achieved partial remission (PR) in HP eradication therapy, three of eight achieved CR, and two of eight achieved PR in PPI monotherapy. No significant differences were observed in the CR + PR of these patients between HP eradication therapy and PPI monotherapy. As for cost comparisons, HP eradication therapy is cheaper than PPI monotherapy, but it is less effective.

Tsutsumi Y, Kanamori H, Yamato H, et al. Randomized study of Helicobacter pylori eradication therapy and proton pump inhibitor monotherapy for idiopathic thrombocytopenic purpura. Ann Hematol 2005;84:807-811. (Reprinted with kind permission of Springer Science and Business Media; http://springerlink.com/home/main.mpx)


Helicobacter pylori (HP) infection has been cited as a contributing factor in several erosive conditions involving the upper gastrointestinal tract mucosa including gastritis and peptic ulcer disease, as well as in malignant conditions like gastric adenocarcinoma and mucosa-associated lymphoid tissue lymphoma.1–3 However, a potential role for HP infection in autoimmune processes like rheumatoid arthritis, autoimmune neutropenia and idiopathic thrombocytopenic purpura (ITP) has also been described.4, 5 Numerous reports and studies have been published regarding the platelet response to the eradication of HP in humans.6–10

How HP infection might modulate the thrombocytopenia in ITP remains highly debated. Some authors have proposed that there may be no link between HP and ITP,11, 12 while there is a great deal of data, particularly from eastern Asia, which would suggest a benefit in determining the HP status of ITP patients and eradicating HP in those with confirmed infection.13–15 This discrepancy in data is possibly explained by the variations in study design, small study cohorts, heterogenous patient populations, and different treatment regimens and documentation of HP clearance. The production of anti-platelet antibodies (PAIgG) in HP-infected patients has been described, although the lack of specificity seen with PAIgG precludes its use in the diagnosis of ITP.6, 16, 17 More recently, antibody production to the more specific platelet surface autoantigen, glycoprotein IIb/IIIa (GP IIb/IIIa), has been documented in HP-infected patients. However, the presence of the anti-GP IIb/IIIa antibodies did not appear to have any significant influence on platelet response to HP eradication therapy in this small study.9 This same study demonstrated that antibodies against the HP cytotoxin-associated gene A (Cag-A) protein will react directly with platelet eluates from patients with ITP.9 Moreover, the prevalence of Cag-A containing strains of HP varies depending on geographic location, with Asia and particularly Japan having high rates of Cag-A positivity.18–20 This theory of molecular mimicry via the production of Cag-A antibodies that cross react with platelet surface antigens has been used as a potential explanation for the disparity in platelet response rates from varying studies conducted at different locations.17

Tsutsumi et al. attempt to clarify the relationship between HP infection and ITP as well as the potential effect of age, gender, presence of anti-HP antibodies and levels of PA-IgG in response to HP eradication therapy and lansoprozale monotherapy in their recent study. The study population included 17 adult patients with thrombocytopenia. These patients were given a diagnosis of ITP based on the presence of platelet counts <100 × 109/L and megakaryocytic bone marrow hyperplasia or normoplasia. Importantly, other causes of thrombocytopenia to include chronic hepatitis C, autoimmune disorders, drugs, and malignancy were excluded.

The first two sets of data summarize the pertinent clinical parameters upon patient entry into the study and confirm that there were no significant differences between the patient arms in the age, gender distribution, platelet count, PA-IgG and post-treatment length of observation. It should be noted that the clinical parameters between the HP eradication therapy and lansoprazole monotherapy study groups were similar. However, there was a significant difference in the levels of anti-HP antibodies at the onset of treatment between the two groups with the antibody levels being lower in the HP eradication therapy group. It is uncertain what effect the lower level of Anti-HP antibodies in the lansoprazole group had on the eventual results, but it is conceivable that this would have had a negative impact on the response rates to this treatment arm.

The remaining data review the treatment outcomes of the two groups. Tsutsumi et al. found no significant difference in the complete response (CR) (platelet count >120 × 109/L) and partial response (PR) (platelet count <120 × 109/L) rates between HP eradication therapy and lansoprazole monotherapy, with CR rates of 45% and 38%, respectively. While there was a trend towards greater platelet response for the HP eradication group one month post-treatment (CR + PR of 67% vs. 13% for lansoprazole monotherapy), this trend was reversed at 3 and 6 months post-treatment. No significant difference in PA-IgG levels between groups after treatment was found. Worthy of mention is that the investigators were only able to achieve an HP eradication rate of 45% in the triple drug, seven day treatment group. Interestingly, the authors conclude that combination therapy is superior to lansoprazole monotherapy. However, while there is a trend toward this effect, there is no significant difference seen in HP eradication between the two groups.

While intriguing, this study was only capable of making an indirect link between HP eradication and improvement in platelet counts. This link would have been stronger had the data shown a significant decrease in the PA-IgG levels between responders and non-responders after treatment. However, the authors note that recent studies have demonstrated that proton pump inhibitor monotherapy alone is capable of reducing anti-HP and anti-platelet antibodies in conjunction with elevation of platelet counts in ITP patients with HP infection.21, 22 It would be of interest to have longer post-treatment data to determine the durability of platelet response, although the follow up period defined by Tsutsumi et al. is one that has been commonly used in prior studies. Serotyping the HP organisms identified in infected patients would also have provided valuable information that might have supported the potential argument of geographic variations in serotypes and the effect of serotype on antigenicity. Given the adverse effects of longer durations of preceding thrombocytopenia on response rates to HP therapy, this characteristic should also have been accounted for.13 Use of control groups for lansoprazole monotherapy and HP eradication therapy in uninfected (non-HP) patients would have also been appropriate and may have helped determine the potential immunosuppressive effects of lansoprazole. Finally, the authors report an HP eradication rate of only 45% with seven days of triple drug therapy. This eradication rate is significantly lower than the historic rates for triple drug therapy.23, 24 Whether the reported resistance is a result of local antibiotic usage patterns or some other factor is unknown and was not reported.

We have found the potential association between HP infection and ITP intriguing. Among patients with cirrhosis, thrombocytopenia does not appear to be solely explained by splenic sequestration and low thrombopoietin secretion. Kajihara et al., has previously demonstrated the presence of platelet associated autoantibodies in patients with cirrhosis.25 Subsequently, in a recent pilot trial at our own institution, we determined the HP serology of 12 patients with HCV-induced cirrhosis and thrombocytopenia. All patients with positive serology were treated for 14 days with amoxicillin, clarithromycin and omeprazole and had platelet counts repeated 4 to 5 weeks after eradication therapy. Repeat platelet counts were obtained in HP negative patients 4 to 5 weeks after initial testing.

Seven patients had positive serology for HP while the remaining 5 patients were HP-negative and served as a control group. No difference was noted in platelet counts in the control group; however, a 32.5% increase (P = .017) in median platelet counts was noted in the HP group with treatment. The average change in platelet count (×1,000) in the control group was 5.2±14.5 and in the HP group was 17.0±14.5.

Five of 12 patients were on HCV antiviral therapy (AVT) with weight based ribavirin and either daily interferon alfacon-1 or weekly pegylated interferon alpha-2a or -2b at time of HP testing; 3 in the HP treated group and 2 in the control group. Mean change in platelet count within the HP group was (+) 13.8, compared with (−) 5.6 in the control group. In the remaining 4 patients treated for HP not on antiviral therapy, platelet counts increased by a mean of 38.9 compared with an increase of 12.8 in the 3 control group patients not on AVT.

This pilot trial suggests that treatment of HP positive patients with HCV-induced cirrhosis and severe thrombocytopenia may improve platelet counts. This improvement may allow for initiation of AVT in chronic HCV patients with cirrhosis who might otherwise be excluded due to severe thrombocytopenia. Limitations of this pilot trial include the small sample size, no clear documentation of HP eradication in the treatment group, and no formal platelet autoantibody testing throughout the study. Larger studies addressing these limitations are warranted. Further studies designed to identify specific immunogenic strains of HP or HP-induced platelet antibodies may, in turn, allow for the focused treatment of those patients with HP infection and thrombocytopenia most likely to benefit from HP eradication. Clearly the complexity of the immune system, the variation in immunogenicity between HP strains and lack of well-defined pathophysiology for HP-induced thrombocytopenia will continue to make research in this area very challenging.