ITP: the bad actors

Comment on McMillan and Durette, page 956

Patients with ITP who relapse after splenectomy or who do not benefit at all from a splenectomy are “bad actors”—they can be most challenging to treat.

Hematologists dealing with immune thrombocytopenic purpura (ITP) patients owe a great debt of gratitude to our Scripps colleague, Dr Robert McMillan. His professional career has been preoccupied with this illness. He wrote a seminal review article on ITP in the early 1980s.¹  For me, it was a privilege to witness his involvement in the promulgation of the American Society of Hematology ITP guideline.²  During those meetings, Dr McMillan would focus on what we did not know about ITP, without having the benefitofa large number of randomized controlled trials. We were forced to come up with a rating system for how we would deal with ITP diagnostic and therapeutic challenges. After the ITP guideline, Dr McMillan assumed responsibility for organizing a group of hematologists interested in ITP. There have been several innovative spin-offs and collaborations from these interactions, which precede the annual ASH meeting.FIG1 

Now, McMillan and Durette give us a treatise on the natural history of postsplenectomy refractory ITP. Admittedly, it is data derived from a prospective study of antiplatelet antibodies, but it is a rich, multiyear follow-up. For patients who fail to develop a remission after splenectomy, if watched long enough, the majority (75 of 105) eventually attain a stable complete or partial remission. It may take more than 4 years, but they get there. However, 17.6% of failures to splenectomy experience recurrent bleeding, hospitalizations, and death.

Dr McMillan arbitrarily separates his patients into 4 groups, based on intensity of treatment. Research in ITP would benefit from a standardized staging system, as in oncology, with the TNM classifications (tumor size, node status, and metastases). Comparing this Scripps series with other population-based studies³  is difficult because of the tertiary center referral bias as well as the fact that there are differences between groups of ITP patients. This arbitrary grouping of McMillan's might be a start toward staging. However, as stated, groups of ITP patients are heterogeneous. Two patients with 5 × 10⁹/L platelets can have dramatically different clinical behaviors. One might be developing deep venous thromboses and the other, intracranial hemorrhage. We are so fixed on the platelet count that we may be missing other, more important biologic markers of risk for bleeding. We have to be able to determine which of our splenectomy failures is destined to have an unfavorable outcome. We need to develop better techniques to predict bleeding risk, so as to avoid overtreatment (with its attendant toxicities) for those severely thrombocytopenic patients who are not bleeding. Newer agents may hold promise for those 17.6% of splenectomy failures in the McMillan series.