Considering Evidence for a New Drug for Immune Thrombocytopenia Purpura
I’ve been wondering, lately, why so many of the medical blogs cover the same topics, like last week’s lung cancer detection trial, which are often the exact same studies as are reported by conventional news outlets. I’ve been trying, here, to sometimes consider new published articles that seem important to me but, for whatever reasons, don’t get so much attention.
Yesterday’s NEJM includes an article Romiplostim or Standard of Care in Patients with Immune Thrombocytopenia.* It’s about a drug, manufactured and sold by Amgen as NPlate, that received FDA approval for treatment of chronic immune thrombocytopenia purpura (ITP) in August, 2008. Some consider ITP a rare disease, and I suppose it is; my perception is skewed by the fact that as a practicing hematologist I saw patients with ITP almost constantly in the clinic and hospital. In patients with this autoimmune condition, the body produces pathologic antibodies that bind and lead the tiny, clot-forming blood cells to destruction. Sometimes patients with chronic ITP need treatment, such as steroids, surgical removal of the spleen (splenectomy) or other drugs that raise the platelet count and reduce their risk of bleeding.
The new-ish drug, romiplostim, is effective in raising the platelet count in most patients who are said to have chronic ITP. It’s similar to another platelet-stimulating agent, eltrombopag, that’s manufactured and sold by GlaxoSmithKline under the brand name Promacta. Eltrombopag received FDA approval for the same condition in November, 2008.
So long before reading the thoughtful editorial* by Dr. James George, this hematologist was impressed by the availability of two agents that work by stimulating platelet production in the bone marrow. Dr. George considers the possibility that these new drugs alter the usual but debatable algorithms that most hematologists use in caring for patients for ITP. He acknowledges having consulted to the study’s sponsor, Amgen, on this drug and serving as an investigator in clinical trials, and considers:
…In this randomized study, 85 investigational sites in 14 countries enrolled 234 patients…The conclusions are clear. The outcomes were better in patients receiving romiplostim than in patients receiving standard care (short of splenectomy): romiplostim was associated with a greater incidence of a sustained platelet response, less bleeding and fewer transfusions, a decreased requirement for other treatments (including splenectomy), and greater improvement in quality of life. The side effects of romiplostim therapy were minimal, but …confidence about the safety of the drug requires that more patients be observed for a longer time. Because patients treated with romiplostim had better outcomes, does the work of Kuter and colleagues establish romiplostim as the new standard of care?
As I read the original study results, this issue – of possibly changing the standard of care in this disease – became evident. His question is on-point.
The problem is this: In the study, the authors treat a total of 234 patients who are said to have chronic ITP. But, if you look closely at the eligibility for enrollment, besides having a low platelet count (defined for the study as less than 50,000 platelets per microliter of blood, which in itself is a questionable criterion), examination of a bone marrow-biopsy specimen was required to confirm the diagnosis of immune thrombocytopenia in patients older than 60 years. According to Table 1 in the paper, the median age for study participants was 57 years. This means that more than half of the participants were under 60 years and, as far as we know, they did not have a diagnostic pathology test for ITP. They might instead have a myelodysplastic syndrome or other bone marrow disorder that was missed.
Why this matters, besides to patients who have low platelets and their hematologists, is that it’s really a problem in the practice of evidence-based medicine. If data are derived and published from fundamentally flawed studies, in which the patient groups who respond or don’t to a particular intervention are not well-defined by careful diagnostic parameters, the conclusions we draw from those studies may be incorrect.
I am definitely in favor of evidence based medicine, but we need to be careful in how we evaluate the results of published studies, even in the best of journals, and how we incorporate those findings into recommendations for care.