Comparison of Clinical Outcomes Between Initial and Relapsed Events of Thrombotic Thrombocytopenic Purpura In a Large TTP Cohort.

Thrombotic thrombocytopenic purpura (TTP) is a serious clotting disorder characterized by high mortality and morbidities. It is known that as many as 50% of acquired cases of TTP will relapse in the future. However, there has been no report yet comparing the clinical outcomes and therapeutic responses between initial and relapsed events of TTP. Such knowledge may be important to guide clinicians in managing TTP patients. More importantly, since TTP is a rare disease, a full evaluation of relapsed vs. initial episodes of TTP would help clinical trialists to decide whether to enroll the relapsed patients into therapeutic trials.

Between 2000 and 2010, the Ohio State University TTP Program has been treating all TTP patients using a uniform protocol for plasma exchange (PE) therapy. There were a total of 112 acute episodes of TTP from 52 study subjects, 43 of whom were treated at OSU for both their initial and relapsed events when they occurred. The other ten patients were referred to OSU for treatment after relapse of their disease. In this study, pre-PE and longitudinal samples were available and ADAMTS13 biomarker data were analyzed in 75% of episodes. Differences in clinical outcomes (e.g. treatment success rate, number of PE procedures to achieve response) and lab/correlative endpoints (e.g. platelets, LDH, ADAMTS13 biomarkers) were evaluated based on episode type (initial vs. relapsed) and episode number (ordinal marker with initial=1, first relapse=2, etc.) using generalized estimating equation (GEE) models. All 112 TTP episodes were analyzed in these models, with the GEE model also assessing the impact of intra-patient correlation. Cox regression models were used to assess differences in time to relapse based on episode type and number.

The majority of our cohort were female (70% pts; 68% episodes), Caucasian (66% pts; 76% episodes), and had not received immunosuppressive therapy within 30 days of the event (80% pts; 70% episodes). No significant differences were observed between relapsed vs. initial episodes in treatment success rate (p=0.15), exacerbation rate (p=0.26), and number of PE procedures required to achieve a clinical response (p=0.15). The only clinical outcome that showed significant intra-patient correlation between TTP episodes was the number of PEs required to achieve a clinical response, although this was a relatively small correlation (r = 0.20). Time to relapse between initial (I) and relapsed (R) groups was also not significantly different when adjusting for intra-patient correlation (p=0.31). At the time of acute presentation, significant differences were seen between I and R groups in terms of platelet count (I<R; p<0.001), LDH (I>R; p<0.001), and hemoglobin (I<R; p<0.001). No significant differences were observed in ADAMTS13 biomarkers (ADAMTS13 activity, antigen, IgG levels and inhibitor titers), although inhibitor titer levels were significantly different based on episode number, where they appeared to decrease with subsequent relapse episodes (p=0.02). Although significant differences in ADAMTS13 antigen were not observed based on TTP episode type, there was notable intra-subject correlation (r = 0.36; p=0.001).

This study is based on a relatively large data set for such a rare disease. Using a uniform TTP treatment protocol, the major clinical outcomes and therapeutic responses are similar between initial and relapsed TTP episodes. Differences in routine laboratory values can be attributed to the tendency for patients to seek medical care sooner given their awareness of their diagnosis. Considering that TTP is a rare disease with associated difficulties in enrolling sufficient study subjects, these data support the strategy to enroll TTP patients with a relapsed event that is clinically distinct from previous episodes. However, we recommend against re-enrollment of patients for subsequent TTP episodes in the same clinical trial given there is some intra-patient correlation. Considering the moderate sample size of this study from statistical perspective, we recommend that future randomized studies stratified based on episode type to balance the proportion of de novo and previously treated patients to avoid any potential bias.

No relevant conflicts of interest to declare.