Factors Associated with Mortality in Immune-Mediated Thrombotic Thrombocytopenic Purpura: Results from the United States Thrombotic Microangiopathy TTP Registry

Background: Thrombotic Thrombocytopenic Purpura (TTP) is a life threatening microangiopathy used by a deficiency in ADAMTS13 (a disintegrin and metalloproteinase with a thrombospondin type 1 motif, member 13). ADAMTS13 cleaves von Willebrand factor, and ADAMTS13 deficiency leads to an excess of high molecular weight multimers of von Willebrand factor which, in turn, causes platelet aggregation and congestion in small arteries and organ damage. TTP occurs in two forms: inherited due to a congenital defect in either production or function of the protein, and immune-mediated due to autoantibody formation either leading to immune clearance or inhibition of the function of the enzyme. TTP is a medical emergency and if untreated carries a mortality rate of approximately 90 percent. Current therapies (plasma exchange, immunomodulatory agents) have improved outcome in this disease, but mortality remains unacceptably high at approximately 10 percent. Attempts have been made to identify factors placing patients at higher risk for adverse outcomes, perhaps signaling the need for more aggressive treatment in these patients.

Design and Methods: We queried the United States Thrombotic Microangiopathy (USTMA) immune-mediated TTP registry which currently contains data from 181 patients with 316 distinct episodes of TTP in order to evaluate the French TMA consortium predictive model and identify additional factors associated with mortality.(Benhamou, 2012) There were 269 TTP episodes in 149 patients with sufficient data available to calculate the prognostic scoring system under review. Clinical and demographic variables from episodes resulting in death less than thirty days from diagnosis were compared with those in which patients survived. In the group of patients succumbing to the disease (n=13), 3 had multiple episodes of TTP captured in the registry. For two patients, episodes of TTP were separated by greater than 12 months. For a third patient, the initial episode of TTP was followed by an early relapse resulting in death 29 days after the first day of the first event. This death was counted once and attributed to the second episode.

Results: There were thirteen deaths in the cohort, yielding a mortality of 4.8%. Demographic factors associated with mortality were increasing age (mean age 55 vs. 44, p=0.01), and male gender (46% vs. 25%, p=0.30). Precipitating factors (surgery, infection, new medication) were more often present in those succumbing to the disease than those surviving (46% vs. 27%, p=0.20). Presenting symptoms seen more frequently in the mortality group were fever (38% vs. 13%, p=0.02) and neurologic symptoms (77% vs. 47%, p=0.02). Patients in the mortality cohort were more likely to have abnormal renal function with a higher mean creatinine (1.91 vs. 1.23, p=0.0004) and a higher proportion of patients with a creatinine over 1.3 (61% vs. 32%, p=0.04). Elevated troponin was also more common in the mortality group (39% vs. 16%, p=0.05). A mortality risk score using LDH 10 times normal, CNS involvement and age was calculated for each patient. When applied to our patient population, we found this score to have a sensitivity (% non-survivors with score ≥ 3) of 23% compared to 52% previously reported. We calculate a specificity (% of survivors scoring < 3) of 87% which is similar to previous reports of 90%. The positive predictive value for mortality with a score greater than or equal to 3 was 8%, and the negative predictive value of a score less than 3 was found to be 96% in the USTMA registry, compared to 41% and 93% reported by the French TMA consortium.

Conclusion: We find lower rates of mortality in our patient population as compared to previously reported series. Increasing age, neurologic involvement, renal dysfunction, and cardiac involvement seem to herald increase risk of early mortality. Sensitive tools to identify those at highest risk remain elusive.