Unresponsive thrombotic thrombocytopenic purpura in critically ill adults.

INTRODUCTION: The prognosis of thrombotic thrombocytopenic purpura (TTP) has considerably improved since the introduction of plasma exchange (PEX) therapy. However, unresponsive thrombotic thrombocytopenic purpura (Un-TTP) still carries high morbidity and mortality rates, indicating a need for early specific treatments.

PATIENTS AND METHODS: In a retrospective study including consecutive adults with TTP admitted between January 1997 and January 2011 in a teaching hospital intensive care unit (ICU), our objective here is to identify early clinical and laboratory features predicting Un-TTP. Patients who responded to plasma exchange and steroids (N = 49) were compared with patients with unresponsive TTP defined as requirement for other treatments, protracted course, or death (N = 37, 43 %).

RESULTS: Hospital mortality was 24.3 % in the Un-TTP group. Variables associated with Un-TTP on univariate logistic regression were older age, cardiac involvement, neurological involvement, higher anti-a disintegrin and metalloproteinase with thrombospondin motifs (ADAMTS13) immunoglobulin G (IgG) titer, lower platelet counts starting on day 2, higher Sequential Organ Failure Assessment (SOFA) scores starting on day 3, need for higher plasma volumes to obtain remission, and greater use of adjuvant treatments and life-sustaining interventions. Multivariate logistic regression identified four factors independently associated with Un-TTP: age over 60 years [odds ratio (OR) 7.90; 95 % confidence interval (95 % CI) 1.06-78.34], cardiac (OR 5.17; 95 % CI 1.63-16.39) or neurological (OR 8.04; 95 % CI 1.27-51.03) manifestations at diagnosis, and day 2 platelet count less than 15 G/l (OR 3.88; 95 % CI 1.30-11.62).

CONCLUSION: Therapeutic intensification starting on day 3 or even earlier in patients with the independent risk factors for unresponsive TTP identified in our study deserves evaluation in a multicenter prospective study.