DOI: /jth.12790

Transcription

1 Journal of Thrombosis and Haemostasis, 13: DOI: /jth ORIGINAL ARTICLE Cardiac troponin-i on diagnosis predicts early death and refractoriness in acquired thrombotic thrombocytopenic purpura. Experience of the French Thrombotic Microangiopathies Reference Center Y. BENHAMOU,* P.-Y. BOELLE, B. BAUDIN, ** S. EDERHY, J. GRAS, L. GALICIER, E. AZOULAY, *** F. PROVÔT, E. MAURY, F. P EN E, J.-P. MIRA, A. WYNCKEL, **** C. PRESNE, P. POULLIN, J.-M. HALIMI, Y. DELMAS, ***** T. KANOUNI, A. SEGUIN, C. MOUSSON, A. SERVAIS, D. BORDESSOULE, ****** P. PEREZ, M. HAMIDOU, A. COHEN, A. VEYRADIER and P. C O P P O, ******* *Service de Medecine Interne, CHU Charles Nicolle; Inserm U1096, Rouen; Centre de Reference des Microangiopathies Thrombotiques, Hôpital Saint-Antoine, AP-HP; Faculte demedecine Saint-Antoine, Inserm U707; UPMC Univ. Paris 06, Sorbonne Universite; **Service de Biochimie, Hôpital Saint-Antoine, AP-HP, Universite Paris-Sud; Service de Cardiologie, Hôpital Saint-Antoine, AP-HP; Service d Hematologie, Hôpital Saint-Antoine, AP-HP; Service d Immunologie Clinique, Hôpital Saint-Louis, AP-HP; Sorbonne Paris Cite, Univ. Paris Diderot; ***Service de Reanimation Medicale, Hôpital Saint-Louis, AP-HP, Paris; Service de Nephrologie, Hôpital Albert Calmette, Lille; Service de Reanimation Medicale, CHU Saint-Antoine, AP-HP; Service de Reanimation Polyvalente, Hôpital Cochin, AP-HP; Universite Paris 5, Paris; ****Service de Nephrologie, Hôpital Maison Blanche, Reims; Service de Nephrologie Medecine Interne, Hôpital Sud, Amiens; Service d Hemapherese, Hôpital de Marseille Conception, Marseille; Service de nephrologie-immunologie clinique, Hôpital Bretonneau, Tours; EA 4245, Universite Francßois-Rabelais, Tours; *****Service de Nephrologie Transplantation Dialyse, Centre Hospitalier Universitaire, Bordeaux; Unite d Hemapherese therapeutique, CHU de Montpellier, Montpellier; Service de Reanimation Medicale, Centre Hospitalier Universitaire de Caen, Caen; Service de Nephrologie, CHU de Dijon, Dijon; Service de Nephrologie, Hôpital Necker-Enfants Malades, AP-HP, Paris; ******Service d Hematologie Clinique et de Therapie Cellulaire, CHU Dupuytren, Limoges; Service de Reanimation, CHU Brabois, Nancy; Service Medecine Interne A, Hôpital Hôtel-Dieu, Nantes; Service d Hematologie Biologique, Hôpital Antoine Beclere, AP-HP; Universite Paris-Sud 11, Paris; and *******Inserm U1009, Institut Gustave Roussy, Villejuif, France To cite this article: Benhamou Y, Boelle PY, Baudin B, Ederhy S, Gras J, Galicier L, Azoulay E, Provôt F, Maury E, Pene F, Mira JP, Wynckel A, Presne C, Poullin P, Halimi JM, Delmas Y, Kanouni T, Seguin A, Mousson C, Servais A, Bordessoule D, Perez P, Hamidou M, Cohen A, Veyradier A, Coppo P. Cardiac troponin-i on diagnosis predicts early death and refractoriness in acquired thrombotic thrombocytopenic purpura. Experience of the French Thrombotic Microangiopathies Reference Center. J Thromb Haemost 2015; 13: Summary. Background: Cardiac involvement is a major cause of mortality in patients with thrombotic thrombocytopenic purpura (TTP). However, diagnosis remains Correspondence: Paul Coppo, Centre de Reference des Microangiopathies Thrombotiques, Service d Hematologie, Hoˆpital Saint- Antoine, Universite Pierre et Marie Curie, 184 rue du Faubourg Saint-Antoine, Assistance Publique Hôpitaux de Paris, Paris 75012, France. Tel.: ; fax: paul.coppo@sat.aphp.fr Please see the Appendix for the Members of the Reference Center for Thrombotic Microangiopathies. Received 14 August 2014 Manuscript handled by: L. Zheng Final decision: F. R. Rosendaal, 6 November 2014 underestimated and delayed, owing to subclinical injuries. Cardiac troponin-i measurement (ctni) on admission could improve the early diagnosis of cardiac involvement and have prognostic value. Objectives: To assess the predictive value of ctni in patients with TTP for death or refractoriness. Patients/Methods: The study involved a prospective cohort of adult TTP patients with acquired severe ADAMTS-13 deficiency (< 10%) and included in the registry of the French Reference Center for Thrombotic Microangiopathies. Centralized ctni measurements were performed on frozen serum on admission. Results: Between January 2003 and December 2011, 133 patients with TTP (mean age, years) had available ctni measurements on admission. Thirty-two patients (24%) had clinical and/or electrocardiogram features. Nineteen (14.3%) had cardiac symptoms, mainly congestive heart failure and myocardial infarction. Electrocardiogram changes, mainly repolarization disorders, were present in

2 294 Y. Benhamou et al 13 cases. An increased ctni level (> 0.1 lg L 1 ) was present in 78 patients (59%), of whom 46 (59%) had no clinical cardiac involvement. The main outcomes were death (25%) and refractoriness (17%). Age (P = 0.02) and ctni level (P = 0.002) showed the greatest impact on survival. A ctni level of > 0.25 lg L 1 was the only independent factor in predicting death (odds ratio [OR] 2.87; 95% confidence interval [CI] ; P = 0.024) and/or refractoriness (OR 3.03; 95% CI ; P = 0.01). Conclusions: A CTnI level of > 0.25 lg L 1 at presentation in patients with TTP appears to be an independent factor associated with a three-fold increase in the risk of death or refractoriness. Therefore, ctni level should be considered as a prognostic indicator in patients diagnosed with TTP. Keywords: ADAMTS13 protein, human; prognosis; thrombotic microangiopathy; thrombotic thrombocytopenic purpura; troponin. Introduction Thrombotic thrombocytopenic purpura (TTP) is a specific form of thrombotic microangiopathy (TMA) characterized by thrombocytopenia, microangiopathic hemolytic anemia, and widespread microvascular thrombi that are responsible for multiorgan failure of variable severity. TTP is related to severe deficiency of the von Willebrand factor-cleaving protease ADAMTS-13 [1]. ADAMTS-13 deficiency has been related to gene mutations in the hereditary form of the disease, and to anti-adamts-13 autoantibodies in the acquired form [2]. A deficiency of ADAMTS-13 results in adhesion and aggregation of platelets to ultralarge von Willebrand multimers, capillaries, and arterioles, resulting in thrombosis and associated microvascular symptoms. Despite recent advances in the management of these patients with the early use of rituximab in association with therapeutic plasma exchange (TPE) [3,4], TTP is still an acute life-threatening disease, with an overall mortality rate of ~ 20% [5 8]. Hence, attempts have been made to determine early predictive factors for death or refractoriness in the acute phase of the disease. We have previously established a predictive risk model of mortality, including age, lactate dehydrogenase (LDH) level, and neurologic involvement, in an attempt to identify patients with a worse prognosis in whom early intensive treatment might be required [9]. In this study, the impact of cardiac involvement on TTP outcome was not assessed. In recent years, an increasing number of reports have suggested that heart involvement is a frequent feature in TMA and, more particularly, in TTP [10], where, typically, widespread myocardial microthrombosis has been observed at autopsy examination [11 13]. Studies on a few limited clinical cohorts have reported that cardiac involvement occurs in 10 40% of cases of TTP, with many unrecognized cardiac symptoms leading to sudden death [14 17]. Typical cardiac features include infarction, congestive failure, arrhythmias, and cardiogenic shock [14 17]. However, on admission, cardiac symptoms are frequently lacking, or may not be specific, such as chest or substernal pain and dyspnea, which may lead to a delay in diagnosis [14]. Nevertheless, diagnosis of cardiac involvement during TTP is mandatory, because of the higher mortality rate in these patients than in those without cardiac injury [18 20]. To improve the diagnosis of subclinical cardiac injury, measurement of troponin-i or troponin-t as a very sensitive biochemical marker of myocardial damage might be relevant [21]. Hughes et al. [22] have reported, in a selected population of 41 patients with measurement of troponin, that mortality is associated with a higher level of troponin-t on admission. However, the predictive value of troponin still remains unclear in non-selected patients with microvascular damage induced by TTP. Here, we wished to assess the prognostic value of cardiac troponin-i (ctni) in TTP. For this, we conducted a study on patients enrolled prospectively in the French Reference Center for Thrombotic Microangiopathies registry. Our results provide evidence that ctni on diagnosis is a reliable prognostic marker associated with death and treatment refractoriness. Materials and methods Patients and treatment We reviewed admission data of all patients in the French Reference Center for Thrombotic Microangiopathies registry enrolled between October 2000 and December 2011 ( Patients with a diagnosis of acquired idiopathic TTP and available frozen serum on diagnosis for ctni measurement were considered for the present study. TTP diagnostic criteria were based on previous studies [4,8], and included: (i) the presence of Coombsnegative microangiopathic hemolytic anemia or microangiopathic hemolysis; (ii) acute peripheral thrombocytopenia (< L 1 ) with the absence of any other identifiable cause of thrombocytopenia and microangiopathic hemolytic anemia, such as severe disseminated intravascular coagulopathy or malignant hypertension; and (iii) severe acquired ADAMTS-13 deficiency (< 10% of activity). In order to work on a homogeneous population of patients, we did not include patients with idiopathic TMA associated with detectable ADAMTS-13 activity > 20% as these forms correspond to a specific subset of patients with a distinct presentation and outcome [8].

3 Prognostic value of troponin-i in TTP 295 Treatment of TTP was performed according to a written protocol based on a previous study and in accordance with international guidelines [23 25]. This protocol was consensually acknowledged nationwide, and has been implemented by all participating centers since October A complete response was defined as full resolution of any neurologic manifestations and platelet count recovery (> L 1 ) for at least 2 days. Durable remission was complete response with no further thrombocytopenia or clinical worsening for > 30 consecutive days from the first day of platelet count recovery (this period included the time on maintenance plasma exchange). Refractoriness was defined as the absence of platelet count doubling after four full days of standard intensive treatment with persistently elevated LDH levels. Relapse was the reappearance of neurologic manifestations and/or thrombocytopenia (< L 1 for at least 2 days) with no other identifiable cause after durable remission had been achieved. Exacerbation was defined as worsening neurologic manifestations and/or recurrent thrombocytopenia (< L 1 for at least 2 days) and/or worsening thrombocytopenia (i.e. platelet count decrease of more than one-third of the highest count, for at least 2 days) with no other identifiable cause, before durable remission had been achieved. In the case of disease exacerbation or relapse, daily TPE was resumed, in association with rituximab as a second-line treatment [3,4]. Patients without features of active infection received steroids (1 mg kg 1 per day for 3 weeks). All patients received folic acid orally or intravenously. Because of a potential bleeding risk, patients who were treated with antiplatelet agents before TTP diagnosis generally had their treatment stopped during the acute phase of the disease; it was resumed when platelet count reached > 50 g L 1. This study was approved by our institutional review board in accordance with the Declaration of Helsinki, and the French Data Protection Authority ( Commission Nationale Informatique et Libertes, CNIL, authorization no , and Comite consultatif sur le traitement de l information en matiere de recherche dans le domaine de la sante, CCTIRS, authorization no , Paris, France). Data collection Each patient s epidemiologic, clinical and biological data were collected by a research study nurse, as previously reported [8], using a standardized form, and transferred to a computerized database. The study nurse visited each center regularly to update clinical and biological data. Collected data included any past history of cardiovascular risk factors (including arterial hypertension and diabetes) and a pre-existing history of ischemic heart disease and ischemic stroke. Collected cardiac symptoms included dyspnea, chest pain, syncope, palpitation, and any changes in electrocardiograms (ECGs). Cerebral involvement included focal deficiency, seizure, headache, vigilance disturbances (inadequate response), and coma (absence of response). Renal function was assessed according to the estimated glomerular filtration rate (egfr). Evaluation of ADAMTS-13 activity and anti-adamts-13 antibodies Blood collection, plasma preparation and ADAMTS-13 activity measurement were performed as previously described [26]. Severe ADAMTS-13 deficiency was defined as no detectable ADAMTS-13 activity (< 10%), moderate ADAMTS-13 deficiency as activity between 10% and 49%, and normal ADAMTS-13 as activity 50%. Anti-ADAMTS-13 antibodies were detected by ELISA (Technozym R ADAMTS13 INH; Technoclone, Vienna, Austria), as recommended by the manufacturer. ADAMTS-13 deficiency was considered to be acquired if it was associated with anti-adamts-13 antibodies. The threshold for antibody positivity was 12 U ml 1. Patients with no detectable anti-adamts-13 antibodies on diagnosis but who partially or completely recovered ADAMTS-13 activity at the time of clinical remission were also considered to have acquired TTP. ctni measurement Blood samples were collected from all patients within 24 h of admission, pre-tpe. As recommended, samples were stored at 80 C and frozen quickly, within < 6h of the blood test [27]. In the present study, the assessment of ctni was centralized in one center (Saint-Antoine Hospital), and ctni was measured in blood serum samples with the same immunoassay on the Access II immunoassay system (Beckman-Coulter, Fullerton, CA, USA) with the AccuTnI (ref. A78803) assay and AccuTnI (ref ) calibrators. Both external and internal quality controls verified the accuracy of the results. The 99th percentile cut-off value of ctni (from a healthy reference population) was set at 0.04 lg L 1. Results for ctni were interpreted as follows: lg L 1, unlikely to have acute coronary syndrome (ACS); lg L 1, possible ACS; and > 0.1 lg L 1, probable ACS [28]. Statistical methods Quantitative variables were summarized as mean standard deviation and compared by use of Wilcoxon s ranksum test. Categorical data were summarized as count (%), and compared by use of the chi-squared test or Fisher s exact test, as appropriate. Risk factors for early mortality were investigated by logistic regression. First, a univariable analysis was carried out. A cut-off value of troponin was obtained by maximizing the Youden score

4 296 Y. Benhamou et al (sensitivity + specificity 1), and the sensitivity, specificity, positive and negative predictive values and their respective 95% confidence intervals (CIs) were determined. The area under the receiver operating characteristic (ROC) curve was used to evaluate the discriminative value of troponin in predicting survival. Variables strongly associated with mortality at P < 0.05 were then included in a backward multivariable analysis. According to the size of our cohort, only four variables were tested, to decrease the possibility of misinterpretation. To investigate a possible relationship between an increasing ctni level and outcome (death and/or refractoriness), we divided the ctni levels into three classes, as follows: <0.1 lg L 1 (a recognized cut-off defining an elevated troponin level); [ lg L 1 ] and >0.5 lg L 1, to have a sufficient number of patients in each class. Variables were compared by use of the chi-squared trend test. Data were analyzed with R software version 2.9. A P- value of < 0.05 was considered to be statistically significant for all statistical tests performed. Results Study group The study flow chart is detailed in Fig. 1. From October 2000 to December 2011, among 2912 patients with TMA included in the national registry, 380 had acquired idiopathic TTP and were treated according to consensual guidelines. Of these patients, 47 (12.4%) died during the acute phase of the disease, leaving 333 survivors. Finally, all patients considered for the present study were enrolled in our registry from 2003, as we started to freeze serum systematically from this date. One hundred and forty-two patients (109 survivors and 33 non-survivors) with available frozen serum on diagnosis were considered for the current study and for ctni measurement. In nine survivors, the ctni result was uninterpretable, probably as a result of ctni degradation during the freezing thawing process. Finally, this study focused on the remaining 100 survivors and 33 non-survivors with available ctni measurements. The other 240 patients had no available frozen serum (in most cases, serum was not sampled on diagnosis, or the volume of available serum was not sufficient for ctni assessment, or the serum was used for previous studies). Patients characteristics on diagnosis and outcome The mean age was years, and the male/female ratio 1 : 2.6. Seven patients (5.2%) had a past history of stroke, and 10 had a history of ischemic heart disease. Thirteen patients had been receiving long-term antiplatelet therapy that was disrupted during the acute phase of the disease, when thrombocytopenia was identified. On admission, 91 patients (68%) had cerebral involvement and 19 (14.3%) had clinical cardiac symptoms, including chest pain (n = 3), syncope (n = 1), congestive heart failure (n = 5), myocardial infarction (n = 8), and cardiac arrest (n = 2). ECG changes were present in 13 of 104 patients (12.5%), with essentially negative T wave (9/13), sinus tachychardia (1/13), and ST elevation (3/13). In total, 32 patients (24%) had clinical cardiac symptoms and/or ECG features. Seventy-eight patients (59%) had a ctni level of > 0.1 lg L 1, and the median ctni value was 0.19 lg L 1 (0 24.5). Among these latter patients, an increased ctni level without any cardiac injury (no clinical symptoms and no ECG changes) was observed in 46 cases (59%). Patients with clinical and/or ECG features of cardiac involvement had a higher ctni level than patients without cardiac involvement ( lg L 1 and ng ml 1, respectively; P = 0.003). Moreover, patients with clinical and/or ECG features of cardiac involvement more frequently had a ctni level of > 0.1 lg L 1 than patients without cardiac involvement (41% and 18%, respectively; P = ). The main outcomes were death in 33 patients (25%) and refractoriness in 19 patients (17%). The most prevalent causes of death were: cardiogenic shock (n = 6), cardiac arrest (n = 9), stroke and cerebral hemorrhage (n = 7), multiorgan failure (n = 6), and sepsis (n = 2). Consistent with our previous report, non-survivors were older (P = 0.02) and more frequently presented with a history of cardiac injury, although the difference did not reach the significance level (P = 0.06). On presentation, cerebral involvement, including coma, stupor, and seizure, tended to be more frequently observed in non-survivors (P = 0.058), whereas cardiac symptoms were not different between groups. However, the ctni level was higher (P = 0.002) and the egfr was lower (P = 0.01) in nonsurvivors, whereas the LDH level was similar in both groups. Finally, as expected, non-survivors more frequently experienced a refractory disease than survivors (50% and 12%, respectively; P = ; Table 1). There was no difference in the use of immunomodulatory drugs between patients with or without cardiac involvement (46% and 30%, respectively; P = 0.15). In particular, the use of rituximab was similar in both groups (52% and 37%, respectively; P = 0.13). Conversely, the ctni level was more frequently increased in patients given antiplatelet therapy than in those who were not (84.5% and 36.6%, respectively; P = 0.016), whereas no difference in the use of antiplatelet therapy was found between survivors and non-survivors (P = 0.50). Age and ctni level clearly showed the greatest impact on survival in our cohort. We then determined a cut-off value of troponin for predicting death, by measuring the Youden score. A 0.25 lg L 1 cut-off was associated with a sensitivity, a specificity, a positive predictive value (mortality among those with a ctni level of > 0.25 lg L 1 ) and a negative predictive value (survival among patients with a ctni level of < 0.25 lg L 1 ) of

5 Prognostic value of troponin-i in TTP 297 TMA registered in the registry (October 2000 to December 2011) 2912 patients TMA with associated conditions and/or TMA with detectable ADAMTS patients Idiopathic acquired TTP 380 patients Survivors 333 patients Non-survivors 47 patients Patients with available serum 109 Patients with available serum 33 Uninterpretable results: 9 patients 133 patients available for study - Survivors: 100 patients - Non-survivors: 33 patients Fig. 1. Flow chart of the study. TMA, thrombotic microangiopathy; TTP, thrombotic thrombocytopenic purpura. 64%, 66%, 38%, and 85%, respectively. The area under the ROC curve was then used to evaluate the discriminatory power of our value. The area under the curve was 0.68, showing that this threshold had discriminatory power (P = 0.001). Patients characteristics on diagnosis according to ctni level Patients with a serum ctni level of > 0.25 lg L 1 on presentation were older (P = 0.01) and more frequently presented with a past history of ischemic heart disease (P = 0.01) and ischemic stroke (P = 0.02). Cerebral and cardiac features, including chest pain, myocardial infarction, cardiac arrest, and ECG changes, were also more frequently observed in this group (P = 0.04 and P = 0.001, respectively). The serum creatinine level was higher (P < ) and the egfr was lower (P < ) in patients with an increased serum ctni level. Then LDH level was higher in this group. Inhibitory anti-adamts-13 antibody and anti-adamts-13 antibody levels were similar in both groups. Patients who experienced a fatal outcome and/or a refractory disease more frequently had a ctni level of > 0.25 lg L 1 (P = and P = 0.008, respectively) (Table 2). ctni as a predictive factor for unfavorable outcome On the basis of univariate analysis, and given an existing interaction between renal function, age, and ctni values, we chose to test age, cerebral involvement, ctni level and egfr by multivariable analysis. We found that a ctni level of > 0.25 lg L 1 was the only independent variable associated with death, with an odds ratio (OR) of 2.87 (95% CI ; P = 0.024) (Table 3). As refractoriness was associated with death and ctni level, we considered a composite criterion including death and/or refractory TTP. The sensitivity, specificity, positive predictive value and negative predictive value of a ctni level of > 0.25 lg L 1 for death and/or refractoriness were 64%, 69%, 49%, and 81%, respectively. In the same statistical analysis procedure, we performed a multivariable analysis, with age, cerebral involvement, ctni level and egfr as tested variables. We found that ctni level remained the only independent factor associated with death and treatment refractoriness (OR 3.03; 95% CI ; P = 0.01). In addition, patients who experienced death and/or refractoriness were divided into three classes of ctni level. Nine patients (16%) were included in the first class

6 298 Y. Benhamou et al Table 1 Clinical characteristics of patients according to outcome Parameters Non-survivors, %(N = 33) Survivors, %(N = 100) Overall population, %(N = 133) P Age (years) Female 70 (23) 66 (66) 67 (89) 0.7 Cardiovascular risk factors and 52 (17) 55 (55) 54 (72) 0.73 pre-existing comorbidities Arterial hypertension 24 (8) 20 (20) 21 (28) 0.6 Diabetes 9 (3) 8 (8) 8 (11) 0.84 Past history of ischemic stroke 3 (1) 6 (6) 5 (7) 0.51 Past history of ischemic heart disease 9 (3) 7 (7) 8 (10) 0.69 Clinical cardiac symptoms 12 (4) 15 (15) 14 (19) 0.85 Electrocardiogram findings (N = 104) 7.7 (2) 14.1 (11) 12.5 (13) 0.51 Neurologic involvement 82 (27) 64 (64) 68 (91) Headache 18 (6) 27 (27) 25 (33) 0.31 Confusion 27 (9) 12 (12) 16 (21) 0.04 Seizure 24 (8) 8 (8) 12 (16) 0.01 Coma/vigilance disturbances 30 (10) 14 (14) 18 (24) Focal deficiency 33 (11) 37 (37) 36 (48) 0.7 Hemoglobin level (g dl 1 ) Reticulocyte count (N = 90) LDH level (9 normal) (N = 124) Platelet count (10 9 L 1 )(N = 132) Serum creatinine (lmol L 1 ) Estimated glomerular filtration rate (ml min 1 ) Serum cardiac troponin-i (lg ml 1 ) Flare-up 100 (3) 63 (63) 64 (66) 0.19 Refractory TTP (N = 112) 50 (8) 12 (11) 17 (19) ADAMTS-13 inhibitor (N = 104) 96 (24) 92 (73) 93 (97) 0.86 Anti-ADAMTS-13 antibodies (N = 124) Antiplatelet therapy 12 (4) 9 (9) 9.7 (13) 0.50 LDH, lactate dehydrogenase; TTP, thrombotic thrombocytopenic purpura. Quantitative values are expressed as mean standard deviation. Qualitative values are expressed as percentage (total number). In cases of missing values, the number of patients tested is specified in parentheses in the left column (N). P < 0.05 was considered to be statistically significant. Significant values appear in bold. (0 0.1 lg L 1 ), 13 (35%) in the second class ( lg L 1 ), and 22 (54%) in the third class ( lg L 1 ). There was a significant increase in death and/or refractoriness with the increase in ctni level (P < ). Discussion We report here that serum ctni level on admission is a reliable marker of cardiac involvement in patients with acquired TTP [20,22,29], and allows the identification of patients with a more aggressive presentation [14,18]. Patients with an increased ctni level of > 0.25 lg L 1 show more cerebral and cardiac involvement at presentation, and a more severe outcome with a higher death rate. An additional original and unexpected finding of our study is that ctni level strongly predicts refractoriness, with an overall three-fold increase in the risk of death or refractoriness. We confirm from this large prospective cohort of patients that cardiac symptoms and/or ECG changes are frequent in patients with acquired TTP on diagnosis, with a prevalence of 10 40% [14 17] (24% in the present study). Importantly, however, up to 60% of the studied patients had an increased ctni level consistent with the diagnosis of ACS, with 75% of them being clinically asymptomatic. These results are in line with previous retrospective studies, and support the view that clinical examination and ECGs clearly underestimate the prevalence of cardiac involvement in TTP [10,22]. Therefore, ctni represents a sensitive marker that allows the identification of subclinical cardiac involvement, and should be systematically assessed in these patients on diagnosis. Notably, we found that clinical features of cardiac involvement were not significantly associated with death or refractoriness. On the contrary, among patients with no clinical features of cardiac involvement, those with only an increased ctni level experienced more death and refractoriness (data not shown), providing further evidence that ctni is a more reliable prognostic factor than clinical features. Notably, we found that ctni also represents a more reliable marker than LDH for predicting death. Indeed, LDH probably reflects severe multiorgan damage, but only very high levels (> 10 times higher than normal) may predict death [9]. A serum troponin level of > 0.20 ng ml 1 at presentation of TMA was previously reported to be an independent risk factor for early myocardial infarction, typically before

7 Prognostic value of troponin-i in TTP 299 Table 2 Clinical characteristics of patients according to cardiac troponin-i level Parameters Troponin-I < 0.25 lg L 1, %(N = 78) Troponin-I > 0.25 lg L 1, %(N = 55) Age (years) Cardiovascular risk 49 (38) 62 (34) 0.14 factor and pre-existing comorbidities Ischemic stroke 1 (1) 11 (6) 0.01 Ischemic heart disease 3 (2) 15 (8) 0.01 Clinical cardiac 8 (6) 24 (13) symptoms Electrocardiogram 8.2 (5) 18.6 (8) 0.82 findings (N = 104) Neurologic 62 (48) 78 (43) 0.04 involvement Headache 28 (22) 20 (11) 0.28 Confusion 10 (8) 24 (13) Seizure 8 (6) 18 (10) 0.07 Coma/impaired 8 (6) 33 (18) < consciousness Focal deficiency 32 (25) 42 (23) 0.25 Hemoglobin level (g dl 1 ) Reticulocyte count (n = 90) LDH level < (9 normal) (n = 124) Platelet count (10 9 L 1 )(n = 132) Serum creatinine < (lmol L 1 ) Estimated glomerular < filtration rate (ml min 1 ) Flare-up 65 (43) 62 (23) 0.76 Refractory 9 (6) 30 (13) TTP (n = 112) Death 15 (12) 38 (21) ADAMTS-13 inhibitor 93 (57) 93 (40) 0.93 (n = 104) Antiplatelet therapy 2.5 (2) 20.7 (11) LDH, lactate dehydrogenase; TTP, thrombotic thrombocytopenic purpura. Quantitative values are expressed as mean standard deviation. Qualitative values are expressed as percentage (total number). In cases of missing values, the number of patients tested is specified in parentheses in the left column (N). P < 0.05 was considered to be statistically significant. Significant values appear in bold. the occurrence of overt symptoms [10]. Similarly, increased levels of troponin (> 0.05 lg L 1 ) were associated with myocardial necrosis [22], and troponin-t levels of > 0.1 lg L 1 with increased mortality and acute morbidity [22]. Here, we determined, from a large number of patients explored homogeneously, that a threshold value of 0.25 lg L 1 is highly predictive of death and refractoriness. Moreover, the negative predictive value is high (81%); therefore, when the troponin level is < 0.25 lg L 1, P Table 3 Association between patients characteristics and death by multivariable analysis Odds ratio (95% CI) Troponin-I > 0.25 lg L ( ) Age (years) ( ) > ( ) Neurologic involvement 1.66 ( ) 0.4 egfr 0.61 ( ) 0.32 egfr, estimated glomerular filtration rate (ml min 1); CI, confidence interval. P < 0.05 was considered to be statistically significant. Significant values appear in bold. physicians will be misled in only 19% of cases, whereas the mortality rate increases from 15% to 38%. CTnI is a very sensitive and specific marker of myocardial damage; nevertheless, high levels of troponin were also found in patients with stroke and were associated with a poorer prognosis, reflecting the systemic severity of acute ischemic stroke [30,31]. This observation is in agreement with our results, showing that patients with cerebral involvement also more frequently had cardiac involvement. The identification of ctni as a predictive marker of death and refractoriness may have consequences for the management of patients; in particular, it raises the question of whether traditional antiplatelet agents should be added to the standard treatment in these patients. However, as microthrombi in patients with TTP mainly involve the glycoprotein Ib von Willebrand factor axis, it is likely that the more suitable antiplatelet treatment for patients with cardiac microthrombi is ADAMTS-13, which could be provided in large amounts through intensive TPE (i.e. twice-daily TPE [32]). Alternatively, compounds aimed at specifically inhibiting the glycoprotein Ib von Willebrand factor interaction may also be of interest [33,34]. The strengths of our study include the homogeneous management (that is, we included patients within a limited period of time, investigated them in the same manner, and recorded their data prospectively) and the large number of centers and medical specialties that participated in patient recruitment. In particular, the determination of ctni level and the determination of ADAMTS-13 activity were centralized, and blood samples were systematically collected before TPE was started, which is mandatory, because this procedure may significantly impact on the measurement of these parameters [29,35]. However, our study also has limitations. In particular, we enrolled a large number of non-survivors, which may account for the high prevalence of ctni increase in our patients (i.e. up to 60% of patients), which is slightly higher than in previous reports [14,20]. Additionally, as troponin is mostly cleared by the kidney, interpretation of troponin requires caution in patients with acute renal P

8 300 Y. Benhamou et al failure. However, the inclusion of the egfr in our multivariable analysis provided strong evidence for troponin as an indicator of death regardless of renal function. Finally, when incorporated with our previous validated score [9], ctni failed to improve the performance of this (data not shown), providing further evidence that ctni gives information that is independent from that given by our previous score. Another limitation of this study is that ctni measurements were available for only one-third of patients, because serum was not systematically frozen for all patients (our group progressively started to freeze serum from patients with TTP from 2003); moreover, the serum of some of the patients may have been used for previous studies. However, our population of patients is large, and, importantly, patients were selected independently of initial presentation; therefore, our population is highly representative of patients with acquired TTP with regard to their different clinical aspects, providing reliable results. In conclusion, we report here that serum ctni is a reliable marker with which to assess cardiac involvement in patients with acquired TTP on diagnosis, and to identify, early in the course of the disease, patients with a higher risk for death and refractoriness. On the basis of our experience, we suggest that ctni should be systematically assessed in all patients with a diagnosis of TTP. A ctni value of > 0.25 lg L 1 should prompt consideration of close monitoring in intensive care units and more intensive treatments that will require evaluation in prospective trials. Addendum Y. Benhamou and P.-Y. Boelle performed the statistical analysis of the French Registry for Thrombotic Microangiopathies. Y. Benhamou, P.-Y. Boelle, J. Gras, P. Coppo, and L. Galicier interpreted the results, wrote the manuscript, designed the study, and critically reviewed the manuscript. B. Baudin performed cardiac troponin-i measurement, interpreted the results, and critically reviewed the manuscript. A. Veyradier performed AD- AMTS-13 analysis, interpreted the results, and critically reviewed the manuscript. Y. Benhamou, S. Ederhy, J. Gras, L. Galicier, E. Azoulay, F. Provoˆt, E. Maury, F. Pene, J.-P. Mira, A. Wynckel, C. Presne, P. Poullin, P. Perez, J.-M. Halimi, Y. Delmas, T. Kanouni, A. Seguin, C. Mousson, A. Servais, D. Bordessoule, A. Cohen, M. Hamidou, and P. Coppo enrolled patients, collected clinical and laboratory information, and critically reviewed the manuscript. Acknowledgements Patients were recruited with the help of the members of the Reference Center for Thrombotic Microangiopathies (CNR-MAT) (listed in Appendix). We thank S. Thouzeau, S. Savigny, S. Capdenat (Laboratoire d Hematologie, Hoˆpital Antoine Beclere, AP-HP, Clamart) and S. Malot (Centre de Reference des Microangiopathies Thrombotiques, Hoˆpital Saint-Antoine, AP-HP, Paris) for technical assistance. Disclosure of Conflict of Interests Y. Delmas reports receiving personal fees and non-financial support from Alexion; and non-financial support from Amgen and Hemotech, outside the submitted work. Y. Benhamou reports receiving grants from Bayer Healthcare; and non-financial support from Baxter and Leo Pharma, outside the submitted work. E. Azoulay reports receiving grants from MSD and Pfizer; and grants and personal fees from Gilead, outside the submitted work. F. Pene reports receiving grants from INSERM and the Societe de Reanimation de Langue Francßaise; and personal fees from MSD and LFB, outside the submitted work. J.-P. Mira reports receiving personal fees from LFB, MSD, Astellas, Gilead, and Novartis, outside the submitted work. L. Galicier reports receiving non-financial support from GlaxoSmithKline, Roche, and Amgen; and personal fees for development of educational presentation from GlaxoSmithKline, Roche, Chugai, Raison et Sante, and Janssen, outside the submitted work. A. Cohen reports receiving grants from RESICARD; and personal fees from AstraZeneca, Bayer Pharma, Boehringer Ingelheim, Daiichi Sankyo, GlaxoSmithKline, and Sanofi, outside the submitted work. P. Coppo is a member of the Clinical Advisory Board for Alexion. The other authors state that they have no conflict of interest. Appendix The members of the Reference Center for Thrombotic Microangiopathies (CNR-MAT) are: E. Azoulay (Service de Reanimation Medicale, Hoˆpital Saint-Louis, Paris); V. Barbay (Laboratoire d Hematologie, CHU Charles Nicolle, Rouen); G. Bonmarchand (Service de Reanimation, CHU Charles Nicolle, Rouen); D. Bordessoule (Service d Hematologie, Hoˆpital Dupuytren, Limoges); C. Charasse (Service de Nephrologie, Centre Hospitalier de Saint-Brieuc); D. Chauveau (Service de Nephrologie et Immunologie Clinique, CHU Rangueil, Toulouse); G. Choukroun (Service de Nephrologie, Hoˆpital Sud, Amiens); J.-P. Coindre (Service de Nephrologie, CH Le Mans); P. Coppo (Service d Hematologie, Hoˆpital Saint- Antoine, Paris); E. Corre (Service d Hematologie, Hoˆpital Saint-Antoine, Paris); Y. Delmas (Service de Nephrologie, Hoˆpital Pellegrin, Bordeaux); G. Deschenes (Service de Nephrologie Pediatrique, Hoˆpital Robert Debre, Paris); A. Devidas (Service d Hematologie, Hoˆpital Sud-Francilien, Corbeil-Essonnes); O. Fain (Service de Medecine Interne, Hoˆpital Jean Verdier, Bondy); V. Fremeaux-Bacchi (Laboratoire d Immunologie, Hoˆpital Europeen

9 Prognostic value of troponin-i in TTP 301 Georges Pompidou, Paris); L. Galicier (Service d Immunopathologie, Hoˆpital Saint-Louis, Paris); B. Guidet (Service de Reanimation Medicale, Hoˆpital Saint-Antoine, Paris); J.-M. Halimi (Service de Nephrologie Pediatrique, Hoˆpital Bretonneau, Tours); M. Hamidou (Service de Medecine Interne, Hoˆtel-Dieu, Nantes); R. Herbrecht (Service d Oncologie et d Hematologie, Hoˆpital de Hautepierre, Strasbourg); F. Jacobs (Service de Reanimation Medicale, Hoˆpital Antoine Beclere, Clamart); B. Joly (Service d Hematologie Biologique, Hoˆpital Antoine Beclere, Clamart); T. Kanouni (Unite d Hemaphrese, Service d Hematologie, CHU de Montpellier); A. Lautrette (Service de Nephrologie Pediatrique B, Hoˆpital Hoˆtel- Dieu, Clermont-Ferrand); V. Le Guern (Unite d Hemapherese, Service de Medecine Interne, Hoˆpital Cochin, Paris); C. Loirat (Service de Nephrologie Pediatrique, Hoˆpital Robert Debre, Paris); J.-P. Mira (Service de Reanimation Medicale, Hoˆpital Cochin); B. Moulin (Service de Nephrologie, Hoˆpital Civil, Strasbourg); C. Mousson (Service de Nephrologie, CHU de Dijon); M. Ojeda Uribe (Service d Hematologie, Hoˆpital Emile Muller, Mulhouse); A. Ouchenir (Service de Reanimation, Hoˆpital Louis Pasteur, Le Coudray); N. Parquet (Unite de Clinique Transfusionnelle, Hoˆpital Cochin, Paris); J. Peltier (Urgences Nephrologiques et Transplantation Renale, Hoˆpital Tenon, Paris); P. Perez (Service de Reanimation polyvalente, CHU de Nancy); P. Poullin (Service d hemapherese et d autotransfusion, Hoˆpital la Conception, Marseille); C. Pouteil-Noble (Service de Nephrologie, CHU Lyon-Sud, Lyon); C. Presne (Service de Nephrologie, Hoˆpital Nord, Amiens); F. Provoˆt (Service de Nephrologie, Hoˆpital Albert Calmette, Lille); J.- A. Ribeil (Service de Therapie Cellulaire, Hoˆpital Necker- Enfants Malades, Paris); E. Rondeau (Urgences Nephrologiques et Transplantation Renale, Hoˆpital Tenon, Paris); S. Saheb (Unite d Hemapherese, Hoˆpital la Pitie- Salpetriere, Paris); B. Schlemmer (Service de Reanimation Medicale, Hoˆpital Saint-Louis, Paris); A. Seguin (Service de Reanimation Medicale, CHU de Caen); A. Stepanian (Laboratoire d Hematologie, Hoˆpital Louis Mourier, Colombes); J.-P. Vernant (Service d Hematologie, Hoˆpital la Pitie-Salpetriere, Paris); A. Veyradier (Service d Hematologie Biologique, Hoˆpital Antoine Beclere, Clamart); C. Vigneau (Service de Nephrologie, Hoˆpital Pontchaillou, Rennes); F. Vrtovsnick (Service de Nephrologie, Hoˆpital Bichat, Paris); A. Wynckel (Service de Nephrologie, Hoˆpital Maison Blanche, Reims); M. Wolf (Service d Hematologie Biologique, Hoˆpital Antoine Beclere, Clamart); and P. Zunic (Service d Hematologie, Groupe Hospitalier Sud-Reunion, la Reunion). References 1 Coppo P, Veyradier A. Thrombotic microangiopathies: towards a pathophysiology-based classification. Cardiovasc Hematol Disord Drug Targets 2009; 9: Tsai HM. Autoimmune thrombotic microangiopathy: advances in pathogenesis, diagnosis, and management. Semin Thromb Hemost 2012; 38: Scully M, McDonald V, Cavenagh J, Hunt BJ, Longair I, Cohen H, Machin SJ. A phase 2 study of the safety and efficacy of rituximab with plasma exchange in acute acquired thrombotic thrombocytopenic purpura. Blood 2011; 118: Froissart A, Buffet M, Veyradier A, Poullin P, Provoˆt F, Malot S, Schwarzinger M, Galicier L, Vanhille P, Vernant JP, Bordessoule D, Guidet B, Azoulay E, Mariotte E, Rondeau E, Mira JP, Wynckel A, Clabault K, Choukroun G, Presne C, et al. Efficacy and safety of first-line rituximab in severe, acquired thrombotic thrombocytopenic purpura with a suboptimal response to plasma exchange. Experience of the French Thrombotic Microangiopathies Reference Center. Crit Care Med 2012; 40: Rock GA, Shumak KH, Buskard NA, Blanchette VS, Kelton JG, Nair RC, Spasoff RA. Comparison of plasma exchange with plasma infusion in the treatment of thrombotic thrombocytopenic purpura. N Engl J Med 1991; 325: George JN. Thrombotic thrombocytopenic purpura. N Engl J Med 2006; 354: Bell WR, Braine HG, Ness PM, Kickler TS. Improved survival in thrombotic thrombocytopenic purpura hemolytic uremic syndrome. N Engl J Med 1991; 325: Coppo P, Schwarzinger M, Buffet M, Wynckel A, Clabault K, Presne C, Poullin P, Malot S, Vanhille P, Azoulay E, Galicier L, Lemiale V, Mira JP, Ridel C, Rondeau E, Pourrat J, Girault S, Bordessoule D, Saheb S, Ramakers M, et al. Predictive features of severe acquired ADAMTS13 deficiency in idiopathic thrombotic microangiopathies: the French TMA reference center experience. PLoS ONE 2010; 5: e Benhamou Y, Assie C, Boelle PY, Buffet M, Grillberger R, Malot S, Wynckel A, Presne C, Choukroun G, Poullin P, Provoˆt F, Gruson D, Hamidou M, Bordessoule D, Pourrat J, Mira JP, Le Guern V, Pouteil-Noble C, Daubin C, Vanhille P, et al. Development and validation of a predictive model for death in acquired severe ADAMTS13 deficiency-associated idiopathic thrombotic thrombocytopenic purpura: the French TMA reference center experience. Haematologica 2012; 97: Patschan D, Witzke O, Duhrsen U, Erbel R, Philipp T, Herget- Rosenthal S. Acute myocardial infarction in thrombotic microangiopathies clinical characteristics, risk factors and outcome. Nephrol Dial Transplant 2006; 21: Hosler GA, Cusumano AM, Hutchins GM. Thrombotic thrombocytopenic purpura and hemolytic uremic syndrome are distinct pathologic entities. A review of 56 autopsy cases. Arch Pathol Lab Med 2003; 127: Ridolfi RL, Bell WR. Thrombotic thrombocytopenic purpura. Report of 25 cases and review of the literature. Medicine (Baltimore) 1981; 60: Ridolfi RL, Hutchins GM, Bell WR. The heart and cardiac conduction system in thrombotic thrombocytopenic purpura. A clinicopathologic study of 17 autopsied patients. Ann Intern Med 1979; 91: Hawkins BM, Abu-Fadel M, Vesely SK, George JN. Clinical cardiac involvement in thrombotic thrombocytopenic purpura: a systematic review. Transfusion 2008; 48: Siersema PD, Kros JM, van den Berg B. Cardiac manifestations of thrombotic thrombocytopenic purpura. Neth J Med 1989; 35: Podolsky SH, Zembowicz A, Schoen FJ, Benjamin RJ, Sonna LA. Massive myocardial necrosis in thrombotic thrombocytopenic purpura: a case report and review of the literature. Arch Pathol Lab Med 1999; 123: James TN, Monto RW. Pathology of the cardiac conduction system in thrombotic thrombocytopenic purpura. Ann Intern Med 1966; 65:

10 302 Y. Benhamou et al 18 Gami AS, Hayman SR, Grande JP, Garovic VD. Incidence and prognosis of acute heart failure in the thrombotic microangiopathies. Am J Med 2005; 118: Gandhi K, Aronow WS, Desai H, Amin H, Sharma M, Lai HM, Singh P. Cardiovascular manifestations in patients with thrombotic thrombocytopenic purpura: a single-center experience. Clin Cardiol 2010; 33: Wahla AS, Ruiz J, Noureddine N, Upadhya B, Sane DC, Owen J. Myocardial infarction in thrombotic thrombocytopenic purpura: a single-center experience and literature review. Eur J Haematol 2008; 81: Peacock WF 4th, De Marco T, Fonarow GC, Diercks D, Wynne J, Apple FS, Wu AH; ADHERE Investigators. Cardiac troponin and outcome in acute heart failure. N Engl J Med 2008; 358: Hughes C, McEwan JR, Longair I, Hughes S, Cohen H, Machin S, Scully M. Cardiac involvement in acute thrombotic thrombocytopenic purpura: association with troponin T and IgG antibodies to ADAMTS 13. J Thromb Haemost 2009; 7: Kremer HJ, Vesely SK, Terrell DR, Lammle B, George JN. Survival and relapse in patients with thrombotic thrombocytopenic purpura. Blood 2010; 115: ; quiz Coppo P, Veyradier A. Current management and therapeutical perspectives in thrombotic thrombocytopenic purpura. Presse Med 2012; 41: e Scully M, Hunt BJ, Benjamin S, Liesner R, Rose P, Peyvandi F, Cheung B, Machin SJ; British Committee for Standards in Haematology. Guidelines on the diagnosis and management of thrombotic thrombocytopenic purpura and other thrombotic microangiopathies. Br J Haematol 2012; 158: Ferrari S, Scheiflinger F, Rieger M, Mudde G, Wolf M, Coppo P, Girma JP, Azoulay E, Brun-Buisson C, Fakhouri F, Mira JP, Oksenhendler E, Poullin P, Rondeau E, Schleinitz N, Schlemmer B, Teboul JL, Vanhille P, Vernant JP, Meyer D, et al. Prognostic value of anti-adamts 13 antibody features (Ig isotype, titer, and inhibitory effect) in a cohort of 35 adult French patients undergoing a first episode of thrombotic microangiopathy with undetectable ADAMTS 13 activity. Blood 2007; 109: Zaman Z, de Spiegeleer S, Gerits M, Blanckaert N. Analytical and clinical performance of two cardiac troponin I immunoassays. Clin Chem Lab Med 1999; 37: Collinson PO, Stubbs PJ, Kessler AC. Multicentre evaluation of the diagnostic value of cardiac troponin T, CK-MB mass, and myoglobin for assessing patients with suspected acute coronary syndromes in routine clinical practice. Heart 2003; 89: McCarthy LJ, Danielson CF, Skipworth EM, Peters SL, Miraglia CC, Antony AC. Myocardial infarction/injury is relatively common at presentation of acute thrombotic thrombocytopenic purpura: the Indiana University experience. Ther Apher 2002; 6: Di Angelantonio E, Fiorelli M, Toni D, Sacchetti ML, Lorenzano S, Falcou A, Ciarla MV, Suppa M, Bonanni L, Bertazzoni G, Aguglia F, Argentino C. Prognostic significance of admission levels of troponin I in patients with acute ischaemic stroke. J Neurol Neurosurg Psychiatry 2005; 76: Anders B, Alonso A, Artemis D, Sch afer A, Ebert A, Kablau M, Fluechter S, Findeisen P, Hennerici MG, Fatar M. What does elevated high-sensitive troponin I in stroke patients mean: concomitant acute myocardial infarction or a marker for highrisk patients? Cerebrovasc Dis 2013; 36: Nguyen L, Li X, Duvall D, Terrell DR, Vesely SK, George JN. Twice-daily plasma exchange for patients with refractory thrombotic thrombocytopenic purpura: the experience of the Oklahoma Registry, 1989 through Transfusion 2008; 48: Jilma-Stohlawetz P, Gorczyca ME, Jilma B, Siller-Matula J, Gilbert JC, Kn obl P. Inhibition of von Willebrand factor by ARC1779 in patients with acute thrombotic thrombocytopenic purpura. Thromb Haemost 2011; 105: Holz JB. The TITAN trial assessing the efficacy and safety of an anti-von Willebrand factor nanobody in patients with acquired thrombotic thrombocytopenic purpura. Transfus Apher Sci 2012; 46: Tutarel O, Golla P, Beutel G, Bauersachs J, David S, Schmidt BM, Lichtinghagen R, Kielstein JT. Therapeutic plasma exchange decreases levels of routinely used cardiac and inflammatory biomarkers. PLoS ONE 2012; 7: e38573.