Risk factors associated with intracranial hemorrhage in adults with immune thrombocytopenia: A study of 27 cases.

Abstract

Translate Article

To the Editor Patients with immune thrombocytopenia (ITP) are at risk for life-threatening bleeding complications, particularly intracranial hemorrhage (ICH). ICH occurs in <1% of patients with ITP 1. No study has assessed the risk factors of ICH in adults with ITP. This case-control study aimed to determine ICH risk factors in adults with primary ITP. A French multicenter case-control retrospective study was conducted over 22 years (1998–2012) using the national hospital discharge database (PMSI), and included patients aged 18 years and older with primary ITP and ICH, with platelet count ≤100 × 109/L at the time of ICH. Each center provided two controls per case. The controls were defined as the next two patients with primary ITP who were hospitalized after the ICH case, whatever was the reason of hospitalization provided that the platelet count was <50 × 109/L at inclusion. Patients with secondary ITP were excluded. Data collected included patients and ITP characteristics before ICH, the events preceding ICH (e.g., head trauma in the 7 days preceding ICH, infections in the month preceding ICH, medications interfering with hemostasis at the time of ICH, cerebral vascular malformations) and the description of ICH. The Committee of Protection of Persons of Ile-de-France X approved this study. Bleeding severity of ITP was classified as mild (only skin bleeding or no bleeding), intermediate (for visible mucosal and skin bleeding), or visceral (including gastrointestinal, hematuria, gynecological and retinal) and was based on the worst bleeding that occurred during ITP history before the ICH for cases and before the last visit for controls. At the time of ICH, bleeding severity was assessed by the Khellaf bleeding score 2. All categorical data were analyzed by chi-square or Fisher's exact test. All continuous data were analyzed by nonparametric Mann–Whitney test. A logistic regression model was used to determine ICH risk factors. A two-sided P < 0.05 was considered statistically significant. Ten cases (37%) presented ICH during newly diagnosed ITP; for five of them, the ICH was diagnosed concomitantly with ITP. Bleeding symptoms before ICH onset are described in Table 1: the frequency of cutaneous bleeding and hematuria differed between cases and controls; the frequency of overall visceral hemorrhage and life-threatening bleeding was greater for cases than controls. Overall, 20/27 cases (74%) received treatment for ITP before ICH as compared with all controls (P = 0.0002). For five cases, ICH was diagnosed concomitantly with ITP, so no previous treatment was administered. For two cases, the platelet count was >30 × 109/L, so no treatment was required before the ICH. Only 16% (3/19) of cases responded to steroids versus 85% (45/53) for controls. Events preceding ICH are described in Table 1: head trauma preceded ICH in five cases and in two of the controls (P = 0.038). The two groups did not differ in consumption of medications that interfered with hemostasis. A total of 10/27 cases (37%) experienced infection within 5 days (range: 0–22 days) before ICH. Infection data for controls were missing; three cases showed intracranial vascular malformations. For 19/27 cases (70%), at least one of the following factors preceded the ICH: cranial trauma, cranial vascular malformation, drugs interfering with hemostasis or infection; 12(44.4%), 5 (18.5%), and 2 (7.4%) patients had 1, 2 or at least 3 factors, respectively; 12/27 (44.4%) cases had precipitating factors excluding infection. According to multivariate analysis, risk of ICH was increased with life-threatening bleeding (OR: 21 [1.9–243], P = 0.0143) and nonresponse to steroids during ITP (OR: 59 [6.7–523.7], P = 0.0002) and reduced with cutaneous bleeding (OR: 0.01 [0–0.34], P = 0.01). At the time of ICH, the median platelet count for cases was 6 × 109/L (range: 1–86 × 109/L); 4 (15%) had a platelet count >30 × 109/L and at least one precipitating condition. The median bleeding score 2 (excluding the ICH score) at the time of ICH was 8 (range: 0–29). Only 12/23 cases (52%) with platelet count <30 × 109/L at the time of ICH were treated with steroids, IVIg and fractionated platelet transfusion combined as recommended in the guidelines published in 2011. The mortality rate was 44% (n = 12). This is the first case–control study of risk factors for ICH in adult patients with ITP. A case-control study of 40 children with ITP and ICH found increased risk of ICH with cranial trauma and hematuria 3. Our controls were from hospitalized patients for easier identification, their platelet count at inclusion <50 × 109/L instead of 100 × 109/L in order to select patients with more active disease but this may represent a bias by selecting controls with more severe disease. Life-threatening bleedings were significantly related to ICH occurrence. Cortelazzo et al. found a previous significant bleeding episode during ITP to be a major risk factor for severe hemorrhage (relative risk 27.5, P < 0.0005) 4. Risk of ICH was reduced with cutaneous bleeding and this might not be simply due to reporting bias, as suggested for children 3, but cutaneous bleeding could be associated with early ITP diagnosis and treatment. We found that nonresponse to steroids during ITP was strongly associated with ICH occurrence. No previous study assessed the response to ITP therapy before ICH. Portielje reported that a lack of response to ITP therapy during the 2 years after ITP diagnosis increased four-fold the risk of death due to hemorrhage or infection 5. In our study, 19 cases (70%) had at least one of the precipitating factors. Some patients with severe thrombocytopenia developed ICH, while others did not, assuming the role of precipitating factors and the role of residual platelets' function 6. Patients with mild bleeding scores seemed at low risk of ICH in the absence of a potential triggering factor. This finding reinforces the decision to treat patients with ITP based on bleeding severity rather than platelet count. The patients over 60 years of age did not exhibit an increased ICH risk. However, three of the four patients who had ICH at platelet count >30 G/L were aged ≥70-years old. In conclusion, this study showed that patients presenting with life threatening bleeding, nonresponse to steroid treatment, and precipitating factors are at high risk of ICH. Contribution: S.M.B. collected and analyzed the data and wrote the manuscript. O.F. designed and supervised all steps in the study and contributed to writing the article. B.G. coordinated, designed the study and helped write the article. A.S.M. contributed to the study design. A.A. reviewed the brain CT or MRI images. M.B. and V.L. analyzed data. F.A. contributed to data analyses. F.B. and M.K. helped write the article. The remaining authors contributed to data collection by identifying cases and selecting controls. The authors thank Mrs. Joan, Meriem, and Hayat Tazir, the American Journal expert and Laura Smales for manuscript corrections. Sara Melboucy-Belkhir,1* Mehdi Khellaf,2 Alexandre Augier,3 Marouane Boubaya,4 Vincent Levy,4 Guillaume Le Guenno,5 Louis Terriou,6 Bertrand Lioger,7Mikaël Ebbo,8 Anne-Sophie Morin,9 Marie-Paule Chauveheid,10 Marc Michel,11Farid Belkhir,12 Frédégonde About,13 Christian Rose,14 Guillaume Moulis,15Arsene Mekinian,16 Jérôme Stirnemann,17 Thomas Papo,10 Stéphane Cheze,18Eric Rosenthal,19 Jean-François Viallard,20 Nicolas Schleinitz,8 Lionel Galicier,21 Daniel Adoue,15 Olivier Lambotte,22 Mohamed Hamidou,23Bertrand Godeau,11 and Olivier Fain16 1Department of Internal Medicine, Saint-Quentin Hospital, Saint-Quentin, France; 2Department of Emergency, Henri Mondor Hospital, AP-HP, Université Paris-Est Créteil, Créteil, France; 3Department of Radiology, Avicenne Hospital, AP-HP, Université Paris XIII, Bobigny, France; 4Department of Clinical Research, Avicenne Hospital, AP-HP, Université Paris XIII, Bobigny, France; 5Department of Internal Medicine, Estaing University Hospital, Clermont Ferrand, France; 6Department of Hematology, Claude Huriez University Hospital, Lille, France; 7Department of Internal Medicine, Bretonneau University Hospital, Tours, France; 8Department of Internal Medicine, La Conception Hospital, Assistance Publique Hôpitaux de Marseille, Université Aix-Marseille, Marseille, France; 9Department of Internal Medicine, Jean Verdier Hospital, AP-HP, Bondy, Université Paris XIII, Bondy, France; 10Department of Internal Medicine, Bichat Hospital, AP-HP, Université Paris VII, Paris, France; 11Department of Internal Medicine, Henri Mondor Hospital, AP-HP, Université Paris-Est Créteil, Créteil, France; 12Department of Onco-radiotherapy, Saint-Quentin Hospital, Saint-Quentin, France; 13Department of Statistics and Public Health, Henri Mondor Hospital, AP-HP, Université Paris-Est Créteil, Créteil, France; 14Department of Hematology, Saint-Vincent de Paul Hospital, Lille, France; 15Department of Internal Medicine, Purpan University Hospital, Toulouse, France; 16Department of Internal Medicine, Saint Antoine Hospital, DHUi2B, Université Paris VI, Paris, France; 17Department of Internal Medicine, Geneva University Hospital, Geneva, Switzerland; 18Department of Hematology, Clémenceau University Hospital, Caen, France; 19Department of Internal Medicine, L'Archet 1 University Hospital, Nice, France; 20Department of Internal Medicine, Haut-Lévêque University Hospital, Pessac, France; 21Department of Clinical Immunology, Saint Louis University Hospital, Paris, France; 22Department of Internal Medicine, Bicêtre Hospital, AP-HP, Université Paris Sud, Le Kremlin-Bicêtre, France; 23Department of Internal Medicine, Hôtel Dieu University Hospital, Nantes, France