CLINICAL INVESTIGATION
The Significance of Platelet Consumption in the Development of Thrombocytopenia in Patients With Cirrhosis Yoshihiro Ikura, MD, Masahiko Ohsawa, MD, Masako Okada, Yasuhiro Iwai, MD and Kenichi Wakasa, MD
Abstract: Introduction: Thrombocytopenia in cirrhosis is mainly explained by accelerated platelet destruction/sequestration because of hypersplenism or by decreased thrombopoietin levels. Excessive platelet consumption because of cirrhosis-related hypercoagulability has also been assumed to be an etiopathologic factor in thrombocytopenia. To clarify whether excessive platelet consumption (eg, venous thrombosis and disseminated intravascular coagulation) contributes to thrombocytopenia in cirrhosis, the following autopsy-based study was performed. Methods: Ninety-nine autopsies of chronic liver disease (80 cirrhosis and 19 noncirrhosis) were examined retrospectively. Platelet count, weight of spleen and thrombotic complications were checked in autopsy protocols. Megakaryocytes in bone marrow were counted under highpower microscopic view. Univariate and multivariate analyses were performed to evaluate significances of these factors in thrombocytopenia. Results: The platelet count was significantly lower in the cirrhosis cases (88 6 51 3 109/L) than in the noncirrhosis cases (150 6 120 3 109/L). The megakaryocyte count was also lower in the cirrhosis cases (1.5 6 0.6 per high-power field) than in the noncirrhosis cases (1.9 6 0.5 per high-power field). The weight of the spleen was greater in the cirrhosis cases (264 6 179 g) than in the noncirrhosis cases (142 6 82 g). Thrombotic complications had been recorded in 29 cases, whose platelet count (70 6 41 3 109/L) was lower than that of those without these complications (113 6 80 3 109/L). Multivariate analysis revealed that these 3 factors (megakaryocyte count, weight of spleen, and thrombotic complications) were independently correlated with the platelet count. Conclusions: These results suggest that the imbalance of platelet production-destruction/sequestration-consumption contributes to thrombocytopenia in cirrhosis. Excessive platelet consumption cannot be ignored to explain this complex condition, especially in patients with major thrombotic events. Key Indexing Terms: Thrombocytopenia; Cirrhosis; Platelet consumption; Thrombosis. [Am J Med Sci 2013;346(3):199–203.]
pathway to this complex condition has not been fully elucidated and is still controversial.1,6,7 Although it seems to be rather paradoxical, patients with cirrhosis are sometimes hypercoagulable as well as with hemorrhagic diathesis.8–10 The production of many functional proteins, including prothrombotic and fibrinolytic factors, from cirrhotic livers markedly deteriorates. Thus, the patient’s hemostatic condition is very sensitive to the prothrombotic/antithrombotic imbalance. The existence of the relatively larger amount of prothrombotic factors than that of antithrombotic factors in patients’ blood will induce a hypercoagulable status in them. The hypercoagulability potentially contributes to the development of thrombotic complications in cirrhosis, such as portal vein thrombosis and pulmonary embolism.8–11 In a very early study,12 excessive platelet consumption was assumed to be an etiopathologic factor in cirrhosis-associated thrombocytopenia. Failure of antiplatelet drugs in the treatment of it13 and discovery of reduced TPO production in diseased livers4 have weakened the authenticity of the hypothesis. However, intravascular platelet consumption (ie, platelet aggregation) indeed occurs and leads to thrombotic complications in patients with cirrhosis and is considered to contribute to progression of the liver disease itself (ie, parenchymal extinction).9,10 Thus, it is still worthy to explore the significance of platelet consumption in the pathophysiology of cirrhosis. To clarify the significance of platelet consumption including venous thrombosis and disseminated intravascular coagulation (DIC) in cirrhosis-associated thrombocytopenia, we reviewed autopsy files of cases of chronic liver disease with or without cirrhosis and analyzed the platelet productiondestruction/sequestration-consumption balance in these subjects.
MATERIALS AND METHODS
T
hrombocytopenia is a common complication of chronic liver diseases, especially cirrhosis.1,2 Bleeding disorders based on the thrombocytopenia are serious and sometimes can be fatal. A decrease in the number of circulating platelets in patients with cirrhosis has been recognized as a multifactorial phenomenon. Accelerated platelet destruction or sequestration in enlarged spleen and suppressed thrombopoiesis because of reduced thrombopoietin (TPO) production from diseased livers largely account for its pathophysiology.3–5 However, the precise
From the Departments of Pathology (YIK, M OKADA), and Diagnostic Pathology (M OHSAWA, KW), Osaka City University Graduate School of Medicine, Osaka, Japan; and Department of Pathology (YIK, YIW), Takatsuki General Hospital, Takatsuki, Japan. Submitted June 20, 2012; accepted in revised form August 8, 2012. The authors have no financial or other conflicts of interest to disclose. Correspondence: Kenichi Wakasa, MD, Department of Diagnostic Pathology, Osaka City University Graduate School of Medicine, 1-4-3 Asahimachi, Abeno-ku, Osaka 545-8585, Japan (E-mail:
[email protected]).
The American Journal of the Medical Sciences
Autopsy Cases From January 2000 to March 2010, 70 cases of chronic liver disease were autopsied at Osaka City University Hospital. From January 2003 to December 2011, 52 cases of chronic liver disease were autopsied at Takatsuki General Hospital. Among these 122 autopsies, 23 cases were excluded from the present analysis because of the following reasons: lack of laboratory data of platelet count within 5 days before death (4 cases), platelet transfusion within 7 days before death (4 cases), administration of antiplatelet agents (acetylsalicylic acid, ticlopidine hydrochloride and cilostazol; 7 cases), thrombocytosis of undetermined cause (.500 3 109/L; 1 case), former splenectomy or lack of autopsy data on the weight of spleen (6 cases) and lack of a bone marrow specimen (1 case). The remaining 99 cases became the subjects of this study. In addition, 11 patients without hepatic and hematologic disorders autopsied at Takatsuki General Hospital were studied as controls. The autopsies were begun 1 to 21 hours (median value, 4.5 hours) after death. In cases when autopsies began $3 hours
Volume 346, Number 3, September 2013
199
Ikura et al
after death, subjects’ bodies had routinely been stored in a refrigerator (4°C) to reduce postmortem tissue degradation. This study was performed in accordance with the principles in the Declaration of Helsinki. Because all of the data were obtained retrospectively from the routine diagnostic procedure of autopsy, ethical approval was not required. Parameters The circulating platelet count, the number of megakaryocytes, the weight of spleen and the presence of major thrombotic complications were adopted as parameters to assess the platelet production-destruction/sequestration-consumption balance. The platelet count in the last laboratory test before death (#5 days) was informed from hospital charts. For comparative analyses, platelet counts measured at 2 to 4 weeks before death were also collected. The weight of spleen was checked in autopsy protocols. Concerning thrombotic complications, we searched the following findings: (1) a large vessel thrombus or embolus, such as portal vein thrombus and pulmonary embolus; (2) autopsy-proven (histologically diagnosed) DIC and (3) clinically diagnosed DIC. The histological diagnosis of DIC had been made by the detection of microthrombi in renal glomerular capillaries or pulmonary alveolar capillaries. It was confirmed that the clinical diagnosis of DIC fulfilled the diagnostic criteria of the International Society on Thrombosis and Haemostasis for DIC (DIC score $ 5).14 The presence of thrombotic complications was numerically expressed as 0 (absent) or 1 (present).
TABLE 1. Clinical characteristics of the patients studied Cirrhosis Noncirrhosis Controls (n 5 80) (n 5 19) (n 5 11) Agea (yr) Sex (male/female) Etiology Hepatitis B Hepatitis C Alcoholic Others ALTa (IU) PT/INRa Child-Pugh class (A/B/C) Malignant complications HCC Extrahepatic malignancies Causes of death GI bleeding Liver failure Ruptured HCC Infections Others
68 6 9 56/24
70 6 11 15/4
80 6 19 9/3
4 54 16 6 237 6 525 2.52 6 1.82b 1/13/66b
2 14 3 0 81 6 98 1.54 6 0.78 3/12/4
— — — — 49 6 43 — —
52 6
7 7
0 5
15 30 8 19 8
1 5 0 4 9
0 0 0 8 3
Data are shown as mean 6 standard deviation. P , 0.05 vs. noncirrhosis (Student’s t test or x2 test). ALT, alanine aminotransferase; GI, gastrointestinal; HCC, hepatocellular carcinoma; PT/INR, prothrombin time/international normalized ratio. a
Megakaryocyte Counting In every autopsy, lumbar vertebrae had routinely been sampled as a bone marrow specimen. Megakaryocytes were counted in 20 randomly chosen high-power (340 objective lens) fields of the bone marrow tissue section stained with hematoxylin-eosin. The result was expressed as the average number of megakaryocytes per high-power field (hpf). Statistics All statistical analyses in this study were performed with StatView version 4.0 (Abacus Concepts, Berkeley, CA). For contingency table analysis, Fisher’s exact test or x2 test was used. Data comparisons between 2 groups were performed by Student’s t test. Correlations between 2 variables were examined with Pearson’s least square regression. Potential determinants of the circulating platelet count, including the weight of spleen, the number of megakaryocytes and the presence of thrombotic complications, were tested by multiple regression analysis. P , 0.05 was considered statistically significant.
RESULTS Clinical Data Among the 99 patients with chronic liver disease, 19 cases were of chronic hepatitis without cirrhosis (noncirrhosis) and 80 cases were of cirrhosis (Table 1). Age, sex and etiologies of the liver diseases were not different between these 2 groups. There were significantly more patients categorized in Child-Pugh class C among the cirrhosis cases than among the noncirrhosis cases (P , 0.0001; Table 1). The prothrombin time/international normalized ratio (PT/INR) value was significantly higher in the cirrhosis cases than in the noncirrhosis cases (P 5 0.045; Table 1). Although the cirrhotic patients had largely died of liver-related causes (53 of 80 cases), including variceal bleeding, liver failure and ruptured hepatocellular
200
b
carcinoma (HCC), only 6 of the 19 noncirrhosis cases were liver-related death (P 5 0.009). Platelet Counts Circulating platelets measured within 5 days before death (114 6 87 3 109/L) were significantly fewer in the cirrhosis cases than in the noncirrhosis cases (P 5 0.001) or the control subjects (P , 0.0001; Table 2). The platelet count of the patients was inversely correlated with the PT/INR value (r 5 20.311, P 5 0.003). Fourteen patients who died of gastrointestinal bleeding showed relatively lower platelet counts (85 6 38 3 109/L), but the difference from that in the other 85 patients (103 6 77 3 109/L) was not statistically significant. The Child-Pugh classification did not relate to the platelet count, and the presence of HCC (59 cases) was associated with neither low nor high platelet counts. In 86 of 110 cases studied, platelet counts had been measured also at 2 to 4 weeks before death (145 6 85 3 109/L), which were significantly higher (P 5 0.014) than those within 5 days before death and were significantly fewer in the cirrhosis cases than in the noncirrhosis cases (P , 0.0001) or the control subjects (P , 0.0001; Table 2). Megakaryocytes and Spleen Megakaryocytes were significantly fewer in the cirrhosis cases than in the noncirrhosis cases (P 5 0.024) or the control subjects (P 5 0.020; Table 2). However, as the data indicated, the megakaryocyte count was mostly above the lower limit of the normal range (0.5–3.415 or 0.8–3.7/hpf [data of the present control subjects]) even in the cirrhosis cases. The megakaryocyte Volume 346, Number 3, September 2013
Platelet Consumption and Thrombocytopenia in Cirrhosis
TABLE 2. Results of the platelet study Cirrhosis (n 5 80) (3109/L)
(#5 days before death) Platelet count Platelet count (3109/L) (2–4 weeks before death) Megakaryocyte count (hpf) Weight of spleen (g)
88 110 1.5 264
6 6 57a,b (n 5 60) 6 0.6a,b 6 179a,b 51a,b
Noncirrhosis (n 5 19) 150 213 1.9 142
6 6 6 6
120 90 (n 5 17) 0.5 82a
Controls (n 5 11) 236 247 2.0 69
6 6 6 6
107 79 (n 5 9) 0.8 20
Data are shown as mean 6 standard deviation. a P , 0.05 vs. controls. b P , 0.05 vs. noncirrhosis (Student’s t test). hpf, high-power field.
count was lower in patients of Child-Pugh class C (1.5 6 0.6/hpf) than in those of Child-Pugh class A or B (1.8 6 0.7/hpf, P 5 0.012) and was decreased in association with lowering of the platelet count (r 5 0.261, P 5 0.006; Figure 1). The weight of spleen was greater in the cirrhosis cases than in the noncirrhosis cases (P 5 0.005) or the controls (P 5 0.001; Table 2) and was greater in patients of Child-Pugh class C (264 6 183 g) than in those of Child-Pugh class A or B (183 6 120 g, P 5 0.034). Despite some exceptions (7 patients with moderate thrombocytopenia [,80 3 109/L] but without splenomegaly [,100 g]), increased weights of spleen were associated with lower platelet counts (r 5 20.267, P 5 0.005). Thrombotic Complications A total of 29 patients (29%) had thrombotic complications: 13 patients had portal vein thrombosis, 6 patients had other large vessel thrombosis (2 cases of inferior vena cava thrombosis and 1 case each of hepatic vein thrombosis, iliac vein thrombosis, iliac artery thrombosis and left atrial thrombosis), 1 patient had pulmonary thromboembolism and 12 patients had DIC (3 cases had autopsy-proven DIC [multiple microthrombi in pulmonary capillaries or renal glomeruli] and 9 cases had clinically diagnosed DIC [International Society on Thrombosis and Haemostasis DIC score, 5–7 points]). Three patients with DIC also had portal vein thrombosis or the other large vessel thrombosis. All of the large vessel thrombi were scarcely organized, suggesting that these thrombi had formed recently. Whereas the incidence of thrombotic complications was significantly higher in the liver disease cases
FIGURE 1. The correlation between megakaryocyte count and platelet count in the patients (cirrhosis and noncirrhosis) and the control subjects (Pearson’s least square test). Black circles, cirrhosis case; gray circles, noncirrhosis case; white circles, control subject. Ó 2012 Lippincott Williams & Wilkins
(29%) than in the nonliver disease autopsies (18%; P 5 0.009) of our institutes of the identical periods, it was not different between the cirrhosis cases (24 [30%] of 80 cases) and the noncirrhosis cases (5 [26%] of 19 cases). Neither advanced Child-Pugh class C nor the presence of HCC linked to the occurrence of thrombotic complications. There was no relationship between the presence of thrombotic complications and the platelet counts 2 to 4 weeks before patient’s death (137 6 84 3 109/L [patients with thrombotic complications; 22 cases] vs. 131 6 76 3 109/L [patients without thrombotic complications; 55 cases]). However, the platelet count measured within 5 days before death was significantly lower (P 5 0.007) in patients with thrombotic complications (70 6 41 3 109/L) than in those without thrombotic complications (113 6 80 3 109/L; Figure 2). The presence of thrombotic complications was also associated with deteriorated prothrombin activity; the PT/INR value was significantly elevated in patients with thrombotic complications (3.03 6 2.05) than in those without thrombotic complications (2.06 6 1.49, P 5 0.013). Multivariate Analysis To assess the significance of these 3 factors, especially the presence of thrombotic complications, as determinants of
FIGURE 2. The difference in platelet count between patients with thrombotic complications and those without thrombotic complications. The mean platelet count is significantly lower (P 5 0.007, Student’s t test) in patients with thrombotic complications (70 6 41 3 109/L) than in those without thrombotic complications (113 6 80 3 109/L). Horizontal bars indicate mean values of individual groups. Black circles, cirrhosis case; gray circles, noncirrhosis case.
201
Ikura et al
the circulating platelet count, we performed the following multiple regression analyses. First, the megakaryocyte count and the weight of spleen were used as parameters, both of which were correlated with the platelet count (regression equation A; Table 3). Next, the presence of thrombotic complications was added as a parameter in the regression equation, in which all 3 parameters were correlated with the platelet count (regression equation B; Table 3). The results indicated that each of these factors was an independent determinant of the platelet count. Finally, we tested whether or not these factors had influenced the platelet count 2 to 4 weeks before death. Although the megakaryocyte count and the weight of spleen remained independent determinants of the platelet count, the presence of thrombotic complications was not a factor in the multiple regression equation (regression equations A9 and B9; Table 3).
DISCUSSION In this autopsy-based human study dealing with thrombocytopenia in cirrhosis, we successfully demonstrated that the circulating platelet count of patients with cirrhosis was determined principally by the rates of platelet production and destruction/sequestration. There have been many studies about the causes of thrombocytopenia in cirrhosis, which have tended to emphasize the significance of only a single factor. The recent clinical study by Pradella et al16 cautioned that thrombocytopenia in cirrhosis should be recognized as a multifactorial phenomenon. The authors suggested that cirrhosis-associated thrombocytopenia was basically attributable to platelet production-destruction imbalance. The present study further provided evidence that platelet consumption associated with major thrombotic events also significantly contributed to this complex pathophysiological phenomenon. Hypersplenism caused by cirrhosis-associated portal hypertension has been thought to play a crucial role in the development of thrombocytopenia.3 In fact, splenectomy or splenic embolization to recover the platelet count has often provided a satisfactory outcome.17 The contribution of hypersplenism via platelet destruction/sequestration to cirrhosisrelated thrombocytopenia was confirmed in the present study. Alternatively, a nonnegligible proportion of patients with moderate thrombocytopenia did not have splenomegaly (7 of 99 cases). This finding suggested that hypersplenism was not the sole etiopathologic factor in cirrhosis-related thrombocytopenia. Suppressed thrombopoiesis because of decreased TPO production from diseased livers has been recognized as another potential cause of thrombocytopenia in cirrhosis.3–5 In this study, we could confirm that suppressed megakaryocytopoiesis was closely related to thrombocytopenia in cirrhosis. However, because the decrease in megakaryocyte count was not so severe even in the cirrhosis cases, insufficiency of thrombopoiesis/ megakaryopoiesis was recognized to be a relative one against
the loss of platelets in periphery. A TPO receptor agonist, eltrombopag, temporarily showed an excellent result in the clinical trial for cirrhosis-associated thrombocytopenia,18 but the trial was halted later because of increased risk of portal vein thrombosis. Nonselective administration of TPO certainly raises the platelet count but on the other hand raises the coagulability.19 The failure of the trial of eltrombopag might have been associated with the relative (not absolute) insufficiency of thrombopoiesis/megakaryopoiesis in cirrhosis. We had considered that the balance among platelet production, destruction/sequestration and consumption was important to explain the multifactorial phenomenon of thrombocytopenia in cirrhosis. The multivariate analyses (Table 3) proved that platelet production represented by the megakaryocyte count and platelet destruction/sequestration represented by spleen weight were almost equally related to platelet count. It was concordant with a previous study that revealed imbalanced platelet kinetics (reduced production and accelerated turnover) in patients with cirrhosis-associated thrombocytopenia.16 These 2 factors were recognized to have changed gradually and to have influenced the platelet count chronically; they were correlated with the platelet count not only within 5 days before death (regression equation A) but also at 2 to 4 weeks before death (regression equation A9). Our multivariate analyses further revealed that platelet consumption represented by the presence of thrombotic complications was a significant cofactor comparable to the prior 2 factors (regression equation B). However, because this parameter was not significant in regression equation B9 for the platelet count at 2 to 4 weeks before death, the effect of platelet consumption on the platelet count is supposed to be of acute. Appropriateness of this interpretation was confirmed by the univariate analysis between platelet count and thrombotic complications. It is thought that, despite acute effect, excessive platelet consumption as well as reduced platelet production and accelerated platelet destruction/sequestration plays a crucial role in cirrhosis-associated thrombocytopenia. An early study by Stein and Harker12 suggested the contribution of excessive platelet consumption to the development of thrombocytopenia in cirrhosis through platelet kinetics analysis. Subsequent failure of recovery of platelet counts by antiplatelet agents cast doubt on this mechanism as a therapeutic target.13 The failure can be attributed to the misunderstanding that cirrhosis-associated thrombocytopenia was entirely explainable by a single factor. Moreover, their original data may imply simply increased platelet clearance. As a result of the present multivariate analysis, we revealed that excessive platelet consumption was one of the multiple factors in cirrhosisassociated thrombocytopenia. Platelet consumption had been forgotten but cannot be ignored for recognition of the accurate pathologic mechanism. In fact, DIC, which is not a rare complication of cirrhosis, always accompanies platelet waste, and
TABLE 3. Results of the multiple regression analysis Regression equationa,b
Adjusted b0 89 (P , 0.001) 102 (P , 0.001) 128 (P , 0.001) 129 (P , 0.001)
A B A9 B9
Adjusted b1 (megakaryocyte count) 0.260 0.250 0.306 0.305
(P (P (P (P
5 5 5 5
0.005) 0.005) 0.002) 0.002)
Adjusted b2 (weight of spleen) 20.265 20.240 20.397 20.396
(P (P (P (P
5 5 , ,
0.004) 0.007) 0.001) 0.001)
Adjusted b3 (thrombotic complications) — 20.274 (P 5 0.002) — 20.022 (not significant)
Platelet count 5 b0 + b1 3 (megakaryocyte count) + b2 3 (weight of spleen) + b3 3 (thrombotic complications). Equations A and B are for the platelet count #5 days before death, and equations A9 and B9 are for the platelet count 2–4 weeks before death. b0, y-intercept; b1–3, regression coefficients. a b
202
Volume 346, Number 3, September 2013
Platelet Consumption and Thrombocytopenia in Cirrhosis
acute platelet decrease in association with portal vein thrombosis has been reported.20,21 It was not surprising that thrombocytopenia was linked to PT/INR deterioration in cirrhotic patients. PT/INR deterioration was strengthened in patients with thrombotic complications. These suggest that, as well as the platelet counts, coagulation factors are affected by either the reduced productions in diseased livers or accelerated kinetics because of excessive consumption in cirrhotic patients. It also seems logical that the deterioration of overall liver function estimated by the Child-Pough classification was associated with suppressed megakaryopoiesis and hypersplenism/splenomegaly. Although the previous investigations found the association of HCC with decrease in platelet count and with portal or other venous thrombosis,22,23 we could not find such associations. Further studies of a larger sample size may be needed to clarify this matter. The present autopsy-based protocol has some inherent limitations. The possibility that agonal changes affected the findings cannot be excluded. A study entirely based on clinical data would be the best approach. The variables can be replaced by clinical data of patients alive. Spleen size can be measured by images of computed tomography or ultrasonography,24 thrombopoiesis can be estimated by reticulated platelet counts (immature platelet fraction),16,25 the platelet consumption rate can be estimated by beta-thromboglobulin, etc.26,27 An equation like the current multiple regression can be composed with these parameters and may be useful for estimating the weight of each parameter in thrombocytopenia.
CONCLUSIONS The present study indicated that both platelet production and platelet destruction/sequestration were independent and almost equally weighted determinants of the circulating platelet count in patients with cirrhosis. In addition, excessive platelet consumption was an important cofactor that provided acute and substantial effect to the platelet count. This suggests the necessity of reconsidering the significance of excessive platelet consumption because of thrombotic complications such as portal vein thrombosis and DIC in cirrhosis-associated thrombocytopenia.
ACKNOWLEDGMENTS We thank Mr. Yuki Nishikawa for his secretarial assistance in this work. REFERENCES 1. Tripodi A, Mannucci PM. The coagulopathy of chronic liver disease. N Engl J Med 2011;365:147–56. 2. Afdhal N, McHutchison J, Brown R, et al. Thrombocytopenia associated with chronic liver disease. J Hepatol 2008;48:1000–7. 3. Aster RH. Pooling of platelets in the spleen: role in the pathogenesis of “hypersplenic” thrombocytopenia. J Clin Invest 1966;45:645–57. 4. Peck-Radosavljevic M, Zacherl J, Meng YG, et al. Is inadequate thrombopoietin production a major cause of thrombocytopenia in cirrhosis of the liver? J Hepatol 1997;27:127–31. 5. Ishikawa T, Ichida T, Matsuda Y, et al. Reduced expression of thrombopoietin is involved in thrombocytopenia in human and rat liver cirrhosis. J Gastroenterol Hepatol 1998;13:907–13. 6. Freni MA, Spadaro A, Ajello A, et al. Serum thrombopoietin in chronic liver disease: relation to severity of the disease and spleenize. Hepatogastroenterology 2002;49:1382–5.
Ó 2012 Lippincott Williams & Wilkins
7. Kajihara M, Okazaki Y, Kato S, et al. Evaluation of platelet kinetics in patients with liver cirrhosis: similarity to idiopathic thrombocytopenic purpura. J Gastroenterol Hepatol 2007;22:112–8. 8. Northup PG, McMahon MM, Ruhl AP, et al. Coagulopathy does not fully protect hospitalized cirrhosis patients from peripheral venous thromboembolism. Am J Gastroenterol 2006;101:1524–8. 9. Northup PG, Sundaram V, Fallon MB, et al. Hypercoagulation and thrombophilia in liver disease. J Thromb Haemost 2008;6:2–9. 10. Lisman T, Caldwell SH, Burroughs AK, et al. Hemostasis and thrombosis in patients with liver disease: the ups and downs. J Hepatol 2010;53:362–71. 11. Parikh S, Shah R, Kapoor P. Portal vein thrombosis. Am J Med 2010; 123:111–9. 12. Stein SF, Harker LA. Kinetic and functional studies of platelets, fibrinogen, and plasminogen in patients with hepatic cirrhosis. J Lab Clin Med 1982;99:217–30. 13. Luzzatto G, Fabris F, Gerunda GE, et al. Failure of two anti-platelet drugs (indobufen and dipyridamole) to improve thrombocytopenia in liver cirrhosis. Acta Haematol 1987;77:101–6. 14. Taylor FB Jr, Toh CH, Hoots WK, et al. Towards definition, clinical and laboratory criteria, and a scoring system for disseminated intravascular coagulation. Thromb Haemost 2001;86:1327–30. 15. Singal R, Belliveau RR. Quantitation of megakaryocytes in normal bone marrow. Anal Quant Cytol Histol 1988;10:33–6. 16. Pradella P, Bonetto S, Turchetto S, et al. Platelet production and destruction in liver cirrhosis. J Hepatol 2011;54:894–900. 17. Akahoshi T, Tomikawa M, Korenaga D, et al. Laparoscopic splenectomy with peginterferon and ribavirin therapy for patients with hepatitis C virus cirrhosis and hypersplenism. Surg Endosc 2010;24:680–5. 18. McHutchison JG, Dusheiko G, Shiffman ML, et al. Eltrombopag for thrombocytopenia in patients with cirrhosis associated with hepatitis C. N Engl J Med 2007;357:2227–36. 19. Dultz G, Kronenberger B, Azizi A, et al. Portal vein thrombosis as complication of romiplostim treatment in a cirrhotic patient with hepatitis C-associated immune thrombocytopenic purpura. J Hepatol 2011; 55:229–32. 20. Zhang D, Hao J, Yang N. Protein C and D-dimer are related to portal vein thrombosis in patients with liver cirrhosis. J Gastroenterol Hepatol 2010;25:116–21. 21. Ohkubo M, Nagino M, Kamiya J, et al. Portal and mesenteric vein thrombosis after portal vein embolization in a patient with protein S deficiency. J Hepatobiliary Pancreat Surg 2004;11:338–41. 22. Kim SJ, Choi IK, Park KH, et al. Serum vascular endothelial growth factor per platelet count in hepatocellular carcinoma: correlations with clinical parameters and survival. Jpn J Clin Oncol 2004;34:184–90. 23. Connolly GC, Chen R, Hyrien O, et al. Incidence, risk factors and consequences of portal vein and systemic thromboses in hepatocellular carcinoma. Thromb Res 2008;122:299–306. 24. Picardi M, Martinelli V, Ciancia R, et al. Measurement of spleen volume by ultrasound scanning in patients with thrombocytosis: a prospective study. Blood 2002;99:4228–30. 25. Barsam SJ, Psaila B, Forestier M, et al. Platelet production and platelet destruction: assessing mechanisms of treatment effect in immune thrombocytopenia. Blood 2011;117:5723–32. 26. Han P, Turpie AG, Genton E. Plasma beta-thromboglobulin: differentiation between intravascular and extravascular platelet destruction. Blood 1979;54:1192–6. 27. Panasiuk A, Prokopowicz D, Zak J, et al. Activation of blood platelets in chronic hepatitis and liver cirrhosis P-selectin expression on blood platelets and secretory activity of beta-thromboglobulin and platelet factor-4. Hepatogastroenterology 2001;48:818–22.
203
