Relation between lupus anticoagulant and splanchnic venous thrombosis in cirrhosis of the liver

BMJ 1994; 309 doi: (Published 23 July 1994) Cite this as: BMJ 1994;309:239
  1. F Violi,
  2. D Ferro,
  3. S Basili,
  4. A D'Angelo,
  5. G Mazzola,
  6. C Quintarelli,
  7. C Cordova
  1. Istituto di I Clinica Medica, Policlinico Umberto I, Rome 00185
  2. Italy Istituto di Terapia Medica, Universita La Sapienza, Rome
  3. Italy Servizio di Coagulazione IRCSS, H S Raffaele, Milan, Italy
  1. Correspondence to: Dr Violi
  • Accepted 10 March 1994

While angiographic studies have indicated that splanchnic venous thrombosis rarely occurs in patients with cirrhosis of the liver, a postmortem study has shown that it may occur in up to a fifth.1 We have shown that patients with cirrhosis have lupus anti-coagulant, which predisposes to venous and arterial thrombosis.2 In this study we determined whether splanchnic venous thrombosis is associated with lupus anticoagulant in patients with cirrhosis.

Patients, methods, and results

From October 1990 to November 1991, 73 consecutive patients (43 men) aged 35-77 with cirrhosis of the liver that had been diagnosed by liver biopsy entered the study. Liver failure was categorised as mild, moderate, or severe according to Child-Pugh's classification.

Twenty nine patients had markers of hepatitis B virus infection, 30 had markers of hepatitis C virus infection, 11 had a history of alcohol misuse, and in three the cause of cirrhosis was unknown. Patients with cancer or acute inflammation who needed immediate plasma or blood transfusions were excluded.

Ultrasonography showed thrombosis in nine patients (portal vein thrombosis in eight, mesenteric vein thrombosis in one). These findings were confirmed by enhanced computed tomography.

Blood samples taken within 48 hours of the patients' admission and mixed with 3.8% sodium citrate were checked for lupus anticoagulant with four coagulation tests. Lupus anticoagulant was regarded as being present if the coagulation time was prolonged in at least two of these tests and if a further, fifth test yielded a positive result.2 Anticardiolipin antibodies were measured as described elsewhere.2

To evaluate whether thrombosis was related to low concentrations of coagulation inhibitors we measured antithrombin III, protein C, and protein S3 concentrations in the patients with thrombosis (four men, five women; age 40-72) and in nine patients without (four men, five women; 45-70). In each group seven patients had moderate and two severe liver failure.

Data were analysed with the unpaired t test, Mann-Whitney U test, X2 test with continuity correction, or (if n<=5) Fisher's test and the two tailed test.

The patients' age and sex and the cause of cirrhosis did not significantly distinguish patients with and without splanchnic venous thrombosis. The proportion of patients with moderate or severe liver failure was greater in those with thrombosis than in those without, although the difference was not significant (table). Thrombosis was significantly associated with the presence of lupus anticoagulant (odds ratio 18.7 (95% confidence interval 2.8 to 143.5); P=0.0008) and anticardiolipin antibodies (6.7 (1.3 to 37.9); P=0.015) (table). Seven of the nine patients with lupus anticoagulant had markers of hepatitis C virus infection. Concentrations of antithrombin III, protein C, bound protein S, and free protein S did not differ between the patients with cirrhosis who had thrombosis and those who did not.

Characteristics of 73 patients with cirrhosis of liver with and without splanchnic venous thrombosis. Figures are numbers (percentages) of patients unless stated otherwise

View this table:


This study shows that 12% (95% confidence interval 5.8% to 22.1%) of patients with cirrhosis of the liver may have splanchnic venous thrombosis. Thrombosis is rare in compensated cirrhosis, which suggests that liver failure is an important factor. The pathogenesis of thrombosis in patients with cirrhosis is not known. Lupus anticoagulant has been reported in such patients2 but has not previously been evaluated as a risk factor for thrombosis.

Our finding that over half of the patients with thrombosis were positive for lupus anticoagulant or anticardiolipin antibodies indicates that antiphospholipid antibodies may be a risk factor. A previous study showed that antiphospholipid antibodies may promote hepatic veno- occlusive disease or splanchnic venous thrombosis.4 Plasma concentrations of antithrombin III, protein C, and protein S, which may be low in patients with cirrhosis, may also promote thrombosis, but no differences were found between patients with and without thrombosis. Although the origin of antiphospholipid antibodies in cirrhosis is unknown, lupus anticoagulant has often been observed in viral infections, particularly hepatitis C virus infection5; in our study three quarters of patients positive for lupus anticoagulant had markers of hepatitis C virus infection. In conclusion, this study shows a significant relation between splanchnic venous thrombosis and lupus anticoagulant, indicating that lupus anticoagulant may be an important risk factor for thrombosis in patients with cirrhosis.

We thank Professor Guido Valesini for his cooperation. This study was funded partly by the Andrea Cesalpino Foundation, Rome.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
View Abstract