A 42-year-old male was admitted to the Department of Hepatobiliary Internal Medicine 10 d following surgery for meningioma with fatigue and abdominal distension for 1 d.
History of present illness
Ten days before admission to the Department of Hepatobiliary Internal Medicine, the patient underwent surgery for atypical meningioma (World Health Organization grade II) resection because of progressive deterioration of vision in the right eye for more than half a year and headache for more than one month. His preoperative hematological parameters were essentially normal: Hemoglobin (Hgb), 132 g/L (130-175 g/L); platelet (PLT) count 99 × 109/L (125-350 × 109/L); alanine aminotransferase (ALT), 34.7 U/L (7-50 U/L); aspartate aminotransferase (AST), 22.5 U/L (13-40 U/L); total bilirubin (TBIL), 12 µmol/L (< 21 mmol/L); direct bilirubin (DBIL), 2.7 µmol/L (< 5 mmol/L); indirect bilirubin (IBIL), 9.3 µmol/L (< 16 mmol/L); creatine kinase (CK), 86 U/L (26-174 U/L); lactate dehydrogenase (LDH), 226 U/L (109-245 U/L); serum creatinine (Scr), 94 µmol/L (53-115 mmol /L); prothrombin time (PT), 11.9 s (9.8-12.1s); international normalized ratio (INR), 1.05 (0.82-1.15); D-dimer (D-D), 0.14 mg/L (< 0.5 mg/L); fibrinogen (FIB), 2.48 g/L (2-4 g/L). Preoperative abdominal color Doppler ultrasound showed that his other organs were normal except a mildly coarse hepatic parenchymal echotexture. Cranial contrast-enhanced magnetic resonance imaging indicated the right sphenoid ridge meningioma (6.6 cm × 5.5 cm). After the surgery, sodium valproate 1.2 g per day was used to treat secondary epilepsy. On postoperative day 3, sodium valproate was increased to 1.6 g per day and was given for 7 d due to the exacerbation of secondary epilepsy. On postoperative day 5, the patient started to develop symptoms of cerebral edema with headaches and intermittent vomiting; therefore, methylprednisolone 1000 mg per day was given for 4 d (postoperative day 5 to day 8). During his surgical hospitalization, because of fever on postoperative day 1 and increased cerebrospinal fluid leukocytes on postoperative day 5, antibiotics were used for anti-infective therapy as follows: Ceftriaxone plus vancomycin (postoperative day 1 to day 4) and biapenem plus linezolid (postoperative day 5 to day 9). The major medications used in the patient during hospitalization are recorded in Figure 1.
Figure 1 Major medications used in the patient during hospitalization.
The dosage of the antibiotics are as follows: Ceftriaxone 2 g/q12h; vancomycin 1000 mg/q12h; biapenem 0.6 g/q12h; linezolid 600 mg/q12h.
On postoperative day 9, the patient began to experience restlessness and progressive abdominal distension. His hematological parameters were as follows: Hgb, 103 g/L; PLT count, 109 × 109/L; ALT, 5713.8 U/L; AST, 7329.5 U/L; TBIL, 71.7 µmol/L; DBIL, 33.6 µmol/L; IBIL, 38.1 µmol/L; CK, 3912 U/L; LDH, 7744 U/L; Scr, 80 µmol/L; PT, 18.3 s; INR, 1.63; and D-D, 10.63 mg/L. Then, he received treatment for liver protection, gastrointestinal motility promotion, and enema; however, he did not recover from aggravated abdominal distension and gradually developed sleepiness, fatigue, oliguria, soy sauce-colored urine, and jaundice. Therefore, sodium valproate was discontinued. and he was transferred to the Department of Hepatobiliary Internal Medicine for further treatment on postoperative day 10.
History of past illness
The patient was diagnosed with hepatitis B surface antigen (HBsAg) positivity without antiviral treatment for 3 years due to continuous normal hepatic function, and hepatitis B virus (HBV) DNA was < 500 IU/mL (< 500 IU/mL) during periodic re-examinations.
Personal and family history
The patient did not have a history of smoking or alcoholism or a remarkable family medical history.
The patient was in a somnolent state with myoclonic jerks in the upper limbs. His skin and sclera were severely yellow. Bilateral petechia and ecchymosis were noted both in the lower limbs and near the injection and puncture sites. His abdomen was distended, with tympanic percussion sounds.
On postoperative day 10, the patient’s laboratory data showed that his Hgb was 61 g/L, hematocrit was 18.3% (40%-50%), PLT count was 56 × 109/L, D-D exceeded 35.2 mg/L, PT was 76.9 s, INR was 7.25, and FIB was 0.88 g/L. His reticulocyte percentage was up to 4.5% (0.5%-1.5%). Although the urinalysis revealed that his urine was positive for occult blood and urobilinogen, there were no fragmented erythrocytes on the peripheral blood smear. His ALT was 16144.0 U/L, AST exceeded 21000.0 U/L, CK was 4471.0 U/L, LDH was 21962.0 U/L, TBIL was 144.1 µmol/L, DBIL was 19.6 µmol/L, IBIL was 124.5 µmol/L, and Scr was 255.0 µmol/L. HBV DNA, ceruloplasmin, and autoimmune antibodies in his serum were all negative. The hematological indices of the patient during hospitalization are shown in Figure 2.
Figure 2 Hematological indices of the patient during hospitalization.
A: Serum aminotransferase; B: Serum bilirubin; C: Renal function; D: Serum myocardial enzymes; E: Coagulation function; F: Routine blood parameters. ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; TBIL: Total bilirubin; DBIL: Direct bilirubin; IBIL: Indirect bilirubin; Scr: Serum creatinine; BUN: Blood urea nitrogen; LDH: Lactate dehydrogenase; CK: Creatine kinase; PT: Prothrombin time; INR: International Normalized Ratio; D-D: D-dimer; FIB: Fibrinogen; Hgb: Hemoglobin; PLT: Platelet; HCT: Hematocrit.
The morphology and density of the liver were normal on preoperative day 5. On postoperative day 9, geographical patterns of low-density shadow in the liver were demonstrated on computed tomography (CT), which suggested extensive hepatic necrosis. No noticeable abnormality was detected in the kidneys and spleen, but a small amount of pelvic and peritoneal effusions. The abdominal CT of the patient during hospitalization is shown in Figure 3.
Figure 3 Abdominal computed tomography of the patient during hospitalization.
A: On preoperative day 5, the morphology and density of the liver were normal; B: On postoperative day 9, the density of several parts of the liver was low.
On the 9th postoperative day for meningioma, the patient developed ALF with extreme fatigue, severe jaundice, prolonged prothrombin time, and acute progressive hepatic coma. Although he was an HBsAg carrier, the HBV DNA in his serum was negative; thus, the possibility of liver injury caused by HBV reactivation can be excluded. In addition, the patient’s history suggested no trauma, exertion, hyperthermia, or infections, which are common causes for liver failure, and his symptoms did not include myalgia, which is typical in rhabdomyolysis. Although he was on several medications during hospital stay, the major adverse effects did not include rhabdomyolysis. Therefore, the liver failure was not considered to be associated with rhabdomyolysis. Because his alcoholism history and hematological indices of autoimmune diseases were negative, we finally considered his ALF to be attributed to the drug.
Sodium valproate, vancomycin, and linezolid are the common drugs associated with DILI[5,7,8]. Among all the adverse reactions of vancomycin, most are rapid onset, and renal injury occurs more frequently than liver injury. Common adverse reactions of linezolid include myelotoxicity and peripheral and optic neuropathy. Studies show that liver injury induced by linezolid usually occurs after two weeks of medication, with a mild elevation of transaminase. The characteristics of our patient who developed severe liver dysfunction after using linezolid for only 5 d were inconsistent with those reported in the literature. Liver injury is one of the most often reported adverse effects of sodium valproate because sodium valproate is metabolized by glucuronidation and mitochondrial beta-oxidation in the liver. ALF due to sodium valproate can still be encountered in the clinic. The RUCAM scoring table is recognized as the primary DILI causality assessment tool. The RUCAM score of this patient was 6 (probable) when we completed the causality assessment for sodium valproate. Thus, we considered that his liver injury was induced by sodium valproate.
Sodium valproate, as one of the most widely used broad-spectrum antiepileptic drugs, has been used as a first-line treatment in clinical practice. It has the characteristics of high bioavailability, good tolerability, and remarkable efficacy except for a narrow therapeutic window. A black box warning of severe hepatotoxicity with sodium valproate was published by the Food and Drug Administration. The pathogenesis of the toxic effects of sodium valproate on the liver has not yet been fully elucidated. The obstacles of cytochrome P450 metabolism and β-oxidation in mitochondria are generally considered the major mechanisms of the hepatotoxicity of sodium valproate, while the former is an important factor for individual differences in liver injury. Studies have shown that risk factors for fatal liver failure caused by sodium valproate include age younger than 2 years, combination therapy with sodium valproate and other antiepileptic drugs, pregnancy, a history of liver diseases, and other neurological diseases[16,17]. It is controversial whether the liver injury induced by sodium valproate is related to the plasma concentration of sodium valproate. The research has shown that patients with high plasma concentrations of sodium valproate are more susceptible to liver injury than those with low plasma concentrations. However, Ghozzi et al supposed that liver injury is independent of the sodium valproate plasma concentration. However, there is currently no research on the correlation between HBV infection and liver injury induced by sodium valproate. Therefore, whether the fatal liver failure induced by sodium valproate in our patient is related to HBV infection needs further research.
Except for ALF, TMA should be considered in diagnosis. Generally, microangiopathic hemolytic anemia is a sine qua non for the diagnosis of TMA. This patient exhibited anuria, hemoglobinuria, progressively decreased levels of hemoglobin and platelets, elevated proportion of reticulocytes, and significantly elevated bilirubin and D-D level during the early stage of his disease, which were consistent with the characteristics of TMA. However, no fragmented erythrocytes were observed on the peripheral blood smear either before or after PE. Cases of TMA without the presence of fragmented erythrocytes on the peripheral blood smear have been reported previously[21,22]. This patient responded to PE and methylprednisolone. Therefore, a diagnosis of TMA was suggested. As far as we are concerned, the presence of fragmented erythrocytes is essential for TMA diagnosis in most cases, but not all.
TMA is defined as a clinical syndrome characterized by thrombocytopenia, hemolytic anemia, and multiple organ dysfunction. The microthrombosis of capillaries and arterioles is the typical pathological feature of TMA. Thrombotic thrombocytopenic purpura (TTP) and hemolytic uremic syndrome (HUS) are the two clinical presentation forms of TMA. TTP mostly involves the nervous system, while there is a preponderance of renal injury in HUS patients. The etiology of TMA is multifactorial, including genetic factors and acquired risk factors. DI-TMA is a type of acquired TMA that is caused by multiple drugs, such as antitumor agents, antiplatelet drugs, and oral contraceptives. Studies show that DI-TMA occurs via two main mechanisms: Immune-mediated reactions or dose-dependent toxicity. TMA mediated by dose-dependent reactions has a slow, progressive onset, while immune-mediated TMA has a rapid onset. Our patient can be classified as having immune-mediated DI-TMA.
For the patient, it was inferred that sodium valproate plus carbapenem antibiotics might increase the risk of hemolysis, which could cause TMA. However, the period of the combination was short. König et al showed that sodium valproate altered the exposure of immunoglobulin receptor and fatty acid content in the erythrocyte membrane. Alteration of the membrane fluidity and receptor proteins on the membrane would facilitate immunoglobulin to destruct erythrocytes. The correlation between destruction of erythrocytes and plasma level of sodium valproate remains unclear. Carbapenem antibiotics can reduce the plasma level of sodium valproate by inhibiting multidrug resistance-associated proteins, which can efflux sodium valproate back to the plasma from erythrocytes and result in an increased erythrocyte distribution of sodium valproate. Therefore, it would be of great significance to study the effects of interaction between carbapenem antibiotics and sodium valproate on hemolysis.
In summary, this is the first case report of ALF with TMA caused by sodium valproate, which is notable. However, there were some limitations to our study. Liver pathology was unavailable. Without results regarding ADAMTS13 activities or anti-ADAMTS13 antibodies, we found it difficult to distinguish between HUS and TTP in this patient. Establishing an animal model can facilitate further studies on pathogenesis and pathophysiological state.