Thrombotic microangiopathy associated with sunitinib, a VEGF inhibitor, in a patient with factor V Leiden mutation.

Haemolytic uraemic syndrome (HUS) and thrombotic thrombocytopenic purpura (TTP) are closely related disorders characterised by non-immune microangiopathic haemolytic anaemia and thrombocytopenia. Thrombotic microangiopathy is the underlying pathologic lesion in both syndromes. Under physiological conditions, platelets bind to endothelium via von Willebrand Factor (vWF) and are released back into the circulation by ADAMTS13, a metalloprotease that cleaves ultra-large, newly synthesized vWF multimers. A deficiency of ADAMTS13 (due to either a decrease in its production or the presence of a circulating inhibitor) may thus result in microvascular thrombi. However, patients with thrombotic microangiopathy may have normal ADAMTS13 levels and no evidence of a circulating inhibitor [1].

Sunitinib malate, an oral multi-targeted tyrosine kinase inhibitor that blocks the receptors for VEGF, has utility in a variety of metastatic tumours [2]. We report a patient who developed proteinuria, hypertension and biopsy-proven thrombotic microangiopathy after treatment with sunitinib malate. This patient had normal ADAMTS13 levels but had factor V Leiden mutation that may have predisposed her to this complication.

Case report

A 57-year-old female with metastatic renal-cell carcinoma began treatment with sunitinib (Sutent [SU11248]; Pfizer) 50 mg/day from January 2006 after failing cytokine therapy with interleukin-2. She had undergone left nephrectomy 2 years previously. Her past medical history included stage III chronic kidney disease, stage I hypertension, Factor V Leiden mutation, history of deep venous thrombosis 32 years earlier and hypothyroidism. Two weeks after starting sunitinib therapy, her haemoglobin decreased from 123 g/L to 117 g/L and the platelet count decreased from 246 × 109/L to 63 × 109/L. She was admitted with complaints of fatigue, anorexia, mild nausea/vomiting and decreased urine output. On exam, she was afebrile and hypertensive (206/114 mmHg). She had 1+ bilateral lower extremity oedema, but otherwise the examination was unremarkable. Laboratory findings on presentation were remarkable for elevated serum urea nitrogen and creatinine, 18.9 mmol/L and 366 μmol/L, respectively, marked thrombocytopenia, 36 × 109/L, and worsening anaemia, 111 g/L. Coagulation tests were normal. Haemolytic anaemia was confirmed by low serum haptoglobin, <0.1 g/L, elevated serum lactate dehydrogenase, 736 IU/L, and schistocytes on blood smear. The Coombs test was negative. Fibrin split products were elevated, 80 μg/ml. Urinalysis showed 3+ proteinuria with 86 red blood cells, 8 white blood cells, 3 granular casts and no red blood cell casts per high power field. The protein to creatinine ratio in a random urine sample was 15 mg/mg. Prior urinalyses had showed trace of 1+ proteinuria, without quantification. Serologic testing for glomerular disease was unrevealing. ADAMTS13 levels were normal and no inhibitor was present.

Sunitinib was discontinued immediately and she was treated with prednisone and daily plasma exchange over the next 7 days. Her anaemia and thrombocytopenia resolved and serum creatinine slowly decreased to near baseline. Transjugular kidney biopsy performed after normalisation of the platelet count showed several glomeruli with fragmented red blood cells within the mesangial areas. There was also thickening of the glomerular capillary wall with a double contour appearance (Figure 1). Pathology from the original nephrectomy showed normal renal parenchyma in non-neoplastic tissue.

Fig. 1

(A) Kidney biopsy with two glomeruli: one showing advanced tuft collapse (left lower corner) and the other with capillaries showing a double contour appearance. PAS stain, magnification 200×. (B and C) Segment of a glomerulus with thickened, double-contoured capillaries and increased mesangial matrix with entrapped fragmented red blood cells. H & E (B) and PAS stain (C), magnification 400×.

Discussion

There is increasing evidence that VEGF inhibitors could cause or at least predispose to thrombotic microangiopathy. For example, administration of anti-VEGF antibodies to animals has been shown to cause proteinuria associated with detachment of glomerular endothelial cells and downregulation of nephrin [3]. VEGF promotes repair of glomerular endothelium in experimental thrombotic microangiopathy [4], and elevated serum levels of VEGF were noted in patients 2 to 3 weeks after the onset of thrombotic microangiopathy, suggesting a reactive increase in VEGF in an attempt to repair glomerular endothelial damage [5]. A naturally circulating soluble form of the VEGF receptor, sFlt-1, binds VEGF in the circulation and acts as an antagonist. Upregulation of sFlt-1 is a characteristic of pre-eclampsia and administration of sFlt-1 to pregnant rats can induce hypertension, proteinuria and endotheliosis [6]. Thus VEGF inhibition may be the common mechanism in both patients receiving VEGF antagonists and patients with severe pre-eclampsia that results in endothelial cell injury and predisposes to thrombotic microangiopathy.

Adverse effects of VEGF inhibitors reported in clinical trials have included hypertension, proteinuria, anaemia, neutropenia and thrombocytopenia [2]. An increased incidence of thrombosis also occurs after VEGF inhibition, which is probably due to impairment of vascular integrity due to diminished regenerative capacity of endothelial cells and exposure of pro-coagulant phospholipids on the luminal plasma membrane or underlying matrix [7]. VEGF increases production of nitric oxide and prostaglandin I2; reduction in these substances after inhibition of VEGF signaling may predispose to thrombosis [7]. VEGF inhibition may also increase erythropoietin production, resulting in an increase in haematocrit and blood viscosity [7].

Thrombotic microangiopathy occurring in single patient treated with bevacizumab [8], sunitinib [9], VEGF trap treatment [10] and sequential bevacizumab and sunitinib [11] has been very recently reported. However, an unanswered question is what predisposes only certain patients to this severe complication of VEGF inhibition.

Our patient had a history of factor V Leiden with prior deep venous thrombosis. This hypercoaguable state, in light of endothelial injury due to sunitinib-induced VEGF inhibition, may have predisposed her to thrombotic microangiopathy. This theory is supported by a recent study involving 30 patients with thrombotic microangiopathy and normal ADAMTS13 activity, in whom 4 (13%) were found to have factor V Leiden mutation, whereas only 3% of control subjects (6/186) possessed the mutation [12].

Although such data provide a plausible hypothesis as to why our patient developed this complication, it should be noted that factor V Leiden is generally associated with venous rather than microvascular thombosis, whereas anti-VEGF agents (as well as other antiangiogenic agents) have been more commonly associated with arterial and microvascular thrombosis [13]. Moreover, the association between factor V Leiden and thrombotic microangiopathy has been recently questioned [14]. Thus it remains unclear as to whether or not factor V Leiden will prove to be a predisposing factor for thrombotic complications of anti-VEGF agents. At our present state of knowledge, the presence of Factor V Leiden should not be considered a contraindication to the use of an anti-VEGF agent.