Systemic capillary leak syndrome following bosentan treatment in a patient with systemic sclerosis

Abstract

A previously healthy 54-year-old man was diagnosed with systemic sclerosis (SSc) based on Raynaud’s phenomenon, puffy hands, sclerodactyly, digital ulceration, oesophageal dysmotility, and positive anti-RNA polymerase III antibody. Nailfold capillaroscopy showed mild, nonspecific abnormalities, and there was no evidence of pulmonary hypertension, interstitial lung disease, or scleroderma renal crisis. He was prescribed bosentan for Raynaud’s phenomenon with digital ulcerations, and instructed to perform daily blood pressure measurements to facilitate early detection of scleroderma renal crisis. Following initiation of bosentan, he developed rapidly progressive dyspnoea, orthopnoea, and 10 kg weight gain, resulting in hospitalization on the third day of treatment. On clinical examination, he had substantial peripheral oedemas, and bedside ultrasound revealed bilateral pleural effusions and a pericardial effusion without cardiac tamponade. There were clinical and echocardiographic signs of intravascular hypovolaemia, with tachycardia, hypotension, and a small and hyperkinetic left ventricle. Laboratory tests revealed haemoconcentration, leucocytosis, acute renal failure, hyponatraemia, hyperkalaemia, hypoalbuminaemia, and metabolic acidosis (Table 1). He stabilized following crystalloid fluid resuscitation and thoracocentesis, and was admitted to the intensive care unit. Initial work-up excluded Addison’s crisis, sepsis, anaphylaxis, and scleroderma renal crisis. Urine analysis showed no proteinuria, and thoracocentesis revealed benign pleural fluid with a protein content of 29 g/L. Based on the constellation of third spacing, intravascular hypovolaemia, and hypoalbuminaemia without proteinuria, a diagnosis of systemic capillary leak syndrome (SCLS) was made. SCLS is characterized by increased capillary permeability, leading to massive extravasation of protein-rich fluid causing intravascular hypovolaemia and hypoalbuminaemia (1). It follows three phases: a prodromal phase with weight gain and fatigue, an extravasation phase, and a recovery phase. The extravasation phase may lead to haemodynamic collapse and hypoperfusion, while in the recovery phase, the extravasated fluid is recruited back to the intravascular space, causing a risk of pulmonary and cerebral oedemas and compartment syndrome. Bosentan was withdrawn, and the patient was treated with crystalloid fluids and low-dose steroids. On the third day of hospitalization, he entered the recovery phase with diminishing need for fluid resuscitation, and normalization of haemoglobin and white blood cell count, indicating a shift of fluid back into the intravascular compartment. Fluid resuscitation was terminated, with close observation of volume status. On the eighth day, he was discharged without complications. The pathophysiology of SCLS is not completely understood, but a role of endothelial cell apoptosis and elevated levels of vascular endothelial growth factor, angiopoietin-2, and pro-inflammatory cytokines has been hypothesized (2, 3). While vasculopathy and capillary leakage are involved in the pathophysiology of SSc (4), only a couple of cases with SSc and SCLS have been reported (5–7). Drugs may trigger SCLS, but bosentan-induced SCLS has not previously been reported. Bosentan is a dual endothelin receptor antagonist (ERA), which inhibits binding of endothelin1 (ET-1) to endothelin receptor A and B (ETA, ETB). The endothelin system is involved in regulation of vasomotor tone, vascular permeability, andfluid homeostasis, and fluid retention and oedema are commonly reported adverse events of ERAs (8). A recent study reported elevated levels of ET-1 and endothelial upregulation of ETA in patients with SCLS (3), which led the authors to speculate that ET-1 agonism at ETA may destabilize endothelial integrity and increase vascular permeability. Another study demonstrated increased vascular permeability and fluid retention with ETA-selective ERAs, probably mediated through ET-1 overstimulation at non-antagonized ETB (8). In the latter study, dual ERAs (such as bosentan) did not induce significant vascular alterations (8). The close temporal association between bosentan initiation and symptom onset makes it tempting to speculate that bosentan triggered SCLS in our patient, which further challenges our current understanding of the role of the ET-1/ETA/ETB axis in the pathophysiology of SCLS. Further mechanistic studies are required to better 426 Scand J Rheumatol 2019;48:426–427