Superior vena cava stenosis in haemodialysis patients with a tunnelled cuffed catheter: prevalence and risk factors

BACKGROUND: Although superior vena cava (SVC) stenosis may be a life-threatening complication of haemodialysis (HD) catheters, its prevalence and risk factors in HD patients are unknown. Our aim was to assess the prevalence and risk factors for SVC stenosis in HD patients with a tunnelled cuffed catheter (TCC) and to describe its clinical presentation. METHODS: In this single-centre, retrospective cohort study, all in-centre chronic HD patients carrying a TCC (1 January 2008-31 December 2012) were included (n = 117 patients, 214 TCC, 80 911 catheter-days). SVC stenosis was defined as a diameter reduction >50% on phlebography or computed tomography. Imaging was triggered by clinical SVC stenosis syndrome or vascular access (VA)-related concerns. We recorded demographics, conditions potentially influencing catheter permeability (medications, carriage of thoracic devices), number of TCCs, total duration of TCC carriage, previous arteriovenous VA and last (in use at time of stenosis detection) TCC details (location, diameter and length). VAs created while a TCC was still used were also recorded. RESULTS: An SVC stenosis was found in 11 patients (9.4%, 0.14/1000 catheter-days), which represents almost one-quarter of patients undergoing imaging, whatever the cause (11/45). Only two presented with clinically obvious SVC stenosis. The number of TCCs per patient was 2.64 ± 1.8 in the SVC stenosis group versus 1.75 ± 0.94 in the negative group (P = 0.13). On multivariate analysis (Poisson), diabetes {incidence rate ratio [IRR] 4.63 [confidence interval (CI) 1.2-17.8]; P = 0.02} and total duration of TCC carriage [IRR 1.47 (CI 1.2-1.8) per year; P = 0.001] were associated with SVC stenosis, whereas age had a slightly protective effect [IRR 0.96 (CI 0.91-1.01); P = 0.01]. Limitations are the retrospective design, detection and survivor bias. CONCLUSION: SVC stenosis is not a rare condition, is mostly asymptomatic in the absence of a peripheral VA, is strongly associated with diabetes and is promoted by long TCC carriage. Age is slightly protective