The percutaneous treatment of stenosis and thrombosis in haemodialysis access achieves patency rates similar to those reported in the surgical literature and confirms that grafts must be avoided as much as possible given their poorer outcome, especially after the first thrombosis. Poorer outcome is also demonstrated in accesses of less than 1 year's duration.

Delayed maturation of native fistulas should lead systematically to imaging as an underlying stenosis is diagnosed in all cases. Interventional radiology can treat the majority of cases and achieve a 97% success rate but early recurrence of stenoses can occur. Multidisciplinary re-evaluation of the patient must, therefore, be performed after radiological salvage of the fistula.

Over 54 months, 70 short stenoses of 63 shunts (32 Brescia-Cimino fistulas, 31 grafts) in 59 patients necessitated a first percutaneous transluminal angioplasty (PTA). Restenosis led to 63 redilations in 38 lesions. Nine stents were inserted in seven grafts and two proximal veins in seven patients, the indication being that stenosis had recurred twice in 6 months. In three of these stenoses, five delayed intrastent redilations were necessary. Three previously dilated occluded grafts were recovered with local thrombolysis. Morbidity was 4.08%, with one immediate rupture, four delayed pseudoaneurysms (1-28 months), and two periprocedural bacteremias. Half (15 of 29) of graft stenoses and only 14% (four of 27) of Brescia-Cimino fistula stenoses had a mean restenosis interval of less than 6 months. The mean restenosis interval increased from 3.6 months +/- 0.5 (standard deviation) before stent placement to 15.2 months +/- 0.4 after stent placement (P < .001). Insertion of a stent can be advised when stenoses of graft venous anastomoses have recurred twice in less than 6 months. The combination of all interventional radiologic procedures allowed a significant improvement in secondary patency rates after PTA, with 82% at 1 year, 79% at 2 years, and 71% at 3 years.

The percutaneous declotting of forearm fistulae by manual catheter-directed thrombo-aspiration was effective in more than 90% of cases and yielded 50% primary and 80% secondary patency rates at one year. The results were poorer in upper arm fistulae. The need for maintenance reinterventions was three times smaller in forearm fistulae than in upper arm fistulae and grafts.

Death after withdrawing from dialysis was the most common cause of death in ESRD patients in our French population cohort. The patients who died after discontinuation of treatment were more often in a poor general condition, near the end of life, and most often the physician decided to stop dialysis treatment.

Compared to a moderate protein restriction (0.65 g/kg/day), a severe protein restriction (0.3 g/kg/day) supplemented by ketoanologues does not limit GFR decrease when GFR is below 20 mL/min/1.73m2, but improves phosphocalcic plasma parameters.

Radiology achieves results comparable with surgery, with minimal invasiveness and better venous preservation. However, wide variations in the results suggest that the degree of commitment of physicians might be as important as the type of technique used.

Our results suggest that the treatment of anastomotic stenoses should be surgical rather than endovascular. Angioplasty and surgery have shown similar results when used to treat other perianastomotic stenoses, but repeat procedures were more frequent with angioplasty.