Peripheral vascular calcification in long-haemodialysis patients: associated factors and survival consequences

Vascular calcifications (VCs) are frequently observed in chronic kidney disease (CKD) and haemodialysis (HD) patients. They have been associated with numerous factors, particularly hyperphosphataemia, excess calcium load, hypertension and increased mortality rate. The purpose of this study is to measure VCs in long-HD patients with good blood pressure and phosphate control, with the occasional use of sevelamer, using a plain radiological score to identify the associated factors and effects on the 1-year survival rate.

We studied HD patients from one centre using a semi-quantitative score ranging from 0 to 3 according to the severity and extent of VCs. The following patients' characteristics were compared according to their VC scores: medical history, treatments, blood pressure, standard biological data, fibroblast growth factor (FGF) 23, osteoprotegerin (OPG), whole PTH, beta-crosslaps, bone alkaline phosphatases and bone mineral density scores. One-year survival analyses were also performed.

Among the 250 HD patients of the centre, 161 were studied; the mean age was 67.2 +/- 13 years, 45% of the subjects were females, 35% were diabetics, and they had been on dialysis for between 1-486 months (median: 45 months) with a 3 x 5-3 x 8 h dialysis schedule using 1.5 mmol/l dialysate calcium and providing a mean 2.25 +/- 0.5 Kt/V. Only 17% of the patients were free from VCs and 11% had severe VCs. The factors associated with VCs were classified into 'classic' (age, diabetes, male gender, tobacco use, inflammation, more frequent warfarin treatment and peripheral vascular and cardiac diseases) and 'non-traditional' (higher FGF-23 and OPG serum levels, low albumin serum levels and low alfacalcidol and CaCO(3) use). In logistic regression, only age, diabetes and FGF-23 serum levels were associated with VC scores of 2 and 3. The patients with a score of 3 had a higher 1-year mortality rate (RR 2.1; P = 0.01) as compared to patients with a 0 score.

A plain radiological score showed the high prevalence (83%) of VCs in HD patients in spite of a long and intensive dialysis strategy and adherence to guidelines. The main associated factors were classic factors such as ageing and diabetes. No relationship was found with blood pressure and phosphataemia that remained well controlled in long dialysis; the association with FGF-23 serum levels may aggregate some non-traditional risk factors. The harmful effects of VCs on survival require their systematic assessment and optimization of the potentially modifiable associated factors in CKD and HD patients.