Twenty-one patients whose diagnosis had been made between 1 and 3 months before the commencement of dialysis was excluded from the analysis. The main clinical features of the late diagnosis group at presentation were dyspnoea/pulmonary oedema (41%), severe hypertension (26%),

severe asthenia (22%) and apathy/mental changes (8%). The rate of pulmonary infections (17.9% vs 5.1%, P < 0.01) and mean systolic blood pressure (172 ± 4 mmHg vs 161 ± 4 mmHg) were significantly higher in the late diagnosis group. All patients in the late diagnosis group required a CVC for initiation of dialysis. In the early diagnosis group, 33% of patients had a vascular access created electively. Creatinine clearance at the time of initiation of dialysis was significantly lower in the late dialysis group (4.4 ± 0.5 mL/min vs 6.4 ± 0.5 mL/min, P < 0.01). Smoothened Agonist mouse Survival at 6 months was significantly decreased (69% vs 87%, P < 0.01) and the risk of death was 2.77 times higher in the late dialysis group. In multivariate

analysis, the most significant predictors of poor outcome were age, intercurrent pulmonary infection and low serum albumin at the commencement of dialysis. In Ratcliffe et al.’s retrospective review of characteristics of all patients accepted for dialysis in the Oxford Unit in 1981, criteria for commencement of dialysis were uraemic symptoms associated with a creatinine clearance PD98059 less than 6 mL/min.31 Thirty-two patients were referred >1 month (early diagnosis Cetuximab group) and 23 patients were referred <1 month (late diagnosis group) before the commencement of dialysis. In the early referral group, 91% of patients commenced dialysis electively, 72% had a functioning fistula at the time of initiation of dialysis and 22% were commenced on continuous ambulatory peritoneal dialysis. Only two patients required initiation of dialysis via a CVC. In the late referral group, 39%

of patients commenced haemodialysis via a CVC. ‘Serious complications’, which significantly prolonged the length of stay in hospital, were significantly more frequent in the late diagnosis group (70% vs 9%, P < 0.001). Jungers et al. retrospectively reviewed records of 250 patients who commenced dialysis at the Necker Hospital between January 1988 and December 1990.32 The records of patients who required emergency dialysis and who had been referred within 4 weeks of commencing dialysis were identified. Of the total cohort, 25% were in this late referral category. From these patients, 20 records were randomly selected and compared with a control group of 20 age- and sex-matched patients who had been regularly followed up at the renal clinics for at least 6 months prior to the commencement of dialysis.