The older patient group had higher 1-year mortality (31% vs 19%). Late referral was associated with greater mortality in both groups (34% vs 9% in the younger group and 42% vs 16% in the older group). The RR for death in the older group was 1.80 and 2.2 in the younger group. Because of the higher frequency of late referral in older patients this accounted for a large proportion of excess mortality. Stoves et al. retrospectively studied all 1260 patients who received dialysis from 1980 to July 1999 at St James Hospital in Leeds.69 Group A commenced dialysis <90 days after referral and group B >90 days. Survival at 4 months beta-catenin assay was 87% in group A and 94% in group

B with survival at 1 year being 74% versus 87% and survival at 5 years being 31% versus 55%. Fewer group A patients were listed for transplantation. By multifactorial analysis, age, diabetes, serum albumin, transplant listing and time of referral were significant predictors of survival. Wasse et al. used Medicare and Medicaid data from 5042 US dialysis patients to analyse reasons for persistent use of CVC 90 days after dialysis initiation.70 At 90 days, 59.4% were still using a CVC, 25.4% an AV graft and only 15.2% a fistula. Age, sex, race and cardiovascular comorbidity were associated with persistence of catheter use. The authors suggested that this could be due to late access referral or primary access

failure. JNK inhibitor screening library White et al. looked at another aspect of timely referral – whether or not allowing participation in a predialysis clinic could improve quality of life.71 A total of 74 patients attended a predialysis multidisciplinary clinic and 46 did not. The former showed improvement in 4 of 8 physical Quality of Life scores at 6 months after start of dialysis, even when adjusted for comorbidities and other variables. Winkelmayer et al. defined late referral as less than 90 days prior to starting dialysis.72 Medicare and Medicaid

data identified all adult patients in New Jersey who commenced dialysis between 1990 and mid-1996 (3014 patients). Late referral was associated with old age, race, lack of comorbidity and management by a general internist rather than a primary care doctor or other subspecialist. Winkelmayer et al. also looked at potential associations between late referral and choice of of dialysis modality.73 Late referral was defined as less than 90 days before first dialysis. Timing of referral did not influence the initial dialysis modality; however, late referral patients commencing predialysis were more likely to switch to haemodialysis than early referred patients (HR 1.47). Winkelmayer et al. performed a propensity analysis of late versus early nephrologist referral and dialysis mortality.74 Late referral was again defined as less than 90 days before initiation of dialysis. There was a 36% excess mortality in late referrals which was, however, limited to the first 3 months (HR 1.75, 95% CI: 1.48–2.