The statistical relationship between solute clearance (expressed as Kt/V) and dietary protein intake (DPI; estimated from the normalized protein catabolic rate [NPCR]) has led to the assumption that dialysis dose is one of the most important determinants of DPI in continuous ambulatory peritoneal dialysis patients. However, there is increasing concern about the clinical value of such a correlation between variables that contain common components. We have investigated the statistical nature of the cross-sectional correlation between Kt/V and NPCR using mathematical modelling. In addition, in a prospective study, the impact of increasing Kt/V on protein intake estimated from the NPCR and from 3-day food diaries (DPI) was evaluated in 59 patients over a 1-year period. Two thousand sets of random numbers were generated for the components, from which Kt/V and NPCR were calculated. The subsequent cross-sectional correlation between the randomly generated Kt/V and NPCR was highly significant (r = 0.64 to 0.74). In contrast, there was a much weaker cross-sectional correlation between relatively independent measures of dialysis dose (urea clearance) and protein intake measured from food diaries (r = 0.36). Analysis of the prospective relationship revealed a significant correlation with a decrease in urea clearance (due to loss of renal function) accompanied by a decrease in both PCR (r = -0.80) and DPI (r = -0.51). In contrast, when Kt/V was increased (by larger volume exchanges) there was no significant increase in protein intake. This study has demonstrated that the strength of the cross-sectional relationship between Kt/V and NPCR is due to mathematical coupling of "like with like," and this correlation cannot be used to substantiate a link between dialysis dose and nutrition. Prospective analysis does show a dependence of protein intake on urea clearance. However, this dependence is only seen in those patients undergoing a reduction in clearance due to loss of renal function. Increasing exchange volume to offset such losses was not accompanied by improvements in protein intake.