HDCN Article Review/Hyperlink

Burkart J et al

An elevated ratio of measured to predicted creatinine production in CAPD patients is not a sensitive predictor of non-compliance with the dialysis prescription

Perit Dial Int (Mar) 16:142-146 1996

Also reviewed is Comparison of measured and predicted creatinine excretion is an unreliable index of compliance in PD patients Blake et al, Perit Dial Int 16:147-153, 1996 (same issue).

The two manuscripts in the March issue of PDI deal with the topic of non-compliance in PD patients and, in particular, with the use of creatinine excretion ratios to detect this important problem. Three years ago Keen et al suggested that non-compliance could be suspected in patients whose measured creatinine excretion was significantly higher than their predicted creatinine excretion. Measured creatinine excretion is a combination of dialysate and urinary creatinine losses combined with an estimate of extra-renal creatinine degradation based on the Mitch-Walser formula. Predicted creatinine excretion is based on the Cockroft-Gault formula. The creatinine excretion ratio is the measured divided by the predicted creatinine excretion. The theory is that in non-compliant patients creatinine accumulates when they are non-compliant but that on the day of the collection, when they are compliant, this accumulated creatinine is washed out. A creatinine excretion ratio of more than 1.24 has been proposed as a marker of non-compliance. Using this methodology, Keen et al, Warren et al and Nolph et al all suggested that non- compliance is quite common in CAPD patients, with a prevalence as high as 26%.

The two articles in the March 1996 PDI, however, raise questions about this methodology. Burkart et al show, in a study of nine patients on CAPD, that creatinine excretion does indeed increase significantly after a three day period of intentional non-compliance. They also demonstrate, however, that at baseline three of their patients have very low creatinine excretion ratios and four have high creatinine excretion ratios. In the three with low ratios, the value still remains below unity, even after the intentional non-compliance has taken place. Thus, these patients would not be picked up in a survey of non-compliance. Conversely, four of their patients had a baseline ratio that was >1.24. This suggests that these patients were either non- compliant before the study began or else that the ratio is unreliable.

The second article in the same edition, from our group, shows that in 43 patients the baseline ratio averaged 1.13 and that 40% had a ratio higher than the suggested cutoff of 1.24. Because these patients denied non-compliance, six of them were investigated more closely and had the ratio repeatedly measured during a four day period of guaranteed, observed compliance. The ratios did not change during that time, suggesting that these patients had high ratios due, not to non-compliance, but rather to higher creatinine excretions than the Cockroft-Gault formula predicted. In one case, where non-compliance was admitted, the ratio was only 1.09.

Comment: These findings suggest that the creatinine excretion ratio is neither a sensitive nor a specific indicator of non-compliance. It also suggests that existing literature based on this technique may not be reliable. We need a better technique to measure non-compliance. One possibility that is suggested in both papers, and in the accompanying editorial by Brandes, is that serial creatinine excretion measurements may be helpful. Thus, a rise in creatinine excretion ratio over time would suggest that the patient is either non-compliant or is increasing his lean body mass. If the high ratio is associated with other features of better nutrition, this might suggest lean body mass is increasing. If, in contrast, the higher ratio is associated with signs of poor nutrition, this might suggest that the problem is non-compliance. This might be the case even in cases where the absolute level of the creatinine excretion ratio was < 1.0; i.e., the increase might be, for example, from 0.7 to 0.85. Thus, there is still a case to be made for serial follow-ups of creatinine excretion ratios, but, as with many of such tests, clinical judgement will be needed to interpret them properly. (Peter G. Blake, M.D., Victoria Hospital, London, Ontario)