How does one measure compliance in PD patients? What about the
measured/predicted creatinine ratio, and how useful is this to
identify non-compliant patients?
Answer
The issue of non-compliance in PD patients has recently received a
lot of attention. It has been speculated in the past that this
problem may be quite common in PD patients and that it may be a major
cause of inadequate dialysis in this group. Up to now, however,
non-compliance has been somewhat of a `black box' because there was no
real way of diagnosing it.
In 1993 Keen et al suggested that non-compliance could be diagnosed by
calculating a creatinine excretion ratio. This is a ratio of the
measured creatinine excretion to the predicted creatinine excretion.
The measured creatinine excretion is the sum of the dialysate and
urinary daily creatinine content, with an estimate of extra-renal
creatinine degradation based on the MacKenzie-Walser formula added in.
The predicted creatinine ratio is based on the classic Cockroft-Gault
formula. If the ratio of the measured:predicted is >1.24 this is
supposed to suggest that the patient has been playing `catch up' on
the day of the collection and has been washing out previously
accumulated creatinine from days when he was non-compliant. Thus, a
ratio higher than 1.24 is supposed to be suggestive of non-compliance.
A number of studies using this technique found that non-compliance was
quite common, involving somewhere between 10-30% of PD patients. More
recent work, however, suggests that the creatinine excretion ratio is
confounded by the patient's nutritional status and, specifically, by
the patient's lean body mass. Thus, patients may have a high
creatinine excretion ratio because they are non-compliant or,
alternately, because they have a higher lean body mass than the
Cockroft-Gault formula predicts. Similarly, a low creatinine
excretion ratio is found in patients with a below-average lean body
mass and may still be less than 1.00 in non-compliant patients; i.e.,
the low lean body mass decreases the ratio more than the
non-compliance increases it.
In addition, it appears that the Cockroft-Gault formulae may be
inappropriate for dialysis patients in that they do not take into
account extra-renal degradation which may be a major factor in these
patients. Cockroft-Gault formulae were initially derived from
patients with relatively mild renal failure where it might have been
reasonable to ignore the extra-renal creatinine degradation
contribution. It is because of this that most studies have shown the
average creatinine excretion ratio to be well above 1.00. This should
not be interpreted as meaning that PD patients have higher lean body
masses than normal; the contrary is the case.
What then can we do to get at this problem? One possibility is to
look at trends in the creatinine excretion ratio with time. In other
words, an increase in creatinine excretion ratio may be a more
significant finding than a single value that is high or low. An
increasing ratio would presumably either mean non-compliance or an
increase in lean body mass. If the latter had occurred one would
expect it to be associated with other features of improved
nutritional status. If the former was the case, however, nutritional
status might actually be expected to have stayed the same or got
worse. Similarly, a fall in creatinine excretion ratio might reflect
an improvement from previously poor non-compliance, but might also
reflect a fall in lean body mass. Obviously, the danger is that the
tendency of non-compliance to push up the ratio may be
counter-balanced by the tendency of decreased lean body mass to push
it down. Thus, a malnourished, non-compliant patient might not show a
significant difference in his creatinine excretion ratio over time.
Nevertheless, serial follow-up of this ratio may be useful both as an
index of nutrition and of compliance. It has to be interpreted in
association with nutritional data and also in association with careful
interviews with the patients. It still may be that the best technique
for detecting non-compliance is a simple question asked of the patient
in a non-threatening, non-judgemental manner.
Formulae to calculate the creatinine excretion ratio in SI Units are
to be found on page 148 of the March issue of Peritoneal Dialysis
International; or if traditional units are preferred, on page 142-143
of the same issue.
April, 1996
(Peter G. Blake, M.D., Victoria Hospital, London, Ontario)