Estimation of Total Creatinine Clearance is Unreliable in Children on Peritoneal Dialysis

Objective

To test the reliability of creatinine clearance in children on peritoneal dialysis (PD).

Design

Longitudinal, case-controlled.

Setting

Routine clinic visits at the pediatric dialysis unit of the Universitätskinderklinik of Vienna. Patients: Eleven children (2 -13 years, 10 -55 kg) with end-stage renal disease on PD.

Interventions

Creatinine clearance (CCr) was determined by measuring creatinine excretion (ECr) over 24 hours in both dialysate and urine. Each child had three to five separate measurements of their CCr. At the same time we also calculated the schwartz formula clearance from the patient's height and serum creatinine, using a modified correlate.

Main Outcome Measures

Reliability of CCr was assessed by two approaches. First, we compared each serial measurement with the mean value for each patient and thereby assessed the “intramethodical” variability. Second, we compared each CCr with the simultaneous formula clearance and assessed the “intermethodical” disagreement.

Results

Twenty-seven percent of the measurements of CCr were classified as unreliable based on a comparison with the mean value for each patient. Reliability was closely correlated with residual renal function (p < 0.01); only 12% of the measurements in the an uric patients were classified as unreliable (vs 31% in the patients with residual renal function). The simultaneous formula clearance was less variable than the CCr. The formula clearance had a sensitivity of 93% and a specificity of 60% for detecting unreliable values of CCr.

Conclusion

Estimation of total CCr is unreliable in pediatric patients on PD. A simultaneous formula clearance can be used to detect which values are unreliable.

Comparison of Measured and Predicted Creatinine Excretion is An Unreliable Index of Compliance in Pd Patients

Objective

To evaluate the use of the ratio of measured to predicted creatinine excretion as an index of compliance in peritoneal dialysis (PD) patients.

Design

A prospective analysis.

Setting

Academic teaching hospital dialysis unit.

Patients

Forty-three patients on PD.

Measurements

Creatinine excretion in daily dialysate and urine collections was measured on one occasion in 10 patients and on two occasions in 33 patients, and, after adding an estimate for extrarenal creatinine degradation, was divided by predicted creatinine excretion to give a creatinine excretion ratio, which has been proposed as an index of compliance with exchanges in PD patients. Values above 1.24 have been suggested to indicate non-compliance. Lean body mass was also estimated from creatinine excretion.

Results

The mean creatinine excretion ratio was 1.12, and 30% of patients had a value above 1.3. Only one patient admitted noncompliance. Studies on four consecutive days of guaranteed compliance in 7 patients with high ratios showed that creatinine excretion remained constant, suggesting that the patients were high creatinine producers rather than noncompliant. Creatinine excretion was stable when measured at intervals of days, but over months it tended to change markedly in many patients. Lean body mass estimations using creatinine excretion were low in most patients.

Conclusion

Comparison of measured and predicted creatinine excretion is not a reliable indicator of noncompliance because many compliant patients consistently excrete more creatinine than predicted. The standard formulas were not validated in dialysis patients and underestimate creatinine excretion significantly in many PD patients. Existing estimates in the literature of non-compliance, using this methodology, may not be accurate. Better methods of detecting this problem are required.

Sources of variation in estimates of lean body mass by creatinine kinetics and by methods based on body water or body mass index in patients on continuous peritoneal dialysis

OBJECTIVE

We identified factors that account for differences between lean body mass computed from creatinine kinetics (LBM(cr)) and from either body water (LBM(V)) or body mass index (LBM(BMI)) in patients on continuous peritoneal dialysis (CPD).

DESIGN

We compared the LBM(cr) and LBM(V) or LBM(BMI) in hypothetical subjects and actual CPD patients.

PATIENTS

We studied 439 CPD patients in Albuquerque, Pittsburgh, and Toronto, with 925 clearance studies.

INTERVENTION

Creatinine production was estimated using formulas derived in CPD patients. Body water (V) was estimated from anthropometric formulas. We calculated LBM(BMI) from a formula that estimates body composition based on body mass index. In hypothetical subjects, LBM values were calculated by varying the determinants of body composition (gender, diabetic status, age, weight, and height) one at a time, while the other determinants were kept constant. In actual CPD patients, multiple linear regression and logistic regression were used to identify factors associated with differences in the estimates of LBM (LBM(cr)<LBM(V), or LBM(cr)<LBM(BMI)).

MAIN OUTCOME MEASURE

We sought predictors of the differences LBM(V) - LBM(cr) and LBM(BMI) - LBM(cr).

RESULTS

Both LBM(V) (regardless of formula used to estimate V) and LBM(BMI) exceeded LBM(cr) in hypothetical subjects with average body compositions. The sources of differences between LBM estimates in this group involved differences in the coefficients assigned to gender, age, height, weight, presence or absence of diabetes, and serum creatinine concentration. In CPD patients, mean LBM(V) or LBM(BMI) exceeded mean LBM(cr) by 6.2 to 6.9 kg. For example, the LBM(V) obtained from one anthropometric formula was 50.4+/-10.4 kg and the LBM(cr) was 44.1+/-13.6 kg (P < .001), whereas among the 925 clearance studies, only 216 (23.3%) had LBM(cr)>LBM(V). The differences in determinants of body composition between groups with high versus low LBM(cr) were similar in hypothetical and actual CPD patients. Multivariate analysis in actual CPD patients identified serum creatinine, height, age, gender, weight, and body mass index as predictors of the differences LBM(V)-LBM(cr) and LBM(BMI)-LBM(cr).

CONCLUSIONS

Overhydration is not the sole factor accounting for the differences between LBM(cr) and either LBM(V) or LBM(BMI) in CPD patients. These differences also stem from the coefficients assigned to major determinants of body composition by the formulas estimating LBM.