Background: Measuring GFR using exogenous markers without a bladder catheter, errors can be easily made due to incomplete urine collection. The aim was to quantify agreement for 125I-iothalamate GFR measurements with and without a correction for inaccurate urine collection using 131I-hippuran.
Methods: The last available GFR measurement of adult patients was included. GFR was measured in two subsequent clearance periods with 125I-iothalamate by the standard method with correction for inaccurate urine collections using 131I-hippuran. The uncorrected and corrected GFR measurements were compared within and between the time periods for each individual patient. To study the agreement between both methods, intraclass correlation coefficients (ICC) were calculated and Bland-Altman plots, with accompanying accuracies and precisions, were used. Cohen's kappa was calculated to analyze the agreement of both methods for classifying patients according to the stages of chronic kidney disease (CKD).
Results: For the 332 stable included patients, the mean GFR of the uncorrected measurements was 77.8 ml/min (34.7) and the mean GFR of the corrected measurements was 81.0 ml/min (34.9). The ICC was 0.80 for the uncorrected measurements with an accuracy of 7.3 ml/min and a precision of 21.7 ml/min. For the corrected GFR measurements the ICC was 0.98, with an accuracy of 2.1 ml/min and a precision of 6.5 ml/min. Comparison between the methods showed an ICC of 0.95, an accuracy of 3.2 and a precision of 11.0. In total, 86% of the patients were classified similarly into CKD stages with both methods, overall Cohen's kappa was 0.81.
Conclusion: Agreement was better for GFR measurements corrected for inaccurate urine collections. Therefore, GFR measurements with 125I-iothalamate should be corrected for inaccurate urine collections using 131I-hippuran. Without such correction the GFR is easily underestimated which may lead to overtreatment.