Comparability of Plasma Iohexol Clearance Across Population-Based Cohorts.

RATIONALE & OBJECTIVE: Glomerular filtration rate (GFR) estimation based on creatinine or cystatin C level is currently the standard method for assessing GFR in epidemiologic research and clinical trials despite several important and well-known limitations. Plasma iohexol clearance has been proposed as an inexpensive method for measuring GFR that could replace estimated GFR in many research projects. However, lack of standardization for iohexol assays and the use of different protocols such as single- and multiple-sample methods could potentially hamper comparisons across studies. We compared iohexol assays and GFR measurement protocols in 3 population-based European cohorts.

STUDY DESIGN: Cross-sectional investigation.

SETTING & PARTICIPANTS: Participants in the Age, Gene/Environment Susceptibility-Kidney Study (AGES-Kidney; n=805), the Berlin Initiative Study (BIS, n=570), and the Renal Iohexol Clearance Survey Follow-up Study (RENIS-FU; n=1,324).

TESTS COMPARED: High-performance liquid chromatography analyses of iohexol. Plasma iohexol clearance calculated using single- versus multiple-sample protocols.

OUTCOMES: Measures of agreement between methods.

RESULTS: Frozen samples from the 3 studies were obtained and iohexol concentrations were remeasured in the laboratory at the University Hospital of North Norway. Lin's concordance correlation coefficient ρ was>0.96 and Cb (accuracy) was>0.99 for remeasured versus original serum iohexol concentrations in all 3 cohorts, and Passing-Bablok regression did not find differences between measurements, except for a slope of 1.025 (95% CI, 1.006-1.046) for the log-transformed AGES-Kidney measurements. The multiple-sample iohexol clearance measurements in AGES-Kidney and BIS were compared with single-sample GFRs derived from the same iohexol measurements. Mean bias for multiple-sample relative to single-sample GFRs in AGES-Kidney and BIS were-0.25 and-0.15mL/min, and 99% and 97% of absolute differences were within 10% of the multiple-sample result, respectively.

LIMITATIONS: Lack of comparison with an independent gold-standard method.

CONCLUSIONS: Agreement between the iohexol assays and clearance protocols in the 3 investigated cohorts was substantial. Our findings indicate that plasma iohexol clearance measurements can be compared across these studies.