Plasma Iohexol Clearance for Assessing Residual Kidney Function in Dialysis Patients

To the editor

Residual kidney function in dialysis patients is strongly associated with improved outcomes(1) but is difficult to assess.(2) The average of urinary urea and creatinine clearance (CL_urea,cr) from carefully performed timed urine collections between dialysis sessions can reliably estimate GFR measured by urinary inulin clearance (the gold standard),(3) but timed urine collections are cumbersome and error prone. Alternative GFR measurement methods include determining plasma clearance of exogenous filtration markers such as iohexol, an 821-Dalton iodinated contrast agent. Plasma and urinary iohexol clearance (CL_iohexol) is highly correlated with urinary inulin clearance and sampling for plasma clearance calculations can be minimized to 2 time points.(4) These characteristics make determining plasma CL_iohexol attractive for GFR measurement. The goal of our study was to assess the accuracy of plasma CL_iohexol versus CL_urea,cr in dialysis patients.

Between November 2011 and October 2014, we recruited 40 dialysis patients, 18 years or older, with self-reported ability to produce 1 cup of urine. We simultaneously measured plasma and urinary CL_iohexol along with urinary CL_urea,cr during a 24-hour inpatient study visit; research nurses closely monitored urine collection. We included a late (24 hour) time point in our calculation of plasma CL_iohexol by a 2-compartment model (plasma sampling times: 10, 30 min; 2, 4, 24 h). In 10 participants, we repeated these measurements at a follow-up visit, resulting in 50 (40 baseline + 10 repeats) clearance measurements. We considered CL_urea,cr the reference test and plasma and urinary CL_iohexol as the index tests, and assessed correlation and median bias between these measures (detailed methods in Item S1).

Mean patient age was 55 years, 65% were male, 78% were African American, and 90% were on hemodialysis (Table 1). All clearance measurements are reported in mL/min/1.73 m², and median [IQR] values were as follows: plasma CL_iohexol, 7.1 [5.3–9.4] at 24 hours, 12.7 [7.8–18.0] at 4 hours; 24-hour urinary CL_iohexol, 2.3 [0.5–4.3]; and 24-hour urinary CL_urea,cr, 3.2 [1.6–5.4]. Plasma CL_iohexol at 24 hours was significantly lower than at 4 hours (bias, −4.6 [−5.9 to −3.2]; p < 0.001; Item S1) but significantly higher than 24-hour urinary CL_iohexol (bias, 4.4 [3.6–5.2]; p < 0.001) and 24-hour urinary CL_urea,cr (bias, 3.7 [3.3– 4.1]; p < 0.001; Figure 1). Also, 24-hour urinary CL_urea,cr was higher than 24-hour urinary CL_iohexol (bias, 0.5 [0.3–0.7]; p < 0.001) but the overestimation was smaller compared with 24-hour plasma CL_iohexol. During follow-up, the correlation and median change in repeat clearance measures was highest for urinary clearances (Item S1). Clearance calculations using a one-compartment model (4- and 24-hour iohexol samples) yielded results similar to the primary analysis (Item S1).

Table 1.

Characteristics of Participants and Clearance Measurements in the Residual Kidney Function Study

Characteristics	Distribution
Demographics
Age, y	56 [49–64]
Male	26 (65%)
African American	31 (78%)
Clinical Characteristics
Cause of ESRD
Diabetes	11 (29%)
Glomerulonephritis	5 (14%)
Other	24 (60%)
Diabetes	21 (53%)
Hemodialysis	36 (90%)
Peritoneal Dialysis	4 (10%)
Dialysis vintage, mo	17 [9–36]
Height, cm	173 [164–181]
Weight, kg	88.3 [74.7–115.2]
Body Mass Index, Kg/m2	30.5 [26.1–35.1]
Body Surface Area, m2	2.0 [1.9–2.3]
24-Hour Urine Volume, ml	783 [370–1154]
Measured Clearances, ml/min/1.73 m2
Urinary CLcr	4.2 [1.9–6.8]
Urinary CLurea	2.6 [0.9–3.9]
Urinary CLurea,cr	3.2 [1.6–5.4]
Urinary CLiohexol*	2.3 [0.5–4.3]
Weekly standard renal Kt/Vurea**	0.6 [0.2–1.1]
4-h plasma CLiohexol***	12.7 [7.8–18.0]
24-h plasma CLiohexol	7.1 [5.3–9.4]
Endogenous GFR Markers (Time-Averaged)
Serum Urea Nitrogen, mg/dL	43 [36–53]
Creatinine, mg/dL	6.9 [5.3–10.2]

Note: N = 40. Values given as median [IQR] or number (percentage). Conversion factors for units: creatinine in mg/dL to µmol/L, ×88.4; Serum urea nitrogen in mg/dL to mmol/L, ×0.357

^*Excludes 3 patients where time from start of urine collection to iohexol injection exceeded 90th percentile (71 min; median was 6 [3–35] min).

^**Weekly urea clearance by residual kidney function divided by anthropometrically calculated volume of distribution of urea.

^***Excludes 1 patient in whom this could not be calculated (the extrapolated iohexol concentrations at 10 and 30 min were higher than measured concentrations).

Figure 1. Comparison of Clearance Measurements in the Residual Kidney Function Study.

A Scatterplot of 24-hour plasma (red dots) or urinary (blue dots) CL_iohexol vs 24-hour urinary CL_urea,cr. The dashed lines and shaded areas represent the linear fit and confidence interval of the regression, respectively, of either 24-hour plasma (red) or urinary (blue) CL_iohexol on 24-hour urinary CL_urea,cr.

B Plot of 24-hour CL_iohexol (plasma, red; urinary, blue) less 24-hour urinary CL_urea,cr vs 24-hour urinary CL_urea,cr. Positive numbers represent overestimation of 24-hour urinary CL_urea,cr; negative numbers, underestimation. The dots represent the difference between the clearance measures and their relation to 24-hour urinary CL_urea,cr. The red and blue lines represent fit from a median quantile regression of the bias on 24-hour urinary CL_urea,cr.

In our study, plasma CL_iohexol was significantly lower at 24 versus 4 hours, a finding consistent with prior observations.(5) At 24 hours, plasma CL_iohexol was significantly higher than urinary CL_iohexol and urinary CL_urea,cr. Possible explanations include hepatic clearance of iohexol, lack of steady-state plasma iohexol concentrations after bolus injection, tubular reabsorption of iohexol, or systematic error in urinary CL_urea,cr. In previous animal studies and smaller human studies, nonrenal CL_iohexol was noted to be ~2–4 mL/min/1.73 m²,(6,7) with high variability reported in one study.(8) Proximal tubular reabsorption of filtered iohexol has been described in animal models and may explain higher plasma versus urinary CL_iohexol.(9) Measurement differences between plasma and urinary CL_iohexol could also be due to a matrix effect (using an iohexol assay calibrated in plasma for urine samples).(10)

Study limitations include possible errors in urine collection (lack of bladder catheterization and assessment of urinary retention by bladder ultrasound); absence of measured urinary inulin clearance as a gold-standard reference test, hepatic iohexol clearance, and direct measurements of extracellular volume; and lesser standardization of iohexol in urine than plasma. Strengths include sampling of iohexol at 24 hours and carefully supervised urine collection. In published studies, carefully supervised urinary CL_urea,cr is highly reliable in estimating GFR measured by urinary inulin clearance.(3)

In summary, our findings suggest that in dialysis patients with residual kidney function, plasma CL_iohexol overestimates urinary CL_iohexol and urinary CL_urea,cr. Consequently, adjusting dialysis dose based on plasma CL_iohexol rather than urine collection may result in dialysis underdosing. Further research is needed to understand the mechanisms and variability in renal handling and extrarenal elimination of iohexol and optimal methods to standardize its measurement. A carefully designed study with simultaneous plasma and urinary measurements of iohexol, inulin, and other markers at various levels of GFR is needed to answer these questions.