Abstract
1242
Objectives Glomerular filtration rate (GFR) is an important element of renal function in patients with acute and chronic kidney disease. In the US, gold standard methods for measuring GFR, inulin and 125-I iothalamate clearance, are rarely used because of technical complexity and cost. Serum creatinine levels are commonly used to estimate GFR but are fraught with inaccuracy in the elderly and patients with reduced muscle mass, in whom serum Cr levels are proportionately low. Plasma clearance of Tc-99m (diethylene triamene pentaacetic acid) DTPA has been validated as equivalent to I-125 iothalamate clearance for measuring GFR in patients with mild to moderate renal insufficiency and has virtually become the gold standard for measuring GFR in the US. However, this method is somewhat inconvenient, requiring two venous access lines, multiple blood samples and hours to completion. Careful attention to laboratory technique is also required. In recent years, improved estimating equations have been developed for measuring GFR from serum markers, both creatinine (Cr) and/or cystatin-C (CysC). These include MDRD, CKD-EPI (Cr), CKD-EPI (CysC) and CKD-EPI (CysC + Cr). However, these improved equations have not been well studied in subgroups of patients with specific diseases. The degree to which these estimating equations compare to Tc-99m DTPA GFR measurements is unknown, particularly in patients with conditions that result in reduced muscle mass. The purpose of this study was to compare the GFR generated by these improved estimating equations to plasma clearance of Tc-99m DTPA in patients with hepatic cirrhosis, who display reduced muscle mass (sarcopenia).
Methods 59 adult patients with hepatic cirrhosis were enrolled under an IRB-approved protocol. All subjects underwent Tc-99m DPTA-based measurements of GFR from two protein-free (ultra-filtered) serum samples obtained at 1 and 3 hours following the intravenous administration of 5.0 mCi (185 MBq) Tc-99m DTPA. This was compared to the GFR generated by 4 prediction formulas that utilize serum creatinine (Cr) and/or cystatin-C (CysC) for estimating GFR: MDRD (from the Modified Diet in Renal Disease Study), and 3 formulas developed by the Chronic Kidney Disease-Epidemiology Collaboration (CKD-EPI), including CKD-EPI (Cr), CKD-EPI (CysC) and CKD-EPI (CysC+Cr). Paired data (estimating equations vs Tc-99m DTPA method) were analyzed by paired t-test, Pearson’s correlation, Passing and Bablock fit, and the Altman Bland agreement test.
Results Based on the Tc-99m DTPA plasma clearance method, the mean GFR among the 59 subjects examined was 76.8 ml/min/m2 (range 38 - 149). There was a positive correlation between the GFR measurements derived from the Tc-99m DTPA method with those of each of the CKD-EPI estimating formulas as well as the MDRD formula. Differences between the Tc-99m DTPA and estimating formula GFR values were statistically significant (p < .01) for all formulas except for the MDRD (p = 0.078). However, when compared to the Tc-99m DTPA method, the MDRD resulted in a standard deviation of differences of 17.3 (95% CI: -29.99 to 38.10) with a fit bias of +15.94 (95% CI -2.32 to 28.91) ml/min/m2. The presence or absence of ascites did not change the degree of agreement between the GFR’s generated by the estimating equations and that of the Tc-99m DTPA method. Equations that utilized CysC showed no better performance than those that utilized serum Cr.
Conclusions The MDRD estimating formula showed better agreement when compared to the Tc-99m DTPA method than any of the CKD-EPI equations. However, all estimating equations were judged inadequate to replace the Tc-99m DTPA method in this group of subjects. Caution is recommended in using the estimating equations for GFR measurements in patients likely to have reduced muscle mass. Whether the MDRD or CKD-EPI estimating equations might perform better in other patient populations is unknown. RESEARCH SUPPORT: NIH R01DK088375-03