Differential Effect of Hospitalization on Cystatin C- and Creatinine-Based Estimated GFR.

Abstract

Critical illness is known to acutely decrease eGFRdiff (estimation of GFR by cystatin C−estimation of GFR from serum creatinine [eGFRCr]), presumably because of loss of muscle mass falsely elevating eGFRCr. It is unknown whether this eGFRdiff effect occurs in all-cause hospitalizations and whether the effect persists after hospitalization. More hospitalization was associated with larger decreases in estimation of GFR by cystatin C compared with eGFRCr on measurements months after hospital discharge. Cystatin C has entered mainstream clinical care as a measure of kidney function, joining serum creatinine, which has been used for almost a century. However, many physicians notice that estimation of GFR from serum creatinine (eGFRCr) and estimation of GFR by cystatin C (eGFRCys) values can differ considerably. Hospitalization with critical illness is known to acutely decrease eGFRdiff (eGFRCys−eGFRCr). However, whether this effect occurs in all-cause hospitalizations and persists after hospitalization is unknown. Among 5599 adult participants in the Chronic Renal Insufficiency Cohort study with serum creatinine and cystatin C measurements, we estimated the association of six categories of total days of hospitalization between annual study visits (never hospitalized, hospitalized <7, 7 to <14, 14 to <28, 28 to <42, and ≥42 days) and changes in eGFRCr, eGFRCys, and eGFRdiff between those study visits. Compared with no hospitalization between study visits, increasing days of hospitalization were associated with decreases in eGFRCys (<i>e.g</i>., −3.30 [95% confidence interval, −5.48 to −1.13] ml/min per 1.73 m² for ≥42 days of hospitalization, test for trend <i>P</i> < 0.001), while eGFRCr remained relatively stable (<i>e.g</i>., −1.12 [−2.77 to 0.53] ml/min per 1.73 m² for ≥42 days of hospitalization, test for trend <i>P</i> = 0.21). The differential effect resulted in eGFRdiff becoming progressively more negative with more total days of hospitalization (test for trend <i>P</i> < 0.001). Prolonged or repeated hospitalization was associated with larger decreases in eGFRCys compared with eGFRCr on measurements months after hospital discharge.