Accurate estimation of kidney function in cirrhosis is crucial for prognosis and decisions regarding dual-organ transplantation. We performed a systematic review/meta-analysis to assess the performance of creatinine-based and cystatin C (CysC)–based eGFR equations compared with measured GFR (mGFR) in patients with cirrhosis. A total of 25 studies (n = 4565, 52.0 years, 37.0% women) comprising 18 equations met the inclusion criteria. In all GFR equations, the creatinine-based equations overestimated GFR (standardized mean difference, SMD, 0.51; 95% confidence interval [CI], 0.31-0.71) and CysC-based equations underestimated GFR (SMD, −0.3; 95% CI, −0.60 to −0.02). Equations based on both creatinine and CysC were the least biased (SMD, −0.14; 95% CI, −0.46 to 0.18). Chronic kidney disease–Epi–serum creatinine–CysC (CESC) was the least biased but had low precision and underestimated GFR by −3.6 mL/minute/1.73 m² (95% CI, −17.4 to 10.3). All equations significantly overestimated GFR (+21.7 mL/minute/1.73 m²; 95% CI, 17.7-25.7) at GFR <60 mL/minute/1.73 m²; of these, chronic kidney disease–Epi–CysC (10.3 mL/minute/1.73 m²; 95% CI, 2.1-18.4) and GFR Assessment in Liver Disease (12.6 mL/minute/1.73 m²; 95% CI, 7.2-18.0) were the least biased followed by Royal Free Hospital (15 mL/minute/1.73 m²; 95% CI, 5.5-24.6) and Modification of Diet in Renal Disease 6 (15.7 mL/minute/1.73 m²; 95% CI, 10.6-20.8); however, there was an overlap in the precision of estimates, and the studies were limited. In ascites, overestimation of GFR was common (+8.3 mL/minute/1.73 m²; 95% CI, −3.1 to 19.7). However, overestimation of GFR by 10 to 20 mL/minute/1.73m² is common in patients with cirrhosis with most equations in ascites and/or kidney dysfunction. A tailored approach is required especially for decisions regarding dual-organ transplantation.

中文翻译：

---

使用基于肌酐和胱抑素 C 的方程估计肝硬化肾小球滤过率：系统评价和荟萃分析

准确评估肝硬化肾功能对于预后和双器官移植决策至关重要。我们进行了一项系统回顾/荟萃分析，以评估基于肌酐和胱抑素 C (CysC) 的 eGFR 方程与肝硬化患者测量的 GFR (mGFR) 相比的性能。包含 18 个方程的总共 25 项研究（n = 4565，52.0 岁，37.0% 女性）符合纳入标准。在所有 GFR 方程中，基于肌酐的方程高估了 GFR（标准化平均差，SMD，0.51；95% 置信区间 [CI]，0.31-0.71），而基于 CysC 的方程低估了 GFR（SMD，-0.3；95% CI， -0.60 至 -0.02）。基于肌酐和 CysC 的方程式偏差最小（SMD，-0.14；95% CI，-0.46 至 0.18）。²（95% CI，-17.4 至 10.3）。在 GFR <60 mL/ min/1.73 m ²时，所有方程均显着高估了 GFR（+21.7 mL /minute/1.73 m ² ；95% CI，17.7-25.7） ；其中，慢性肾病–Epi–CysC（10.3 mL/minute/1.73 m ²；95% CI，2.1-18.4）和肝病 GFR 评估（12.6 mL/minute/1.73 m ²；95% CI，7.2- 18.0) 偏差最小，其次是皇家自由医院 (15 mL/minute/1.73 m ²；95% CI，5.5-24.6) 和肾脏疾病饮食改良 6 (15.7 mL/minute/1.73 m ²；95% CI , 10.6-20.8); 然而，估计的精确度存在重叠，研究也很有限。在腹水中，高估 GFR 很常见 (+8.3 mL/min/1.73 m ²; 95% CI，-3.1 至 19.7）。然而，GFR 高估 10 至 20 mL/min/1.73m ²在肝硬化患者中很常见，大多数等式为腹水和/或肾功能不全。需要量身定制的方法，尤其是对于有关双器官移植的决策。