Supplementary Material for: Cystatin C-Creatinine Based Glomerular Filtration Rate Equation in Obese Chronic Kidney Disease Patients: Impact of Deindexation and Gender

<i>Background:</i> Cystatin C is considered an alternative to creatinine to estimate glomerular filtration rate (GFR). However, studies have reported that increased adiposity is associated with a higher level of circulating cystatin C questioning the performance of estimation of GFR using cystatin C in obese subjects. <i>Methods:</i> We prospectively included 166 obese stages 1-5 chronic kidney disease (CKD) patients between 2013 and 2015. GFR was measured with a reference method without (measured GFR [mGFR]) and with adjustment to body surface area (mGFRr) and estimated (eGFR) or de-indexed eGFR using the Chronic Kidney Disease and Epidemiology (CKD-EPI) equation using creatinine (CKD-EPIcreat), cystatin (CKD-EPIcyst) and the combination of cystatin and creatinine (CKD-EPIcyst-creat). <i>Results:</i> The biases between mGFR and de-indexed CKD-EPIcyst-creat were significantly lower than de-indexed CKD-EPIcreat (p = 0.001). Accuracies were significantly better with de-indexed CKD-EPIcyst-creat compared to CKD-EPIcreat and CKD-EPIcyst, respectively (p = 0.04 and 0.03). Bland and Altman plot showed a great dispersion of all formulae when patients had a GFR &gt;60 ml/min. Interestingly, there is a gender difference; biases, precisions and accuracies of de-indexed CKD-EPIcyst-creat were significantly lower in obese women. These results may be related to a difference in the change of body composition during obesity in men versus women and in fact only waist circumference (WC) was positively and significantly correlated with cystatin C (p &lt; 0.0001) whereas body mass index (BMI; p = 0.3) was not; bias for CKD-EPIcyst-creat was related with WC. <i>Conclusion:</i> Cystatin C-creatinine-based GFR equations outperform creatinine-based formula in obese CKD patients especially those with BMI ≥35 and in obese women.

<i>背景：</i> 胱抑素C（cystatin C）被视为估算肾小球滤过率（glomerular filtration rate, GFR）的肌酐替代标志物。然而已有研究显示，体脂升高与循环胱抑素C水平升高存在关联，这对肥胖人群中基于胱抑素C估算GFR的应用效能提出了质疑。
<i>方法：</i> 本研究于2013至2015年前瞻性纳入166例1~5期慢性肾脏病（chronic kidney disease, CKD）肥胖患者。采用参考方法检测肾小球滤过率，包括未校正体表面积的实测GFR（measured GFR, mGFR）、校正体表面积的实测GFR（mGFRr）；同时基于慢性肾脏病流行病学合作组（CKD-EPI）方程，分别通过肌酐、胱抑素C以及两者联合计算估算GFR（eGFR）与去体表面积校正的eGFR，即基于肌酐的CKD-EPI方程（CKD-EPIcreat）、基于胱抑素C的CKD-EPI方程（CKD-EPIcyst）以及胱抑素C与肌酐联合的CKD-EPI方程（CKD-EPIcyst-creat）。
<i>结果：</i> 去体表面积校正的CKD-EPIcyst-creat与mGFR之间的偏差显著低于去体表面积校正的CKD-EPIcreat（p=0.001）。去体表面积校正的CKD-EPIcyst-creat的准确率分别显著优于CKD-EPIcreat与CKD-EPIcyst（p=0.04和0.03）。Bland-Altman图显示，当患者GFR＞60ml/min时，所有公式的计算结果均存在较大离散度。值得注意的是，本研究存在性别差异：去体表面积校正的CKD-EPIcyst-creat的偏差、精密度与准确率在肥胖女性中均显著更低。上述结果或与男女肥胖患者的身体组成变化差异相关；事实上，仅腰围（waist circumference, WC）与胱抑素C水平呈显著正相关（p＜0.0001），而体重指数（body mass index, BMI, p=0.3）则无此关联；CKD-EPIcyst-creat的偏差与腰围相关。
<i>结论：</i> 基于胱抑素C-肌酐联合的GFR估算公式在肥胖CKD患者中，尤其是BMI≥35的患者以及肥胖女性中，其表现优于基于肌酐的估算公式。