Department of Endocrinology and Metabolism, Hospital Universitario Fundación Jiménez Díaz, Madrid, Spain.
Department of Gastroenterology, Liver Unit, Hospital Universitario La Paz, Madrid, Spain.
Obes Surg. 2020 Jul;30(7):2538-2546. doi: 10.1007/s11695-020-04509-0.
Nonalcoholic fatty liver disease (NAFLD) is the most common cause of chronic liver disease. It is a spectrum of progressive alterations, with the final step in liver fibrosis which carries a high burden of long-term mortality. The scores used to predict liver fibrosis are not properly validated in morbid obesity (MO). Our aim was to evaluate the performance of seven risk scores in bariatric surgery (BS) patients.
Cross-sectional analysis in a cohort of 60 patients with MO undergoing BS. Liver biopsy (LB) was taken and compared with fibrosis risk assessed by noninvasive scores: APRI, FIB-4, Forns, NFS (NAFLD fibrosis score), BARD, BAAT, and Hepamet. The area under the receiver operator characteristic curve (AUROC) and measures of diagnostic accuracy were calculated; performance of fibrosis scores was evaluated at standard threshold vs those suggested by ROC analysis.
LB was available in 50 patients; 9 (18%) had significant fibrosis (F2-F4). The BARD and Forns scores best predicted the absence of fibrosis, both with negative predictive value (NPV) of 95.5%, with AUROC of 0.761 and 0.667, respectively. Modification of standard thresholds (2 for BARD and 6.9 for Forns) to those suggested by ROC analysis (3 and 3.6, respectively) improved performance of scores. Basal glucose, glycated hemoglobin (HbA1c), aspartate transaminase (AST), and gamma glutamyl transferase (GGT) were identified by logistic regression analysis as independent predictor of fibrosis.
Existing scoring systems are unable to stratify fibrosis risk in MO using established thresholds; its performance is improved if these cutoffs are modified.
非酒精性脂肪性肝病（NAFLD）是慢性肝病最常见的原因。它是一系列进行性改变，最终步骤是肝纤维化，这给长期死亡率带来了很高的负担。用于预测肝纤维化的评分在病态肥胖（MO）中没有得到适当验证。我们的目的是评估七种风险评分在减肥手术（BS）患者中的表现。
对接受 BS 的 60 例 MO 患者队列进行横断面分析。进行肝活检（LB）并与非侵入性评分评估的纤维化风险进行比较：APRI、FIB-4、Forns、NFS（非酒精性脂肪性肝病纤维化评分）、BARD、BAAT 和 Hepamet。计算接收者操作特征曲线（AUROC）下的面积和诊断准确性的度量；在标准阈值和 ROC 分析建议的阈值下评估纤维化评分的性能。
50 例患者的 LB 可用；9 例（18%）有明显纤维化（F2-F4）。BARD 和 Forns 评分最能预测无纤维化，两者的阴性预测值（NPV）均为 95.5%，AUROC 分别为 0.761 和 0.667。将标准阈值（BARD 为 2，Forns 为 6.9）修改为 ROC 分析建议的阈值（分别为 3 和 3.6）可提高评分的性能。逻辑回归分析确定基础葡萄糖、糖化血红蛋白（HbA1c）、天冬氨酸转氨酶（AST）和γ谷氨酰转移酶（GGT）是纤维化的独立预测因子。
现有的评分系统无法使用既定阈值对 MO 中的纤维化风险进行分层；如果修改这些截止值，其性能会得到提高。
