Interdepartment of Critical Care Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro 173 Beon-gil, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-707, Korea.
Department of Anesthesiology and Pain Medicine, Seoul National University Bundang Hospital, 82, Gumi-ro 173 Beon-gil, Bundang-gu, Seongnam-si, Gyeonggi-do, 463-707, Korea.
Sci Rep. 2018 Mar 15;8(1):4614. doi: 10.1038/s41598-018-22886-6.
Serum albumin and conventional BMI (cBMI) are commonly used indices of malnutrition status. Because cBMI does not reflect fluid accumulation, modified body mass index (mBMI, serum albumin × cBMI) is a more accurate measure of malnutrition status. This study investigated the association between preoperative mBMI and postoperative 1-year mortality, in comparison with serum albumin and cBMI. Medical records of 80,969 adult patients who underwent surgical procedures in a tertiary care hospital between 1 January, 2011 and 31 December, 2015 were retrospectively reviewed. Postoperative 1-year mortality increased with reduction in cBMI, mBMI, and albumin separately (P < 0.001). When considering interaction between cBMI and albumin, albumin was the only significant factor of postoperative 1-year mortality [odds ratio: 0.377, 95% confidence interval (0.245-0.579), P < 0.001], while cBMI or interaction (cBMI * albumin) was not significant (P > 0.05). Adjusted area under the curve (AUC) was highest (0.885) in the overall model (cBMI + albumin + cBMI * albumin); adjusted AUC between mBMI and albumin did not differ significantly (P = 0.558). Low albumin is the strongest independent predictor of postoperative 1-year mortality among the three variables (albumin, cBMI, mBMI). Adding cBMI to albumin does not increase the validity of the AUC of albumin.
血清白蛋白和常规 BMI（cBMI）通常被用作营养不良状况的指标。由于 cBMI 不能反映液体蓄积，因此改良 BMI（mBMI，血清白蛋白×cBMI）是衡量营养不良状况的更准确指标。本研究比较了术前 mBMI 与术后 1 年死亡率的关系，并与血清白蛋白和 cBMI 进行了比较。回顾性分析了 2011 年 1 月 1 日至 2015 年 12 月 31 日期间在一家三级保健医院接受手术的 80969 名成年患者的病历。术后 1 年死亡率随着 cBMI、mBMI 和白蛋白的降低而增加（P<0.001）。当考虑 cBMI 和白蛋白之间的交互作用时，白蛋白是术后 1 年死亡率的唯一显著因素[比值比：0.377，95%置信区间（0.245-0.579），P<0.001]，而 cBMI 或交互作用（cBMI白蛋白）则不显著（P>0.05）。总体模型（cBMI+白蛋白+cBMI白蛋白）的校正曲线下面积（AUC）最高（0.885）；mBMI 和白蛋白之间的校正 AUC 差异无统计学意义（P=0.558）。在这三个变量（白蛋白、cBMI 和 mBMI）中，低白蛋白是术后 1 年死亡率的最强独立预测因子。在白蛋白中添加 cBMI 不会增加 AUC 的有效性。
