| Literature DB >> 34937811 |
Yuka Kotake1, Shigehiro Karashima2, Masaki Kawakami3, Satoshi Hara4, Daisuke Aono5, Seigo Konishi5, Mitsuhiro Kometani5, Hiroyuki Mori6, Yoshiyu Takeda7, Takashi Yoneda5,8,9, Hidetaka Nambo3, Kenji Furukawa10,11.
Abstract
Diabetic kidney disease is an important and common cause of end-stage renal disease. Measurement of urinary albumin excretion (UAE) requires the diagnosis of the stage of diabetic nephropathy and the prognosis of renal function. We aimed to analyze the impact of lifestyle modification on UAE in patients with stage 2 and 3 type 2 diabetic nephropathy who received comprehensive medical care, using a generalized additive model (GAM), an explanatory machine learning model. In this retrospective observational study, we used changes in HbA1c, systolic blood pressure (SBP), and diastolic blood pressure (DBP) levels; body mass index (BMI); and daily salt intake as factors contributing to changes in UAE. In total, 269 patients with type 2 diabetic nephropathy were enrolled (stage 2, 217 patients; stage 3, 52 patients). The rankings that contributed to changes in UAE over 6 months by permutation importance were the changes in daily salt intake, HbA1c, SBP, DBP, and BMI. GAM, which predicts the change in UAE, showed that with increase in the changes in salt intake, SBP, and HbA1c, the delta UAE tended to increase. Salt intake was the most contributory factor for the changes in UAE, and daily salt intake was the best lifestyle factor to explain the changes in UAE. Strict control of salt intake may have beneficial effects on improving UAE in patients with stage 2 and 3 diabetic nephropathy.Entities:
Keywords: Explainable artificial intelligence; Generalized additive model; Salt intake; Type 2 diabetes; Urinary albumin excretion
Year: 2021 PMID: 34937811 DOI: 10.1507/endocrj.EJ21-0447
Source DB: PubMed Journal: Endocr J ISSN: 0918-8959 Impact factor: 2.349