Recent small subcortical infarct (RSSI) is one of the manifestations of lacunar infarction. It is a common brain disease and can lead to the clinical outcome of disability or dementia in many patients. However, the relationship of infarction burden and cognitive impairment with blood glucose fluctuation in type 2 diabetes mellitus (T2DM) patients with RSSI is not very clear.

To explore the correlation of blood glucose variability (GV) with infarction burden and cognitive impairment in T2DM patients with RSSI, and based on this, to build a risk prediction model.

A total of 140 patients with T2DM and RSSI who were treated in the Second Affiliated Hospital of Zhengzhou University from January 2021 to June 2022 were retrospectively selected. The basic clinical data of the patients were collected. The 72-hour continuous blood glucose monitoring was performed. The infarct burden was evaluated by the magnetic resonance imaging performance (the study subjects were divided into the high infarction burden group including 45 cases and the low infarction burden group including 95 cases according to the imaging performance). The cognitive function was evaluated by the Montreal Cognitive Assessment (MoCA). Spearman correlation analysis was used to explore the correlation between GV and cognitive function (MoCA score). Multivariate Logistic regression analysis was used to explore the influencing factors of infarction burden and cognitive dysfunction in T2DM patients with RSSI. The receiver operating characteristic (ROC) curve was drawn to evaluate the predictive value of GV on cognitive impairment in T2DM patients with RSSI, and the nomogram predictive model was constructed and the predictive value was analyzed.

In terms of GV-related indicators, the high infarction burden group had higher standard deviation (SD) and percentage of coefficient of variation (%CV) and lower time in range (TIR) than the low infarction burden group, with statistically significant differences (P<0.05). The results of Spearman correlation analysis showed that SD (r_s=0.272, P=0.001) and %CV (r_s=0.391, P<0.001) were directly proportional to MoCA score, and TIR (r_s=-0.325, P<0.001) was inversely proportional to the MoCA score in T2DM patients with RSSI. The results of multivariate Logistic regression analysis showed that elevated SD〔OR=4.201, 95%CI (1.380, 12.788), P=0.011〕 and %CV〔OR=1.218, 95%CI (1.096, 1.354), P<0.001〕were risk factors for high infarction burden in patients with T2DM and RSSI, while increased TIR〔OR=0.866, 95%CI (0.814, 0.921), P<0.001〕 was a protective factor. Elevated SD〔OR=2.947, 95%CI (1.150, 7.548), P=0.024〕 and %CV〔OR=1.174, 95%CI (1.072, 1.287), P=0.001〕were risk factors for cognitive impairment, while elevated TIR〔OR=0.954, 95%CI (0.917, 0.992), P=0.018〕 was a protective factor in T2DM patients with RSSI. The area under the curve (AUC) of %CV for predicting cognitive impairment in patients with T2DM and RSSI was 0.758〔95%CI (0.660, 0.856), P<0.001〕, with an optimal cut-off value of 29.5%, 66.7% sensitivity and 76.0% specificity. The AUC of TIR in predicting cognitive impairment in T2DM patients with RSSI was 0.714〔95%CI (0.624, 0.804), P<0.001〕, with an optimal cut-off value of 60.5%, 97.2% sensitivity and 44.2% specificity. The nomogram prediction model based on SD, %CV, and TIR for the risk of cognitive impairment in T2DM patients with RSSI demonstrated great clinical benefits, and the internal correction suggested that the actual prediction results were similar to the ideal prediction results.

Elevated GV indicators such as SD and %CV may be independent risk factors, and increased TIR may be a protective factor for high infarct burden and cognitive dysfunction in T2DM patients with RSSI. %CV and TIR had good predictive value for cognitive dysfunction in T2DM patients with RSSI.

近期皮质下小梗死（RSSI）是腔隙性脑梗死的表现形式之一，是一种常见脑部疾病，可以使许多患者走向残疾或痴呆的临床结局。但是2型糖尿病（T2DM）合并RSSI患者的梗死负担及认知功能障碍与血糖波动之间关系尚未明确。

探索T2DM合并RSSI患者血糖变异性（GV）与梗死负担、认知功能障碍之间的相关性并构建风险预测模型。

回顾性选取2021年1月至2022年6月就诊于郑州大学第二附属医院的T2DM合并RSSI患者140例为研究对象。收集患者临床基本资料并进行72 h动态血糖检测，使用磁共振影像表现评估梗死负担（根据影像学表现将研究对象分为高梗死负担组45例和低梗死负担组95例）；使用蒙特利尔认知评估量表（MoCA）评估认知功能。采用Spearman秩相关分析探讨T2DM合并RSSI患者GV与认知功能（MoCA评分）的相关性。采用多因素Logistic回归分析探讨T2DM合并RSSI患者梗死负担及认知功能障碍的影响因素。绘制受试者工作特征（ROC）曲线评价GV对T2DM合并RSSI患者发生认知功能障碍的预测价值，构建列线图预测模型并分析预测价值。

高梗死负担组患者GV相关指标中血糖浓度标准差（SD）、变异系数百分比（%CV）高于低梗死负担组，葡萄糖目标范围内时间（TIR）则低于低梗死负担组（P<0.05）。Spearman秩相关分析结果显示，T2DM合并RSSI患者SD（r_s=0.272，P=0.001）、%CV（r_s=0.391，P<0.001）与MoCA评分呈正相关，TIR（r_s=-0.325，P<0.001）与MoCA评分呈负相关。多因素Logistic回归分析结果表示，SD〔OR=4.201，95%CI（1.380，12.788），P=0.011〕、%CV〔OR=1.218，95%CI（1.096，1.354），P<0.001〕升高是T2DM合并RSSI患者高梗死负担的危险因素；TIR〔OR=0.866，95%CI（0.814，0.921），P<0.001〕升高则是保护因素；SD〔OR=2.947，95%CI（1.150，7.548），P=0.024〕、%CV〔OR=1.174，95%CI（1.072，1.287），P=0.001〕升高是T2DM合并RSSI患者认知功能障碍的危险因素，TIR〔OR=0.954，95%CI（0.917，0.992），P=0.018〕升高则是保护因素。%CV预测T2DM合并RSSI患者发生认知功能障碍的ROC曲线下面积（AUC）为0.758〔95%CI（0.660，0.856），P<0.001〕，最佳截断值为29.5%，灵敏度和特异度分别为66.7%和76.0%；TIR预测T2DM合并RSSI患者发生认知功能障碍的AUC为0.714〔95%CI（0.624，0.804），P<0.001〕，最佳截断值为60.5%，灵敏度和特异度分别为97.2%和44.2%。基于SD、%CV、TIR建立T2DM合并RSSI患者发生认知功能障碍风险的列线图预测模型有较大临床获益，进行内部矫正提示，实际预测结果与理想预测结果相近。

GV指标中SD、%CV升高可能是T2DM合并RSSI患者高梗死负担以及认知功能障碍的独立危险因素；TIR升高可能是T2DM合并RSSI患者高梗死负担以及认知功能障碍的保护因素。%CV和TIR对T2DM合并RSSI患者发生认知功能障碍有较好的预测价值。