血壓變化趨勢之模型分析

Blood Pressure Variation Trend Analysis Based on Model Study

丁浩恁(Hao-Jen Ting)

指導教授 ： 林康平

中原大學/電機資訊學院/電機工程研究所/碩士(2018年)

https://doi.org/10.6840/cycu201800775

摘要

血壓變異度是引發中風及心血管疾病重要風險因子，但因為連續血壓量測的不方便性使這項參數時常被忽略。現今的血壓變異度量測僅能以動態血壓量測來得到血壓的非連續趨勢變化，但此方法需要在上臂安裝袖套，運作時完全阻斷手臂血液造成不適且干擾睡眠。

如今有一種血壓量測方法是透過心電圖(Electrocardiogram, ECG)和光體積描計圖(Photoplethysmography, PPG)來推算血壓，其原理為計算脈波到達時間，即血液在動脈兩端點的輸送時間與血壓呈現負相關，此量測方法能提高量測舒適度也俱備設計成穿戴式裝置的潛力。

本論文收集了MIMIC II資料庫中含有ECG、PPG與動脈血壓訊號的十個病患來驗證血壓變異度估測結果，提出以脈波到達時間為主要參數，心率、脈波間期以及脈波寬做為輔助參數的不同血壓回歸模型，特徵值計算以10秒、30秒與60秒內平均取代逐拍代入的方式。

在相關性與RQA分析時以60秒平均的結果為最佳，在平均60秒內特徵值的狀況下，只將主要參數代入血壓回歸模型的相關係數為0.66，RQA分析後之一致性為0.51，加入輔助模參數進行血壓回歸時，能將相關係數提升至0.84，一致率提升至0.69。

關鍵字

血壓變異度 ； 脈波到達時間 ； 無袖式血壓量測

並列摘要

Blood pressure variability is an important risk factor of stroke and cardiovascular diseases, but people often ignored it because of the inconvenient of continuous blood pressure measurement. For the last decade people use ambulatory blood pressure measurement to estimate the trend of blood pressure curve, but this method need to set a cuff around the upper arm and fully occluding the arm’s blood circulation during the recording period, makes people feel uncomfortable and affects the quality of sleep.

Nowadays an innovative method can estimate blood pressure using electrocardiogram(ECG) and photoplethysmopraphy(PPG). The time delay between R peak and PPG feature point is reversely related to blood pressure. This is a potential method to improve comfort during measurement and the sensor can be design as a wearable device.

This study collects ten patient’s data from the MIMIC II database including ECG, PPG and arterial blood pressure to verify the blood pressure estimate result. We build up several blood pressure regression models with PAT as key parameter, and heart rate, interval of pulse and pulse width as minor parameter, and calculate the feature using difference widow length average (10s, 30s and 60s).

The 60s widow length give the best result in both correlation and recurrence quantification analysis. The correlation coefficient is 0.66 and the recurrence rate ratio is 0.51 when only estimate the blood with key parameter PAT, and the correlation coefficient and recurrence rate ratio becomes 0.84 and 0.69 after adding minor parameter into the blood pressure regression model.

並列關鍵字

Blood Pressure Variability ； Pulse Arrival Time ； Cuff-less Blood Pressure Measurement

參考文獻

[1]Lewington S, et al. Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies. Lancet. 2002 Dec 14;360(9349):1903-13.

[2]Walter N. Kernan, et al. Guidelines for the Prevention of Stroke in Patients With Stroke and Transient Ischemic Attack. Stroke. 2014;45:2160-2236.

[3]Wein T, et al. Canadian stroke best practice recommendations: Secondary prevention of stroke, sixth edition practice guidelines, update 2017. Int J Stroke. 2018 Jun;13(4):420-443.

[4]Rothwell, P. M. et al. Prognostic significance of visit-to-visit variability, maximum systolic blood pressure, and episodic hypertension. Lancet 375, 895–905 (2010).

[5]World Health Organization, http://www.who.int/