心率變異性的短幅碎形相關於不同麻醉方法下的探討

Abstract

一、研究背景及目的: 利用時域(time domain)及頻域(frequency domain)分析法，分析心率變異性(heart rate variability，HRV)，不論是在健康者或心血管疾病患者，此線性(linear)分析法為最常使用於評估心臟自律神經調節的狀況。然而，人類的心率變動卻隱藏著非線性(nonlinear)的特色，故新的分析技術也被發展來探測心率行為，希望能彌補傳統分析方法所未能偵測到的部分。去勢波動分析法(detrended fluctuation analysis，DFA)即是其一，它描述了一系列心跳間距(R-R interval)的碎形相關性(fractal-like correlation)。另外取樣熵(sample entropy，SampEn)為另一種非線性方法，可以用來定量複雜度。 當碎形相關性降低，也就是碎形的模式被破壞，代表著心血管事件及死亡率的風險性增加，然而，針對碎形的心率變動性背後之病生理基礎並未清楚了解。文獻中雖已有繁多麻醉相關所造成的頻譜研究，來探討自律神經的變化，但是卻少有以麻醉介入後所造成的短幅碎形變化之研究，故我們的研究目標為探討半身麻醉及全身麻醉對自律神經的調節所造成碎形相關性的影響，並輔以傳統線性分析分式及SampEn加以比較及說明其對HRV造成影響之生理意義。 二、研究方法: 本研究經人體試驗委員會同意及病人簽署同意書，我們選100位美國麻醉醫學會體位分級 (ASA classification) 第一級，年紀介於20-50歲中間，預定接受常規手術的病患。排除有嚴重心肌梗塞、鬱血性心臟病、糖尿病、聽覺障礙以及其他影響自主神經系統的疾病，同時沒有服用影響心臟血管的藥物。為了觀察短幅碎形尺度指數α1(short-term fractal scaling exponent)、傳統線性頻譜分析及SampEn，實驗將病人分成(1)半身麻醉組:1)一般劑量組(HM)(n=19)，2)低劑量組(LM)(n=20)，3)低劑量+吩坦尼組(LMf)(n=20);(2)全身麻醉組:1)靜脈麻醉劑propofol組(n=15)，2)吸入性麻醉劑desflurane 組(n=18)。每位病患由於須進行手術的緣故，皆於術前空腹八小時，且在手術前兩天禁止劇烈運動及飲用含酒精或咖啡因之飲料，到達手術室時，先利用心電圖記錄器(MSI&reg; Portable ECG Recorder and analyzer, E3-8010)，讓病患平躺於安靜的環境中，先經5分鐘的測試讓心電圖波形達到平穩，之後連續紀錄心電圖變化。半身麻醉組:HM及LM組分別接受0.5% Hyperbaric bupivacain 12mg及6mg，LMf組接受0.5% Hyperbaric bupivacain 6mg及20μg fentanyl ;而全身麻醉組皆以面罩式100%氧氣給予持續 2-3分鐘，propofol組接受300μg/kg/min propofol持續靜脈注射，desflurane組接受fentanyl 1μg/kg的靜脈注射，並給予O2及N2O (1:1)，以3%-6%-9%-12% desflrane漸進方式來誘導，直到麻醉深度偵測器之指標AAI (A-Line ARX Index)降到35以下，其中持續偵測血氧飽合濃度和吐氣末二氧化碳，並以溫和間歇正壓呼吸 (IPPV) 從面罩式給予氧氣，以維持正常呼吸。 將基礎及麻醉後的心電圖資料，傳輸至電腦進行非線性方法—DFA、SampEn，及線性方法--LF/HF ratio (low-/high-frequency power ratio)、SDNN(Standard deviation of the NN interval)、RMSSD(Square root of the mean squared differences of consecutive NN interval)及SI(Similarity index)分析。 三、研究結果: 研究結果發現在給予半身麻醉的三組，其α1皆下降(HM由1.24±0.15&agrave;0.78±0.11;LM由1.32±0.25&agrave;0.98±0.21;LMf由1.28±0.17&agrave;0.8±0.21，P皆<0.0001);在全身麻醉propofol 及desflurane組，AAI介於35及60時，α1增加(分別P<0.05,及P<0.001)，至AAI小於35時，α1皆呈現下降(propofol組1.14±0.2&agrave;0.94±0.35;P<0.05; desflurane組1.1±0.26&agrave;0.7±0.31;P<0.0001)。在半身麻醉三組及desflurane組中，normalized high frequency(nHF)能量增加，normalized low frequency(nLF)能量減少，LF/HF ratio (LHR)降低(p<0.05)。SampEn在LM、LMf及desflurane組呈現下降。利用ROC (receiver operating characteristic) curve 分析結果顯示，以DFAα1相較於其它線性方法，用來區分麻醉前後具有較佳的敏感度及特異性。 四、結論 在半身麻醉不同劑量及合併fentanyl，與吸入性氣體麻醉劑desflurane及靜脈麻醉劑propofol達AAI小於35的麻醉深度時，α1與麻醉前基礎值相比，皆顯著降低，顯示不同麻醉方法介入後造成心率的非線性訊號產生動態變化，皆使心率短幅碎形相關性降低，而趨於更隨機的變化。此外，在全身麻醉組發現隨著麻醉深度加深的過程中α1有雙向性的改變(由增加到減低)，在未達AAI<35時，α1增加意味著此階段的短幅碎形關係增強。對半身麻醉及全身麻醉而言，DFAα1比傳統線性方法對於區分麻醉前後具有更佳的指標性。

Key words: Heart rate variability, spinal anesthesia, general anesthesia, detrended fluctuation analysis, sample entropy, linear analysis Background Time and frequency domain analyses of heart rate variability (HRV) are the most commonly used noninvasive methods to evaluate autonomic regulation of heart rate in healthy subjects as well as in patients with cardiovascular disorders. Because nonlinear phenomena are involved in the genesis of human heart rate fluctuations, new analysis techniques have been developed to probe features in heart rate behavior that are not detectable by traditional analysis methods. Analysis of fractal scaling exponents by detrended fluctuation analysis (DFA) is one such method that describes the fractal-like correlation properties of R-R interval data. Sample entropy (SampEn) is another nonlinear method that quantifies the amount of complexity in the time-series data. Breakdown of short-term fractal-like behaviour of heart rate indicates an increased risk for adverse cardiovascular events and mortality, but the pathophysiological background for altered fractal heart rate dynamics is not known. Despite a large body of data concerning the changes in spectral characteristics of HRV during anesthesia, there is little information on the effects of theses physiological interventions on non-linear characteristics of heart rate behavior. This study was designed to assess the changes in the nonlinear features of HRV caused by the spinal anesthesia and general anesthesia. The main purpose was to gain insight into the physiological background for fractal and complexity characteristics of heart rate dynamics. Short-term fractal scaling exponent (α1)along with spectral components of HRV were analyzed during the following anesthesia interventions in patients : (1) spinal anesthesia group : 1)normal dose (Group HM, n=19), 2)low dose (Group LM, n=20), 3) low dose combine fentanyl (Group LMf, n=20); (2) general anesthesia group: 1)total intravenous propofol infusion (Group P, n=15),2) inhalation induction with desflurane (Group D, n=18) Method After institutional ethical approval and getting informed consent, we recorded the electrocardiogram of 100 ASA class I (American Society of Anesthesiologist physical status class I) patients proposed to receive elective surgery. Patients were excluded if they suffered from severe ischemic heart disease, congestive heart failure, diabetes mellitus, or other disorders known to affect autonomic function. None of the patient was taking medications that affect cardiovascular function. Each patient fasted at least 8h prior to testing. Vigorous exercise, alcohol and coffee were also forbidden for 48 h before the operation. On arrival to the operating room, the patients lay in a supine position in a quiet room at least 5 min prior to data collection. In Group HM and LM, 12mg and 6mg of 0.5% hyperbaric bupivacain were injected respectively. In Group LMf, 6mg of 0.5% hyperbaric bupivacaine was supplemented with 20μg of intrathecal fentanyl. All patients received 100% oxygen via face mask for 2 to 3 min prior to induction of general anesthesia. In Group P, patients received propofol infusion at a rate of 300ug/kg/min . In Group D, anesthesia was induced with 3-6-9-12% desflurane increasing gradually in 2L/min O2 and 2L/min N20. Arterial oxygen saturation (SpO2) and end-tidal carbon dioxide (ETCO2) were monitored, and normoventilation was maintained with gentle IPPV via mask if required. Depth of anesthesia was monitor by AAI (A-Line ARX Index) continuously until the value reached 35. Therefore, the HRV measurements were performed at AAI values of 60 to 35 and less than 35. The electrocardiogram data was transferred into the hard disk in a personal computer and offline analysis was performed. Results Short-term fractal scaling exponent (α1) decreased during spinal anesthesia in three groups ( Group HM:from 1.24±0.15 to 0.78±0.11;Group LM:from 1.32±0.25 to 0.98±0.21;Group LMf:from 1.28±0.17 to 0.8±0.21，P<0.0001).α1 increased during both general anesthesia group at AAI value of 60 to 35. Thenα1 decreased during the AAI value less than 35 (Group P: from 1.14±0.2 to 0.94±0.35,P<0.05; Group D:from 1.1±0.26 to 0.7±0.31,P<0.0001). Conventional HRV indices did not show the dynamic changes in Group P.Group HM, LM, LMf and Group D decreased the normalized low frequency spectral power and LF/HF ratio and increased normalized high frequency spectral power (p<0.05). SampEn value decreased in Group LM, LMf and Group D. In addition, the receiver operating characteristic (ROC) was used to estimate the sensitivity and specificity of classification of subjects in awake and after anesthesia states using different parameters. The results show that the DFAα1 is a better indicator for distinguishing baseline from anesthesia state. Conclusion Spinal and deep general anesthesia result the breakdown of short-term fractal-like behaviour of heart rate. Incremental depth of anesthesia until AAI less than 35 results in bidirectional changes in correlation properties of R-R interval dynamics. The results suggest that decrease sympathetic outflow at the same time activation of vagal outflow explains the breakdown of fractal-like behaviour of human heart rate dynamics. Change in α1 can be detected also in light anesthesia levels, when the conventional measures of HRV can not be applied. In addition, α1 is a better indicator for distinguishing baseline from spinal anesthesia state.