機器輔助療法於中風復健之密集度效應與劑量反應關係研究

Abstract

背景與目的 中風目前仍是世界各國關注的重要公共健康議題之一。而許多中風患者皆伴隨動作功能缺損，驅使尋求有效的復健療法以提升患者之動作功能復原。機器輔助治療是近年來一個極具潛力的復健療法，它融入一些有效的中風動作治療策略於設計當中。然而，目前對於機器輔助療法在臨床指標之療效及治療對生理反應之影響等科學性證據仍然有限。此外，探究適切的治療密集度以闡述治療劑量與療效間關係亦有其必要性。因此，本論文主要包含兩部分的研究：(一)機器輔助治療之療效試驗與(二)療效工具之臨床計量特性研究。第一部分之研究目的為: (1)探究機器輔助療法相對於對照組運用於中風患者之成效、(2)使用較高與較低密集度兩組不同治療強度之組別以檢視治療劑量與成效之關係、及(3)探究密集性機器輔助療法對於細胞氧化壓力之生物指標8-hydroxy-2''- deoxyguanosine (8-OHdG)的影響。而第二部分的研究旨在探討影響機器輔助療法成效之預測因子以界定適當的介入族群、及檢驗機器輔助療法成效工具之臨床計量特性。

方法 此研究共募集54位中風患者參與，並隨機分派至高密集度機器輔助療法組、低密集度機器輔助療法組、或對照組之其中1組，每位患者皆接受4週療程。主要成效評量工具為傅格梅爾動作量表(Fugl-Meyer Assessment, FMA)與醫學研究會議之肌力量表(Medical Research Council scale, MRC)，患者於治療前、期中(治療2週後)、與治療後各接受1次評估。次級成效評量工具包含箱子與積木測驗(Box and Block Test, BBT)、改良版艾斯渥式量表(modified Ashworth scale, MAS)、動作活動日誌量表(Motor Activity Log, MAL)、中風影響量表(Stroke Impact Scale, SIS)之生理層面，分別於治療前、後進行評估。此外，患者之尿液樣本8-OHdG檢驗、疲勞與疼痛評估則做為不良反應指標。資料分析使用重複量數二因子共變數分析比較3種治療組別在主要成效評量工具3個評估點之差異；另使用共變數分析檢驗3組間在次級成效工具之治療效應。另以復原速率及劑量與反應曲線檢視2組不同密集度機器輔助療法之劑量與反應間關係。此外，使用多重迴歸分析找出主要成效工具之預測因子；而成效評量工具於偵測治療後最小臨床重要改變之能力及反應性亦將被檢驗。

結果 每組各18名患者參加此研究。對於主要成效評量工具，在FMA總分上具顯著的組別與時間之交互作用(F3.4, 83.8 = 3.95, P = 0.01)。事後檢定發現3組在FMA總分從治療前到期中、與治療前到治療後皆呈現顯著的組內進步量(all P < 0.05)；組間比較發現高密集度機器輔助療法組於期中與治療後相較其他兩組在FMA總分上皆具顯著較高的進步量(all P < 0.05)。而在MRC，組別與時間並無交互作用(F4, 100 = 1.41, P = 0.24)，分析呈現顯著的時間主要效應(F2, 100 = 4.54, P = 0.01)，代表所有個案在MRC皆隨時間顯著進步；但各組別間則無顯著差異(F2, 50 = 0.87, P = 0.43)。在次級成效工具結果中發現3組間於BBT分數具有顯著差異(F2, 50 = 4.68, P = 0.01)，事後分析結果為高密集度機器輔助療法組和對照組兩組相較低密集度機器輔助療法組有較佳的進步量。另3組間於MAS、MAL與SIS生理層面量表之分數則無顯著差異。在不良反應指標中發現3組患者在疲勞與疼痛之評分上具輕微疲累或疼痛(平均分數小於3分)。而3組間於8-OHdG數值並無達顯著差異(P = 0.24)，且3組在治療前、後之8-OHdG平均數值皆在正常參考範圍值內。此外，高密集度機器輔助療法組之復原速率較低密集度機器輔助療法組顯著較高，尤其是在FMA (P < 0.05)和BBT (P = 0.05)。且發現患者治療前之損傷嚴重程度會影響其治療成效，具中等程度之動作缺損或肌肉無力者經高密集度機器輔助療法後的進步量最多。在第二部份的研究中發現接受機器輔助療法之患者，上肢遠端動作能力高低、治療組別、與治療前患手使用量多寡為影響FMA治療後分數之顯著預測因子(adjusted R2 = 0.662, P < 0.01)；而上肢遠端動作能力高低與治療組別則為影響MRC治療成效之顯著預測因子(adjusted R2 = 0.597, P < 0.01)。此外，約20%到40%接受機器輔助療法之病患其進步量達到臨床重要意義；且本研究所使用之成效工具於偵測接受機器輔助療法後之進步量皆具高度反應性(standardized response mean = 0.96 to 1.69)。

結論 此研究結果顯示高密集度機器輔助療法組相較其他兩組有較佳的治療成效，尤其在上肢動作能力之成效。高密集度機器輔助療法於動作成效上之復原速率亦較低密集度機器輔助療法為快。亦發現接受高密集度機器輔助療法之患者，損傷嚴重程度之不同會影響其治療成效。而影響機器輔助療法動作與肌力成效之較佳預測因子為患者之上肢遠端動作能力優劣與治療組別。本研究中使用之成效工具可反映出中風患者接受機器輔助療法後之進步量。治療後患者並無明顯的疲勞或疼痛感上升之情形，且生理氧化壓力指標亦無上升趨勢，代表此研究之治療方案應可被參與者接受與容忍。總和以上結果，高密集度機器輔助療法應可被建議運用於具中、輕度動作損傷之慢性中風患者以協助提升其動作復原。此研究之整體結果更充實我們對於機器輔助復健療法運用於中風病人之治療成效、治療劑量與成效間關係、細胞氧化壓力影響、成效預測因子、及成效工具之臨床計量特性等瞭解。此具潛質之研究結果與經驗將提供未來持續探究中風復健重要議題的洞悉力與理解，並有助於神經復健實證之發展。

Background and purposes Stroke remains a compelling public health issue worldwide. With high percentages of stroke survivors left with motor deficits, motivating the search for effective rehabilitation to improve motor recovery. Robot-assisted therapy (RT) has emerged as a prominent approach in the last decade that incorporates successful therapeutic elements of motor rehabilitation into its design. However, scientific evidence for the effects of the RT on clinical outcomes and physiological responses in stroke patients remains limited. Also, there is a need to identify the proper level of treatment intensity to elucidate the dose-response relations. This dissertation consisted of two parts of study: Efficacy Study of Robot-Assisted Therapy, and Clinimetric Study of Outcomes. In the first part of study, the purposes were (1) to investigate the treatment effects of RT relative to a comparison treatment (CT) in patients with stroke on clinical outcomes, (2) to test the dose-response relations by using 2 RT groups with higher- and lower-intensity, and (3) to examine the effects of intensive RT on a biomarker of oxidative stress (i.e., 8-hydroxy-2''-deoxyguanosine [8-OHdG]). In the second part of study, the purposes were to define the appropriate populations for RT and to examine the clinimetric properties of outcomes used in RT.

Methods A total of 54 patients with stroke were recruited in this study and were randomized into the higher-intensity RT, the lower-intensity RT, or the CT group for a 4-week intervention. Primary outcome measures, including the Fugl-Meyer Assessment (FMA) and Medical Research Council scale (MRC), were administered to patients before intervention, at midterm (2 weeks after intervention), and immediately after intervention. Secondary outcomes, including Box and Block Test (BBT), modified Ashworth scale (MAS), Motor Activity Log (MAL), and physical domains of the Stroke Impact Scale (SIS), were administered to patients before and after intervention. In addition, urinary 8-OHdG levels of patients, pain and fatigue evaluation were also investigated as adverse responses. Two-way repeated measures analysis of covariance (ANCOVA) was used to evaluate the effects of primary outcomes among the 3 groups at 3 assessments. ANCOVA was used to examine treatment effects of secondary outcomes among the 3 groups. To represent the dose-response relations between the 2 RT groups with different intensities, the recovery rates and dose-response curves of each outcome were examined. Moreover, multiple regression models were used to identify the significant predictors for primary outcomes. The capacity of outcome measures to capture minimal clinically important changes after RT and the responsiveness of outcomes were also examined.

Results Each group had 18 patients who participated in the study. For the primary outcomes, there was a significant group × time interaction effect (F3.4, 83.8 = 3.95, P = 0.01) on the FMA-total score. All 3 groups showed significant within-group gains in the FMA from baseline to midterm and from baseline to posttreatment (all P < 0.05). The higher-intensity RT group had significantly higher improvements in the FMA than the other 2 groups at midterm and posttreatment (all P < 0.05). On the MRC, no significant group × time interaction effect was found (F4, 100 = 1.41, P = 0.24). The analysis revealed a significant time main effect (F2, 100 = 4.54, P = 0.01), but there was no significant differences for the group main effect (F2, 50 = 0.87, P = 0.43). For the secondary outcomes, a significant difference in the BBT among the 3 groups was observed (F2, 50 = 4.68, P = 0.01). Post hoc analysis revealed that the higher-intensity RT and CT groups had greater gains in the BBT than the lower-intensity RT group. However, the improvements in the MAS, MAL, and physical domains of the SIS did not show significant difference among the 3 groups. For the adverse responses, the mean ratings of fatigue and pain of the 3 groups were mild (mean scores < 3). Further, there were no significant differences in the 8-OHdG levels among the 3 groups (P = 0.24) and the mean 8-OHdG levels of the 3 groups were in the normal reference range. In addition, recovery rates of the higher-intensity RT group were significantly higher than those of the lower-intensity RT group at midterm and at posttreatment, particular in the FMA (P < 0.05) and the BBT (P = 0.05). The initial severity levels of the patients were found to affect their treatment effects on the primary outcomes. The patients in a middle range of motor deficits or muscle weakness gained most improvements after the higher-intensity RT. For the second part of study, motor ability of the distal part of the upper extremity, RT treatment group, and amount of affected hand use in daily activities were significantly predictive of the FMA model (adjusted R2 = 0.662, P < 0.01) after RT. Motor ability of the distal part of the upper extremity and RT treatment group were the significant predictors for the MRC model (adjusted R2 = 0.597, P < 0.01) after RT. Moreover, there were about 20% to 40% of the patients receiving RT with clinically meaningful improvement on the outcomes. The outcome measures used in this study had large responsiveness in detecting improved changes after RT (standardized response mean = 0.96 to 1.69).

Conclusions The findings of this study suggest that the higher-intensity RT intervention had better treatment effects, especially in upper-extremity motor ability, than the other 2 interventions. Recovery rates of the higher-intensity RT group were greater than those of the lower-intensity RT group on motor outcomes. The initial severity levels of the patients were found to affect their treatment outcomes after the higher-intensity RT. The better predictors for motor ability and muscle power outcomes after RT were motor ability of the distal part of the upper extremity and RT treatment group. The outcome measures used in this study are responsive to improvements of stroke patients after RT. The intervention protocols in this study generally can be tolerated by the participants without exacerbation of pain or fatigue and did not increase more oxidative stress after treatment. Based on the results, the higher-intensity RT is suggested to deliver in chronic stoke with moderate to mild motor deficits to improve motor recovery. The overall results enrich our understandings of treatment effects, dose-response relations, oxidative responses, prediction models, and clinimetrics of outcomes after robot-assisted rehabilitation in stroke patients. The promising results and experiences provide insights for continued study of these critical issues in stroke rehabilitation to contribute to evidence-based neurorehabilitation.