台灣版神經性疼痛症狀量表之研發與驗證

Abstract

背景：臨床上疼痛往往很難區分傷害感受性疼痛和神經性疼痛。不同的疼痛描述可能可以幫助我們區分疼痛的病因和病生理機轉。我們的目標是設計改良的繁體中文版本的神經性疼痛症狀量表，以提高診斷能力並執行驗證測試。此外，透過此量表評估個體疼痛描述與疼痛類型和神經生理學參數的相關性。 方法：本研究實驗對象包括20-85歲，患有神經性疼痛（周圍神經病和中樞神經系統疾病）和非神經性疼痛（骨關節炎，肌肉痛，頭痛）的患者。神經性疼痛症狀量表將進行中文翻譯與反向翻譯(NPSI-T)。短版麥吉爾疼痛問卷2，貝克抑鬱量表II，貝克焦慮量表和生活品質簡短量表36作為同時效度與其他標準效度之評量。主成分分析用於建構效度。Mcdonald’s omega用於測試整體內在信度，而Cronbach’s alpha用於依照建構效度分組後的個別信度測試。在兩次調查之間無疼痛改變的患者則計算其組內相關係數以確定再驗信度。如果患者之疼痛指數有所改變，則將重複測量並計算其相關性以確定變化敏感性。此外，使用羅吉斯回歸比較各個述詞的神經性疼痛和非神經性疼痛患者，計算勝算比。在多發性神經病變與神經根病變患者，透過計算神經傳導檢查的下肢動作電位，定量感覺測驗、與表皮內神經纖維密度之相關性以探討其神經生理價值。 結果： 從2018年9月到2020年4月，總共招募了144名患者。 其中NPSI-T之總分與短版麥吉爾疼痛問卷2為高度相關，與貝克抑鬱量表II，貝克焦慮量表和生活品質簡短量表36的心理成分為中度至輕度相關。 建構效度方面顯示了四種疼痛組別。內在信度的Mcdonald’s omega係數為0.806，各組別的alpha值以依照建構效度分類最高。再驗信度的組內相關係數為0.61。變化敏感性之Spearman相關係數為0.503。燒灼痛，電擊痛、針刺、麻刺、搔癢、感覺低落感表示神經性疼痛，而緊捏和壓迫則表示非神經性疼痛。電擊痛、刀刺痛、溫覺異痛、針刺、麻刺、癢、感覺低落與與大纖維神經之生理數據相關（神經傳導檢查的下肢動作電位，定量感覺測驗振動類別）有關。緊捏痛、壓迫痛、電擊痛、各種異痛、針刺、麻刺、癢、感覺遲鈍與小神經纖維異常（溫度定量感覺測驗、與表皮內神經纖維密度）有關。 結論： 改良的繁體中文版本的神經性疼痛症狀量表表現出足夠的信效度，其總體得分與其他問卷具有良好的相關性，尤其是在神經病患者中。個體的疼痛描述因素可以反映出不同的疼痛病理生理學。

Background: The clinical issue of pain is the difficulty in differentiation of nociceptive pain from neuropathic one. Different pain descriptors have potential for differentiate etiologies and pathophysiology of pain. We aimed to design a traditional Chinese version of neuropathic pain symptom inventory (NPSI-T) with modification to improve the diagnostic power and perform a validating test. The correlation of individual pain descriptor with pain types and neurophysiological parameters were also evaluated. Methods: Subjects between 20-85 years old were included. Patients with neuropathic pain (peripheral neuropathy and central neurologic lesion) and non-neurogenic pain (osteoarthritis, myalgia, headache) were recruited. The NPSI questionnaire were translated into NPSI-T and back-translated. Short-form Mcgill Pain Questionnaire-2 (SF-MPQ2), Beck Depression InventoryⅡ(BDI-Ⅱ), Beck Anxiety Inventory (BAI), and short form-36 (SF-36) were conducted as criterion-related and other criteria validities. Principle component analysis (PCA) was used for constructive validity. McDonald’s Omega was used to test internal consistency for the whole group, and Cronbach’s alpha was used for individual pain group generated from PCA. Patients with no subjective alternation in symptoms between the two surveys was taken into calculation of intraclass correlation coefficients (ICCs). To establish the sensitivity to change (STC), repeated NPSI-T was performed if patient had subjective alternation of pain severity, measured by NRS. Patient with neuropathic pain and nociceptive pain were compared or correlated with subcomponents of NPSI-T by logistic regression. Pearson’s correlation was used for each score with lower limb action potential in nerve conduction study (NCS), vibratory and thermal threshold in quantitative sensory testing (QST) and intraepidermal nerve fiber density (IENFD) for neurophysiology survey. Results: From Sep. 2018 to Apr. 2020, a total number of 144 patients were recruited. The total score of NPSI-T had strong correlation with MPQ2, and mild to moderate correlation with BDI-II, BAI, and mental component of SF-36. PCA showed four loadings in constructive validity. McDonald’s Omega coefficient was 0.806 for the whole group, and highest alpha was calculated from individual components from PCA. ICCs was 0.61. Spearman’s rho was 0.503 in STC. Burning and electric shock pain, pin-needle, tingling, itchy sensations and hypoesthesia were suggestive for neuropathic pain, while squeezing and pressure pain were suggestive for non-neuropathic pain. Shock, stabbing pain, thermal allodynia, pin-needle, tingling, itchy sensations, and hypoesthesia were correlated with large nerve fiber abnormalities (NCS, vibratory-QST). Squeezing, tight, electric shock pains, all allodynia, pin-needle, tingling, itchy sensations, and hypoesthesia were correlated with small nerve fiber studies (thermal-QST, IENFD). Conclusion: The NPSI-T showed adequate validity and reliability, and its global score had good correlation to other questionnaires, especially in neuropathic patients. Individual pain descriptors were able to reflect different pathophysiology of pain.