並列摘要

Background: Postoperative pain is distressful, and it imposes adverse effects on multi-systems. Early intervention and effective postoperative pain management had always been major concerns of clinical anesthesiologists. For pain is subjective, psychological factors had been taken into considerations to make predictions in several studies. Temporal changes of heart rate variability (HRV) across the perioperative period, which reflects the dynamic activities of the autonomic nervous system (ANS), is another important part we want to incorporate into the prediction model. Our goal was to develop a better prediction model of pain severity based on both the demographic factors and intraoperative indices. Method: We enrolled 80 women ≥ 20 years of age scheduled for gynecological surgeries under general anesthesia. All participants were American Society of Anesthesiologists classification of physical status 1 to 3 without using drugs affecting HRV. Questionnaires including Insomnia Severity Index (ISI) and Beck Depression Inventory-II (BDI-II) were used to evaluate participants’ sleep qualities and severity of depression, respectively. Physiological signals were recorded perioperatively. After surgery, the numeric rating scale (NRS) for pain was measured as a patient’s arrival at the postanesthesia care unit (PACU). The HRV indices of frequency-domain and nonlinear-domain were computed and analyzed offline. The demographic factors and intraoperative indices were included to build a prediction model of postoperative pain severity by using the stepwise linear regression. Results: We used the stepwise linear regression to build a model for the initial NRS scale on arrival at the PACU. The formula of the final multivariable model is as follows: NRS = -0.784 × Surgery Type (1 for laparoscopic surgery and 0 for open surgery) + 0.086 × ISIscore - 0.044 × Age + 0.002 × Volume of blood loss + 0.006 × deltaVLF + 0.014 × deltaSD1 - 0.006 × deltaSD2 - 0.003 × deltaEntropy. (delta in the formula denotes the change ratio from the midpoint of the surgery to before the end of surgery) The results showed that this model is a significant predictor of the initial pain score in the PACU (F8,71 = 3.798, P = .0009). The adjusted square of R was .22. Conclusions: With sleep quality, demographic factors, and changes in measures of intraoperative HRV, we develop a prediction model of initial NRS on arrival at the PACU. Further research is required to validate the results of this pilot study.

摘要 Summary

背景：術後疼痛不僅造成病人的不適，所帶來的壓力亦會產生不利影響。由於疼痛受多重因子調節，心理與生理層面皆須納入考量。先前的研究曾利用憂鬱或焦慮的問卷來預測術後的疼痛，但術中的動態生理指標的納入，才能反映不同個體面對手術壓力的反應差異。我們的研究目標是融合術前的身心理評估與術中的自主神經系統活性變化，來建立一個預測術後疼痛程度的模型。方法：此研究收錄80位年滿20歲，接受婦科手術的女性。我們於術前完成失眠嚴重度量表和貝克憂鬱量表的評估，作為個案睡眠品質和憂鬱程度的指標。手術當中全程收錄生理監測資料並以Matlab來計算心率變異度。之後以病人基本資料、問卷結果以及心率變異度變化來進行多因子線性迴歸。結果：我們所建立的術後疼痛（NRS）的模型如下：術後疼痛分數= -0.784 × 手術種類 + 0.086 × 睡眠問卷總分 - 0.044 × 年紀+ 0.002 × 出血量 + 0.006 × VLF頻域變化 + 0.014 × SD1 變化 - 0.006 × SD2變化 - 0.003 × Entropy 變化。結論：本前瞻研究建立了有意義的疼痛預測模型。後續的主研究可證實這個模型的可信度。