建立臺灣野鳥救傷在初步穩定時期之檢傷分數及處置建議模型

Abstract

野生動物救傷的最終目的在於幫助野生動物康復並返回野外，然而，在野鳥入院後的48小時，具有很高的死亡率，為了幫助獸醫師評估剛入院患禽的生理狀態，並給予治療建議，本研究針對入院時所進行的初步評估項目，與入院後48小時的死亡率進行分析，期建立一個初期檢傷系統，並分析不同治療方式與死亡率的差異，以提高傷救的成功率。 本研究於2016年十月至2018年八月間，共分析200隻野鳥，於入院時記錄鳥種、年齡以及身體體態指數、呼吸次數、心跳次數、意識狀態、黏膜顏色、拉皮測試、微血管回充時間、創傷等級等八個理學檢查項目。以及48小時內對該患禽所進行的醫療處置及最後結果。所有野鳥分為161隻的訓練樣本及39隻的驗證樣本，將訓練樣本以邏輯斯回歸進行分析，得到野鳥檢傷分數，並以此得到預估的死亡率。訓練樣本中每個個體的野鳥檢體以野鳥檢傷分數進行排序後，以四分位數進行分組，再利用卡方檢定比較分數相近下，不同處置間的結果是否有顯著差異。 透過回歸分析，本研究發現心跳速率、意識狀態、黏膜顏色與野鳥入院48小時死亡率有關，可利用回歸係數得到野鳥檢傷分數(Wild bird triage score, WST)公式，WST=0.451*(心跳分數)+0.794*(意識狀態分數)+0.513*(黏膜顏色分數)-2.909 (P<0.05)，由此分數預估救傷野鳥在入院後48小時的死亡率之公式為P(死亡率)=eWTS/(1+ eWTS )，並可得到分數與死亡率之對照表，分數越高代表死亡率越高。本系統之接收者特徵曲線下面積(area under the curve, AUC)為0.729，達到可接受標準。經過Youden指數輔助，在切點等於-1.4940時，敏感度為67.4%及特異度為70.3%，其陽性似然比為2.30。 將卡方檢定的結果顯示在分數較低的組別中，不同處置間的結果並沒有顯著差異，只有分數最高的第四組具有顯著差異，在該組中，上針給予輸夜的組別死亡率低於皮下給予的組別。 本研究所提出的三項因子：心跳速率、意識狀態、黏膜顏色，分別包含了循環系統及神經系統的評估，與過去人類醫學及小動物急診加護所提出的檢傷系統相類似。過去的野生動物救傷文獻中，也建議野生動物入院後首要被評估的系統應為心血管系統，在相關鳥類的動物實驗中也顯示失血會導致心跳的速率改變。本研究結果與過去野鳥救傷長期預後因子的研究結果有所不同，顯示在野鳥救傷的不同階段，需要不同面向的評估。而治療策略的比較則顯示在嚴重程度較高的組別中，上針給予輸液也許可以幫助死亡率的降低。 本研究嘗試建立一完整的野鳥檢傷系統，以期提供一個非侵入性並可快速取得資料的指標，幫助臨床醫師預測野鳥在入院初期的死亡率，並給予相對應的治療建議。

To overcome high mortality rate encountered in wild bird rescue and achieve early prediction of prognosis, establishing a triage system and management criteria for birds is mandatory. A scoring system was set up to evaluate the outcomes of different levels of stabilization managements under different patient factors. Signalment and physical examination parameters including body condition score, heart rate, respiratory rate, mentation, severity of symptoms and hydration state of 200 rescued birds from several wildlife rescue facilities were collected from October, 2016 to August, 2018. Prognosis and stabilization manner based on invasiveness including overnight stabilization, oxygen support, subcutaneous (SQ) fluid therapy and intravenous(IV) or intraosseous(IO) fluid therapy during the first 48-hour-hospitalization were also recorded. The Wild Bird Triage Score was set up by logistic regression analysis to predict mortality of the first 48 hours. Chi-Square Tests was performed to compare result of different management in similar score. Logistic regression model indicates the heart rate score, tenting skin score and mucus membrane color score have positive correlation with mortality rate (P<0.05). The model was tested for receiver operating characteristic curve, and the area under the curve was 0.729, with 67.4% sensitivity, 70.3% specificity at the cut-off point -1.4049 The positive likelihood was 2.30. The result of Chi-Square Tests demonstrate no significant differences in the first second and third class, while significant differences are noted in the fourth class. In the fourth class, the average mortality of IV/IO class are lower than SQ class, The three factors are related to cardiovascular system and neurological system, similar to previous literatures of human and small animal medicine. Therefore, evaluation of cardiovascular system and neurological system should be the first priority after animal admission. In a previous study, mallard ducks showed heart rate alternation after acute hemorrhage. These risk factors were not investigated in previous wild bird prognostic indicators studies. It emphasized different rehabilitation stages may require different assessments. As for therapy strategy, intravenous or intraosseous fluid therapies could reduce mortality rate in patients of different severity, while subcutaneous fluid therapy is more appropriate for patient of medial severity. The study established a general wild bird triage score system which offered a guild line based on non-invasive and rapidly assessable examinations. Apply to the scoring system, clinicians can estimate the first-48-hour mortality rate of hospitalized wild birds and provide appropriate treatments.