加護病房末期病患有簽署撤除維生醫療的現況暨主要影響因素以及維生醫療對存活時間之影響

背景：撤除維生醫療的精神為尊重末期病人不願仰賴機器維生的醫療意願，提供以舒適為主的緩和醫療。末期病人的家屬雖已簽署撤除「末期病人終止或撤除維生醫療」，但仍期待有相關醫療處置在進行。國外近年來對於末期病患維生醫療終止或撤除後所進行相關醫療處置之相關研究較多，反觀國內研究甚少，因此引發本篇作者希冀藉由此次的研究，能了解目前台灣在撤除維生醫療的現況。研究目的：本研究目的探討內科加護病房末期病人，一、家屬有簽署「撤除維生醫療」同意書之現況和簽署時間，以及主要影響因素；二、於內科加護病房所接受相關醫療處置之項目和治療時間；三、接受相關醫療處置是否增加於加護病房中以及一年內死亡的風險。研究方法及工具: 本研究採回溯縱貫性研究設計以查閱病歷方式分析北區大型醫學中心內科四個內科加護病房(internal medical intensive care unit, MICU)，自2016年1月份至2016年12月份，以入內科加護病房後已簽署「不施行心肺復甦術」且有兩位專科醫師診斷為末期病患為對象，以病歷查閱方式收集病患入加護病房的日期、死亡或轉出日期、人口學、疾病資料、簽署「撤除維生醫療」同意書日期、入內科加護病房後所接受之相關醫療處置之項目和治療時間。研究結果：本研究呈現326位末期病患的分析結果，有簽署撤除維生醫療的比例低10.4% (34/326)，同意簽署撤除維生醫療的平均時間為入加護病房後13.09天(SD±5.57)，簽署後同意書後至實際撤除維生醫療的平均時間為1.74天(SD±1.52)。相較沒有簽署的病患，有簽署「撤除維生醫療」同意書的病患年紀較年輕，平均為58.97歲(SD±16.44)，較少社會資源補助(58.8%)，癌症(29.4%)及心臟疾病(20.6%)為主要診斷比例較高，疾病嚴重度平均分數較高，平均住院天數較短(14.7天)。迴歸分析(Logistic regression)結果發現年齡及疾病嚴重度對於有無簽署「撤除維生醫療」同意書的解釋力達顯著意義，個別變項勝算比(Odds ratio)的檢定結果，平均年齡越高的病患簽署「撤除維生醫療」同意書的機率為平均年齡較低病患的0.965倍(p=0.004)；疾病嚴重度較高的病患簽署「撤除維生醫療」同意書的機率為疾病嚴重度較低病患的1.045倍(p=0.017)。末期病患的家屬決定簽署「撤除維生醫療」同意書的主要原因為不願生病的家人痛苦，比例為58.8%，其次為預後不佳(23.5%)以及生病之前意願(14.7%)，不願撤除的原因87.5%希望積極治療；而撤除維生醫療項目中最多為人工呼吸器(侵襲性及非侵襲性)，其比例為64.7%，其次為撤除強心升壓藥物、血液透析、葉克膜。相較沒有簽署病患，有簽署「撤除維生醫療」同意書的病患接受較多高濃度的強心升壓藥物治療(x2=4.253, p= 0.039)，持續靜脈輸液(x2=4.686, p= 0.03)，鎮靜藥物治療(緩解疼痛) ( x2=5.713, p= 0.017)，以及葉克膜治療 (x2=7.567, p= 0.014)；接受較少的治療為人工氣道(氣切手術) (x2=4.631, p=0.031)及腸道營養的給予(x2=11.45, p= 0.001)。以COX存活迴歸分析(COX regression)末期病患於加護病房接受相關醫療處置是否增加於加護病房中以及一年內死亡的風險。在控制疾病嚴重度的因素以及有無簽署「撤除維生醫療」同意書，影響於加護病房中死亡勝算比(Odds ratio)的風險結果為相較沒有接受以下治療，有接受氧氣治療死亡風險為0.451倍，有接受腸道營養治療死亡風險為0.275倍，有接受非腸道營養治療死亡風險為0.434倍，有接受點滴持續輸液治療死亡風險為0.625倍，有接受輸血治療死亡風險為0.487倍；相反的有接受強心升壓劑治療較沒有接受治療病患死亡風險為2.524倍。於一年內死亡的風險，相較沒有接受以下治療，有接受氧氣治療死亡風險為0.545倍，有接受腸道營養治療死亡風險為0.381倍，有接受非腸道營養治療死亡風險為0.562倍，有接受點滴持續輸液治療死亡風險為0.677倍，有接受輸血治療死亡風險為0.666倍；相反的，有接受強心升壓劑治療較沒有接受治療病患死亡風險為2.096倍。而沒有影響死亡風險的治療為人工呼吸器(侵襲性及非侵襲性)、血液透析、抗生素、抽血檢驗以及鎮靜藥物。研究結論及應用：目前加護病房末期病患簽署「撤除維生醫療」同意書現況仍偏低。影響簽署的主要原因為年齡較輕和疾病嚴重度較高，有簽署病患較年輕且癌症及心臟疾病的比例較高，疾病嚴重度也較高，而沒有簽署病患平均年齡較年長且多為長期慢性疾病，主要決定撤除維生醫療的原因為不願看到生病家人痛苦。接受強心升壓劑治療與末期病患死亡風險有正相關，而接受氧氣治療、腸道營養治療、非腸道營養治療、點滴輸液以及輸血治療與死亡風險為負相關。長期而言，接受強心升壓劑治療末期病患於一年內之死亡風險增加有關。短期和長期而言人工呼吸器(侵襲性及非侵襲性)、血液透析、抗生素、抽血檢驗以及鎮靜藥物等治療不會影響死亡風險的治療。未來需有更多維生醫療對於末期病患之存活分析期能預測醫療處置對病患生存時間的影響，以幫助末期病患避免無效的維生醫療處置。

關鍵字

末期病患；不施行心肺復甦術；撤除維生醫療；末期照護；醫療處置

並列摘要

Background: The withdrawing life support for terminally ill patients is developed to meet their wishes for not maintaining their lives with machines, and providing palliative care to enhance their comfort. However, even after families have given the consent for withholding or withdrawal of life support for their ill family members, they are reluctant to withdrawal some medical intervention. While in the recent years, this issue has been done in overseas, very few researchers explore this issue in Taiwan. Research purpose: This study aims to explore the following issues among the terminally ill patients in medical intensive care units (MICU): 1. the incidence and time of withdrawal of life support decisions and the main predictors associated with their decisions; 2. type and lengths of receiving life support treatments; 3. the impacts of life support medical interventions on the mortality in period of MICU and one-year after discharge from MICU. Methods: This study adopts a retrospective longitudinal study. We collected data from January 2016 to December 2016 by chart review of medical servicesamong the terminally ill patients from four MICUs in the general hospital. The inclusion criteria for this study were the patients who had signed “Do Not Resuscitate” forms and are diagnosed as terminally ill patients by two specialists. The date of admission to the MICU, date of death or discharge from the MICU, epidemiological information, disease information, date of signing a “withdrawal of life support consent”, and types and lengths of medical interventions received in the MICU. Results: The results of this study are based on the 326 terminal patients. The rates of consent for withdrawal of life support were 10.4% (34/326), average time of consenting to withdrawal is13.09 days (SD±5.57) after MICU admission, and average time from providing the written consents to implementing withdrawal of life support were 1.74 days (SD±1.52). Compared to the patients without the written consent for withdrawal, the patients who had signed “withdrawal of life support consents” were younger at an average age of 58.97 years old (SD±16.44), had less supports from social welfare (58.8%), had the higher percentages of cancer or cardiac diseases as the major diagnosis (20.6%), had higher average scores of Apache II for disease severity, and had a shorter length in MICU (14.79 days). Logistic regression model showed that age and Apache II disease severity were associated with the providing writeen consent for “withdrawal of life support”. The older patients were 0.965 times (p=0.004) likely provided the consent for withdrawal of life support than the younger patients. Patients with higher Apache II scores of disease severity were 1.045 times more likely to give consent for withdrawal of life support than patients with lower Apache II scores (p=0.017). The major reason for the families providing the written consent for “withdrawal of life support” was wishing no more suffering for their illfamily members (58.8%). The major reason for refusing to give the written consent for withdrawal of life support was to have the hope for being alive after the medical treatments (87.5%). The most common withdrawn life support was the ventilator (64.7%). Compared to the patients without providing the written consent, patients with written consent for “withdrawal of life support” received more high-dose inotropic agent and vasopressors (x2=4.253, p= 0.039), continuous intravenous fluid support (x2=4.686, p= 0.03), sedative drugs treatment (x2=5.713, p= 0.017), and ECMO treatment (x2=7.567, p= 0.014) while they received less tracheostomy operations (x2=4.631, p=0.031) and enteral nutrition (x2=11.45, p= 0.001). The results of COX regression model showed that after controlling Apache II scores, compared with the patients without receiving the following life support treatments, the odds ratios of the mortality for the patients who received these treatments were 0.451 for oxygen therapy, 0.275 for enteral nutrition, 0.434 for non-enteral nutrition, 0.625 for continuous intravenous fluid support, 0.487 for blood transfusion, and inotropic agents and vasopressors. Compared with the patients without receiving these treatments, the odds ratios for one-year mortality for the patients who received these treatments were: 0.545 for oxygen therapy, 0.381 for enteral nutrition, 0.562 f or non-enteral nutrition, 0.677 for continuous intravenous fluid support, 0.666 for blood transfusions, and 2.096 for inotropic agents and vasopressors. Theree were no significant differences in the mortality between with and without receiving the treatments of ventilator use (invasive and non-invasive), hemodialysis, antibiotics treatment, blood draws, and sedative drugs use. Conclusions: The incidence rates of providing written consent for withdrawal of life support are low. The main factors associated with providing written consent were younger age and higher levels of disease severity. The main reason for families providing the written consent was wishing no more suffering for their ill family members. The treatments of inotropic agents and vasopressors likely increase the risks of mortalityat the time of ICU and one year after discharging from ICU for the terminally ill patients. The treatments ofoxygen therapy, enteral nutrition, non-enteral nutrition, intravenous fluid support and blood transfusions likely reduce the risk of their mortality. No significant impacts of ventilator use (invasive and non-invasive), hemodialysis, antibiotics treatment, blood draws, and sedative drugs on the risks of mortality were found at the time of ICU and one year after discharge from ICU. In the future, more survival studies on the impacts of life support treatments on the risk of mortality are required to help the terminally ill patients away from ineffective treatments.