透過您的圖書館登入
IP:54.221.159.188
  • 學位論文

利用nomogram分析缺血性中風的危險因子及預測治療預後

Risk Factors Analysis and Treatment Prognosis Predictions of Stroke by Nomogram

指導教授 : 邱泓文

摘要


背景 腦中風會造成殘障及死亡;是世界上第三大死因,預防中風的可能因子對中風的照顧是相當重要的事情。中風與傳統動脈硬化因子相關,但有些病人並沒有明顯血管阻塞的危險因子,所以必須找出其他可調整的危險因子,給予適當的預防措施以達到預防中風的效果。關於急性缺血性中風的治療,發病三小時內可以使用靜脈溶栓治療,雖可以改善預後,但仍然有一些預後不良的病患。當病患有嚴重頸動脈狹窄的狀況,可以選擇頸動脈支架來預防後續腦梗塞發生。找出中風的危險因子及治療預後的預測可以幫助臨床預防與治療的決策。為了找出少見但可治療的危險因子及預後狀況,我們分析了懷孕及敗血症造成中風的情形,並建立靜脈溶栓治療預後和頸動脈支架危險因子的分析,以了解治療後恢復及心血管不良預後的機率。 先前文獻發現懷孕造成中風、心肌梗塞或靜脈栓塞的情形會影響到產後12周。在中老年婦女的回溯性研究發現懷孕胎次越多越容易造成心血管事件,然而亞洲族群的資料較為不足。敗血症會導致在住院期間及長期缺血性腦梗塞及死亡。而新發生的心房顫動是敗血症造成缺血性腦梗塞的重要因子。敗血症病患出院後造成缺血性腦梗塞短期再住院與新發生的心房顫動的關係尚待研究。敗血症亦會造成心血管事件發生,然而長期敗血症發生缺血性腦梗塞的機率仍需進一步研究。於發病3小時內的急性缺血性腦梗塞使用靜脈溶栓治療可以改善預後,然而之前研究只分辨預後好或不好,並未顯示預後不良的機會;必須找出恢復不良者給予積極的動脈內治療。頸動脈支架用來預防嚴重性頸動脈狹窄的病患發生續發性腦梗塞。必須使用方法找出適合的病患以增加預防效果,當心血管事件發生率過高的病患強烈建議藥物治療。 材料及方法 本研究使用台灣健保資料庫歸人檔,觀察懷孕婦女發生中風,敗血症長期追蹤發生缺血性腦梗塞的情形。以2010的台灣健保資料庫全住院檔研究敗血症出院後短期發生缺血性腦梗塞的情形。並利用2011-2015三軍總醫院靜脈溶栓治療病患的資料建立模型,並使用國軍桃園總醫院靜脈溶栓治療病患驗證模型之表現。因為頸動脈支架治療於2004年7月通過健保給付,因此我們使用2004-2009的台灣健保資料庫歸人檔,觀察頸動脈支架治療的病患發生心血管事件的狀況,以建立預後模型,並使用2004-2014三軍總醫院頸動脈支架治療的病患來驗證建立模組的正確性。 我們運用邏輯式回歸及存活分析找出各種影響因子,建立預測模型。頸動脈支架手術後的預後、懷孕婦女發生中風、長期敗血症發生缺血性腦梗塞的研究有經過時間的追蹤,利用存活分析找出有統計差異的相關因子,建立危險因子。敗血症存活者新發生心房顫動並再入院缺血性腦梗塞的因子、靜脈溶栓治療預後使用多變項邏輯式回歸的往前步驟去找出個別因子的勝算比。所使用的統計軟體為SPSS version 18。Nomogram使用R軟體的 rms 套件,利用各因子的分數加總以建立長期敗血症發生缺血性腦梗塞的機率、靜脈溶栓治療的預後機會及頸動脈支架手術後發生心血管事件的可能性。 結果 雖然懷孕中風的發生率並不高,懷孕會在產後一年內造成明顯增加中風發生機會,但長期追蹤10年的中風發生率卻較低。如果以胎次來看,第三胎及第四胎較第一胎易發生中風。多變量邏輯式回歸分析發現約3個月短期追蹤的敗血症存活者造成缺血性中風,新發生心房顫動的勝算比為1.74(95% C.I.:1.255-2.413)。在長期追蹤的敗血症存活者中年齡、新發生心房顫動、高血壓、糖尿病、冠心症、慢性腎臟病、慢性阻塞性肺病及地區醫院住院和後續發生缺血性中風有關。並以nomogram建立後續發生缺血性中風機率。我們以年齡及中風量表分數建立簡單nomogram來預測靜脈溶栓治療預後,驗證表現良好。而頸動脈支架治療的病患發生心血管事件的危險因子為年齡、充血性心衰竭、惡性腫瘤、糖尿病及有臨床症狀的病患,並以這5個因子建立nomogram,並以單一醫院追蹤一年的資料驗證,獲得可接受的模型表現。 結論 臨床醫師必須了解中風之可治療的危險因子,提供預防的策略。其中懷孕增加中風機會影響至產後一年,經由研究結果,必須告知懷孕婦女可能發生中風的風險,控制懷孕併發症以降低中風發生,以達到預防懷孕相關中風的目的。敗血症存活者會因新發生心房顫動及其他共病症造成出院後短期及長期缺血性中風發生,故須告知敗血症存活病患,並積極提供中風預防策略。新發生心房顫動病人可經由抗凝血劑治療達到預防中風效果,在敗血症存活者則尚待進一步研究。整合性方法可以支援醫師與病患於靜脈溶栓治療及頸動脈支架治療前的溝通,以降低焦慮及協助臨床治療決策,減少醫院內的治療延遲。當發現靜脈溶栓治療預後不良病患須建議更積極的動脈內溶栓治療。頸動脈支架治療的年老病患合併多種共病症會造成發生心血管事件的危險過高時,建議內科藥物治療。我們利用nomogram建立視覺化的分數量表以提供中風相關發生機率及預後資訊,可以方便在臨床上使用。

並列摘要


Background Since stroke is the third leading disease that causes mortality throughout the world, the prevention of stroke from potential risk factors is important issue. Stroke related to traditional risk factors, but it still have some uncommon risk factors. To perform risk factors analysis and treatment outcome prediction can help clinical prevention and treatment strategies. To find out treatable risk factors, we designed studies to evaluate stroke linked to the conditions of pregnancy, septicemia and outcome predictions after intravenous thrombolysis and carotid stent treatment. The pregnancy related stroke was noted within postpartum 12 weeks at most. The previous study found more deliveries with more risk of cardiovascular risk in middle and older-aged retrospective studies. But there is lack of population data of pregnant related stroke in Asia. Stroke related to septicemia within hospitalization and long-term follow up. New-onset atrial fibrillation (AF) is an important factor of ischemic stroke in sepsis. But it is lack of evidence of short-term ischemic stroke risk of new-onset atrial fibrillation after sepsis discharge, we find out the relationship between ischemic stroke and new-onset AF within 3 months. In addition, we designed a study to realize the ischemic stroke chances of septicemia survivors in long-term period with nomogram. Intravenous thrombolysis treatment for acute ischemic stroke can reduce disability and increase survival rate. Previous studies just classify well or poor recovery situations without quantifying degree of prognosis, we constructed functional recovery probabilities. When patients with more than 50% carotid stenosis will increase ischemic stroke risk, but high risk patients it may cause some events after the carotid stent (CAS) procedure that reduces the benefit of stroke prevention. We tried to develop prognostic model to discriminate high risk patients to enhance CAS preventive effect. Materials and methods We investigated the period of pregnant related stroke, long-term ischemic stroke risk of septicemia survivors, major adverse cardiovascular events (MACE) of carotid stent using longitudinal National Health Insurance Database (LHID). The short-term ischemic stroke in septicemia survivors was using inpatient dataset of National Health Insurance Research Dataset (NHIRD) in 2010. Intravenous thrombolysis prognosis prediction was using data in Tri-Service General Hospital from 2011-2015, and validated using data in Army Taoyuan General Hospital. We validated to MACE of carotid stent using data from Tri-service general hospital from 2004-2014. The statistical analysis performed by SPSS version 18. The multivariate Cox regression was performed for pregnant related stroke, long-term ischemic stroke risk of septicemia survivors, MACE of carotid stent. The multivariate logistic regression was assessed for short-term ischemic stroke in septicemia survivors and intravenous thrombolysis prognosis. The nomogram based on factors on multivariate Cox and logistic regression in long-term ischemic stroke risk of septicemia survivors, intravenous thrombolysis prognosis and MACE of carotid stent. The nomogram constructed by R language rms package. Results The pregnancy related stroke is rare frequency, but we found increasing risk within one year after delivery and reducing risk for long-term period. The women with more than 3 deliveries carried more risk. The risk of short-term IS after septicemia survivors was noticed adjusted OR 1.74 (95% C.I.: 1.26-2.41) for new-onset atrial fibrillation within 3 months. The advanced age, new-onset atrial fibrillation, hypertension, diabetes mellitus, coronary artery disease, chronic kidney disease, chronic obstructive pulmonary disease and local hospital admission related to subsequent long-term ischemic stroke and developed subsequent ischemic stroke chances over 6 years by nomogram. The intravenous thrombolysis was performed acute ischemic stroke within 3 hours, although some prediction scores were constructed without probabilities definition. We used age and National Institute of Health Stroke Scale (NIHSS) score to develop simple nomogram from one medical center and validated with good performance in one community hospital. Poor prognosis patients were suggested intra-artery procedure with recanalization. We constructed nomogram for major adverse cardiovascular disease risk analysis of carotid stent in population dataset and validated with dataset of one medical center showed acceptable performance. The risk factors included advance age, congestive heart failure, malignant disease, diabetes mellitus, and symptomatic status. The advanced age with more comorbid conditions carried out more risk. Before carotid stent procedure, high risk patients were suggested conservative treatment. Conclusions The clinicians must understand other uncommon treatable risk factors, preventive strategies must perform. We investigated pregnancy increasing stroke risk within one year after delivery; the new-onset atrial fibrillation is a mediator of ischemic stroke in septicemia survivors. We developed nomograms with a visual scale method and prognostic information, and it is easy to use in clinical practice. The integer-base method may support communication between clinicians and patients before intravenous thrombolysis and CAS in order to reduce the anxiety about making a treatment decision. However, insofar as elderly patients with multiple comorbidities are at high risk, the option of an alternative treatment strategy with conservative therapy should be suggested. The findings need prospective study to confirm the performance in the future.

參考文獻


Asuzu, D., Nyström, K., Schindler, J., Wira, C., Greer, D., Halliday, J., et al. (2015). TURN score predicts 90-day outcome in acute ischemic stroke patients after IV thrombolysis. Neurocritical care, 23(2), 172-178.
Ban, L., Sprigg, N., Sultan, A. A., Nelson‐Piercy, C., Bath, P. M., Ludvigsson, J. F., et al. (2017). Incidence of First Stroke in Pregnant and Nonpregnant Women of Childbearing Age: A Population‐Based Cohort Study From England. Journal of the American Heart Association, 6(4), e004601.
Bremme, K. A. (2003). Haemostatic changes in pregnancy. Best practice & research Clinical haematology, 16(2), 153-168.
Brott, T. G., Hobson, R. W., Howard, G., Roubin, G. S., Clark, W. M., Brooks, W., et al. (2010). Stenting versus endarterectomy for treatment of carotid-artery stenosis. New England Journal of Medicine, 363(1), 11-23.
Cheng, C.-A., Cheng, C.-G., Lin, H.-C., Lee, J.-T., Cheng, C.-C., Chien, W.-C., et al. (2017). New-onset atrial fibrillation-related ischemic stroke occurring after hospital discharge in septicemia survivors. QJM: An International Journal of Medicine, 110(7), 453-457.

延伸閱讀