晚期肝癌住院病人嚴重藥物不良反應預測因子與存活之探討

Predictive factors and survival for hospitalized patients with advanced hepatocellular carcinoma encountering serious adverse drug reactions

何香毅(Shiang-Yih Ho)

指導教授 ： 黃貴薰

國立臺灣大學/醫學院/護理學研究所/碩士(2012年)

https://doi.org/10.6342/NTU.2012.01935

摘要

【背景】肝癌佔全球癌症發生率的第七位，癌症死因的第三位。在台灣，肝癌佔所有癌症死因的第一位，每年有七千多人因肝癌而死亡；此類患者有高住院率及使用多種治療藥物，而藥物不良反應常是使人罹病、住院及死亡的主要原因之一。
【目的】瞭解晚期肝細胞癌住院病人發生嚴重藥物不良反應之現況、預測因子及存活。
【方法】本研究採橫斷式描述相關性研究設計，以病歷回溯方式及結構式表單進行資料收集。以台灣北部某醫學中心2010年1 月1日至2010年12月31日接受抗腫瘤藥物治療之晚期肝癌住院病人為研究對象。探討嚴重藥物不良反應的現況以及人口學特性、疾病及治療狀況、醫療照護型態對嚴重藥物不良反應的影響、及嚴重藥物不良反應對存活的影響。以“廣義估計方程式（GEE）模式分析嚴重藥物不良反的危險因子和預測因子。以Kaplan-Meier存活分析，分析發生嚴重藥物不良反應之存活時間及整體存活時間及以 Cox迴歸分析及風險比例模型分析死亡的危險因素。
【結果】本研究結果顯示，晚期肝癌住院時嚴重藥物不良反應的發生率為27.74%，而住院中嚴重藥物不良反應發生率為1.29%。發生嚴重藥物不良反應者的平均年齡較大 (60.36歲±SD=13.55 vs. 50.9±SD=15.55)、有門靜脈栓塞 (53.5% vs. 33%)、C型肝炎 (23.3% vs. 7.1%)、Child-Pugh Class B (48.8% vs. 25%)、中度或重度腹水 (25.6% vs. 8%)、心臟病( 4人次 9.3% vs. 2人次1.8%)、其他癌症( 6人次 14.0% vs 8人次7.1%)、或胃腸潰瘍(60.5% vs. 36.6%)者的比率較多，而沒有發生嚴重藥物不良反應者有較高比率的B型肝炎(69% vs. 84.8%)、肺部轉移(20.9% vs. 418%)。
存活時間：研究對象發生嚴重藥物不良反應後之存活時間有差異，但整體存活時間無顯著差異。運用Kaplan-Meier 存活分析結果：(1)發生嚴重藥物不良反應者中位存活時間(17±8.66天，95% CI 0.028- 33.97))，明顯少於沒有發生嚴重藥物不良反應者的中位存活時間(61±9.34天， 95% CI 42.69-79.31), p=0.019。(2)沒有發生嚴重藥物不良反應者之整體存活時間中位數為196±34.43天( 95% CI 128.52-263.47)，而發生嚴重藥物不良反應者之整體存活時間中位數為130±77.48天(95% CI 0.00- 281.86)，Log-Rank (Mantel-Cox)檢定χ2= 2.212, p=0.137。
嚴重藥物不良反應之影響因素：以GEE 逐步迴歸分析結果發現三個嚴重藥物不良反應預測模式，分別為：
(1) 出血嚴重藥物不良反應預測因子：研究對象治療前白蛋白指數越高(OR=5.27, 95% CI 1.42-19.59)、有門靜脈栓塞 (OR=17, 95% CI 3.41-84.79)、腹水越嚴重(OR=4.56, 95% CI 2.12-9.81) 、有心臟病(OR=45.2, 95% CI 8.76-233.11)、有胃腸潰瘍(OR=6.71, 95% CI 1.74-25.78)、有其他癌症 (OR=168.51, 95% CI 22.51-1261.33)、及此次肝癌治療藥物為Sorafenib或其他標靶藥物±節拍式化學治療(OR=10.17, 95% CI 3.37-30.65)，比較會發生出血嚴重藥物不良反應；而有糖尿病者比較不會發生出血嚴重藥物不良反應。Nagelkerke R 2= 0.503，the modified Hosmer-Lemeshow 檢定 p = 0.976，AUC=0.903。
(2) 感染嚴重藥物不良反應預測因子：研究對象治療前的肝功能狀態為Child-Pugh Class B (OR=1.14, 95% CI 0.30-4.31)、有C型肝炎(OR=5.06，95% CI 1.4-18.32)、年齡較大(OR=1.11，95% CI 1.06-1.16)及此次肝癌治療藥物為化療±標靶治療 (OR=2.99, 95% CI 0.63-14.08)者，比較會發生感染嚴重藥物不良反應；而有高血壓(OR=0.04，95% CI. 0.01-0.17)或曾經接受肝癌局部治療(OR=0.22，95% CI 0.07-0.66)者，比較不會發生感染嚴重藥物不良反應。Nagelkerke R 2 = 0.356 (>0.3)，The modified Hosmer-Lemeshow 檢定 p = 0.455 (>0.05)，AUC=0.848。
(3) 發生任一型態的嚴重藥物不良反應預測因子： 研究對象治療前的肝功能狀態為Child-Pugh Class B (OR=5.13, 95% CI 2.28-11.56)、有門靜脈栓塞(OR=2.28，95% CI 1.0-5.22)、α-fetoprotein>400ng/ml (OR=2.26, 95% CI 0.79-6.50)、有心臟病(OR=5.09, 95% CI 0.86-30.32)、年齡較大(OR=1.08，95% CI 1.04-1.11)、及此次肝癌治療藥物為其他標靶±節拍式化學治療(OR=2.44, 95% CI 0.94-6.34)者，比較會發生任一型態的嚴重藥物不良反應；而有高血壓(OR=0.22，95% CI. 0.07-0.68)者，比較不會發生任一型態的嚴重藥物不良反應。Nagelkerke R 2 = 0.324，The modified Hosmer-Lemeshow 檢定 p = 0.591，AUC=0.808。
研究對象之存活的影響因素：以單變項分析發現，研究對象治療前ECOG身體功能狀態>1、曾接受肝癌外科手術、有胃腸潰瘍、及此次肝癌治療為化學±標靶治療與病人的存活相關(p 皆< 0.05)。以Cox迴歸模式分析結果顯示，接受全性藥物治療之晚期肝癌患者，治療前α-FP>400ng/ml (HR=3.942 95% CI 2.04-7.614)或腫瘤>5cm (HR=4.084, 95% CI 1.909-8.735.)、曾經接受肝癌相關外科手術(HR=3.331, 95% CI 1.648-6.743)或全身性治療(HR=2.280, 95% CI 1.184-4.392 )、及發生SADR (HR=2.912, 95% CI 1.595-5.316)，其發生死亡風險相對的比較高。而治療前白蛋白指數越高(HR=0.238, 95% CI 0.127-0.446)或此次肝癌治療為化學治療±標靶治療(HR=0.207, 95% CI 0.067-0.638)，其發生死亡風險相對的比較低。 Adjusted generalized R2=0.5011，Concordance: 0.77 (SE: 0.045)。
【結論】晚期肝癌住院病人嚴重藥物不良反應已是臨床上重要的議題，約有30%的住院與嚴重藥物不良反應有關，且高致死率，主要死亡原因為出血及感染。本研究結果發現不同的嚴重藥物不良反應受多種不同的因素影響，而嚴重藥物不良反應的發生也會影響病人的存活。本研究之結果可以提供臨床醫護人員在照護晚期肝癌病人時的參考依據，以及提供未來發展相關研究及政策的參考，以預防或減少嚴重藥物不良反應的發生。

關鍵字

肝細胞癌 ； 藥物不良反應 ； 危險因子 ； 預測因子 ； 存活

並列摘要

Background: Hepatocellular carcinoma (HCC) is the seventh most common malignancy and the third most common cause of canceru-related deaths worldwide. In Taiwan, HCC is the first leading cause of cancer-related death, accounting for more than 7000 deaths annually. HCC patients usually have high hospitalization rates and use a variety of therapeutic agents; and adverse drug reaction was one of the major causes of hospitalization, morbidity and mortality.
Objective: To understand the current status of serious adverse drug reactions (SADRs), SADRs’ predictors, and survival of advanced HCC inpatients.
Methods: This was a retrospective and cross-sectional descriptive correlational study in which data were retrieved from medical charts. Patients with advanced HCC in treated with anticancer drugs and admitted to the medical center in northern Taiwan between January 1, 2010 and December 31, 2010 were included in this study. To explore the current status of serious adverse drug reactions, “demographic characteristics”, “disease and treatment status”, “medical care patterns” of study population effects on SADRs, and SADRs effects on survival. Risk and predictive factors of SADRs was analysis by using “Generalized estimating equation” (GEE) model. Survival time of post SADRs and overall survival was analyzed by using Kaplan-Meier method. Cox regression and proportional hazard model was used for analyzed the risk factors of mortality.
Results: A total 155 hospitalizations of 81 patients were included, in which 43 hospitalized and 2 prolonged hospitalizations were due to SADRs. The incidence rate of SADRs of advanced HCC inpatients at admission was 27.74%, and during hospitalization was 1.29%. Compared with the “non-serious adverse drug reactions” (non-SADRs) the group had a significantly older average age, high percentage of protal vein thrombosis (PVT), hepatitis B, Child-Pugh Class B, ascites, “gastrointestinal ulcer”(GI ulcer) history, and a low percentage of hepatitis B infection or lung metastasis. In Kaplan-Meier survival analysis, we found that the median survival time after SADRs was significant shorter in the SADRs group compared with non-SADRs group (17±8.66 days, 95%CI 0.028- 33.97 vs. 61±9.34 days, 42.69-79.31, p=0.007), and the median overall survival time was also decreased in SADRs group, but no significant difference (130±77.48 days, 95% CI 0.00- 281.86 vs. 196±34.43days, 95% CI 128.52-263.47, p=0.137).
In multivariate analysis of the SADR predictors by multiple logistic regression model with GEE method, we found three serious adverse drugs reactions predictive models: development of bleedings, infections and any kind of SADRs.
For the development of bleeding SADRs, initial with higher serum albumin level (OR=5.27, 95% CI 1.42-19.59), PVT (OR=17, 3.41-84.79), more serious ascites (OR=4.56, 2.12-9.81), with cardiovascular disease (OR=45.2, 8.76-233.11), GI ulcer (OR=6.71, 1.74-25.78), non- Hodgkin’s lymphoma or gastric cancer history (OR=168.51, 22.51-1261.33), current treatment with Sorafenib base therapy or other target drug with/without metronomic therapy (OR=10.17, 3.37-30.65) pose a high risk for bleeding SADRs, except diabetes mellitus.
For the development of infection SADRs, initial hepatic function was Child-Pugh Class B (OR=1.14, 95% CI 0.30-4.31), with hepatitis C infection (OR=5.06, 1.4-18.32), older (OR=1.11, 1.06-1.16) and current treatment with chemotherapy with/without target drug (OR=2.99, 0.63-14.08) were at a high risk for infection related serious adverse drug reactions, except hypertension and history of local therapy for HCC.
For the development of any kind of SADRs, we found it was almost fully affected by bleeding and infection SADRs model, initial hepatic function was Child-Pugh Class B (OR=5.13, 95% CI 2.28-11.56), PVT (OR=2.28, 1.0-5.22), α-fetoprotein>400ng/ml (OR=2.26, 0.79-6.50), history of cardiovascular disease (OR=5.09, 0.86-30.32), older (OR=1.08, 1.04-1.11), and current treatment with other target drug ± metronomic therapy (OR=2.44, 0.94-6.34) were at a high risks for developing of any serious adverse drug reactions, except hypertension (OR=0.22, 0.07-0.68).
Survival analysis: In univariate analysis, we found that survival rate was significant lower in patient with Eastern Cooperative Oncology Group (ECOG) performance status >1, history of surgery for HCC and with gastrointestinal ulcer, but high in current treatment with chemotherapy with/without target drug (p<0.05). In Cox regression model, initial α-FP >400ng/ml (HR=3.942 95% CI 2.04-7.614), tumor >5cm (HR=4.084, 1.909-8.735.), history of surgery (HR=3.331, 1.648-6.743) or systemic therapy for HCC (HR=2.280, 1.184-4.392 ), and development of a SADR (HR=2.912, 1.595-5.316) were high risks for mortality; in contrast initial with higher serum albumin level or current treatment with chemotherapy ± target drug were at a low risk for mortality..
Conclusion: SADRs of in-patients with advanced HCC has became a critical clinical issue. Approximately 30% of the advanced HCC hospitalizations are due to SADRs, as is high mortality. Bleeding and infection were the two most common cause of SADRs related deaths in our study group. Different SADRs were affected by various factors, and SADRs also affect the survival of patients. Our study results can be a reference for clinical staff who care for patients with advanced HCC, in order to identify the patients who are at risk for admission due to SADRs, to prevent or minimize the incidence of SADRs, as well as for future development of relevant research and policy.

並列關鍵字

Hepatocellular carcinoma ； serious adverse drug reactions ； risk factor ； predictive factor ； survival

參考文獻

林志凌、高嘉宏(2008)•肝癌的流行病學•中華癌醫會誌，24(5)， 277-281。

張家銘, 盧豐華, & 柯文謙. (2005).老年人的感染症概論. 台灣老年醫學雜誌, 民國94年第1卷第2期, 51-60

邵幼雲、林宗哲、洪敏瑛 (2009)•晚期肝細胞癌全身性治療的新趨勢•腫瘤護理雜誌, 9 (增訂刊)，1-12。

Hsu, C., Cheng, J. C., & Cheng, A. L. (2004). Recent advances in non-surgical treatment for advanced hepatocellular carcinoma. Journal of the Formosan Medical Association, 103(7), 483-495.

行政院衛生署國民健康局(2010)•中華民國97年癌症登記報告•2011年7月18日取至：http://www.bhp.doh.gov.tw/Download/97Statistics/1.癌症登記年度報告(全)/ Y97癌症登記年度報告(全).pdf