心臟衰竭病人使用sacubitril-valsartan腎功能惡化與死亡或心臟移植之相關性

The Association between Worsening Renal Function and Death or Heart Transplantation in Heart Failure Patients Treated with Sacubitril-Valsartan

張朝凱(Chao-Kai Chang)

指導教授 ： 徐莞曾

國立臺灣大學/醫學院/臨床藥學研究所/碩士(2021年)

https://doi.org/10.6342/NTU202103985

摘要

研究背景
心臟衰竭 (heart failure) 病人合併慢性腎臟病 (chronic kidney disease, CKD) 或腎功能惡化 (worsening renal function, WRF) 皆具有較高全因性死亡及心臟衰竭住院風險。心臟衰竭治療藥物sacubitril/valsartan (S/V) 於上市前大型臨床試驗中證實能降低射出分率低下心臟衰竭 (heart failure with reduced ejection fraction, HFrEF) 病人之心血管死亡或心臟衰竭住院之綜合事件風險；但在S/V使用過程中可能導致WRF，使醫師不敢貿然調升劑量，甚至可能因而減量或停藥，造成臨床治療之決策困境。然而目前尚無治療指引建議心臟衰竭病人使用S/V發生WRF應如何處置，又相關預後之研究鮮少且有諸多限制。
研究目的
本研究目的為分析心臟衰竭病人啟用S/V前腎功能參數與預後之相關性；以及分析啟用S/V初期不同時間內發生WRF與預後之相關性；接續分析啟用S/V初期肌酸酐 (creatinine, Cre) 上升程度與預後之相關性；最後控制或校正可能影響預後之因素，分析啟用S/V初期發生WRF及追蹤期間腎功能參數與預後之相關性。
研究方法
本研究為回溯性世代追蹤研究，共收納臺大醫院333位於2017年3月1日至2019年2月28日間啟用S/V且用藥 ≥ 30天及左心室射出分率 (left ventricular ejection fraction, LVEF) < 40% 之HFrEF病人，並蒐集其人口學特性、心臟衰竭特性、共病症、腎功能、心臟功能及用藥等參數，其中腎功能參數包含Cre、估算之腎絲球過濾率 (estimated glomerular filtration rate, eGFR)、血清尿素氮 (blood urea nitrogen, BUN) 及BUN與Cre之比值 (BUN/Cre)。本研究參考既有文獻將CKD定義為啟用S/V前eGFR < 60 mL/min/1.73 m2；WRF之腎功能變化程度定義為：(1) eGFR下降 ≥ 20% (WRFeGFR)；(2) Cre上升 ≥ 0.3 mg/dL且 ≥ 25% (WRFCre)；(3) Cre上升 ≥ 30% (Cre30%)；(4) Cre上升 ≥ 50% (Cre50%)；WRF之時間區間定義為：(1) 1個月內 (1M)；(2) 3個月內 (3M)；(3) 6個月內 (6M)。本研究追蹤起日為S/V啟用日，追蹤迄日為2020年12月31日，研究終點為全因性死亡或進行心臟移植之綜合事件。
本研究首先利用Kaplan–Meier method之存活分析合併log-rank test探討啟用S/V前腎功能參數與研究終點之相關性；接續分析啟用S/V初期不同時間內發生WRF及不同Cre上升程度與研究終點之相關性；最後則利用Cox’s proportional hazards model with time-dependent covariates (Cox’s model) 進行多變項分析 (multivariable analysis)，控制可能影響預後之變項，以分析啟用S/V初期發生WRF及追蹤期間腎功能參數與研究終點之相關性。多變項分析共有兩模型，Cox’s model 1納入之變項包含啟用S/V前臨床特性、6M WRFeGFR及追蹤期間用藥；Cox’s model 2則額外納入追蹤期間腎功能相關檢驗數據、心臟功能與住院。
研究結果
本研究333位心臟衰竭病人經過中位數33.1個月之追蹤時間，存活分析顯示合併CKD之病人研究終點風險較高 (log-rank test, p = 0.0167)；Cre、eGFR、BUN與BUN/Cre較其平均值差者研究終點風險亦較高。存活分析亦顯示6M WRFeGFR及6M WRFCre病人具有較高研究終點風險，然而3M WRFeGFR及3M WRFCre病人不具有較高風險。後續針對3個月內之腎功能變化情形探討，存活分析顯示3M Cre50% 病人具有較高研究終點風險 (log-rank test, p = 0.0165)。
在挑選適當變項放入Cox’s model及排除有部分變項缺失之病人後，最終共291人進行多變項分析。Cox’s model 1結果未顯示6M WRFeGFR與研究終點相關，而與研究終點獨立相關之變項共有22個，大致可分為4種類型：(1) 啟用S/V前心臟衰竭特性、共病症與住院史；(2) 啟用S/V前心臟功能；(3) 啟用S/V前用藥；(4) 追蹤期間用藥或醫療處置，包含S/V日劑量 (hazard ratio [HR], 0.9890; 95% confidence interval [CI], 0.9846–0.9935; p < 0.0001)、環利尿劑 (loop diuretics) 平均日劑量 (HR, 1.0160; 95% CI, 1.0060–1.0261; p = 0.0017)、doxazosin總累積劑量 (HR, 1.0013; 95% CI, 1.0004–1.0021; p = 0.0045) 等。
Cox’s model 2結果亦未顯示6M WRFeGFR與研究終點相關，而與研究終點獨立相關或接近統計顯著差異之變項共有11個，大致可分為4種類型：(1) 啟用S/V前共病症；(2) 追蹤期間腎功能相關檢驗數據，包含每7天Cre變化百分比 (HR, 1.0150; 95% CI, 1.0103–1.0198; p < 0.0001) 及BUN/Cre平均值 (HR, 1.0357; 95% CI, 0.9999–1.0728; p = 0.0506) 等；(3) 追蹤期間心臟功能，包含LVEF (HR, 0.8993; 95% CI, 0.8623–0.9378; p < 0.0001) 及每月LVEF變化 (HR, 0.9423; 95% CI, 0.8982–0.9886; p = 0.0151) 等；(4) 追蹤期間用藥或醫療處置，包含S/V日劑量 (HR, 0.9941; 95% CI, 0.9901–0.9981; p = 0.0037) 及nicorandil日劑量 (HR, 1.2195; 95% CI, 1.0724–1.3867; p = 0.0025) 等。
結論
啟用S/V前後皆須定期監測腎功能，且應同時檢測Cre與BUN。啟用S/V後3個月內Cre上升 < 50% 為可接受之範圍。然而腎功能變化與預後之相關性同時受到心臟功能、S/V及利尿劑等因素影響，因此除了評估腎功能變化程度，更重要的是應同時注意心臟功能、利尿劑與血管擴張劑等影響預後之危險因子，並避免心臟功能惡化、低血壓、renal hypoperfusion以及neurohormonal activation等可能導致預後不佳之情形。若能維持理想血壓與體液狀態，S/V便有機會調升至最佳可耐受劑量，進而獲得較佳之預後。

關鍵字

心臟衰竭 ； 腎功能惡化 ； sacubitril/valsartan ； 肌酸酐 ； 血清尿素氮 ； 全因性死亡 ； 心臟移植

並列摘要

Background
Chronic kidney disease (CKD) and worsening renal function (WRF) in heart failure (HF) were associated with an increased risk of all-cause death and hospitalization for HF. Sacubitril/valsartan (S/V) reduced the risk of cardiovascular death and hospitalization for HF in patients with heart failure and reduced ejection fraction (HFrEF). However, after the initiation or titration of S/V, WRF may occur and then contribute to the therapeutic dilemma. Unfortunately, there was no definite recommendation about S/V-related WRF in the guidelines, as well as studies analyzing the prognosis in this condition were scarce and full of limits.
Objective
We have three main aims in the present study. First, we want to analyze the association between renal function parameters before initiating S/V and the prognosis in patients with HF receiving S/V. Second, we aim to investigate the prognosis of early WRF in various periods of time and to various extents. Lastly, we intend to assess the independent association between early WRF, renal function parameters during follow-up, and the prognosis.
Methods
We performed a retrospective cohort study that included 333 patients prescribed S/V ≥ 30 days with HFrEF and left ventricular ejection fraction (LVEF) < 40% in National Taiwan University Hospital (NTUH) between March 1, 2017, and February 28, 2019. Then, we collected clinical parameters, e.g., comorbidities, renal function, cardiac function, and medications. The renal function contained creatinine (Cre), estimated glomerular filtration rate (eGFR), blood urea nitrogen (BUN), and BUN-to-Cre ratio (BUN/Cre). We defined CKD as eGFR < 60 mL/min/1.73 m2. The definitions of WRF extent were (1) a decrease in eGFR ≥ 20% (WRFeGFR); (2) an increase in Cre ≥ 0.3 mg/dL and ≥ 25% (WRFCre); (3) an increase in Cre ≥ 30% (Cre30%); (4) an increase in Cre ≥ 50% (Cre50%). We also defined the WRF period as (1) within one month (1M); (2) within three months (3M); (3) within six months (6M). The follow-up period was from the date of initiating S/V to December 31, 2020. The study endpoint was the composite event of all-cause death or heart transplantation.
Survival analysis was used to analyze the association between renal function parameters before initiating S/V and the prognosis, followed by investigating the prognosis of early WRF. Cox’s proportional hazards model with time-dependent covariates (Cox’s model) was used to assess the independent association between early WRF, renal function parameters during follow-up, and prognosis. Cox’s model 1 enrolled variables including baseline characteristics, 6M WRFeGFR, and medications during follow-up; Cox’s model 2 also enrolled variables including renal function-related laboratory data, cardiac function, and hospitalizations.
Results
Through a median follow-up of 33.1 months, survival analysis indicated patients with CKD had a higher risk of the study endpoint compared to patients without CKD (log-rank test, p = 0.0167). Patients with 6M WRFeGFR and 6M WRFCre had a higher incidence of the study endpoint than those without 6M WRFeGFR and 6M WRFCre, respectively. The aforementioned poor prognosis did not occur in patients with 3M WRFeGFR and 3M WRFCre. However, patients with 3M Cre50% were at risk of more study endpoints compared with those without 3M Cre50% (log-rank test, p = 0.0165).
There were 291 patients analyzed in Cox’s models. Cox’s model 1 did not indicate that 6M WRFeGFR was associated with the study endpoint. Instead, there were 22 independent predictors in this model, divided into four categories: (1) baseline HF characteristics, comorbidities, and hospitalizations; (2) baseline cardiac function; (3) medications before initiating S/V; (4) medical treatments during follow-up. Significantly, patients with higher loop diuretics average daily dose and doxazosin accumulative dose during follow-up had a higher risk of the study endpoint; patients with higher S/V daily dose during follow-up were at lower risk.
Cox’s model 2 also did not present 6M WRFeGFR as a predictor of the prognosis. There were 11 predictors in this model, divided into four categories: (1) baseline comorbidity; (2) renal function-related laboratory data during follow-up; (3) cardiac function during follow-up; (4) medical treatments during follow-up. Primarily, higher Cre change percentages per 7 days, BUN/Cre average value, and nicorandil daily dose during follow-up predicted a worse prognosis; higher LVEF, LVEF change per month, and S/V daily dose during follow-up predicted fewer study endpoints.
Conclusion
After initiating S/V, we should monitor Cre and BUN regularly in patients with HF; an increase in Cre < 50% within three months could be acceptable. However, cardiac function and medications may influence the prognosis of renal function change. Therefore, we should evaluate renal function change and concurrently pay attention to deterioration in cardiac function, inappropriate diuretics, and vasodilators. If patients maintain stable blood pressure and optimal volume status, we will have the potential to titrate S/V to the maximum tolerated dose.

並列關鍵字

heart failure ； worsening renal function ； sacubitril/valsartan ； creatinine ； blood urea nitrogen ； all-cause death ； heart transplantation

參考文獻

Fernandez-Jimenez R, Hoit BD, Walsh RA, et al. Normal physiology of the cardiovascular system. In: Fuster V, Harrington RA, Narula J, eds. Hurst's The Heart, 14e. New York, NY: McGraw-Hill Education; 2017.

Overview of the cardiovascular system. In: Mohrman DE and Heller LJ. Cardiovascular Physiology, 9e. New York, NY: McGraw-Hill Education; 2018.

Thomas DL, Lythgoe MF, Pell GS, et al. The measurement of diffusion and perfusion in biological systems using magnetic resonance imaging. Physics in medicine and biology. 2000;45:R97.

Parker RB, Nappi JM, and Cavallari LH. Chronic heart failure. In: DiPiro JT, Talbert RL, Yee GC, eds. Pharmacotherapy: A Pathophysiologic Approach, 10e. New York, NY: McGraw-Hill Education; 2017.

Gaggin HK and Dec GW. Pathophysiology of heart failure. In: Fuster V, Harrington RA, Narula J, eds. Hurst's The Heart, 14e. New York, NY: McGraw-Hill Education; 2017.