Heart failure and a preserved ejection fraction (HFpEF) is characterized in many patients by the coexistence of a systemic metabolic or inflammatory disorder that causes coronary endothelial dysfunction, microvascular rarefaction, and cardiac fibrosis, leading to impaired left ventricular distensibility. Neurohormonal antagonists that are effective in patients with heart failure and a reduced ejection fraction have generally not been useful in patients with HFpEF. Until recently, large-scale trials of patients with HFpEF have reported no benefit or only a modest reduction in the risk of heart failure outcomes, with borderline levels of statistical significance.^1,2 The trials have also noted meaningful subgroup interactions, which have further complicated interpretation of the results.

The EMPEROR-Preserved (Empagliflozin Outcome Trial in Patients With Chronic Heart Failure With Preserved Ejection Fraction) investigators recently reported the effects of the sodium-glucose cotransporter 2 inhibitor empagliflozin in patients with heart failure and an ejection fraction >40%.³ Empagliflozin reduced the risk of the composite of cardiovascular death or hospitalization for heart failure by 21% (hazard ratio, 0.79 [95% CI, 0.69–0.90]; P<0.001), which was related mainly to a 29% decrease in the risk of hospitalization for heart failure. The effect on the primary end point was consistent across all prespecified subgroups. The magnitude of the effect on heart failure hospitalizations with empagliflozin in HFpEF was similar to that previously reported in patients with heart failure and a reduced ejection fraction (29% versus 31% risk reduction, respectively).^3,4

Are the results of EMPEROR-Preserved truly different from the results of earlier trials in HFpEF? Two authors of this commentary (M.P. and F.Z.) served on the Executive Committee of both the PARAGON-HF trial (Prospective Comparison of ARNI with ARB Global Outcomes in HF With Preserved Ejection Fraction) with sacubitril/valsartan and the EMPEROR-Preserved trial with empagliflozin; thus, we are well positioned to examine the findings of the 2 landmark HFpEF trials.

The PARAGON-HF and EMPEROR-Preserved trials were designed to recruit similar groups of patients. Both trials required patients to have symptoms of heart failure and an ejection fraction indicative of HFpEF, defined as ≥45% in PARAGON-HF and >40% in EMPEROR-Preserved. Both trials specified that patients had evidence of structural heart disease (typically left atrial enlargement) and increased levels of natriuretic peptides, using the same threshold for NT-proBNP (N-terminal pro-B-type natriuretic peptide; ie, >300 pg/mL). The baseline characteristics were similar in the 2 trials for most variables, including renal function and prevalence of diabetes. Because of differences in ejection fraction eligibility, compared with EMPEROR-Preserved, patients in the PARAGON-HF trial had higher ejection fractions (57.5±8.0% versus 54.3±8.8%). Compared with PARAGON-HF, patients in the EMPEROR-Preserved trial were somewhat more likely to be treated with β-blockers (85.9% versus 79.5%) and mineralocorticoid receptor antagonists (37.6% versus 27.1%).

Although PARAGON-HF is often referred to as an active control trial, the study compared sacubitril/valsartan with valsartan, thus allowing an unconfounded comparison of neprilysin inhibition in 1 group with no neprilysin inhibition in the other. Similarly, EMPEROR-Preserved compared patients receiving empagliflozin in 1 group with patients receiving no sodium-glucose cotransporter 2 inhibitor in the other.

The PARAGON-HF trial randomized 4796 patients who were followed up for a median of 35 months, whereas the EMPEROR-Preserved trial randomized 5988 patients who were followed up for a median of 26 months. For the composite of cardiovascular death and total hospitalizations for heart failure, the incidence rates for the control groups were comparable, that is, 14.6 versus 12.5 per 100 patient-years of follow-up in PARAGON-HF and EMPEROR-Preserved, respectively. However, because of its longer follow-up, more of these events were recorded in PARAGON-HF than in EMPEROR-Preserved. The PARAGON-HF trial recorded more total (first and recurrent) hospitalizations for heart failure than EMPEROR-Preserved (1487 versus 948) and more deaths (691 versus 463), thus providing PARAGON-HF with more statistical power to discern a treatment effect. In PARAGON-HF, the addition of neprilysin inhibition reduced the rate ratio for the composite of cardiovascular death and total hospitalizations for heart failure by 13% (rate ratio, 0.87 [95% CI, 0.75–1.01]; P=0.059).² In contrast, in EMPEROR-Preserved, the addition of empagliflozin reduced the rate ratio for the same end point by 21% (rate ratio, 0.79 [95% CI, 0.68–0.92]; P=0.003). For the primary end point of time to cardiovascular death or hospitalization for heart failure, the effect size for empagliflozin was identical (hazard ratio, 0.79 [95% CI, 0.69–0.90]; P<0.0001).³

Comparison of effect sizes across trials is fraught with difficulties because of potential differences in patient populations or end points. Therefore, to evaluate the results of PARAGON-HF and EMPEROR-Preserved appropriately, we examined the effect of active treatment on the same end points using the same definition of ejection fraction subgroups (Figure). The cut points for the ejection fraction subgroups were defined by the PARAGON-HF investigators in an earlier publication,⁵ and we reanalyzed the data from EMPEROR-Preserved to match these cut points. We analyzed 4 conventional end points, which focused on cardiovascular death and heart failure hospitalizations, analyzed individually or as a composite.

Figure. Effect of randomized treatment in the PARAGON-HF trial (Prospective Comparison of ARNI with ARB Global Outcomes in HF With Preserved Ejection Fraction) and EMPEROR-Preserved trial (Empagliflozin Outcome Trial in Patients With Chronic Heart Failure With Preserved Ejection Fraction) according to ejection fraction in subgroups previously defined by the PARAGON-HF investigators. Effect of neprilysin inhibition with sacubitril/valsartan in PARAGON-HF side by side with the effect of sodium-glucose cotransporter 2 inhibition with empagliflozin in EMPEROR-Preserved for 4 conventional end points: (1) the composite of cardiovascular death (analyzed as time to first event), (2) hospitalization for heart failure (time to event), (3) total (first and recurrent) hospitalizations for heart failure, and (4) cardiovascular death. Results from PARAGON-HF and EMPEROR-Preserved are shown in blue and red, respectively. Number of patients in each subgroup for both treatment groups combined is as follows: for PARAGON-HF, >42.5% to ≤52.5% (n=1427), >52.5% to ≤62.5% (n=2166), and >62.5% (n=1202); for EMPEROR-Preserved, >42.5% to ≤52.5% (n=2340), >52.5% to ≤62.5% (n=2130), and >62.5% (n=1106).

Neither sacubitril/valsartan nor empagliflozin exerted a significant effect on cardiovascular death. Nevertheless, for all outcomes that included hospitalization for heart failure, the effect size was larger for empagliflozin than for sacubitril/valsartan in most ejection fraction subgroups (Figure). Of particular note, with respect to time to first hospitalization for heart failure, for patients with an ejection fraction of >42.5% to ≤52.5%, the hazard ratios were 0.83 (95% CI, 0.65–1.06) for PARAGON-HF and 0.65 (95% CI, 0.50–0.85) for EMPEROR-Preserved. For this same end point, for patients with an ejection fraction of >52.5% to ≤62.5%, the hazard ratios were 0.87 (95% CI, 0.71–1.07) for PARAGON-HF and 0.68 (95% CI, 0.51–0.89) for EMPEROR-Preserved.

A side-by-side examination of the pattern of effects in 2 large-scale outcomes trials in patients with HFpEF affords an opportunity to evaluate the benefits of neprilysin inhibition and sodium-glucose cotransporter 2 inhibition using the same end points in the same ejection fraction subgroups in comparable patient populations. The magnitude of the reduction in the risk of serious heart failure outcomes appears to be greater with sodium-glucose cotransporter 2 inhibition than with neprilysin inhibition for most patients with HFpEF.

The PARAGON-HF trial was funded by Novartis, and the EMPEROR-Preserved trial was funded by Boehringer Ingelheim and Eli Lilly and Co.

Disclosures Dr Packer reports consulting fees from Abbvie, Actavis, Amgen, Amarin, AstraZeneca, Boehringer Ingelheim, Bristol Myers Squibb, Casana, CSL Behring, Cytokinetics, Johnson & Johnson, Lilly, Moderna, Novartis, ParatusRx, Pfizer, Relypsa, Salamandra, Synthetic Biologics, and Theravance, outside the submitted work. He was a member of the Executive Committee of the PARAGON-HF and EMPEROR-Preserved trials. Dr

Zannad has received steering committee or advisory board fees from Amgen, AstraZeneca, Bayer, Boehringer Ingelheim, Boston Scientific, Cardior, CVRx, Janssen, Livanova, Merck, Mundipharma, Novartis, Novo Nordisk, and Vifor Fresenius. He was a member of the Executive Committee of the PARAGON-HF and EMPEROR-Preserved trials.

Dr Anker reports grants and personal fees from Vifor International and Abbott Vascular, as well as personal fees from AstraZeneca, Bayer, Brahms, Boehringer Ingelheim, Cardiac Dimensions, Novartis, Occlutech, Servier, and Vifor International.

The opinions expressed in this article are not necessarily those of the editors or of the American Heart Association.

https://www.ahajournals.org/journal/circ

The podcast and transcript are available as a Data Supplement at https://www.ahajournals.org/doi/suppl/10.1161/CIRCULATIONAHA.121.056657.

This work was presented in a Late-Breaking Clinical Trial session at the European Society of Cardiology Congress, August 27, 2021 to August 30, 2021.

For Sources of Funding and Disclosures, see page 1195.

在许多患者中，心力衰竭和射血分数保留 (HFpEF) 的特征是全身代谢或炎症疾病并存，导致冠状动脉内皮功能障碍、微血管稀疏和心脏纤维化，导致左心室扩张性受损。对射血分数降低的心力衰竭患者有效的神经激素拮抗剂通常对 HFpEF 患者无效。直到最近，对 HFpEF 患者进行的大规模试验报告说，心力衰竭结果的风险没有益处或仅适度降低，具有统计学意义的临界水平。^1,2试验还注意到有意义的亚组相互作用，这使得对结果的解释更加复杂。

EMPEROR-Preserved（射血分数保留的慢性心力衰竭患者的恩格列净结局试验）研究人员最近报告了钠-葡萄糖协同转运蛋白 2 抑制剂恩格列净对射血分数 > 40% 的心力衰竭患者的影响。³ Empagliflozin 将心血管死亡或因心力衰竭住院的复合风险降低了 21%（风险比，0.79 [95% CI，0.69–0.90]；P<0.001)，这主要与因心力衰竭住院的风险降低 29% 相关。在所有预先指定的亚组中，对主要终点的影响是一致的。HFpEF 中恩格列净对心力衰竭住院治疗的影响程度与之前报道的心力衰竭和射血分数降低患者相似（风险分别降低 29% 和 31%）。^3,4

EMPEROR-Preserved 的结果与早期 HFpEF 试验的结果真的不同吗？本评论的两位作者（MP 和 FZ）曾在 PARAGON-HF 试验（ARNI 与 ARB 保留射血分数保留的 HF 总体结果的前瞻性比较）与沙库巴曲/缬沙坦和 EMPEROR-Preserved 试验与恩格列净的执行委员会任职; 因此，我们可以很好地检查 2 项具有里程碑意义的 HFpEF 试验的结果。

PARAGON-HF 和 EMPEROR-Preserved 试验旨在招募相似的患者组。两项试验都要求患者有心力衰竭症状和射血分数表明 HFpEF，定义为 PARAGON-HF ≥45% 和 EMPEROR-Preserved ≥40%。两项试验均使用 NT-proBNP（N 端前 B 型利钠肽；即 >300 pg/毫升）。在 2 项试验中，大多数变量的基线特征相似，包括肾功能和糖尿病患病率。由于射血分数资格的差异，与 EMPEROR-Preserved 相比，PARAGON-HF 试验中的患者具有更高的射血分数（57.5±8.0% 对 54.3±8.8%）。

尽管 PARAGON-HF 通常被称为主动对照试验，但该研究将沙库巴曲/缬沙坦与缬沙坦进行了比较，因此可以对一组中的脑啡肽酶抑制与另一组中的脑啡肽酶抑制进行无混杂的比较。同样，EMPEROR-Preserved 将一组接受恩格列净的患者与另一组未接受钠-葡萄糖协同转运蛋白 2 抑制剂的患者进行了比较。

PARAGON-HF 试验随机分配了 4796 名随访中位时间为 35 个月的患者，而 EMPEROR-Preserved 试验随机分配了 5988 名随访中位时间为 26 个月的患者。对于心血管死亡和心力衰竭总住院率的复合数据，对照组的发生率相当，即在 PARAGON-HF 和 EMPEROR-Preserved 中，每 100 患者年随访分别为 14.6 和 12.5。然而，由于其更长的随访时间，PARAGON-HF 中记录的这些事件多于 EMPEROR-Preserved。与 EMPEROR-Preserved 相比，PARAGON-HF 试验记录的心力衰竭住院总数（首次和复发）更多（1487 人对 948 人）和更多死亡（691 人对 463 人），从而为 PARAGON-HF 提供了更多的统计能力来辨别治疗效果。在 PARAGON-HF 中，P = 0.059）。²相比之下，在 EMPEROR-Preserved 中，添加 empagliflozin 使相同终点的比率降低了 21%（比率比率，0.79 [95% CI，0.68–0.92]；P = 0.003）。对于心血管死亡或因心力衰竭住院的主要终点时间，恩格列净的效应大小相同（风险比，0.79 [95% CI，0.69–0.90]；P <0.0001）。³

由于患者人群或终点的潜在差异，跨试验比较效应大小充满困难。因此，为了适当评估 PARAGON-HF 和 EMPEROR-Preserved 的结果，我们使用相同的射血分数亚组定义检查了积极治疗对相同终点的影响（图）。射血分数亚组的分界点由 PARAGON-HF 研究人员在较早的出版物^{5 中定义}，我们重新分析了来自 EMPEROR-Preserved 的数据以匹配这些分界点。我们分析了 4 个传统终点，重点是心血管死亡和心力衰竭住院，单独分析或作为复合分析。

数字。 PARAGON-HF 试验（射血分数保留的 HF 的 ARNI 与 ARB 总体结果的前瞻性比较）和 EMPEROR-Preserved 试验（射血分数保留的慢性心力衰竭患者的恩格列净结果试验）中随机治疗的效果根据射血分数在 PARAGON-HF 研究人员先前定义的亚组中。PARAGON-HF 中沙库巴曲/缬沙坦脑啡肽酶抑制作用与 EMPEROR-Preserved 中钠-葡萄糖协同转运蛋白 2 抑制作用与恩格列净对 4 个常规终点的影响：（1）心血管死亡的复合（分析为时间到首次事件），（2）因心力衰竭住院（事件发生时间），（3）因心力衰竭住院（首次和复发），以及（4）心血管死亡。PARAGON-HF 和 EMPEROR-Preserved 的结果分别以蓝色和红色显示。合并两个治疗组的每个亚组中的患者数如下：对于 PARAGON-HF，>42.5% 至 ≤52.5% (n=1427)、>52.5% 至 ≤62.5% (n=2166) 和 >62.5% (n=1202); 对于 EMPEROR-Preserved，>42.5% 至 ≤52.5% (n=2340)、>52.5% 至 ≤62.5% (n=2130) 和 >62.5% (n=1106)。

沙库巴曲/缬沙坦和恩格列净均未对心血管死亡产生显着影响。然而，对于包括因心力衰竭住院在内的所有结局，在大多数射血分数亚组中，恩格列净的效果大于沙库巴曲/缬沙坦（图）。特别值得注意的是，对于射血分数 >42.5% 至 ≤52.5% 的患者，至首次因心力衰竭住院的时间，PARAGON-HF 的风险比为 0.83（95% CI，0.65-1.06）和 0.65 (95% CI, 0.50–0.85) 对于 EMPEROR-Preserved。对于同一终点，对于射血分数 >52.5% 至 ≤62.5% 的患者，PARAGON-HF 的风险比为 0.87（95% CI，0.71–1.07）和 0.68（95% CI，0.51–0.89）为 EMPEROR-Preserved。

在 HFpEF 患者中进行的 2 项大规模结果试验中的效应模式的并排检查提供了一个机会，使用相同的射血分数中的相同终点来评估脑啡肽酶抑制和钠-葡萄糖协同转运蛋白 2 抑制的益处可比患者群体中的亚组。对于大多数 HFpEF 患者而言，钠-葡萄糖协同转运蛋白 2 抑制剂在降低严重心力衰竭风险方面的幅度似乎大于脑啡肽酶抑制剂。

PARAGON-HF 试验由诺华资助，EMPEROR-Preserved 试验由勃林格殷格翰和礼来公司资助。

披露Packer 博士报告了来自 Abbvie、Actavis、Amgen、Amarin、阿斯利康、勃林格殷格翰、百时美施贵宝、Casana、CSL Behring、Cytokinetics、强生、礼来、Moderna、诺华、ParatusRx、辉瑞、Relypsa、Salamandra、Synthetic 的咨询费Biologics 和 Theravance，在提交的工作之外。他是 PARAGON-HF 和 EMPEROR-Preserved 试验的执行委员会成员。博士

Zannad 已从 Amgen、阿斯利康、拜耳、勃林格殷格翰、波士顿科学、Cardior、CVRx、Janssen、Livanova、默克、Mundipharma、诺华、Novo Nordisk 和 Vifor Fresenius 收到指导委员会或咨询委员会的费用。他是 PARAGON-HF 和 EMPEROR-Preserved 试验的执行委员会成员。

Anker 博士报告了 Vifor International 和 Abbott Vascular 的资助和个人费用，以及 AstraZeneca、Bayer、Brahms、Boehringer Ingelheim、Cardiac Dimensions、Novartis、Occlutech、Servier 和 Vifor International 的个人费用。

本文中表达的观点不一定是编辑或美国心脏协会的观点。

https://www.ahajournals.org/journal/circ

播客和成绩单可在 https://www.ahajournals.org/doi/suppl/10.1161/CIRCULATIONAHA.121.056657 作为数据补充获得。

这项工作于 2021 年 8 月 27 日至 2021 年 8 月 30 日在欧洲心脏病学会大会的最新临床试验会议上进行了介绍。

有关资金来源和披露信息，请参见第 1195 页。