Both American and European heart failure (HF) guidelines currently recommend initiation and optimization of guideline-directed medical therapy (GDMT), which is composed of renin–angiotensin system (RAS) inhibitors, β-blocker, mineralocorticoid receptor antagonist (MRA), and sodium–glucose cotransporter 2 (SGLT2) inhibitor, during the hospitalization for acute HF (AHF).^1-3 In previous clinical trials, it was found that initiation of sacubitril/valsartan (Sac/Val) early in stabilized patients after an AHF episode requiring hospitalization resulted in a greater reduction in N-terminal pro-B-type natriuretic peptide (NT-proBNP) concentration than the use of standard RAS inhibitors over 8 weeks of therapy.⁴ However, since those trials and another AHF trial assessing intensive and rapid up-titration of GDMT included few patients being treated by SGLT2 inhibitors,^{4, 5} it is currently unclear whether the treatment effect of early Sac/Val initiation on NT-proBNP concentration differs according to the combination status of GDMT in this patient population.

In the recent Program Angiotensin–Neprilysin Inhibition in Admitted Patients with Worsening Heart Failure (PREMIER) study (NCT05164653),⁶ in-hospital initiation of Sac/Val, compared with the use of angiotensin-converting enzyme inhibitors or angiotensin receptor blockers (ACEI/ARB), also triggered a greater NT-proBNP concentration reduction over 8 weeks in Japanese patients admitted with AHF. In that study, background usage of individual GDMT other than RAS inhibitors was more frequent than in previous studies.⁴ We herein examined the NT-proBNP response according to the combination status of GDMT usage as a post hoc secondary analysis of the PREMIER study.⁶

The PREMIER study was an investigator-initiated, multicentre, prospective, randomized controlled, open-label, blinded-endpoint design that included haemodynamically stabilized Japanese inpatients after an AHF event, regardless of left ventricular ejection fraction (LVEF) status and acute de novo or decompensated chronic HF. The study participants on standard ACEI/ARB therapy were allocated within 7 days of an index hospitalization to receive either switched Sac/Val or continued ACEI/ARB therapy for 8 weeks. The study protocol was approved by the ethics committee of each site and individual informed consent was obtained before study entry.

Study participants were sub-classified according to the combined use of background GDMT, excluding the study drugs, at baseline (week 0). Patients who were on three background GDMT (β-blocker, MRA, and SGLT2 inhibitor) were categorized into quadruple (three GDMT and study drug) HF therapy recipients. In contrast, patients, who were on two or fewer background GDMT plus study drugs, were categorized into non-quadruple HF therapy recipients. The NT-proBNP concentration was analysed in a manner similar to the primary analysis,⁶ using a mixed-effects model for repeated measurements that included an interaction term between treatment and GDMT status, adjusted by baseline NT-proBNP levels and other variables with a large treatment-group difference (a standardized mean difference >0.200).

Among the full analysis set (n = 376; Sac/Val, n = 183; ACEI/ARB, n = 193),⁶ 188 patients (Sac/Val, n = 91; ACEI/ARB, n = 97) received quadruple HF therapy and 188 patients (Sac/Val, n = 92; ACEI/ARB, n = 96) received non-quadruple HF therapy. The background characteristics of the patients according to their GDMT status are shown in Table 1. Quadruple recipients were younger, had more de novo HF, and had a lower LVEF than non-quadruple recipients. In the quadruple recipients, percent reductions in the geometric means of NT-proBNP concentration at week 8, compared to the baseline value, were − 54% (95% confidence interval [CI], −61% to −45%) for the Sac/Val group and − 37% (95% CI, −47% to −25%) for the ACEI/ARB group; a group ratio with change (Sac/Val vs. ACEI/ARB) was 0.78 (95% CI, 0.62 to 0.98; p = 0.034). In the non-quadruple recipients, the reductions were −35% (95% CI, −45% to −22%) for the Sac/Val group and −26% (95% CI, −38% to −12%) for the ACEI/ARB group; the group ratio was 0.87 (95% CI, 0.69 to 1.09; p = 0.232) (Figure 1). This was similar in the exclusive analysis targeted at subgroup with background LVEF <40% (group ratio 0.62 [95% CI, 0.47 to 0.83; p = 0.001] for the quadruple recipients) (mean LVEF at randomization 27.7 ± 6.4%) and 0.77 (95% CI, 0.54 to 1.09; p = 0.140) for the non-quadruple recipients (mean LVEF at randomization 28.9 ± 6.0%) (Figure 1).

A key finding of this secondary analysis of the PREMIER study was that the treatment effect of Sac/Val, relative to ACEI/ARB, on NT-proBNP reduction in patients admitted for AHF was potentially more evident in quadruple recipients, although there was no clear statistical heterogeneity between the GDMT subgroups (p_interaction = 0.505). In this context, the background LVEF was lower in quadruple therapy recipients than in non-quadruple therapy recipients. Studies have shown that the beneficial effect of Sac/Val therapy after AHF was generally evident in patients with a lower LVEF spectrum.^{4, 6} Hence, we assumed that our finding was due to the large proportion of participants with lower LVEF in the quadruple therapy recipients. However, in the exclusive analysis of patients with a background LVEF <40%, the NT-proBNP response to Sac/Val therapy was still evident in quadruple recipients. This may enhance the clinical significance of the early implementation of Sac/Val-based quadruple GDMTs after an AHF episode, especially in patients with a lower LVEF.

Our findings should be interpreted as indicative, considering several limitations but not as definitive proof of the evidence. This was a post hoc exploratory analysis from a randomized PREMIER study, which was not specifically designed to assess the difference in the treatment effect of Sac/Val according to the baseline combination status of GDMT. The relatively small number of subclassified patients reduced the statistical power to detect the differences. Additionally, this analysis was performed only to evaluate the original primary endpoint of the PREMIER study, and the implementation of GDMT along with other clinical confounding factors may have partially affected the impact of the study interventions on the endpoint. Finally, we did not consider the doses of individual GDMT and their changes, due to limited information after initiating the study drugs.

Collectively, our findings suggest that the Sac/Val-based full-combination use of GDMT, relative to ACEI/ARB-based use, was incremental in reducing NT-proBNP concentrations in haemodynamically stabilized patients after an AHF episode. Further studies are warranted to establish an optimized clinical strategy for contemporary GDMT implementation in patients with AHF.

美国和欧洲心力衰竭 （HF） 指南目前都建议在急性 HF （AHF） 住院期间开始和优化指南指导的药物治疗 （GDMT），该治疗由肾素-血管紧张素系统 （RAS） 抑制剂、β阻滞剂、盐皮质激素受体拮抗剂 （MRA） 和钠-葡萄糖协同转运蛋白 2 （SGLT2） 抑制剂组成。^1-3 在之前的临床试验中，发现在需要住院治疗的 AHF 发作后，稳定患者早期开始使用沙库巴曲/缬沙坦 （Sac/Val） 导致 N 末端 B 型利钠肽前体 （NT-proBNP） 浓度降低幅度大于使用标准 RAS 抑制剂超过 8 周的治疗。⁴ 然而，由于这些试验和另一项评估 GDMT 强化和快速滴定的 AHF 试验包括很少接受 SGLT2 抑制剂治疗的患者，^{4， 5} 目前尚不清楚早期 Sac/Val 开始对 NT-proBNP 浓度的治疗效果是否根据该患者群体中 GDMT 的组合状态而有所不同。

在最近的 血管紧张素-脑啡肽酶抑制项目 （PREMIER） 研究中 （NCT05164653），⁶ 与使用血管紧张素转换酶抑制剂或血管紧张素受体阻滞剂 （ACEI/ARB） 相比，在院内开始使用 Sac/Val 也引发了 AHF 入院的日本患者在 8 周内更大的 NT-proBNP 浓度降低。在该研究中，除 RAS 抑制剂外的单个 GDMT 的背景使用比以前的研究更频繁。⁴ 我们在此根据 GDMT 使用的组合状态检查 NT-proBNP 反应，作为 PREMIER 研究的事后二次分析。⁶

PREMIER 研究是一项研究者发起的、多中心、前瞻性、随机对照、开放标签、盲法终点设计，包括 AHF 事件后血流动力学稳定的日本住院患者，无论左心室射血分数 （LVEF） 状态和急性新发或失代偿性慢性 HF。接受标准 ACEI/ARB 治疗的研究参与者在指数住院后 7 天内被分配接受转换 Sac/Val 或继续 ACEI/ARB 治疗 8 周。研究方案由每个站点的伦理委员会批准，并在研究开始前获得个人知情同意。

研究参与者根据基线（第 0 周）背景 GDMT 的联合使用（不包括研究药物）进行细分。接受 3 种背景 GDMT （β 受体阻滞剂、 MRA 和 SGLT2 抑制剂） 的患者被分为四重 （3 种 GDMT 和研究药物） HF 治疗接受者。相比之下，接受两种或更少背景 GDMT 加研究药物的患者被归类为非四联 HF 治疗接受者。NT-proBNP 浓度的分析方式与初步分析相似，⁶ 使用混合效应模型进行重复测量，其中包括治疗和 GDMT 状态之间的交互项，通过基线 NT-proBNP 水平和其他具有较大治疗组差异的变量进行调整（标准化平均差异 >0.200）。

在完整的分析集 （n = 376;Sac/Val，n = 183;ACEI/ARB，n = 193），⁶ 188 名患者 （Sac/Val，n = 91;ACEI/ARB，n = 97）接受四联 HF 治疗，188 例患者 （Sac/Val，n = 92;ACEI/ARB，n = 96）接受非四联 HF 治疗。根据 GDMT 状态的患者背景特征如 表 1 所示。与非四联接受者相比，四联接受者更年轻，新发 HF 更高，LVEF 更低。在四联接受者中，与基线值相比，第 8 周时 NT-proBNP 浓度几何平均值的降低百分比为 -54%（95% 置信区间 [CI]，-61% 至 -45%），ACEI/ARB 组为 -37%（95% CI，-47% 至 -25%）;变化组比 （Sac/Val vs. ACEI/ARB） 为 0.78 （95% CI，0.62 至 0.98;p = 0.034）。在非四联受者中，Sac/Val 组的减少为 -35% （95% CI，-45% 至 -22%），ACEI/ARB 组的减少为 -26% （95% CI，-38% 至 -12%）;组比为 0.87 （95% CI，0.69 至 1.09;p = 0.232）（图 1）。这在针对背景 LVEF <40% 的亚组的独家分析中相似（组比 0.62 [95% CI，0.47 至 0.83;p=0.001]）（随机分组时的平均LVEF为27.7±6.4%）和0.77（95%CI，0.54至1.09;p = 0.140） 对于非四联接受者 （随机分组时的平均 LVEF 28.9 ± 6.0%） （图 1）。

PREMIER 研究的二次分析的一个关键发现是，相对于 ACEI/ARB，Sac/Val 对 AHF 入院患者 NT-proBNP 降低的治疗效果在四重接受者中可能更明显，尽管 GDMT 亚组之间没有明显的统计异质性 （p_交互作用 = 0.505）。在这种情况下，四联治疗接受者的背景 LVEF 低于非四联治疗接受者。研究表明，AHF 后 Sac/Val 治疗的有益效果在 LVEF 谱较低患者中通常很明显。^4、6因此，我们假设我们的发现是由于四联治疗接受者中 LVEF 较低的参与者比例很大。然而，在对背景 LVEF <40% 患者的独家分析中，NT-proBNP 对 Sac/Val 治疗的反应在四联接受者中仍然很明显。这可能会增强在 AHF 发作后早期实施基于 Sac/Val 的四联 GDMT 的临床意义，尤其是在 LVEF 较低的患者中。

考虑到一些局限性，我们的研究结果应被解释为指示性，但不是证据的明确证据。这是一项来自随机 PREMIER 研究的事后探索性分析，该研究并非专门设计用于根据 GDMT 的基线组合状态评估 Sac/Val 治疗效果的差异。相对较少的亚组患者数量降低了检测差异的统计能力。此外，进行该分析仅用于评估 PREMIER 研究的原始主要终点，GDMT 的实施以及其他临床混杂因素可能部分影响了研究干预对终点的影响。最后，由于开始研究药物后的信息有限，我们没有考虑单个 GDMT 的剂量及其变化。

总的来说，我们的研究结果表明，相对于基于 ACEI/ARB 的使用，基于 Sac/Val 的 GDMT 完全联合使用在降低 AHF 发作后血流动力学稳定的患者的 NT-proBNP 浓度方面是递增的。需要进一步的研究，以建立在 AHF 患者中实施当代 GDMT 的优化临床策略。