Blood pressure (BP) is a significant risk factor implicated in the progression of atherosclerosis and the formation of vulnerable plaques, contributing to increased mortality in acute coronary syndrome (ACS) patients [1, 2]. Various BP parameters are validated as essential prognostic and therapeutic markers in managing cardiovascular diseases across different clinical settings. Guidelines are recommended for rigorous BP control in individuals with hypertension to minimize the morbidity and mortality of cardiovascular outcomes [3, 4]. Evidence from randomized controlled trials confirmed that reducing BP in those with hypertension decreased the risk of cardiovascular events in the future [5, 6].

Previous research suggests that while elevated BP is linked to a higher mortality risk following ACS, lower BP does not consistently demonstrate the same association. Nevertheless, achieving optimal BP control in patients after ACS remains essential for reducing subsequent cardiovascular events [7]. The SPRINT trial showed that an intensive treatment strategy aiming for a systolic blood pressure (SBP) 120 mmHg in high-risk patients significantly decreased cardiovascular mortality, thereby highlighting the benefits of aggressive BP management [8]. However, evidence indicates that universal BP management may not be appropriate for all patient populations. For example, elderly patients with an SBP below 125 mmHg had nearly a twofold risk of cardiovascular death within one year compared to those with an SBP above 125 mmHg [9]. Similarly, data from the Chinese ST-segment elevation myocardial infarction (STEMI) PPCI Registry indicated that although STEMI patients undergoing percutaneous coronary intervention (PCI) with SBP 120 mmHg experienced higher spontaneous reperfusion, the lowest all-cause mortality was observed among those with an SBP ranging from 121–150 mmHg [10]. These suggest that the correlation between BP and outcomes in acute cardiovascular settings appears complicated, as studies show inconsistent findings regarding BP levels and the risk of mortality. These studies indicated that a U- or J-shaped relationship between SBP/diastolic blood pressure (DBP) and adverse outcomes [11, 12, 13, 14, 15], suggesting that both lower BP and higher BP levels increase the risk of adverse cardiovascular outcomes. The therapeutic benefit of lowering BP may be reversed if BP is reduced below a certain threshold [16, 17].

These conflicting results highlight persistent uncertainties regarding the association between BP levels and the cardiovascular mortality risk of ACS patients. Also, the precise BP levels that correlate with the minimal mortality risk remain poorly defined. Currently, evidence on the prognostic impact of admission SBP/DBP on in-hospital mortality and long-term outcomes in STEMI patients undergoing PCI is limited. Utilizing the data sourced from the large-scale China Acute Myocardial Infarction (CAMI) registry, our study represents the first investigation to employ restricted cubic spline (RCS) analysis to systematically determine the association between levels of admission SBP/DBP and the risk of both in-hospital mortality and long-term endpoints in a large Chinese STEMI cohort. Based on prior evidence, we hypothesize that a J-shaped relationship between admission SBP/DBP and in-hospital mortality, aiming to identify the optimal admission BP levels linked to the lowest mortality risk in STEMI patients undergoing PCI.

This CAMI study is a large-scale, prospective, multicenter observational study in China, aimed at gathering real-world clinical data from acute myocardial infarction (AMI) patients (NCT01874691) [18]. This project was approved by the central institutional review board at Fuwai Hospital and by the ethics committees at each participating institution. Written informed consent was obtained from every enrolled participant. The Data Monitoring Committee was set to supervise data and ensure its quality. The registry includes 108 hospitals across 31 provinces and municipalities in mainland China, with Hong Kong and Macau not included. Eligible patients primarily diagnosed with AMI were consecutively recruited in the registry from January 2013 to September 2014. The AMI diagnosis was rigorously defined based on the third Universal Definition for Myocardial Infarction, including types 1, 2, 3, 4b, and 4c [19]. Type 4a and type 5 AMI were not included in this registry.

Data collected encompassed patient demographics, clinical risk factors, physical clinical examination findings, discharge medications and laboratory results [18]. The discharge medications specifically encompassed $\beta$-blockers, dual antiplatelet therapy (aspirin/clopidogrel), statins, and angiotensin-converting enzyme inhibitors (ACEis)/angiotensin receptor blockers (ARBs). The recorded laboratory parameters included total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), triglycerides (TG), high-density lipoprotein cholesterol (HDL-C), N-terminal pro-B-type natriuretic peptide (NT-pro BNP), white blood cell count (WBC), high-sensitivity C-reactive protein (Hs-CRP), and Troponin I (TnI). Extensive information on managing data and ensuring its quality has been previously detailed in the methodological sections of an earlier publication on the CAMI registry [18].

2.1 Study Population and Definitions

The CAMI registry recorded 26,648 AMI patients from January 2013 to September 2014. The inclusion criteria encompassed those diagnosed with AMI. The excluded criteria included: those with non-STsegment elevation myocardial infarction (NSTEMI) or an uncertain diagnosis with STEMI/NSTEMI (n = 7294); patients without primary or selective PCI (n = 6908); missing or outlier SBP/DBP values (n = 153); or those lacking clear survival status during hospitalization or any follow-up data (n = 1811). Ultimately, a total of 10,482 STEMI patients undergoing PCI were included in subsequent analysis (Fig. 1).

Fig. 1.

The flowchart of criteria for including and excluding participants. CAMI, China Acute Myocardial Infarction; NSTEMI, Non-ST segment elevation myocardial infarction; STEMI, ST-segment elevation myocardial infarction; PCI, percutaneous coronary intervention; SBP, systolic blood pressure; DBP, diastolic blood pressure.

Diabetes mellitus was defined as a documented history of the condition, prior hypoglycemic treatment or an admission hemoglobin A1c (HbA1c) level of 6.5% or higher [20]. Upon first admission to the Cardiology Department of FuWai Hospital, patients were instructed to rest quietly for at least five minutes before BP measurement. BP was measured by professional cardiologists with a validated electronic BP monitor. During the measurement, the upper arms of the patients were placed at the same height as the heart. Measurements were taken from both arms, and the higher value was recorded. Hyperlipidemia was characterized by plasma triglyceride levels of at least 200 mg/dL, total cholesterol levels of at least 240 mg/dL, or a recorded history of lipid-lowering medication use prior to admission [21].

In this study involving 10,482 individuals from a large-scale Chinese cohort of consecutive STEMI patients undergoing PCI, a J-shaped association between SBP/DBP and in-hospital mortality was identified using RCS analysis. This J-curve relationship persisted as significant even after controlling for possible confounding variables. Our data indicate the following: (1) the optimal admission SBP/DBP for minimizing risk of in-hospital mortality is 157/94 mmHg; (2) Additionally, the RCS analysis reveals a J-shaped trend, where patients with lower admission SBP/DBP are associated with an elevated risk for 2-year incidence of HF, severe bleeding, 2-year MACCEs and all-cause mortality. However, no J-shaped association was found for recurrent MI. (3) As a categorical variable, the HRs for in-hospital mortality were significantly higher in groups with lower admission SBP/DBP levels compared to those with an admission SBP/DBP of 120–129/70–79 mmHg. This association remained even after adjusting for potential confounders. According to our knowledge, this is the first investigation of the relationship between SBP/DBP and various long-term adverse clinical outcomes in STEMI patients receiving PCI within a large Chinese population cohort.

The majority of previous studies were conducted as retrospective or post hoc analysis of randomized trials within hypertension cohorts. These studies often explore the association between BP and adverse endpoints but often fail to reach consistent results due to the variability in patient history and underlying fragility. A systematic review and meta-analysis, excluding ACS patients, confirmed the benefits of reducing SBP [5, 6]. However, it is crucial to recognize that lower BP is not universally better, particularly for certain patient populations like those with ACS. In the ONTARGET and TRANSCEND trials, patients treated for hypertension with an SBP 120 mmHg or a DBP 70 mmHg exhibited increased cardiovascular risk, compared to those patients admitted with an SBP ranging from 120–140 mmHg [24]. A J-shaped relationship between BP and all-cause mortality was also identified in patients with both hypertension and coronary artery disease (CAD), with the lowest risk occurring at 119/84 mmHg [16]. Notably, these studies did not include the population with known ACS, which may limit the generalizability of the findings to this specific population.

Limited studies have investigated the phenomenon of J-shaped or U-shaped curves in ACS individuals. Consistent with our observations, a J-shaped curve was shown for the relationship between admission SBP and 2-year cardiovascular mortality in elderly ACS patients [14]. The Acute Coronary Syndrome Israel Survey (ACSIS) revealed that patients admitted with an SBP below 110 mmHg experienced significantly higher all-cause mortality at both 7-day and 1-year compared to those with admission SBP (110–140 mmHg) [25]. Additionally, both lower and higher SBPs were related to increased risks for different outcomes in AMI patients, with the risk associated with lower SBP being greater than that of a higher SBP [26, 27, 28]. However, it is crucial to highlight that the aforementioned studies primarily concentrated on the relationship between BP and adverse clinical outcomes in ACS patients, without investigating the optimal BP levels related to the minimal risk of such adverse endpoints.

The optimal levels of BP in patients with ACS have not been adequately defined. Previous studies identified a J-shaped curve between SBP/DBP and cardiovascular risks, with a nadir of 136/85 mmHg in ACS patients [29]. Similarly, the J-shaped curve was observed for SBP in relation to AMI prognosis, with a nadir at 114 mmHg [26]. Our findings are concordant with previous studies indicating a J- or U-shaped curve between BP and adverse endpoints. However, our results diverge somewhat from earlier results. One possible explanation for this divergence could be differences in data quality, sample sizes and confounders adjusted in the models. Furthermore, few studies have investigated the direct impact of admission BP levels on both short-term and long-term outcomes in ACS patients receiving PCI. Unlike prior research focusing on hypertension cohorts or broader ACS populations, our study specifically targeted STEMI patients. Given the widespread availability of PCI, the majority of ACS patients can receive reperfusion therapy in a timely manner. Therefore, our study specifically examined STEMI patients undergoing PCI and also found that the J- or U- shaped curve phenomenon was not alleviated by reperfusion therapy.

Both lower and higher SBP are associated with increased risks for the prognosis of AMI, with a lower SBP creating a greater risk compared to that of a higher SBP [27, 28]. In contrast to studies conducted by US and European researchers, a small Japan–based study found an average admission SBP in the range of 141–159 mmHg, while ventricular rupture–related deaths were more frequently observed in both the group with SBP $\ge$160 mmHg and the group with lower SBP [27]. Furthermore, a higher SBP might increase the ventricular afterload, leading to diastolic dysfunction due to ventricular hypertrophy, myocardial damage induced by increasing oxygen consumption, and promotion of atherosclerotic plaque rupture [30, 31, 32].

Interestingly, paradoxical findings have emerged, suggesting that elevated SBP levels might have protective effects concerning short-term mortality and improved in-hospital prognosis among ACS patients [33, 34]. Patients with an extremely elevated SBP ($>$160 mmHg) on admission exhibited a reduced risk for the same endpoints [25]. Similarly, a study involving 3943 AMI patients from an Austrian tertiary care hospital found that an admission SBP of 120 mmHg or lower was linked to poorer outcomes compared to a normal SBP range of 121–140 mmHg. In contrast, the admission SBP ($>$160 mmHg) was related to the most favorable 1-year outcomes compared to a normal admission BP [35]. A study of 119,151 patients admitted for acute chest pain in the medical care unit (ICU) found that those with admission SBP in the highest quartile ($>$163 mmHg) had the lowest 1-year mortality compared to those in the second quartile (128–144 mmHg) [36]. SBP has historically been viewed as an indicator of peripheral resistance and cardiac output. Higher SBP may lead to improved prognosis in AMI patients, which may arise from the better-preserved cardiac functions and reduced myocardial damage in AMI patients. Another possible explanation is that the potentially beneficial effects of higher admission SBP might be attributed to the cardioprotective functions of anti-hypertension medications including ACEis and $\beta$ receptor-blockers.

The J-shaped or U-shaped relationship, particularly the higher rate of adverse events at lower BP levels observed in this study can be attributed to several reasons: (i) Coronary reperfusion is influenced by two major factors: coronary arterial pressure and myocardial oxygen consumption [37]. A lower DBP often leads to decreased coronary reperfusion, and this phenomenon is more significant in coronary atherosclerotic patients with impaired coronary flow reserve [38]. In addition, it could be proposed that STEMI patients with reduced BP suffer poor systematic health status, overstimulation of the sympathetic nervous system and severe coronary microvascular dysfunction [39]. Consequently, the poorer outcomes in patients with reduced SBP/DBP might be attributed to the compromised reperfusion of the ischemic myocardium; (ii) Low BP may represent an inability to generate a hypertensive response, potentially reflecting an epiphenomenon resulting from comorbidity burden and frailty [40]. Regarding the J-shaped concept, it has raised concerns about reverse causality, suggesting that a low SBP/DBP might merely reflect the unhealthy condition of patients rather than directly causing worse cardiovascular outcomes. Increased risk at lower DBP demonstrated, at least in part, reverse causation due to arterial aging, stiffening, or other conditions that contribute to a lower DBP. However, the CLARIFY trial argued against this viewpoint, as the authors excluded certain conditions that affect life expectancy and other serious diseases. Even after adjusting for factors like heart failure, peripheral artery disease, and specific baseline characteristics, the link between low SBP/DBP and a higher risk of cardiovascular events remains [17]. (iii) Reduced SBP/DBP might also be an epiphenomenon of damaged cardiac function [41, 42]. Nonetheless, the study indicated that low DBP remained a significant predictor of adverse events, even after adjusting for left ventricular function [43]. Although different studies have identified varying BP thresholds based on different demographic characteristics, the findings consistently indicated that patients with a low SBP/DBP experienced an elevated risk of cardiovascular outcomes compared to reference groups.

Considering age and diabetes as key risk factors in hypertension patients, the management strategies for these populations should not be overlooked. For the elderly, particularly those with cardiovascular disease, arterial stiffness and multiple organ dysfunction are inevitable. However, there are no clear recommendations for very elderly hypertension patients. The very elderly ($\ge$80 years) were recommended to control their BP to below 150/90 mmHg, while those aged 65–79 years were recommended to aim for a BP below 140/90 mmHg [44, 45]. Excessive reduction of BP might bring adverse outcomes, including reperfusion reduction in target organs and cognitive decline [46, 47]. For diabetic patients, strict BP control targeting 130/80 mmHg should be appropriate for patients with both diabetes and CAD [3, 48, 49]. Thus, when initiating BP management strategies for these populations, a comprehensive assessment of risk factors should be conducted. Dynamic BP monitoring and real-time drug adjustment are expected to avoid adverse clinical outcomes associated with excessive BP reduction. Further research should focus on the elderly and diabetic population to explore the optimal management strategy and avoid being extrapolated directly to these populations without caution. There is an urgent need for a specific BP-targeted threshold, particularly in high-risk populations.

This study, sourced from the CAMI database, is a retrospective observational analysis, primarily involving patients from China. Therefore, the reported relationship in STEMI patients receiving PCI between admission SBP/DBP and the risk of endpoint events should not be extrapolated to other populations, such as those with different comorbidities or from different geographical regions. Despite the adjustments for numerous baseline confounders, our multivariable model failed to account for several unmeasured factors potentially affecting outcomes, including frailty, socioeconomic status, and mental health. The underlying mechanisms in the relationship between a low SBP/DBP and adverse clinical outcomes are multifactorial and not yet fully understood, necessitating caution when generalizing our observations. Furthermore, some patient data were not fully recorded, potentially influencing the validity of the findings. Considering that the admission BP levels may have been influenced by analgesic drugs and vasoactive drugs, future studies should account for these medications and their timing as confounders in the adjusted models.

Additionally, in our studies, the nadir points of SBP/DBP for different clinical outcomes were considered preliminary exploratory results. Although these findings could provide initial clues regarding the relationship between SBP/DBP and various clinical outcomes, further research is needed in different populations to confirm the generalizability and reliability of our findings. Moreover, we observed wide confidence intervals appeared near some SBP/DBP values, indicating greater uncertainties in risk estimates for adverse clinical outcomes at these BP values. As BP levels change, the width of the confidence interval may also vary, reflecting the instability or uncertainty of the risk estimates in these regions. In future studies, the accuracy of risk prediction could be improved by increasing the sample size and refining the group criteria.

In conclusion, our observational study of admitted STEMI patients undergoing PCI revealed a J-shaped relationship between admission SBP/DBP and in-hospital mortality risk with the lowest risk at 157/94 mmHg. The optimal values fluctuated around 140/85 mmHg in the relationship between admission SBP/DBP and long-term outcomes (2-year heart failure, MACCEs, and all-cause death). Further studies will focus on several key areas. First, multicenter, prospective cohort studies will be initiated to assess the effects of different types of hypertension, such as permissive hypertension and resistant hypertension on adverse endpoints like all-cause mortality, cardiovascular outcomes and cardiovascular-kidney-metabolic syndrome. Additionally, it is worth exploring the association between long-term BP variation and adverse clinical outcomes through advanced wearable devices. Finally, with the continued advancement of artificial intelligence, the analysis of large-scale population data on BP can be facilitated through artificial intelligence, machine learning and other advanced techniques for personalized and automatic identification, classification, and prediction.

The data supporting the findings of this study are available at https://www.chictr.org.cn/ with the registration number ChiCTR-ONC-12002636. However, the data is not publicly accessible at the moment. It can be obtained from the author, Yuejin Yang (yangyj_fw@126.com), upon reasonable request.

ZS and YJ designed the research study. CD and ZS conducted the research. CD performed the formal statistical analysis. ZS and CD were responsible for writing, reviewing, and revising the original draft, and for providing the visualizations for this research. YY, HQ, WY and JY conducted an initial review of the article’s theme and content, offering valuable help and suggestions for revisions. YY provided the data. All authors contributed to the interpretation of the data and writing of the manuscript. All authors read and approved the final manuscript. All authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.

The study was approved by the central institutional review board at Fuwai Hospital and by the ethics committees of all participating institutions (431). Written informed consent was obtained from every enrolled participant. The study was carried out in accordance with the guidelines of the Declaration of Helsinki.

The work was supported by the grants from the National High Level Hospital Clinical Research Funding (2023-GSP-GG-32).

The authors declare no conflict of interest.

Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/RCM33512.