Abstract
AIMS: To determine whether lipoprotein(a) [Lp(a)] and cumulative Lp(a) (CumLp(a)) are associated with adverse outcomes in patients with acute myocardial infarction (AMI). METHODS: This cohort study included 2,634 hospitalized patients diagnosed with AMI who underwent coronary angiography at Zhongda Hospital, Southeast University, from July 2013, to December 2021. The main outcome was major adverse cardiac and cerebrovascular events (MACCE), defined as cardiovascular (CV) death, non-fatal myocardial infarction, non-fatal stroke, or unplanned revascularization—occurring singly or in combination. We used Cox proportional hazards models, with subgroup and sensitivity analyses, restricted cubic spline (RCS) modeling, and threshold-effect assessment to evaluate the relationships between Lp(a), CumLp(a), and prognosis. RESULTS: Across a median 55.2-month follow-up, 907 participants (34.40%) experienced a MACCE, 342 (13.00%) patients had CV death, 177 (6.70%) patients had non-fatal MI, 202 (7.70%) patients had non-fatal stroke, 399 (15.10%) patients underwent unplanned revascularization, and all-cause death occurred in 547 (20.80%) patients. Multivariable Cox regression models demonstrated a significantly increased risk of MACCE, CV death, non-fatal MI, and non-fatal stroke in both the higher Lp(a) and higher CumLp(a) groups compared with the lower groups (HRs for Lp(a): 1.652, 2.157, 3.455, and 1.930; HRs for CumLp(a): 1.697, 1.675, 3.759, and 2.032), and every one-unit rise in CumLp(a), the risk of MACCE, CV death, non-fatal MI and non-fatal stroke increased by 1.3%, 1.4%, 1.9% and 1.2%, respectively. The majority of subgroup and sensitivity checks consistently supported a stable link between Lp(a)/CumLp(a) and the risks of MACCE, CV death, non-fatal MI, and stroke. Analyses using RCS and threshold models revealed that Log(10)(Lp(a)) had both nonlinear and threshold effects on the risks of MACCE and each single event other than unplanned revascularization, while log(10)-transformed CumLp(a) showed nonlinear and threshold associations only with MACCE, CV death, and non-fatal MI (nonlinearity P < 0.05). CONCLUSIONS: Higher levels of Lp(a) and CumLp(a) are linked to a greater risk of poor outcomes among patients with AMI as the index event, highlighting their potential value for risk stratification and guiding clinical decision-making. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12944-025-02800-6.