Atrial fibrillation(AF) is one of the most prevalent cardiovascular arrhythmia, and its incidence is increasing globally[1]. Approximately 1–3% of the general population is affected by this condition, with a higher prevalence observed among elderly individuals, males, and those with comorbidities[2–6]. In 2019, global epidemiological data indicated that there were approximately 59.7 million individuals affected by AF, with an estimated 4.72 million new cases reported during that year[7]. AF is linked to a range of serious complications and constitutes a significant contributor to the global burden of cardiovascular diseases[8–11]. Despite advancements in current treatment approaches, the management of AF continues to pose significant challenges, primarily due to the intricate pathophysiology of the condition and the heterogeneous responses observed among patients. Therefore, the identification of reliable prognostic markers is of critical importance in mitigating complications and enhancing clinical outcomes.

Estimated pulse wave velocity (ePWV) serves as a convenient surrogate marker for carotid-femoral pulse wave velocity (cfPWV)[12]. Research has demonstrated that ePWV is significantly associated with the prognosis of various clinical conditions[13–16]. It remains uncertain whether ePWV is associated with the prognosis of critically ill patients diagnosed with AF. To address this knowledge gap, the present study conducted a retrospective cohort analysis using data from patients with AF extracted from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. The primary objective was to investigate the potential association between ePWV and short-term mortality in this patient population. Furthermore, this research aims to facilitate the development of a straightforward and reliable prognostic assessment tool, which may support the formulation of novel strategies to improve outcomes for critically ill patients with AF.

In this investigation, we initially identified 18,805 critically ill patients with AF who were first admitted to ICU between 2008 and 2019. We excluded 2,996 patients with length of ICU stay shorter than 24 hours, 144 patients without records of mean arterial pressure (MAP), and 6,486 patients lacking body-mass index (BMI) data. Ultimately, 9,179 patients with critical AF were retained and then classified into low- and high-ePWV groups according to the optimal cut-off derived from the ROC curve. The patient selection flowchart is presented in Fig. 1. Data extraction was performed with PostgreSQL (v13.7.2) and Navicat Premium (v16) by executing structured query language (SQL) scripts. Potential confounders considered in this study included: (1) baseline demographic information: age, sex, race, body mass index (BMI); (2) vital signs: systolic blood pressure (SBP), diastolic blood pressure (DBP) and mean arterial pressure (MAP); (3) Comorbidities: chronic kidney disease(CKD), type 2 diabetes mellitus(T2DM), heart failure(HF), myocardial infarction(MI), stroke, hypertension, liver cirrhosis(LC), chronic obstructive pulmonary disease (COPD); (4) Treatments and Medications: mechanical ventilation(MV), antibiotics, corticosteroids, antihypertensive agents utilisation; (5) Laboratory parameters: white blood cell(WBC), red blood cell (RBC), hemoglobin, platelet, glucose, anion gap(AG), potassium, calcium; (6) Illness-severity scores: Acute Physiology ScoreIII (APSIII), Sequential Organ Failure Assessment (SOFA), Systemic Inflammatory Response Syndrome (SIRS). Missing data were present, however, for every variable the proportion of missing values did not exceed 5%. We retained the missing values without imputation. ePWV was calculated using the formula[14]:ePWV = 9.587 − 0.402×age + 4.560×10 − 3×age²−2.621×10 − 5×age²×MAP + 3.176×10 − 3×age×MAP − 1.832×10 − 2×MAP.The primary outcome was 30-day mortality after ICU admission.

Univariate cox regression analysis was conducted for 30-day mortality, and the results were summarized in Table 2. Table 3 presents a comparison from one group to another using a multifactor Cox regression model. In the unadjusted model, the mortality risk in the high ePWV group was significantly higher than in the low ePWV group (HR = 1.97; 95% CI: 1.78–2.19; P < 0.001). In Model 3, after adjusting for additional covariates, the high ePWV group still exhibited a significantly higher mortality risk compared to the low ePWV group (HR = 1.70; 95% CI: 1.51–1.91; P < 0.001). Furthermore, the covariate-adjusted restricted cubic spline (RCS) model revealed a nonlinear association between ePWV levels and 30-day mortality risk (nonlinear P < 0.001), as illustrated in Fig. 3. The 30-day mortality risk among AF patients increased progressively with higher ePWV levels (P < 0.001).

To further substantiate the relationship between 30-day mortality and ePWV, subgroup analyses were executed,partitioned by clinical conditions such as hypertension, T2DM, HF, MI and AKI. As depicted in Fig. 4, the outcomes of these subgroup examinations align closely with the overarching results. Importantly, ePWV exhibited marked disparities in 28-day mortality rates, regardless of the presence or absence of hypertension, T2DM, HF, MI and AKI, with statistical significance (p < 0.001). The interactions within these subgroups did not reach statistical significance, indicating that the prognostic impact of ePWV is consistent across various patient subpopulations. (P for interaction > 0.05).

BMI, body mass index; SOFA, Sequential Organ Failure Assessment; APSIII, Acute Physiology ScoreIII (APSIII); SIRS, Systemic Inflammatory Response Syndrome; CKD, chronic kidney disease; T2DM, type 2 diabetes mellitus; HF, heart failure; MI, myocardial infarction; LC, liver cirrhosis; WBC, white blood cells; RBC, red blood cells; AG, anion gape.

Model 1: Unadjusted; Model 2: Adjusted for sex, race, and BMI; Model 3: Adjusted for all covariates. HR, hazard ratio; CI, confidence interval.

This research delves into the relationship between ePWV and the likelihood of dying within 30 days in patients suffering from AF. Our results indicate that ePWV could be a prognostic marker independently for 30-day mortality in this distinct patients.Even after adjusting for potential confounding factors, elevated ePWV levels continued to exhibit a robust predictive capacity for 30-day mortality. Furthermore, among patients with AF, a non-linear association was observed between ePWV and the predicted 30-day mortality risk (non-linear p < 0.001).

ePWV serves as a simplified alternative indicator to cfPWV. The clinical application of cfPWV is limited due to its complex operational procedure, which requires coordination among multiple professionals[19]。However, ePWV can be estimated soely based on the patient's age and blood pressure. In certain critical clinical scenarios, its performance in detecting arterial stiffness is comparable to that of cfPWV[20]. Therefore, to conserve limited medical resources, ePWV can serve as a simplified alternative to cfPWV. Cui identified a statistically significant association between ePWV and increased mortality risk associated with AKI during ICU stays[21]. Wei demonstrated that ePWV serves as an independent predictor of 28-day mortality among patients with sepsis-associated AKI[22]. ePWV has been identified as an independent predictor of both short-term and long-ter[13]m mortality among critically ill patients with coronary heart disease. Integrating ePWV into conventional risk assessment models can substantially enhance the predictive accuracy of all-cause mortality within one year as well as during hospitalization for this patient population[23]. Chen conducted an 11-year follow-up study and found that elevated ePWV is associated with an increased risk of new-onset AF. Among the overall population, the incidence of new-onset AF in the medium, medium-high, and high ePWV groups was 1.64-fold, 1.92-fold, and 2.64-fold higher, respectively, compared to the low ePWV group[13]. There may be a causal relationship between arterial stiffness and AF. Research on atrial tissue has demonstrated that atherosclerosis can influence the progression of atrial fibrillation[24]. In addition, atherosclerosis and its associated risk factors contribute to structural and electrical remodeling of the atrium. Atrial fibrosis is recognized as both a consequence and a promoter of the initiation and progression of atrial fibrillation[25]. Furthermore, emerging evidence indicates that greater severity of atherosclerosis may lead to left ventricular hypertrophy, impaired left ventricular diastolic function, and left atrial enlargement, all of which can elevate the risk of developing atrial fibrillation[26].

In this study, based on a retrospective cohort of 9,179 ICU patients diagnosed with AF from the MIMIC-IV database, we confirmed that the 30-day all-cause mortality rate was 22.6% among patients with ePWV ≥ 11.97 m/s, which was significantly higher than that of patients in the low ePWV group. Multivariate Cox regression analysis revealed that ePWV was an independent risk factor for 30-day mortality (HR = 1.70, 95% CI: 1.51–1.91), with a non-linear association observed. Subgroup analyses demonstrated consistent results across various stratifications, indicating the robustness of the findings. Therefore, ePWV, which can be rapidly calculated at the bedside, may serve as a non-invasive and independent predictor for early mortality risk among ICU patients with AF, enabling timely identification of high-risk individuals and facilitating optimized clinical interventions.

Nevertheless, this study has several limitations. As the data were derived from an observational cohort within the MIMIC-IV database, potential confounding and selection biases may exist. Although multivariate Cox regression and restricted cubic spline (RCS) analyses were employed to adjust for known confounders, residual or unmeasured confounding factors cannot be entirely ruled out. Additionally, the MIMIC-IV database primarily includes data from patients in the United States, which may limit the generalizability of the findings to AF patients in other regions or countries. Furthermore, this study focused solely on the initial ePWV measurement following ICU admission and did not account for potential dynamic changes in ePWV over time. Given that ePWV may fluctuate in response to changes in a patient’s clinical condition, future research should investigate the longitudinal behavior of ePWV and further validate its predictive value and underlying mechanisms.

This study represents the first to identify a significant nonlinear relationship between ePWV and 30-day mortality among AF patients admitted to the ICU. These findings indicate that ePWV may function as an independent and novel biomarker for prognostic evaluation within this patient cohort, offering a theoretical foundation for early risk stratification and tailored therapeutic interventions. Further research is warranted to confirm the generalizability of these results through multicenter cohort studies, as well as to investigate the underlying pathophysiological mechanisms, including the interplay between arterial stiffness, atrial fibrillation, and multi-organ dysfunction.