In total, 364 individuals were selected for this retrospective cohort study as per corresponding diagnostic criteria and were diagnosed as AF at the First Hospital of Quanzhou Affiliated to Fujian Medical University, between January 2019 and June 2022. In addition, all individuals met the following inclusion criteria: (a) All of these individuals met the guideline-recommended indications for RFCA (2). (b) All individuals underwent RFCA for the first time and were successfully converted to sinus rhythm by ablation. (c) These patients underwent preoperative collection of basic information as well as electrocardiography, echocardiography, and laboratory tests. Patients with incomplete baseline information, unavailable follow-up information, and taking oral lipid-lowering drugs before the collection of fasting blood specimens were not included in this research. Ultimately, 320 individuals were selected for follow-up analysis. This research was approved by the Ethics Review Committee of First Hospital of Quanzhou Affiliated to Fujian Medical University (approval No.: 2023K031). The study was retrospective in design and therefore exempted from informed consent. Additionally, all procedures in this research were executed according to the Declaration of Helsinki and all relevant guidelines and regulations.

Baseline characteristics of individuals comprising age, gender, histories of alcohol abuse and smoking, body mass index (BMI), type of AF, and histories of hypertension, diabetes mellitus, coronary artery disease, and hyperthyroidism were collected at the time of admission.

Fasting blood samples were acquired from individuals early in the morning before ablation on the second day after their admission to the hospital. A set of biochemical markers, encompassing total cholesterol (TC), low-density lipoprotein cholesterol (LDL-C), very low-density lipoprotein cholesterol (VLDL-C), high-density lipoprotein cholesterol (HDL-C), and triglycerides (TG) were assessed. There is no consensus on the measurement of RLP-C because RLP is difficult to separate from its precursors. The available methods include the formula and direct measurement techniques (including direct automated analysis, immunoseparation, ultracentrifugation, etc.) (22). The direct measurement technique has not been recommended in clinical practice due to high equipment requirements, operational difficulties, and high costs. In contrast, the formula technique has gained widespread applications in clinical studies due to its convenience, rapidity, and economical efficiency (19–21). The RLP-C derived by the formula technique can accurately reflect the actual measurement level of RLP-C (23). Therefore, the formula technique was used for RLP-C measurement in this research: RLP-C = TC-(LDL-C + HDL-C).

Transthoracic echocardiography was performed by an experienced sonographer and recorded with a nationally approved diagnostic echocardiography system (Philips IE Elite, Philips Ultrasound, Inc., Bothell, WA, USA) with a probe frequency of 2.0–4.0 MHz. Specifically, the individual was placed in a left lateral position to fully expose the thorax. Subsequently, cardiac ultrasound data were collected in parasternal short and long-axis views, as well as apical four-chamber and two-chamber views in sequence. The left atrium anterior-posterior diameter (LAD) was measured at the end of the left ventricular systole. Finally, the left ventricular ejection fraction (LVEF) was measured using a dual-plane method.

Three-dimensional reconstruction of the left atrium and bilateral pulmonary ostium venosum sites was performed with the aid of an electroanatomical landmarking system (Carto3, Biosense Webster, Diamond Bar, CA, USA). Circumferential pulmonary vein ablation (CPVA) was performed in individuals with paroxysmal AF utilizing an adjustable curved radiofrequency ablation catheter (Thermcool ST, Biosense Webster, Diamond Bar, CA, USA) (target temperature 43°C, maximum power 45 W). Individuals with persistent AF underwent CPVA combined with top-line as well as bottom-line ablation of the left atrium. Additionally, other linear ablations, such as mitral isthmus addition, tricuspid isthmus addition, and linear ablation of the upferior vena cava were added at the discretion of the operator, depending on whether atrial arrhythmias were induced. Details of the CPVA procedure have been illustrated in an expert consensus statement (24). The endpoint of successful ablation is defined as the completion of the circumferential ablation trail of the pulmonary veins, as well as other additional ablation lines while achieving a bidirectional conduction block on both sides of the ablation line.

The 1-year follow-up data of individuals with AF after surgery were recorded, in which 3-month recurrence was defined as the endpoint for follow-up. The initial follow-up visit was scheduled on day 7 after discharge. Subsequent follow-up visits were then recommended every 3 months thereafter. All individuals were recommended to undergo immediate reexamination if they developed symptoms similar to AF. Patients who were not followed up promptly were contacted by telephone to record the onset of arrhythmia symptoms including palpitations, chest distress, and fatigue, and were instructed to seek prompt medical care. A long-term AF recurrence was defined as the presence of atrial tachyarrhythmia that last for more than 30 s (recorded by ECG or 24-h-Holter) after 3 months of ablation. In our practice, in order to improve the detection rate of atrial tachyarrhythmia, we routinely carry out 12-lead ECG on patients undergoing review, and further 24-h-Holter are performed on patients in whom atrial tachyarrhythmia is not recorded. A recurrence within 3 months is referred to as a blanket recurrence or early recurrence (ER). In the past, AF recurrence within 3 months was considered to be a recurrence without clinical significance, and it was thought that ER was associated with inflammation and edema of tissues and cells after RFCA. However, recent research evidence suggests a higher risk of long-term recurrence among patients with early recurrence (3). Therefore, it was used as an important covariate in the follow-up analysis. Postoperative antiarrhythmic drug (AAD) administration and anticoagulation were maintained for at least 3 months. After 3 months, AADs were discontinued in the absence of atrial tachyarrhythmia recurrence. Furthermore, anticoagulants were discontinued or reduced according to the CHA₂DS₂-VASc score.

The mean ± standard deviation was utilized for expressing normally distributed data, and the t-test was employed for comparisons in this case. Non-normally distributed data, on the other hand, were expressed as the median and interquartile range (IQR), and comparisons were made utilizing the Mann–Whitney U-test in such cases. Percentages were utilized for the expression of categorical variables, and comparative assessment was done via the χ² test. Univariate Cox regression analysis was conducted to identify risk factors for long-term AF recurrence and calculate hazard ratios (HRs) and 95% confidence intervals (95% CIs). Independent variables with a P < 0.1, as well as lipid factors, gender and age, were incorporated in the multivariate Cox regression analysis. Backward partial likelihood estimation was utilized for eliminating collinearity and constructing a multivariate regression model. The link between RLP-C levels and AF recurrence was analyzed using these models, followed by the calculation of HRs and 95% CIs. To evaluate the robustness of the findings, subgroup analyses were conducted for gender, ER, and type of AF, and forest plots were generated. ROC curves were plotted to determine the predictive value of RLP-C in long-term AF recurrence after RFCA and calculate the sensitivity and specificity. Moreover, the Youden index was employed to evaluate the optimal cut-off point. The study subjects were then categorized into two groups: the high RLP-C level group and the low RLP-C level group, as per the aforementioned cut-off point. Kaplan–Meier curves were then generated. Subsequently, the log-rank test was performed to assess any significant variation between both groups. All the analyses described above were executed utilizing SPSS 23.00 (SPSS, Inc., Chicago, IL, USA) and GraphPad Prism 9.0.2 for Windows (GraphPad Software, San Diego, CA, USA). A P < 0.05 denoted the statistical significance level.

Figure 1 illustrates that 364 individuals who satisfied the diagnostic criteria for AF and underwent their first RFCA procedure between January 2019 and June 2022, were initially enrolled in the research. Subsequently, 44 individuals were excluded, including 20 individuals taking lipid-lowering drugs before blood specimen collection, 16 individuals with no follow-up information, and 8 individuals whose lipid levels were not measured before the procedure. As per the inclusion and exclusion criteria, 320 individuals were selected. Among them, 103 individuals were classified into the recurrence group, while 217 individuals were categorized into the non-recurrence group. Recurrence herein was defined as long-term recurrence occurring after 3 months.

During the 12-month follow-up, 103 individuals experienced recurrence. As presented in Table 1, the recurrence group was older (65.0 vs. 62.0; P = 0.046) and showed a greater proportion of ER (37.9% vs. 14.3%; P < 0.001) in comparison to the non-recurrence group. The recurrence group had a greater proportion of diabetes mellitus (30.0% vs. 20.7%; P < 0.001) and CAD (18.9% vs. 6.9%; P < 0.001) and longer LAD (43.0 vs. 37.0; P < 0.001). However, the proportion of paroxysmal AF was reduced in the recurrence group in comparison to that in the non-recurrence group (47.6% vs. 74.7%; P < 0.001). In addition, the levels of TC (3.9 vs. 5.1; P < 0.001), LDL-C (2.4 vs. 3.1; P < 0.001), HDL-C (1.0 vs. 1.1; P < 0.001), and RLP-C (0.4 vs. 0.7; P < 0.001) were decreased in the recurrence group in comparison to the non-recurrence group. Moreover, no statistically significant variations in age, gender, BMI, smoking, alcohol abuse, hypertension, hyperthyroidism, LVEF, and TG were noted across the two groups in this research (P > 0.05).

As presented in Table 2, in the univariate Cox regression analysis, statistically significant predictors for AF recurrence included paroxysmal AF (HR = 0.41, 95% CI: 0.28–0.60, P < 0.001), ER (HR = 2.58, 95% CI: 1.72–3.83, P < 0.001), CAD (HR = 2.46, 95% CI: 1.49–4.05, P < 0.001), LAD (HR = 1.11, 95% CI: 1.08–1.14, P < 0.001), LVEF (HR = 0.98, 95% CI: 0.96–0.99, P = 0.007), TC (HR = 0.93, 95% CI: 0.91–0.95, P < 0.001), TG (HR = 1.34, 95% CI: 1.01–1.79, P = 0.049), LDL-C (HR = 0.43, 95% CI: 0.33–0.56, P < 0.001), and every 0.1 mmol/L rise in HDL-C (HR = 0.88, 95% CI: 0.80–0.93, P < 0.001) and RLP-C (HR = 0.69, 95% CI: 0.64–0.76, P < 0.001).

The variables with P < 0.1 in the above univariate Cox regression analysis and gender underwent a multivariate Cox regression model and were analyzed by employing a backward stepwise screening technique. The statistically significant independent predictors in the final analysis included ER (HR = 1.57, 95% CI: 1.04–2.36, P = 0.032), CAD (HR = 2.03, 95% CI: 1.22–3.38, P = 0.006), LAD (HR = 1.07, 95% CI: 1.03–1.10, P < 0.001), TG (HR = 1.51, 95% CI: 1.16–1.96, P = 0.002), LDL-C (HR = 0.74, 95% CI: 0.55–0.98, P = 0.036) and every 0.1 mmol/L rise in RLP-C (HR = 0.75, 95% CI: 0.68–0.83, P < 0.001).

To assess the stability of the results, validation was performed in three subgroups. As presented in Figure 2, RLP-C exhibited an inverse relation with AF recurrence post-RFCA in all subgroups in terms of male (HR = 0.69, 95% CI: 0.62–0.78, P < 0.001), female (HR = 0.70, 95% CI: 0.61–0.79, P < 0.001), ER (HR = 0.77, 95% CI: 0.67–0.88, P < 0.001), non-ER (HR = 0.68 95% CI: 0.61–0.76, P < 0.001), paroxysmal AF (HR = 0.67, 95% CI: 0.59–0.76, P < 0.001), and persistent AF (HR = 0.75, 95% CI: 0.67–0.84, P < 0.001). In addition, no considerable variations were observed at the respective subgroup levels of sex, presence of ER, and AF type, with P-values for the interaction of 0.467, 0.144, and 0.129, respectively.

ROC analysis was executed to further assess the predictive value of RLP-C for AF recurrence. As presented in Figure 3, the area under the curve (AUC) value of RLP-C was 0.81 (95% CI: 0.77–0.86, P < 0.001), with the maximum Youden index of 0.51, the optimal threshold value of RLP-C of 0.645 mmol/L, sensitivity of 87.4%, and specificity of 63.6%.

The 320 patients were regrouped into two groups (≤0.645 mmol/L group and >0.645 mmol/L group) according to the optimal RLP-C threshold value of 0.645 mmol/L, and Kaplan–Meier survival curves were plotted (Figure 4). A total of 103 patients (32.2%) experienced recurrence during the follow-up period. Of these, 90 (53.3%) recurrences occurred at the end of follow-up in the ≤0.645 mmol/L group and 13 (8.6%) occurred in the >0.645 mmol/L group. According to the Log-rank test, the AF recurrence rate was reduced in the >0.645 mmol/L group than in the ≤0.645 mmol/L group (Log-rank P < 0.001). After adjusting for ER, CAD, LAD and LDL-C, multivariate Cox regression analysis showed that lower recurrence rates of AF in the high RLP-C level (>0.645 mmol/L) group compared with ≤0.645 mmol/L group (HR = 0.21, 95% CI: 0.11–0.39, P < 0.001).