This study was approved by the Health Research and Ethics Committee of Lagos State University Teaching Hospital in Nigeria (reference number: LREC/06/10/1706). Before participation, all individuals provided written informed consent. All procedures followed the ethical principles of the 2024 Declaration of Helsinki. To ensure anonymity, a barcode system was employed, and all data were handled with strict confidentiality. The collected data were used solely for this study.

This cross-sectional case-control study determined a sample size of 138 based on prevalence data from a previous regional study,²⁷ which reported an 8% – 12% prevalence of cardiovascular disease in sub-Saharan African countries.²⁸ In total, 172 confirmed ASCVD patients, aged 20 years to 64 years, were enrolled. Participants were selected using a simple random sampling technique at the Lagos State University Teaching Hospital, Lagos, Nigeria. We enrolled participants diagnosed with hypertensive heart disease with significant atherosclerosis, ischaemic heart disease (IHD), myocardial infarction (MI), ischaemic stroke, or congestive heart failure linked to prior IHD or MI. Recruitment took place during routine outpatient visits at the Cardiology Clinic, Lagos State University Teaching Hospital, Lagos, Nigeria, from November 2021 to November 2022. Eligible participants were identified through consecutive clinic record reviews and were enrolled at most six months post-diagnosis, with clinical stability confirmed by no acute cardiovascular events (e.g. MI or stroke) in the prior three months, as verified by a cardiologist and medical records. The IHD group comprised patients with stable conditions, such as stable angina, without a documented MI. The MI group included those with a prior MI, defined per the 2012 Universal Definition of Myocardial Infarction,²⁹ based on clinical signs of ischaemia, elevated cardiac biomarkers (e.g. troponin) and echocardiographic findings. To minimise confounding effects on lipid-related biomarkers, we excluded patients on lipid-lowering therapy at recruitment or within the prior six months, as well as those on hormonal therapy, pregnant women, or individuals who declined participation. In our setting, some patients were not on lipid-lowering therapy owing to clinical contraindications (e.g. statin intolerance) or non-adherence, verified through medical records and patient interviews. We also excluded patients with acute MI, stroke, or other cardiovascular events within three months before sampling, to focus on early-stage ASCVD and to support our aim of evaluating plasma biomarkers for early detection.

The control group consisted of 55 healthy volunteers with no history of chronic or degenerative diseases, recruited from relatives of patients and staff at the Lagos State University Teaching Hospital. All participants completed a standardised, structured paper questionnaire (Online Supplementary Questionnaire) to capture general characteristics, inclusion/exclusion criteria, and potential confounding factors that could influence study outcomes. Control participants were further evaluated for clinical examination, and basic laboratory screening, including fasting plasma glucose, lipid profile, and blood pressure measurements, to confirm their apparently healthy status and exclude subclinical chronic or metabolic conditions

Blood samples from ASCVD patients were collected within 0–6 months of diagnosis to capture early-stage disease, consistent with the aim of the study to evaluate apolipoprotein A1, apolipoprotein B, and HDL phospholipids as potential early diagnostic markers. Approximately 7 mL of venous blood was drawn from each participant after an overnight fast (8 h – 12 h). Samples were collected in potassium-ethylenediaminetetraacetic acid (1 mg/mL) tubes, maintained at 4 °C, and processed within two hours. Plasma was separated by centrifugation at 3000 rpm for 5 min, aliquoted into labelled vials, and stored at –20 °C until analysis, conducted within 1–12 months. The use of potassium-ethylenediaminetetraacetic acid minimised lipid peroxidation and was appropriate for apolipoprotein and phospholipid assays. Although minor effects on total cholesterol and LDLC might occur, consistent use across all samples ensured data comparability and analyte stability.

Lipid profile parameters (total cholesterol, high-density lipoprotein cholesterol [HDLC] and triglycerides), were measured via commercial kits (Randox Laboratories Ltd., Crumlin, Northern Ireland, United Kingdom).

Total cholesterol and HDLC (following the removal of all non-HDL lipid fractions with phosphotungstic acid and magnesium ions) were estimated using the methods of Allain et al.³⁰ Triglycerides were measured via the method developed by Bucolo and David.³¹ The calculation of LDLC was performed using the formula provided by Friedewald, Levy and Fredrickson,³² selected for its validation, cost-effectiveness, and suitability in resource-limited settings such as sub-Saharan Africa, where direct methods (e.g. ultracentrifugation or the Martin-Hopkins equation) are often impractical. All samples had triglyceride levels below 4.52 mmol/L, confirming the validity of the equation. Based on these measurements, we further evaluated derived lipid indices, including LDLC, non-HDLC, Castelli’s risk indices (CRI-I, CRI-II), atherogenic coefficient (AC), and atherogenic index of plasma (AIP), using standard lipid panel measurements (total cholesterol [TC], triglycerides, HDLC). Formulas and clinical relevance are provided in Online Supplementary Table 1.^32,33,34

Plasma total phospholipids were analysed via the Wako Phospholipid C Assay Kit (Fujifilm Wako Pure Chemical Industries, Ltd., Osaka, Japan), which employs the N-ethyl-N-(2-hydroxy-3-sulfopropyl) 3,5-dimethoxy aniline sodium salt method.³⁵ High-density lipoprotein phospholipids were assayed via this method after treatment of participants’ plasma with phosphotungstic acid and magnesium ions (to remove all non-HDL phospholipids fractions).³⁶

The plasma levels of human Apo A1 and Apo B in the participants were determined using the sandwich enzyme-linked immunosorbent assay technique,³⁷ with assay kits sourced from Elabscience Biotechnology Limited, Wuhan, China. Analyses were performed on a Bio-Rad iMark™ Microplate Reader (Bio-Rad Laboratories, Hercules, California, United States) at 450 nm. The system performs optical calibration and data processing automatically, following the manufacturer’s instructions. Quality control sera were included in each batch to ensure analytical precision and reliability.

Sandwich enzyme-linked immunosorbent assay was selected for apolipoprotein analysis because of its high specificity, sensitivity and practical advantages over Western blotting and radioactive immunoassays. Spectrophotometry was chosen for lipid profiling because of its efficiency and cost-effectiveness, providing direct measurement of lipid fractions. Chromatographic methods, although precise, require more resources and technical skill, making spectrophotometry a more accessible option for routine analysis in sub-Saharan Africa.

Statistical analyses were conducted using IBM SPSS Statistics (version 21, IBM Corp., Armonk, New York, United States). Data completeness and normality were verified, with continuous variables reported as means ± standard deviation, and categorical variables as frequencies (percentages). Normality of lipid and apolipoprotein outcomes (Apo A1, Apo B, HDL phospholipids, TC, LDLC, HDLC, triglycerides, non-HDLC, CRI-I, CRI-II, AC, AIP) was assessed using Shapiro–Wilk tests and Q–Q plots (n = 227). Non-normal variables were log-transformed. Levene’s test confirmed homogeneity of variances (p > 0.05).

Group differences (ASCVD vs control) were analysed using independent samples t-tests and one-way analysis of variance with Bonferroni correction (p < 0.001). Effect sizes were reported as Cohen’s d, r², and partial η². Separate analysis of covariance models, with age as the covariate, were used for each lipid and apolipoprotein outcome owing to its significant correlation with outcomes (p < 0.05, Pearson’s correlation; Online Supplementary Table 2)

Other potential confounders (sex, BMI, smoking, hypertension) were non-significant (p > 0.05) and excluded, along with diabetes and lipid-lowering medication use.

Analysis of covariance assumptions were met: linearity (via scatterplots and Pearson’s correlation), homogeneity of regression slopes (group × covariate interaction, p > 0.05), normality of residuals (Shapiro–Wilk and Q–Q plots), homoscedasticity (Breusch-Pagan test, p > 0.05), and independence of residuals (Durbin–Watson ≈ 2.0). Age-adjusted means and partial eta-squared (η²) values for lipid parameters are presented in Online Supplementary Table 3 to Online Supplementary Table 9.

A total of 172 confirmed ASCVD patients (73 men, 99 women) aged 20–64 years were enrolled, including those with IHD (n = 40), ischaemic stroke (n = 25), hypertensive heart disease with significant atherosclerosis (n = 45), MI (n = 20), and congestive heart failure with a history of IHD or MI (n = 42). The control group comprised 55 healthy volunteers (24 men, 31 women) aged 20–64 years. Atherosclerotic cardiovascular disease patients showed significantly higher mean plasma TC, LDLC, non-HDLC, and triglycerides levels than controls (all p < 0.001, Table 1). Similarly, all conventional lipid indices (CRI-I, CRI-II, AC, and AIP) (p < 0.001), Apo B (p = 0.002), and the Apo B/A1 ratio (p < 0.001) were elevated significantly.

Conversely, mean plasma HDLC, Apo A1, and HDL phospholipid levels were significantly lower in ASCVD patients (all p < 0.001). After adjusting for age using analysis of covariance, significant differences remained in all lipid and apolipoprotein parameters between ASCVD patients and controls. Full statistical details are provided in Online Supplementary Table 3 to Online Supplementary Table 10.

Significant increases were observed in the mean plasma levels of TC, LDLC, non-HDLC, and triglycerides (p < 0.001) across ASCVD subclasses and controls (Table 2). Similarly, all conventional lipid indices (CRI-I, CRI-II, AC, and AIP) showed marked elevations (p < 0.001). In contrast, plasma HDLC levels were significantly reduced (p < 0.001) across the ASCVD subclasses and controls. Patients with IHD and MI had the highest levels of atherogenic lipids (triglycerides and TC) and lipoproteins (LDLC and non-HDLC), with mean values significantly elevated compared to controls and other ASCVD subgroups (p < 0.001). These groups also presented the lowest HDLC levels (p < 0.001), indicating a more atherogenic lipid profile.

Significant increases in mean plasma Apo B, and the Apo B/A1 ratio (p < 0.001) were observed across ASCVD subclasses and controls (Table 3). Conversely, mean plasma Apo A1 and HDL phospholipid levels were significantly reduced (p < 0.001). The highest Apo B and Apo B/A1 ratios were found in patients with MI and IHD, who also had the lowest levels of Apo A1 and HDL phospholipids. The Apo B/A1 ratio proved to be a superior marker for ASCVD stratification compared to Apo A1 or Apo B alone.³⁸ For instance, Apo B levels did not differ significantly when patients suffering from stroke and heart failure were compared with the control group (Online Supplementary Table 10), whereas the Apo B/A1 ratio showed statistically significant differences across all ASCVD subclasses when each subclass is compared with the control group (Online Supplementary Table 11).

A significant negative correlation was observed between mean plasma Apo A1 levels and atherogenic lipids, including triglycerides (r = –0.386, p < 0.001), TC (r = –0.471, p < 0.001), LDLC (r = –0.483, p < 0.001), and non-HDLC (r = –0.500, p < 0.001) (Table 4). In contrast, Apo A1 showed a strong positive correlation with antiatherogenic HDLC (r = 0.386, p < 0.001).

Apo B was positively correlated with atherogenic lipids and lipoproteins (triglycerides, TC, LDLC, and non-HDLC). In addition, the Apo B/A1 ratio exhibited strong positive correlations with conventional lipid indices, including CRI-I (r = 0.626, p < 0.001), CRI-II (r = 0.601, p < 0.001), AC (r = 0.613, p < 0.001), and AIP (r = 0.604, p < 0.001). High-density lipoprotein phospholipid levels correlated positively with Apo A1 (r = 0.524, p < 0.001) but negatively with Apo B (r = –0.174, p = 0.009) and the Apo B/A1 ratio (r = –0.586, p < 0.001).

Graded variations were observed in mean plasma Apo B concentrations across ASCVD subclasses (Figure 1), with the highest levels found in MI patients. Similarly, the mean plasma Apo B/A1 ratio showed a graded increase (Figure 2), peaking in IHD patients. In contrast, mean plasma Apo A1 levels exhibited a graded decline (Figure 3), with the lowest levels recorded in MI patients.

The mean total plasma phospholipid levels remained within a narrow range across ASCVD patients and controls (Figure 4), with no significant difference between the groups (Table 1). However, the mean plasma HDL phospholipid fraction was significantly higher in controls than in ASCVD patients, with the lowest levels observed in those with MI and IHD.

The findings of this study carry significant clinical and public health relevance for Nigerian and sub-Saharan African populations. The elevated apolipoprotein B/A1 ratio, coupled with changes in apolipoprotein A1, apolipoprotein B, and HDL phospholipids, indicates a higher risk of premature ASCVD. The apolipoprotein B/A1 ratio emerges as a key marker for ASCVD risk, reinforcing the need for earlier diagnosis and targeted interventions. The findings also underscore the importance of early screening and prevention programmes tailored to these regions.

Future research should focus on tracking changes in the apolipoprotein B/A1 ratio over time and assessing targeted interventions for ASCVD prevention in sub-Saharan Africa. Developing tailored screening programmes and investigating genetic and environmental factors influencing ASCVD risk are crucial. Comparative studies across sub-Saharan Africa can refine risk assessment, while public health initiatives and policy research may help to reduce ASCVD mortality in these regions.

While the findings provide important insights, certain factors warrant consideration. Unmeasured variables such as diet, physical activity and genetic predisposition may have influenced apolipoprotein and HDL phospholipid levels. The mean age difference of about 10 years between cases and controls, despite statistical adjustment, may also introduce residual confounding. In addition, the moderate areas under curve observed in preliminary analyses for Apo A1, Apo B, and HDL phospholipids probably reflect the sample size and exploratory design of the study. Further research with larger, diverse cohorts is needed to validate these findings and to refine clinically relevant cut-off values through advanced modelling approaches.

This study identified associations between lipid profile disruptions and ASCVD in Nigerian patients, characterised by elevated atherogenic lipids (triglycerides, TC, LDLC, and non-HDLC) and reduced HDLC, particularly in those with MI and IHD. Apolipoprotein analysis showed decreased Apo A1 and increased Apo B, leading to a higher Apo B/A1 ratio and indicating a greater atherogenic burden. Reduced HDL phospholipid levels further suggest impaired HDL functionality. The Apo B/A1 ratio emerged as a key marker for lipid-related ASCVD risk, highlighting its potential for early diagnosis and risk stratification. These findings underscore the need for targeted screening and prevention strategies to improve cardiovascular outcomes in this population.