Non-specific ST-T abnormalities are associated with major adverse cardiovascular and cerebrovascular events (MACCE) in hospitalized systemic lupus erythematosus patients

Abstract

Objective

This study aims to explore the prevalence of non-specific ST-segment and T-wave abnormalities (NSSTTA) in patients with SLE and evaluate the predictive value of NSSTTA indices for major adverse cardiovascular and cerebrovascular events (MACCE) in SLE patients.

Methods

This study retrospectively enrolled 395 hospitalized SLE patients and 420 controls from health examinations, during the period from January 1, 2014, to December 31, 2024. After propensity score matching (PSM), ECG indices were compared between the SLE group and the control group. Multivariable Cox regression analysis was used to assess the association between NSSTTA and MACCE.

Results

Compared with the control cohort, patients with SLE demonstrated significantly higher prevalence rates of NSSTTA (16.7% vs. 6.4%, p < 0.001). Multivariable Cox proportional hazards analysis further identified advancing age (HR = 1.027, 95% CI: 1.002–1.053, p = 0.033), comorbid interstitial lung disease (ILD, HR = 3.024, 95% CI: 1.477–6.193, p = 0.002), non-specific ST-T abnormalities (HR = 3.082, 95% CI: 1.209–7.853, p = 0.018), and elevated serum uric acid (HR = 1.004, 95% CI: 1.001–1.007, p = 0.021) as independent predictors for MACCE in the SLE cohort.

Conclusions

NSSTTA are independently associated with adverse outcomes and facilitate cardiovascular risk stratification in patients with SLE.

Supplementary Information

The online version contains supplementary material available at 10.1186/s41927-026-00641-z.

Introduction

Cardiovascular disease (CVD) is a leading cause of death in patients with systemic lupus erythematosus (SLE) [1]. Compared with the controls, SLE patients have an approximately twofold increased risk of cardiovascular events [2]. However, conventional cardiovascular risk factors cannot adequately account for the excess cardiovascular risk burden in SLE patients [3]. Currently, some cardiovascular risk stratification tools, such as QRISK3 and mFRS, have been found to be applicable to SLE patients [4, 5]. These tools have a broad scope of application but lack specificity. For instance, in the scoring criteria of QRISK3, only two items are directly related to SLE disease: a history of glucocorticoid exposure and the presence of SLE/rheumatoid arthritis. Therefore, exploring novel indicators and constructing more specific cardiovascular risk assessment models for SLE remains an urgent and necessary task.

As an inexpensive and non-invasive diagnostic modality, the electrocardiography (ECG) can effectively detect various cardiac abnormalities, including arrhythmias, myocardial ischemia, and structural cardiac changes. Given these advantages, ECG holds significant potential for cardiovascular risk management in SLE populations. Previous studies have demonstrated that approximately 37.5%-44% of SLE patients exhibit minor non-specific ST-segment and T-wave abnormalities (NSSTTA) on ECG [6, 7]. While these abnormalities are recognized markers of myocardial injury and possess established predictive value for cardiovascular events in general populations [8], their specific prognostic significance for CVD risk stratification in SLE patients remains to be elucidated.

Ventricular repolarization abnormalities are a well-established risk factor for ventricular arrhythmias and sudden cardiac death. In addition to the QT interval, some novel electrocardiographic indices have been identified in recent years for assessing ventricular repolarization, including the T-peak to T-end interval (T_p−e), T_p−e/QT ratio, and the corrected index of cardiac electrophysiological balance (iCEBc) [9]. While these novel indices have been extensively investigated in other clinical contexts, their exploration in SLE populations remains limited.

Our study aims to compare the prevalence of electrocardiographic abnormalities between SLE patients and matched controls, and to investigate whether specific ECG markers can predict the occurrence of major cardiovascular and cerebrovascular events (MACCE) in this high-risk population.

Methods

Study population

In this study, 395 SLE patients were collected from the database of the Department of Integrated Traditional Chinese and Western Medicine/Rheumatology and Immunology at the First Affiliated Hospital of Chongqing Medical University, from January 2014 to December 2024 (Fig. 1.). The diagnosis of SLE was based on the 2019 European League Against Rheumatism (EULAR)/American College of Rheumatology (ACR) Classification Criteria. The inclusion criteria were an age ≥ 18 years and the availability of 12-lead resting ECG records during hospitalization. The exclusion criteria included (1) pregnancy; (2) acute coronary syndrome (ACS) or acute cerebrovascular accident (CVA); (3) acute infections; (4) severe hepatic/renal dysfunction; (5) malignancy; (6) electrolyte disturbances; (7) incomplete clinical data. Additionally, 420 individuals who underwent routine health examinations at the same hospital during the study period were recruited as the control group.

Fig. 1.

Flowchart of recruiting study cohorts. PSM, propensity score matching

Outcome measures

ECG abnormalities

The 12-lead ECGs (25 mm/s paper speed and 10 mm/mV amplitude) were analyzed by two experienced cardiologists who were blinded to the clinical information. In this study, first-degree atrioventricular block and various bundle branch blocks were collectively categorized as conduction blocks. Sinus tachycardia (heart rate ≥ 100 bpm), sinus bradycardia (heart rate < 60 bpm), and sinus irregularity were categorized as sinus arrhythmia. Atrial premature beat and atrial fibrillation were categorized as atrial arrhythmia. Ventricular arrhythmia only included ventricular premature beat. Other ECG abnormalities included: clockwise rotation, left/right axis deviation, short PR interval, and right ventricular hypertrophy. When more than one abnormality was observed in the same patient, each was recorded separately. And NSSTTA were defined according to MCs 4 − 3, 4–4, 5 − 3, and 5 − 4.

The QT interval was calculated and adjusted for the heart rate using Bazett’s formula (QTc = QT/√RR). The T_p−e was measured as the duration between the T-wave peak and its end. The end point is defined as the intersection point of the isoelectric baseline with the tangent line drawn at the maximal descending slope of the T-wave. The corrected T-peak to T-end interval (cT_p−e) was calculated as the ratio of T_p−e to √RR. The iCEBc was calculated as the ratio of QTc to QRS duration.

Covariates

Demographic information was collected, including age, sex, disease duration, smoking status, comorbidities/complications, medication history, body mass index (BMI), pulse, systolic and diastolic blood pressure. Based on the Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2K) scores, SLE patients were stratified into mild activity (0–6 score) and moderate-to-severe activity (> 6 score) groups.

Laboratories included a complete blood count, serum creatinine, serum cystatin C, C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), complements C3 and C4, as well as an extensive autoantibody profile comprising ANA, anti-U1 RNP, anti-Sm, anti-SSA, anti-SSB, anti-Scl-70, anti-dsDNA, AnuA, anti-P antibodies and anti-histone antibody (AHA).

Clinical endpoints and follow-up protocol

The endpoint of this study was MACCE, including non-fatal myocardial infarction (MI), CVA, heart failure (HF), cardiac death, and all-cause mortality. Follow-up was conducted through telephone interviews and electronic medical record review until March 2025, with the endpoint defined as the first occurrence of any adverse event post-discharge.

Statistical analysis

PSM was performed at a 1:1 ratio to minimize confounding effects. Continuous variables with normal distribution were compared using Student’s t-test, while non-normally distributed quantitative data were analyzed with the Wilcoxon rank-sum test. Categorical variables were assessed by chi-square or Fisher’s exact tests as appropriate. Univariable and multivariable Cox proportional hazards regression models identified risk factors for cardiovascular outcomes in SLE patients. Analyses were conducted using SPSS 26.0 (IBM Corp.).

Results

Baseline characteristics of SLE patients

The baseline characteristics of the patients with SLE stratified by disease activity according to the SLEDAI-2K was shown in Supplementary Table 1. A total of 395 patients with SLE were enrolled. The median age was 42 years, and 29 (7.34%) of the participants were men. The median SLEDAI-2K was 6 (IQR, 4–9), and the median disease duration was 12 (IQR, 3–78) months. Patients with a higher SLEDAI-2K were younger (p = 0.013), and demonstrated elevated R_V5+S_V1 amplitude (p = 0.021). And there was a trend toward higher male representation was observed in the moderate-to-severe activity group (p = 0.051). The overall incidence of NSSTTA was 72 (18.2%). Except for a significantly longer QT interval in patients with mild disease activity (p = 0.025), no statistically significant differences were observed in other ventricular repolarization indices between patients with mild and moderate-to-severe disease activity, including QTc, Tp-e, Tp-e/QT, and iCEBc.

Baseline characteristics after PSM: SLE group vs. control group

The covariate balance between the SLE and control groups is shown in Table 1 PSM (1:1 ratio) based on age, sex, CHD, HTN, DM, hyperthyroidism, hypothyroidism, and BMI resulted well-balanced groups (n = 311 per group).

Table 1.

Covariate balance between the SLE and control groups before and after PSM

Characteristics	Before matching			After matching
	SLE (N = 395)	Control (N = 420)	p value	SLE (N = 311)	Control (N = 311)	SMD
Age, years	42 (30, 51)	41 (32.75, 51.25)	0.078	42 (29.5,51)	40 (31, 51)	0.099
Male, n(%)	29 (7.34%)	38 (9.05%)	0.376	26 (8.36%)	37 (11.9%)	0.107
BMI, kg/㎡	21.49 (19.63, 23.86)	21.83 (19.99, 24.01)	0.079	21.63 (19.57, 23.85)	21.74 (19.95, 23.99)	0.112
HTN, n(%)	55 (13.92%)	53 (12.62%)	0.583	35 (11.25%)	43 (13.83%)	0.078
DM, n(%)	15 (3.8%)	19 (4.52%)	0.604	10 (3.22%)	13 (4.18%)	0.051
CHD, n(%)	13 (3.29%)	2 (0.48%)	0.003	0 (0%)	2 (0.64%)	0.114
Hyperthyroidism, n(%)	9 (2.28%)	2 (0.48%)	0.026	3 (0.96%)	2 (0.64%)	0.036
Hypothyroidism, n(%)	45 (11.39%)	8 (1.9%)	< 0.001	8 (2.57%)	7 (2.25%)	0.021

Abbreviation: SMD, standardized mean differences

Key laboratory and electrocardiographic indices were compared between the SLE and control groups after PSM (Table 2). Compared with controls, SLE patients exhibited significantly higher prevalence of NSSTTA (p < 0.001), left ventricular high voltage (p = 0.04), and low voltage (p < 0.001). And SLE patients demonstrated increased ventricular repolarization indices, including T_p-e, cT_p-e and T_p-e/QT (all p < 0.001). Faster resting heart rates, elevated R_V5+S_V1 amplitude, and more negative QRS axis angles (all p < 0.05) were also observed in the SLE cohort.

Table 2.

Characteristics of SLE and control groups after PSM

Characteristics	SLE (N = 311)	Control (N = 311)	p value
Pulse, bpm	85 (78,96.5)	79 (72,85)	< 0.001
SBP, mmHg	120 (112,131)	113 (103,124)	< 0.001
DBP, mmHg	77 (69,86)	72 (65,78)	< 0.001
Hemoglobin, g/L	117 (101, 128)	135 (127, 141)	< 0.001
White blood cell count, 109/L	5 (3.62, 6.85)	5.43 (4.57, 6.37)	< 0.001
Platelet count, 109/L	178 (135.25, 228)	230 (194.5, 272.5)	< 0.001
Serum albumin, g/L	39 (36, 43)	44 (43, 46)	< 0.001
Serum uric acid, umol/l	313 (274,376)	283 (241.5,351)	< 0.001
Serum creatinine, mmol/l	60 (52, 70)	61 (54, 69)	0.278
Serum cystatin C, mg/L	1.02 (0.86, 1.23)	0.77 (0.7, 0.84)	< 0.001
Sinus arrhythmia, n(%)	37 (11.9%)	44 (14.15%)	0.404
Atrial arrhythmia, n(%)	2 (0.64%)	3 (0.96%)	1.0
Ventricular arrhythmia, n(%)	3 (0.96%)	1 (0.32%)	0.616
NSSTTA, n(%)	52 (16.7%)	20 (6.4%)	< 0.001
Left ventricular high voltage, n(%)	6 (1.93%)	0 (0%)	0.04
Conduction disorders, n(%)	6 (1.9%)	8 (2.6%)	0.589
Low voltage, n(%)	24 (7.72%)	5 (1.61%)	< 0.001
Other types of ECG abnormality, n(%)	14 (4.5%)	6 (1.93%)	0.069
Heart rate, bpm	78 (70, 86)	69 (63, 75)	< 0.001
QT interval, ms	368 (346, 388)	391 (375.5, 411)	< 0.001
QTc, ms	416.89 (404.25, 430.82)	419.45 (407.57, 432.76)	0.101
QRS axis, °	51 (23, 71)	65 (45.5, 77)	< 0.001
T-wave axis, °	38 (30, 54)	47 (21.5, 57)	0.101
Tp−e, ms	78 (71, 84)	66 (61, 72)	< 0.001
cTp−e, ms	88.08 (80.47, 97.94)	70.79 (64.53, 76.83)	< 0.001
Tp−e/QT	0.21 (0.19, 0.23)	0.17 (0.15, 0.18)	< 0.001
iCEBc	4.78 (4.41, 5.26)	4.74 (4.43, 5.04)	0.058
RV5+SV1, mV	1.88 (1.46, 2.41)	1.72 (1.39, 2.09)	0.001

Baseline characteristics of SLE patients with and without NSSTTA

The baseline characteristics of SLE patients stratified by the presence of NSSTTA are detailed in Supplementary Table 2. Compared to patients without NSSTTA, those with NSSTTA were significantly older (p = 0.002), had a longer disease duration (p = 0.009), and exhibited higher systolic and diastolic blood pressures (p = 0.004 and p = 0.008, respectively). The NSSTTA subgroup also demonstrated a greater burden of comorbidities, including HTN (p = 0.001), CHD (p = 0.022), HF (p = 0.002), and NPSLE (p = 0.018). Electrocardiographically, patients with NSSTTA had faster resting heart rates (p = 0.048), prolonged QTc (p < 0.001) and cTp-e intervals (p = 0.003), and a more leftward QRS axis (p = 0.003). Glucocorticoid use was more frequent in patients with NSSTTA (p = 0.035).

Univariable and multivariable Cox regression analyses

During a median follow-up period of 42 months, 7 patients with SLE were lost to follow-up. Among the remaining patients, a total of 37 MACCE were observed. Specifically, there were 17 readmissions for HF, 10 cases of myocardial infarction, 3 CVA, and 7 deaths. Regarding mortality, 3 fatalities were attributed to lupus disease activity, 1 resulted from sudden cardiac death, 1 was due to unrelated comorbidities, and 2 cases had undetermined etiology.

Univariable Cox regression analysis was performed for all laboratory and electrocardiographic parameters, demographic characteristics, comorbidities, and significant medication exposure, including glucocorticoids, antiplatelet agents, and statins. Variables significant in univariable Cox regression (p < 0.05) were considered for multivariable analysis. As a result, age (HR = 1.027, 95% CI: 1.002–1.053, p = 0.033), ILD (HR = 3.024, 95% CI: 1.477–6.193, p = 0.002), NSSTTA (HR = 3.082, 95% CI: 1.209–7.853, p = 0.018) and serum uric acid (HR = 1.004, 95% CI: 1.001–1.007, p = 0.021) were determined to be independent prognostic factors for adverse outcome in patients with SLE (Table 3).

Table 3.

Univariable and multivariable Cox regression for MACCE in SLE patients

Characteristics	Univariable analysis		Multivariable analysis
	HR (95% CI)	p value	HR (95% CI)	p value
Age, years	1.046 (1.022–1.071)	< 0.001	1.027 (1.002–1.053)	0.033
HTN, n(%)	3.116 (1.556–6.240)	0.001
ILD, n(%)	3.301 (1.437–7.583)	0.005	3.024 (1.477–6.193)	0.002
HF, n(%)	4.378 (1.915–10.008)	< 0.001
CVA, n(%)	3.916 (1.199–12.792)	0.024
NSSTTA, n(%)	2.937 (1.506–5.727)	0.002	3.082 (1.209–7.853)	0.018
QRS axis, °	0.986 (0.977–0.995)	0.003
T-wave axis, °	0.987 (0.977–0.996)	0.007
Serum uric acid, umol/l	1.004 (1.000–1.007)	0.032	1.004 (1.001–1.007)	0.021
Serum cystatin C, mg/L	2.516 (1.142–5.540)	0.022
Anti-Scl-70, n(%)	8.838 (1.136–68.759)	0.037

In sensitivity analysis, we built one additional model to examine the stability of these associations: Model 2 further adjusted Model 1 for traditional cardiovascular risk factors (gender, LDL, BMI, and smoking history), and SLE-specific factors (SLEDAI-2K and LN). The result of the sensitivity analysis, presented in Supplementary Table 3.

Discussion

This study identified two major findings. First, after adjusting for several cardiovascular risk factors, the prevalence of NSSTTA remained significantly higher in patients with SLE compared to the control group. Second, multivariate Cox analysis identified NSSTTA as an independent risk factor for MACCE in hospitalized SLE patients.

It is widely accepted that the ventricular wall comprises epicardial, endocardial, and mid-myocardial cells. These three cell types exhibit differences in action potential characteristics, ion channel expression, and pharmacological responses. Any factor that alters currents such as the transient outward potassium current, the rapid delayed rectifier potassium current, and the slow delayed rectifier potassium current in epicardial, endocardial, or M cells can increase transmural dispersion of repolarization, subsequently manifesting as ST-segment and/or T-wave changes on the electrocardiogram [10]. Primary ST-T changes are mainly attributed to intrinsic myocardial repolarization abnormalities, such as myocardial ischemia or electrolyte imbalances, whereas secondary ST-T changes often result from structural or electrical conduction abnormalities, like bundle branch block or ventricular hypertrophy [11]. NSSTTA represents a minor, nonspecific electrocardiographic alteration that can also be present in the general population. Its underlying mechanisms are complex and likely multifactorial. Nevertheless, numerous studies have indicated an association between NSSTTA and an increased risk of cardiovascular diseases [12–14]. Consequently, NSSTTA may serve as an electrocardiographic indicator for cardiovascular risk stratification in SLE patients.

Our study also identified ILD and serum uric acid levels as independent risk factors for MACCE in SLE patients. Increased afterload due to pulmonary vasoconstriction is a significant mechanism contributing to right ventricular remodeling and dysfunction in SLE patients with ILD [15]. Beyond HF, ILD is also recognized as a risk factor for myocardial infarction [16, 17]. Separately, elevated serum uric acid levels have been reported as a predictor of multi-system morbidity, including cardiovascular, pulmonary, neurological, and renal complications, in the SLE population [18].

In recent years, novel indices such as the T_p-e interval, T_p-e/QT ratio, cT_p-e, and iCEB have been proposed as markers of ventricular repolarization heterogeneity, demonstrating predictive value for ventricular arrhythmias, cardiac death, and all-cause mortality [19–23]. In the present study, these ventricular repolarization parameters were significantly increased in SLE patients compared to controls. However, they failed to show significant predictive value in the survival analysis, which might be attributable to the relatively low incidence of malignant arrhythmias in our SLE cohort.

This study has several limitations. First, it was a single-center, retrospective study with a relatively small sample size. Second, although we adjusted for some relevant covariates using exclusion criteria and propensity score matching, potential residual confounding factors may persist, such as individual and cumulative doses of corticosteroid and HCQ therapy, and the potential for ECG changes being influenced by the heightened stress of the hospitalization setting itself. Third, the study lacked comprehensive cardiac assessments, including echocardiography, magnetic resonance imaging, or biomarkers of myocardial injury, which precluded the determination of the underlying causes for the observed ECG abnormalities in SLE patients.

Conclusion

In conclusion, the present study indicated the increased prevalence of NSSTTA in SLE patients, and firstly proposed NSSTTA as the predictor for CVD risk assessment in this population.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

SLE

Systemic lupus erythematosus

NSSTTA

Non-specific ST-segment and T-wave abnormalities

ECG

Electrocardiography

MACCE

Major adverse cardiovascular and cerebrovascular events

PSM

Propensity score matching

ACS

Acute coronary syndrome

CVA

Cerebrovascular accident

BMI

Body mass index

SLEDAI-2K

Systemic Lupus Erythematosus Disease Activity Index 2000

AHA

Anti-histone antibody

MI

Myocardial infarction

HF

Heart failure

NPSLE

Neuropsychiatric SLE

LN

Lupus nephritis

ILD

Interstitial lung disease

LMV

Lupus mesenteric vasculitis

HTN

Hypertension

DM

Diabetes mellitus

CHD

Coronary heart disease

LVH

Left ventricular hypertrophy

LAD

Left axis deviation

Author contributions

Chen Jie: Conceptualization, Data curation, Methodology, Statistical analysis, Writing, review and editing; Zhou Xiaoli: Conceptualization, Writing, review and editing; Lu Shidie: Data collecton.

Funding

This research did not benefit from specific funding.

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author or the first author on reasonable request.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the Declaration of Helsinki and was approved by the Ethics Committee of the First Affiliated Hospital of Chongqing Medical University (protocol code: CY2024-319-02, date of approval: 11 November 2024). Due to the retrospective nature of the study, which involved the analysis of de-identified data from medical records, the requirement for written informed consent from participants was waived by the same ethics committee, in accordance with the applicable national guidelines and regulations.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.