N-terminal pro-brain natriuretic peptide is a biomarker for cardiovascular damage in systemic lupus erythematous: a cross-sectional study

Abstract

Objectives

Prediction models based on traditional risk factors underestimate cardiovascular (CV) risk in systemic lupus erythematosus (SLE). In a large sample of unselected SLE patients, we investigated cross-sectional associations of NT-proBNP with cardiovascular damage (CVD).

Methods

Serum NT-proBNP was measured in SLE patients enrolled in the MUHC Lupus Clinic registry. Serum was collected between March 2022 and April 2023 at annual research visits. The primary outcome was CVD identified on the SLICC Damage Index. Factors associated with CVD and NT-proBNP levels were determined.

Results

Overall, 270 SLE patients [female 91%, median age 50.7 (first quartile to third quartile: 39.6–62.1) years] were analysed for the primary outcome. Among them, 33 (12%) had CVD. The ROC curve for NT-proBNP demonstrated strong associations with CVD (AUC 0.78, 95% CI 0.69–0.87) with a threshold of 133 pg/ml providing the best discrimination for those with/without CVD. Hypertension (OR 3.3, 95% CI 1.2–9.0), dyslipidaemia (OR 3.6, 95% CI 1.3–9.6) and NT-proBNP >133 pg/ml (OR 7.0, 95% CI, 2.6–19.1) were associated with CVD in the multivariable logistic regression model. Increased NT-proBNP levels were associated with age (OR 4.2, 95% CI 2.2–8.3), ever smoking (OR 1.9, 95% CI 1.0–3.5), reduced eGFR (4.1, 95% CI 1.3–13.1), prior pericarditis/pleuritis (OR 2.5, 95% CI 1.4–4.5) and aPL antibodies (OR 2.6, 95% CI 1.4–4.9).

Conclusion

NT-proBNP is a biomarker for CV damage in SLE. The novel associations of NT-proBNP levels with prior pericarditis/pleuritis and aPL antibodies suggest new avenues for research to better understand what drives CV risk in SLE.

systemic lupus erythematosus, brain natriuretic peptide, biomarkers, cardiovascular diseases

Rheumatology key messages

• Identification of biomarkers able to better stratify cardiovascular (CV) risks in SLE is crucial.

• NT-proBNP with a cut-off value of 133 pg/mL is a biomarker for CV damage in SLE.

• Elevated NT-proBNP level is associated with prior pericarditis/pleuritis and antiphospholipid antibodies.

Introduction

Despite progress in the long-term management of systemic lupus erythematosus (SLE), the mortality of SLE patients remains over 2-fold higher than in the general population [1–3]. This excess mortality is, partly, attributed to cardiovascular (CV) disease in these patients [4, 5].

Models for predicting CV events in the general population, such as the Framingham score, underestimate CV risk in SLE, potentially because they do not incorporate important risk factors such as chronic kidney disease, body mass index or lupus disease activity scores [6–9]. It is thus important to identify biomarkers that help better stratify CV risk in SLE.

N-terminal pro-brain natriuretic peptide (NT-proBNP) is a key biomarker for the diagnosis of heart failure (HF) [10, 11] and predicts CV death and cardiac events in non-lupus populations, even in the absence of heart failure [12–14]. Interestingly, increased levels of NT-proBNP have been reported in SLE [15–17] but little is known about its association with CV disease in this population. Therefore, we sought to examine the association of NT-proBNP with presence of CV damage (CVD) in a large cohort of SLE patients.

Patients and methods

Study population

We completed a cross-sectional cohort study of SLE patients to evaluate the association of NT-proBNP with CVD. Data were extracted from the McGill University Health Centre (MUHC) Lupus Clinic registry [18]. Patients with a clinical SLE diagnosis meeting American College of Rheumatology (ACR) criteria [19] have been consecutively enrolled in this cohort for over 20 years. Patients are followed with annual research visits, during which data are systematically collected on demographics, drugs, SLE-related features, disease activity (using the SLE Disease Activity Index, SLEDAI-2K), and lupus organ damage (using the Systemic Lupus International Collaborating Clinics Damage Index, SDI) [20]. Information on traditional cardiovascular risk factors, including smoking status, being overweight (body mass index >25 kg/m²), hypertension, diabetes mellitus (DM), dyslipidaemia, and reduced estimated glomerular filtration rate (eGFR), is updated annually. For our analyses, we defined hypertension, DM, and dyslipidaemia as requiring drug therapy for these conditions.

Upon enrolment into the MUHC Lupus Clinic registry, each patient’s written informed consent to publish the data was obtained. The research project has been approved by the McGill University Health Centre (MUHC) Research Ethics Board (REB). Neither patients nor public were involved in the design nor the conduction of the study.

NT-proBNP measurement

NT-proBNP was measured on a Cobas e411 analyser (Roche Diagnostics, Laval (Québec), Canada). The assay was calibrated and quality controlled using the manufacturer’s reagents. Coefficients of variation for NT-proBNP were 4.5% for the low control and 2.7% for the high control. The limit of detection of the NT-proBNP assay is set to 10 pg/ml by the manufacturer, and in the current study anything less than the limit of detection was reported as 8 pg/ml for continuous analysis. NT-proBNP measurements were undertaken between March 2022 and April 2023 on fresh serum aliquots.

Primary outcome

The primary outcome was the presence of cardiovascular damage (CVD). CVD was defined as having any of the following SDI item: angina or coronary bypass, myocardial infarction, cardiomyopathy, valvular disease, pericarditis for >6 months, cerebrovascular accident, and/or claudication.

Statistical analysis

Continuous variables are expressed as median [first quartile to third quartile]. Categorical variables are expressed as frequencies and percentages. ROC analysis was used to calculate the optimal threshold of NT-proBNP concentration associated with CVD. Clinically relevant predictor variables for the primary outcome identified in univariable analysis with a P-value <0.05 were used in a multivariable logistic regression model to identify those independently associated with CVD, with estimation of odds ratios (OR) and 95% confidence intervals (CIs). Univariable and multivariable logistic regression analyses were also performed to identify factors associated with NT-proBNP. Statistical analyses were performed with R software version 4.3.0.

Results

Characteristics of SLE patients

All SLE patients (n = 270) involved in the MUHC Lupus Clinic registry had annual research visits between March 2022 and April 2023 and were analysed. Demographic and clinical factors at assessment are shown in Table 1.

Table 1.

Characteristics of SLE patients

	Missing data	n = 270
Age, years	0 (0.0)	50.7 [39.6–62.1]
Ethnic origins:
White	0 (0.0)	161 (59.6)
Black	0 (0.0)	44 (16.3)
Asian	0 (0.0)	35 (13.0)
Others	0 (0.0)	30 (11.1)
Male sex	0 (0.0)	24 (8.9)
SLE features:
Duration of SLE disease, years	0 (0.0)	17.7 [11.6–27.2]
Mucocutaneous	0 (0.0)	210 (77.8)
Arthritis	0 (0.0)	216 (80)
Serosal (pericarditis/pleuritis)	0 (0.0)	113 (41.8)
Neuropsychiatric	0 (0.0)	32 (11.8)
Autoimmune hemolysis	0 (0.0)	47 (17.4)
Thrombocytopenia	0 (0.0)	49 (18.1)
Renal	0 (0.0)	112 (41.5)
aPL antibodies	0 (0.0)	106 (39.3)
SLEDAI score	6 (2.2)	2.0 [0.0–4.0]
SDI score	0 (0.0)	1.0 (0.0–3.0)
Cardiovascular risk factors:
Smoker ever	1 (0.4)	117 (43.3)
Hypertension	2 (0.7)	78 (28.9)
Overweight	2 (0.7)	151 (53.9)
Dyslipidaemia	2 (0.7)	52 (19.2)
Diabetes mellitus	2 (0.7)	12 (4.4)
eGFR <50%	0 (0.0)	19 (7.0)
NT-proBNP, pg/mL	0 (0.0)	95 [54–185]
Treatment at study time:
Steroids	0 (0.0)	25 (9.3)
Hydroxychloroquine	0 (0.0)	216 (80.0)
Immunosuppressive drugs	0 (0.0)	124 (45.9)
Belimumab	0 (0.0)	20 (7.4)

Results are shown as median [first quartile to third quartile] or n (%).

Immunosuppressive drugs included azathioprine, methotrexate, mycophenolate and tacrolimus. Overweight defined as BMI >25 kg/m².

Others (ethnic origins) included Hispanic (n = 12), Middle Eastern or North African (n = 9) and Indigenous (n = 6) peoples.

aPL: antiphospholipid; BMI: body mass index; eGRF: estimated glomerular filtration rate; NT-proBNP: N-terminal pro-brain natriuretic peptide; SLEDAI: SLE Disease Activity Index.

NT-proBNP as a biomarker for cardiovascular damage in SLE patients

Overall, 33 (12.2%) patients had at least one CVD item recorded on the SDI. CVD included coronary artery disease (n = 14), cerebral vascular accident (n = 12), chronic pericarditis (n = 6), valvular disease (n = 5), cardiomyopathy (n = 3) and peripheral artery disease (n = 1). Eight patients had more than one CVD.

The median level of NT-proBNP in serum was 95 pg/ml (54–185). NT-proBNP level was higher in SLE patients with CVD [281 (140–856) as compared with 84 (50–147) pg/ml in those with no CVD, P = 0.002] and was even higher in SLE patients with more than one CVD [989 (734–1725) pg/ml as compared with 194 (122–327) in those with one CVD P < 0.001] (Fig. 1A). The corresponding ROC curve showed that NT-proBNP concentration was predictive for CVD (AUC 0.78, 95% CI 0.69–0.87). The cut-off value for NT-proBNP that provided the best separation between the SLE patients with/without CVD was 133 pg/ml (Fig. 2). NT-proBNP level was >133 pg/ml in 30% (n = 71/237), 72% (n = 18/25) and 100% (n = 8/8) of SLE patients with 0, 1 and >1 CVD, respectively (Fig. 1B).

Figure 1.

NT-pro-BNP in SLE patients with cardiovascular damage. (A) Boxplots of serum level of NT-proBNP (pg/mL) in SLE patients who experienced 0 [NT-proBNP, 84 pg/mL (50–147)], 1 [194 pg/mL (122–327)] or more than 1 CVD [989 pg/mL (734–1725) pg/mL]. (B) Percentage of SLE patients with 0 (30%, n=71/237), 1 (72%, n=18/25) and more than 1 CVD (100%, n=8/8) who had a serum level of NT-proBNP >133 pg/mL. ***P<0.01, ****P<0.001

Figure 2.

Receiver operating characteristics (ROC) curve of NT-proBNP for prediction of cardiovascular damage. Area under the curve 0.78 [95% CI (0.69–0.87)] with a serum NT-proBNP level of 133 pg/mL showing a sensitivity of 79% (95% CI (64–91)] and a specificity of 70% [95%CI (64–76)] for cardiovascular damage

Univariable analysis of factors associated with CVD (Table 2) revealed that age (OR 3.6, 95% CI 1.6–8.3), White ethnicity/race (OR 3.5, 95% CI 1.4–8.7), male sex (OR 3.5, 95% CI 1.3–9.2), history of neuropsychiatric (OR 3.5, 95% CI 1.4–8.4) and renal (OR 2.4, 95% CI 1.2–5.1) SLE involvement, hypertension (OR 5.5, 95% CI 2.5–11.8), dyslipidaemia (OR 7.1, 95% CI 3.3–15.4) and NT-proBNP >133 pg/ml (OR 8.7, 95% CI 3.6–20.9) were significantly associated with CVD. In a multivariable logistic regression analysis involving six covariates, hypertension (OR 3.3, 95% CI 1.2–9.0), dyslipidaemia (OR 3.6, 95% CI 1.3–9.6) and NT-proBNP >133 pg/ml (OR 7, 95% CI 2.6–19.1) were independently associated with CVD. We still observed a clear association between high level of NT-proBNP and CVD (univariable analysis: OR 9.4, 95% CI 3.6–24.3; multivariate analysis: OR 9.8, 95% CI 3.2–30.0) after exclusion of patients with serum creatinine level above 100 µmol/l (n = 30). Interestingly, odds of CVD were lower in SLE patients receiving hydroxychloroquine (OR 0.3, 95%CI 0.1–0.8) (Table 2).

Table 2.

Multivariate analyses for independent predictor of cardiovascular damage in SLE

	OR [95% CI]*	P-value	OR [95% CI]**	P-value
Age >50 years	3.6 [1.6–8.3]	0.002	1.1 [0.4–3.5]	0.881
Ethnic origins:
White	3.5 [1.4–8.7]	0.007
Black	0.1 [0.02–1.0]	0.024
Asian	0.4 [0.09–1.7]	0.275
Others	0.8 [0.2–2.7]	0.999
Male sex	3.5 [1.3–9.2]	0.016
SLE features:
Mucocutaneous	0.9 [0.4–2.1]	0.823
Arthritis	1.1 [0.4–2.9]	0.999
Serosal (pericarditis/pleuritis)	2.1 [1.0–4.3]	0.060
Neuropsychiatric	3.5 [1.4–8.4]	0.008
Autoimmune hemolysis	0.8 [0.3–2.3]	0.999
Thrombocytopenia	0.8 [0.3–2.1]	0.811
Renal	2.4 [1.2–5.1]	0.023	2.4 [0.9–6.1]	0.062
aPL antibodies	1.8 [0.9–3.7]	0.132
SLEDAI score >2	0.5 [0.2–1.0]	0.060
Cardiovascular risk factors:
Smoker ever	1.7 [0.8–3.5]	0.192
Hypertension	5.5 [2.5–11.8]	<0.001	3.3 [1.2–9.0]	0.018
Overweight	0.9 [0.4–1.9]	0.853
Dyslipidaemia	7.1 [3.3–15.4]	<0.001	3.6 [1.3–9.6]	0.012
Diabetes mellitus	1.4 [0.3–6.9]	0.649
eGFR <50%	2.8 [0.9–8.5]	0.066
NT-proBNP> 133 pg/mL	8.7 [3.6–20.9]	<0.001	7.0 [2.6–19.1]	<0.001
Treatment:
Steroids	1.8 [0.6–5.8]	0.302
Hydroxychloroquine	0.3 [0.1–0.7]	0.005	0.3 [0.1–0.8]	0.015
Immunosuppressive drugs	1.5 [0.7–3.1]	0.352
Belimumab	0.0 [0.0–NaN]	0.147

Analysis was performed on 270 patients; 33 patients had cardiovascular damage according to SDI; six factors identified in univariable analysis (*) with a P-value <0.05 were used in a multivariable logistic regression model.

Immunosuppressive drugs included azathioprine, methotrexate, mycophenolate and tacrolimus. Overweight defined as BMI >25 kg/m².

aPL: antiphospholipid; BMI: body mass index; CI: confidence interval; eGRF: estimated glomerular filtration rate; NT-proBNP: N-terminal pro-brain natriuretic peptide; OR: odds ratio; SLEDAI: SLE Disease Activity Index.

Association of NT-proBNP with baseline clinical features

We identified NT-proBNP as a biomarker for CVD in SLE patients. We next determined which factors were associated with NT-proBNP levels. Univariable analyses (Table 3) showed that age (OR 5.0, 95% CI 2.9–8.7), White ethnicity/race (OR 2.0, 95% CI 1.2–3.4), smoking (OR 2.2, 95% CI 1.3–3.6), hypertension (OR 2.5, 95% CI 1.5–4.3), dyslipidaemia (OR 2.5, 95% CI 1.3–4.6), reduced eGFR (OR 3.3, 95% CI 1.3–8.8), history of pericarditis/pleuritis (OR 2, 95% CI 1.2–3.3) and aPL antibodies (OR 2.7, 95% CI 1.6–4.6) were associated with NT-proBNP>133 pg/ml. Multivariable logistic regression analysis involving eight covariates showed that age (OR 4.2, 95% CI 2.2–8.3), smoking (OR 1.9, 95% CI 1.0–3.5), reduced eGFR (OR 4.1, 95% CI 1.3–13.1), past pericarditis/pleuritis (OR 2.5, 95% CI 1.4–4.5) and aPL antibodies (OR 2.6, 95% CI 1.4–4.9) were associated with increased NT-proBNP (Table 3). Of note, the associations of NT-proBNP levels with age, smoking, reduced eGFR, pericarditis/pleuritis and aPL antibodies were also observed in SLE patients who had no CVD (Supplementary Table S1, available at Rheumatology online).

Table 3.

Multivariate analyses for independent predictor of high NT-proBNP in SLE

	OR [95% CI]*	P-value	OR [95% CI]**	P-value
Age >50 years	5.0 [2.9–8.7]	<0.001	4.2 [2.2–8.3]	<0.001
Ethnic origins:
White	2.0 [1.2–3.4]	0.010	1.2 [0.6–2.2]	0.625
Black	0.5 [0.3–1.1]	0.122
Asian	0.3 [0.1–0.8]	0.014
Others	1.2 [0.6–2.6]	0.688
Male sex	1.8 [0.7–4.6]	0.274
SLE features:
Mucocutaneous	1.1 [0.6–1.9]	0.999
Arthritis	1.8 [0.9–3.5]	0.112
Serosal (pericarditis/pleuritis)	2.0 [1.2–3.3]	0.010	2.5 [1.4–4.5]	0.003
Neuropsychiatric	1.3 [0.6–2.7]	0.561
Autoimmune hemolysis	0.8 [0.4–1.6]	0.617
Thrombocytopenia	0.9 [0.5–1.8]	0.871
Renal	0.8 [0.5–1.3]	0.304
aPL antibodies	2.7 [1.6–4.6]	<0.001	2.6 [1.4–4.9]	<0.001
SLEDAI score >2	0.8 [0.5–1.3]	0.360
Cardiovascular risk factors:
Smoker ever	2.2 [1.3–3.6]	0.003	1.9 [1.0–3.5]	0.040
Hypertension	2.5 [1.5–4.3]	0.001	1.1 [0.6–3.2]	0.839
Overweight	0.8 [0.5–1.4]	0.523
Dyslipidaemia	2.5 [1.3–4.6]	0.004	1.1 [0.5–2.2]	0.909
Diabetes	0.9 [0.3–3.0]	0.999
eGFR <50%	3.3 [1.3–8.8]	0.013	4.1 [1.3–13.1]	0.018
Treatment:
Steroids	1.3 [0.6–3.1]	0.510
Hydroxychloroquine	0.8 [0.4–1.4]	0.430
Immunosuppressive therapy	0.8 [0.5–1.4]	0.527
Belimumab	0.6 [0.2–1.6]	0.341

Analysis was performed on 270 patients; 97 patients had high NT-proBNP (i.e. >133 pg/mL); eight factors identified in univariable analysis (*) with a P-value <0.05 were used in a multivariable logistic regression model (**).

Immunosuppressive drugs included azathioprine, methotrexate, mycophenolate and tacrolimus. Overweight defined as BMI >25 kg/m².

aPL: antiphospholipid; BMI: body mass index; CI: confidence interval; eGRF: estimated glomerular filtration rate; NT-proBNP: N-terminal pro-brain natriuretic peptide; OR: odds ratio; SLEDAI: SLE Disease Activity Index.

Discussion

Prediction models based on traditional risk factors are less accurate at identifying CV disease in SLE as compared with the general population. Our study demonstrates that a higher level of NT-proBNP—with a cut-off value of 133 pg/ml—is associated with CVD in SLE.

According to recently published international consensus definitions, the diagnosis of HF is corroborated, in patients with symptoms or signs of the disease, by an ambulatory NT-proBNP of 125 pg/ml or higher [21]. NT-proBNP ≥125 pg/ml has been recently analyzed in non-lupus cohort study from Scotland involving 18 356 individuals without previous CV disease. In a population with a mean age of 46.1 years (SD 14.7 years), NT-proBNP ≥125 pg/ml was detected in 19.8% of female participants [22]. In our cohort of 220 females with a mean age of 51.2 years, the prevalence of NT-proBNP ≥125 pg/ml was 35% (n = 77). This prevalence approximated those observed in females older than 60 years enrolled in the above Scottish study [22].

In our study, NT-proBNP level was marked increased in patients with more than one CVD. Several biomarkers have been identified to have some predictive values for CV disease in SLE [23–25]. As a risk stratification tool, NT-proBNP measurement has the advantage of being simple, relatively inexpensive, reproducible, non-invasive, and widely available. Measured with high-sensitivity (HS) assays, cardiac troponin T has proven predictive value for both subclinical atherosclerosis and incident cardiovascular event in SLE [24, 26]. The combination of both biomarkers (i.e. NT-proBNP and HS-cTnT) might improve the identification of SLE patients at high risk of CVD. An important concern of using NT-proBNP in SLE patients is the impact of renal insufficiency on the test, particularly when eGFR is <30 mL/min/1.73 m² [27]. Although the median serum creatinine level was 68 [first quartile to third quartile: 59–83] µmol/l in our cohort, it might be useful in further studies to adopt different cut-offs stratified by renal function in SLE.

NT-proBNP is a powerful predictor of major adverse CV events in individuals without heart failure [12–14, 28]. In SLE patients, elevated NT-proBNP levels might reflect latent ongoing cardiac myocyte damage and ‘smoldering’ cardiac dysfunction. Interestingly, increased NT-proBNP levels were strongly associated with aPL antibodies in our study while a high prevalence of occult myocardial scarring—identified by means of cardiac MRI—has been reported in patients with antiphospholipid syndrome [29]. In the same line, the association of NT-proBNP levels with prior pericarditis/pleuritis might indicate past and unnoticed myocarditis [30–32].

In accordance with literature [33], the lower odds of CVD in patients receiving hydroxychloroquine suggest that HCQ is protective on cardiovascular disease in SLE. Aberrant secretion of type I interferon (IFN) such as IFNα by plasmacytoid dendritic cells (pDCs) is a key factor of SLE pathogenesis. Evidence indicates that pDCs play also an important role in atherosclerosis development. Indeed, IFNα-producing pDCs are increased in atherosclerotic plaques and pDCs depletion is found protective in murine atherosclerosis models [34, 35]. SLE patients with raised type I IFNs signature have increased carotid intima media thickness and coronary calcification score [36]. IFNα produced by pDCs drives the expansion and migration of T cells through the arterial wall leading to the development of atherosclerosis in atherosclerosis/lupus-prone mice [37]. Interestingly, HCQ has been shown to impair the ability of pDCs from patients with SLE to produce IFNα [38] and might thus contribute to a specific cardioprotection through pDCs inhibition.

Our study has several strengths. First, it is an ancillary study of a large prospective cohort of unselected SLE patients with annual research visits dedicated to the global evaluation of lupus disease and outcomes. Second, the primary end point focuses on CV events in contrast to most studies in SLE analysing surrogate markers of accelerated atherosclerosis only. Third, we determined the ideal cut-off point for NT-proBNP values through the ROC curve.

Our study has some potential limitations. First, NT-proBNP was only assessed once and serial testing may be superior to a single value to assess the risk for CV damage over time. Second, considering interlaboratory variation for NT-proBNP assessment [39], our study should be repeated at centers that use different methods of measuring NT-proBNP. Third, we do not have direct evidence of the predictive value of NT-proBNP for future CV events, as our study was cross-sectional. Fourth, the choice of the independent predictor variables included in the multivariable regression model was impeded by the limited number of CVD. Fifth, NT-proBNP is closely related to left ventricular function but echocardiograms were not done on all patients. Future larger and adequately powered prospective studies are warranted to clarify the assay standardization, the optimal cut-off, and the prognostic value of NT-proBNP for CV events in SLE patients.

In conclusion, NT-proBNP is an independent biomarker for CV damage in SLE. The novel associations of NT-proBNP levels with pericarditis/pleuritis and aPL antibodies suggest new avenues for research to better understand whether myocardial damage caused by these factors also drives CV risk in SLE. Future evaluations are needed to identify the usefulness of NT-proBNP in tailoring treatments to prevent CV burden in SLE.

Declaration of generative AI and AI-assisted technologies in the writing process

Authors did not use AI and AI-assisted technologies in the writing process.

Supplementary material

Supplementary material is available at Rheumatology online.

Data availability

The available data supporting the findings of this study are available within the article and its supplementary materials.

Contribution statement

C.A.P. is the MUHC SLE clinic director. S.B., E.V. and K.S. designed the study analyses. K.S. conducted the analysis. All authors contributed to the finalization of the manuscript.

Funding

This work was supported by the Departement du Developpement Professionnel Continu Medical, Assistance Publique Hôpitaux de Paris and the Singer Family Fund for Lupus Research.

Disclosure statement: The authors have declared no conflicts of interest.