Association Between Type 2 Diabetes Mellitus and Heart Failure: A Retrospective Study from a Tertiary Care Diabetes Centre in India

Abstract

Introduction

The study aimed to explore the association between type 2 diabetes (T2D) and heart failure (HF) using echocardiography and NT-proBNP. The study also derived an NT-proBNP cut-off for diagnosing HF by echo in Asian Indians with T2D.

Methods

A retrospective study was performed using data from individuals with T2D, aged ≥ 18 years, who visited diabetes clinics in India between March 2019 and December 2023. NT-proBNP levels were quantified by chemiluminescence, and left ventricular ejection fraction (LVEF) was assessed from echo using two-dimensional (2D) echocardiography. Heart failure was classified based on the European Society of Cardiology (ESC) guidelines. Receiver operating characteristic (ROC) curve was performed to determine the optimal NT-proBNP cut-off for diagnosing HF by echo.

Results

Among the 1189 study individuals included in the study (714 men and 475 women), 5.9% were identified as having HF with reduced ejection fraction (HFrEF), 5.5% had mildly reduced ejection fraction (HFmrEF), and 14.1% had HF with preserved ejection fraction (HFpEF) while the rest (74.5%) had LVEF > 50%. Elevated NT-proBNP levels were observed in those with reduced ejection fraction. ROC analysis identified an optimal NT-proBNP threshold of 398 pg/mL for diagnosing HF, with 87% sensitivity and 78% specificity. HF prevalence increased with age, peaking at 30.6% in individuals aged 61–70 years. Women with HF had higher NT-proBNP levels than men.

Conclusions

In this diabetes clinic population, 11.5% of individuals with T2D had moderate to reduced LVEF. Early identification of HF using echocardiography and NT-proBNP in a diabetes clinic could help improve prognosis.

Supplementary Information

The online version contains supplementary material available at 10.1007/s13300-025-01746-3.

Key Summary Points

Why carry out this study?

Heart failure is an under-recognized complication of type 2 diabetes (T2D), contributing to significant morbidity and healthcare burden, especially in Asian Indian populations.

There is a need for simple, clinic-based tools to identify heart failure (HF) early in individuals with T2D to enable timely intervention and improve outcomes.

What was learned from the study?

Among 1189 individuals with T2D (714 men and 475 women), 11.5% had moderate to reduced left ventricular ejection fraction (LVEF), and an N-terminal pro B-type natriuretic peptide (NT-proBNP) cut-off of 398 pg/mL showed 87% sensitivity and 78% specificity for detecting HF.

Early use of echocardiography and NT-proBNP in diabetes clinics could improve HF diagnosis and guide timely initiation of therapies to improve prognosis.

Introduction

The prevalence of heart failure (HF) and type 2 diabetes (T2D) are rising globally and they frequently coexist. With an estimated global prevalence of 537 million individuals with diabetes [1] and 64.3 million individuals with HF [2], both these conditions place a growing strain on healthcare systems worldwide [3, 4]. In India, approximately 101 million individuals are affected by T2D [5], and the incidence of T2D is particularly high [6] and it also occur at a younger age [7].

T2D significantly increases the risk of HF, and individuals with T2D are estimated to have 2–4 times greater risk of HF compared with those without T2D [6, 8–10]. Additionally, 30–40% of individuals with HF have coexisting T2D or impaired glucose tolerance, and this figure rises to over 50% among those hospitalized for HF [11, 12]. In India, an 18% annual rise in HF is anticipated and the prevalence of HF is about 50% in individuals with T2D [13].

A meta-analysis of various randomized clinical trials on cardiovascular outcomes involving individuals with T2D revealed that HF is one of the most prevalent non-fatal cardiovascular events, second only to myocardial infarction, with a mean prevalence between 13 and 30% [14]. In addition, people with T2D have a 33% higher likelihood of being hospitalized for HF compared with those without the condition [15]. The occurrence of undiagnosed HF among individuals with T2D is also substantial [16]. Consequently, according to the recent global definition/classification of HF, T2D has been identified as a major risk factor for developing HF, thereby placing those with T2D in the early stage of HF (stage A) [17].

B-type natriuretic peptide (BNP) and its pro‐peptide, N-terminal proBNP (NT-proBNP) serve as valuable biomarkers for supporting clinical decisions in HF diagnosis [17, 18]. Both biomarkers demonstrate prognostic value for short- and intermediate-term cardiovascular outcomes in individuals with diabetes [19, 20]. NT-proBNP is particularly useful in correlating HF risk among individuals with T2D [21] and in predicting adverse outcomes in patients with HF [22]. Thus, the 2021 ESC guidelines [23] and the revised 2017 American College of Cardiology (ACC)/American Heart Association (AHA)/Heart Failure Society of America (HFSA) guidelines [24] recommend the measurement of natriuretic peptides (NT-proBNP or BNP) to identify “pre-HF” in individuals with diabetes.

There are few studies that have reported on HF among individuals with T2D in India. Acute decompensated heart failure (ADHF) is a critical medical emergency with a high mortality rate. Advanced age, elevated serum creatinine levels, and a low left ventricular ejection fraction (LVEF) have been identified as significant contributors to readmission in hospital and increased mortality among those with ADHF [25]. The present study aimed at exploring the relationship between T2D and HF, by looking at echocardiography and NT-proBNP levels in Asian Indians adults with T2D attending a diabetes clinic. We also attempted to derive a cut point for NT-proBNP to predict HF in Indians with T2D.

Methods

This study retrospectively investigated individuals with T2D who had been treated at a network of diabetes centres where electronic records were available. This research involved de-identified information of 1189 individuals with T2D aged 18 years and above and comprised both sexes, seen between March 2019 to December 2023 and for whom data on NT-proBNP levels and echocardiogram test were available. Individuals who were pregnant or lactating, those who has an eGFR < 30 mL/min/1.73 m² and those who had not provided informed consent for the use their data in research studies were excluded. In addition, patients with a documented episode of acute HF at the time of NT-proBNP measurement were also excluded from the study as they would be mostly treated in the intensive care unit and also to avoid confounding elevations in biomarker levels. The study was approved by the institutional ethics committee of Madras Diabetes Research Foundation (approval no. MDRF/NCT/01/02/2024). Written informed consent was obtained from all participants to the use of their anonymized data for research purposes.

Every individual is enrolled and given a distinct identification number during their initial visit to the diabetes centre, enabling accurate tracking over time as described in an earlier publication [26]. After registration, a comprehensive assessment is conducted by a dietitian or diabetes educator, gathering detailed medical histories that include information on family history of diabetes and medications. Anthropometric measurements (height, weight, and waist circumference) and clinical assessment (blood pressure [BP]) were assessed using standard protocols. Weight was assessed in kilograms and height in centimetres using an electronic weighing scale and a stadiometer, respectively. Body mass index (BMI) was calculated using the formula: weight (kgs)/[height (m)]². Blood pressure readings were taken from the right arm while sitting, using a mercury sphygmomanometer. Biochemical estimations included fasting plasma glucose (FPG), postprandial plasma glucose (PPPG), HbA1c, lipid profile blood urea and serum creatinine. Data of microvascular complications (retinopathy, nephropathy and neuropathy) and macrovascular complications (cardiovascular disease [CVD] were obtained from the diabetes electronic medical records (DEMR) as described elsewhere [26]. Symptoms of HF were also obtained from DEMR, which included, breathlessness, fatigue, palpitation, orthopnea, paroxysmal nocturnal dyspnea and swelling/pain/numbness in legs.

A fasting blood sample was collected ensuring an overnight fast of 8–10 h. Following the standard Indian breakfast, which consisted of approximately 60 g of carbohydrates, participants provided a venous blood sample 90 min later to assess postprandial glucose levels. All biochemical analyses were conducted in the laboratory of the centre that had accreditation from National Accreditation Board for Hospitals and Healthcare Providers (NABH). Using the GOD-POD method the plasma glucose levels were measured, while serum cholesterol, high-density lipoprotein (HDL) cholesterol and triglycerides were analyzed using the CHOD-PAP, GPO-PAP, and direct methods, respectively. All assessments were done on a Beckman Coulter AU 2700/480 Autoanalyzer (Ireland) using commercially available kits. Using the Friedewald formula, the levels of Low-density lipoprotein (LDL) cholesterol were calculated. Using the Bio-Rad Variant machine (Hercules, CA), HbA1c was measured through high-pressure liquid chromatography. Serum creatinine levels were determined using Jaffe’s method on the Beckman Coulter AU2700 analyser (Fullerton, CA, USA) and Beckman kits. The standardized eGFR (mL/min per 1.73 m²) was calculated using the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) 2009 equation without the race correction [27].

NT-proBNP Measurements

NT-proBNP levels were estimated by the CLIA (chemiluminescence immunoassay) method using the Siemens ADVIA Centaur XPT immunoassay system. The assay offers a broad dynamic range from 35–35,000 pg/mL, with a lower detection threshold of 33 pg/mL and coefficients of variation was < 5.0%, allowing precise measurement across a wide spectrum of clinical presentations. The normal thresholds were set to ensure age-specific diagnostic accuracy [28].

Complications

Using an FF 450 Plus model mydriatic fundus camera (Carl Zeiss, Jena, Switzerland) digital fundus colour photography was conducted following dilation of the pupil by a trained photographer/optometrist to assess diabetic retinopathy. Photographs were evaluated by trained retina specialists and final diagnosis was made on the basis of the grading of the worse eye on the basis of the Early Treatment Diabetic Retinopathy Study (ETDRS) criteria [29]. The diagnosis of diabetic retinopathy required the presence of at least one identifiable microaneurysm. The neuropathy assessment involved measuring the vibratory perception threshold (VPT) of the big toes with a biothesiometer in a standardized manner, as previously described and if the mean VPT was 20 V or greater, neuropathy was diagnosed [30].

HF Measurements

Assessment of HF was carried out using two-dimensional (2D) echocardiography. LVEF was determined and the results were documented in accordance with 2023 guidelines of the European Society of Cardiology (ESC) [31]. The detection of HF with reduced EF (HFrEF), was by definition of HF symptoms ± signs (may not be always present especially in optimally treated patients) and/or LVEF ≤ 40% (Reduced ejection fraction, n = 71, 5.9%). The detection of HF with mid-range EF’ (HFmrEF), comprised of individuals with HF symptoms ± signs and/or a LVEF of 41–49% (Mildly reduced ejection fraction, (n = 65, 5.5%). The diagnosis of HF with preserved EF (HFpEF) was by definition of symptoms ± signs of HF and/or a LVEF ≥ 50% along with raised NT-proBNP following ESC guidelines [< 50 years: NT-proBNP > 450 pg/mL; 50–75 years: > 900 pg/mL; > 75 years: > 1800 pg/mL] (preserved ejection fraction, n = 219, 18.4%) and no HF (LVEF ≥ 50% and NT-proBNP ≤ 450 pg/mL (< 50 years), ≤ 900 pg/mL (50–75 years) and ≤ 1800 pg/mL (> 75 years) [31].

Statistical Methods

Estimates were expressed as mean ± standard deviation or proportions. One-way ANOVA with post hoc Tukey HSD or Student’s t test were used to compare groups for continuous variables and Chi-square test was used to compare proportions between four groups. Pearson’s correlation analysis was carried out to determine the correlation between NT-proBNP, eGFR and EF. Receiver operating characteristics (ROC) curves were constructed to identify NT-proBNP cut points with best sensitivity/specificity for predicting HF. A p < 0.05 was considered statistically significant. Statistical Package for Social Sciences version 27.0 (SPSS Inc., Chicago, IL) was used for performing analyses.

Results

Among the 1189 individuals with T2D, the average age was 56 ± 11 years with 60% (714) of them being men and 40% (475) being women. The median duration of diabetes was 6 years (Interquartile range: 1–12). FPG was 179 ± 85 mg/dL, PPPG, 274 ± 113 mg/dL and HbA1c, 9.4 ± 2.2%. The mean eGFR was 80 ± 28 mL/min/1.73 m². Statins were used by 66% of individuals (n = 789) and hypertension drugs by 27% (n = 330). Retinopathy was present in 32% (n = 194/601), and neuropathy in 79% (n = 510/643) of those assessed for these complications.

Majority (74.5%) of the individuals had no HF, 14.1% had HFpEF, 5.5% had HFmrEF and 5.9% had HFrEF. The age adjusted clinical and biochemical characteristics based on EF and NT-proBNP levels are presented in Table 1. Individuals with reduced EF (LVEF ≤ 40%, n = 71) were older and the median duration of diabetes was significantly higher compared with the no heart failure group. Individuals with reduced EF had significantly lower eGFR (65 ± 5 mL/min/1.73 m²) compared with those without heart failure (88 ± 23 mL/min/1.73 m²). Use of anti-hypertensive medications varied from 23–35% without any significant differences between groups. Statin use ranged between 60–82%. Use of antidiabetic drugs was higher among those with LVEF < 50%.

Table 1.

Clinical and biochemical characteristics of study individuals based on LVEF on Echo and NT-proBNP levels (n = 1189)

Variables	No heart failure (based on NT-proBNP) (n = 885, 74.5%)]	Preserved ejection fraction (based on NT-proBNP) (n = 168, 14.1%)	p value*	Mildly reduced ejection fraction (n = 65, 5.5%)	p value*	Reduced ejection fraction (n = 71, 5.9%)	p value*
Age (years)	58 ± 11	64 ± 12	< 0.001	60 ± 12	0.295	61 ± 10	0.068
Men n (%)	524(59)	96 (57)	–	45(69)	–	49(69)	–
Duration of diabetes (years) median (IQR)	7 (2–11)	10 (6–14)	< 0.001	10 (5–15)	< 0.001	10 (4–13)	0.022
BMI (kg/m2)	27.2 ± 4.9	27.9 ± 4.7	0.251	26.4 ± 3.7	0.277	26.1 ± 5.7	0.210
Systolic blood pressure (mmHg)	131 ± 18	132 ± 23	0.909	132 ± 25	0.808	121 ± 18	< 0.001
Diastolic blood pressure (mmHg)	79 ± 10	77 ± 12	0.084	79 ± 12	0.669	79 ± 14	0.866
HbA1c (%)	9.4 ± 2.2	9.5 ± 2.3	0.469	9.2 ± 2.3	0.546	9.4 ± 2.2	0.973
Fasting plasma glucose (mg/dL)	178 ± 84	185 ± 94	0.338	170 ± 91	0.524	168 ± 73	0.369
Post prandial glucose (mg/dL)	275 ± 111	281 ± 122	0.557	267 ± 129	0.609	271 ± 102	0.800
Serum cholesterol (mg/dL)	172 ± 51	157 ± 56	< 0.001	165 ± 66	0.379	149 ± 60	< 0.001
Serum triglyceride (mg/dL)	147 ± 41	155 ± 56	0.0299	172 ± 66	< 0.001	182 ± 60	< 0.001
Serum LDL cholesterol (mg/dL)	96 ± 40	89 ± 43	0.053	91 ± 40	0.071	96 ± 46	0.063
Serum HDL cholesterol (mg/dL)	44 ± 12	38 ± 16	< 0.001	39 ± 12	< 0.001	37 ± 11	< 0.001
eGFR (mL/min/1.73m2)	88 ± 23	56 ± 31	< 0.001	62 ± 3	< 0.001	65 ± 5	< 0.001
Retinopathy n (%)	128 (14)	40 (24)	–	14 (21.0)	–	12 (17.0)
Neuropathy n (%)	356 (40)	98 (58)	–	26 (40)	–	30 (42.2)
Signs and symptoms of heart failure n (%)	88 (10)	85 (50)	–	32 (49.2%)	–	42 (59%)
Antidiabetic drug treatment
Insulin (alone or in combination with OHA) n (%)	509 (57)	140 (83)	–	49 (75)	–	47 (66)
Metformin (alone or in combination) n (%)	647 (73)	7 (4)	–	29 (45)	–	27 (38)
Sulphonylurea (alone or in combination) n (%)	603 (68)	110 (65)	–	33 (50)	–	33 (46)
SGLT-2 (alone or in combination) n (%)	114 (13)	43 (25)	–	25 (38.4)	–	42 (59)
Other OHA (alone or in combination) n (%)	541 (61)	96 (57)	–	31 (47)	–	23 (32)
Hypertension drugs n (%)	237 (26)	60 (35)	–	15 (23.0)	–	18 (25.3)
Statin n (%)	560 (63)	138 (82)	–	48 (73.8)	–	43 (60.6)

Data are presented as mean ± SD

BMI body mass index, eGFR estimated glomerular filtration rate, HDL high-density lipoprotein, IQR Interquartile range, LDL, low-density lipoprotein, LVEF left ventricular ejection fraction, LVIDd left ventricular internal diameter at end-diastole, LVIDs left ventricular internal diameter at end-systole, OHA oral hypoglycemic agent, SGLT-2 sodium-glucose co-transporter 2

*All p value compared with no heart failure

Table 2 presents the echocardiograph parameters across different LVEF groups. As LVEF decreases, both left ventricular internal diameter at end-diastole (LVIDd) and left ventricular internal diameter at end-systole (LVIDs) progressively increased, reflecting left ventricular dilation. Individuals with reduced ejection fraction (LVEF ≤ 40%) exhibit the largest ventricular dimensions (LVIDd: 56 ± 6.7 mm, LVIDs: 45 ± 9.1 mm. Individuals with mildly reduced ejection fraction (LVEF 41–49%) also show enlarged ventricles (LVIDd: 54 ± 6.4 mm, LVIDs: 41 ± 5.4 mm). However, those with preserved ejection fraction showed normal ventricular dimension (LVIDd: 47 ± 6.9 mm, LVIDs: 31 ± 5.3 mm), and normal ejection fraction. All heart failure subtypes showed significant differences compared with the no heart failure group.

Table 2.

Echo parameters of study individuals based on LVEF on echo and NT-proBNP levels (n = 1189)

Variables	No heart failure (n = 885, 75.5%)	Preserved ejection fraction (n = 168, 14.1%)	p value*	Mildly reduced ejection fraction (n = 65, 5.5%)	p value*	Reduced ejection fraction (n = 71, 5.9%)	p value*
LVEF (%)	66 ± 4.4	60 ± 5.9	< 0.001	45 ± 2.5	< 0.001	35 ± 4.6	< 0.001
LVIDd (mm)	46 ± 5.9	47 ± 6.9	< 0.001	54 ± 6.4	< 0.001	56 ± 6.7	< 0.001
LVIDs (mm)	28 ± 4.2	31 ± 5.3	< 0.001	41 ± 5.4	< 0.001	45 ± 9.1	< 0.001

*All p value compared with no heart failure;

Data presented as mean ± SD

LVEF left ventricular ejection fraction, LVIDd left ventricular internal diameter at end-diastole, LVIDs left ventricular internal diameter at end-systole

Figure 1 presents the NT-proBNP levels based on LVEF categories. Individuals with LVEF ≤ 40% had the highest NT-proBNP levels, followed by those with (LVEF between 41–49% and the lowest levels were observed in individuals with preserved EF (LVEF ≥ 50%) and the trend was significant (p < 0.001), suggesting that worsening EF was associated with higher NT-proBNP levels. The levels of NT-proBNP increased progressively with age among both the sexes as shown in Supplementary Material (Figure S1).

Fig. 1.

Geometric mean NT-proBNP levels based on Left Ventricular Ejection Fraction (LVEF) on echocardiography

We also constructed ROC curves for NT-proBNP levels in relation to HF (defined by echocardiography) and this showed an area under the curve (AUC) of 0.869 for HF (p < 0.001) (Fig. 2). NT-proBNP cut-off of 398 pg/mL provided an optimal balance between sensitivity and specificity, with a sensitivity of 87% and a specificity of 78%. Thus, we report that a NT-proBNP cut point of 398 pg/mL can be used to differentiate individuals with T2D, with and without HF with a fair sensitivity and specificity.

Fig. 2.

ROC curve of NT-proBNP levels among individuals with and without heart failure

We next looked at the age-wise prevalence of HF among individuals with T2D defined as LVEF < 50% and NT-proBNP cut points (Fig. 3a). This showed that HF prevalence increased with age. Among the overall cohort, the prevalence of HF was highest in those aged 61 years and above. A similar pattern was observed in both men and women, with men having slightly higher rates of HF in the 61–70 age group (33.2%) compared with women (26.3%). After the age of 70 years, the prevalence declined slightly to 27.4%. When the distribution of HF was assessed with duration of diabetes (Fig. 3b), the prevalence of HF was significantly higher among those with > 10 years duration of diabetes (70.7%).

Fig. 3.

Distribution of heart failure in individuals with T2D based on age (a) and diabetes duration (b) using Echo and NT-proBNP levels (n = 304)

Pearson correlation analysis between NT-proBNP levels namely the standard cut point of ≥ 450 pg/mL mentioned in the literature and the cut point of ≥ 398 pg/mL derived by us, were tested against the LVEF on echocardiography and the results are presented in Table 3. A significant inverse relationship was observed between NT-proBNP levels and EF. For NT-proBNP levels ≥ 450 pg/mL, the correlation coefficient was −0.493 (p < 0.001), while for NT-proBNP levels ≥ 398 pg/mL, the correlation coefficient was stronger −0.599 (p < 0.001), suggesting that the latter is more suitable for the Asian Indian population especially for those with T2D.

Table 3.

Correlation between NT-proBNP and EF

		EF (echo)
NT-proBNP < 450 (pg/mL)	Pearson correlation	−0.493
	Sig. (2-tailed)	< 0.001
NT-proBNP < 398 (pg/mL)	Pearson correlation	−0.599
	Sig. (2-tailed)	< 0.001

Discussion

This paper reports on the association between T2D and HF from a tertiary care diabetes centre in India using echocardiography and NT-pro BNP estimations. The main findings from this study are as follows: (i) A total of 11.5% of individuals with T2D had moderate to reduced EF; (ii) Individuals with reduced EF had the highest NT-proBNP levels (median, 4669 pg/mL), followed by those with mildly reduced EF (1785 pg/mL) and individuals with preserved EF (162 pg/mL); (iii) A NT-proBNP cut-off value of 398 pg/mL had a sensitivity of 87%, and a specificity of 78%, for differentiating individuals with and without HF in our patients with T2D, (iv) there was an inverse correlation between EF and (v) NT-proBNP and a NT-proBNP cut off of ≥ 398 pg/mL had the best correlation with reduced EF in this study population of individuals with T2D.

According to clinical and epidemiological evidence, in addition to myocardial infarction and atherosclerosis-related CVD, HF can contribute significantly to morbidity/mortality among individuals with T2D [32]. The underlying mechanisms of T2D and HF are intricately connected. Diabetes can lead to HF, by causing myocardial dysfunction through mechanisms that contribute to the development of atherosclerosis and coronary artery disease (CAD), as well as through direct effects on the myocardium i.e. diabetic cardiomyopathy [33]. Visceral fat levels have been shown to increase with worsening glucose intolerance, which is linked to reduced adiponectin levels and elevated TNF-alpha, hs-CRP, visfatin and oxidized LDL levels, insulin resistance, and carotid intima-media thickness (IMT), further highlighting the complex relationship between fat distribution, glucose metabolism and cardiovascular risk [34]. Physical inactivity is another contributing factor to both T2D and heart disease [35]. Early asymptomatic changes in diabetic heart disease/diabetic cardiomyopathy, such as inflammation, increased stiffness, LV hypertrophy, cardiac fibrosis and subclinical diastolic dysfunction, can evolve into systolic dysfunction and the onset of symptoms associated with HF [36].

Research has shown a higher prevalence of LV dysfunction among individuals with T2D [37, 38]. Individuals with both T2D and HFrEF have been found to have up to a tenfold higher mortality [39]. An earlier observational study reported that 28% of individuals with diabetes had HFrEF and 25% had HFpEF [40]. A study conducted in older individuals with T2D reported the prevalence of HFrEF to be 4.8% and that of HFpEF to be 22.9% and the prevalence of HF increased steeply with age [38]. A study conducted in north India reported that 13.3% of the individuals with T2D had HFmrEF and 10.1% had HFrEF [41]. Our study reports a lower prevalence of HFmrEF (5.5%) and HFrEF (5.9%) among individuals with T2D. These differences could be because many of our patients underwent routine echo and NTproBNP estimations and also those with acute HF were excluded. It is possible that in the earlier studies mostly those with symptoms of HF were referred for these investigations.

Plasma levels of BNP and NT‐pro‐BNP are linked to the risk of cardiovascular mortality, HF, and stroke in individuals with CAD/T2D [42–44]. Additionally, hypertension, the prevalence of which is high in both urban and rural India, is a major risk factor for CVD [45]. Strong correlations were found between hypertension and various factors including obesity, diabetes, age, male sex, residing in urban areas and behavioural factors, such as physical inactivity and excessive alcohol consumption [45]. Furthermore, India is experiencing a diabetes epidemic, with urban regions, particularly in southern India, showing a sharp increase in prevalence. The rising rates of diabetes, even in rural areas, are largely driven by environmental and lifestyle changes owing to urbanization. The increasing burden of T2D and its complications necessitates effective preventive programmes [46]. The Universal Statement [47] and AHA/ACC/HFSA Guidelines [24], suggest that an elevation in either of the natriuretic peptide (NP) biomarkers can help confirm a diagnosis of Stage B HF in individuals at high risk for the condition. Additionally, BNP and NT-proBNP act as makers of cardiac myocyte stretch, volume overload and wall stress, and they are strongly linked to the mechanisms underlying HF [40]. NP concentrations are typically elevated in patients with HFrEF and may also be elevated in those with HFmrEF [48]. A study conducted by Salah et al. [48] which looked at individuals who had both echocardiography and NT‐pro‐ BNP levels, reported that the median NT‐pro‐ BNP levels significantly increasing from 4436 pg/mL in the group with HFpEF to 7173 pg/mL in the HFrEF group. In our study also, we report that among individuals with T2D median NT‐pro‐ BNP levels were significantly higher (4669 pg/mL) in the HFrEF group compared with 162 pg/mL in the HFpEF group. Considering the high prevalence of underdiagnosed HF in individuals with T2D, elevated NT-proBNP levels may aid in early diagnosis and help identify individuals at greater risk of developing HF.

Our findings indicate that women with T2D and HF exhibit significantly elevated levels of NT-proBNP compared with their male counterparts. This points to need for screening for HF in both men and women with T2D. It appears that T2D removes the protection that women generally have from heart disease. Although LVIDd and LVIDs remain within normal ranges, the elevated BNP levels classify this group under preserved ejection fraction, suggesting a distinct heart failure phenotype despite normal ventricular dimensions. A growing body of evidence indicates that measuring plasma BNP and NT-proBNP levels is a valuable complement to echocardiography, in evaluating individuals suspected of HF [49–52]. Bay et al. [53] demonstrated that a NT pro BNP level > 357 pmol/L effectively identified patients experiencing acute LVF with an ejection fraction of less than 40%, achieving a sensitivity and a specificity of 73% and 82%, respectively. Another study which assessed the relationship between ventricular dysfunction and BNP levels among individuals with diabetes, whether or not they had a history of congestive HF, reported that a BNP level of 75 pg/mL had a sensitivity and specificity of 85% and 97%, respectively, with 90% accuracy for predicting LV dysfunction [50]. The PRIDE study which used age-adjusted cut-points for NT-proBNP reported a cut-point of 300 pg/mL for the diagnosis of acute HF in individuals with diabetes with a sensitivity and specificity of 92% and 90%, respectively [54]. In our study, we report that a NT-proBNP cut-off level of 398 pg/mL is useful for differentiating Indian individuals with T2D, with and without HF.

Our study has several limitations. Firstly, as this was a retrospective study, limited data was available on other confounding factors that could have influenced the findings. Secondly, the findings may not be generalizable to the broader population owing to potential referral bias among individuals with T2D seeking care at a diabetes centre (hospital-based study design). Thirdly, repeated NT-proBNP and echocardiography measurements were not available. Fourthly, this study did not incorporate the H2FPEF score as an alternative for diagnosing heart failure with preserved ejection fraction (HFpEF) owing to the unavailability of some parameters, such as atrial fibrillation status and estimated pulmonary artery systolic pressure (PASP ≥ 35 mmHg). Fifthly, we excluded patients with acute HF. Future studies could validate the H2FPEF score in the Indian population to improve the precision of HFpEF diagnosis. Fifthly, the absence of high-sensitivity troponin T (hs-TnT) measurements is another limitation, as it could provide valuable insights into cardiac stress and myocardial injury, particularly in asymptomatic diabetic individuals where its utility is well established. Although we attempted to derive age-specific NT-proBNP cut-off values, the limited sample size within certain age subgroups restricted our ability to establish thresholds for each age category. Additionally, while the exclusion of individuals with an eGFR < 30 mL/min/1.73 m² reduced confounding from severe renal dysfunction, it limits the generalizability of our findings to individuals with advanced chronic kidney disease, who are at higher risk for heart failure.

Conclusions

In summary, this real-world data from a diabetes centre demonstrates a substantial burden of LV dysfunction and HF in individuals with T2D in India.

NT-proBNP, with a cutoff of ≥ 398 pg/mL, showed the best correlation with reduced EF in this study population, highlighting its potential as a key biomarker for early detection. It is crucial to develop strategies for the early detection of LV dysfunction in individuals with T2D to prevent and effectively manage HF. Additionally, these findings may advocate for the routine monitoring of echo and NT-proBNP at diabetes centres to detect early stages of HF in individuals with T2D to improve the prognosis.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Conceptualization: Rajendra Pradeepa, Ranjit Mohan Anjana and Viswanathan Mohan; methodology: Rajendra Pradeepa and Thyparambil Aravindakshan PramodKumar; formal analysis and investigation: Thyparambil Aravindakshan PramodKumar, Rajendra Pradeepa and Abdul Subhan Naziyagulnaaz; writing—original draft preparation: Rajendra Pradeepa and Thyparambil Aravindakshan Pramod Kumar; writing—review and editing: Viswanathan Mohan, Ranjit Mohan Anjana, Saravanan Jebarani, Abdul Subhan Naziyagulnaaz, Sadasivam Ganesan, Natarajan Premananth, Abraham Oomman, Soumitra Kumar, P.B. Jayagopal, Gurpreet Singh Wander, Ajit Mullasari, Jagat Narula, Sanjay Jain and Onkar C Swami; funding acquisition: Viswanathan Mohan; supervision: Ranjit Mohan Anjana and Viswanathan Mohan.

Funding

The work was commissioned and supported by Alembic Pharmaceuticals Limited (#12026725) including Rapid Service Fee in order to provide data on heart failure in diabetes. However, the funders had no role in collection and analysis of the data.

Data Availability

The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.

Declarations

Conflict of Interest

Rajendra Pradeepa, Thyparambil Aravindakshan PramodKumar, Ranjit Mohan Anjana, Saravanan Jebarani, Abdul Subhan Naziyagulnaaz, Sadasivam Ganesan, Natarajan Premananth, Abraham Oomman, Soumitra Kumar, P.B.Jayagopal, Gurpreet Singh Wander, Ajit Mullasari, Jagat Narula and Viswanathan Mohan do not have any conflicts of interest. Sanjay Jain and Onkar Swami are employees of Alembic Pharmaceuticals Limited.

Ethics Approval

The study was approved by the institutional ethics committee of Madras Diabetes Research Foundation (approval no. MDRF/NCT/01/02/2024). Written informed consent was obtained from all participants to the use of their anonymized data for research purposes.