Prevalence and Predictors of Subclinical Cardiomyopathy in Patients With Type 2 Diabetes in a Health System

Abstract

Introduction:

Diabetic cardiomyopathy (DbCM) is characterized by subclinical abnormalities in cardiac structure/function and is associated with a higher risk of overt heart failure (HF). However, there are limited data on optimal strategies to identify individuals with DbCM in contemporary health systems. The aim of this study was to evaluate the prevalence of DbCM in a health system using existing data from the electronic health record (EHR).

Methods:

Adult patients with type 2 diabetes mellitus free of cardiovascular disease (CVD) with available data on HF risk in a single-center EHR were included. The presence of DbCM was defined using different definitions: (1) least restrictive: ≥1 echocardiographic abnormality (left atrial enlargement, left ventricle hypertrophy, diastolic dysfunction); (2) intermediate restrictive: ≥2 echocardiographic abnormalities; (3) most restrictive: 3 echocardiographic abnormalities. DbCM prevalence was compared across age, sex, race, and ethnicity-based subgroups, with differences assessed using the chi-squared test. Adjusted logistic regression models were constructed to evaluate significant predictors of DbCM.

Results:

Among 1921 individuals with type 2 diabetes mellitus, the prevalence of DbCM in the overall cohort was 8.7% and 64.4% in the most and least restrictive definitions, respectively. Across all definitions, older age and Hispanic ethnicity were associated with a higher proportion of DbCM. Females had a higher prevalence than males only in the most restrictive definition. In multivariable-adjusted logistic regression, higher systolic blood pressure, higher creatinine, and longer QRS duration were associated with a higher risk of DbCM across all definitions.

Conclusions:

In this single-center, EHR cohort, the prevalence of DbCM varies from 9% to 64%, with a higher prevalence with older age and Hispanic ethnicity.

Introduction

Type 2 diabetes mellitus is increasing in prevalence and is associated with an increased risk of heart failure (HF), including heart failure with preserved ejection fraction (HFpEF).^1–3 However, the progression from an at-risk phenotype to overt HF does not occur in a single step. The intermediate stage, termed diabetic cardiomyopathy (DbCM), is characterized by abnormalities in cardiac structure and/or function, despite normal ejection fraction and an absence of overt cardiovascular disease (CVD). Although the prevalence of DbCM in pooled community cohort studies ranges from 11% to 67% depending on the definition, ⁴ the prevalence in a contemporary, real-world cohort is not well characterized. Accordingly, we aimed to evaluate the prevalence of DbCM among individuals with diabetes free of CVD in a real-world electronic health record (EHR).

Methods

Deidentified patient-level data were obtained from a single-center EHR from the University of Texas Southwestern Medical Center. Patients with an active problem of type 2 diabetes mellitus who were registered in the institution’s diabetes registry as of December 31, 2014 were included. ⁵ Patients with a diagnosis of ischemic heart disease (myocardial infarction or coronary revascularization) on the problem list and patients who were also included in the institution’s HF registry, based on having an active problem of HF listed or having a previous clinical encounter for HF, were excluded.

Demographic data (age, sex, race, and ethnicity) were self-reported at the time of patient registration. Vital signs (systolic and diastolic blood pressure, height, and weight) were entered by the clinical care team and recorded in the patient flow sheet. Medical history was obtained from the patient’s problem list. Laboratory data (serum creatinine and hemoglobin A1c [HbA_1c]) were obtained as part of the clinical workup through standardized assays. Clinical echocardiographic data were also obtained as part of clinically indicated assessment by trained sonographers, with measurements confirmed by the attending cardiologist in accordance with the American Society of Echocardiography (ASE) guidelines. ⁶ Left ventricular (LV) mass index (LVMi) was calculated using the Devereux formula. ⁷ Established cutoffs were used to define echocardiographic abnormalities as previously described: ⁴ (1) LV hypertrophy: LVMi ≥ 115 g/m² in males or ≥ 95 g/m² in women; (2) left atrial (LA) volume index (LAVi) ≥ 34 mL/m²; or (3) presence of diastolic dysfunction as defined by the ASE criteria. ⁸

The DbCM phenotype was defined using 3 separate definitions: (1) least restrictive: ≥ 1 echocardiographic abnormality (LVH, LA enlargement, or diastolic dysfunction); (2) intermediate restrictive: ≥ 2 echocardiographic abnormalities; (3) most restrictive: ≥ 3 echocardiographic abnormalities. The prevalence of DbCM was compared across age, sex, race, and ethnicity-based subgroups with differences assessed using the χ² test. Adjusted logistic regression models were constructed to evaluate significant predictors of DbCM. Analyses were performed using R 4.0.3 (R Foundation) with a 2-sided P < .05 indicating significance. The present analysis was considered exempt from the Institutional Review Board (IRB) approval at UT Southwestern Medical Center.

Results

Among 4752 patients, 2831 were excluded because of the presence of ischemic heart disease (n = 303) or HF (n = 651), ejection fraction < 45% (n = 351), or missing data (n = 1526). The final 1921 patients had a mean age of 65 ± 7 years, 53% females, 25% Black race, and 17% Hispanic or Latino ethnicity. LA enlargement was observed among 806 patients, diastolic dysfunction among 742 patients, and LVH among 425 patients (Table 1). The prevalence of DbCM was 64.4% in the least restrictive, 29.6% in the intermediate restrictive, and 8.7% in the most restrictive definitions. With the least restrictive definition, the prevalence of DbCM significantly increased from 54.4% in adults < 60 years to 74.6% in adults ≥ 75 years (P < .001) (Figure 1a). Similar results were observed for the intermediate restrictive (<60 years: 23.2%, ≥75 years: 35.2%; P < .001) and most restrictive definitions (<60 years: 6.7%, ≥75 years: 11.5%; P = .03) (Figure 1a). Differences in DbCM prevalence by sex were only observed in the most restrictive definition, with females having a higher prevalence than males (11.2% vs 5.9%; P < .001) (Figure 1b). No significant differences were observed among Black vs non-Black individuals across all definitions (Figure 1c). Finally, Hispanic (vs non-Hispanic) individuals had a higher prevalence of DbCM across all definitions (Figure 1d).

Table 1.

Distribution of Echocardiographic Abnormalities Among the Study Population.

	LV hypertrophy	LA enlargement	Diastolic dysfunction
N	425	806	742
Least restrictive	425 (100.0)	806 (100.0)	742 (100.0)
Intermediate restrictive	342 (80.5)	510 (63.3)	453 (61.1)
Most restrictive	167 (39.3)	167 (20.7)	167 (22.5)

LV hypertrophy was defined as an LV mass index ≥ 115 g/m² in men or ≥ 95 g/m² in females; LA enlargement as a left atrial volume index (LAVi) ≥ 34 mL/m²; and diastolic dysfunction was defined by the ASE criteria. ⁸

Least restrictive: ≥ 1 echocardiographic abnormality (LV hypertrophy, LA enlargement, or diastolic dysfunction); intermediate restrictive: ≥ 2 echocardiographic abnormalities; and most restrictive: ≥ 3 echocardiographic abnormalities.

Abbreviations: LA, left atrial; LV, left ventricular.

Figure 1.

The proportion of individuals with DbCM across all definitions in (a to d) age-, sex-, race-, and ethnicity-based subgroups. *P < .05.

In multivariable-adjusted logistic regression, older age, higher systolic blood pressure, higher creatinine, and longer QRS duration on electrocardiogram were associated with a higher risk of DbCM in the least restrictive definition (Table 2). With the most restrictive definition, age was no longer associated with DbCM. Higher diastolic blood pressure and higher HbA_1c were associated with a lower risk of DbCM based on the most restrictive definition. Body mass index (BMI) and high-density lipoprotein cholesterol (HDL-c) were not associated with DbCM across all definitions.

Table 2.

Adjusted Association Between Baseline Risk Factors and Odds of DbCM Across All Definitions.

	Least restrictive		Intermediate restrictive		Most restrictive
	Odds ratio [95% CI]	P	Odds ratio [95% CI]	P	Odds ratio [95% CI]	P
Age, y	1.14 [1.02, 1.27]	.02	1.02 [0.91, 1.16]	.70	0.96 [0.79, 1.19]	.72
BMI, kg/m2	1.04 [0.95, 1.14]	.42	0.96 [0.87, 1.06]	.45	1.12 [0.95, 1.30]	.18
SBP, mm Hg	1.19 [1.10, 1.28]	<.001	1.20 [1.11, 1.29]	<.001	1.29 [1.14, 1.45]	<.001
DBP, mm Hg	0.76 [0.67, 0.86]	<.001	0.72 [0.64, 0.82]	<.001	0.65 [0.53, 0.79]	<.001
HbA1c, %	0.99 [0.91, 1.07]	.74	0.92 [0.84, 1.01]	.10	0.83 [0.69, 0.98]	.04
Creatinine, mg/dL	1.23 [1.13, 1.35]	<.001	1.25 [1.17, 1.34]	<.001	1.18 [1.10, 1.26]	<.001
HDL-c, mg/dL	0.98 [0.94, 1.02]	.42	0.97 [0.92, 1.01]	.12	0.94 [0.86, 1.01]	.10
QRS, ms	1.13 [1.06, 1.21]	<.001	1.14 [1.07, 1.21]	<.001	1.19 [1.09, 1.29]	<.001

Age is per 10 years, BMI is per 5 kg/m², SBP and DBP are per 10 mm Hg, HbA_1c is per 1%, creatinine is per 1 mg/dL, HDL-c is per 5 mg/dL, and QRS duration is per 10 ms increase.

Abbreviations: BMI, body mass index; DBP, diastolic blood pressure; HbA_1c, hemoglobin A1c; HDL-c, high-density lipoprotein cholesterol; SBP, systolic blood pressure.

Discussion

In this analysis of over 1900 patients with type 2 diabetes mellitus from a single-center, EHR-based cohort, we observed several key findings. First, the prevalence of DbCM varied from 64% when only 1 echocardiographic abnormality was required for inclusion to nearly 9% with the use of a restrictive definition requiring 3 abnormal echocardiographic findings. Second, irrespective of the definition, the prevalence of DbCM increased with older age and Hispanic ethnicity. Finally, across all definitions, significant predictors of DbCM in multivariable logistic regression models included higher systolic blood pressure, higher creatinine, and longer QRS duration. Taken together, these findings provide support for the DbCM phenotype being a distinct clinical entity in patients with type 2 diabetes.

Type 2 diabetes is a well-recognized risk factor for HF, and patients with diabetes are twice as likely to develop incident HF compared with the general population. ⁹ Notably, this increased risk persists despite control of traditional cardiovascular risk factors. ³ Type 2 diabetes is associated with an increased risk of structural heart disease, and echocardiographic changes can be observed even in the absence of ischemia, hypertension, and obesity.^4,10 Several epidemiological and clinical studies have identified echocardiographic changes in LV structure and function in patients with type 2 diabetes. ¹¹ In this study of patients with diabetes in a health system, LA enlargement and diastolic dysfunction were more prevalent echocardiographic abnormalities followed by LVH, and this is similar to the pattern of echocardiographic abnormalities observed among community-based cohorts.^12,13 However, the magnitude of diastolic dysfunction is greater among patients with diabetes than among the general adult population, ¹⁴ and findings from our study add to the existing literature that DbCM is a unique, intermediate-stage clinical phenotype detected before the development of overt HF.

The pathophysiologic basis underpinning the increased risk of HF among patients with diabetes centers around adverse cardiac remodeling due to hyperglycemia and insulin resistance.^15,16 Hyperglycemia generates reactive oxygen species, leading to DNA damage and the formation of advanced glycation end products, which initiate inflammation. ¹⁷ Hyperglycemia also activates the renin-angiotensin-aldosterone system, leading to microvascular dysfunction and coronary small vessel arteriosclerosis. ¹⁸ Insulin resistance is associated with increased myocardial free fatty acid uptake, which leads to increased oxygen demand, oxidative stress, and lipotoxicity. ¹⁰ Parasympathetic denervation from autonomic neuropathy can lead to a dominant sympathetic tone, further increasing myocardial oxygen demand and contributing to diastolic dysfunction. ¹⁹ Finally, activation of sodium-hydrogen exchangers within the heart and in the kidneys can lead to myocardial microvascular dysfunction and hypertrophy, hyperfiltration, and sodium retention in the kidneys. ²⁰ These pathophysiologic changes ultimately result in cardiac fibrosis, hypertrophy, diastolic dysfunction, and eventually systolic dysfunction.

Clinical Implications

Dedicated efforts are needed to promote systematic screening for DbCM. This study highlights that pre-existing echocardiogram data can be routinely identified within the EHR. In addition, cardiovascular biomarkers of heart disease, including natriuretic peptides or high-sensitivity cardiac troponin, can be used to identify patients with type 2 diabetes at risk of clinical HF, and multiple guidelines have now recommended biomarker testing in patients with type 2 diabetes mellitus.^21,22 However, from a population health perspective, identification of prevalent DbCM with a primarily biomarker-based approach may suffer from data missingness as these guidelines are relatively new. A pragmatic strategy may include initial screening with an HF clinical risk score such as WATCH-DM or Thrombolysis in Myocardial Infarction Risk Score for Heart Failure in Diabetes (TRS-HF_DM), with further testing with cardiovascular biomarkers or echocardiography on the basis of the patient risk profile for additional clinical information. ²³

Identification of DbCM may allow for early intervention with preventive therapies. Lifestyle interventions—including increasing cardiorespiratory fitness, and combating comorbid obesity—are associated with lower CVD events. ²⁴ Optimizing medical therapy for type 2 diabetes with sodium/glucose cotransporter-2 inhibitor (SGLT2i) therapy, which has been shown to reduce first hospitalization for HF, may also help mitigate HF risk. ²⁵ More recently, treatment with the mineralocorticoid receptor antagonist finerenone has been shown to reduce the risk of HF among patients with type 2 diabetes and chronic kidney disease. ²⁶ Finally, 2 ongoing studies investigating the effect of an aldose reductase inhibitor (ARISE-HF, NCT04083339) on exercise capacity as measured by peak VO₂ among patients with DbCM and investigating N-terminal prohormone of brain natriuretic peptide (NT-proBNP)-guided intensification of medical therapy for prevention of cardiovascular events in a high-risk population with type 2 diabetes (ADOPT, NCT04286399) will help identify interventions to modify the natural history of progression from DbCM to overt HF.

Our study has several strengths, including a detailed echocardiographic assessment on nearly 2000 patients in a single-center EHR. Notably, our study was limited by the lack of available biomarkers in the EHR cohort. As such, the definitions of cardiomyopathy could only be assessed based on echocardiographic parameters. ⁴ In addition, outcomes following development of HF were not recorded in the registry, and HF prognosis could not be assessed. Finally, our results do not establish a causal effect between type 2 diabetes mellitus and cardiomyopathy.