Heart failure phenotypes in people with type 1 diabetes

Type 1 diabetes is an increasingly-recognized risk factor for incident heart failure (HF), with epidemiological studies reporting a relative risk ranging anywhere from 2.681 to 4.322 versus controls.1-4 However, data defining the prevalence of HF phenotypes and associated demographic characteristics in people with type 1 diabetes are sparse and, to our knowledge, no such data have been reported from the United States. Herein, we addressed this gap by investigating the Premier Healthcare Database (PHD) to determine HF classifications in this population. The PHD is a US hospital-based, service-level, all-payer database that includes more than 1030 healthcare institutions with >93 million outpatient encounters annually. People with type 1 diabetes with and without HF were identified using ICD-10 codes entered during any outpatient encounter that occurred between 1 January 2022 and 31 December 2023. Of note, demographic and biochemical data were only obtained from the patient's first encounter in the database during the specified timeframe and data from all subsequent encounters were excluded. With this, we ensured that each patient was only included in the analysis once. Specific ICD-10 codes for type 1 diabetes included: E10.9, E10.65, E10.10, E10.69, E10.11, E10.641, E10.29, E10.21, E10.36, E10.319, E10.39, E10.311, E10.40, E10.51, E10.649, E10.621, E10.628, E10.630, E10.638, E10.620, E10.622, E10.618, E10.8, E10.22, E10.3211, E10.3212, E10.3213, E10.3213, E10.3219, E10.3291, E10.3292, E10.3293, E10.3299, E10.3311, E10.3312, E10.3313, E10.3319, E10.3391, E10.3392, E10.3393, E10.3399, E10.3411, E10.3412, E10.3413, E10.3419, E10.3491, E10.3492, E10.3493, E10.3499, E10.3511, E10.3512, E10.3513, E10.3519, E10.3521, E10.3522, E10.3523, E10.3529, E10.3531, E10.3532, E10.3533, E10.3539, E10.3541, E10.3542, E10.3543, E10.3551, E10.3552, E10.3553, E10.3559, E10.3591, E10.3592, E10.3593, E10.3599, E10.3599, E10.37X1, E10.37X2, E10.37X3, E10.37X9, E10.37X1, E10.41, E10.42, E10.43, E10.44, E10.49, E10.52, E10.59, and E10.610. Specific ICD-10 codes for HF included: I50.1, I50.20, I50.21, I50.22, I50.23, I50.30, I50.31, I50.32, I50.33, I50.40, I50.41, I50.42, I50.43, I50.810, I50.811, I50.812, I50.813, I50.82, I50.83, I50.84, I50.89, and I50.9. The University of Virginia Institutional Review Board granted an exemption for this study given that all data were de-identified and HIPAA-compliant. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines,5 and all statistical analyses for this observational study were performed with SAS version 9.4 (SAS; Cary, NC). We identified 90 647 unique people with type 1 diabetes: 53.4% identified as female, 75.2% identified as White, 16.7% identified as Black, 5.3% were of other or unknown race, and 0.7% identified as Asian. Regarding baseline complication rates of these 90 647 people, 11 067 (12.2%) had ICD-10 codes for type 1 diabetes with chronic kidney disease, 8205 (9.05%) had ICD-10 codes for type 1 diabetes with neuropathy, 5459 (6.02%) had ICD-10 codes for type 1 diabetes with hypoglycaemia, 2229 (2.46%) had ICD-10 codes for type 1 diabetes with retinopathy, 1443 (1.59%) had ICD-10 codes for type 1 diabetes with angiopathy and/or other circulatory complications, 980 (1.08%) had ICD-10 codes for type 1 diabetes with nephropathy, and 448 (0.49%) had ICD-10 codes for type 1 diabetes with other kidney complications. Within this total cohort, 8836 (9.7%) people also had HF and were further stratified by HF type: 2868 (3.2% of total cohort; 32.5% of HF subgroup) had HF with reduced ejection fraction (HFrEF), 2801 (3.1% of total cohort; 31.7% of HF subgroup) had HF with preserved ejection fraction (HFpEF), 2318 had unspecified HF (2.6% of total cohort; 26.2% of HF subgroup), 796 had combined systolic and diastolic HF (0.9% of total cohort; 9.0% of HF subgroup), 31 had right HF (<0.1% of total cohort; 0.4% of HF subgroup), 12 had biventricular HF (<0.1% of total cohort; 0.1% of HF subgroup), 9 had end-stage HF (<0.1 of total cohort; 0.1% of HF subgroup), and 1 had high output HF (<0.1% of total cohort; 0.01% of HF subgroup). We then focused our investigation on demographic and biochemical differences between those who developed HFrEF versus HFpEF (Table 1). In general, those with HFrEF were older and more likely to identify as white and/or male. Conversely, those with HFpEF were younger and more likely to identify as Black or African-American and/or female. Anthropomorphic results showed that those with HFpEF had greater BMI. Biochemical results demonstrated that those with HFpEF had significantly higher creatinine, lower eGFR, and lower HbA1c compared to those with HFrEF. As expected, median values of NT-proBNP and BNP (natriuretic peptides that serve as biomarkers of HF) were elevated in both groups, though those with HFrEF had higher BNP and lower NT-proBNP levels (Table 1). Importantly, numerous factors (e.g. sex, age, obesity, etc.) can impact natriuretic peptide levels6-8 and there were significant differences in some of these variables between the HFpEF and HFrEF cohorts in the current study. This should be kept in mind when interpreting our results. Few studies have attempted to clarify the prevalence of HF types in people with type 1 diabetes. A 7-year prospective cohort study in Poland followed 1685 people with type 1 diabetes and utilized echocardiography to report that the prevalence of HF was 3.7%.9 Of those who developed HF, 85% had HFpEF and 15% had HFrEF.9 Conversely, we identified a slight predominance of HFrEF in the current study. This raises the question of whether the prevalence of HF phenotypes in people with type 1 diabetes may differ by nation and/or geographic region. We also identified demographic and biochemical differences between HF phenotypes in the current study. For example, people with HFpEF were younger, more likely to identify as Black or African-American, and more likely to identify as female. Similar differences have been identified in other populations10; however, to our knowledge, we are the first to investigate such differences in people with type 1 diabetes. Several potential limitations of the current study warrant consideration. First, we recognize that defining type 1 diabetes and HF by ICD-10 codes is an inherent limitation of our work. Second, we did not have access to clinical documentation or echocardiogram results and thus could not confirm the accuracy of ICD-10 HF codes in our study population. However, we note that a recent study by Bates et al.11 reported that ICD-10 codes reliably identified HFrEF and HFpEF when an EF of 50% was the cutoff. Third, we did not have access to medication lists and this prevented us from determining whether certain medications that lower natriuretic peptide levels (e.g. diuretics) were administered at a higher frequency in certain HF populations. Finally, there is no ICD-10 code for HF with mildly reduced EF and thus this HF type was unidentifiable in the current study. This study is the first to quantify specific phenotypes of people with type 1 diabetes who developed HF within the United States. We identified several demographic and biochemical disparities between those who developed HFpEF versus HFrEF; thus, future work could investigate whether type 1 diabetes disproportionately impacts the relationship between these characteristics and HF risk when compared to other high-risk populations (e.g. people with type 2 diabetes). APE wrote the first draft of the manuscript. JTP researched data and reviewed/edited the manuscript. WBH researched data and reviewed/edited the manuscript. All authors approved the final version of the manuscript. JTP is the guarantor of the work and, as such, had full access to all data in the study and takes responsibility for the integrity of the data and accuracy of the data analysis. The authors thank Dr. Kaitlin Love for her helpful review and subsequent invaluable input on stylistic writing of this manuscript. WBH was supported by a career development award from the American Heart Association (Grant Number: 941481) and a strategic research agreement with Breakthrough T1D (3-SRA-2023-1236-M-B). The authors have no conflicts of interest to disclose. The peer review history for this article is available at https://www.webofscience.com/api/gateway/wos/peer-review/10.1111/dom.16133. The data that support the findings of this study are available from Premier Inc. but restrictions apply to the availability of these data, which were used under licence for the current study and therefore are not publicly available. Data are however available from the authors upon reasonable request and with permission of Premier Inc.