OBJECTIVE

To explore risks and associated mediation effects of developing chronic kidney disease (CKD) and heart failure (HF) in young- and usual-onset type 2 diabetes (T2D) between White Americans (WAs) and African Americans (AAs).

RESEARCH DESIGN AND METHODS

From U.S. medical records, 1,491,672 WAs and 31,133 AAs were identified and stratified by T2D age of onset (18–39, 40–49, 50–59, 60–70 years). Risks, mediation effects, and time to CKD and HF were evaluated, adjusting for time-varying confounders.

RESULTS

In the 18–39, 40–49, 50–59, 60–70 age-groups, the hazard ratios (of developing CKD and HF in AAs versus WAs were 1.21 (95% CI 1.17–1.26) and 2.21 (1.98–2.45), 1.25 (1.22–1.28) and 1.86 (1.75–1.97), 1.21 (1.19–1.24) and 1.54 (1.48–1.60), and 1.10 (1.08–1.12) and 1.11 (1.07–1.15), respectively. In AAs and WAs aged 18–39 years, time in years to CKD (8.7 [95% CI 8.2–9.1] and 9.7 [9.2–10.2]) and HF (10.3 [9.3–11.2] and 12.1 [10.6–13.5]) were, on average, 3.6 and 4.0 and 3.1 and 4.1 years longer compared with those diagnosed at age 60–70 years. Compared with females, AA males aged <60 years had an 11–49% higher CKD risk, while WA males aged <40 years had a 23% higher and those aged ≥50 years a 7–14% lower CKD risk, respectively. The mediation effects of CKD on the HF risk difference between ethnicities across age-groups (range 54–91%) were higher compared with those of HF on CKD risk difference between ethnicities across age-groups (13–39%).

CONCLUSIONS

Developing cardiorenal complications within an average of 10 years of young-onset T2DM and high mediation effects of CKD on HF call for revisiting guidelines on early diagnosis and proactive treatment strategies for effective management of cardiometabolic risk.

In the U.S., ∼13% of adults have type 2 diabetes (T2D) and ∼15% of adults have chronic kidney disease (CKD), and the prevalence of heart failure (HF) is projected to increase by 46% from 2.4% in 2012 to 3% by 2030 (1,2). Nearly 30% of people with T2D in the U.S. were found to have HF, four times higher than the general population (3). Approximately 40–50% of people with HF have CKD, with the severity of renal dysfunction being associated with a graded increased risk of mortality and morbidity (4). The 2013–2016 U.S. registry study in people with established T2D reported that 51% have three or more cardiorenal metabolic conditions, with hypertension (83%), hyperlipidemia (81%), coronary artery disease (32%), and CKD (20%) being the most prevalent (5).

In people with T2D, the presence of HF and/or CKD is associated with increased risks of both death and further cardiovascular disease (CVD) risk. The most common causes of death in people with T2D and CKD are atherosclerotic CVD (ASCVD) and HF (6). Among people with T2D, those with CKD have significantly shorter times to modified major adverse cardiovascular events, HF, and all-cause mortality than those without CKD (7). A recent multinational cohort study confirmed a vicious cycle between failing cardiorenal organs in people with T2D: HF manifestation before and after the T2D diagnosis was associated with a 2.3-fold increased risk of incident CKD, while CKD manifestation before T2D diagnosis was associated with a twofold increased risk of HF (6). Complex multidirectional associations among T2D, CKD, and HF challenge treatment strategies and result in a significant loss of years of life (4). The lifetime risk associated with CKD and HF is likely to be higher in people with young-onset T2D versus usual-onset T2D.

Recent epidemiological evidence shows different patterns of increasing trends in young-onset T2D in different ethnic groups and by sex, with increasing prevalence of cardiometabolic multimorbidity and high-risk factors at T2D onset (6,810). African Americans (AAs) are at a higher risk of developing T2D and its complications at an earlier age compared with White Americans (WAs), while hypertension and kidney disease are also disproportionately prevalent among AAs, with significant sex differences (2,1013). However, there is a lack of evidence-based guidance on the pathway of the development of end-stage renal disease and CKD in various age-groups and its management in AAs, particularly those with T2D (11,14). While studies have discussed the increased risk of CKD in people with HF and T2D, we are not aware of any study evaluating the mediation effects of HF on CKD and vice versa in young- and usual-onset T2D in different ethnicities (6). In addition, while sex differences in the prevalence or incidence of CKD and HF in various ethnic groups have been reported in U.S. national surveys, we are not aware of any nationally representative population-level data from the U.S. on the prevalence or incidence of CKD and/or HF in incident T2D in different ethnic groups by age and sex at diagnosis, particularly how the cardiorenal risk dynamics work in WA and AA young-onset T2D (2,12,1517).

Using large, nationally representative electronic medical records (EMRs) from the U.S., the aims of this study were to estimate, by age-group at T2D onset in AAs and WAs, the population-level 1) prevalence, incidence, and risk of developing CKD and/or HF, 2) time to developing these conditions, and 3) mediation effect of CKD on the risk of developing HF and vice versa.

Data

The Centricity EMR (CEMR) database incorporates deidentified patient-level data from >40,000 care providers from all U.S. states. The similarity of the general population characteristics and cardiometabolic risk factors in the CEMR database to those reported in U.S. national health surveys has been reported (18) and has been used extensively for academic research (10,18,19). Longitudinal EMRs were available for >46 million individuals from 1995 until September 2018, with comprehensive patient-level information on demographics, anthropometric measures, disease events, medications, and clinical and laboratory measures (19).

Study Design and Variables

Study cohort identification criteria were 1) data available on age and sex, 2) T2D diagnosis from 1 January 2000 to 31 September 2018, 3) age 18–70 years at T2D diagnosis (index date, baseline), and 4) known WA or AA ethnicity. The clinically driven machine learning–based algorithms to identify patients with T2D and extraction of longitudinal prescription and risk factor data from the CEMR database have been described previously (2022).

Ethnicity in the CEMR is coded according to the U.S. Census Bureau categorization (18,23). HbA1c measures at baseline were obtained as the nearest measure within 3 months of either side of the diagnosis. Baseline body weight, systolic blood pressure, blood lipid levels, and estimated glomerular filtration rate (eGFR) were calculated as the average of available measures within 3 months of diagnosis. The CEMR provides an automated MDRD equation–based eGFR. In the absence of this estimate, the MDRD eGFR was determined using available background data otherwise declared missing. A detailed account of glucose-lowering, antihypertensive, lipid-lowering, and other drug use in the CEMR have been previously reported (10,18,19,21).

The presence of comorbidities was assessed by relevant disease identification codes (ICD-9, ICD-10, or SNOMED CT). ASCVD was defined by the presence of available diagnostic codes for ischemic heart disease (myocardial infarction, unstable angina, or coronary revascularization, excluding stable angina), cerebrovascular disease (ischemic/hemorrhagic stroke, transient ischemic attack, or carotid revascularization), or peripheral arterial/vascular disease. Patients with CVD included those with ASCVD and HF. The CKD definition included disease identification codes and the following laboratory measures: CKD stages 2–5, end-stage renal disease, dialysis, transplant, nephropathy, proteinuria, albuminuria, nephrotic syndrome, nephritis, eGFR <60 mL/min/1.73 m2, or urine albumin-to-creatinine ratio >30 mg/g (two elevated measures within a 12-month window). Microvascular disease was defined by a diagnostic code for neuropathy, retinopathy, or CKD (as defined previously). Hypertension and dyslipidemia were defined by the presence of a clinical diagnosis or use of antihypertensive and lipid-lowering drugs before diabetes diagnosis. Antihypertensive therapy was identified using the Anatomical Therapeutic Classification system and included diuretics, peripheral vasodilators (excluding nicotinic acid), β-blockers, calcium channel blockers, and agents acting on the renin-angiotensin system. Lipid-lowering drugs included statins, bile acid sequestrants, fibrates, nicotinic acid, proprotein convertase subtilisin/kexin type 9 inhibitors, and potent (≥1 g) forms of n-3, fish, or krill oil. Depression was defined using a clinically guided machine learning algorithm (24), where the definition included a diagnostic code or at least two prescriptions for antidepressants that are commonly prescribed for depression (10,25). Cancer was defined as any malignant neoplasm, excluding melanoma (10,19). A disease was considered prevalent if its first available diagnostic date was on or before the T2D diagnosis.

Statistical Methods

Baseline characteristics were summarized separately by age-groups for AAs and WAs as number and percent, mean with SD, or median with first and third quartiles, as appropriate. Age-groups at T2D diagnosis were 18–39, 40–49, 50–59, and 60–70 years.

By ethnicity and age-groups, among those without existing CKD and/or HF at baseline, incidence rates with 95% CIs per 1,000 person-years were calculated, and the risk of CKD and/or HF was evaluated using flexible parametric survival models with time from index date to the event or censoring (31 September 2018) (26). Models were adjusted for age, sex, BMI, and smoking status and accounted for the use of insulin, retinopathy or neuropathy, hypertension, dyslipidemia, depression, and ASCVD before the end of follow-up as time-dependent confounders. The use of noninsulin antidiabetic drugs as a potential confounder was excluded from the model on the basis of statistical information criteria. A separate assessment was conducted by adding individual use of glucagon-like peptide 1 receptor agonists (GLP-1RAs) and sodium–glucose cotransporter 2 inhibitors (SGLT-2is) before the event as time-varying covariates. The multivariable model allowed the estimation of hazard ratios (HRs) and restricted mean survival time (RMT) to event. As sensitivity analyses, a treatment effects modeling approach was used to estimate time to events, with balancing of and adjusting for the covariates as described above (9,10). Based on the multivariable flexible parametric survival modeling approach, the mediation effects of CKD on HF and vice versa were evaluated using natural effect model techniques (27), where the existing CKD/HF before developing incident HF/CKD variable was used as a mediator.

Statistical significance level was set at 5%, with all CIs set at 95%. This study has been conducted following Reporting of Studies Conducted Using Observational Routinely Collected Data (RECORD) guidelines.

Data Provider and License

The U.S. General Electric CEMR used for this study was licensed through IQVIA Inc. for academic research, with S.K.P. as the licensee principal investigator. This study involved the use of patient-level EMRs where the subjects could not be identified directly or through identifiers linked to the subjects. Thus, according to U.S. Department of Health and Human Services Exemption 4 (CFR 46.101[b][4]), this study is exempt from ethics approval from an institutional review board and from informed consent.

Data Availability

Aggregated data are available upon request. Complete data cannot be shared because of license restrictions.

Patient characteristics at T2D diagnosis along with missing data in relevant risk factors are presented in Table 1. In the cohorts of 311,333 AAs and 1,491,672 WAs, 39,219 (13%) and 131,288 (9%), respectively, were diagnosed with T2D at 18–39 years. AAs had significantly higher mean HbA1c at diagnosis across all age-groups compared with WAs (all P < 0.05), while both ethnic groups had the highest HbA1c in the youngest age-group compared with the older age-groups (Table 1). AAs had significantly higher BMI (81% obese) than WAs (76%) in the 18–39 age-group (P = 0.03), while the BMI distributions were similar between the ethnic groups in all age-groups ≥40. Hypertension was present in 57% of AAs and 48% of WAs, with 38% and 29% prevalence in the youngest age-group, respectively (Table 2). Dyslipidemia was present in 32% of AAs and 39% of WAs, with 16% and 21% prevalence in the youngest age-group, respectively. The prevalence of ASCVD in the 18–39, 40–49, 50–59, and 60–70 age-groups was 3, 6, 10, and 15% in AAs and 3, 7, 12, and 19% in WAs, while HF was present in 2, 3, 4, and 5% of AAs and 1, 1, 2, and 3% of WAs, respectively (Table 2). The prevalence of microvascular disease in the 18–39, 40–49, 50–59, 60–70 age-groups was 12, 18, 22, and 26% in AAs and 11, 17, 19, and 23% in WAs, where CKD was present in 5, 9, 12, 17% and 3, 6, 9, and 14%, respectively.

Table 1

Baseline characteristics of the study cohorts at the time of T2D diagnosis by age-group

Age-group, years
18–3940–4950–5960–70
CharacteristicWAAAWAAAWAAAWAAATotal
Patients, n 131,288 39,217 247,622 63,429 480,782 107,371 631,980 101,316 1,803,005 
Follow-up, mean (SD), years 5.1 (4.3) 4.7 (4.1) 5.1 (4.3) 4.7 (4.1) 4.9 (4.2) 4.4 (3.9) 5.0 (4.3) 4.5 (4.0) 4.9 (4.2) 
Male sex, n (%) 50,441 (38) 13,141 (34) 122,982 (50) 26,456 (42) 245,543 (51) 42,998 (40) 324,226 (51) 38,975 (38) 864,762 (48) 
Age, mean (SD), years 33 (5) 32 (6) 45 (3) 45 (3) 55 (3) 55 (3) 65 (3) 64 (3) 55 (11) 
Smoking status, n (%)          
 Ever-smoker 50,028 (38) 12,243 (31) 100,396 (41) 21,210 (33) 198,534 (41) 42,354 (39) 247,785 (39) 37,493 (37) 710,043 (39) 
 Unknown 43,358 (33) 12,901 (33) 79,333 (32) 20,500 (32) 154,194 (32) 33,739 (31) 218,705 (35) 35,466 (35) 598,196 (33) 
HbA1c          
n (% of nonmissing) 60,456 (46) 19,168 (49) 104,708 (42) 27,985 (44) 179,041 (37) 41,896 (39) 206,083 (33) 34,596 (34) 673,933 (37) 
 Mean (SD) % 8.2 (2.3) 8.6 (2.6) 8.1 (2.2) 8.4 (2.5) 7.8 (2.1) 8.1 (2.3) 7.4 (1.7) 7.7 (2.0) 7.8 (2.1) 
 ≥7.5%, n (% of nonmissing) 31,395 (52) 10,381 (54) 50,373 (48) 14,015 (50) 76,458 (43) 18,756 (45) 68,104 (33) 13,062 (38) 282,544 (42) 
Weight          
n (%) 113,596 (87) 34,400 (88) 208,250 (84) 54,200 (85) 398,091 (83) 90,666 (84) 517,809 (82) 84,026 (83) 1,501,038 (83) 
BMI          
n (% of nonmissing) 111,448 (85) 33,812 (86) 204,744 (83) 53,364 (84) 391,331 (81) 89,076 (83) 506,953 (80) 82,244 (81) 1,472,972 (82) 
 Mean (SD) kg/m2 37.2 (10.0) 38.9 (10.4) 36.6 (8.8) 37.1 (9.2) 35.3 (8.2) 35.1 (8.4) 33.6 (7.4) 33.4 (7.8) 35.1 (8.4) 
 Obese, n (% of nonmissing) 84,976 (76) 27,462 (81) 160,183 (78) 41,994 (79) 289,343 (74) 63,868 (72) 335,787 (66) 52,812 (64) 1,056,425 (72) 
Systolic blood pressure          
n (% of nonmissing) 112,977 (86.1) 34,339 (88) 206,005 (83) 53,748 (85) 390,427 (81) 89,018 (83) 506,904 (80) 82,473 (81) 1,475,891 (82) 
 Mean (SD) mmHg 125.5 (14.0) 129 (16) 129 (15) 133 (17) 131 (16) 135 (18) 133 (17) 137 (18) 132 (17) 
 ≥140 mmHg, n (% of nonmissing) 16,364 (14) 7,477 (22) 44,332 (22) 16,389 (30) 104,503 (27) 31,439 (35) 159,265 (31) 33,400 (40) 413,169 (28) 
 ≥130 mmHg, n (% of nonmissing) 112,977 (86) 34,339 (88) 206,005 (83) 53,748 (85) 390,427 (81) 89,018 (83) 506,904 (80) 82,473 (81) 1,475,891 (82) 
LDL cholesterol          
n (% of nonmissing) 38,609 (29.4) 11,644 (30) 80,114 (32) 20,044 (32) 142,898 (30) 30,905 (29) 168,498 (27) 25,315 (25) 518,027 (29) 
 Mean (SD) mg/dL 111 (36) 112 (37) 112 (36) 114 (38) 108 (37) 113 (39) 101 (36) 108 (38) 106.9 (37.2) 
 ≥100 mg/dL, n (% of nonmissing) 23,597 (61) 7,099 (61) 49,289 (62) 12,629 (63) 80,122 (56) 18,770 (61) 78,379 (47) 13,845 (55) 283,730 (55) 
 ≥70 mg/dL, n (% of nonmissing) 34,207 (89) 10,321 (89) 70,652 (88) 17,847 (89) 121,807 (85) 27,003 (87) 134,665 (80) 21,401 (85) 437,903 (85) 
Non-HDL cholesterol          
n (% of nonmissing) 49,813 (38) 15,453 (39) 102,155 (41) 26,271 (41) 180,375 (38) 40,497 (38) 209,806 (33) 33,545 (33) 657,915 (37) 
 Mean (SD) mg/dL 154.1 (50) 143 (46) 155 (49) 144 (45) 148 (47) 142 (45) 136 (44) 134 (43) 144 (47) 
 ≥130 mg/dL, n (% of nonmissing) 34,177 (69) 8,891 (58) 70,877 (69) 15,700 (60) 112,174 (62) 23,141 (57) 106,933 (51) 16,529 (49) 388,422 (59) 
 ≥100 mg/dL, n (% of nonmissing) 45,110 (91) 13,127 (85) 92,630 (91) 22,560 (86) 156,493 (87) 33,883 (84) 167,621 (80) 26,307 (78) 557,731 (85) 
Triglycerides          
n (% of nonmissing) 49,946 (38) 15,597 (40) 102,256 (41) 26,527 (42) 180,584 (38) 40,870 (38) 208,725 (33) 33,934 (34) 658,439 (37) 
 Median (Q1, Q3), mg/dL 177 (119, 273) 118 (83, 176) 180 (124, 273) 120 (85, 177) 169 (119, 247) 119, (86, 172) 156 (111, 221) 112 (82, 156) 156 (109, 231) 
 ≥150 mg/dL, n (% of nonmissing) 30,580 (61) 5,277 (34) 64,253 (63) 9,218 (35) 106,893 (59) 13,656 (33) 111,042 (53) 9,422 (28) 350,341 (53) 
eGFR          
n (% of nonmissing) 64,202 (49) 20,098 (51) 124,285 (50) 32,348 (51) 221,282 (46) 50,601 (47) 272,602 (43) 44,076 (44) 829,494 (46) 
 Median (Q1, Q3) mL/min/1.73 m2 106 (90, 125) 109 (92, 127) 95 (8, 112) 97 (80, 114) 86 (71.1 101.5) 87.5 (71, 103) 76 (60, 90) 78 (58, 94) 86 (69, 103) 
 <60 mL/min/1.73 m2, n (% of nonmissing) 1,561 (2) 857 (4) 7,017 (6) 2,801 (9) 25,171 (11) 7,777 (15) 70,207 (26) 11,843 (27) 127,234 (15) 
Age-group, years
18–3940–4950–5960–70
CharacteristicWAAAWAAAWAAAWAAATotal
Patients, n 131,288 39,217 247,622 63,429 480,782 107,371 631,980 101,316 1,803,005 
Follow-up, mean (SD), years 5.1 (4.3) 4.7 (4.1) 5.1 (4.3) 4.7 (4.1) 4.9 (4.2) 4.4 (3.9) 5.0 (4.3) 4.5 (4.0) 4.9 (4.2) 
Male sex, n (%) 50,441 (38) 13,141 (34) 122,982 (50) 26,456 (42) 245,543 (51) 42,998 (40) 324,226 (51) 38,975 (38) 864,762 (48) 
Age, mean (SD), years 33 (5) 32 (6) 45 (3) 45 (3) 55 (3) 55 (3) 65 (3) 64 (3) 55 (11) 
Smoking status, n (%)          
 Ever-smoker 50,028 (38) 12,243 (31) 100,396 (41) 21,210 (33) 198,534 (41) 42,354 (39) 247,785 (39) 37,493 (37) 710,043 (39) 
 Unknown 43,358 (33) 12,901 (33) 79,333 (32) 20,500 (32) 154,194 (32) 33,739 (31) 218,705 (35) 35,466 (35) 598,196 (33) 
HbA1c          
n (% of nonmissing) 60,456 (46) 19,168 (49) 104,708 (42) 27,985 (44) 179,041 (37) 41,896 (39) 206,083 (33) 34,596 (34) 673,933 (37) 
 Mean (SD) % 8.2 (2.3) 8.6 (2.6) 8.1 (2.2) 8.4 (2.5) 7.8 (2.1) 8.1 (2.3) 7.4 (1.7) 7.7 (2.0) 7.8 (2.1) 
 ≥7.5%, n (% of nonmissing) 31,395 (52) 10,381 (54) 50,373 (48) 14,015 (50) 76,458 (43) 18,756 (45) 68,104 (33) 13,062 (38) 282,544 (42) 
Weight          
n (%) 113,596 (87) 34,400 (88) 208,250 (84) 54,200 (85) 398,091 (83) 90,666 (84) 517,809 (82) 84,026 (83) 1,501,038 (83) 
BMI          
n (% of nonmissing) 111,448 (85) 33,812 (86) 204,744 (83) 53,364 (84) 391,331 (81) 89,076 (83) 506,953 (80) 82,244 (81) 1,472,972 (82) 
 Mean (SD) kg/m2 37.2 (10.0) 38.9 (10.4) 36.6 (8.8) 37.1 (9.2) 35.3 (8.2) 35.1 (8.4) 33.6 (7.4) 33.4 (7.8) 35.1 (8.4) 
 Obese, n (% of nonmissing) 84,976 (76) 27,462 (81) 160,183 (78) 41,994 (79) 289,343 (74) 63,868 (72) 335,787 (66) 52,812 (64) 1,056,425 (72) 
Systolic blood pressure          
n (% of nonmissing) 112,977 (86.1) 34,339 (88) 206,005 (83) 53,748 (85) 390,427 (81) 89,018 (83) 506,904 (80) 82,473 (81) 1,475,891 (82) 
 Mean (SD) mmHg 125.5 (14.0) 129 (16) 129 (15) 133 (17) 131 (16) 135 (18) 133 (17) 137 (18) 132 (17) 
 ≥140 mmHg, n (% of nonmissing) 16,364 (14) 7,477 (22) 44,332 (22) 16,389 (30) 104,503 (27) 31,439 (35) 159,265 (31) 33,400 (40) 413,169 (28) 
 ≥130 mmHg, n (% of nonmissing) 112,977 (86) 34,339 (88) 206,005 (83) 53,748 (85) 390,427 (81) 89,018 (83) 506,904 (80) 82,473 (81) 1,475,891 (82) 
LDL cholesterol          
n (% of nonmissing) 38,609 (29.4) 11,644 (30) 80,114 (32) 20,044 (32) 142,898 (30) 30,905 (29) 168,498 (27) 25,315 (25) 518,027 (29) 
 Mean (SD) mg/dL 111 (36) 112 (37) 112 (36) 114 (38) 108 (37) 113 (39) 101 (36) 108 (38) 106.9 (37.2) 
 ≥100 mg/dL, n (% of nonmissing) 23,597 (61) 7,099 (61) 49,289 (62) 12,629 (63) 80,122 (56) 18,770 (61) 78,379 (47) 13,845 (55) 283,730 (55) 
 ≥70 mg/dL, n (% of nonmissing) 34,207 (89) 10,321 (89) 70,652 (88) 17,847 (89) 121,807 (85) 27,003 (87) 134,665 (80) 21,401 (85) 437,903 (85) 
Non-HDL cholesterol          
n (% of nonmissing) 49,813 (38) 15,453 (39) 102,155 (41) 26,271 (41) 180,375 (38) 40,497 (38) 209,806 (33) 33,545 (33) 657,915 (37) 
 Mean (SD) mg/dL 154.1 (50) 143 (46) 155 (49) 144 (45) 148 (47) 142 (45) 136 (44) 134 (43) 144 (47) 
 ≥130 mg/dL, n (% of nonmissing) 34,177 (69) 8,891 (58) 70,877 (69) 15,700 (60) 112,174 (62) 23,141 (57) 106,933 (51) 16,529 (49) 388,422 (59) 
 ≥100 mg/dL, n (% of nonmissing) 45,110 (91) 13,127 (85) 92,630 (91) 22,560 (86) 156,493 (87) 33,883 (84) 167,621 (80) 26,307 (78) 557,731 (85) 
Triglycerides          
n (% of nonmissing) 49,946 (38) 15,597 (40) 102,256 (41) 26,527 (42) 180,584 (38) 40,870 (38) 208,725 (33) 33,934 (34) 658,439 (37) 
 Median (Q1, Q3), mg/dL 177 (119, 273) 118 (83, 176) 180 (124, 273) 120 (85, 177) 169 (119, 247) 119, (86, 172) 156 (111, 221) 112 (82, 156) 156 (109, 231) 
 ≥150 mg/dL, n (% of nonmissing) 30,580 (61) 5,277 (34) 64,253 (63) 9,218 (35) 106,893 (59) 13,656 (33) 111,042 (53) 9,422 (28) 350,341 (53) 
eGFR          
n (% of nonmissing) 64,202 (49) 20,098 (51) 124,285 (50) 32,348 (51) 221,282 (46) 50,601 (47) 272,602 (43) 44,076 (44) 829,494 (46) 
 Median (Q1, Q3) mL/min/1.73 m2 106 (90, 125) 109 (92, 127) 95 (8, 112) 97 (80, 114) 86 (71.1 101.5) 87.5 (71, 103) 76 (60, 90) 78 (58, 94) 86 (69, 103) 
 <60 mL/min/1.73 m2, n (% of nonmissing) 1,561 (2) 857 (4) 7,017 (6) 2,801 (9) 25,171 (11) 7,777 (15) 70,207 (26) 11,843 (27) 127,234 (15) 
Table 2

Prevalence of comorbidities at the time of T2D diagnosis, by age and ethnicity

Age-group, years
18–3940–4850–5960–70
DiseaseWAAAWAAAWAAAWAAATotal
Patients, n 131,288 39,217 247,622 63,429 480,782 107,371 631,980 101,316 1,803,005 
ASCVD 4,069 (3) 1,036 (3) 17,489 (7) 3,940 (6) 58,544 (12) 11,259 (10) 117,699 (19) 15,674 (15) 229,710 (13) 
Myocardial infarction 690 (1) 177 (0) 2,944 (1) 601 (1) 8,875 (2) 1,419 (1) 15,042 (2) 1,661 (2) 31,409 (2) 
Stroke 386 (0) 154 (0) 1,479 (1) 634 (1) 4,560 (1) 1,859 (2) 8,342 (1) 2,408 (2) 19,822 (1) 
Peripheral arterial disease 824 (1) 258 (1) 3,283 (1) 803 (1) 11,858 (2) 2,504 (2) 24,012 (4) 3,950 (4) 47,492 (3) 
Cancer 2,021 (2) 443 (1) 5,569 (2) 1,256 (2) 15,937 (3) 3,507 (3) 34,234 (5) 5,887 (6) 68,854 (4) 
Depression 31,307 (24) 5,022 (13) 64,723 (26) 9,463 (15) 125,461 (26) 16,862 (16) 146,716 (23) 13,272 (13) 412,826 (23) 
Hypertension 37,576 (29) 14,906 (38) 104,770 (42) 34,047 (54) 236,819 (49) 64,948 (60) 337,488 (53) 64,635 (64) 895,189 (50) 
Dyslipidemia 27,201 (21) 6,141 (16) 83,553 (34) 17,468 (28) 191,728 (40) 37,303 (35) 273,089 (43) 39,800 (39) 676,283 (38) 
Microvascular disease 14,689 (11) 4,677 (12) 41,186 (17) 11,531 (18) 92,755 (19) 23,887 (22) 147,141 (23) 25,931 (26) 361,797 (20) 
CKD at baseline 4,583 (3) 2,081 (5) 14,768 (6) 5,496 (9) 41,234 (9) 12,637 (12) 90,819 (14) 16,919 (17) 188,537 (10) 
HF at baseline 813 (1) 638 (2) 3,306 (1) 1,869 (3) 10,000 (2) 4,084 (4) 20,889 (3) 4,823 (5) 46,422 (3) 
CKD or HF 5,239 (4) 2,568 (7) 17,377 (7) 6,821 (11) 48,728 (10) 15,430 (14) 104,562 (17) 19,902 (20) 220,627 (12) 
Age-group, years
18–3940–4850–5960–70
DiseaseWAAAWAAAWAAAWAAATotal
Patients, n 131,288 39,217 247,622 63,429 480,782 107,371 631,980 101,316 1,803,005 
ASCVD 4,069 (3) 1,036 (3) 17,489 (7) 3,940 (6) 58,544 (12) 11,259 (10) 117,699 (19) 15,674 (15) 229,710 (13) 
Myocardial infarction 690 (1) 177 (0) 2,944 (1) 601 (1) 8,875 (2) 1,419 (1) 15,042 (2) 1,661 (2) 31,409 (2) 
Stroke 386 (0) 154 (0) 1,479 (1) 634 (1) 4,560 (1) 1,859 (2) 8,342 (1) 2,408 (2) 19,822 (1) 
Peripheral arterial disease 824 (1) 258 (1) 3,283 (1) 803 (1) 11,858 (2) 2,504 (2) 24,012 (4) 3,950 (4) 47,492 (3) 
Cancer 2,021 (2) 443 (1) 5,569 (2) 1,256 (2) 15,937 (3) 3,507 (3) 34,234 (5) 5,887 (6) 68,854 (4) 
Depression 31,307 (24) 5,022 (13) 64,723 (26) 9,463 (15) 125,461 (26) 16,862 (16) 146,716 (23) 13,272 (13) 412,826 (23) 
Hypertension 37,576 (29) 14,906 (38) 104,770 (42) 34,047 (54) 236,819 (49) 64,948 (60) 337,488 (53) 64,635 (64) 895,189 (50) 
Dyslipidemia 27,201 (21) 6,141 (16) 83,553 (34) 17,468 (28) 191,728 (40) 37,303 (35) 273,089 (43) 39,800 (39) 676,283 (38) 
Microvascular disease 14,689 (11) 4,677 (12) 41,186 (17) 11,531 (18) 92,755 (19) 23,887 (22) 147,141 (23) 25,931 (26) 361,797 (20) 
CKD at baseline 4,583 (3) 2,081 (5) 14,768 (6) 5,496 (9) 41,234 (9) 12,637 (12) 90,819 (14) 16,919 (17) 188,537 (10) 
HF at baseline 813 (1) 638 (2) 3,306 (1) 1,869 (3) 10,000 (2) 4,084 (4) 20,889 (3) 4,823 (5) 46,422 (3) 
CKD or HF 5,239 (4) 2,568 (7) 17,377 (7) 6,821 (11) 48,728 (10) 15,430 (14) 104,562 (17) 19,902 (20) 220,627 (12) 

Data are n (%) unless otherwise indicated.

Exposure to different antidiabetic drugs, except for GLP-1RAs and SGLT-2is, was similar in both ethnicities across age-groups (Supplementary Table 1). Overall, 8 and 11% of AAs and WAs, respectively, received GLP-1RAs and 5 and 8% received SGLT-2is at any time during follow-up.

Risk of CKD and HF

Among patients without a history of CKD at T2D diagnosis, the CKD incidence rose with age and was significantly higher in AAs than in WAs in all age-groups (Supplementary Table 2). The HRs of developing CKD in AAs compared with WAs were 1.2 (95% CI 1.2–1.3), 1.3 (1.2–1.3), 1.2 (1.2–1.2), and 1.1 (1.1–1.1) in the 18–39, 40–49, 50–59, and 60–70 age-groups, respectively (Table 3).

Table 3

Risk of developing CKD, HF, and composite of CKD and HF from time of T2D diagnosis by age, ethnicity, and sex

Adjusted risk (95% CI)
CKDHF
Age-group, yearsAA vs. WAAA male vs. femaleWA male vs. femaleAA vs. WAAA male vs. femaleWA male vs. femaleComposite AA vs. WC
18–39 1.21 (1.17, 1.26) 1.49 (1.39, 1.60) 1.23 (1.19, 1.28) 2.21 (1.98, 2.45) 1.82 (1.53, 2.15) 1.39 (1.22, 1.59) 1.23 (1.18, 1.28) 
40–49 1.25 (1.22, 1.28) 1.28 (1.23, 1.34) 1.01 (0.98, 1.03) 1.86 (1.75, 1.97) 1.57 (1.42, 1.73) 1.50 (1.41, 1.60) 1.26 (1.23, 1.29) 
50–59 1.21 (1.19, 1.24) 1.11 (1.07, 1.15) 0.93 (0.91, 0.94) 1.54 (1.48, 1.60) 1.54 (1.44, 1.65) 1.49 (1.44, 1.55) 1.22 (1.20, 1.24) 
60–70 1.10 (1.08, 1.12) 0.97 (0.94, 1.00) 0.86 (0.85, 0.87) 1.11 (1.07, 1.15) 1.36 (1.28, 1.45) 1.33 (1.30, 1.36) 1.10 (1.08, 1.12) 
Adjusted risk (95% CI)
CKDHF
Age-group, yearsAA vs. WAAA male vs. femaleWA male vs. femaleAA vs. WAAA male vs. femaleWA male vs. femaleComposite AA vs. WC
18–39 1.21 (1.17, 1.26) 1.49 (1.39, 1.60) 1.23 (1.19, 1.28) 2.21 (1.98, 2.45) 1.82 (1.53, 2.15) 1.39 (1.22, 1.59) 1.23 (1.18, 1.28) 
40–49 1.25 (1.22, 1.28) 1.28 (1.23, 1.34) 1.01 (0.98, 1.03) 1.86 (1.75, 1.97) 1.57 (1.42, 1.73) 1.50 (1.41, 1.60) 1.26 (1.23, 1.29) 
50–59 1.21 (1.19, 1.24) 1.11 (1.07, 1.15) 0.93 (0.91, 0.94) 1.54 (1.48, 1.60) 1.54 (1.44, 1.65) 1.49 (1.44, 1.55) 1.22 (1.20, 1.24) 
60–70 1.10 (1.08, 1.12) 0.97 (0.94, 1.00) 0.86 (0.85, 0.87) 1.11 (1.07, 1.15) 1.36 (1.28, 1.45) 1.33 (1.30, 1.36) 1.10 (1.08, 1.12) 

Among patients without a history of HF at T2D diagnosis, the HF incidence rates were significantly higher among AAs than WAs in all age-groups (all P < 0.01) (Supplementary Table 2), except for the 60–70 age-group. In the 18–39, 40–49, 50–59, and 60–70 age-groups, the adjusted HRs of developing HF in AAs compared with WAs were 2.2 (95% CI 2.0–2.5), 1.9 (1.8–2.0), 1.5 (1.5–1.6), and 1.1 (1.1–1.2), respectively (Table 3).

Among 12,906 and 59,180 AAs and WAs, respectively, who had no HF history at T2D onset, 48 and 52% had CKD before incident HF (P < 0.01). Among 59,995 and 279,381 AAs and WAs who had no CKD history at T2D onset, 8 and 7% had HF before incident CKD. The CKD rates were significantly higher among those with prior HF than those without across all age-groups in both ethnicities (Supplementary Table 2).

Within the AA cohort, males aged 18–59 years had 11–49% significantly higher rates of and adjusted CKD risk (all P < 0.01) (Supplementary Table 3 and Table 3) than females. Within the WA cohort, males aged 18–39 years had 23% significantly higher CKD risk than females (P < 0.01), while males aged ≥50 years had 7–14% significantly lower CKD risk than their female counterparts. Males had consistently higher rates and adjusted HF risk across all age-groups in both ethnicities.

Among patients without a history of HF and CKD at T2D diagnosis, the composite of CKD and HF incidence rates per 1,000 person-years in AAs and WAs in the 18–39, 40–49, 50–59, 60–70 age-groups were 30 and 24, 48 and 37, 67 and 53, and 94 and 82, respectively (all P < 0.01) (Supplementary Table 2). In these age-groups, the adjusted HRs of developing the composite of CKD or HF in AAs compared with WAs were 1.2 (95% CI 1.2–1.3), 1.3 (1.2–1.3), 1.2 (1.2.-1.2), and 1.1 (1.1–1.2), respectively (Table 3).

Time to Developing CKD and HF

In the 18–39, 40–49, 50–59, and 60–70 age-groups, the adjusted RMT to incident CKD in the AA and WA cohorts was 8.7 (95% CI 8.2–9.1) and 9.7 (9.2–10.2) years, 7.0 (6.7–7.3) and 8.2 (7.9–8.5) years, 6.1 (5.8–6.4) and 7.1 (6.8–7.3) years, and 5.1 (4.9–5.3) and 5.7 (5.5–5.8) years, respectively (Fig. 1 and Supplementary Table 4). This time to event was, on average, 3.6 and 4.0 years longer in AAs and WAs aged 18–39 at T2D diagnosis compared with those aged 60–70.

Figure 1

Time of T2D diagnosis by age-group and ethnicity (WA and AA) and adjusted RMT (in years) to developing CKD and HF.

Figure 1

Time of T2D diagnosis by age-group and ethnicity (WA and AA) and adjusted RMT (in years) to developing CKD and HF.

Close modal

In the 18–39, 40–49, 50–59, and 60–70 age-groups, the adjusted RMT to incident HF in the AA and WA cohorts was 10.3 (9.3–11.2) and 12.1 (10.6–13.5) years, 8.6 (8.1–9.1) and 9.6 (9.1–10.0) years, 7.5 (7.0–8.1) and 8.9 (8.4–9.5) years, and 6.9 (6.5–7.4) and 8.0 (7.8–8.3) years, respectively (Fig. 1 and Supplementary Table 4). This time to event was, on average, 3.4 and 4.1 years longer in AAs and WAs aged 18–39 at T2D diagnosis compared with those aged 60–70. The average time to the composite of CKD and HF is presented in Supplementary Table 4.

Mediation Effect of CKD and HF

The overall independent and interactive mediation effects of CKD in the context of risk difference in HF between AAs and WAs were 54 (95% CI 40–67), 40 (30–50), 41 (33–49), and 91% (65–96) in the 18–39, 40–49, 50–59, and 60–70 age-groups, respectively. In these age-groups, the mediation effects of HF in the context of risk difference in CKD between AAs and WAs were 13 (3–66), 39 (17–62), 35 (18–52), and 15% (3–49).

The mediation effect of CKD in the observed higher HF risk in usual-onset (≥50 years) compared with young-onset (<50 years) T2D in AAs (63% [95% CI 55–71], P < 0.01) and WAs (52% [48–57], P < 0.01) was similar. In both ethnicities, 49% (47–52) of the estimated higher HF risk in males compared with females was mediated through prior CKD. The mediation effect of HF in the observed higher CKD risk in usual-onset compared with young-onset T2D in AAs (74% [67–80], P < 0.01] and WAs (66% [61–71], P < 0.01) was also similar.

The novelty of this large U.S. nationally representative real-world–based study is the holistic exploration of cardiorenal risk patterns in an incident T2D population stratified by age at disease onset and ethnicity. To our knowledge, the prevalence of CKD and HF by age at T2D diagnosis, time to incident CKD and HF, and mediation effects of CKD in HF risk (and vice versa) in the young-onset (<50 years) and usual-onset (≥50 years) T2D population have not been previously explored. Large national surveys do not report age with ethnicity and sex interactions for cardiorenal risk, and this study presents valuable epidemiological insights with regard to this major public health problem.

Specifically, we observed in this study 1) a consistently higher risk of cardiorenal diseases in AAs than in WAs in all age-groups; 2) a higher CKD risk in AA males aged <40 years compared with female counterparts, while females aged ≥50 years had a higher CKD risk; 3) similar mediation effects of HF (CKD) in terms of explaining ethnic and sex differences; and 4) a consistently shorter time to developing cardiorenal complications for AAs than WAs, while for both ethnicities, the youngest age-groups could develop diseases on average ∼4 years later than in the oldest age-groups.

We observed that AAs had consistently higher cardiorenal risk compared with WAs across all age-groups. However, the ethnicity-specific difference in CKD risk was lower in the 60–70 age-group (95% CI 1.08–1.12), while the risk difference was similarly higher across age-groups <60 years (95% CI 1.17–1.28). The ethnicity-specific HF risk difference linearly reduced with increasing age, with differences between AAs and WAs aged <40 years being twofold higher than that those aged ≥60 years (HR 2.21 [95% CI 1.98–2.45] vs. 1.11 [1.07–1.15]). AA males aged <60 years had a significantly higher CKD risk than females, while WA males had a higher CKD risk only in the 18–39 age-group, suggesting similar CKD risk between WA males and females aged ≥40 years at T2D diagnosis. Males across all age-groups from both ethnicities had higher HF risk than females. Accounting for confounders, patients with young-onset T2D could develop CKD or HF on average within 10 years of diagnosis, with 3–4 years average time difference between usual and young onset. Previous findings suggested increased morbidity and mortality in people who develop either CKD or HF (6), and our findings therefore highlight potential life-years lost in those who are diagnosed with T2D at younger ages. A recent U.S. study investigating HF hospitalization trends in people with diabetes reported a declining proportion of hospitalization due to HF in patients aged ≥65 years during 2005–2015, while the incidence of HF hospitalization among younger patients remained relatively stable (12). In addition, the probability of death as a result of diabetic CKD has significantly increased in younger people (aged 20–54 years) in the U.S. between 2002–2016, with diabetic CKD contributing 27 and 26% of the increased probability of death as a result of CKD in patients with young- and usual-onset diabetes, respectively (28).

A recent multinational study in people with T2D reported that the presence of HF was associated with a more than twofold increased risk of incident CKD and that incident CKD was associated with about a twofold increased risk of HF (6). We have observed that in the context of higher adjusted HF and CKD risk in usual- compared with young-onset T2D, mediative and interactive effects of the CKD in HF risk and that of HF in CKD risk were statistically similar between ethnicities (95% CI 48–80%), although with a higher magnitude in the AA cohort. Interestingly, in the context of explaining ethnic differences, the mediation of CKD in HF risk was higher in the youngest and oldest age-groups, while the magnitude of overall mediative effects of HF in CKD risk was lower in the youngest and oldest age-groups. This suggests possible nonlinear ethnic differences in the risk of cardiorenal diseases in different age-groups. While 50% of WAs and 39% of AAs in our cohort were male, the mediative effects of CKD in HF risk and HF in CKD risk in the context of sex differences were similar between ethnicities. Interestingly, the risk of CKD in WA males aged <40 years was higher than in females, while WA females aged >50 years had higher CKD risk than males.

SGLT-2is are now recognized for their cardiorenal benefits in patients with and without T2D. However, most of the associated clinical trials had only 4–5% representation from the AA community (29). While the primary aim of these large randomized controlled trials was to demonstrate drug efficacy on cardiorenal outcomes, in the context of recent debates around individualized treatment strategies and recognized ethnic differences, minority populations (young-onset T2D and AA) remain underrepresented, even though they have higher life-years lost when at least one comorbidity is present. Furthermore, the scarcity of trials investigating patients with HF with reduced renal function has been previously discussed (30). While prescription patterns for SGLT-2is and GLP-1RAs were different between AAs and WAs, the adjusted CKD and HF risk differences between the ethnicities across age-groups were similar irrespective of adjustments for these therapies. A differently designed real-world study would be of great value to evaluate whether the mediation effects of these therapies could be different in different ethnicities. Similarly, individual explorations of the mediation effects of agents acting on the renin-angiotensin system in terms of CKD and HF risk in different ethnicities would be of great interest.

This study has several strengths, including the use of nationally representative population-based data with a mean of 5 years of follow-up and identification of people with T2D using a robust, clinically guided machine learning approach, reducing bias due to underidentification and misclassification. We have defined CKD using clinical diagnosis and laboratory measurements and robustly extracted longitudinal patient-level medication data (22). This EMR-based study also has several limitations. Coding of conditions is a common limitation when using EMRs. The HF event estimates would be underrepresented with no definitive data on hospitalization due to HF. Similarly, identification of people with CKD from EMRs and claims data is always challenging, with a significant underestimation risk. Other limitations include unavoidable indication bias and residual confounding, which are common problems in any EMR-based outcome studies, and lack of data on socioeconomic characteristics, physical activity, medication adherence, nature of insurance, education, income, and possible cultural drivers. Finally, Asian and Hispanic populations were not studied.

Despite these limitations, we believe that the large cohort size, robust study design, use of advanced data mining methods, and appropriate assessment of confounders ensure reliable population-level estimates and reflect real-world patterns of CKD and HF risk in people with incident T2D in a detailed evaluation by age, ethnicity, and sex. Our findings relative to different risk paradigms in AAs and WAs, including a mediation effect difference of HF and CKD, reflect the inherent differences in how the pathway of development of cardiometabolic diseases differ in these ethnic groups (31).

In conclusion, both CKD and HF were apparent in some individuals within a few years of a T2D diagnosis, even among those diagnosed at age <40 years and particularly within the AA population. An urgent need exists to incorporate multidisciplinary care in the identification of high-risk patients from T2D onset, along with legislative support promoting equitable access to therapies and care, especially for young, vulnerable, and underrepresented patient populations.

This article contains supplementary material online at https://doi.org/10.2337/figshare.19747249.

Funding. Melbourne EpiCentre gratefully acknowledges support from the National Health and Medical Research Council and the Australian Government’s National Collaborative Research Infrastructure Strategy initiative through Therapeutic Innovation Australia.

Duality of Interest. S.K.P. is currently an employee of AstraZeneca and has acted as a consultant and/or speaker for Novartis, Sanofi Aventis, GI Dynamics, Roche, AstraZeneca, Guangzhou Zhongyi Pharmaceutical, and Amylin Pharmaceuticals. He has received grants in support of investigator and investigator-initiated clinical studies from Merck, Novo Nordisk, AstraZeneca, Hospira, Amylin Pharmaceuticals, Sanofi, and Pfizer. J.E.S. has received honoraria for advisory boards and lectures from AstraZeneca, Eli Lilly, Novo Nordisk, Sanofi, Mylan, Boehringer Ingelheim, MSD, Abbott, and Pfizer. P.F. is a full-time employee of AstraZeneca, venture partner at HT180, and chairman of hemostatics and contract teaching professor at Catholic University in Rome. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. S.K.P. and O.M. conceptualized and designed the study, conducted the statistical analyses, and wrote the first draft of the manuscript. J.E.S. and P.F. contributed to the clinical interpretation of the results and manuscript writing. O.M. conducted the data extraction. S.K.P. and O.M. are the guarantors of this work and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Prior Presentation. Parts of this study were presented in abstract form at the 81st Scientific Sessions of the American Diabetes Association, Virtual, 25–29 June 2021.

1.
Centers for Disease Control and Prevention
.
Chronic Kidney Disease in the United States, 2021
.
2.
Virani
SS
,
Alonso
A
,
Aparicio
HJ
, et al.;
American Heart Association Council on Epidemiology and Prevention Statistics Committee and Stroke Statistics Subcommittee
.
Heart disease and stroke statistics - 2021 update: a report from the American Heart Association
.
Circulation
2021
;
143
:
e254
e743
3.
Kenny
HC
,
Abel
ED
.
Heart failure in type 2 diabetes mellitus: impact of glucose-lowering agents, heart failure therapies, and novel therapeutic strategies
.
Circ Res
2019
;
124
:
121
141
4.
Aguilar
D
.
Heart failure, diabetes mellitus, and chronic kidney disease: a clinical conundrum
.
Circ Heart Fail
2016
;
9
:
e003316
5.
Arnold
SV
,
Kosiborod
M
,
Wang
J
,
Fenici
P
,
Gannedahl
G
,
LoCasale
RJ
.
Burden of cardio-renal-metabolic conditions in adults with type 2 diabetes within the Diabetes Collaborative Registry
.
Diabetes Obes Metab
2018
;
20
:
2000
2003
6.
Birkeland
KI
,
Bodegard
J
,
Eriksson
JW
, et al
.
Heart failure and chronic kidney disease manifestation and mortality risk associations in type 2 diabetes: a large multinational cohort study
.
Diabetes Obes Metab
2020
;
22
:
1607
1618
7.
Cherney
DZI
,
Repetto
E
,
Wheeler
DC
, et al
.
Impact of cardio-renal-metabolic comorbidities on cardiovascular outcomes and mortality in type 2 diabetes mellitus
.
Am J Nephrol
2020
;
51
:
74
82
8.
Ling
J
,
Koye
D
,
Buizen
L
,
Khunti
K
,
Montvida
O
,
Paul
SK
.
Temporal trends in co-morbidities and cardiometabolic risk factors at the time of diagnosis of type 2 diabetes in UK
.
Diabetes Obes Metab
2021
;
23
:
1150
1161
9.
Koye
DN
,
Ling
J
,
Dibato
J
,
Khunti
K
,
Montvida
O
,
Paul
SK
.
Temporal trend in young-onset type 2 diabetes-macrovascular and mortality risk: study of U.K. primary care electronic medical records
.
Diabetes Care
2020
;
43
:
2208
2216
10.
Dibato
JE
,
Montvida
O
,
Zaccardi
F
, et al
.
Association of cardiometabolic multimorbidity and depression with cardiovascular events in early-onset adult type 2 diabetes: a multiethnic study in the U.S
.
Diabetes Care
2021
;
44
:
231
239
11.
Kountz
DS
.
The use of sodium-glucose cotransporter 2 inhibitors in patients with type 2 diabetes and hypertension: a focus on African-American populations
.
Postgrad Med
2017
;
129
:
421
429
12.
Honigberg
MC
,
Patel
RB
,
Pandey
A
, et al
.
Trends in hospitalizations for heart failure and ischemic heart disease among US adults with diabetes
.
JAMA Cardiol
2021
;
6
:
354
357
13.
Nayak
A
,
Hicks
AJ
,
Morris
AA
.
Understanding the complexity of heart failure risk and treatment in black patients
.
Circ Heart Fail
2020
;
13
:
e007264
14.
Chaturvedi
N
,
Eastwood
S
.
Prescribing by ethnicity: (im)precision medicine?
Diabetes Care
2020
;
43
:
1687
1689
15.
Dunlay
SM
,
Givertz
MM
,
Aguilar
D
, et al.;
American Heart Association Heart Failure and Transplantation Committee of the Council on Clinical Cardiology
;
Council on Cardiovascular and Stroke Nursing
;
Heart Failure Society of America
.
Type 2 diabetes mellitus and heart failure: a scientific statement from the American Heart Association and the Heart Failure Society of America
.
Circulation
2019
;
140
:
e294
e324
16.
Centers for Disease Control and Prevention
.
Health, United States Spotlight: Racial and Ethnic Disparities in Heart Disease, April 2019
.
17.
Ziaeian
B
,
Kominski
GF
,
Ong
MK
,
Mays
VM
,
Brook
RH
,
Fonarow
GC
.
National differences in trends for heart failure hospitalizations by sex and race/ethnicity
.
Circ Cardiovasc Qual Outcomes
2017
;
10
:
e003552
18.
Montvida
O
,
Verma
S
,
Shaw
JE
,
Paul
SK
.
Cardiometabolic risk factor control in Black and White people in the United States initiating sodium-glucose co-transporter-2 inhibitors: a real-world study
.
Diabetes Obes Metab
2020
;
22
:
2384
2397
19.
Paul
SK
,
Bhatt
DL
,
Montvida
O
.
The association of amputations and peripheral artery disease in patients with type 2 diabetes mellitus receiving sodium-glucose cotransporter type-2 inhibitors: real-world study
.
Eur Heart J
2021
;
42
:
1728
1738
20.
Owusu Adjah
ES
,
Montvida
O
,
Agbeve
J
,
Paul
SK
.
Data mining approach to identify disease cohorts from primary care electronic medical records: a case of diabetes mellitus
.
Open Bioinform J
2017
;
10
:
16
27
21.
Montvida
O
,
Shaw
J
,
Atherton
JJ
,
Stringer
F
,
Paul
SK
.
Long-term trends in antidiabetes drug usage in the US: real-world evidence in patients newly diagnosed with type 2 diabetes
.
Diabetes Care
2018
;
41
:
69
78
22.
Montvida
O
,
Arandjelović
O
,
Reiner
E
,
Paul
SK
.
Data mining approach to estimate the duration of drug therapy from longitudinal electronic medical records
.
Open Bioinform J
2017
;
10
24.
Deferio
JJ
,
Levin
TT
,
Cukor
J
, et al
.
Using electronic health records to characterize prescription patterns: focus on antidepressants in nonpsychiatric outpatient settings
.
JAMIA Open
2018
;
1
:
233
245
25.
U.S. Food and Drug Administration
.
FDA approved drugs
2019
.
Accessed 5 May 2022. Available from https://www.accessdata.fda.gov/scripts/cder/daf/
26.
Lambert
PC
,
Royston
P
.
Further development of flexible parametric models for survival analysis
.
Stata J
2009
;
9
:
265
290
27.
Discacciati
A
,
Bellavia
A
,
Lee
JJ
,
Mazumdar
M
,
Valeri
L
.
Med4way: A Stata Command to Investigate Mediating and Interactive Mechanisms Using the Four-Way Effect Decomposition
.
New York
,
Oxford University Press
,
2019
28.
Bowe
B
,
Xie
Y
,
Li
T
, et al
.
Changes in the us burden of chronic kidney disease from 2002 to 2016: an analysis of the global burden of disease study
.
JAMA Netw Open
2018
;
1
:
e184412
29.
Tuttle
KR
,
Brosius
FC
3rd
,
Cavender
MA
, et al
.
SGLT2 inhibition for CKD and cardiovascular disease in type 2 diabetes: report of a scientific workshop sponsored by the National Kidney Foundation
.
Am J Kidney Dis
2021
;
77
:
94
109
30.
Damman
K
,
Valente
MA
,
Voors
AA
,
O’Connor
CM
,
van Veldhuisen
DJ
,
Hillege
HL
.
Renal impairment, worsening renal function, and outcome in patients with heart failure: an updated meta-analysis
.
Eur Heart J
2014
;
35
:
455
469
31.
Yaghootkar
H
,
Whitcher
B
,
Bell
JD
,
Thomas
EL
.
Ethnic differences in adiposity and diabetes risk – insights from genetic studies
.
J Intern Med
2020
;
288
:
271
283
Readers may use this article as long as the work is properly cited, the use is educational and not for profit, and the work is not altered. More information is available at https://www.diabetesjournals.org/journals/pages/license.