Long-term impact of diabetes in patients hospitalized with ischemic and non-ischemic heart failure

Predicting 1-year mortality of patients with diabetes mellitus in Kazakhstan based on administrative health data using machine learning

Diabetes mellitus (DM) affects the quality of life and leads to disability, high morbidity, and premature mortality. DM is a risk factor for cardiovascular, neurological, and renal diseases, and places a major burden on healthcare systems globally. Predicting the one-year mortality of patients with DM can considerably help clinicians tailor treatments to patients at risk. In this study, we aimed to show the feasibility of predicting the one-year mortality of DM patients based on administrative health data. We use clinical data for 472,950 patients that were admitted to hospitals across Kazakhstan between mid-2014 to December 2019 and were diagnosed with DM. The data was divided into four yearly-specific cohorts (2016-, 2017-, 2018-, and 2019-cohorts) to predict mortality within a specific year based on clinical and demographic information collected up to the end of the preceding year. We then develop a comprehensive machine learning platform to construct a predictive model of one-year mortality for each year-specific cohort. In particular, the study implements and compares the performance of nine classification rules for predicting the one-year mortality of DM patients. The results show that gradient-boosting ensemble learning methods perform better than other algorithms across all year-specific cohorts while achieving an area under the curve (AUC) between 0.78 and 0.80 on independent test sets. The feature importance analysis conducted by calculating SHAP (SHapley Additive exPlanations) values shows that age, duration of diabetes, hypertension, and sex are the top four most important features for predicting one-year mortality. In conclusion, the results show that it is possible to use machine learning to build accurate predictive models of one-year mortality for DM patients based on administrative health data. In the future, integrating this information with laboratory data or patients' medical history could potentially boost the performance of the predictive models.

Lifetime Risk of Heart Failure and Trends in Incidence Rates Among Individuals With Type 2 Diabetes Between 1995 and 2018

Background

There are limited data on the lifetime risk of heart failure (HF) in people with type 2 diabetes and how incidence has changed over time. We estimated the cumulative incidence and incidence rates of HF among Danish adults with type 2 diabetes between 1995 and 2018 using nationwide data.

Methods and Results

In total, 398 422 patients (49% women) with type 2 diabetes were identified. During follow‐up, 36 400 (9%) were diagnosed with HF and 121 459 (30%) were censored due to death. Using the Aalen‐Johansen estimators, accounting for the risk of death, the estimated residual lifetime risk of HF at age 50 years was calculated as 24% (95% CI 22%–27%) in women and 27% (25%–28%) in men. During the observational period, the proportion of patients treated with statins, angiotensin‐converting enzyme inhibitors or angiotensin II receptor blockers, and metformin increased from <30% to >60%. Similarly, the annual incidence rates of HF decreased significantly, with declines being greater in older versus younger individuals (5% versus 2% in age >50 versus ≤50 years, respectively; P<0.0001) and in women versus men (5% versus 4%, P=0.02), but similar in patients with and without IHD (4% versus 4%, P=0.53).

Conclusions

The current lifetime risk of HF in type 2 diabetes approximates 1 in 4 for men and women. Paralleled by an increase in use of evidence‐based pharmacotherapy over the past decades, the risk of developing HF has declined across several subgroups and regardless of underlying IHD, suggesting that optimal diabetes treatment can mitigate HF risk.

Long-Term Mortality Associated With Use of Carvedilol Versus Metoprolol in Heart Failure Patients With and Without Type 2 Diabetes: A Danish Nationwide Cohort Study

Background

Carvedilol may have favorable glycemic properties compared with metoprolol, but it is unknown if carvedilol has mortality benefit over metoprolol in patients with type 2 diabetes (T2D) and heart failure with reduced ejection fraction (HFrEF).

Methods and Results

Using Danish nationwide databases between 2010 and 2018, we followed patients with new‐onset HFrEF treated with either carvedilol or metoprolol for all‐cause mortality until the end of 2018. Follow‐up started 120 days after initial HFrEF diagnosis to allow initiation of guideline‐directed medical therapy. There were 39 260 patients on carvedilol or metoprolol at baseline (mean age 70.8 years, 35% women), of which 9355 (24%) had T2D. Carvedilol was used in 2989 (32%) patients with T2D and 10 411 (35%) of patients without T2D. Users of carvedilol had a lower prevalence of atrial fibrillation (20% versus 35%), but other characteristics appeared well‐balanced between the groups. Totally 11 306 (29%) were deceased by the end of follow‐up. We observed no mortality differences between carvedilol and metoprolol, multivariable‐adjusted hazard ratio (HR) 0.97 (0.90–1.05) in patients with T2D versus 1.00 (0.95–1.05) for those without T2D, P for difference =0.99. Rates of new‐onset T2D were lower in users of carvedilol versus metoprolol; age, sex, and calendar year adjusted HR 0.83 (0.75–0.91), P<0.0001.

Conclusions

In a contemporary clinical cohort of HFrEF patients with and without T2D, carvedilol was not associated with a reduction in long‐term mortality compared with metoprolol. However, carvedilol was associated with lowered risk of new‐onset T2D supporting the assertion that carvedilol has a more favorable metabolic profile than metoprolol.

Long-term mortality is increased in patients with undetected prediabetes and type-2 diabetes hospitalized for worsening heart failure and reduced ejection fraction

Background

We assessed the prevalence of newly diagnosed prediabetes and type-2 diabetes mellitus (T2DM), and their impact on long-term mortality in patients hospitalized for worsening heart failure with reduced ejection fraction (HFrEF).

Methods

We included patients hospitalized with HFrEF and New York Heart Association (NYHA) functional class II–III. Baseline two-hour oral glucose tolerance test was used to classify patients as normoglycaemic or having newly diagnosed prediabetes or T2DM. Outcomes included post-discharge all-cause and cardiovascular mortality during the median follow-up of 2.1 years.

Results

At baseline, out of 150 patients (mean-age 57 ± 12 years; 88% male), prediabetes was diagnosed in 65 (43%) patients, and T2DM in 29 (19%) patients. These patients were older and more often with NYHA class III symptoms, but distribution of comorbidities was similar to normoglycaemic patients. Taking normoglycaemic patients as a reference, adjusted risk of all-cause mortality was significantly increased both in patients with prediabetes (hazard ratio, 2.6; 95% confidence interval (CI), 1.1–6.3; p = 0.040) and in patients with T2DM (hazard ratio, 5.3; 95% CI, 1.7–15.3; p = 0.023). Likewise, both prediabetes (hazard ratio, 2.9; 95% CI, 1.1–7.9; p = 0.041) and T2DM (hazard ratio, 9.7; 95% CI 2.9–36.7; p = 0.018) independently increased the risk of cardiovascular mortality compared with normoglycaemic individuals. There was no interaction between either prediabetes or T2DM and heart failure aetiology or gender on study outcomes (all interaction p-values > 0.05).

Conclusions

Newly diagnosed prediabetes and T2DM are highly prevalent in patients hospitalized for worsening HFrEF and NYHA functional class II–III. Importantly, they impose independently increased long-term risk of higher all-cause and cardiovascular mortality.

Type 2 diabetes mellitus and heart failure: a position statement from the Heart Failure Association of the European Society of Cardiology

The coexistence of type 2 diabetes mellitus (T2DM) and heart failure (HF), either with reduced (HFrEF) or preserved ejection fraction (HFpEF), is frequent (30-40% of patients) and associated with a higher risk of HF hospitalization, all-cause and cardiovascular (CV) mortality. The most important causes of HF in T2DM are coronary artery disease, arterial hypertension and a direct detrimental effect of T2DM on the myocardium. T2DM is often unrecognized in HF patients, and vice versa, which emphasizes the importance of an active search for both disorders in the clinical practice. There are no specific limitations to HF treatment in T2DM. Subanalyses of trials addressing HF treatment in the general population have shown that all HF therapies are similarly effective regardless of T2DM. Concerning T2DM treatment in HF patients, most guidelines currently recommend metformin as the first-line choice. Sulphonylureas and insulin have been the traditional second- and third-line therapies although their safety in HF is equivocal. Neither glucagon-like preptide-1 (GLP-1) receptor agonists, nor dipeptidyl peptidase-4 (DPP4) inhibitors reduce the risk for HF hospitalization. Indeed, a DPP4 inhibitor, saxagliptin, has been associated with a higher risk of HF hospitalization. Thiazolidinediones (pioglitazone and rosiglitazone) are contraindicated in patients with (or at risk of) HF. In recent trials, sodium-glucose co-transporter-2 (SGLT2) inhibitors, empagliflozin and canagliflozin, have both shown a significant reduction in HF hospitalization in patients with established CV disease or at risk of CV disease. Several ongoing trials should provide an insight into the effectiveness of SGLT2 inhibitors in patients with HFrEF and HFpEF in the absence of T2DM.

Heart failure patients with prediabetes and newly diagnosed diabetes display abnormalities in myocardial metabolism

BACKGROUND

In type 2 diabetes, a decrease in myocardial glucose uptake (MGU) may lower glucose oxidation and contribute to progression of chronic heart failure (CHF). However, it is unsettled whether CHF patients with prediabetes have abnormal MGU and myocardial blood flow (MBF) during normal physiological conditions.

METHODS AND RESULTS

We studied 35 patients with CHF and reduced left ventricular ejections fraction (34 ± 9%) without overt T2D (mean HbA1c: 40 ± 4 mmol/mol) using echocardiography and quantitative measurements of MGU by 18F-FDG-PET and perfusion by 15O-H2O-PET. An oral glucose tolerance test (OGTT) was performed during the FDG-PET, which identified 17 patients with abnormal and 18 patients with normal glucometabolic response. Global MGU was higher in patients with normal OGTT response (0.31 ± 0.09 µmol/g/min) compared with patients with abnormal OGTT response (0.25 ± 0.09 µmol/g/min) (P = 0.05). MBF (P = 0.22) and myocardial flow reserve (MFR) (P = 0.83) were similar in the study groups. The reduced MGU in prediabetic patients was attributable to reduced MGU in viable myocardium with normal MFR (P < 0.001).

CONCLUSION

CHF patients with prediabetes have reduced MGU in segments with preserved MFR as compared to CHF patients with normal glucose tolerance. Whether reversal of these myocardial abnormalities can improve outcome needs to be investigated in large-scale studies.

Prognostic Impact of Diabetes on Long-term Survival Outcomes in Patients With Heart Failure: A Meta-analysis

OBJECTIVE

Several studies have explored the impact of diabetes on mortality in patients with heart failure (HF). However, the extent to which diabetes may confer risk of mortality and hospitalization in this patient population remains imperfectly known. Here we examine the independent prognostic impact of diabetes on the long-term risk of mortality and hospitalization in patients with HF.

RESEARCH DESIGN AND METHODS

PubMed, Scopus, and Web of Science from January 1990 to October 2016 were the data sources used. We included large (n ≥1,000) observational registries and randomized controlled trials with a follow-up duration of at least 1 year. Eligible studies were selected according to predefined keywords and clinical outcomes. Data from selected studies were extracted, and meta-analysis was performed using random-effects modeling.

RESULTS

A total of 31 registries and 12 clinical trials with 381,725 patients with acute and chronic HF and 102,036 all-cause deaths over a median follow-up of 3 years were included in the final analysis. Diabetes was associated with a higher risk of all-cause death (random-effects hazard ratio [HR] 1.28 [95% CI 1.21, 1.35]), cardiovascular death (1.34 [1.20, 1.49]), hospitalization (1.35 [1.20, 1.50]), and the combined end point of all-cause death or hospitalization (1.41 [1.29, 1.53]). The impact of diabetes on mortality and hospitalization was greater in patients with chronic HF than in those with acute HF. Limitations included high heterogeneity and varying degrees of confounder adjustment across individual studies.

CONCLUSIONS

This updated meta-analysis shows that the presence of diabetes per se adversely affects long-term survival and risk of hospitalization in patients with acute and chronic HF.

Prognostic Impact of Diabetes Mellitus in Chronic Heart Failure According to Presence of Ischemic Heart Disease – With Special Reference to Nephropathy

BACKGROUND

It is unclear whether the prognostic impact of diabetes mellitus (DM) in chronic heart failure (CHF) is influenced by ischemic heart disease (IHD) and/or nephropathy.

METHODS AND RESULTS

We enrolled 4,065 consecutive patients with stage C/D CHF (mean age, 69.0 years; 68.7% male) in the CHART-2 Study (n=10,219). We defined DM as current history of DM treatment or HbA1c ≥6.5% (National Glycohemoglobin Standardization Program [NGSP]), and nephropathy as urine albumin:creatinine ratio ≥30 mg/g or urine dipstick test ≥(±) at enrollment. Impacts of DM and nephropathy on the composite of death, myocardial infarction, stroke, and HF admission were examined. Among the 4,065 patients, 1,448 (35.6%) had DM, while IHD and nephropathy were also noted in 1,644 (40.4%) and in 1,549 (38.1%), respectively. During the median follow-up of 2.88 years, 1,025 (25.2%) reached the composite endpoint. On multivariate Cox regression, DM was significantly associated with the composite endpoint in all patients (HR, 1.17; P=0.02), and in those with IHD (HR, 1.38; P=0.004), but not in those without IHD (HR, 1.12; P=0.22; P for interaction=0.12). Furthermore, when the patients were stratified by nephropathy, DM was associated with worse prognosis only in the IHD patients with nephropathy.

CONCLUSIONS

The prognostic impact of DM was more evident in patients with IHD than in those without IHD, particularly when complicated with nephropathy.

Association of heart failure severity with risk of diabetes: a Danish nationwide cohort study

AIMS/HYPOTHESIS

Heart failure has been suggested to increase the risk of developing diabetes. We investigated the relation between heart failure severity, defined by loop-diuretic dosage, and the risk of developing diabetes in a nationwide cohort of patients with heart failure.

METHODS

We followed all Danish patients discharged from hospitalisation for first-time heart failure in 1997-2010, without prior use of hypoglycaemic agents, until a claimed prescription for hypoglycaemic agents, death or 31 December 2010. The association of loop-diuretic dosage (furosemide equivalents) 90 days after discharge (study baseline) with risk of diabetes was estimated by multivariate Cox regression models.

RESULTS

In total, 99,362 patients were included and divided into five loop-diuretic dose groups: 30,838 (31%) used no loop diuretics; 24,389 (25%) used >0-40 mg/day; 17,355 (17%) used >40-80 mg/day; 11,973 (12%) used >80-159 mg/day; and 14,807 (15%) used ≥160 mg/day. A total of 7,958 patients (8%) developed diabetes. Loop-diuretic dosages were associated with an increased risk of developing diabetes in a dose-dependent manner. Concomitant use of renin-angiotensin system inhibitors (RASis) attenuated the risk (p value for interaction <0.0001). Compared with patients using no loop diuretics (group 1), the adjusted HRs (95% CI) for developing diabetes for groups 2-5 respectively were 1.16 (1.07, 1.26), 1.35 (1.24, 1.46), 1.48 (1.35, 1.62) and 1.76 (1.61, 1.92) with RASi treatment, and 2.06 (1.83, 2.32), 2.28 (2.01, 2.59), 2.88 (2.52, 3.30) and 3.02 (2.66, 3.43) without RASi treatment.

CONCLUSIONS/INTERPRETATION

In a nationwide cohort of patients with heart failure, severity of heart failure was associated with a stepwise increased risk of developing diabetes. Increased awareness of risk of diabetes associated with severe heart failure is warranted.

Association between type-2 diabetes mellitus and post-discharge outcomes in heart failure patients: findings from the RICA registry

AIMS

Heart failure (HF) and diabetes are common clinical conditions that may coexist. The main objective was to analyze the association of type-2 diabetes mellitus (T2DM) on prognosis in hospitalized patients with HF.

METHODS

We evaluated the association between T2DM with all-cause mortality and readmissions in the Spanish National Registry on Heart Failure-"Registro Nacional de Insuficiencia Cardiaca" (RICA). This is a multi-center and prospective cohort study that includes patients admitted for decompensated HF from 2008 to 2011. Study endpoints were all-cause mortality and hospital readmission. We determined the adjusted hazard ratio (HR) by a multivariable Cox regression model.

RESULTS

A total of 1082 patients (mean age 77.6±8.5) were included of whom 490 (45.3%) had diabetes and 592 patients (54.7%) had preserved left ventricular ejection fraction (LVEF). During one-year follow-up, 287 patients died (151 patients with diabetes) and 383 patients were readmitted (197 patients with diabetes). After adjusting for baseline characteristics T2DM was significantly associated with all-cause mortality (HR 1.54; 95%CI 1.20-1.97, p=0.001) and readmissions (HR 1.46; 95%CI 1.18-1.80, p<0.001). Age, dementia, peripheral vascular disease, NYHA class, renal insufficiency, hyponatremia and anemia were also independently associated with outcomes. There were no differences in mortality (p=0.415) and readmissions (p=0.514) according to preserved or reduced LVEF.

CONCLUSION

T2DM is very common in patients hospitalized for HF. This condition is a strong and independent co-morbidity of all-cause mortality and readmission for both HF with preserved and reduced LVEF.

The prognostic importance of a history of hypertension in patients with symptomatic heart failure is substantially worsened by a short mitral inflow deceleration time

BACKGROUND

Hypertension is a common comorbidity in patients with heart failure and may contribute to development and course of disease, but the importance of a history of hypertension in patients with prevalent heart failure remains uncertain.

METHODS

3078 consecutively hospitalized heart failure patients (NYHA classes II-IV) were screened for the EchoCardiography and Heart Outcome Study (ECHOS). The left ventricular ejection fraction (LVEF) was estimated by 2 dimensional transthoracic echocardiography in all patients and a subgroup of 878 patients had additional data on pulsed wave Doppler assessment of transmitral flow available. A restrictive filling (RF) was defined as a mitral inflow deceleration time ≤140 ms. Patients were followed for a median of 6.8 (Inter Quartile Range 6.6-7.0) years and multivariable Cox regression models were used to assess the risk of all-cause mortality associated with hypertension.

RESULTS

The study population had a mean age of 73 ± 11 years. 39% were female, 27% had a history of hypertension and 48% had a RF. Over the study period, 64% of the population died. Hypertension was not associated with increased risk of mortality, hazard ratio (HR) 0.95 (0.85-1.05). LVEF did not modify this relationship (p for interaction = 0.7), but RF pattern substantially influenced the outcomes associated with hypertension (p for interaction < 0.001); HR 0.75 (0.57-0.99) and 1.41 (1.08-1.84) in patients without and with RF, respectively.

CONCLUSIONS

In patients with symptomatic heart failure, a history of hypertension is associated with a substantially increased relative risk of mortality among patients with a restrictive transmitral filling pattern.

Diabetes reduces left ventricular ejection fraction--irrespective of presence and extent of coronary artery disease

BACKGROUND

It is not clear whether diabetes reduces systolic left ventricular function (left ventricular ejection fraction, LVEF) irrespective of coronary artery disease (CAD). The aim of this study was to compare the LVEF between diabetic and non-diabetic patients with respect to the extent of CAD.

METHODS AND RESULTS

Consecutive patients undergoing stress myocardial perfusion SPECT (MPS) were evaluated. MPS was interpreted using a 20-segment model with a five-point scale to define summed stress score (SSS), summed rest score, and summed difference score. LVEF was measured by gated SPECT and then compared with respect to diabetic status and SSS categories. Of 2635 patients, data of 2400 was available. Of these, 24% were diabetic, mean age was 64±11y, and 31% were female. Diabetics had a significantly lower LVEF compared with non-diabetics regardless of the extent of CAD: 53±13 and 55±13% respectively (P=0.001). Diabetics and non-diabetics did not differ significantly in the distribution of SSS categories. Diabetes was an independent predictor of decreased LVEF (odds ratio 1.6, 95% confidence interval 1.2-2.0; P<0.001).

CONCLUSION

Diabetics had a lower LVEF than non-diabetics. This difference could be demonstrated regardless of CAD extent and might in part explain their generally worse cardiac survival compared with non-diabetics on an epidemiological level. In addition, this finding points to discussed mechanisms other than CAD lowering LVEF in diabetics.

Oral glucose tolerance testing in an outpatient heart failure clinic reveals a high proportion of undiagnosed diabetic patients with an adverse prognosis

AIMS

We evaluated the applicability and prognostic importance of oral glucose tolerance testing (OGTT) among outpatients with systolic heart failure (SHF).

METHODS AND RESULTS

Consecutive patients with SHF and left ventricular ejection fraction (LVEF) ≤ 45% referred to a heart failure clinic (n= 413) were included in this study. An OGTT was conducted in patients without a history of diabetes. Information on NYHA class, aetiology of SHF, LVEF, treatment, and biochemical parameters were collected at baseline. The survival status was obtained after a median follow-up time of 591 days. Of the 413 patients, 82 (20%) had known diabetes. Of the remaining 331 patients, 227 (69%) agreed to undergo an OGTT. Among the tested subjects, 136 (60%) were classified as having normal glucose tolerance (NGT), 51 (23%) impaired glucose tolerance (IGT), and 40 (18%) newly diagnosed diabetes. Assuming a similar prevalence of unrecognized diabetes among the patients who refused OGTT, the prevalence of diabetes in the total population was 34%. If only fasting blood glucose had been used, 16 of the 40 newly diagnosed diabetic patients would have been undiagnosed. During follow-up, 24 (29%) patients with known diabetes, 6 (15%) of the newly diagnosed diabetic patients, 9 (18%) of those with IGT, and 13 (9%) patients with NGT died. Patients with diabetes had higher mortality rate compared with non-diabetic patients [multivariate hazard ratio 1.89 (1.02-3.59); P = 0.047].

CONCLUSION

It is feasible to perform diabetes screening using OGTT in outpatients with SHF. A substantial proportion of patients tested were found to have unrecognized diabetes. The presence of diabetes was associated with a higher mortality rate.

Metformin treatment is associated with a low risk of mortality in diabetic patients with heart failure: a retrospective nationwide cohort study

AIMS/HYPOTHESIS

The safety of metformin in heart failure has been questioned because of a perceived risk of life-threatening lactic acidosis, though recent studies have not supported this concern. We investigated the risk of all-cause mortality associated with individual glucose-lowering treatment regimens used in current clinical practice in Denmark.

METHODS

All patients aged ≥ 30 years hospitalised for the first time for heart failure in 1997-2006 were identified and followed until the end of 2006. Patients who received treatment with metformin, a sulfonylurea and/or insulin were included and assigned to mono-, bi- or triple therapy groups. Multivariable Cox proportional hazard regression models were used to assess the risk of all-cause mortality.

RESULTS

A total of 10,920 patients were included. The median observational time was 844 days (interquartile range 365-1,395 days). In total, 6,187 (57%) patients died. With sulfonylurea monotherapy used as the reference, adjusted hazard ratios for all-cause mortality associated with the different treatment groups were as follows: metformin 0.85 (95% CI 0.75-0.98, p = 0.02), metformin + sulfonylurea 0.89 (95% CI 0.82-0.96, p = 0.003), metformin + insulin 0.96 (95% CI 0.82-1.13, p = 0.6), metformin + insulin + sulfonylurea 0.94 (95% CI 0.77-1.15, p = 0.5), sulfonylurea + insulin 0.97 (95% CI 0.86-1.08, p = 0.5) and insulin 1.14 (95% CI 1.06-1.20, p = 0.0001).

CONCLUSIONS/INTERPRETATION

Treatment with metformin is associated with a low risk of mortality in diabetic patients with heart failure compared with treatment with a sulfonylurea or insulin.