Abstract

Background
People with diabetes have an increased risk of heart failure (HF), compared to those without diabetes. However, no comprehensive systematic review and meta-analysis has explored whether these associations could differ in relation to prevalent cardiovascular disease (CVD).

Aims
To estimate the association between diabetes and incident heart failure (HF), compared to without diabetes, in individuals with and without CVD.

Methods
PubMed, Scopus, and Web of Science were searched for observational cohort studies from the earliest dates to 22nd March 2023. A random-effects model calculated the pooled relative risk (RR).

Results
Of 11,609 articles, 31 and 6 studies reported data in people with type 2 diabetes (T2D) and type 1 diabetes (T1D) respectively. Individuals with T2D had an increased risk of HF irrespective of CVD prevalence: 1.61 (95% CI: 1.35–۱٫۹۲) in those with CVD; 1.78 (1.60–۱٫۹۹) without CVD; and 2.02 (1.75–۲٫۳۳) with unspecified CVD prevalence. Meta-regression did not identify a significant difference comparing HF risk in T2D individuals with vs. without CVD (p = 0.232).

Conclusion
People with T2D, compared to those without diabetes, have similar increased risk of HF, regardless of CVD prevalence. Strategies proven to lower HF risk in T2D individuals should be prioritized for those with and without CVD.

Introduction

Diabetes mellitus and heart failure (HF) tend to exist as common comorbid conditions [1], with approximately 4.3 to 28 % of HF individuals with prevalent type 2 diabetes (T2D) and 12–۵۷ % of T2D individuals presenting with HF [2]. People with diabetes also have an increased risk of HF, which subsequently increases their risk of hospitalizations [3], [4] and mortality [5], [6] to further burden global economic and healthcare resources [7]. Nearly 537 million adults are currently living with diabetes world-wide, estimated to rise to 643 million by 2030 [8]; HF currently affects 64.3 million people globally [9], predicted to rise by 50 % in the next 20 years [10]. Hence, early prevention and management strategies to lower HF risk in individuals with diabetes is of public health importance [11], [12]. This is further reinforced in evidence showing HF as the most common initial presentation of cardiovascular disease (CVD) in people with diabetes [13].

A previous meta-analysis from 2018 [14] showed that persons with diabetes have an increased risk of HF, relative to without diabetes (relative risk, RR: 2.06, 95 % confidence interval, CI: 1.73 to 2.46). However, this study pooled estimates from heterogeneous study designs (e.g., case-control and cohort), which could have led to biased estimates. To address this limitation, a recent 2020 meta-analysis [15], exclusively using data from observational cohort studies, reported that individuals with diabetes have an increased risk of new-onset HF, relative to without diabetes (RR: 2.14; 95 % CI: 1.96 to 2.34). However, the author’s included studies that did not adjust for at minimum age and sex, which could have led to biased results. The study also found an attenuated association when restricting the population to individuals with prevalent coronary heart disease (CHD) (RR: 1.94, 95 % CI: 1.77 to 2.12) [15]. It remains unclear whether, and to what extent, the association between diabetes and incident HF could differ in populations with and without broad CVD.

Results

۳٫۱٫ Study characteristics
Of the 11,609 studies, 31 articles [6,13,22–۵۰] consisting of 34 cohorts for T2D were identified (prevalent CVD: 11; without CVD: 12; unspecified CVD prevalence: 11) and 6 articles [6,24,27,29,46,51] consisting of 7 cohorts were identified for T1D (prevalent CVD: 1; without CVD: 3; unspecified CVD prevalence: 3) (Fig. 1). The total number of individuals were 501,536 for cohorts of T2D with CVD, 27,855,580 for T2D without CVD, 9,046,486 for T2D with unspecified CVD prevalence; 251,538 for T1D with CVD, 18,580,888 for T1D without CVD, and 3,690,911 for T1D unspecified CVD prevalence (Table S3). Of studies investigating T2D, mean age was 57.99, 61.25 and 64.74 in those with CVD, without CVD and unspecified CVD prevalence respectively. Mean baseline recruitment (years) was higher in those with prevalent CVD (2007), compared to without CVD (2004) and unspecified CVD (2001). Most studies were from Europe at 45.45 %, 50.00 % and 54.55 % in those with CVD, without CVD and unspecified CVD prevalence. In T2D, most studies had low risk of bias, with 90.91 %, 91.67 % and 72.73 % reported in populations with CVD, without CVD, and unspecified CVD prevalence. Table S3 summarises the characteristics of all included studies.