Infective Endocarditis in Diabetic Patients: A Different Profile with Prognostic Consequences

Background. Infective Endocarditis (IE) is a severe condition. Diabetes mellitus (DM) has been associated with a poor prognosis in other settings. Our aim was to describe the profile and prognosis of IE with and without DM and to analyze the prognostic relevance of DM-related organ damage. Methods. Retrospective analysis of the Spanish IE Registry (2008–2020). Results. The cohort comprises 5590 IE patients with a mean age of 65.0 ± 15.5 years; 3764 (67.3%) were male. DM was found in 1625 patients (29.1%) and 515 presented DM-related organ damage. DM prevalence during the first half of the study period was 27.6% vs. 30.6% in the last half, p = 0.015. Patients with DM presented higher in-hospital mortality than those without DM (521 [32.1%] vs. 924 [23.3%], p < 0.001) and higher one-year mortality (640 [39.4%] vs. 1131 [28.5%], p < 0.001). Among DM patients, organ damage was associated with higher in-hospital (200 [38.8%] vs. 321 [28.9%], p < 0.001) and one-year mortality (247 [48.0%] vs. 393 [35.4%], p < 0.001). Multivariate analyses showed an independent association of DM with in-hospital (odds ratio [OR] = 1.34, 95% confidence interval [CI]: 1.16–1.55, p < 0.001) and one-year mortality (OR = 1.38, 95% CI: 1.21–1.59, p < 0.001). Among DM patients, organ damage was independently associated with higher in-hospital (OR = 1.37, 95% CI: 1.06–1.76, p = 0.015) and one-year mortality (OR = 1.59, 95% CI = 1.26–2.01, p < 0.001) Conclusions. The prevalence of DM among patients with IE is increasing and is already above 30%. DM is independently associated with a poor prognosis, particularly in the case of DM with organ damage.


Introduction
Infective Endocarditis (IE) is a severe disease with high in-hospital mortality [1,2]. Almost one third of IE patients present diabetes mellitus (DM) [2][3][4][5]. DM has been associated with a poor prognosis in sepsis [6][7][8]. An association of DM with prognosis in IE patients has also been described [1,4]. However, DM is associated with advanced age, comorbidities, atypical clinical presentation, and longer IE diagnosis time, among other characteristics that have a strong prognostic influence in IE [2][3][4][5]. Due to that reason, the independent association of DM with IE mortality is unclear [3,4]. Some authors have suggested an independent association [1] while other data do not support it [3].
Our aim was to describe the profile and prognosis of IE with and without DM and to analyze the prognostic relevance of DM-related organ damage. We also studied the evolution of the yearly prevalence of DM in these patients.

Methods
The Spanish Collaboration on Endocarditis-GAMES (Grupo de Apoyo al Manejo de la Endocarditis Infecciosa en España)-is a national observational registry that has been previously described [9][10][11]. Multidisciplinary teams that compose this group, including infectious disease physicians, cardiologists, cardiac surgeons, microbiologists, echocardiographers, and other imaging specialists, prospectively completed standardized case report forms with information regarding IE episodes and follow-up data. A complete list of GAMES members is shown in Acknowledgments. IE patients were consecutively included at 38 Spanish hospitals between January 2008 and December 2020. Inclusion criteria were the diagnosis of definite or probable IE by modified Duke criteria [12]. IE management, including the decision to perform surgery and the type of surgery, was done by the local medical team following the 2009 and 2015 European Society of Cardiology recommendations [13]. DM was diagnosed based on the American Diabetes Association criteria [14]. DM-associated organ damage was considered to be present after analyzing clinical and laboratory techniques, as well as image data. For instance, renal disease with albuminuria and/or reduced glomerular filtration rate in the absence of signs or symptoms of other primary causes of kidney damage, neuropathy with loss of protective sensation, and neovascularization and/or vitreous/preretinal bleeding (in addition to non-proliferative retinopathy) [15].
This study complies with the principles outlined in the Declaration of Helsinki and was approved by the ethics committee of participating centers.

Statistical Methods
Continuous variables are summarized as means ± standard deviations (SD) or medians, and interquartile ranges, when a normal distribution was not observed, as per the Kolmogorov-Smirnov goodness-of-fit test; categorical variables are expressed as numbers and percentages. Student's t-test, Mann-Whitney U test, or paired t-test were used to compare continuous variables. Categorical variables were compared using the χ 2 test or Fisher's exact test. Multivariable logistic regression analyses (backward selection) were performed to determine mortality predictors and to assess the independent association of DM, with and without organ damage, with mortality. All variables with p value < 0.10 in univariate analyses were included in the multivariable analyses. Statistical analysis was performed using SPSS, version 22.0 (IBM, Armonk, NY, USA).

Results
The cohort comprises 5590 IE patients with a mean age of 65.0 ± 15.5 years; 3764 (67.3%) were male. DM was found in 1625 patients (29.1%) and 515 presented DM-related organ damage (Figure 1).    Table 1 shows the clinical characteristics of patients with and without DM. Compared with those without DM, DM patients presented more frequently advanced age, cardiac implantable electronic device location, nosocomial origin, cardiovascular and renal disease, and had a higher Charlson Comorbidity index. Mean age on the first 6.5 years of the study period was lower than in the last 6.       Table 1 shows the clinical characteristics of patients with and without DM. Compa with those without DM, DM patients presented more frequently advanced age, card implantable electronic device location, nosocomial origin, cardiovascular and renal ease, and had a higher Charlson Comorbidity index. Mean age on the first 6.5 years of study period was lower than in the last 6.5 years, both in the global population and diabetic patients (65.0 ± 15.6 years vs. 66.2 ± 15.3 years, p = 0.004; 69.3 ± 11.2 vs. 71.0 ± 1 p = 0.003, respectively).  Table 1 shows the clinical characteristics of patients with and without DM. Compared with those without DM, DM patients presented more frequently advanced age, cardiac implantable electronic device location, nosocomial origin, cardiovascular and renal disease, and had a higher Charlson Comorbidity index. Mean age on the first 6.5 years of the study period was lower than in the last 6.5 years, both in the global population and in diabetic patients (65.0 ± 15.6 years vs. 66.2 ± 15.3 years, p = 0.004; 69.3 ± 11.2 vs. 71.0 ± 10.5, p = 0.003, respectively). Staphylococcus and enterococcus etiology were more common among diabetics ( Table 2). Table 3 shows the clinical outcome according to the presence of diabetes. Compared with those without DM, DM patients presented complications more frequently and had higher in-hospital and one-year mortality. Among diabetics, patients with DM-related organ damage were a high-risk population with a poor prognosis (Table 4 and Figure 3). Multivariate analyses showed an independent association of DM with in-hospital and oneyear mortality (Table 5). In addition, among diabetics, organ damage was an independent predictor of mortality.

Discussion
Our data show that the prevalence of DM among patients with IE is increasing and about 30% of patients with IE have DM. Diabetics had a poor prognosis, particularly in the case of DM with organ damage. Compared with non-diabetics, diabetic patients had comorbidities more frequently, mainly cardiovascular and renal disease. Diabetics also had a high-risk profile with more nosocomial and healthcare-related IE and more frequent S. aureus etiology. As expected, diabetics had a poor prognosis. Even after correcting for confounding factors, the association of DM with in-hospital and one-year mortality remained significant.
The prevalence of DM in the general population is increasing [16]. In our country the prevalence of DM in the general population aged 65-75 years increased from 17% in 2006 to 21% in 2017 [17]. Our study also found a similar trend in IE patients. Previous authors have suggested an increase in DM prevalence in IE patients [3,5]. Abe et al. [5] found a prevalence of 22% in 2004 that increased to 30% in 2014. The reasons that explain the increase in the prevalence of DM are unknown. Population aging may play a role. In our sample, the mean age on the last half of the study period was higher than in the first half, and this was also true in diabetic patients.
Although some previous studies suggested an association of DM with IE mortality [1,4], others did not identify DM as a prognostic factor [18][19][20][21][22][23][24]. The independent association of DM with prognosis is unclear and a significant age interaction could be a confounding factor [4]. On the other hand, even prediabetes has been associated with a higher mortality risk [25,26]. Although no previous studies have focused on the prognostic implications of DM-related organ damage, DM has been associated with higher rates of heart and renal failures [26] and more advanced DM stages, such as longer DM duration, insulin-treated DM [25], and higher Diabetes Complications Severity Index [27], which have been related to higher IE risk and a poor prognosis.
The relation of DM with the prognosis of IE might have several explanations. DM is associated with endothelial dysfunction which can promote stronger bacterial adhesion [28,29]. In addition, diabetics have an impaired immune response [30] and more common bacteremia with aggressive bacteria such as S. aureus [31]. Moreover, diabetics have defects of neutrophil activities [32][33][34]. Immune system dysfunction due to chronic low-grade inflammation seen in DM favors micro-organism growth, a process that contributes to sepsis progression [6][7][8].
Our work could have relevant clinical implications. Due to the poor prognosis of IE in diabetics, it might be reasonable to consider earlier and more aggressive treatments and interventions in these patients, particularly in those with previous DM-related organ damage. Close follow-up and correct glycaemia control might improve the outcome.
The limitations of this study should be noted. The retrospective design justifies that relevant variables such as type of DM, DM duration, level of glycated hemoglobin, presence of diabetic cardiomyopathy, and DM-therapy were not collected systematically. Local medical teams were responsible for IE management, including deciding on surgery, and any judgements may have been influenced by factors not registered in this study. Finally, cause of death during follow-up was not available for a large number of patients.
In any case, our data come from a large national database and show a clear association of DM with IE prognosis. Moreover, ours is the first study to compare the prognosis of IE in diabetics with and without organ damage.

Conclusions
The prevalence of DM among patients with IE is increasing and is already above 30%. DM is independently associated with a poor prognosis, particularly in the case of DM with organ damage.