Comparative analysis of infective endocarditis in hemodialysis versus non-hemodialysis patients in Iran: implications for clinical practice and future research

Male patients comprised the majority of our study population in both HD and non-HD groups. Gender differences are reported heterogeneously in other studies. Pericas et al. and Kwon et al. found predominantly female hemodialysis IE patients, whereas Bentata et al., Stahl et al., and Zhang et al. noted mostly males in both groups [19, 20, 23‐25]. This may be due to regional differences in the cohorts.

In the present study, the mean ages of HD and non-HD patients were not significantly different, which is in line with studies by Bentata et al., Kwon et al., Hsiao et al., and Zhang et al., who have not noted any statistically significant age difference between the groups [20, 23, 24, 26]. In contrast, Stahl et al. and Bhatia et al. reported significantly lower mean ages in the hemodialysis group [8, 25]. This diversity is likely due to the smaller number of older hemodialysis patients in their cohorts. However, our numbers are noticeably lower than those reported in a review of other studies on the epidemiological features of IE. The disparity mentioned earlier can be an important issue for future research.

Our results highlight chronic IE as a potential concern in the HD population. Overall, the IE onset types indicate that dialysis status does not overtly alter the typical acute presentation but may confer a higher risk for indolent chronic infection.

In this study, the prevalence rates of diabetes mellitus, hypertension, and congestive heart failure were significantly higher in the HD group than in the non-HD group. This finding is in accordance with prior epidemiological observations by Stahl et al., who reported diabetes mellitus, ischemic heart disease, and hypertension as significantly prevalent comorbidities in HD patients [25]. Our observations did not show any meaningful differences in the prevalence of cerebrovascular disease, intravenous drug use, history of cardiac valve surgeries, or immunosuppressant use between the study groups, contrary to some previous studies. For instance, Pericas et al. found that peripheral vascular disease and cerebrovascular disease were notably higher in the HD-IE group. However, intravenous drug use and congenital heart disease were more frequently found in the non-HD-IE group [19]. Stahl et al. reported that prosthetic heart valves prior to IE were significantly more common in the non-HD group [25]. Zhang et al. and Wang et al. consistently reported that predisposing cardiac anomalies were more frequently observed in the non-HD-IE group than in the HD-IE group [20, 27]. Accordingly, Bhatia et al. found a lower prevalence of most IE risk factors, including congenital heart disease, rheumatic heart disease, valvular heart disease, prior valve replacement, and history of drug abuse in dialysis patients [8].

In contrast to previous studies that reported rheumatic valve diseases as a notable risk factor for IE, only 4% of our studied patients suffered from such diseases, with no statistically significant difference between the HD and non-HD groups [28, 29]. Our study’s low percentage of rheumatic valve diseases could be because we considered only echocardiographic findings and did not analyze pathological findings. Furthermore, valve destruction caused by IE might mask rheumatic changes and render them undetectable by echocardiography.

The most common congenital heart defect in the HD group was BAV. While the prevalence of BAV in the general population is 1–2%, it was approximately 11.7% in our HD group [30]. Zegri-Reiriz et al. and Zhu et al. reported a higher risk of IE in patients with BAV and recommended antibiotic prophylaxis [31, 32].

Negative blood cultures occurred in 50% of HD patients and 54.3% of non-HD patients. Similarly, Bentata et al. and Zhang et al. reported negative blood cultures in nearly half of their study population [20, 23]. These results differ from Stahl’s 2023 study, which reported 7.1% and 19.4% negative blood cultures in HD and non-HD patients with IE, respectively [25]. In the studies by Wang et al. and Hsiao et al., the reported numbers for positive cultures were over 70%. This discrepancy could be partially attributable to referral bias because, patients are often transferred to our tertiary center after antibiotic initiation [26, 27].

The current study found that coagulase-negative staphylococci were the most common pathogen, with Staphylococcus aureus being the second most common pathogen in the HD group. In the non-HD group, Staphylococcus aureus and enterococci were the most common and second-most common pathogens, respectively. The incidence of methicillin-resistant Staphylococcus aureus (MRSA) was observed to be similar across the various cohorts. On the contrary, almost all previous studies have documented Staphylococcus aureus as the prevailing causative pathogen responsible for endocarditis in individuals undergoing dialysis [8, 12, 18, 33‐35]. Pericas et al. reported that the viridans group streptococci and Streptococcus bovis were found to be more commonly observed among non-hemodialysis patients, which is quite unexpected [19]. The epidemiological variability of infective endocarditis (IE) and the lack of research in Iran necessitate a comprehensive examination of antibiotic coverage for coagulase-negative staphylococci. Our HD cohort comprised of a limited patient population. Consequently, conducting future prospective investigations is warranted to assess microbiological characteristics more broadly.

In accordance with the study by Zhang et al.’s, our investigations showed that the serum hemoglobin level was significantly lower in the HD group than in the non-HD group [20]. In addition, white blood cell (WBC), erythrocyte sedimentation rate (ESR), and C-reactive protein (CRP) levels were not significantly different between our studied groups. In contrast to our finding, investigations by Hsiao et al. showed that while CRP levels did not significantly differ between the hemodialysis and non-hemodialysis groups, ESR was markedly higher in the hemodialysis patients than in the non-HD patients [26].

This study showed that the mitral valve is the most frequently involved cardiac valve, followed by the aortic valve, in all patients with IE. The aforementioned dominance remains unchanged in both HD and non-HD patients. This finding is backed up by pieces of evidence from most previous observations by Rekik et al., Nori et al., Zhang et al., Doulton et al., and Gülmez et al., which could be due to the higher susceptibility of the mitral valve to fluid overload and calcification [11, 12, 20, 34, 35]. In addition, our investigations revealed that right-sided heart valve vegetation, especially tricuspid valve vegetation, was significantly more frequent in the HD group than in the non-HD group. However, this result has not been previously described in similar investigations. Therefore, further studies are required to confirm and validate this finding. Simultaneous left and right-sided valve vegetations were more common in the HD patients, suggesting the aggressive nature of the disease in this group of patients. In addition, the mean vegetation size was significantly larger in HD versus non-HD patients (18.6 ± 11.8 mm vs. 12.1 ± 8.7 mm).

Studies by Kwon et al., Zhang et al., and Durante-Mangoni et al. could not demonstrate any significant differences in terms of vegetation of size and location between HD and non-HD patients [20, 24, 36].

Similar to Rekik et al., we found a higher frequency of extracardiac embolism in the HD group [12]. Based on our results, only pulmonary artery embolism was significantly more common in the HD group. The fact that tricuspid valve vegetation was more frequent in the HD group may have caused this cohort’s higher prevalence of septic embolism in the pulmonary arteries. Nonetheless, in a recent study, Wang et al. reported no significant difference between HD and non-HD groups regarding embolic and neurological complications [27].

In our study, valve surgeries were performed on 52.94% of the HD patients, similar to the non-HD group (57.14%). In line with several previous investigations by Durante-Mangoni et al., Hsiao et al., and Kwon et al., there was no significant difference in the surgical requirements between the studied groups [24, 26, 36]. However, other studies by Stahl et al., Pericas et al., and Zhang et al. reported cardiac surgery being performed in a significantly lower number of HD patients than in non-HD patients [19, 20, 25]. Our study population was selected among patients referred to a tertiary cardiovascular center who may have required special treatment or been in a worse condition than those at primary or secondary centers. Hence, such differences between our findings and those of other studies are expected and can help understand all aspects of IE in different clinical settings.

Consistent with most prior studies, in-hospital mortality was significantly higher in the HD group than in the non-HD group [5, 19, 20, 24, 27]. However, Stahl et al. and Durante-Mangoni et al. reported no significant differences between the two groups [25, 36]. In addition, further analysis revealed that the rate of in-hospital mortality was significantly higher in surgically treated HD patients compared to surgically treated non-HD patients. This finding is in line with the result of a study by Raza et al. who reported 13% hospital mortality for HD patients versus 5% for non-HD patients following surgical intervention [37].

Moreover, within HD patients, in-hospital mortality rates were similar for medically versus surgically managed patients (43.8% vs. 38.9%). An investigation by Jones et al. reported numerically lower but statistically indistinguishable in-hospital mortality in HD patients with surgery (11.1% vs. 15.2%) [18]. A recent meta-analysis by Ting et al. examining surgery versus medical therapy specifically in HD infective endocarditis patients also found no difference in the survival outcomes between these interventions [38].

The mean length of hospital stay for HD patients was lower than that for non-HD patients. This is probably because of the higher mortality rate in the former group.

This study has several limitations worth acknowledging. First, the retrospective observational design based on registry data restricts our ability to infer causality between hemodialysis status and clinical outcomes. Second, our cohort was restricted to patients admitted to a single tertiary academic medical center. Results may not be generalizable to other healthcare settings or geographic regions. Third, longer-term follow-up data was not available in the Iranian Registry of Infective Endocarditis (IRIE), including details on readmissions, post-discharge mortality, other long-term outcomes, and specifics on surgical interventions performed. Thus, our data and analyses were limited by the constraints of the registry. Fourth, the sample size, while determined to be adequately powered for the primary outcome examined, remains relatively small overall and in the hemodialysis subgroup specifically. Larger multi-center prospective studies are warranted to validate findings.