For patients undergoing renal replacement therapy, infection is the leading cause of morbidity and the second leading cause of mortality [1]. In Brazil, it is estimated that 97,586 patients currently receive chronic dialysis treatment [26]. According to North American data, the rates of BSI among hemodialysis patients range between 0.5 and 27.1 per 100 patients/month depending on the type of venous access used [2]. Results of a Brazilian study reported that rates varied between 0.55 and 7.32 per 1000 venous accesses/day [27].
In the present study, risk factors for BSI were evaluated in patients on hemodialysis at a hemodialysis service in São Paulo, Brazil. The study was a case–control design, which can analyze risk factors related to a determined condition. Patients presenting with BSI during the study period were included as cases, and controls were those patients who did not develop the studied event (pairing criteria: age, previous time on hemodialysis, and diagnosis of diabetes mellitus).
Use of a CVC compared with use of an AVF, and previous hospitalization were independent risk factors for the occurrence of BSI among patients on hemodialysis treatment. Use of a CVC was associated with an increased risk of developing BSI (OR: 11.2; CI 95%: 5.17–24.29; p < 0.001) compared with the use of an AVF. This finding aligns with previous trials of antimicrobial locks, which reported a higher occurrence of BSI among hemodialysis patients with CVC (3-4 cases per 1000 catheter-days) [28]. In a longitudinal cohort study, Xui et al. found that the rates of BSI in patients using CVC were three times higher than in patients using AVF (p < 0.001) [8]. A case–control study conducted at a university hospital in Greece demonstrated, through multivariate logistic regression, that patients using CVC had a higher risk of BSI compared with patients using AVF (OR: 2.93; p = 0.047) [10].
In the current study, only five patients (four cases and one control) used a non-tunneled hemodialysis CVC. In these patients, three had started hemodialysis within 60 days and were awaiting the manufacture of an AVF, and the other two had been undergoing treatment for a longer period of time and had spent longer with a non-tunneled hemodialysis CVC because of difficulties in the insertion of a tunneled hemodialysis CVC or the manufacture of an AVF.
Previous hospitalization increased the risk of BSI 5.33-fold (p = 0.003). There are few reports in the literature on the association between BSI and previous hospitalizations in patients undergoing hemodialysis treatment. Nguyen et al. in a study performed between 2005 and 2011, evaluated BSI from MRSA in patients on dialysis treatment, and found that 70% of patients who developed BSI were hospitalized at least once in the year prior to the infection [6]. In a study by Barbosa et al. the researchers reported that, in 320 chronic renal patients, previous hospitalization was an important risk factor for colonization of multi-resistant microorganisms (e.g., VRE) [29]. And Aktas et al. in a study of 70 patients on dialysis treatment, reported a higher colonization for MRSA in patients who had been hospitalized within the previous 6 months compared with those who had not [30].
In univariate logistic regression analysis, systemic arterial hypertension, peritoneal dialysis as a prior treatment, duration of venous access, presence of previous venous access, use of tunneled hemodialysis CVC as a previous access, and previous antimicrobials were associated with a higher occurrence of BSI. No data was found in the literature on the possible relationship between systemic arterial hypertension and previous peritoneal dialysis with the occurrence of BSI in this population.
For examination of time of duration of current venous access as a risk factor, BSI was found to be associated with access that had been gained within 30 days. Duration of current venous access was analyzed using three categories (0–30 days, 30–180 days, and >180 days). BSI was higher in those who had had the CVC inserted within 30 days, which underscores the need for greater care during the CVC insertion procedure. Napalkov et al. in a study that evaluated infectious and noninfectious complications in CVC, reported that the majority of BSI occurred within the first 90 days, with an incidence rate/1000 catheter-days of 5.1 (CI 95%: 3.7–4.3) [31]. There was also a higher risk of infection for catheters inserted within the 6 months following infection, reinforcing the need for care during this period.
The use of previous venous access and, primarily, previous access of tunneled hemodialysis CVC, was associated with a higher occurrence of BSI. As mentioned above, studies have reported that CVC use is associated with a higher incidence of BSI compared with the use of an AVF [8-10]. References to previous venous access as a risk factor for BSI were not found in the literature.
Previous use of antimicrobials was associated with a higher occurrence of BSI (OR: 2.53; p = 0.013). As with other variables, data were collected on the use of antimicrobials within 6 months before the event. We found no specific studies that related prior antimicrobial exposure to BSI in patients undergoing hemodialysis treatment. However, Lim et al. in a case–control study conducted with patients seen in the emergency department of a tertiary hospital in Australia, showed through multivariate analysis that prior antimicrobial use increased the risk of developing BSI by multi-resistant microorganisms (OR: 5.49; p < 0.001) [32]. According to the Centers for Disease Control and Prevention, the rational use of antimicrobials is an important measure for controlling the spread of multi-resistant microorganisms [23].
Gram-positive microorganisms were the most prevalent (72.8%) found in the current study. Among them, S. aureus was the most frequently isolated (32.1%), with about 38.5% resistant to oxacillin. Other studies have also reported a high prevalence of Gram-positive microorganisms in patients undergoing hemodialysis treatment, primarily S. aureus [7,10,11]. Despite the high prevalence of Gram-positive microorganisms observed in the current study, Gram-negative microorganisms accounted for 25.9% of isolates in blood cultures. In a university hospital in Spain, microorganisms isolated in blood cultures from patients with kidney disease were: Gram-negative bacteria, 52.3%; Gram-positive bacteria, 46.5%; and fungi, 1.2%. Escherichia coli was the most frequent microorganism (27%) [33].
With regard to resistance profiles, microorganisms such as MRSA and VRE have been isolated in surveillance cultures collected from patients on dialysis treatment [29,34,35]. Aktas et al. used molecular typing to show a high similarity between MRSA strains isolated from surveillance cultures and clinical cultures collected from patients on dialysis treatment [30]. And a study conducted at long-term care facilities in Hong Kong reported a significant dissemination of MRSA, reinforcing the need to adopt measures to reduce the transmission of this microorganism [36].
The independent risk factors for BSI identified in the current study were use of a CVC compared with use of an AVF, and previous hospitalization. Although previous studies have indicated a relationship between the presence of CVC and BSI, our study expands the characteristics of this group, highlighting the importance of care during the initial period and the period of up to 6 months of catheter duration. Furthermore, the study highlights the importance of previous hospitalization as a risk factor for BSI. To the best of the authors’ knowledge, this factor has not been reported in previous studies and requires further research.
This study had some limitations inherent in the retrospective character of the data collection. However, patient records at the study institution were completed with significant caution, given the characteristics of a service dedicated to the treatment of chronic renal patients, and followed the high standards of the organization. Additionally, as there are standardized protocols for performing blood cultures for these patients, it is unlikely that events were underdiagnosed in the study period.
Preventative measures against BSI related to CVC should be strengthened and effectively applied to hemodialysis treatment units, as well as reducing the use of this device whenever possible, prioritizing the use of AVF. Considering that previous hospitalization was an independent risk factor for BSI, patients who developed BSI after hospitalization are probably infected by microorganisms during this period. Standard precautions should be enforced to prevent microorganism dissemination during hospitalization.
A reduction in the number of infections in this population may contribute to a decrease in hospital admissions, since infections are a leading cause of hospitalization in patients on hemodialysis treatment.