Background

Improvement in exit site care and the use of antibiotic agents at the exit site have significantly decreased the incidence of Gram-positive exit site infection (ESI) [1,2,3]. The International Society for Peritoneal Dialysis (ISPD) Guidelines recommend daily application of topical antimicrobial agents (including mupirocin or gentamicin) at the exit site to prevent ESI [4]. Corynebacterium species, which are Gram-positive bacilli, are uncommon but recognizable causes of peritoneal dialysis (PD)-associated peritonitis. Corynebacterium species have also been increasingly recognized over the past decades, largely due to improved recognition and microbiological techniques [5,6,7,8]. Previous studies have reported that Corynebacterium infection accounts for a major proportion of PD-associated peritonitis [5,6,7,8]. On the other hand, we have seen only a few recent reports about how many ESIs due to Corynebacterium lead to PD-associated peritonitis and tunnel infection [9, 10]. ESI is one of the most common causes of PD-associated peritonitis, and it can lead to tunnel infection (TI) and catheter loss [1, 4]. We conducted a single-center, retrospective study of ESI due to Corynebacterium.

Methods

Study design and participants

We retrospectively reviewed the clinical data of 125 patients with end-stage renal disease who received PD and were followed up at the International University of Health and Welfare (IUHW) Hospital, Tochigi, Japan, from April 2018 to January 2023. This study protocol was approved by the ethics committee of the IUHW (approval number: 23-Im-028). Patients without submission rate data and those who opted out were excluded. This study was registered at inception in the UMIN Clinical Trials Registry under identification number UMIN000052200. In this study, we included patients aged ≥ 20 years who selected PD based on shared decision-making in renal replacement therapy selection.

Data collection and patient evaluation

Patients’ demographic data were obtained from medical records. Baseline characteristics at PD initiation included sex, age, weight, primary cause of kidney disease, estimated glomerular filtration rate (eGFR) (ml/min/1.73 m2), information about self-care PD or assisted PD, diseases that caused end-stage renal disease, and blood pressure. The Body mass index (BMI) was calculated using height (m) and body weight (kg) data at the time of PD initiation. Moreover, PD-related parameters were obtained at the first onset of ESI. PD vintage and the Charlson comorbidity index were obtained and compared between patients with ESI due to Corynebacterium and those with ESI associated with other pathogens. Unfortunately, markers of small solute clearance from PD—such as PD Kt/V or creatinine clearance—and those of residual kidney function—such as residual GFR—were not assessed systematically in our clinical practice. Therefore, we measured PD fluid volume as a marker of small solute clearance from PD and the serum urea-to-creatinine ratio as a marker of residual kidney function, which has been identified as an indicators of residual kidney function in patients on PD [11]. Moreover, the treatment period of each ESI was obtained.

Follow‑up

All participants were followed up until PD cessation (i.e., transfer to HD alone), death, kidney transplantation, transfer to another hospital, or study completion (31 January 2023). Transfer to HD alone was defined as a transition to thrice-weekly HD.

Exit site care and definitions of PD-associated infection

We have applied gentamicin sulfate ointment at the exit site as usual care for all patients on PD. ESI diagnosis is based on the presence of a purulent discharge, with or without skin erythema at the catheter–epidermal interface [4]. The persistence of ESI due to the same pathogens for more than 4 weeks was defined as persistent pustular discharge and a positive culture from the exit site [4]. If the culture was negative for three consecutive months, the ESI was considered cured. TI is defined when we confirm the presence of clinical inflammation with or without ultrasonographic evidence of a fluid collection anywhere along the PD catheter tunnel or the presence of a low-density area around the PD catheter by computed tomography [4, 8]. PD-associated peritonitis is defined when at least two of the following three conditions are satisfied: (1) presence of fever and abdominal pain and (2) total PD effluent cell count > 100/cm3, of which more than 50% were polymorphonuclear neutrophils, and (3) a positive culture of pathogens in the PD fluid [4,5,6,7,8]. In cases of polymicrobial PD-associated infection (ESI, TI, and PD-associated peritonitis), multiple isolations of the same cultured organisms with different susceptibility patterns were evaluated separately. If Corynebacterium was detected along with other major pathogens, such as Staphylococcus, the latter bacteria would be considered pathogenic. Therefore, this type of ESI was defined as “ESI due to other pathogens with Corynebacterium isolated” and categorized as “other pathogens-associated ESI,” along with “ESI due to other pathogens alone.” ESI in which only Corynebacterium was detected from the exit site was defined as “ESI due to Corynebacterium” or “Corynebacterium ESI.” Topical antibiotic treatment of Corynebacterium ESI was mupirocin ointment and minocycline ointment. Systemic antibiotics were initiated when we confirmed poor response to topical treatment. The standard systemic antibiotic treatment for Corynebacterium ESI includes oral levofloxacin, oral sulfamethoxazole, and intravenous vancomycin. When we confirm poor response to this series of antibiotics, we perform unroofing surgery, subcutaneous pathway diversion, or PD catheter removal. Response to medical treatment was defined as the disappearance of pustular discharge from the exit site after treatment.

Statistical analysis

Continuous variables were presented as mean values ± standard deviations or medians (25–75th percentiles) based on the normality of data distribution assessed by the Shapiro–Wilk test. Categorical variables were presented as frequencies and percentages. Normally and non-normally distributed continuous variables were evaluated using the one-way analysis of variance and Kruskal–Wallis test, respectively, and the chi-square test was used for categorical variables to compare parameters between groups. Infection rates were calculated as the number of infections divided by the total duration of exposure and expressed as episodes per patient-year. All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan), a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [12]. A two-tailed P-value of < 0.05 was considered statistically significant.

Results

Patient characteristics

A total of 125 patients underwent PD from April 2018 to January 2023. They were followed for 179.6 patient dialysis years (median follow-up period: 1.40 years per patient). A total of 191 cases of ESI, 10 cases of TI, and 32 cases of PD-induced peritonitis occurred in 125 patients with PD during the study period. The baseline characteristics of our patients are presented in Table 1.

Table 1 Clinical characteristics of patients at the initiation of peritoneal dialysis
Full size table

Exit site infection characteristics

The incidence rates of all ESI and Corynebacterium ESI were 1.06 and 0.35 per patient dialysis year, respectively. A total of 62 cases of Corynebacterium ESI were detected in 24 patients during the study period. Of the 24 patients who developed Corynebacterium ESI, 16 had a combination of ESI due to Corynebacterium and ESI associated with other pathogens. We found that the eGFR during PD introduction was relatively lower in patients with Corynebacterium associated ESI than in patients without ESI (6.4 versus 7.9 mL/min/1.73m2, respectively; P = 0.052). Meanwhile, there was no significant difference in the serum urea-to-creatinine ratio between patients with ESI due to Corynebacterium and patients with ESI associated with other types of pathogens (Table 2). These data are presented in Tables 1 and 2 and Fig. 1.

Table 2 Comparison of clinical characteristics of ESI between the two groups
Full size table
Fig. 1
figure 1

The Venn diagram for ESI due to Corynebacterium alone, ESI due to a combination of Corynebacterium and other types of pathogens, and EIS due to other types of pathogens alone. ESI exit site infection, PD peritoneal dialysis

Full size image

A total of 36 cases of ESI associated with other pathogens were identified on separate days in 14 of 24 ESI patients with Corynebacterium ESI. Of these, 30 cases of ESI associated with other types of pathogens, three cases of Corynebacterium ESI were followed by the onset of ESI associated with other pathogens, and 27 cases of Corynebacterium ESI were developed after the onset of ESI associated with other pathogens. However, 41 patients had only ESI associated with other pathogens. In this group of 41 patients, 90 cases of ESI were detected (including ESI due to other pathogens with Corynebacterium isolated). We observed a tendency that once Corynebacterium was detected at the exit site, this pathogen continued to be detected. In PD catheter exit site culture, 16 of 24 patients tested positive for Corynebacterium multiple times. We identified patients who tested positive on exit site culture with Corynebacterium up to eight times and compared the recurrence rate of ESI between ESI due to Corynebacterium and ESI associated with other pathogens (including ESI due to other pathogens with Corynebacterium isolated). Consequently, the recurrence rate of ESI due to Corynebacterium was 35.5%, which is significantly higher than the rate of ESI associated with other pathogens (8.5%, P < 0.001). There was no significant difference in the treatment period of ESI between ESI due to Corynebacterium and ESI associated with other pathogens (P = 0.404). PD vintage during the onset of ESI was significantly longer in patients with Corynebacterium ESI than in those with other pathogen-associated ESI (P < 0.050).

Characteristics of tunnel infections

We detected ten cases of TI. The incidence rates of all TI and Corynebacterium TI were 0.06 and 0.02 per patient dialysis year. TI did not recur in any of these patients. Three cases of TI were caused by Corynebacterium, five were caused by other types of pathogens (including ESI due to other pathogens with Corynebacterium isolated), and two were caused by unknown bacteria. Of these ten cases of TI, eight required PD catheter removal. Of the 62 cases of ESI due to Corynebacterium, 3 (4.9%) progressed to TI. Among these three cases, two (3.2%) required PD catheter removal, and the other was relieved by conservative treatment with intravenous vancomycin. Of the two cases that required PD catheter removal, both had diabetes. In these two cases, one had the PD catheter removed on the fifth day of intravenous vancomycin administration and the other on the eighth day of intravenous vancomycin administration and oral combination of trimethoprim and sulfamethoxazole administration. One case that was relieved by intravenous vancomycin administration for 14 days was a nondiabetic patient case. Among the 129 cases of ESI associated with other types of pathogens, 7 (5.4%) progressed to TI directly. Among these seven cases, four required PD catheter removal followed by PD catheter replacement later. Among these 129 cases of ESI associated with other types of pathogens, Corynebacterium was isolated in 20 cases, of which 1 (5.0%) progressed to TI followed by PD catheter removal. There was no significant difference in the frequency of direct progression to TI between ESI due to Corynebacterium and ESI associated with other types of pathogens (P = 1).

Characteristics of peritoneal dialysis-associated peritonitis

We also detected a total of 32 cases of PD-associated peritonitis. Of the 129 cases of ESI associated with other pathogens (including ESI due to other pathogens with Corynebacterium isolated), 2 (1.6%)—from which Corynebacterium was not isolated—directly progressed to PD-associated peritonitis. However, no cases of ESI due to Corynebacterium directly progressing to PD-associated peritonitis were confirmed. There was no significant difference in the frequency of direct progression to peritonitis between ESI due to Corynebacterium and ESI associated with other pathogens (P = 1).

Discussion

Generally, ESI control is critical in PD treatment. However, the nature of ESI due to Corynebacterium has hardly been investigated to date. Corynebacterium species are Gram-positive bacilli and belong to the natural flora of the skin [4,5,6,7]. Gentamicin prophylaxis previously demonstrated efficacy in preventing most ESIs [4, 13, 14]. In the past, some researchers worried that using gentamicin cream at the exit site would increase the transmission of gentamicin-resistant pathogens; however, this concern has now largely been dispelled [15, 16]. Although some previous reports suggested that gentamicin ointment might adversely affect some specific pathogens (including Corynebacterium) [9, 10, 17,18,19,20], these past reports only dealt with about 20 cases of Corynebacterium-associated ESI at most. Additionally, the study with the most patients followed up among these ones was published more than 15 years ago, and the ISPD guidelines have changed between then and now, implying that exit site care protocols have changed in that time.

In the case of ESI in which both Corynebacterium and other pathogens are isolated, the presence of Corynebacterium in exit-site samples has been thought to indicate just contamination and not full-blown infection [9, 21]. Although we confirmed one case (5.0%) of TI following ESI due to other pathogens with Corynebacterium isolated, this does not necessarily indicate that Corynebacterium is the “true” causative pathogen. However, three (4.9%) cases of ESI due to Corynebacterium alone directly progressed to TI, suggesting that Corynebacterium cannot be overlooked as the “true” causative pathogen in ESI [21]. Moreover, the eGFR during PD introduction was relatively lower in patients with Corynebacterium ESI than in patients without ESI (6.4 versus 7.9 mL/min/1.73 m2, respectively; P = 0.052). Furthermore, PD vintage was significantly longer in patients with ESI due to Corynebacterium, suggesting that the residual kidney function during the onset of ESI may be decreased from baseline in patients with Corynebacterium ESI compared with those with other pathogen-associated ESI. Although there was no significant difference in serum BUN/Cre during the onset of each ESI between the Corynebacterium ESI group and the ESI associated with other types of pathogens, five of six patients on hybrid PD/HD therapy were categorized in the Corynebacterium ESI group, and the BUN/Cre ratio was not considered an accurate indicator of residual kidney function in such patients [11]. Considerably, lower residual kidney function may increase ESI because of Corynebacterium. Reportedly, skin barrier function decreases as renal function decreases [22, 23]. Additionally, Corynebacterium infection is more frequent in pathologies and diseases in which the skin barrier function is impaired [24,25,26]. Another report suggests that the risk of Corynebacterium infection may increase because of antibiotic use, continuous catheter placement, and increased Corynebacterium colonization [21]. In patients with a longer PD vintage, the periods for the application of gentamicin ointment naturally extend to a longer period. Herein, PD vintage was significantly longer in patients with Corynebacterium ESI. Therefore, we hypothesized that the long-term use of antibiotics such as gentamicin sulfate ointment in patients with reduced residual kidney function may increase the number of cases of ESI due to Corynebacterium. However, a large-scale, multicenter study is needed to validate these results and determine whether the avoidance of the ointment decreases the rate of ESI due to Corynebacterium. The estimated incidence of ESI is 0.53–0.6 and 0.40 per patient dialysis year in the USA and Japan [13, 27, 28]. Herein, the total incidence of all ESI—i.e., Corynebacterium ESI and non-Corynebacterium ESI—was higher (1.06 per patient-year) than that in previous reports. Because the incidence of ESI varies from country to country and region to region—typically between 0.13 and 1.28 per patient-year—the results of our study may fall into the category of such variations [29,30,31]. Moreover, diagnoses of each ESI were based on subjective observations [4, 30, 32]. At our facility, there is continuous education and coaching on exit site management by nurses who specialize in PD. This established system allows patients to directly visit the PD outpatient clinic if they experience any abnormality at the exit site. Therefore, culture samples collected from the exit site were possibly collected more frequently than in general facilities, which may have overestimated ESI incidence.

We found no case of ESI where Corynebacterium ESI directly progressing to PD-associated peritonitis in our study. Therefore, we assume that Corynebacterium-associated ESI may be less likely to develop into Corynebacterium-associated peritonitis. In other words, Corynebacterium may show different properties in ESI- and PD-associated peritonitis. This result is consistent with those of previous studies [4, 5, 7].

The present study has several limitations. First, since this was a retrospective study, we may have overlooked other ESIs. Second, since it was a single-center study, our findings are inconclusive and not generalizable. However, despite these limitations, we still think that the findings of this study reflect the actual situation of Corynebacterium-associated ESI.

Conclusions

In our current study, the incidence of Corynebacterium-associated ESI was higher than that in similar past studies. In patients with Corynebacterium-associated ESI, the eGFR at the time of PD initiation tended to be relatively low. However, such ESIs were thought to rarely develop into PD-associated peritonitis due to Corynebacterium.