Clinical Outcomes of Dialysis Catheter–Related Candidemia in Hemodialysis Patients

Materials and Methods

Results

All patients’ catheter-related candidemia was diagnosed by a standardized protocol for management of catheter-related bacteremia at our institution and treated by individual nephrologists using objective outcome measures. The prospective access database documenting treatment and outcomes allowed formal analysis of success rate and complication rate in the subset of patients with candidemia.

The mean age of the study population was 55 ± 14 yr, and 60% were female. Black patients constituted 80% of the study population, reflecting the demographic distribution of our dialysis population. Diabetes was present in approximately half the patients, and hypertension was present in >80%. Vascular disease was common in the study population. The mean catheter age was approximately 7 mo. Hypoalbuminemia was common, with a mean serum albumin concentration of 2.7 g/dl. A total of 58% of the study patients had received antibiotics in the preceding 2 mo for treatment of catheter-related bacteremia. In addition, 25% of the patients had a concomitant bacteremia with their candidemia. The two treatment groups did not differ significantly in their demographics, clinical and laboratory features, or catheter characteristics (Table 1).

Of the 40 patients with dialysis catheter–related candidemia, 45% grew C. parapsilosis, 22% had C. albicans, 15% grew C. tropicalis, 12% had C. glabrata, and 5% grew other (unknown) Candida species. The distribution of Candida species did not differ significantly between the two treatment groups (Table 2).

Of the 40 study patients with dialysis catheter–related candidemia, 27 (67%) underwent guidewire exchange of their infected catheter, and 13 (33%) had catheter removal with delayed placement of a new, tunneled dialysis catheter. The two treatment groups (guidewire exchange and catheter removal with delayed replacement) were similar in age, gender, race, diabetes, hypertension, vascular disease, catheter age, and serum albumin concentration (Table 1). The frequency of patients with previous or concomitant bacteremia was also similar for both treatment groups, as was the relative frequency of Candida species (Table 2).

Only one patient (2.5% of the total) developed a serious complication of candidemia (Table 3). This patient received a diagnosis of Candida endophthalmitis. A recurrent episode of candidemia within 90 d of the initial infection was observed in 15% of the patients in each treatment group. Finally, patient survival was 75% at 6 mo after the episode of candidemia and was similar in both treatment groups.

Discussion

We found that dialysis catheter–related candidemia is rare, representing only 2% of all episodes of catheter-related bloodstream infections. In comparison, in a national surveillance program of 4725 bloodstream infections in hospitalized patients, candidemia accounted for 8% of the total (18). The lower frequency of candidemia in dialysis outpatients likely reflects the smaller number of risk factors, such as recent surgery, corticosteroids, malignancy, or organ transplantation (5). Hypoalbuminemia and recent treatment with antibiotics were common in our hemodialysis patients with candidemia and presumably represented risk factors for this complication.

The distribution of Candida species in our hemodialysis population differed substantially from that reported in several large series of hospitalized patients. C. albicans accounted for 48 to 62% of all Candida infections in those series (1,5,7,18,19). In contrast, C. albicans accounted for only 23% of candidemia episodes in our series. It is not clear why the distribution of Candida bloodstream infections differs in hemodialysis patients as compared with the general population. Candidemia caused by C. parapsilosis has been associated with a lower failure to respond to fluconazole as compared with infection caused by C. albicans (23 versus 51%) (7). This difference may explain, in part, the benign course of candidemia in our study population, because C. parapsilosis was the predominant organism. Oral fluconazole is the first-line antifungal agent for treating this infection in outpatients.

Candidemia in our hemodialysis population was associated with a low frequency of serious complications, which occurred in only 2.5% of patients. In contrast, major complications were observed in 15% of 427 hospitalized inpatients with candidemia, including 9% with chorioretinitis. Our hemodialysis patients with candidemia experienced a 25% mortality within 6 mo of diagnosis. Again, this is a much lower rate than that reported in three large series of hospitalized patients with candidemia. Those studies observed a 45% mortality at 30 d (1), 30% mortality at 60 d (5), and 40% mortality at 90 d (7). This difference again suggests that catheter-related candidemia is more benign in hemodialysis outpatients than in hospitalized patients with major risk factors.

Finally, guidewire exchange of the infected catheter was equally effective as catheter removal with delayed replacement while offering the advantage of minimizing the need for catheter procedures. A recurrence of candidemia at 90 d was seen in only 15% of patients, confirming the efficacy of this treatment strategy.

This study has some limitations. First, it was a retrospective study; however, all episodes of catheter-related candidemia, medical management, clinical outcomes, and complications were recorded prospectively in a computerized database by two full-time access coordinators. Thus, it is unlikely that we missed relevant information. Second, the patients were treated at a single medical center, and the results obtained may not generalize to all dialysis centers. Third, there was no uniform treatment plan in terms of antifungal agent or catheter management. Rather, the treating nephrologists managed each episode of candidemia according to their medical judgment. Given that it took 8.5 yr to identify 40 episodes of candidemia in a dialysis center with approximately 500 patients, even a multicenter, randomized clinical trial would require several years to compare the two catheter management strategies.

Conclusions

Our study demonstrated that catheter-related candidemia in hemodialysis patients has a low complication rate of 2.5%. Catheter exchange over a guidewire in conjunction with antifungal therapy offers a high cure rate and a low likelihood of complications. This strategy also simplifies the medical management of catheter-related candidemia by minimizing the number of required interventions and avoiding disruption of the dialysis schedule.

Disclosures

None.