Effectiveness of Fluconazole Prophylaxis in a Targeted High-Risk Group in a Surgical Intensive Care Unit

Patients and Methods

In March 2019, a review was conducted of central line-associated blood stream infections at our medical center. The review, dating from January 2017 through March 2019 (“pre-intervention period”), also included a review of patients receiving TPN in the surgical ICU. This review identified seven cases of candidemia occurring in 88 patients undergoing TPN administration. Six candidemia episodes occurred in 58 patients receiving TPN for ≥6 days. Consistent with Infectious Diseases Society of America guidelines, an intervention was developed to initiate fluconazole prophylaxis in patients receiving TPN for ≥6 days. The typical dose was 200 mg intravenously per day, with adjustment for renal function. A 17-month follow-up period (April 1, 2019–August 31, 2020) comprised the post-intervention period.

A treatment course of TPN was considered as continuous if there were ≤3 consecutive days of interrupted therapy. If an interruption of therapy occurred that was ≥4 consecutive days, a new course of therapy was considered to begin with the re-initiation of TPN. An episode of TPN administration was considered “high-risk” when it exceeded 6 days. All positive blood cultures during TPN therapy were recorded.

The Student t-test was used to compare the number of days of TPN therapy during the pre-intervention and post-intervention treatment periods. Statistical comparison of infection rates was performed using the OpenEpi software program [13]. A two-tailed p ≤ 0.05 was considered significant.

Results

The 20-bed surgical ICU is located in a 627-bed Level 1 trauma center. During the 27-month pre-intervention period, there was an average of 481 patient days and 139 central line days per month. During the 17-month intervention period, there was an average of 493 patient days and 156 central line days per month.

During the pre-intervention period, there were 129 treatment courses of TPN in 88 patients. In the post intervention period, there were 49 treatment courses in 27 patients. The average length of courses of TPN therapy was similar for the pre- and post-intervention periods (13 ± 22 days versus 18 ± 31 days; p = not significant [NS]). In the pre-intervention period, there were 81 treatment courses involving TPN that were considered “high-risk” (> 6 days); during the post-intervention period, there were 28 high-risk treatment courses. The average duration of high-risk TPN therapy in the pre-intervention period was 13 ± 26 days compared with 24 ± 38 days in the post-intervention group (p = NS).

In the pre-intervention period, seven patients developed candidemia during TPN administration. Of the seven Candida, three were C. parapsilosis, two were C. albicans, and one each were C. tropicalis and C. glabrata. Of the seven cases, six involved patients who underwent extensive small and/or large bowel resection, including three with high-output enterocutaneous fistulas; the seventh patient suffered extensive trauma during a motor vehicle accident and had multi-system organ dysfunction with septic shock. Four of the seven patients had a Sequential Organ Failure Assessment (SOFA) score ≥8. Six episodes occurred during high-risk TPN therapy.

During the pre-intervention period, there were seven episodes of candidemia during 1,277 days of TPN therapy (Table 1); in comparison, there were zero episodes during 852 days of TPN therapy (p = 0.03). Similarly, during the pre-intervention period, there were six episodes of candidemia during 867 high-risk days of TPN therapy (Table 1) compared with zero during 643 days of high-risk therapy in the post-intervention period (p = 0.04). During the pre-intervention period, there were 30 bacteremias during 1,277 days of TPN therapy compared with 13 during 852 days in the post-intervention period (p = NS). During the intervention period, there were 26 cultures positive for fungi from all patients residing in the surgical ICU: 21 were C. albicans, two were C. tropicalis, two were C. glabrata, and one was Aspergillus versicolor. Of the two patients with C. glabrata, one was known to be colonized prior to admission to the ICU, and the second did not receive TPN or fluconazole before or at the time of the culture.

In the pre-intervention period, there were 118 days of fluconazole therapy during the 1,277 days of TPN administration; in the post-intervention period, there were 569 days of therapy during the 852 days of TPN treatment (p < 0.0001). During the pre-intervention period, there were 76 days of fluconazole treatment during the 867 high-risk days of TPN therapy; during the post-intervention period, there were 495 days of fluconazole therapy during the 643 high-risk days of TPN administration (p < 0.0001).

Finally, there were no clear instances of fluconazole-induced hepatotoxicity. Of 35 patients receiving both TPN and fluconazole, there were five instances in four patients of transaminase concentrations that reached five times the upper limit of normal. In all five instances, the abnormal liver function test results were attributed to sepsis/bacteremia or hypotension, not the drug.

Discussion

The cost associated with a central line catheter infection has been estimated to be approximately $48,000 [14]. This likely underestimates the cost associated with Candida-related infections, given that short courses of therapy are not recommended; and severe complications (e.g., endophthalamitis) can occur. During the 27-month baseline period, there were six candidemias in patients having prolonged TPN administration; therefore, during the 17-month intervention period, we estimate that approximately three central line-associated candidemias were prevented. We also estimate that the cost of fluconazole used during the intervention was approximately $1,500, suggesting the intervention was highly cost-effective.

Several studies have identified critically ill patients receiving TPN to be at high risk for candidemia [2,8], especially patients receiving an extended course of therapy [12]. Prompt transition to enteral feeding has been recommended to reduce this risk [12]; however, in many patients, this is not feasible. At our medical center, this focused intervention was associated with a significant decrease in candidemias, without emergence of species other than C. albicans. Additional follow-up will be needed to assess the durability of these findings.