LY303366

Hemodynamic changes in surgical intensive care unit patients undergoing echinocandin treatment

Abstract
Background Echinocandins are well-established agents for the treatment of patients with fungal infections, but growing evi- dence questions their safety in special patient populations prone to systemic inflammatory responses. Objective The study aimed to analyse early hemodynamic changes during echinocandin therapy in critically ill surgical patients. Setting The study was conducted at the surgical intensive care unit at the University Hospital of Giessen, Germany. Methods This single-centre retrospective study includes data from critically ill patients who underwent primary antifungal treatment during 2009–2013. Main outcome measures Hemodynamic parameters, need for vasopressor/inotropic therapy, and dose of vasopressor/inotropic therapy were recorded 2 h before and 2 h after the onset of antifungal treatment. Comparisons of echinocandins to azoles and analysis of a combined endpoint (decrease of mean arterial pressure ≥ 10 mmHg and/or new or increased dosages of norepinephrine, epinephrine, or dobutamine) were performed. Results We found 342 episodes of intravenous antifungal treatment (33 [9.6%] anidulafungin, 116 [33.9%] caspofungin, 132 [38.6%] fluconazole, 17 [5%] micafungin, 44 [12.9%] voriconazole). Group comparisons revealed no significant differences of hemodynamic parameters, need for vasopressor/ inotropic therapy, and dose of vasopressor/inotropic therapy, expect for a decreased dose of norepinephrine in the flucona- zole group (p < 0.001). The combined endpoint occurred in 58 (50%) caspofungin-, 16 (48.5%) anidulafungin-, 4 (23.5%) micafungin-, 23 (17.4%) fluconazole-, and 15 (34.1%) voriconazole treatment episodes. Secondary analysis of the combined anidulafungin/caspofungin group to the azoles group (fluconazole, voriconazole) showed a significant decrease of mean arterial pressure ≥ 10 mmHg (n = 37 [25%] vs. n = 27 [15%], OR = 1.8, p = 0.04), increased use of norepinephrine (n = 38 [26%] vs. n = 12 [7%], OR = 4.7, p ≤ 0.001), increased use of dobutamine (n = 12 [8%] vs. n = 4 [2%], OR = 3.8, p = 0.02), and the combined endpoint (n = 74 [50%] vs. n = 38 [21%], OR = 3.6, p ≤ 0.001). Conclusion Our retrospective data might demonstrate clinically relevant hemodynamic-depressing effects of anidulafungin and caspofungin. Further prospective acquisition of clinical data will be necessary to evaluate their impact on hemodynamic function. Introduction Although bacteria remain the main cause of severe infec- tions, fungal infections are an emerging problem in the treatment of critically ill patients, resulting in significant risk for patients in the surgical intensive care unit (ICU) [1–3]. Several treatment guidelines address antifungal therapy for severely ill patients, including echinocan- dins (anidulafungin, caspofungin, and micafungin) for the empiric and specific treatment of invasive Candida infections [4, 5]. Furthermore, echinocandins are also rec- ommended as salvage therapy for patients suffering from invasive aspergillosis [6]. Echinocandins are well-estab- lished agents for the prophylaxis and treatment of patients with proven or suspected fungal infections, but growing evidence questions their safety in special patient popu- lations prone to systemic inflammatory responses (e.g., critically ill patients, burn patients, patients with severe organ dysfunction). These concerns are hypothesized to be due to the high pharmacokinetic variability of echino- candins related to age, body surface area, disease status, and body integrity [4, 7, 8]. Adverse cardiac effects as a result of echinocandin administration were first described in case reports of critically ill and neutropenic patients [9–11]. In contrast, a recent prospective study of fifteen medical patients suffering from septic shock showed a significant increase in mean arterial and diastolic blood pressures immediately after application of echinocandins, while 4 h after administration differences were no longer present. Further analysis of the amount of catecholamine therapy as well as all transpulmonary thermodilution monitoring parameters did not indicate any significant changes immediately and 4 h after echinocandin admin- istration [12]. Previous experimental studies analysing isolated cardiomyocytes (using the Langendorff model) and isolated rat hearts confirmed an echinocandin-induced cardiac impairment [13, 14]. Central venous high dose administration of anidulafungin or caspofungin in adult rats led to a significant reduction of cardiac output and was associated with a reduced survival rate compared to con- trol animals, which further studies showed to be a dose- dependent mechanism [14, 15]. In contrast, micafungin did not cause cardiac impairment in rats or isolated cardiac myocytes as shown by Stover et al., who demonstrated that even high-dose micafungin treatment was free from car- diac toxicities [14–17]. One possible explanation for these findings might be the different pharmacological properties of micafungin compared to the other echinocandins. Since micafungin is water-soluble, in contrast to the lipophilic properties of anidulafungin and caspofungin, it does not penetrate the cell membrane and might therefore cause less cell damage [13]. This retrospective observational single-centre cohort study was performed at the surgical ICU of the University Hos- pital of Giessen, Germany. We included the data of each adult patient (≥ 18 years) who received intravenous antifun- gal treatment during the period from 2009 to 2013. Patients with suspected or proven aspergillosis were excluded. All patients were treated according to the local standard of the surgical ICU, University Hospital of Giessen, Germany, and according to the guidelines of the European Society of Clini- cal Microbiology and Infectious Disease for the diagnosis and management of Candida diseases in non-neutropenic adult patients [4]. Empiric antifungal therapy was defined as the initiation of antifungal therapy prior to documented microbiological detection of Candida species, while specific treatment was defined as the initiation of antifungal therapy after documented microbiological detection of Candida spe- cies. Prophylactic treatment was referred to a pre-emptive therapy in patients with high risk for fungal infections (e.g., after lung transplantation). Drug dosage and speed of intra- venous administration of antifungal agents were carried out according to manufacturers’ recommendations. Study patients were reviewed and validated for inclusion and exclusion criteria as well as for plausibility by analys- ing the ICU Patient Data Management System (PDMS) ICUData (IMESO® GmbH, Giessen, Germany). Data were recorded in an external database using Statistical Package for the Social Sciences (SPSS) Statistics, version 24 (IBM, Armonk, NY, USA). Baseline parameters included age, body mass index (BMI), gender, and history of surgical treatment. We also recorded the severity of illness using the sequential organ failure assessment (SOFA) score, simplified acute physiology score (SAPS II) score, and acute physiol- ogy and chronic health evaluation score (APACHE II) score. Hemodynamic parameters including systolic, diastolic, and mean arterial blood pressure (MAP), heart rate (HR), car- diac index (CI), central venous pressure (CVP), need for vasopressor/inotropic therapy, and dosage of vasopressor/ inotropic therapy were recorded 2 h before and 2 h after the onset of antifungal treatment. We also recorded length of ICU stay, length of hospital stay, and 30-days mortality.We observed positive fungal cultures in 89.8% of all patients included in the study. C. albicans was most fre- quently found within the microbiological results (60.1%) (Supplement 1). We found that 58.8% of patients suffered from invasive Candida infections including severe infec- tions such as candidemia (7.3%) and intra-abdominal infections (29.5%), while 29.5% were treated with anti- fungals without proven Candida infection. Addition- ally, 11.7% received prophylactic antifungal treatment after lung transplantation. Invasive Candida infections were associated with a significantly elevated mortality between the study groups were analysed using paired t tests or Wilcoxon signed-rank tests for matched samples. P values of statistical differences for single- and combined endpoints were calculated by using the Fisher’s exact test. A two-tailed value of p < 0.05 was considered to be statistically signifi- cant. Furthermore, odds ratios were calculated for all group comparisons. All statistical analyses were performed using R statistical software version 3.4.2 (www.r-project.org) and Graphpad Prism version 5.0 for Mac (GraphPad Software, La Jolla, CA, USA).Data are shown as number of treatment episodes n (%), mean (± SD) or median [IQR]BMI Body Mass Index, ICU Intensive Care Unit, APACHE II: Acute Physiology and Chronic Health Evaluation Score, SOFA Score Sequential Organ Failure Assessment Score, SAPS II Simplified Acute Physiology Score compared to patients without proven Candida infection (14.6% vs. 27.5%; p = 0.048).Hemodynamic measurementsWe analysed 342 episodes of intravenous antifungal treat- ment at baseline 2 h before administration of antifungal therapy and 2 h after the administration of antifungal medi- cation. Patients receiving anidulafungin, caspofungin, flu- conazole, micafungin, or voriconazole showed no changes in hemodynamic function (CI, HR, blood pressure, CVP) (Table 2). Except for a decreased dose of norepinephrine in the fluconazol group (p < 0.001), analysis of single echino- candins found that the median infusion rate of continuously administrated norepinephrine, epinephrine, and dobutamine also did not differ between baseline values and 2 h after anti- fungal administration within each group (Table 3).Analysing these factors, the combined endpoint occurred in 58 caspofungin treatment episodes (50%). Additionally, fol- lowing anidulafungin administration the combined endpoint occurred in 16 episodes (48.5%) (Fig. 1a). In contrast, the combined endpoint was achieved only in 4 micafungin treat- ment episodes (23.5%), 23 fluconazole treatment episodes (17.4%) and 15 voriconazole treatment episodes (34.1%) (Fig. 1a). Combining the echinocandins anidulafungin and caspofungin resulted in a significant decrease of MAP ≥ 10 mmHg (n = 37 [25%] vs. n = 27 [15%], OR = 1.8,p = 0.04), increased use of norepinephrine (n = 38 [26%] vs. n = 12 [7%], OR = 4.7, p ≤ 0.001), and increased use of dobu- tamine (n = 12 [8%] vs. n = 4 [2%], OR = 3.8, p = 0.02) com- pared to the azole group (Fig. 1b). Furthermore, the com- bined endpoint occurred in 74 cases (50%) of the combined anidulafungin/caspofungin group compared to 38 cases (21%) of the azoles group (OR = 3.6, p < 0.001) (Fig. 1b).Comparison of the micafungin group (n = 17) to the azole group (n = 176) did not reveal significant differences in the decrease of MAP ≥ 10 mmHg (n = 2 [11%] vs. n = 27 [15%], OR = 0.74, p = 1), the use of norepinephrine (n = 3 [18%] vs. n = 12 [7%], OR = 2.91, p = 0.13), the use of dobutamine (n = 0 vs. n = 4 [2%], OR = not applicable, p = 1) or the com- bined endpoint (n = 4 [24%] vs. n = 38 [22%], OR = 1.11, p = 0.77) (data not shown). Discussion In the current study our aim was to analyse early hemody- namic changes during echinocandin therapy in 342 antifungal treatment episodes from 270 surgical ICU patients. Consider- ing the severity of illness reflected by the SOFA, SAPS, and Values are shown as median with interquartile range [IQR] or numbers with percentage n (%), n: number of patients per group, n (dose > 0): Number of patients receiving medication of interest per group, % (dose > 0): Percentage of patients receiving medication of interest per group, (all): entire group including patients without receiving medication of interest NA: not applicable, not enough data for statistical test. Data were analyzed using the Wilcoxon signed-rank test associated with an increase in morbidity and mortality. Since the ideal individual blood pressure of critically ill patients has not yet been well-defined, various definitions of intra- and postoperative hypotension have been published. Futier et al. found that a reduction of MAP of 6 mmHg in a cohort of patients undergoing major surgery lasting 2 h or longer was significantly associated with elevated rates of postoperative organ dysfunction [18]. Furthermore, Salmasi et al. assessed that in a cohort of patients after non-cardiac surgery MAP values below an absolute threshold of 65 mmHg or a rela- tive threshold of 20% reduction of MAP from preoperative baseline values were progressively related to both myocar- dial- and kidney injury [19]. The European Best Practice Guidelines on Hemodynamic Instability by Kooman et al. defined an absolute decrease of MAP ≥ 10 mmHg as a clinically relevant hemodynamic instability in patients dur- ing hemodialysis [20]. Similarly, Khanna et al. determined that in a cohort of postsurgical ICU patients, an absolute decrease of MAP ≥ 10 mmHg was significantly associated with the occurrence of acute kidney injury [21]. Further- more, a pre-existing history of hypertension as well as the duration of the hypotensive episode might affect patients’ outcomes. Therefore, we chose the combination of endpoints MAP ≥ 10 mmHg and changes of hemodynamic treatment in order to reflect the clinicians’ daily routine evaluation and management of hemodynamic alterations. Due to the retrospective study design, we also defined the start of a new hemodynamic medication or the increase (independent from the relative percentage change) in continuously administered catecholamines as clinically relevant outcome parameters that might reflect clinicians’ treatment of possible echinocandin- induced hemodynamic suppression.

Expect for a minor but significantly decreased dose of nor- epinephrine in the fluconazol group, the individual analyses of hemodynamic parameters (blood pressure, CI, HR, CVP) as well as the use of vasopressors or inotropes revealed no significant changes of hemodynamic stability within 2 h after echinocandin administration. It is possible that short-term clinical interventions might have ameliorated the effect of echinocandins on blood pressure, CI, HR, and CVP. There- fore, fluid administration, vasopressor therapy, or other therapeutic interventions might have attenuated echinocan- din-induced hemodynamic depression. Additionally, flucona- zole- and voriconazole therapy caused no changes in hemo- dynamic parameters. However, examining the combined endpoint consisting of the reduction in MAP (≥ 10 mmHg) and/or the need for new or increased vasopressor/inotrope therapy could reveal important differences between theFig. 1 Combined endpoint analysis a comparison of individual anti- fungals. Number of antifungal treatment episodes n (%) per group: anidulafungin 33 (9.6), caspofungin: 116 (33.9), fluconazole 132(38.6), micafungin 17 (5), voriconazole 44 (12.9). MAP: mean arte- rial pressure. * = p < 0.05; ** = p < 0.01; *** = p < 0.001. The p values have been obtained from applying fisher’s exact test for count data. b Comparison of anidulafungin/caspofungin vs. fluconazole/vori- conazole group. Number of antifungal treatment episodes n (%) per group: anidulafungin/caspofungin (total n = 149; combined endpoint: n = 74; MAP ≥ 10 mmHg: n = 37; epinephrine: n = 3; norepinephrine: n = 38; dobutamine: n = 12), fluconazole/voriconazole (total n = 176, combined endpoint: n = 38; MAP ≥ 10 mmHg: n = 27; epinephrine: n = 1; norepinephrine: n = 12; dobutamine: n = 4). Combined end- point included decline of MAP ≥ 10 mmHg and/or new- or increased dosages of norepinephrine, epinephrine or dobutamine. Norepineph- rine group: new- or increased dosages, epinephrine group: new- or increased dosages, dobutamine group: new- or increased dosages.* = p < 0.05; ** = p < 0.01; *** = p < 0.001. The p values have been obtained from applying fisher’s exact test for count data different antifungal agents. The combined endpoint was found in 48.5% of the anidulafungin treatment episodes and in 50% of the caspofungin treatment episodes. In contrast, the combined endpoint was only found in 24% after micafungin treatment. These results are in line with animal studies [13, 14, 17, 22, 23]. One possible explanation might be the dif- ferent pharmacological features of micafungin compared to the other echinocandins. Since micafungin is water-soluble in contrast to the lipophilic properties of anidulafungin and caspofungin, it does not penetrate the cell membrane and might therefore cause less cell damage [14–16]. Furthermore, Stover et al. suggest from their studies that micafungin alone represents an effective, high-dose regimen that is compared to anidulafungin and caspofungin free from cardiac toxicities [17]. Therefore, we combined anidulafungin and caspofungin for further analysis and identified a significant decrease of MAP (≥ 10 mmHg) and a significant increase of the use of norepinephrine, epinephrine, and dobutamine in the echino- candin group compared to the azole group (voriconazole- and fluconazole-treated patients). Previously, we were able to demonstrate a dose-dependentdecrease of contractility after echinocandin administration in isolated cardiomyocytes using the Langendorff model, which has been confirmed in another ex vivo study [13, 15, 16]. Subsequently, we infused echinocandins in dosages equivalent to standard clinical human therapy as well as in high dosages via central lines into adult rats while monitor- ing the hemodynamic changes with an in vivo assessment device. The results indicated that high dose administration of anidulafungin [25 mg/kg bodyweight (BW)] and caspo- fungin (8.75 mg/kg BW) effectively reduced cardiac activity, and survival time in contrast to the lower dose study group (anidulafungin group: 2.5 mg/kg BW; caspofungin group: 0.875 mg/kg BW) [14]. Similar results have been described following administration of clinically relevant doses of anid- ulafungin and caspofungin in an in vivo endotoxemic shock rat model [22]. However, the detailed pathomechanism of echinocandin-induced hemodynamic and cardiac impair- ment remains unclear. Cleary et al. identified toxic effects of echinocandins on cardiac mitochondrial function and adeno- sine triphosphate (ATP) synthesis [13, 23, 24]. In contrast to these results, our study group was not able to detect any altered mitochondrial enzyme activity in spectrophotomet- ric analyses of rat left ventricular cardiac tissue [22]. We assumed an echinocandin-induced dysregulation of intra- cardial calcium homeostasis based on previous study results and similar findings concerning septic cardiac failure [22, 25, 26]. Recently, our group was able to demonstrate that caspofungin therapy induces a dose-dependent increase of intracellular calcium in human cardiac myocytes. However,calcium ions were found to be released from intracellular caffeine-sensitive stores likely via activation of ryanodine receptors [27]. Therefore, dysregulation of calcium home- ostasis might be one possible explanation for our current results. In order to shed light into the pathomechanism of echinocandin-toxicity further experimental and clinical stud- ies are necessary, especially to differentiate echinocandin effects from the influence of other drugs and the underlying severe disease in the clinical context.In contrast to our findings, Lahmer et al. were not able to detect a significant impairment of hemodynamic status after echinocandin administration but observed a transient increase of the diastolic and mean arterial pressure [12]. Other hemodynamic parameters deriving from transpulmo- nary thermodilution and the need for vasopressors did not change significantly, even though the dosage of echinocan- dins were identical to our recent study. However, it is possi- ble that differences observed between our study and Lahmer et al.’s study may be due to differences in the sizes of the populations studied [12]. Our study has several limitations. First, due to the retro- spective design, no causal effect of echinocandins on nega- tive hemodynamic function can be verified. Second, data concerning the type of intravenous application (central vs. peripheral line) is lacking, particularly since our data sup- ports a possible aggravating effect of centrally administered echinocandins as well as fast injection of echinocandins. Third, due to the dose-dependent prolonged infusion scheme (90–180 min) that is suggested for anidulafungin treatment, some of the study patients might not have received the com- plete medication at the observation point (2 h after onset of antifungal treatment). Fourth, we analysed an unsorted cohort of postsurgical patients with diverse comorbidities and severity of illness. Due to the retrospective character of our study, we were not able to detect special cohorts of patients at risk. Fifth, sedation, fluid management, and other drugs might also have had an impact on our hemodynamic findings that we were not able to exclude. Last, due to the exploratory character of the study, we did not predefine the comparison between echinocandins and azoles, even though the severity of illness might have affected the choice of anti- fungal therapy and therefore influenced the study´s results. However, we believe that our findings might affect daily clinical awareness during antifungal therapy in critically ill postsurgical patients and justify further prospective studies. Conclusion In summary, this retrospective study yields data on critically ill postsurgical patients that might demonstrate an important hemodynamic-depressing effect of anidulafungin and caspo- fungin. These findings LY303366 are in line with published case reports and experimental data. Further prospective acquisition of clinical data will be necessary to evaluate their impact on hemodynamic function.