Introduction

Pandemics have shaped our genomics, politics, geography, and history, and continue to do so in the 21st century. According to Professor John C. Marshall (Toronto, Canada), intensive care units (ICUs) became the frontline for epidemiological studies of emerging infectious disease for the first time in public health history during the influenza A (H1N1) pandemic of 2009, when the infection was first characterized and the spread of the pandemic traced among critically ill patients [1, 2].

A pandemic confronts public health authorities with the immediate need for information regarding the epidemiology and natural history of the pathogen, the availability of treatment resources, and optimal therapies to maximize survival of the largest number of patients. Professor Marshall reported that the critical care community is responding to this challenge through the International Forum for Acute Care Trialists (InFACT), an informal consortium of over 20 investigator-led clinical trial groups and professional societies [3].

InFACT is now collaborating with the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC) to design a clinical trial that could readily be used to test multiple inexpensive interventions and generate real-time information during the next pandemic. In the meantime, Professor Marshall urged ISICEM delegates to support InFACT and other collaborative investigator-led initiatives, as clinical trials are not only key to pandemic preparedness, but are also one of the best ways to discover how to optimize care of ICU patients.

Candida

ICUs are also at the frontline of managing the adverse effects of other serious infections such as invasive candidiasis, which is associated with high mortality in critically ill patients, especially if treatment is delayed for more than 12 hours [4]. Antifungal prophylaxis has been suggested as a promising approach among at-risk patients in the ICU. However, the large randomized placebo-controlled INTENSE study, which was presented during ISICEM, found no difference between preemptive therapy with micafungin or placebo in the incidence of invasive fungal infection (IFI) or time to confirmed IFI in patients with intra-abdominal infection [5].

The European Society of Clinical Microbiology and Infectious Diseases (ESCMID) recommends prophylaxis with fluconazole to prevent intra-abdominal candidiasis in patients with recent abdominal surgery and recurrent perforations or anastomotic leakages. The recommendation is based on limited evidence, however, and to date, no drug has been shown to reduce overall mortality from IFIs when used prophylactically in non-neutropenic patients [6]. According to Dr. Thierry Calandra (Lausanne, Switzerland), this underscores the importance of early empirical antifungal therapy at first suspicion of candidemia. ESCMID strongly recommends first-line therapy with one of the echinocandins rather than an azole antifungal agent or liposomal amphotericin B [6].

Echinocandins are the drug of choice because they are broad-spectrum, fungicidal, and efficacious in reducing the risk of death [7]. There is also evidence that median time to first negative blood culture is shorter in patients treated with an echinocandin compared to those treated with fluconazole (two days versus five days) [8]. The ideal time at which to step down therapy remains unclear, but Dr. Calandra endorsed ESCMID’s advice to continue echinocandin for 14 days after the end of candidemia, as determined by one blood culture per day until results are negative, switching to fluconazole if the patient is stable and tolerates the oral dose and if the species is susceptible to the drug [6].

Candida peritonitis presents a particular challenge in the ICU. Isolation of Candida species is an independent risk factor for mortality in nosocomial peritonitis (although not in community-acquired infection) [9], but the decision to initiate therapy is complicated by the difficulty in differentiating between infection and colonization. According to Professor Philippe Montravers (Paris, France), conventional Candida risk scores are of limited value in the ICU. However, a high rate of Candida isolation in peritoneal fluid is associated with the presence of at least three of four risk factors: female gender, upper gastrointestinal tract origin of peritonitis, intraoperative cardiovascular failure, and previous antimicrobial therapy at least 48 hours before the onset of peritonitis [10]. Biomarkers may also be helpful, with the combination of ß-D-glucan > 259 pg/mL and positive test for Candida albicans germ tube antibodies (CAGTA) accurately differentiating between Candida colonization and invasive candidiasis in non-neutropenic critically ill patients [11].

Professor Montravers reported that, at this time, only the 2010 Infectious Disease Society of America (ISDA) guidelines address the management of intra-abdominal infections caused by Candida [12]. If Candida is grown from intra-abdominal cultures, antifungal therapy is recommended in patients with severe community-acquired or healthcare-associated infection. Fluconazole is the treatment of choice, with echinocandins reserved for resistant Candida spp. or as first-line therapy in critically ill patients. Because of its toxicity, amphotericin B is not recommended as first-line therapy [12].

Dr. Calandra urged ISICEM delegates to collaborate with hospital colleagues to ensure that appropriate therapy for Candida infection is also introduced in a timely manner outside the ICU, especially in internal medicine wards, which he described as a “hot spot” for invasive candidiasis. Since the definition of “internal medicine” varies depending on the hospital and health service, the precise incidence of candidemia in these patients is uncertain. However, the caseload is likely to be similar or greater in internal medicine than in the ICU, with one-third to one-half of patients with blood cultures positive for Candida found in these wards [13–15].

There is also some indication that while the incidence of candidemia remains stable or is declining in traditionally high-risk areas of the hospital such as the ICU, the infection may be becoming more common in internal medicine wards [16]. As in the ICU, the risk of mortality rises with each day’s delay in starting antifungal therapy, ranging from 15 % for internal medicine patients who begin therapy on the day of the first positive blood culture to 41 % among those starting treatment three or more days after the first positive sample [17]. The risk of death may be even higher among patients in internal medicine wards than among high-risk patients elsewhere in the hospital. An Italian study identified 30-day mortality rates of 54.1 % in internal medicine patients compared to 47.6 % among those in ICU and hematology/oncology wards [18].

Early-Onset Candidemia

Professor Francesco De Rosa (Turin, Italy) agreed that early initiation of antifungal therapy for internal medicine patients is critical. There are, however, many risk factors for Candida infection in hospitalized patients, and the challenge of diagnosing invasive candidiasis underscores the importance of addressing modifiable risk factors. Professor De Rosa regarded this strategy as especially important for patients outside the ICU, where management of central venous catheters and long courses of antibiotics may select for endogenous Candida. It is also important to bear in mind the association between candidemia and severe infection with Clostridium difficile, possibly due to the elimination of bowel flora following antibiotic treatment [19].

The timing of onset of candidemia may also be an important risk factor for adverse outcomes in internal medicine patients. Professor De Rosa reported results of a study involving 779 patients in two large Italian hospitals comparing early-onset (183 patients) and late-onset (596) infection. In this study, "early onset" was defined as within 10 days of hospital admission in order to include patients with healthcare-associated infections and to clearly differentiate from the usual timing of the onset of candidemia at 22–27 days after admission [20].

The investigators had anticipated that mortality would be higher in early-onset patients than in the late-onset group. The overall risk of death, however, was significantly lower in early-onset candidemia (38.8 % versus 47.5 %, p = 0.03), although mortality due to inadequate initial antifungal therapy was three times higher (OR 3.02, 95 % CI, 1.40–6.51; p = 0.005). Other independent risk factors for mortality in early-onset patients were C. albicans etiology (OR 2.17, 95 % CI, 1.11–4.26; p = 0.02) and older age (OR 1.05, 95 % CI, 1.02–1.07; p = 0.001). In late-onset candidemia, liver disease (OR 2.46, 95 % CI, 1.36–4.43; p = 0.003), inadequate antifungal therapy (OR 2.01, 95 % CI, 1.28–3.15; p = 0.002), and older age (OR 1.03, 95 % CI, 1.02–1.04; p = 0.001) were independently associated with higher mortality, while treatment with caspofungin was associated with improved survival (OR 0.42, 95 % CI, 0.26–0.67; p = 0.001).

Early-onset candidemia was more frequent in internal medicine wards. Professor De Rosa commented that these findings may identify lack of awareness as an important risk factor for mortality in these patients. As such, it is vitally important that patients are stratified according to their risk factors and clinical severity, while maintaining a high index of suspicion to ensure early diagnosis and treatment, as well as de-escalation when a patient is clinically stable [21].

Invasive Pulmonary Aspergillosis

While Candida infection is a well-recognized phenomenon in the ICU, Dr. Elie Azoulay (Paris, France) submitted that there is less recognition of the risk of invasive Aspergillus infection. Current evidence suggests that the fungus is present in the respiratory tract in approximately 1–2 % of patients in the ICU [22, 23]. Since Aspergillus is an air-borne fungus, the degree of immunocompetence is a crucial factor in determining a patient’s risk. An abnormal response may begin with colonization of the lung, which could be the result of hypersensitivity and is seen primarily in patients with allergies or asthma, but the fungus can enter the respiratory tract or lung vasculature to cause invasive disease in immunocompromised patients.

As such, while ICU patients do not have classical risk factors for invasive aspergillosis such as neutropenia, they are at risk because they have become immunocompromised due to septic shock, respiratory failure, or treatment with corticosteroids. Dr. Azoulay advised that in critically ill patients, ventilator-associated pneumonia in the absence of proven bacterial infection should raise the possibility of aspergillosis. This presentation differs from that seen in hematology/oncology and solid-organ transplant patients, in whom aspergillosis may be suspected in the presence of clinical signs such as fever.

Dr. Azoulay went on to describe other differences in presentation as a potential source of confusion and delay in diagnosing invasive aspergillosis in ICU patients. The “halo sign” on high-resolution computed tomography of the lung is among the clinical criteria for IFIs in hematology/oncological and other transplant patients [24], but is of limited value in mechanically ventilated patients. In these patients, the presence of galactomannan in brocheoalveolar lavage fluid is a more reliable means of ensuring the early diagnosis of invasive aspergillosis [25].

According to Dr. Azoulay, the European Organization for Research and Treatment of Cancer/Mycosis Study Group (EORTC/MSG) definitions of proven, probable, and possible invasive fungal disease are very reliable in daily practice because they are focused on the clinician at the bedside. These criteria have recently been adapted to produce a simple clinical algorithm that is able to discriminate between respiratory tract colonization by Aspergillus and invasive pulmonary aspergillosis in ICU patients, and is especially helpful in ruling out invasive disease [26].

Since critically ill patients have been largely excluded from clinical trials of antifungal therapies, there are currently no evidence-based recommendations specifically for the ICU. Dr. Azoulay advised adapting current recommendations on the treatment of invasive aspergillosis in hematology/oncology patients, taking into account contraindications and drug–drug interactions likely in critically ill patients. He concluded by reiterating the importance of clinical trials in the ICU. In particular, studies should include patients on long-term (more than 10 days) mechanical ventilation and should investigate early treatment strategies in patients with risk factors for invasive aspergillosis.