Over one year has gone by since the description of the first cases of respiratory infection due to the SARS-CoV-2 coronavirus.1 The Spanish national COVID-19/SEMICYUC registry of cases admitted to Intensive Care Units (ICUs) in Spain has evidenced a high percentage of patients who develop organ failure (100% respiratory, 44% cardiovascular and 30% renal), with long ICU stays (14 days [range 8–26]) and a mortality rate of 32.6% according to the results corresponding to the first 2200 patients published.2 Potere et al., in a systematic review and meta-analysis of the first series of patients in the ICU, documented a mortality rate of 34% that increased to 83% among patients on mechanical ventilation (MV)3 – a figure significantly higher than that observed in the COVID-19/SEMICYUC registry.2

It can be affirmed that the empirical treatments with scant scientific evidence used in the early stages of the pandemic unfortunately yielded no benefits and did not modify the ominous course of the severe cases seen in the ICU.4 Likewise, the first data generated by the ENVIN-COVID registry, based on the analysis of 1525 patients with COVID-19 admitted to intensive care between 15 March and 30 May, showed that 50% of the patients had suffered one or more infections related to medical care, and that the frequencies of the different controlled infections had multiplied between two- and four-fold (ENVIN-COVID report, 2020).5 It therefore seems prudent to place emphasis on measures for the prevention of infections in the ICU, and on their early detection, with the aim of providing adequate and early treatment in the difficult challenge of combating the high associated mortality.

With this aim in mind, the Working Group on Infectious Diseases and Sepsis (Grupo de Trabajo de Enfermedades Infecciosas y Sepsis [GTEIS]) of the Spanish Society of Intensive and Critical Care Medicine and Coronary Units (Sociedad Española de Medicina Intensiva, Crítica y Unidades Coronarias [SEMICYUC]), on occasion of its latest ordinary meeting, agreed to carry out a literature review on the state of the art and offer a series of treatment recommendations established by the authors of this document, which were selected by the coordinators of the Working Group on the basis of their experience in infectious diseases.

During the first months of the pandemic, the lowering of a dysregulated immune response was considered to be a potential therapeutic strategy. As a result, and even in the absence of clear scientific evidence, frequent use was made of a range of antiinflammatory drugs and immune modulators such as glucocorticoids, inhibitors of interleukins and/or their receptors, antimalarials, antimitotic agents, calcineurin inhibitors, etc. However, these drugs are not without side effects – one of the most serious of which is an enhanced susceptibility to superinfections in the hospital setting.

Knowledge about the effect of corticosteroids upon the development of superinfections in patients with SARS-CoV-2 infection is presently limited. A meta-analysis of the clinical trials that have evaluated corticosteroid use in critical patients with SARS-CoV-2 infection carried out by the World Health Organization (WHO) concluded that the included studies contributed little information on the adverse effects, and that no meta-analysis of the data was possible.6 Nevertheless, some studies suggested that the use of corticosteroids may be associated to an increased risk of certain infections such as bacteremias, particularly when these drugs are combined with other immunosuppressors.7

Based on the evidence available at the time of drafting of this document, and following the guidelines of the WHO, among others, the use of corticosteroids – preferably dexamethasone – is recommended for the treatment of patients with severe COVID-19.8 It should be considered that although patients with milder respiratory failure could benefit from such treatment, the maximum benefits have been observed in critical patients subjected to mechanical ventilation (odds ratio [OR] 0.66 [95%CI: 0.53−0.82]). Monedero et al.9 reported that the early administration of corticosteroids, within the first 48 h of admission to the ICU is associated to a lesser risk of nosocomial infection (56.3% of cases; OR 1.39, p = 0.05) than when such treatment is provided on a later basis (69.9% of cases; OR = 2.5 p < 0.001) – though the risk of infection is even lower among those patients who never received the drug (48.2% of the cases, OR = 1). Table 110–19 summarizes the main studies on the use of corticosteroids and their association to the development of infections. As regards the type of drug and dose, dexamethasone is preferentially advised, at a dose of 6–7.5 mg/day between 7 and 10 days, since it is the drug substance used in a larger number of studies and which has been found to afford statistically significant clinical benefits.8

The adoption of the recommendations of the Zero Projects has proven to be effective in reducing the appearance of infections related to medical care during patient admission to the ICU in Spain.40,41 The statistics on the different infections related to the use of invasive devices published by the ENVIN registry in the year 201942 were more than adequately compliant with the quality standards of the SEMICYUC.43 During the early phases of the COVID-19 pandemic, the application of contingency plans designed to address the growing healthcare demands, implied the introduction of important structural, functional and human changes that posed barriers for complying with the established prevention measures. In recent months (Table 3), important increments have been reported in the incidence of infections related to medical care among the COVID-19 patients admitted to different Spanish ICUs.44–48 The Critical Patients Safety Projects Advisory Council has recently published a document49 proposing adaptation of the recommendations of the Zero Projects during the SARS-CoV-2 pandemic. Appendix B summarizes the measures contemplated in the mentioned document, and which are complemented by adaptation of the management of invasive devices during patient transfer or during placement in prone decubitus, the prevention of pressure ulcers, and the application of body hygiene.

Despite the recording of a low incidence of bacterial coinfection in hospitalized patients with SARS-CoV-2,50 the incidence recorded in the ICU according to the COVID-19/SEMICYUC registry was 9.3%.2 The recommendation provided in this document seeks to draw attention to the possibility of such infection, in the awareness of the risk and increase in mortality associated with delays in treating complications of this kind.51

A proactive diagnosis of coinfection in clinically severe presentations should be a priority concern. The classical clinical criteria of fever, respiratory failure and radiological infiltration are common to viral pneumonia, and therefore are unable to differentiate the possibility of bacterial coinfection. The distinguishing clinical feature is probably the fact that secretions are typically scarce or inexistent in the early stages of viral pneumonia without coinfection. Patients admitted to the ICU may come directly from the Emergency Department or from the hospital ward. In the latter case, and depending on the number of days spent in the ward, the previous treatments used and the invasive devices employed, we may consider the possibility of infection related to medical care, though in any case the recommendation to actively assess infectious complications upon arrival in the Unit remains the same.

Although procalcitonin values alone are not advised for deciding the start of antibiotic treatment,52 procalcitonin is a more specific tool than C-reactive protein, and in conjunction with the clinical and radiological data may help in considering the possibility of coinfection. Accordingly, it is advisable to determine procalcitonin upon admission to the UCI, and in this regard some authors have proposed a cut-off concentration of <0.1 µg/l as suggestive of viral pneumonia,53,54 in view of its high negative predictive value. In addition, it will be useful to know the subsequent clearance of this marker with a view to deciding the suspension of empirical antibiotic treatment. We thus advise the collection of respiratory samples upon admission to the ICU (at least a tracheal aspirate in intubated patients and a sputum sample in non-intubated patients). Table 4 details the recommended microbiology laboratory tests for studying the respiratory samples.

Langford et al. evaluated the presence of bacterial coinfection in a meta-analysis55 of over 3000 patients. A total of 3.5% of the patients presented coinfection – a figure lower than that reported in patients with infection due to influenza virus. In relation to the critically ill, the authors analyzed the data of 5 studies documenting coinfection in 14 out of a total of 144 patients (9.7%). No etiological data in critical cases were provided, though it seems reasonable to consider Streptococcus pneumoniae and S. aureus as possibly being the most frequent bacteria in community-acquired pneumonia. Based on these data, it has been debated whether empirical treatment should be started or not. What is clear is that the potential presence of bacterial coinfection should be investigated in all cases. A strategy which in our opinion seems reasonable is to start antibiotic treatment considering the most frequent pathogens in community-acquired pneumonia, and to adjust or suspend treatment after 48–72 h conditioned to the microbiological results obtained. If empirical antibiotic therapy is started, we recommend combination treatment in the form of a third-generation cephalosporin such as ceftriaxone or cefotaxime associated to levofloxacin or azithromycin, though a fifth-generation cephalosporin such as ceftaroline could also be prescribed.

Ventilator-associated respiratory infection (VARI), including ventilator-associated tracheobronchitis and ventilator-associated pneumonia (VAP), are common in critical patients subjected to mechanical ventilation56. Patients with pneumonia due to SARS-CoV-2 could be at an increased risk of VARI,57 due to acute respiratory distress syndrome (ARDS) and the long duration of mechanical ventilation. Acute respiratory distress syndrome is a well-known risk factor for VAP,58,59 and its incidence in ventilated patients with SARS-CoV-2 pneumonia is reportedly high (42–89%).60,61 Likewise, SARS-CoV-2 is responsible for alterations of host immune response.

Recently, two Spanish groups62,63 have published their experience, with an incidence of almost 52% referred to infections acquired in the ICU – respiratory foci being the most common presentation in both series, and implying significantly higher mortality in one of the studies.62Pseudomonas aeruginosa was the most frequently isolated microorganism.30,40

Rouze et al.64 carried out a multicenter study to determine the relationship between SARS-CoV-2 infection and the incidence and etiology of VARI. The study was carried out in 36 ICUs in Europe (28 in France, 3 in Spain, 3 in Greece, 1 in Portugal and 1 in Ireland). The interesting feature of this study is that it included the same number of patients on mechanical ventilation per center with SARS-CoV-2 infection, influenza infection or no viral infection. The main finding was that the incidence of ventilator-associated tracheobronchitis and VAP was greater in the patients with SARS-CoV-2 than in the other two groups (influenza infection or no viral infection). Gramnegative bacilli were responsible for most of the first VARI episodes in all three study groups. P. aeruginosa, Enterobacter spp. and Klebsiella spp. were the most commonly identified bacteria. The percentage of patients with VARI involving multiresistant bacteria was lower in pneumonia due to SARS-CoV-2 than in the other two groups.

The choice of empirical and/or guided treatment should be based on the clinical entity and the severity of presentation, the presence of risk factors for infection due to a multiresistant microorganism, the existence of comorbidities, and knowledge of the local flora and multiresistance rates of the center involved, according to the proposed algorithm.65 The increase in bacterial resistances and the high incidence of infections involving gramnegative bacilli such as Klebsiella spp. and Pseudomonas spp., lead us to consider empirical treatments with activity against this etiological profile, including carbapenems (for which the appearance of resistances is increasing), ceftazidime-avibactam and ceftolozane-tazobactam, whose spectrum of action covers these possibilities. The duration of treatment has not been clearly established, and in this type of infection in particular, the decision must be individualized, in view of the important lung damage caused. Likewise, we always must consider the possibility of shortening both the duration of treatment and tapering of the regimen, based on the use of scales such as the Clinical Pulmonary Infection Score (CPIS), and biomarkers such as C-reactive protein and particularly procalcitonin65 in combination may be of help.

Table 5 summarizes the recommendations on the management of VAP in patients with COVID-19.

The Bacteremia Zero Project has reduced the incidence of catheter-related bacteremia (CRB),40 though in the moments of maximum critical care bed occupation and expansion it has been difficult to maintain compliance with the measures recommended by this and by other similar projects in our Units.49 A greater incidence of nosocomial bacteremia has been observed in COVID-19 patients admitted to hospital than in non-COVID-19 patients (5.3% versus 1.4%; p < 0.001), and admission to the ICU has moreover been identified as a risk factor.66 The ENVIN-COVID report of 2020 has described an incidence of 6.32 cases per 1000 days of central venous catheter,5 which is much higher than usual in our Units. Similar and even poorer results have been obtained in other studies. Giacobbe et al. documented an incidence of nosocomial bacteremia of 47 episodes per 1000 patients-day, with 73% corresponding to primary bacteremias and CRB.67 In contrast, Buetti et al., in their case-COntrol study, found that COVID-19 patients have no greater risk of CRB than non-COVID-19 patients.68 Both studies do coincide in describing a greater risk of CRB in patients treated with corticosteroids, tocilizumab or anakinra – observations that could not be confirmed in the clinical trials with these drugs.

The situations in which CRB should be suspected, and the required diagnostic procedures, are no different from those referred to non-COVD-19 patients.69 With regard to management, there is no evidence to establish treatment regimens different from the usual ones – though some aspects need to be considered. In relation to the choice of antimicrobial therapy and catheter withdrawal, patients treated with corticosteroids or tocilizumab are to be regarded as immunosuppressed individuals, at least during treatment with corticosteroids and between two and four weeks after the administration of tocilizumab. We should adapt the antimicrobial treatment to the local resistances profile, focusing special attention on the changes in microbiology that might appear derived from the peculiarities of COVID-19 patients (Table 3), the high frequency of treatment with cephalosporins upon admission, and care of critical patients in areas different from the usual areas.

According to the data of the ENVIN-HELICS report of the year 2020,5 which describe the analysis of 1525 patients with SARS-CoV-2 pneumonia in the first wave of the pandemic, 19% of the infections related to medical care corresponded to bladder catheter-related urinary tract infections (C-UTI), with an incidence of 6.54 C-UTI episodes per 1000 days of catheter. A series of factors may predispose to the development of C-UTI, particularly a prolonged ICU stay, immune alterations, the need for an indwelling bladder catheter, and antibiotic treatment. The strategy for the prevention of C-UTI in the ICU has been extensively developed in the UTI-ZERO Project,70 and there are no differential aspects applicable to patients with SARS-CoV-2 infection. The most frequently isolated organism has been Enterococcus faecalis (Table 3). The change in microbiology observed in C-UTI in patients with SARS-CoV-2 pneumonia must be interpreted with caution, since it may be explained by the initial management approach involving cephalosporins, which since then has changed. Accordingly, until new data are forthcoming, the antibiotic treatment of these patients should ensure the coverage of gramnegative bacilli, considering the significant percentage of P. aeruginosa isolates. Antifungal treatment can be provided by fluconazole in patients with symptomatic candiduria and in those with asymptomatic candiduria and a risk of disseminated disease, as in neutropenic individuals, immunosuppressed patients or patients subjected to urological manipulations. In the case of infection due to Candida krusei or Candida glabrata resistant to azoles, use can be made of amphotericin B deoxycholate via the intravenous route. The echinocandins have been shown to be effective in some cases despite their scant elimination in urine. The current tendency in relation to the duration of treatment is to prescribe short cycles of between 3–7 dias.71,72

In its more severe presentations, pneumonia due to SARS-CoV-2 has been associated with an increase in the incidence of opportunistic infections. Although a prolonged ICU stay or treatments with corticosteroids and/or immune modulators can be identified as risk factors in these patients, there is growing evidence that infection due to this virus and the development of ARDS give rise to lymphocyte dysfunction,73 which in turn can favor the development of opportunistic infections. We will focus on the two most common opportunistic infections in patients with COVID-19: invasive pulmonary aspergillosis and the reactivation of Herpesviridae. A marked increase in Candida spp. isolates has been observed, though the treatment in such cases is no different from that indicated in patients without SARS-CoV-2 infection.

The main limitation of this document is that the recommendations made are not fundamented upon any analysis of the evidence, since at this time there is still much to be learned about how the pandemic should be managed, and very few studies with an adequate methodological design have been able to contribute solid information in this respect. The existing evidence therefore must be classified as being of low level. Nevertheless, it can be concluded that the scarce evidence on effective treatments applicable to SARS-CoV-2 viral infection in critical patients focuses on supportive measures, the prevention of infections related to medical care, and the early detection and treatment of those infectious complications that have been described as being common in these patients.

All the authors form part of the GTEIS of the SEMICYUC, and were selected by its coordinators for development of the present document.

They have participated in the preparation of all the document sections and have approved the final version. AE and RZ coordinated the project that gave rise to this document.

AE have participated in speeches/training activities organized by Pfizer, Gilead, MSD.

PVC has participated in speeches/training activities organized by Pfizer, Shionoghi, MSD.

RZ has participated in speeches/training activities organized by Pfizer, Shionoghi, MSD, Gilead.

The rest of the authors declare that they have no conflicts of interest.