Adjunctive treatment in COVID-19 and sepsis—What did we learn?

Sepsis is the leading killer worldwide. A worldwide survey in 2017 revealed 48.9 million incident cases per year globally of which almost 11 million died [1]. By applying the new Sepsis‑3 definitions, it became obvious that almost 80% of hospitalized patients infected by severe acute respiratory syndrome coronavirus type 2 (SARS-CoV-2) had viral sepsis [2]. Contrary to sepsis, several drugs which modulate the immune response were registered for severe coronavirus disease 2019 (COVID-19), namely anakinra, baricitinib and tocilizumab as adjunctive treatment to antivirals. The present review aims to provide insights into how lessons learnt from the use of adjunctive treatment for COVID-19 might be applied in sepsis in order to advance the field.

Despite the high incidence and mortality, the global sepsis problem is ignored by most global health authorities. In response to this, five European societies, the European Sepsis Alliance, the European Society of Clinical Microbiology and Infectious Diseases, the European Society for Intensive Care Medicine, the European Society of Anesthesiology and Intensive Care and the European Society of Pediatric and Neonatal Intensive Care framed a call for action to the European Center for Disease Prevention and Control [3]. It might be asked why sepsis is so profoundly ignored. A clear-cut answer is that there is no drug registered for sepsis and all drugs administered to patients aim to eliminate the infection which triggered sepsis sequelae. In recognition of the need for antisepsis drugs, several randomized controlled trials (RCT) were conducted throughout the years 1990–2003 with conflicting results. Drotrecogin-alpha was registered for severe sepsis with an APACHE II score 25 or more following the results of the PROWESS RCT; however, when this RCT was repeated the results were negative and the drug was retracted from the market [4]. The experiences from the past generate the profound question on what we expect from adjunctive immunotherapy. There are two answers to this question: adjunctive therapy potentiates the effect of co-administered antimicrobics towards better elimination of bacteria by modulation of the immune response or adjunctive therapy attenuates the exacerbated proinflammatory phenomena to minimize the deleterious effects on the organs. In order to decide which is the mechanism of benefit, the current review is following a three-step approach. At the first step, all available evidence on how the successful immunotherapies tested for severe COVID-19 impacted on viral load is presented. Then, existing evidence on how persistence of bacteria drive outcome of sepsis is presented. At the last step, teachings of precision immunotherapy for COVID-19 are transferred to the sepsis field.

In a prospective study patients infected by SARS-CoV‑2 were split into those who achieved clearance of the virus from the nasopharyngeal swab in the first 16 days from onset of symptoms and into those who had prolonged presence of the virus lasting more than 16 days. The extent of bilateral lung involvement in the chest computed tomography was broader among patients with a long persistence of the virus. The absolute counts of CD4 lymphocytes and CD8 lymphocytes were lower among patients with a long viral persistence [5]. The key to the elimination of SARS-CoV‑2 is the lymphocytes and among them the natural killer (NK) cells. A study on 168 patients early during the pandemic proved that viral shedding was prolonged when the absolute NK cell count was less than 100 cells/mm³; this was associated with poor survival [6]. A positive association was found between circulating interleukin (IL) 6 and the viral load [7]. This drives the question if inhibition of the IL‑6 receptor by tocilizumab may improve viral clearance through an effect on the lymphocytes. Cumulative clinical evidence is presented in Table 1 [8‐12] showing that inhibition of the IL‑6 pathway is associated with better viral clearance. The suggested mechanisms are through the increase of the absolute NK cell count or through up-regulation of the human leukocyte antigen-DR (HLA-DR) on CD14 monocytes which improves antigen presentation and virus-induced lymphopenia [11, 12]. It is even suggested that circulating IL‑6 antagonizes the pharmacokinetics of the antiviral drugs [9], so that blocking of IL‑6 may improve the delivery of the standard of care (SoC) antiviral agents to tissues.

In the SAVE-MORE RCT [13, 14] which led to the approval of anakinra for patients at risk for progression into severe respiratory failure, the primary endpoint was the allocation of the two groups of treatment (SoC and placebo versus SoC and anakinra) into the different strata of the World Health Organization clinical progression scale (WHO-CPS). The comparison between the two groups showed that patients randomized to treatment with SoC and anakinra had an odds ratio (OR) of 0.36 for worse outcome than patients treated with SoC and placebo. This means that the benefit of adjunctive anakinra treatment was exerted towards both spectra of the WHO-CPS, i.e., severe disease and death or full disease resolution with viral elimination [13, 14]. There is a double interpretation of the findings; anakinra treatment prevented disease progression and the deleterious effect of the proinflammatory phenomena and, in parallel, facilitated better viral elimination.

Early antibiotic treatment is considered the cornerstone of sepsis management. This is suggested in all versions of the Surviving Sepsis Campaign guidelines and it is a direct expression of the need to eliminate the bacterial pathogen in order to improve the outcomes [15]; however, contrary to COVID-19 no clinical evidence is available on whether adjunctive immunotherapy in bacterial sepsis is enhancing pathogen clearance or not; however, it is self-explanatory that if the patients do not improve in the first 72 h, the administered antibiotics need to be changed and it is confirmed [16] that appropriate antibiotic treatment is a sine qua non for sepsis management. In a propensity-matched analysis, patients with bacteremia due to enterobacteria were switched after an initial intravenous course of antibiotics to oral antibiotics. The switch was carried out provided that an appropriate source control was included, orally administered antibiotics were in vitro active against the pathogen according to antimicrobial susceptibility testing and the Pitt bacteremia score was 5 or less (n = 739). These patients were compared to another fully matched 739 patients completely treated with intravenous antibiotics. The mortality after 30 days was similar [17]. Some treating physicians may consider this switch from intravenous to oral treatment as risky for patients with sepsis and bacteremia; however, the lack of a difference in mortality underlines that bacterial elimination using appropriate antibiotics is the most important strategy for sepsis management.

However, clinical evidence to associate a change of bacterial load with immunomodulatory treatment is not available. The current understanding of the pathophysiology of sepsis is that bacterial recognition by the immune system leads to an initial proinflammatory response which is followed by anti-inflammatory phenomena. Scarce evidence from animal studies suggests that experimental treatment which modulates either the proinflammatory or the anti-inflammatory phase and leads to improvement in survival is also accompanied by better containment of bacteria in tissues. A recent example is experimental treatment with a recombinant antibody targeting IL-36alpha in mice which developed sepsis after cecal ligation and puncture (CLP). The CLP sepsis led to an increase of IL-36alpha, but not of IL-36beta and IL-36gamma in blood and tissues. Anti-IL-36alpha after CLP prolonged survival, decreased bacterial load and improved the macrophage capacity for phagocytosis in the tissues [18]. Intratracheal instillation of Pseudomonas aeruginosa in mice 5 days after CLP leads to an experimental condition of immunoparalysis greatly resembling the human situation. Mice are not able to achieve potent proinflammatory responses and the death rate depends on the instilled inoculum of P. aeruginosa [19]. Sepsis-induced immunoparalysis of tissue macrophages exists in mice not only in the model of pneumonia following CLP but even when the first infection is pneumonia. In this setting alveolar macrophages are not able to achieve appropriate long-term phagocytosis [20]. Our group studied a model of immunoparalysis induced in rabbits after acute pyelonephritis by Escherichia coli. The aim was to restore immunoparalysis with recombinant human interferon-gamma (rhIFNγ). Rabbits received amikacin antibiotic treatment and co-administration with rhIFNγ which prolonged survival, decreased circulating counts of bacteria and improved the function of splenocytes for ex vivo production of tumor necrosis factor-alpha (TNF-alpha) [21].

There are several lessons from the success of adjunctive immunotherapy for severe COVID-19 which must be followed in order to increase the chances that immunomodulatory treatment succeeds in sepsis. These lessons are summarized in the next sections.

The introduction of anakinra, baricitinib and tocilizumab for the treatment of severe COVID-10 reinforced the concept of immunotherapy in sepsis; however, it seems that it is high time that instead of testing drugs in all patients, well-characterized populations should be treated using biomarkers which identify the prevailing immune mechanism in the body of the host. This approach may maximize efficacy and bring new advances in the field.