• COVID-19
• inflammation
• outcome assessment
• health care

We thank Tampe et al 1 for their interest in our recent work, in which we describe criteria to predict worse outcome in hospitalised patients with COVID-19 and an hyperinflammatory response that we defined as cytokine storm.2 A limit of our work is that it was based on patients treated in a single hospital. It is, therefore, important that Tampe et al 1 successfully validated our criteria in an independent cohort, from a different continent and a different health system, and therefore it is a step toward criteria that can be applied to other cohorts around the world.

We find interesting that when Tampe et al applied the criteria without ferritin and C-reactive protein (CRP), it further improved the ability of the same criteria to predict longer ICU stay and risk of death in this very severe population. This is perfectly in line with our original work, in which we included ferritin and CRP only for clinical reassurance of an ongoing systemic inflammation and despite they did not add predictive power.2 These results diminish the role of ferritin and CRP and highlight that COVID-19 induces a very atypical cytokine storm, as already reported by others.3 4 Moreover, patients who met the criteria in our own cohort had increased levels of serum interleukin 6 although far lower than those reported in macrophage activation syndrome or cytokine release syndrome.5 These results suggest that the COVID-19 infection leads to a storm where damage may be as relevant as the cytokine-induced inflammation.

Effective treatments of COVID-19-induced cytokine storm have yet to be found; numerous rigorous clinical trials have shown marginal or no benefits of cytokine neutralisation, while the only drug so far shown to substantially improve survival is the pan anti-inflammatory corticosteroids.6 7 Most if not all performed clinical trials have in common very broad criteria of inclusion, such as elevation of some inflammatory markers, lung involvement by imaging studies and requirement for oxygen supplementation. Such entry criteria might have allowed the enrolment of patients with a mild disease or with features that did not focus on a cytokine/damage storm, therefore altering significantly the outcome of the trials.8 An asset of the work by Tampe et al is that they selected severe patients and applied the criteria at the time of the admission to ICU, therefore focusing on a very specific group of patients. The fact that in this group the ‘Temple Criteria’4 identified the patients bound to die suggests that this cytokine-damage storm is the main cause of death in patients with COVID-19. We hope that our criteria, although imperfect and in need of improvement as demonstrated by Tampe et al, will help better stratifying patients for anti-inflammatory treatment during COVID-19 infection. We will await for further validation and improvements on larger cohorts.