Numerous efforts have been made to develop effective treatments for individuals who have been hospitalized with Coronavirus disease 2019 (COVID-19) (1). The pathogenesis of severe COVID-19 is related to SARS-CoV-2 infection and secondary to a dysregulated host immune response (2). In line with the excessive inflammatory response, coagulation activation plays a critical role (3, 4). Within the coagulation pathways, the importance of targeting the complement system (CS), the kinin–kallikrein system (KKS), and the contact activation system (CAS) to curb the progression of COVID-19 to severe disease has been supported by several observational studies, and the rationale to inhibit the CS with a therapeutic candidate has been outlined by us and others (5–9). Lower complement C1q values have been reported in patients with severe COVID-19, indicating relevant complement activation and consumption of this protein (10, 11). Within the CS, C1 inhibitor (C1INH) is a strong inhibitor of C1r/C1s and MASP-1/-2. In addition, C1INH also serves as an inhibitor of factor XII of the CAS as well as kallikrein of the KKS (12). Hence, C1INH may contribute to the deterrence of COVID-19 disease progression by prevention of microthrombi generation, inhibition of leucocyte activation, and interaction with endothelial cells and microorganisms (13, 14).

Conestat alfa (Ruconest^®, Pharming, Leiden, The Netherlands, abbreviated as ConA in this manuscript) is a recombinant human C1INH and shares an identical protein structure with plasma-derived C1INH (pdC1INH). It has a different glycosylation pattern, which is responsible for a shorter half-life (3 h) than pdC1INH (30 h) (15, 16), and seems to target the activation of the lectin pathway more effectively (17). A pilot study using conestat alfa in five patients with severe COVID-19 showed promising results with immediate defervescence, and decrease of inflammatory markers and oxygen supplementation in four of five patients (18). To evaluate the efficacy and safety of ConA for SARS-CoV-2 infection, we conducted a randomized, open-label clinical trial in adult patients hospitalized with COVID-19 (PROTECT-COVID-19 trial).

This randomized trial found that C1INH treatment added to SOC was not associated with clinical improvement or faster recovery from COVID-19 in comparison with SOC alone in hospitalized, non-critically ill patients. The overall mortality in this trial (7%) was substantially lower than the one previously reported in the dexamethasone trial (23%), published by the RECOVERY collaborative group (23). The current trial was terminated prematurely because of a lack of efficacy. The power of the study is low because the sample size needed for adjustment after the interim analyses could not be realized. Although there were more AEs and SAEs in the intervention arm than in the control arm, the trial was not terminated because of safety concerns.

The conduction of this trial was justified by scientific rationales and a previous pilot study (18). The former included the idea that C1INH may dampen uncontrolled complement activation and collateral lung damage, may reduce capillary leakage and subsequent pulmonary edema and the generation of microthrombi by inhibition of the KKS and the CAS, and may preserve the regulatory role of endothelial cells. Despite these arguments, we did not observe a clinical benefit in preventing COVID-19 progression when using ConA plus SOC. The following reasons may explain the study outocomes observed.

First, the administration of ConA may have been too late in the course of COVID-19. Because of the short half-life of ConA, repeated administration of ConA over 3 days was chosen. The dosage chosen was supposed to increase the C1INH concentrations by at least 50%. Pharmacokinetic analyses revealed that C1INH concentrations were already markedly elevated at baseline in both arms, which is in agreement with our previous data (24). It is also a potential hint that activation of plasma cascades is already significant and potentially beyond the “point of no return” for inhibition by C1INH, when patients with COVID-19 seek for medical help because of their symptoms. C1INH targets plasma cascades at an early activation level.

Second, the dosing regimen may have been unable to inhibit the CS in COVID-19 sufficiently and for a prolonged period of time. C1INH concentrations doubled initially after the first administration of ConA but returned to almost baseline before the 4th administration. The exact concentration required to inhibit the complement and other cascades and the optimal target are unknown both in the specific context of COVID-19 and in hyperinflammation secondary to infection in general. Of note, COVID-19 is characterized by a local activation of plasma cascades that progresses to a systemic activation early in the disease (25). As such, our data demonstrate that even a twofold increase in C1INH concentration is not sufficient to significantly inhibit all potential downstream effects of the CS, KKS, CAS, and endothelial cells (26, 27).

Third, in COVID-19, the potential inhibitory role of C1INH may have been overestimated when considering activation through the classical and alternative pathway (11, 24, 28, 29). The lectin pathway of complement has been initially implicated as a major driver of complement activation, and binding of MBL or MASP-2 to SARS-CoV-2 was demonstrated with subsequent complement activation (30–32). Subsequently, it became clear that the S protein of SARS-CoV-2 directly activates the alternative pathway (33), which is not efficiently inhibited by C1INH, and circulating immune complexes may activate the classical pathway (28), in line with results from autopsy studies (34).

Fourth, cross-activation of the CS downstream of the lectin and classical pathway proteases may play a considerable role in COVID-19 disease progression. For example, factor XIa inhibited the regulatory complement factor H of the alternative pathway enhancing alternative pathway activity (35). Activated platelets may trigger alternative pathway activation (36–38), and the proteases of the coagulation and fibrinolysis cascades (both activated during COVID-19 but not or only poorly influenced by C1INH) are capable of cross-activating the CS at the level of C3 or C5 (39, 40). In line, the course of serum markers of inflammation and activation of the coagulation system was similar in both arms. Last, it is unclear if different SARS-CoV-2 variants circulating during the study period may have differential effects on the activation of plasma cascade and hence the inhibitory effect of C1INH.

The CS has been a target of a number of randomized controlled trials in COVID-19. However, only retrospective data from patients requiring continuous positive airway pressure support (41) and a single phase 3 trial reported a survival benefit when administering an anti-C5a antibody, albeit in mechanically ventilated COVID-19 patients for the latter (42), whereas another large phase 3 trial in the same population using a C5 inhibitor was futile (43). Clinical trials targeting the CAS and KKS are scarce, and consist of small sample sizes (44, 45). In the only published randomized controlled trial of C1INH in COVID-19, C1INH did not influence clinical improvement and length of stay compared to the SOC arm, in line with our study results (46). The present trial, despite reporting negative findings, is important to direct future clinical interventions targeting the CS, KKS, and CAS in other respiratory viral diseases.

This trial was designed early during the pandemic. As such, it was decided that instead of defining futility margins for a rather small-scale trial including a diverse patient population, a regular review of the DSMB might be more appropriate, as the included patients would potentially receive various new treatments with poorly known safety profiles as time evolved. Instead, two adaptive interim analyses after 40 and 80 patients were carried out. Although more SAEs occurred in the intervention arm, the majority of them were noted after the intervention period. In agreement with external evaluations, we attributed most SAEs to COVID-19 or a complication of it (e.g., pneumonia or sepsis). This interpretation is also supported by the observation of more severe COVID-19 at baseline and throughout the study period in the intervention arm. Of note and in contrast to C5 inhibitory strategies, C1INH treatment has not been associated with an increased risk for bacterial or viral infections or any effect on viral replication (47). Despite being implemented in the electronic case report form and using variable block sizes, which impairs correct forecasting of the allocation, randomization was not associated with a balanced distribution of severity. Reasons include the 2:1 randomization ratio associated with a rather small number of individuals included in the control arm. Moreover, stratification by center may have contributed to the imbalances observed given that some centers only enrolled a limited number of patients. Indeed, we observed a much larger imbalance after the first interim analysis (40 patients included), which was reduced after including more patients. The DSMB reviewed all safety data and deaths in detail during three meetings. For several reasons including the imbalances in disease severity at baseline favoring the control group and the overall lower-than-expected mortality in the intervention arm in the Swiss centers with the majority of enrolled patients, termination of the study for safety was not recommended.

Our study has limitations. These include the premature closure of the trial, the lack of prespecified futility margins, the small number of patients in the control arm (due to the 2:1 randomization), and the imbalances in baseline characteristics and disease severity observed between the groups at baseline (potentially due to the 2:1 randomization and stratification in the setting of a limited number of enrolled patients). Almost 50% of patients were recruited in one center, reflecting an uneven distribution between the institutions. It was an open-label trial, and as such, it may have been prone to bias. The SOC treatment allowed the use of remdesivir and dexamethasone, which may have diminished the possible effect of ConA. Last but not least, follow-up phone calls might potentially have been less accurate than in-person follow-up visits when assessing for AEs. Despite these limitations, this is the largest study to investigate a treatment strategy that interferes with the CS, KKS, and CAS in hospitalized COVID-19 patients. In addition, we provide patient-matched laboratory data regarding the inability of ConA to limit the activation of these plasma cascades in COVID-19.

In conclusion, ConA did not significantly accelerate clinical improvement or reduce mortality in hospitalized patients with non-critically ill COVID-19 infection. The chosen dose and regimen of ConA used in this study are insufficient to limit complement activation in COVID-19. It remains unclear whether or not higher dosages, a longer treatment regimen, or an earlier intervention time point with ConA in COVID-19 could have prevented disease progression.

The raw data supporting the conclusions of this article will be made available by the authors, on reasonable request.

The studies involving humans were approved by the ethics committees «Ethikkommission Nordwest- und Zentralschweiz», «Ethikkommission Ostschweiz», and “Kantonale Ethikkommission Zürich» in Switzerland, “Comissâo Nacional de Ética Em Pesquisa” (CONEP) in Brazil and “Comisión Federal Para La Protección Contra Riesgos Sanitarios” (COFEPRIS) in Mexico. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

PU: Conceptualization, Data curation, Formal Analysis, Investigation, Methodology, Project administration, Supervision, Visualization, Writing – original draft, Writing – review & editing. ML: Investigation, Project administration, Writing – review & editing. PC: Investigation, Project administration, Writing – review & editing. SM: Investigation, Project administration, Writing – review & editing. IH: Conceptualization, Investigation, Methodology, Writing – review & editing. MT: Investigation, Project administration, Supervision, Writing – review & editing. RT: Investigation, Project administration, Writing – review & editing. JS: Investigation, Project administration, Writing – review & editing. AC: Investigation, Project administration, Supervision, Writing – review & editing. MB: Investigation, Project administration, Supervision, Writing – review & editing. LH: Investigation, Project administration, Supervision, Writing – review & editing. MS: Investigation, Project administration, Writing – review & editing. WA: Investigation, Project administration, Supervision, Writing – review & editing. PS: Conceptualization, Formal Analysis, Investigation, Methodology, Project administration, Supervision, Validation, Writing – original draft, Writing – review & editing. MO: Conceptualization, Data curation, Formal Analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Supervision, Validation, Visualization, Writing – original draft, Writing – review & editing.