Background

Coronavirus 2019 (COVID-19) is a life-threatening disease that is caused by SARS-CoV-2 (WHO 2020). Disease spectrum of COVID-19 is ranged from asymptomatic to severe pneumoniae leading acute respiratory failure and even death (Lai et al. 2020). Hyperinflammatory state (cytokine storm) developed by excessive release of several pro-inflammatory cytokines such as interleukin 1 (IL-1), tumor necrosis factor alpha (TNF-alpha) and interleukin-6 (IL-6) (Hu et al. 2021). In COVID-19, severe disease course such as need of intensive care unit (ICU) as well as mortality is mainly due to development of cytokine storm (Tang et al. 2020).

COVID-19-associated cytokine storm develops with decreasing of type 1 interferon production via inhibition toll-like receptor-3 (TLR-3) signaling by virus accessory proteins and, therefore, virus can escape the immune response. Defective clearance of SARS-CoV-2 and virus–host interaction induces the overproduction of immune mediators that leads to increase in secretion of chemokines and cytokines including IL‐1, IL‐6, IL‐8, IL‐21 and TNF in infected cells (Seyed Hosseini et al. 2020). There is also another pathway into the development of COVID-19-associated cytokine storm. SARS-CoV-2 activates an intracellular multiprotein complex that is called ‘inflammasome’ after binding TLRs into the development of cytokine storm in severe COVID-19 (López-Reyes et al. 2020). Inflammasomes have essential role in the host defense against microorganisms including viruses that are present in multiple innate immune cells such as neutrophils, macrophages and dendritic cells. Activation of inflammasomes is leading to the cleavage of pro-IL-1β to produce active IL-1β (Vora et al. 2021). Anakinra, an IL-1 receptor antagonist, is commonly used in the treatment of autoinflammatory diseases such as hereditary periodic fever syndromes (HPFS), gout and adult-onset still disease (AODS) (Saag et al. 2021; Atas et al. 2021; Hong et al. 2014). Safety and efficacy of anakinra was established in several hyperinflammatory conditions such as hemophagocytic lymphohistiocytosis (HLH) and macrophage activation syndrome (MAS) secondary to various diseases including malignancies, viruses and rheumatological conditions (Bami et al. 2020; Gilboa et al. 2020). Recently, anakinra also became an emerging treatment option against to COVID-19-associated cytokine storm (Bozzi et al. 2021).

In this study, we aimed to evaluate the high-dose intravenous anakinra treatment response and outcome in patients with severe and critical COVID-19 compared to standard of care.

Methods

Patients and data

This retrospective observational study was carried out at a tertiary referral center in Turkey. All patients had positive polymerase chain reaction (PCR) and delta variant (B.1.617.2) proven by variant analysis as well as typical computer tomography (CT) findings in addition to clinical signs and symptoms. Patients with negative PCR results and/or variants rather than delta and/or inconsistent CT findings with COVID-19 were excluded from the study.

The study population consisted of two groups as follows: the patients receiving high-dose intravenous anakinra (anakinra group) added to background therapy between 01.09.2021 and 01.02.2022 and the patients treated with standard of care (SoC, control group) as historical control group who were hospitalized between 01.07.2021 and 01.09.2021. COVID-19 disease severity was evaluated according to National Institute of Health (NIH) severity scale and only severe (NIH score 3; patients with SpO2 < 94% on room air at sea level, PaO2/FiO2 < 300 mm Hg, a respiratory rate > 30 breaths/min, or lung infiltrates > 50%) and critically ill (NIH score 4; patients have acute respiratory distress syndrome, septic shock that may represent virus-induced distributive shock, cardiac dysfunction, an exaggerated inflammatory response, and/or exacerbation of underlying comorbidities, as well as requirement of high-flow nasal oxygen therapy [HFNO] or invasive mechanical ventilation [IMV]) patients in the ward were included into the study (COVID-19 Treatment Guidelines Panel. Coronavirus Disease (2019) (COVID-19) Treatment Guidelines 2022).

Individual written patient consent and local ethic committee approval were obtained for this study (date/number: 24.02.2022, 2022/04-09).

Laboratory evaluation

Laboratory values such as hemogram, liver enzymes, C-reactive protein (CRP), ferritin, d-dimer, lactate dehydrogenase (LDH) at admission; the peak levels of CRP, ferritin, d-dimer and LDH levels were recorded. Inflammatory state of the patients was evaluated and derived based on COVID hyperinflammatory syndrome score (cHIS) and it was calculated according to combination of neutrophil and lymphocyte counts at the admission and the peak levels of CRP, ferritin, d-dimer and LDH during to the follow-up (Webb et al. 2020). The item of fever was removed due to presence of lower frequency (< %10) of the patients in both arms. Therefore, maximum score of the new version of cHIS score was 5 points (modified cHIS [mcHIS] score) which were calculated in both the groups.

Treatment protocol and outcome

All patients received background corticosteroid therapy with 80 mg/day methylprednisolone (or its equivalent) and enoxaparin 0.4 mg/day at the admission and consecutive days (SoC). Anakinra was started in patients who did not respond to steroid therapy at least 2 days or concomitantly with steroids in patients with higher risk and/or critical illness at admission. Average starting dose of anakinra was 400 mg/day intravenously and increased gradually to maximum 1600 mg/day if necessary (10 mg/kg/day). Anakinra dose adjustment was performed by the same rheumatologist (MB) according to daily clinical and laboratory findings. Severe infection was defined as development of opportunistic infection, intravenous antibiotics, sepsis or requirement of intensive care unit (ICU) admission or development of death due to secondary infection.

Statistical analysis

In our study, 22.0 version (IBM, Armonk, NY, USA) of SPSS (Statistical Package for the Social Sciences) program was used for statistical analysis of data. Descriptive statistics, discrete and continuous numerical variables were expressed as mean  ± standard deviation or median (minimum–maximum). Categorical variables were expressed as number of cases (%). Cross table statistics were used to compare categorical variables (Chi-square, Fisher’ exact test). Normally distributed parametric data were compared with Student’s t-test and non-parametric data that did not meet normal distribution were compared with Mann–Whitney U and Kruskal–Wallis tests. Correlation analysis was performed by Pearson or Spearman method according to normality distribution. Kaplan–Meier and log-rank methods were used for survival analysis. Multivariate analysis was performed using logistic regression. p < 0.05 value was considered statistically significant.

Propensity score matching

The first step in propensity score matching (PSM) is to identify the covariates from which to calculate propensity scores (PS). These variables, which can explain the output variable, should be included in the propensity score creation process (Fan and Nowell 2011). Age, gender, baseline NLR, ferritin, CRP, D-dimer, LDH levels and NIH COVID disease scores of the patients were determined as the variables to be matched. The PS matching was done as 1:1 with nearest neighbor method. Caliper value was 0.2. When matching, we performed this analysis by assigning values according to the averages of the parameters with missing data. PSM was performed with SPSS package program 28.0.1 using R package program and an auxiliary plugin (PS matching 3.0 SPE). Dotplot of standardized mean differences for all covariates before and after PS matching is shown in Fig. 1. Jitter plots for trend scores and lineplot of standardized differences are described in supplemental figure 1 and 2, respectively.

Fig. 1
figure 1

Dotplot of standardized mean differences for all covariates before and after PS matching

Full size image

Results

Initial analysis with study group before PS matching

Data of 148 patients in anakinra and 114 patients in control group were analyzed. Flow-chart of the study design and participants is shown in Fig. 2. Mean ± standard deviation (SD) patient age was 66.8 ± 17 and 63.1 ± 17 years in anakinra and SoC groups, respectively (p = 0.09). Male gender was 78 (52.7%) in anakinra and 45 (39.5%) SoC (p = 0.033; Odds ratio [OR] 4.5). Median (IQR) duration of hospitalization was 11 (12) days and 9 (7.3) days in anakinra and SoC groups, respectively (p = 0.02). Diabetes mellitus (DM) was in 28% (n = 41) and 34.2% (n = 39), hypertension (HT) in 58.7% (n = 84) and 56% (n = 64), coronary heart disease (CHD) in 27% (n = 19) and 24% (n = 21), heart failure (HF) in 12.6% (n = 18) and 20% (n = 23), chronic kidney failure (CKD) in 21% (n = 31) and 5.3 (n = 6), chronic obstructive pulmonary disease (COPD) in 16% (n = 23) and 16.7% (n = 19), dementia in 12.8% (n = 15) and 1.8% (n = 2), malignancy in 11% (n = 16) and 7% (n = 8), and immunosuppressive usage in 12.3% (n = 18) and 1.8 (n = 2) in anakinra and control groups, respectively. Baseline clinical and laboratory findings of participants in anakinra and control groups before and after PSM are described in Table 1.

Fig. 2
figure 2

Flow-chart of the study design and participants

Full size image
Table 1 Baseline clinical and laboratory features of patients receiving standard of care (SoC) and anakinra before and after propensity score (PS) matching
Full size table

Fifty-seven (38.5%) and 68 (59.6%) patients had severe, 91 (61.5%) and 46 (40.4%) had critical disease in anakinra and control groups, respectively (p = 0.001; OR 11.5). Mean mcHIS score was 3.4 ± 1.2 and 2.64 ± 1.5 in patients receiving anakinra and SoC, respectively (p = < 0.001). Overall, pneumothorax was developed in 2.2 (n = 3) and 0, myocardial infarction (MI) in 2.3% (n = 3) and 5.3 (n = 6), pulmonary embolism (PE) in 3% (n = 4) and 9.6% (n = 11), ICU admission was in 40.5% (n = 60) and 22% (n = 25), intubation was in 36.5% (n = 54) and 11.4% (n = 13) patients, and 37.8% (n = 56) and 23.7% (n = 27) patients died in anakinra and control groups, respectively (Table 2).

Table 2 Outcomes of patients receiving SoC and anakinra before and after PS matching
Full size table

In univariable analysis, higher patient age (p < 0.001 and p = 0.002), higher frequency of critical illness (p < 0.001; OR 52.3 and p < 0.001; OR 51.3), higher mcHIS score (p < 0.001 and p < 0.001), higher peak levels of CRP (p < 0.001 and p < 0.001), ferritin (p < 0.001 and p < 0.001), d-dimer (p < 0.001 and p < 0.001) and LDH (p < 0.001 and p < 0.001) were observed in patients who had mortality in SoC and anakinra groups, respectively. Higher frequency of male gender (p = 0.004; OR 4.8), coronary artery disease (p = 0.001; OR 11.6) and ALT levels (p = 0.009) in SoC, and higher frequency of dementia (p = 0.002; OR 9.8), higher NLR (p < 0.001), baseline CRP (p = 0.004) and d-dimer (p = 0.016) levels were observed in anakinra group in patients who had mortality (Table 3).

Table 3 Univariable analysis of mortality in patients receiving SoC and anakinra before PS matching
Full size table

Analysis of study group after PS matching

After PS procedure, 78 patients in anakinra and 78 patients in SoC matched and were included into the analysis. Mean ± SD patient age was 67.4 ± 16.7 and 67.1 ± 16.3 years in anakinra and SoC groups, respectively (p = 0.9). Male gender was 38 (48.7%) in anakinra and 36 (46.2%) SoC (p = 0.8). Median (IQR) duration of hospitalization were 7.5 (9) days and 11 (8) days in anakinra and SoC groups, respectively (p = 0.01). Frequency of DM was in 23% (n = 18) and 39.7% (n = 31), HT in 61.5% (n = 30) and 64% (n = 50), CHD in 23% (n = 18) and 25.6% (n = 20), HF in 18% (n = 14) and 25.6% (n = 20), CKD in 19% (n = 15) and 7.7 (n = 6), COPD in 18% (n = 14) and 19% (n = 15), dementia in 5% (n = 3) and 2.6% (n = 2), malignancy in 11.5% (n = 9) and 7.7% (n = 6), and immunosuppressive usage in 6.5% (n = 5) and 2.6 (n = 2) in anakinra and control groups, respectively. The differences were only significant in patients who had DM (p = 0.025; OR 5) and CRF (p = 0.035; OR 4.5) and those who did not among comorbidities after PSM (Table 1).

Forty-eight (61.5%) and 44 (56.4%) patients had severe, 30 (38.5%) and 34 patients (43.6%) had critical disease in anakinra and control groups, respectively (p = 0.5). Mean mcHIS score was 2.9 ± 1 and 3.1 ± 1.3 in patients receiving anakinra and SoC, respectively (p = 0.2). Overall, pneumothorax was developed in 2.7 (n = 2) and 0 (p = 0.5), myocardial infarction (MI) in 2.8% (n = 2) and 3.6 (n = 2) (p = 1), pulmonary embolism (PE) in 4.1% (n = 3) and 9% (n = 7) (p = 0.3), ICU admission was in 14.1% (n = 11) and 30.8% (n = 24) (p = 0.013; OR 6.2), intubation in 12.8% (n = 10) and 16.7% (n = 13) patients (p = 0.5), and 14.1% (n = 11) and 32.1% (n = 25) patients died in anakinra and control groups, respectively (p = 0.008; OR 7.1) (Table 2).

In univariable analysis, higher patient age (p = 0.005 and p < 0.001), higher frequency of critical illness (p < 0.001; OR 47.6 and p < 0.001; OR 14.9), higher mcHIS score (p = 0.005 and p = 0.036), higher peak levels of ferritin (p = 0.005 and p = 0.04) and LDH (p = 0.002 and p = 0.01) were observed in patients had mortality in SoC and anakinra groups, respectively. Higher frequency of CHD (p = 0.002; OR 9.6), higher peak CRP (p = 0.006) and d-dimer levels (p < 0.001) in SoC; higher frequency of CRF (p = 0.005; OR 10.3) were observed in anakinra group in patients who had mortality (Table 4).

Table 4 Univariable analysis of mortality in patients with SoC and anakinra after PS matching
Full size table

Higher patient age (OR 1.048, p = 0.018 [95% confidence interval (CI) 1.008–1.09]), male gender (OR 3.2, p = 0.048 [95% CI 1.013–10.3]), critical disease (compared to severe) (OR 141.9, p < 0.001 [95% CI 16.8–1198]), and SoC treatment (compared to anakinra) (OR 6.7, p = 0.002 [95% CI 1.98–22.3]) were associated with higher mortality in multivariable analysis after PSM (Table 5).

Table 5 Multivariable analysis of mortality in patients receiving anakinra after PS matching
Full size table

In survival analysis, mortality rate was tended to be lower in patients with anakinra compared to SoC group after PSM (Log-Rank: p = 0.14) (Fig. 3).

Fig. 3
figure 3

Comparison of mortality between patients receiving anakinra and standard of care (SoC) in survival analysis

Full size image

Discussion

In COVID-19, mortality rates differ according to the features of the study population (hospitalized vs outpatient patients or ward vs ICU), inflammatory burden and disease severity of patients. Several studies revealed higher mortality rate among severe and critically COVID-19 patients. In a retrospective study from Saudi Arabia with 352 critically ill patients, 28-day mortality was 32.1%. In this study, mortality was associated with older age, higher LDH and d-dimer levels (Alharthy et al. 2021). In a multicenter observational study from Mexico with 164 critically ill patients, mortality was 51.8%, and higher age as well as CRP levels were also associated with higher mortality (Ñamendys-Silva et al. 2021). In Zhao et al. study, mortality was 2% and 47% among severe and critically ill patients, respectively (Zhao et al. 2021). Beyond the demographic features, there is a close association between poor outcome and higher inflammatory response. A hyperinflammatory score (cHIS score) reflecting cytokine storm was defined by Webb et al. (Hu et al. 2021). In this study, mortality increased gradually with higher scores; those were 1%, 15% and 33% in patients who had < 2, ≥ 2 and ≥ 3 scores, respectively. The fact that an average higher mcHIS score of current study population than original study even removing item of fever reflects a higher inflammatory burden in our study compared to the previous one. Therefore, our study group had more severe disease and higher mortality risk at admission which was expected to have a higher mortality rate. In our study, mortality rate of control group was consistent with previous results, although mortality rate was higher in patients receiving anakinra compared to SoC before PS matching, which was lower in anakinra group than SoC after PS matching. This finding emphasizes the importance of making comparisons in terms of relevant treatment efficacies between similar disease groups according to patient characteristics and disease severity.

Several risk factors for higher mortality were defined such as age and gender, presence of cancer or on immunosuppression in COVID-19 (Alwani et al. 2021; Aboueshia et al. 2021). In addition to traditional risk factors, some laboratory parameters such as elevated CRP, ferritin, d-dimer, LDH and IL-6 levels were associated with poor outcomes including higher mortality (Terpos et al. 2020; Hanff et al. 2020; Feld et al. 2020). These laboratory values are also compatible with other reason of MAS beyond the COVID-19 and reflect higher inflammatory and/or prothrombotic state (Zou et al. 2020; Gavand et al. 2017). Higher baseline and peak levels of NLR, CRP, ferritin, d-dimer and only peak levels of LDH were associated with mortality both in two groups before PS matching. These findings are compatible with previous results. On the other hand, after PS matching, only peak levels of CRP, ferritin, d-dimer and LDH levels remained higher in patients who deceased compared to the survivors. These results indicate that peak levels of such laboratory parameters are more relevant than those baseline levels into the development of mortality in patients with COVID-19-associated cytokine storm. These results were consistent in both treatment groups in our study and further studies are needed to clarify this issue.

IL-1 has a driven role in several diseases such as HPFS, AODS and also several hyperinflammatory conditions such as MAS and HLH due to various immune-related disorders, infections and malignancies (Gabay et al. 2010; Dinarello 2011). In addition, safety and efficacy of IL-1 inhibitors such as anakinra were established in many studies in cytokine storms due to various diseases (Giavridis et al. 2018; Grom et al. 2016; Mehta et al. 2020). Furthermore, anakinra became one of the standard targeted therapies for the treatment of COVID-19-associated cytokine storm during the pandemic (Kyriazopoulou et al. 2021). In a comparative study from Italy with 392 patients which consisted of 62 anakinra and 55 anti-IL-6 treatments, overall mortality was 25%, 32% in biologic-naïve, 14% in anakinra and 18% in tocilizumab-receiving patients (Cavalli et al. 2021). In this study, higher CRP levels at baseline and decreasing levels of LDH with treatment predicted higher IL-1 and IL-6 inhibitor response and reduced mortality. In a prospective comparative study, mortality was lower in patients receiving anakinra compared to those with standard of care (29% vs 46%; p = 0.082) (Balkhair et al. 2021). These studies revealed higher efficacy of anakinra which were consistent with our results in patients with COVID-19.

Intravenous and high-dose anakinra is an emerging therapeutic option both in rheumatology and COVID-19 practice. Intravenous administration of anakinra enables higher and fast maximum plasma concentration compared to subcutaneous administration (Mehta et al. 2020). Recently, there is sufficient evidence with high-dose intravenous anakinra administration in various hyperinflammatory syndromes (Nigrovic et al. 2011; Phadke et al. 2021). High-dose intravenous anakinra treatment was safe and effective in a preliminary retrospective study from Italy in patients with COVID-19 (Cavalli et al. 2020). In another prospective controlled study (ESCAPE open label study) with critically ill COVID-19 patients, high-dose intravenous anakinra had lower mortality rate than standard of care as well as tocilizumab treatment (Karakike et al. 2022). No increase in the frequency of severe infection as well as other complications such as myocardial infarction in anakinra group compared to controls in our study indicates that high-dose intravenous anakinra is safe in patients with COVID-19. Daily dose adjustment of anakinra may allow early intervention of the hyperinflammatory state according to daily clinical and laboratory parameters, as well as withdrawing the drug in case of infection and/or other potential complications.

This study has some limitations. Retrospective design of the study was the main limitation. Having missing data is another limitation of the study. On the other hand, controlled design of the study with PS matching and the fact that the study is conducted in a single center ensures homogeneity in terms of patient population and treatment decisions that are made by a single physician.

Conclusion

In our study, mortality was lower in patients receiving anakinra compared to SoC. Our study also indicates that hyperinflammatory response is one of the most important risk factor for the development of mortality in patients with COVID-19. Intravenous high-dose anakinra is safe and effective treatment in patients with severe and critical COVID-19.