The Use of Corticosteroids or Tocilizumab in COVID-19 Based on Inflammatory Markers.

BACKGROUND: The inflammatory cascade is the main cause of death in COVID-19 patients. Corticosteroids (CS) and tocilizumab (TCZ) are available to treat this escalation but which patients to administer it remains undefined.
OBJECTIVE: We aimed to evaluate the efficacy of immunosuppressive/anti-inflammatory therapy in COVID-19, based on the degree of inflammation.
DESIGN: A retrospective cohort study with data on patients collected and followed up from March 1st, 2020, to May 1st, 2021, from the nationwide Spanish SEMI-COVID-19 Registry. Patients under treatment with CS vs. those under CS plus TCZ were compared. Effectiveness was explored in 3 risk categories (low, intermediate, high) based on lymphocyte count, C-reactive protein (CRP), lactate dehydrogenase (LDH), ferritin, and D-dimer values. PATIENTS: A total of 21,962 patients were included in the Registry by May 2021. Of these, 5940 met the inclusion criteria for the present study (5332 were treated with CS and 608 with CS plus TCZ). MAIN MEASURES: The primary outcome of the study was in-hospital mortality. Secondary outcomes were the composite variable of in-hospital mortality, requirement for high-flow nasal cannula (HFNC), non-invasive mechanical ventilation (NIMV), invasive mechanical ventilation (IMV), or intensive care unit (ICU) admission. KEY
RESULTS: A total of 5940 met the inclusion criteria for the present study (5332 were treated with CS and 608 with CS plus TCZ). No significant differences were observed in either the low/intermediate-risk category (1.5% vs. 7.4%, p=0.175) or the high-risk category (23.1% vs. 20%, p=0.223) after propensity score matching. A statistically significant lower mortality was observed in the very high-risk category (31.9% vs. 23.9%, p=0.049).
CONCLUSIONS: The prescription of CS alone or in combination with TCZ should be based on the degrees of inflammation and reserve the CS plus TCZ combination for patients at high and especially very high risk.

INTRODUCTION

It has become evident that most of the mortality due to this COVID-19 results from the so-called cytokine storm,[1] an exaggerated and uncontrolled inflammatory response that frequently leads to death if not aborted. The 2 most commonly used drugs in the management of COVID-19 are corticosteroids (CS)[2, 3] and tocilizumab (TCZ).[3-17] While CS have been uniformly found helpful,[2, 3] TCZ efficacy has been mixed.[3-17] However, most studies have evaluated efficacy based on oxygenation/ventilation of the patients rather than inflammatory markers. Our study aims to evaluate the efficacy of these drugs based on the degree of inflammation.[18]

METHODS

Study Design, Patient Selection, and Data Collection

This is a retrospective cohort study with data on patients collected and followed up from March 1st, 2020, to May 1st, 2021, from the nationwide Spanish SEMI-COVID-19 Registry. The characteristics of the patients included in this registry have been extensively described previously.[19] This is a multicenter, nationwide registry with over 150 hospitals. All included patients were diagnosed by polymerase chain reaction (PCR) test taken from a nasopharyngeal sample, sputum, or bronchoalveolar lavage. The collection of data from each patient in terms of laboratory data, treatments, and outcomes was verified by the principal investigator of each center through the review of clinical records.

All participating centers in the register received approval from the relevant Ethics Committees, including Bellvitge University Hospital (PR 128/20).

Inclusion Criteria

The group that received only CS was considered the standard of care (SOC) for hospitalized patients. We included patients whose CS use started within the first 72 h after hospital admission and before the onset of high-flow nasal cannula (HFNC), non-invasive mechanical ventilation (NIMV), invasive mechanical ventilation (IMV), or the requirement of intensive care unit (ICU) admission. The CS plus TCZ group included patients who received both drugs in the first 72 h after hospital admission and also before the onset of HFNC, NIMV, IMV, or ICU admission.

Exclusion Criteria

We excluded patients who did not receive CS, or received it more than 3 days after hospitalization, those with a nosocomial infection, and those who died within 24 h.

Treatments Prescribed and Definitions of Groups

We divided the cohort into 2 groups: patients who received solely CS, and patients who received both CS and TCZ. The usual dose of TCZ in our country was 4–8 mg/kg iv, generally in a single dose, although some additional doses are allowed at the discretion of the responsible physician.

Regarding antiviral treatment, the use of antivirals (lopinavir/ritonavir,[20] remdesivir[21]), hydroxychloroquine,[22] and azithromycin[22] was allowed according to the recommendations of the Spanish Ministry of Health.

Degrees of Inflammation

We previously reported the 3 categories of risk: (low, intermediate, and high risk) based on the total lymphocyte count, and the C-reactive protein (CRP), lactate dehydrogenase (LDH), ferritin, and d-dimer values taken at the time of admission (Table S1).[18] The high-risk category was defined based on only 1 of the 5 criteria described above the previously defined cutoff. In addition, for the present study, a very high–risk category was added, defined as the presence of 3 or more high-risk upon admission criteria (Table S1).

Outcome Definition

The primary outcome of our study was in-hospital mortality. Secondary outcomes included length of stay (LOS), and the requirement of HFNC, NIMV, IMV, and ICU admission.

Statistical Analysis

Categorical variables were expressed as absolute numbers and percentages. Continuous variables are expressed as mean plus standard deviation (SD) in the case of parametric distribution or median [IQR] in the case of non-parametric distribution. Differences among groups were assessed using the chi-square test for categorical variables and the t-test or Mann-Whitney test as appropriate for continuous variables. p values< 0.05 indicated statistical significance.

For the study of risk factors associated with in-hospital mortality, univariate and multivariate binary logistic regression was performed. For the latter, variables with p<0.10 in the univariate study plus age and gender were included. Differences in mortality were shown graphically using Kaplan-Meier curves with their log-rank test (event: death; censored data: hospital discharge). Missing data were treated with multiple imputations. To improve the comparability of the groups, propensity score matching (PSM) was performed. This included age, sex, body mass index (BMI), race, smoking behavior, days from onset to admission, all comorbidities, Charlson index, heart rate on admission, tachypnea on admission, PaO2/FiO2, lymphocyte count, CRP, LDH, ferritin, d-dimer, remdesivir treatment, and prescription of low-molecular-weight heparins (LMWH) during admission.

Statistical analysis was performed by IBM SPSS Statistics for Windows, Version 26.0. Armonk, NY, USA: IBM Corp.

RESULTS

General Data and Symptoms Between Groups

A total of 21,962 patients were included in the Registry by May 2021; 9430 were treated with CS or CS plus TCZ and 5940 met our inclusion criteria (5332 were treated with CS and 608 with CS plus TCZ) (Fig. 1). Table 1 shows the differences between the two groups. Those in the isolated CS group were older (72.4 vs. 63.6 years), with less male predominance (58.9% vs. 72.9%), higher prevalence of Caucasian population (90.8% vs. 82.1%), higher degree of dependency (21.8% vs. 4. 7%), hypertension (57.8% vs. 48.8%), chronic heart failure (42.3% vs. 38.5%), chronic renal failure (7.7% vs. 1.1%), dementia (11.3% vs. 1.8%), chronic obstructive pulmonary disease (COPD) (9.8% vs. 6.4%), and a higher Charlson index (1 vs. 0). The groups were similar after propensity score matching (Table 1).

Fig. 1

Flow chart.

Table 1

General Data

	All cohort			Matched-cohort
	CS	CS+TCZ	p value	CS	CS+TCZ	p value
n	5,332	608		607	608
Age, median [IQR]	72.4 [59.7–82.7]	63.6 [54.8–74.2]	<0.001	63.5 [52.4–74]	63.6 [54.8–74.2]	0.542
Gender (males), n (%)	3,140 (58.9)	443 (72.9)	<0.001	419 (69)	443 (72.9)	0.141
Days from onset to admission, median [IQR]	7 [4–9]	7 [5–10]	<0.001	7 [4–10]	7 [5–10]	0.053
BMI, median [IQR]	28.9 [25.7–32.6]	29.1 [25.8–32.5]	0.477	29.2 [25.9–32.8]	29.1 [25.8–32.5]	0.967
Race, n (%)			<0.001			0.300
Caucasian	4,840 (90.8)	499 (82.1)		506 (83.4)	499 (82.1)
Black	35 (0.7)	5 (0.8)		8 (1.3)	5 (0.8)
Hispanic	375 (7)	83 (13.7)		70 (11.5)	83 (13.7)
Asian	17 (0.3)	2 (0.3)		7 (1.2)	2 (0.3)
Others	65 (1.2)	19 (3.1)		16 (2.6)	19 (3.1)
Smoking behavior, n (%)			0.793			0.505
Never smoker	3,545 (66.5)	409 (67.3)		410 (67.5)	409 (67.3)
Former smoker	1,527 (28.6)	173 (28.5)		163 (26.9)	173 (28.5)
Current smoker	260 (4.9)	26 (4.3)		34 (5.6)	26 (4.3)
Degree of dependency, n (%)			<0.001			0.072
None or mild	4,169 (78.2)	580 (95.4)		579 (95.4)	580 (95.4)
Moderate	676 (12.7)	21 (3.5)		27 (4.4)	21 (3.5)
Severe	487 (9.1)	7 (1.2)		1 (0.2)	7 (1.2)
Arterial hypertension, n (%)	3,084 (57.8)	297 (48.8)	<0.001	304 (50.1)	297 (48.8)	0.667
Dyslipidemia, n (%)	2,255 (42.3)	234 (38.5)	0.072	247 (40.7)	234 (38.5)	0.432
Diabetes mellitus, n (%)	1,218 (22.8)	148 (24.3)	0.405	149 (24.5)	148 (24.3)	0.934
Ischaemic cardiopathy, n (%)	460 (8.6)	53 (8.7)	0.940	62 (10.2)	53 (8.7)	0.373
Chronic heart failure, n (%)	451 (8.5)	31 (5.1)	0.004	32 (5.3)	31 (5.1)	0.892
Chronic liver disease, n (%)	208 (3.9)	17 (2.8)	0.176	26 (4.3)	17 (2.8)	0.161
Severe chronic renal failure, n (%)	413 (7.7)	13 (2.1)	0.001	18 (3)	13 (2.1)	0.361
Dementia, n (%)	605 (11.3)	11 (1.8)	<0.001	5 (0.8)	11 (1.8)	0.132
Cancer, n (%)	498 (9.3)	56 (9.2)	0.917	61 (10)	56 (9.2)	0.620
COPD, n (%)	522 (9.8)	39 (6.4)	0.007	33 (5.4)	39 (6.4)	0.470
Asthma, n (%)	416 (7.8)	44 (7.2)	0.621	52 (8.6)	44 (7.2)	0.390
OSAS, n (%)	348 (6.5)	39 (6.4)	0.915	35 (5.8)	39 (6.4)	0.637
Charlson index, median [IQR]	1 [0–2]	0 [0–1]	<0.001	1 [0–2]	0 [0–1]	0.299

BMI body mass index, IQR interquartile range, COPD chronic obstructive pulmonary disease, OSAS obstructive sleep apnea syndrome, CS corticosteroids, CS+TCZ corticosteroids + tocilizumab. Severe chronic renal failure: creatinine >300 mg/dl or dyalisis

The CS group presented less frequently (Table 2) with cough (69% vs. 79.8%), arthromyalgia (26.7% vs. 34.7%), sore throat (8% vs. 11.8%), headache (11.4% vs. 15.8%), fever (77.2% vs. 86), and diarrhea (23.1% vs. 29.8%). There were also differences in heart rate (87 vs. 91 bpm) and tachypnea (39.6% vs. 53.9%).

Table 2

Symptoms and Physical Examination Upon Admission

	All cohort			Matched-cohort
	CS	CS+TCZ	p value	CS	CS+TCZ	p value
Cough, n (%)	3677 (69)	485 (79.8)	<0.001	434 (71.5)	485 (79.8)	0.001
Arthromyalgias, n (%)	1425 (26.7)	211 (34.7)	<0.001	182 (30)	211 (34.7)	0.079
Ageusia, n (%)	574 (10.8)	71 (11.7)	0.493	88 (14.5)	71 (11.7)	0.145
Anosmia, n (%)	484 (9.1)	68 (11.2)	0.090	69 (11.4)	68 (11.2)	0.920
Sore throat, n (%)	428 (8)	72 (11.8)	0.001	69 (11.4)	72 (11.8)	0.796
Headache, n (%)	608 (11.4)	96 (15.8)	0.002	74 (12.2)	96 (15.8)	0.071
Fever, n (%)	4118 (77.2)	523 (86)	<0.001	505 (83.2)	523 (86)	0.173
Dyspnea, n (%)	3625 (68)	433 (71.2)	0.105	434 (71.5)	433 (71.2)	0.913
Diarrhea, n (%)	1231 (23.1)	181 (29.8)	<0.001	163 (26.9)	181 (29.8)	0.259
Vomiting, n (%)	360 (6.8)	44 (7.2)	0.653	42 (6.9)	44 (7.2)	0.829
Abdominal pain, n (%)	285 (5.3)	34 (5.6)	0.798	30 (4.9)	34 (5.6)	0.612
Heart rate, bpm median [IQR]	87 [76–100]	91 [80–104]	<0.001	90 [80–101]	91 [80–104]	0.079
Respiratory rate >20 rpm, n (%)	2109 (39.6)	328 (53.9)	<0.001	314 (51.7)	328 (53.9)	0.439

IQR interquartile range, CS corticosteroids, CS+TCZ corticosteroids + tocilizumab

Lab Tests Between Groups

The CS group had higher PaO2/FiO2 (285.7 vs. 271), lymphocytes (900×106 vs. 820×106), and d-dimer (869.5 ng/ml vs. 717 ng/ml) at admission. Alternatively, they had lower CRP (84.1 mg/l vs. 131 mg/l), LDH (344 U/l vs. 382.5 U/l), and ferritin (857.4 mcg/l vs. 1120.7 mcg/l) (Table 3). These differences disappeared after PSM except for d-dimer, which remained higher in the CS group (842 ng/ml vs. 717 ng/ml).

Table 3

Lab Tests Upon Admission

	All cohort			Matched-cohort
	CS	CS+TCZ	p value	CS	CS+TCZ	p value
PaO2/FiO2, median [IQR]	285.7 [231.9–338.1]	271 [214.3–320.4]	<0.001	266.7 [209.7–319.1]	271 [214.3–320.4]	0.484
Lymphocytes ×106/l, median [IQR]	900 [612.5–1260]	820 [600–1120]	<0.001	850 [600–1200]	820 [600–1120]	0.196
CRP mg/l, median [IQR]	84.1 [34–151.1]	131 [63.3–201]	<0.001	120 [62–192.7]	131 [63.3–201]	0.477
LDH U/l, median [IQR]	344 [262–460.4]	382.5 [306–499]	<0.001	393 [290–530]	382.5 [306–499]	0.797
Ferritin mcg/l, median [IQR]	857.4 [395.9–1596]	1120.7 [636.2–1830.9]	<0.001	1127.6 [528–1990.8]	1120.7 [636.2–1830.9]	0.855
d-Dimer ng/ml, median [IQR]	869.5 [465.3–1992.8]	717 [400–1421.3]	<0.001	842 [487–1702]	717 [400–1421.3]	0.003

CRP C-reactive protein, LDH lactate dehydrogenase, IQR interquartile range, CS corticosteroids, CS+TCZ corticosteroids + tocilizumab

Treatments Between Groups

The treatments received in both groups are shown in Table S2. The CS group less frequently received remdesivir (9.8% vs. 14.3%) as well as intermediate (12.2% vs. 22.5%) or full doses of LMWH (14.3% vs. 22.2%). These differences disappeared after PSM.

The CS regimen was not standard in all patients. There were significant differences between both groups in the maximum dose of prednisone or equivalent (75 mg vs. 100 mg), days of treatment (7 days vs. 8 days), and cumulative dose (400 mg vs. 600 mg).

Outcomes Between Groups

Compared to patients receiving CS alone, in-hospital mortality was significantly lower for combination of CS + TCZ for high risk (27.4% vs. 20%, p<0.001) and very high–risk patients (40.9% vs. 23.9%, p<0.001) (Figs. 2 and 3). There was no difference for those with low/intermediate risk (8.9% vs. 7.4%, p=1.000) (Table 4, Figs. 2 and 3). Our PSM analysis found significantly lower mortality only in the very high–risk category (31.9%, vs. 23.9%, p=0.049) (Table 5; Figure S1). After PSM, we found no differences in the low/intermediate-risk (1.5% vs. 7.4%, p=0.175) or the high-risk category (23.1% vs. 20%, p=0.223).

Fig. 2

In-hospital mortality (%) between groups in the general cohort.

Fig. 3

In-hospital mortality (%) between groups. Kaplan-Meier. Log-rank test=44.3; p<0.001.

Table 4

Outcomes in the General Cohort

	Low/Intermediate risk			High risk			Very high risk
	CS, N=951	CS+TCZ, N=54	p value	CS, N=4381	CS+TCZ, N=554	p value	CS, N=1542	CS+TCZ, N=238	p value
In-hospital death	85 (8.9)	4 (7.4)	1.000	1201 (27.4)	13 (20)	<0.001	630 (40.9)	57 (23.9)	<0.001
HFNC	59 (6.2)	16 (29.6)	<0.001	452 (10.4)	156 (28.3)	<0.001	192 (12.5)	74 (31.2)	<0.001
NIMV	32 (3.4)	10 (18.5)	<0.001	266 (6.1)	103 (18.6)	<0.001	130 (8.4)	39 (16.4)	<0.001
IMV	30 (3.2)	9 (16.7)	<0.001	268 (6.1)	123 (22.2)	<0.001	149 (9.7)	68 (28.6)	<0.001
ICU	47 (4.9)	11 (20.4)	<0.001	375 (8.6)	168 (30.3)	<0.001	188 (12.2)	84 (35.3)	<0.001
Length of stay (days), median [IQR]	8 [5–12]	12 [8–18.5]	<0.001	9 [6–13]	14 [10–20]	<0.001	9 [5–14]	15 [10–23]	<0.001

HFNC high-flow nasal cannula, NIMV non-invasive mechanical ventilation, IMV invasive mechanical ventilation, ICU intensive care unit, IQR interquartile range, CS corticosteroids, CS+TCZ corticosteroids + tocilizumab

Table 5

Outcomes in the Matched-Cohort

	Low/Intermediate risk			High risk			Very high risk
	CS, N=65	CS+TCZ, N=54	p value	CS, N=542	CS+TCZ, N=554	p value	CS, N=257	CS+TCZ, N=238	p value
In-hospital death	1 (1.5)	4 (7.4)	0.175	125 (23.1)	111 (20)	0.223	82 (31.9)	57 (23.9)	0.049
HFNC	3 (4.6)	16 (29.6)	<0.001	87 (16.1)	156 (28.2)	<0.001	50 (19.5)	74 (31.1)	0.003
NIMV	1 (1.5)	10 (18.5)	0.002	57 (10.5)	103 (18.6)	<0.001	34 (13.2)	39 (16.4)	0.322
IMV	2 (3.1)	9 (16.7)	0.022	75 (13.8)	123 (22.2)	<0.001	46 (17.9)	68 (28.6)	0.005
ICU	4 (6.2)	11 (20.4)	0.026	102 (18.8)	168 (30.3)	<0.001	59 (23)	84 (35.3)	0.002
Length of stay (days), median [IQR]	8 [4.5–12]	12 [8–18.5]	0.001	10 [6.8–16]	14 [10–20]	<0.001	11 [6–20]	15 [10–23]	<0.001

HFNC high-flow nasal cannula, NIMV non-invasive mechanical ventilation, IMV invasive mechanical ventilation, ICU intensive care unit, IQR interquartile range, CS corticosteroids, CS+TCZ corticosteroids + tocilizumab

HFNC use was lower in CS patients in all 3 risk categories: 6.2% vs. 29.6% (p<0.001) for the low/intermediate-risk category, 10.4% vs. 28.3% (p<0.001) for the high-risk category, and 12.5% vs. 31.2% (p<0.001) for the very high–risk category (Table 4). After PSM, we found the same differences between groups in the 3 risk categories: 4.6% vs. 29.6% (p<0.001), 16.1% vs. 28.2% (p<0.001), and 19.5% vs. 31.1% (p=0.003) (Table 5).

NIMV use was also lower in CS patients in all 3 risk categories: 3.4% vs. 18.5% (p<0.001) for the low/intermediate-risk category, 6.1% vs. 18.6% (p<0.001) for the high-risk category, and 8.4% vs. 16.4% (p<0.001) for the very high–risk category (Table 4). After PSM, we found similar differences between groups in the 3 risk categories: 1.5% vs. 18.5% (p<0.001), 10.5% vs. 18.6% (p<0.001), and 13.2% vs. 16.4% (p=0.322).

The use of IMV was also lower in CS patients in all 3 risk categories: 3.2% vs. 16.7% (p<0.001) for the low/intermediate-risk category, 6.1% vs. 22.2% (p<0.001) for the high-risk category, and 9.7% vs. 28.6% (p<0.001) for the very high–risk category (Table 4). After PSM, we found similar differences between groups in the 3 risk categories: 3.1% vs. 16.7% (p<0.001), 13.8% vs. 22.2% (p<0.001), and 17.9% vs. 28.6% (p=0.005) (Table 5).

The need for ICU admission was lower in patients with CS in all 3 risk categories: 4.9% vs. 20.4% (p<0.001) for the low/intermediate-risk category, 8.6% vs. 30.3% (p<0.001) for the high-risk category, and 12.2% vs. 35.3% (p<0.001) for the very high–risk category (Table 4). After PSM, we found similar differences between groups in the 3 risk categories: 6.2% vs. 20.4% (p<0.001), 18.8% vs. 30.3% (p<0.001), and 23% vs. 35.3% (p=0.002) (Table 5).

Finally, median LOS was higher in patients with CS+TCZ in all 3 risk categories: 8 days vs. 12 (p<0.001) for the low/intermediate-risk category, 9 days vs. 14 (p<0.001) for the high-risk category, and 9 days vs. 15 (p<0.001) for the very high–risk category (Table 4). After PSM, we found similar differences between groups in the 3 risk categories: 8 days vs. 12 (p=0.001), 10 days vs. 14 (p<0.001), and 11 days vs. 15 (p<0.001) (Table 5).

Risk Factors for In-Hospital Mortality

The independent risk factors for mortality in the high-risk category were age, male sex, moderate/severe dependency, higher Charlson index, tachypnea on admission, and lower PaO2/FiO2 (Table 6). The use of TCZ showed a trend of benefit that did not reach statistical significance as an independent protective factor. The very high–risk category showed similar results (data not shown). The AUC of the final model was 0.792 (Figure S2).

Table 6

Risk Factors in the Matched-Cohort. High-Risk Category

	Univariate analysis		Multivariate analysis
	OR (95% CI)	p value	OR (95% CI)	p value
Age	1.07 (1.05–1.09)	<0.001	1.08 (1.06–1.09)	<0.001
Gender (female)	1.16 (0.74–1.81)	0.525	0.69 (0.47–0.99)	0.045
BMI	1.02 (0.98–1.06)	0.405
Race			NS
Caucasian (ref.)	1 ref.
Black	1.10 (0.90–1.15)	0.999
Hispanic	0.46 (0.21–1.01)	0.054
Asian	2.37 (0.15–38.14)	0.543
Others	0.79 (0.16–3.96)	0.774
Moderate/severe dependency	4.90 (2.77–8.68)	<0.001	1.98 (1.01–3.90)	0.048
Arterial hypertension	2.33 (1.55–3.49)	<0.001	NS
Dyslipidemia	1.75 (1.18–2.59)	0.006	NS
Diabetes mellitus	1.81 (1.15–2.83)	0.010	NS
Ischaemic cardiopathy	2.06 (1.08–3.93)	0.028	NS
Chronic heart failure	7.59 (2.90–19.83)	<0.001	NS
Chronic liver disease	1.10 (0.42–2.93)	0.845
Severe chronic renal failure	4.04 (1.41–11.57)	0.009	NS
Dementia	5.24 (0.95–28.97)	0.057	NS
Cancer	2.54 (1.41–4.59)	0.002	NS
COPD	2.49 (1.11–5.61)	0.027	NS
Asthma	0.58 (0.25–1.34)	0.202
OSAS	1.67 (0.80–3.39)	0.176
Charlson index	1.28 (1.15–1.43)	<0.001	1.14 (1.04–1.24)	0.003
Respiratory rate >20 rpm	2.29 (1.49–3.54)	<0.001	2.01 (1.41–2.87)	<0.001
PaO2/FiO2	0.99 (0.99–0.99)	<0.001	0.99 (0.99–0.99)	<0.001
Tocilizumab	0.67 (0.45–0.99)	0.050	0.78 (0.56–1.09)	0.150

BMI body mass index, NS not significant, COPD chronic obstructive pulmonary disease, OSAS obstructive sleep apnea syndrome

DISCUSSION

We found that higher degrees of inflammation responded to combination therapy, consistent with COVID-19 as an inflammatory disease. Treatment should be risk-stratified based on inflammation. At present, the approach to the disease has been heterogeneous and often based on oxygenation/ventilation status. In order to evaluate the efficacy of immunosuppressive/anti-inflammatory treatments, we have to include the degree of inflammation in patients to judge efficacy. The degree of inflammation in most studies is difficult to assess and appears to include many patients with low degrees of inflammation. It is thus difficult to know the real efficacy of these drugs and explain differences in efficacy between observational studies and clinical trials [23].

Our results suggest that the greater the inflammation, the more effective these drugs will be. Our group previously described 3 categories of inflammation based on 5 parameters at admission (lymphopenia, CRP, LDH, ferritin, and d-dimer).[18] Since the low-risk category rarely requires hospital admission, it is the least numerous in our national series.

Our study shows that the addition of TCZ does not provide benefit in the low/intermediate-risk category. While the combination reduced mortality in the high-risk group, we did not achieve statistical significance, due to inadequate power. Patients classified as very high risk (3–5 high-risk criteria) had statistically significant reduction in death.

Our secondary outcomes (use of HFNC, NIMV, and IMV, and admission to the ICU) suggest that the CS+TCZ group had more severe disease despite PSM. The sociodemographic, clinical, and analytical data included could not fully capture patient severity. Our patients were on oxygen therapy (not high-flow) at the time of treatment initiation (CS patients or CS+TCZ patients). However, we do not know the exact FiO2 they were receiving; it is possible that the combination of CS+TCZ was used in patients requiring higher amounts of oxygen.

Our study strengths include that it is large and nationally representative. In addition, the therapeutic approach based on degree of inflammation is a good approximation to clinical practice decision-making.

Our study also has some limitations. First, it is a retrospective study. Second, it comes from a multicenter registry, with the heterogeneity that this implies, though we used standardized definitions. Another limitation to be taken into account is the heterogeneity in CS dosage and administration time as well as lack of information of important variables that might trigger addition of TCZ, such as oxygen requirement.

In conclusion, the prescription of CS alone or in combination with TCZ should be based on the degrees of inflammation and reserve the CS plus TCZ combination for patients at high and especially very high risk.

(DOCX 26 kb)

(DOCX 74 kb)