Early clinical experience with imatinib in COVID-19: Searching for a dual effect

Graphical abstract

Dear editor

We have read with interest the recent articles by Chen et al.¹ and Chan and colleagues,² in which the antiviral action of favipiravir and sofosbuvir/daclatasvir against SARS-CoV-2 is assessed. Data reported by the authors are certainly instructive, but, in our opinion, finding drugs not only with antiviral effects, but also with immunomodulatory properties, would be more advantageous, since a heightened immune response has been observed in severe COVID-19.³ In this regard, tyrosine kinase inhibitor (TKI) imatinib, originally designed to treat chronic myeloid leukaemia (CML), has been also proposed as a possible therapy for COVID-19⁴ given its antiviral and immunomodulatory effects observed in preclinical models, as well as the lower incidence of SARS-CoV-2 infection reported among CML patients who were being treated with TKIs.⁵

In March 2020, due to lack of robust data favouring any specific therapy for COVID-19, and after reviewing the theoretical framework about the potential utility and safety of imatinib, it was considered as a possible off-label treatment in some of our patients in order to perform a preliminary analysis of both its clinical effects and safety profile that could subsequently support the design of the COVINIB Study Group (CSG) clinical trial (NCT04346147), which is currently ongoing.

In this letter, we would like to briefly describe the characteristics of twenty consecutive COVID-19 patients admitted to our hospital from March to May 2020 who received imatinib as part of their treatment for SARS-CoV-2 infection. As in other studies, our series involved the off-label administration of a drug for a transmissible disease which had become pandemic even before any therapy was found effective. In such scenarios, WHO offers ethical guidance. In order to meet its recommendations, previous evidence regarding potential effectiveness and safety of imatinib in COVID-19 was carefully reviewed, leading to the creation of the CSG. Informed consent was obtained from each patient before imatinib was started, and the study was approved by the Hospital Ethics Committee. A dosage of 400 mg daily was chosen since it is a usual initial dose in hematological disorders. Treatment duration, however, depended on the clinical course, the appearance of possible adverse reactions and the criteria of the attending physician. Data regarding previous medical history and clinical, radiological and laboratory information were collected.

Table 1 summarises the main characteristics of patients. The median age was 73 years, and 45% of patients were male. The prevalence of hypertension, hyperlipidaemia, diabetes mellitus and previous cardiac and pulmonary disease was 65%, 45%, 30%, 20% and 20%, respectively; four subjects were ex-smokers, and only one was a current smoker. All subjects but one (number 7) had pneumonia, which was bilateral in eighteen patients. Polymerase chain reaction test for SARS-CoV-2 in nasopharyngeal swab was positive in all cases except one (number 13), in whom the diagnosis of COVID-19 was made on the basis of clinical, radiological and laboratory features, according to WHO case definitions at that time.

Table 1.

Characteristics of COVID-19 patients treated with imatinib. Subjects are arranged chronologically according to imatinib start date.

Case	Age (yr)	Sex	Days of symptoms* (Md: 8; IQR: 6–9.3)	Oxygen therapy†	C-reactive protein (Highest, mg/dL) (Md: 14.8; IQR: 11–23.5)	Lymphocytes (Lowest, 109/L) (Md: 0.74; IQR: 0.5–1.18)	D-dimer (Highest, ng/mL) (Md: 1230.5; IQR: 523–1922.5)	Days of imatinib (Md: 9.5; IQR: 7–11.8)	Days of hospitalization (Md: 10; IQR: 7–18)	Other therapies		Final vital status
										Before imatinib	After/with imatinib
1‡	38	F	12	LFNC	21.2	0.84	1129	5	11	HCQ, LPV/r	–	Recovered
2	60	F	9	Mask	16.7	0.88	5650	15	27	HCQ, LPV/r	C	Recovered
3	87	M	5	Mask	30.7	0.36	1501	4	13	HCQ, LPV/r, AZM	C	Dead
4	76	M	8	Mask	18.1	0.31	20,539	5	10	HCQ, LPV/r	C	Dead
5	74	M	9	Mask	30.8	0.52	359	7	10	HCQ, LPV/r	C	Dead
6	58	M	9	LFNC	11.5	1.20	488	13	7	HCQ, LPV/r	–	Recovered
7	75	F	6	LFNC	0.9	1.45	392	8	6	–	HCQ	Recovered
8	40	M	7	LFNC	26.3	1.41	628	15	7	LPV/r	HCQ	Recovered
9	76	F	6	IMV	24.1	0.29	22,787	3	28	HCQ, LPV/r	C, TCZ	Recovered
10	67	F	8	LFNC	1.5	0.77	1234	10	8	HCQ, LPV/r	C	Recovered
11	83	M	6	LFNC	11.0	0.66	281	13	12	HCQ, LPV/r	–	Recovered
12	76	M	13	No need	12.5	0.71	2388	9	3	HCQ, LPV/r	–	Recovered
13	71	F	8	Mask	21.6	0.50	1837	11	22	HCQ, LPV/r, AZM, C, TCZ	–	Recovered
14	74	M	11	Mask	30.2	0.45	1927	11	21	HCQ, LPV/r, AZM, C, TCZ	–	Recovered
15	72	M	8	LFNC	5.3	1.20	1227	10	4	HCQ, LPV/r	–	Recovered
16	65	F	10	Mask	17.9	1.11	669	8	10	HCQ, LPV/r	C, TCZ	Recovered
17	52	F	8	LFNC	11.0	0.52	292	12	10	HCQ, LPV/r	–	Recovered
18	83	F	2	Mask	12.9	0.55	1909	11	17	HCQ, AZM, C	–	Recovered
19	79	F	8	LFNC	12.5	1.01	1666	7	18	–	HCQ	Recovered
20	71	F	8	No need	1.8	1.38	720	7	11	–	HCQ	Recovered

Abbreviations: yr = years; Md = median value; IQR = interquartile range; F = female; M = male; LFNC = low-flow nasal cannula; Mask = Venturi mask or nonrebreather face mask; IMV = invasive mechanical ventilation; HCQ = hydroxychloroquine; LPV/r = lopinavir/ritonavir; AZM = azithromycin; C = corticosteroids; TCZ = tocilizumab. Reference range for C-reactive protein: <0.5 mg/dL. Reference range for D-dimer: <500 ng/mL.

*Days from symptom onset to treatment with imatinib.

†Highest oxygen support needed during admission.

‡This case has been previously published (Morales-Ortega et al.)⁴.

Hydroxychloroquine (HCQ) and lopinavir/ritonavir (LPV/r), were the most common concomitant therapies. On the basis of the latest evidence, these two drugs seem to have had slight impact, if any, on patients’ outcomes. On the contrary, corticosteroids have been found useful in COVID-19;⁶ however, according to our protocol at that time, they were indicated in patients whose clinical course deteriorated despite other treatments. Half of the cases were considered to have a favourable clinical course, and subsequently did not receive corticosteroids.

Six patients reported mild gastrointestinal symptoms (which are among the most common side effects of imatinib): all of them had nausea, and one had also diarrhea. No other adverse events were recorded, and this TKI was not discontinued in any patient because of side effects. These data seem to be in line with the safety profile of imatinib observed in CML patients: adverse reactions are infrequent and usually related to long-term therapy. Furthermore, digestive complaints are also associated with other treatments (such as HCQ or LPV/r), and could even be part of the clinical spectrum of COVID-19.

Three patients died in our series. All of them were male, over the age of 70 years and had previous comorbidities. In addition, they all needed supplemental oxygen and had bilateral pneumonia and poor prognostic laboratory features at imatinib onset. All these factors might also explain their fatal outcome, which could not be directly linked to imatinib administration.

Six patients received imatinib within the first week from the beginning of the symptoms. In this scenario, the drug might display antiviral properties (observed in in vitro studies against coronaviruses closely related to SARS-CoV-2).⁷ However, there are conflicting data regarding the real antiviral effect of imatinib against SARS-CoV-2.^{8 , 9}

Another rationale for the potential beneficial role of imatinib in COVID-19 is its immunomodulatory effect, which would probably become more valuable from the second week after symptom onset; 70% of our patients received the drug at that time. Evidence from animal and human-cell studies suggests that imatinib can attenuate inflammatory cytokine release, including interleukin-6 and tumor necrosis factor-alpha, probably by inhibiting NF-κB signaling pathway,¹⁰ which seems to play a prominent role in the immune-mediated lung injury observed in severe COVID-19.³ Moreover, imatinib has also been found to prevent pulmonary damage by reducing tissue edema and maintaining endothelial barrier integrity in murine models of acute inflammatory lung injury.¹⁰

This preliminary study, which is limited by its observational design and a low number of cases, did not find differences when comparing patients who received imatinib before or after day 7 from the onset of symptoms in terms of age, sex, duration of imatinib therapy, concomitant treatments, C-reactive protein and D-dimer levels, lymphocyte count, days of hospitalization and death.

To our knowledge, this is the first case series of COVID-19 patients in whom imatinib was used as a treatment for this condition. This report cannot provide enough evidence for the effectiveness of this drug against SARS-CoV-2 infection, but the safety of a short-term treatment with imatinib, as well as its potential antiviral and immunomodulatory properties, suggests that it could be an acceptable option to explore in controlled clinical trials.

Funding

None

Ethical statement

This study was approved by the Ethical Committee of Hospital Universitario de Fuenlabrada (IRB protocol 20/42).

Author contributions

Conception and design: AMO, BFP, FB, MGG, JVSL and DBB. Data analysis and acquisition: AMO, LRP, BFP, and DBB. Interpretation of the data: AMO, BJB and DBB. Drafting or revision of the manuscript: AMO, BJB, AIFS, FB and DBB. Final approval of the manuscript: all authors.

Declaration of Competing Interests

David Bernal-Bello is the principal investigator of a non-sponsored randomised trial investigating the therapeutic role of imatinib and baricitinib in COVID-19 patients (NCT04346147). The rest of the authors are sub-investigators in this project. All authors declare no other competing interests.