INTRODUCTION

Patients with haematologic malignancies (HM), including haematopoietic stem cell transplant (HSCT) and chimaeric antigen receptor T-cell therapy (CAR-T) recipients infected with SARS-CoV-2 have higher risk of progressing to severe COVID-19 compared to the general population.¹ Although the prognosis of COVID-19 has significantly improved over time due to better management, vaccination, available treatment options and changes in viral virulence, patients with HM remain among those with worse prognosis and an early case fatality rate of approximately 14%.^2-5 Indeed, during Omicron predominance, 52% (309/593) of HM patients with COVID-19 were admitted to hospital, with an overall mortality of 16.5% among hospitalized patients and 8.6% in the whole cohort.⁶ This is still much higher than in the general population with Omicron infection (0.11% at day 28).⁷ In addition, the negative impact of COVID-19 in the immunocompromised should be considered on a timeframe longer than 28 days, since delayed mortality might be present.⁸

New treatment options have been developed and introduced in Italy starting from March 2021 for early treatment of mild to moderate COVID-19 in patients at risk for progression to severe disease: anti-spike monoclonal antibodies (MABs),^9-11 and antivirals such as molnupiravir, nirmatrelvir/ritonavir and remdesivir.^12-14 Although the immunocompromised might be the population that is expected to gain maximum benefit from early treatment, data are limited since these patients have been excluded or underrepresented in the randomized trials.^{15, 16}

Negative impact of SARS-CoV-2 infection in the HM population is not limited to a high rate of progression to severe COVID-19 and high mortality, but comprises also prolonged viral shedding, with possible negative consequences for withdrawing or postponing chemotherapy.¹⁷ In 77 studies in the general population the median duration of SARS-CoV-2 RNA positivity was 18 days after the onset of symptoms.^{18, 19} There are numerous reports of persistent RNA positivity and persistent or recurring symptomatic SARS-CoV-2 infection in HM patients, mainly in those treated with B-cell-depleting therapies.^{17, 20-23} In 382 HM patients infected between March 2020 and March 2021, the rate of SARS-CoV-2 RNA detection beyond day 30 was 13.9%.²² Little is known about the rate and predictors of prolonged SARS-CoV-2 shedding in HM patients receiving early therapy.

The aim of our study is to report the outcomes of early treatment of mild/moderate COVID-19 in patients with HM, and to identify the predictors of poor outcomes.

METHODS

RESULTS

Patients

Overall, 331 patients received early treatment during the study period. Since three patients received both antiviral and MABs, 328 patients were finally included. Non-Hodgkin lymphoma (NHL) was the most frequent diagnosis (33.8%), 27% had undergone HSCT or CAR-T, and 65% had received three or four doses of COVID-19 vaccine (Table 1).

TABLE 1. Clinical characteristics of included patients, divided into treated with MABs and antivirals.

	Total	Patients treated with MABs	Patients treated with an antiviral	p
	N = 328	N = 120 (36%)	N = 208 (64%)
Age, years, median (min–max)	66 (16–89)	62 (16–85)	68 (22–89)	0.007
Male sex n (%)	195 (59.5)	72 (60)	123 (59.1)	0.87
Diabetes mellitus	33 (10.1)	9 (7.5)	24 (11.5)	0.242
Chronic renal failure	34 (10.4)	12 (109)	22 (10.6)	0.869
Number of comorbidities	1 (0–5)	1 (0–5)	1 (0–5)	0.019
Underlying disease
Acute myeloid leukaemia	51 (15.5)	24 (20)	27 (13)
Acute lymphoid leukaemia	13 (4)	7 (5.8)	6 (2.9)
Non-Hodgkin lymphoma	111 (33.8)	40 (33.3)	71 (34.1)
Hodgkin disease	14 (4.3)	5 (4.2)	9 (4.3)
Chronic lymphocytic leukaemia	33 (10.1)	7 (5.8)	26 (12.5)
Multiple myeloma	78 (23.8)	28 (23.3)	50 (24)
Myelodysplastic syndrome	9 (2.7)	3 (2.5)	7 (2.9)
Myelofibrosis	9 (2.7)	5 (4.2)	4 (1.9)
Other (aplastic anaemia, 4; CML, 3; other 3)	10 (3)	1 (0.8)	9 (4.3)
Underlying disease, acute leukaemia versus other	64 (19.2)	31 (25.8)	33 (15.9)	0.028
Underlying disease in remission	144 (43.9)	53 (44.2)	91 (43.8)	0.942
Previous HSCT/CAR-T	89 (27.1)	39 (32.5)	50 (24)	0.097
Allogeneic	49 (14.9)	22 (18.3)	27 (13)
Autologous	37 (11.3)	16 (13.3)	21 (10.1)
CAR-T therapy	3 (1)	1 (1)	2 (1)
GvHD, if applicable	18/49 (36.7)	10/22 (45.5)	8/27 (29.6)
Time from HSCT/CAR-T, months	13 (0–225.5)	10.2 (1–201)	18.4 (0–225.5)	0.244
Recent (within 12 months) HSCT/CAR-T	42 (12.8)	21 (17.5)	21 (10.1)	0.053
Recent (within 12 months) chemotherapy	218 (66.5)	86 (71.7)	132 (63.5)	0.13
Recent (within 12 months) anti-CD20 treatment	66 (20.2)	26 (21.8)	40 (19.2)	0.57
Anti-SARS-CoV-2 vaccination	330 (91.7)	98 (81.7)	202 (97.6)	<0.001
Median number of doses	3 (0–4)	2 (0–4)	3 (0–4)	0.014
Hospital admission at the time of SARS-CoV-2 diagnosis	42 (12.8)	30 (25)	12 (5.8)	<0.001
For chemotherapy	27 (8.2)	18 (15)	9 (4.3)
For other reasons	26 (4.6)	12 (10)	3 (1.4)
Time from symptoms to treatment, days, median (min–max)	2 (0–13)	2 (0–10)	2 (0–13)	0.769
Time from SARS-CoV-2 diagnosis to treatment, median (min–max)	1 (0–9)	1 (0–9)	1 (0–7)	0.161
Treatment with antiviral		-		-
Nirmatrelvir/r	116 (55.8)		116 (55.8)
Remdesivir	59 (28.4)		59 (28.4)
Molnupiravir	33 (15.9)		33 (15.9)
MABs			-	-
B + E	24 (20)	24 (20)
C + I	23 (19.2)	23 (19.2)
Sotrovimab	73 (68.8)	73 (68.8)
Period of treatment				<0.0001
Before Omicron	19 (5.8)	19 (15.8)	0
Omicron predominance	309 (94.2)	101 (84.2)	208 (100)

• Abbreviations: B + E, bamlanivimab and etesevimab; C + I, casirivimab and imdevimab; CAR-T, chimaeric antigen receptor T-cell therapies; CML, chronic myeloid leukaemia, HSCT, haematopoietic stem cell transplant; MABs, monoclonal antibodies against spike protein of SARS-CoV-2; nirmatrelvir/r, ritonavir-boosted nirmatrelvir.

Most infections occurred after December 2021, confirming higher transmissibility and partial immune escape of the new Omicron variants (Figure 1).

Predictors of treatment failure

The univariate analysis of predictors of failure is shown in Table 2. Multivariable analysis confirmed higher risk of failure in case of older age, lower number of vaccine doses and treatment with MABs (Table 3). Figure 2 shows risk of failure according to treatment type in an adjusted survival curve.

TABLE 2. Univariate analysis of predictors of failure of early COVID-19 therapy (in bold variables included in the multivariable model, p < 0.2).

	Rate of failure, n = 31 (9.5%)	Unadjusted cause specific HR	95% CI	p
Age, years, median (min–max)	71 (22–85)	1.035	1.004–1.067	0.027
Male sex n (%)	18 (9.2)	0.934	0.458–1.907	0.852
Diabetes mellitus	5 (15.2)	1.807	0.694–4.707	0.226
Chronic renal failure	5 (14.7)	1.745	0.67–4.545	0.254
Number of comorbidities	1 (0–5)	1.232	0.922–1.644	0.158
Underlying diseasea				0.056
AML/MDS	10 (16.7)	1.00
NHL/CLL	15 (10.4)	0.624	0.28–1.389
Other	6 (4.8)	0.289	0.105–0.795
Active diseases versus in remission	21 (11.4)	1.711	0.805–3.633	0.162
Hospital admission at SARS-COV-2 diagnosis	9 (21.4)	2.962	1.364–6.433	0.006
Previous HSCT/CAR-T	9 (10.1)	1.089	0.501–2.365	0.829
Allogeneic	6 (12.2)
Autologous	2 (5.4)
CAR-T	1 (33.3)
Recent (within 1 year) HSCT/CAR-T	6 (14.3)	1.618	0.664–3.943	0.209
Recent (within 1 year) chemotherapy	19 (8.8)	0.769	0.373–1.587	0.477
Recent (within 12 months) anti-CD20 treatment	4 (6.1)	0.556	0.194–1.589	0.273
Days from symptoms to treatment, median	3 (0–7)	1.053	0.866–1.279	0.606
Treatment				0.004
MABs	19 (15.8)	1.00
B + E	4 (16.7)
C + I	6 (26.1)
Sotrovimab	9 (12.3)
Antivirals	12 (5.8)	0.348	0.169–0.717
Nirmatrelvir/r	6 (5.2)
Remdesivir	2 (3.4)
Molnupiravir	4 (12.1)
Anti-SARS-CoV-2 vaccination, n (%)				0.099
No	5 (18.5)	1.00
Yes	26 (8.7)	0.447	0.172–1.165
Number of doses, median	2 (0–4)	0.661	0.492–0.888	0.006
Number of doses				0.001
0–1	10 (23.8)	1.00
2–4	21 (7.3)	0.282	0.133–0.6
Period of treatment				<0.0001b
Before Omicron	7/19 (36.8%)	1.00
Omicron	24/309 (7.8%)	0.18	0.078–419

• The variables which were included in the multivariate analyses are indicated in bold.
• Abbreviations: AML, acute myeloid leukaemia; B + E, bamlanivimab and etesevimab; C + I, casirivimab and imdevimab; CAR-T, chimaeric antigen receptor T-cell therapies; CLL, chronic lymphocytic leukaemia; HM, haematological malignancy; HR, hazard ratio; HSCT, haematopoietic stem cell transplant; MABs, anti-spike protein monoclonal antibodies; MDS, myelodysplastic syndrome; NHL, non-Hodgkin lymphoma; nirmatrelvir/r, ritonavir-boosted nirmatrelvir. Continuous variables are expressed as median and min–max.
• ^a Rate of failure for specific HM: MDS 22% (n = 2); AML 15.7% (n = 8); CLL 15.2% (n = 5); NHL 9% (n = 10); acute lymphoblastic leukaemia 7.7% (n = 1); multiple myeloma 6.4% (n = 5).
• ^b Variable not included in the multivariable model since separate analysis was performed for the Omicron period.

TABLE 3. Multivariable Cox regression model for failure of early treatment in the whole population and in the patients during the Omicron predominance (from December 2021).

Model covariate	Adjusted cause specific HR	95% CI	p	Adjusted cause specific HR	95% CI	p
	All patients (n = 328)			Patients during the Omicron period (n = 309)
Age	1.055	1.02–1.09	0.002	1.052	1.014–1.092	0.007
Underlying disease			0.063			0.037
AML/MDS	1.00			1.00
NHL/CLL	0.772	0.336–1.776		0.699	0.281–1.74
Other	0.303	0.11–0.839		0.177	0.047–0.668
Treatment with antivirals versus MABs	0.432	0.193–0.968	0.042	0.549	0.23–1.31	0.177
Number of doses of vaccine	0.628	0.428–0.921	0.017	0.561	0.359–0.877	0.011

• Abbreviations: AML, acute myeloid leukaemia; CI, confidence interval; CLL, chronic lymphocytic leukaemia; HR, hazard ratio; MABs, anti-spike protein monoclonal antibodies; MDS, myelodysplastic syndrome; NHL, non-Hodgkin lymphoma.

The rate of failure was significantly higher in pre-Omicron versus Omicron period (7/19, 36.8% vs 24/309, 7.8%; p < 0.001). When the analyses were restricted to the Omicron period, the rate of failure was 5.8% (12/208) for antivirals and 11.8% (12/101) for MABs (p = 0.06). In multivariate analysis the independent predictors of failure in the Omicron period were older age, number of vaccine doses, and diagnosis of acute myeloid leukaemia/myelodysplastic syndrome (AML/MDS) (Table 3).

Considering only 97 patients infected during Omicron BA.1 predominance, when all but one MABs-treated patients received sotrovimab which is active against BA.1, the difference in failure rate was no longer significant: 13.1% (8/61) for MABs vs 5.6% (2/36) for antivirals, p = 0.177.

Length of viral shedding

Data on follow-up swabs were available for 318 (97%) patients. Among them, 12 never resolved SARS-CoV-2 infection, and all died on median 29 days after the diagnosis (min–max, 6–159). In 301 patients who resolved SARS-CoV-2 infection, the median time to the first negative swab was 14 days from SARS-CoV-2 infection diagnosis (min–max, 1–112): 81 were still positive by day 21 (17%), and 37 by day 30 (12%).

Patients treated with antivirals, compared to MABs, had significantly shorter viral shedding: median 12 (min–max, 1–104) vs 20 days (min–max, 3–159).

Independent predictors of longer time to first negative swab were: male sex, comorbidities, having NHL or chronic lymphocytic leukaemia (CLL), hospital admission at the time of SARS-CoV-2 diagnosis and early treatment with MABs (Table 4). The same independent predictors were identified also when only the Omicron period was considered (data not shown).

TABLE 4. Univariate and multivariable analysis of characteristics associated with time to first negative SARS-CoV-2 swab. Variables with p < 0.2 at the univariate analysis were included in the multivariable model.

	Unadjusted cause specific HR	95% CI	p	Adjusted cause specific HR	95% CI	p
Age, years	0.997	0.99–1.004	0.420
Male sex	0.789	0.627–0.994	0.044	0.780	0.619–0.985	0.037
Underlying diseasea			0.021			0.001
NHL/CLL	0.763	0.606–0.961		0.664	0.524–0.841
Other	1.00
Active diseases versus in remission	0.989	0.786–1.243	0.922
No. of comorbidities	0.931	0.839–1.033	0.178	0.894	0.803–0.996	0.042
Hospital admission at the time of SARS-COV-2 diagnosis	0.618	0.439–0.870	0.006	0.660	0.461–0.944	0.023
Recent (within 1 year) HSCT/CAR-T	0.809	0.572–1.145	0.232
Recent (within 1 year) chemotherapy	1.062	0.834–1.353	0.626
Anti-CD20 within 6 months	0.822	0.618–1.094	0.178	-
Days from symptoms to treatment	1.006	0.946–1.007	0.852
Treatment			<0.0001			<0.001
Antivirals	1.00
MABs	0.599	0.472–0.761		0.575	0.445–0.742
Number of doses of anti-SARS-CoV-2 vaccine	1.142	1.014–1.286	0.028	-
Period of treatment			0.13	-
Before Omicron	1.00
Omicron	1.514	0.885–2.592

• Note: Analysis performed for 318 patients for whom the information on follow-up testing was available. In bold values that are statistically significant. Higher HR refers to probability of shorter time to the first negative result, hence HR < 1 identifies predictors of longer shedding.
• Abbreviations: CAR-T, chimaeric antigen receptor T-cell therapies; CI, confidence interval; CLL, chronic lymphocytic leukaemia; HR, hazard ratio; HSCT, haematopoietic stem cell transplant; MABs, anti-spike protein monoclonal antibodies; NHL, non-Hodgkin lymphoma.
• ^a With AML/MDS patients as reference group, NHL/CLL patients had longer time to first negative and patients with other diseases had shorter time to first negative, thus, NHL/CLL were compared to all other patients.

The effect of the type of underlying disease on the time to negativity is shown in an adjusted cumulative incidence curve derived by a competing-risk analysis using a multivariable Fine̶–Grey model (Figure 3).

DISCUSSION

This study analysed the efficacy of early treatments for COVID-19 in a large cohort of HM patients over a 16-month period during which different viral variants emerged and became predominant. Looking at the distribution of included patients, we noted an important increase in the number of patients receiving early treatment for COVID-19 during Omicron predominance, despite unchanged policies of early testing and early treatment since March 2021, which demonstrates higher transmissibility and capacity to evade prior vaccine-induced humoral immunity.²⁸ These data are in line with national data on rates of new SARS-CoV-2 cases in the same period, which showed a marked increase in infections from December 2021 onwards, when the Omicron variant became dominant in Italy.²⁹ In accordance with the international and national recommendations, the use of available MABs was progressively discontinued due to the reduction in their activity against the emerging variants.³⁰

Clinical features of patients included were comparable to the data available in the literature.^{1, 4} The rate of progression to severe COVID-19 requiring oxygen was significantly lower than during the pretreatment periods, but still reached 9.5% in the whole population, and 21% in the pre-Omicron period. Indeed, even though the rate of early treatment failure was lower during the Omicron period, it was still 7.8%, which is much higher than in the general population in the same period (1.2%–1.4%), and even higher than in the treatment arm of pivotal studies performed before the Omicron period in the general at-risk population (7.3% for molnupiravir, 0.77% for nirmatrelvir/ritonavir, 0.7% for remdesivir and 1% for sotrovimab).^11-14 Our rate was only slightly lower to what was previously reported in more heterogeneous cohorts of patients with HM treated with MABs.^{31, 32}

The probability of failure of early treatment was higher in case of older patients, lower number of previous doses of COVID-19 vaccine and treatment with MABs compared to antivirals. The reasons for lower efficacy of MABs compared to antivirals might be related to antiviral potency; however, during the study period the following changes took place and might have contributed to the lower rate of failure in the more recent period: (1) more patients vaccinated with booster doses; (2) availability of antivirals for early treatment much later than the availability of MABs; and (3) emerging variants with lower virulence but also lower susceptibility to available MABs. However, the rate of failure in case of MABs was high even before the emergence of variants associated with their inefficacy and was present (although not statistically significant) even when treatments were compared in a subgroup of patients treated during the Omicron BA.1 period with sotrovimab, which is active against this variant. Therefore, several issues warrant consideration. First, some patients developed failure on the same day of infusion of MABs, raising a question if a possible worsening of respiratory function due to MAB administration could be the reason.³³ Second, the selective pressure of monoclonal antibodies which persist in patient's system, in combination with the lack of an effective endogenous immune response, may promote the emergence of SARS-CoV-2 escape mutations. In a recent prospective study, the use of sotrovimab in immunocompromised patients was associated with prolonged viral shedding, and after longitudinal viral sequencing, the authors found that 32.6% (14/43) of immunocompromised patients who received sotrovimab developed spike protein mutations that substantially reduced susceptibility to sotrovimab in a neutralization assay, while patients who received combination treatment with remdesivir had a significant reduction in the selection of escape variants.³⁴ Similar findings were noted in a previous study reporting that two of five immunocompromised patients treated with sotrovimab developed resistance.³⁵ Finally, we cannot exclude that despite the predominance of certain variants, some patients might be infected with another variant, resistant to the MABs used. In fact, in a general population of patients during the Omicron BA.2 predominance, similar efficacy of bebtelovimab (active against all main Omicron subvariants) and nirmatrelvir/ritonavir was noted (1.4% and 1.2%), but such a low rate of progression, and low rate of severely immunocompromised patients might not allow detection of any differences applicable to subgroups such as the HM population.³⁶ Whatever the reason for the lower efficacy of MABs observed in our cohort, considering a possible negative effect of emerging variants, antivirals seem to provide a more reliable treatment option.

Analysing specifically the predictors of failure during Omicron infection, the protective role of younger age and higher number of vaccine doses was confirmed, but an effect of the underlying disease emerged. The highest rate of failure was observed in case of AML/MDS, while HR was 30% lower in lymphoid disorders such as NHL/CLL and almost 80% lower for all other diseases. The negative outcome in AML/MDS patients might be related to the aggressiveness of their underlying disease and high intensity chemotherapy, and was present also in initial studies in HM patients as a risk factor for higher mortality.^{1, 37}

COVID-19-related mortality was 3.4% in the whole population, but as high as 21% in the pre-Omicron period despite early treatment, and 2.3% during the Omicron period, with age and AML/MDS being other independent predictors. COVID-19-associated mortality and overall mortality were lower than in previous reports in HM patients (e.g., 18% in 3801 patients from a European survey), but still significantly higher than in the general population considered at risk for severe COVID-19 and thus deserving early treatment.⁴ Moreover, in our cohort, the overall mortality doubled from day 30 to day 90, confirming the previous observation from 936 solid-organ transplant recipients reporting delayed mortality in that population.⁸ This might be partially due to the indirect impact of COVID-19 on patients’ survival including delayed access to chemotherapy or higher risk of late complications. Interestingly, the early treatment seemed to change the previously observed timing of COVID-19 worsening, which typically occurred early during infection, while in this cohort, half of the failures occurred more than 10 days and up to 35 days from the onset of infection.

The optimal management of patients who progress to severe COVID-19 remains to be established but highly effective strategies are required, since the mortality in those who progressed to severe COVID-19 remains very high (57% in the pre-Omicron and 26% in the Omicron period) despite provided treatment in accordance with current guidelines.^{16, 25} The difficulty in managing these patients also stems from the uncertainty which component — viral or inflammatory — is predominantly responsible for clinical worsening, and which treatment strategies might be most effective.

Prolonged shedding of SARS-CoV-2 virus is frequent in HM patients, and this study confirmed that even in case of early treatment during the Omicron period, 12% were still SARS-CoV-2-positive beyond day 30, which is almost identical to what was previously reported.²² Among the predictors of prolonged shedding, we confirmed the predisposing role of NHL/CLL, male sex and comorbidities.^{22, 38, 39} In addition, MABs were also associated with longer shedding, while the role of the Omicron variant and anti-CD20 treatment was not confirmed in multivariate analysis. The possible reasons for longer shedding in case of MABs might be lower efficacy against the Omicron variants and higher probability of inducing resistance due to prolonged exposure to MABs remaining in the patient's system.^{34, 35} The concerns regarding persistent SARS-CoV-2 infection include the risk of clinical relapse, limited access to healthcare and delay of chemotherapy and the threat of intrahost viral evolution. Viral dissemination and accelerated evolution, which may be responsible for immune escape from the initial infection, have been described in immunosuppressed patients with persistent or severe SARS-CoV-2 infection.^{21, 40, 41}

Limitations of this study include retrospective design and change over time in circulating variants overlapping with the change in early treatment and increase in administered vaccine doses, making it difficult to ascertain the independent role of one factor versus another. Additionally, data on infecting variants and posttreatment sequencing were not routinely available. However, inclusion of the whole cohort provided a complete overview of over 300 consecutive HM patients diagnosed with SARS-CoV-2 infection during the time when early treatment was available and promptly administered. In addition, analyses of the Omicron-limited subgroup offered more updated information, while not excluding the pre-Omicron period showed how a change in the predominant variant can rapidly increase the prevalence of SARS-CoV-2 infection also in HM setting.

In conclusion, patients with HM are still at high risk for negative outcomes in case of SARS-CoV-2 infection, with 7.8% rate of progression to severe COVID-19 despite early treatment, almost universal vaccination and the Omicron variant. Moreover, those who experience disease progression continue to be at high risk for mortality. Our findings strongly support the recommendations to vaccinate with boosters and provide effective early treatment, keeping in mind that the efficacy of antivirals is currently not affected by the emergence of new variants.¹⁶

AUTHOR CONTRIBUTIONS

Malgorzata Mikulska, Chiara Sepulcri, Diletta Testi, Chiara Russo and Michele Bartoletti performed research and wrote the manuscript; Elisa Balletto, Linda Bussini, Chiara Dentone, Federica Magne, Sílvia Policarpo, Caterina Campoli, Federica Magne, Chiara Ghiggi, Alessandro Cilli, Sara Aquino, Chiara Ghiggi performed data collection; Malgorzata Mikulska, Maddalena Giannella, Daniele Roberto Giacobbe, Michele Bartoletti performed data analysis; Anna Maria Raiola, Francesca Bonifazi, Pierluigi Zinzani, Michele Cavo, Roberto Lemoli, Emanuele Angelucci, Pierluigi Viale, and Matteo Bassetti supervised the study. All Authors reviewed and approved the manuscript.

FUNDING INFORMATION

No funding was provided for this study.

CONFLICT OF INTEREST STATEMENT

Emanuele Angelucci reports research grants and/or personal fees for advisor/consultant and/or speaker/chairman from Vertex Pharmaceuticals Incorporated and Celgene (BSM), Vifor Pharma, Novartis, Blue Bird Bio, Menarini Stemline, Glaxo, Regeneron, GILEAD and Novartis. Matteo Bassetti reports research grants and/or personal fees for advisor/consultant and/or speaker/chairman from Bayer, BioMérieux, Cidara, Cipla, Gilead, Menarini, MSD, Pfizer, and Shionogi. The other authors have no conflict of interest to disclose.