Comparison of two antifibrotic treatments for lung fibrosis in post-COVID-19 syndrome: A randomized, prospective study

Abstract

Background

Although pulmonary fibrosis secondary to COVID-19 infection is uncommon, it can lead to problems if not treated effectively in the early period. This study aimed to compare the effects of treatment with nintedanib and pirfenidone in patients with COVID-19-related fibrosis.

Methods

Thirty patients who presented to the post-COVID outpatient clinic between May 2021 and April 2022 with a history of COVID-19 pneumonia and exhibited persistent cough, dyspnea, exertional dyspnea, and low oxygen saturation at least 12 weeks after diagnosis were included. The patients were randomized to receive off-label treatment with nintedanib or pirfenidone and were followed up for 12 weeks.

Results

After 12 weeks of treatment, all pulmonary function test (PFT) parameters, 6MWT distance, and oxygen saturation were increased compared to baseline in both the pirfenidone group and nintedanib groups, while heart rate and radiological score levels were decreased (p < 0.05 for all). The changes in 6MWT distance and oxygen saturation were significantly greater in the nintedanib group than in the pirfenidone group (p = 0.02 and 0.005, respectively). Adverse drug effects were more frequent with nintedanib than pirfenidone, with the most common being diarrhea, nausea, and vomiting.

Conclusion

In patients with interstitial fibrosis after COVID-19 pneumonia, both nintedanib and pirfenidone were observed to be effective in improving radiological score and PFT parameters. Nintedanib was more effective than pirfenidone in increasing exercise capacity and saturation values but caused more adverse drug effects.

Resumen

Introducción

Aunque la fibrosis pulmonar secundaria a la infección por COVID-19 es poco común, puede generar problemas si no se trata de manera efectiva en el período inicial. Este estudio tuvo como objetivo comparar los efectos del tratamiento con nintedanib y pirfenidona en pacientes con fibrosis relacionada con COVID-19.

Métodos

Se incluyeron 30 pacientes que acudieron a la consulta externa post-COVID entre mayo de 2021 y abril de 2022 con antecedentes de neumonía por COVID-19 y presentaron tos persistente, disnea, disnea de esfuerzo y baja saturación de oxígeno al menos 12 semanas después del diagnóstico. Los pacientes fueron aleatorizados para recibir un tratamiento no aprobado con nintedanib o pirfenidona y fueron seguidos durante 12 semanas.

Resultados

Después de 12 semanas de tratamiento, todos los parámetros de la prueba de función pulmonar (PFT), la distancia de la PM6M y la saturación de oxígeno aumentaron en comparación con los valores basales tanto en el grupo de pirfenidona como en el de nintedanib, mientras que la frecuencia cardíaca y los niveles de puntuación radiológica disminuyeron (p < 0,05 para todos). Los cambios en la distancia de la PM6M y la saturación de oxígeno fueron significativamente mayores en el grupo de nintedanib que en el grupo de pirfenidona (p = 0,02 y p = 0,005, respectivamente). Los efectos adversos del fármaco fueron más frecuentes con nintedanib que con pirfenidona, siendo los más comunes diarrea, náuseas y vómitos.

Conclusión

En pacientes con fibrosis intersticial después de neumonía por COVID-19 se observó que tanto nintedanib como pirfenidona son efectivos para mejorar la puntuación radiológica y los parámetros de la PFT. Nintedanib fue más eficaz que pirfenidona para aumentar la capacidad de ejercicio y los valores de saturación, pero provocó más efectos adversos del fármaco.

Introduction

The most common pulmonary sequelae of COVID-19 are a significant reduction in diffusing capacity of the lung for carbon monoxide (DLCO) and associated pulmonary interstitial damage. One year after moderate COVID-19, the prevalence of reduced DLCO and permanent lung injury exceeds 30%, with one-third of patients presenting with severe DLCO impairment and fibrotic lung injury.^1,2

The incidence of COVID-19-induced pulmonary fibrosis can be only estimated based on a 15-year observational study of post-SARS lung pathology. While most SARS patients with fibrotic lung injury recovered within the first year and remained healthy thereafter, 20% of cases showed significant fibrosis progression within 5–10 years.³ Based on these data, the incidence rate of post-COVID pulmonary fibrosis after moderate disease can be estimated as 2%–6%. However, there is evidence that fibrosis may be one of the major long-term complications of COVID-19 even in people with asymptomatic infections.^3,4 Although there are recommendations for the treatment for COVID-19-related fibrosis, there is no definitive treatment.

Pirfenidone and nintedanib are two agents that were first indicated for idiopathic pulmonary fibrosis. Both drugs have been observed to prevent the progression of idiopathic pulmonary fibrosis and have a positive effect on predicted life expectancy. Pirfenidone (5-methyl-1-phenyl-2-[1H]-pyridone) may inhibit apoptosis, down-regulate ACE receptor expression, reduce inflammation by various mechanisms, and reduce oxidative stress.^5,6 Nintedanib acts by inhibiting tyrosine kinase. This allows strong inhibition of vascular endothelial growth factor (VEGF) receptors, fibroblast growth factor (FGF) receptors, and platelet-derived growth factor (PDGF) receptors, which play an important role in fibrosis.⁷ There are no studies in the literature comparing the effectiveness of these two drugs in the prevention of lung fibrosis after COVID-19.

In this study, we aimed to observe and compare the effects of two antifibrotic treatments on pulmonary function test (PFT) parameters, 6-min walk test (6MWT) distance, and clinical and radiological improvement in patients presenting with chronic COVID-19, defined as persistent respiratory symptoms after 12 weeks, and pulmonary fibrosis detected on chest computed tomography (CT).

Materials and methods

This prospective study included 30 patients who presented to the post-COVID outpatient clinic of the Atatürk University Medical Faculty Hospital, Department of Chest Diseases between May 2021 and April 2022, had a previous diagnosis of COVID-19 pneumonia, and exhibited cough, dyspnea, exertional dyspnea, and low oxygen saturation at least 12 weeks after diagnosis (Fig. 1). The patients were randomly assigned to two treatment groups (15 patients per group) by an independent physician using the randomization function in Microsoft® Excel. The aim of the study was to conduct clinical laboratory and radiological follow-up of these patients for 12 weeks and evaluate the results.

Fig. 1.

CONSORT diagram.

Approval to conduct the study was obtained from the Erzurum Atatürk University Faculty of Medicine Ethics Committee. Before starting the research, participating patients were informed about the study objectives, methods, and required time commitment. They were also informed that participating in the study carried no risk, that participation was completely voluntary, and that they could leave the study at any time.

Inclusion criteria

All patients in the study were diagnosed with COVID-19 by real-time PCR analysis of a nasopharyngeal swab at least 3 months ago and presented with chronic COVID-19 symptoms. Other inclusion criteria were:

• 1.

  Being older than 18 years of age.

• 2.

  Having fibrosis secondary to COVID-19 on radiological imaging.

• 3.

  Not requiring intubation and mechanical ventilation during acute COVID-19.

• 4.

  Agreeing to attend follow-up visits during the 12-week study period.

Exclusion criteria

Patients with any of the following were excluded:

• 1.

  Any conditions that may contraindicate PFT (recent myocardial infarction, pulmonary embolism, cerebral aneurysm, active hemoptysis, pneumothorax, nausea, vomiting, recent thoracic, abdominal and eye surgery).

• 2.

  Cognitive disorder/disability or inability to cooperate.

• 3.

  Previously known or newly detected lung pathology.

Pulmonary function testing

As per the ATS/ERS 2019 guideline, patients were informed of the rules they should follow before spirometry. The test maneuver was explained to the patient by the technician and the patient performed three acceptable spirograms. Tests complying with the PFT reproducibility and acceptability criteria published by ATS/ERS in 2019 were included in the study.⁸ The lower limit of normal parameters determined for the healthy population are presented by calculating on a spirometry device in accordance with the criteria in this declaration. Spirometry was performed by the same technician with Plusmed MIR SpiroLab III device.

6-Min walk test (6MWT) and fingertip oxygen saturation measurement

The patients rested for at least 15 min at the beginning of the 30-m track, and their oxygen saturation, heart rate was measured using a fingertip pulse oximeter and recorded. Under the supervision of a physician, patients were instructed to walk along the level corridor as fast as they could for 6 min. In the event of any symptoms such as excessive fatigue, dyspnea, or palpitations during the test, it was ended early to avoid endangering the patient. At the end of the test, the patient rested while the distance walked was recorded in meters.

Radiological assessment

All patients underwent contrast-enhanced CT scans of the chest on a second-generation Somatom Definition Flash 256-slice dual-source multidetector CT scanner (Siemens Healthcare, Forchheim, Germany). CT examinations were performed with breath holding during deep inspiration. The images were transferred to a commercial workstation (Singo via. Workstation, Siemens, Erlangen, Germany) and assessed by a radiologist who were blinded to the patients’ identities. The reader had 18 years of experience in chest radiology of experience in radiology. Radiological scoring of the lungs was performed using the scoring system applied in our previous study. Accordingly, each lobe was given a score between 1 and 5 points, for a total score between 5 and 25 points.⁹

Medical treatment

All of the patients who presented to our outpatient clinic for chronic COVID-19 were using oxygen concentrators at home due to low oxygen saturation levels, and all patients had continued methylprednisolone treatment for at least 8 weeks after discharge. PFT, 6MWT, and high-resolution computed tomography (HRCT) examinations were performed at the time of presentation.

Consent for the initiation of antifibrotic therapy (pirfenidone or nintedanib) was obtained from patients with a restrictive pattern in PFT, DLCO < 80%, and >5% fibrosis with profibrotic areas on HRCT. A report including the patients’ test results and tomography images was sent to the Turkish Ministry of Health for approval to administer off-label treatment. It was planned to approve the drug administration for a period of 3 months for all patients and then extend it as appropriate based on their clinical course. In the pirfenidone group, treatment was started at 600 mg/day in the first week, 1200 mg/day in the second week, and 1800 mg/day in the third week. In the nintedanib group, treatment was started at a dose of 300 mg/day from the beginning. In patients still using methylprednisolone, the treatment was tapered off over 2 weeks.

Liver function tests were evaluated monthly in all patients using pirfenidone and nintedanib. For patients with diarrhea as an adverse effect of nintedanib use, loperamide tablets were used to control the symptoms and treatment was continued. Antihistaminic tablets and sunscreen cream were used to control rash and photosensitivity that occurred as adverse effects in patients using pirfenidone.

Statistical analysis

Analyses were performed using IBM SPSS version 20.0 software (IBM Corp, Armonk, NY). Data were presented as mean, standard deviation, number, and percentage. Shapiro–Wilk test and Kolmogorov–Smirnov test were used to determine whether continuous variables were normally distributed. Continuous variables were compared between more than two dependent groups using analysis of variance (ANOVA) if normally distributed. Post hoc tests after ANOVA were performed using Tukey's test when variances were homogeneous and Tamhane's T2 test when variances were not homogeneous. Post hoc analysis after Kruskal–Wallis test was performed using the Kruskal–Wallis one-way ANOVA (k samples) test. Relationships between two quantitative variables were examined using Pearson correlation analysis if normally distributed and Spearman correlation analysis if non-normally distributed. p values <0.05 were considered statistically significant.

Results

The mean ages of the pirfenidone and nintedanib groups were 62.6 ± 8.1 years and 68.4 ± 9.8 years, respectively (p = 0.08). In each group, 9 patients (60%) were women.

The groups’ mean PFT values, oxygen saturation levels, heart rates, radiological scores, and 6MWT distances at the start (week 0) and week 12 of treatment are shown in Table 1. There were no statistically significant differences between the two groups in any of the parameters at the start of treatment. After 12 weeks of treatment, there were significant increases in all PFT parameters, 6MWT, and oxygen saturation compared to week 0 and decreases in heart rate and radiological score in both the pirfenidone group and nintedanib group (p < 0.05 for all).

Table 1.

Comparison of demographic data, pulmonary function tests, oxygen saturation, heart rate, and 6-min walk test distance of the groups before and after antifibrotic treatment.

	Pirfenidone group (n = 15)		p^a	Nintedanib group (n = 15)		p^b	p^c
	Week 0Mean ± SD	Week 12Mean ± SD		Week 0Mean ± SD	Week 12Mean ± SD
Age	62.6 ± 8.1		–	68.4 ± 9.8		–	0.08
BMI	32.5 ± 4.9		–	30.9 ± 6.4		–	0.45
FVC (L)	1.8 ± 0.5	2.1 ± 0.8	0.01	2.1 ± 0.7	2.5 ± 0.7	0.02	0.35
FVC%	67.2 ± 13.9	74.4 ± 19.6	0.01	70.9 ± 18.1	82.6 ± 20.8	0.02	0.53
FEV1 (L)	1.6 ± 0.5	1.7 ± 0.7	0.01	1.8 ± 0.6	2.01 ± 0.5	0.02	0.26
FEV1%	68.8 ± 13.2	76.3 ± 18.6	0.02	75.3 ± 17.9	84.1 ± 16.3	0.008	0.27
DLCO%	53.3 ± 10.1	62.2 ± 14.3	0.004	57.9 ± 11.1	72.6 ± 9.5	0.02	0.24
DLCO/VA%	60.9 ± 9.1	69.4 ± 12.5	0.001	63.8 ± 14.1	76.1 ± 11.3	0.001	0.51
6MWT (m)	210.1 ± 110.4	239.9 ± 121.4	0.001	201.3 ± 79.5	271.3 ± 112.1	0.001	0.62
SO2	77.8 ± 10.1	83.3 ± 10.6	0.001	79.7 ± 4.3	90.3 ± 4.8	0.001	0.5
Heart rate	100.4 ± 9.9	98.4 ± 12.3	0.001	96 ± 7.8	85.7 ± 6.6	0.001	0.33
Radiological score	14.1 ± 2.8	8.1 ± 3.2	0.001	13.4 ± 3.4	8.3 ± 4.1	0.001	0.56

FVC: forced vital capacity, FEV₁: forced expiratory volume in 1 s, DLCO/VA: diffusing capacity divided by the alveolar volume, 6MWT: 6-min walk test distance, SO₂: fingertip oxygen saturation in room air.

a

Pirfenidone group, week 0 vs. week 12.

b

Nintedanib group, week 0 vs. week 12.

c

Pirfenidone vs. nintedanib group, week 0.

Changes in PFT, oxygen saturation, heart rate, radiological score, and 6MWT values between week 0 and week 12 are shown in Table 2. Comparisons between the groups showed that the change in 6MWT distance and oxygen saturation levels were significantly greater in the nintedanib group than in the pirfenidone group (p = 0.02 and 0.005, respectively).

Table 2.

Comparison of the changes in pulmonary function tests, saturation, heart rate, and 6-min walking test distances of the groups after 12 weeks of antifibrotic treatment.

	Pirfenidone group (n = 15)Mean ± SD	Nintedanib group (n = 15)Mean ± SD	p
ΔFVC (L)	0.2 ± 0.3	0.4 ± 0.3	0.17
ΔFVC%	7.4 ± 13.5	11.7 ± 13.4	0.38
ΔFEV1 (L)	0.2 ± 0.3	0.2 ± 0.2	0.66
ΔFEV1%	7.7 ± 12.3	8.8 ± 8.8	0.78
ΔDLCO%	9.1 ± 10.3	14.6 ± 15.8	0.26
ΔDLCO/VA%	8.6 ± 7.2	12.2 ± 16.6	0.44
Δ6MWT (m)	29.8 ± 27.2	70 ± 48.4	0.02
ΔSO2	5.6 ± 4.8	10.6 ± 4.1	0.005
ΔHeart rate	−12.9 ± 11.6	−10.2 ± 7.4	0.46
ΔRadiological score	5.5 ± 2.7	5.6 ± 3.7	0.95

FVC: forced vital capacity, FEV₁: forced expiratory volume in 1 s, DLCO/VA: diffusing capacity divided by the alveolar volume, 6MWT: 6 min walking test, SO₂: fingertip saturation in room air.

The linear regression analysis of the change in pulmonary function test values, oxygen saturation, heart rate, radiological score, and 6MWT values of the groups at the 12th week of treatment with the initial radiological score is shown in Table 3. Accordingly, it was observed that lower radiological score at week 0 was associated with significantly greater changes in DLCO%, DLCO/VA% and oxygen saturation values (p = 0.001, 0.003, and 0.017, respectively).

Table 3.

Linear regression analysis of the effect of radiological score before the start of antifibrotic therapy on the change in pulmonary function test parameters, oxygen saturation, heart rate, 6-min walk test distance, and radiological score.

Coefficients^a
	Unstandardized coefficients		Standardized coefficients	t	p	95% Confidence interval for B
	B	Std. error	Beta			Lower bound	Upper bound
(Constant)	17.376	1.113		15.615	.000	15.047	19.705
ΔFVC	−.882	4.790	−.108	−.184	.856	−10.908	9.144
ΔFVC%	−.013	.132	−.056	−.097	.924	−.289	.263
ΔFEV1	−13.948	7.495	−1.174	−1.861	.078	−29.636	1.739
ΔFEV1%	.354	.180	1.226	1.970	.064	−.022	.730
ΔDLCO%	−.291	.077	−1.281	−3.782	.001	−.451	−.130
ΔDLCO/VA%	.192	.056	.803	3.462	.003	.076	.309
Δ6MWT	.026	.016	.392	1.663	.113	−.007	.059
ΔSO2	−.302	.115	−.502	−2.622	.017	−.544	−.061
ΔHeart rate	.009	.057	.027	.149	.883	−.112	.129
ΔRadiological score	−.099	.254	−.102	−.388	.702	−.630	.433

FVC: forced vital capacity, FEV₁: forced expiratory volume in 1 s, DLCO/VA: ratio of diffusing capacity of the lung for carbon monoxide to alveolar volume, 6MWT: 6-min walk test, SO₂: fingertip oxygen saturation in room air.

a

Dependent variable: radiological score.

The drug-related adverse effects observed in the groups during treatment are shown in Table 4. Diarrhea, nausea/vomiting, and loss of appetite were observed frequently in the nintedanib group, while rash and photosensitivity were observed more frequently in the pirfenidone group.

Table 4.

Adverse effects related to the antifibrotic drugs.

	Pirfenidone group (n = 15)n (%)	Nintedanib group (n = 15)n (%)
Diarrhea	–	12 (80%)
Nausea-vomiting	1 (6.6%)	10 (66.6%)
Loss of appetite	1 (6.6%)	4 (26.6%)
Rash	1 (6.6%)	–
Photosensitivity	1 (6.6%)	–

Discussion

In this study comparing antifibrotic treatments for fibrosis in post-COVID-19 syndrome, we observed increases in PFT parameters and oxygen saturation values and decreases in radiological score and heart rate compared to baseline in both the nintedanib and pirfenidone groups, but 6MWT performance and saturation levels at 12-week follow-up were better in the nintedanib group. However, adverse drug effects were reported more frequently by patients using nintedanib than patients using pirfenidone, with diarrhea being the most common side effect. The improvement in PFT parameters and oxygen saturation at 12 weeks was greater in patients with lower radiological scores at baseline.

Fibrosis is the end result of almost all chronic inflammatory diseases of the heart, liver, lungs, and kidneys.¹⁰ In response to tissue injury, myofibroblasts from multiple sources (resident fibroblasts, mesenchymal cells, circulating fibroblasts, and other cell types) can initiate wound healing by remodeling the extracellular environment to restore tissue integrity and support the replacement of parenchymal cells. Usually, this profibrotic process ends once the tissue has healed. However, repeated damage and repair (as in severe COVID-19) causes instability in this process, resulting in excessive pathological accumulation of extracellular matrix (ECM) proteins. Downregulation of myofibroblast activity in this region causes an influx of immune cells, primarily macrophages, and activates the chronic inflammatory process.¹¹ The subsequent massive release of proinflammatory and profibrotic cytokines activates fibrotic pathways such as the transforming growth factor-beta (TGF-β), Wnt (wingless-related integration site), and YAP/TAZ (Yes-associated protein/transcriptional co-activator with PDZ-binding domain) signaling pathways.¹²

Although COVID-19 infection manifests with a mild disease course in a large part of the population, it can cause severe illness characterized by macrophage activation syndrome and acute respiratory distress in some patients, especially those with comorbidities. The cytokine storm, which involves overproduction of interleukin-6, tumor necrosis factor-α, TGF-β, PDGF, and VEGF, causes intense fibroblast and myofibroblast activity in the ECM.^13–15 This process is initially beneficial for tissue repair but may result in fibrosis in the lung parenchyma. Systemic corticosteroid therapy and anti-cytokine treatments administered to prevent this cytokine discharge in the acute period have yielded successful results. However, in the follow-up of some patients, it was observed that this process continued despite maintenance therapy with systemic steroids.^16,17 This persistence of symptoms at least 12 weeks after acute COVID-19 infection was defined as chronic COVID-19. One of the most important problems observed in these patients is respiratory symptoms, which have been associated with diffusion restriction on PFT and the presence of prefibrotic and fibrotic areas on radiological examination.²

Data on the use of antifibrotic therapy for lung fibrosis and the effectiveness of these drugs is constantly emerging. Nintedanib and pirfenidone are among the most commonly used agents, although the mechanism of action of pirfenidone is not fully known. The results of an in vitro study suggested it acts by inhibiting the TGF-β pathway, preventing fibroblast proliferation and differentiation into myofibroblasts.^18,19 Nintedanib is a tyrosine kinase inhibitor that exerts its antifibrotic effect by inhibiting PDGF receptor alpha, FGF receptors 1, 2, and 3, and VEGF receptors 1 and 2, thereby preventing fibroblast proliferation and ECM formation.⁷ These drugs were initially approved for idiopathic pulmonary fibrosis but have started to be used for indications in rheumatological lung diseases that cause fibrosis, with evidence of their effectiveness.^20,21 In the first stage of our study, which evaluated the antifibrotic efficacy of pirfenidone in post-COVID-19 patients, we showed that there was a significant improvement in both pulmonary function parameters and radiological scores over 12-week follow-up.⁹

In the second stage of our study, we aimed to compare the efficacy of antifibrotic treatments in a randomized trial. The results of our study showed that both treatment arms provided improvement in PFT results, 6MWT distance, and oxygen saturation levels. In addition, there was also a reduction in the patients’ radiological scores. This may have occurred because the antifibrotic treatments resolved prefibrotic areas. When these factors were evaluated together in regression analysis, we determined that lower initial radiological score was associated with greater improvement in PFT parameters and saturation values after antifibrotic therapy. These findings suggest that antifibrotic therapy is more effective early in the development of fibrosis.

In terms of adverse effects related to the antifibrotic treatment, diarrhea was reported by most patients using nintedanib. Diarrhea is one of the most commonly observed side effects of nintedanib but can be controlled with oral loperamide, as seen in our study. Therefore, we think that patients with diarrhea should use loperamide tablets without discontinuing treatment. For rash and photosensitivity reactions associated with pirfenidone, the use of sunscreen cream and administration of antihistaminic medication for rash may be beneficial, as we observed in our study.

In our study, approval for off-label drug use was obtained from the Ministry of Health for all patients started on antifibrotic treatment. Observation of more than 5% fibrosis in the lung was among the most important criteria in the ministry granting this approval. According to this criterion, the patients were referred to antifibrotic treatment by us. The lack of early follow-up of the patients was our most important limitation. In addition, the limited number of patients was another important limitation.

In conclusion, assessing parenchymal involvement is crucial in the evaluation of patients with long-term dyspnea complaints after COVID-19. Early detection and treatment of pulmonary fibrosis may yield more effective results from treatment. Our comparison of the effectiveness of antifibrotic therapy with pirfenidone and nintedanib showed that both agents were equally effective in improving PFT parameters, while nintedanib seems to be more effective in increasing both exercise capacity and oxygen saturation levels. However, pirfenidone has a slightly better side-effect profile because of the frequent adverse gastrointestinal effects of nintedanib.

Ethical consideration

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Author contributions

B.K., G.Ç.: Conceptualization, methodology, software, validation, formal analysis; B.K., F.A.: Investigation, resources, data curation; B.K., Ö.A.: Manuscript writing, review & editing; F.A, M.A: Visualization, supervision, project administration

Funding

The authors received no financial support for the research and/or authorship of this article.

Conflict of interest

The authors declare that they have no conflict of interest to the publication of this article.