CYP1A2 Polymorphism and Drug Co-administration Affect the Blood Levels and Adverse Effects of Pirfenidone

Supplemental Digital Content is Available in the Text.

Abstract

Background:

Mutations in metabolic enzymes and co-administration of drugs may affect the blood concentration of pirfenidone effective in pulmonary fibrosis. To provide a basis for the precise clinical use of pirfenidone, the authors analyzed the correlation between steady-state pirfenidone trough concentration and adverse drug reactions (ADRs) and examined the impact of CYP1A2*1C (rs2069514) and *1F (rs762551) variants and co-administration on pirfenidone blood concentrations and ADRs.

Methods:

Forty-four patients were enrolled. The blood concentration of pirfenidone was determined using high-performance liquid chromatography. CYP1A2*1C and *1F genotypes were determined using direct SNP sequencing. Additional information related to drug associations was collected to screen factors affecting drug metabolism.

Results:

The highest predictive value of ADRs was observed when the steady-state trough concentration of pirfenidone was 3.18 mcg·mL⁻¹ and the area under the receiver operating characteristic curve was 0.701 (P = 0.024). The pirfenidone concentration-to-dose ratio (C/D) in CYP1A2*1F homozygous AA mutants was lower than that in C carriers (CC+AC) (1.28 ± 0.85 vs. 2.03 ± 1.28 mcg·mL⁻¹; P = 0.036). Adverse drug reaction (ADR) incidence in the homozygous AA mutant group (28.0%) was significantly lower than that in the C carriers (CC+AC) (63.2%; P = 0.020), and ADR incidence in the A carriers (AC+AA) was considerably lower than that in the CC group (85.7%; P = 0.039). The C/D value of the combined lansoprazole/rabeprazole group was lower than that of the noncombination group (P < 0.05).

Conclusions:

The ADR incidence was positively correlated with pirfenidone blood concentration. The CYP1A2 (rs762551) AA genotype is associated with lower pirfenidone concentrations and fewer ADRs. Lansoprazole/rabeprazole co-administration reduced pirfenidone concentrations. Randomized controlled trials should further explore personalized dosing of pirfenidone and combination therapies.

INTRODUCTION

Pulmonary fibrosis (PF) is a common irreversible interstitial lung disease that involves multiple immune cells and inflammatory factors, eventually leading to respiratory failure or even death. There are 3 main types of PF: secondary, idiopathic pulmonary fibrosis (IPF), and hereditary. IPF is the most severe and common type (65%).¹ IPF's incidence and prevalence in Asia-Pacific are 0.35–1.30 and 0.57–4.51 per 10,000 people, respectively.¹ Among untreated patients, the median survival is 2–3 years,² and IPF is referred to as “a cancer that is not cancer.”³

Pirfenidone (PFD) is an effective treatment for PF. It inhibits fibrotic activity by affecting the intracellular signaling pathway of transforming growth factor-β, the transformation of fibroblast phenotype to myofibroblasts, and fibrotic activity by suppressing the development of Nod-like receptor P3 inflammatory vesicles in the myocardium.^4,5 However, adverse drug reaction (ADR) incidence in PFD therapy is high, often resulting in the discontinuation or change of therapy. A multicenter study of 502 patients with long-term PFD found that 186 patients (37.1%) terminated treatment within a year. ADRs were the primary cause of treatment termination (77 patients, 41.1%) and the second leading cause of death (60 patients, 32.3%).⁶

PFD efficacy necessitates long-term use, and there is notable individual variability in response to PFD. PFD is metabolized by cytochrome P450 enzymes in the body, including CYP1A2 (72.8%), 2C19 (11.8%), and 2D6 (8.9%).⁷ Owing to gene mutations, the activity of metabolic enzymes may increase, decrease, or become null, leading to significant differences in the treatment efficacy and ADRs at the same dose.⁸ Therefore, knowledge of the relevant gene loci before treatment is beneficial for selecting antifibrotic drugs and determining effective maintenance doses for PFD.

In addition, many drugs interact with PFD and co-administration often affects the activity of metabolic enzymes. Therefore, the precise dosing of PFD is crucial.

We examined the effects of CYP1A2 polymorphisms and a real-world combination therapy on PFD blood concentrations and ADRs to address these objectives.

MATERIALS AND METHODS

Study Design

This was a prospective, randomized, observational, single-center study.

Patients

Patients were diagnosed with PF diseases, such as IPF, interstitial PF, interstitial pneumonia, and secondary PF, between January 2022 and December 2022 at Sichuan Provincial People's Hospital, China. The study was approved by the Ethics Committee of Sichuan Provincial People's Hospital [no. LSE (Research) 2022-152]. The patients and their families signed an informed consent form.

All participants met the following inclusion criteria: (1) patients diagnosed with PF diseases, including IPF, interstitial PF, interstitial pneumonia, and secondary PF, who had been taking PFD regularly and (2) complete patient information, including disease and medication, biochemical indicators such as liver and kidney function, ADRs, and PFD blood concentration.

Patients were excluded based on the following criteria: (1) children and pregnant or lactating women; (2) patients who lacked liver and kidney function indicators before and after concentration testing or had severe liver and kidney dysfunction; (3) abnormal blood test results and extreme abnormalities, including those below the lower limit of quantification of the assay or abnormally high values; (4) patients who did not agree to a blood test or could not cooperate, including those with mental illnesses; and (5) other patients who did not meet the inclusion criteria.

Dosing Regimen

Patients were administered PFD in capsule or tablet form. Each capsule (Aisirui) contained 100 mg of PFD (Beijing Contini Pharmaceutical, Beijing, China); the medication was approved by the China Food and Drug Administration (approval number: H20133376). Each tablet (Anbosi) contained 200 mg of PFD (Beijing Kain Technology, Beijing, China). The tablets were approved by the China Food and Drug Administration (approval number: H20193259). Depending on the condition and tolerance of the patient's body, patients could take a single dose of 200–600 mg thrice a day; the same dose was taken regularly for at least 2 days before the concentration was measured.

ADR Evaluation and Sample Testing

The Naranjo ADR Probability Scale was used to determine the correlation between ADRs and PFD.⁹ This study recognized ADRs that occur with PFD as “definite” and “probable.”

PFD concentration was determined by high-performance liquid chromatography, with a measurement range of 0.15–37.44 mcg·mL⁻¹. Three milliliters of venous blood was collected from the median cubital vein of the patient 0.5 hours before taking PFD in the morning and were processed and analyzed using an established PFD blood concentration detection method.¹⁰ The quality control of samples showed that the method displayed good precision (coefficient of variation not exceeding 5.72%) and accuracy (measured concentration not exceeding ±6.24% of the target concentration).¹⁰

Direct SNP sequencing was used to identify CYP1A2*1C (rs2069514) and *1F (rs762551). DNA was extracted using the Trelief Animal Genomic DNA Kit and the Animal Genome DNA Extraction Kit (Tsingke Biotech, Beijing, China). Synthesized primers and other related reagents were purchased from Tsingke Biotech.

Statistical Analysis

Statistical analysis was performed using SPSS software (version 27.0; IBM, Armonk, NY). Count data were expressed as frequency (n), percentage, or rate (%), and between-group comparisons were made using the χ² test or Fisher exact probability method. The normal distribution and skewed distribution were expressed as the mean ± SD and median (lower quartile and upper quartile), respectively. Between-group differences were compared using t tests, Mann–Whitney U tests, and Kruskal–Wallis H tests. Population genetic balance was tested using Hardy–Weinberg equilibrium. Statistical significance was set at P < 0.05. The drug blood concentration, expressed as concentration-to-dose ratio (C/D) was used as the evaluation index. The efficacy of blood PFD concentrations in predicting the occurrence of ADRs was analyzed using receiver operating characteristic (ROC) curves.

RESULTS

Patient Genotype

Forty-four patients were included in the study. The patient inclusion process is illustrated in Supplemental Digital Content 1 (see Figure, http://links.lww.com/TDM/A738). The patients were followed up for 3 months. The G and A allele frequencies in CYP1A2*1C were 62.50% and 37.50%, respectively, whereas those of C and A in CYP1A2*1F were 29.55% and 70.45%, respectively. All CYP1A2 genotype frequencies were in Hardy–Weinberg equilibrium (P > 0.05). This study examined patients who did not smoke or drink alcohol during the medication period. Patient characteristics are shown in Table 1.

TABLE 1.

General Information for Patients

Parameter	Values/Frequencies (n = 44)
Sex (male, %)	19 (43.18)
Age (yrs)	65.05 ± 11.81
BMI (kg·m−2)	22.81 ± 3.54
Inclusion of disease type
Secondary PF (frequency, %)	21 (47.73)
IPF (frequency, %)	10 (22.73)
Interstitial PF (frequency, %)	9 (20.45)
Interstitial pneumonia (frequency, %)	4 (9.09)
Drug varieties
Pirfenidone capsules (AiSiRui) (frequency, %)	41 (93.18)
Pirfenidone tablets (AnBosi) (frequency, %)	3 (6.82)
CYP1A2*1C (%)
GG	15 (34.09)
AA	4 (9.09)
AG	25 (56.82)
CYP1A2*1F (%)
CC	7 (15.91)
AA	25 (56.82)
AC	12 (27.27)

PFD Blood Concentration

The number of patients taking 200 mg in a single dose in this study was 15, followed by 12 each in the 300 and 400 mg groups; 500 and 600 mg were administered to 2 and 3 patients, respectively. The concentration of PFD in the included patients was 1.65 ± 1.46 mcg·mL⁻¹, with a maximum concentration value of 5.67 mcg·mL⁻¹ and a minimum value of 0.16 mcg·mL⁻¹. When using the C/D value as the evaluation index, we did not find any significant effects of sex, age, body mass index (BMI), disease type, or various biochemical indicators on PFD blood concentration (see Table, Supplemental Digital Content 1, http://links.lww.com/TDM/A738).

ADRs

Nineteen individuals experienced 24 ADRs, which were mostly mild-to-moderate and tolerable. Gastrointestinal ADRs were the most frequent (n = 21) (see Table, Supplemental Digital Content 1, http://links.lww.com/TDM/A738). We did not find any significant influence of dose, sex, age, or BMI on ADRs (see Table, Supplemental Digital Content 1, http://links.lww.com/TDM/A738). As most of the patients' biochemical indicators were within normal limits, no statistical analyses were performed.

Correlation Between PFD Blood Concentration and ADRs

In this study, ADR incidence in the high concentration group (≥1.15 mcg·mL⁻¹ at steady state) was notably higher than that in the low concentration group (<1.15 mcg·mL⁻¹) (59.10% vs. 27.30%; P = 0.033). ADR incidence in the high concentration group (≥0.96 mcg·mL⁻¹) of C/D was significantly higher than that in the low concentration group (<0.96 mcg·mL⁻¹) (55.20% vs. 20.00%; P = 0.026; Table 2).

TABLE 2.

Influence of PFD Blood Drug Concentration on ADRs

Projects	Groups	Non-ADRs	ADRs	χ 2	P
Steady-state trough concentration	Low-concentration	16 (72.70)	6 (27.30)	4.539	0.033*
	High-concentration	9 (40.90)	13 (59.10)
PFD C/D	Low-concentration	12 (80.00)	3 (20.00)	4.985	0.026*
	High-concentration	13 (44.80)	16 (55.20)

^*p < 0.05, indicating statistical significance.

The ROC curve for predicting ADRs based on PFD concentration was plotted. The area under the curve (AUC) was 0.701 (P = 0.024; 95% CI, 0.546–0.856; Fig. 1).

FIGURE 1.

ROC curve for predicting ADRs based on PFD concentration.

Correlation Between Gene Polymorphisms and PFD Blood Concentration

PFD C/D and ADR occurrence rates were compared after grouping patients according to CYP1A2*1C and *1F genotypes. For CYP1A2*1F, regardless of whether proton-pump inhibitors (rabeprazole and lansoprazole) were used, the PFD C/D ratio of C allele carriers (CC+AC) was higher than that of AA homozygous individuals (Fig. 2C). However, the difference was statistically significant in all patients (Fig. 2D); (CC+AC) individuals had a significantly higher PFD C/D ratio than those carrying AA (2.03 ± 1.28 vs. 1.28 ± 0.85 mcg·mL⁻¹, P = 0.036).

FIGURE 2.

The effect of CYP1A2 on PFD concentration. A, Comparison of the concentration differences between the AA and GG/GA genotypes of CYP1A2*1C within each group, grouped based on whether patients use proton-pump inhibitors. B, Overall concentration differences between the AA and GG/GA genotypes of CYP1A2*1C in all patients. C, Comparison of the concentration differences between the AA and CC/AC genotypes of CYP1A2*1F within each group, grouped based on whether patients administered proton-pump inhibitors. D, Overall concentration differences between the AA and CC/AC genotypes of CYP1A2*1F in all patients. *, P < 0.05, indicating statistically significant differences. Circles represent outliers.

However, the difference in PFD C/D was not statistically significant (P > 0.05) when compared between genotypes within each subgroup of CYP1A2*1C (Figs. 2A, B).

Correlation Between Gene Polymorphisms and ADRs

For CYP1A2*1F, the incidence of ADR in the AA group was lower than that in the CC group (28.00% vs. 85.70%, P = 0.010). In the CC group, the ADR incidence was significantly higher than that in the A allele carriers (AC+AA) (85.7% vs. 35.1%, P = 0.039). In the AA group, the ADR incidence was notably lower than that in the C allele carriers (CC+AC) (28.0% vs. 63.2%, P = 0.020; Table 3). No correlation was observed between CYP1A2*1C and the ADRs.

TABLE 3.

Association Between CYP1A2 Polymorphism and ADRs of PFD

Gene	Groups	Genotypes	Non-ADRs	ADRs	χ2	P
CYP1A2*1C rs2069514	Group 1	GG	9 (60.00)	6 (40.00)	0.094	0.759
		GA+AA	16 (55.20)	16 (44.80)
	Group 2	AA	2 (50.00)	2 (50.00)	0.000	1.000
		GG+GA	23 (57.50)	17 (42.50)
CYP1A2*1F rs762551	Group 1	CC	1 (14.30)	6 (85.70)	7.501	0.018*
		AC	6 (50.00)	6 (50.00)
		AA	18 (72.00)	7 (28.00)†
	Group 2	CC	1 (14.30)	6 (85.70)	4.249	0.039*
		AC+AA	24 (64.90)	13 (35.10)
	Group 3	AA	18 (72.00)	7 (28.00)	5.439	0.020*
		CC+AC	7 (28.00)	12 (63.20)

^*p < 0.05, indicates statistical significance.

^†The comparison between AA type and CC type, with p = 0.010, indicating a statistically significant difference.

Correlation of Combination Drugs with PFD Concentrations

In this study, many patients had multiple comorbidities and received combination therapies with various drugs. Of the 44 patients prescribed PFD, 35 were on concurrent medications, with an average of 4.47 medications per patient and a range of 1–10 medications. Only 9 patients were on PFD monotherapy. Commonly used medications included glucocorticoids (27 individuals, including prednisone for 15 individuals and methylprednisolone for 8 individuals), proton-pump inhibitors (rabeprazole and lansoprazole) for 18 individuals, calcium supplements for 16 individuals, and statins for 8 individuals. It was found that the PFD C/D was 1.48 (1.02–3.21) mcg·mL⁻¹ in patients not receiving proton-pump inhibitors, which was significantly higher than that in the co-administered group 0.96 (0.51–1.88) mcg·mL⁻¹ (P = 0.03, see Figure, Supplemental Digital Content 1, http://links.lww.com/TDM/A738). However, we did not find any effect of the other drugs on PFD blood concentrations.

Correlation of Combination Medications with ADRs

The differences in ADR incidence between the groups treated with or without a combination of corticosteroids, proton-pump inhibitors, or calcium supplements were not statistically significant (P > 0.05; see Table, Supplemental Digital Content 1, http://links.lww.com/TDM/A738).

DISCUSSION

PFD is an effective drug to treat PF. However, considerable individual variability exists in treatment efficacy and ADRs, possibly related to mutations in enzymes that metabolize PFD and drug co-administration affecting the PFD blood concentration. Here, we examined the correlation between PFD trough steady-state concentration and ADRs and investigated the effect of CYP1A2 polymorphisms and real-world combination therapy on PFD blood concentrations and ADRs.

We found a significant correlation between the PFD steady-state trough concentration and the dose. When the C/D value was used as the evaluation index, we did not find any influence of sex, age, BMI, disease type, or various biochemical indicators on PFD blood concentration. Similarly, this study did not find any influence of dose, sex, age, BMI, or disease type on PFD-associated ADRs. The steady-state trough concentration and C/D of PFD were, however, found to be positively associated with ADR occurrence. Moreover, proton-pump inhibitor (lansoprazole/rabeprazole) co-administration was associated with lower PFD concentrations, and the CYP1A2 (rs762551) AA genotype was associated with lower PFD concentrations and fewer ADRs.

Gastrointestinal ADRs were the most common, consistent with the findings of other studies.¹¹ Furthermore, ADR incidence was positively associated with the steady-state trough concentration and C/D of PFD, consistent with the results of Li et al,¹² that the overall incidence of adverse events was higher in the high concentration group (86.0%) than in the low concentration group (62.5%). However, Li et al used AUC monitoring analysis of PFD with poor clinical compliance. The results of the present study predict that both steady-state trough concentration and C/D value monitoring of PFD are easy to manipulate clinically, and that both may replace AUC monitoring to guide the clinical dosing regimen of PFD.

Lansoprazole is an effective CYP2C19 inhibitor. Rabeprazole thioether, a metabolite of rabeprazole, can effectively inhibit CYP2C19 activity in vivo,¹³ thus reducing drug metabolism and increasing PFD concentration. However, lansoprazole and rabeprazole inhibit gastric acid secretion. Because PFD is an acidic drug (acid dissociation constant: −0.16 ± 0.63), its absorption in the stomach may be reduced or delayed, leading to a lower blood drug concentration.

Clinical relevance of lansoprazole/rabeprazole in combination with PFD has not been reported in the literature, but 1 study found similar results.¹⁴ GDC-3280 is a benzopyridone scaffold drug based on PFD designed to improve the pharmacokinetics, antifibrotic activity, and tolerability of PFD.¹⁴ A trial found that co-administration of 20 mg of rabeprazole led to a reduction in GDC-328 exposure of approximately 13%–25% compared with that seen with the administration of GDC-3280 alone, under fasting and low-fat feeding conditions.¹⁴ Our findings also emphasize the importance of monitoring the use of combinations, especially proton-pump inhibitors, recommended in clinical treatment guidelines¹⁵ for patients with IPF treated with PFD. However, the underlying mechanism remains unclear.

In previous studies, the homozygous mutation type AA of CYP1A2*1F was often found to be associated with accelerated clearance of drugs such as caffeine, carbamazepine, and antipsychotics and fewer ADRs.^16–18 We found similar results for PFD in our study. This is because the mutation allele A of rs762551 can lead to increased CYP1A2 enzyme activity; studies have shown that enzyme activity in individuals with AA is approximately 40% higher than that in C carriers.^19–21 Therefore, PFD clearance is accelerated, resulting in lower PFD concentrations and fewer ADRs. Smoking is a well-known inducer of CYP1A2 expression; however, none of the patients in this study smoked during the medication period.

CYP1A2*1C can reduce enzymatic activity; however, this study did not observe statistically significant differences in PFD C/D between mutation carriers and noncarriers. This discrepancy may be attributed to the small sample size, insufficient inclusion of mutation carriers, or other confounding factors.

In addition to the metabolizing enzyme CYP1A2 (72.8%), PFD is metabolized in vivo by 2C19 (11.8%) and 2D6 (8.9%), with 2C9 and 2E1 playing a minor role.⁷ Mutation sites for other metabolic enzymes similarly affect the in vivo concentration of PFD. CYP2C19 *2 (rs4244285), CYP2C19 *3 (rs4986893), CYP2D6 *5 (rs3892097), CYP2D6 *10 (rs1065852), CYP2C9 *2 (rs1799853), and CYP2C9 *3 (rs1057910) are common mutation sites in Chinese patients. Except for rs1057910, which increased the activity of the corresponding enzyme, the first 5 mutation sites led to a decrease in the activity of the corresponding enzymes and slowed down the metabolism of PFD in vivo. However, the mutation rates of rs4986893, rs3892097, rs1799853, and rs1057910 were <10% in the Chinese population. More importantly, except for CYP1A2 (72.8%), the metabolic proportions of other enzymes, such as 2C19 (11.8%), 2D6 (8.9%), 2C9, and 2E1, account for an even lower percentage.⁷ Therefore, if these enzymes are mutated, changes in PFD concentration may also not be easily observed.

This clinical study had some limitations. It had a small sample size and was confounded by factors, such as patients taking PFD with meals, differences in meal type and quantity, patient compliance, and the impact of the original diseases. In addition, we did not collect data on patients who continued taking PFD after experiencing severe ADRs. Furthermore, the study did not categorize or grade the ADRs or include patient-efficacy indicators.

CONCLUSIONS

The primary significance of this study is showing that the blood concentration of PFD is reduced when co-administered with lansoprazole/rabeprazole. Prescribing physicians may be required to monitor PFD blood concentrations more closely when using PFD in combination with these drugs or to avoid co-administration when possible. Moreover, the correlation between the homozygous AA phenotype of the CYP1A2*1F (rs762551) mutation and reduced PFD C/D and ADR incidences suggests the need for well-designed randomized controlled trials to further explore the personalized use of PFD and combination therapies.