Pirfenidone

oral antifibrotic (pyridone), IPF

Also sold as Esbriet, Pirespa.

Analyze another medication Medication directory

Findings for this medication

Research about this medication, graded for how strongly each finding can be attributed to it. Nothing here describes your own levels.

Pirfenidone has no classical depletion machinery. It is a small molecule cleared by hepatic CYP1A2 oxidation (with minor CYP2C9, 2C19, 2D6 and 2E1 contributions) and then excreted in urine mostly as 5-carboxy-pirfenidone, with a half-life of about 3 hours [1][2]. The label sections read describe no chelation, transporter, vitamin-cofactor or mineral mechanism, and the only mention of a mineral in the US label sections read is ferric oxide as a tablet colorant [1]. The two real signals are not classical depletions. They are a labeled intake-reducing effect and a labeled, uncommon water-balance disturbance. The water-balance disturbance reaches orange with its limits stated in the entry title; the intake-reducing effect is an outcome entry (Hard Rule 2) and is graded yellow.

oral antifibrotic (pyridone), IPF · ATC L04AX05 (WHO, as listed on Wikipedia [24]) · also sold as Esbriet and Pirespa and Etuary; many generics · reviewed 2026-10-08. This describes research about the medication. It does not say anything about your own levels or health.

How to read the evidence grades

These colours describe how strongly a finding can be attributed to the medication in the research. They are not a severity score and they say nothing about your own health.

Grades describe how strongly a finding can be attributed to Pirfenidone in the research — not how common it is, how serious it would be, and not anything about you personally.

Red — Severe: Two or more independent Tier A sources, this drug in the cohort, the nutrient directly measured, consistent direction, clinically meaningful magnitude, a named mechanism, no substantial contradicting Tier A evidence.
Orange — Moderate: At least one Tier A source with this drug in the cohort showing a real but bounded deficit: an exposure window, a subgroup, low frequency, or a functional rather than circulating deficit.
Yellow — Low: Mechanistic plausibility only (Tier B), or class extrapolation (Tier C) with no drug-specific human status data; also used for context flags and interactions running the other direction.
Not graded — disclosed evidence gap: Nothing has been measured here, so this finding cannot be graded. That does not establish either safety or harm, and it is not the same as a White result.Source definition: Not gradeable under the rules above because nothing has been measured. Published as a documented gap, never as reassurance that there is no effect.

What the research found

Sodium: dilutional hyponatraemia (retention-type disturbance), uncommon, SIADH-type mechanism suggested

Orange — Moderate

See study details

Pirfenidone can lower serum sodium through water retention, listed as an uncommon post-marketing reaction on the EU/UK label and seen in 3 of 52 starters in one centre. Direction: disturbance (excess water), not a sodium deficiency.

Full detail and limits

What the grade means here: At least one Tier A source with this drug in the cohort showing a real but bounded deficit: an exposure window, a subgroup, low frequency, or a functional rather than circulating deficit. Orange

Limits on what this shows

Null findings: No hyponatraemia in 29 nintedanib starters at the same centre [9]. The frequency differs by source: under 1% on the SmPC, 3.4% (vs 0.3% on placebo) in trial supplementary data, 5.8% for "severe" cases locally, about 2% grade 3 in Ogawa [32], and absent from the US label and the published main adverse-event tables. Reconciliation attempted and not achieved: the rates cannot be reconciled from the sources opened because none reports the sodium threshold behind its count. Plausible explanations, not shown by any source: the trial supplementary figure may count any laboratory-detected low sodium while the SmPC counts only attributed adverse reactions (a lower figure); the local rate is from a small retrospective sample with a wide interval (2.0% to 15.6%); and the Ogawa figure counts grade 3 only. All of these frequencies are therefore stated with their source and none is used as the population rate.
At-risk subgroups: Not established. The SmPC singles out people with nausea, headache or dizziness as the ones in whom symptoms should prompt a sodium check [2]. Concomitant sodium-lowering drugs were not assessed in the abstract.

Regulatory status

Two EU/UK labels, both opened. UK SmPC (pirfenidone 534 mg tablets) [2] and the EMA Esbriet SmPC [33]: hyponatraemia is "uncommon" (1 in 1,000 to under 1 in 100) in the section 4.8 table, footnoted as identified through post-marketing surveillance; section 4.4 states "Hyponatraemia has been reported in patients treated with Esbriet" and recommends regular laboratory monitoring because symptoms may be subtle and masked by concomitant morbidities, highlighting nausea, headache or dizziness [33]. The US generic label [1] was re-opened for this repair and specifically queried: it does not mention hyponatremia anywhere (section 6.2 postmarketing lists only agranulocytosis, drug-induced liver injury, angioedema and severe cutaneous adverse reactions; section 6.1 has no laboratory-abnormality table). The US and EU labels therefore differ on this reaction.

Population evidence

Silhan 2017 [9], retrospective single-centre review of pirfenidone or nintedanib starts: severe hyponatraemia in 3 of 52 pirfenidone patients (5.8%; Wilson 95% CI 2.0% to 15.6%, calculated here) vs 0 of 29 nintedanib patients. The abstract states hyponatraemia was absent from the main adverse-event tables of CAPACITY-1, CAPACITY-2 and ASCEND but present in their supplementary data at 3.4% on pirfenidone vs 0.3% on placebo. That trial figure is still secondhand: the abstract wording was confirmed in three independent records of the same paper (the BMJ page, an OpenAlex record and a PubMed listing, which agree word for word on 3.4% vs 0.3%, 3 of 52 and 0 of 29), but they are all the same abstract. The BMJ full-text page is abstract-only (paywalled) and the PMC copy (PMC5720250) returned a bot-verification page, so the sodium values, the severe-hyponatraemia threshold, onset and the reference for the supplementary data could not be read. The CAPACITY/ASCEND supplementary tables were not opened. Second independent cohort: Ogawa 2018 [32], retrospective, 46 Japanese IPF patients on pirfenidone (30 treated over a year, 16 shorter): "only one patient exhibited grade 3 hyponatremia" among grade 3 to 5 adverse events (about 2%, 1 of 46, abstract only; the sodium value, the CTCAE definition and the all-grade rate were not in the abstract).

Mechanism

Present but not proven. The authors state laboratory findings were suggestive of SIADH induced by pirfenidone [9]. The label states the mechanism of action in IPF is not established [1].

Counter-evidence held on record

Counter-evidence and confounding The cohorts are retrospective, small and read as abstracts only; the main trial publications and the US label do not list the event; the SmPC frequency is much lower than the local rate. Nausea and vomiting, which are common on pirfenidone [1], can independently lower sodium. Why orange and not red: the evidence is a bounded, uncommon reaction (label frequency under 1%, one centre at 5.8%, one at about 2%) with the direction consistent across sources. It rests on serum sodium measured in two retrospective cohorts and two labels, so the orange grade does not depend on the secondhand 3.4% trial figure. It does not reach red because the two cohorts are small abstract-level retrospectives with no reported threshold, the trial supplementary data were not opened, frequencies disagree, and the US label is silent. Framing caution This is an excess-water (retention) disturbance, not a dietary sodium deficiency. Eating more salt or taking sodium supplements is not the remedy and could be the hazard. Management is a prescriber decision.

Appetite loss and weight loss (outcome entry: reduced intake effect, no nutrient or intake measured)

Yellow — Low

See study details

Pirfenidone roughly doubles the rate of weight decrease and more than doubles decreased appetite versus placebo, and weight loss of 5% or more per year is common in treated patients. No study measured dietary intake or any nutrient, so this is an outcome entry and not a nutrient deficiency.

Full detail and limits

What the grade means here: Mechanistic plausibility only (Tier B), or class extrapolation (Tier C) with no drug-specific human status data; also used for context flags and interactions running the other direction. Yellow

Limits on what this shows

Null findings: No source measured calories, protein, albumin or micronutrients in pirfenidone users, so no nutrient deficiency is shown. Proesmans [8] found no significant appetite or weight-loss difference between pirfenidone and nintedanib users.
At-risk subgroups: Older patients and those with lower baseline BMI lost more weight in Kim [6]; low BMI predicted early nintedanib discontinuation after prior pirfenidone (multivariate P = 0.0111 in the text, 0.0239 in the abstract) [7].

Regulatory status

US generic label adverse reactions table, pooled Studies 1 to 3 (CAPACITY 004, CAPACITY 006, ASCEND; 623 on 2,403 mg/day vs 624 on placebo, up to 118 weeks) [1]: decreased appetite 21% vs 8%, weight decreased 10% vs 5%, nausea 36% vs 16%, dyspepsia 19% vs 7%, vomiting 13% vs 6%, abdominal pain 24% vs 15%. Dose reduction or interruption for GI events 18.5% vs 5.8%; discontinuation for GI events 2.2% vs 1%. From the label percentages (which are rounded), the risk difference is about 5 points for weight decrease (approximate 95% CI 2.1 to 7.9) and about 13 points for decreased appetite (9.2 to 16.8), calculated here. The EU/UK SmPC lists anorexia as very common (11.4% vs 3.5%) [2].

Population evidence

Kim 2022 [6], retrospective, 215 IPF patients on pirfenidone, Korea: 54 (25.1%) lost 5% or more of weight per year; whole-cohort weight 64.1 to 62.8 kg over a year (P under 0.001); weight loss of 5% or more without pirfenidone was 7.8% (data not shown). Weight loss of 5% or more per year was associated with all-cause mortality: multivariable HR 2.358 (95% CI 1.572 to 3.537) and time-dependent HR 1.751 (1.229 to 2.494). Albumin and nutrition data were not reported; dose reduction was not related to weight loss (P = 0.957). Ikeda 2019 [7], 30 patients who switched from pirfenidone to nintedanib: anorexia during pirfenidone in 63.3% (grade 2 or higher 53.3%); weight loss of 5% or more in 56.7% and of 10% or more in 23.3% (a selected switching group, so likely an overestimate). Proesmans 2019 [8], survey of 71 pirfenidone users, 85 nintedanib users and 20 non-users: anorexia associated with pirfenidone (P under 0.01); the weight-loss result for pirfenidone was internally inconsistent in the text. The CAPACITY abstract lists anorexia among the events more common on pirfenidone [5]; the ASCEND and pooled-analysis abstracts report GI and skin events generally without item-level figures [3][4].

Mechanism

Present (mechanism class H, changed intake): nausea, dyspepsia and early satiety-type GI effects, most frequent in the first 3 months [1]. A cause for the appetite loss itself is not established in any source read. The expert panel notes preclinical data that split dosing partly reduced an effect on gastric motility [22].

Counter-evidence held on record

Counter-evidence and confounding IPF itself causes weight loss; the 5% placebo rate and the 7.8% non-pirfenidone figure in Kim bound that. Kim has no comparator arm and excluded patients who died or worsened early; Ikeda is a selected switching group. Why yellow and not orange or red: weight and appetite are outcomes, no nutrient or dietary intake was directly measured in any cohort (Hard Rule 2), so the entry cannot carry a nutrient-deficiency grade however strong the outcome data are; the cause of appetite loss is also not established. Framing caution Do not restate this as a protein, calorie or vitamin deficiency. It is a monitoring prompt: track weight and appetite, and involve a dietitian when loss is large (the survey authors suggest referral above 10% loss [8]). No supplement is recommended from this entry.

Food and grapefruit juice: nutrient-to-drug timing interaction (direction: food affects the drug)

Yellow — Low

See study details

Food slows pirfenidone absorption and lowers peak concentration; the label tells patients to take it with food to reduce nausea and dizziness, and the EU/UK SmPC says to avoid grapefruit juice. This is the nutrient affecting the drug, not the drug depleting a nutrient.

Full detail and limits

Regulatory status

US label [1]: with food, Tmax about 0.5 to about 3 hours, Cmax about 49% lower and AUC about 16% lower ("food decreased the rate and extent of absorption"); counseling to take with food to help reduce nausea and dizziness. EU/UK SmPC [2]: fed Cmax about 50% lower for capsules and about 40% for tablets; fed AUC about 80 to 85% of fasted in older adults on a single 801 mg dose; grapefruit juice inhibits CYP1A2 and should be avoided; fluvoxamine contraindicated. A consumer reference repeats the food and grapefruit statements [23] (Tier D, not used for grading).

Population evidence

Label pharmacokinetic studies in volunteers; the primary food-effect study was not opened. An expert panel advises dosing at the end of the meal or splitting the three capsules across the meal and notes food lowers peak concentration without changing overall exposure much [22].

Mechanism

Delayed absorption rate (lower, later peak); grapefruit juice inhibition of CYP1A2 raises exposure.

Separation interval

None. The labeled instruction is to take with food, not to separate from food. No mineral or vitamin separation interval appears in the label sections read [1][2].

Counter-evidence held on record

Counter-evidence and confounding The AUC effect is small (about 16 to 20% lower), so reduced total exposure with food is modest; the benefit is tolerability. Taking food with the drug may modestly reduce peak exposure, a trade-off the labels accept. Framing caution This is a timing and tolerability instruction, not a depletion and not a reason to change diet or supplements. Do not skip food or alter dose timing without the prescriber.

Body composition and sarcopenia: indication-linked outcome, no pirfenidone-specific effect shown

Yellow — Low

See study details

Muscle mass is low in IPF cohorts treated with pirfenidone and falls on antifibrotic therapy, but muscle loss is present before treatment and no study separates pirfenidone from disease.

Full detail and limits

Limits on what this shows

Null findings: Faverio [10]: no difference in antifibrotic type, dose reduction or GI events between sarcopenic and non-sarcopenic patients at follow-up. No pirfenidone-specific loss was isolated in any cohort.

Regulatory status

Not mentioned in the labels read [1][2].

Population evidence

Suzuki [11], two Japanese retrospective antifibrotic cohorts (106 and 102 patients; pirfenidone 84 and 58, nintedanib 22 and 44) plus 72 controls: erector spinae muscle area and attenuation on CT at treatment start were lower than controls (P under 0.0001 for area in both cohorts); muscle area predicted death (multivariable HR 0.957, P = 0.0037); no analysis by drug, no longitudinal data, fat not measured. Faverio 2022 [10], prospective, 83 patients before antifibrotics: sarcopenia 22.9% at diagnosis; at about 6 months (51 evaluable, mostly nintedanib among the sarcopenic) no differences by antifibrotic received, dose reduction or GI events; no numeric P values in the text. O'Mahony 2025 [12], 84 patients on pirfenidone, nintedanib or a switch: pectoralis and erector spinae areas fell significantly in all three groups; in BMI under 30, pectoralis loss was greater on nintedanib than pirfenidone; abstract gives no numbers.

Mechanism

No drug mechanism is named. Candidate contributors in the sources are disease stage, a sedentary lifestyle (63.2% vs 41.0% in Faverio [10]) and, as an untested hypothesis here, the intake reduction in the appetite entry.

Counter-evidence held on record

Counter-evidence and confounding Confounding by indication is presumed (Hard Rule 6): low muscle mass predates the drug, and there is no untreated matched comparator for change over time. Sarcopenia is an outcome, not protein or vitamin deficiency (Hard Rule 2). Framing caution Not a pirfenidone-caused deficiency. At most a context flag that weight and muscle should be followed in treated patients.

Vitamin D: hypothetical pathway through photosensitivity and sun avoidance

Yellow — Low

See study details

Photosensitivity is a labeled, common reaction and patients are told to avoid sun, which could in principle lower vitamin D synthesis; no study measured 25-hydroxyvitamin D in pirfenidone users.

Full detail and limits

Limits on what this shows

Null findings: None directly; the question has not been measured.

Regulatory status

US label [1]: photosensitivity reaction 9% vs 1% (mostly in the first 6 months); rash 30% vs 10%. EU/UK SmPC [2]: photosensitivity very common (9.3% vs 1.1%). Neither label mentions vitamin D.

Population evidence

CAPACITY, per an expert-panel review [22]: photosensitivity 12.2% vs 1.7%, rash 32.2% vs 11.5%. The panel advises sun avoidance after dosing, high-SPF broad-spectrum sunscreen and protective clothing, and suggests considering vitamin D deficiency, with a plasma assay, when photosensitivity persists despite good photoprotection. This is a narrative recommendation, not a measurement. One web search (pirfenidone, photosensitivity, 25-hydroxyvitamin D) found no study of 25-hydroxyvitamin D in pirfenidone-treated patients; a follow-up query was blocked when the shared web-search budget ran out. A Mendelian randomization study found no causal link between genetically predicted 25-hydroxyvitamin D and IPF risk (OR 0.891, 95% CI 0.523 to 1.518) [28] but did not study pirfenidone or deficiency in treated patients.

Mechanism

Present as a behavioral pathway only (reduced UVB exposure); untested. The drug has no named effect on vitamin D metabolism in the labels read.

Counter-evidence held on record

Counter-evidence and confounding Older age, winter, low outdoor activity and chronic lung disease all lower vitamin D independent of the drug; baseline prevalence in IPF was not retrieved. No dose appears in this entry. Framing caution Do not present as vitamin D deficiency caused by pirfenidone. A 25-hydroxyvitamin D check is a reasonable monitoring prompt only if photosensitivity persists despite photoprotection, as the expert panel suggests.

The five body-system areas

Each finding is counted once, in a single area. Where a study is mentioned in a second area it is context only.

Mitochondrial Burden

Each finding below is counted once, in this area only.

Mitochondrial burden: preclinical, reverse direction (mitophagy), main in vitro dose far above clinical levels

Yellow — Low

See study details

In lung fibroblasts and mice pirfenidone increases PARK2-dependent mitophagy and lowers mitochondrial ROS in a PARK2-knockdown setting; no human mitochondrial data exist.

Full detail and limits

Limits on what this shows

Null findings: No mitochondrial toxicity signal was found in any source read. The mitochondrial link to the antifibrotic effect was not established [15].

Regulatory status

Not mentioned in the label sections read [1].

Population evidence

Kurita 2017 [15]: primary human lung fibroblasts (including IPF-derived) and bleomycin mice (n = 6 to 10 per group). Pirfenidone raised PARK2 and mitophagy and lowered MitoSOX and total ROS during PARK2 knockdown (P under 0.05; exact values not reported). Main in vitro concentration 500 ug/mL, which the authors describe as about 50-fold above maximum blood concentrations; PARK2 rose significantly at 10 ug/mL, near clinical levels, but other endpoints were not tested there. PINK1 did not change; membrane potential was not measured. The antifibrotic effect in PARK2 knockout mice was PARK2-independent. Plataki 2019 [16]: aged mouse macrophages and lungs improved mitochondrial function with pirfenidone during pneumococcal infection (abstract only).

Mechanism

Named (PARK2-mediated mitophagy) [15]. Direction: reverse or protective.

Counter-evidence held on record

Counter-evidence and confounding Supraclinical concentrations, one bleomycin model and one knockout strain; a news headline overstated this as proof in patients (see Fails verification). Framing caution Not a mitochondrial-damage claim and not a mitochondrial-benefit claim for patients.

Oxidative Stress

Each finding below is counted once, in this area only.

Oxidative stress: reverse direction (antioxidant signalling in preclinical work; no significant change in a 10-patient pirfenidone arm)

Yellow — Low

See study details

Pirfenidone raises antioxidant defences and lowers ROS and lipid peroxidation in cell and mouse models; in 10 treated patients no oxidative stress marker changed significantly.

Full detail and limits

Limits on what this shows

Null findings: Fois [13]: no pirfenidone-arm oxidative stress marker reached significance; the authors call the study preliminary and note the glutathione change is not explained by a mechanism shown in the study. Whole-cohort changes mixed both drugs.

Regulatory status

No oxidative stress statement in the US label sections read [1].

Population evidence

Fois 2020 [13], open-label before-and-after, single centre, 24 weeks, 10 pirfenidone (2,403 mg/day) and 8 nintedanib patients vs 18 matched controls. Pirfenidone arm: kynurenine/tryptophan ratio 0.030 to 0.025 (P = 0.048, an inflammation marker); glutathione +23%, protein thiols +9%, taurine +23% and TBARS -17% all non-significant (no P values given for the arm). At baseline, IPF patients had lower blood glutathione than controls (457 vs 880 umol/L, P under 0.001), before any drug. Preclinical: Liu 2017 [14] (mouse lung fibroblasts and bleomycin mice; abstract only, via a reconstructed OpenAlex abstract): pirfenidone lowered Bach1, raised Nrf2, HO-1 and GPx1, and lowered MDA and ROS. Kurita 2017 [15]: oxidative damage markers (8-OHdG, 4-HNE) reduced in PARK2 knockout mice.

Mechanism

Named: Nrf2/Bach1 rebalancing [14] and PARK2-linked mitophagy [15]. Direction is protective (reverse), not depleting.

Counter-evidence held on record

Counter-evidence and confounding Tiny sample, no placebo, no multiplicity correction, mostly mild disease; baseline glutathione is already low from the disease. A journalistic headline presenting the result as reduced oxidative stress is rebutted under Fails verification. Framing caution Not an oxidative-stress deficiency or harm claim, and not evidence that pirfenidone improves redox status in patients. No antioxidant supplement is implied.

Copper and Iron Balance

Copper and iron are assessed separately. Nothing here is counted twice.

Iron: unattributable signal in a mixed antifibrotic cohort (hepcidin fell, hemoglobin unchanged), CLASS EXTRAPOLATION

Yellow — Low

See study details

The only iron-axis human data come from a 38-patient antifibrotic cohort (13 on pirfenidone) in which hepcidin fell overall and hemoglobin did not change; in the pirfenidone subgroup the hepcidin fall was not statistically significant, and no iron or ferritin value was reported by drug.

Full detail and limits

Limits on what this shows

Null findings: Hemoglobin unchanged at 3 months in the whole cohort [19]; pirfenidone-subgroup hepcidin change not significant (P = 0.073) [19]; no anemia in the US label adverse-reaction tables [1], and the words anaemia, haemoglobin and haemolysis do not appear anywhere in the EMA Esbriet SmPC [33].

Regulatory status

Neither label mentions iron absorption or iron status; the US label lists ferric oxide only as a tablet colorant [1][2].

Population evidence

Yilmaz Gunes 2026 [19], prospective observational cohort, 38 clinically stable IPF patients on antifibrotic therapy, 3 months, full text now read on the MDPI page: 13 on pirfenidone (34.2%) and 25 on nintedanib. Serum hepcidin overall 33.00 +/- 16.74 to 24.41 +/- 12.42 ng/mL (P under 0.001); pirfenidone subgroup 34.89 +/- 12.26 to 27.02 +/- 10.84 ng/mL (P = 0.073, not significant, underpowered); nintedanib 32.01 +/- 18.82 to 23.05 +/- 13.17 (P under 0.001); change did not differ between groups (P = 0.817), which the authors say is not evidence of equivalence. Hemoglobin overall 14.15 +/- 1.51 to 14.28 +/- 1.61 g/dL (P = 0.7024), not reported by drug. Anemia (WHO criteria) was present at baseline in 8 of 38 (21.1%); at baseline, ferritin and transferrin saturation were lower in the anemic group (ferritin median 8.20 vs 39.80 ug/L, P = 0.037; transferrin saturation 10.25% vs 22.81%, P under 0.001). Ferritin, iron and transferrin saturation were not reported over time or by drug; copper and ceruloplasmin were not mentioned. The authors list the single-centre design, small sample, short follow-up, test familiarization and regression to the mean as limitations and state the hepcidin fall cannot be attributed exclusively to antifibrotic therapy. ERJ Open Res 2024 [20], pooled ASCEND and CAPACITY data (pirfenidone 623, placebo 624): baseline hemoglobin was associated with hospitalization; the abstract does not report a drug effect on hemoglobin.

Mechanism

Absent for a depleting effect. Hepcidin is a regulator, not an iron-status measure, and a fall is directionally neutral or even favors absorption; it more plausibly tracks inflammation.

Class cohort test

CLASS EXTRAPOLATION (antifibrotics) for the iron-status question. Pirfenidone was in the cohort (13 of 38) and hepcidin was reported for it, but hepcidin is a regulator and not an iron-status measure, and iron, ferritin, transferrin saturation and hemoglobin were not reported for the pirfenidone subgroup, so no direct iron measurement attributable to this drug exists. Capped at yellow.

Counter-evidence held on record

Counter-evidence and confounding No placebo, short follow-up, 13 pirfenidone patients, 21% anemic at baseline, FVC and DLCO gains attributed by the authors to test familiarization or regression to the mean, which suggests the cohort's physiology shifted for non-drug reasons. Why not higher: no iron or ferritin value is available for pirfenidone patients, and the one drug-specific number (hepcidin) is non-significant and not an iron-status measure. Framing caution Not an iron-deficiency claim. Do not recommend iron on the basis of this entry.

Copper: indirect preclinical signal only (reverse direction, copper-export pathway in colitis mice); no copper status measured in any pirfenidone study

Yellow — Low

See study details

No study measured copper or ceruloplasmin in pirfenidone-treated people. The only Tier B source is a mouse colitis paper in which pirfenidone raised glutaredoxin 1, a protein reported to stabilize the copper exporter ATP7B; this is a pathway signal in the protective direction, not evidence that pirfenidone changes copper status.

Full detail and limits

Limits on what this shows

Null findings: No ceruloplasmin or serum copper measurement in pirfenidone users was found in any channel tried (see Re-check list). The hepcidin cohort [19] did not report copper or ceruloplasmin. Cohort test: pirfenidone is the studied compound in [34] (Tier B, animal, a different disease and tissue from IPF); [35] is Tier C and does not include pirfenidone patients.

Regulatory status

No copper statement in the US label [1] (re-opened: the only mineral-type words are sodium starch glycolate and iron oxide colorants) or in the EMA SmPC [33].

Population evidence

Zhang 2026 [34], Redox Biology, ulcerative colitis tissue and mouse colitis models. From the abstract (PubMed listing and an OpenAlex record, two views of the same abstract): loss of Grx1 caused impaired metal-ion transport, copper overload and cuproptosis in vitro and in vivo; copper chelation reversed colitis in Grx1-deficient mice; GRX1 binds and stabilizes ATP7B, a copper exporter; "pirfenidone increased GRX1 levels and reduced DSS-induced colitis". Pirfenidone dose, route, group sizes and effect sizes could not be read (the PMC copy returned bot-verification and ScienceDirect disallows automated access). The abstract does not state that copper content or ATP7B was measured after pirfenidone. Human IPF context, not pirfenidone-specific: a 2025 narrative review (Nechifor, Medicina) [35] states that a higher copper/zinc ratio is frequently encountered in IPF without numbers or cited studies, names pirfenidone and nintedanib only as drugs that minerals might accompany, and reports no study of copper in pirfenidone-treated patients.

Mechanism

Indirect and unproven for copper status: GRX1 upregulation, then ATP7B stabilization and lower copper toxicity in intestinal epithelium of colitis mice. The copper link is the authors' inference from Grx1-deficient mice, not a measured pirfenidone effect on copper. Direction would be protective (reverse), not depleting.

What was searched

PubMed listing "pirfenidone copper": 7 records, of which only [34] studies pirfenidone against a copper pathway (the others are synthetic chemistry, nanoparticle drug delivery and a hamster pancreatic cancer model with disulfiram plus copper, none measuring copper status in people); [35] was the IPF mineral review. Earlier searches (two web searches, ferroptosis review [27], which cites pirfenidone only for Nrf2, HO-1 and GPX1 upregulation and copper only as cuproptosis background). A combined PubMed query for IPF with serum copper or ceruloplasmin and antifibrotics returned the PubMed home page, and "pirfenidone ceruloplasmin" returned a client error (possibly zero results); neither is counted as a null. Genuine gap for human copper status plus a tooling limit on full text of [34].

Counter-evidence held on record

Counter-evidence and confounding Pirfenidone is a tool compound in a colitis-mechanism paper; IPF itself is associated with a higher copper/zinc ratio [35], which is disease context (confounding by indication) and not a drug effect. The abstract cannot show a copper change after pirfenidone. Framing caution Not a copper deficiency or copper excess claim. No copper or zinc supplement is implied; copper and iron are separate entries and share no source here.

Gut Microbiome Risk

Each finding below is counted once, in this area only.

Gut microbiome risk: mouse-model data only, reverse direction (diversity recovery in diabetic mice)

Yellow — Low

See study details

Pirfenidone shifted gut bacterial diversity toward normal in diabetic mice; no human study of pirfenidone and the gut microbiome was found.

Full detail, limits and 1 references

Limits on what this shows

Null findings: Two other fibrosis-microbiome studies read did not include pirfenidone (a cannabidiol rat study [29] and a germ-free mouse study [30]) and are excluded from the grade.

Regulatory status

Not addressed. The label records GI adverse reactions (nausea 36% vs 16%, diarrhea 26% vs 20%) without microbiome data [1].

Population evidence

Singh 2020 [17], db/db mice, cecal 16S, n = 6 to 7 per group, about 4 weeks of short-acting or long-acting pirfenidone in food pellets: Shannon index db/m vs db/db P = 0.002; short-acting pirfenidone vs db/m P = 0.008; long-acting pirfenidone vs db/m not different (P = 0.149); Akkermansia about 15% in db/db vs about 5% with long-acting pirfenidone; dose not stated; SCFA pathways predicted, not measured. Human IPF microbiome work [18] (conference abstract, 184 IPF vs 146 controls): Streptococcus enriched and linked to worse progression-free survival, with reflux treatment and age also shaping composition; antifibrotic use was not mentioned.

Mechanism

Not established. The authors infer a link between diversity recovery and Akkermansia, which the data do not show.

Counter-evidence held on record

Counter-evidence and confounding Diabetic mouse model with tiny groups, not IPF; human IPF microbiome differences are confounded by disease, reflux drugs and age. Framing caution Not a dysbiosis claim and not a probiotic indication.

Singh et al. Gut microbial changes in diabetic db/db mice and recovery of microbial diversity upon pirfenidone treatment. Microorganisms 2020;8(9):1347. — Mouse 16S study, n = 6 to 7 per group. MDPI page read via a summarizing tool; first-name initials and full author list not captured.

Oxygen and ATP Output

Oxygen delivery and ATP production are assessed separately. Nothing here is counted twice.

ATP production: reverse direction in an aged-mouse infection model, not IPF

Yellow — Low

See study details

Pirfenidone increased ATP production in aged mouse macrophages and lungs during pneumococcal infection; nothing is known in humans or in IPF.

Full detail, limits and 1 references

Limits on what this shows

Null findings: No human ATP, respiration or energy-metabolism measurement with pirfenidone was found.

Regulatory status

Not mentioned in the label sections read [1].

Population evidence

Plataki 2019 [16], aged mice and macrophages (PMID 30700745), abstract read via a reconstructed OpenAlex record: aging lowered ATP production and healthy mitochondria; pre-treatment of aged macrophages with pirfenidone improved mitochondrial function and reduced oxidative stress; in aged mouse lungs pirfenidone reduced superoxide, increased healthy mitochondria and ATP production, and reduced inflammatory cell recruitment and edema. Numbers, doses and group sizes were not retrieved because the PMC full text returned a bot-verification page.

Mechanism

Improved mitochondrial function and lower superoxide (same underlying study as the mitochondrial and oxidative entries).

Counter-evidence held on record

Counter-evidence and confounding Single animal study in an infection model. Pirfenidone's slowing of FVC decline is an intended effect and says nothing about ATP. Framing caution Not an energy-deficit claim and not evidence of a patient benefit.

Plataki ME, et al. Mitochondrial dysfunction in aged macrophages and lung during primary Streptococcus pneumoniae infection is improved with pirfenidone. Sci Rep 2019;9:971. — Aged mouse and macrophage study. PMID 30700745; DOI 10.1038/s41598-018-37438-1. Abstract only, reconstructed from an OpenAlex record; the PMC full text returned bot-verification.

Oxygen delivery/utilization

Not graded — disclosed evidence gap

See study details

No evidence was found that pirfenidone lowers hemoglobin, red-cell mass or tissue oxygen delivery, and no study was found that supports even a mechanistic yellow grade.

Full detail, limits and 2 references

What the grade means here: Nothing has been measured here, so this finding cannot be graded. That does not establish either safety or harm, and it is not the same as a White result. Ungraded

Limits on what this shows

Why no grade: Dual condition met: no Tier B or C evidence supports a yellow grade (a null hemoglobin result in a mixed cohort is not support for an effect) and no circulating oxygen-carrying claim for pirfenidone was found to rebut. Absence of a signal here is not a finding of safety.

What was searched

PubMed listing, pirfenidone AND (anemia OR hemoglobin OR methemoglobin OR hemolysis): 13 records [21]. Repeated in this repair with abstracts (10 of 13 visible): none reports anemia, hemolysis or hemoglobin change caused by pirfenidone; they were trial protocols using hemoglobin-adjusted DLCO, a pooled-trial paper with baseline hemoglobin as a prognostic marker [20], pharmaceutics and case reports (a child with lupus ILD and anemia already present, a dyskeratosis congenita case with pre-existing anemia), and a French fibroid review. The 3 records not shown by the listing tool were not read. The 2001 myelofibrosis phase II record (PMID 11472354) was opened this time through an OpenAlex single-record lookup [36]: 28 patients, 12 completed a year, three withdrew for intolerance, and only one patient had a clinically relevant benefit with respect to anaemia, so pirfenidone neither improved nor, as reported in the abstract, worsened anaemia (this is an efficacy null, not a safety measurement). Hepcidin cohort full text [19]: whole-cohort hemoglobin 14.15 to 14.28 g/dL (P = 0.7024) with 13 of 38 on pirfenidone, not stratified by drug, and anemia present in 8 of 38 at baseline. Labels: the EMA Esbriet SmPC [33] contains no anaemia, haemoglobin or haemolysis wording and lists only uncommon agranulocytosis as a blood event; the US label [1] lists only agranulocytosis in section 6.2 and no hematologic laboratory table. A Saudi Med J 2017 meta-analysis of five RCTs (1,568 participants) [37] was opened at abstract/metadata level only; it reports a higher overall adverse-event rate on pirfenidone without item-level data. A Greek real-world cohort of 80 patients [31] reported skin and GI events only. Crossref title searches for "pirfenidone anemia hemoglobin adverse events" and "pirfenidone erythropoiesis red blood cell hemolysis methemoglobin" returned nothing that studied pirfenidone and red cells.

Channel outcomes

PubMed listing: opened (13 results reported, 10 read). PubMed retry for "pirfenidone AND anemia AND adverse" returned 429 and was not repeated; OpenAlex, Crossref, J-STAGE, MDPI, EMA and DailyMed opened. WebSearch had no budget. Mostly a genuine gap; minor tooling limits remain for the 3 unread listing records and for item-level adverse-event tables in the CAPACITY/ASCEND supplementary data. Slowing of FVC decline and DLCO are disease measures and an intended effect, not oxygen delivery (Hard Rule 2).

Claims checked and not supported

These circulate online for this medication. They are listed here only so they can be answered, never as findings, and they are not graded.

A claim with no supporting evidence is not necessarily disproven. Where evidence actually argues against a claim, that is stated below.

Mitochondrial repair in patients ("Esbriet stops lung fibrosis by improving mitochondria")
Where it comes from: Patient-news article summarizing Kurita 2017 [25][15].
Why it fails: The underlying work is human lung fibroblasts and bleomycin mice, with a main concentration about 50-fold above maximum clinical blood levels, no membrane potential measurement, and a PARK2-independent antifibrotic effect in knockout mice [15]. No human mitochondrial measurement exists. The narrower preclinical finding is graded yellow above; the human-level claim fails. Not a depletion claim.
Lower systemic oxidative stress in IPF patients on Esbriet
Where it comes from: Patient-news article on Fois 2020 [26][13].
Why it fails: In the pirfenidone arm (n = 10) only the kynurenine/tryptophan ratio changed significantly (P = 0.048), and that is an inflammation marker. Glutathione, protein thiols, taurine and TBARS did not change significantly [13]. The significant glutathione and ADMA changes in the article belong to the nintedanib arm. Rebutted as a pirfenidone human claim.

How the body handles this medication

Metabolism and enzymes: Hepatic. CYP1A2 is the main enzyme (about 70 to 80% per the EU/UK SmPC), with minor CYP2C9, 2C19, 2D6 and 2E1. The main plasma metabolite is 5-carboxy-pirfenidone. CYP induction or inhibition by pirfenidone itself was not extracted from the label sections read.[1] US generic label 12.3; [2] SmPC 5.2
Elimination and impairment: Terminal half-life about 3 hours; about 80% of the dose is recovered in urine, mostly as the 5-carboxy metabolite. Child-Pugh B hepatic impairment raises pirfenidone AUC about 1.6-fold. Mild, moderate and severe renal impairment raise parent AUC about 1.4, 1.5 and 1.2-fold and metabolite AUC about 1.7, 3.4 and 5.6-fold.[1] 12.3; [2] 4.4
Absorption and food effect: Fasted Tmax about 0.5 hours; with food about 3 hours. Food lowers Cmax about 49% and AUC about 16% (US label). The EU/UK SmPC gives Cmax about 50% lower for capsules and about 40% lower for tablets. No named absorption transporter appears in the sections read. Plasma protein binding about 58%.[1] 12.3; [2] 5.2
CYP1A2 interactions and smoking: Fluvoxamine raises exposure about 4-fold in non-smokers (about 7-fold in smokers per the US label) and is contraindicated in the EU/UK SmPC. Ciprofloxacin raises exposure about 81%. Smokers have roughly half to two-thirds of non-smoker exposure. The SmPC advises avoiding grapefruit juice, a CYP1A2 inhibitor.[1] sec 7, 12.3; [2] 4.5, 5.2
Hepatic and skin safety signals: ALT or AST at least 3 times the upper limit of normal: 3.7% vs 0.8% on placebo; at least 10 times: 0.3% vs 0.2%; fatal liver injury has been reported post-marketing. Photosensitivity 9% vs 1%. These are drug toxicities and outcome measures, not nutrient findings.[1] sec 5.1, 5.2; [2] 4.4
Water and electrolyte handling: Hyponatraemia is listed as uncommon and identified post-marketing in the EU/UK SmPC, with a recommendation to monitor relevant laboratory parameters, especially with nausea, headache or dizziness. The label states the mechanism of action in IPF has not been established.[2] 4.4, 4.8; [1] 12.1

What the research says is worth watching

General information from the medication label and the studies above. It is not personal medical advice, and any change belongs with your prescriber.

Lab / symptoms
Serum sodium
Worth checking if nausea, headache, dizziness or confusion develop, as the EU/UK SmPC advises monitoring relevant laboratory parameters [2]. A fall is a water-balance problem for the prescriber to manage, not a prompt to add salt.
Lab / label
Liver tests
ALT, AST and bilirubin at baseline, monthly for the first 6 months, then every 3 months per the SmPC [2]. This is drug toxicity monitoring, not a nutrient finding.
Clinic visit
Weight and appetite
Record weight at each visit. Weight loss of 5% or more per year was associated with higher mortality in one cohort [6]. A survey-based algorithm suggests dietitian referral above 10% loss [8]. Dose reduction and re-titration for GI effects are prescriber decisions.
Timing
Food, grapefruit, separation intervals
Take with food as labeled [1]. Avoid grapefruit juice [2]. No vitamin or mineral separation interval is required by the labels read. No supplement dose is recommended or appears anywhere in this dossier, so no upper-limit comparison was needed.
Skin / lab
Photoprotection and vitamin D
Sun protection is the labeled measure. A 25-hydroxyvitamin D check is reasonable only if photosensitivity persists despite good photoprotection [22].
Lab / low priority
Iron indices and blood count
A blood count and ferritin are indicated only for clinical reasons such as fatigue or pallor. No drug-attributed anemia signal was found [1][2].
Special populations
Renal, hepatic, pregnancy, pediatrics
Renal impairment raises the 5-carboxy metabolite up to about 5.6-fold; Child-Pugh B raises parent AUC about 1.6-fold, and severe hepatic impairment is contraindicated in the EU/UK SmPC [1][2]. Pregnancy and pediatric nutrient considerations were not assessed from the label sections read.
Re-check list
Items to revisit
CAPACITY/ASCEND supplementary tables and Silhan full text [9] for the sodium threshold and rate; Ogawa full text [32]; PMC full text of [34] for the pirfenidone dose and any copper measurement; a ceruloplasmin or serum copper study in pirfenidone users; full text of [16]; a 25-hydroxyvitamin D study in treated patients; the Esbriet brand label on DailyMed.

Overlap map, to prevent double counting

Mitochondrial burden
Shares evidence with: ATP production; Oxidative stress
Note: Plataki 2019 [16] supports both mitochondrial and ATP entries, and Kurita 2017 [15] supports both mitochondrial and oxidative stress. These are two papers, not four independent lines.
Oxidative stress
Shares evidence with: Fails verification: oxidative stress news claim
Note: The same Fois study [13] underlies the yellow entry and the rebuttal of the news claim.
Iron
Shares evidence with: Oxygen delivery/utilization (ungraded)
Note: The hepcidin cohort [19] (hemoglobin overall unchanged, anemia in 8 of 38 at baseline) and the pooled-trial hemoglobin paper [20] are cited in both. Any hemoglobin effect would run through iron, so oxygen delivery is not counted separately.
Appetite loss and weight loss
Shares evidence with: Body composition and sarcopenia; Food and grapefruit juice
Note: Kim 2022 [6] and the label table [1] underlie the appetite entry only. The sarcopenia entry shares the intake hypothesis (untested) and the label GI data. The food entry shares the label [1] and expert panel [22].
Sodium
Shares evidence with: Appetite loss and weight loss
Note: Nausea is a common confounder of both, and the SmPC [2][33] ties sodium checks to nausea. Silhan [9] and Ogawa [32] are independent cohorts used only for sodium; the 3.4% vs 0.3% trial figure is quoted within [9] and is not a third source.
Vitamin D
Shares evidence with: Food and grapefruit juice
Note: The expert panel [22] supports both for different statements (sun protection and vitamin D versus dosing with food).
Copper
Shares evidence with: Iron; Oxidative stress; Gut microbiome risk
Note: No shared source with Iron (graded separately per Hard Rule 7). The copper entry rests on one mouse colitis paper [34] (a redox-enzyme and intestinal model) that is not cited in the oxidative stress or gut entries; the shared theme is only redox biology and the intestine, so it is not independent confirmation of either. The Nechifor review [35] is context and supports no other entry.
Gut microbiome risk
Shares evidence with: Appetite loss and weight loss
Note: No shared source; the label GI reactions [1] are context only and are not microbiome evidence.

All references used above

Pirfenidone

Findings for this medication

What the research found

Sodium: dilutional hyponatraemia (retention-type disturbance), uncommon, SIADH-type mechanism suggested

Appetite loss and weight loss (outcome entry: reduced intake effect, no nutrient or intake measured)

Food and grapefruit juice: nutrient-to-drug timing interaction (direction: food affects the drug)

Body composition and sarcopenia: indication-linked outcome, no pirfenidone-specific effect shown

Vitamin D: hypothetical pathway through photosensitivity and sun avoidance

The five body-system areas

⚡Mitochondrial Burden

Mitochondrial burden: preclinical, reverse direction (mitophagy), main in vitro dose far above clinical levels

💥Oxidative Stress

Oxidative stress: reverse direction (antioxidant signalling in preclinical work; no significant change in a 10-patient pirfenidone arm)

⚖️Copper and Iron Balance

Iron: unattributable signal in a mixed antifibrotic cohort (hepcidin fell, hemoglobin unchanged), CLASS EXTRAPOLATION

Copper: indirect preclinical signal only (reverse direction, copper-export pathway in colitis mice); no copper status measured in any pirfenidone study

🌿Gut Microbiome Risk

Gut microbiome risk: mouse-model data only, reverse direction (diversity recovery in diabetic mice)

💨Oxygen and ATP Output

ATP production: reverse direction in an aged-mouse infection model, not IPF

Oxygen delivery/utilization

Claims checked and not supported

Mitochondrial repair in patients ("Esbriet stops lung fibrosis by improving mitochondria")

Lower systemic oxidative stress in IPF patients on Esbriet

How the body handles this medication

What the research says is worth watching

Overlap map, to prevent double counting

Mitochondrial Burden

Oxidative Stress

Copper and Iron Balance

Gut Microbiome Risk

Oxygen and ATP Output