Analysis of real-world data and a mouse model indicates that pirfenidone causes pellagra

Results of the real-world data analysis. The reported proportions of photosensitivity were compared between a) patients who used pirfenidone (PFD) and/or niacin and b) those who used PFD and/or a multivitamin preparation including niacin. Gastrointestinal events, such as nausea, fatigue, decreased appetite and diarrhoea, were compared between c) patients who used PFD and/or niacin and d) those who used nintedanib and/or niacin. *: significant difference (p<0.05); ^#: no significant difference, but there was a tendency; ns: nonsignificant.

Pellagra-related photosensitivity caused by pirfenidone (PFD). Histological and clinical features of mice treated with or without PFD at day 16 are shown. a, c) Histological and clinical features of mice a) without or c) with ultraviolet (UV) irradiation at day 16. b) No dermal response, such as oedema (ear swelling) and erythema, was observed in mice fed the normal diet with PFD. d) Some dermal responses, such as oedema (ear swelling) and erythema, to UV irradiation were observed in mice fed the normal diet with PFD or the low-niacin diet with PFD. p-values are indicated. e) Skin levels of arachidonic acid (AA) metabolites were analysed by liquid chromatography-mass spectrometry-based methods in mice fed the normal diet (N) or the low-niacin diet (L) with 0 (vehicle), 1 or 3 mg PFD per mouse without UV irradiation (UV–) or with UV irradiation (UV+). Vertical bars indicate the percentage change in each metabolite versus the average of mice fed the normal diet with vehicle without UV irradiation (N0 UV–). The results of the statistical analysis are shown in supplementary table S1. Multiple comparisons were carried out by two-way ANOVA (Tukey–Kramer post hoc test). ^#: we focused on 18-hydroxyeicosatetraenoic acid (18-HETE), 11,12-dihydroxyeicosatrienoic acid (11,12-DHET) and 14,15-DHET. PG: prostaglandin; EET: epoxyeicosatrienoic acid. Data represent mean±se (n=5). Statistical analysis was carried out as described in the supplementary material. The experiments were repeated twice.

Pellagra-related nausea caused by pirfenidone (PFD) (3 mg per mouse). a) Growth curves (body weight % change) in mice fed the normal diet with PFD or the low-niacin diet with PFD (n=5 in each group). b) Each comparison was carried out between both groups. AUC: area under the curve. c) The effect of pica on the shape of faeces from day 10 to 15 (D10–D15) was evaluated in mice treated with PFD kept under the low-niacin diet. Some faeces are indicated and pica began at 5 days onward after the first administration. d) Mice were kept under paper strips stained with carminic acid and coloured faeces were used to evaluate the severity of nausea. Some faeces from mice (n=4) at days 5 and 6 onwards are shown, and nausea severity (nausea score) was evaluated quantitatively using the coloured faeces (n=5). e) Multiple comparisons were carried out by two-way ANOVA (Tukey–Kramer post hoc test). Data are presented as mean±se (n=5). Statistical analysis was carried out as described in the supplementary material. The experiments were repeated twice.

Mouse liver gene expression profiles. a) A cluster dendrogram from gene analysis was prepared to confirm genetic similarity among mice fed the low-niacin diet with vehicle or pirfenidone (PFD) (3 mg per mouse) and mice fed the normal diet with vehicle or PFD. b) Volcano plots between mice fed the normal diet with vehicle and those treated with PFD, mice fed the low-niacin diet with vehicle and those treated with PFD, mice fed the normal diet with vehicle and those fed the low-niacin diet, and mice fed the normal diet with PFD and those fed the low-niacin diet. Each dot represents one gene. Red dots indicate genes with significantly increased or decreased expression between the comparisons.