Idiopathic pulmonary fibrosis (IPF) is a chronic, progressive interstitial lung disease with poor prognosis [1,2]. Antifibrotic therapies such as pirfenidone slow disease progression but are frequently limited by gastrointestinal (GI) adverse events (AEs), which may compromise treatment adherence [3–6]. Despite their clinical relevance, the role of dietary factors in modulating these events remains unclear. In particular, the type of added fat may influence GI tolerance. We therefore examined the association between predominant added fat consumption (olive oil, butter, margarine, and sunflower oil) and GI outcomes within the MADIET clinical trial.

In this post-hoc analysis of the MADIET clinical trial (NCT03539289), a total of 86 patients with IPF treated with pirfenidone were included and followed for 16 weeks [6]. Of the 95 patients initially screened, six were excluded due to not following a clearly defined dietary pattern based on monounsaturated (MUFA) or saturated fatty acids, and three additional patients were excluded due to study withdrawal. Participants were instructed to maintain their habitual diet, including both food choices and cooking practices, and to follow standard pirfenidone dosing recommendations. Adherence to dietary habits was monitored using patient diaries, which recorded daily food intake, changes in treatment, and the occurrence of AEs, with overall compliance exceeding 80%. Habitual dietary intake was assessed using a validated food frequency questionnaire (FFQ) [7], and participants were categorized according to the predominant type of added fat used for cooking and dressing (olive oil, butter, margarine, or sunflower oil). In cases where more than one type of fat was consumed, classification was based on the most frequently reported fat. Olive oil was used as the reference category in all analyses. GI AEs were prospectively collected during follow-up and classified according to standardized criteria. The primary outcomes were the presence and total number of GI AEs recorded throughout the study period, while the secondary outcome was the occurrence of nausea and vomiting. Descriptive analyses were performed to summarize baseline characteristics across added fat groups. Continuous variables were expressed as mean±standard deviation, and categorical variables as frequencies and percentages. Associations between type of added fat and GI outcomes were evaluated using regression models according to outcome type. A total of nine primary statistical comparisons were performed across the three gastrointestinal outcomes evaluated. Logistic regression was applied to binary outcomes (presence of at least one AE), and Poisson regression with robust variance was used for count data (number of AEs), with results expressed as odds ratios (ORs) and incidence rate ratios (IRRs) with 95% confidence intervals, respectively. Multivariable models were adjusted for age, sex, and body mass index (BMI), as inclusion of additional variables such as alcohol consumption did not materially modify the associations.

Forty eight (55.8%) out of 86 patients included in the study experienced at least one GI AE during the follow-up. Participants were classified according to their predominant added fat intake into olive oil (n=41), butter (n=20), margarine (n=14) and sunflower oil (n=11). The mean age of the overall cohort was 72.9 years, and no significant differences in age, sex, smoking status, race or BMI categories were observed across added fat groups. Almost all participants were of Caucasian origin, and therefore variability in ancestry was minimal and could not be meaningfully explored. A higher prevalence of alcohol consumption before starting pirfenidone was reported in the butter and margarine groups compared with olive oil and sunflower oil groups, but adjustment for this variable did not materially affect the associations, and final models were, therefore, only adjusted for age, sex and BMI (Table 1).

When examining the risk of experiencing at least one GI AE, both margarine and sunflower oil groups were associated with a significantly higher risk compared to the olive oil group (OR 3.81, 95% CI 1.06–15.13, p=0.045; and OR 6.97, 95% CI 1.45–51.4, p=0.026, respectively), whereas the butter group showed a non-significant trend (OR 2.84, 95% CI 0.90–9.41, p=0.078). In terms of the number of events, patients consuming butter, margarine or sunflower oil reported a significantly greater incidence of GI AEs than those in the olive oil group, with IRRs of 2.19 (95% CI 1.09–4.43, p=0.028), 3.52 (95% CI 1.80–6.88, p<0.001) and 2.32 (95% CI 1.24–4.35, p=0.008), respectively. Moreover, the risk of nausea and vomiting was also significantly increased among patients consuming butter (OR 5.63, 95% CI 1.37–27.08, p=0.021), margarine (OR 8.07, 95% CI 1.72–44.25, p=0.010) and sunflower oil (OR 11.34, 95% CI 2.13–71.84, p=0.006) compared to olive oil (Table 2). These estimates should be interpreted cautiously given the very limited sample size of the sunflower oil subgroup.

In this post-hoc analysis of the MADIET clinical trial, we found that patients with IPF treated with pirfenidone and preferentially consuming olive oil tended to experience a lower incidence of GI AEs compared with those who used either butter, margarine, or sunflower oil for cooking and seasoning. These results suggest that the quality of added fat may influence GI tolerance and may represent a modifiable factor to improve treatment adherence. These findings should be interpreted with caution given the limited sample size in some subgroups, the exploratory nature of this post-hoc analysis and the close association between fat type and country of origin. Rather than differences between non–olive oil fats, the most consistent finding was the lower risk observed with olive oil. To our knowledge, no previous study has specifically investigated the impact of individual added fat types on pirfenidone-related GI AEs beyond the broader MUFA versus SFA dietary pattern comparison previously reported in the original MADIET trial. However, our results are consistent with existing evidence linking olive oil and the Mediterranean dietary pattern to improved digestive discomfort and lower risk of GI AEs [8]. In pivotal clinical trials of pirfenidone, such as ASCEND and CAPACITY, GI AEs, nausea, dyspepsia, vomiting, and diarrhoea were among the most frequently reported reasons for dose reduction or treatment discontinuation [3,4]. Several biological mechanisms may explain the observed associations. First, olive oil is a rich source of MUFA, particularly oleic acid, and contains bioactive compounds such as (poly)phenols with anti-inflammatory and gastroprotective effects [9,10], whereas saturated and trans fats have been associated with impaired GI function [11,12]. Second, added fat composition may modulate the pharmacokinetics of orally administered drugs. Co-administration of pirfenidone with food is known to reduce peak plasma concentrations without altering overall exposure, thereby improving tolerability [13], with unsaturated fats potentially enhancing this effect. These mechanisms were not directly assessed in our study and should therefore be considered hypothesis-generating. Given the exploratory post-hoc nature of the analyses and the number of statistical comparisons performed, the possibility of type I error and chance findings cannot be excluded. The strengths of this analysis include its novelty, particularly the evaluation of individual added fats beyond the traditional MUFA versus SFA classification, and the systematic collection of dietary and AEs data within a multinational clinical trial framework. Limitations include its observational post-hoc design, which precludes causal inference. Dietary intake was assessed through self-reported FFQ, which may introduce recall bias and potential under- or overreporting of certain foods. Patients who reported consumption of more than one type of fat were assigned to the group corresponding to the fat most frequently consumed, which may not fully capture mixed intake patterns or combined effects of different added fats. Although this approach simplifies complex dietary patterns, it reflects predominant habitual intake based on frequency of consumption, as commonly applied in nutritional epidemiology. The sample size was small, especially in the margarine and sunflower oil groups, limiting the precision of estimates. Moreover, fat type was strongly associated with geographic distribution, which may introduce residual confounding related to country-specific factors such as dietary patterns, healthcare systems, and clinical management. In addition, residual confounding by unmeasured factors such as fibre intake, other dietary components, gut microbiota composition and other healthy/unhealthy lifestyle habits further than smoking history cannot be ruled out. Although effect sizes were relatively large, confidence intervals were wide and overlapping across non–olive oil fats, limiting conclusions regarding clinically meaningful differences among them. Importantly, dietary intake was assessed prior to treatment initiation, reducing the likelihood of reverse causation. Finally, the follow-up period was limited to 16 weeks, and longer-term effects remain to be determined. These findings support the potential role of dietary counselling in patients with IPF receiving pirfenidone, although they should be interpreted cautiously given the exploratory nature of the study and the close association between fat type and country of origin. Further studies, including randomized trials and mechanistic research, are needed to confirm these observations.

This post-hoc analysis of the MADIET clinical trial showed that patients with IPF treated with pirfenidone and consuming diets predominantly based on olive oil were associated with a lower incidence of GI AEs compared to those consuming butter, margarine, or sunflower oil. These findings suggest a possible association between predominant added fat type and GI tolerance to pirfenidone, although residual geographic confounding cannot be excluded. Given the exploratory nature of this analysis and the limited sample size, these results should be interpreted with caution and require confirmation in larger studies.