This was a non-sponsored, retrospective, observational study, conducted in 4 hospitals in Argentina that have a specialized ILD team. We included patients with a diagnosis of IPF according to the ATS/ERS/JRS/ALAT 2011 consensus, who received pirfenidone in the period between June 2013 and September 2016.1 Patients who did not complete at least 3 months of treatment were excluded. Clinical and demographic variables included age, sex, smoking habit, the presence of symptoms of gastroesophageal reflux disease (GERD), lung pattern on high resolution computed tomography (HRCT), immunological laboratory results, need for lung biopsy for diagnosis, and development of acute disease exacerbation. HRCT pattern was categorized as typical usual interstitial pneumonia (UIP), possible UIP, and pattern inconsistent with NIU, following the recommendations of the ATS/ERS/JRS/ALAT 2011 consensus. The presence of associated emphysema was determined, and considered significant when extension was greater than 10% on HRCT. All patients were studied for rheumatoid factor (RF) (nephelometry, cut-off point 35U/ml) and antinuclear antibodies (ANA) (indirect immunofluorescence, cut-off point 1/80). Acute exacerbation was defined according to criteria proposed recently by an international study group.10 The following lung function and exercise tolerance parameters were studied: % of predicted FVC; meters walked on the 6-minute walk test; and predicted DLCO adjusted for hemoglobin. All of these variables were evaluated at the time of diagnosis, on the date of starting pirfenidone, and during treatment, according to the criteria of each medical group. Tests were conducted according to the proposals of the reference societies.11 The percentage of patients receiving treatment with proton pump inhibitors and/or N-acetylcysteine was calculated. The percentage of patients receiving systemic corticosteroids, azathioprine, and N-acetylcysteine for at least 3 months was also studied.12 To analyze the period between onset of disease and diagnosis, the number of days between the onset of symptoms (dyspnea and/or cough) and date of diagnosis in a reference center was calculated. We also analyzed the time between diagnosis and the date of starting pirfenidone. Tolerance and safety of pirfenidone were evaluated from data reported in the clinical records, such as adverse effects (AE) in the treatment period between the date of starting the drug and the date of the last visit, death or withdrawal. A significant rise in liver enzymes was defined as more than 3 times the normal value in patients who reported symptoms, or more than 5 times in those who reported no symptoms. The different clinical, functional, tomographic, and progress variables were compared between patients with and without reported AEs. To study the effectiveness of pirfenidone, the change in FVC in the period prior to receiving the drug (before pirfenidone) was compared with the change in the same variable during the period in which the patients were exposed to the drug (after starting pirfenidone). For this analysis, only patients with FVC results in both periods, with a difference of no more than 6 months between the determinations, were included. Finally, we compared the change in FVC during treatment with pirfenidone among subgroups of patients with FVC≥75% and those with FVC<75% at the time of starting treatment.

Categorical variables were described by frequency, and continuous variables by mean and standard deviation (SD) or median and interquartile range (IQR), depending on whether distribution was normal or not, respectively. For comparative analysis, the Chi square test, Fisher's exact test, Student's test or Wilcoxon test were used as appropriate, depending on the type of variable and its distribution. A P-value of <.05 was considered statistically significant. Results were communicated according to the indications of the STROBE initiative.13

Cohort characteristics are shown in Table 1. In total, 50 patients were included. The distribution of patients by hospitals was: 33 patients (66%) from the Hospital de Rehabilitación Respiratoria María Ferrer in the city of Buenos Aires, 8 patients (16%) from the Hospital Privado de Comunidad in Mar del Plata, 5 patients (10%) from the Hospital de Clínicas José de San Martin in the city of Buenos Aires, and 4 patients (8%) from the Laboratorio de Función Pulmonar de Alta Complejidad in the city of Bahía Blanca. Most patients were men (76%), with a mean (SD) age of 67.8 (8.36) years. In 12 cases (24%), surgical lung biopsy was required for diagnosis, of which 8 showed a pattern of possible UIP on CT, one had a pattern inconsistent with UIP, and 3 had a pattern of typical UIP. Thirteen patients had emphysema on HRCT, all of whom were smokers. Of the 37 remaining patients who did not show emphysema, only 12 (32.4%) had never smoked. In all cases, the diagnosis of IPF was reached after discussion among the members of the multidisciplinary committee, after carefully ruling out hidden connective tissue diseases and/or chronic hypersensitivity.14 Only 8 patients (16%) were ANA positive, and all had low titers (less than or equal to 1/160). Fourteen patients (28%) were included in the list of possible candidates for lung transplantation. Lung function parameters were recorded in 49 patients. Mean (SD) FVC at baseline was 69.7% (14.22). DLCO was measured in 37 patients, with a mean (SD) of 44.9% (77.76). PO2 was obtained in 29 patients, with a mean (SD) of 71.6mmHg (25.03). Forty-two patients performed the 6-minute walk test, with a mean (SD) of 403.69 (98.33) meters walked. Thirteen patients (26%) died during the study period, all for causes related to IPF. Five of them (38.46%) died after an exacerbation and 8 (61.53%) due to disease progression. Characteristics of patients who died were compared with those of patients who survived. Among those who died, 5 (38.46%) had an exacerbation, while none of the patients who survived presented an exacerbation. This difference was statistically significant (P=.0006). Furthermore, 9 patients (69.23%) in the group of patients who died discontinued treatment temporarily vs 9 (24.32%) in the group of patients who survived, a difference that was also statistically significant (P=.005) (Table 2). Finally, only 1 patient received single-lung transplant due to IPF progression.

Mean (SD) number of days between onset of symptoms and IPF diagnosis was 677.7 (563.4), while the mean (SD) number of days between diagnosis and starting pirfenidone treatment was 283.40 (201.7).

The mean (SD) number of days of pirfenidone treatment was 645.68 (428.19), with a mean (SD) daily dose of 2064.56 (301.49)mg. Twenty-six patients (52%) received treatment with pirfenidone for longer than 1 year, and 14 (28%) for more than 2 years. Nineteen AEs were reported in a total of 15 patients (30%). Of these, 11 had only 1 adverse event and 4 patients had 2 adverse events (Table 3). Analysis of the different variables that could influence the development of AEs did not reveal statistically significant differences between both groups (Table 4). A total of 18 patients (36%) discontinued treatment: 17 discontinued temporarily for less than 1 month before resuming the full dose; and 1 discontinued definitively. Reasons for discontinuation are listed in Fig. 1. Of these 18 patients, only 3 discontinued treatment due to AEs. Two patients interrupted treatment for less than 1 month, 1 due to persistent nausea and asthenia, and the other due to severe diarrhea. The third patient discontinued definitively due to a phototoxic skin reaction involving a severe burn, requiring admission. Subsequent progress was good. This patient had a Fitzpatrick II phototype (burns easily on sun exposure and develops a light tan), predisposing him to such phototoxicity, and he failed to take the indicated protective measures. Finally, none of the patients developed hepatotoxicity.

Fig. 1.

Reasons for pirfenidone discontinuation.

(0.07MB).

We included 28 patients in this analysis. No significant differences were found between the periods before and after starting pirfenidone. We obtained a median (IQR) functional follow-up of 173.5 days in the period before pirfenidone (127.5–265) and 172 days (144.5–270.5) in the period after starting pirfenidone (P=.247). In the comparative analysis between both periods (Fig. 2), we observed a decline in mean (SD) FVC of 4.03% (7.63) during the pre-pirfenidone period, and 2.64% (7.1) during the period after starting pirfenidone; this difference was not statistically significant (P=.534). Nor was this difference statistically significant in the subgroup of patients with concomitant emphysema. To complete this analysis, we classified the patients according to severity and compared the decline in FVC during pirfenidone treatment (Fig. 3). We found 11 patients with FVC≥75% (mild) and 17 patients with FVC<75% (moderate and severe). We found a difference in FVC behavior with a trend toward statistical significance in favor of the subgroup with FVC≥75%. This group showed a change in median (IQR) FVC of 1% (−6 to 10), compared with a median (IQR) of −5 (−6 to −1) in the <75% subgroup (P=.09). We found no differences in the proportion of patients with concomitant emphysema between both subgroups, with 3 (27.2%) in the FVC≥75% group (17.65%) and 3 in the FVC<75% group (P=.65).

Fig. 2.

Change in FVC during the follow-up period Comparison between periods before and after starting pirfenidone (P=.534).

(0.05MB).

Fig. 3.

Change in FVC during the follow-up period Comparison between groups with baseline FVC 75% or more or less than 75% (P=.09).

(0.08MB).