Introduction/objectives Idiopathic Pulmonary Fibrosis is progressive with poor outcomes. Evidence from the CAPACITY and ASCEND trials suggests Pirfenidone slows disease progression (Noble et al, 2011; King et al, 2014). We reviewed all patients with more than a year since initiation to assess the proportion that discontinued Pirfenidone for adverse drug reactions (ADRs) or due to unsuccessful treatment (>10%/year%FVC decline or died on therapy), to assess temporal patterns of these stoppages and assess any demographic or disease extent predictors of success (age, sex, BMI, smoking history, %FVC prediction and FEV1).

Methods 155 patients have been referred to our consultant pharmacist for Pirfenidone initiation since August 2013, of which 65 started and have more than 1 year of data (i.e., we excluded those starting post July 2015). We reviewed hospital databases and medical notes with subsequent data analysis using appropriate parametric statistical methods.

Results 42/65 (64.6%) stopped therapy overall divided between ADRs (24/42, 57.1%), FVC decline >10%/year (11/42, 26.2%), dying on treatment (5/42, 11.9%) and unclear (2/42, 4.8%). Table 1 demonstrates association between demographics and those that discontinued, suggesting there is one subgroup associated with treatment continuation (patient previously smoked P ≤ 0.001). The temporal pattern of stopping treatment for ADRs vs those failing for clinical reasons (death on treatment or FVC decline) demonstrated a median cessation time of 40 days vs 226 respectively.

Conclusions Our data suggests two important findings. Firstly, patients’ chances of continuing therapy is intriguingly affected positively by smoking history, causation for which remains unclear. Sex, age, BMI or disease extent does not seem to be associated with outcome success. Secondly, those that discontinued for ADRs did so much earlier than those that discontinued due to FVC decline or dying on treatment.

Potential implications for when prescribing Pirfenidone are that we cannot predict those at risk of discontinuation from their demographics. However, a smoking history intriguingly seems to be positively associated with continuation. Additionally, most patients affected by ADRs ceased treatment early, suggesting that if side effects are not noted post early follow up, they are unlikely to occur later. This information improves understanding of prescribing practice for Pirfenidone.