The clinical course of idiopathic pulmonary fibrosis and its association to quality of life over time: longitudinal data from the INSIGHTS-IPF registry

Idiopathic pulmonary fibrosis (IPF) is a severe disease characterised by debilitating symptoms such as cough, fatigue, increasing and immobilising dyspnoea and is associated with a limited survival [1, 2]. IPF symptoms and outcomes are also strongly influenced by co-morbidities [3, 4]. The psychosocial impact of IPF is substantial for both patients and their caregivers [1, 5]. Recent reports have shown that in IPF patients health-related quality of life (QoL) is associated with clinical symptoms, duration of disease, physical activity, comorbidities, and disease severity [6‐10] and may be prognostic [11].

While current therapies aim to improve the disease course and to prolong life expectancy, less well understood is how best to improve an individual’s health-related quality of life and psychosocial needs [12]. To address this, patient-reported outcomes incorporating determination of health-related quality of life (QoL) assessments are increasingly used as important outcome measures, in clinical practice and in clinical trials [6]. However, data on longitudinal determinations of QoL in IPF patients, especially in the real-life setting are sparse. Previously, we have reported on baseline characteristics and QoL status in patients enrolled in a large nationwide observational IPF registry (INSIGHTS-IPF) [2, 7]. In the present analyses, we report longitudinal data (with up to 36 months follow-up) for a range of QoL measures and examine changes in QoL outcomes and their relationship to clinical parameters.

According to the present data, collected from a representative setting of centres with expertise in the management of patients with IPF, a close relationship between QoL and clinical outcomes were found; including clinically meaningful changes in lung function, comorbidities, disease duration and clinical course of IPF (including hospitalisation, acute exacerbations and mortality). For these analyses, we used a range of QoL assessments which, although not originally designed for IPF, have data to support their validity in this patient population, as shown in our previous report and by others [6, 7, 9, 15‐18]. The optimal tool for QoL assessment in IPF patients remains to be determined, with an adaptation of the SGRQ specifically for use in IPF not yet fully validated or in widespread use [19], while other more focused IPF QoL-measures are also in use or being developed for use (e.g. the Kings’ Brief Interstitial Lung Disease health status [K-BILD] and the ‘A Tool to Assess Quality of Life in Idiopathic Pulmonary Fibrosis [ATAQ-IPFcA] questionnaires) [9, 20, 21]. It is likely that, while subtle differences may exist, most current tools assess QoL with similar sensitivity and with high or very high correlation between most questionnaires [7, 9].

To our knowledge, the only previous report on longitudinal assessments of QoL in IPF patients in clinical practice comes from the Australian IPF registry in a somewhat smaller dataset [6]. In that study, a strong relationship between decline in FVC % predicted and decline in QoL was shown, with QoL impairment mainly due to dyspnoea (measured using the SGRQ and UCSD-SOBQ scores) although cough and depression (assessed by a VAS scale and the Hospital Anxiety and Depression Scale (HADS)) also contributed [6]. While in that study, no association between baseline QoL measures and mortality was found, a more recent report from the Australian registry indicates that in those patients with mild physiological impairment (FVC ≥ 80%), baseline SGRQ and UCSD-SOBQ scores are predictive of mortality, with HRs of 1.11 (95% CI: 1.06 to 1.15) and 1.23 (95% CI: 1.13 to 1.33) respectively (p < 0.001 for both) [22]. In contrast to the Australian Registry, we found a significant relationship between lower QoL at last available follow-up but no significant association (on multivariable analysis) between mortality and change in QoL for the total IPF population. Number of comorbidities as well as hospitalisations were driving this QoL decline. Notably, significant changes in both FVC and DLco % predicted values closely reflected changes in QoL indicating that stabilising lung function may also halt decline of QoL in IPF.

The overall pattern we observed across the cohort was a longitudinal decline in QoL. How then can we seek to improve upon or at least maintain patient QoL in clinical practice? At present, only sparse data on the long-term effects of available therapies on QoL outcomes exist. While disease stabilisation by antifibrotic therapies may offer some potential in maintaining QoL (or retarding decline), available data to support this goal remain limited. In the pivotal phase 3 trials evaluating nintedanib and pirfenidone (each versus placebo), no significant differences in changes in SGRQ scores (from baseline to 52 weeks) were seen between the treatment and placebo arms [23, 24] a finding which could be due to side effects or to short observation times. Furthermore, this has also to be interpreted in light of the here reported differences in associations of QoL to lung functional decline which was statistically significant comparing stable to significantly progressing patients, i.e. with a decrease of FVC > 10%. However, the difference to moderately progressive patients, i.e. a decline between 0 and 10% was only by trend. In the INPULSIS trials for example adjusted absolute mean change from baseline in FVC% was also within this range with − 2.8 for nintedanib and − 6.0 for placebo. Still, a subsequent pooled analysis of the TOMORROW and INPULSIS trials comparing nintedanib vs. placebo has shown that the reduction in annual FVC % predicted decline seen in the nintedanib arms was accompanied by lower increases in SGRQ compared to placebo (with a mean difference in adjusted mean change from baseline of − 2.05 [95% CI: –3.59 to − 0.50]; p < 0.0095) [25]. In the step-IPF study, sildenafil was reported to positively impact upon QoL in IPF patients with severe impairment of gas exchange (i.e. DLco ≤35%) [26]. In a recent randomised controlled trial using nintedanib and sildenafil in combination (INSTAGE study) a positive, while non-significant trend on QoL was observed, but conclusive data is still lacking [27]. On balance, it is likely that optimising pharmacotherapy in IPF patients (and reducing decline in lung function) may also assist in improving QoL in IPF patients. As groups in cohort studies like ours, e.g. with and without antifibrotic therapies, inevitably differ in their clinical characteristics comparisons of therapy effects on QoL in registries may be subject to bias by indication. We therefore refrained from performing comparative analyses with regards to therapy, which however should be assessed in future research.

Non-pharmacotherapy treatments are also important. As physical activity is a strong predictor of IPF disease course, methods to increase physical activity e.g., pulmonary rehabilitation are important in prolonging a more positive life-span, in particular as data suggests that pulmonary rehabilitation may increase QoL [28‐30]. Allied to this, greater attention to assessment and treatment of comorbid conditions may improve or at least alleviate longitudinal decline in QoL, although as of yet this remains unproven [4]. For several patients, early palliative care may be an appropriate approach [5, 12].

Our study has a number of limitations (above that of using non-IPF specific QoL measures as discussed above). Typically for data collection under clinical practice conditions, sites showed considerable variance in terms of number of documented visits and visit time-points, and in terms of completeness of data (each depending on their routine clinical schedule) including QoL reports. Furthermore, a potential bias cannot be excluded, such as underreporting or underdiagnosing of subacute deterioration and any influence of social factors (e.g., family situation) and other aspects external to the disease process. A high number of patients enrolled were excluded from the present analyses due to incomplete data of longitudinal QoL assessment. As some differences had been noted between the group of in- and excluded patients, e.g. a more severe disease and a higher mean number of comorbidities, a bias of our results toward patients with a more severe disease course cannot be excluded. In addition, not every patient completed each of the QoL questionnaires at least yearly as stipulated. However, the relatively large numbers of included patients and completed QoL measures in connection with a long follow-up period do provide a substantially large sample and exposure time from which we found significant differences in a number of outcomes, and association of QoL changes with clinical status. A final limitation is that we have not included potential effects of treatment (including use of anti-fibrotic agents) in the present analyses and cannot report on benefits of such treatment.

In conclusion, we found strong associations of changes of QoL to important outcomes in IPF; principally decline in lung function, hospitalisation for acute exacerbations and mortality. Pharmacological and non-pharmacological therapies used to reduce these clinically important events should also aim to maintain or even improve QoL as assessed by these equally important patient-reported outcomes.