Survival in patients with pulmonary hypertension associated with fibrotic interstitial lung disease (ILD) is independent of ILD subtype

Extract

Pulmonary hypertension (PH) is common in patients with interstitial lung disease (ILD), observed in 7–15% of those with early idiopathic pulmonary fibrosis (IPF) [1] and in over 70% in those with late-stage disease [2]. PH drives an increase in mortality which appears independent of haemodynamic severity in patients with ILD-PH [3]. Historic ILD registries suggest that survival is dependent upon ILD subtype; for instance, overall survival in non-IPF subtypes, including nonspecific interstitial pneumonia (NSIP), greatly exceeds that of IPF [4].

Shareable abstract

This is the first study to compare well-phenotyped patients with ILD-PH in a single centre. Unlike broader cohorts, once PH is severe enough to warrant RHC, survival appears independent of the underlying ILD subtype when comparing IPF-PH with non-IPF-PH. https://bit.ly/3IimPxW

To the Editor:

Pulmonary hypertension (PH) is common in patients with interstitial lung disease (ILD), observed in 7–15% of those with early idiopathic pulmonary fibrosis (IPF) [1] and in over 70% in those with late-stage disease [2]. PH drives an increase in mortality which appears independent of haemodynamic severity in patients with ILD-PH [3]. Historic ILD registries suggest that survival is dependent upon ILD subtype; for instance, overall survival in non-IPF subtypes, including nonspecific interstitial pneumonia (NSIP), greatly exceeds that of IPF [4]. However, once PH develops, survival is considered to be driven by PH regardless of the underlying ILD subtype. To date, this statement has been based on studies of individual cohorts [5, 6] rather than comparison between ILD-PH subtypes within centres. As such, we studied the survival of patients with fibrotic ILD and right heart catheterisation (RHC)-confirmed PH managed in a tertiary PH centre.

Survival was assessed in consecutive patients with PH associated with IPF, chronic hypersensitivity pneumonitis (cHP), idiopathic NSIP, or unclassifiable ILD referred to the Royal Brompton PH Centre (London, UK) between 2006 and 2023. ILD-PH was defined as mean pulmonary artery pressure >20 mmHg, pulmonary capillary wedge pressure ≤15 mmHg, and pulmonary vascular resistance (PVR) >2 Wood units at baseline diagnostic RHC. Patients with PH predominantly due to left-sided heart disease, thromboembolic disease, combined pulmonary fibrosis-emphysema syndrome (CPFE), and connective tissue disease-related ILD (CTD-ILD) were excluded. Baseline data comprised demographic, haemodynamic (RHC), pulmonary function, 6-min walk distance (6MWD), emPHasis-10 score (collected since 2014), World Health Organization (WHO) functional class, brain natriuretic peptide (BNP), and echocardiographic measurements. Echocardiographic right ventricular (RV) dysfunction was defined by any of the following: tricuspid annular plane systolic excursion <1.7 cm, tissue Doppler S’ wave <9 cm·s⁻¹, or fractional area change <35% [7]. Baseline data were defined as those nearest to the time of ILD-PH diagnosis or within two months afterwards (provided no PH treatment had been commenced).

Continuous variables are presented as mean (±sd) or median (interquartile range). ANOVA or Kruskal–Wallis tests were used to assess differences in baseline characteristics for parametric and non-parametric data, respectively. Categorical variables were analysed using Chi-squared tests. Survival analysis used time from ILD-PH diagnosis to death or lung transplantation (or were censored as alive at last follow-up), with comparison between ILD-PH subtypes using the log-rank test. Univariate and multivariate Cox proportional hazard regression models identified variables associated with mortality. Statistical analyses were performed using R (Version 4.3.3). This study received institutional ethics review board approval (RSID: 195474).

Data on 125 patients were analysed: IPF-PH (n=83, 66.4%), cHP-PH (n=26, 20.8%), NSIP-PH (n=12, 9.6%), and unclassifiable ILD-PH (n=4, 3.2%). Around two-thirds (63.2%) of patients were male, with a median age of 68.4 (60.7–74.8) years at ILD-PH diagnosis, and a median follow-up period of 1.1 (0.5–2.2) years. Over half of the cohort (n=78, 62.4%) had severe PH as defined by a PVR >5 Wood units.

Patients with IPF-PH were mostly male (77.1%) while those with cHP-PH and NSIP-PH were mostly female (65.4% and 66.7%, respectively, p<0.0001), with equal sex distribution among other subtypes. Patients were in WHO functional class II in 4 (3.2%), WHO III in 102 (82%), and WHO IV in 18 (14%). Mean 6MWD was 224±120 m (n=80) and mean emPHasis-10 score was 31±11 (out of a total of 50; n=43). Mean pulmonary artery pressure was 34 (29–43) mmHg, PVR was 6.1 (4.5–9.1) Wood units, pulmonary capillary wedge pressure was 8 (5.8–12) mmHg, forced vital capacity (FVC) was 57.9 (46.3–73.2) % predicted, total lung capacity was 54.2 (48.0–61.6) % predicted, transfer factor of the lung for carbon monoxide (T_LCO) was 23.1 (18.7–29.0) % predicted, and transfer coefficient of the lung for carbon monoxide (K_CO) was 51.4±17.1 % predicted. Median BNP was 142 (45.5–437.5) ng·L⁻¹ (n=112). RV dysfunction at echocardiography was present in 65 (50.4%) of patients. Of note, when a comparison was made between IPF-PH and non-IPF-PH groups, there was no statistical difference in their baseline clinical characteristics in terms of markers of PH or ILD severity.

Overall, 1-, 3-, and 5-year transplant-free survival were 51.8%, 18.5% and 6.2%, respectively, with a median time to death or lung transplant of 1.1 (0.5–2.2) years. No difference in survival was detected when comparing IPF and non-IPF (log-rank p=0.2), as illustrated in figure 1. Comparison of survival between IPF-PH and non-IPF-PH groups was also non-significant using the multivariable Cox regression analysis. A total of nine patients underwent lung transplantation. Two of these had IPF-PH while the remainder had non-IPF-PH.

FIGURE 1.

Kaplan–Meier curves comparing survival of idiopathic pulmonary fibrosis pulmonary hypertension (IPF-PH) (blue line) versus non-IPF-PH subtypes (red line) (log-rank p=0.2). Vertical lines represent censored patients. ILDs: interstitial lung diseases.

Univariate analysis identified five clinical parameters associated with a risk of death or transplant in this cohort: FVC% (Hazard ratio (HR) 0.99, 95% CI 0.98–1.00, p=0.04); T_LCO% (HR 0.95 (0.93–0.98), p<0.001); 6MWD below the median of <230 m (HR 1.83, 95% CI 1.15–2.93, p=0.01); BNP>100 ng·L⁻¹ (HR 1.67, 95% CI 1.11–2.52, p=0.01); and echocardiographic RV dysfunction (HR 1.81, 95% CI 1.20–2.73, p<0.01). On stepwise backward multivariate analysis (excluding 6MWD due to missingness), T_LCO% and BNP>100 ng·L⁻¹ remained predictors of death or transplantation. Using thresholds, T_LCO <25% increased mortality by 139% (HR 2.39, 95% CI 1.53–3.75, p<0.001) and FVC<60% increased mortality by 48% (HR 1.48, 95% CI 1.00–2.19, p<0.05). Following PH diagnosis, the use of oral PH therapies in 67 (53.6%) patients was associated with improved survival (HR 0.50, 95%CI: 0.34–0.74, p<0.001) as compared to never-treated patients, in keeping with a previous study in which the treated group had better outcomes despite more severe PH [7].

This report is the first to compare outcomes between multiple subtypes of ILD-PH within a single referral centre and demonstrates that, upon onset of PH sufficiently severe to warrant right heart catheterisation, survival is independent of ILD subtype when comparing IPF-PH with non-IPF-PH groups. This finding diverges from the well-established survival difference between IPF and non-IPF (notably NSIP) patients in cohorts that did not undergo RHC and were, therefore, agnostic to the presence or absence of PH [4]. The interval between ILD diagnosis and PH diagnosis is likely to differ between IPF and non-IPF groups, with PH arising earlier in IPF. This should be the focus of future prospective studies, as PH in patients with IPF may progress more rapidly than in those with non-IPF-PH. Indeed, a subset of IPF patients awaiting transplant and undergoing serial RHC have recently been described to show rapid deterioration in PVR [8]. The notion that the underlying ILD subtype is not the driver of mortality in ILD-PH was previously highlighted in the COMPERA registry, where no difference in survival was observed when comparing patients with IPF-PH and NSIP-PH [9]. However, determining the exact cause of death in patients with advanced IPF-PH is challenging since most succumb to cardiorespiratory failure, typically involving a combination of respiratory decompensation and right ventricular failure. Both of these components were highlighted in this study, with factors reflecting both the lung and pulmonary vascular compartments as risk factors for death. In previously published studies of patients with ILD-PH, right heart failure accounted for 29% of all fatalities, while the leading cause of death was respiratory failure (49%) [9].

Survival in this ILD-PH cohort was associated with low gas transfer, which is likely to reflect both the degree of lung function impairment as well as pulmonary vascular disease. Notably, patients with concomitant emphysema (CPFE))–in whom gas transfer impairment is profound and mortality may be higher–were excluded from this analysis. BNP was predictive of mortality in this cohort, likelyreflecting RV dysfunction, a recognised cause of death in ILD-PH and other PH subgroups [9, 10]. Monitoring BNP as a marker of RV dysfunction is this setting is clinically relevant beyond its prognostic value since it may also aid in screening for the onset of PH in patients with ILD. Emerging evidence suggests that early initiation of PH therapies, prior to RV dysfunction, may improve clinical outcomes in ILD-PH [7, 11], although this requires further validation.

In conclusion, this is the first study to compare survival between ILD-PH subtypes within a single centre. We demonstrate that once PH is sufficiently advanced to warrant RHC, survival in ILD-PH due to IPF, cHP, idiopathic NSIP, and unclassifiable ILD is independent of the underlying ILD subtype. Of note, this analysis excluded patients with CTD-ILD and CPFE. We demonstrate on multivariate analysis that mortality is associated with gas transfer impairment but not with other lung function parameters. We also highlight the importance of BNP and RV dysfunction in this high-risk patient group and encourage monitoring with serial measurements of these parameters in the routine care of ILD-PH patients–and arguably in all patients with ILD–to further inform prognosis. Future studies in larger cohorts will be important to clarify the role of PH therapies and RV functional stratification on survival in this high-risk patient population.