Staging of idiopathic pulmonary fibrosis: past, present and future

Introduction

The median survival of patients with idiopathic pulmonary fibrosis (IPF) ranges from 2.5 years to 3.5 years [1]. However, the clinical course of disease can vary considerably, from slow progression over many years to acute exacerbation, rapid loss of lung function and early death [2–4]. A number of studies have shown that selected features commonly observed in clinical practice are associated with an increased risk of mortality. These include clinical history and physical examination, as well as radiographic, physiological, pathological and biomarker findings [3]. However, there is no widely accepted, standardised method to combine such predictors in order to define prognosis or stage of disease.

The importance of staging in disease management

In general, staging systems for patients with a particular disease aim to define a classification schema that is practical, informs prognosis and impacts choice of therapeutic modality [5]. Staging systems are widely used in other diseases for guiding management decisions, with examples including those used in lung cancer [6], HIV/AIDS [7] and chronic obstructive pulmonary disease (COPD) [8]. In COPD, for example, the Global Initiative for Chronic Obstructive Lung Disease (GOLD) staging system categorises severity of disease into four stages based on forced expiratory volume in 1 s. Mortality risk rises with increasing stage, and treatments are added as patients progress from one stage to the next [8, 9].

Staging IPF is widely acknowledged to be an important goal as it: 1) would allow clinicians to use standardised nomenclature in daily practice; and 2) would guide treatment options and facilitate patient counselling. For IPF, several baseline features of the disease may be useful in staging including age, level of dyspnoea, forced vital capacity (FVC), diffusing capacity of the lung for carbon monoxide (D_LCO), desaturation during the 6-min walk test, extent of honeycombing on high-resolution computed tomography (HRCT), or presence of pulmonary hypertension or emphysema [1]. Longitudinal factors such as increase in level of dyspnoea, decrease in FVC over time, decrease in D_LCO or worsening of fibrosis on HRCT have also been recommended [1]. Importantly, for any staging system to be useful in daily practice, it must be simple and reproducible.

Traditional approaches to staging IPF: mild, moderate and severe

Traditionally, terms such as “mild”, “moderate”, “severe”, “early” and “advanced” have been used to loosely stage IPF. These stages have been primarily based on pulmonary function test results. Several pulmonary function values have been associated with prognosis and are believed to reflect disease severity and/or progression, most reliably FVC and D_LCO [10–13]. While there is no standardised definition for mild, moderate and severe disease, clinical trials have generally agreed on a FVC threshold of 50–55% predicted and a D_LCO threshold of 35–40% pred to separate mild-to-moderate patients from those with severe disease [14–20]. This approach to staging provides physicians with a general framework to monitor disease progression and advise patients. Certain stages in the current model are used to guide treatment; for example, in several countries pirfenidone is indicated for patients with mild-to-moderate IPF, while lung transplantation is considered in severe IPF.

Limitations of the traditional approach to IPF staging

While the traditional approach to IPF staging has been useful, it is arbitrary and is not based on epidemiological or biological data. It remains unclear if these stages are truly relevant to the management of IPF. Critically, these traditional stages are not known to reflect distinct biological or clinical phenotypes and the true therapeutic and prognostic relevance of these stages remains undetermined.

IPF does not progress in a linear pattern and using baseline physiological parameters, such as FVC, alone probably oversimplifies the staging process. As mentioned previously, a number of studies have shown that selected features of IPF, which are commonly observed in clinical practice, are associated with an increased risk of mortality [3]. Incorporation of these factors into a multidimensional staging model may enhance the validity and clinical applicability of staging in IPF, but requires validation.

Novel IPF staging systems

Several groups have provided multivariate models that attempt to provide a basis for a more sophisticated approach to staging of IPF. Using a large clinical trial population of patients with IPF, du Bois et al. [21] developed a scoring system comprised of four predictors: age, recent respiratory hospitalisation, baseline FVC and 24-week change in FVC. As shown in table 1, the sum of individual scores for the four parameters are used to obtain a composite score that can accurately predict the 1-year risk of death (table 2).

More recently, Ley et al. [22] developed a multidimensional risk prediction model and staging system using data from three large, geographically distinct cohorts of IPF patients in California (USA), Minnesota (USA) and Northern Italy. This GAP model consists of four baseline variables: gender (G), age (A) and two lung physiology variables (P) (FVC and D_LCO). A regression model using continuous predictors (GAP calculator) and a simplified scoring system (GAP index) (table 3) were developed (www.acponline.org/journals/annals/extras/gap). Three stages, I, II and III, were identified based on the GAP index with 1-year mortalities of 6%, 16% and 39%, respectively (table 4). It was proposed that the GAP index and staging system be used as a quick and simple screening method for estimating risk in patients with IPF; the GAP calculator can then be used to estimate the individual risk for patients in whom a more precise estimate may affect clinical decisions.

Future directions

As highlighted previously, it will also be important in future staging systems to integrate baseline information with longitudinal parameters such as changes in dyspnoea and categorical decline in FVC of ≥10% pred, both of which are strong dynamic predictors [23]. Visual scoring of the extent of fibrosis on HRCT at baseline and over time may also be valuable in staging disease [24]. Moving forward, probably the greatest need in staging is the integration of biological data and the inclusion of this data offers the greatest potential improvements. Aggressive development of molecular and other biological biomarkers is just beginning, and additional work in this area is greatly needed. Markers being investigated include serum markers, such as KL-6, surfactant protein (SP)-A/D, CCL18, brain natriuretic peptide and matrix metalloproteinase-7, and markers from bronchoalveolar lavage, such as SP-A and neutrophilia [25, 26].

Ultimately, the development of improved staging systems in IPF requires balancing of two key priorities: 1) identifying the most accurate predictors of IPF prognosis; and 2) determining a practical and clinically useful method with which to integrate these parameters.

Conclusion

Staging systems offer important benefits to medical practice and clinical trial design. Traditional approaches to staging in IPF using mild, moderate or severe classifiers based on baseline pulmonary physiology leave substantial room for improvement and should be replaced by more integrative and evidence-based methods. Improved staging of IPF would allow clinicians to use standardised nomenclature for medical practice, enable better definition of clinical trial populations and end-points, guide treatment options and enhance patient counselling. We hope that the IPF community can develop a standard evidence-based approach to staging that can then be improved as new knowledge of disease biology, IPF phenotypes and behaviour emerges.

Undoubtedly, our increasing understanding of IPF pathophysiology will result in more sophisticated IPF staging systems, which will better inform the management of this devastating condition.