Paul W Noble Department of Medicine, Division of Pulmonary, Allergy and Critical Care, Duke University Medical Center, Durham, North Carolina 27710, USA , Eric B Meltzer Department of Medicine, Division of Pulmonary, Allergy and Critical Care, Duke University Medical Center, Durham, North Carolina 27710, USA.

Prognostic indicators
Early studies looking at the prognosis of IPF identified older age, male sex, significant dyspnea, severe physiologic abnormalities, advanced fibrosis and a poor response to therapy as factors predicting shortened survival [83]. Early studies were limited by the use of retrospective study designs.

Recently, systematic studies have evaluated specific features of IPF that are predictive of mortality. Features of the surgical biopsy specimen have been evaluated and it was found that neither the degree of cellularity nor degree of fibrosis could predict survival ^[84]. However, the presence of “young” connective tissue, characterized by multiple fibroblast foci, was found to correlate with shortened survival. Other investigators have confirmed this link between fibroblast foci and mortality ^[85]. It was also found that the extent of fibroblast foci can predict physiologic functions such as vital capacity and diffusion capacity.

Three separate groups of investigators observed a relationship between physiologic measures and survival amongst well-defined cohorts of IPF patients ^[3-5]. Of particular interest, one study reported that physiologic measures were more accurate than histopathology (NSIP vs. UIP) in predicting mortality ^[5]. Twelve-month trends in diffusion capacity were shown to predict survival. In this study, physiologic measures were taken at baseline and at a twelve month follow-up visit. The patients were grouped into two categories, demonstrating significant decline (more than 15% of baseline) versus demonstration of stability or improvement. Mortality was shown to be substantially higher in the group demonstrating decline.

Other investigators have examined IPF patients and identified interval decline of forced vital capacity as a characteristic that is predictive of survival. A 10% decline of forced vital capacity, at either six or twelve months, had poor prognostic implications and was more accurate than predictions based upon baseline physiologic parameters alone ^[3].

The ability of HRCT to predict the outcome of IPF was also demonstrated. When biopsyproven IPF patients were followed for three years by HRCT, it was found that radiographic honeycombing predicted the worst survival [86]. In addition, when radiographic fibrosis and histopathologic fibrosis were assigned scores, they were found to be equivalent with respect to predicting death or clinical worsening.

Another interesting study examined the relationship between the “confident” IPF pattern and survival. A cohort of patients with biopsy-proven IPF were analyzed by HRCT and divided into groups based upon radiographic/pathologic concordance. UIP pathology was shown to confer a worse prognosis when seen in combination with the “confident” IPF pattern. It was found that the an “indeterminate” pattern of HRCT conferred better prognosis despite the presence of UIP on biopsy ^[87].

Pharmacotherapy
In the past, unremitting inflammation was thought to cause progressive pulmonary fibrosis. Therefore treatment regimens were designed to suppress the immune system with the goal of halting subsequent fibroproliferation. However, large randomized and placebo-controlled trials were never performed to assess the efficacy of this strategy. The only evidence in support of immunosuppressive therapy for IPF is a handful of small studies. Nonetheless, the ATS/ERS consensus statement recommends the use of corticosteroids combined with a cytotoxic agent for carefully selected IPF patients ^[1]. The consensus statement recommends prednisone (starting at 0.5 mg/kg and tapered to a maintenance level of 0.125 mg/kg), combined with either azathioprine or cyclophosphamide (the dose targeted to 2-3 mg/kg). Combination therapy is suggested for a period of at least six months with clinical and physiological response used to guide further management.

The best evidence in support of the prednisone/azathioprine regimen comes from a prospective, randomized, double-blind study of only 27 IPF patients ^[88]. This study examined survival and lung function over a period of several years. While no statistically significant difference was measured using unadjusted data at the end of the first year, a marginal advantage to the prednisone/azathioprine regimen was demonstrated by examining age-adjusted survival curves. This study was performed in the era prior to the initial description of NSIP, which has a better prognosis than IPF. It is most probable that some of the 27 cases included in this study were, actually, unrecognized cases of NSIP. Further inspection of the survival curves from this study reveal that the difference in survival occurs after the fourth year of treatment. Since most IPF patients die within the first three years following diagnosis, such a late treatment benefit would be useless.

A single randomized study of 43 IPF patients compared prednisolone alone to prednisolone/cyclophosphamide ^[89]. This study followed lung function, radiographic images, dyspnea scores and survival. While the study could not demonstrate a survival benefit from combination therapy, an analysis of time to treatment failure was favorable toward the prednisone/cyclophosphamide group. Because this study was also performed in the era before NSIP, its results are confounded by possible inclusion of cases other than IPF. Two recent studies re-evaluated the utility of prednisone/cyclophosphamide in the treatment of IPF but both studies failed to support its use ^{[90, 91]}. The first study followed a series of 19 IPF patients, treated with cyclophosphamide following a corticosteroid taper. This study contained no control group and only one patient was shown to stabilize while undergoing cytotoxic therapy. The study concluded that cyclophosphamide conferred no benefit in the treatment of IPF. The other study utilized a retrospective design to evaluate survival time amongst 82 IPF patients receiving prednisone/cyclophosphamide. The IPF patients were compared to an untreated age-matched, lung-function matched cohort that served as the control group. No survival benefit was observed. Overall, given its considerable toxicity and lack of support for efficacy, it seems unreasonable to prescribe prednisone/cyclophosphamide for the treatment of IPF.

Another potential therapy for IPF is N-acetylcysteine (NAC) which is a molecular precursor to the naturally occurring antioxidant glutathione. Glutathione is known to be depleted in the lungs of patients with IPF ^[92]. Theoretically, oral NAC should replete glutathione stores and restore natural oxidant/anti-oxidant balance to prevent oxidative injury that precedes fibroproliferation ^[93]. A small prospective, but uncontrolled, study following 18 IPF patients receiving NAC demonstrated restoration of glutathione levels and improvement of lung function measures ^[94]. This pilot study prompted a large, randomized, double-blinded study of 182 IPF patients to compare the efficacy of a regimen that included prednisone, azathioprine and NAC to a regimen of only prednisone and azathioprine ^[23]. A statistically significant difference in lung function was found to favor treatment with prednisone/azathioprine/NAC. However, the study is flawed because of a high rate of drop-out among the study subjects. Drop-out was addressed in statistical analysis by deriving and using imputed data for missing data points. Both groups in the study experienced decline of lung function over the 12-month study period; the difference was that one group declined more. The regimen of prednisone/azathioprine/NAC was shown to be possibly superior to prednisone/azathioprine alone. However, the efficacy of prednisone/azathioprine/NAC has not been compared to placebo and has never been shown to improve or stabilize patients with IPF.

Recently, a panel of experts were asked to rate the evidence for the various treatment options for IPF ^[95]. The panel concluded that the most appropriate pathway was to enroll eligible patients in clinical trials or refer for lung transplantation as indicated. Patients without access to clinical trials or lung transplantation could be offered other therapy but the sole use of corticosteroids was deemed inappropriate. The use of corticosteroids in conjunction with azathioprine was deemed acceptable. Given the evidence for prednisone/azathioprine/NAC, this regimen could be considered with little risk attributable to NAC.

Several clinical trials are presently assessing the utility of novel agents in the treatment of IPF. One promising drug is pirfenidone which has already been tested in phase I and phase II studies in the United States and Japan. A study examining the use of pirfenidone enrolled 105 IPF patients to receive either the study drug or placebo ^[21]. The primary endpoint of this study was gas exchange as measured by pulse oximetry during a six-minute walk. This study was discontinued prematurely due to concerns over excess mortality in the placebo group. Analysis revealed no difference between groups when assessed for the primary endpoint. However, pirfenidone was shown to confer benefit in measures of forced vital capacity and survival. A word of caution: experience has shown that, although a drug may appear promising in small phase II trials, large trials with additional power to determine efficacy may, in fact, reveal that a drug is ineffective. This was recently demonstrated by trials investigating the medication interferon-γ (IFN-γ). In the first multicenter study, 330 patients with IPF were randomized to receive either IFN-γ or placebo ^[22]. Patients were treated for 48 weeks with study drug and the primary endpoint measured was the effect on progression-free survival (a composite measure that included death and physiologic decline). The study showed no benefit from IFN-γ as measured by the primary endpoint. However, analysis of secondary endpoints revealed a trend toward improved survival in the group receiving IFN-γ. This trend did not reach statistical significance (p=0.08), but the study was not powered to detect an effect on survival. Therefore, a second trial was designed to specifically evaluate survival, with a plan to enroll over 800 IPF patients. Unfortunately, this second trial was recently discontinued after a planned interim analysis determined a lack of benefit from IFN-γ relative to placebo (unpublished data).

Non-pharmacological treatments

Lung transplantation
A survival benefit has been demonstrated for lung transplantation in IPF patients ^[96]. However, transplantation is only appropriate for carefully selected patients. Currently, five-year post-transplant survival approaches 50% ^[97]. Rejection remains a common and formidable problem leading to significant post-transplant morbidity and mortality. Following transplantation, patients require lifelong treatment with a combination of immunosuppressants in order to prevent rejection. Patients must also submit to frequent surveillance bronchoscopy, for the purpose of identifying infectious and inflammatory complications.

The timing of pulmonary transplantation poses additional challenges. Until recently, early referral was advocated for all patients with IPF because of long pre-transplant waiting times exceeding the median survival time of patients with IPF. However, new allocation scores in the U.S., devised to alleviate transplant waiting list mortality, have dramatically reduced waiting times for patients with IPF thus removing the impetus for early referral ^[98]. Now the decision to refer a patient for transplant revolves around the identification of the small subgroup of patients with IPF that might survive longer without transplant ^[96]. The judicious use of prognostic indicators, as discussed above, can inform such judgment.

The decision to perform lung transplantation in a patient with IPF requires careful consideration of the risks and benefits of such an undertaking. Advanced age precludes many patients with IPF from serious consideration of lung transplantation. Lung transplantation should be reserved for those with adequate social support and limited comorbidities, in order to face the rigors of post-transplantation medical management. Lung transplantation, on the whole, is best performed at specialized centers that employ experienced surgeons and physicians who are familiar with post-transplantation management.

Supportive measures
Regardless of the primary therapy, patients with IPF need to be treated with supportive measures as clinically indicated by their condition. For example, exercise-induced hypoxemia warrants a prescription for supplemental oxygen. While supplemental oxygen has been shown to improve exercise performance in chronic obstructive pulmonary disease (COPD), it has not been rigorously evaluated for the treatment of IPF. Nevertheless, a study that examined quality of life (QOL) in patients with IPF found no difference in QOL between patients receiving supplemental oxygen compared with those not receiving oxygen [99]. This is in spite of the fact that patients requiring oxygen were sicker.

Patients with IPF should be encouraged to enroll in a program for pulmonary rehabilitation. Pulmonary rehabilitation has not been rigorously examined in IPF, though quadriceps strength has been correlated with exercise capacity amongst patients with IPF ^[100]. This implies that training of the lower extremities could increase exercise capacity of IPF patients, as it does for patients with COPD. Because overall QOL is impaired in IPF, with specific deficits in the areas of physical health and perceived independence, it is reasonable to assume that rehabilitation programs, designed to increase physical well being and independence, will improve QOL ^[99].

Unresolved issues
IPF remains a disease for which the etiology is unknown. The pathogenesis is only poorly understood and the natural history of the disease is just beginning to reveal itself through observation of placebo groups from several large multi-center clinical trials. There is no definitive approach to the treatment of IPF because evidence for effective medical therapy is still lacking. Future directions for research should include programs that encourage the search for new molecular targets for therapy; and research to identify genetic susceptibility factors ^{[101, 102]}. Several centers are banking tissues from IPF patients that will enable translational research in the field. A multicenter clinical network, sponsored by the United States’ National Institute of Health, was recently established to facilitate the study of novel therapeutic agents as appropriate. In the next decade, it is likely that considerable progress will be made toward understanding and treating this devastating illness.