Idiopathic pulmonary fibrosis (IPF) is an incurable progressive lung disease characterised by cough and exertional dyspnoea. Although disease course varies, the majority develop progressive respiratory failure, culminating in death within five years of diagnosis unless lung transplantation is possible.1 Worldwide prevalence is estimated to be between 1.25 and 63 per 100,000, depending on the population studied and definition of IPF,2 while incidence appears to be increasing globally.3 At present the epidemiology of IPF in New Zealand is unknown. In Canterbury, outpatient IPF diagnoses are not coded, making us reliant on inpatient coding which underestimates the true impact of disease.

Although there is no cure for IPF, the anti-fibrotic medications pirfenidone and nintedanib have been shown to reduce the decline in forced vital capacity (FVC) in patients with IPF in phase III trials.4,5 Data has since emerged about their effectiveness in real-world IPF patients, including from the Australian IPF registry.6 In New Zealand, pirfenidone has been available since January 2017 under a special authority scheme for patients with a diagnosis of IPF and FVC between 50 and 80% predicted, while nintedanib only became funded in October 2018. However, at present there is no IPF registry or outpatient coding allowing us to capture data on their efficacy.

In addition to anti-fibrotic therapies, supportive and non-pharmacological care maintains a central role in the management of this condition. Pulmonary rehabilitation is beneficial in terms of six-minute walk test (6MWT) and health-related quality of life in patients with IPF.7 Domiciliary oxygen therapy benefits patients with IPF, and its use is recommended where appropriate.8 The role of palliative care is under-recognised but is crucial given the unremitting symptoms of this disease which significantly impacts on quality of life.9

Recently published position papers from the Thoracic Society of Australia and New Zealand highlight the importance of making an accurate diagnosis of IPF in the context of an interstitial lung disease (ILD) multidisciplinary meeting (MDM) with the aim of optimising the holistic care of patients with IPF.8,10 In response to this, Canterbury District Health Board (DHB) instituted a dedicated monthly ILD MDM in 2017, with the ability to accept referrals both locally and from the surrounding regions. One year on from the institution of the MDM, and the availability of pirfenidone in New Zealand, the aims of this study were therefore; to establish an estimate of the prevalence of IPF in the Canterbury and neighbouring regions; to assess current practice with respect to the aforementioned position statements;8,10 and to review the early experience with pirfenidone in our community.

Methods

We performed a retrospective observational study of patients with idiopathic pulmonary fibrosis managed in secondary or tertiary care across the four Northern district health boards (DHB) in New Zealand’s South Island: Canterbury DHB, South Canterbury DHB, West Coast DHB and Nelson and Marlborough DHB. Patients were identified through keyword searches for “usual interstitial pneumonia” (UIP), “idiopathic pulmonary fibrosis” and “cryptogenic fibrosing alveolitis” in radiology reports and clinical communications stored in the electronic record, through ICD-10 coding of discharge summaries over the last 10 years up to April 2018 and from a database of patients discussed in the ILD MDM. Private providers in Canterbury were also invited to provide patients for inclusion in the study.

In order to capture all patients who had been treated with pirfenidone, those who were alive on 1 January 2017 then underwent manual review of the case notes and were included if they had a radiological or histological finding of definite UIP and documented clinical correlation consistent with a diagnosis of IPF. Patients with a radiological finding of possible UIP were included without lung biopsy, provided the treating clinician felt the clinical picture was one of IPF.11 Follow up was complete until April 2018 or the time of death. The prevalent cohort included those alive in April 2018 and prevalence was estimated using the Ministry of Health projected populations for the regions studied.

Clinical, physiological and demographic data were collected using a pre-formatted spreadsheet. Physiological data collected included forced vital capacity (FVC) and where available; diffusion capacity (DLCO), six-minute walk test (6MWT) and oxygen saturation (Sp02). This included the most recent measurements and, in those treated with pirfenidone, measurements immediately prior to commencing treatment. Clinical data collected included diagnostic investigations (radiology and histology), MDM discussion; the presence of complications including hospitalisation and death; documentation of end-of-life discussion or advance care plans (ACP); and prescriptions for pirfenidone and domiciliary oxygen therapy. Palliative care, nurse specialist and pulmonary rehabilitation referrals were collected for those patients in the Canterbury DHB region only.

Descriptive data is presented as n (%), mean (standard deviation) or median (interquartile range). Comparative statistics were performed using students t test for continuous data and Fisher’s Exact test for categorical data. Results were considered significant with a p value <0.05.

As a retrospective audit, this study was exempted from ethical review by the New Zealand Health and Disability Ethics Committee. Approval for this study was granted by the Canterbury DHB Clinical Audit Department.

Results

Three hundred and forty-nine records were reviewed and 68 patients who met the inclusion criteria were identified. Fifty (73.5%) were male and the median age was 80 years (72–84). Seven (10.3%) patients were of non-European descent (three (4.4%) New Zealand Māori, four (5.9%) Asian, 52 (76.5%) New Zealand European, nine (13.2%) other European). Median follow-up was 33 months (14–49). Fifteen (22.1%) patients died during follow-up, therefore 53 patients made up the prevalent cohort in April 2018, providing an estimated prevalence of IPF of 6.53 cases/100,000 persons.

All patients underwent high-resolution computed tomogrophy as the initial investigation of choice. Sixty-five patients had radiological features consistent with UIP (63 definite, two possible) and required no further investigation. One patient underwent a non-diagnostic transbronchial biopsy and proceeded to surgical lung biopsy. Two further patients underwent surgical lung biopsy to establish the diagnosis of IPF. One patient had an incidental finding of UIP in a lobectomy sample following surgery for non-small cell lung cancer. Overall, six (8.8%) patients from this cohort had been discussed in the ILD MDM at the end of follow-up. This included three out of 10 patients diagnosed since 1 January 2017. Four out of the six patients discussed were given a new diagnosis of IPF on the basis of MDM discussion; the remaining two patients had an established diagnosis supported by the MDM.

Eight (11.8%) patients had undergone a 6MWT at the end of follow up. Resting oxygen saturations were documented in 44 (64.7%) patients. Twenty-eight (41.2%) patients had documented resting hypoxia (Sp02<90%), and a further six, without resting hypoxia, had evidence of exertional hypoxia. All patients with resting hypoxia received a discussion regarding the role of oxygen therapy and domiciliary oxygen therapy was prescribed to 15 (22.7%) patients in total. Nineteen (27.9%) patients had clearly documented end-of-life discussions or an advance care plan in place.

Thirty-six (52.9%) patients were eligible for treatment with pirfenidone on the basis of their physiological parameters. The majority of ineligible patients (22/32, 68.8%) had an FVC above the treatment threshold of 80% predicted. In total 20 (29.4%) patients were prescribed pirfenidone between 1 January 2017 and the end of follow-up, 18 of whom met the special authority criteria and two who were treated outside of this guidance with an FVC <50% predicted. Patients prescribed pirfenidone were no more likely to have been discussed in the ILD MDM (2/20 vs 4/48, p>0.05) than those not prescribed pirfenidone, but were more likely to have undergone a 6MWT (5/20 vs 3/48, p<0.05) or to have been hospitalised in the last 12 months (12/20 vs 11/48, p<0.05) (Table 1).

Table 1: Demographics and clinical characteristics of patients who were prescribed, and were not prescribed pirfenidone during the follow-up period.

‍

* p<0.05

Seven patients were still taking pirfenidone at the end of follow-up. Among those who discontinued therapy, five died during follow-up and eight (40%) discontinued therapy due to side effects. No patients had discontinued therapy due to progressive disease at the end of follow-up.

Eighteen patients eligible for pirfenidone were not prescribed treatment. Reasons for this were documented in the clinical record in nine cases and included active decisions not to treat due to ‘stable/slowly progressive disease’ (n=4), ‘comorbidity’ (n=1), ‘patient decision’ (n=1), and ‘treatment decision still under consideration’ (n=3).

Fifteen (22.1%) patients died during follow-up with a median survival of 19 months (6.5–54). These patients had significantly lower FVC (2.16L vs 2.64L, p<0.05) than those who survived. They were also more likely to have been prescribed domiciliary oxygen therapy (6/15 vs 9/53, p<0.05), have had end-of-life plans made (8/15 vs 11/53, p<0.05) and to have been hospitalised in the last 12 months (9/6 vs 14/53, p<0.05).

Among patients in the Canterbury DHB region (n=53), 10 (18.9%) had been referred to a respiratory nurse specialist, 10 (18.9%) for pulmonary rehabilitation and nine (16.9%) to palliative care services respectively. Five patients in this region died without a referral to palliative care services.

Discussion

In this study we have estimated the prevalence of IPF in secondary and tertiary care centres across Northern regions of the South Island of New Zealand to be 6.53/100,000. This is comparable with worldwide prevalence rates of between 1.25 and 63/100,000.2 Our estimate is at the lower end of this range, and may reflect the fact that patients were only captured if they had come into contact with a physician in a secondary or tertiary centre, thereby underestimating the true prevalence of IPF in the community. Nevertheless, the population captured is likely to have included the majority of symptomatic or severely affected patients and therefore this estimate remains relevant for the planning of hospital-based services.

Guidelines recommend discussion of all new cases of IPF in an ILD MDM to confirm the diagnosis.8 In our study only 8% of patients had been discussed in such a forum. No doubt this reflects the fact that an ILD MDM has only been available at Christchurch Hospital since 2017, though even among patients diagnosed since its inception, the rate is only 30%. The lack of uptake likely reflects capacity issues, as the ILD MDM was only held monthly initially and triage priority given to new or complex ILD cases. However, four patients were given a new diagnosis of IPF on the recommendation of the MDM, in keeping with international literature whereby changes in diagnosis and management have been shown to occur in >50% of ILD cases discussed in such a forum.12 Improving the capacity and uptake of this service should therefore be considered a priority, particularly as multidisciplinary discussion is a requirement for funded pirfenidone therapy in New Zealand.

In this study 25% of patients prescribed pirfenidone died during follow-up. Among those surviving, less than half achieved long-term adherence. The majority reported gastrointestinal intolerance although reasons for discontinuation were not always documented. This is in contrast to real world experience of pirfenidone elsewhere in which discontinuation rates due to adverse events are reported as between 16 and 20%.13,14 In these studies, dose adjustment was associated with improved adherence, and elsewhere dedicated close follow-up, patient support and education has been of benefit.15 Currently no specific protocol for treatment initiation exists in our institution, but this may be an option, incorporating the above factors, that needs to be explored to improve adherence. Alternatively, the poor tolerance may reflect the age of our patients, who had a median age of almost 10 years greater than those described elsewhere.

The provision of supportive care and end-of-life planning are important considerations in what remains a terminal condition. Fifteen (22.7%) patients in this study were prescribed either ambulatory or long-term oxygen therapy, which is comparable to data from the British Thoracic Society IPF Registry in which 26% of patients were prescribed oxygen.16 However, documentation of oxygen saturations was only complete in 64.7% of patients, and exertional testing was infrequent, so it is possible that patients who may benefit from this therapy are not being identified in the clinic. Meanwhile, documentation of end-of-life planning was complete in just 27.9% of patients.

It is reassuring to see that patients who died were more likely to have been prescribed oxygen and to have had end-of-life care plans made, but the overall proportion of patients receiving these interventions was low, as it was for those referred to specialist nursing services and palliative care in the Canterbury DHB region. Our rate of palliative care referrals are similar to those (13.7%) identified in one large retrospective cohort,17 but this would appear to be inadequate for a disease process with high mortality and symptom burden. In fact, data from the Swedish Registry of Palliative Care would suggest that patients with ILD receive poorer access to end-of-life care than patients with lung cancer.18 There is clearly an unmet need in this area, and it may be the case that use of a decision support tool to prompt end-of-life discussions could improve documentation of these issues and referral for appropriate supportive services where indicated.19

Limitations of this study include the retrospective nature of data collection and case identification. These may have introduced a risk of ascertainment bias, or an underestimate of disease prevalence respectively. Furthermore, data on referrals to palliative care, pulmonary rehabilitation and specialist nursing was not available outside Canterbury DHB.

To our knowledge this is the first study to estimate the prevalence of IPF in New Zealand and is an important step forward in planning for the future in a disease in which worldwide incidence is increasing, therapeutic options are widening and pathways of care are becoming increasingly complex, yet standardised.8 Furthermore we have been able to identify limitations in current practice locally, in terms of the uptake of MDM discussion, physiological measurements and supportive care, along with high discontinuation rates of therapy. Improvements in these areas may be beneficial for patients but will also have impacts in terms of the resources required to achieve this.

‍