If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
No clear correlation is found between prednisone dose and FVC change in newly-treated pulmonary sarcoidosis patients.
•
Prednisone therapy aimed at improving/preserving FVC in newly-treated pulmonary sarcoidosis can often be reduced in dose.
•
A treatment strategy leading to a low cumulative prednisone dose was mainly defined by rapid dose tapering to 10 mg/day.
Abstract
Background
Prednisone is used as first-line therapy for pulmonary sarcoidosis. What dosing strategy has the best balance between effect and side-effects is largely unknown. We analyzed change in forced vital capacity (FVC) and weight during different prednisone doses used in daily practice for treatment naïve pulmonary sarcoidosis patients.
Methods
Multilevel models were used to describe FVC and weight change over time. Correlations were calculated using linear regression models.
Results
Fifty-four patients were included. FVC changed over time (p < 0.001), with an average increase of 9.6% predicted (95% CI: 7.2 to 12.1) at 12 months. Weight changed significantly over time (p < 0.001), with an average increase of 4.3 kg (95% CI: 3.0 to 5.6) at 12 months. Although FVC and weight changed significantly over time, there was little correlation between prednisone dose and FVC change, while weight increase correlated significantly with cumulative prednisone dose at 24 months. In patients treated with a high cumulative prednisone dose, baseline FVC was on average lower (p = 0.001) compared to low dose treated patients, while no significant differences were observed in need for second/third-line therapy or number of exacerbations. A strategy leading to a low cumulative dose at 12 months was defined by rapid dose tapering to 10 mg/day within 3.5 months.
Conclusions
These results suggest that prednisone therapy aimed at improving or preserving FVC in newly- treated pulmonary sarcoidosis can often be reduced in dose, using a treatment regimen that is characterized by early dose tapering.
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
Interstitial lung disease guideline: the british thoracic society in collaboration with the thoracic society of Australia and New Zealand and the Irish thoracic society.
]. Although prednisone treatment in pulmonary sarcoidosis is reported to induce short-term benefits on clinical symptoms and inflammation, it remains unclear whether the therapy modifies long-term progression of the disease [
]. Therapy should therefore primarily be aimed at symptom relief, inflammation control to prevent (further) organ damage and improving patient's quality of life while avoiding unnecessary side-effects [
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
Interstitial lung disease guideline: the british thoracic society in collaboration with the thoracic society of Australia and New Zealand and the Irish thoracic society.
] per day for 1–3 months. Subsequently, prednisone dose should be tapered to a maintenance dose of 5–10 mg/day, which is commonly continued for 6–12 months before discontinuation [
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
Interstitial lung disease guideline: the british thoracic society in collaboration with the thoracic society of Australia and New Zealand and the Irish thoracic society.
]. As these guidelines include a broad range in recommended prednisone doses, variation in treatment regimen in clinical practice is suspected, varying from low- to high-dose treated patients. Prolonged high dose corticosteroid therapy is associated with numerous side-effects, including weight gain, diabetes and osteoporosis [
What dosing strategy has the best balance between effect and side-effects is largely unknown. Therefore, in this study we aimed to evaluate treatment effect on forced vital capacity (FVC) (effect) and weight (side-effect) of different prednisone doses used in daily practice.
2. Materials & methods
2.1 Study design
This study is a multicenter retrospective study, performed in one academic sarcoidosis referral center (Erasmus MC) and three regional training hospitals (Franciscus Gasthuis & Vlietland, Ikazia hospital and Amphia hospital) in the Netherlands. Medical records were reviewed for demographic and diagnostic data, organ involvement, radiographic Scadding stage, prednisone dose, weight, pulmonary function parameters and exacerbations. An exacerbation was determined as an increase in daily prednisone dose from 5 to 10 mg/day maintenance dose to ≥ 20 mg/day or if patients restarted prednisone after prior discontinuation. Data that was available up to 5 years following therapy initiation were collected per patient. Individual prednisone regimens were analyzed. Weight was collected from pulmonary function records.
2.2 Patients
Treatment naïve sarcoidosis patients, in whom prednisone therapy was started for a pulmonary indication between January 2000 and December 2013, were included in this study. Patients were identified by screening medical records of patients that were in hospital-specific databases that track sarcoidosis diagnosed patients over time. Patients were included when they met standard criteria for diagnosis of the disease [
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
], were treatment naive and had a treatment indication for pulmonary sarcoidosis as determined by the treating physician. Patients were excluded when they were: 1. primarily treated for a non-pulmonary indication; 2. treated solely with methotrexate for a pulmonary indication; and 3. when there were less than two hospital visits documented.
2.3 Ethical requirements
Formal consultation with the Medical Ethical Committee of the Erasmus MC learnt that, under the Dutch act for medical research involving human subjects (Wet Medisch Onderzoek), approval of this study by the Medical Ethical Committee was not required. The local institution review board of all participating centers approved with registration number MEC-2014-089.
2.4 Statistical analysis
The comparison of means of continuous variables were tested with the student t-test, the categorical variables were tested with the X2 or the Fisher exact test. Absolute changes in FVC and weight were used as outcome in multilevel models. Log time appeared to be an adequate transformation to enter as fixed factor in the model, while patient functioned as a random intercept. For different points in time we analyzed the correlation between cumulative prednisone dose and the absolute change in outcome between start of treatment and the specific time point. Correlations were calculated using linear regression models.
FVC is shown as mean percent (%) predicted (± standard deviation (SD)) or as mean absolute change of % predicted (including a 95% confidence interval (CI)) compared to baseline. Weight is shown as mean kg (± SD) or as mean absolute change (including 95% CI) in kg compared to baseline. Prednisone dose is shown as mean daily dose in mg or as cumulative dose in mg.
For pulmonary function tests the European Community for Steel and Coal 1993 prediction equations were used in all hospitals.
Statistical analyses were performed using SPSS (version 21.0.0.1) and R software (version 3.2.2). Figures were created with R software. A p-value < 0.05 was considered as statistically significant.
3. Results
3.1 Patients
A total of 54 treatment naïve sarcoidosis patients that were initiated on prednisone therapy for a pulmonary indication were identified for this study; two third from regional hospitals and one third from an academic referral center for sarcoidosis. Mean % predicted FVC at start of prednisone treatment was 83.4 ± 20.4, and mean % predicted diffusing capacity of the lung for carbon monoxide (DLCO) (corrected for hemoglobin levels) was 69.9 ± 25.0 (Table 1). Average weight was 79.2 ± 19.2 and average body mass index (BMI) was 26.6 ± 5.9 kg/m2; 48.1% of the patients were men and 71.1% of the patients had Scadding stage II sarcoidosis. Additional baseline characteristics are shown in Table 1.
Table 1Baseline characteristics of study cohort.
Characteristics
Patients (n = 54)
Age (n = 54)
44 ± 13
Gender
Male
26 (48,1)
Female
28 (51,9)
Treated in
Academic center
18 (33,3)
Training hospital
36 (66,7)
Race/Ethnicity
Unknown
17
White
23 (62,2)
Black
7 (18,9)
Hispanic
7 (18,9)
BMI in kg/m2 (n = 52)
26,6 ± 5,9
Weight in kg (n = 52)
79,2 ± 19,2
Smoking
Unknown
8
Never
21 (45,7)
Current
7 (15,2)
Former
18 (39,1)
Scadding stage
Unknown
9
I
8 (17,8)
II
32 (71,1)
III
5 (11,1)
Extra pulmonary organ involvement*
No/Yes
26/28
Skin
7 (13)
Eyes
14 (25,9)
Joints
15 (27,8)
Other (e.g. liver, heart, neural)
9 (16,7)
Prednisone dose at start
32,6 ± 8,7
>10 ≤ 20 mg
6 (11,2)
>20 ≤ 30 mg
32 (59,3)
>30 ≤ 40 mg
12 (22,2)
>40 ≤ 50 mg
2 (3,7)
>50 ≤ 60 mg
2 (3,7)
Pulmonary function tests
% predicted FVC (n = 42)
83,4 ± 20,4
% predicted DLCOc (n = 34)
69,9 ± 25,0
Categorical data is presented as No. (% of total patients with available data) and continuous data as mean ± standard deviation of patients with available data (No. of patients with available data is presented behind the continuous variable). Abbreviations: DLCOc: diffusing capacity of lung for carbon monoxide (corrected for hemoglobin levels), FVC: forced vital capacity, kg: kilograms, mg: milligrams. *As assessed and described by the treating physician.
Mean initial prednisone dose was 32.6 ± 8.7 mg (Table 1). On average, prednisone was tapered to 10 mg/day at approximately 6 months (e-Fig. 1). Mean FVC (% predicted) change over time was calculated using a multilevel model that incorporated regression lines of all 54 individual patients (e-Fig. 2). FVC changed significantly over time (p < 0.001), with an average increase of 7.4% predicted (95% CI: 5.5 to 9.3) at 3 months and 9.6% predicted (95% CI: 7.2 to 12.1) at 12 months (Fig. 1). At 24 months, an average increase of 10.8 (95% CI: 8.0 to 13.5) was observed (Fig. 1), which was largely preserved in patients with data available at 3 and 5 years following therapy initiation (e-Fig. 2A). FVC increase over time significantly depended on baseline FVC (p = 0.002), whereby a higher increase was observed in patients with a lower FVC at start than patient with a higher FVC at start. Interestingly, the major increase in FVC occurred within 1–3 months of treatment (Fig. 1).
Fig. 1FVC change after initiation of prednisone therapy.
FVC change over time during prednisone treatment. Percent predicted FVC change compared to baseline (absolute change) is shown. Statistics: Data is calculated using a multilevel model with FVC over time as outcome. Abbreviations: FVC: forced vital capacity, % pred: percent predicted.
The association between change in FVC and prednisone dose used, was determined. Although all correlations were weakly positive, no significant correlation was found between FVC change and cumulative prednisone dose used in the short-term (3 (Fig. 2A), 6 (not shown) and 9 months (not shown)) and long-term (12 (Figure 2B) and 24 months (Fig. 2C)). At 12 months, the correlation became almost zero after correction for baseline FVC (correlation: −0.07, p = 0.76). Similarly, no significant correlation was found between the average daily prednisone dose used in the first 3 months and FVC change (e-Fig. 3). In addition, no correlation was found on any given time point between change in DLCO and prednisone dose (data not shown).
Fig. 2Association between FVC and prednisone dose.
Correlation between percent predicted FVC change and cumulative prednisone dose in milligrams used in the short-term at A. 3 months, and long-term at B. 12 months and C. 24 months. The cumulative prednisone dose used was calculated per individual at time of an available FVC around month 3, 12 and 24 following treatment initiation. The correlation was analyzed using Spearman's rank-order correlation test. The regression line with the correlation and p-value are shown in the plot. Abbreviations: FVC: forced vital capacity, % pred: percent predicted, mg: milligrams.
To gain insight in side-effects occurring during prednisone treatment, we captured weight change over time using a multilevel model (e-Fig. 4). Weight changed significantly over time (p < 0.001), with an average increase of 3.3 kg (95% CI: 2.3 to 4.3) at 3 months and 4.3 kg (95% CI: 3.0 to 5.6) at 12 months (Fig. 3). After tapering of the prednisone dose to a mean of ≤ 5 mg/day, weight remained increased compared with baseline with an average of 4.8 kg (95% CI: 3.3 to 6.3) at 24 months, which persisted up to 5 years in patients with data available (Fig. 3 and e-Fig. 1). Similar to FVC change (Fig. 1), the largest proportion of weight gain occurred during the first 1–3 months of treatment (Fig. 3).
Fig. 3Weight change after initiation of prednisone therapy.
Weight change in kilograms over time during prednisone treatment compared to baseline (absolute change). Statistics: Data is calculated using a multilevel model with weight over time as outcome. Abbreviations: kg: kilograms.
We aimed to determine whether weight change depends on total prednisone dose used. No significant correlation was found between weight change and cumulative prednisone dose used in the short-term at 3 (Fig. 4A), 6 (not shown) and 9 months (not shown). Also, no significant correlation was found between daily prednisone dose used and weight change in the first 3 months (e-Fig. 5). However, an association between cumulative prednisone dose and weight change became apparent and significant in the long-term at 12 (Fig. 4B) and 24 months (Fig. 4C), respectively.
Fig. 4Association between weight change and prednisone dose.
Correlation between weight change in kilograms and cumulative prednisone dose in milligrams used in the short-term at A. 3 months, and long-term at B. 12 months and C. 24 months. The cumulative prednisone dose used was calculated per individual at time of an available weight around month 3, 12 and 24 following treatment initiation. The correlation was analyzed using Spearman's rank-order correlation test. The regression line with the correlation and p-value are shown in the plot. Abbreviations: kg: kilograms, mg: milligrams.
In order to gain insight in clinical practices that are associated with reduced prednisone use, we aimed to determine the average daily prednisone dose regimen administered to patients that received a relatively low versus a higher cumulative dose at 12 months. Therefore, patients with available data on prednisone therapy at 12 months (n = 39) (e-Fig. 1) were divided into a low dose and a high dose group. The median cumulative dose given at 12 months was 4000 mg (range 1050–13090 mg) thus patients who had received less than 4000 mg prednisone were assigned to the low dose group (n = 20), and patients who had received 4000 mg prednisone or more were assigned to the high dose group (n = 19).
The average start dose of prednisone was lower in the low dose treated group than the high dose treated group, but this did not reach statistical significance (30.3 ± 9.1 mg versus 36.1 ± 9.8 mg; p = 0.06) (Table 2). The treatment strategy leading to a lower cumulative dose prednisone at 12 months seemed mainly characterized by earlier dose tapering, i.e. less than 10 mg/day at 3.5 months, whereas the high cumulative dose group was characterized by later dose tapering i.e. less than 10 mg/day at 6.9 months (Fig. 5).
Table 2Baseline characteristics of low versus high dose treated patients.
Low dose (n = 20)
High dose (n = 19)
P-value
Age
46 ± 14,0 (20)
43 ± 12,5 (19)
0.54
Gender
Male
9 (45)
9 (47,4)
0.88c
Female
11 (55)
10 (52,6)
Treated in
Academic center
3 (15)
6 (31,6)
0.27f
Regional hospital
17 (85)
13 (68,4)
Race/Ethnicity
Unknown
8
5
White
8 (66,7)
7 (50,0)
0.76f
Black
2 (16,7)
4 (28,6)
Hispanic
2 (16,7)
3 (21,4)
BMI in kg/m2
27,5 ± 5,1 (18)
27,9 ± 6,8 (19)
0.87
Weight in kg
82,1 ± 17,6 (18)
83,6 ± 23,1 (19)
0.83
Smoking
Unknown
4
1
Never
7 (43,8)
9 (50,0)
0.34f
Current
4 (25,0)
1 (5,6)
Former
5 (31,2)
8 (44,4)
Scadding stage
Unknown
1
3
I
4 (21,1)
3 (18,8)
1.00f
II
13 (68,4)
11 (68,8)
III
2 (10,5)
2 (12,5)
Extra pulmonary organ involvement*
12 (60)
7 (36,8)
0.15c
Skin
4$ (20)
0 (0)
0.11f
Eye
4 (20)
5 (26,3)
0.72f
Joint
6 (30)
7 (36,8)
0.65c
Other
1 (5)
4$$ (21,1)
0.18f
Multiple organs
2 (10)
7 (36,8)
0.07f
Prednisone dose use at start
30,3 ± 9,1
36,1 ± 9,8
0.06
>10 ≤ 20 mg
5 (25)
0 (0)
0.08f
>20 ≤ 30 mg
9 (45)
12 (63,2)
>30 ≤ 40 mg
5 (25)
4 (21,1)
>40 ≤ 50 mg
1 (5)
1 (5,3)
>50 ≤ 60 mg
0
2 (10,5)
Pulmonary function tests
% predicted FVC
96,5 ± 20,8 (16)
70,7 ± 15,7 (14)
0.001
% predicted DLCOc
76,9 ± 20,6 (13)
64,7 ± 18,2 (13)
0.12
Categorical data is presented as No. (%) and continuous data as mean ± standard deviation (No. of patients with available data and that were included in the analyses). Statistics: We tested the continuous outcomes for residuals and they had a normal distribution. Significance between continuous data was analyzed with an unpaired student t-test. Categorical data was analyzed with a X2 (=c) or the Fisher (=f) exact test. P values are given for significant differences between the two groups. Abbreviations: DLCOc: diffusing capacity of lung for carbon monoxide (corrected for hemoglobin levels). FVC: forced vital capacity, kg: kilograms, mg: milligrams. $ Two patients had erythema nodosum. $$ One patient had heart involvement, one had parotids involved and two patients had peripheral neurological involvement. *As assessed and described by the treating physician.
The mean prednisone dose given per day in milligrams (mg) to patient who had received a low (< 4000 mg; n = 20) or high (≥ 4000 mg; n = 19) cumulative dose prednisone at 12 months. Patients who received a low or high cumulative prednisone dose at 12 months were treated with more than 10 mg/day up to 3.5 or up to 6.9 months, respectively.
We examined baseline and therapy-response characteristics in order to determine factors associated with a treatment strategy leading to a high or low cumulative dose.
No significant differences were observed between the high and low dose groups in baseline characteristics such as race/ethnicity (p = 0.76), Scadding stage (p = 1.00), weight (p = 0.83), BMI (p = 0.87) or DLCO (% predicted) (p = 0.12) (Table 2). Furthermore, no statistical significant difference was observed in the presence of extra pulmonary organ involvement (p = 0.15), but the low dose treated group tended to have more skin involvement (including 2 patients with erythema nodosum) than the high dose treated group (20% versus 0%, p = 0.11), while high dose treated patients seemed to have more involvement of other organs, such as cardiac and neurological involvement (5% versus 21.1%, p = 0.18). Also, a trend was observed that high dose treated patients had more often multiple organs affected (10% versus 36.8%, p = 0.07) (Table 2). Interestingly, the baseline percent predicted FVC was significantly lower in the high dose treated group (70.7 ± 15.7) than the low dose treated group (96.5 ± 20.8) (p = 0.001) (Table 2). When we looked into therapy-response characteristics, no significant differences were however observed between the two groups in the number of patients that experienced an exacerbation during tapering (p = 0.66) or patients that received additional second or third line drugs (p = 0.34) (Table 3).
Table 3Response to therapy of low versus high dose treated patients.
Low dose (n = 20)
High dose (n = 19)
P-value
Exacerbation
No
18 (90)
16 (84,2)
0.66
Yes
2 (10)
3 (15,8)
Additional 2nd or 3th line drugs
No
19 (95)
16 (84,2)
0.34
Yes
1 (5)
3 (15,8)
Data are presented as No. (%). Categorical data was analyses with the Fisher exact test. P values are given for significant differences between the two groups.
It currently remains unknown what dosing strategy for first-line prednisone treatment in pulmonary sarcoidosis has the best balance between effect and side-effect. In this study, we aimed to determine treatment outcome of different prednisone regiments used in current clinical practice for newly-treated pulmonary sarcoidosis. No strong correlation was found between prednisone dose used and FVC (and DLCO) change at different time points, while weight increased significantly more in patient receiving a higher cumulative prednisone dose in the long-term. These data suggest that a treatment strategy leading to a lower cumulative dose prednisone in the long-term has the potential to be equally effective in treating pulmonary sarcoidosis patients as a higher dose regimen, while reducing side-effects.
In this study, we gained more insight in prednisone therapy-induced effects. Serial FVC is currently the best end-point to monitor pulmonary sarcoidosis patients' response to therapy [
]. We observed an increase of approximately 10% predicted FVC after prednisone treatment initiation. Strikingly, the major part of the FVC change seems to occur within the first 1–3 months of therapy, as has been described in exacerbation patients [
], although we cannot formally discount the notion that this improvement is in part due to the natural history of the disease. Importantly, no clear correlation was found between average daily or cumulative prednisone dose given and FVC change.
Weight gain is a well-known side-effect of prednisone therapy [
]. In this study, we found a correlation between weight increase and cumulative dose used, specifically in the long-term. A treatment strategy leading to a lower cumulative dose at 12 months was characterized by early dose tapering, i.e. to less than 10 mg/day at 3.5 months, while the strategy leading to a higher dose was characterized by prolonged use of more than 10 mg/day up to 6.9 months. Prolonged prednisone therapy for pulmonary sarcoidosis at a dose of more than 10 mg a day was reported to induce significant more side-effects than a lower dose [
], which is now further supported by the results of this study. These data highlight that early dose tapering can be essential in reducing cumulative prednisone dose and hazardous side-effects such as weight gain.
Taken together, our data suggest that a lower prednisone dose increases and/or maintains FVC similar to a higher dose, thus meeting an important therapy objective, i.e. preserving organ function, while avoiding side-effects [
There are limitations to our study. Our study included a small number (n = 7) of black patients, thus it remains to be determined whether these patients, who more often suffer from a severe chronic form of sarcoidosis, respond identically to first-line prednisone therapy. While a number of characteristics such as radiographic Scadding stage, BMI and DLCO were similar, baseline FVC was significantly lower in patients treated with a higher cumulative prednisone dose. Also, these patients tended to have features associated with severe disease, such as multiple organ involvement more often. We can therefore not exclude that these patients suffered of worse disease, which may have prohibited early dose tapering for other reasons. Indeed, other factors are often taken into account by the treating physicians when determining a treatment strategy, such as change in symptoms, radiological features or extra-thoracic organ involvement [
]. Alternatively, physicians may have had a tendency to treat patients with a lower FVC with higher doses of prednisone from the start. Hence, the retrospective design of our study does not allow for firm conclusion on whether the high dose treated patients may have benefitted from a lower dose. Nevertheless, increase in FVC over time was significantly higher in patients with a lower baseline FVC in our study population. The weak, non-significant, association between cumulative prednisone dose and FVC change at 12 months (r = 0.203) was no longer found (r = −0.07) after correction for baseline FVC. These data suggest that FVC response to prednisone is not dependent on the (cumulative) dose even for patients with a low baseline FVC. Furthermore, high dose treated patients did not more often experience an exacerbation during tapering or receive additional second or third line drugs than low dose treated patients.
In conclusion, we did not find a clear association between prednisone dose and FVC change in newly-treated pulmonary sarcoidosis patients, while weight gain positively correlated with cumulative prednisone dose used in the long-term. These results support that prednisone therapy that is mainly aimed at maintaining and/or improving FVC in pulmonary sarcoidosis may be dose-reduced in current clinical practice by early dose tapering and highlight the need for prospective clinical trials on treatment strategies that carefully balance effects-side-effects.
Conflicts of interest
None.
Funding
None.
Author contributions
CB, LP, LvdT, MW, HH, MK, MSW, BvdB contributed to the study concept and design; CB, LP, JV, MG, RH, LvdT, MW, HH, MK, MSW, BvdB patient recruitment and data collection; CB, LP, CL, BvdB data analysis and interpretation; CB, LP, CL, BvdB manuscript preparation, and manuscript drafting and all authors gave final approval on the manuscript.
Acknowledgements
The authors gratefully acknowledge patients, research nurses and physicians participating in this study from the Erasmus MC, Franciscus Gasthuis & Vlietland, Ikazia hospital and Amphia hospital in The Netherlands.
Appendix A. Supplementary data
The following are the supplementary data related to this article:
Joint statement of the american thoracic society (ATS), the european respiratory society (ERS) and the world association of sarcoidosis and other granulomatous disorders (WASOG) adopted by the ATS board of directors and by the ERS executive committee, february 1999.
Interstitial lung disease guideline: the british thoracic society in collaboration with the thoracic society of Australia and New Zealand and the Irish thoracic society.
30361-X&id=gr1.jpg){kind=link}
30361-X&id=gr2.jpg){kind=link}
30361-X&id=gr3.jpg){kind=link}
30361-X&id=gr4.jpg){kind=link}
30361-X&id=gr5.jpg){kind=link}