First, the reality of pirfenidone treatment was investigated. The baseline characteristics of 85 patients are listed in Table 1.
The mean administered dose of pirfenidone was 1,242 mg/day; 25.9%, 54.1%, and 20% of patients received a dose of 1,800, 1,200, and <1,200 mg/day, respectively, showing that the proportion of patients treated at a full dose was limited. The reasons for not being able to increase the dose to a full dose (Table 2) included adverse events (54.0%), which was the most common reason and followed by decision by attending physician (36.5%) and poor adherence (7.9%). In the group receiving <1,200 mg/day, the proportion of adverse events was higher (70.6% vs 47.8%), and the proportion of doctor's decisions was lower (23.5% vs 41.3%) than those of the 1,200 mg/day group.
The mean duration of pirfenidone treatment was 16.3 months. During the observation period, pirfenidone was discontinued in 24 patients (28.2%) for the following reasons: adverse events (45.8%), poor adherence (33.3%), financial difficulties (8.3%), decision by the attending physician (8.3%), and insufficient treatment effect (4.2%).
Next, we compared the characteristics of patients treated with different doses pirfenidone (Table 3). Regarding the age at the time of pirfenidone treatment initiation, the 1,200 mg/day group was significantly older than the 1,800 mg/day group (75.3 vs 69.8 years). The FVC or diffusing capacity of the lung for carbon monoxide (DLCO) level was significantly lower in patients receiving <1,200 mg/day than in those receiving 1,800 or 1,200 mg/day. The Krebs von den Lungen-6 (KL-6) level was significantly lower in the 1,200 mg/day group than in the 1,800 mg/day group.
Table 1. Patients’ baseline characteristics
The number of patients
|
85
|
The age of diagnosis (years)
|
70.4 ± 0.9
|
The age of initiation of pirfenidone (years)
|
73.5 ± 0.8
|
Male sex (%)
|
65 (76.5%)
|
BMI (kg/m2)
|
22.15 ± 0.46
|
BSA (m2)
|
1.57 ± 0.19
|
Current or ex / never smoker
|
63 (74.1%)/ 22 (25.9%)
|
Smoking index
|
753.0 ± 80.1
|
HRCT pattern
|
UIP
|
39 (45.9%)
|
Probable UIP
|
29 (34.1%)
|
Indeterminate for UIP
|
17 (20.0%)
|
FVC (L, % predicted)
|
2.27 ± 0.09, 76.5% ± 2.7%
|
DLCO (% predicted)
|
58.3% ± 2.9%
|
Alb (g/dL)
|
3.80 ± 0.05
|
KL-6 (U/mL)
|
1,160 ± 83
|
GAP stage (I/ II/ III)
|
37 (50.7%)/ 20 (27.4%)/ 16 (21.9%)
|
Treatment with corticosteroids
|
22 (25.9%)
|
Dose of pirfenidone
|
<1,200 mg/day
|
17 (20.0%)
|
1,200 mg/day
|
46 (54.1%)
|
1,800 mg/day
|
22 (25.9%)
|
Mean treatment duration (months)
|
16.3 ± 2.3
|
Median treatment duration (months, IQR)
|
10.0 (2.8–18.9)
|
Alb: serum albumin; BMI: body mass index; BSA: body surface area calculated by the Du Bois formula; DLCO: diffusing capacity of the lung for carbon monoxide; FVC: forced vital capacity; GAP: Gender–Age–Physiology; HRCT: high-resolution computed tomography; IQR: interquartile range; KL-6: Krebs von den Lungen-6; UIP: usual interstitial pneumonia. The data at the time of the initiation of pirfenidone treatment is presented as number (%) or mean ± SEM. BMI and BSA: n=84, FVC: n=80, DLCO: n=65, Alb: n=82, GAP stage: n=73.
Table 2. Reasons for not increasing the pirfenidone dose to 1,800 mg/day
Dose of pirfenidone
|
Total
|
< 1,200 mg/day
|
1,200 mg/day
|
N
|
63
|
17
|
46
|
Adverse events
|
34 (54.0%)
|
12 (70.6%)
|
22 (47.8%)
|
Digestive symptoms
|
28 (82.4%)
|
11 (91.7%)
|
17 (77.3%)
|
Rash
|
5 (14.7%)
|
1 (8.3%)
|
4 (18.2%)
|
Dizziness
|
1 (2.9%)
|
0
|
1 (4.5%)
|
Decision by the attending physician
|
23 (36.5%)
|
4 (23.5%)
|
19 (41.3%)
|
Poor adherence
|
5 (7.9%)
|
1 (5.9%)
|
4 (8.7%)
|
Patient’s preference
|
1 (1.6%)
|
0
|
1 (2.2%)
|
p = 0.430, the proportion of reasons among the groups stratified by the different doses of pirfenidone.
Table 3. Comparison of characteristics among patients treated with different doses of pirfenidone
Dose of pirfenidone
|
< 1,200 mg/day
|
1,200 mg/day
|
1,800 mg/day
|
p value
|
N
|
17 (20.0%)
|
46 (54.1%)
|
22 (25.9%)
|
|
Age at initiation of
Pirfenidone (years)
|
73.2 ± 1.8
|
75.3 ± 1.1†
|
69.8 ± 1.6
|
0.018§
|
The duration of treatment of pirfenidone (months)
|
11.1 ± 4.8
|
13.0 ± 2.9†
|
27.1 ± 4.3
|
0.015§
|
Male sex (%)
|
76.5%
|
80.4%
|
68.2%
|
0.54
|
Smoking index
|
447.5 ± 181.7
|
854.1 ± 107.2
|
764.0 ± 154.9
|
0.16
|
BMI (kg/m2)
|
20.9 ± 1.1
|
22.7 ± 0.6
|
21.9 ± 0.9
|
0.35
|
BSA (m2)
|
1.56 ± 0.04
|
1.54 ± 0.03
|
1.64 ± 0.04
|
0.12
|
Dose/BSA (mg/m2)
|
414.0 ± 24.2††‡‡
|
789.0 ± 14.9‡‡
|
1109.1 ± 21.3
|
<0.0001§§
|
UIP pattern in HRCT
|
70.6%
|
39.1%
|
40.9%
|
0.23
|
FVC (% predicted)
|
63.6 ± 5.6††
|
80.5 ± 3.6
|
78.8 ± 5.1
|
0.040§
|
DLCO (% predicted)
|
41.0 ± 5.8‡‡
|
67.2 ± 3.5
|
53.3 ± 5.1
|
<0.001§§
|
Alb (g/dL)
|
3.65 ± 0.11
|
3.86 ± 0.07
|
3.77 ± 0.10
|
0.27
|
KL-6 (U/mL)
|
1,190 ± 170
|
900 ± 103††
|
1,680 ± 150
|
<0.001§§
|
GAP stage (%)
I/ II/ III
|
21.4/ 42.9/ 35.7
|
53.7/ 24.4/ 22.0
|
66.7/ 22.2/ 11.1
|
0.13
|
Treatment with corticosteroids
|
17.7%
|
23.9%
|
36.4%
|
0.38
|
Alb: serum albumin; BMI: body mass index; BSA: body surface area calculated by the Du Bois formula; DLCO: diffusing capacity of the lung for carbon monoxide; FVC: forced vital capacity; GAP: Gender–Age–Physiology; HRCT: high-resolution computed tomography; KL-6: Krebs von den Lungen-6; UIP: usual interstitial pneumonia.
p < 0.05†, p < 0.01††: vs. 1,800 mg/day; p < 0.05‡, p < 0.01‡‡: vs. 1,200 mg/day;
p < 0.05§, p < 0.01§§: between groups by analysis of variance test.
The data at the time of pirfenidone treatment initiation is presented as number (%) or mean ± SEM. BMI and BSA: n=84, FVC: n=80, DLCO: n=65, Alb: n=82, GAP stage: n=73.
Next, we analyzed the contributing factor to the pirfenidone treatment dose of ≤1,200 mg/day (Table 4). The mean administered dose of pirfenidone in the ≤1,200 mg/day group was 1047.6 ± 255.8 mg/day (n = 63), and 73.0% of patients in this group were treated with 1,200 mg/day. In the univariate analysis, significant contributions of age at the time of pirfenidone treatment initiation, body surface area (BSA) and the KL-6 level were observed. Multivariate analysis was performed by factors with p < 0.20 in the univariate analysis and age at initiation of pirfenidone and the level of KL-6 also significantly contributed to the dose of ≤1,200 mg/day.
We further analyzed the contributing factors to the pirfenidone dose of <1,200 mg/day (Table 5). The mean administered dose of the group was 635.3 ± 78.6 mg/day, with 82.4% of the patients receiving 600 mg/day. In the univariate analysis, significant contributions of FVC, DLCO, usual interstitial pneumonia (UIP) pattern on HRCT, and GAP stage was observed. Multivariate analysis was performed by factors with p < 0.20 in the univariate analysis. As FVC, DLCO, and GAP stage were highly correlated with each other (p < 0.0001), three models were analyzed (Model 1: Table 5, Models 2 and 3: Additional Table 1). Serum albumin (Alb) significantly (Model 1 and 2, p < 0.05) and tended to (Model 3, p = 0.20) contribute to the pirfenidone dose of < 1,200 mg/day. DLCO significantly (Model 1, p < 0.05) and GAP stage (Model 2, p = 0.27) and FVC (Model 3, p = 0.16) tended to contribute to a low dose of pirfenidone. These results showed that hypoalbuminemia and pulmonary dysfunction were related to <1,200 mg/day of pirfenidone treatment.
Table 4. The contributing factors of low-dose pirfenidone of ≤1,200 mg/day
|
Univariate analysis
|
Multivariate analysis
|
|
HR
|
95% CI
|
p
|
HR
|
95% CI
|
p
|
Age at initiation
of pirfenidone (years)
|
1.095
|
1.02–1.18
|
<0.01**
|
1.14
|
1.038–1.261
|
<0.01**
|
Treatment with
corticosteroids
|
0.5
|
0.17–1.43
|
0.2
|
|
|
|
BMI (kg/m2)
|
1.017
|
0.91–1.14
|
0.77
|
|
|
|
BSA (m2)
|
0.059
|
0.004-0.961
|
0.039*
|
0.39
|
0.006-25.5
|
0.655
|
FVC
(% predicted)
|
0.994
|
0.97–1.02
|
0.599
|
|
|
|
DLCO
(% predicted)
|
1.013
|
0.988–1.039
|
0.29
|
|
|
|
UIP pattern on
HRCT
|
1.31
|
0.49–3.51
|
0.586
|
|
|
|
Alb (g/dL)
|
1.198
|
0.401–3.58
|
0.747
|
|
|
|
KL-6 (U/mL)
|
0.999
|
0.998–1.000
|
<0.01**
|
0.999
|
0.998–1.000
|
0.017*
|
GAP stage
|
1.85
|
0.87–3.95
|
0.092
|
1.80
|
0.67–4.86
|
0.231
|
The result of the multivariate analysis is composed of Age, BSA, KL-6, and GAP stage. Alb: albumin; BMI: body mass index; BSA: body surface area calculated by the Du Bois formula; CI: confidence intervals; DLCO: diffusing capacity of the lung for carbon monoxide; FVC: forced vital capacity; GAP: Gender–Age–Physiology; KL-6: Krebs von den Lungen-6; UIP: usual interstitial pneumonia. p < 0.05*, p < 0.01**. BMI and BSA: n=84, FVC: n=80, DLCO: n=65, Alb: n=82, GAP stage: n=73.
Table 5. The contributing factors of a pirfenidone dose of <1,200 mg/day
|
Univariate analysis
|
Multivariate analysis (Model 1)
|
|
HR
|
95% CI
|
p
|
HR
|
95% CI
|
p
|
Age at initiation of pirfenidone (years)
|
0.995
|
0.928–1.066
|
0.88
|
|
|
|
Treatment with corticosteroids
|
0.553
|
0.143–2.142
|
0.37
|
|
|
|
BMI (kg/ m2)
|
0.912
|
0.791–1.051
|
0.19
|
1.120
|
0.891–1.409
|
0.33
|
BSA (m2)
|
0.694
|
0.038–12.60
|
0.80
|
|
|
|
FVC
(% predicted)
|
0.968
|
0.942–0.994
|
<0.01**
|
|
|
|
DLCO
(% predicted)
|
0.951
|
0.919–0.985
|
<0.01**
|
0.963
|
0.923–1.005
|
0.048*
|
UIP pattern on HRCT
|
3.644
|
1.153–11.52
|
0.021*
|
3.036
|
0.478–19.30
|
0.24
|
Alb (g/dL)
|
0.420
|
0.128–1.381
|
0.15
|
0.095
|
0.010–0.875
|
0.022*
|
KL-6 (U/mL)
|
1.000
|
0.999–1.001
|
0.86
|
|
|
|
GAP stage
|
2.226
|
1.074–4.616
|
0.028*
|
|
|
|
The result of the multivariate analysis shows that Model 1 is composed of BMI, DLCO, UIP pattern, and Alb. Alb: albumin; BMI: body mass index; BSA: body surface area calculated by the Du Bois formula; CI: confidence intervals; DLCO: diffusing capacity of the lung for carbon monoxide; FVC: forced vital capacity; GAP: Gender–Age–Physiology; KL-6: Krebs von den Lungen-6; UIP: usual interstitial pneumonia. p < 0.05*, p < 0.01**. BMI and BSA n=84, FVC n=80, DLCO n=65, Alb n=82, GAP stage n=73, other n=85.
Next, we analyzed the relationship between OS and treatment doses of pirfenidone. The patients’ OS is displayed in Figure 2A. The MST was 66.3 months, and the 2- and 5-year survival rates were 65.0% and 56.0%, respectively. The difference in the survival rate stratified by the dose of pirfenidone was investigated (Figure 2B). Although there was no significant difference between the 1,200 and 1,800 mg/day groups, survival was significantly shorter in the <1,200 mg/day group than in the 1,800 mg/day group (p < 0.01). MST was 16.7 months, 66.5 months, and unreachable in <1,200, 1,200, and 1,800 mg/day groups, respectively.
Finally, we investigated the factors related to OS (Table 6). A pirfenidone dose of <1,200 mg/day, low FVC, DLCO, and high GAP stage were significantly associated with a poor prognosis in the univariate analysis. Significant factors in the univariate analysis were selected for the multivariate analysis. Since FVC, DLCO, and GAP stage were highly correlated with each other (p < 0.001), three models were investigated. In Model 1 (Table 6), the contributions of FVC and pirfenidone dose of <1,200 mg/day were significantly associated with the prognosis. In other models (Additional Table 2), a pirfenidone dose of <1,200 mg/day also showed HR >2.0 (Models 2 and 3), and the GAP stage significantly related to the prognosis in Model 2. These results showed that a pirfenidone dose of <1,200 mg/day and presence of pulmonary dysfunction were related to a poor prognosis.
Table 6. Prognostic factors in overall patients
|
Univariate analysis
|
Multivariate analysis (Model 1)
|
|
HR
|
95% CI
|
p
|
HR
|
95% CI
|
p
|
Male sex
|
1.625
|
0.628–4.752
|
0.33
|
|
|
|
Age at initiation of pirfenidone (years)
|
1.024
|
0.963–1.089
|
0.45
|
|
|
|
Smoking Index
|
1.000
|
0.999–1.000
|
0.46
|
|
|
|
Pirfenidone dose <1,200 mg/day
|
2.803
|
1.024–7.184
|
0.045*
|
3.117
|
1.079–8.814
|
0.036*
|
Treatment with corticosteroids
|
1.780
|
0.656–4.435
|
0.23
|
|
|
|
BMI (kg/ m2)
|
0.952
|
0.858–1.045
|
0.31
|
|
|
|
BSA ( m2)
|
0.416
|
0.032–4.627
|
0.49
|
|
|
|
FVC
(% predicted)
|
0.957
|
0.929–0.982
|
<0.01**
|
0.952
|
0.919–0.981
|
<0.01**
|
DLCO
(% predicted)
|
0.971
|
0.940–1.000
|
0.046*
|
|
|
|
UIP pattern
|
1.602
|
0.652–4.286
|
0.62
|
|
|
|
Alb (g/dL)
|
0.622
|
0.244–1.665
|
0.33
|
|
|
|
KL-6 (U/mL)
|
1.000
|
0.999–1.000
|
0.45
|
|
|
|
GAP stage
|
3.093
|
1.511–7.099
|
<0.01**
|
|
|
|
The prognostic factors in the overall patients were analyzed by Cox proportional hazard analysis. Model 1 was composed of pirfenidone at dose of <1,200 mg/day and FVC. Alb: albumin; BMI: body mass index; BSA: body surface area calculated by the Du Bois formula; CI: confidence intervals; DLCO: diffusing capacity of the lung for carbon monoxide; FVC: forced vital capacity; GAP: Gender–Age–Physiology; HR: hazard ratio; KL-6: Krebs von den Lungen-6; UIP: usual interstitial pneumonia. p < 0.05*, p < 0.01**. BMI and BSA: n=84, FVC: n=80, DLCO: n=65, Alb: n=82, GAP stage: n=73.