Characteristics of Study Subjects
Demographics and clinical characteristics of the studied subjects are shown in Table 1. Comorbidities are shown separately in Supplement Table 1 (Table S1). Follow up time of IPF patients was 7.5±4.5 yrs. Forty-five patients (72.5%) received antifibrotic treatment, 15 (24%) received N-acetylcysteine (NAC) monotherapy and 2 no treatment over time. Sixteen patients (26%) died and 12 (19%) underwent lung transplantation during follow-up. 19 (31%) patients experienced acute exacerbation. Median survival time, defined as time to death or lung transplantation, was 5.9 ± 3.2 years.
MUC5B and TOLLIP SNPs allele and genotypes distribution
Allele and genotype distribution for each SNP in IPF and HC is shown In Table 2. MUC5B rs35705950, TOLLIP rs5743890 and rs3750920 alleles were all in HWE. The frequency of MUC5B rs35705950 minor allele (T) was significantly different between IPF patients and HC (p < 0.001). For TOLLIP rs5743890 and rs3750920 the frequency of minor alleles did not differ between IPF and HC (Table 2). Genotype distribution did not differ significantly between IPF and HC for the three SNPs studied.
Four major haplotypes with three polymorphic sites were established. The haplotype TTT occurred more frequently and GTT less frequently in IPF patients than HC (Table S2).
Measures of pair-wise linkage disequilibrium between MUC5B and TOLLIP SNPs are shown in Table S3. MUC5B rs35705950 and TOLLIP rs5743890 were in high linkage disequilibrium (D´-1, R2 0.038, p=0.0035).
Correlations between SNPs and clinical characteristics
None of the SNP genotypes was associated with gender, BMI, smoking habits, lung function, BAL cell differential or comorbidities at baseline (data not shown). No difference in the frequency of antifibrotic or NAC treatment was observed between genotypes.
MUC5B und TOLLIP SNPs and pulmonary dysfunction at baseline and over time
Lung function tests at baseline in patients with TOLLIP rs5743890 C/T genotype (n=14) did not differ from those in patients with T/T genotype (n= 48) (Table S4). Patients with TOLLIP rs5743890 C/T genotype tended to have a greater decline in FVC, DLco % pred. and PaO2 mmHg per year compared to those with T/T genotype (Table S4). No differences were seen in lung function tests at baseline or decline over time according to MUC5B and TOLLIP rs3750920 genotype (data not shown).
Correlation between MUC5B and TOLLIP genotype and survival
The frequency of death or lung transplantation did not vary across MUC5B and TOLLIP genotypes. Kaplan-Meier analysis for MUC5B and TOLLIP rs3750920 did not show any correlation with survival, and median survival time did not vary according to genotype (Fig 1a and 1c). Patients with TOLLIP rs5743890 C/T genotype had worse survival compared to T/T genotype, T being the major allele (p=0.014) (Fig 1b, Table S4).
Correlation between MUC5B and TOLLIP genotype and disease progression
Disease progression as defined in the methods occurred in a total of 37 (62%) of 60 patients where data were available. Median time to progression was 50 months (IQR: 34-68). No association was observed between MUC5B and TOLLIP rs3750920 genotype and disease progression.
The frequency of disease progression significantly varied across TOLLIP rs5743890 genotype: all patients with C/T genotype had disease progression compared to 50% of patients carrying T/T (p=0.001) (Tab S5) and median time to progression tended to be shorter in patients with the C/T genotype, 43.5±23 months vs 63±27 months compared to those with the T/T genotype (p=0.058). Kaplan Kaplan-Meier analysis of disease progression according to TOLLIP rs5743890 genotype in IPF patients is shown in Figure 2a.
Correlation between MUC5B and TOLLIP SNP genotype and acute exacerbation
Acute exacerbation occurred in 19/62 (31%) patients. Median time to first acute exacerbation was 47 months (IQR: 38-73). 15 patients had AE after disease progression had started and 4 patients during a stable disease course. We did not analyze this small subgroup separately.
No association between MUC5B or TOLLIP genotype and frequency of acute exacerbation was seen (Table S5). Whereas Kaplan-Meier analysis did not show any significant association between AE and TOLLIP rs5743890 genotype (Log rank p=0.326), a tendency for an unfavorable effect of the minor allele can be observed (Figure 2b).
Uni and multivariate analysis for predictors of survival, disease progression and AE
We performed univariate and multivariate analysis by using Cox regression to investigate the role of several factors as predictors for survival and disease outcome. We did not find any association between MUC5B or TOLLIP rs3750920 and survival or AE (Table 3 and S6). Disease progression was found to be a strong predictor of AE in the univariate analysis (HR 4.627, CI: 0.105-20.175, p=0.041) but this association was not confirmed in the multivariate analysis (Table S6). In the uni- and multivariate analysis for identifying predictors of death or lung transplantation, TOLLIP rs5743890 (C/T genotype) was the strongest one, followed by Dlco % pred., also after including a number of covariates in the model (HR, 3.434, 95% CI 1.360-8.671, p=0.009) (Table 3).
The results of uni- and multivariate analysis for predictors of disease progression are shown in Table 4. In the univariate analysis, FVC % pred. and TOLLIP rs5743890 C/T genotype were the strongest predictors of disease progression (Table 3) but in the multivariate analysis only TOLLIP rs5743890 C/T genotype was identified as a predictor of disease progression after including age, gender, BMI, smoking status and lung function at baseline as covariates in the multivariate model (HR 2.949, 95% CI 1.327-6.556, p = 0.008) (Table 3).