Skip to main content
. 2018 Feb 20;19:31. doi: 10.1186/s12931-018-0734-y

Table 3.

Multiple regression analysis predicting FEV1%, TLCO%, RV/TLC, 6MWD, desaturation on exertion, CRP, fibrinogen and sputum neutrophils

B coefficient Standardised B coefficient significance
FEV1%
 %LAA<− 950 − 0.42 − 0.34 < 0.001
 E/I MLD − 180.8 − 0.56 < 0.001
 BMI −0.56 − 0.20 0.010
TLCO%
 %LAA<− 950 −0.62 −0.37 < 0.001
 FEV1% 0.49 −0.39 < 0.001
 Gender −7.5 −0.19 0.014
RV/TLC
 E/I MLD 0.75 0.30 0.003
 FEV1% −0.003 −0.32 0.002
 Gender 0.07 0.28 < 0.001
6MWD (metres)
 FEV1% 3.03 0.41 < 0.001
 BMI −5.19 −0.26 0.003
 Gender −38.2 −0.17 0.044
Desaturation on exertion (%)
 %LAA<−950 0.041 0.73 < 0.001
 BMI 0.032 0.248 0.004
 Gender 0.233 0.164 0.037
CRP (mg/L)
 %LAA<− 950 −0.026 −0.34 < 0.001
Fibrinogen (g/L)
 %LAA<−950 −0.020 −0.28 0.003
 Age 0.022 0.216 0.020
Sputum Neutrophils (%)
 Current smoker −21.1 −0.29 0.018

%LAA<−950, E/I MLD, PI10, bronchiectasis, FEV1% and demographics variables (age, gender, current smoking status, pack years and BMI) were combined into a regression model to predict each of the dependent variables. Only variables that made a significant difference to the model are included in the results. The natural log of CRP and desaturation on exertion (+ 1) were used for this analysis to normalise the distribution of the residuals