Table 3:

Univariate and Logistic Regression analysis. Conditional logistic regression was done for COPD and Control groups, which was matched for age and gender. Unconditional Logistic Regression was done for risk of exacerbation in subjects with COPD.

	COPD (Control vs COPD)				Exacerbation of COPD in previous year (No vs Yes)
	Univariate		Conditional Logistic Regression		Univariate		Unconditional Logistic Regression
Variable	OR (95% CI)	P-value	Adj OR (95% CI)	P-value	OR (95% CI)	P-value	Adj OR (95% CI)	P-value
Smoking Habits (180 males):	Ref Non-smokers
Ex-smokers	8 (1.7–48.4)	0.003#	59.1 (2.8–1260)	0.01	1.1 (0.05–70)	0.999#	8×10^9$	0.999
Current smokers	3.48 (0.86–20.1)	0.058#	16.5 (1.1–248)	0.043	1.53 (0.07–94)	0.999#	18×10^8$	0.999
Occupation:	Ref Non Heavy Work
Heavy Work	1.15 (0.5–2.41)	0.86	4.26 (0.53–34)	0.17	1.98 (0.67–5.9)	0.3	2.87 (0.73–11.22)	0.131
Charlson’s comorbidity index:	Ref Score 0–1
Score 2–4	1.38 (0.79–2.41)	0.32	3.8 (0.5–28)	0.19	1.69 (0.75–3.78)	0.21	1.49 (0.6–3.8)	0.41
Standard of living index (SLI)	Ref >=20
<20	0.7 (0.39–1.22)	0.26	0.5 (0.15–28)	0.3	0.48 (0.21–1.1)	0.08	0.51 (0.21–1.27)	0.15
Body Mass Index (BMI):	Ref Normal
Underweight	6.23 (2.65–14.65)	<0.001	2.71 (0.65–11.2)	0.17	0.8 (0.32–1.96)	0.7	0.68 (0.23–1.95)	0.47
Overweight	1.02 (0.5–2.09)	0.95	0.39 (0.08–1.7))	0.22	1.35 (0.45–4.1)	0.6	1.44 (0.43–4.82)	0.56
Vitamin D levels:	Ref >20 ng/ml				Ref >=19 ng/mL
<=20 ng/mL (deficiency)	3.09 (1.68–5.68)	<0.001	5.05 (1.4–17.8)	0.012	5.38 (2.06–14.1)	<0.001	3.51 (1.27–9.67)	0.016
Six Minute Walk Distance, % predicted	ref ≥80%
<80	17.67 (7.1–44.1)	<0.001	17.12 (3.8–76.8)	<0.001	undefined	0.039	350000 (0–10^12)$	0.97

^#Fisher exacts values.

^$the values are >100,000