Table 5:

Ordinal logistic regression analysis for dyslipidemia and serum vitamin D level groups

Model 1	Estimate	Standard error	Wald	Odd ratio (95% confidence interval)	P-value*
Vitamin D status 1	1.97	1.36	2.10	7.15 (0.50-102.01)	0.15
Vitamin D status 2	3.47	1.40	6.16	32.21 (2.08-499.97)	0.01
Dyslipidemia  Normal lipids (Ref.)	-0.54	0.62	0.74	0.59 (0.17-1.99)	0.39
Age	0.03	0.02	1.27	1.03 (0.98-1.07)	0.26
Male gender  Female gender (Ref.)	-013	0.51	0.07	0.88 (0.33-2.36)	0.79
Education  Below high school  High school or above (Ref.)	0.11	0.49	0.06	1.12 (0.43-2.90)	0.82
Model 2
Vitamin D status 1	-3.44	1.93	3.18	0.03 (0.00-1.41)	0.08
Vitamin D status 2	-1.84	1.93	0.91	0.16 (0.00-6.96)	0.34
Dyslipidemia  Normal lipids (Ref.)	-0.51	0.67	0.58	0.60 (0.16-2.24)	0.45
Smoking  Non-smoking (Ref.)	-0.16	0.58	0.08	0.85 (0.27-2.64)	0.78
HbA1c	-064	0.33	3.84	0.53 (0.28-1.00)	0.05

* Ordinal logistic regression analysis (ordinal dependent variable was vitamin D status: deficient, insufficient and sufficient vitamin D levels in order). We used two models of logistic regression using two groups of co-variables. Model 1 included age, gender, education level and type of lipid as co-variables while model 2 included smoking, HbA1c and type of lipid as co-variables. P-values < 0.05 were considered statistically significant. TGs; triglycerides, HDL-C; high density lipoprotein cholesterol, LDL-C; low density lipoprotein cholesterol, HbA1c; hemoglobin A1c.