Table 1 presents the clinical variables of the subjects, and they were compared among the different groups. Except for current smoking, current drinking and AST levels, all variables, including SBP, DBP, ALT, BMI, ALP, γ-GT, BUN, SCR, UA, FBG, HGB, TC, and TG levels, exhibited significant variations among all groups (P < 0.05). Subjects in Group 2 had significantly decreased ALT, γ-GT, BUN, FBG, HGB, TC, and TG levels compared to the other groups (P < 0.01). Subjects in Group 3 had significantly increased ALT, γ-GT, BUN, SCR, HGB, TC, and TG levels compared to the other groups (P < 0.05). In Group 2, serum ALP and UA levels were significantly higher than those in the other groups (P < 0.05).

Correlation analysis was conducted for DBil, IBil, and lipid levels, and Table 2 lists the results. Significant negative associations were found between DBil and TC and TG levels (P < 0.01). The correlation coefficient outcomes were -0.291 and -0.299, respectively. Significant positive associations were observed between IBil and TC (r = 0.360) and TG (r = 0.117) levels (P < 0.01). Additionally, a partial correlation analysis was performed, controlling for DBil and IBil levels as covariates (Table 3). The results showed a more obvious negative association between DBil and TC levels (r = -0.777, P < 0.01), while the association between DBil and TG levels did not change appreciably (r = -0.397, P < 0.01). More obvious positive correlations were found between IBil and TC (r = 0.790, P < 0.01) and TG (r = 0.302, P < 0.01) levels.

The patients with abnormal laboratory variables were analysed in each group, and significant differences were found in four laboratory variables, including TC, TG, HGB and BUN levels (Table 4). The number of patients with abnormal TC, TG, HGB and BUN levels was the lowest in Group 2, and the frequencies were 0.6%, 1.2%, 34.7% and 0.6%, respectively. The number of patients with abnormal TC, TG, HGB and BUN levels was the highest in Group 3, and the frequencies were 17.4%, 14.8%, 52.5% and 4.9%, respectively. The variations among all groups were significant (P < 0.05). No significant variations of other abnormal laboratory variables were found among the three groups. Further analysis showed mildly elevated DBil levels were associated with decreased TC and TG levels, with ORs (95% CIs) of 0.097 (0.013–0.700) and 0.112 (0.027–0.458), respectively (P < 0.01), and mildly elevated IBil levels were associated with increased TC and TG levels, with ORs (95% CIs) of 3.436 (2.398–4.924) and 1.636 (1.163–2.303), respectively (P < 0.01) (Table 5). The results of univariate logistic analysis confirmed the association of bilirubin with TC and TG levels (Table 6). DBil levels were negatively associated with TC and TG levels, with ORs (95% CIs) of 0.685 (0.600–0.782) and 0.592 (0.552–0.671), respectively. IBil levels were positively associated with TC and TG levels, with ORs (95% CIs) of 1.298 (1.228–1.373) and 1.089 (1.037–1.143), respectively. After adjusting for age, BMI, smoking status, drinking status, SBP, DBP, HGB, ALT, AST, ALP, γ-GT, BUN, SCR, UA and FBG levels, DBil levels were an independent protective factor against high TC (OR = 0.702, 95% CI 0.602–0.817, P < 0.01) and TG (OR = 0.632, 95% CI 0.541–0.739, P < 0.01) levels, and IBil levels were an independent risk factor for high TC levels, with an OR (95% CI) of 1.251 (1,176–1.331).