Figure 3.

Adjusted incidence rate ratios (95% confidence intervals [CIs]) of all-cause, cardiovascular, and infection-related hospitalizations according to baseline smoking status. The reference group is never-smokers. Multivariable estimates were adjusted for age, sex, race (black versus nonblack), HEMO Study Kt/V and flux group assignments, postdialysis weight, sex-by-postdialysis weight cross-product terms, diabetic status, access (fistula, graft, catheter), urine volume (≤200 ml/d, >200 ml/d), dialysis session length (≤180, 181–209, 210–239, ≥240 minutes), serum sodium level, and alcohol and drug abuse (never, former, current). Associations with all-cause mortality were estimated by negative binomial regression; associations with cardiovascular and infection-related hospitalization were estimated by Poisson regression.