Assessing the effect of oral activated vitamin D on overall survival in hemodialysis patients: a landmark analysis

Table 4 Multivariate Cox proportional hazard models examining activated vitamin D treatment as compared with no treatment by landmark time in hospital-setting hemodialysis patients

Model	HR (95% CI)
Unadjusted	0.69 (0.66–0.73)
Adjusted
Urbanization and hospital level	0.72 (0.68–0.76)
Age and sex	0.78 (0.74–0.82)
Age, sex, urbanization, and hospital level	0.80 (0.76–0.84)
Age, sex, vascular access type, and comorbidities	0.89 (0.85–0.94)
Age, sex, comorbidities, and baseline medications	0.90 (0.85–0.95)
Age, sex, urbanization, hospital level, vascular access, comorbidities, and baseline medications	0.91 (0.87–0.96)
Propensity score (PS) method
PS trimming (1–99%)	0.70 (0.67–0.74)
PS trimming + IPTW	0.95 (0.92–0.97)
PS matching (1: 3)	0.95 (0.89–1.00)

Propensity score (PS): PS was calculated with logistic regression using covariates of age, sex, vascular access type, baseline comorbidities, medications, and levels of hospital and urbanization. The PS matched methods was employed compared vitamin D users versus non-users without further adjustment of baseline covariates
Abbreviations: HR hazard ratio, CI confidence intervals, PS propensity score, IPTW inverse probability treatment weighting

ISSN: 1471-2369