Aims: There is a trend toward deferring the initiation of chronic dialysis until absolutely indicated. This strategy, however, might lead to increased uncertainties in the timing of dialysis access creation prior to dialysis onset for patients approaching end-stage renal disease (ESRD), and the impact of which on hard end points remains largely unclear. We hereby investigated the effect of varied patterns of dialysis initiation on outcomes of new-onset hemodialysis (HD) patients. Methods: Four hundred sixty-two prospectively recruited patients were stratified into planned elective (n = 117, 25%), planned urgent (n = 65, 14%) or unplanned urgent (n = 280, 61%) starters based on the timing of access creation with respect to dialysis initiation. The outcome measures were all-cause mortality, hospitalization and access reconstruction over 2 years. Results: The mean estimated glomerular filtration rate (eGFR) was higher in the planned elective than in the planned urgent or unplanned urgent starters at access creation (5.3 vs. 4.4 or 4.3 ml/min/1.73 m2), but not at dialysis initiation (4.2 vs. 3.9 or 4.3 ml/min/1.73 m2). During the follow-up, the planned elective population exhibited the lowest rates of overall mortality and hospitalization, but not access reconstruction. Multivariate Cox's regression analysis showed that the planned urgent and the unplanned urgent groups, comparing to the planned elective population, displayed a greater risk of early death (hazards ratio [HR] 3.324, 95% CI 1.409-7.840; HR 2.510, 95% CI 1.177-5.355, respectively) and early hospitalization (sub-hazards ratio [SubHR] 2.238, 95% CI 1.530-3.274; SubHR 1.529, 95% CI 1.096-2.133, respectively). Conclusion: Incident ESRD patients undergoing planned elective start of HD, compared to their planned or unplanned urgent counterparts, showed reduced risk of overall mortality and hospitalization in the first 2 years after commencing long-term dialysis at a mean eGFR <5 ml/min/1.73 m2.