Frail older adults are at increased risk of poor clinical outcomes. Frailty assessment is therefore important in clinical trials to understand the benefits and harms of interventions. However, consensus is lacking on how frailty should be assessed.We developed a prospectively specified index using a battery of formal tests and instruments and a retrospectively generated index using medical comorbidities and patient reported outcomes (PROs) within an adjuvanted recombinant zoster vaccine (RZV) trial (NCT02979639). For both frailty indices (FIs), a total deficit score was calculated as the accumulation of deficits and participants were categorized as non-frail, pre-frail and frail. We assessed (1) the feasibility and validity of both FIs; (2) the impact of RZV vaccine reactogenicity by frailty status on Short Form-36 [SF-36] physical functioning (PF) scores.Of 401 participants, aged ≥50 years, 236 (58.9%) were categorized non-frail, 143 (35.7%), pre-frail, and 22 (5.5%) frail using the prospective FI. Corresponding numbers for the retrospective FI were 192 (47.9%), 169 (42.1%) and 40 (10.0%), respectively. Strong concordance was observed between the frailty status assessments (P < .001). The proportion defined as frail increased from 1.5%, to 10.4% in participants aged 50-59, and ≥70 years, respectively, for the prospective FI. Corresponding numbers for the retrospective FI were 3.7%, and 17.2%, respectively. RZV vaccination was associated with a transient, non-clinically meaningful, decrease on the SF-36 PF score in frail participants.Both frailty indices provided similar results. The retrospectively generated FI offers the advantage of being easier to incorporate into vaccine clinical trials of older adults.