ATBC 1998.

Methods	Method of allocation: random. Sponsor provided coded capsules. Masking: participant: yes; provider: yes; outcome: yes Exclusions after randomisation: no Losses to follow‐up: 31%. Random sample for maculopathy study: 9%. 2 x 2 factorial design. Maculopathy add‐on random sample in 2 regions.
Participants	Country: Finland Number of participants randomised: 29,133. Random sample of 1035 selected for maculopathy study. Age: 50 to 69 years in 1984. Maculopathy study 1992‐3 in people aged 65 plus. Sex: male Inclusion criteria: 5 or more cigarettes daily Exclusion criteria: history of cancer or serious disease limiting ability to participate; those taking supplements vitamin E, A, or beta‐carotene in excess of predefined doses; those treated with anticoagulants
Interventions	Intervention: alpha‐tocopherol (50 mg/day) N = 237 beta‐carotene (20 mg/day) N = 234 alpha‐tocopherol (50 mg/day) and beta‐carotene (20 mg/day) N = 257 Comparator: placebo N = 213 Duration: 5 to 8 years (median 6.1)
Outcomes	AMD: 4 grades: Grade I: dry maculopathy with hard drusen, pigmentary changes, or both Grade II: soft macular drusen Grade III: disciform degeneration Grade IV: geographic atrophy Quote "A person was considered to have ARM if he had a class I or higher change in either eye, and severity was classified according to the worst eye."
Notes	Compliance with treatment excellent; 4/5 active participants took more than 95% of scheduled capsules. Drop‐out rate and compliance similar between all 4 groups. Funding source: Quote "This study was supported by the Juho Vainio Foundation, Helsinki, Finland. The ATBC study was supported by Public Health Service Contract N01‐CN‐45165 from the Division of Cancer Prevention and Control, National Cancer Institute of the United States." Declarations of interest: NR Date study conducted: Quote "The ophthalmological examination took place during their final follow‐up trial visit between December 1992 and March 1993" Trial id: NCT00342992
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote "The participants were randomly assigned to one of four treatment groups: AT alone, AT and BC, BC alone, or placebo in a complete 2 x 2 factorial design" and "Randomization was performed in blocks of eight within each of the study areas."
Allocation concealment (selection bias)	Low risk	Quote "A coded reserve supply of capsule packs..." Not clearly stated that allocation concealed, but the study was described as being "double‐blind"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Judgement comment: placebo‐controlled study.
Blinding of outcome assessment (detection bias) AMD	Low risk	Quote "The retinal specialist [...] examined six photographs (three per eye) of each participant without knowledge of the subject's treatment group"
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote "A total of 941 persons participated (91%) and non‐participation rates were similar across the intervention groups."
Selective reporting (reporting bias)	Unclear risk	Visual acuity measured but not reported, but as the main results for AMD showed no difference between groups, it is not clear whether this was an example of selective reporting or whether, in fact, the investigators considered that visual acuity in this age group might be attributed to a variety of causes, and therefore, was not a relevant outcome.