Rakusin 1972.

Methods	Study design: Quasi‐randomized controlled series. Exclusions after allocation: 59 patients in the series with large hyphemas underwent surgery and were not included in the analysis. Losses to follow‐up: 20. Intention‐to‐treat: Not all participants were accounted for in the final analyses, thus intention‐to‐treat analysis was not performed. Sample size calculations: Not reported.
Participants	Country: South Africa. Dates: 1966‐69. Number allocated: 390 consecutive patients. Age: Not reported. Sex: Not reported. Sickle cell disease: Not reported. Race: 90% African origin, 10% Asiatic origin. Inclusion criteria: Traumatic hyphema. Exclusion criteria: Surgical treatment indicated.
Interventions	Series of comparisons based on 6 variable factors: Bed rest (n = 26) vs ambulatory treatment (n = 26) Eye pads: bilateral eye pads (n = 27) vs single eye pads (n = 26) vs no eye pads (n = 10) Topical antibiotics (0.5% chloramphenicol, n = 21) vs corticosteroids (0.5% hydrocortisone acetate, n = 13) vs neither (n = 3) Mydriatics (1% homatropine, n = 17) vs miotics (4% pilocarpine, n = 17) vs neither (n = 19) vs both (n = 17) Enzymes: oral trypsin (n = 15) vs oral papase (n = 18) vs neither (n = 10) Ocular hypotensive agents: acetazolamide 250 mg (n = 18) vs oral glycerol 1 mL/kg (n = 18) vs neither (n = 10) Treatment and control groups followed the same regimen, except even‐numbered participants received the variable factor, and odd‐numbered participants did not. Excluding the variable factor for each series, all participants received bed rest, single pad over the injured eye, and topical chloramphenicol.
Outcomes	Primary outcomes: Speed of absorption of blood from the anterior chamber Risk of secondary hemorrhage Complications of the hyphema Final VA Follow‐up: Range 1 to 2 weeks to 3 years.
Notes	Funded by the University of Witwatersrand, the South African Medical Research Council, Leo Laboratories, Mer‐National, and Warner Pharmaceutical Co. In the third comparison group, antibiotics versus corticosteroids, 3 participants were assigned to receive neither treatment, but this group was discontinued after all 3 participants developed a mucous conjunctival discharge.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Method of allocation unclear, not all patients in the series were allocated to the 6 comparisons under study; 59 patients were selected for surgery. Also, even and odd patient number allocation is not applicable to comparison with 3 treatment groups.
Allocation concealment (selection bias)	High risk	Method of allocation concealment not reported, not randomized.
Blinding (performance bias and detection bias) Participants	High risk	Masking of participants was not possible for some variables (i.e. bed rest and eye patching). Use of placebo for other variables was not mentioned.
Blinding (performance bias and detection bias) Personnel and outcome assessors	Unclear risk	Masking was not reported.
Incomplete outcome data (attrition bias) Primary outcome	Unclear risk	79 participants were not included in the analyses, and the intention‐to‐treat principle was not followed.
Incomplete outcome data (attrition bias) Secondary outcomes	Unclear risk	79 participants were not included in the analyses, and the intention‐to‐treat principle was not followed.
Selective reporting (reporting bias)	Low risk	Reported results for primary and secondary outcomes
Other bias	Unclear risk	The primary interventions of interest for this study were not clear. Although the majority of the patients in the series were assigned to 1 of 6 conservative‐treatment comparison groups, 59 recruited patients were selected for surgery.