Guidozzi 1999.

Methods	Randomized controlled trial
Participants	Country: South Africa Setting: Department of Obstetrics and Gynecology, Johannesburg Hospital and Medical School of the University of the Witwatersrand, Johannesburg, South Africa. Sealed opaque envelopes were used to random the participants in equal number of HRT and no HRT, from January 1987 to June 1994. Baseline characteristics: of the 130 participants included, there were participants who were lost to follow‐up: three participants in the HRT and two participants in the no HRT groups, therefore the total number of participants who were analysed were 59 participants in the HRT and 66 participants in the no HRT groups. The participants aged less than 59 years old were included. Most participants were 56 to 59 years old (44.0%), of serous type (68.0%), grade 1 (42.4%), FIGO stage III (67.2%), and had undergone optimal surgery (72.0%). Inclusion criteria: all participants were younger than 59 years old with invasive EOC treated with total abdominal hysterectomy, bilateral salpingo‐oophorectomy, omentectomy, and tumour debulking, following chemotherapy 6 cycles of cisplatin 100 mg/m2 and cyclophosphamide 500 mg/m2, then 2 cycles of cisplatin only and oral chlorambucil for 1 year thereafter. Exclusion criteria: participants with ovarian carcinoma of low malignant potential and those who had ever taken conjugated oestrogens were excluded.
Interventions	Intervention characteristics: HRT Conjugated oestrogens 0.625 mg daily: 59 patients Control: no HRT (N=66) There were no data on length of time participants received HRT.
Outcomes	The median follow‐up time was 90 months. There were participants who were assigned to HRT but did not receive or discontinued HRT within 9 months (9 participants (13.6%)) and participants who were assigned to no HRT but received HRT (five participants (8.5%)). However, all analyses were performed on an intention‐to‐treat basis. Death rate: 73 participants (58.4%) died: 32 (54%) in the HRT group and 41 (62%) participants in the no HRT group. The median overall survival was 44 months in the HRT group and 34 months in the no HRT group. Recurrent rate: 73 participants (58.4%) had recurrence: 32 (54%) in the HRT group and 41 (62%) in the no HRT group. The median disease‐free interval was 34 months in the HRT group and 27 months in the no HRT group.
Notes	Sponsorship source: there were no data of funding source and conflict of interest.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "All participants were randomised at the routine assessment consultation held 6 – 8 weeks postoperatively." Comment: no randomization method was identified.
Allocation concealment (selection bias)	Low risk	Quote: "Randomization involved participant choice of a sealed opaque envelope from a predetermined equal number of similarly sealed opaque envelopes that contained directions for the randomisation to either continuous conjugated equine oestrogen replacement (ERT), consisting of Premarin 0.625 mg daily (Wyeth‐ Ayerst, Philadelphia, PA), or no supplementation (non‐ERT)."
Blinding of participants and personnel (performance bias) All outcomes	High risk	Placebo tablets were not used, and prior to randomization all participants were fully counselled about the risks and benefits of oestrogen replacement as well as the aims and limitations of the study.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	No information who were outcome assessor and whether they knew the treatment group.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Quote: "One hundred thirty participants met the criteria and were randomised into their respective groups: 62 to post‐operative ERT and 68 to non‐ERT. Three participants in ERT group and two participants in non‐ERT group were lost to follow‐up, so that the final analysis involved 59 participants in ERT group and 66 participants in non‐ERT."
Selective reporting (reporting bias)	Low risk	Quote: "For the purposes of analysis, the participants remained in the treatment group to which they were originally allocated, irrespective of whether they elected to commence, stop taking, or refuse ERT, i.e., intention‐to‐treat analyses are reported."
Other bias	Low risk	We do not suspect any other source of bias.