Jensen 2009.

Methods	Case‐cohort.
Participants	The cohort comprised 54,449 women with subfertility problems referred to Danish hospitals or private fertility clinics between 1965 and 1998, as well as recorded in the nationwide National Patient Registry. Through linkage of the cohort to the population‐based civil registration system, 87 women with invalid civil registry numbers were excluded, leaving an analytic cohort of 54,362 women. The cohort was followed for endometrial cancer occurrence from the date of the first subfertility evaluation to the date of emigration, death, or hysterectomy or June 30, 2006, whichever occurred first. At the time of linkage, 101 women had been diagnosed with endometrial cancer during the follow‐up period. For comparison, a subcohort of 1360 women was randomly selected from the subfertility cohort. Hospital files and medical records on all subfertility‐related medical visits were collected. For 18 cases, records could not be found, leaving 83 women with endometrial cancer for analysis. In the subcohort, 78 women for whom the hospital files could not be found, 8 women for whom a diagnosis of subfertility could not be confirmed, and 33 women for whom the cause of subfertility was previous sterilization were excluded, leaving 1241 women. Two of the subcohort members were diagnosed with endometrial cancer during the follow‐up period; therefore, they were included both as cases and as members of the subcohort in the analyses. Median follow‐up, 16 years.
Interventions	Gonadotropins, clomiphene citrate, human chorionic gonadotropin, gonadotropin‐releasing hormone analogs. Information on number and duration of fertility treatment cycles, and, for a minority of the women, dosage of fertility drugs, was available.
Outcomes	Endometrial cancer cases identified through cohort linkage to the nationwide Danish Cancer Registry and the Danish Registry of Pathology from January 1, 2004, onward, because the Danish Cancer Registry was updated only until December 31, 2003, at the time of analysis.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Selection bias (comparability)	High risk	Comment: Exclusion of endometrial cancer cases at the beginning of the study was not reported. Comment: Non‐exposed women drawn from the same population as exposed cohort. Comment: No comparability on cause of subfertility, diabetes mellitus, polycystic ovary syndrome (PCOS), and obesity was ensured.
Selection bias (confounding)	High risk	Quote: "All analyses were stratified according to calendar year and age at start of follow‐up. Rate ratios were adjusted for parity (nulliparous/parous) and number of additional births (linear)". Comment: Analyses inadequately controlled for potential confounding factors.
Performance bias	High risk	Quote: "We collected hospital files and medical records on all infertility‐related medical visits for all infertile women in whom uterine cancer developed and for members of the subcohort". Comment: Exposure to ovary‐stimulation drugs was ascertained by medical records. Blindness regarding the allocated interventions not guaranteed.
Detection bias	Low risk	Quote: "Information on uterine cancer status was obtained through cohort linkage to the Danish Cancer Registry and the Danish Registry of Pathology". Comment: Outcome was ascertained by record linkage.
Attrition bias	Low risk	Quote: "Losses to follow‐up were virtually absent in our study as a result of the precise linkage between our cohort and the Danish population‐based registers, and uterine cancer diagnoses were completely ascertained through linkage with the Danish Cancer Registry and the Danish Registry of Pathology". Comment: Follow‐up was considered complete.
Selective reporting (reporting bias)	Low risk	Comment: All of the study’s prespecified (primary and secondary) outcomes that were of interest in the review have been reported in the prespecified way.
Other bias	High risk	Quote: "The median length of follow‐up was 16.0 years". Comment: Length of follow‐up was considered adequate. Comment: Non‐RCT study.