28 Chapter 2 Third, as patient characteristics were similar between allocation groups at randomization but might differ later in follow-up, the following characteristics were adjusted for, and the following steps from the primary statistical analysis repeated: female age, duration of infertility, primary or secondary infertility, percentage of progressive motile spermatozoa, volume of semen sample, referral status by general practitioner or specialist, abnormal HSG result in terms of blockages, female smoking status and female body mass index. Supplementary analysis As a supplementary analysis, the study continued with the question of whether the effect of oil contrast versus water contrast would be different for couples undergoing expectant management or receiving medically assisted reproduction (MAR), i.e. intrauterine insemination (IUI)/intrauterine insemination with ovarian stimulation (IUI-OS) or IVF/ intracytoplasmic sperm injection (ICSI) as the Fallopian tubes were bypassed in IVF/ICSI. First, the follow-up data were reformatted by dividing the follow-up time into periods when couples where either pursuing expectant management, receiving IUI/IUI-OS or receiving IVF/ICSI. The start of follow-up for MAR was defined as 14 days before the first day of last menstruation previous to commencing treatment. The end of MAR follow-up was defined as the first day of the last menstruation before either ongoing pregnancy or the final insemination or embryo transfer. This aligned with the definition of time to natural conception. A treatment indicator was created that denoted which treatment (expectant, IUI/IUI-OS or IVF/ICSI) was received in which time period. Next, a Cox model was fitted with the oil versus water allocation, the MAR indicator and the interaction between these two. This model estimates the effect of oil versus water separately for expectant management, IUI/IUI-OS and IVF/ICSI. RESULTS The H2Oil study randomized 1119 couples. After excluding couples who conceived before receiving HSG and couples with missing follow-up or pregnancy data, data on 1107 couples were available, of which 550 couples received oil-based contrast and 557 couples received HSG with water-based contrast. Ongoing pregnancy rates at 3 years were 426 (77%) and 394 (71%), respectively. Figure 1 displays ongoing pregnancy rates as a Kaplan–Meier curve including the sample size over time. Average female age at randomization was 32.7 years (5th–95th percentile: 26.1–38.9) and median duration of infertility was 1.61 years (5th–95th percentile: 0.91–3.89). A total of 746 (67%) couples had primary infertility. The median follow-up for all couples including up to conception was 9–10 months (5th–95th percentile: <1 to 36).
RkJQdWJsaXNoZXIy MTk4NDMw