Intrauterine insemination with gonadotropin stimulation or in vitro fertilization for the treatment of unexplained subfertility: a randomized controlled trial

The singleton live birth rate with one cycle of in vitro fertilization was not significantly different than three cycles of Intrauterine inseminationþcontrolled ovarian hyperstimulation.

Volume 107, Issue 6, Pages 1329–1335.e2


Anupa Nandi, M.R.C.O.G., Priya Bhide, M.R.C.O.G., Richard Hooper, Ph.D., Anil Gudi, M.R.C.O.G., Amit Shah, M.R.C.O.G., Khalid Khan, Ph.D., Roy Homburg, F.R.C.O.G.



To evaluate the best first line management option for the treatment of unexplained subfertility-controlled ovarian hyperstimulation (COH) with gonadotropins and IUI or IVF.


Randomized controlled trial.


Single center trial in a tertiary referral unit.


Couples with unexplained subfertility.


Couples were randomized to receive either three cycles of IUI + COH or one cycle of IVF.

Main Outcome Measure(s)

Singleton pregnancy rate (PR) per couple.


A total of 207 couples were randomly assigned to three cycles of IUI + COH (n = 101) or one cycle of IVF (n = 106). There were 25 (24.7%) singleton live births for the IUI + COH group and 33 (31.1%) for the IVF group (relative risk, 1.3; 95% confidence interval [CI] 0.81–1.96) with an absolute risk difference of 6.4% (95% CI −5.8% to 18.6%). The multiple pregnancies per live birth were 4 (13.8%) for the IUI + COH group and 3 (8.3%) for the IVF group (relative risk, 0.6; 95% CI 0.14–2.4). There were no cases of ovarian hyperstimulation syndrome (OHSS) in the IUI group and three cases of OHSS (3.7%) in the IVF group. There were 17 live births from spontaneous conception in between treatment cycles (8.2%).


The singleton live birth rate with one cycle of IVF was not significantly different than three cycles of IUI + COH.

Clinical Trial Registration Number


Read the full text here.

Please sign in or register for FREE

Your Fertility and Sterility Dialog login information is not the same as your ASRM or EES credentials. Users must create a separate account to comment or interact on the Dialog.

Go to the profile of Micah J Hill
over 5 years ago
I read your article with interest and have a couple of comments. With a trial that had so many spontaneous pregnancies and 8 people crossing over from IUI to IVF early, the differences between the ITT and per protocol analyses are very interesting. How did you come up with the per protocol numbers? It seems the denominators should be the number of patients who actually got each treatment, which is 93 for IUI and 89 for IVF (81 in the IVF arm and 8 in the IUI arm who switched to IVF). If you do it this way, the birth rate is 18/93 in the IUI arm and 30/89 in the IVF arm, which turns out to be 19.3% versus 33.7% P=0.03. Clearly, this gives a very different interpretation to the results. How did you calculate the per protocol data the way you reported it? I think this information is particularly important in a study where 21% of patients did not receive their intervention or crossed over to the other treatment. The conclusion of the paper is that there is no difference between 1 IVF cycle and 3 IUI cycles in birth rates. However, this is very limited by the sample size. With this sample size, the IVF arm would have had to have a live birth rate of 43.7% compared to the 25% in the IUI arm to reach statistical significance. While early termination of the study was unavoidable due to coverage changes, we are left with data that is underpowered to truly say the outcomes between the two groups are similar. For example, if we believe that a true difference of 10% in live birth is clinically important in our counseling, then this study only had 34% power to detect that difference. I look forward to your thoughts and replies.
Go to the profile of Anupa Nandi
over 5 years ago
Many thanks for showing interest and your comments. Since there were a large number of spontaneous pregnancies, we also performed a per-protocol analysis as a secondary outcome. We calculated the per protocol analysis based on the people who completed all three cycles of IUI vs the people who completed one cycle of IVF. Excluding the dropouts and the spontaneous conceptions, there were 18 live births from IUI (17 in the IUI group and one from those who withdrew from IVF and underwent IUI instead). 69 couples in the IUI group completed all three cycles of IUI and out of three couples who withdrew from IVF group and had IUI; only one completed all three cycles. Hence as per protocol, there were 18/70 (25.7%) live births per three cycles of IUI. There were 23 live births from 81 IVF cycles in the IVF group and 7 live births from 18 IVF cycles in the IUI group, 30/99 (30.3%) live births per cycle. According to per protocol analysis, three cycles of IUI had a similar live birth rate as one cycle of IVF, RR 1.17 (95% CI 0.71 – 1.93). One of the main limitations of our trial is that we had to close our trial after recruiting 207 couples. This was a purely administrative decision, which was beyond the investigator’s control. Hence we report the results of the trial as they stand. We believe that the 95% confidence interval in our trial would have automatically taken into account the size of the trial. Kind regards Anupa Nandi
Go to the profile of Micah J Hill
over 5 years ago
Thank you for the clarification on the per protocol analysis.