Study question: Does the use of diagnostic criteria in the hysteroscopic diagnosis of a septate uterus improve inter-observer agreement?

Summary answer: Pre-set diagnostic criteria slightly improve the inter-observer reproducibility of hysteroscopy in diagnosing a uterine septum, although agreement remains moderate.

What is known already: The inter-observer agreement on the hysteroscopic diagnosis of the septate uterus has been reported to be poor.

Study design, size, duration: From April 2013 until May 2014, a randomized controlled comparative inter-observer study was performed. A total of 191 gynecologists from 43 countries took part.

Participants/materials, settings, methods: Each gynecologist was asked to assess 10 video recordings of hysteroscopy procedures with a specific focus on the internal uterine shape. The hysteroscopies had been performed in subfertile women and women with recurrent miscarriage. The recordings contained images of uterine cavities primarily diagnosed as septate, arcuate or normal. Participating gynecologists were randomized into two groups: one group received diagnostic criteria for a septate uterus before assessment of the videos (DC group), whereas the other group assessed the recordings without instruction (no DC group). The inter-observer agreement, expressed as the intra-class correlation coefficient (ICC), was compared between groups. Main outcomes were the inter-observer agreement on the uterine shape and the necessity of surgical correction.

Main results and the role of chance: Eighty-six observers were randomized to the DC group and 105 to the no DC group. The ICCs in the diagnosis of a septum were 0.59 versus 0.52, in the DC group and the non-DC group, respectively (P-value: 0.002). The overall agreement on the need for surgical correction was found to be moderate (DC ICC 0.43 versus no DC 0.39, P-value: 0.70). Most importantly, once a septate uterus had been diagnosed, the agreement on the need for surgery was poor in both groups (DC ICC 0.05 versus no DC ICC 0.02, P-value: 0.78).

Limitations, reasons for caution: We used video recordings rather than studying real-time hysteroscopic procedures, which may have influenced the accuracy of the assessments.

Wider implications of the findings: The reproducibility of hysteroscopy for the diagnosis of a septate uterus is moderate, even with the use of standardized criteria. The fact that the agreement among physicians on both the diagnosis of a uterine septum, as well as the decision to resect such septum after hysteroscopy is moderate, may imply that hysteroscopy is insufficient as single tool to diagnose and decide on treatment of a septate uterus.

Study funding/competing interests: No study funding was received and no competing interests are present.