Mersilene

All patients were operated on in a supine position receiving general anesthesia, with the table in anti-Trendelenburg position. Figure 1 shows the anatomic locations of the incisions. Five trocars were inserted abdominally, of which three were occupied by the robotic arms (no. 1 and 2), and two were used for conventional laparoscopic instruments (no. 3 and no. 4). The camera was placed in position 1 (umbilical incision in smaller patients, and supra-umbilical incision in larger patients), and the two working trocars were in position 2. Position 3 was used for the liver retractor and position 4 (which is optional) was used as an extra instrument for the assistant or supervisor to give some additional help when needed.
The RNF was similar to the LNF method. The procedures started with mobilization of the esophagus and gastric fundus as far as deemed necessary, and transection of the upper short gastric vessels in all cases. A retrogastric opening was created. Hiatal repair was performed using a single stitch (2-0/3-0 Mersilene, Ethicon, Johnson & Johnson, Raritan, NJ, USA). The construction of a floppy 360-degree NF was achieved with the placement of 2–3 sutures (2-0/3-0 Mersilene).

The monofilament non-braided suture used is Ethilon®, (Ethicon, UK) and the multifilament braided suture is Mersilene® tape (Ethicon, UK) for the intervention and control groups, respectively. The stitch is inserted using a McDonald technique. Planned removal of the suture would occur at 37 (±1) weeks of gestation. When the stitch is removed it is retained for microbiological investigations. A vaginal swab is taken before the cerclage procedure to ensure that any infection is treated before inserting the suture. A vaginal swab will also be taken at the time of suture removal. The follow-up should not involve additional contact with the patient beyond routine local clinical management protocols. The study will be deemed complete when the last recruited woman has delivered and, if applicable, her baby is discharged from the neonatal unit.
In view of logistic difficulties in ensuring blinding of participants, and the fact that all the COTS outcomes are objective measures that would be easily and independently retrieved from hospital records (and hence it is unlikely that the lack of blinding will be a cause of serious bias), we do not intend to attempt blinding participants or assessors.