Berndsen 2007.

Methods	One‐thousand one‐hundred and eighty‐three participants (1183) were randomised in a multi‐centre trial with the primary aim of comparing recurrence rates after laparoscopic transabdominal preperitoneal patch (TAPP)/mesh repair (586) and Shouldice/suture (597) repair. In the Berndsen 2007 study 431 participants from the Shouldice group and 436 from TAPP/mesh group successfully returned for follow‐up at 5 years. In the Arvidsson 2005 study, 454 participants from the TAPP/mesh group and 466 participants from the Shouldice group were followed‐up at 5 years. The participants were clinically examined after 1 and 5 years, and answered questionnaires 2 and 3 years postoperation.
Participants	Male participants between 30 and 70 years with unilateral primary inguinal hernia were eligible for randomisation. Seven surgical centres in Sweden participated in the study: 2 university hospitals and 5 regional/county hospitals. Participants with a history of multiple abdominal operations or concomitant disease contraindicating general anaesthesia were excluded, as also were participants regarded as unable to participate in postoperative follow‐up because of drug abuse, psychiatric disorders, or language difficulties.
Interventions	The TAPP procedure was used to facilitate a mesh repair, and modified Shouldice procedure was used to facilitate a suture repair. TAPP procedure In the TAPP procedure a peritoneal pocket was created and the hernia sac was either dissected and inverted (direct and short indirect sacs) or divided (long indirect sacs). The Cooper’s ligament and pubic bone were exposed as was the spermatic cord for a distance of at least 4 cm from the internal ring. A 7 cm x 12 cm polypropylene mesh (Prolene; Ethicon, Somerville, NJ, USA) was inserted and fixed by use of an endoscopic stapler (EMS Multifeed Staplegun; Ethicon). Care was taken not to apply staples below the inguinal ligament. Modified Shouldice procedure In the modified Shouldice procedure the cremaster muscle was transected at the internal ring. Indirect sacs were dissected to the internal ring, ligated, and extirpated. The posterior inguinal wall was opened from the internal ring to the pubic tubercle. A three‐layer repair was performed using a running 2‐0 polypropylene suture. A relaxation incision was optional. No prophylactic antibiotics were used in either approaches.
Outcomes	In the 2007 study Postoperative pain or discomfort Measured at 1, 2, 3, and 5 years following the procedure, using study‐defined mild, moderate, or severe pain grades Measured at days 1, 2, 3, 5, and 7 following the procedure, using the acute pain visual analogue scale (VAS) Recurrence of hernia Measured at 5 years following procedure, subdivided according to the Nyhus' classification of inguinal hernias Haematoma formation Measured at 5 years following procedure Sick‐leave days taken from work Measured at 5 years following procedure, calculated as a median time (with the range given) amongst the group In the 2005 study Recurrence Recurrence was defined as a bulge in the operated groin when standing and straining. Herniography was performed if the examination was inconclusive. Only participants who completed 5‐year follow‐up or who developed recurrence during the study period were included in the final analysis. Risk factors assessed were American Society of Anaesthesiology (ASA) grade, age, smoking habit, occupation, hernia size (visible versus palpable only), Nyhus classification, operating time, complications at 1 week, pain the first week after operation (combined visual analogue scale (VAS) score on days 1, 2, 3, 5 and 7 days), duration of sick leave, complaints at 3 months’ follow‐up, hospitals and surgeons.
Notes	In the 2007 study Of 1,068 operated participants, 867 were eligible for analysis after 5 years (81.2%). The percentage of participants experiencing discomfort of any kind were 8.5% in the TAPP group and 11.4% (P = 0.156) in the Shouldice group. Although discomfort was usually mild, it was severe for 0.2% in the TAPP group and 0.7% in the Shouldice group. Severe pain the first postoperative week was a risk factor for late discomfort in the Shouldice group (odds ratio 2.25, P = 0.022) but not in the TAPP group. No other risk factor for late discomfort was found. Overall, there was no difference between late discomfort at five‐year follow‐up after laparoscopic TAPP and Shouldice repair. In the 2005 study Recurrences were evenly distributed between groups throughout the follow‐up period. The cumulative recurrence rate after 5 years was 6.6 per cent in the TAPP group and 6.7 per cent in the Shouldice group. Postoperative pain was a risk factor for recurrence after Shouldice operation but not after TAPP repair. There was a correlation between a low surgeon’s performance score and recurrence.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation and allocation of participants into either TAPP/mesh or Shouldice/suture repair groups, that ensured equal baseline demographics. Noted that out of the original 1183 intended for randomisation, only 1068 proceeded to be operated on; dropouts during randomisation were accounted for and openly disclosed as randomisation error, patient refusal of operation, patient moving to another area of residence, emergency operations that had to be done before the study, and other miscellaneous reasons.
Allocation concealment (selection bias)	Low risk	Randomisation and allocation was done so within the confidence of the computer programme and its operator who was impartial towards the participants and had nothing further to do with the study. Allocations were delivered via a confidentially sealed and opaque envelope.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	There was no information in the published journal of the study regarding attempts at blinding either the participants or personnel assessing outcomes.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	There was no information in the published journal of the study regarding attempts at blinding either the participants or personnel assessing outcomes.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Majority of participants were successfully able to be followed‐up by the 5th year. In the 2007 study 431 out of the 538 operated on in the TAPP/mesh group, and 436 out of the 530 from the Shouldice/suture group (> 80%). In the 2005 study 454 from TAPP/mesh and 466 successfully returned for follow‐up at 5 years. There is no clear explanation for this discrepancy in data. Generally those that failed to follow‐up were unable to attend the follow‐up appointment; it is nonetheless noted that 20 people died within the 5 years follow‐up period, and it is not noted whether or not this is directly attributable to the procedure.
Selective reporting (reporting bias)	Low risk	All outcomes were fairly reported and discussed in all participants that were able to be followed‐up.
Other bias	Low risk	Funding noted from Ethicon Endosurgery and Johnson & Johnson company