Fleckenstein 2016.

Study characteristics
Methods	Study design: parallel‐group randomized controlled trial Sample size: 45 Country: Germany Setting: secondary care Dates conducted: February 2006 to December 2011 Postoperative opioid used and delivery: morphine PCA Pain score collection: not reported Concurrent postoperative analgesics: none reported
Participants	Inclusion criteria Aged ≥ 18 years old Abdominal or thoracic surgery ASA 1 or 2 Exclusion criteria Severe cardiac, pulmonary, renal or neurologic diseases Type 1 diabetes Diseases influencing the peripheral nervous system (e.g. polyneuropathy, chronic pain syndromes) Regional anaesthesia Use of analgesics Pregnancy or lactation Uncontrolled hypertension Contraindications listed in the product information of etoricoxib
Interventions	Group 1 (unclear number of participants): PO etoricoxib 120 mg on morning of surgery then PO etoricoxib 120 mg/day for 3 days and postoperative placebo Group 3 (unclear number of participants): placebo on morning of surgery then PO etoricoxib 120 mg/day etorocoxib for 3 days
Outcomes	Morphine consumption (mg consumed at 1, 2, 4 hours and day 1‐3) Postoperative pain (on 0‐10 cm VAS and DGSS pain questionnaire during the first 48 hours) Quantitative sensory testing (various parameters at 48 hours)
Notes	Funding: MSD Sharp and Dome (industry) Declarations of interest: none declared Authors contacted: yes Other: standard deviation from post‐incision group estimated from preventive group. Data extracted from graph. Pain score data not reported. Participant numbers unclear but estimated from total number of participants receiving etoricoxib. Some information extracted from published protocol
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated
Allocation concealment (selection bias)	Low risk	Pharmacy‐controlled. Quote: "sequentially numbered envelopes containing two boxes of study medication for pre‐ and postoperative use performed by pharmacy"
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐dummy placebo used. The pills were similar in appearance.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Blinded. Quote: "analysis of all records is performed by blinded evaluators".
Incomplete outcome data (attrition bias) All outcomes	High risk	Patients lost to follow‐up and some for high pain. Quote: "six out of the eight dropouts (75%) because of increased pain were part of the placebo group".
Selective reporting (reporting bias)	High risk	NCT00716833 and published protocol. Pain data not fully reported
Other bias	Unclear risk	No separate data to assess. Industry funding but stated not involved in study