Koppert 2006.

Methods	Randomized, double‐blind, controlled, multiple dose study; evaluated outcomes over at least 3 days. Medication was administered immediately after surgery upon arrival in the PACU.
Participants	Type of surgery: hip replacement or surgery of the femoral shaft Paracetamol group Entered/completing: 27/25 Age (mean, SD): 76.7 +/‐ 8.9 Sex (male, %): 15 (55.6%) Parecoxib group Entered/completing: 28/25 Age (mean, SD): 76 +/‐ 8 Sex (male, %): 11 (39.3%) Placebo group (saline) Entered/completing: 28/25 Age (mean, SD): 76.7 +/‐ 8.6 Sex (male, %): 12 (42.9%)
Interventions	Paracetamol: IV infusion of 1000 mg paracetamol (Perfalgan) over 10 min; admin at 6‐h intervals for at least 3 days Parecoxib: 40 mg IV over 10 min (Dynastat); admin at 12‐h intervals for at least 3 days Placebo: saline IV over 10 min
Outcomes	Primary: renal function: blood samples (serum Cystatin C, creatinine, blood urea nitrogen, liver biochemistry); urine samples (creatinine clearance, urinary excretion of sodium, potassium, albumin, alpha1‐microglobulin; fluid balance (CVP) Secondary: pain intensity (NRS 0 to 10), rescue medication usage including morphine equianalgesic dosages
Source of funding	Supported by Bristol‐Meyers Squibb
Were treatment groups comparable at baseline?	“All groups were comparable with regard to age, weight, height, distribution of sex, preexisting diseases, and ASA status…. Type and lengths of surgical and anesthetic procedures across the treatment groups were similar… Furthermore, consumption of crystalloids and colloids were similar.”
Details of preoperative pain	Not reported
Notes	“If a patient had received NSAIDs or COX‐2 inhibitors, there was a washout period of at least 72 h and the weak opioid, tramadol, was provided as a substitute.”
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Randomization of the study medication (parecoxib versus paracetamol versus saline) was performed by computer‐generated codes maintained in sequentially numbered, opaque envelopes. Additional envelopes were provided if participants had to be excluded after recruitment and randomization.”
Allocation concealment (selection bias)	High risk	Despite use of sequentially numbered envelopes, participants and nursing staff on ward were unblinded
Blinding (performance bias and detection bias) All outcomes	High risk	Anesthesiologist, nursing staff, and investigators were blinded. All study medication solutions were prepared by a hospital pharmacist who was not involved in the data collection. On the ward, participants and nursing staff were unblinded. How blinding was maintained was not described.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No missing outcome data. All participants that did not complete the study were accounted for.
Selective reporting (reporting bias)	Low risk	All outcomes were reported in Results
Size	High risk	Fewer than 50 participants per arm of the study (27 paracetamol, 28 parecoxib, 28 placebo)