Bartolucci 2009.

Methods	Allocation: randomized Controlled: placebo and active comparator Blinding: double‐blind Arm: 2‐arm, parallel treatment groups Centre: single centre Study duration: 39‐month period trial (August 2000 to March 2003)
Participants	Inclusion criteria: homozygous SCD participants' who had a severe VOC requiring hospitalization in the internal medicine department. Exclusion criteria: VOC lasting > 72 hr or < 24 hr; parenteral hydration > 24 hr; blood transfusion during the previous month; any NSAID intake during the previous 7 days; pregnancy; history of drug abuse; hypertension; leukocyte count > 30 × 109/L or < 4 × 109/L; presence at inclusion of an acute chest syndrome, defined as the association of 2 criteria among chest pain, radiological infiltrate and auscultatory abnormality; severe anaemia requiring a blood transfusion at inclusion; psychiatric disorder or progressive visceral disease; and ketoprofen allergy or NSAID contraindication; peptic ulcer or treated with 1 of the following drugs the week before enrolment: aspirin, valproic acid, macrolides, anti‐H2, imidazole, rifampicin, phenobarbital, carbamazepine, phenytoin, heparin, vitamin K antagonist, ticlopidine, lithium, methotrexate, interferon‐α, diuretics or antihypertensive agents Baseline characteristics N: 54 participants; 66 VOC episodes Gender: F 20, M 34 Age (mean): ≥ 15 years overall: 27 (SD 7) years intervention; 26 (SD 7) years control Number randomized: 33 intervention; 33 control (VOC episodes) Number completed: 26 intervention; 26 control (VOC episodes) Setting and location: Adult Sickle Cell Referral Centre, Paris University Hospital, France
Interventions	Treatment group (33 VOC episodes): Days 1 and 2: IV infusion ketoprofen 300 mg, 2 days IV morphine 0.1 mg/kg every 5 minutes until pain relief achieved, followed by continuous morphine infusion at initial dose of 2 mg/hr with repeated pulses until pain was well controlled Days 3–5: oral ketoprofen 100 mg, every 8 hr Control group (33 VOC episodes): Days 1 and 2: continuous IV physiological saline (placebo), for 2 days IV morphine 0.1 mg/kg every 5 minutes until pain relief achieved, followed by continuous morphine infusion at initial dose of 2 mg/hr with repeated pulses until pain was well controlled Days 3–5: oral placebo, every 8 hr Duration of treatment: 5 days Follow‐up period: up to 14 days postdischarge Cointerventions/additional analgesia: adjunctive standardized treatment: bed rest, < 3 L of 5% glucose infusion, 1 L/day oral alkaline water, 20 mg/day folic acid, morphine dose not reported and IV paracetamol 2 g every 8 hr for 48 hr, then 1 g every 8 hr)
Outcomes	Primary outcomes Duration of VOC from inclusion (in hr); when at least ¾ were met: absence of fever for 8 hr, absence of pain progression and no requirement of IV infusion of opioid analgesics for the last 8 hr Secondary outcomes Morphine consumption during VOC Pain intensity by VAS; at baseline, every 4 hr Severity of pain assessed by CPS to grade pain at 7 body sites; at baseline and 12 hr Treatment failure, toxicity and adverse effects; reported on a standardized case report form
Notes	Sources of funding: Inserm was the trial promoter. Publication costs of the article were defrayed in part by page charge payments. The article was marked as advertising.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "patients were randomly assigned in a 1:1 ratio to either the intervention or control group." Comment: no details of randomization method.
Allocation concealment (selection bias)	Unclear risk	Comment: insufficient information of concealment.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Comment: no information of details of the blinding process for either IV saline or oral placebo.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Comment: no statement of outcome assessment of blinding.
Incomplete outcome data (attrition bias) All outcomes	High risk	Quote: "7 VOCs each group were excluded from the analysis because of treatment failures." Comment: 20% attrition, and similar in each group. But outcome reporting ignored these randomized participants, and analyses did not seem to be ITT.
Selective reporting (reporting bias)	Low risk	Comment: all end points listed in the methods were reported in the results.
Size	High risk	Comment: total participants 54 (66 events), < 50 participants per treatment arm.
Other bias	Unclear risk	Comment: unit of analysis error. Unit of randomization was the painful crisis (66 randomized crises among 54 participants), but unit of analysis ignored the clustered nature of the data; analysed as if all randomized events were independent.