Mueller 2005

Methods	Randomized controlled cross‐over study with 2 arms. Duration: 16 months in total (oral Mg pidolate or placebo for 6 months, followed by 2‐month washout period, then cross over to 6 months period of alternate agent (placebo/Mg) and 2 months wash out). Multicentre: 2 major paediatric sickle cell centres. Location: Boston and Texas, USA.
Participants	Over 100 children invited to participate between January 2002 and December 2004, enrolled participants or parents/guardian gave informed consent. Only 12 participants with HbSC disease and at least 1 pain crisis within last year were enrolled. Age: range 3.9 to 16.8 years. Gender split: 7 male, 5 female. Of 12 participants, 5 were evaluated for efficacy assessment, 7 dropped out of trial for several reasons: 3 for non‐compliance, 2 for study violation (pharmacy dispense wrong formulation), 2 withdrew for personal reasons.
Interventions	Intervention: a liquid containing 0.6 m Eq Mg pidolate/kg/body weight per day, divided into 2 daily doses; Mg pidolate (45 g) distributed as a pre‐mixed powder containing Koolaid Tropical Punch powder (9 g), and sucrose (67 g). Control: a placebo liquid containing an equivalent amount of placebo to the study medication, divided into 2 daily doses; contained the same amount of sucrose and Tropical Punch powder as the intervention as well as 45 g of lactose.
Outcomes	Primary outcome: change in the number of painful crises. Secondary outcomes: tolerance of long‐term treatment with oral Mg pidolate; change in the intracellular Mg content of erythrocytes; effect of Mg pidolate therapy on the K‐Cl co transport system activity. Follow‐up undertaken every 2 weeks, then every 4 weeks.
Notes	Clinical trials.gov identifier: NCT00040456 The study was terminated due to poor participant enrolment. The study was opened to 2 major sickle cell centres and had relatively non‐restrictive enrolment criteria, accrual was extremely low, presumably due to sporadic and rather low disease intensity of HbSC disease. We have requested further details from the study investigators (randomization, blinding, missing data and efficacy results of outcomes measure). Abstract was published 10 years ago and a search using authors' name did not reveal full publication; to date no data received from study investigators.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Described as randomized, but no details given for how the sequence was generated. Insufficient information to judge.
Allocation concealment (selection bias)	Unclear risk	Insufficient information to judge.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Entry on clinicaltrials.gov NCT00040456 states there was double‐blind masking (participant, caregiver, investigator). Placebo described as a liquid containing an equivalent amount of placebo to the study medication, divided into the same number of daily doses and containing the same amount of sucrose and Tropical Punch powder as the intervention as well as 45 g of lactose.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Entry on clinicaltrials.gov NCT00040456 states there was double‐blind masking (participant, caregiver, investigator); however, the method of blinding the outcome assessors was not described.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Of 12 participants, 5 were evaluated for efficacy assessment, 7 dropped out of trial for several reasons: 3 for non‐compliance, 2 for study violation (pharmacy dispense wrong formulation), 2 withdrew for personal reasons.
Selective reporting (reporting bias)	Unclear risk	There was insufficient information to judge.
Other bias	Unclear risk	In this study expected to enrol 20 participants; 100 eligible potential participants were contacted and invited to participate in the study. According to the current recruiting status on clinicaltrials.gov, the number of participants recruited is 12 and the study was terminated due to lack of accrual. A full paper has not been published and the results for 12 participants have only been presented narratively without data analyses.

HbF: foetal haemoglobin HbSC: haemoglobin sickle cell HC: hydroxyurea HRQoL: health‐related quality of life IQR: interquartile range IV: intravenous LOS: length of stay m Eq: morphine equivalent units Mg: magnesium MgSO4: magnesium sulphate Na: sodium RBC: red blood cell SCD: sickle cell disease SD: standard deviation VOC: vaso‐occlusive crisis