Pollack 1999.

Study characteristics
Methods	Study design: parallel‐group randomised trial Study dates: randomisation June 29, 1993 to December 14, 1995 Setting: inpatient procedure, outpatient follow‐up, multicentre, national Country: United States
Participants	Inclusion criteria: patients who required placement of ventriculoperitoneal or ventriculo‐atrial shunt for the treatment of hydrocephalus (first placement or replacement) Exclusion criteria: patients under treatment with drugs for anticoagulation, coagulopathies, active infection, valve explantation because of infection within the past 3 months, requiring two valves at the same time, any contraindication for ventriculo‐atrial or ventriculo‐peritoneal shunting, expected life‐span less than 24 months, and unwillingness or inability to return for required follow‐up evaluation at the investigational site Sample size: 377 participants randomised Group 1: Experimental: 194 Group 2: Control: 183 Age (years): Group 1: Experimental: median 14 (+/‐ 23) Group 2: Control: median 12 (+/‐ 23) Sex (M/F): Group 1: 59% male, 41% female Group 2: 54% male, 46% female Diagnosis:  Group 1: communicating hydrocephalus 59%, non‐communicating 41% Group 2: communicating hydrocephalus 57%, non‐communicating 43%
Interventions	Group 1: Experimental (n = 194): Codman‐Hakim programmable valve system (Codman/Johnson & Johnson, Raynham, MA)  Group 2: Control (n = 183): Any non‐programmable valve pressure with fixed pressure available on the market: Delta (Medtronic PS Medical, Goleta Ca): 10%, other Medtronic PS Medical: 57%, other Codman/Johnson & Johnson: 8%, Orbis‐Sigma (Cordis, Miami Lakes, FL): 3%, others: 22 % Co‐interventions: The study required that the programmable valve in all patients in whom it had been placed be reprogrammed after any imaging study because of concern that the valve setting would alter in the magnetic field. Other than this instruction, decisions regarding the programmable valve pressures were left entirely to the respective investigator's judgement.
Outcomes	Primary outcomes Treatment failure [shunt system failure]: Definition: a surgical intervention to the original shunt system on‐study, for any reason, was considered a shunt system failure. Primary outcome measure: survival of the shunt system. Time points measured: clinical evaluations 1, 3, 6, 12, 18 and 24 months Time points reported: 24 months Treatment failure [infection necessitating explantation]: Definition: not reported Time points measured: clinical evaluations 1, 3, 6, 12, 18 and 24 months Time points reported: Not reported (probably 104 weeks) Adverse events: Definition: "complications included the following: haemorrhage, malabsorption of CSF, subdural haematoma/hygroma, neurological deficit, catheter perforation, foreign body reaction, infection, and “other” complications". Time points measured: clinical evaluations 1, 3, 6, 12, 18 and 24 months and radiological images at 3, 12 and 24 months Time points reported: Not reported (probably 104 weeks) Adverse events [subdural hematoma/hygroma]: Definition: not reported Time points measured: clinical evaluations 1, 3, 6, 12, 18 and 24 months Time points reported: Not reported (probably 104 weeks) Mortality for all causes:  not reported Quality of life: not reported Ventricular size reduction Definition: ventricle size was reported as big, average, small, or slit, according to qualitative assessment of image studies (computed tomography, MRI, or ultrasound studies). Time points measured: evaluated at 3, 12 and 24 months with radiological images (ultrasound, CT scan and MRI) Time points reported: radiological images at 3, 12 and 24 months Head circumference: not reported
Funding sources	“The authors have no financial involvement with any of the products used in this study.”
Declarations of interest	Not available
Notes	1) The Codman‐Hakim programmable valve is similar to the original Hakim valve in its ball‐cone and pressure‐inducing spring mechanism to achieve differential pressure; however, the pressure setting can be adjusted by a stepper motor and spiral cam, which are rotated noninvasively using a electromagnetic programmer. It can be set while the valve is packaged and sterile, or after implantation. Pressure settings from 30 to 200 mm H₂O in increments of 10 mm H2O can be achieved. 2) At a scheduled follow‐up visit, an investigator recorded patient status on the following scale: 1) well, 2) symptomatic, relating to hydrocephalus, 3) symptomatic, relating to shunt system, 4) symptomatic, cause unknown, and 5) symptomatic, unrelated to hydrocephalus or shunt system. 3) Analyses were performed on the first shunt systems for each patient to incorporate the protection from bias provided by randomisation at study entry.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	“Randomisation was stratified by study site, and within each site randomisation was stratified further by patient shunt history. Patients with a history of one or more previous shunts before placement of a study valve were classified as “replacement” patients. Each of these two groups was randomised separately within each site to maintain balance design”. No information available
Allocation concealment (selection bias)	Unclear risk	See above. No other information available
Blinding of participants and personnel (performance bias) Subjective outcomes: treatment failure, adverse events, quality of life	High risk	Participants and personnel were not blinded. There was no detail on co‐interventions.
Blinding of participants and personnel (performance bias) Objective outcomes: mortality, head circumference, ventricular size	Low risk	Participants and personnel were not blinded. There was no detail on co‐interventions. Outcomes were unlikely to be affected by blinding.
Blinding of outcome assessment (detection bias) Subjective outcomes: treatment failure, adverse events, quality of life	High risk	Participants and personnel were not blinded.
Blinding of outcome assessment (detection bias) Objective outcomes: mortality, head circumference, ventricular size	Low risk	Participants and personnel were not blinded. Outcomes were unlikely to be affected by blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	7 patients (2/194  experimental, 5/183 control) were lost to follow‐up, 19 died (11/194 experimental, 8/183 control) before the 24‐month follow‐up point was reached. More than 90% of patients completed the 3, 12 and 24‐month postoperative evaluations.
Selective reporting (reporting bias)	Unclear risk	Insufficient information to permit judgement of ‘Low risk’ or ‘High risk’ (no protocol available)
Other bias	Low risk	No other sources of bias were identified.