Drake 1998.

Study characteristics
Methods	Study design: parallel‐group randomised trial Study dates: October 1 1993 to October 31 1995 Setting: inpatient, multicentre, international Country: Canada, United States, France
Participants	Inclusion criteria: Patients between birth and 18 years of age, newly diagnosed hydrocephalus with documented ventriculomegaly in CT, MRI or ultrasonographic image scans and requiring a first ventriculo‐peritoneal CSF shunt insertion. Exclusion criteria: 1‐ previous indwelling CSF shunt (patients with subcutaneous reservoirs for aspiration or ventricular catheters draining externally or to the subcutaneous scalp were eligible); 2‐ active abdominal or CSF infection; 3‐ diffuse spread of tumour in the subarachnoid space; 4‐ marked prematurity with skin at risk for erosion from shunt hardware; 5‐ systemic disorders precluding shunt insertion; 6‐ septated loculations within the ventricular system requiring more than one shunt;. 7‐ Dandy Walker malformation; 8‐ arachnoid cyst as a cause of hydrocephalus; 9‐ inability to be monitored for 1 year baseline. Sample size: 344 participants randomised Group 1: Delta valve 115 Group 2: Standard valve 114  Group 3: Orbis‐Sigma valve 115  Age (years): Group 1: Delta valve: mean 1.83 years, median 0.2 years Group 2: Standard valve: mean 1.65 years; median 0.21 years Group 3: Orbis‐Sigma valve: mean 1.56 years; median 0.4 years Sex (M/F): Group 1: Delta valve 64/51 Group 2: Standard valve 70/44 Group 3: Orbis‐Sigma valve 57/58 Hydrocephalus cause Group 1: Delta valve: intraventricular haemorrhage 19.1%,  myelomeningocele 24.3%, tumour 8.7%, aqueduct stenosis 7.8%, CSF infection 6.1%, head injury 2.6%, two or more causes 9.6%, other 9.6%, unknown 12.2% Group 2: Standard valve: intraventricular haemorrhage  25.9%,  myelomeningocele 24.1%, tumour 8.0%, aqueduct stenosis 8.0%, CSF infection 4.5%, head injury 1.8%, two or more causes 6.3%, other 11.6%, unknown 9.8% Group 3 Orbis‐Sigma valve: intraventricular haemorrhage  28.1%,  myelomeningocele 15.8%, tumour 10.5%, aqueduct stenosis 5.3%, CSF infection 5.3%, head injury 0.0%, two or more causes 10.5%, other 13.2%, unknown 11.4% Previous surgery: Group 1 Delta valve: subcutaneous reservoir 6.1%, ventricular drain 8.0% Group 2 Standard valve: subcutaneous reservoir 6.1%, ventricular drain 11.9% Group 3 Orbis‐Sigma valve: subcutaneous reservoir 7.1%, ventricular drain 18.0%
Interventions	Group 1 (n = 115): Delta valve (Medtronic PS Medical): a standard diaphragm valve with an additional anti‐syphon device that consists of paired flexible diaphragms, which narrow an orifice as the shunt pressure becomes negative Group 2 (n = 114): Standard (conventional) valve  Group 3 (n = 115): Cordis Orbis‐Sigma valve. This is a flow‐regulation valve with a flexible diaphragm which moves along a piston of variable diameter. It operates through three stages, so that the pressure curve is sigmoid in shape which limits flow by progressively narrowing the flow orifice with increasing pressure, as a pressure‐sensitive ring moves along a variable‐diameter rod with a single opening pressure of approximately 5 cm H2O. Co‐interventions: The details of the surgical technics and postoperative care were established according to the discretion of the surgeon and were recorded (including the prophylactic use of antibiotics, hair removal, the site of hardware insertion, the use of technical aids such as ventriculoscopes, hardware configurations, opening pressure designations for the standard and Delta valves, and postoperative head elevation or compressive dressings). Suggestions from the manufacturers for insertion techniques were solicited at the beginning of the trial and were distributed to all participating centres.
Outcomes	Treatment failure  How measured: defined as shunt malfunction due to shunt obstruction, shunt over‐drainage, loculated ventricles, or shunt infection. Each of these end points were defined by specific criteria in the trial protocol (see below). An adjudication committee, blinded to the treatment group, reviewed each patient´s imaging studies, operative notes and hospital discharge summary when a primary end point was reported. Time points measured:  clinical evaluation at 3 months, 1 year, and yearly until the end of the study Time points reported: complete follow‐up Treatment failure [shunt obstruction] Definition:  1. If a patient has at least one symptom or sign and at least one positive ancillary test Symptoms: headache, nausea, vomiting, decreased level of consciousness, irritability, decreased school performance, loss of developmental milestones Signs: papilledema, bulging fontanelle, nuchal rigidity, 6th nerve(s) paresis, loss of upward gaze, new seizures (or increased seizure frequency), increasing head circumference, fluid tracking along the shunt tubing, umbilication of the shunt reservoir, inability to depress the shunt reservoir. Fluid accumulation around the burr hole site in the early weeks following surgery will not be considered indicative of shunt failure unless it is extreme and progressive or results in leakage of CSF through the wound. Small fluid collections are common and normally resolve spontaneously. Ancillary tests: (a) CT scan, ultrasound, or MRI scan showing enlarged ventricles compared to the 3‐month study or ventricles that have failed to decrease in size compared to the preoperative study (normalisation of ventricle size is not a mandatory criterion for shunt function) (b) Disruption or migration of the shunt system on plain radiographs (c) Radionucleotide or iodinated contrast study showing shunt obstruction (d) ICP monitoring showing persistent elevation of pressure with or without plateau waves (e) Shunt tap in which fluid cannot be aspirated or high pressure is recorded or symptoms/signs of shunt obstruction are relieved 2. When there are no symptoms or signs of shunt obstruction but the ventricles are increased in size, shunt obstruction is said to have occurred if there is no clinical or radiographic suggestion that atrophy is the cause of the ventricular enlargement. 3. A CSF leak that does not resolve and which requires a shunt revision 4. In the rare event of an emergent shunt revision without any ancillary tests, or revision prior to the 3‐month follow‐up scan, obstruction will be judged to be present or absent using clinical information and the operative findings (by the Monitoring/Adjudication Committee) Time points measured:  clinical evaluation at 3 months, 1 year, and yearly until the end of the study Time points reported: complete follow‐up Treatment failure [shunt over‐drainage] Definition:  in the presence of (a) subdural fluid collections ‐ large subdural fluid collections associated with brain compression or symptoms and signs otherwise indicative of shunt obstruction; or (b) slit ventricle syndrome ‐ smaller than normal ventricles associated with postural headache, chronic headache, intermittent headache of an incapacitating nature and documentation of one of the following: (1) transient enlargement of the ventricles as seen on imaging; (2) extreme negative pressure with associated headache in the upright position; (3) sustained elevations of pressure above normal associated with headache Time points measured:  clinical evaluation at 3 months, 1 year, and yearly until the end of the study Time points reported: complete follow‐up Treatment failure [loculated ventricles] Definition: The presence of a loculated portion of a ventricular system that is enlarged above normal and compressing surrounding brain and that requires reoperation Time points measured:  clinical evaluation at 3 months, 1 year, and yearly until the end of the study Time points reported: complete follow‐up Treatment failure [shunt infection] Definition:   1. In the presence of purulent discharge through the wound or erosion of the shunt material through the skin 2. In the presence of one of the following symptoms or signs with at least one of the following ancillary tests Symptoms and signs: fever, meningismus, wound erythema, abdominal pain and/or distention, abdominal mass or peritonitis Ancillary tests: culture or identification of organisms on Gram stain of CSF taken from shunt lumen or abdominal fluid collection, if present, withdrawn under sterile conditions or from purulent material around the shunt Time points measured:  clinical evaluation at 3 months, 1 year, and yearly until the end of the study Time points reported: complete follow‐up Adverse events [surgical complications] How measured: not specified. The following events were included in this outcome (obtained from the Hospital Discharge Form): transient CSF leak, intracranial haemorrhage, new or increased neurological deficit, neck injury, chest injury, abdominal injury, inadvertent skin perforation along shunt tract, wound dehiscence, postoperative subcutaneous fluid collection, perioperative death, and "other" complications.  Time points measured: time of hospital discharge Time points reported: time of hospital discharge Mortality for all causes How measured: not specified. Cause of death was recorded on a death form and classified as probably related to shunt malfunction; neurologic, but probably not related to the shunt malfunction; or non‐neurological. Time points measured: complete follow‐up Time points reported: complete follow‐up Quality of life: not reported Ventricular size reduction: not reported Head circumference: not reported
Funding sources	Medtronic PS Medical (Goleta, CA), Cordis Corp (Miami FL), British Columbia Health Research Foundation and The Science Council of British Columbia (Burnaby Canada), Mr. and Mrs. PA Woodward´s Foundation (Vancouver Canada) British Columbia´s Children´s Hospital Telethon Clinical Investigative Initiatives (Vancouver, Canada) British Columbia´s Medical Services Foundation (Vancouver, Canada) Spina Bifida Association of Canada (Winnipeg, Canada) and Spina Bifida Association of British Columbia (Surrey, Canada)
Declarations of interest	Not reported
Notes	1) Standard differential pressure valves can be classified into four categories: silicone rubber slit valves, silicon rubber mitre valves, metallic spring ball valves. Below a  threshold difference in pressure across the valve, it remains closed, and above the threshold difference in pressure, it opens (the opening pressure). Conventional valves are supplied usually as low, medium and high pressure. Although there are no uniform standards for these designations, typically their pressure differentials are 5, 10 and 15 mm H₂O, respectively. Despite the variations in the standard valve design, the pressure/flow characteristics are very similar.  2) Assessment at 3 months, 1 year and yearly until the end of the study were performed by the attending neurosurgeon during an outpatient hospital visit by the patient, using a specific form. Follow‐up included imaging studies at 3 months and thereafter according to the surgeons' individual practices. A 12‐month scan was encouraged if possible. A completed telephone follow‐up form was required at months 18 and 30. A minimum follow‐up of 1 year was required. The details of the surgical techniques and postoperative care were established according to the discretion of the surgeon, including the prophylactic use of antibiotics, hair removal, the site of hardware insertion, the use of technical aids such as ventriculoscopes, hardware configurations, opening pressure designations for the standard and Delta valves, and postoperative head elevation or compressive dressings. Suggestions from the manufacturers for insertion techniques were solicited at the beginning of the trial and were distributed to all participating centres.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Patients were stratified by centre and by age (less than or more than 6 mo), and the randomisation scheme was blocked with randomly varying block size.”
Allocation concealment (selection bias)	Low risk	“Consecutive opaque envelopes located in each operating room.”
Blinding of participants and personnel (performance bias) Subjective outcomes: treatment failure, adverse events, quality of life	High risk	Study personnel were aware of the type of valve that the patient received. The shunts were distinctly different in shape, can be easily palpated through the skin, and were radio‐opaque.
Blinding of participants and personnel (performance bias) Objective outcomes: mortality, head circumference, ventricular size	Low risk	No information about blinding available. Outcomes were likely to be unaffected by blinding.
Blinding of outcome assessment (detection bias) Subjective outcomes: treatment failure, adverse events, quality of life	Low risk	"Determination of outcome will be made by the neurosurgeon in charge of the care of the child at the time of the malfunction. The outcome will be reviewed by the adjudication committee, which will be blinded to the type of shunt in place."
Blinding of outcome assessment (detection bias) Objective outcomes: mortality, head circumference, ventricular size	Low risk	No information about blinding available. Outcomes were likely to be unaffected by blinding.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Eight patients did not complete the minimum 1‐year follow‐up period. “8/344 patients did not complete the minimum 1‐year follow‐up period.”
Selective reporting (reporting bias)	High risk	Some outcomes listed in the protocol were not reported, e.g. change in ventricular size, hospital stay. Data on mortality was not disaggregated by group.
Other bias	Low risk	No other sources of bias were detected.