Corticosteroids for acute traumatic brain injury

 

Characteristics of included studies [ordered by study ID]

Alexander 1972

---

Methods	Paper states that patients were randomized, half of them receiving dexamethasone.
Participants	Those admitted to hospital with acute non missile head injuries who responded to painful stimuli by withdrawal or rigidity but who would not respond to voice command.
Interventions	Dexamethasone 10mg IV; then 4mg IM every 6 hours for 10 days. No steroid, no mention of placebo.
Outcomes	Death (time of assessment not stated). Complications.
Notes	States all patients uniformly supervised. Author contacted, no further details of trial available.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Braakman 1983

---

Methods	Randomized into two groups using identical vials in identical boxes prepared by a Pharmacy Department. Code not broken until all patients were evaluated for outcome.
Participants	Patients of any age with severe non missile related head injury who were in a coma on admission to hospital. Coma was defined as no eye opening, no spoken response to painful stimuli and not obeying commands. Exclusions were those already brain dead (apnoea, flaccidity, dilated pupils not reacting to light, absence of reflex eye movements) or expected to become so within 1 hour, those who regained consciousness during initial examination, those who had already been given steroids and those with diabetes mellitus or a history of peptic ulcer.
Interventions	Dexamethasone phosphate: initial dose 100mg IV (less than 6 hours from injury), days 1 to 4 100mg/day IV, days 5 to 7 16mg/day IV or IM, day 8 12mg IV/IM, day 9 8mg IV/IM, day 10 4mg IV/IM. Placebo.
Outcomes	Glasgow Coma Scale at 6 months after injury. Complication rate.
Notes	Other care was determined by the result of a CT scan: those with a mass lesion had immediate operation, those without had monitored control of intracranial pressure using controlled ventilation and/or osmotic diuretic. Cerebrospinal fluid was drained in some patients. Barbiturates were not used.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Braun 1986

---

Methods	Patients randomly assigned in emergency room to one of four groups.
Participants	Those head trauma victims admitted from the emergency room to the adult neurosurgical intensive care unit.
Interventions	Dexamethasone 6mg/kg IV in the emergency room; then 6mg/kg IV 6 hourly for 1 day, then 1mg/kg IV 6 hourly for 4 days followed by a tapering dose. Dexamethasone 6mg/kg IV in the emergency room; then 0.1mg/kg IV 6 hourly for 5 days followed by a tapering dose. Dexamethasone 6mg/kg IV in the emergency room: single dose only. No steroids, no mention of placebo.
Outcomes	Incidence of pneumonia defined as a new alveolar or alveolar-interstitial infiltrate identified by a radiologist and one of the authors and two of: Fever greater than 100F. White cell count greater than 15,000/ml. Increased sputum with a predominant organism on either Gram stain or culture of sputum.
Notes	Otherwise treated with a standardised protocol including ICP monitoring, initial hyperventilation, hyperosmotic agents. Thiopentone in some.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Chacon 1987

---

Methods	Patients randomized 'by lottery method' to two groups. Blinding not clear.
Participants	Children admitted to hospital with severe head trauma, a Glasgow Coma Score of 7 or less and evidence of cerebral oedema on computerised tomography.
Interventions	Dexamethasone - dose not stated. No dexamethasone.
Outcomes	Mortality at discharge.
Notes	Other interventions - fluid restriction guided by central venous pressure and urine output, hyperventilation to PCO2 of 25 to 30 torr, head up position of 35 to 40 degrees.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Cooper 1979

---

Methods	Random allocation using identical coded vials prepared and supplied by a drug company, who retained the code until the 6 month outcome assessment had been made.
Participants	All patients with head injury admitted with a Grady Coma Grade of 3, 4 or 5. Glasgow Coma Scale also measured and in all except two patients GCS was 8 or less. Patients were excluded due to inability to obtain informed consent, previous administration of steroids or arrival at hospital more than 6 hours after the injury.
Interventions	High dose dexamethasone phosphate: 60mg initial dose, 24mg every 6 hours thereafter for 6 days. Low dose dexamethasone phosphate: 10mg initial dose, 4 mg every 6 hours thereafter for 6 days. Placebo (water, sodium bisulphite, methylparaben, propylparaben). All doses were contained in the same volume of fluid. Medication decreased gradually from day 7 and all were finished by day 11. Children aged 16 or younger given weight related doses.
Outcomes	GOS at 6 months after injury.
Notes	CT scan used or diagnosis in most patients followed by operation for mass lesions. Other co-interventions not specifically described.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

CRASH 2005

---

Methods	Stratified random allocation via central randomisation service or identical coded treatment packs.
Participants	Adults (16 or older) less than 8 hours after head trauma with GCS of 14 or less on admission to hospital.
Interventions	High dose methylprednisolone for 48 hours. 2g over 1 hour, then 0.4g per hour for 48 hours. Identical placebo.
Outcomes	Mortality and GOS at 6 months. Complications (admission to intensive care, gastrointestinal bleeding, infections, neurosurgical operation).
Notes
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Dearden 1986

---

Methods	Randomly allocated to receive treatment or placebo.
Participants	All ages, patients with severe head injury (no definition given). Exclusions were due to incorrect administration of steroid or placebo during the trial and those receiving steroids before admission.
Interventions	Dexamethasone: initial dose 50mg IV (0.75mg/kg for children), days 1, 2, 3 100mg/day IV, day 4 50mg/day IV, day 5 25mg/day IV. Placebo.
Outcomes	Glasgow Outcome Scale at 6 months after injury. Duration of artificial ventilation.
Notes	Other interventions guided by CT and ICP monitoring. Controlled hypocapeic ventilation and osmotic diuretics used to control ICP.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Faupel 1976

---

Methods	Randomly allocated using identical, coded vials containing the same volumes of clear solution.
Participants	All adult patients with severe (not defined) closed head injury. Exclusions were due to being a child (age cut off not given), impression fractures, missile injuries, imminent brain death and open head injuries. After randomization, 3 patients considered to be brain dead on angiography, one given steroids outside the trial and one who died of other injuries were excluded.
Interventions	Dexamethasone: initial dose 100mg IV, then 100mg IM after 6 hours, then 4mg IM every 6 hours for 8 days, then tapered off by daily reduction of 4mg. Dexamethasone: initial dose 12mg IV, then 4mg IM every 6 hours for 8 days, then tapered off by daily reduction of 4mg. Placebo.
Outcomes	Disability status scale at discharge from hospital. Frequency of complications.
Notes	Mass lesions removed following angiography. Other interventions not specified.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Gaab 1994

---

Methods	Patients were randomized in blocks of 6. Stated to be double blind design.
Participants	Age 15 to 55 with moderate central nervous system injury (two or more of disturbed consciousness, eye opening to stimulation, no adequate verbal response, disorientation in time and place) or severe central nervous system injury (comatose on injury and/or on admission to hospital, no eye opening to painful stimuli). Patients also were required to have obvious neurological symptoms (e.g. hemiparesis or hemiplegia) or CT evidence of lesions requiring surgical intervention or hypodense area(s) in the brain. Reasons for exclusion were: time to treatment was more than 3 hours, they had already had steroids, penetrating head injury, primary bulbar symptoms present, prognosis considered hopeless, known malignancy, peptic ulcer, tuberculosis, Cushing's syndrome, non traumatic neurological or psychiatric disease, spinal injury, suspected coagulation defects.
Interventions	Dexamethasone: initial dose 200mg IV, then 300mg over next 3 hours, then 200mg 3 hours later, then 200mg 6 hourly for 8 doses. Placebo.
Outcomes	Modified GOS at 10 to 14 months (good recovery classified as fit for work or for rehabilitation, extra category indicating degree of care required for disabled). GCS on day 5. Interval to regaining consciousness.
Notes	Co-interventions are not specified, but the study states that other care was not controlled.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Giannotta 1984

---

Methods	Randomly allocated to treatment or placebo using identical, coded vials. Code not broken until all outcomes had been collected.
Participants	Patients with blunt head trauma and a GCS of 8 or less 6 hours after the injury. Exclusions were due to history of peptic ulcer, an undiagnosed or untreated medical condition or who had been taking steroids during the two weeks before injury, penetrating brain injuries, other injuries expected to cause rapid death, and pregnancy.
Interventions	Methylprednisolone 30mg/kg IV every 6 hours for 2 doses, then 250mg IV every 6 hours for 8 doses, then tapering off over the next 8 days. Methylprednisolone 1.5mg/kg IV every 6 hours for 2 doses, then 25mg IV every 6 hours for 8 doses, then tapering off over 8 days. Placebo. NB Patients randomized in 2:2:1 ratio for these groups.
Outcomes	Glasgow Outcome Scale at 6 months.
Notes	CT guided management. Controlled ventilation, cerebrospinal fluid drainage, osmotic diuretics and barbiturates were used to control ICP.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Grumme 1995

---

Methods	Random allocation to treatment or control using identical vials.
Participants	Patients of all ages admitted with head injury. Excluded groups were those with contraindication to steroids (not specified), impaired level of consciousness not due to trauma and absence of relevant brain damage (not specified).
Interventions	Triamcinolone acetonide: 200mg IV, then 40mg 8 hourly for 4 days, then 20mg 8 hourly for 4 days. Placebo.
Outcomes	GOS at discharge from hospital and at approximately 1 year from the injury.
Notes	CT was used to guide treatment. Osmotic diuretics, controlled ventilation and cerebrospinal fluid drainage were used to control ICP.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Hernesniemi 1979

---

Methods	Random allocation using sealed opaque envelopes. Double blind.
Participants	Age 15 or above with severe closed brain injury. Five exclusions; three not head injury, one 14 years old, one imminent death.
Interventions	Betamethasone 100mg IV on admission, then 80mg/day IV for 7 days, then tapering off over further 7 days. Placebo.
Outcomes	Glasgow Outcome Scale at 6 or 12 months. Complications.
Notes
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Hoyt 1972

---

Methods	Assigned "sequentially in random order".
Participants	Patients with cranial trauma.
Interventions	Dexamethasone, dose not stated. Triamcinolone, dose not stated. Placebo.
Outcomes	Proportion demonstrating a "marked improvement".
Notes	Unable to contact author for further details.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Pitts 1980

---

Methods	"Prospectively randomized" into three groups.
Participants	Head injured adults admitted to hospital who were comatose on admission or who lapsed into coma for six hours or more.
Interventions	Dexamethasone: initially 24mg IV per day, for a maximum of 7 days, tapering of for the last 5. Dexamethasone: initially 16mg IV daily, for a maximum of 7 days, tapering off for the last 5. Placebo.
Outcomes	Glasgow Outcome Scale at 6 months. Gastrointestinal bleeding. Infections (signs of pneumonia, wound infection, urinary infection, blood culture, CSF culture).
Notes
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Ransohoff 1972

---

Methods	Stated to be "random steroid studies" and a double blind study of steroids against placebo.
Participants	Critically ill acute closed head injury patients without evidence of angiographic shift or significant clots, but with documented increase in intracranial pressure. Age not stated.
Interventions	Methylprednisolone 125mg IV every 6 hours for 4 days, starting within 24 hours of admission. Placebo.
Outcomes	Death. Complications (no data presented).
Notes	Also received fluid restriction.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Saul 1981

---

Methods	States that 100 patients were "randomized into two groups of 50". One group received steroid, there was no placebo.
Participants	Patients admitted with craniocerebral trauma within 6 hours of injury. No other body systems injured, Glasgow Coma Scale 7 or less on admission. Average age 31 years.
Interventions	250mg methylprednisolone IV initially, then 125mg every 6 hours. Some in steroid group received dexamethasone in equivalent dose. No steroid, no placebo.
Outcomes	Glasgow Outcome Scale at 6 months.
Notes	Also received mechanical hyperventilation and surgery if indicated.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Stubbs 1989

---

Methods	"Randomized and double-blind" study. Allocation concealment not described. Separate randomization for patients over and under 40 years old.
Participants	Patients were more than 6 years old with closed head trauma within the previous 48 hours and a GCS of 9 to 12. The first dose had to be administered within 6 hours of the GCS measurement. Exclusions were pregnancy, hypersensitivity to steroids and presence of infectious disease.
Interventions	High dose MPSS: 30mg/kg IV twice in 6h, then 250mg IV every 6h until 48 h after the first dose. Then a tapering schedule: days 3 to 5 25mg IV/IM 6 hrly, day 6 10mg IV/IM 6 hrly, day 7 5mg IV/IM 6 hrly days 8 and 9 10mg IV/IM daily. Low dose MPSS: 1mg/kg IV twice in 6h followed by 25mg IV every 6h until day 6. Then the tapering schedule above. Placebo
Outcomes	Glasgow Outcome Scale, Glasgow Coma Scale, Karnofsky Rating Scale at discharge, 3 and 6 months. Complications listed.
Notes	Glasgow Outcome Scale reported as mean score. Attempting to locate authors for further information.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Tahara 1972

---

Methods	Random allocation to one of three groups.
Participants	Seriously head injured.
Interventions	Prednisolone 2,680mg over 2 weeks. Prednisolone 160mg on first day, tapering over two weeks, total dose 1,000mg. No steroid.
Outcomes	None reported.
Notes	Trial of 100 participants; author contacted but unable to provide details.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

Zagara 1987

---

Methods	Randomization by table into steroid or no therapy groups.
Participants	Average age 29. Severe isolated head trauma with average GCS of 5.7 (sd 1.2) in one group and 5.8 (sd 1.2) in the other.
Interventions	Dexamethasone 0.36mg/kg/day IV for 9 days. No steroid, no placebo mentioned.
Outcomes	Glasgow Outcome Scale at 3 months. Nitrogen balance.
Notes	Also received mechanical hyperventilation, surgery if required, mannitol infusion and benzodiazepine sedation.
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Yes	Adequate

---

Zarate 1995

---

Methods	"Randomly allocated", no further details given. 60 patients, 30 in each group.
Participants	Children admitted with head injury with a Glasgow Coma Scale of 9 to 15.
Interventions	Corticosteroids - no details given. Symptomatic treatment.
Outcomes	Mortality at discharge.
Notes
Risk of bias
Item	Authors' judgement	Description
Allocation concealment?	Unclear	Unclear

---

 CT: computerised tomography
GCS: Glasgow Coma Scale
CSF: Cerebrospinal Fluid
GOS: Glasgow Outcome Scale
ICP: intracranial pressure
IM: Intramuscular
IV: Intravenous
MPSS: methylprednisolone sodium succinate

 

Characteristics of excluded studies [ordered by study ID]

---

Study	Reason for exclusion
Cheng 1991	Randomised trial of high versus low dose steroids. No group with no steroids.
Fanconi 1988	Judged to have inadequate allocation concealment after contact with authors.
Gobiet 1976	Study was retrospective.
James 1979	Allocation consisted of the first four patients being given no or low dose steroids and the next five given high dose steroids. There was therefore no concurrent control group and no concealment of allocation.
Robertson 1985	Patients were alternately allocated.

---

 

Comparison 1. Any steroid administered in any dose against no steroid

---

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 Death at end of follow up period	17		Risk Ratio (M-H, Fixed, 95% CI)	Totals not selected
2 Death or severe disability at the end of the study period	10		Risk Ratio (M-H, Fixed, 95% CI)	Totals not selected
3 Infections of any type	5	10798	Risk Ratio (M-H, Fixed, 95% CI)	1.03 [0.99, 1.07]
4 Major or significant gastrointestinal bleed	10	11302	Risk Ratio (M-H, Fixed, 95% CI)	1.23 [0.91, 1.67]

---