Megadose steroids in severe head injury

Results of a prospective double-blind clinical trial

Full access

✓ A prospective double-blind clinical trial was performed on 161 patients to determine the effectiveness of high-dose steroid therapy in patients admitted comatose after a non-missile-related head injury. Patients were randomized into a high-dose dexamethasone phosphate group and a placebo group. The initial dose of 100 mg of dexamethasone was administered within 6 hours of the accident. For statistical analysis, a sequential test was chosen, using survival at 1 month as a basic criterion of effectiveness. No significant difference was found in the 1-month survival rate or in the distribution of outcome after 6 months, either within the group as a whole, or in subgroups with varying severity of brain damage on admission. The authors conclude that dexamethasone in high doses has no statistically significant effect on morbidity or mortality in head-injured patients who are comatose on admission.

Abstract

✓ A prospective double-blind clinical trial was performed on 161 patients to determine the effectiveness of high-dose steroid therapy in patients admitted comatose after a non-missile-related head injury. Patients were randomized into a high-dose dexamethasone phosphate group and a placebo group. The initial dose of 100 mg of dexamethasone was administered within 6 hours of the accident. For statistical analysis, a sequential test was chosen, using survival at 1 month as a basic criterion of effectiveness. No significant difference was found in the 1-month survival rate or in the distribution of outcome after 6 months, either within the group as a whole, or in subgroups with varying severity of brain damage on admission. The authors conclude that dexamethasone in high doses has no statistically significant effect on morbidity or mortality in head-injured patients who are comatose on admission.

There are conflicting opinions about the effectiveness of corticosteroid therapy in severely head-injured patients. Gobiet, et al.,7 in their studies of head injury, could not detect any significant beneficial effect with conventional doses of dexamethasone (16 mg/day). However, following administration of high doses (48 mg initially, and a total dose of 96 mg during the 1st day), they achieved a significant reduction in mortality, in the complication rate, and in the frequency with which intracranial pressure (ICP) rose. In their study, the control group was not managed concurrently, but consisted of patients treated in previous years. On the basis of a double-blind trial, Faupel, et al.,6 reported a favorable dose-related effect on mortality, even in patients who had received treatment some 6 hours or more after the accident. A few years earlier, Gutterman and Shenkin9 and Alexander1 had denied any beneficial effect from the administration of conventional doses of dexamethasone phosphate (16 mg/day).

We considered it important that the controversy be resolved. Therefore, we initiated a prospective double-blind controlled clinical trial of the effectiveness of high-dose steroids in head-injured patients who were comatose on admission. This paper presents our findings.

Clinical Material and Methods
Patient Population

The study includes 161 patients with a severe, non-missile-related head injury, who were in coma when admitted to the Department of Neurosurgery of the Academic Hospital Dijkzigt, Erasmus University, Rotterdam, or the Department of Neurology, University of Groningen between 1978 and 1981. Coma was defined as a state in which the patient did not open his eyes, did not utter words to standardized painful stimuli, and did not obey commands. We excluded patients who were already brain-dead or were expected to be so within 1 hour, as determined by apnea, flaccidity, dilated pupils not reacting to light, and the absence of reflex eye movements. We also excluded a few patients who recovered consciousness during the initial examination in the emergency room. Some patients had received steroid therapy prior to the initiation of management in our hospital (for example, in a referring hospital or in the ambulance); these patients were not entered into the study, irrespective of the dose of steroid they had received. The administration of steroids (or placebo) was stopped in three patients who later appeared to be suffering from diabetes mellitus or who had a history of peptic ulcer, and these patients were also removed from the study. The study includes only those patients who were given their first intravenous dose of either steroid or placebo within 6 hours of the accident; in most patients it was within 3 hours.

Management

All patients were examined and resuscitated in the emergency room. Subsequently, computerized tomography scans were obtained, usually within 1 hour of admission. Patients with mass lesions (midline shift in excess of 5 mm) usually had an immediate operation via a bone flap. We did not make burr holes in the emergency room. In most patients, ICP was recorded continuously. In patients who did not have a mass lesion, increased pressure was managed by controlled ventilation and/or the administration of an osmotic diuretic, and occasionally by drainage of cerebrospinal fluid. High doses of barbiturates were not used.

To the best of our knowledge no major changes in management protocol occurred during the period of this study, either in Rotterdam or in Groningen. Nevertheless, management in Rotterdam differed from that in Groningen in that in Rotterdam more patients were artificially ventilated and more had their ICP measured.

Randomization to Groups

Patients were randomized into two groups: a placebo group and a high-dose dexamethasone group. The regimen for administration of dexamethasone* in adults was as follows: initial dose: 100 mg intravenously; Days 1 to 4: 100 mg/day intravenously; Days 5 to 7: 16 mg/day intravenously or intramuscularly; and Days 8, 9, and 10: 12 mg, 8 mg, and 4 mg/day, respectively, intravenously or intramuscularly. Children between 10 and 14 years of age were given half the adult dose, and those under 10 years of age received 25% of the adult dose. All patients received a daily 2000-cc intravenous infusion of 2.5% glucose and 0.45% NaCl in solution, unless abnormalities in electrolyte levels or other factors demanded another fluid regime.

The vials with dexamethasone and placebo were indistinguishable and were kept in identical, numbered boxes, one box for each patient, ready for use. The vials were prepared by the Department of Pharmacy, Erasmus University, Rotterdam.

Statistical Analysis

We decided in advance that the number of patients studied should depend on the effectiveness of the high-dose steroids: a strong effect would result in a short trial. To achieve this aim we used a sequential statistical test, as described by Armitage2 (Fig. 1). Survival at 1 month was taken as the measure of effectiveness of high-dose steroids.

Fig. 1.
Fig. 1.

Sequential scheme for this high-dose corticosteroid trial. When the upper (or lower) boundary is reached, the administration of high-dose dexamethasone results in a significantly higher (or lower) 1-month survival rate. For further description, see the text.

The sequential scheme was used as follows: The first patient on steroids was compared with the first placebo patient, and so on, with the Rotterdam and Groningen patients being evaluated separately. The course of the trial is represented in Fig. 1 by a zig-zag line indicating the cumulative results in successive pairs of patients. Thus, if after 1 month the patient on steroids was still alive but the placebo patient had died, there was preference for high-dose steroids. This is represented by extending the curve in the graph one unit to the right in an upward direction. On the other hand, if the patient on steroids died but the placebo patient survived, there was a preference for placebo, represented by an extension of the graph one unit to the right and downward. If both patients of a pair survived or died, this pair was not included in the sequential scheme.

Under the “null” hypothesis that high doses of steroids and placebo have exactly the same effect, there is 1) a 2.5% probability of reaching the upper boundary, indicating that high doses of steroids result in a significantly higher survival rate; 2) also a 2.5% probability of reaching the lower boundary, which would indicate that high doses of steroids result in a significantly lower survival rate; and 3) a 95% probability of reaching the middle boundary (Fig. 1), where no significant difference is found. The trial was to be stopped if one of these three boundaries was reached. In practice, the trial was stopped when it appeared impossible to reach the upper boundary.

We assessed outcome of each individual patient at 6 months, according to the Glasgow Outcome Scale.10 Complications were also recorded. Furthermore, on admission, we calculated for each patient the probability of surviving 6 months (Psurv). The 161 patients were categorized into five subgroups according to their estimated severity of brain damage; those with a Psurv of < 0.10, 0.10 to 0.30, 0.30 to 0.70, 0.70 to 0.90, and > 0.90. This calculation was based on the admission scores for age, pupil reactivity to light, and best motor response of the arms. These features had proven to be the most useful combination in predicting outcome at 6 months in 305 patients who had been treated previously in either Rotterdam or Groningen.4 This procedure allowed us to check whether the distribution of severity of brain damage was similar in both groups. The Mann-Whitney U-test was performed separately for the Rotterdam and Groningen patients; after which, both tests were combined. The procedure also provided an opportunity to compare the mortality in the subgroups with differing degrees of injury severity. The Mantel-Haenszel test was used to combine evidence from two by two tables.

The code of the trial remained with the statistician until the 6-month outcome of the last patient was known.

Results

When the trial was stopped, 161 patients had been entered into the study, 120 from Rotterdam and 41 from Groningen. No significant difference was found in the survival rate between the steroid and placebo groups, either in Rotterdam or in Groningen (Table 1), or in the outcome at 6 months after injury in the two groups. The number of deaths in the placebo group was slightly higher than in the steroid group, but conversely the number of dependent patients (vegetative and severely disabled) was higher in the steroid group than in the placebo group (Table 2).

TABLE 1

One-month survival in the steroid and placebo groups in this study*

Status at 1 MonthRotterdam GroupsGroningen Groups
SteroidPlaceboSteroidPlacebo 
alive
 no.28311210
 percent47525759
dead
 no.3229910
 percent53484341
total cases60602120

There was no significant difference in survival according to the Mantel-Haenszel test (x2 = 0.01).

TABLE 2

Outcome at 6 months following injury in the steroid and placebo groups in this study*

Outcome at 6 MonthsRotterdam GroupsGroningen Groups
SteroidPlaceboSteroidPlacebo 
recovered8864
moderately disabled101004
severely disabled8431
vegetative0121
dead34371010
total60602120

There was no significant difference in outcome at 6 months between the steroid and placebo groups according to the Mann-Whitney U-test of standard normal values (Rotterdam: z = 0.40; Groningen: z = 0.10).

There was no difference in distribution of severity of injury between the steroid and placebo groups and no significant difference in the 6-month survival rate in any of the subgroups (Table 3). Only two patients suffered from a major episode of gastrointestinal bleeding which required blood transfusion; one in the steroid group and one in the placebo group. The number of cases of severe pulmonary infection developing 3 days or more after the injury was higher in the steroid group than in the placebo group, but this difference was not statistically significant (p = 0.07, Mantel-Haenszel test) (Table 4). The effect on ICP could not be studied reliably, because not all patients were subjected to ICP recording. The effects of steroids on electrolytes, levels of blood sugar, prothrombin time, and other aspects of clotting did not form part of this study.

TABLE 3

Survival at 6 months compared with the probability of survival assessed on admission*

Survival at 6 MonthsProbability of Survival (Psurv)
 < 0.100.10–0.300.30–0.700.70–0.90> 0.90Total
SteroidPlaceboSteroidPlaceboSteroidPlaceboSteroidPlaceboSteroidPlaceboSteroidPlacebo 
Rotterdam
 alive0001651515522623
 total9119717191819746060
Groningen
 alive10114242151110
 total21568553152120
Rotterdam & Groningen
 alive10121071917673733
 total1112141325242322898180

Mantel-Haenszel test: x2 = 0.08.

TABLE 4

Number of patients with complications developing more than 72 hours after injury*

ComplicationsRotterdam GroupsGroningen Groups
SteroidPlaceboSteroidPlacebo 
gastrointestinal bleeding1100
severe pulmonary infection15710
total alive after 72 hours40381210

Mantel-Haenszel test: x2 = 3.18; p = 0.07.

Discussion

The study shows that administration of high-dose steroids within 6 hours after injury does not influence outcome of patients with head injuries from causes other than missiles who were comatose on admission. There is no significant difference in the 1-month survival rate, or in the distribution of outcome after 6 months, either in the group as a whole or in subgroups with varying severity of brain damage on admission. The number of severe pulmonary infections appearing 3 days or more after injury is, however, higher in the steroid group, although the difference is not statistically significant. Moreover, we did not define in advance the criteria for “severe pulmonary infection,” and classified these patients at the end of the study, albeit without the information as to whether they had received steroids or a placebo.

In this study, 1-month survival was taken as the basic criterion by which effectiveness of high-dose steroids was measured. In many other trials, some interim analyses have been performed to determine whether an observed difference is statistically significant. This may be very misleading if one does not use a sequential test.2,3 Toward the end of the trial, some physicians became impatient, as it was annoying not knowing when the trial would end. This seems to be the only serious disadvantage of sequential tests.

Since the reports of Gobiet, et al.,7 and Faupel, et al.,6 other studies on the effectiveness of corticosteroids have been published, during the course of our present study. Gudeman, et al.,8 reported that they were unable to detect a significant reduction in ICP and cerebral elastance in 20 consecutive patients with severe head injury who received high-dose methylprednisolone. Saul, et al.,11 found that high-dose methylprednisolone did not reduce the frequency of high levels of ICP or improve outcome. These authors did suggest that different patient groups might have variable responses to steroids. They recommended the initiation of steroid therapy in all severely head-injured patients, and suggested that it should be continued if after 3 days a patient's response, based on a sensitive neuro-index, seemed to be favorable. Our study did not support this view: in the patients with less severe initial brain damage, there was no difference in outcome between the steroid and placebo groups.

Cooper, et al.,5 performed a prospective double-blind study of 76 patients and concluded that dexamethasone in either high or low doses had no significant effect on morbidity or mortality following severe head injury. They stressed that human head injury is a more complex process than that produced in experimental models, from which most favorable results have been reported. In their view, a bad outcome was commonly due to severe diffuse injuries with gross anatomical disruption not amenable to any treatment, including the administration of steroids.

Another prospective double-blind trial of 260 patients established that there was not a significant difference in the mortality, morbidity, or occurrence of complications between groups receiving placebo, conventional dose, or high-dose steroids: cessation of corticosteroid therapy in these patients was advised (Pitts LH, et al.: unpublished data, 1981). On the basis of our results, we too think that administration of high-dose or low-dose corticosteroids in patients with severe head injury is not warranted.

Researchers who in the future want to perform trials to determine the efficacy of other drugs or other aspects of management in severely head-injured patients might consider omitting from their studies patients who can clearly be predicted to die or to survive regardless of treatment.12 These patients can be identified on admission by using the scores of age, pupil reactivity to light, and best motor response of the arms. One might also omit patients who have a definitely good or bad outlook, (that is, those with a probability of survival (Psurv) of > 0.70 or < 0.30). Such trials should involve various centers, and the efforts should be directed toward the patients whose chances of dying are between 30% and 70%, as predicted on admission.

Acknowledgments

The following of our co-workers also took part in this study: C. J. J. Avezaat, J. H. van den Berge, A. I. R. Maas, and T. Segeren. We thank Professor Staal and his colleagues, Department of Neurology, Erasmus University, Rotterdam, and Professor Minderhoud's co-workers from University Hospital, Groningen, who entrusted their patients to us. We are also grateful to the nursing staffs of our intensive care units. We thank Dr. P. J. H. Eggels from the Department of Pharmacy, Erasmus University, who prepared the vials of dexamethasone. Thanks are also due Mr. Graham Teasdale for criticism of the manuscript, to Mrs. B. S. Vollers-King for editorial help, and to Miss B. A. A. Koehorst who typed the manuscript.

References

  • 1.

    Alexander E Jr: Medical management of closed head. injuries. Clin Neurosurg 19:2402501972Alexander E Jr: Medical management of closed head. injuries. Clin Neurosurg 19:

  • 2.

    Armitage P: Sequential Medical Trials. London: Blackwell19756667Armitage P: Sequential Medical Trials.

  • 3.

    Armitage P: Statistical Methods in Medical Research. London: Blackwell1971Chap 15Armitage P: Statistical Methods in Medical Research.

  • 4.

    Braakman RGelpke GJHabbema JDFet al: Systematic selection of prognostic features in patients with severe head injury. Neurosurgery 6:3623701980Neurosurgery 6:

  • 5.

    Cooper PRMoody SClark WKet al: Dexamethasone and severe head injury. A prospective double-blind study. J Neurosurg 51:3073161979J Neurosurg 51:

  • 6.

    Faupel GReulen HJMüller Det al: Double-blind study on the effects of steroids on severe closed head injuryPappius HMFeindel W (eds): Dynamics of Brain Edema. Berlin/Heidelberg/New York: Springer-Verlag1976337343Dynamics of Brain Edema.

  • 7.

    Gobiet WBock WJLiesegang Jet al: Treatment of acute cerebral edema with high dose of dexamethasoneBeks JWFBosch DABrock M (eds): Intracranial Pressure III. Berlin/Heidelberg/New York: Springer-Verlag1976231235Intracranial Pressure III.

  • 8.

    Gudeman SKMiller JDBecker DP: Failure of high-dose steroid therapy to influence intracranial pressure in patients with severe head injury. J Neurosurg 51:3013061979J Neurosurg 51:

  • 9.

    Gutterman PShenkin HA: Prognostic features in recovery from traumatic decerebration. J Neurosurg 32:3303351970J Neurosurg 32:

  • 10.

    Jennett BBond MR: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:4804841975Lancet 1:

  • 11.

    Saul TGDucker TBSalcman Met al: Steroids in severe head injury. A prospective randomized clinical trial. J Neurosurg 54:5966001981J Neurosurg 54:

  • 12.

    Teasdale G: Prognosis of coma after head injuryWMG Tunbridge (ed): Advanced Medicine. London: Pitman1981106116Advanced Medicine.

Dexamethasone phosphate was kindly supplied by Merck, Sharp and Dohme, The Netherlands.

Article Information

Address reprint requests to: Reinder Braakman, M.D., Department of Neurosurgery, Academisch Ziekenhuis Rotterdam-Dijkzigt, Erasmus University, 3015 GD Rotterdam, The Netherlands.

© AANS, except where prohibited by US copyright law.

Headings

Figures

  • View in gallery

    Sequential scheme for this high-dose corticosteroid trial. When the upper (or lower) boundary is reached, the administration of high-dose dexamethasone results in a significantly higher (or lower) 1-month survival rate. For further description, see the text.

References

1.

Alexander E Jr: Medical management of closed head. injuries. Clin Neurosurg 19:2402501972Alexander E Jr: Medical management of closed head. injuries. Clin Neurosurg 19:

2.

Armitage P: Sequential Medical Trials. London: Blackwell19756667Armitage P: Sequential Medical Trials.

3.

Armitage P: Statistical Methods in Medical Research. London: Blackwell1971Chap 15Armitage P: Statistical Methods in Medical Research.

4.

Braakman RGelpke GJHabbema JDFet al: Systematic selection of prognostic features in patients with severe head injury. Neurosurgery 6:3623701980Neurosurgery 6:

5.

Cooper PRMoody SClark WKet al: Dexamethasone and severe head injury. A prospective double-blind study. J Neurosurg 51:3073161979J Neurosurg 51:

6.

Faupel GReulen HJMüller Det al: Double-blind study on the effects of steroids on severe closed head injuryPappius HMFeindel W (eds): Dynamics of Brain Edema. Berlin/Heidelberg/New York: Springer-Verlag1976337343Dynamics of Brain Edema.

7.

Gobiet WBock WJLiesegang Jet al: Treatment of acute cerebral edema with high dose of dexamethasoneBeks JWFBosch DABrock M (eds): Intracranial Pressure III. Berlin/Heidelberg/New York: Springer-Verlag1976231235Intracranial Pressure III.

8.

Gudeman SKMiller JDBecker DP: Failure of high-dose steroid therapy to influence intracranial pressure in patients with severe head injury. J Neurosurg 51:3013061979J Neurosurg 51:

9.

Gutterman PShenkin HA: Prognostic features in recovery from traumatic decerebration. J Neurosurg 32:3303351970J Neurosurg 32:

10.

Jennett BBond MR: Assessment of outcome after severe brain damage. A practical scale. Lancet 1:4804841975Lancet 1:

11.

Saul TGDucker TBSalcman Met al: Steroids in severe head injury. A prospective randomized clinical trial. J Neurosurg 54:5966001981J Neurosurg 54:

12.

Teasdale G: Prognosis of coma after head injuryWMG Tunbridge (ed): Advanced Medicine. London: Pitman1981106116Advanced Medicine.

TrendMD

Metrics

Metrics

All Time Past Year Past 30 Days
Abstract Views 1 1 1
Full Text Views 84 84 19
PDF Downloads 56 56 38
EPUB Downloads 0 0 0

PubMed

Google Scholar