Carlos
A good example. Best to think of this as a 2 step process:
1. Use the level to guide the *initial* search (Levels 1-4)
2. Appraise and Grade the evidence (which may be different to the
initial level - Grade High -> Low)
(See Atkins D, et al Grading quality of evidence and strength of
recommendations. BMJ. 2004 Jun 19;328(7454):1490.)
So in your case
1. You find a few trials, so then you check
2. Yes - there is a systematic review
So the "levels" tell us to look at the review first (level 1).
But the review needs to be appraised, and checked if it is up to date.
If it is poorly done or very out of date, we may go back and check the
trials.
In this case even without further meta-analysis, the SR + CRASH trial
gives a pretty clear answer.
So its Grade A evidence.
The "level" alone is *not* sufficient to grade the evidence - its just
for an initial sorting.
Best Wishes
Paul Glasziou
Dr. Carlos Cuello wrote:
> I would like to note the timing of doing the evaluation of the evidence.
> Consider the case of head injury and corticosteroids use. Ten years
> ago, as a novice resident I was looking for evidence on this topic and
> a meta-analysis could not give me a definitive answer. [Alderson P,
> Roberts I. BMJ 1997 Jun 28;314(7098):1855-9]. However, at that time,
> this would have been level I evidence.
>
> Consider the example below of the a large CRASH-RCT (#1) vs a
> systematic review including the RCT (#2).
>
> The RCT is much better for giving an answer than the previous and
> perhaps the updated SR, but the SR still should be on level 1... don´t
> you think?
>
>
>
> 1: Lancet. 2005 Jun 4-10;365(9475):1957-9.
>
> Final results of MRC CRASH, a randomised placebo-controlled trial of intravenous
> corticosteroid in adults with head injury-outcomes at 6 months.
>
> Edwards P, Arango M, Balica L, Cottingham R, El-Sayed H, Farrell B, Fernandes J,
> Gogichaisvili T, Golden N, Hartzenberg B, Husain M, Ulloa MI, Jerbi Z, Khamis H,
> Komolafe E, Laloë V, Lomas G, Ludwig S, Mazairac G, Muñoz Sanchéz Mde L, Nasi L,
> Olldashi F, Plunkett P, Roberts I, Sandercock P, Shakur H, Soler C, Stocker R,
> Svoboda P, Trenkler S, Venkataramana NK, Wasserberg J, Yates D, Yutthakasemsunt
> S; CRASH trial collaborators .
>
> MRC CRASH is a randomised controlled trial (ISRCTN74459797) of the effect of
> corticosteroids on death and disability after head injury. We randomly allocated
> 10,008 adults with head injury and a Glasgow Coma Scale score of 14 or less,
> within 8 h of injury, to a 48-h infusion of corticosteroid (methylprednisolone)
> or placebo. Data at 6 months were obtained for 9673 (96.7%) patients.
> The risk of
> death was higher in the corticosteroid group than in the placebo group (1248
> [25.7%] vs 1075 [22.3%] deaths; relative risk 1.15, 95% CI 1.07-1.24; p=0.0001),
> as was the risk of death or severe disability (1828 [38.1%] vs 1728 [36.3%] dead
> or severely disabled; 1.05, 0.99-1.10; p=0.079). There was no evidence that the
> effect of corticosteroids differed by injury severity or time since
> injury. These
> results lend support to our earlier conclusion that corticosteroids
> should not be
> used routinely in the treatment of head injury.
>
> Publication Types:
> Clinical Trial
> Multicenter Study
> Randomized Controlled Trial
> Research Support, Non-U.S. Gov't
>
> PMID: 15936423 [PubMed - indexed for MEDLINE]
>
> 2: Cochrane Database Syst Rev. 2005 Jan 25(1):CD000196.
>
> Update of:
> Cochrane Database Syst Rev. 2000;(2):CD000196.
>
> Corticosteroids for acute traumatic brain injury.
>
> Alderson P, Roberts I.
>
> UK Cochrane Centre, NHS R&D Programme, Summertown Pavilion, Middle Way, Oxford,
> UK, OX2 7LG. [log in to unmask]
>
> BACKGROUND: Traumatic brain injury is a leading cause of death and disability.
> Corticosteroids have been widely used in treating people with traumatic brain
> injury. OBJECTIVES: To quantify the effectiveness and safety of corticosteroids
> in the treatment of acute traumatic brain injury. SEARCH STRATEGY: Electronic
> sources: MEDLINE, EMBASE, Cochrane Library and specialised database searches.
> Additional hand searching and contact with trialists. Date of the most recent
> search October 2004. SELECTION CRITERIA: All randomised controlled trials of
> corticosteroid use in acute traumatic brain injury with adequate or unclear
> allocation concealment. DATA COLLECTION AND ANALYSIS: Quality of allocation
> concealment was scored. Data on numbers of participants randomised, numbers lost
> to follow up, length of follow up, case fatality rates, disablement, infections
> and gastrointestinal bleeds were extracted independently and checked. MAIN
> RESULTS: We identified 20 trials with 12303 randomised participants. The effect
> of corticosteroids on the risk of death was reported in 17 included trials. Due
> to significant heterogeneity we did not calculate a pooled estimate of the risk
> of death. The largest trial, with about 80% of all randomised
> participants, found
> a significant increase in the risk ratio of death with steroids 1.18 (1.09 to
> 1.27). For the nine trials that reported death or severe disability, the pooled
> relative risk was 1.01 (0.91 to 1.11), although this does not yet contain data
> from the largest trial. For infections the pooled risk ratio from five
> trials was
> 1.03 (0.99 to 1.07) and for the ten trials reporting gastrointestinal bleeding
> 1.23 (0.91 to 1.67). AUTHORS' CONCLUSIONS: In the absence of a meta-analysis, we
> feel most weight should be placed on the result of the largest trial. The
> increase in mortality with steroids in this trial suggest that
> steroids should no
> longer be routinely used in people with traumatic head injury.
>
> Publication Types:
> Review
>
> PMID: 15674869 [PubMed - indexed for MEDLINE]
>
>
>
>
--
Paul Glasziou
Director, Centre for Evidence-Based Medicine,
Department of Primary Health Care,
University of Oxford www.cebm.net
ph - +44-1865-289298 fax +44-1865-289287
|