Exercise: Meta-analysis of endarterectomy versus stenting for carotid artery stenosis using CMA2

Stenosis or narrowing of the carotid artery in the neck greatly increases the risk of stroke by blocking the flow of blood to the brain. The standard treatment is to graft a piece of vein in place of the blocked section of artery, endarterectomy. Most of the patients are frail and elderly. This is an invasive procedure carried out under general anaesthesia and it might be preferable if this can be avoided. Several trials have been done of angioplasty, where a wire is inserted into the artery and use to guide an inflatable balloon to the site of the stenosis. There it can be inflated to widen the narrowed artery. To keep the artery open, a stent is often left in place. This is a wire mesh tube which sits in the artery.

There are short-term risks to both procedures and there may be peri-operative deaths related to anaesthesia problems and there may be strokes because a piece of plaque is dislodged from the artery and blocks a blood vessel leading to the brain.

Meier and colleagues (2010) reported a meta-analysis of randomised trials comparing endarterectomy versus stenting for carotid artery stenosis to prevent death or stroke.

They identified ten trials which provided data on the short-term outcome of death or stroke, and one of their meta-analyses is shown in the following figure, taken from the BMJ:


(CAVATAS and ICSS are my own trials — Martin.)

In this exercise we shall use CMA2 to carry out this meta-analysis.

Start CMA2.


Question 1: How should we put in the data from the BMJ graph? Answer:

We first need to put in the study names. Click "Insert" "Column for" "Study names". Type the names in the column. We next put in the data. Click "Insert" "Column for" "Effect size data". A dialogue box appears. Click "Next". We have two groups, endarterectomy and stenting, so click the button for "Comparison of two groups ...", click "Next". Our data are dichotomous, so click that, and unmatched and prospective, so click that. We have number of events and sample size in each group, so click that. We are ready to go, so click finish. You can put in the names for the two groups, "Endarterectomy" and "Stenting", but CMA2 does not seem to use them for anything. Now put in the data for each study.

Question 2: On the CMA2 data entry screen, what is missing for the study of Brooks et al. (2004)? Why is it missing?

Check suggested answer 2.

Question 3: On the CMA2 analysis screen, what is missing for the study of Brooks et al. (2004)? Why is it missing?

Check suggested answer 3.

Question 4: How can we same this output onto a file for future use?

Check suggested answer 4.

Question 5: Is the default fixed effects model the right one to use?

Check suggested answer 5.

Draw the high resolution forest plot.

Question 6: How could we improve the forest plot?

Check suggested answer 6.

Question 7: How could we improve the forest plot further?

Check suggested answer 7.

Question 8: How could we label the combined estimate as "Fixed" on the forest plot?

Check suggested answer 8.

Question 9: How could we change the combined estimate to the random effects estimate and label it "Random" on the forest plot?

Check suggested answer 9.

Question 10: How can we save the forest plot image for future use in a word processing file?

Check suggested answer 10.

Question 11: How could we check for any evidence of publication bias?

Check suggested answer 11.

In the CAVATAS/ICSS team, the hypothesis was put forward that, in the time between these two studies, surgical techniques had improved. This had increased the difference between endarterectomy and stenting in the 30 day follow-up.

Question 12: What method could we use to examine the effect of callendar time on treament effect?

Check suggested answer 12.

Question 13: How can we do this in CMA2?

Check suggested answer 13.


Meier P, Knapp G, Tamhane U, Chaturvedi S, Gurm HS. Short term and intermediate term comparison of endarterectomy versus stenting for carotid artery stenosis: systematic review and meta-analysis of randomised controlled clinical trials BMJ 2008; 340: c467.

To Systematic Reviews index.

To Martin Bland's M.Sc. index.

To Martin Bland's home page.

This page maintained by Martin Bland.
Last updated: 17 February, 2010.

Back to top.