Rapid Responses to:

PRACTICE: John Fletcher What is heterogeneity and is it important? BMJ 2007; 334: 94-96 [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

There is no heterogeneity between steroids and antihistamines

Christopher J Cates   (14 January 2007)

Re: There is no heterogeneity between steroids and antihistamines

John Fletcher   (15 January 2007)

There is no heterogeneity between steroids and antihistamines 14 January 2007
Christopher J Cates, GP Manor View Practice, Bushey, WD23 2NN Send response to journal: Re: There is no heterogeneity between steroids and antihistamines	Editor, The heterogeneity documented by Tramer et al[1], is not between the subgroups treated with steroids and anti-histamines. The authors calculated the combined odds ratio for both groups as 0.36. These are shown as combined results for each subgroup in Figure 1, but for some reason the usual convention of a diamond for the combined results has not been used. The significant heterogeneity described in the legend refers to the within subgroup differences in the anti-histamine trials, and did not defer the original authors from combining the results for anti-histamines! The suggestion that heterogeneity between the classes of drugs was significant and that this dissuaded the authors from combining the two types of treatment cannot be correct. [1] Tramer MR, von Elm E, Loubeyre P, Hauser C. Pharmacological prevention of serious anaphylactic reactions due to iodinated contrast media: systematic review. 2006:675-8. Competing interests: None declared
Re: There is no heterogeneity between steroids and antihistamines 15 January 2007
John Fletcher, Clinical Epidemiologist British Medical Journal Send response to journal: Re: Re: There is no heterogeneity between steroids and antihistamines	I thank Christopher Cates for the clarification. The greatest apparent heterogeneity is on the effect of antihistamines on cutaneous symptoms and not between treatments as my wording implies. I wanted to illustrate when it is appropriate not to summarise effects across trials. The authors have not produced a summary measure for “the effect of pharmacological prevention of serious anaphylactic reactions due to iodinated contrast media”. This seems appropriate and the figure shows that there are a range of effects for hypotension, respiratory symptoms and cutaneous symptoms. Competing interests: None declared