Skip to main content
NCBI home page
The BMJ logoLink to The BMJ
. 2000 Feb 12;320(7232):446. doi: 10.1136/bmj.320.7232.446

Treatment of venous leg ulcers

Nice study, pity about the sample size

Frank Dobbs 1
PMCID: PMC1117556  PMID: 10669458

Editor—Dale et al based their calculation of the sample size (200 patients) on the minimum size required to find the large effect observed in a previous very small study.1 The confidence interval for this large effect would have been very large.

A better strategy would have been to decide on a detectable difference that would be clinically significant and design the study to be capable of detecting this difference. The difference that they found (64% healing with pentoxifylline v 53% healing with placebo) would probably be clinically significant in view of the high material and labour cost of continuing treatment with pressure bandaging and the unpleasantness of leg ulcers.

The study described only had a 30% power to detect this magnitude of difference. To have an 80% power to detect this difference would require a study with 332 in each group. We need larger groupings than single hospital clinics for research into conditions like this. The rapidly developing primary care research networks are ideal structures for research into common conditions that are usually managed in the community.

References

  • 1.Dale JJ, Ruckley CV, Harper DR, Gibson B, Nelson EA, Prescott RJ. Randomised, double blind placebo controlled trial of pentoxifylline in the treatment of venous leg ulcers. BMJ. 1999;319:875–878. doi: 10.1136/bmj.319.7214.875. . (2 October.) [DOI] [PMC free article] [PubMed] [Google Scholar]
BMJ. 2000 Feb 12;320(7232):446.

Authors' reply

C V Ruckley 1, R J Prescott 1, J J Dale 1

Editor—We note with interest Dobbs's comments on our trial and agree that primary care research networks are promising vehicles with which to conduct trials of conditions normally managed in the community. We are, however, unconvinced by his comments concerning the trial size in relation to our particular study.

Our trial was of factorial design in order to compare three types of treatment within a single trial, in order to derive the maximum amount of data from the minimum number of patients. The treatments compared were: two types of dressing; two types of bandage; and a pharmacological treatment (pentoxifylline) against placebo.1-1 The paper under discussion presented only the pharmacological data.

Trial size is always a compromise between sophistication of design, cost, and the difference we would like to detect. Patients with leg ulcers constitute a highly heterogeneous group. The recruitment of large numbers of adequately assessed patients is very difficult. Our trial size and 20% margin of benefit were carefully derived from the known outcomes of alternative treatments, such as bandaging regimens.

The trial required precise definition of patient groups by means of detailed patient screening including vascular laboratory tests and duplex scanning. Altogether 525 patients were screened to arrive at our study population of 200. Also required were precise standardisation of treatment methods by specialist nurses and tight day to day supervision of the protocol. The cost of an equivalent trial in 634 patients conducted in the community as suggested by Dobbs would have been enormous, and funding would have been unlikely, even by a pharmaceutical company. Furthermore, if such a large number of patients were required to demonstrate a benefit it would raise serious questions about the cost effectiveness of such a treatment in this particular group of patients.

References

  • 1-1.Prescott RJ, Nelson EA, Dale JJ, Harper DR, Ruckley CV. Design of randomised controlled trials in the treatment of leg ulcers: more answers with fewer patients. Phlebology. 199;13:107–112. [Google Scholar]

Articles from BMJ : British Medical Journal are provided here courtesy of BMJ Publishing Group

RESOURCES