Copyright © 2004 by the European Society of Cardiology.
Letter to the Editor
Meta-analysis of long-term studies to assess the effect of acarbose on cardiovascular risk reduction – scientifically credible: Reply
Zentrum fur Klinische Studien, Fiedlerstr. 34, GWT, TU Dresden, Dresden 01307, Germany
* Tel.: +49-351-4400-580; fax: +49-351-4400-581
E-mail address: hanefeld{at}gwt-tud.de
The authors criticised our meta-analysis of seven long-term studies to assess the effect of acarbose on cardiovascular risk reduction. Two different approaches are generally accepted as standard procedures for meta-analyses.1 These are systematic literature reviews or meta-analyses of all randomised controlled trials. Our meta-analysis is based on the latter approach. As clearly stated in our article, data from all randomised, double-blind, placebo-controlled clinical trials with a minimum treatment duration of 52 weeks were included. The source was the Bayer Acarbose clinical database. As a consequence, publication and selection bias (regarded as substantial risks in the situation of systematic literature review)2 can be excluded.
Whilst it is true that two studies included in the analyses were not formally published and therefore not subject to peer review, both studies were part of the New Drug Application for acarbose. Although we highly respect the quality and rigour of the science journal peer review procedure, we doubt that the drug approval process is any less strict.3 The study by Bachmann et al., has in the meanwhile recently been published.
We agree with Dr. van de Laar et al., that heterogeneity should not be investigated by statistical testing alone. We consider only heterogeneity in the treatment effect to be problematic for statistical analysis (in accordance with CPMP). Some deviations concerning other characteristics could be seen as an advantage for increasing its generalisation and external validity.2,4 This said, co-intervention was balanced between acarbose and placebo because of the randomisation. It should be mentioned that in addition to formal statistical testing, graphical and descriptive methods were also employed to confirm homogeneity of treatment effect.
The use of individual patient data has the advantage of enabling a common outcome variable to be defined across studies.4 Therefore it seems straightforward to use safety data to define an outcome based on cardiovascular events. Indeed, we are surprised to note the concerns of Dr. van de Laar et al., regarding the safety data collection procedure. As a matter of fact, in clinical trials performed according to Good Clinical Practice the occurrence of adverse events is monitored carefully and recorded in detail during the trial.
We would like to point out that in the Cox Proportional Hazards model, patients contribute only for the time that they are observed under treatment and only to the corresponding treatment group. Moreover, not only the actual number of events under treatment and placebo contribute to the analysis, but also the event-free times in both groups.
Finally, the `any cardiovascular event' outcome is indeed still significant, even after exclusion of myocardial infarctionfrom the cardiovascular events (HR=0.69, 95% CI: 0.50–0.95).
In summary, our meta-analysis was conducted based on scientifically sound and credible principles. Thus the data published in the European Heart Journal suggest that acarbose, in addition to its effect on glycemic control, could have a beneficial effect on cardiovascular complications in patients with established type 2 diabetes. These results are in line with the beneficial effect of acarbose on cardiovascular events, as found in the STOP-NIDDM trial.5
References
- Melander H, Ahlquist-Rastadt J, Meijer G et al. Evidence b(i)ased medicine-selective reporting from studies sponsored by pharmaceutical industry: review of studies in new drug applications. BMJ. 2003;326:1171–1173.
[Abstract/Free Full Text] - CPMP/EWP/2330/99 Points to Consider on Application with (1) Meta-analyses and (2) One Pivotal study (adopted by CPMP May 2001). Available from: http://www.emea.eu.int/pdfs/human/ewp/23099en.pdf.
- Senn S. Statistical aspects of research done outside pharmaceutical industry could be improved. BMJ. 1998;316:228.
[Free Full Text] - Kübler J. Validity and interpretation of meta-analyses and one pivotal study. DIJ. 2001;35:1507–1515.
- Chiasson JL, Josse RG, Gomis R, Hanefeld M, Karasik A, Laakso M. Acarbose treatment and the risk of cardiovascular disease and hypertension in patients with impaired glucose tolerance. The STOP-NIDDM Trial. JAMA. 2003;290:486–494.
[Abstract/Free Full Text]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  J. McCormack, K. Johns, and H. Tildesley Metformin's contraindications should be contraindicated Can. Med. Assoc. J., August 30, 2005; 173(5): 502 - 504. [Full Text] [PDF]
|
|
|
|
 |
|
 M. Hanefeld, R. G. Josse, and J.-L. Chiasson {alpha}-Glucosidase Inhibitors for Patients With Type 2 Diabetes: Response to van de Laar et al. Diabetes Care, July 1, 2005; 28(7): 1840 - 1840. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||