Author + information
- Pierluigi Costanzo, MD⁎ ( and )
- Pasquale Perrone-Filardi, MD, PhD
- ↵⁎Department of Internal Medicine, Cardiovascular Science and Immunology, Federico II University of Naples, via S. Pansini 5, 80131 Naples, Italy
We write this letter to clarify some issues that have arisen regarding our study (1). The purpose of our meta-analysis was mostly driven by the fact that it has been claimed that interventions that reduce intima-media thickness (IMT) progression also reduce cardiovascular events. However, although studies that have shown a relationship between a high IMT and increased risk of cardiovascular events do exist (2–4), there is no study that has ever been powered and published to show such a relationship. IMT prognostic value has always relied on its relationship with other proven surrogate endpoints and never with cardiovascular events. The only study that is often quoted to show a putative relationship between IMT changes and cardiovascular risk is the review of Espeland et al. (2). However, it should be stated that, technically, the Espeland et al. study (2) merely incorporated the addition of change in carotid IMT as a covariate in a regression model of a meta-analysis of some statin trials. In particular, they showed that the change in carotid IMT raised the summary odds ratio of developing a cardiovascular event on statin therapy from 0.48 to 0.64. In addition, they included few studies and only those that directly compared statin therapy with placebo. Therefore, this study cannot be quoted as clear evidence of a significant relationship between change in carotid IMT and cardiovascular risk.
Technical aspects concerning the reproducibility of serial intra-individual changes and lack of standardization of IMT measurements are crucial in understanding the role of this tool in cardiovascular risk assessment. Despite the fact that carotid IMT measurements are prone to generate variability in follow-up studies, in controlled clinical trials, measurement variability has been decreasing due to technical improvements, standardization, and training (5). Nevertheless, we could not find any significant influence of the year of publication. Furthermore, in our meta-analysis, some source of variability could also have come from the inclusion of studies with different imaging protocols. However, we substantially identified 2 major groups: multicenter trials in which images were handled and IMT measurements recorded off-line in a core ultrasound laboratory. However, after performing a sensitivity analysis by excluding studies that did not measure IMT in a central core laboratory, our results again did not significantly change.
The short follow-up of studies included in our meta-analysis may be used to justify the lack of association between IMT changes and cardiovascular risk; however, it becomes troublesome then to explain why short-term changes in low-density lipoprotein do predict cardiovascular risk. Thus, although we recognize that IMT changes perhaps may predict long-term cardiovascular risk, they do not perform as low-density lipoprotein changes in the short term, as we have shown in our paper.
Finally, we would like to point out that, as for all meta-analyses, particularly for those not based on an individual dataset, findings need to be interpreted with caution and are intended to be hypothesis-generating. Nonetheless, we believe that the conclusion from our study can help achieve a better understanding of clinical trials having IMT as endpoint and the planning of future studies.
- American College of Cardiology Foundation
- Costanzo P.,
- Perrone-Filardi P.,
- Vassallo E.,
- et al.
- Lorenz M.W.,
- Markus H.S.,
- Bots M.L.,
- Rosvall M.,
- Sitzer M.