Biomarkers to Predict the Response to Cardiac Resynchronization Therapy

Ward Heggermont; Angelo Auricchio; Marc Vanderheyden


Europace. 2019;21(11):1609-1620. 

In This Article

Clinical Perspectives, Limitations, and Future Outlook

In this review, we gave an overview of the currently available evidence on biomarkers predicting the response to CRT. How should we interpret the amount of evidence presented here? As far as the classical HF biomarkers are concerned (NT-pro-BNP, BNP, hsTnT, sST2, Gal-3), a beneficial response to CRT coincides with lower values of these molecules. This finding however does not mean that these biomarkers can predict the CRT response before implantation! Rather these molecules reflect a beneficial reverse remodelling process after implantation. With respect to the biomarkers related to inflammation (hsCRP, complement C3a), it is interesting to note that inflammation plays an important role in the remodelling process in HF. Therefore, targeting the inflammatory pathway to find predictive biomarkers for CRT, is reasonable. Of particular interest seems complement C3a, because it also brings prognostic information on mortality after CRT implantation, beyond NT-pro-BNP levels. Closely linked to inflammation is the extracellular matrix, a paramount player in the remodelling process. In general, sample sizes of the studies on ECM-related biomarkers, adrenergic signalling, and others are too small to make relevant conclusions. Interestingly, high CysC levels at baseline coincided with adverse 6-month response to CRT, making it a potentially interesting biomarker. However, until now, it is unclear whether the predictive value of CysC is preserved when adjusting for the grade of chronic kidney failure, and this should be investigated in a large patient population.

It is clear that the need to better predict the response to CRT reflects an important clinical question as many attempts to find a useful biomarker have been undertaken. On the other hand, we conclude from our literature search that not a single biomarker is currently able to better predict the CRT response, on top of already known clinical markers. First, we have to acknowledge that a lot of research is performed, which is very positive. However, outcomes of the studies discussed are rather diffuse and difficult to interpret. This is due to the fact that (i) different definitions are applied to define the CRT response, (ii) it is sometimes unclear whether a biomarker predicts the response to CRT, vs. coincides with a beneficial CRT response, and (iii) sample sizes are generally quite small, except for some larger cohorts. Moreover, most of the papers discussed are to some extent descriptive and hypothesis-generating, rather than systematically testing or validating a predefined hypothesis. Also, most molecules are tested because they are known to be involved in the pathophysiology of HF, potentially implying some sort of selection bias.

The lack of a universal conclusion on which is the 'best' biomarker to predict the CRT response, is partly to be explained by some limitations, with which we were also confronted when listing the biomarkers studies. First, all studies used blood sampling as the option of choice to obtain human material. Related to the site of sampling, although conventional venous puncture is routinely used, CS sampling might be a favourable option.[54] This CS sampling is perfectly feasible in patients already selected for CRT, but will be hard to perform in patients who did not undergo eligibility screening. Furthermore, when sampling the CS, one should also have an idea about the cardiac output, since the transcoronary gradient is most likely not the only parameter explaining differences in biomarker levels. Moreover, the fact that CS sampling would be used is a bit in contradiction with the fact that a biomarker should be easy to procure. On the other hand, taking into account the invasiveness and the impact of the implantation procedure, CS sampling might still be acceptable in this patient group. Second, an important limitation of this review is related to the definition of a CRT response which varies significantly in the different studies in terms of the parameter(s) measured and the time point. Third, a lot of pilot data are available that did not undergo extensive validation. Fourth, the often-unclear distinction between ischaemic and non-ischaemic cardiomyopathy as the underlying pathology resulting in CRT implantation, may influence the predictive value of certain biomarkers, as the pathophysiology of non-ischaemic cardiomyopathy (NICMP) and ischemic cardiomyopathy (ICMP) is largely different. And last but not least, a careful distinction has to be made between studies investigating a single, vs. multiple, biomarkers. With respect to the latter drawback, the MARC study[55] nicely showed one of the problems with biomarker research: several of 16 investigated biomarkers were independently associated with the CRT response, but their significance diminished in a multivariate analysis.