Ventricular fibrillation and the proteome problem: can we solve it?

This editorial refers to ‘Discovery of plasma proteins associated with ventricular fibrillation during first ST-elevation myocardial infarction via proteomics’, by N.K. Stampe et al., https://doi.org/10.1093/ehjacc/zuad125.

Survival from sudden cardiac arrest (SCA) depends on the presenting rhythm.¹ Those manifesting with ventricular fibrillation (VF) are most likely to be successfully resuscitated. Timely defibrillation is an effective intervention, but is arguably a ‘sledgehammer’ approach that was first conceived almost 125 years ago.² The discovery of novel treatments as well as improved methods of prediction and prevention await a better mechanistic understanding of VF.³ This process of discovery can be facilitated by proteomic analysis of blood samples obtained from individuals who suffer VF.⁴ However, as we will discuss, the inherently unexpected nature of VF-SCA poses some unique challenges for interrogation of the proteome.

In this issue of the Journal, Stampe et al.⁵ from the celebrated sudden cardiac death investigative team at Copenhagen University Hospital report their findings from a proteomic analysis conducted in patients who suffered SCA manifesting with VF. What makes this study unique is their focus on VF that occurred within 12 h of symptoms in the setting of a first ST-segment elevation myocardial infarction (STEMI). Whole blood samples that were included in this analysis were obtained from the ongoing Danish GEVAMI study (GEnetic causes of Ventricular13 Arrhythmias in patients with first ST-elevation Myocardial Infarction) prior to a primary percutaneous coronary intervention procedure (PCI) or shortly afterward, all performed within 24 h from the onset of symptoms. Comparisons were made with a control group of STEMI patients that did not suffer VF during the same time frame.

The mass spectrometry (MS) analysis showed that a total of 26 proteins were associated with VF when compared with controls. The majority (n = 24) was up-regulated. These proteins are involved in multiple pathways including blood coagulation, haemostasis, as well as immune and inflammatory responses. However, when corrected for multiple comparisons, only two up-regulated proteins, actin beta-like 2 (ACTBL2) and coagulation factor XIII-A (F13A1) remained significantly associated with VF, both displaying a moderate effect size. The authors have discussed a broad range of possible roles for these two proteins in the pathobiology of VF, all of which will need to be evaluated in future studies.

As the authors clearly acknowledge, there are a few weaknesses inherent in their study design that limit the conclusions that can be drawn from their findings. Most importantly, blood samples were collected in the setting of VF, making it difficult to determine whether the two biomarkers significantly associated with VF played a potential mechanistic role or if these alterations were simply a consequence of the near-death event itself and/or administered treatments/resuscitation efforts. In 16% of the patients, the samples were drawn prior to primary percutaneous coronary intervention (PPCI) (14% of cases and 17% of controls), which implies that in >80%, blood samples were drawn after PPCI. This leaves open possibility that the PPCI procedure and revascularization may have affected protein levels. Finally, Stampe et al. were not able to independently validate their findings in a separate group of individuals, which limits both confirmation and generalizability.

Stampe and colleagues kindly bring up our work on the proteomics of SCA also using MS, published earlier in 2023 from the Center for Cardiac Arrest Prevention at Cedars-Sinai, Los Angeles (Norby et al.⁶). A more detailed comparison of the two investigations may help to put the current findings in context. The Norby study also investigated VF-SCA (90% of cases had documented VF, unpublished data) but there were important differences in the clinical settings. Norby et al. conducted their study among SCA survivors with blood samples collected within approximately one year of their event, with the assumption that any potential effects of VF-SCA on the proteome had resolved. Their study had the additional advantage of including both an apparently healthy as well as a coronary artery disease control group in their discovery sample that allowed for identification of proteins uniquely associated with VF-SCA (and not with coronary disease). They also incorporated separate validation groups of cases and controls. Coincidentally, they also identified 26 proteins significantly associated with SCA in the discovery sample. There were three proteins identified, all related to fibrinogen that overlapped with the more recent Stampe study. However, the direction of change was opposite in the two studies, upward in Stampe et al. and downward in Norby et al. Neither of the two proteins that were significantly associated with VF-SCA (ACTBL2 or F13A1) was observed among the proteins identified in the Norby study. In their final results, Norby et al. reported a total of eight proteins that were successfully replicated, two associated with SCA on the coronary disease pathway and six of which were uniquely associated with SCA. These proteins can potentially augment the VF-SCA risk stratification strategy when combined with other biomarkers such as reduced ejection fraction,⁷ novel risk prediction algorithms,⁸ and high-risk electrocardiographic abnormalities.⁹ Clearly, these findings will also need to be tested in additional larger and more diverse populations.

From a broader perspective, what have we learned from these two studies that can help to design future investigations of VF-SCA biomarkers? There are some unique hurdles due to the inherently complex, dynamic, and temporally uncertain nature of SCA that need to be overcome. However, a collaborative, multi-pronged strategy following one uniform definition of VF-SCA has the potential to address these challenges. Blood samples could be collected at four possible timepoints temporally related to the VF-SCA event (Figure 1). (1) Cohort studies collect blood samples at initiation, generally remote from the occurrence of the VF-SCA event. For a largely acquired, complex trait like SCA, findings from such blood samples may only partially reflect the ultimate occurrence of the event a decade or more later but do provide a proteomic baseline at the individual level. (2) If prediction of VF-SCA is the ultimate objective, then the best time to draw blood samples for analysis is immediately prior to the SCA event even before symptoms have manifested, and then perform comparisons with blood samples collected after VF-SCA. Due to the unpredictable and unexpected quality of SCA, this seems like a difficult task. However, a large study of remnant blood from laboratory test samples that were drawn prior and relatively close to the SCA event could potentially accomplish this. (3) As discussed above, collection of blood immediately following SCA also has drawbacks but comparisons with timepoint (2) would be useful. (4) Finally, sample collection in VF-SCA survivors also involves two major assumptions that (a) the proteome has normalized following the event, and (b) with no permanent effects of the event. If the sample size is large enough, the comparisons between findings at all three timepoints, especially augmented by artificial intelligence tools,¹⁰ have the potential to tease out and validate components of the proteome that could contribute both to prediction as well as mechanistic understanding of VF.

Figure 1.

Suggested approach for interrogation of the proteome in individuals with ventricular fibrillation/sudden cardiac arrest.

Stampe et al.⁵ are to be commended for joining the quest to learn more about the pathobiology of VF, and their innovative analysis of blood samples archived in an ongoing case-control study of STEMI patients. Complex problems have complex solutions, and it will take all the collective and collaborative innovation and teamwork that we can muster to solve this VF-proteome problem.

Funding

This work is funded, in part, the by National Institutes of Health, National Heart Lung and Blood Institute Grants R01HL145675 and R01HL147358 to Dr Chugh. Dr Chugh holds the Pauline and Harold Price Chair in Cardiac Electrophysiology at Cedars-Sinai, Los Angeles, CA, USA.

Data availability

Data will be made available on request.