|dc.description.abstract||Inflammation is vital in the repair following an acute coronary syndrome (ACS). Studies suggest that excess inflammation is pathological and may lead to major adverse cardiovascular events (MACE). Hence, inflammatory markers may be potential MACE predictors, although it is currently unknown which marker(s) best predict MACE.
We firstly investigated common inflammatory markers as MACE predictors in two acute myocardial infarction (AMI) cohorts, using simple methods of combining markers to determine if such approaches could better predict MACE. The first study with 860 patients investigated white blood cell subtypes, markers routinely used in clinical practice. Individual subtypes were not associated with MACE. Ratios of these subtypes were predictive on univariate, but not multivariate analysis. The second study measured a small panel of inflammatory markers (C-reactive protein, interleukin (IL) -6 and tumour necrosis factor alpha) in a case-control study. Cases were defined as patients who developed MACE within one year. No significant association was found with MACE when using individual or combined markers. This finding might have been influenced by time as a confounder, as we did not strictly control the sampling time.
Next, we investigated cytokines because these markers are more reflective of acute inflammatory changes. We conducted a systematic review to investigate whether a combined cytokine approach would be superior to individual cytokines to predict MACE in ACS. Analysis of 10 studies meeting our eligibility criteria revealed inconsistent associations with MACE for an identical cytokine. The four studies using simple combined cytokine approaches found significant associations with MACE, suggesting that a combined approach could be superior to an individual one for MACE prediction. However, the 10 studies could not answer how best to combine the non-independent cytokines, or when it was best to measure them.
We next examined temporal variation of six cytokines in 23 AMI patients. This was to determine if time was a potential confounder and if an optimal timepoint existed to capture peak cytokine levels (representing excess inflammation). There was significant variation in cytokine concentrations over time, so no optimal timepoint could be determined. Only IL-6 had a clear trend with a peak on Day 1, and this peak was produced by half the cohort. The observed variation could not be explained by clinical risk factors or surrogate markers of infarct size. We concluded sampling time might be an important consideration when designing future studies.
Finally, we trialled the use of principal component analysis (PCA), a mathematical technique allowing non-independent variables to be reduced into combined scores, to create a combined cytokine score and tested its prognostic utility. Six cytokines were measured in a cohort of 320 patients. We found that IL-6, IL-8, all cytokines combined into a PCA score, and an IL-6-IL-8 PCA score univariately predicted MACE. Only the IL-6-IL-8 score and IL-6 were independently associated with MACE, with the IL-6-IL-8 score having a stronger relationship than IL-6 alone.
Collectively, these studies demonstrate the complexity of inflammation in ACS. While we could not answer how best to characterise inflammation, we showed that using single biomarkers was unlikely to be sufficiently representative of pathological inflammation. As a proof-of-concept, we demonstrated PCA was one mathematical approach that could be used to combine collinear markers. However, further studies are required to determine the best method to create a prognostic inflammatory score and the best time to measure inflammatory markers.||