Acute coronary syndrome (ACS) is the leading cause of death in industrialized countries and is thus a major health problem.1,2

Obesity is a growing epidemic and a global healthcare challenge.3,4 It is associated with cardiovascular mortality and with advanced coronary artery disease (CAD).3–6 However, obesity is heterogeneous in terms of metabolic profile and risk implications.7 In this regard, abdominal obesity is increasingly recognized as an important risk factor for cardiovascular disease as well as a major correlate of atherogenic and metabolic abnormalities.7–10 Visceral fat accumulation is a better marker of the cardiometabolic risk associated with obesity than body mass index (BMI).8,11 Adipose tissue is known to be an active endocrine organ. Due to its metabolic activity and anatomic location, abdominal fat secretes large quantities of adipokines and inflammatory cytokines directly into the portal circulation that contribute to the proinflammatory and prothrombotic state of visceral obesity, which may be responsible for the higher attributed risk in abdominally obese subjects.9 Moreover, abdominal obesity is the most prevalent feature of the metabolic syndrome, which is predictive of cardiovascular mortality.12–14 Metabolic syndrome is highly prevalent in ACS patients, and therefore the associated cardiometabolic risk should be assessed in this population.

Direct measures of abdominal obesity, such as waist circumference (WC), might help to better assess fat location and thus improve understanding of obesity-related risk.15 At the population level, increased WC has been independently related to all-cause mortality, myocardial infarction (MI) and CAD risk.16 Although the link between abdominal adipose depot and mortality has been extensively demonstrated in the general population, this relationship appears more complex in ACS patients and requires further clarification. Despite the prevalence of abdominal obesity, few studies have addressed its impact on prognosis after ACS. Recent evidence has demonstrated that elevated WC is associated with increased mortality in patients with CAD.3 It may thus be hypothesized that assessment of WC could improve risk prediction in ACS patients. However, to be a useful tool to predict risk in ACS, it has to demonstrate incremental utility beyond that of an established risk model.17

Clinical management of ACS patients should be guided by a risk estimate.18,19 ACS covers a large spectrum of clinical conditions and for this reason patients are heterogeneous in terms of risk profile for recurrent events. Risk stratification enables tailoring of aggressive medical therapies or invasive procedures towards higher-risk individuals.20,21 Evidence and current guidelines suggest that the benefit of more intensive treatment is related to individual patients’ risk of clinical events.22

To optimize risk assessment, several prognostic factors must be simultaneously taken into account. Risk scores are critical for this purpose. Several risk scores developed in the past decade are currently available to help clinicians with risk stratification. Of those, the Global Registry of Acute Coronary Events (GRACE) risk score (GRACE RS) is probably the most accurate, robust and widespread clinical tool.1 The GRACE RS can be used to estimate the risk of mortality or the combined risk of mortality or MI at admission or at six months. Although the GRACE RS has been extensively evaluated and validated in ACS patients, accurate risk stratification is a difficult and complex task. Little is known about the additional value of abdominal obesity for risk stratification beyond other prognostic variables already used in practice.23 The purpose of the present study was to assess the incremental prognostic value of WC beyond the GRACE RS in a cohort of consecutive patients admitted to a cardiac care unit (CCU) with a diagnosis of ACS.

The present study set out to assess whether the inclusion of WC measurement in the multivariate model derived from the GRACE RS improved risk prediction in ACS patients. To the best of our knowledge, this is the first study to quantify the additional predictive value of WC in relation to the GRACE RS.

Our data demonstrated that WC assessment did not improve the predictive accuracy of the GRACE RS. Furthermore, this analysis did not confirm WC measurement as an independent outcome predictor in multivariate analysis. An additional unanticipated result was that the predictive performance of the calculated GRACE RS was shown to be poor.

The study population was mostly composed of males (78.9%); similar proportions were reported in previous studies of ACS populations. Mean age and previous medical conditions were similar to those of the GRACE cohort.16 A higher proportion of STEMI patients was observed (57.5%), compared to ACS registries.18,21 However, our cardiac center receives patients from peripheral hospitals unable to perform PCI, which might explain the number of STEMI patients observed for whom immediate revascularization procedures were required.25

The results revealed that higher CRP and NT-pro-BNP and lower hemoglobin level at admission and lower GFR were related to outcome. All of these biomarkers are considered established prognostic markers, although none are currently used in risk scores.26

Aggressive treatment of ACS has become widespread, as shown by the PCI rate in our study (67.3%). Six-month mortality was 3.5%, lower that observed in GRACE or in the contemporary MASCARA registry.21 Nevertheless, unlike the present analysis, both registries had a PCI rate lower than 15%, which may explain our lower mortality. Besides, overall evidence-based cardiovascular therapy rates were very high, which has been associated with improvements in outcome.25

The prevalence of general and abdominal obesity was very high in the present analysis (26.9% and 44.6%, respectively), highlighting the obesity epidemic. These results confirm those of studies addressing the impact of obesity on ACS prognosis.3,4,12 It is recognized that abdominal obesity correlates with excessive visceral adiposity, which is associated with several proatherogenic features, including insulin resistance, hypertriglyceridemia, diabetes and hypertension.3 Despite the unquestionable utility of BMI at the population level, it appears to be a poor marker of adiposity-related risk in ACS, and there is currently a debate as to whether an “obesity paradox” exists in ACS.27 The association of adipose tissue with mortality in ACS is complex, and may rely more on the location than on the total amount of adipose tissue.3

In a recent meta-analysis, Coutinho et al. demonstrated that in CAD patients, increasing abdominal obesity (represented by WC or waist-to-hip ratio [WHR]) was associated with high mortality, independently of BMI,3 while Lee et al. found that WHR was an independent predictor of six-month mortality in STEMI patients.11 However, our results did not support the prognostic value of abdominal obesity in ACS. Zeller et al. also concluded that neither BMI nor WC independently predicted mortality after acute MI. However, they observed that patients with high WC but normal BMI (supposedly reflecting the presence of visceral adiposity) had a poorer prognosis.12 Similar results were reported by Kadakia et al.4

We decided to use a combined approach, using the GRACE RS as a prognostic reference. However, our data demonstrated that the GRACE RS had unexpectedly poor predictive accuracy, meaning that it represents a random classification of the risk of event occurrence. Despite its broad acceptance, it should be recognized that the use of the GRACE RS in different populations from its original cohort may compromise its accuracy. Besides, Beygui et al. stated that the discriminatory value of the GRACE RS might be limited in a low-event population.28 Although the GRACE RS has been validated in a Portuguese cohort of NSTEMI patients by De Araújo Gonçalves et al., these authors studied different time points and used a more selected population.22

Most of the GRACE RS variables did not reach statistical significance in multivariate analysis. However, ST-segment deviation was shown to be an independent predictor of outcome. This feature is actually a powerful risk predictor and has been incorporated in several risk scores. Yan et al. found that continuous ST-segment monitoring in NSTEMI patients confers additional prognostic value beyond the GRACE RS.29 However, Fox et al. stated that although the ECG is helpful for guiding treatment, it is unable to predict late complications, and a multivariate approach is therefore needed.2 Nevertheless, almost all the predictive information associated with the GRACE RS in the study population was due to ST changes, which raises doubts concerning the usefulness of the GRACE RS in clinical practice. Moreover, elevated cardiac enzymes had a significant protective prognostic effect. The mechanism behind this relation to outcome cannot be ascertained, since any measurable cardiac troponin is related to an unfavorable prognosis and even minor CK-MB enzyme elevations appear to have prognostic implications.17 Nevertheless, it should be pointed that 90.2% of our study population had elevated cardiac enzymes, and therefore the statistical analysis could have been skewed.

AUC is a commonly used statistic. However, it simply and directly addresses discriminatory ability, a major but not unique component of accuracy. Recent and more refined statistical methods, such as net reclassification improvement and integrated discrimination improvement, have been proposed as better tools for assessment of the additional value of a risk marker.30

The study has several limitations, and consequently its results should be interpreted with caution. Abdominal obesity data were not available for 34% of the cohort. The prognosis of these individuals was poorer (with higher mortality), although their general baseline characteristics did not differ from those of the study population. Severe clinical status might have limited anthropometric measurement in several individuals and therefore a selection bias cannot be excluded.

Second, this was a relatively small retrospective analysis in a single institution, based on nonrandomized data. Accordingly, the negative results obtained could be due to insufficient statistical power, and therefore this issue should be revisited in larger population-based multicenter studies, perhaps with extended follow-up, to assess the delayed prognostic impact of abdominal obesity.

Our results should be considered exploratory and need to be confirmed by future studies.

Routine early risk stratification remains an essential part of ACS care, and is supported by risk scores, such as the GRACE RS. However, risk scores need to be continually re-evaluated, since accurate risk stratification is still a challenge. In this study, abdominal obesity did not improve the discriminatory ability of the GRACE RS, and we can therefore conclude that abdominal obesity assessment does not increase the prognostic information provided by the GRACE RS. However, given the high prevalence of general and abdominal obesity observed in the study population, clarification of obesity-related risk is of major importance and should be reassessed in further research.

The authors have no conflicts of interest to declare.