Author + information
- Received November 29, 2009
- Revision received October 28, 2010
- Accepted November 23, 2010
- Published online July 19, 2011.
- Diego Ardissino, MD⁎,⁎ (, )
- Carlo Berzuini, PhD†,
- Piera Angelica Merlini, MD‡,
- Pier Mannuccio Mannucci, MD§,
- Aarti Surti, BA∥,
- Noel Burtt, BS∥,
- Benjamin Voight, PhD¶,#,⁎⁎,
- Marco Tubaro, MD††,
- Flora Peyvandi, MD§,
- Marta Spreafico, PhD§,
- Patrizia Celli, MD‡‡,
- Daniela Lina, MD⁎,
- Maria Francesca Notarangelo, MD⁎,
- Maurizio Ferrario, MD§§,
- Raffaela Fetiveau, MD§§,
- Giorgio Casari, PhD∥∥,
- Michele Galli, MD¶¶,
- Flavio Ribichini, MD##,
- Marco L. Rossi, MD⁎⁎⁎,
- Francesco Bernardi, MD†††,
- Nicola Marziliano, PhD§§,
- Pietro Zonzin, MD‡‡‡,
- Francesco Mauri, MD‡,
- Alberto Piazza, MD, PhD§§§,
- Luisa Foco, PhD∥∥∥,
- Luisa Bernardinelli, PhD∥∥∥,¶¶¶,
- David Altshuler, MD, PhD§,¶,#,⁎⁎,###,⁎⁎⁎⁎,
- Sekar Kathiresan, MD∥,¶,###,
- Italian Atherosclerosis, Thrombosis and Vascular Biology Investigators
- ↵⁎Reprints requests and correspondence:
Dr. Diego Ardissino, Unità Operativa di Cardiologia, Azienda Ospedaliero-Universitaria di Parma, Via Gramsci 14, 43100 Parma, Italy
Objectives The purpose of this study was to test whether the 9p21.3 variant rs1333040 influences the occurrence of new cardiovascular events and coronary atherosclerosis progression after early-onset myocardial infarction.
Background 9p21.3 genetic variants are associated with ischemic heart disease, but it is not known whether they influence prognosis after an acute coronary event.
Methods Within the Italian Genetic Study of Early-onset Myocardial Infarction, we genotyped rs1333040 in 1,508 patients hospitalized for a first myocardial infarction before the age of 45 years who underwent coronary angiography without index event coronary revascularization. They were followed up for major cardiovascular events and angiographic coronary atherosclerosis progression.
Results Over 16,599 person-years, there were 683 cardiovascular events and 492 primary endpoints: 77 cardiovascular deaths, 223 reoccurrences of myocardial infarction, and 383 coronary artery revascularizations. The rs1333040 genotype had a significant influence (p = 0.01) on the primary endpoint, with an adjusted hazard ratio of 1.19 (95% confidence interval [CI]: 1.08 to 1.37) for heterozygous carriers and 1.41 (95% CI: 1.06 to 1.87) for homozygous carriers. Analysis of the individual components of the primary endpoints provided no significant evidence that the rs1333040 genotype influenced the hazard of cardiovascular death (p = 0.24) or the reoccurrence of myocardial infarction (p = 0.57), but did provide significant evidence that it influenced on the hazard of coronary revascularization, with adjusted heterozygous and homozygous ratios of 1.38 (95% CI: 1.17 to 1.63) and 1.90 (95% CI: 1.36 to 2.65) (p = 0.00015), respectively. It also significantly influenced the angiographic endpoint of coronary atherosclerosis progression (p = 0.002).
Conclusions In early-onset myocardial infarction, the 9p21.3 variant rs1333040 affects the progression of coronary atherosclerosis and the probability of coronary artery revascularization during long-term follow-up.
Epidemiological and family data suggest that genetic factors have an impact on the occurrence of myocardial infarction, particularly at an early age (1,2). Genome-wide association studies have revealed an association between common genetic variants in chromosomal region 9p21.3 and ischemic heart disease (3–7), a finding that has been replicated in a number of case-control studies of populations of different ethnic origin (8–16).
The mechanism by which 9p21.3 genetic variants influence ischemic heart disease is still unknown. Indirect data suggest that they may affect the development and progression of coronary atherosclerosis (17–19), but no clinical results are available currently concerning their influence on prospectively observed cardiovascular events or the progression of coronary atherosclerosis.
The Italian Genetic Study of Early-onset Myocardial Infarction was designed to study the genetics of susceptibility to myocardial infarction. It was a prospective, nationwide project involving patients who were hospitalized because of a first myocardial infarction before the age of 45 years and who were followed up for the occurrence of cardiovascular events. We assessed the influence of the 9p21.3 variant tagged by rs1333040 on the occurrence of major adverse cardiovascular events and coronary atherosclerosis progression during long-term follow-up.
The Italian Genetic Study of Early-onset Myocardial Infarction involved 125 coronary care units. It was divided into 2 parts: a case-control study and a prospective follow-up of the cases. The cases and controls were enrolled between 1988 and 2002. Cases were considered eligible if they were hospitalized for a first myocardial infarction before the age of 45 years and underwent coronary angiography without coronary revascularization for the index event; controls were eligible if they were blood donors without a history of thromboembolic disease and were unrelated to the patients, but were matched individually with the patients by age, gender, and geographic origin.
The protocol was approved by the institutional review boards of all of the participating centers, and all of the patients gave their written informed consent. After identifying the patients suitable for enrollment, the investigators collected their written informed consent, completed a standardized case report form covering their family history of cardiovascular diseases, cardiovascular risk factors, lifestyles, and medications and obtained a blood sample for plasma separation and deoxyribonucleic acid extraction. The patients were followed up by means of standardized telephone contacts, outpatient visits, and subsequent hospital admissions for the occurrence of endpoints. An attempt was made to telephone all of the patients (or their families or primary care physicians if they could not be contacted); if they reported an event, their medical records were obtained.
Blood collection, processing, and storage
Blood was drawn from the antecubital vein into 3 tubes containing 0.106 M trisodium citrate and was separated into plasma and red cells by means of centrifugation. The plasma was divided into 5 aliquots and was stored at −80°C. Deoxyribonucleic acid was isolated from the white blood cells using the salting-out method and was stored in alcohol at −20°C.
The primary endpoint was the composite of cardiovascular death, the reoccurrence of myocardial infarction, and coronary artery revascularization by means of a percutaneous coronary intervention or coronary artery bypass surgery. Only the event that occurred first was counted.
All deaths were investigated using the death certificate specifying the cause of death; all other events were investigated by means of source data verification. Cardiovascular death was defined as any death attributed to a cardiovascular cause on the death certificate. The reoccurrence of myocardial infarction was defined as subsequent hospitalization ending with a discharge diagnosis of myocardial infarction. Rehospitalization for revascularization was defined as any period of hospitalization during which the patient underwent coronary revascularization percutaneously or by means of coronary artery bypass surgery. Other thromboembolic events were defined as any thromboembolic event requiring hospitalization.
All of the events were adjudicated by an events committee of 2 cardiologists who were unaware of the genotyping results. In the case of disagreement, the opinion of a third cardiologist was required.
Assessment of angiographic coronary artery disease
The absence of any narrowing in coronary diameter was considered evidence of normal coronary arteries; a narrowing of <70% (50% in the case of the left main coronary artery) was considered nonsignificant coronary artery stenosis, and a narrowing of 70% or more (50% or more in the case of the left main coronary artery) was considered significant coronary artery stenosis. A significant coronary artery stenosis involving 1 of the 3 major coronary arteries was defined as single-vessel disease; if significant coronary artery stenoses affected 2 or 3 of the major coronary arteries, it was defined as multivessel disease.
The extent of coronary artery disease was evaluated using the Duke Coronary Artery Disease Index (20), which was established in all patients on the basis of the coronary angiography performed during hospitalization for the index myocardial infarction. In the subset of patients who underwent coronary angiography during follow-up, the progression of coronary artery disease was evaluated by comparing the Duke Coronary Artery Disease Index of the first coronary angiography with that of the follow-up coronary angiography. If the patient underwent more than one coronary angiography during follow-up, the last one was used. Angiographic progression of coronary atherosclerosis was defined as the need for angiographic assessment during follow-up accompanied by an increase in the Duke Coronary Artery Disease Index of more than the median value of the observed changes.
Genotyping was performed using matrix-assisted laser desorption ionization time-of-flight mass spectrometry and a Sequenome MassARRAY platform (San Diego, California). The genotyping success rate was 99.8% (range 99.6% to 99.9%); the genotyping error rate estimated using 15 samples genotyped in quadruplicate was <0.7% (21).
This study focused on a single locus (the single nucleotide polymorphism rs1333040) and used it as a tag for 9p21.3 genetic variants that may influence the occurrence of cardiovascular events after early-onset myocardial infarction. The locus was chosen on the basis of the results of an earlier case-control study (also part of the Italian Genetic Study of Early-onset Myocardial Infarction) in which 5 9p21.3 genetic variants that have shown a consistent signal of association with ischemic heart disease in previous genome-wide association studies were tested for their association with early-onset myocardial infarction. These variants were: rs10116277, rs1333040, rs2383206, rs10757278, and rs1333049. The single nucleotide polymorphism rs1333040 showed the strongest association with early-onset myocardial infarction, with no residual association being found at any of the remaining 4 loci, conditional on rs1333040. These findings prompted the present investigation of the effect of rs1333040 on prospectively observed cardiovascular events after early-onset myocardial infarction.
The primary endpoint of the analysis was the composite of cardiovascular death, the reoccurrence of myocardial infarction, and coronary artery revascularization by means of a percutaneous coronary intervention or coronary artery bypass surgery. The primary endpoint-free time was defined as the time between the index myocardial infarction and the earliest of the above events, or the time of censoring.
The angiographic endpoint of progression of coronary atherosclerosis was considered to have been met when the increase in the Duke Coronary Artery Disease Index from baseline to follow-up coronary angiography was more than the median value of the observed changes. A first assessment of the Duke Coronary Artery Disease Index was made in all patients at the time of the baseline coronary angiography examination, and a subsequent assessment was performed only in those patients who repeated coronary angiography during follow-up. The progression of coronary atherosclerosis was defined as the need for an angiographic assessment during follow-up accompanied by an increase in the Duke Coronary Artery Disease Index of at least 19 (the median increase in our sample) over baseline.
The dependence of primary endpoint-free time on rs1333040 was analyzed using Cox's regression model for censored failure time data, adjusting for the extent of coronary artery disease at the time of the coronary angiography performed during the index hospitalization and the main risk factors under a proportional hazard assumption.
Step-wise variable selection was used to isolate minimally adequate models. This involved the interleaving of short sequences of backward variable elimination and short sequences of forward inclusion. In the case of a significant association between rs1333040 and the primary endpoint, subsequent Cox analyses were made to test the association with each component of the primary endpoint. The corresponding estimates of heterozygous and homozygous relative risk were obtained from each of the analyses. Kaplan-Meier estimates of the genotype-specific cumulative curves for the various endpoints were also calculated. The dependence of the number of significant coronary artery stenoses on the number of copies of the rs1333040 risk allele was analyzed by means of logistic regression. The association between the 9p21.3 variant rs1333040 and coronary artery disease, as measured at the time of the first infarction, was tested by means of the Cochran-Armitage test of trend and the chi-square test of association.
Demographic characteristics and traditional risk factors
The case sample consisted of 1,508 patients (1,334 men and 174 women) whose median age at the time of their first myocardial infarction was 41 years (interquartile range: 37 to 43 years). The genotype distribution of rs1333040 in the study population was: 141 patients had no risk allele (CC), 587 patients had 1 risk allele (CT), and 780 patients had 2 risk alleles (TT). There was no evidence of any departure from Hardy-Weinberg equilibrium (p > 0.05). Table 1 shows the demographic and angiographic characteristics of the patients and the frequency distribution of the traditional risk factors by genotype.
rs1333040 genotype and long-term clinical outcomes
The patients were followed up for a median of 9.95 years (interquartile range: 8 to 11.8 years), a total of 16,599 person-years. The vital status of 1,497 patients (99%) was ascertained, and 1,434 (95%) completed the follow-up. During the follow-up, there were 683 cardiovascular events and 492 primary endpoints: 77 patients died of cardiovascular death, 223 experienced reoccurrences of myocardial infarction, and 383 underwent coronary artery revascularization (230 by means of a percutaneous coronary intervention and 153 by means of coronary artery bypass surgery). During the follow-up, 53 patients experienced more than 1 myocardial infarction, 85 patients underwent more than 1 percutaneous coronary revascularization, and 2 patients underwent more than 1 coronary artery bypass intervention.
The rs1333040 genotype had a significant (p = 0.01) influence on the primary endpoint, with an adjusted heterozygous hazard ratio of 1.19 (95% confidence interval [CI]: 1.08 to 1.37) and homozygous hazard ratio of 1.41 (95% CI: 1.06 to 1.87). Table 2 shows the estimated hazard ratios, 95% CIs, and p values of the independent predictors of the primary endpoint.
The analyses of the individual components of the primary endpoint provided no significant evidence of an influence of the rs1333040 genotype on the hazard of cardiovascular death (p = 0.24) or the reoccurrence of myocardial infarction (p = 0.47). The adjusted hazard ratio for cardiovascular death was 1.1 (95% CI: 0.8 to 1.4) for heterozygous carriers of the risk allele and 1.2 (95% CI: 0.6 to 2.5) for homozygous carriers of the risk allele; the adjusted hazard ratio for the reoccurrence of myocardial infarction was 1.05 for heterozygous carriers of the risk allele (95% CI: 0.86 to 1.28) and 1.10 for homozygous carriers of the risk allele (95% CI: 0.74 to 1.64), respectively. However, there was significant evidence (p = 0.00015) that the rs1333040 genotype influenced the hazard of undergoing coronary artery revascularization during follow-up, with an adjusted heterozygous hazard ratio of 1.38 (95% CI: 1.17 to 1.63) and a homozygous hazard ratio of 1.90 (95% CI: 1.36 to 2.65). Kaplan-Meier curves of the cumulative incidence of the primary endpoint and the individual components of the primary endpoint are shown in Figure 1 by rs1333040 genotype. Table 3 shows estimated hazard ratios, 95% CIs, and p values for the independent predictors of the primary endpoint.
rs1333040 genotype and coronary artery disease
Coronary angiography at the time of the index myocardial infarction showed normal coronary arteries or coronary atherosclerosis without significant stenoses in 282 patients and significant coronary artery disease in 1,226 patients.
There was significant evidence of an association between the rs1333040 genotype and the presence of significant coronary artery stenoses (p = 0.014) and the Duke Coronary Artery Disease Index (p =1 e−06) at the time of the index coronary angiography (Table 1).
During the follow-up, 405 patients underwent at least 1 repeat coronary angiography. The median time between the coronary angiography performed during the index hospitalization and the coronary angiography performed during follow-up was 6.6 years (interquartile range: 4 to 9 years), for a total of 2,625 person-years of angiographic follow-up. The median change in the Duke Coronary Artery Disease Index was 19 (interquartile range: 0 to 24). Individual and median changes in the Duke Coronary Artery Disease Index by rs1333040 genotype are shown in Figure 2.
The rs133340 genotype had a significant influence (p = 0.002) on the angiographic endpoint of coronary atherosclerosis progression. The adjusted heterozygous hazard ratio was 1.5 (95% CI: 1.17 to 2.02), and the adjusted homozygous hazard ratio was 2.2 (95% CI: 1.3 to 2.7). The Kaplan-Meier curves of the cumulative incidence of the angiographic endpoint of coronary atherosclerosis progression by rs1333040 genotype are shown in Figure 3.
This study assessed the influence of 9p21.3 rs1333040 on the risk of incident cardiovascular events and the progression of coronary atherosclerosis in individuals who experienced a first myocardial infarction at a young age. Its major finding is that 9p21.3 rs1333040 influences the probability of undergoing subsequent coronary artery revascularization after a first myocardial infarction, even when accounting for the extent of coronary artery disease at the time of the initial acute coronary event. We observed that the variant was also associated with the progression of coronary atherosclerosis, thus suggesting a mechanism for the increased future revascularization rate in carriers of the risk allele.
These results confirm the findings of genome-wide association studies (3–7) suggesting that genetic variations in the 9p21.3 chromosomal region represent a novel risk factor for coronary artery disease. Multiple single nucleotide polymorphisms in strong linkage disequilibrium have been identified in this genomic region of approximately 50,000 bases, but they probably represent the same genetic signal (5,6). The functional effect of this signal remains unknown, because there are no annotated genes in this region.
The prospective nature of our study allows a true assessment of the risk of cardiovascular events after early-onset myocardial infarction related to the 9p21.3 variation. The rs1333040 genotype significantly influenced the hazard of the combined endpoint of cardiovascular death, myocardial infarction, or the need for coronary artery revascularization, with an estimated increased risk of 19% for heterozygous and 41% for homozygous carriers of the risk allele. Despite the strong influence of the genotype on the need for subsequent coronary artery revascularization, no evidence was found to support its influence on the risk of cardiovascular death or the reoccurrence of myocardial infarction.
Why the 9p21.3 genetic variant rs1333040 influences the need for coronary artery revascularization, but does not affect cardiovascular death or the reoccurrence of myocardial infarction, remains unknown. It can be surmised that this genetic variant influences only the development and progression of coronary atherosclerosis, and not plaque instability or coronary thrombosis. A similar observation was made by Horne et al. (17), who found that the 9p21.3 genetic variant was not associated with incident events or prevalent myocardial infarction in a population of patients undergoing coronary angiography, although it did predict a diagnosis of coronary artery disease.
The 9p21.3 region has been associated with the calcium score (5,14), severe premature atherosclerosis (5), the prevalence of angiographic coronary artery disease (18,22), and the progression of carotid atherosclerosis (19). In keeping with these findings, we observed an association between the 9p21.3 genetic variant rs1333040 and the prevalence of coronary artery disease at the time of the coronary angiography performed during the index hospitalization. In addition, the variant influenced the progression of coronary atherosclerosis as measured by changes in the Duke Coronary Artery Disease Index. Carriers of 2 risk alleles showed more rapid progression of coronary atherosclerosis, whereas carriers of 1 risk allele showed intermediate progression between that of wild-type and homozygous carriers, thus indicating a gene gradient for the progression of coronary atherosclerosis.
One possible limitation of this study is the use of early-onset myocardial infarction as a clinical model to study the influence of 9p21.3 sequence variants on cardiovascular events and the progression of coronary atherosclerosis. It is known that early-onset myocardial infarction is greatly influenced by genetics (1,2), and so the effect of genetic factors may be overestimated in this population. However, we focused on the narrow phenotype of early-onset myocardial infarction because its degree of inheritability makes it very promising for fruitful genetic mapping, although future studies are clearly needed to assess the real influence of 9p21.3 genetic variants in broader populations of patients with all of the manifestations of ischemic heart disease.
Another possible limitation may be related to the selection of patients with early-onset myocardial infarction who did not undergo coronary revascularization at the time of the index event. We did this to be able to assess the influence of the 9p21.3 genetic variant rs1333040 on the cumulative probability of cardiovascular events and the progression of atherosclerosis by eliminating the interference of coronary revascularization. However, because it may have had an impact on the incidence of cardiovascular events, our conclusions apply only to patients with early-onset myocardial infarction not undergoing coronary revascularization for the index event.
Moreover, the data relating to the progression of atherosclerosis came only from the patients who underwent repeat coronary angiography for clinical reasons. This selection bias may have confounding potential, but performing coronary angiography for scientific reasons has ethical implications that are difficult to overcome.(21)
This study was supported by the Italian charitable foundation “Associazione per lo Studio della Trombosi in Cardiologia” and the Atherosclerosis, Thrombosis, and Vascular Biology Italian Study Group. The Italian foundation “Associazione per lo Studio della Trombosi in Cardiologia” and the Atherosclerosis, Thrombosis, and Vascular Biology Italian Study Group were not involved in designing or conducting the study; in data collection, management, analysis, or interpretation; or in preparing, reviewing, or approving the manuscript. Dr. Galli is employed by Mundipharma Italy. Dr. Kathiresan has received research grants from Pfizer, Merck, and Almylam. All other authors have reported that they have no relationships to disclose.
- Abbreviations and Acronyms
- confidence interval
- Received November 29, 2009.
- Revision received October 28, 2010.
- Accepted November 23, 2010.
- American College of Cardiology Foundation
- Helgadottir A.,
- Thorleifsson G.,
- Manolescu A.,
- et al.
- McPherson R.,
- Pertsemlidis A.,
- Kavaslar N.,
- et al.
- Myocardial Infarction Genetics Consortium
- Shen G.Q.,
- Li L.,
- Rao S.,
- et al.
- Schunkert H.,
- Gotz A.,
- Braund P.,
- et al.
- Broadbent H.M.,
- Peden J.F.,
- Lorkowski S.,
- et al.,
- PROCARDIS Consortium
- Assimes T.L.,
- Knowles J.W.,
- Basu A.,
- et al.
- Horne B.D.,
- Carllquist J.F.,
- Muhlestein J.B.,
- Bair T.L.,
- Anderson J.L.
- Ye S.,
- Willeit J.,
- Kronenberg F.,
- Xu Q.,
- Kiechl S.
- Mark D.B.,
- Nelson C.L.,
- Califf R.M.,
- et al.
- Gabriel S.B.,
- Schaffner S.F.,
- Nguyen H.,
- et al.