Author + information
- Received June 19, 2017
- Revision received July 30, 2017
- Accepted August 2, 2017
- Published online December 13, 2017.
- Ahmed N. Mahmoud, MDa,∗ (, )
- Siva Sagar Taduru, MDb,
- Amgad Mentias, MDc,
- Dhruv Mahtta, MDd,
- Amr F. Barakat, MDe,
- Marwan Saad, MD, PhDf,
- Akram Y. Elgendy, MDa,
- Mohammad K. Mojadidi, MDa,
- Mohamed Omer, MDb,
- Ahmed Abuzaid, MDg,
- Nayan Agarwal, MDa,
- Islam Y. Elgendy, MDa,
- R. David Anderson, MDa and
- Jacqueline Saw, MDh
- aDivision of Cardiovascular Medicine, University of Florida, Gainesville, Florida
- bDepartment of Medicine, University of Missouri Kansas City, Kansas City, Missouri
- cDivision of Cardiovascular Medicine, University of Iowa, Iowa City, Iowa
- dDepartment of Medicine, University of Florida, Gainesville, Florida
- eUPMC Heart and Vascular Institute, University of Pittsburgh, Pittsburgh, Pennsylvania
- fDepartment of Medicine, Division of Cardiovascular Medicine, University of Arkansas for Medical Sciences, Little Rock, Arkansas
- gDepartment of Cardiovascular Medicine/Heart and Vascular Institute, Sidney Kimmel Medical College at Thomas Jefferson University-Christiana Care Health System, Newark, Delaware
- hDivision of Cardiology, University of British Columbia, Vancouver General Hospital, Vancouver, British Columbia, Canada
- ↵∗Address for correspondence:
Dr. Ahmed N. Mahmoud, Department of Medicine, Division of Cardiovascular Medicine, University of Florida, 1600 SW Archer Road, Gainesville, Florida 32610.
Objectives The authors sought to determine the clinical characteristics and in-hospital survival of women presenting with acute myocardial infarction (AMI) and spontaneous coronary artery dissection (SCAD).
Background The clinical presentation and in-hospital survival of women with AMI and SCAD remains unclear.
Methods The National Inpatient Sample (2009 to 2014) was queried for all women with a primary diagnosis of AMI and concomitant SCAD. Iatrogenic coronary dissection was excluded. The main outcome was in-hospital mortality. Propensity score matching and multivariable logistic regression analyses were performed.
Results Among 752,352 eligible women with AMI, 7,347 had a SCAD diagnosis. Women with SCAD were younger (61.7 vs. 67.1 years of age) with less comorbidity. SCAD was associated with higher incidence of in-hospital mortality (6.8% vs. 3.4%). In SCAD patients, a decrease in in-hospital mortality was evident with time (11.4% in 2009 vs. 5.0% in 2014) and concurred with less percutaneous coronary intervention (PCI) (82.5% vs. 69.1%). Propensity score yielded 7,332 SCAD and 14,352 patients without SCAD. The odds ratio (OR) of in-hospital mortality remained higher with SCAD after propensity matching (OR: 1.87, 95% confidence interval [CI]: 1.65 to 2.11) and on multivariable regression analyses (OR: 2.41, 95% CI: 2.07 to 2.80). PCI was associated with higher mortality in SCAD patients presenting with non–ST-segment elevation myocardial infarction (OR: 2.01, 95% CI: 1.00 to 4.47), but not with STEMI (OR: 0.62, 95% CI: 0.41 to 0.96).
Conclusions Women presenting with AMI and SCAD appear to be at higher risk of in-hospital mortality. Lower rates of PCI were associated with improved survival, with evidence of worse outcomes when PCI was performed for SCAD in the setting of non with ST-segment elevation myocardial infarction.
Spontaneous coronary artery dissection (SCAD) is an infrequent disease that presents as acute coronary syndrome or sudden cardiac death, with a wide clinical spectrum of severity. Since it was first described in 1931 on an autopsy of a young woman with sudden cardiac death (1), the true prevalence of this uncommon entity has been difficult to establish. The incidence of SCAD ranges from 0.07% to 1.1% and is considered predominantly a disease of women (>90% in some cohorts) (2–6). There is a paucity of data on the outcomes of SCAD patients, especially in those presenting with acute myocardial infarction (AMI). Moreover, controversies exist on the optimum management strategies in this patient population, due to the challenges encountered during percutaneous coronary intervention (PCI) and risk of intramural hematoma propagation with angioplasty or stent deployment (7). Thus, the concrete benefit of PCI for SCAD in the setting of AMI has yet to be established.
The aim of this study was to conduct a population-based analysis, using the National Inpatient Sample (NIS) database, to determine the prevalence, trends, and in-hospital mortality of women with SCAD in the setting of AMI, and to evaluate the impact of PCI on in-hospital mortality.
The NIS database is considered the largest all-payer inpatient database in the United States, including data on more than 7 million patient discharge records each year (8). The NIS is a part of the Healthcare Cost and Utilization Project (HCUP), sponsored by the Agency for Healthcare Research and Quality (AHRQ). It is constructed from billing data submitted by hospital to data organizations across the United States. NIS includes patients under Medicaid, Medicare, and private insurance, and the uninsured. The data comprise approximately 20% of stratified sample discharges from U.S. hospitals; long-term acute care and rehabilitation hospitals are not included (8). In 2012, the NIS database was redesigned for more acute representation of national estimates, where the sample records represented a random sample of discharge records from all HCUP participating hospitals rather than a random sample of hospitals from which all discharges were obtained.
The NIS data include a primary (principal) diagnosis and 24 secondary diagnoses together with 15 procedural diagnoses, all in the International Classification of Diseases, Ninth Edition, Clinical Modification [ICD-9-CM] coding format. The primary discharge diagnosis is usually considered the main reason for hospitalization. Other variables include sex, age, race, primary payer (e.g., Medicare, Medicaid, private, or uninsured), hospital characteristics (e.g., location, bed size, etc.), median home income (in percentile groups), day of admission (weekend or weekday), length of hospital stay, total hospital charges, and discharge status (e.g., dead or alive). A discharge weight variable is also available to calculate the national estimates of the various variables previously stated. Each record included in the NIS database is deidentified with absence of any personal identifying information.
Validation of data
The NIS data are crosschecked by the AHRQ annually to ensure the internal validity of the data. Data from the NIS have been compared with the American Hospital Association Annual Survey Database, the National Hospital Discharge Survey from the National Center for Health Statistics, and the Med-PAR inpatient database from Centers for Medicare & Medicaid Services in prior studies with comparable estimates (9).
Patient selection and inclusion criteria
The NIS database years 2009 to 2014 were queried for women with a primary diagnosis of AMI, defined as either non–ST-segment elevation myocardial infarction (NSTEMI) (ICD-9-CM code of 410.7x) or ST-segment elevation myocardial infarction (STEMI) (ICD-CM 9 codes of 410.1x, 410.2x, 410.3x, 410.4x, 410.5x, 410.6x, 410.8x, and 410.9x). In order to ensure the diagnosis of SCAD (ICD-9-CM 414.12), only patients with a procedural diagnosis of coronary angiography (ICD-9-CM codes of 88.53, 88.54, 88.55, 88.56, 37.22, or 37.23) and/or PCI (ICD-9-CM codes of 00.66, 36.06, and 36.07) were included. To decrease the chances of coding errors, patients with a concomitant diagnosis code of accidental puncture or laceration during a procedure (ICD-9-CM 998.2) were excluded.
Covariables and comorbidities definitions
To provide for a robust analysis and to minimize confounders, a large number of covariables were included in the analysis. Most of the included variables were readily supplied by the NIS database including age, sex, race (white, African American, Hispanic, Asian, and other), insurance type, median home income (0 to 25th percentile, 26th to 50th percentile, 51st to 75th percentile, and 76th to 100th percentile), day of admission, hospital bed size (large, medium, small), location (urban, urban teaching, rural), and region (Northeast, Midwest/North Central, South, and West). Other covariables and comorbidities included in the analysis (e.g., hypertension, diabetes, etc.) were either derived from the Elixhauser list of comorbidities supplied by the NIS (10) or manually coded using their specific ICD-9-CM codes. A list of ICD-9-CM codes for the covariables included in the current analysis is described in the Online Table 1.
The main outcome of interest was in-hospital mortality (referred to as the “died” variable in the NIS database). In-hospital mortality was compared between patients with and without SCAD in both the unadjusted and propensity-matched samples.
National weighted estimates were calculated using the discharge weight variable supplied by NIS. For descriptive purposes, frequencies were used to estimate categorical variables, means and standard deviations for non-skewed continuous variables, and medians with 25th to 75th percentiles ranges for skewed continuous variables. Frequencies of covariables were compared between patients with and without SCAD using the Pearson chi-square test. Means were compared using independent sample Student t tests, whereas medians were compared using the Mood median test. A linear-by-linear association trend test (Mantel-Haenszel test for trend) was used to assess trends of categorical variables.
A propensity score match was constructed using 55 patient and hospital covariables to choose a group of controls with similar characteristics to the SCAD patient population. A nearest neighbor, 1:2 matching was adopted. The propensity score match robustness was tested by evaluation of standardized mean differences (bias %) between the unmatched and matched variables with a cut level of 0.1 (11). The incidence of in-hospital mortality was compared in both groups by the Pearson chi-square test and univariable logistic regression. A subgroup analysis was performed in the matched population to evaluate the impact of different covariables on the outcome of in-hospital mortality. Covariables included in the subgroup analyses were age (above and below the median), hypertension, diabetes mellitus, acute decompensated heart failure, clinical presentation (i.e., STEMI vs. NSTEMI), and PCI.
To ensure accuracy of our estimates, we conducted a secondary analysis using multivariable backward selection logistic regression with in-hospital mortality being the dependent variable and the previously stated 58 covariates being as independent ones, using 0.05 probability of stepwise entry and 0.1 for removal. We used multiple statistical analyses models to ensure the accuracy of the effect size (12). To account for hospital cluster effect, multiple hospital variables, for example, hospital size, location, and region, were included in both the propensity score matching and multivariable regression analysis. To assess the impact of PCI on in-hospital mortality in the SCAD group, a multivariable logistic regression analysis was performed in both NSTEMI and STEMI SCAD patients with in-hospital mortality being the dependent variable and the rest of covariables together with PCI being independent variables.
All statistical analyses were performed by SPSS software (version 23.0, IBM, Armonk, New York) with a 2-sided p value of <0.05 as the cutoff for statistical significance and odds ratio (OR) with 95% confidence interval (CI) as a measure of effect size. A list of all covariables included in the propensity match construction and multivariable regression analyses is described in Online Table 2.
Of 1,259,158 women discharged with a primary diagnosis of AMI in the years 2009 to 2014, 752,352 women with AMI who underwent coronary angiography were included. Of these, 7,347 women had a co-diagnosis of SCAD, representing 0.98% of the women presenting with AMI (Figure 1). And 3,296 SCAD women had the diagnosis of STEMI, representing 44.9% of the total SCAD women with AMI. Compared with women without SCAD, women with SCAD were younger (mean age 61.7 ± 15.1 vs. 67.1 ± 13.2 years of age), with fewer risk factors for coronary artery disease. Table 1 illustrates the incidences of various patients’ and hospital characteristics of the women with SCAD and concomitant AMI.
There was a slight increase in the incidence of SCAD with time (929 vs. 961 per 100,000 women with AMI in years 2009 and 2014, respectively, ptrend = 0.042) (Figure 2). On the other hand, the trends of PCI use in SCAD women with AMI had decreased significantly (82.5% vs. 69.1% for 2009 and 2014 years, respectively, ptrend <0.0001) (Figure 3). This was true for women with both NSTEMI and STEMI SCAD patients, with a larger trend toward lower PCI use in NSTEMI SCAD patients (Online Figure 1). The incidence of in-hospital mortality was higher with SCAD compared with patients without SCAD (6.8% vs. 3.4%, ORunadjusted: 2.11, 95% CI: 1.92 to 2.31; p < 0.0001). However, the incidence of in-hospital mortality in SCAD patients appeared to be decreasing with time, 11.4% versus 5.0% for years 2009 and 2014, respectively; ptrend <0.0001 (Figure 3, Online Figure 2). Of 7,347 women with SCAD in our cohort, 1,687 were <50 years of age. The different characteristics of women <50 or ≥50 years of age with SCAD are described in Online Table 3; as expected, older women had greater comorbidities compared with younger women. Table 2 illustrates the patients’ characteristics and in-hospital mortality trends of women with AMI in the setting of SCAD.
Propensity-matched in-hospital mortality
Propensity score matched 7,332 women with SCAD and 14,352 women without SCAD (Figure 1). Both groups had well-matched patient and hospital characteristics (Table 1). The propensity match appeared to be adequately balanced with standard mean differences <10% for all included covariables (Table 1, Figure 4). The incidence of in-hospital mortality was higher in SCAD patients compared with those without SCAD (6.8% vs. 3.8%, ORadjusted: 1.87, 95% CI: 1.65 to 2.11; p < 0.0001). Subgroup analysis for the pre-defined covariables, illustrated that age >62 years, hypertension, and undergoing PCI were all associated with a higher incidence of in-hospital mortality in women with SCAD and AMI (Figure 5).
Multivariable regression-adjusted in-hospital mortality
After adjustment, the incidence of in-hospital mortality remained higher in the SCAD patients compared with patients without SCAD (ORadjusted: 2.41, 95% CI: 2.07 to 2.80; p < 0.0001). Table 3 illustrates the covariables associated with higher OR of in-hospital mortality after multivariable adjustment.
Impact of PCI on survival in SCAD patients
In the SCAD group, PCI was associated with increased in-hospital mortality in NSTEMI patients by both unadjusted (ORunadjusted: 2.11, 95% CI 1.39-3.21, p < 0.0001) and multivariable regression adjusted analyses (ORadjusted: 2.01, 95% CI: 1.00 to 4.47; p = 0.02), but not STEMI patients (ORunadjusted: 1.37, 95% CI: 0.99 to 1.89; p = 0.06; and ORadjusted: 0.62, 95% CI: 0.41 to 0.96; p = 0.03).
In this large population-based study, the reported prevalence of SCAD in women with AMI undergoing coronary angiography was ∼1%. Women with SCAD in the setting of AMI had a higher likelihood of in-hospital mortality compared with those without SCAD. Additionally, PCI in SCAD patients appeared to be associated with an increased incidence of in-hospital mortality compared with conservative management, especially in patients presenting with NSTEMI.
To the best of our knowledge, this study is the largest study to date on women with SCAD in the setting of AMI (6,13–22). The incidence of SCAD in women with AMI in our study population was similar to prior published registries, ranging from 0.2% to 1.5%. However, this number may have underestimated the true incidence of SCAD, because SCAD might be missed by standard coronary angiography, especially if associated with an intramural hematoma and/or in the absence of an intimal entry point. Although the diagnostic yield of coronary angiography for SCAD is increased by the use of intravascular ultrasound or optical coherence tomography (23,24), we believe that both modalities were not widely utilized for AMI at the time these study data were collected. It is highly likely that the coronary dissection most commonly detected by angiographers in 2009 to 2014 is type I, where there is an obvious dissection plane or dye hang-up (25). This type of angiographic SCAD represents <30% of total SCAD detected on coronary angiography (22). The type 2 and 3 angiographic variants of SCAD are more likely to be missed because of lack of familiarity with these SCAD appearance, and likelihood of misdiagnosis as atherosclerotic changes (25).
Although traditionally considered a rare cause of AMI, SCAD has now been shown to be the underlying etiology in 10% to 24% of myocardial infarctions in women younger than 50 years of age (3,20,26). The average age of SCAD patients in our study was higher than previous reports; however, this finding is not surprising because SCAD had been reported before in elderly patients, with the oldest reported being 84 years of age (22). It is possible that the incidence of SCAD is underreported in elderly women, because the likelihood of having concomitant atherosclerotic coronary artery disease is higher, and operators may associate the dissection with the atherosclerotic process.
SCAD patients commonly present either as NSTEMI or STEMI infarction (27). The results of the current study question the benefit of PCI in SCAD patients presenting with AMI and suggest that conservative management might be associated with better short-term survival. The current findings are different from previous reports on women presenting with NSTEMI that showed a benefit of routine invasive strategy (28), and they concur with prior reports on women with SCAD illustrating that a conservative approach may be preferable (14) because the majority of SCAD lesions heal spontaneously when assessed in follow-up (22). In an observational study of 134 SCAD patients, the decision of a conservative versus revascularization approach was influenced predominantly by location of the dissection (proximal vs. distal) and level of flow impairment (Thrombosis In Myocardial Infarction [TIMI] flow grade 0 to 1 vs. 2 to 3); clinical presentation was not an influential factor. Furthermore, successful PCI was achieved in only 72% of patients who underwent PCI in that study (18). Technical failure of PCI was a major problem occurring in up to 35% of SCAD patients who underwent PCI (13). In another registry, PCI failure occurred in up to 53% of the cases, with one-half of these being in nonoccluded vessels with preserved flow (14). At the same time, SCAD lesions usually heal in the majority of patients who underwent conservative management, resulting in restoration of normal coronary flow. This also explains why long-term results with CABG may be suboptimal as venous and arterial grafts become occluded or atretic with competitive flow (13). Another major problem in SCAD patients is recurrence. Dissection recurrence usually presents in a new vessel or location different from the originally dissected vessel (14,29), with recurrence rates of 10% in the first week following presentation and up to 27% in 5 years (14). Unlike with atherosclerosis, PCI does not prevent recurrence of AMI with SCAD because it usually occurs in a different segment. In one study, SCAD recurrence was similarly high between PCI and conservatively managed patients (14). Furthermore, the majority of patients with conservatively managed SCAD have normalization of segmental wall motion abnormality and left ventricular dysfunction at follow-up (30). All of this supports the feasibility of adopting a conservative approach in SCAD patients presenting with AMI, with the option of invasive or non-invasive follow-up by computerized topography angiogram to document healing of the lesions (21).
Optimal medical therapy plays a cardinal role in the management of SCAD patients with AMI as previously described (31). Antiplatelet therapy with aspirin and adenosine di-phosphate inhibitors are frequently administered following the acute event for 1 to 12 months, followed by aspirin indefinitely. Beta-blockers are routinely administered long term to reduce arterial shear stress, especially because recent data showed lower risk of recurrent SCAD with beta-blockade (32). Selective use of statins for patients with pre-existing dyslipidemia, and selective use of angiotensin-converting enzyme inhibitor/angiotensin receptor blocker are also indicated for patients with left ventricular dysfunction after AMI (31).
The large sample size from the NIS database provided an excellent source for further evaluation of rare diseases, such as SCAD, in a high-risk patient population presenting with AMI. The NIS database had been used to evaluate outcomes of various special patient populations on a national scale, providing accurate estimates of both the clinical presentations and outcomes (28,33).
First, being an administrative database, the data are subject to errors in coding or misdiagnosis. For example, the definition of SCAD was based on an ICD-9-CM code; such a code could be misplaced for a patient with iatrogenic coronary artery dissection. In an attempt to lower the chances of such error, we excluded any patient record that had a simultaneous ICD-9-CM code of “procedure injury” with the SCAD ICD-9-CM code. Second, although we evaluated all outcomes of interest after adjusting for more than 50 patient and hospital covariables, there is always the possibility of confounding biases that were not included in our analysis. Third, the NIS database lacks information regarding medical therapy, echocardiographic findings, or angiographic findings that could be used to further stratify the risk of poor outcomes in the current patient population. Fourth, the reason for choosing or deferring PCI could not be determined from the current data, with the lack of intravascular ultrasound or optical coherence tomography use data that could guide this decision. It is conceivable that patients who underwent PCI had worse geographic anatomic dissections (e.g., left main or proximal artery dissection, multivessel dissection) that were not captured in the NIS database, and may have contributed to bias for worse outcomes in this cohort. These findings support the current recommendations that conservative management should be the preferred initial strategy for management SCAD in women presenting with AMI.
Women presenting with AMI and concomitant SCAD have higher likelihood of in-hospital mortality compared with those without SCAD. A lower incidence of in-hospital mortality was evident in these patients with time and concurring with the decreased utilization of PCI. And thus, supporting the current recommendations that conservative management should be the preferred initial strategy for management of AMI women presenting with SCAD.
WHAT IS KNOWN? SCAD is an uncommon cause of AMI in women but more frequently encountered in those <50 years of age. PCI is commonly used to treat women with SCAD; however, there are insufficient data on the short-term outcomes of these patients.
WHAT IS NEW? In women admitted for AMI, the presence of SCAD appears to be an independent predictor of increased in-hospital mortality. The incidence of mortality might be higher with PCI compared with conservative management especially in patients presenting with NSTEMI. Thus, an initial conservative approach could be considered for management of SCAD in the setting of AMI.
WHAT IS NEXT? Further prospective studies are required to identify a subset of AMI women presenting with SCAD who could benefit from PCI. Given the rarity of this condition, registry-based randomized trials might be the most adequate method to investigate the best therapeutic approach for AMI women with SCAD.
Dr. Anderson is a consultant for Biosense Webster, a Johnson & Johnson Company. Dr. Saw has received research grants for SCAD research from the Canadian Institutes of Health Research, Heart and Stroke Foundation of Canada, University of British Columbia Division of Cardiology, AstraZeneca, Abbott Vascular, St. Jude Medical, and Servier; has been a consultant for Boston Scientific, Abbott Vascular, and St. Jude Medical; has received honoraria for proctoring from Boston Scientific and St. Jude Medical; and has received speaker honoraria for SCAD from AstraZeneca, St. Jude Medical, and Sunovion. All other authors have reported that they have no relationships relevant to the contents of this paper to disclose. The first two authors contributed equally to this work.
- Abbreviations and Acronyms
- acute myocardial infarction
- confidence interval
- International Classification of Diseases-9th Edition-Clinical Modification
- National Inpatient Sample
- non–ST-segment elevation myocardial infarction
- odds ratio
- percutaneous coronary intervention
- spontaneous coronary artery dissection
- ST-segment elevation myocardial infarction
- Received June 19, 2017.
- Revision received July 30, 2017.
- Accepted August 2, 2017.
- 2017 American College of Cardiology Foundation
- Pretty H.
- Saw J.,
- Ricci D.,
- Starovoytov A.,
- Fox R.,
- Buller C.E.
- ↵Healthcare Cost and Utilization Project (HCUP). Overview of the National (Nationwide) Inpatient Sample (NIS). Available at: http://www.hcup-us.ahrq.gov/nisoverview.jsp. Accessed May 8, 2017.
- ↵Elixhauser A, Steiner C, Kruzikas D. Comorbidity software documentation. 2004. HCUP Methods Series Report# 2004-1. ONLINE February 6, 2004. US Agency for Healthcare Research and Quality. Available at: http://www.hcup-us.ahrq.gov/toolssoftware/comorbidity/comorbidity.jsp. Accessed May 8, 2017.
- Elze M.C.,
- Gregson J.,
- Baber U.,
- et al.
- Tweet M.S.,
- Hayes S.N.,
- Pitta S.R.,
- et al.
- Tweet M.S.,
- Eleid M.F.,
- Best P.J.,
- et al.
- Alfonso F.,
- Paulo M.,
- Lennie V.,
- et al.
- McGrath-Cadell L.,
- McKenzie P.,
- Emmanuel S.,
- Muller D.W.,
- Graham R.M.,
- Holloway C.J.
- Rogowski S.,
- Maeder M.T.,
- Weilenmann D.,
- et al.
- Nakashima T.,
- Noguchi T.,
- Haruta S.,
- et al.
- Roura G.,
- Ariza-Solé A.,
- Rodriguez-Caballero I.F.,
- et al.
- Saw J.,
- Aymong E.,
- Sedlak T.,
- et al.
- Motreff P.,
- Malcles G.,
- Combaret N.,
- et al.
- Ho S.W.,
- Lin W.,
- Chan K.H.,
- Seow S.C.
- Elgendy I.Y.,
- Mahmoud A.N.,
- Mansoor H.,
- Bavry A.A.
- Eugène M.,
- Siam-Tsieu V.,
- Pillière R.,
- Deblaise J.,
- Dubourg O.,
- Mansencal N.
- Franco C.,
- Starovoytov A.,
- Heydari M.,
- Mancini G.B.,
- Aymong E.,
- Saw J.
- Saw J.,
- Mancini G.B.,
- Humphries K.H.
- Saw J.,
- Humphries K.H.,
- Aymong E.,
- et al.
- Mahmoud A.N.,
- Elgendy I.Y.,
- Mansoor H.,
- et al.