In contrast to atherosclerotic acute myocardial infarction (AMI), conservative therapy is considered preferable in the acute management of spontaneous coronary artery dissection (SCAD) if clinically possible. The present study aimed to investigate factors associated with treatment strategy for SCAD. Women aged ≤60 years with AMI and SCAD were retrospectively identified in the Nationwide Readmissions Database 2010 to 2015 and were divided into revascularization and conservative therapy groups. The revascularization group (n = 1,273, 68.0%), compared with the conservative therapy group (n = 600, 32.0%), had ST-elevation AMI (STEMI) (anterior STEMI, 20.3% vs 10.5%; inferior STEMI, 25.1% vs 14.5%; p <0.001) and cardiogenic shock (10.8% vs 1.8%; p <0.001) more frequently. Multivariable logistic regression analysis demonstrated that anterior STEMI (vs non-STEMI, odds ratio 2.89 [95% confidence interval 2.08 to 4.00]), inferior STEMI (2.44 [1.85 to 3.21]), and cardiogenic shock (5.13 [2.68 to 9.80]) were strongly associated with revascularization. Other factors associated with revascularization were diabetes mellitus, dyslipidemia, smoking, renal failure, and pregnancy/delivery-related conditions; whereas known fibromuscular dysplasia and admission to teaching hospitals were associated with conservative therapy. Propensity-score matched analyses (546 pairs) found no significant difference in in-hospital death, 30-day readmission, and recurrent AMI between the groups. In conclusion, STEMI presentation, hemodynamic instability, co-morbidities, and setting of treating hospital may affect treatment strategy in women with AMI and SCAD. Further efforts are required to understand which patients benefit most from revascularization over conservative therapy in the setting of SCAD causing AMI.
Graphical abstract
Spontaneous coronary artery dissection (SCAD) predominantly affects younger women, with an incidence of 1% to 4% in patients with acute coronary syndrome. , In contrast to atherosclerotic acute myocardial infarction (AMI), , conservative therapy is considered preferable in the acute management of SCAD because a dissected artery often heals after conservative management and unsuccessful revascularization is common (10% to 40%). , , Revascularization should be considered in high-risk patients with ongoing ischemia, hemodynamic instability, or left main trunk dissection. The rate of revascularization for SCAD varied among previous studies (12% to 100%). , , No study has examined the factors associated with treatment strategy for SCAD in a nationwide setting. Additionally, although readmission is occasionally required in SCAD patients due to recurrent ischemia, data on readmission outcomes stratified by treatment strategy are limited. Thus, the present study sought to investigate the determinants and outcomes of treatment strategy in AMI with SCAD using a United States population-based database.
Methods
The present study is a retrospective analysis using the Nationwide Readmissions Database (NRD). NRD is a publicly available, population-based database of all-age, all-payer discharges released by the Healthcare Cost and Utilization Project of the Agency for Healthcare Research and Quality (AHRQ). The database includes data on approximately 17 million discharges per year with verified patient linkage numbers in more than 20 states, representing approximately 50% of all hospitalizations in the United States. The present study used the NRD data between January 1, 2010 and September 30, 2015. Diagnoses and procedures have been recorded using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. This study was exempted from the review of the institutional review board due to the anonymized and de-identified nature of the data in the NRD.
We selected study patients based on the following inclusion and exclusion criteria. Inclusion criteria were (1) female gender (∼90% of SCAD patients were women in previous relatively large studies), , , and (2) patients who had diagnoses of AMI (ICD-9-CM Codes 410.0 to 410.9) and coronary artery dissection (ICD-9-CM Code 414.12, as in previous studies. The exclusion criteria were (1) >60 years, (2) elective admissions or missing data on type of admission, (3) patients who underwent coronary angiography, (4) patients with a history of myocardial infarction, (5) patients with a history of percutaneous coronary intervention (PCI), (6) patients with a history of coronary artery bypass grafting (CABG), and (7) concomitant accidental puncture or laceration during a procedure (to avoid iatrogenic coronary artery dissection in the same way as previous SCAD studies using ICD-9-CM; , Supplementary Table 1). Eligible patients were categorized into 2 groups according to the in-hospital treatment: revascularization group and conservative therapy group. The revascularization group was defined as patients who underwent PCI or/and CABG during index hospitalization. The conservative therapy group was defined as patients who underwent neither PCI nor CABG during index hospitalization.
The following patient characteristics were identified: age, weekend admission, AMI category, atrial fibrillation, heart failure, cardiogenic shock, number of Elixhauser co-morbidities, hypertension, diabetes mellitus, dyslipidemia, obesity, smoking, renal failure, fibromuscular dysplasia (FMD), migraine, pregnancy/delivery-related condition, bed size of hospital (small, medium, or large), and teaching status of hospital (metropolitan teaching hospital or others). AMI category was determined based on the ICD-9-CM codes and was classified into 5 categories: non–ST-elevation myocardial infarction (NSTEMI), anterior STEMI, inferior STEMI, lateral/posterior/other STEMI, or unspecified AMI. Data on the 29 Elixhauser co-morbidities were provided by the AHRQ co-morbidity measures (Supplementary Table 2). Other co-morbidities used in the present study were identified using the corresponding ICD-9-CM codes or the AHRQ co-morbidity measures (Supplementary Table 1). The variable of ≥4 of Elixhauser co-morbidities was used to indicate a high co-morbidity burden.
The outcomes that we assessed were in-hospital all-cause death, 30-day any-cause readmission, and 30-day recurrent AMI. In the NRD, patients’ admissions were tracked across hospitals within a state during the same calendar year using a de-identified unique patient linkage number, and the time between admissions was calculated. Thus, 30-day outcomes were evaluated in patients discharged alive from index SCAD hospitalization before December in 2010 to 2014 or before September in 2015 to allow for 30-day follow-up after discharge in the NRD. The 30-day recurrent AMI was identified from the diagnoses recorded during readmissions that occurred within 30 days of discharge.
Categoric variables were presented as numbers and percentages and were compared using Fisher’s exact test. Categoric variable data with n ≤10 were suppressed in compliance with the privacy protection policy of the Healthcare Cost and Utilization Project Data Use Agreement. Continuous variables were presented as mean ± standard deviation or median (interquartile range [IQR]) compared using t test or the Mann–Whitney U test. We compared patient characteristics between the revascularization and conservative therapy groups in all eligible patients (overall cohort) as well as in subcohorts (STEMI cohort and NSTEMI cohort). Factors associated with revascularization were examined using multivariable logistic regression models in the overall cohort as well as the STEMI and NSTEMI cohorts. All variables of patient characteristics were included as covariates in the multivariable analyses. In the analysis of the overall cohort, the category of unspecified AMI is considered missing value because it cannot be included in either STEMI or NSTEMI. Thus, multiple imputation was performed to produce 20-copy datasets in which the missing data (unspecified AMI) were imputed, and the pooled results were obtained from the 20-copy datasets based on Rubin’s rules. In addition, as a sensitivity analysis, another multivariable analysis for revascularization was performed using the cohort after excluding CABG cases because CABG cases without PCI would likely be cases with challenging dissected lesions (e.g., dissection of left main trunk or proximal lesion of left anterior descending artery) for which PCI was not appropriate, and cases with both PCI and CABG would likely be cases of PCI failure followed by bailout CABG. , ,
If outcomes are compared in this observational study, confounding by indication for treatment strategy must be considered because conservative therapy is likely selected in less severe SCAD cases with stable hemodynamic and maintained coronary artery flow. Thus, a propensity score matching was conducted to adjust for measured confounding factors. Propensity scores were estimated using a logistic regression model including all patient characteristics as covariates. 1 : 1 nearest-neighbor matching was performed based on the estimated propensity scores within a caliper of 0.2 times the pooled standard deviation of the logit of the propensity scores for the cohort. Balances of patient characteristics between the groups were examined assessed by calculating standardized differences; a difference of <0.10 was considered to indicate good balance.
All hypothesis tests had a 2-sided significance level of 0.05. All statistical analyses were conducted using IBM SPSS Statistics, version 26 (IBM Corp., Armonk, New York) and Stata/IC 15.0 (StataCorp, College Station, Texas).
Results
We identified 1,873 patients (mean age, 47.4 ± 8.7 years) for the present study, of whom 68% underwent revascularization ( Figure 1 ). The revascularization group included 1,051 patients treated with PCI alone (56.1%), 127 with CABG alone (6.8%), and 95 with PCI and CABG (5.1%). The proportion of conservative therapy increased over the 6 years ( Figure 2 ). The proportion of conservative therapy was greater in the patients treated at hospitals in large metropolitan areas (37.0%), followed by hospitals in small metropolitan areas (27.2%), compared with those in micropolitan areas (13.9%) (Supplementary Table 3). Overall, intravascular ultrasound (IVUS) was used in 167 patients (8.9%), but it was used in only 13 patients (0.69%) who did not receive revascularization. The rate of IVUS use did not differ between teaching hospitals and nonteaching hospitals (9.4% vs 8.0%, p = 0.347). Among the 1,051 patients treated with PCI alone, coronary stent was implanted in 92.3% (drug-eluting stent was implanted in 70.5%) of the patients. Overall, the median length of stay was 3 days (IQR 2 to 4 days), and the length of stay was significantly longer in the revascularization group than the conservative therapy group (median 3 days [IQR 2 to 4 days] vs 4 days [IQR 2 to 6 days]; p <0.001). Eligible patients were also categorized into 784 with STEMI (41.9%), 986 with NSTEMI (52.6%), and 103 with unspecified AMI (5.5%). In the STEMI cohort, 78.8% underwent revascularization (65.4% PCI alone, 6.3% CABG alone, and 7.1% PCI and CABG), whereas in the NSTEMI cohort, 58.3% underwent revascularization (48.2% PCI alone, 6.7% CABG alone, and 3.4% PCI and CABG).
The revascularization group had STEMI, atrial fibrillation, cardiogenic shock, ≥4 of Elixhauser co-morbidities, renal failure, diabetes mellitus, dyslipidemia, and smoking significantly more frequently than the conservative therapy group ( Table 1 ). The conservative therapy group had FMD and migraine significantly more frequently and were treated at metropolitan teaching hospitals significantly more frequently. Almost all results in the STEMI and NSTEMI cohorts were consistent with those in the overall cohort. Propensity score matching in the overall cohort created a cohort of 546 pairs, in which the characteristics were well-balanced between the 2 groups. The propensity-score matched patients, compared without the unmatched patients, had several different characteristics, with lower rates of in-hospital death and 30-day any-cause readmission (Supplementary Table 4).
| Overall Cohort (n=1,873) | ST-Segment Elevation Myocardial Infarction Cohort (n=784) | Non-ST-Elevation Myocardial Infarction Cohort (n=986) | Propensity Score-Matched Cohort (n=1,092) | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Revascula-rization | Conservative therapy | Revascula-rization | Conservative therapy | Revascula-rization | Conservative therapy | Revascula-rization | Conservative therapy | Absolute standardized difference, % | |||||
| (n=1,273) | (n=600) | p value | (n=618) | (n=166) | p value | (n=575) | (n=411) | p value | (n=546) | (n=546) | p value | ||
| Age category (years) | 0.089 | 0.165 | 0.129 | 0.778 | |||||||||
| ≤40 | 284 (22.3%) | 138 (23.0%) | 150 (24.3%) | 45 (27.1%) | 107 (18.6%) | 89 (21.7%) | 129 (23.6%) | 127 (23.3%) | 0.9 | ||||
| 41–50 | 432 (33.9%) | 230 (38.3%) | 206 (33.3%) | 64 (38.6%) | 202 (35.1%) | 158 (38.4%) | 200 (36.6%) | 211 (38.6%) | 4.2 | ||||
| 51–60 | 557 (43.8%) | 232 (38.7%) | 262 (42.4%) | 57 (34.3%) | 266 (46.3%) | 164 (39.9%) | 217 (39.7%) | 208 (38.1%) | 3.4 | ||||
| Weekend admission | 322 (25.3%) | 152 (25.3%) | >0.99 | 171 (27.7%) | 49 (29.5%) | 0.628 | 129 (22.4%) | 97 (23.6%) | 0.701 | 128 (23.4%) | 137 (25.1%) | 3.8 | |
| Acute myocardial infarction category | <0.001 | 0.270 | 0.984 | ||||||||||
| Non-ST-elevation | 575 (45.2%) | 411 (68.5%) | (-) | (-) | 575 (100.0%) | 411 (100.0%) | 363 (66.5%) | 361 (66.1%) | 0.8 | ||||
| Anterior ST-elevation | 259 (20.3%) | 63 (10.5%) | 259 (41.9%) | 63 (38.0%) | (-) | (-) | 65 (11.9%) | 61 (11.2%) | 2.3 | ||||
| Inferior ST-elevation | 320 (25.1%) | 87 (14.5%) | 320 (51.8%) | 87 (52.4%) | (-) | (-) | 84 (15.4%) | 86 (15.8%) | 1.0 | ||||
| Lateral, posterior or other ST-elevation | 39 (3.1%) | 16 (2.7%) | 39 (6.3%) | 16 (9.6%) | (-) | (-) | 14 (2.6%) | 16 (2.9%) | 2.2 | ||||
| Unspecified acute myocardial infarction | 80 (6.3%) | 23 (3.8%) | 20 (3.7%) | 22 (4.0%) | 1.9 | ||||||||
| Atrial fibrillation | 49 (3.8%) | ≤10 (≤1.7%) * | 0.010 | 22 (3.6%) | ≤10 (≤6.0%) * | 0.135 | 25 (4.3%) | ≤10 (≤2.4%) * | 0.047 | ≤10 (≤1.8%) * | ≤10 (≤1.8%) * | 0.813 | 2.9 |
| Heart failure | 23 (1.8%) | ≤10 (≤1.7%) * | 0.706 | 11 (1.8%) | ≤10 (≤6.0%) * | 0.535 | 6 (1.0%) | ≤10 (≤2.4%) * | 0.480 | ≤10 (≤1.8%) * | ≤10 (≤1.8%) * | 0.579 | 5.1 |
| Cardiogenic shock | 138 (10.8%) | 11 (1.8%) | <0.001 | 93 (15.0%) | ≤10 (≤6.0%) * | <0.001 | 34 (5.9%) | ≤10 (≤2.4%) * | <0.001 | 15 (2.7%) | 11 (2.0%) | 0.552 | 4.8 |
| ≥4 of Elixhauser comorbidities | 205 (16.1%) | 55 (9.2%) | <0.001 | 84 (13.6%) | 12 (7.2%) | 0.032 | 106 (18.4%) | 42 (10.2%) | <0.001 | 48 (8.8%) | 48 (8.8%) | 1.000 | 0.0 |
| Hypertension | 601 (47.2%) | 263 (43.8%) | 0.180 | 256 (41.4%) | 71 (42.8%) | 0.790 | 301 (52.3%) | 185 (45.0%) | 0.024 | 241 (44.1%) | 235 (43.0%) | 0.760 | 2.2 |
| Diabetes mellitus | 229 (18.0%) | 43 (7.2%) | <0.001 | 93 (15.0%) | ≤10 (≤6.0%) * | 0.002 | 125 (21.7%) | 32 (7.8%) | <0.001 | 46 (8.4%) | 43 (7.9%) | 0.825 | 2.0 |
| Dyslipidemia † | 558 (43.8%) | 192 (32.0%) | <0.001 | 253 (40.9%) | 40 (24.1%) | <0.001 | 268 (46.6%) | 145 (35.3%) | <0.001 | 187 (34.2%) | 184 (33.7%) | 0.898 | 1.2 |
| Obesity ‡ | 282 (22.2%) | 123 (20.5%) | 0.435 | 119 (19.3%) | 24 (14.5%) | 0.175 | 147 (25.6%) | 94 (22.9%) | 0.367 | 112 (20.5%) | 108 (19.8%) | 0.821 | 1.8 |
| Smoker | 531 (41.7%) | 167 (27.8%) | <0.001 | 267 (43.2%) | 46 (27.7%) | <0.001 | 236 (41.0%) | 115 (28.0%) | <0.001 | 173 (31.7%) | 165 (30.2%) | 0.647 | 3.2 |
| Renal failure | 59 (4.6%) | ≤10 (≤1.7%) * | <0.001 | 26 (4.2%) | ≤10 (≤6.0%) * | 0.170 | 30 (5.2%) | ≤10 (≤2.4%) * | 0.001 | ≤10 (≤1.8%) * | ≤10 (≤1.8%) * | 0.789 | 3.3 |
| Fibromuscular dysplasia | ≤10 (≤0.8%) * | 11 (1.8%) | <0.001 | 0 (0.0%) | ≤10 (≤6.0%) * | <0.001 | 3 (0.5%) | ≤10 (≤2.4%) * | 0.289 | ≤10 (≤1.8%) * | ≤10 (≤1.8%) * | 0.687 | 5.0 |
| Migraine | 66 (5.2%) | 52 (8.7%) | 0.006 | 28 (4.5%) | 18 (10.8%) | 0.004 | 37 (6.4%) | 34 (8.3%) | 0.318 | 38 (7.0%) | 41 (7.5%) | 0.815 | 2.1 |
| Pregnancy/delivery related condition | 88 (6.9%) | 34 (5.7%) | 0.633 | 46 (7.4%) | 12 (7.2%) | >0.99 | 31 (5.4%) | 20 (4.9%) | 0.772 | 28 (5.1%) | 32 (5.9%) | 0.691 | 3.2 |
| Hospital status | |||||||||||||
| Bed size of hospital | 0.238 | 0.650 | 0.142 | 0.964 | |||||||||
| Small | 68 (5.3%) | 35 (5.8%) | 35 (5.7%) | ≤10 (≤6.0%) * | 27 (4.7%) | 27 (6.6%) | 31 (5.7%) | 29 (5.3%) | 1.6 | ||||
| Medium | 271 (21.3%) | 147 (24.5%) | 122 (19.7%) | 37 (22.3%) | 129 (22.4%) | 107 (26.0%) | 129 (23.6%) | 128 (23.4%) | 0.4 | ||||
| Large | 934 (73.4%) | 418 (69.7%) | 461 (74.6%) | 122 (73.5%) | 419 (72.9%) | 277 (67.4%) | 386 (70.7%) | 389 (71.2%) | 1.2 | ||||
| Metropolitan teaching hospital | 792 (62.2%) | 444 (74.0%) | <0.001 | 377 (61.0%) | 122 (73.5%) | 0.003 | 356 (61.9%) | 304 (74.7%) | <0.001 | 391 (71.6%) | 393 (72.0%) | 0.946 | 0.8 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
