In-hospital and long-term outcomes in spontaneous coronary artery dissection with concurrent cardiac arrest: Systematic review and meta-analysis

Abstract

Background

Our understanding of factors predisposing patients with spontaneous coronary artery dissection (SCAD) to worse outcomes, such as concurrent sudden cardiac arrest (CA) and secondary prevention of sudden cardiac death in those patients, is limited.

Objective

We conducted the largest systematic review of studies assessing clinical outcomes in SCAD with concurrent CA.

Methods

This study was performed according to Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines. PubMed, Cochrane, and Scopus were searched using relevant search terms including “Spontaneous Coronary Artery Dissection,” “Ventricular Tachycardia,” “Ventricular Fibrillation,” “Sudden Cardiac Death,” and “Cardiac Arrest.” The search was conducted from database inception to January 2025.

Results

Out of 269 studies that underwent screening, 10 were included (n = 3978). In-hospital mortality, postdischarge mortality, recurrent myocardial infarction (MI) and recurrent SCAD occurred in 20%, 3%, 12%, and 9% of patients with SCAD and CA, respectively. When compared with patients with SCAD without CA, patients with SCAD and CA were at significantly higher risk of in-hospital mortality (risk ratio [RR] 6.7, 95% confidence interval [CI] 4.5–10.1, P < .00001), postdischarge mortality (RR = 5.9, 95% CI 1.7–19.9, P = .005), recurrent MI (RR = 3.3, 95% CI 2.0–5.4, P < .00001), and recurrent SCAD (RR = 1.9, 95% CI 1.1–3.3, P = .02). Out of a pooled 35 implanted cardiac defibrillators (ICDs) and wearable cardiac defibrillators (WCDs), there was only 1 appropriate and 1 inappropriate defibrillator discharge recorded over the follow-up period.

Conclusion

SCAD with concurrent CA is associated with worse in-hospital and long-term outcomes, although long-term rate of administered defibrillator therapies was low, supporting a conservative approach.

Graphical abstract

Key Findings.

• ▪Randomized data on clinical outcomes among patients with spontaneous coronary artery dissection (SCAD) and concurrent cardiac arrest (CA) are scarce, and guidelines on prevention of secondary sudden cardiac death in this patient group remain unclear.
• ▪We performed the largest systematic review on outcomes in SCAD with concurrent CA.
• ▪In-hospital mortality, postdischarge mortality, recurrent myocardial infarction (MI), and recurrent SCAD occurred in 20%, 3%, 12%, and 9% of patients with SCAD and CA, respectively.
• ▪SCAD with concurrent CA was associated with worse in-hospital and long-term outcomes including higher likelihood of mortality, acute heart failure, recurrent MI, and recurrent de novo SCAD compared with SCAD without CA.
• ▪Defibrillator therapy rate on long-term follow-up was low, with only 1 appropriate implantable cardiac defibrillator (ICD) discharge among a pooled 35 patients with wearable cardiac defibrillators (WCDs) or ICDs.

Introduction

Spontaneous coronary artery dissection (SCAD) is a form of acute coronary syndrome (ACS) characterized by the nontraumatic, noniatrogenic intimal tear or separation of the coronary arterial tunica intima and media by an intramural hematoma.^1,2 This is often in the setting of a pre-existing arteriopathy weakening the arterial wall, with additional precipitating emotional or physical stressors that incite the dissection.3, 4, 5, 6 Although its true prevalence is unknown, the increasing use of advanced intracoronary imaging techniques and greater awareness has led to SCAD being recognized as an important etiology of ACS. A disease of young-to-middle–aged women, SCAD is responsible for 24% to 35% of cases of ACS in women <60 years of age.⁷ Predisposing conditions to SCAD include fibromuscular dysplasia, inheritable connective tissue diseases, and pregnant/postpartum status.^3,4 The most common presentation of SCAD is non-ST segment elevation myocardial infarction (NSTEMI) followed by STEMI.⁸ The presence of ventricular arrhythmia (VA) has been reported in 3% to 13% of patients with SCAD and is associated with worse in-hospital and long-term outcomes, including mortality and major adverse cardiac events (MACE).⁹ Risk stratification for sudden cardiac death (SCD) in SCAD remains poorly explored, and guidelines are unclear on whether or not to pursue implantable cardiac defibrillator (ICD) placement in patients with a reversible cause of SCD.⁹ In addition, data remain scarce pertaining to optimal revascularization strategies, timing of ICD placement, long-term ICD outcomes, and role of wearable cardiac defibrillators (WCDs) at hospital discharge in patients with SCAD and concurrent cardiac arrest (CA).

To consolidate published literature on the topic, we performed a systematic review to explore the impact of concurrent CA in SCAD on in-hospital and long-term clinical outcomes and to investigate the burden of defibrillator therapies among patients received ICDs or WCDs.

Methods

This meta-analysis was carried out and reported per the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA).¹⁰ The study was registered on PROSPERO with registration ID CRD42024511286. Our study did not require ethics committee or institutional review board approval, as it was a systematic review, and no new clinical data were collected.

Literature search

We conducted a search for studies reporting outcomes of people with SCAD complicated by CA, which included ventricular tachycardia/ventricular fibrillation (VT/VF), pulseless electrical activity (PEA), and asystole. Searches were performed on PubMed/MEDLINE, Cochrane, and Scopus, using key words and standardized index terms such as “Spontaneous Coronary Artery Dissection,” “Ventricular Tachycardia,” “Ventricular Fibrillation,” “Sudden Cardiac Death,” and “Cardiac Arrest.” The search was conducted from date of database establishment to January 2025.

Inclusion and exclusion criteria for identification of studies

Articles were considered eligible for inclusion if they were written in English, were peer-reviewed observational studies or randomized clinical trials; included adults ≥18 years of age, reported on outcomes of patients with SCAD and concurrent CA, and included information on hospital follow-up and longer. Studies that did not report on the distinct outcomes of SCAD with CA included participants below 18 years of age or those with no hospital or postdischarge follow-up were excluded. We also excluded meta-analyses, systematic reviews, case reports, abstracts, editorials, commentaries, and letters to the editor.

Data screening

Two reviewers (S.K. and H.T.) screened the title and abstracts of all retrieved studies using the predefined selection criteria. The full texts of studies meeting the criteria were reviewed by O.B. and R.Q. Any disagreements regarding article inclusion were resolved after discussing with the senior reviewer (D.S.).

Data extraction

Two authors (H.T. and O.B.) extracted data simultaneously into a standard data extraction sheet. The data collected included in-hospital overall mortality, postdischarge overall mortality, major MACE, acute heart failure, cardiogenic shock, left ventricular ejection fraction (LVEF) (%), recurrent myocardial infarction (MI), and recurrent SCAD, which predominantly included recurrent de novo SCAD. Postdischarge mortality was defined as mortality after hospital discharge over follow-up period. CA either on presentation (including out-of-hospital CA) or index in-hospital CA was considered. In-hospital and long-term recurrent MI and SCAD on follow-up were included. Any discrepancies were resolved after consulting the senior reviewer (D.S.).

Assessment of methodologic quality and risk of bias

The reviewers then assessed the risk of bias for each study using the Newcastle-Ottawa Scale (NOS) for prospective and retrospective cohort studies (total score of 9 indicating lowest risk of bias), and adapted NOS for studies without control arm (total score of 6). The NOS rates observational cohort studies based on the selection of participants, comparability between the exposed and unexposed groups, and the assessment of the association between exposure and outcome. NOS scores are categorized as 7–9 for low risk of bias, 4–6 for moderate risk of bias, and ≤ for high risk of bias. All included studies scored ≥5 (Supplementary Material).

Statistical analysis

Data were extracted, and the results were presented as risk ratio (RR) and pooled percentage. For single-arm meta-analysis; we applied the formula: SE = √p (1-p)/ n, in which p stands for prevalence to calculate SE. Heterogeneity was assessed using the I² test, which calculates percentage variability in endpoints that is caused by heterogeneity rather than chance. The larger the I², the more likely that any observed statistical outcomes are caused by heterogeneity. An I² <30% for an outcome of interest was set as the standard for low heterogeneity. All statistical tests were 2-sided, and a P value < .05 was considered statistically significant. We used the random-effects model in the presence of significant heterogeneity (P < .1). The statistical analysis was performed using RevMan 5.4. The primary outcome of our study was to compare in-hospital overall mortality between patients with SCAD with and without CA. Secondary outcomes of interest included postdischarge overall mortality, acute heart failure, cardiogenic shock, recurrent MI, recurrent SCAD, and LVEF ≤40%. Finally, to assess for publication bias, we constructed funnel plots to assess asymmetry for each effect estimate and Egger’s test as a statistical test of funnel plot asymmetry (Supplementary Material).

Results

Search results and study characteristics

A total of 269 nonduplicate study abstracts and titles were screened. Following the predefined exclusion and inclusion criteria, 10 studies were selected for inclusion (Figure 1).^5,9,11, 12, 13, 14, 15, 16, 17, 18, 19 Detailed study characteristics are outlined in Table 1. Of 10 included studies, 2 were prospective,^13,15 and 8 were retrospective.^5,9,12,14,16, 17, 18, 19 The studies represented various geographic regions, with 4 studies from the United States,^5,9,12,14,19 2 from Canada,^13,15 1 from Switzerland,¹¹ 1 from Italy,¹⁷ 1 from Italy and Spain,¹⁸ and 1 from the Middle East (Kingdom of Saudi Arabia, United Arab Emirates, Kuwait, and Bahrain).¹⁶ A total of 3978 patients with SCAD were included, of whom 357 (9%) had concurrent CA. In addition to VT/VF, Giacobbe et al¹⁸ and Phan et al¹⁴ included PEA and asystole in their definition of cardiac arrest. Two studies reported CA on presentation, 3 studies reported CA on presentation and during hospitalization, and 5 studies did not specify CA timing. Six studies included SCAD confirmed by coronary imaging, including angiography. Daoulah et al¹⁶ (51%), Krittanawong et al⁵ (64%), and Tan et al¹² (78%) reported the lowest proportions of patients with SCAD of female sex. Pooled patient sample was predominantly of female sex and <65 years of age.

Figure 1.

PRISMA flow diagram detailing the identification and screening process used for study selection. PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses.

Table 1.

Characteristics of included studies^∗

Study	Type of study	Definition of CA	Total number of patients with SCAD	Patients with CA (n, %)	CA on presentation (n, % of total CA)	Follow-up duration (months)	Duration of hospital stay (days)	Angiographic confirmation of SCAD? (Y/N)	Age (years)	Female n (%)	Defibrillator therapies	Defibrillator outcomes	SCAD type (% overall)†	Overall revascularization (n, %)
Giacobbe 202418	Retrospective	PEA, asystole, VF, pulseless VT	375	20 (5)	6 (30) at presentation 14 (70) in-hospital	21 (10.7–46) (median)	7.5 ± 0.49	Y	52.6 ± 10.2	315 (84)	1 ICD 2 WCDs	-	Type 1: 23% Type 2: 50% Type 3: 6% Type 4: 26%	PCI: 130 (35%) CABG: 1 (0.2%)
Tan 202312	Retrospective (NRD)	ICD-10 codes for VT and VF	877	118 (13)	Not specified	1	7.8 ± 8.3	N	52.6 ± 13.8	687 (78)	-	-	-	-
Antonutti 202117	Retrospective	Sustained VT/VF	70	7 (10)	Not specified	39.1 (IQR 13.9–86.6) (median)	-	Y	52 (47-58)	60 (86)	-	-	Type 1: 56% Type 2: 43% Type 3: 1%	PCI: 13% CABG: 3%
Cheung 202013	Prospective	Sustained VT/VF	1056	84 (8)	Not specified	57.6 ± 39.6 (mean)	-	Y	49.3 ± 11.3	941 (89)	8 ICDs	1 shock for VT/VF	-	Overall: 174 (16.5%)
Phan 202014	Retrospective	Cardiac arrest, including PEA, asystole, VT/VF	208	11 (5)	Not specified	56.4 ± 37.2† (mean)	-	Y	48.53 ± 12.06	186 (94)	2 ICDs 1 WCD	0 shocks	-	PCI: 23 (11%) CABG: 9 (4.6%)
Krittanawong 20205	Retrospective	Not specified	375	20 (5)	Not specified	-	-	N	52.2 ± 12.8 (overall)	241 (64)	-	-	-	-
Daoulah 202016	Retrospective	Sustained VT/VF	83	10 (12)	10 (100)	18.8 (IQR 9.0–40.1) (median)	-	Y	46 (26–81)	42 (51)	1 ICD	1 death confirmed 10 months post-discharge, unknown cause	Type 1: 52% Type 2: 42% Type 3: 4%	Overall: 60% PCI: 53% CABG: 7%
Chen 20209	Retrospective	VT/VF	349	20 (6)	17 (85)	69.6 ± 54 (mean)	-	Not specified	47 ± 12	19 (95) for CA group	5 ICDs 4 WCDs	0 shocks	-	PCI: 7 (2%)
Saw 201715	Prospective	VT/VF	327	29 (9)	29 (100)	37.2 (IQR 17.9–65.9) (median)	3 (IQR 2–5) (median)	Y	52.5 ± 9.6 (overall)	297 (91)	9 “cardioversion or ICD”	-	Type 1: 26% Type 2: 70% Type 3: 5%	Overall: 61 (18.7%) PCI: 16.5% CABG: 2.1%
Sharma 201719	Retrospective	Not specified	102	14 (14)	7 (50) OOH 4 (29) in ED 3 (21) during/shortly after PCI	42 (mean)	-	Not specified	44 ± 11	88 (86)	7 ICDs 4 WCDs	1 shock for SVT	Not specified	PCI: 37 (36%) CABG: 3 (3%)

CA = Cardiac arrest; CABG = coronary artery bypass graft; ED = emergency department; IQR = interquartile range; NRD = National Readmissions Database; OOH = out of hospital; PCI = percutaneous coronary intervention; PEA = pulseless electrical activity; SCAD = spontaneous coronary artery dissection; SVT = supraventricular tachycardia; VF = ventricular fibrillation; VT = ventricular tachycardia.

^∗Data cited are for SCAD patients with CA unless specified otherwise.

^†Saw J. Coronary angiogram classification of spontaneous coronary artery dissection. Catheter Cardiovasc Interv 2014;84:1115–1122.

Clinical outcomes

A single-arm meta-analysis was performed to assess the prevalence of study endpoints among patients with SCAD and concurrent CA. A pooled analysis revealed that in-hospital mortality, postdischarge mortality, recurrent MI, and recurrent SCAD occurred in 20%, 3%, 12%, and 9% of patients with SCAD and CA, respectively (Figure 2). Of a pooled 24 ICDs and 11 WCDs, there was only 1 appropriate ICD discharge reported for VT/VF and 1 inappropriate ICD discharge for supraventricular tachycardia (SVT) (Table 1).

Figure 2.

Single-arm meta-analyses for clinical outcomes among patients with SCAD and concurrent CA. CA = cardiac arrest; SCAD = spontaneous coronary artery dissection.

Outcomes among patients with SCAD and CA compared with patients with SCAD without CA

In-hospital overall mortality

Seven studies were included in the meta-analysis, and the pooled results showed that patients with SCAD and CA had higher in-hospital overall mortality compared with patients with SCAD without CA (27 of 295 vs 47 of 2900, RR = 6.75, 95% CI 4.50, 10.14, P < .00001, I² = 0%) (Figure 3A).

Figure 3.

Outcomes among patients with SCAD and CA compared with patients with SCAD and without CA. Abbreviations as in Figure 2.

Postdischarge overall mortality

Four studies were included in the meta-analysis. Follow-up ranged from 18.8 to 39.1 months for studies reporting median follow-up and 42.0 to 69.6 months for studies reporting mean follow-up. A Nationwide Readmissions Database (NRD)-based study by Tan et al¹² assessed 1-month postdischarge outcomes. The pooled results showed that patients with SCAD and CA had higher postdischarge overall mortality compared with patients with SCAD without CA (3 of 127 vs 7 of 1488, RR = 5.86, 95% CI 1.72, 19.91, P = .005, I² = 0%) (Figure 3B).

MACE

Four studies were included in the meta-analysis and the pooled results showed that patients with SCAD and CA had higher risk of MACE compared with patients with SCAD without CA (29 of 121 vs 180 of 1463, RR = 1.91, 95% CI 1.11, 3.27, P = .02, I² = 77%). High heterogeneity was noted, potentially attributable to varying definitions of MACE across studies (Figure 3C).

Acute heart failure

Three studies were included in the meta-analysis, and the pooled results showed that patients with SCAD and CA had higher risk of acute heart failure compared with patients with SCAD without CA (34 of 212 vs 49 of 1804, RR = 4.82, 95% CI 3.22, 7.20, P < .00001) (I² = 42%) (Figure 3D).

Cardiogenic shock

Two studies were included, and the pooled results showed that patients with SCAD and CA had higher risk of cardiogenic shock compared with patients with SCAD without CA (37 of 128 vs 39 of 832, RR = 6.11, 95% CI 4.07, 9.19, P < .00001) (I² = 64%) (Figure 3E).

Recurrent MI

Three studies were included, and the pooled results showed that patients with SCAD and CA had higher risk of recurrent MI compared with patients with SCAD without CA (16 of 114 vs 134 of 1400, RR = 3.31, 95% CI 2.03, 5.39, P < .00001), I² = 56%) (Figure 3F).

Recurrent SCAD

Recurrent SCAD primarily included recurrent de novo SCAD and not SCAD extension. Five studies were included, and the pooled results showed that patients with SCAD and CA had higher risk of recurrent SCAD compared with patients with SCAD without CA (14 of 135 vs 84 of 1551, RR = 1.9, 95% CI 1.1–3.3, P = .02) (I² = 43%) (Figure 3G).

LVEF ≤40%

Three studies were included, and the pooled results showed that patients with SCAD and CA had higher risk of LVEF ≤40% compared with patients with SCAD without CA (20 of 105 vs 52 of 1242, RR = 5.26, 95% CI 2.78, 9.94, P < .00001) (I² = 0%) (Figure 3H).

Discussion

We present, to the best of our knowledge, the largest analysis on the clinical outcomes in SCAD with concurrent cardiac arrest. Patients with SCAD and CA were at a higher risk of in-hospital and postdischarge mortality, acute heart failure, cardiogenic shock, recurrent MI, and recurrent SCAD. Importantly, our results are mostly applicable to patients with SCAD and CA who survive at least up to hospital arrival to undergo medical evaluation and additional testing. Those who do not survive to hospital presentation may have unique clinical characteristics and risk factors that are likely largely understudied.

Although SCAD is generally associated with good long-term outcomes with low overall mortality and incidence of MACE,⁴ our pooled double-arm meta-analysis highlights that patients who develop concurrent CA should be considered high-risk for both acute and long-term complications. Patients with SCAD and CA are more likely to present with STEMI, which is commonly associated with larger infarct size and increased risk of LV dysfunction and adverse cardiovascular outcomes.⁹ Conservative management at index presentation is typically pursued, although Waterbury et al^20,21 highlighted that 17.5% of conservatively managed patients with SCAD had early progression of SCAD based on angiographic evaluation: Specifically, patients with intramural hematoma at baseline were at higher risk of progression and early deterioration.

Studies on the community-based Canadian SCAD cohort,⁴ the Nationwide Readmissions Database,¹² the Mayo Clinic SCAD Virtual Multi-Center registry,²² and a multinational study covering 4 Gulf countries (KSA, UAE, Kuwait, and Bahrain)¹⁶ have recorded rates of CA between ∼3% and 13% at index hospitalization and ∼4% of patients during postdischarge follow-up, corresponding with our pooled estimates. The reported prevalence of CA in patients with SCAD is more variable, which could be explained by a multitude of reasons, including baseline patient features and severity of SCAD, and time frame over which CA was assessed (eg, immediate postevent, in-hospital, and long-term follow-up).

Consensus statements by the European Society of Cardiology (ESC) and American Heart Association (AHA) on SCAD recommend medical management when possible for patients with SCAD who are hemodynamically stable and do not show signs of ongoing myocardial ischemia.^2,23 Long-term management of SCAD involves maintenance on aspirin and beta blockers, the latter associated with reduced incidence of recurrence of SCAD in 1 observational study.¹⁵ This guidance is based on findings that the dissection will most likely heal spontaneously, and recurrence typically does not involve the initial culprit coronary artery nor does PCI prevent future recurrence of SCAD.^21,24 In our meta-analysis, only 3 studies had data on beta-blocker use, although specific data on patients with SCAD and concurrent CA was lacking, and use of antiarrhythmic medications was unavailable.^11,15,17 In a prospective multicenter study by Saw et al⁴ of 750 patients with SCAD of whom 8.3% had VT/VF, aspirin and beta-blocker use was high with low long-term mortality. Tweet et al¹ reported an uncomplicated hospital course among 31 patients with SCAD managed conservatively, with 2 deaths occurring on long-term follow-up. In contrast, the failure rate of percutaneous coronary intervention (PCI) was 53%, and it did not significantly reduce rates of revascularization and recurrence of SCAD.²⁴ One population very much underrepresented in the included studies is obstetric patients; pregnancy and the peripartum period have been posited as major risk factors for SCAD in multiple studies with a more severe presentation.²⁵ Among 13 deaths attributable to pregnancy-associated SCAD (P-SCAD) in the Mothers and Babies: Reducing Risk Through Audits and Confidential Enquiries Across the UK (MBRRACE-UK) audit, 12 suffered from out-of-hospital CAs.²⁶ Phan et al¹⁴ found that patients with SCAD and CA were more likely to have P-SCAD. However, Krittanawong et al,²⁷ in their analysis based on NRD data, noted a similar incidence of CA between P-SCAD and non–P-SCAD. Therefore, further research is needed to clarify possible risk-enhancing factors for CA and CA-related outcomes in this patient group.

Those who receive PCI show no statistical difference in long-term outcomes compared with those who received no PCI, and the former may expose patients to greater risks of iatrogenic complications such as propagation of intramural hematoma and dissection and abrupt vessel occlusion.^4,24,28,29 Nearly one-half of the VT/VF arrests reported in their cohort by Phan et al¹⁴ occurred during PCI. A meta-analysis by Bocchino et al³⁰ also showed that the mean success rate of PCI was less than 50% and that revascularization therapies did not decrease rates of all-cause mortality, cardiovascular mortality, MI, heart failure, and recurrence of SCAD.³⁰ In light of uncertain benefits and increased risk of bleeding complications associated with antiplatelet therapy in patients with SCAD who do not undergo PCI, experts recommend individualized decision making considering the risk–benefit ratio.² Indications for the use of more advanced imaging modalities such as intravascular ultrasound (IVUS) and optical coherence tomography (OCT) are also not well defined but may optimize the use of PCI in high-risk patients, although their uptake in studies has been low.^30,31

Notably, based on our pooled sample which included placement of 24 ICDs and 11 WCDs, there was only 1 appropriate ICD discharge reported for VT/VF by Cheung et al¹³ over a mean follow-up of 4.8 ± 3.3 years, and 1 inappropriate discharge for SVT.¹⁹ In a community-based study of 27 patients with VA caused by SCAD, Chen et al⁹ demonstrated that no patient developed recurrent VA caused by SCAD, and none required defibrillation therapies over a 5.8-year mean follow-up. Over a mean follow up period of 3.5 years following ICD placement among 7 patients with SCAD and SCD as part of the Massachusetts General Hospital (MGH)-SCAD registry, there was 1 inappropriate ICD discharge for supraventricular tachycardia.¹⁹ The unclear risk–benefit ratio of ICD placement for secondary SCD prevention in SCAD was highlighted by Garg et al³² in their meta-analysis, noting 1.2% (95% CI 0%–15.8%, I² = 0%) and 1% (95% CI 0%–15.3%, I² = 0%) patients received appropriate and inappropriate ICD therapies, respectively, during the follow-up period. Given these findings and considering that most SCAD lesions heal spontaneously on follow-up, current guidelines do not support early ICD placement for patients after an episode of cardiac arrest with a reversible cause.^33,34 Our results support the idea of SCAD-mediated CA being a transient substrate with a low recurrent event rate, as long as the LVEF recovers. Factors such as incomplete coronary revascularization, persistently reduced LVEF, and recurrent VA can also be used to risk stratify SCD and guide decision making regarding ICD placement.³⁵ It would be of value for future studies to stratify patient outcomes data based on timing of CA such as out-of-hospital, in-hospital and procedural.

We also found that the incidence of recurrent de novo SCAD was higher in patients with SCAD and CA. Management of modifiable factors such as smoking cessation, avoidance of pregnancy, and adherence to long-term medical therapy with beta blockers may decrease the risk of recurrent SCAD. The development of VA in patients with SCAD can be related to multiple factors, with only 1 of them being recurrent SCAD. For instance, myocardial fibrosis after SCAD-induced MI may also increase the long-term risk of VAs. Hence, future studies evaluating the association between myocardial fibrosis as assessed by imaging techniques such as cardiac magnetic resonance (CMR) imaging and recurrent VAs may reveal new strategies for identifying candidates that may benefit from ICD placement.

Indeed, it is well recognized that solely relying on LVEF is inadequate in stratifying future risk of patients with SCAD of SCD; the cutoff of LVEF <35% that guides ICD placement may correctly predict only ∼20% of SCD events in other cardiomyopathies.³⁶ CMR—particularly late gadolinium enhancement (LGE) to detect myocardial fibrosis—provides additional prognostic information about risk of SCD over LVEF solely.37, 38, 39 Replacement of viable myocardial tissue by fibrosis in SCAD is likely to significantly enhance risk of post-SCAD arrhythmias. Hence, characterizing such an arrhythmogenic substrate by CMR-LGE may enhance SCD risk stratification and further guide appropriateness of ICD placement. In this context, 1 study of 14 patients with acute or recurrent SCAD demonstrated LGE in the distribution of the culprit coronary artery in all cases.⁴⁰ Patient populations that would benefit most from CMR-based characterization of scar tissue require further defining. For instance, Androulakis et al⁴¹ demonstrated that SCAD involving the RCA was predictive of myocardial fibrosis detected by LGE (odds ratio [OR] 5.2; P = .034). In addition, most patients with SCAD may not present with large infarct sizes and subsequent myocardial fibrosis; however, those with pregnancy-related SCAD, multivessel SCAD, STEMI, CA, extracoronary arteriopathies (eg, fibromuscular dysplasia, focal stenoses, and arterial tortuosity) are more likely to present with larger infarct sizes and may therefore benefit from further anatomic and functional characterization by CMR and LGE.^41,42

Limitations

A considerable limitation was inherent to available data on the topic in the included studies: lack of appropriate stratification of data on patients with SCAD: specifically, with CA to allow for analysis of relevant clinical and demographic subgroups. All included studies were observational in nature, with results affected by the completeness of collected and analyzed data. Tan et al¹² was a database study using the National Readmissions Database and used ICD coding for coronary artery dissection due to lack of specific ICD-10 codes for SCAD, possibly contributing to heterogeneity of studied patients caused by potential inclusion of iatrogenic coronary artery dissection and atherosclerotic coronary artery dissection. Similarly, Krittanawong et al⁵ was an ICD-9 or -10 codes-based study without angiogram confirmation, which may contribute to heterogeneity. These studies had a higher proportion of men, suggesting that some of the patients may have had dissection in the setting of atherosclerosis. As our study was a systematic review and meta-analyses of published observational studies, some results may be hindered by publication biases. Our ability to perform sensitivity or subgroup analyses was limited because of lack of stratified data available on cardiac arrest across studies. Heterogeneity was noted among several variables, which can be attributed to differences in study populations, definitions of endpoints, and other clinical or procedural variables. The inclusion of various forms of SCD including PEA and asystole could introduce heterogeneity in outcomes studied. Similarly, these cohorts only included patients who survived long enough to be diagnosed with SCAD and did not include autopsy data. SCAD is a particularly important clinical entity among obstetric patients, although our review noted data on this patient group to be lacking, and so was literature on sex-based differences in outcomes. Further research is needed to explore outcomes in women and obstetric patients with SCAD and CA. Because of limitations in the reported literature, our study did not focus on sex-based differences or pregnancy associated SCAD. Although our study is, to the best of our knowledge, the largest to study clinical outcomes in patients with SCAD and CA, it is important to acknowledge that these findings are hypothesis generating and may not be powered sufficiently to guide clinical care in a diverse patient population with varying characteristics and risk factors.

Conclusion

Our study highlights that concurrent CA in SCAD is associated with worse in-hospital and postdischarge mortality, and higher risk of acute heart failure, recurrent MI, and recurrent SCAD. Pooled follow-up ICD and WCD data noted a low recurrent CA event rate. Our findings support SCAD-mediated CA to be a transient substrate, supporting a conservative approach to prevention of secondary CA in SCAD, consistent with the current recommendation for atherosclerotic coronary artery disease/MI-mediated CA. Further research is needed on various aspects of management relevant to this high-risk patient population, including patient selection for ICD placement and the role of WCDs in managing early recurrent SCD.

Disclosures

Dr Marysia Tweet is supported by a grant from the National Heart, Lung, and Blood Institute (NHLBI), NIH K23H155506. The Mayo Clinic “Virtual” Multicenter SCAD Registry receives support from SCAD Research, Inc. All other authors have no conflicts of interest to disclose.