Post-myocardial infarction pericarditis occurs within several days following ST-elevation myocardial infarction (STEMI) (1–4). Its reported incidence has decreased dramatically from 20% to <5% since the early reperfusion era (1–6). Most of the reports addressing its incidence were based on the combination of clinical, electrocardiographic, and echocardiographic data (1–6).

The current practice of CMR performed early in the post-STEMI period suggested that pericardial involvement is more frequent than previously reported. This led us to conduct a dedicated, early post-STEMI CMR study to document the actual incidence and clinical significance of pericardial involvement. Cardiac MRI using the late Gadolinium enhancement (LGE) sequence is the modality of choice for assessing myocardial and pericardial pathologies (7, 8). The reported sensitivity and specificity of CMR for detecting pericardial involvement is 94–100% (9). Thus, CMR is superior to echocardiography in assessing and characterizing both myocardium and pericardium (9).

The current study documents the incidence, location, and extent of pericardial involvement using CMR in consecutive STEMI patients (PISTEMI). Clinical follow-up was conducted as well in order to evaluate the long-term clinical significance of PISTEMI.

Involvement of the pericardium (LPE+) overlying the ventricles and\or the atria was demonstrated in 145/187 patients (77.5 %). Among these 145 patients, spatial pericardial involvement was demonstrated as follows: LPE overlying the LV (120/187, 64.2%), RV (60/187, 32.1%), RA (52/187, 27.8%) and LA (11/187, 5.9%) (Figure 3).

The study cohort was divided accordingly into LPE+ and LPE– subgroups. Baseline characteristics, laboratory and echocardiographic findings are presented in Table 1. No significant differences in demographics, baseline characteristics, co-morbidities, MI location, or ECG features were found between the two groups (Table 1, Supplementary Table 1).

LPE+ patients had higher peak CPK levels (1,700 [IQR: 900–3,133] vs. 1,067 [IQR: 482–2,709], p = 0.03) and a lower LVEF on the first echocardiogram performed following primary PCI (44.6% ± 9.6 vs. 48.1% ± 10.3, p = 0.03).

Clinical signs of pericarditis were documented and included: pericardial chest pain in 5/187 patients (2.6%), typical ECG changes in 4/187 patients (2.1%), and pericardial effusion at trans-thoracic echocardiography in 6/187 patients (3.2%). Thus, a total of 13/187 patients demonstrated at least two clinical signs consistent with the diagnosis of pericarditis. Out of these patients, two patients only were treated with Colchicine (Table 2).

A robust association was found between the LGE score, expressing the degree and extent of the myocardial damage and the presence of LPE (P < 0.01) (Table 2; Figures 4, 5). Furthermore, a strong association was found between LPE overlying the left ventricle (LV LPE) and the severity of the myocardial damage expressed as LGE and MVO scores (p = 0.031 and p = 0.002, respectively).

No statistically significant correlation was found between the location of the myocardial damage and LPE presence. The prevalence of pericardial effusion detected by CMR was similar in the presence and absence of LPE (47.6 vs. 51.2%, p = 0.81) (Table 2). Moreover, the presence of pericardial effusion was not correlated with the LPE score (3 [IQR: 1–5] vs. 3 [IQR: 2–6], p = 0.55).

A binomial multivariable logistic regression model was performed to ascertain the effects of age, gender, LGE score, and MVO score on the likelihood of LPE+. Late Gadolinium enhancement score was found to be a robust predictor for pericardial involvement (p = 0.002). The logistic regression model was statistically significant [χ52 = 15.8, p = 0.007], and the area under the ROC curve was 0.74 (95% CI, 0.64–0.81) (Table 3). No significant difference in complications during hospitalization or length of stay was demonstrated between the LPE+ and LPE– groups (Table 1).

Post-discharge clinical follow-up was available in 85% of patients (159/187) for a median period of 3 years. Overall, 46/159 patients (28.9%) sustained at least one MACE, including recurrent MI (9/159, 5.6%), acute stroke (3/159, 1.8%), acute coronary syndrome necessitating urgent hospitalization and\or percutaneous coronary intervention (15/159, 9.4%), hospitalization for heart failure (18/159, 11.3%), and cardiovascular death (1/159, 0.6%). The presence and extent of LGE, MVO, and LPE in the follow-up patients were not different from those lost to follow-up. A detailed comparison of MACE between subgroups of the cohort can be found in the Supplementary Material (Supplementary Table 2).

Patients with a higher LGE score and MVO % of LV mass were more likely to sustain MACE during the follow-up period (43 vs. 22%, p = 0.006 and 40 vs. 23%, p = 0.024, respectively). Patients with LPE+ involving more than one pericardial segment demonstrated a lower MACE incidence than patients without pericardial enhancement (LPE–) (23 vs. 38%, p = 0.042). Pericardial involvement adjacent to the LV (LV LPE) was also associated with a significantly lower MACE rate (22 vs. 41%, p = 0.009). A univariate log-rank survival analysis revealed a statistically significant difference in survival distributions between LV LPE (+) and LV LPE (–) patients (p = 0.04). However, significance was not maintained when comparing whole myocardium LPE (p = 0.12).

A multivariable Cox proportional hazard model adjusted for age, gender, LGE score, and MVO% demonstrated that LPE+, involving more than one pericardial segment, was associated with a significantly lower MACE rate [HR 0.46, 95% CI (0.25–0.86), p = 0.02] (Table 4; Figure 6). Pericardial involvement adjacent to the LV (LV LPE) was also associated with a significantly lower MACE rate [HR 0.39, 95% CI (0.21–0.7), p = 0.002] (Figure 7).

The integrated approach for diagnosing post-MI pericarditis (PMIP), which combines the CMR findings of LPE together with either elevated CRP levels and/or the presence of pericardial effusion, demonstrates that 78 patients (41.7%) had PMIP. No significant differences in demographics, baseline characteristics and co-morbidities, were found between the two groups (Table 5). PMIP+ patients have significantly higher levels of inflammation and myocardial injury biomarkers (leucocyte count, CRP, CPK, and Troponin) (Table 5).

A multivariable Cox proportional hazard model adjusted for age, gender, LGE score, and MVO% demonstrated that PMIP, as suggested in the integrated approach, was also associated with a significantly lower MACE rate [HR 0.46, 95% CI (0.23–0.91), p = 0.02] (Table 6).

The present CMR study demonstrated an unexpectedly high incidence of PISTEMI. Late pericardial enhancement (LPE+), the MRI manifestation of pericardial involvement, was documented in 77.5% of the patients on early post-STEMI. Cardiac MRI LPE was documented adjacent to the injured and non-injured LV myocardium, the RV, and both atria.

Post-myocardial infarction pericarditis is traditionally classified as early and late (Dressler syndrome). Its reported incidence has markedly declined in the reperfusion era from 20% to <5% (1–6). Dressler syndrome almost disappeared in the reperfusion era occurring in <0.1% of patients (12). As per current clinical practice, pericarditis is diagnosed by the presence of chest pain, pericardial friction rub, typical ECG changes, and echocardiographic evidence of pericardial effusion (2, 3, 6). Cardiac MRI is considered the modality of choice for myocardial and pericardial characterization, offering superior imaging compared with echocardiography (7–9). However, routine pericardial imaging using CMR is not performed and is recommended only in equivocal cases (13). Pericardial enhancement on the CMR LGE series is considered as evidence of active pericardial inflammation, with reported sensitivity ranging from 94 to 100% (7–9, 14, 15).

In the current study, clinical evidence of pericarditis was present in only 13/187 patients (6.9%), while PISTEMI CMR evidence was documented in 145/187 patients (77.5%).

In addition, we documented a strong correlation between PISTEMI and infarct severity. A similar correlation was reported both in the pre and post-reperfusion era. However, the reported incidence of post-STEMI pericarditis was much lower in comparison (4). Lador et al. reported post-STEMI pericarditis as relatively rare, with larger infarct size and worse short-term outcomes (5). The currently accepted concept is that post-STEMI pericarditis has become infrequent since early reperfusion (4, 5). Our CMR-based data documented early pericardial involvement in over 3/4 of our STEMI cohort. The utilization of CMR as the imaging modality of choice in our study, instead of the classical approach to post STEMI pericarditis diagnosis, explains this significant difference in prevalence. All patients in our cohort underwent early CMR, which, as stated above, is highly sensitive in detecting pericardial involvement, while the reports mentioned above are based on the combination of clinical and echocardiographic data only (2–6).

A single group of investigators reported the utilization of CMR in pericardial involvement diagnosis in two letters to the editor (16, 17). In accordance with our findings, these authors also found an increased incidence of pericardial involvement in STEMI patients, correlating with infarct size and transmurality (16). These reports defined pericardial injury as the presence of pericardial effusion and/or pericardial enhancement on CMR. The reported incidence of pericardial injury was 48%, while LPE was present in only 30% of their patients. Our study, which defined PISTEMI by the presence of pericardial enhancement, demonstrated a higher pericardial enhancement incidence of 77.5%. This discrepancy is related to the fact that the current study assessed the pericardium globally, addressing the pericardium adjacent to all four cardiac chambers. In contrast, these previous reports addressed only the pericardium adjacent to the left ventricle. In addition, differences in LGE acquisition parameters, including in-plane resolution (which were not specified in the aforementioned reports), could also contribute to this discrepancy (16, 17). The present study documented a meaningful correlation between LPE presence and infarct extent and severity. This correlation was also documented in the previous letters to the editor (16, 17). As mentioned, we have documented pericardial involvement adjacent to all four cardiac chambers, with no distinct correlation to the segmental location of the injured myocardial segments. To the best of our knowledge, this is the first report describing a global pericardial response early after STEMI. These findings might suggest that the pericardial involvement is not just an indicator of infarct severity but a more complex process related to other factors.

Post-STEMI pericarditis in the pre-reperfusion era was associated with increased long-term mortality, although it was not found to be an independent prognostic factor in follow-up studies (1–4, 6). Our study is the first CMR-based series reporting follow-up data concerning pericardial involvement in STEMI patients. As expected, patients with increased LGE and MVO sustained more MACE events. In contrast, patients with PISTEMI, defined as LPE+, had a better long-term prognosis with significantly lower MACE incidence at a median follow-up of 3 years. In addition, patients with CMR evidence of LV LPE had significantly better outcomes than those without LV LPE. While the current hypothesis-generating manuscript is not designed to provide a pathophysiological explanation to the reported findings, we assume that regenerative/reparative processes involving the left ventricle are more likely to be correlated to prognosis than the other myocardial chambers. This unexpected result raises the question of whether pericardial enhancement should still be considered a complication of acute MI.

Moreover, our data introduces a novel idea suggesting that PISTEMI results from a physiologic rather than a pathologic mechanism triggered by the myocardial injury. Experimental animal studies focusing on myocardial damage in Zebrafish and low mammalians have recently suggested that cardiac injury may activate the epicardium as a source of a regenerative or/and reparative process (18–21). A similar mechanism was not demonstrated in higher mammalians. In addition, a hint that a reparative myocardial mechanism still exists in the human heart may be found in two recent papers in pediatric patients, reporting better outcomes following cardiac surgery and viral myocarditis in children with increased circulating micro- RNA (22, 23). The presence of elevated CRP in the LPE+ group (and LV LPE+) in our study might additionally suggest that pericardial enhancement is in part a result of an immune reaction. Recently, a possible benefit from anti-inflammatory drug administration (Colchicine) in acute MI was reported (24). Close attention to the endpoint events in this study emphasizes that the anti-inflammatory treatment lowered vascular (strokes and urgent hospitalization for angina) rather than myocardial complications (24). Indeed, previous studies have warned against the utilization of anti-inflammatory drugs (Methylprednisolone and Indomethacin) in acute MI. Both drugs were associated with scar thinning, infarct expansion, and ventricular functional impairment (25–27).

Thus, our observation of frequent PISTEMI, together with those mentioned above in experimental and clinical studies, raises the challenging hypothesis that pericardial involvement is a part of a yet undefined residual myocardial repair process in the human heart. The current study did not measure specific biochemical factors and therefore cannot support this assumption with a defined physiological mechanism.

The major clinical implication of our findings is the clear understanding that pericardial involvement detected by CMR is common, occurring in more than 3/4 of STEMI patients. Thus, it should not be regarded as a post STEMI complication.

Moreover, this study documented that pericardial involvement was associated with pericardial effusion detected by echocardiography in only 3% of the cases. This is of high clinical importance since it challenges the traditional criteria of pericardial involvement diagnosis, which rely mainly on echocardiographic evidence of pericardial effusion. Cautioning that CMR detection of pericardial involvement should not be “automatically” considered as a clinical complication.

This is a retrospective single-center study with the related inherent limitations. The study cohort included scans performed both on 1.5 and 3 T MR scanners. Still, there was no significant difference between scanner field strength and the documented LPE presence.

Although CMR offers high spatial resolution on LGE images, there is an inability to separate the epicardial and parietal layers due to their anatomic proximity. Since the thickness of pericardium is usually <4 mm and the LPE signal is subtle, LPE quantification (similar to LGE quantification) cannot be performed. Compared to acute inflammatory pericarditis, pericardial involvement is local rather than global, rendering T2 sequences non-contributing to pericardial involvement diagnosis.

Factors involved in myocardial repair described in the animal studies were not studied in the current study.

The current CMR study demonstrated that early pericardial involvement detected by CMR is common, occurring in more than 3/4 of STEMI cases, involving the pericardium adjacent to all four cardiac chambers. PISTEMI is correlated with the extent and severity of the myocardial injury. PISTEMI is associated with a decrease in the major adverse events rate on follow-up.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

The studies involving human participants were reviewed and approved by Sheba IRB-Helsinki Committee (International Review Board for human and animal trials) 5897-19-SMC. Written informed consent for participation was not required for this study in accordance with the national legislation and the institutional requirements.

EM and OG: conceptualization, data curation, project administration, analysis, supervision, visualization, validation, methodology, software, writing-original draft, review, and editing. YB and EK: conceptualization, supervision, CMR analysis, investigation, and validation. DO: methodology, software, writing-original draft, review, and editing. AF: methodology, software, analysis, supervision, and validation. SN, IM, AS, and RBein: conceptualization, investigation, supervision, and validation. RBeig: supervision, investigation, and validation. RG: CMR analysis, investigation, and validation. ES and SM: conceptualization, data curation, analysis, supervision, validation, writing-review, and editing. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.