• pericarditis
• constrictive
• echocardiography
• MRI
• inflammation

Despite a long history, our understanding of the aetiology and pathophysiology of pericardial diseases is still evolving. Pericardial diseases were first described in the 17th century with the description of a disease characterised by dyspnoea and intermittent pulse.1 Pericardial syndromes include pericarditis, pericardial effusion, effusive–constrictive pericarditis, pericardial tamponade and constrictive pericarditis (CP).2 In high-income countries, the most common causes of CP are idiopathic or viral, followed by postcardiac surgery; however, in low-income countries, tuberculous pericarditis is still the most common aetiology of CP.2 Only relatively recently, in 1987, transient constrictive pericarditis (TCP) was first described by Sagristà-Sauleda et al in patients with effusive acute idiopathic pericarditis.3 Subsequently, TCP was described in chemotherapy, collagen vascular disease and bacterial endocarditis, and the natural history of TCP was well elucidated in 2004 by Haley et al, who described a transient form of CP which resolves over time.4 Resolution of TCP without surgery offers a therapeutic opportunity but raises the question—which patients with CP are most likely to benefit from intensive medical treatment, which comes with its own set of potential complications? Recent studies have examined treatment response using laboratory and/or imaging indicators of inflammation.5 6 However, available evidence is insufficient to adopt specific parameters in order to predict the response to anti-inflammatory therapy, especially in patients with minimal or no symptoms (New York Heart Association class I or II).

CP is a diagnosis made on clinical and invasive haemodynamic criteria. However, non-invasive multimodality imaging is the first step in the investigation of pericardial diseases in general and CP in particular.2 Whether imaging is useful for guidance of treatment is a separate question. Using echocardiography to diagnose CP is detailed in the American Society of Echocardiography guidelines7 and cardiac MRI (CMR) is the recommended modality to assess the presence and extent of inflammation.7 CMR has been used to assess the response of patients with CP to anti-inflammatory therapy,5 6 but results were inconsistent and the correlation between delayed hyperenhancement (DHE) of the pericardium and the response to anti-inflammatory therapy was only modest.5 The role of echocardiography in this context has not been investigated; this might be due to the perceived limitations of echocardiography in tissue characterisation, especially with a low signal-to-noise ratio of the pericardium. Moreover, the feasibility of using speckle tracking echocardiography (STE) to quantify inflammation in patients with CP is yet to be investigated. The change in left ventricular (LV) speckle tracking strain parameters has been shown to exhibit specific reproducible patterns in CP,8 but again the association between these parameters as a predictor of the response to anti-inflammatory therapy in patients has not been investigated.

In this issue of Heart, Kimi et al investigated the response of patients with CP to anti-inflammatory therapy, with a focus on LV mechanics.9 The authors retrospectively identified patients with CP treated with anti-inflammatory therapy who had both baseline and follow-up laboratory and imaging data. Similar to previous studies, idiopathic/viral and postoperative aetiologies accounted for the vast majority of the study cohort, and nearly all were treated with intensive anti-inflammatory therapy. After a median follow-up of 323 days, they found that patients with TCP (defined by clinical resolution of symptoms) showed improvement in inflammatory markers, echocardiographic parameters, and CMR pericardial DHE. STE strain parameters improved in patients with TCP compared with patients with chronic CP. Using a logistic regression model, the authors found that only erythrocyte sedimentation rate (ESR) was associated with TCP. When a second time point at a median follow-up of 98 days was added, changes in tissue Doppler parameters (lateral/septal e’) and STE strain parameters (LSlateral/LSseptal) were observed in the patients with TCP, but not chronic CP. Interestingly, the ratio of lateral to septal e’, used to assess the specific finding of annulus reversus, improved in those with TCP, but this was due to a decrease in septal e’ rather than an increase in lateral e’. A similar pattern has been seen after pericardectomy (ie, stable lateral e’, with a decrease in septal e’),10 and may suggest that at baseline the lateral wall myocardial tethering leads to exaggerated compensatory early diastolic relaxation, which then resolves as the haemodynamics improve.

There are limitations to the study, many identified by the authors, such as its retrospective nature, the small numbers, lack of standardised treatments between the groups and even the definition of TCP being purely based on clinical symptoms. However, this represents a relatively large cohort of patients with CP who are traditionally difficult to recruit as they present infrequently with significant heterogeneity of presentation, and so, while not a perfect study, the authors are to be commended on their findings. By measuring segmental STE strain as well as traditional tissue Doppler parameters, they have integrated both systolic and diastolic functions into the assessment of CP.

How might these findings be translated to clinical practice? It is important to note that no imaging finding at baseline, either on echocardiography or on CMR, was able to accurately predict subsequent resolution of symptoms. Quantitative DHE was different between groups at baseline, but this is rarely performed in clinical practice, and moderate to severe DHE was not found to be a predictor of subsequent TCP. The only statistically significant predictor was ESR, which showed very modest discriminant ability, with a sensitivity of 78% and a specificity of 70%. Therefore, there remain large gaps in our ability to assess the prognosis (as opposed to the diagnosis) of CP. A corollary of this is that follow-up assessment, including imaging, is likely to be a necessity to guide treatment.

Interval imaging and biomarker assessment (figure 1) may identify those patients who have responded to treatment, but possibly more importantly, may identify those unlikely to gain benefit from ongoing medical therapy. Intensive medical treatment itself carries considerable risk, with prolonged high-dose steroid therapy likely to lead to significant side effects, and the addition of non-steroidal anti-inflammatories, as was the case in the majority of patients in this study, can be expected to substantially increase the risk of gastrointestinal bleeding. Therefore, identifying those unlikely to improve with ongoing therapy would have significant clinical benefits.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Potential role of interval imaging in transient pericarditis. The current clinical role of multimodality imaging in constrictive pericarditis is outlined in blue. more research is needed to support the potential role for multimodality imaging in predicting response to anti-inflammatory therapy as well as following up response to treatment, outlined in orange box. image contains material licensed under CC-BY 3.0 from https://smart.servier.com/. 2D, two-dimensional; CMR, cardiac MRI; DHE, delayed hyperenhancement; LV, left ventricular; NSAID, non-steroidal anti-inflammatory drug; STE, speckle tracking echocardiography; TDI, tissue Doppler imaging.

There remain a number of outstanding questions. The study by Kimi et al is a hypothesis-generating study, paving the way for future prospective studies looking at the feasibility of using interval STE deformation parameters to predict patients’ response to anti-inflammatory treatment. Another important question would be the benefit of interval CMR, and whether any information derived is incremental to that obtained via STE. Notably, previous studies looking at predicting TCP included patients with CP due to different aetiologies; the impact of a specific aetiology on the response to anti-inflammatory treatment should be investigated. Along the same lines, the feasibility of predicting TCP in each aetiological category might show significant differences.

The concept of TCP is relatively new and our understanding of the underlying pathophysiological mechanisms is evolving. Using anti-inflammatory therapy without specific predictors of which patients would benefit leaves many patients with an unfavourable risk to benefit ratio. With a lack of baseline predictors of response, interval imaging will be required, which, for most patients, will remain echocardiography due to practical and resource limitations. Future studies are required to develop better multimodality methods of predicting response to medical therapy in CP at baseline and to determine the optimal combination of follow-up duration and treatment response to subsequent imaging findings.