Search strategy and study selection

Electronic searches were performed using PubMed database, combining search terms (“pericarditis” OR “pericardial effusion” OR “cardiac tamponade” OR “pericardial disease”) AND (“malignant” OR “malignancy” OR “neoplastic” OR “cancer”) as both keywords and MESH terms. Filters were applied, selecting only articles published in the last 10 years. All publications were limited to those involving human subjects and written in English. Research was ended on 31 December 2018.

Abstract, case reports, conference presentations, editorials and expert opinions were excluded. Moreover, only studies reporting a minimum of 10 patients were eligible for inclusion. Review articles were excluded. A total of 186 records were identified through the electronic search. After analysis of the abstract and text, 104 were excluded because they were not pertinent to the present review. An additional 28 papers were included from analysis of text and references from originally selected publications. In total, 43 articles were included in the present review (figure 1).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Selection of studies.

Prevalence and aetiology

Primary pericardial malignancies, either benign (fibromas and lipomas) or malignant (mesothelioma, angiosarcomas, fibrosarcomas), rarely occur. In contrast, malignant pericardial effusions secondary to metastatic cancer are more common. A direct cancer involvement of the pericardium is seen in 7%–44% of pericardial effusions (table 1).1 4–8 13–15

The prevalence and extent of neoplastic pericardial involvement differ depending on the type of cancer (table 2).3 5–8 11–18

Lung cancer is the most common malignancy involving the pericardium, in both men and women. Primary cancer type is a predictor of long-term survival. The median survival was shorter for patients with lung carcinoma than for patients with breast cancer or haematological malignancy.5 10 16

Pathophysiology

In patients with cancer, pericardial involvement results from several different pathophysiological mechanisms: (1) direct extension or metastatic spread of the underlying disease, (2) complication of systemic tumour treatment (radiation therapy and/or chemotherapeutic toxicity), (3) opportunistic infection in the setting of antineoplastic therapies, and (4) metabolic derangements in the liver and/or renal and/or cardiac function from cancer or cancer therapy (less common) (figure 2).3 19 Malignant pericardial effusion can be explained either by direct local extension to the pericardium (ie, lung and breast cancer, oesophageal carcinoma) or by haematogenous and lymphangitic tumour spread (ie, haematological malignancies, melanoma, Kaposi sarcoma).5–15 Additionally, pericardial effusion may arise from mediastinal lymphatic drainage obstruction by tumour infiltration.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2

Pathophysiology of pericardial involvement in patients with cancer (see text for explanation).

The most common type of pericardial involvement is represented by moderate-to-large pericardial effusions.4 13 The prevalence of malignancy-associated effusion is about one-third in patients with pericardial tamponade, with this specific population having poor outcome, reflecting advanced stage of the disease.1 6 8

Several chemotherapy agents may lead to pericardial and/or myocardial disease, mediated by toxic metabolism and the generation of reactive oxygen species. These drugs include anthracyclines, cyclophosphamide, busulfan, cytarabine, tyrosine kinase inhibitors (dasatinib, imatinib mesylate, interferon-α), arsenic trioxide, all-trans retinoic acid, methotrexate, 5-fluorouracil and docetaxel. Direct cardiotoxicity of chemotherapeutic agents and the mechanisms of causing injury may vary by antineoplastic drugs (eg, increased capillary permeability with interleukin-2 therapy, pericardial and endomyocardial fibrosis with busulfan). Immune-mediated adverse effects of newer antineoplastic agents such as ‘immune checkpoint inhibitors’ are also emerging and should be considered (eg, increased risk of myocarditis, heart failure). Many of these patients may also be on anticoagulants, antiplatelets or non-steroidal anti-inflammatory drugs, which may lead to bleeding as cause of pericardial effusion.20–22

Radiation therapy is also responsible for cardiac injury. Acute radiation pericarditis typically occurs within days to few weeks after treatment, being one of the earliest complications of radiation therapy, and is often self-limiting. Mediastinal irradiation is even associated with chronic effusive-constrictive pericarditis, which may occur in up to 30% of patients who underwent radiation therapy, with a latency of 20 years.20 23 24 Nowadays radiation is a less common cause of constriction in developed countries due to improved radiation and shielding regimens.

On these basis, acute, late and very late consequences of irradiation can be seen in different clinical scenarios, ranging from few weeks to years between radiation therapy and symptom onset. In addition, antineoplastic agents and immunosuppressive therapy make oncological patients susceptible to autoimmune pathologies or viral, bacterial and fungal opportunistic infections, being potential cause of pericardial disease.3 19

Diagnostic work-up and management

Malignant pericardial effusions are usually moderate to large, compared with non-malignant ones (neoplastic pericardial involvement is responsible for one-fifth of all moderate-to-large pericardial effusions).4 25 Once pericardial effusion is detected, diagnostic work-up should focus on the potential cause and its haemodynamic impact. Cardiac tamponade, causing haemodynamic compromise, occurs in 50% of patients with malignant pericardial effusions, expression of advanced, terminal disease.8 14–16 Moderate-to-large pericardial effusions often had specific aetiologies in up to 90% of cases (eg, neoplasms, tuberculosis, systemic inflammatory diseases), and clinical presentation is important to suspect them. A suggested approach may consider the presence/absence of cardiac tamponade, any inflammatory signs including C reactive protein and/or erythrocyte sedimentation rate (ie, pericarditis), the presence of a known medical condition (in up to 60% of cases), and the size of the effusion.26

Physical examination may show signs of complicated pericardial involvement with cardiac tamponade and/or constriction. ECG usually presents non-specific findings that could be related to the presence of a large pericardial effusion (eg, low QRS voltages and electrical alternans, ie, a beat-to-beat variation of QRS amplitude in a swinging heart within a large effusion). Chest X-ray may show signs of the underlying cancer in case of pleuropulmonary involvement and increased cardiac silhouette for large pericardial effusions if >300 mL.27

Transthoracic echocardiography is the first-line imaging test in patients with suspected pericardial disease because it accurately detects pericardial effusion and its haemodynamic impact, as well as biventricular function. Its diagnostic accuracy is nearly 100%.27 28 Echocardiography has multiple roles: first evaluation, follow-up and monitoring of therapy. According to guidelines, the simplest effusion size measure is based on the largest end-diastolic echo-free space between the epicardium and parietal pericardium. The effusion is classified as small (<10 mm), moderate (10–20 mm) and large (>20 mm). Signs of cardiac tamponade include respiratory variation of mitral E velocities >25% and tricuspid E velocities >40%, and inferior vena cava plethora. In addition, hepatic vein Doppler may help to assess cardiac tamponade. Normal hepatic vein flow is biphasic, with systolic greater than diastolic velocity, with reduced forward flow or small reversal during atrial contraction. Both flow velocities increase during inspiration. In case of cardiac tamponade, forward flow velocities decrease to 20–40 cm/s. Such findings have high positive and negative predictive values (82% and 88%, respectively); however, they cannot be recorded in more than one-third of cases.27 28 In addition, echocardiography allows guidance and monitoring of pericardiocentesis, identifying the best site for pericardial puncture according to the shortest distance to the largest collection of pericardial fluid, the trajectory for needle entry, as well as the needle and catheter positions.5 Chest CT and cardiac magnetic resonance have a complementary role to characterise pericardial effusions, especially in the presence of haemorrhagic effusion, pericardial thickening or in the suspicion of effusive-constrictive pericarditis.2 27

Although pericardial involvement is common in patients with known active malignancy, large pericardial effusion can also be the initial presentation of an occult malignancy.4 7 25 Thus, malignancy should be excluded in cases of acute pericardial diseases presenting as cardiac tamponade but also moderate-to-large effusions of unexplained origin, without inflammatory signs and not responding to conventional empirical anti-inflammatory therapy.1 Some cases with pericardial effusions or apparent pericarditis may be related to cancer involvement. According to a Danish nationwide cohort study, among 13 759 patients with acute pericarditis, 1550 were subsequently diagnosed with cancer during follow-up. The overall cancer standardised incidence ratio was 1.5 (95% CI 1.4 to 1.5), driven predominantly by increased rates of lung, kidney and bladder cancer, lymphoma, leukaemia, and unspecified metastatic cancer. The less than 3-month cancer risk among patients with pericarditis was 2.7%, and the standardised incidence ratio was 12.4 (95% CI 11.2 to 13.7).29

In another population-based matched cohort study conducted using primary care data from the UK Clinical Practice Research Datalink, pericarditis was associated with an elevated subsequent risk of cancer (HR 3.03, 95% CI 2.74 to 3.36), and this association was particularly evident within 3 months from the diagnosis of pericarditis.30 It is not clear if the finding is related to a misdiagnosis of pericarditis in the presence of pericardial effusion (in clinical practice, pericardial effusions are considered sufficient for the diagnosis of pericarditis without additional European Society of Cardiology (ESC) criteria)27 or true pericarditis; however, it seems reasonable to warrant targeted investigations for cancer in patients presenting with pericarditis/pericardial effusion and a combination of older age, obesity and a need for hospitalisation.30

According to the most recent ESC guidelines, pericardiocentesis (if technically feasible) is an absolute indication for pericardial effusion when neoplastic involvement of the pericardium is suspected.27 A definite diagnosis requires cytological fluid analysis and cytopathological examination in patients undergoing pericardial drainage. Cytological evaluation is the gold standard for diagnosis of neoplastic pericardial effusion with a sensitivity of 71%–92.1% and specificity of nearly 100%, as reported in different studies. The sensitivity and specificity of cytology are markedly superior to the sensitivity (55%–64%) and specificity (85%) reported for pericardial biopsy. Conversely the rate of false negative results for cytology (7.9%–14.7%) was lower than pericardial biopsy (40%–44.7%).7 14 15 A possible explanation may relate to the process leading to the genesis of pericardial fluid: cancer infiltration causes obstruction of lymphatic drainage, culminating in pericardial effusions. Therefore, it is more likely for cytology to sample neoplastic cells suspended in an effusion than it is for a biopsy, especially when there is only focal pericardial involvement.14

In a recent study, Rooper and colleagues7 demonstrated that the volume of fluid submitted to cytology has a profound influence on the diagnosis of malignancy. In particular, the sensitivity of effusion cytology for fluids of 60 mL or less was 70.0%, slightly worse than that of pericardial biopsy, while the sensitivity with >60 mL of pericardial fluid reached 91.7%. On this basis, it is recommended to collect the largest possible amount of pericardial fluid for cytological analysis whenever a cytological diagnosis is sought.

After drainage of the pericardial fluid, samples of fluid are sent both to microbiology and pathology laboratories for culture, chemical tests and neoplastic markers testing. The remaining fluid should be sent immediately to the pathology laboratory for centrifugation and cytological diagnosis, or refrigerated at 4°C.31 Laboratory fluid analyses should include basic biochemical parameters (specific gravity, protein content, glucose, urea, lactate dehydrogenase, cholesterol), cytology, bacterial cultures and PCR analysis for identification of infectious agents according to clinical suspicion and availability of specific treatments to offer after the diagnosis. Immunohistochemical markers should be performed primarily to confirm the presence of malignant cells and to distinguish them from the reactive mesothelial cells. Common tested markers include those for adenocarcinoma (MOC-31, BerEP4, B72.3), mesothelial markers (calretinin, cytokeratin 5/6, Wilms tumour-1), specific markers such as thyroid transcription factor 1, cancer antigen 19-9, oestrogen and progesterone receptors, BRST-2, p63, neuron-specific enolase, and carbohydrate antigens CA 125 and CA 15-3. Carcinoembryonic antigen and serum cytokeratin 19 fragments (CYFRA 21-1) are useful for the diagnosis of lung carcinoma.15 32 33 There is increasing interest in the potential diagnostic utility of such immunohistochemical markers, although their real diagnostic utility is not well established.

In conclusion, the use of both modalities, cytology and pericardial biopsy, may provide the highest diagnostic yield. The submission of a pericardial biopsy alongside effusion cytology increases the overall sensitivity for detecting malignancy by approximately 8% and may be especially useful in the setting of low-volume pericardial effusion.14

Therapeutic scenarios

Possible treatment scenarios vary from pericardiocentesis and balloon pericardiotomy to cardiothoracic surgery, with the primary aim of relieving symptoms and improving quality of life, then preventing recurrent effusion for a long-term symptomatic benefit.

Furthermore, treatment of the local neoplastic disease is advisable with the aim of prolonging survival, on the basis of the patient’s clinical status and prognosis. In the presence of malignancy, optimal treatment of neoplastic pericardial effusion should balance treatment efficacy with life expectancy.34

Prompt treatment may decrease short-term risk of death in patients with cardiac tamponade, but many of these patients have major comorbidities associated with widespread metastatic disease, limiting both quality of life and survival even with optimal management of pericardial disease. On the other hand, some cases of pericardial effusion might be unrelated with the underlying malignancy and long-term survival can be expected.34 Usually pericardial effusion due to cancer involvement is moderate to large and malignant pericardial effusion is less prevalent in small pericardial effusions. On this basis, management of neoplastic pericardial disease is complex and requires a multidisciplinary approach.

In emergency cases with cardiac tamponade or significant effusion, immediate relief of symptoms may be obtained with percutaneous drainage or with a surgical approach.27

Systemic radiotherapy or chemotherapy should be reserved to radiosensitive and chemosensitive tumours (such as lymphomas and breast cancer). For the long-term prevention of recurrences, various approaches have been proposed according to the 2015 ESC guidelines,27 as well as other recommendations: pericardiocentesis with extended drainage5 6 12 34–36; pleuropericardial or pleuroperitoneal window either by percutaneous balloon pericardiotomy or by surgery3 26 34–40; intrapericardial instillation of cytostatic or sclerosing agents; and local and/or systemic chemotherapy and radiation therapy.9 41

In malignant pericardial effusions, the combination of systemic chemotherapy plus pericardial window is probably the most effective treatment option for the prevention of recurrences.3

Outcomes and prognosis

The development of pericardial effusion in oncological patients is a complication that significantly worsens prognosis and can be ultimately fatal.39 Once neoplastic involvement of the pericardium becomes clinically apparent, morbidity and mortality rates are very high either in patients with known malignancies or in patients with new diagnoses of cancer, as it represents an advanced disease.1–8 The mean survival time of these patients rarely exceeds 12 months.10 16 25 39 Malignancy is an independent predictor of recurrent pericardial tamponade and death.6 One of the most consistent prognostic factors related with poor prognosis is positive cytology of the pericardial effusion.13 42 43 Cancer type and stage and bloody pericardial effusion are other predictors of long-term survival.5 10 16 42 43 Patients with malignant pericardial effusion from lung cancer showed a median overall survival time significantly lower compared with those of patients with breast cancer3 5 11 13 16 or haematological malignancy.3 10 13 Moreover pathologically proven malignant effusion is independently associated with poor prognosis in lung cancer (table 3).3 5 10 11 13 16 18 43 Another feared complication is constrictive pericarditis, which can be a direct consequence of the disease (eg, pericardial mesothelioma) or a consequence of treatments, especially radiation. Late-onset radiation-induced pericardial constriction may be especially difficult to treat due to the presence of mixed disease (valvular disease, cardiomyopathy).