Article Text

Download PDFPDF

Original article
Myopericarditis in tuberculous pericardial effusion: prevalence, predictors and outcome
  1. Faisal F Syed1,
  2. Mpiko Ntsekhe2,
  3. Freedom Gumedze3,
  4. Motasim Badri2,4,
  5. Bongani M Mayosi2
  1. 1Division of Cardiovascular Diseases, Mayo Clinic, Rochester, Minnesota, USA
  2. 2Department of Medicine, The Cardiac Clinic, Groote Schuur Hospital and University of Cape Town, Cape Town, South Africa
  3. 3Department of Statistical Sciences, University of Cape Town, Cape Town, South Africa
  4. 4King Saud bin Abdulaziz University for Health Sciences, Riyadh, Saudi Arabia
  1. Correspondence to Dr Faisal F Syed, Division of Cardiovascular Diseases, Mayo Clinic, 200 First Street SW, Rochester, Minnesota, 55905 USA; syed.faisal{at}


Objective The prevalence, predictors and outcome of myopericarditis in patients with tuberculous (TB) pericarditis are unknown.

Methods Eighty-one patients (mean age±SD, 36.1±13.3 years; 54 (66.7%) men; 58 (71.6%) HIV seropositive) with TB pericarditis were recruited between January 2006 and September 2008. Myopericarditis was defined as echocardiographic LV systolic dysfunction (immediately after pericardiocentesis), elevated peripheral blood troponin T (>0.03 ng/mL), or elevated peripheral blood creatine kinase (CK >174 IU/L) with a CK:CK-myocardial band (MB) mass ratio of >6%. The outcome measure was case fatality rate at 6 months of follow-up.

Results Myopericarditis was present in 43 (53.1%) patients. Patients with myopericarditis, as compared with those without, were more likely to be HIV seropositive (35 (81.4%) vs 23 (60.5%) respectively, p=0.038) and have lower peripheral CD4 count (median (IQR) 98 (54–290) vs 177 (104–429), p=0.026). Electrocardiographic ST segment elevation was more common in myopericarditis (15 (36.6%) vs 4 (10.8%), p=0.008) and predicted myopericarditis independently of CD4 count on multiple logistic regression analysis (OR 4.36, 95% CI 1.34 to 17.34, p=0.0132). At 6 months, 14 (18%) patients had died with no significant difference between those with or without myopericarditis (6/42 (14%) vs 8/36 (22%), respectively (p=0.363)).

Conclusions Myopericarditis is common in TB pericardial effusion and associated with HIV-related immunosuppression. It can be identified by electrocardiographic ST-elevation, particularly when peripheral CD4 count is low. There was no significant difference in case fatality rate in those with or without myopericarditis.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


Tuberculous (TB) pericarditis is the most common form of pericardial disease worldwide.1 There has been a dramatic resurgence in the face of the HIV pandemic, with increased extrapulmonary or disseminated involvement.2 Close to 100% of pericardial effusions in those infected with HIV in sub-Saharan Africa are due to TB, compared with 50–70% without HIV and less than 5% in the developed world.3 Our preliminary observations suggest that in patients with TB pericarditis, co-infection with HIV is associated with increased clinical features of heart failure, haemodynamic instability, electrocardiographic ST elevation and an increased mortality, as compared with HIV-uninfected patients.4 ,5 These findings suggest that HIV-associated TB pericarditis is associated with an aggressive myopericarditis. In this study, we prospectively evaluated the prevalence of myopericarditis in TB pericarditis in sub-Saharan Africa, its underlying associations and impact on the case fatality rate after 6 months of follow-up.


Patient recruitment

Informed consent was obtained from each patient and the study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the human research committee of the University of Cape Town. Over a 3-year period, consecutive patients aged 18 years or above who presented with an effusive pericardial syndrome proven to be definite or probable TB pericarditis were prospectively recruited from three secondary care and one tertiary care hospitals in Cape Town as part of the ongoing Investigation of the Management of Pericarditis in Africa registry.4 TB pericarditis was defined by the exclusion of other causes of pericardial effusion and fulfilment of one of the following criteria: (1) presence of tubercle bacilli on smear or culture of pericardial fluid, (2) TB proven elsewhere in the presence of pericarditis or (3) elevated adenosine deaminase activity in a lymphocytic pericardial exudate defined using Light's criteria (whereby an exudate is defined as having one or more of the following: fluid protein divided by serum protein >0.5; fluid lactate dehydrogenase (LDH) divided by serum LDH >0.6; and/or fluid LDH level >66% of the upper limit of normal for serum LDH).1 Exclusion factors were inability to obtain informed consent, inability to gain pericardial access, alternative aetiology of pericardial effusion, concurrent malignancy, cardiomyopathy or structural heart disease present previously, renal failure requiring dialysis, pregnancy, established anti-TB chemotherapy for more than 1 week or a recent course of anti-TB chemotherapy (within 3 months).

Clinical protocol

All patients underwent fluoroscopically guided pericardiocentesis via the subxiphoid approach and a standard Seldinger technique at the Cardiac Catheter Laboratory of Groote Schuur Hospital, University of Cape Town. Upon complete drainage of pericardial fluid, the pigtail catheter was withdrawn. Twelve-lead electrocardiography and 2D transthoracic echocardiography were performed on arrival to the hospital and immediately after pericardiocentesis. Patients were subsequently admitted overnight for observation before being transported to the referring institution. Patients were treated with antituberculosis chemotherapy. HIV-positive patients were treated with antiretroviral therapy according to local protocols. Follow-up was performed for up to 6 months and those not attending follow-up were actively traced to determine vital status at 6 months.

Laboratory evaluation

Simultaneous samples of blood and pericardial fluid were taken in all cases at the time of pericardiocentesis. Sampling of sputum, pleural or peritoneal fluid and tissue biopsy was determined by the clinical indication. All samples were subject to the necessary biochemical analyses to ascertain pericardial adenosine deaminase activity and Light's criteria, Gram and fungal stains, bacterial and fungal cultures and cytological analysis according to routine hospital protocol. A standard assay was used to measure serum creatine kinase isoenzyme MB mass (CK-MB, reference range <6% of CK-MB/CK (38–174 IU/L)) and troponin T (<0.03 ng/mL). All patients had measurement of serum antibodies to HIV 1 and determination of peripheral blood CD4 T cell counts.

Data interpretation

All ECGs and echocardiographic images were reviewed and interpreted by one investigator (FFS) and confirmed by a second investigator (MN). LVEF was measured by 2D transthoracic echocardiography and used to define LV function as normal (LVEF ≥55% or fractional shortening (FS) 25–43%) or abnormal (further defined as mild (EF 45–54% or FS 20–24%), moderate (EF 36–44% or FS 15–19%) or severe (EF ≤35% or FS <15%)). Myopericarditis was defined, in line with criteria previously described, as the presence of abnormal LV function, elevated peripheral blood troponin T (>0.03 ng/mL) or elevated peripheral blood CK (>174 IU/L) with a CK:CKMB mass ratio of >6%.6 Electrocardiographic ST segment elevation was noted when there was elevation of the J-point by at least 0.5 mm relative to the isoelectric line in two or more contiguous leads.

Statistical analysis

The Shapiro-Wilk test was used to determine distribution of continuous data. Data were compared using the Student t test, Wilcoxon test or χ2 test, as appropriate. All tests were two-sided and p value <0.05 was considered statistically significant. Mechanistically independent variables which were significantly associated with myopericarditis in univariate analysis were entered into a final multivariate model, as long as they were not used for defining myopericarditis and there was no colinearity. The JMP software V.8 (SAS Institute, Cary, North Carolina, USA) was used for statistical analysis.


Between January 2006 and September 2008, 81 patients (54 (66.7%) men) were recruited of whom 58 (71.6%) had positive serology for HIV (figure 1). HIV infected patients were younger than non-HIV infected patients (median (IQR) age 31 (28–39) years vs 41 (28–58) years, respectively, p=0.033), and had, as expected, a lower CD4 count (95 (54–223) cells/µL vs 429 (177–651) cells/µL, p<0.0001) compared with non-HIV infected patients with TB pericarditis.

Figure 1

Summary of prevalence and 6-month all-cause mortality according to HIV serology and presence of myopericarditis (*Differences did not reach statistical significance (p>0.05 for all comparisons). Access the article online to view this figure in colour.

Myopericarditis was present in 43 (53.1%) patients on presentation (table 1). Abnormal LV function was present in 24 (29.6%) and CKMB elevation in 26 (32.1%) patients (table 2). However, there was only one patient who had troponin T elevation. Patients with myopericarditis were more likely to be HIV seropositive (35 (81.4%) vs 23 (60.5%), p=0.038) and have lower CD4 T cell counts (median (IQR) 98 (54–290) cells/µL vs 177 (104–429) cells/µL, p=0.026) than those without myopericarditis. The frequency of myopericarditis in those with CD4 count <100 cells/µL was 51% compared with 14% in those with CD4 count >400 cells/µL (figure 2). Although there was no difference in symptoms, signs, New York Heart Association class, blood pressure or respiratory rate (table 1), patients with myopericarditis were more likely to have a shorter duration of symptoms of TB pericarditis (3 (2–5) weeks vs 4 (4–8) weeks, p=0.032) and a higher pulse rate on presentation (mean±SD 119±17 rate/min vs 109±20 rate/min, p=0.035).

Table 1

Characteristics of participants with and without myopericarditis

Table 2

Performance of echocardiography and cardiac enzymes in defining myopericarditis

Figure 2

Frequency of myopericarditis by CD4 count stratifications. Access the article online to view this figure in colour.

Electrocardiographic ST segment elevation was more frequently seen with myopericarditis than without, before (15 (36.6%) vs 4 (10.8%), p=0.008) and after pericardiocentesis (15 (34.2%) vs 5 (13.9%), p=0.04). Those with ST segment elevation were more likely to have CKMB elevation (prepericardiocentesis: 12/19 (63.2%) with ST elevation vs 12/59 (20.3%) with no ST elevation, p=0.0004; postpericardiocentesis: 10/19 (52.6%) vs 14/58 (24.1%), p=0.020).

Vital statistics at 6 months could not be obtained on three patients (figure 1). Of the remainder, 14 (18%) patients had died at 6 months with no significant difference in case fatality rate between those with or without myopericarditis (6/42 (14%) vs 8/36 (22%), p=0.363).

In multivariate logistic regression analysis, only electrocardiographic ST segment elevation was independently and significantly predictive for the presence of myopericarditis (OR for ST elevation pre pericardiocentesis=4.36, 95% CI 1.34 to 17.34, p=0.0132). On receiver operating characteristics analysis, a CD4 T cell count cut-off of 98 cells/µL provided best performance at predicting myopericarditis, with 51.3% sensitivity and 79.4% specificity. When ST segment elevation was present, a CD4 T cell count cut-off of 103 cells/µL provided the best performance at predicting myopericarditis with 53.3% sensitivity and 100% specificity (figure 3).

Figure 3

Receiver operator characteristics of CD4 count with and without electrocardiographic ST segment elevation. Access the article online to view this figure in colour.


The main finding of the study is that the prevalence of myopericarditis, as determined by reduced LV systolic function or elevated circulating myocardial enzymes, in TB pericarditis was 53.1%. This is the first prospectively evaluated data on myopericarditis in TB pericarditis. The results are comparable with other series from sub-Saharan Africa in the post-HIV era. In Burkina Faso, LV systolic dysfunction was present in 40% of patients with HIV and TB pericarditis.7 In the contemporary Heart of Soweto study, the prevalence of cardiomyopathy (defined as isolated LV systolic dysfunction) in patients with HIV was 38%, with median CD4 counts of 180 cells/µL.8

Myopericarditis was found in increased frequency in those who were HIV seropositive and in those with marked immunosuppression, defined as a CD4 count of <100 cells/µL, in whom prevalence was 51%. It was independently predicted by electrocardiographic ST elevation, which increased risk of myopericarditis 4.3-fold. In those with HIV-associated immunosuppression (CD4 <103 cells/µL), the presence of ST elevation had 53.3% sensitivity and 100% specificity for myopericarditis. We noted only one positive troponin T assay and suspect the presence of an assay inhibitor as this finding was unexpected and out of keeping with the CK and CKMB results.

Infection with HIV results in altered cell-mediated immunity, including a phenotypical and functional change in Mycobacterium TB-specific CD4 T cells recruited to the sites of TB invasion and local cytokine responses.9 In advanced HIV with severe CD4 depletion, there is an altered histological pattern, with less granuloma formation.10 HIV infected patients also have a reduced risk of subsequent constrictive pericarditis as compared with non-HIV infected patients.11 The current findings extend these observations to a model of pericarditis, in which infection with TB is unchecked due to HIV immunosuppression, resulting in increased myocardial inflammation, cardiac enzyme release, electrocardiographic ST elevation and LV systolic dysfunction.

The current study was conducted prior to the wide availability and access to highly active antiretroviral therapy (HAART) in South Africa. This is reflected in the fact that only 2 of the 58 participants (3.4%) who were HIV infected were on HAART. While there is no evidence to date that HAART influences the outcome of TB myopericarditis per se, there is evidence that the timely introduction of HAART improves tuberculosis related outcomes in HIV infected patients.12 ,13 The low proportion of participants on HAART meant it was not possible to explore the possible impact of HAART on myocarditis in this study.

The distinction between myopericarditis, in which there is pericarditis with evidence of some myocardial involvement (usually without LV dysfunction), and perimyocarditis, in which there is myocarditis with pericardial involvement (often with depressed LV function), may have implications for management and outcomes.14 In our study 15% (7/43) of participants labelled as myopericarditis had moderate or severe LV dysfunction; all were alive at 6 months suggesting that the LV dysfunction had little if any negative impact on mortality.

Whether the LV dysfunction was de facto caused by an associated TB myocarditis was not demonstrated definitively by this study. We did not biopsy or culture myocardial tissue due to ethical considerations. HIV-associated cardiomyopathy is strongly associated with a CD+ cell count less than 100 cells/µL and other potential causes include myocarditis from opportunistic infections or sepsis syndrome and cytokine mediated LV dysfunction.15 This is an area of ongoing research, with modalities such as cardiac MRI which can provide superior tissue characterisation to echocardiography and identify the exact nature of TB myocardial involvement,16 similar to what has been reported on myopericarditis in Western European series.14

Although the exact cause of death was not possible to be determined due to inherent challenges in loss to follow-up of patients in South Africa,17 the country's vital registration system, which has been shown to record over 80% of deaths in recent years,18 allowed for an assessment of all-cause mortality. The 6-month case fatality rate of 18% was consistent with outcome in TB pericarditis,5 and there was no statistically significant difference between those with and without myopericarditis. Myocarditis is not associated with increased mortality in non-TB forms of pericarditis,14 and it is possible that this also applies to TB pericarditis. However, the small sample size of this study may not provide a definitive answer to this question. Moreover, pericardiocentesis was performed for all patients and therefore the natural course of the disease may have been altered.19

These results are demographic specific and determined largely by the background degree of HIV-immunosuppression and population at risk from TB. In a contemporary Western European series of acute pericarditis, with myopericarditis diagnosed on similar criteria to the current study, the prevalence was 25% and over 90% of cases were idiopathic or autoimmune.14 Similar to our findings, ST elevation was more commonly seen with myopericarditis (OR = 5.4, 95% CI 1.4 to 20.5, p = 0.013). In the study of Imazio and colleagues, there was an increased risk of cardiac arrhythmia in patients with myopericarditis (65% vs17%, p<0.001). By contrast, TB pericarditis is associated with transient atrial fibrillation in a quarter of cases with no evidence of ventricular arrhythmia.20

In summary, we describe a high prevalence of myopericarditis in TB pericardial effusion which is associated with HIV-related immunosuppression. Myopericarditis can be identified with high specificity by the presence of electrocardiographic ST elevation, especially if the CD4 count is low. The presence of myopericarditis had no impact on the case fatality rate at 6 months in patients presenting with TB pericardial effusion.

Key messages

  • What is already known on this subject

  • In the era of HIV, there has been a resurgence of tuberculous pericarditis in sub-Saharan African and other areas of the world where tuberculosis is endemic. Significantly increased rates of heart failure and mortality have been reported in HIV-associated tuberculous pericarditis, as compared with patients without HIV. Whether this is related to increased myopericarditis is currently unknown.

  • What this study adds

  • This is the first prospective study to investigate the prevalence of myopericarditis in tuberculous pericarditis, its association with HIV and HIV-related immunosuppression, clinical predictors and outcome.

  • How might this impact on clinical practice

  • This knowledge is important for determining optimal management of patients with tuberculous pericarditis with confirmed or suspected myopericarditis.



  • FFS and MN contributed equally.

  • Contributors FFS and MN contributed equally to this work and were primarily involved in all stages of the research from initiation, data collection, analysis and manuscript preparation. BMM acted in a supervisory capacity for project conceptualisation, oversight of execution, critical review of the data and writing of parts of the manuscript. FG and MB were involved in data analysis and critical review of manuscript.

  • Funding This work was funded by the South African Medical Research Council and the Lily and Ernst Hausmann Research Trust. FFS was a Lancet International Fellow when this project was initiated in 2005–2006.

  • Conflicts of interest None.

  • Competing interests None.

  • Patient consent Obtained.

  • Ethics approval Human research committee of the University of Cape Town.

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Data sharing statement Please contact the corresponding author with requests for data sharing, stating the purpose for which the data is intended to be used.

  • Disclaimer FFS and MN take responsibility for all aspects of the reliability and freedom from bias of the data presented and their discussed interpretation.