Article Text

other Versions

Original article
Acute catecholamine cardiomyopathy in patients with phaeochromocytoma or functional paraganglioma
  1. Alessandra Giavarini1,2,
  2. Antoine Chedid1,
  3. Guillaume Bobrie1,
  4. Pierre-François Plouin1,
  5. Albert Hagège3,
  6. Laurence Amar1
  1. 1Hypertension Unit, Université Paris-Descartes, Sorbonne Paris Cité, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris cedex 15, France
  2. 2Department of Internal Medicine, Università degli Studi di Milano, Ospedale San Giuseppe, Milan, Italy
  3. 3Department of Cardiology, Université Paris-Descartes, Sorbonne Paris Cité, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, Paris cedex 15, France
  1. Correspondence to Professor Pierre-François Plouin, Hypertension Unit, Université Paris-Descartes, Sorbonne Paris Cité, Assistance Publique-Hôpitaux de Paris, Hôpital Européen Georges Pompidou, 20 rue Leblanc, 75908 Paris cedex 15, France; pierre-francois.plouin{at}


Objective Phaeochromocytomas and paragangliomas (PPGL) can cause acute catecholamine cardiomyopathy (ACC). We assessed the prevalence of ACC and compared the presentation of cases with and without ACC in a large series of PPGL.

Design Single centre retrospective study.

Setting Hypertension Unit, University Hospital, Paris.

Patients 140 consecutive patients with PPGL, referred from January 2003 to September 2012.

Main outcome measures Left ventricular ejection fraction (LVEF), perioperative mortality.

Results Fifteen patients (11%) had suffered an ACC, occurring in 14 cases before the diagnosis of PPGL. Precipitating factors were identified in 11 cases. Twelve patients presented with acute pulmonary oedema, including 10 with cardiogenic shock, requiring life support in eight cases. Seven patients (five with pulmonary oedema) presented with acute chest pain and cardiac dysfunction. Electrocardiographic abnormalities were present in 14 cases: ST segment elevation or pathological Q waves, ST segment depression, and/or diffuse T wave inversion. Six patients displayed classical (apical ballooning) or inverted (basal/mid ventricular stunning) takotsubo-like cardiomyopathy. Coronary arteries were always normal on angiography. In patients with ACC, median LVEF rose from 30% (IQR 23–33%) during ACC to 71% (50–72%) before surgery (n=11, p<0.001). Median LVEF before PPGL surgery was 65% (51–72%) and 65% (60–70%) in patients with and without a history of ACC, respectively (not significant).

Conclusions PPGL may present as ACC in 11% of cases, excluding patients dying from undiagnosed tumours. Left ventricular dysfunction is usually reversible before surgery. PPGL should be suspected in patients with acute heart failure without evidence of valvular or coronary artery disease.

Statistics from


Phaeochromocytomas and paragangliomas (PPGL) are rare tumours of the paraganglia, neural crest derived organs associated with the autonomic nervous system. Adrenal tumours (phaeochromocytoma proper) and other thoracic/abdominal/pelvic paragangliomas are neoplasms of the sympathetic nervous system, and most secrete catecholamines.1 PPGL may occur sporadically, or as part of a hereditary syndrome.2 About 10% are metastatic, malignancy being frequently associated with germline mutations of the gene encoding subunit B of the mitochondrial complex II succinate dehydrogenase (SDHB).3 Clinical features may range from asymptomatic tumours discovered fortuitously (incidentalomas)4 to severe cardiovascular complications induced by the abrupt release of large amounts of catecholamines,5 sudden cardiac death representing the most dramatic presentation of these tumours with unpredictable secretion.6 Acute complications may occur spontaneously or following exposure to stress, effort, and various drugs.7 Acute catecholamine cardiomyopathy (ACC), resulting in severe left ventricular dysfunction, has been described,5 and a takotsubo-like contractile pattern has occasionally been reported.8 ,9

In this study, we assessed the prevalence, features at presentation, and outcome of ACC in a large series of consecutive patients with PPGL.

Patients and methods

Study population

PPGL was diagnosed in 202 patients referred to our unit between January 2003 and September 2012. We excluded 31 patients referred for a recurrence, 25 patients undergoing surgery for a non-secreting PPGL, and six patients who did not undergo surgery because they had non-secreting or metastatic PPGL or refused surgery. Thus, 140 patients undergoing surgery for a catecholamine secreting PPGL were available for analysis (figure 1). The procedures used for PPGL diagnosis and management were consistent with institutional guidelines and have been described elsewhere.4 All patients gave written consent for the collection of data and screening for genes associated with PPGL.

Figure 1

Patient flowchart. ACC, acute catecholamine cardiomyopathy; PPGL, phaeochromocytomas and paragangliomas.

Diagnostic criteria

A functional PPGL was diagnosed as an adrenal or extra-adrenal tumour associated with high concentrations of metanephrine in urine and confirmed by histology.1 PPGL were classified as malignant if metastases were present.1–4 ACC was diagnosed, in the context of PPGL, in the presence of the following combination: (1) acute chest pain and/or cardiac failure requiring hospitalisation; (2) biochemical, ECG and/or echocardiographic evidence of myocardial ischaemia with left ventricular systolic dysfunction; and (3) absence of obstructive epicardial coronary artery disease.10

Clinical assessment

Data concerning ACC presentation were obtained by reviewing reports from the intensive care units to which the patients were initially admitted. Given the emergency context, imaging and biochemical tests were not standardised during ACC. Clinical data and the results of biochemical and imaging tests obtained after the referral of surviving patients to our unit were collected prospectively on a standardised, computerised form.11 Diabetes was defined as current antidiabetic treatment or two fasting blood glucose concentrations exceeding 7 mmol/L.12 Catecholamine hypersecretion was assessed by determining metanephrine and normetanephrine concentrations, the determination of plasma or urinary metanephrines being a more reliable diagnostic test for PPGL than the determination of plasma or urinary catecholamines.13 Plasma-free metanephrine concentrations were not available in all cases, so we report here the urinary excretion of metanephrines, determined with high performance liquid chromatography, as previously described.14 The upper limits of the normal range for the 24 h urinary excretion of metanephrine and normetanephrine are 1.2 and 2.5 µmol, respectively. Cardiac biomarkers were assessed during ACC in most patients, but most clinical report files did not specify the biochemical methods or diagnostic thresholds used. We therefore analysed cardiac biomarkers as categorical variables (normal or high). Quantitative and qualitative data for left ventricular systolic function were extracted from transthoracic echocardiogram reports: left ventricular ejection fraction (LVEF) was calculated by the biplane Simpson's rule, after the delineation of endocardial contours on the two dimensional apical four and two chamber views.15 Global or regional left ventricular wall systolic thickening was assessed semiquantitatively, according to the recommendations of the American Society of Echocardiography, and described qualitatively as global left ventricular systolic dysfunction or segmental hypokinesia, akinesia or dyskinesia, using a 16 segment model; left ventricular dimensions and indexed ventricular mass were determined in line with current recommendations.16 Tumour location and the largest tumour dimension were determined by CT. In cases with multiple tumours, the largest diameter of the largest tumour was reported.

Preoperative management

Patients with diagnosed PPGL were treated as previously described.17 In brief, we aimed to reduce mean office blood pressure (BP) to below 140/90 mm Hg or 24 h ambulatory BP to below 125/80 mm Hg. However, the total abolition of hypertensive paroxysms is not possible, and patients therefore underwent surgery after 1–2 weeks of preparation. BP control requires α- and ß-adrenergic antagonists. Non-competitive α-blockers such as phenoxybenzamine are not available in France, so we used the competitive α-blocker prazosine. As the initial dose of prazosine may induce a sharp drop in BP, the dose is gradually increased from 0.5 to 5 mg three times daily. α-Adrenergic blockade generally gives rise to tachycardia, secondary to catecholamine ß-receptor stimulation. This requires the subsequent addition of a ß-blocker, such as 25–100 mg atenolol daily. If adrenergic blockade proves insufficient to control BP, then a dihydropyridine calcium channel blocker is added.

Data analysis

We compared features during the acute episode and just before PPGL resection in patients with ACC. We also compared preoperative features between patients with and without ACC. Continuous variables are presented as medians (IQR) and were compared in Wilcoxon–Mann–Whitney non-parametric tests. The Wilcoxon signed rank test with continuity correction was used for paired data. Binary variables are shown as absolute numbers (percentage; 95% CI) and were analysed with Fisher's exact test. Differences were considered statistically significant if the p value obtained was less than 0.05.


Of the 140 patients with a new PPGL at risk of ACC, 15 (11%)—four of whom were men—had at least one episode of ACC (figure 1 and table 1).

Table 1

Patients with acute catecholamine cardiomyopathy (ACC): presentation and outcome

Patients with ACC

PPGL was diagnosed after ACC in 14 cases, with a median time from ACC to diagnosis of 6 (2–60) days. One patient diagnosed with PPGL suffered ACC during the up-titration of β-blockers, with insufficient α-blocker therapy. ACC was precipitated by non-adrenal surgery in five cases (arthrodesis, cholecystectomy, mastectomy, plastic surgery, and hernia repair, in one patient each), by physical exercise in two patients, and by metoclopramide or steroid treatment in one case each. Two patients had respiratory tract infections in the days preceding ACC. One patient experienced two episodes of acute heart failure, 7 years apart. The second of these episodes, leading to PPGL diagnosis, was included in this analysis. The previous episode had been classified as acute myopericarditis, but, with hindsight, was probably due to catecholamine cardiotoxicity caused by an undiagnosed catecholamine secreting tumour.

ACC presented as abrupt cardiorespiratory failure in all 15 subjects. Twelve patients suffered acute pulmonary oedema, associated with chest pain in five cases. In 10 of these 12 patients, ACC progressed to cardiogenic shock requiring inotropic treatment and invasive ventilation, associated with intra-aortic balloon counter pulsation, extracorporeal membrane oxygenation or continuous veno-venous haemofiltration in three patients. The three patients without pulmonary oedema presented at the emergency room with sustained oppressive chest pain with stable haemodynamics (two) or heart failure (one). Twelve patients had high BP (170 (163–220)/110 (108–110) mm Hg) and 10 had a high heart rate (140 (129–149) beats/min (bpm)). Three patients were normotensive (BP 120 (118–127)/69 (69–73) mm Hg), including two with tachycardia (heart rate 132 and 140 bpm). Seven patients presented with sinus tachycardia, one with paroxysmal supraventricular tachycardia, and one with atrial flutter. Severe ventricular arrhythmias occurred in three patients, two with non-sustained episodes of ventricular tachycardia and one with ventricular tachycardia that evolved to ventricular flutter; this last patient and another patient who developed asystole had been successfully resuscitated after cardiac arrest. ECGs were available for 14 patients: signs of acute myocardial ischaemia/infarction (ST segment elevation or abnormal Q waves in the absence of bundle branch block) were present in seven cases, whereas ST segment depression and/or diffuse T wave inversion were present in the remaining patients. Troponin values were determined in 13 patients and were found to be high in 12.

Echocardiography was performed during ACC in 13 patients. It showed severe left ventricular systolic dysfunction in all cases, with either segmental but extended hypokinesia/akinesia or global hypokinesia, and severe LVEF reduction (30% (24–32%)). Echocardiography, cardiac MRI or ventriculography (figure 2) revealed wall motion abnormalities suggestive of classical or inverted takotsubo cardiomyopathy in six patients and global left ventricular hypokinesia in seven patients. Two patients had an apical left ventricular thrombus and were successfully treated with oral or subcutaneous anticoagulants. Coronary angiography was normal in the 11 patients in whom this procedure was performed.

Figure 2

Angiography in a patient with acute catecholamine cardiomyopathy. Angiography was performed 3 h after the abrupt onset of a pulmonary oedema precipitated by mastectomy. (A and B) Left ventriculography at end diastole and end systole (right oblique projection), showing severe basal and mid left ventricular dysfunction with the preservation of apical wall motion (similar to the inverted takotsubo pattern). Left ventricular ejection fraction  was estimated at 18%. (C and D) Coronary angiography showing normal arteries.

Median time from ACC to referral to our unit was 25 (20–58) days. Clinical recovery was observed in the 15 patients with ACC, either spontaneously or following preoperative α-blockade followed by ß-blockade therapy. LVEF was 65% (51–72%) in the 14 patients who underwent echocardiography before surgery. In the 11 patients for whom LVEF determinations were carried out during ACC and before surgery, LVEF rose from 30% (23–33%) to 71% (50–72%) (p<0.001) (figure 3).

Figure 3

Left ventricular ejection fraction (LVEF) during acute catecholamine cardiomyopathy (ACC), before surgery and at follow-up. Nine patients had echographic LVEF assessments at the three time points (open circles). Two patients had echocardiographic LVEF assessments during ACC and before surgery (closed circles). *Difference between LVEF during ACC and after recovery, before surgery.

Comparison of patients with and without ACC

The preoperative characteristics of the patients with and without ACC were compared (table 2).

Table 2

Patient characteristics before surgery: comparison of patients with and without acute catecholamine cardiomyopathy (ACC)

As all these patients were selected on the basis of functional PPGL (figure 1), most of the patients in both groups had a history of hypertension. There was no difference between the groups in terms of sex, age at diagnosis of PPGL or hypertension, history of alcohol abuse, or the prevalence of adrenergic symptoms and cardiovascular risk factors. Tumour number, site and size, metanephrine and normetanephrine excretion, and the distribution of mutations, where present, did not differ between patients with and without ACC either. Median systolic BP levels on referral to this unit were lower in patients with than in those without a history of ACC (127 (116–138) vs 141 (128–156) mm Hg, respectively, p=0.01). The two groups did not differ in left ventricular end diastolic dimensions (46 (43–52) vs 46 (43–50) mm, respectively), indexed mass (89 (66–102) vs 88 (74–105) g/m2, respectively) or LVEF (65% (51–72%) vs 65% (60–70%), respectively), but segmental (20% (5–50%) vs 3% (0–9%), p=0.02) and global (13% (2–41%) vs 1% (0–6%), p=0.04) left ventricular dysfunction was more frequent in patients who had suffered ACC.

Subsequent outcome

No perioperative deaths occurred in either group. Thirteen patients with and 93 patients without a history of ACC attended a follow-up outpatient visit 104 (96–157) and 155 (91–195) days, respectively, after tumour resection (figure 1). There was no difference in BP levels (124 (110–128)/78 (71–85) vs 127 (115–141)/77 (72–87) mm Hg) or treatment score (0 (0–1) vs 0 (0–1)) between groups. Follow-up echocardiograms were available for 13 patients with (figure 3) and 42 patients without a history of ACC; in these two groups, LVEF reached 62% (53–65%) and 65% (61–73%), respectively (p=0.01).


The increase in catecholamine production in patients with functional PPGL causes symptoms (mostly headaches, palpitations, and excess sweating) and signs (mostly hypertension and diabetes), reflecting the effects of catecholamines on α- and β-adrenergic receptors.1 PPGL may be diagnosed late for several reasons: (1) these tumours are rare; (2) patients may have long periods without hypertension, as active catecholamines can be converted into biologically inactive metanephrines within the tumour; and (3) the symptoms and signs are non-specific and common to catecholamine release by the tumour (in PPGL) and neurons (during stress). For these reasons, the mean time from the onset of hypertension, when present, and tumour diagnosis typically exceeds 3 years.4 The tumour may be discovered fortuitously4 ,18 or, conversely, during potentially lethal complications.5 ,6 ,19 We show here that ACC occurs in >10% of patients with functional PPGL and is often preceded by an identifiable precipitating factor. The tumour is often not diagnosed before the acute cardiovascular event occurs. Despite the severe clinical presentation, ACC usually resolves spontaneously or following adrenergic blockade, resulting in an uneventful postoperative outcome.

Presentation and prevalence

Acute cardiac complications of PPGL include dilated cardiomyopathy or hypertrophic obstructive cardiomyopathy, myocardial ischaemia, supraventricular or ventricular arrhythmia, aortic dissection, ACC, and sudden cardiac death.5–10 ACC in PPGL may be associated with segmental wall motion abnormalities suggestive of takotsubo cardiomyopathy, also known as stress induced cardiomyopathy or apical ballooning.9 ,20 In this study, we defined ACC as an acute and life-threatening cardiac condition combining chest pain and/or heart failure, evidence of acute myocardial ischaemia, left ventricular systolic dysfunction, and an absence of significant coronary stenosis. Based on this definition, the prevalence of ACC was 11%, with a takotsubo-like pattern observed in six patients. A previous, smaller series reported three cases of takotsubo-like cardiomyopathy in 36 patients with PPGL.20 In a series of 145 patients with PPGL, Zelinka et al21 reported cardiovascular complications in 28 cases. Five patients suffered heart failure, including two with takotsubo-like cardiomyopathy. Another seven patients presented with chest pain and high cardiac troponin concentration, without significant stenosis of the coronary vessels on angiography. As no information about LVEF was provided, we cannot confirm that these 12 patients fulfilled our criteria for ACC. If they did, this would give a prevalence of 8%, consistent with our own estimate.

Our objective was to assess the prevalence of ACC in patients with PPGL, but we also tried to estimate the frequency of takotsubo cardiomyopathy at our institution, so we searched all reports from the Cardiology Intensive Care Unit for the terms takotsubo, tako tsubo, and tako-tsubo; this was possible starting from 2006 to June 2013 and we retrieved 35 confirmed cases, including one patient with PPGL already included in our series. This confirms that ACC is a very rare event.

Diagnosing PPGL in the context of ACC

Only one of our 15 patients was known to have PPGL before ACC occurred. PPGL diagnosis remains difficult in patients admitted for an acute cardiac event because: (1) adrenergic signs and symptoms are similar whether the hypersecretion of catecholamines is mediated by a tumour or by neurons (stress induced); (2) acute coronary events and heart failure induce an increase in the excretion of catecholamines and metanephrine14; (3) determinations of catecholamine or catecholamine metabolites are not rapidly available. The attribution of an acute cardiac event to acute cardiomyopathy is generally based on the presence of very poor systolic myocardial function on echography and an absence of severe stenosis or obstruction on coronary angiography. The presence of ACC is then suggested by a positive history of hypertension and adrenergic symptoms without a history of exposure to toxic agents or infectious signs (which might suggest myocarditis). In this series, most of our patients with ACC had a history of hypertension and adrenergic symptoms. None had clinically relevant alcohol abuse. An absence of coronary obstruction was documented in 11 of 15 cases. Patients were referred 20–58 days after the onset of ACC and we could not determine precisely the timing of biochemical and imaging tests carried out before referral to our unit. However, given the emergency context, we consider the presence of a retroperitoneal tumour to be the most rapid and reliable clue to the diagnosis of PPGL.

Risk factors for ACC in PPGL

We identified no clinical, biological, tumour or genetic (table 2) characteristic associated with ACC in patients with PPGL. Patients with and without ACC were similar in terms of their urinary metanephrine and normetanephrine excretion. This does not rule out a role for catecholamine cardiotoxicity, however, as catecholamine release from PPGL is variable and may be triggered by tumour haemorrhage or necrosis, an increase in abdominal pressure during exercise, micturition or defecation, and by various drugs including anaesthetic agents.1 ,5 ,7 ,22 Such triggers—effort, non-adrenal surgery, metoclopramide, and steroids—were identified in nine of our 15 patients. The sudden onset of an acute cardiac event after routine surgery or exposure to drugs interfering with catecholamine metabolism in a previously healthy subject is probably the best clinical clue for the diagnosis of PPGL related ACC.

Management and outcome

No specific therapeutic approach has been proposed for stress related cardiomyopathy. The emergency management of arrhythmia, pulmonary oedema, and cardiogenic shock is based on conventional therapies, such as antiarrhythmic drugs, electrical shock, oxygen administration, diuretics, vasoactive amines, inotropes, and/or the prophylaxis of thromboembolism.23 Mechanical circulatory support, such as intra-aortic counter-pulsation, ventricular assist devices24 and/or ultrafiltration, may, however, be preferred over sympathomimetic drugs, as adrenergic stimulation is already maximal.10 α-Blockade is an option in stress related cardiomyopathy,25 and for the elective treatment of cardiac dysfunction or takotsubo-like cardiomyopathy associated with PPGL.1 ,9 ,24 ,26 Progressive α-blockade and subsequent β-blockade restored LVEF to normal levels before surgery in our ACC patients (figure 3). A recent study of a series of patients with phaeochromocytoma crisis, defined as severe hypertension or hypotension resulting in end organ damage, also recommended that patients with PPGL and a severe presentation should undergo medical stabilisation by α-blockade before surgery.27

Study limitations

This retrospective study has several limitations. Patients with ACC were first seen in various emergency departments and subsequently referred to this unit. The estimated prevalence of ACC in PPGL does not take into account lethal cases, potentially leading to underestimation, but this may be cancelled out by a referral bias that might lead to overestimation. The figure of 11% should therefore be taken as a reasonably accurate estimate. Early management and diagnostic procedures before referral were not standardised, and were not precisely documented in some cases. Most patients with and without ACC underwent preoperative management and surgery at our institution, but some of the patients in both groups were referred to us from distant areas and could not be followed up.


ACC is a non-rare complication of catecholamine secreting tumours, generally occurring before the diagnosis of these tumours. The clinical clues suggestive of PPGL related ACC are a history of hypertension and adrenergic symptoms, the onset of symptoms during non-adrenal surgery or exposure to drugs interfering with catecholamine metabolism, and an absence of alternative causes of acute cardiac failure, such as coronary occlusion. In emergency conditions, PPGL diagnosis is confirmed by imaging tests disclosing a retroperitoneal tumour. Left ventricular dysfunction may resolve spontaneously or following α-blockade, making it possible to carry out surgery safely, in non-emergency conditions.


This study was supported in part by PHRC grant AOM 06 179 and by grants from INSERM and Ministère Délégué à la Recherche et des Nouvelles Technologies for the COMETE Network. We thank I Cherrak for compiling all reports from the Cardiology Intensive Care Unit for takotsubo and related terms.


View Abstract


  • Contributors All the authors contributed equally to the drafting of this article. AC, GB, PFP and LA conceived and AG, AC, GB, PFP and LA designed the study, and AG, AC, GB, PFP and LA participated in the acquisition and presentation of data. AG, AC, GB, PFP, AH and LA contributed to the analysis and interpretation of data and to the drafting of the manuscript; PFP, AH and LA were substantially involved in revising the article. All authors read and approved the final version of the manuscript.

  • Funding This study was supported in part by PHRC grant AOM 06 179 and by grants from INSERM and Ministère Délégué à la Recherche et des Nouvelles Technologies for the COMETE Network.

  • Competing interests None.

  • Ethics approval All patients gave written informed consent for the collection of data and screening for genes associated with PPGL in the setting of the National French Register COMETE (cortico- medullo tumeurs endocrines), that was approved by the national ethics committee.

  • Patient consent Obtained.

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

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.