|Home | About | Journals | Submit | Contact Us | Français|
Cardiac tamponade is a life-threatening clinical syndrome that requires timely diagnosis. Herein, we present an instructive case of a patient who had cardiac tamponade. The condition went undiagnosed and resulted in the patient's death because almost all of the pathognomonic clinical findings of tamponade were unrecognized or not manifest. To better prepare health care professionals for similar challenges, we discuss the symptoms and clinical signs typical of cardiac tamponade, review the medical literature, and highlight current investigative methods that enable quick, efficient diagnosis and treatment.
Cardiac tamponade is a clinical syndrome caused by an increase in intrapericardial pressure due to the accumulation of blood, pus, other fluid, or gas in the pericardial space. Cardiac tamponade is prevalent in 25% to 30% of large pericardial effusions,1–3 which can manifest themselves as exudative or (rarely) tran-sudative accumulations. Cardiac tamponade typically leads to a crisis by decreasing venous return, which impairs diastolic ventricular filling. If the tamponade is untreated, hemodynamic compromise ensues consequent to the diminishing cardiac output.1–3 Therefore, as a continuum, cardiac tamponade constitutes a medical emergency. The patient's risk of death depends on the speed of diagnosis and treatment, the volume and rate of fluid accumulation, and the cause of the tamponade (Table I).3,4 An atypical, instructive case of cardiac tamponade frames our review of this severe medical condition and enables us to point out several substantial pitfalls of diagnosing cardiac tamponade solely on the basis of classic clinical findings.
A 42-year-old man, a cigarette smoker who had appeared healthy until he was diagnosed with stage IIIB lung adenocarcinoma 4 months earlier, presented at the emergency room in October 2004 after having experienced 3 days of increasing dyspnea. The patient had previously undergone thoracentesis, pleurodesis, and 6 cycles of chemotherapy with gemcitabine and carboplatin (the last cycle was 3 weeks earlier). He reported mild pain in the lateral chest wall and in his back. He had also had a minimally productive cough for 3 days, which exacerbated the pain, as did deep inspiration and sitting up. Lying on his left side alleviated the pain.
A review of the patient's systems revealed a 6-day history of mild, generalized weakness and 2 days of moderate periumbilical pain. The patient had not undergone radiation therapy or implantation of intravenous catheters or ports. He was afebrile, noncyanotic, and mildly diaphoretic without severe distress (heart rate, 100 beats/min; blood pressure, 120/53 mmHg; respiratory rate, 20–22 breaths/min, oxygen saturation, 97% on 2 L oxygen). The pain was reproducible, and tenderness existed more on the right side of the chest and back than on the left; there was no evidence of trauma. Also noted was bibasilar decreased air entry with crackles. The patient had a soft, nondistended abdomen with mild tenderness over the periumbilical region, but without rebound or guarding, and without palpable masses or organomegaly.
Laboratory findings showed elevations in the white blood cell count, creatinine levels, and cardiac and liver enzymes. A chest radiograph revealed a right lower-lobe infiltrate and moderate right pleural effusion; results of electrocardiography (ECG) were unremarkable except for tachycardia. Thoracentesis was scheduled. The patient was admitted to the hospital for empiric treatment of pneumonia and to receive maintenance intravenous fluids at 70 mL/hr.
Six hours after admission, the patient's lactate level was 11.6 mmol/L. However, a physical examination revealed no changes, and, notably, pulsus paradoxus was not detected. Repeat ECG was unremarkable. Measurements of arterial blood gas on 2 L of oxygen showed mild alkalosis that was consistent with tachypnea and could be attributed to the pneumonia and pleural effusion. Lactate levels were tested again to preclude laboratory error, and close monitoring of the patient continued.
Ten hours after his initial presentation, the patient described increased periumbilical pain; within minutes, he experienced cardiac arrest. Cardiopulmonary resuscitation (CPR) and intubation were performed. During resuscitation, arterial blood gas revealed severe decompensated metabolic acidosis, and lactate levels had risen sharply. His blood pressure fell to 48/16 mmHg and his heart rate rose to 188 beats/min. His now low-voltage ECG revealed atrial fibrillation, and a chest radiograph showed an enlarged cardiac silhouette separate from the known infiltrate and pleural effusion. Amid ongoing CPR for rapidly declining vital signs, emergent 2-dimensional transthoracic echocardiography confirmed a massive, global, tamponading pericardial effusion. The echocardiogram showed inferior vena caval plethora without collapse during ventilation, and a right ventricular diastolic invagination. Emergency subxiphoid pericardiocentesis was performed, and >300 cc of dark, unclotted, bloody fluid was drained. The patient's vital signs returned to normal levels almost immediately. Despite the creation of a pericardial window, however, the patient relapsed repeatedly into cardiac tamponade over the next 24 hours. Microscopic analysis of resected tissue confirmed direct pericardial invasion from the lung adenocarcinoma with diffuse metastases (stage IV).
Over the next 48 hours, lactate levels continued to rise, other blood levels (including serum creatinine and liver transaminases) worsened, and physical findings deteriorated. Progressive abdominal distention suggested possible bowel ischemia. A laparotomy revealed a congested liver with multiple mesenteric nodules and metastases positive for adenocarcinoma on frozen section. The patient was made comfortable in accordance with his advance directives, and he died upon the withdrawal of resuscitation.
This unusual case of cardiac tamponade went undiagnosed even during ongoing evaluation, because the clinical findings commonly associated with tamponade1–3,5,6 were absent or unapparent. Of note, signs of tamponade were not looked for specifically during the CPR mandated by the patient's collapse; the focus was on reviving the patient. Even had signs then been present, meaningful intervention might have come too late, and the ongoing CPR would have complicated or obscured recognition of the signs. Regardless, this cautionary case prompts a review of cardiac tamponade and underscores the imprudence of diagnosing the condition only when its signs and symptoms are classic and overt.
An attenuated apical heartbeat (noted in 34% of cardiac tamponade cases), distant heart sounds, and narrow pulse pressures are 3 common signs of cardiac tamponade7,8 that were not detected in this presentation. Also notably absent was pulsus paradoxus (a drop in systolic blood pressure ≥10 mmHg upon inspiration), which is found in 77% of cardiac tamponade cases—although, for optimal evaluation of pulsus paradoxus, respiration should be normal,9–12 and that of our patient was not. Pulsus paradoxus can be absent in low-pressure or right atrial cardiac tamponade,10–12 but our echocardiographic evidence and the amount of pericardial drainage (>300 mL) do not support this in our patient. Pulsus paradoxus can also be seen in nontamponade presentations that involve labored breathing (as in pneumonias and asthma) or right or left ventricular dysfunction.9–11 During CPR, our patient was in atrial fibrillation, which would have rendered difficult the reliable detection of pulsus paradoxus.
Jugular venous distention (JVD) is almost always noted in cardiac tamponade cases; however, it was not documented in our patient. Tachypnea is present in 80% of tamponade cases and tachycardia in 77%7,10,11; although both were distinct in our patient, they are highly nonspecific for cardiac tamponade. Kussmaul's sign (paradoxical JVD and systolic increase on inspiration),10,11 although usually occurring in conjunction with constrictive pericarditis and rarely with cardiac tamponade, was not documented in our patient. Also not found was Beck's triad (first described in 1935 as JVD, hypotension, and muffled heart sounds), which is present in 10% to 40% of tamponade cases.3,7 Without appreciable JVD, neither the rapid “x” descent nor the attenuated or abolished “y” (diastolic) descent of the jugular vein could have assisted in our diagnosis. Ewart's sign (an area of dullness with bronchial breath sounds and increased tactile fremitus below the left scapular angle due to compressive atelectasis by the pericardial sac—also known as Pins' or Bamberger-Pins-Ewart sign) is often observed in large pericardial effusions,13 but it was either not present or was masked by crackles from the incidental right-lower-lobe pneumonia.
Chest pain in tamponade is usually more severe or only on the patient's left side, and lying on the left side exacerbates it. The pain is rarely reproducible with chest wall palpation, except in cases of connective-tissue disease7 or trauma. Our patient presented with right-sided pain, which was relieved when he lay on his left side; this was a peculiar aspect of his presentation.
The periumbilical pain was puzzling, because abdominal pain is uncommon in cardiac tamponade; in addition, when pain is present, it most often occurs in the right upper quadrant. Reddy and colleagues described cardiac tamponade5 and its progression through 3 phases.6 In Phase I, accumulation of pericardial fluid causes increased stiffness of the ventricle, requiring a higher filling pressure. During this phase, the left and right ventricular filling pressures are higher than that of the intrapericardial pressure. Phase II occurs upon further fluid accumulation, which causes an increase in pericardial pressure above the ventricular filling pressure, resulting in reduced cardiac input and output. A further decrease in cardiac output occurs in Phase III due to equilibration of the pericardial and left ventricular filling pressures, and this precipitates severe deterioration of end-organ perfusion and leads ultimately to shock. Therefore, our patient's unusual presentation of abdominal pain could be explained by liver engorgement due to the back-pressure of blood into the vena cava from a poorly filling right heart, and to end-organ ischemia.
The existence of ischemia would also help explain the patient's worsening renal function, among other abnormal laboratory measurements that were indicative of either a more severe underlying disease process or (in retrospect) the patient's gradual progression through the 3 phases of cardiac tamponade.6 The notably elevated lactate levels further support the ischemia hypothesis, but, because the lactate levels were markedly disproportionate to the patient's nondistressed state at that time, it was plausible that they were an indirect result of the patient's highly aggressive lung adenocarcinoma. An alternative explanation, aside from the possibility of chemotherapy's cumulative adverse effects on the liver, is that gradual centrilobular hepatic congestion could have impaired lactate metabolism. The liver congestion on laparotomy and the elevated liver transaminases are perhaps consistent with this last explanation. An elevated white blood cell count is not usually associated with cardiac tamponade unless the tamponade is of infectious origin1–4; in this case, the elevated count could have been caused by the patient's pneumonia.
Chest radiography initially failed to show either an enlarged cardiac silhouette (observed in more than 90% of cardiac tamponade cases) or a silhouette in the shape of a water bottle, which suggests cardiomegaly.7,14 Interestingly, as seen in our patient, cases of pericardial effusion and cardiac tamponade frequently feature large pleural effusions,15,16 usually left-sided. These effusions can transmit the increased intrapleural pressure to the pericardial space and further impair ventricular filling, thus altering the intrapericardial-pressure equilibrium; and tamponade can arise even from modest pericardial effusions.17,18 Electrocardiographic changes such as PR-segment depressions, ST-segment elevations, and T-wave flattening (highly nonspecific)19 can help direct clinicians toward a diagnosis of cardiac tamponade in the presence of other clinical findings, but these changes were not noted in our patient. Electrocardiograms in cardiac tamponade may show low-amplitude QRS complexes, signifying low voltage, and may also depict quasi-specific electrical alternans (caused by “swinging” of the oscillating heart in the buoyant pericardial sac); this latter effect is seen in 10% to 20% of tamponade cases, most of which are neoplastic in origin.19–21 Moreover, variants of electrical alternans can also accompany pulmonary embolism, myocardial ischemia, and certain tachyarrhythmias.22,23 In our patient, the low-voltage ECG did not confirm the diagnosis of cardiac tamponade in the absence of other phenomena; therefore, emergency echocardiography was performed.
In this instructive, albeit humbling, case of cardiac tamponade of neoplastic origin,4 the most plausible scientific explanation for the initially unnoticed tamponade is that when the intrapericardial pressure from the accumulating fluid reached an almost vertical ascent in the pressure–volume (strain–stress) curve (Fig. 1), a small increment of fluid in the pericardial sac caused acute cardiovascular decompensation and led to shock.1–3 Slowly collecting, subacute cardiac tamponade is common in malignant cases, in contrast with the typically acute cardiac tamponade that arises from other causes. Jugular venous distention, if not prominent, could indicate subacute cardiac tamponade: it is acute tamponade—especially acute hemopericardium—that often produces exaggerated jugular pulsations, because there is insufficient time for the venous blood volume to increase and equilibrate.3,12 In our patient, gradual expansion after the filling of the pericardial reserve allowed more time for compensatory mechanisms to effectively react, due to pericardial stretching. This is how the slow accumulation rate of the malignant effusion—and not volume alone—played a crucial role in the slow rise in intrapericardial pressure (Fig. 1), and it is why intraventricular pressure was preserved until the sudden cardiovascular collapse. The patient thus remained in essentially stable condition, and the clinicians were un-aware of the built-up intrapericardial pressure and imminent cardiovascular embarrassment until the “last-drop phenomenon,” when the final increment produced critical cardiac chamber compression.3 Through similar mechanisms, the 1st decrement during drainage therefore produced the largest relative decompression.24
The pressures transmitted by the left pleural effusion, as mentioned earlier,17,18 could have attenuated the pulsus paradoxus by altering ventricular and overall hemodynamic equivalence and could have negated or masked most of the findings, such as an attenuated apical beat, distant heart sounds, and even electrical alternans on ECG. It is also possible (although uncommon) that the invading adenocarcinoma facilitated sudden bleeding into the pericardial sac.
Chest radiographs and ECGs are often unreliable for the prompt diagnosis of cardiac tamponade, because findings are not specific and may not even exist.3,14 Other methods, such as computer-assisted tomography, cardiac magnetic resonance imaging,25 and fluoroscopic imaging, are less than ideal as diagnostic methods because they are expensive and time-consuming. Doppler echocardiography is currently the most useful diagnostic tool for cardiac tamponade. Echocardiography is not only often readily available, inexpensive, and easy to use, but it can, in most cases, pinpoint the cardiac tamponade. This method enables direct views of infe-rior vena caval and right ventricular changes, and it can exclude the diagnosis of tamponade when fluid or gas is absent.26,27 Additional causes of tamponade, such as infiltration, mass rupture, or heart wall rupture, can be ascertained, as can the degree of cardiac dysfunction. Then too, echocardiography performed after pericardiocentesis can aid in evaluating the effectiveness of the procedure in relieving the tamponade or effusion. When used concomitantly, echocardiography can help to guide pericardiocentesis.27
Diagnosis of almost any cardiac disease by relying only on classic signs is an imprudent clinical approach in general, and worse in atypical or challenging situations. It is important to evaluate the overall clinical picture and to consider the underlying pathophysiology of the disease process. A comprehensive approach to cardiac tamponade can enable a timely, appropriate clinical diagnosis whereby hemodynamic embarrassment or crisis can be minimized or averted. Emergent echocardiography plays a key role in confirming the diagnosis and can also assist in treatment when pericardiocentesis is performed.
We thank Maureen Kuppe from St. Boniface General Hospital for her kind assistance in preparing the figure.
Address for reprints: Vignendra Ariyarajah, MD, C5 Cardiology, St. Boniface General Hospital, 409 Tache Avenue, Winnipeg, Manitoba R2H 2A6, Canada. E-mail: moc.liamtoh@ardnengiv