Cardiac tamponade has appeared in the fellowship exam directly in a couple of scenarios; for instance Question 11 from the second paper of 2015 , and  Question 15 from the first paper of 2011 where the college asked the candidates to describe how they would perform blind pericardiocentesis.

From peer-reviewed journals, if one had to choose one article it would have to be the NEJM paper by Spodick from 2003. It of course is paywalled. Spodick co-authored a subsequent paper in 2007, which was thankfully published in the somewhat less prestigious Texas Heart Institute Journal, and is therefore free. Unless otherwise stated, most of the information here comes from these two excellent articles. For minutiae regarding the various eponymous physical signs, a great article by Lange et al (1966) is also freely available online. 

Causes of pericardial effusion

The chronicity of the effusion plays a role in tamponade physiology. The pericardium is a fibrous sack which is described as "inextensible", that is to say at any given moment there is little stretch. Ergo, small volumes quickly will give rise to a rapid increase in pericardial pressure, giving rise to tamponade physiology. This produces decreased cardiac output because the stroke volume is limited by the presence of incompressible pericardial fluid. Interaction of respiratory variation in right atrial venous return with this decreased diastolic filling gives rise to all the interesting clinical and echocardiographic signs of cardiac tamponade.

Clinical features of cardiac tamponade

• Nonspecific findings which include tachycardia and tachypnoea

• Beck's Triad: Muffled heart sounds, hypotension and raised CVP. Claude Beck actually wrote about two  triads in 1935, of which the one everybody remembers is for acute cardiac tamponade. The other was chronic, and consisted of "(1) a high venous pressure, (2) ascites and (3) a small quiet heart." 
• Kussmaul's sign: the neck veins distend with inspiration, instead of collapsing. Adolf Kussmaul described this in 1873 in the case of "a sturdy 21-year-old farming boy" with pericardial effusion. "The jugular veins became considerably swollen, those on the right side were coiling; by each inspiration, a slight increase of its contents could be noted" he wrote. The mechanism behind this is probably the flush of blood produced by the contracting diaphragm as it pushes on intra-abdominal contents. Normally this would be accepted by the right heart and pumped normally, but in tamponade the non-compliant right ventricle is unable to accomodate the extra blood flow. With this, the right atrial pressure exceeds the inspiratory fall in pleural pressure - and therefore the central veins distend.
• Friedreich's sign: an exaggerated early drop in diastolic CVP, remarked upon by Friedreich in his 1864 publication, Zur Diagnose der Herzbeutelverwachsungen. This is also seen in constrictive pericarditis. The usual manifestation on the CVP trace is a rapid y descent, followed by a rapid rise in pressure.  It happens when the tricuspid valve opens in early right ventricular diastole; the right ventricle fills rapidly because of increased right atrial pressure, but is unable to take all of the right atrial volume and so diastole is cut short.  Lange et al (1966) found this sign was often absent in cardiac tamponade, and was more associated with restrictive and heavily calcified pericardial disease.
• Pulsus paradoxus - an exaggeration of the normal inspiratory fall in blood pressure, which has for some reason attracted a considerable amount of college examiner attention and therefore merits its own section, below.
• Pericardial rub  if the tamponade is associated with some sort of pericardial inflammation
• Pericardial "knock", first described by Maynard Smith  (consulting surgeon of the British Expeditionary Force, 1918)  as a sound which  "may be compared to that heard in the ear-piece of a telephone when the lever is moved up and down". Though Smith felt it was "unlike anything met with in civil practice", others (Tyberg et al, 1980) found that in civil practice it is also associated with constrictive pericarditis, and is probably the result of a sudden cessation of ventricular filling (i.e. it is the audible epiphenomenon of the same pathology which produces Friedreich's sign).
• A third heart sound and the pericardial knock are frequently described together and may in fact be either confluent or represent the same sounds (Hancock, 1971). However, the third heart sound is also seen in many other conditions (MR, VSD, dilated cardiomyopathy and so forth). Ozawa et al (1983) discovered that "S3 is due to a sudden intrinsic limitation of longitudinal expansion of the left ventricular wall during early diastolic filling, resulting in a negative jerk that is transmitted to the skin surface".
• Displaced apex beat  is not mandatory - it tends to happen with chronic effusions, and radiologically the cardiac silhouette is unaffected until at least 200 ml of fluid has accumulated (Spodick, 2003)
• Characteristric CVP findings: classically, a sawtooth "M" or "W"  configuration of a raised CVP. Chapter 18 (Pericardial Constriction and Cardiac Tamponade, p.313 ) from Jonathan Marks' Atlas of Cardiovascular Monitoring (1993) has an entire section on the CVP trace in cardiac tamponade. The waveform patterns below were abducted from Marks and severely molested with vector graphics to improve clarity.
  
  In summary
  • The CVP is raised
  • All CVP waveform components are elevated
  • a and v waves are tall
  • x descent is steep
  • y descent is (usually) absent

The y descent is usually absent in tamponade because early diastolic blood flow from the right atrium to the right ventricle is impaired by the compressive effect of the surrounding pericardial fluid.

Realistically, these changes are often lost in the noise. Published waveforms (eg. those grainy scans reproduced below from Shabetai et al, 1970) and real bedside observations (on the right) demonstrate that though the a and v components are objectively elevated the whole waveform is elevated by an even greater degree, so that these fine structures are obliterated by the respiratory pressure variation and the "zoomed out" effect of having to scale your waveform into the 20-30 mmHg range.

These images are probably more interesting in their representation of pulsus paradoxus, which will be the next topic of discussion.

Pulsus paradoxus

Definition of pulsus paradoxus

For whatever reason, pulsus paradoxus was a dominant feature of Question 11 from the second paper of 2015. The candidates were called upon to define it, and to describe the mechanism by which it occurs. Spodick (2003) defined it as 

"an inspiratory systolic fall in arterial pressure of 10 mm Hg or more during normal breathing" 

This is also the definition seen in such sources as Wikipedia, UpToDate and LITFL. It differs slightly from the way the college examiners described it in Question 11 :

"an exaggeration (> 12 mmHg or 10%) of the normal inspiratory decrease in systemic blood pressure"

Of course the CICM examiner's definition as definitive as the official Scrabble dictionary. But why is it different and where did they get it? Moreover,  how does a 2 mmHg difference in the definition influence management (or marking of your SAQ)? Well. It turns out, 12 mmHg and 9% systolic variation (not 10%) are the 95% confidence limits for diagnosis of moderate or severe tamponade (Curtiss et al, 1988). In the vast majority of patients with these findings, pericardiocentesis increased cardiac output by 50%. 10mmHg is a convenient rounding figure, described by Swami and Spodick (2003) as "a quasi arbitrary but practical level".

So, where is the paradoxus? Apparently, the paradox in the pulsus is the fact that during inspiration (of a spontaneously breathing patient) you can hear the heart sounds but can't feel the radial pulse. The term was coined by Adolf Kussmaul in 1873 on the basis of the observations he made in three patients with constrictive pericarditis (Wagner, 1973).  

Mechanism of pulsus paradoxus

According to the college examiners, the main mechanism of inspiratory pulsus paradoxus is related to ventricular interdependence.

• Normally, during inspiration, intrathoracic pressure decreases and the venous return to the heart increases
• This results in the RV filling with increased volume, and the interventricular septum bulges to the left. 
• The bulge of the septum into the LV results in a slightly decreased LV filling volume, and therefore slightly decreased systolic blood pressure (usually no more than 5-6 mmHg lower in inspiration)
• The reason for this pressure difference being so low is the ability of the left ventricle to expand in the pericardium, to compensate for the septal shift.
• In cardiac tamponade, the pericardium is too full and the LV cannot go anywhere. The inspiratory increase in RV filling results in such a bulge of the septum that the LV stroke volume is greatly diminished, with the resulting decrease in systolic blood pressure.

This importance of right ventricular venous return was demonstrated by Shabetai et al (1970), who found that pusus paradoxus disappeared when the right atrial pressure was kept constant during respiration. 

The college examiners also make allusions to other mechanisms in the confusing sentence, "the relatively higher negative pressure in the pulmonary circulation compared to the left atrium in patients with pericardial pathology pooling of blood in pulmonary veins during inspiration resulting in decreased LV stroke volume". This mechanism was well described by Ruskin et al (1973). "It is possible that pooling of blood in the pulmonary veins during inspiration may additionally decrease left ventricular filling and contribute to the paradoxical pulse seen in patients with pericardial tamponade", they wrote. Likely, this pooling occurs due to the decreased intrathoracic pressure in inspiration. The pressure in the thoracic cavity (and thus in the pulmonary circulation) ends up being lower than the pressure in the left atrium, which means that pulmonary venous blood does not feel compelled to flow into the left side of the heart.

Obviously all this is relevant only for the spontaneously breathing patient. In mechanical ventilation, everything is backwards. Positive pressure ventilation decreases preload to the RV, which means the ventricular bulge occurs during expiration. The resulting inspiratory increase in systolic blood pressure has been called  "reverse" pulsus paradoxus. 

Again, it is something which occurs in all mechanically ventilated patients, but to a lesser extent. The excellent diagram to the left is from the 1979 study by Möller et al. These  investigators got a hold of five Cape Chacma baboons (Papio ursinus) and then introduced saline into their pericardia. Saline was added until there was significant pulsus paradoxus and the arterial pressure was halved.The baboons were ventilated using a model 607 Harvard Animal Respirator, which generated a sine-wave inspiratory pressure pattern. The waveforms carefully recorded by the investigators clearly show that whenever pleural pressure was positive, the arterial systolic pressure was at its highest. 

Methods by which pulsus paradoxus can be elicited clinically 

In Question 11 from the second paper of 2015, the college wanted trainees to list several methods by means of which one may be able to elicit pulsus paradoxus. If the trainee did not come to the exam with a prefabricated list of such methods committed to memory, a zero mark was virtually guaranteed. The following list is an expansion of the college answer.

• Invasive arterial pressure trace: that's the classical ICU technique of demonstrating pulsus paradoxus, and is colloquially described as a "swing" of the arterial line.
• Palpation of the radial pulse:  the disappearance of the radial pulse on inspiration was the original sign described by Kussmaul.
• Sphygmomanometry: with the blood pressure measurement cuff inflated to the level of the systolic blood pressure, one ought to hear Korotkoff sounds.  Because the systolic blood pressure falls during spontaneous inspiration, the Korotkoff sounds disappear during inspiration. 
• Pulse oximetry is "particularly useful in paediatrics" according to the college examiners; they probably said this on the basis of a study by Tamburro et al (2002). The pulse oximeter waveform does something similar to the waveform of an arterial line, i.e "a decrease in the highest value of the upper plethysmographic peak of the pulse-oximetry waveform was observed during inspiration in each patient". Tamburro et al  observed this phenomenon in eight children and adolescents, which might give rise to the impression that this technique is "particularly useful in paediatrics". This might be in reference to the practical difficulties of using invasive blood pressure monitoring in children; otherwise the technique is probably equally useful in adults. 

Conditions which lead to the absence of a diagnostic pulsus paradoxus

Khasnis et al (2002) list the following conditions in which pulsus paradoxus will be absent even in cardiac tamponade:

• Aortic regurgitation: because the LV can fill from the aorta during inspiration
• Atrial septal defect: RV volume will be unchanged because increased RA volume will "spill over" into the left side of the circulation, equilbrating the pressures (.e. there will be no interventricular septal bulge).
• Isolated right heart tamponade (eg. following cardiac surgery)
• Raised LV diastolic pressure
• Chest wall immobility, eg. severe kyphoscoliosis preventing chest expansion

Electrocardiographic features of cardiac tamponade

The college asked for four electrocardiographic findings suggestive of pericarditis with cardiac tamponade, rather than tamponade on its own. Unlike virtually everything else in this chapter, this area can be answered with a single reference. The 1974 article by Howard Friedman describes the echocardiographic features of acute cardiac tamponade in great detail. Another more recent article (Eisenberg, 1996) describes their diagnostic value. These features are as follows are:

• Tachycardia is a nonspecific sign and does not merit any additional discussion
• Low QRS voltage trace - which develops as the result of a large volume of fluid in the way between the heart and the electodes, a fluid which has relatively poor conductivity. Not surprisingly, this feature is found more often in patients with large effusions. However, you do not have to have a large effusion to develop cardiac tamponade (less than 200ml could be fatal if i develops rapidly). Conversely, truly humongous effusions can be present without any tamponade physiology. Ergo, low QRS voltage is a sign with poor sensitivity. Eisenberg found it to have a sensitivity of just 25% for cardiac tamponade.
• Electrical alternans  is the presence of alternating high and low QRS complexes. LITFL has a nice example. It is usually associated with massive pericardial effusion, and is caused by the mechanical movement of the contracting heart, which is rhythmic and which changes its position within the dilated pericardial sack relative to the ECG electrodes. Alternans is merely the alternating high-low pattern of QRS complexes; there is also the phenomenon known as total alternans, where the QRS complexes are opposite (i.e. alternating up and down).  Therefore, it is associated with cardiac tamponade only insofar as massive effusion is associated with tamponade. Small effusions or the relatively immobile blood clot following cardiac surgery will not produce this effect. 
• Global concave ST elevation results from the current of injury which develops from direct pressure on the myocardium. It is rarely greater than 5mm. Badiger et al (2012) describes this very well, and offers discriminating features which help to distinguish it from myocardial infarction.
• PR depression - this is usually asociated with pericarditis, and because pericarditis is often associated with pericardial effusion the PR segments are often depressed in cardiac tamponade. Obviously, cardiac tamponade which is not due to pericarditis will probably have normal-looking PR segments. Eisenberg found that PR segments were unremarkable in more than half of all tamponade cases.  
• T wave inversion may develop as a result of pericardial irritation, but is by no means unique to cardiac tamponade. Pericarditis produces these changes, and they tend to resolve as the condition improves.

Echocardiographic features of cardiac tamponade

Pérez-Casares et al (2017) describes this topic very well. 

• A visible pericardial effusion is certainly a helpful finding but is by no means mandatory. Particular examples of tamponade without a significant effusion might include blood clot following cardiac surgery, where the clot does not present as a classical black echolucency you'd normally expect of a pericardial effusion.
• Diastolic collapse of right atrium and right ventricle: this happens when the intra-chamber pressures are at their lowest. In diastole, there will be a timer where the chamber pressures are actually lower than the pericardial fluid pressure. In this situation the chambers will collapse. Atrial collapse is usually seen before ventricular collapse
• Right atrial collapse in systole:  in early systole, atrial cavity pressure  is lower than the pericardial fluid pressure, and there is collapse of the thin free wall. Duration of this phenomenon is important: apparently, collapse for longer than one-third of the cardiac cycle is 100% specific for clinical cardiac tamponade
• Right ventricular collapse in diastole: during the early stages, this is only present in expiration when venous return is at its poorest. Again, the loger the duration of collapse, the more severe the tamponade.
• Diastolic ventricular size variability with respiratory cycle  is visually demonstrated using M-mode. Inspiration brings venous return to the RV and the RV dilates, pushing the septum into the LV. The opposite occurs in expiration. 
• Septal "bounce" is the colloquial-sounding name given to the inspiratory movement of the septum towards the LV. 
• IVC dilatation is seen because all the veins are dilated, and is essentially the echocardiographic equivalent of a raised JVP.
• Mitral flow is decreased on inspiration:  in cardiac tamponade the peak E-wave velocity is decreased by 25% on inspiration.
• Peak E-wave tricuspid valve physiological variation is larger than the mitral valve fluctuations -  in tamponade the peak E-wave velocity will drop by 40% in expiration compared to inspiration.
• RVOT/LVOT flow velocity fluctuation:  during normal respiration the physiologic variation of flow in these regions is less than 10%, but in tamponade the fluctuation is greater. During inspiration the aortic peak velocity will drop by 10%, and a rise of 10% will be seen in the pulmonary trunk.
• Hepatic vein flow reversal: in diastole, before atrial contraction the flow is either slowed or reversed. 
• Pulmonary vein flow reversal: again the flow is either slowed or reversed before the atria contract (usually there would be some flow reversal when the atria contract, which is perfectly normal physiological phenomenon)