A 68-year-old woman with a history of schizophrenia, severe coronary artery disease, hypertension, and type 2 diabetes mellitus was found in her bed minimally responsive by staff at the group home where she lived. She had been discharged from the hospital two days earlier with a diagnosis of segmental pulmonary embolism and on Coumadin anticoagulation. Lower extremity Doppler ultrasounds were negative for deep vein thrombosis during that hospitalization.
EMS brought her to the emergency department, and had intubated for airway protection. She was febrile, tachycardic, and hypotensive, and had a hemoglobin of 4 g/dL. An abdomen/pelvis CT revealed a large retroperitoneal hematoma. She was admitted to the ICU in critical condition. She was fluid resuscitated, and started on norepinephrine. A central venous catheter was placed, and the EKG strip shown was obtained.
Central venous pressure (CVP) reflects the right atrial pressure and by inference the right ventricular filling pressure. Using the Frank-Starling relationship, higher filling pressures are required to generate larger right ventricular stroke volumes or to maintain the same stroke volume in the setting of right ventricular dysfunction. Unfortunately, there is a tendency to concentrate only on this average value. A tracing of the central venous pressure shows a series of waves. If we understand the origins of these waves, close observation of the waveform can provide invaluable additional hemodynamic information.
It is helpful to evaluate a CVP tracing in conjunction with an ECG. The mechanical events during the cardiac cycle are responsible for the sequence of waves seen in the CVP. Coinciding with the ECG p wave and atrial contraction is an increase in right atrium pressure seen as the a wave. Blood leaves the atrium through the tricuspid valve into the ventricle, and the pressure in the atrium then decreases. This is denoted as the x descent on the waveform. The ventricles contract at the beginning of systole, marked by the QRS complex, and the tricuspid valve closes, causing a temporary interruption in the x descent seen as the c wave. The atrium begins filling in late systole by passive return from the vena cavae. The pressure peak, v wave, is released when relaxation of the ventricle occurs to the point that the tricuspid valve opens and blood from the atrium fills the ventricle. This decrease in pressure is designated the y descent. The waveform features are summarized in Table 1 and shown in Figures 1 and 2.
Table 1. Features of the CVP waveform.
Figure 1. Patient strip recording of leads II and V, CVP, and respirations.
Figure 2. Timing between the ECG and CVP waveforms.
It is important to remember that the CVP waveform reflects changes in right atrial pressure, not volume. Right atrial pressure rises because of increasing volume at a constant chamber stiffness (v wave) or increasing chamber stiffness at a constant volume (a wave). Venous return to the heart in contrast depends on pressure gradients between the periphery and the right atrium. Flow from the vena cavae into the atrium will be maximal during right atrial pressure troughs (x and y descent) and minimal during pressure peaks.
We are accustomed to using the CVP measurement to provide an estimate of volume status and right ventricular preload. While the actual driving pressure is the atrial transmural pressure — the difference between atrial pressure and intrathoracic/pericardial pressure — CVP measurements, are often relative to ambient air pressure. Pericardial and intrathoracic pressures are unknown and not practically measurable. End-expiratory values for cardiac filling pressures should be used to provide the best estimate of transmural pressure. At the end of expiration, intrathoracic and juxtacardiac pressures approach atmospheric pressure regardless of ventilatory status, and the CVP values will coincide. CVP measured relative to atmospheric pressure decreases during the inspiratory phase of spontaneous ventilation, but transmural CVP may actually increase slightly as more blood is drawn into the right atrium.
The opposite pattern is observed during positive-pressure ventilation, in which inspiration increases intrathoracic pressure while raising the measured CVP but decreases transmural CVP because the elevated intrathoracic pressure reduces venous return. Variations in the waveform, often with distinctive patterns, can give clues to underlying cardiac pathology as shown in Table 2.
Table 2. Changes to the CVP waveform.
The a and c waves in our critically ill patient can easily be identified along with the fairly prominent x descent. A clear v wave can be seen occurring in early diastole. The presence of a y descent can effectively rule out hemodynamically significant tamponade. (Figure 3.)
Figure 3. Simultaneous ECG and CVP waveforms.
When discussing pressure tracings, it is easy to confuse arterial and CVP waveforms. A strip recording from our patient shows the arterial and CVP waveforms. (Figure 4.) This recording was made after fluid resuscitation, and you will notice that the tachycardia has resolved. The shape of the CVP waveform is similar to that shown in Figure 3, and again the a, c, and v waves are identifiable.
Figure 4. Strip recording with lead II (top), lead V (second), arterial blood pressure (third) and CVP (bottom).
An arterial waveform is distinctly different, however. The timing of the upstroke of the arterial pulse is delayed about 200 ms from the QRS complex. A distinct dicrotic notch and small subsequent wave are seen at the end of systole. This has traditionally been attributed to the closure of the aortic valve, but it more likely represents a reflection wave of the original pulse against the high-impedance venules. The pressure then continues to decrease in diastole until the pressure rises again in the next systole. Our patient had a double peaked initial arterial pulse. This pulsus bisferiens (literally, striking twice) is classically taught to occur in aortic insufficiency, but is actually seen much more commonly in patients with high-output states such as sepsis and severe anemia in our patient.
Unfortunately for our patient, her condition worsened. She remained hypotensive with inadequate perfusion as demonstrated by increasing lactate levels, despite maximal hemodynamic support. Given her poor prognosis and severe comorbidities, her family elected to focus on her comfort. She was treated with palliative care, and allowed to have a natural death.