Arterial catheters should be inserted only when they are specifically required and removed immediately when no longer needed. Too often, they are left in place for convenience to allow easy access to blood sampling. Many studies have documented that arterial catheters are associated with an increased number of laboratory blood tests, leading to greater costs and excessive diagnostic blood loss [6,7]. Protocols incorporating guidelines for arterial catheterization and alternative noninvasive monitoring, such as pulse oximetry and end tidal CO₂ monitoring, have realized significant improvements in resource utilization and cost savings, without impacting the quality of care [8].

The indications for arterial cannulation can be grouped into four broad categories (Table 3-1): (a) hemodynamic monitoring (including pulse contour cardiac output monitoring); (b) frequent arterial blood gas sampling; (c) arterial administration of drugs, such as thrombolytics; and (d) intraaortic balloon pump use.

Noninvasive, indirect blood pressure measurements determined by auscultation of Korotkoff sounds distal to an occluding cuff (Riva-Rocci method) are generally accurate, although systolic readings are consistently lower compared to a simultaneous direct measurement. In hemodynamically unstable patients, however, indirect techniques may significantly underestimate blood pressure [9]. Automated noninvasive blood pressure measurement devices can also be inaccurate, particularly in rapidly changing situations, at the extremes of blood pressure, and in patients with dysrhythmias [10]. For these reasons, direct blood pressure monitoring is usually required for unstable patients. Rapid beat-to-beat changes can easily be monitored and appropriate therapeutic modalities initiated, and variations in individual pressure waveforms may prove diagnostic. Waveform inspection can rapidly diagnose electrocardiogram (ECG) lead disconnect, indicate the presence of aortic valve disease, help determine the effect of dysrhythmias on perfusion, and reveal the impact of the respiratory cycle on blood pressure (pulsus paradoxus). Additionally, the responsiveness of cardiac output to fluid boluses may be predicted by calculating the systolic pressure variation (SPV) or pulse pressure variation (PPV) through the respiratory cycle. Using PPV has been shown to be superior to SPV, with a PPV greater than 13% predicting an increase in cardiac output (CO) greater than or equal to 15% in response to a fluid challenge [11].

Recent advances allow continuous CO monitoring using arterial pulse contour analysis. This method relies on the assumption that the contour of the arterial pressure waveform is proportional to the stroke volume [12]. This, however, does not take into consideration the differing impedances between the arteries of individuals, and therefore requires calibration with another method of determining cardiac output [13]. Calibration is usually done with lithium or transpulmonary thermal dilution methods. A new pulse contour analysis instrument has been introduced that does not require an additional method of determining CO for calibration, but instead estimates impedance based on the waveform and patient demographic data [3]. This method has significant limitations (i.e., atrial fibrillation), and further data and comparison to other methods in authentic clinical situations are required before recommendations for widespread adoption can be made.

Management of patients in critical care units typically requires multiple laboratory determinations. Unstable patients on mechanical ventilators or in whom intubation is contemplated may need frequent monitoring of arterial blood gases. In these situations, arterial cannulation prevents repeated trauma by frequent arterial punctures and permits routine laboratory tests without multiple needle sticks. In our opinion, an arterial line for blood gas determination should be placed when a patient will require three or more measurements daily.

The equipment necessary to display and measure an arterial waveform includes: (a) an appropriate intravascular catheter; (b) fluid-filled noncompliant tubing with stopcocks; (c) transducer; (d) a constant flush device; and (e) electronic monitoring equipment. Using this equipment, intravascular pressure changes are transmitted through the hydraulic (fluid-filled) elements to the transducer, which converts mechanical displacement into a proportional electrical signal. The signal is amplified, processed, and displayed as a waveform by the monitor. Undistorted presentation of the arterial waveform is dependent on the performance of each component. A detailed discussion of relevant pressure monitoring principles is beyond the scope of this chapter, but consideration of a few basic concepts is useful to understand the genesis of most monitoring inaccuracies.

The major problems inherent to pressure monitoring with a catheter system are inadequate dynamic response, improper zeroing and zero drift, and improper transducer/monitor calibration [14]. Most physicians are aware of zeroing techniques but do not appreciate the importance of dynamic response in ensuring system fidelity. Catheter-tubing-transducer systems used for pressure monitoring can best be characterized as underdamped second-order dynamic systems with mechanical parameters of elasticity, mass, and friction [14]. Overall, the dynamic response of such a system is determined by its resonant frequency and damping coefficient (zeta). The resonant or natural frequency of a system is the frequency at which it oscillates when stimulated. When the frequency content of an input signal (i.e., pressure waveform) approaches the resonant frequency of a system, progressive amplification of the output signal occurs, a phenomenon known as ringing [15]. To ensure a flat frequency response (accurate recording across a spectrum of frequencies), the resonant frequency of a monitoring system should be at least 5 times higher than the highest frequency in the input signal [14]. Physiologic peripheral arterial waveforms have a fundamental frequency of 3 to 5 Hz, and therefore the resonant frequency of a system used to monitor arterial pressure should ideally be greater than 20 Hz to avoid ringing and systolic overshoot.

The system component most likely to cause amplification of a pressure waveform is the hydraulic element. A good hydraulic system will have a resonant frequency between 10 and 20 Hz, which may overlap with arterial pressure frequencies. Thus, amplification can occur, which may require damping to accurately reproduce the waveform [16].

The damping coefficient is a measure of how quickly an oscillating system comes to rest. A system with a high damping coefficient absorbs mechanical energy well (i.e., compliant tubing), causing a diminution in the transmitted waveform. Conversely, a system with a low damping coefficient results in under damping and systolic overshoot. Damping coefficient and resonant frequency together determine the dynamic response of a recording system. If the system’s resonant frequency is less than 7.5 Hz, the pressure waveform will be distorted no matter what the damping coefficient. On the other hand, a resonant frequency of 24 Hz allows a range in the damping coefficient of 0.15 to 1.1 without resultant distortion of the pressure waveform [14,17].

Although there are other techniques [18], the easiest method to test the damping coefficient and resonant frequency of a monitoring system is the fast-flush test (also known as the square wave test). This is performed at the bedside by briefly opening and closing the continuous flush device, which produces a square wave displacement on the monitor followed by a return to baseline, usually after a few smaller oscillations (Fig. 3-1). Values for the damping coefficient and resonant frequency can be computed by printing the wave on graph paper [14], but visual inspection is usually adequate to ensure a proper frequency response. An optimum fast-flush test results in one undershoot followed by small overshoot, then settles to the patient’s waveform (Fig. 3-1).

For peripheral pulse pressure monitoring, an adequate fast-flush test usually corresponds to a resonant frequency of 10 to 20 Hz coupled with a damping coefficient of 0.5 to 0.7 [17]. To ensure the continuing fidelity of a monitoring system, dynamic response validation by fast-flush test should be performed frequently: at least every 8 hours, with every significant change in patient hemodynamic status, after each opening of the system (zeroing, blood sampling, tubing change), and whenever the waveform appears damped.

With consideration of the above concepts, components of the monitoring system are designed to optimize the frequency response of the entire system. The 18- and 20-gauge catheters used to gain vascular access are not a major source of distortion but can become kinked or occluded by thrombus, resulting in overdamping of the system. Standard, noncompliant tubing is provided with most disposable transducer kits and should be as short as possible to minimize signal amplification [15]. Air bubbles in the tubing and connecting stopcocks are a notorious source of overdamping of the tracing and can be cleared by flushing through a stopcock. Currently available disposable transducers incorporate microchip technology, are very reliable, and have relatively high resonant frequencies [19]. The transducer is attached to the electronic monitoring equipment by a cable. Modern monitors have internal calibration, filter artifacts, and print the display on request. The digital readout display is usually an average of values over time and therefore does not accurately represent beat-to-beat variability. Monitors provide the capability to freeze a display with on-screen calibration to measure beat-to-beat differences in amplitude precisely. This allows measurement of the effect of ectopic beats on blood pressure, PPV, SPV, or assessment of the severity of pulsus paradoxus.

When presented with pressure data or readings believed to be inaccurate, or which are significantly different from indirect readings, a few quick checks can ensure system accuracy. Improper zeroing of the system is the single most important source of error. Zeroing can be checked by opening the transducer stopcock to air and aligning with the midaxillary line, confirming that the monitor displays zero. Zeroing should be repeated with patient position changes, when significant changes in blood pressure occur, and routinely every 6 to 8 hours because of zero drift. Disposable pressure transducers incorporate semiconductor technology and are very small, yet rugged and reliable, and due to standardization, calibration of the system is not necessary [19]. Transducers are faulty on occasion, however, and calibration may be checked by attaching a mercury manometer to the stopcock and applying 100, 150, and/or 200 mm Hg pressure. A variation of plus or minus 5 mm Hg is acceptable. If calibration is questioned and the variation is out of range, or a manometer is not available for testing, the transducer should be replaced.

If zero referencing and calibration are correct, a fast-flush test will assess the system’s dynamic response. Overdamped tracings are usually caused by air bubbles, kinks, clot formation, overly compliant tubing, loose connections, a deflated pressure bag, or anatomical factors affecting the catheter. All of these are usually correctable and should be assessed if the system is overdamped. An underdamped tracing results in systolic overshoot and can be secondary to excessive tubing length or patient factors such as increased inotropic or chronotropic state. Many monitors can be adjusted to filter out frequencies above a certain limit, which can eliminate frequencies in the input signal that causes ringing. However, this may also cause inaccurate readings if important frequencies are excluded.