Placement and management of arterial catheters

Nagarajan Ramakrishnan, MBBS Gilles Clermont, MDCM, MSc

Jan 4, 1999

Arterial cannulation, a commonly performed procedure in the intensive care unit, simplifies the monitoring and management of many critically ill patients. Although the procedure has been performed for over 60 years, there have been significant technological advances in the monitoring system. These include internal calibration systems, better filtering capabilities, and visually pleasing color displays. Transparent and disposable solid-state systems have also replaced troublesome wire elements in transducers.

An arterial line may be used for the following purposes:

  •  Monitoring of blood pressure in the setting of hypotension or hypertension, particularly if the problem is acute or the blood pressure is labile
  •  Monitoring response to vasoactive drugs
  •  Obtaining frequent arterial specimens for blood gas determination
  •  Obtaining frequent blood specimens in lieu of venipuncture
  •  Emergent administration of medications when venous access cannot be obtained
  •  Intraaortic balloon pump use

TECHNIQUE — The required equipment for arterial cannulation is listed in Table 1 (show table 1). Commonly used sites are the radial, femoral, axillary, dorsalis pedis, brachial, and, in children, temporal and umbilical arteries. Although each site has unique complications, currently available data do not support the superiority of any particular site [1]. The traditional bias against femoral artery cannulation is contradicted by multiple studies demonstrating a complication rate no higher, and in some studies lower, than radial artery cannulation [2]. One study comparing medical and surgical intensive care units (ICU) concluded that the preferred site for arterial cannulation in medical ICUs is femoral and in surgical ICUs is radial. In this study, the site of arterial catheter placement, and the timing and number of catheter/site changes made no significant difference in terms of complications [3].

The anatomy and landmarks of the radial, brachial, femoral, and dorsal pedis arteries are shown in Figures 1 through 4, respectively (show figure 1A-1B, show figure 2A-2B, show figure 3A-3B, and show figure 4A-4B).

When placing a radial catheter, the radial artery is palpated between the distal radius and the tendon of the flexor carpi radialis; it can usually be more easily accessed with the wrist extended. The modified Allen's test should be performed to demonstrate collateral flow through the superficial palmar arch prior to cannulation (show figure 5).

Direct puncture or guidewire approaches may be employed for arterial cannulation. One study of 69 critically ill patients reported that guidewire techniques yielded a lower failure rate, fewer passes, and more rapid cannulation than the direct-puncture method [4]. The guidewire technique is depicted in Figure 6 (show figure 6). We recommend the guidewire technique during training years and also if there are difficulties encountered with the direct puncture technique.

Successful cannulation is more likely in males (because of larger vessel size), and in patients with bounding pulses [5]. Devices which incorporate a disposable ultrasound transducer within the lumen of a thin-wall vascular needle (Smart Needle) are used infrequently, but may be helpful in certain situations. As an example, hypotensive patients with nonpalpable pulses may be difficult to cannulate unless instruments or techniques are employed which use Doppler ultrasonography to localize the artery [6].

COMPLICATIONS — Arterial cannulation is a relatively safe procedure, but is not without complications. Regardless of site, the most common complications of arterial lines are:

  •  Bleeding
  •  Distal limb ischemia from thrombosis or thromboembolism
  •  Air embolism
  •  Catheter-related infection
  •  Pain and swelling
  •  Pseudoaneurysm
  •  Arteriovenous fistula

Additional site specific complications are listed in Table 2 (show table 2).

Thrombosis — Thrombosis can be detected by Doppler ultrasonography in 5 to 25 percent of patients following arterial line insertion, but clinically significant thrombosis occurs in less than 1 percent [7]. The incidence of thrombosis increases with longer duration of cannulation and larger catheter diameters, and is more common in radial and dorsalis pedis arteries because of their smaller size [8].

Continuous or intermittent heparin flush systems have significantly reduced the incidence of catheter-related thrombosis [9]. Continuous flush systems deliver 3 to 15 units of heparin in approximately 3 mL per hour from an infusion bag pressurized to 300 mmHg. A placebo-controlled, randomized trial of 40 ICU patients reported that a 1.4 percent solution of sodium citrate is as effective and safe as heparin, and may be useful when heparin-associated thrombocytopenia is suspected [10]. (See "Clinical use of heparin and low molecular weight heparin"). Saline solutions are used to maintain patency in some centers, but may result in higher catheter failure rates when compared with heparin [11].

Infection — Catheter-related infections are discussed in depth separately. (See "Prevention of intravascular catheter-associated infections"). Reviewed briefly, there is a lower infection rate with arterial than venous catheters, and colonized arterial catheters result less often in bacteremia than colonized venous catheters. The reported incidence of arterial catheter colonization varies from 5 to 10 percent, and appears similar for radial and femoral lines [12,13,14].

The risk for bacterial colonization of the catheter increases with the duration of cannulation, rising to approximately 17 percent after six days [15,16,17]. Nevertheless, scheduled changes of arterial lines have not been documented to result in fewer infections or better patient outcomes. Guidewire exchange of arterial catheters through the same insertion site has not been well studied and is not recommended.

Controversy exists regarding the optimal interval at which arterial line tubing and flush systems should be changed. While some authors recommend scheduled replacement of such equipment every 24 to 96 hours to minimize the opportunity for bacterial contamination and growth, others believe that this practice is unnecessary or may be harmful. One study of 333 monitoring kits and lines found that four out of six documented catheter infections occurred within 48 hours after a scheduled flush bag change [18]. The investigators hypothesized that changes of tubing may predispose to infection because they require violations of the closed monitoring system. Our approach is summarized in Table 3 (show table 3).

Air embolism — Air embolism can occur when the flush solution contains gas. Because the infusate is under a pressure of 300 mmHg, flushing may force air bubbles in a retrograde direction through the radial, brachial, axillary, and subclavian arteries, where they can then travel to the cerebral circulation via the vertebral arteries. In a primate model, as little as 2 mL of air injected into the radial artery with a standard pressurized infusion apparatus can result in significant cerebral air emboli [19]. Such emboli are more likely when patients are smaller and sitting upright. (See "Air Embolism").

Air introduced via arterial lines bypasses the pulmonary capillaries and thus may cause more severe sequelae than air which is embolized through a peripheral venous catheter. The latter occurs fairly commonly but rarely is clinically apparent unless an anatomic right to left shunt is present. As an example, one study performed electron-beam computed tomography of the chest on 208 patients following insertion of a peripheral venous catheter and detected small air emboli in 5 percent, most commonly in the larger pulmonary arteries [20]. All patients were asymptomatic. (See "Electron beam (ultrafast) computed tomography for the evaluation of cardiac disease and function").

Diagnostic blood loss — Prior to obtaining a blood specimen for laboratory analysis, 3 to 12 mL of blood typically is wasted in order to avoid saline contamination. If frequent sampling is required, this practice can result in substantial blood loss and necessitate transfusion [21,22].

The degree of diagnostic blood loss can be minimized by using a proximal sampling port so that heparin-free blood can be obtained without wastage [23]. Patients who require frequent arterial blood gases may also benefit from the use of intraarterial blood gas monitoring. This is a new, fluorescent optode-based technique which can perform arterial pH, PCO2, and PO2 measurements as frequently as clinically required without violating the integrity of the arterial catheter tubing system, or removing blood from the patient [24].

QUALITY OF INVASIVE AND NONINVASIVE BLOOD PRESSURE DATA — Direct measurement by an arterial line is the gold standard for determining blood pressure, provided the pressure transducer is free of technical problems (see below). Values of mean arterial pressure determined by direct measurement and sphygmomanometry generally correlate well in healthy patients. (See "Technique of blood pressure measurement in the diagnosis of hypertension").

Both manual and automated cuff pressures appear less reliable in hemodynamically unstable patients. As an example, one study of 26 critically-ill patients found that indirect techniques underestimated systolic pressure by an average of 17 mmHg, and overestimated diastolic pressure by 3 to 5 mmHg [25]. Cuff pressures lose accuracy in the presence of shock, arrhythmias, vasoconstrictor drugs, or calcified arteries [26,27].

Noninvasive finger artery blood pressure monitoring correlates well with radial artery blood pressure monitoring in various ambulatory settings [28]. However, several small studies suggest that such devices have limited utility when hemodynamic instability is present, and should not be relied upon in settings where this is likely to occur [29,30].

TECHNICAL SOURCES OF BLOOD PRESSURE MEASUREMENT ERROR — Despite the overall superiority of direct blood pressure measurement methods, technical problems may result in systematic errors. The major problems inherent to pressure monitoring with a catheter system are dynamic response, zeroing drift, and transducer/monitor calibration [31].

Dynamic response — Dynamic response is determined by two factors, the resonant frequency and the damping coefficient:

  •  The resonant or natural frequency of the system is the frequency at which it oscillates when stimulated. Physiologic peripheral arterial waveforms have a fundamental frequency of 3 to 5 Hz, although some components may range up to 20 Hz [32]. Thus, the resonant frequency of the system used to monitor arterial pressure must be greater than 20 Hz to avoid ringing and systolic overshoot [33]

  •  The damping coefficient is a measure of how quickly an oscillating system comes to rest [33]. A system with a high damping coefficient (eg, compliant tubing) absorbs mechanical energy well and causes a diminution in the transmitted waveform.

The damping coefficient and resonant frequency of a monitoring system can be assessed at the bedside by the fast-flush test, performed by briefly opening and closing the valve in the continuous flush device (show figure 7). This produces a square wave displacement on the oscilloscope, followed by ringing and a return to baseline. Connecting tubing with stopcocks, excessive tubing lengths, and patient factors (such as tachycardia and high output states) are common causes of underdamping. These factors should be systematically addressed. On the other hand, air bubbles in the tubing are a common source of overdamping; the bubbles can be cleared by flushing the system through the stopcock.

Transducer position — The transducer position can be checked (the transducer can be "zeroed") by turning the stopcock off to the patient and open to air, aligning the stopcock with the level of the heart (usually approximated with the mid-axillary line), and confirming that the monitor displays zero. This procedure should be repeated when patient position is changed, when significant changes in blood pressure are noted, and routinely every six to eight hours [34].

Monitor and transducer calibration — Routine calibration of the monitor and transducer is no longer necessary because currently used, disposable transducers are standardized [35].