Clinical review: complications and risk factors of peripheral arterial catheters used for haemodynamic monitoring in anaesthesia and intensive care medicine.

In order to evaluate the complications and risk factors associated with peripheral arterial catheters used for haemodynamic monitoring, we reviewed the literature published from 1978 to 2001. We closely examined the three most commonly used arterial cannulation sites. The reviewed papers included a total of 19,617 radial, 3899 femoral and 1989 axillary artery catheterizations. Factors that contribute to higher complication rates were investigated. Major complications occurred in fewer than 1% of the cases, and rates were similar for the radial, femoral and axillary arteries. We conclude that arterial cannulation is a safe procedure.

Indwelling arterial catheters are used routinely for continuous haemodynamic monitoring in the operating room during major surgery and in critically ill patients. Arterial cannulation provides easy and convenient access, allowing multiple blood samplings and blood gas analysis. Arterial cannulation has generally been found to be a safe procedure, with few serious complications [1][2][3][4][5][6]. Approximately 8 million and 2.5 million arterial catheters are placed yearly in the USA and Europe, respectively [7]. Despite the frequency with which arterial catheters are employed, there is little information on the impact of cannulation site on risk for complications [8]. We reviewed reports concerning arterial cannulation sites such as the radial, femoral, axillary, brachial, ulnar, dorsal pedis, tibial posterior and temporal arteries with regard to thrombotic, infectious and vascular complications.

Material and methods
The PubMed, Medline Express and Winspirs databases (publication years 1978-2001) were searched to identify pertinent articles. The keywords 'artery catheter', 'radial artery catheter', 'femoral artery catheter', 'axillary artery catheter', 'catheter complication', 'hemodynamic monitoring' and 'peripheral methods' were used. We reviewed the papers and identified further articles from the references of the papers found in the initial search.
We reviewed all studies concerning the use of radial, femoral, axillary, brachial, ulnar, dorsal pedis, tibial posterior and temporal artery catheters for haemodynamic monitoring that dealt with complications and risk factors in adults. We excluded studies conducted in the paediatric patient population.
We opted not to perform statistical analyses because the data selection of the published reports do not follow the same criteria, and most of them give no specific information regarding patient population, catheter material, cannulation technique and times. Therefore, comparison between studies and statistical analyses are unlikely to be valid.

Results
We identified 78 studies that met our criteria.

Radial artery
The radial artery is the most common site for arterial cannulation, and the studies that reported complication rates are listed in Table 1  . The most common complication was temporary occlusion of the artery, the incidence of which ranged from 1.5% [17] to 35% [10] (mean 19.7%), the variability being due to the different methods employed and the degree of effort invested in detecting this complication.

Clinical review: Complications and risk factors of peripheral arterial catheters used for haemodynamic monitoring in anaesthesia and intensive care medicine
Generally, temporary occlusion of the artery has no serious sequelae. Permanent occlusion of the radial artery appears to be rare because it was reported in only four patients (mean incidence 0.09%).
Another serious complication is pseudoaneurysm, which was reported in 14 patients (mean incidence 0.09%). Pseudoaneurysm poses a risk for infection, sepsis, rupture [33][34][35] and formation of an extracorporeal pseudoaneurysm [36]. Radial catheterization was associated with sepsis in eight patients (mean incidence 0.13%), whereas local infection at the cannulation site was identified in 45 patients (mean incidence 0.72%).
Pseudoaneurysm formation occurred in six patients (mean incidence 0.3%), sepsis was observed in 13 patients (mean incidence 0.44%) and local infection was reported in five patients (mean incidence 0.78%). Bleeding (generally minor) was observed in five patients (mean incidence 1.58%), and haematoma formation was reported in 28 (mean incidence 6.1%). One patient developed an infected haematoma and needed blood transfusion [17] and another patient eventually died from massive retroperitoneal bleeding [57].

Axillary artery
The axillary artery was cannulated in a total of 1989 reported cases. The complication rates at this site are summarized in Table 3  Cases 333 permanent ischaemic damage in two patients (mean incidence 0.20%), pseudoaneurysm formation in one patient (mean incidence 0.1%) and sepsis in five patients (mean incidence 0.51%). Paresthesia of the hand due to pressure on the brachial nerve plexus was also described [54,59].

Other sites for arterial cannulation
Other less frequently reported cannulation sites are the brachial, dorsal pedis, ulnar, tibial posterior and temporal arteries.
Only one serious complication was found in a study of 1000 patients [62] in which the brachial artery was used for invasive monitoring in ambulatory patients. This complication was an infected haematoma arising from a pseudoaneurysm. Another study that employed the brachial artery for arterial blood sampling in 6185 patients [63] also showed a small number of complications (incidence 0.2%), mainly paresthesias.
The dorsal pedis artery has also been cannulated without any major complications [24,54]. However, because of its distance from the central circulation and associated difficulties in hypotensive patients it is generally avoided [64]. The ulnar artery has been used without serious complications, and in a direct comparison [29] this site exhibited similar complication rates to those encountered with the radial artery.
The tibial posterior artery has been used in paediatric patients, without major complications [65,66]. However, there also exists a case report of limb amputation secondary to a tibial artery catheterization [67]. The temporal artery is an end artery of the external carotid artery, and is usually avoided for fear of serious complications. There is a report of three cases of cerebral embolization in paediatric patients [68].

Further studies
Other studies that examined arterial catheter complications but that did not specifically attribute their occurrence to one specific site are listed in Table 4 [49,[69][70][71]. One study [69] used mainly the radial (n = 210), ulnar (n = 158) and dorsalis pedis arteries (n = 82), whereas another [49] used mainly the femoral (n = 46), brachial (n = 9), axillary (n = 3) and radial arteries (n = 1). One serious complication was found at the femoral site, with compromised circulation to the leg [70]; eventually, thrombectomy was required. Another reported serious complication is air embolism caused by improper use of the flushing device [70].
A recent study [71] examined complication rates, comparing medical with surgical intensive care units. It also directly compared radial and femoral artery catheters. There were no significant differences in terms of sepsis and local infection rates between the femoral and radial arteries, and no significant difference between the medical and surgical intensive care units.
Another nine studies specifically looked for septic complications associated with arterial catheter placement, and included a total number of 1681 patients (Table 5) [52,54,[72][73][74][75][76][77][78]. No septic complications were identified in seven of those studies [52,[72][73][74][75][76][77], and nine cases of sepsis were reported in the remaining two studies [54,78]. All of the cases of sepsis were identified after the line had been in place for longer than 4 days. The studies in which sepsis was reported predominately used the radial artery.

Risk factors
A further objective of the present review was to consider risk factors that might have impacted on complication rates.

Thrombotic complications
Thrombus formation and occlusion of the artery appear to be caused by changes in the integrity of the vessel wall induced by the presence of the catheter [79]. Recannulation of the occluded artery generally occurs, but this may take up to 75 days [80]. The incidence of thrombus appears to be related to the degree to which the catheter fills the arterial lumen [3]. It has been shown that the incidence of radial obstruction increases linearly with the ratio of outer diameter to vessel lumen diameter [81]. This might explain the higher incidence of radial occlusion in female patients, who generally have vessels of smaller diameter [5,10,14,22,25]. This has led to a preference for 20-gauge catheters for radial artery cannulation [11,14]. Teflon catheters also appear to be associated with a lower incidence of occlusion [11], but in another study [5] this could not be confirmed. Patients with a low cardiac output have a higher incidence of radial artery occlusion [5,14], but no difference was found among patients with episodes of hypotension [21]. Multiple punctures for catheter insertion was shown to be a risk factor for complications in one study [14], but this could not be confirmed in another [5]. A higher incidence of occlusion is found in the presence of a haematoma [5,14,21]. Patients who are treated with aspirin or low-dose heparin have fewer vessel occlusions [13,14]. The age of adult patients does not appear to be a risk factor [14]. An important factor with regard to arterial occlusion is the duration of cannulation. A higher incidence of occlusion was reported when the cannula was left in place longer than 48 or 72 hours [5,9,14,16,25].

Infectious complications
Important factors in decreasing the incidence of catheterrelated infections are aseptic insertion technique and adequate disinfection of the insertion site [82]. The risk for infectious complications and sepsis increases with the duration of cannulation, especially if the catheter is left in place for longer than 96 hours [78,83]. The catheter material also appears to be of importance, because Teflon has a greater resistance to Escherichia coli and Pseudomonas aeruginosa [84] and a lower affinity for stapylococci [85]. Prophylactic treatment with antibiotics does not appear to reduce the risk for catheter-related infections [78,83], but antibiotic bonding of the catheter may confer a degree of protection [86,87]. Arterial catheter placement via the cut-down technique is associated with a higher risk for infection [83]. Furthermore, sepsis occurs more frequently in the presence of local site inflammation [83]. Infection may also occur through contamination by caregivers of the monitoring system, the flushing device and the infusion system [88][89][90].

Discussion
The use of peripheral arterial catheters for haemodynamic monitoring is widespread. The most frequently used site is the radial artery because of its well documented low complication rates and easy access . In the present review we report on 19,617 radial artery cannulations, and the main complication was temporary occlusion of the artery (incidence 19.70%). This complication rate is close to the finding of Wilkins [91], who reviewed reports of catheterization of the radial artery with regard to ischaemic complications published between 1969 and1983. He reported a temporary occlusion rate of 23%. In an extensive study of 1699 patients, Slogoff and coworkers [5] reported a 21.2% rate of temporary occlusion. Although temporary occlusion is reported quite frequently, serious ischaemic damage was reported in only two studies [9,16], with a mean complication rate of 0.09%. Nevertheless, there have been reports of serious ischaemic damage after radial artery cannulation that led to necrosis, and amputation of fingers or the whole hand [92][93][94][95][96][97][98][99].
Other major complications such as pseudoaneurysm and sepsis were reported to occur in a mean of 0.09% and 0.13% of cases, respectively.
The second most cannulated artery for haemodynamic monitoring is the femoral artery, and we reviewed 3899 cases. Temporary occlusion was reported in only 1.18% of the reviewed cases. The incidence of this complication in the femoral artery is much lower than that in the radial artery. We attributed this to the larger vessel diameter of the femoral Available online http://ccforum.com/content/6/3/198  artery. Serious ischaemic complications were reported in only one study [56], and the mean complication rate was 0.18%. The incidence rates for other major complications such as pseudoaneurysm and sepsis (0.3% and 0.44%, respectively) were similar to our findings in the radial artery. Some authors caution against use of the femoral artery for cannulation for fear of higher sepsis rates because of the close proximity to the perianal region [52]. We are unable to corroborate this on the basis of the reviewed literature, and recent studies that directly compared the radial and femoral arteries with regard to septic complications identified similar rates of sepsis [53,71]. In particular, in an extensive study of 4932 patients, Frezza and Mezghebe [71] were unable to identify a difference in complication rates between the radial and femoral arteries. In that study the femoral artery was actually the preferred site in the medical intensive care unit. Other authors prefer the femoral artery over the radial artery because the femoral artery is usually palpable even in hypotensive patients and may be the only accessible route for haemodynamic monitoring [64,[100][101][102][103]. The blood pressure curve that is obtained from this larger artery is generally more accurate and gives a closer estimation of the aortic blood pressure [64,[100][101][102][103].
The third most cannulated artery in the present review of the literature was the axillary artery, with almost 2000 reported cases. Some authors prefer not to cannulate the axillary artery because of its close location to the carotid artery and because of fear of embolism to the brain [101]. On the basis of the data summarized here we cannot confirm this, and no case of embolism to the brain was reported. Sometimes the axillary artery is avoided because of the more difficult approach required, although (particularly in anaesthesia) it is a well known route because it requires the same approach as that for axillary nerve block [3]. The major complications encountered with the axillary artery were similar to those for the radial and femoral arteries, and we conclude that it is a safe route for arterial cannulation.
Other reviewed arteries employed for catheterization, such as the brachial, dorsal pedis, ulnar, tibial and temporal arteries, have been used without serious complications, but published reports of their use are limited.

Conclusion
Incidence rates for major complications such as permanent ischaemic damage, sepsis and pseudoaneurysm formation are low and similar for the radial, femoral and axillary arteries. They occur in fewer than 1% of cases.
On the basis of the present systematic review we can conclude that serious complications of the radial, femoral and axillary artery are rare, and that arterial cannulation is a relatively safe procedure. Nevertheless, one must be aware of the possible serious complications that might arise. Therefore, it is important to use an arterial catheter only when it is indicated and to select carefully the best cannulation site individually for each patient, because the different cannulation sites have their specific advantages and attendant risks.

Competing interests
UJP is President of Pulsion Medical Systems AG, and AP is on the medical advisory board of Pulsion Medical Systems AG, whose PiCCO product uses a femoral arterial line.