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 [
9,
10,
11,
12,
13,
14,
15,
16,
17,
18,
19,
20,
21,
22,
23,
24,
25,
26,
27,
28,
29,
30,
31,
32]. 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 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.