Ideally, prior to obtaining any ABG sample (Figure 13.12), the patient should be in a respiratory steady state for 15–20 minutes. The oxygen should not be reduced, and, if the patient is ventilated, suction should not be attempted and ventilation parameters should not be adjusted (Keogh [114], Pagana and Pagana [177]). Arterial sampling and cannulation are associated with a considerable amount of pain due to adjacent nerves and the need for deeper puncture than venous sampling or cannulation. ABG sampling may be difficult to perform in patients who are unco‐operative or in whom pulses cannot be easily identified (Danckers and Fried [41]).

Pain can be reduced by using local anaesthetic, which will also allow the healthcare professional to perform sampling without the patient moving their arm too much. In an emergency, there may be no time for local anaesthetic application. However, when it is feasible to wait for local anaesthetic to take effect before attempting arterial puncture, studies suggest higher success rates for arterial sampling (McSwain and Yeager [140]).

Despite its controversial value in predicting post‐procedure ischaemia due to a radial arterial thrombosis, the modified Allen test (Figure 13.11) is still widely used and should be performed prior to accessing the radial artery to confirm patent collateral circulation to the hand. Adjuncts to the Allen test (e.g. pulse oximetry) have not been shown to increase the diagnostic power of the test (Pagana and Pagana [177]). The modified Allen test is performed by firmly compressing the radial and ulnar arteries while the patient clenches their fist (Step 1). The patient is then asked to open their hand (Step 2) and the arteries are released one at a time to check their ability to return blood flow to the hand, turning it pink or flushing again (Step 3) (Pagana and Pagana [177]). If blood flow to the hand returns within 15 seconds after release of only the ulnar artery (flushing), it is presumed the radial artery can be used for sampling. If after 15 seconds there is no blood flow returning to the hand (no flushing), the radial artery cannot be used (Pagana and Pagana [177]).

Where a sample is taken from an established arterial cannula, certain volumes of blood should be withdrawn and discarded. This is to ensure that any 0.9% sodium chloride, blood and small emboli from the dead space within the cannula and the three‐way tap have been removed. Discarded volumes will vary according to arterial cannula placement due to the length of the arterial cannula:

When using a vacuum system, take one bottle and discard prior to obtaining requested samples.

ABG sampling may be difficult to perform in patients who are unco‐operative or in whom pulses cannot be easily identified. Challenges arise when healthcare personnel are unable to position the patient properly for the procedure. This situation is commonly seen in patients with cognitive impairment, advanced degenerative joint disease or essential tremor. The amount of subcutaneous fat in overweight and obese patients may limit access to the vascular area and obscure anatomical landmarks. Arteriosclerosis of peripheral arteries, as seen in elderly patients and those with end‐stage kidney disease, may cause increased rigidity in the vessel wall (Danckers and Fried [41]).

Various types of blood gas syringe are available: some are pre‐heparinized, whereas some are vented and self‐fill on arterial blood sampling. If an intra‐arterial cannula is inserted in a critical care environment then a sodium chloride 0.9% infusion is required to keep the cannula open (NPSA [168]). If continuous arterial monitoring is done via a cannula, further equipment is required, including a transducer connected to a cardiac monitor and pressure bag set normally at 300 mmHg for arterial lines (Osborne [174]).