Intended for healthcare professionals


Arterial versus capillary sampling for analysing blood gas pressures

BMJ 1995; 310 doi: (Published 07 January 1995) Cite this as: BMJ 1995;310:24
  1. Khavar Dar, medical registrara,
  2. Tim Williams, consultant physiciana,
  3. Richard Aitken, top grade biochemista,
  4. Kent L Woods, senior lecturerb,
  5. Susan Fletcher, research assistantb
  1. a Department of Medicine and Clinical Biochemistry, Kettering General Hospital NHS Trust, Kettering NN16 8UZ
  2. b Department of Medicine and Therapeutics, University of Leicester, Leicester
  1. Correspondence to: Dr Williams.
  • Accepted 17 October 1994

Arterial puncture is carried out to obtain samples for analysis of blood gas pressures. Although painful1 and not without hazard,2 arterial puncture is done routinely despite reports that similar information about blood gas pressures can be obtained from capillary samples.3 We quantified patients' perception of the discomfort of sampling from radial arteries compared with that of capillary sampling and compared the biochemical data obtained with these two methods.

Methods and results

We recruited 55 patients requiring measurement of blood gas pressures from patients admitted urgently to Kettering General Hospital. The study's protocol was approved by the district's ethics committee, and informed written consent was obtained for the blood sampling. The clinical diagnoses were exacerbation of chronic obstructive airways disease (22 patients), asthma (10), pneumonia (5), pulmonary oedema (4), pulmonary embolism (4), diabetic ketoacidosis (4), bronchiectasis (3), fibrosing alveolitis (2), hyperventilation (1).

We took a capillary and arterial sample in random order from each patient. For capillary samples we rubbed a proprietary rubefacient on the ear lobe and after about three minutes made a stab about 1–2 mm deep with the point of a size 11 scalpel in the fleshy part of the ear lobe. We collected the blood (which had become more like arterial blood because of the rubefacient) in a heparinised capillary tube held horizontally against the puncture site. A small metal rod was placed in each tube, and after both ends had been sealed with plastic plugs the sample was agitated with a magnet. We took arterial samples from the radial artery using a 22 gauge needle. For the first 29 patients no anaesthetic was used. For the next 26, 1% lignocaine was infiltrated over the radial artery before puncture.

As soon as a patient's second blood sample had been taken, the patient completed a simple questionnaire, rating the discomfort associated with each procedure on a scale from 0 (“no pain at all”) to 10 (“worst pain imaginable”) and stating which of the two techniques he or she preferred. The pain scores were analysed with non-parametric tests.

Blood gas pressures and acid-base state in the arterial samples were compared with those in the capillary samples with the paired t test.

Pain scores and preferences of blood sampling method of 55 patients whose blood was collected with both arterial and capillary puncture

View this table:

The table shows the results of the patients' pain scores and preferences of sampling technique. Arterial puncture without local anaesthetic was much more painful than capillary puncture; most patients preferred the capillary technique. Even with local anaesthesia arterial puncture was significantly more painful, with half the patients preferring the capillary technique; only two preferred the arterial technique. Mean differences (arterial minus capillary) were trivial and non-significant for partial pressure of oxygen (0.09 kPa, SD 0.59, P>0.2) and partial pressure of carbon dioxide (0.01 kPa, 0.3, P>0.7 and significant but clinically unimportant for acidity (0.007 pH, 0.02, P<0.01) and standard bicarbonate (0.64 mmol/l, 1.17, P<0.001).


Our findings indicate that radial arterial blood sampling is painful, and even with local anaesthesia the procedure causes more pain than sampling from the ear lobe. The pain scores for capillary sampling may have been lower in the patients given local anaesthetic because of a “carry over” effect—that is, patients who did not receive local anaesthetic found the whole process considerably more uncomfortable and marked up the pain scores for capillary sampling as an indirect consequence of this. In addition, a small but definite risk exists of local complications from arterial puncture.2 As similar biochemical data are obtained from capillary sampling from the ear lobe this method should be used in routine clinical practice. The method of sampling may be particularly important to patients with recurrent respiratory illness such as asthma. Several studies have shown delay to be an important factor contributing to deaths from asthma.4 5 Reluctance to present to hospital because of a previous painful experience could be catastrophic.

We thank Dr James Falconer-Smith for helpful discussion and Mrs Helen James for typing the manuscript.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.