Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Capillary blood gases as an alternative to arterial puncture in shock

Three Part Question

In [shocked patients] can [capillary blood be used as an alternative to arterial puncture] for [guiding appropriate therapy]?

Clinical Scenario

A patient is brought to the Emergency Department with hypovolaemic shock. It is difficult to obtain arterial blood gases, so you wonder if capillary blood can be used as an alternative to arterial puncture.

Search Strategy

Medline using the OVID interface 1966-04/2004
[exp capillaries/ OR] AND [exp blood gas analysis/ OR blood] AND [arterial blood] LIMIT TO [human AND (children 2+ years and adults)].

Search Outcome

271 papers were found of which 1 was relevant to the question (reference 2 below). In addition, 2 further papers were discovered from scrutiny of the references of articles retrieved from the Medline search (references 1 & 3).

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Cooper EA
27 samples on 10 unselected patients undergoing major abdominal surgery. 21 samples taken from patients in normal state, 9 from shocked patients.Non-randomised, uncontrolled clinical trial.Difference between simultaneously drawn arterial blood (taken from the intra-abdominal arteries by the surgeon) and capillary blood taken from the earlobe (after it had been rubbed with histamine cream for 1 min then flicked for 30s).Normal: Capillary pCO2 +/-2mmHg of arterial (18-40mmHg). Shocked: Capillary samples 4-13mmHg higher (25-35mmHg)1. Small study 2. No statistical analysis 3. No note of patients' preoperative CVS or respiratory state. 4. Neither the underlying pathologies nor the definition of shock were described.
Koch G
85 patients referred to a lung function laboratory for assessment of respiratory or acid-base status. 80 were stable & 5 haemo-dynamically compromised.Non-randomised, uncontrolled clinical trial.Difference between arterial blood (brachial/ femoral) and capillary blood (taken from the earlobe which had been massaged with nicotinate paste. The measurements were taken with the patients at rest breathing both air and O2 and during exercise (breathing air). Some patients had capillary blood taken from their fingers (warmed for 10 mins) at rest breathing air.No statistical difference (t-test) was detected between arterial and earlobe pO2 when the patients were breathing air at rest (p>0.7) and during exercise (p>0.3). A significant difference was noted for shocked patients and when capillary blood was taken from the fingers, but no figures were given for these.1. Clinical heterogeneity. 2. Small subgroups 3. Not enough data presented to assess clinical (as opposed to the statisical) significance of the differences for assessing pCO2 and the use of arterialised finger blood. 4. Neither the underlying pathologies nor the definition of shock were described.
Keller R
32 patients (16 with shock, 16 without).Non-randomised, controlled clinical trial.Difference between simultaneously drawn arterial blood (site not specified) and capillary blood taken from an earlobe (after warming for a few minutes).Arterial – capillary differences: Stable: pH +/- 0.01 (7.30-7.64) pCO2 +/- 0.92mmHg (27-69mmHg) pO2 +/- 4mmHg (25-500mmHg). Shocked: pH +/- 0.013 (7.00-7.40) pCO2 +/- 1.5mmHg (26-68mmHg) pO2 +/-50.8mmHg (25-280mmHg)1. Small study 2. No statistical analysis 3. Inadequate clinical details. 4. Neither the underlying pathologies nor the definition of shock were described.


The 3 studies were small and insufficient clinical details were given to ascertain the effect of confounding factors. In addition, the studies were not randomised and the study populations were significantly different, making comparison between them difficult. The overall trend of all of the studies was that in the shocked state, capillary blood seemed to be less reliable at measuring pO2, pCO2, and pH in spite of vigorous attempts to 'arterialise' the blood (using warmth and application of topical vasodilators).

Clinical Bottom Line

Although the evidence base is limited, it would suggest that capillary blood should not be used for the assessment of blood gas and acid-base values in shocked patients. However, further research needs to be undertaken to clarify this conclusion.


  1. Cooper EA and Smith H. Indirect estimation or arterial pCO2. Anaesthesia. 1961; 16; 445-460.
  2. Koch G. The validity of pO2 measurement in capillary blood as a substitute for arterial pO2. Scand J Clin Lab Invest. 1968; 21; 10-13.
  3. Keller R and Herzog H. Comparison of arterial and capillary blood gas analyses in subjects with clinical evidence of shock. Klin Wschr. 1971; 49; 54-55.