Arterial versus capillary blood gases: A meta-analysis

https://doi.org/10.1016/j.resp.2006.07.002Get rights and content

Abstract

A meta-analysis determined whether capillary blood gases accurately reflect arterial blood samples. A mixed effects model was used on 29 relevant studies obtained from a PubMed/Medline search. From 664 and 222 paired samples obtained from the earlobe and fingertip, respectively, earlobe compared to fingertip sampling shows that the standard deviation of the difference is about 2.5× less (or the precision is 2.5× better) in resembling arterial PO2 over a wide range of arterial PO2's (21–155 mmHg). The lower the arterial PO2, the more accurate it is when predicting arterial PO2 from any capillary sample (p < 0.05). However, while earlobe sampling predicts arterial PO2 (adjusted r2 = 0.88, mean bias = 3.8 mmHg compared to arterial), fingertip sampling does not (adjusted r2 = 0.48, mean bias = 11.5 mmHg compared to arterial). Earlobe sampling is slightly more accurate compared to fingertip sampling in resembling arterial PCO2 (arterial versus earlobe, adjusted r2 = 0.94, mean bias = 1.9 mmHg; arterial versus fingertip, adjusted r2 = 0.95, mean bias = 2.2 mmHg compared to arterial) but both sites can closely reflect arterial PCO2 (880 total paired samples, range 10–114 mmHg). No real difference between sampling from the earlobe or fingertip were found for pH as both sites accurately reflect arterial pH over a wide range of pH (587 total paired samples, range 6.77–7.74, adjusted r2 = 0.90–0.94, mean bias = 0.02). In conclusion, sampling blood from the fingertip or earlobe (preferably) accurately reflects arterial PCO2 and pH over a wide range of values. Sampling blood, too, from earlobe (but never the fingertip) may be appropriate as a replacement for arterial PO2, unless precision is required as the residual standard error is 6 mmHg when predicting arterial PO2 from an earlobe capillary sample.

Introduction

Measurements of arterial blood gas tensions, pH, and arterial oxyhemoglobin saturation (%SaO2) are commonly obtained in patients suffering from acute and chronic respiratory disorders. Arterial blood gas tensions are also used in physiological exercise testing for endurance athletes to determine the extent of gas exchange abnormalities (Dempsey and Wagner, 1999). Currently, direct arterial catheterization of the radial artery is the widely accepted gold standard technique for obtaining the most accurate assessment of pulmonary gas exchange. However, this technique can be painful and, in most cases, can only be done by a physician (Pitkin et al., 1994). Alternative non-invasive methods have been proposed such as pulse oximetry that are used to assess arterial oxyhemoglobin saturation; however, obtained values correlated poorly with arterial PO2 (PaO2) values (Pitkin et al., 1994). Pulse oximetry is a poor predictor of PaO2 because of the sigmoidal shape of the oxyhemoglobin dissociation curve and because the curve can be shifted in various clinical and physiological conditions. For measures of pulse oximetry, current studies report a mean absolute difference (bias or accuracy) between 0 and 0.3% with a forehead oximetry sensor, and −0.5 to 2.5% when a sensor is placed on the index finger (Kolb et al., 2004, Yamaya et al., 2002). The precision (S.D. of the differences) of pulse oximetry compared to arterial measurement via co-oximetry is around ±2.5% (forehead) or ±3.0 to 7.3% (finger)(Kolb et al., 2004, Yamaya et al., 2002). Furthermore, both bias and precision values for pulse oximetry worsen at saturations levels below 85% (Kolb et al., 2004). Thus, pulse oximetry is appropriate to calculate oxygen delivery; however, it is poor when evaluating pulmonary gas exchange.

The sampling of arterialized capillary instead of arterial blood may be a reliable substitute to arterial sampling. If adequate vasodilation can be achieved from either applying a topical vasodilatory substance to the skin, and/or warming the area (with warm water, towel, or heater), arterial and venous PO2 should converge and the arterialized capillary sample should closely reflect arterial blood. Arterialized sampling can be a more practical option compared to arterial sampling because it is less invasive (Fajac et al., 1998) and can be performed by non-medical staff (Hughes, 1996) with minimal discomfort (Spiro and Dowdeswell, 1976). The advantages also reduce the cost of this procedure. Small sample amounts are required (only a maximum of about 125 μL of blood is required per sample) and there are minimal procedural, ethical, and safety concerns for performing such a technique in a university setting/exercise physiology lab or hospital. This technique is not commonly practiced today due to equivocal findings; some studies show agreement with PaO2 and arterial oxyhemoglobin saturation compared to arterialized samples at rest in adults (Langlands and Wallace, 1965, Laughlin et al., 1964, MacIntyre et al., 1968, McEvoy and Jones, 1975, Pitkin et al., 1994, Sadove et al., 1973, Spiro and Dowdeswell, 1976) and children (Godfrey et al., 1971), while, other studies do not show such agreement (Eaton et al., 2001, Fajac et al., 1998, Sauty et al., 1996).

Over the last eight decades, several studies have been performed assessing the accuracy, precision, and usefulness of arterialized earlobe and fingertip blood sampling in comparison to arterial blood sampling (Begin et al., 1975, Canny and Levison, 1986, Christoforides and Miller, 1968, Dall’Ava-Santucci et al., 1996, Davis et al., 1975, Fajac et al., 1998, Gambino, 1959, Gambino, 1961, Godfrey et al., 1971, Hale and Nattrass, 1988, Harrison et al., 1997, Kirubakaran et al., 2003, Knudsen and Hansen, 1962, Koch, 1965, Langlands and Wallace, 1965, MacIntyre et al., 1968, McEvoy and Jones, 1975, Nonaka et al., 1971, Olivia et al., 1973, Osugi and Shibata, 1971, Pitkin et al., 1994, Roehr and Roncoroni, 1965, Sadove et al., 1973, Sauty et al., 1996, Spiro and Dowdeswell, 1976, Verges et al., 2005, Wallman et al., 1968, Zavorsky et al., 2005), with the first study dating as far back as 1922 (Lundsgaard and Moller, 1922). A mini-review (Murphy and Harrison, 2001), an editorial (Hughes, 1996) and letters to the editor (Barry et al., 1995, Bernhardt, 1995, Pandit, 1995) have been published on the accuracy of capillary sampling to that of arterial sampling showing conflicting views. A meta-analysis would contribute to our understanding in the controversy arising from conflicting studies on the accuracy, precision, and usefulness of fingertip or earlobe capillary blood samples as a substitute to arterial samples.

Therefore, the purpose of this paper is to estimate the relationship between arterialized capillary blood samples from the earlobe and fingertip to arterial samples for PO2, PCO2, pH and oxygen saturation (SO2) from a review of original research studies. The specific objectives of this study were to determine whether PaO2, arterial PCO2 (PaCO2), arterial pH (pHa), and arterial oxyhemoglobin saturation (SaO2) are similar to arterialized fingertip or earlobe blood samples over a wide physiological range, and to determine which capillary sampling site more closely reflects arterial samples. Our hypothesis was the following: predicting PaCO2 and pHa from either capillary site would be accurate, however, predicting PaO2 from capillary sampling would not be accurate over a wide physiological range. Our reason for suggesting that PaO2 could not be predicted from a capillary sample was that the PaO2 would be higher than the PO2 of arterialized blood flowing from the fingertip or earlobe due to the utilization of oxygen from the skin capillary bed of the fingertip and earlobe. Therefore, the fluid collected from the pierced fingertip and earlobe is a mixture of blood from capillaries and venules and that the arteriovenous difference for PO2 would be large (60 mmHg at rest, 75 mmHg during light exercise (Hughes, 1996)). As such, not even vasodilation of the surrounding skin through rubbing or a vasoactive ointment would be sufficient to raise the capillary PO2 to that of PaO2.

Section snippets

Methods

A PubMed/Medline search strategy (from 1965 to present) was used for this research, and it included different short sentences considered to be directly related to this meta-analysis. The sentences typed and the number of articles displayed for each one of them were as follows: arterialized blood gases (121), arterialized sample (23); artery capillary sample (64); ear lobe puncture (26); arterialized blood for PO2 PCO2 and pH determination (30), arterialized blood for PO2 determination (55),

Results

The mean differences of PO2, PCO2, pH, and SO2 between arterial and capillary sites are presented in a forest plot format in Fig. 1a–g, respectively. The summary data breakdown of each variable is presented in Table 1, Table 2, Table 3, Table 4, respectively. Irrespective of the capillary site, the accuracy and precision improves for PO2 as arterial PO2 decreases. This is not so apparent for PCO2 (Table 2) or pH (Table 3). For SO2 reported in Table 4, we could only obtain individual data from

Discussion

This is the first complete meta-analysis examining the validity of obtaining a blood sample from two different capillary sites to estimate arterial blood-gas values. The data convincingly shows that the blood sampled from the earlobe is more accurate compared to blood sampled from the fingertip for PO2 and PCO2. Sampling blood from earlobe (but never the fingertip) may be appropriate as a replacement for arterial PO2, unless precision is required as the residual standard error in the regression

Acknowledgements

We would like to thank Adriana Decker, MD, for help with the initial statistical analyses of the data, De Tran, MD, and Andrew Owen, MD, for reviewing this manuscript. This manuscript was initially presented as an abstract at the 2005 Canadian Society for Exercise Physiology National Conference (Can. J. Appl. Physiol. 2005, 30 Suppl.; S87, S88).

References (51)

  • C. Christoforides et al.

    Clinical use and limitations of arterialized capillary blood for PO2 determination

    Am. Rev. Respir. Dis.

    (1968)
  • J. Dall’Ava-Santucci et al.

    Is arterialized earlobe blood PO2 an acceptable substitute for arterial blood PO2?

    Eur. Respir. J.

    (1996)
  • K. Dar et al.

    Arterial versus capillary sampling for analysing blood gas pressures

    BMJ

    (1995)
  • J.A. Dempsey et al.

    Exercise-induced arterial hypoxemia

    J. Appl. Physiol.

    (1999)
  • I. Fajac et al.

    Blood gas measurement during exercise: a comparative study between arterialized earlobe sampling and direct arterial puncture in adults

    Eur. Respir. J.

    (1998)
  • R. Ferlinz et al.

    Comparative studies of PO2 PCO2 and pH values in capillary blood from the ear lobe and finger tip

    Prax. Pneumol.

    (1970)
  • L. Fungi et al.

    Arterial sample and arterialized capillary sample in pediatric resuscitation

    Minerva Anestesiol.

    (1976)
  • S.R. Gambino

    Normal values for adult human venous plasma pH and CO2 content

    Tech. Bull. Regist. Med. Technol.

    (1959)
  • S.R. Gambino

    Collection of capillary blood for simultaneous determinations of arterial pH, CO2 content, PCO2, and oxygen saturation

    Am. J. Clin. Pathol.

    (1961)
  • P.J. Hale et al.

    A comparison of arterial and non-arterialized capillary blood gases in diabetic ketoacidosis

    Diabet. Med.

    (1988)
  • A.M. Harrison et al.

    Comparison of simultaneously obtained arterial and capillary blood gases in pediatric intensive care unit patients

    Crit. Care Med.

    (1997)
  • J.M. Hofford et al.

    Arterial blood gas measurements. The clinical usefulness and limitations of arterialized capillary bloods

    Del. Med. J.

    (1973)
  • W.G. Hopkins

    Bias in Bland–Altman but not regression validity analyses

    Sportscience

    (2004)
  • J.M.B. Hughes

    Blood gas estimations from arterialized capillary blood versus arterial puncture: are they different?

    Eur. Respir. J.

    (1996)
  • C. Kirubakaran et al.

    Comparison of blood gas values in arterial and venous blood

    Indian J. Pediatr.

    (2003)
  • Cited by (0)

    View full text