Evaluation of the clinimetrics of transcutaneous oxygen measurement and its application in wound care

Abstract

Transcutaneous oxygen (TcPO₂) measurement is a non‐invasive diagnostic tool increasingly being used in clinical practice. Existing evidence suggests that reduction of TcPO₂ is correlated with an increase in skin ischaemia, and its measurement has been shown to have predictive value for wound healing in diabetic patients, success rate of hyperbaric oxygen therapy, amputation rate in patients suffering from peripheral vascular disease or with ischaemic ulcers. This article evaluates the clinimetrics of TcPO₂ measurement in adults with discussion of its potential application in clinical practice.

Introduction

Oxygen supports certain vital processes of wound healing such as angiogenesis, cell motility and extracellular matrix formation 1. It is believed that improving tissue perfusion with approaches to correct local partial pressure of oxygen (pO₂) would serve as an effective adjunct in wound management 2. Hence, measurement of pO₂ in the body is important 3. Traditionally, pO₂ was measured by invasive polarographic electrodes 4. Nowadays, transcutaneous oxygen (TcPO₂) measurement uses a non‐invasive diagnostic tool applied at the epidermal surface to determine the pO₂ level in tissues 5. This article will critically evaluate the clinimetrics of TcPO₂ measurement in adults and discuss its potential application in clinical practices.

Clinimetrics

Clinimetrics was first introduced by Feinstein 6 as a methodological discipline focusing on the quality of measurement issues in clinical medicine. Possible clinical phenomena to be measured using this approach include symptoms, patho‐physiological findings and disease status or severity. Both the quality of the measurement instruments and the quality of performance of the actual measurements are important. Evaluation of the quality of measurement instruments is the main issue in choosing or designing an instrument, while the quality of performance is important in applying such measurement instruments in clinical practice 7.

Validity is important in the assessment of the quality of a measurement instrument. A measurement is valid if it measures what it is intended to measure 8. On the other hand, reliability or precision is the extent to which repeated measurements yield the same result 8. Another essential clinimetric property is responsiveness, which means that a specific measurement instrument should be able to detect clinically relevant changes over time 9. These are all important features for the measurement of TcPO₂.

Transcutaneous oxygen measurement

This technique was originally applied for monitoring pO₂ in newborn infants 10. TcPO₂ is measured by placing a probe against the skin which is then heated to about 45°C causing vasodilation and local reactive hyperthermia. Oxygen diffusing from the vessels across the skin and towards the cathode in the probe is measured as oxygen pressure 11.

Baumberger and Goodfriend 12 found that the pO₂ measured by a dropping‐mercury electrode immediately surrounding a finger placed in a phosphate buffered electrolyte solution heated to 45°C was similar to the partial pressure of oxygen in arterial blood (PaO₂), while Huch et al. 13 developed a miniaturised and heated electrode with in situ calibration to measure the pO₂ on the skin surface. Both of their findings formed the basis for the development of modern TcPO₂ measurement.

Early research by Lubbers 14 demonstrated that epidermal warming using an electrode leads to vasodilatation and increased blood flow in the underlying skin. Moreover, the lipid structures within the cells in the stratum corneum are liquefied, causing improvement in gaseous diffusion. The transmission of heat from the electrode to the underneath blood also shifts the oxyhaemoglobin dissociation curve to the right, causing increased oxygen unloading in the tissues 15. Such changes ‘arterialise’ adjacent capillary blood, raising the capillary pO₂ towards arterial level, where oxygen then diffuses from the arterialised capillary bed to the electrode, raising dermal pO₂ level and the TcPO₂ value can thus be obtained.

The TcPO₂ reading is affected by both methodological and physiological factors. These procedural issues include the temperature in the tissues, shift of haemoglobin dissociation curve, degree of oxygen metabolism in tissues and response time of the electrodes in the equipment, while the physiological considerations include circulatory status, peripheral blood perfusion, and local skin and anatomical conditions 16. A normal TcPO₂ value regardless of electrode location is approximately 60 mmHg 17. Acceptable correlation between changes in pO₂ and TcPO₂ is well established in some well‐controlled systematic clinical investigations 18, 19.

Assessment of the clinimetrics of TcPO₂ measurement includes not only the quality of instruments but also the quality of performance 20, which includes the expertise of the persons carrying out the measurements, the cooperation of the patients or various comorbidities which may affect the readings obtained. The clinimetric properties related to TcPO₂ measurement are now discussed.

Validity of TcPO₂ measurement

Hutchison et al. 21 compared TcPO₂ measured using a heated electrode with the direct measurement of arterial oxygen tension (PaO₂) by obtaining arterial blood samples in three groups of adult patients: those with chronic respiratory disease (n = 20), hypothermic patients studied immediately after cardiopulmonary bypass surgery (n = 8) and those in an intensive care unit (ICU) (n = 14). They found that in patients with chronic respiratory disease, the TcPO₂ measurement reasonably estimated PaO₂ in the range of 50–100 mmHg, with a standard error measurement of 5·2 mmHg. Moreover, for ICU patients, the agreement was less with a standard error of 12 mmHg. The authors found that TcPO₂ had a tendency to underestimate PaO₂ at higher values. However, this finding was not statistically significant. Conversely, no correlation was found between TcPO₂ and PaO₂ in patients studied immediately after cardiopulmonary bypass surgery, due to intense vasoconstriction associated with such major operation under hypothermia.

The sample sizes recruited in this study were small, and the influencing factors affecting TcPO₂ measurement, such as surrounding temperature and patients' dermal and physiological conditions 3, were not strictly controlled. Moreover, the site of electrode placement was not standardised in all patients. The internal and external validity of the results may be affected in view of the presence of such confounders.

A later study by Green et al. 22 compared the simultaneous measurements of TcPO₂ and PaO₂ in seven critically ill adult patients with varying concentrations of inspired oxygen, and a total of 78 simultaneous values were obtained. They found that despite TcPO₂ correlating highly with PaO₂ (r = 0·94), the standard error of the estimate was substantial [standard error of the estimate (SEE) = 15·2]. Therefore, its clinical reliability may be limited.

Despite recruiting a larger sample size only critically ill adults were studied. The external validity of the findings to other groups of patients is questionable, especially to stable patients with wounds.

In general, the concurrent validity of TcPO₂ measurement, in correlation with PaO₂ measurement, is doubtful. The accuracy of TcPO₂ in the estimation of true PaO₂ is unsatisfactory with contradictory results from different studies under different clinical conditions. Such variations are probably due to different influencing factors affecting the diffusion of oxygen from the capillaries to the skin surface for accurate measurement 3. Furthermore, there is a lack of acceptable human studies comparing TcPO₂ with local pO₂ in clinical situations related to wound healing. This may be due to technical difficulties in obtaining the pO₂ values in the settings where most wound patients are treated, for example, outpatient visits or clinics with limited facilities or equipment available. Moreover, as the measurement is an invasive procedure, ethical issues also exist in performing such tests.

Reliability of TcPO₂ measurement

Controversy exists over the reliability of TcPO₂ readings. Bouye et al. 23 conducted a comparative test–retest experiment in 15 ambulatory patients complaining of claudication with estimated maximal walking distance ranging from 100 to 1000 m in the walking test. Measurements were taken from 1 to 13 days apart. TcPO₂ was measured at rest and during exercise at the calf and buttock during the two treadmill tests. The authors found that the reliability of resting absolute values of TcPO₂ before exercise at the two occasions was low, but a high precision of this measurement during exercise was demonstrated when it was corrected for exercise‐induced systemic pO₂ changes with delta‐from‐rest of oxygen pressure calculation (limb TcPO₂ changes minus chest TcPO₂ changes).

However, all subjects recruited in this study had healthy skin; therefore it may not be appropriate to apply the results to patients with wounds or other skin conditions with local hypoxia, such as infection. Moreover, the duration between the two measurements on the same subject could be up to 13 days. Such a long period may allow changes in some parameters, thus affecting the results; for example, change in surrounding temperature, or changes in local tissues or skin conditions due to injury or other medical conditions.

Contradictory results were demonstrated in a later study by Gélis et al. 24 The authors also conducted a test–retest study. Their sample included 30 patients with spinal cord injury (SCI) and the authors evaluated the reproducibility of TcPO2 measurements at the sacral area of these patients during loading in the supine position over 20‐minute monitoring sessions, with two TcPO₂ monitoring sessions performed at two 24‐hour sessions. They found that both absolute TcPO₂ values and their adjusted values were both reproducible for studying cutaneous microcirculation in SCI subjects during loading. This study only recruited a small sample size of a particular group of patients. Therefore, the results may not be appropriate to be extrapolated to other patients. Moreover, TcPO₂ values were only measured at sacral areas, which may not be applicable to most patients with chronic wounds, as their wounds are often located at distal lower limbs.

These two studies tested the test–retest reliability of TcPO₂ measurement, which is the variation in measurements taken by a particular method on the same item under the same condition 8. Reliability can be considered to be high if the results collected from the same subjects but at different times are close to each other. Overall, the evidence from these two studies suggests that controversy exists concerning the precision of TcPO₂ measurement because small sample sizes were recruited in a limited range of clinical situations, and there appears to be a lack of good quality human studies related to wound healing.

TcPO₂ measurement and skin ischaemia

An early study by Dowd et al. 25 sought to identify a correlation between normal and ischaemic skin with TcPO₂ measurement at the dorsum of the foot. Sixty‐two patients with varying degrees of skin ischaemia were graded with ascending severity, that is, ‘those with intermittent claudication without signs of ischaemia in the foot’, ‘those with skin changes in the foot’ and ‘those with signs of gangrene’. Histograms relating the percentage of patients in each group to the TcPO₂ levels were reported, and compared with that of a group of 73 normal volunteers within a comparable age range (44–85 years old). The authors found that, compared with healthy volunteers, patients with skin ischaemia clearly showed a shift to the left of normal with lower TcPO₂ levels, and the degree of shift was related to the severity of ischaemia. Moreover, they also found that TcPO₂ levels below 40 mmHg were associated with profound degrees of skin ischaemia. The authors concluded that the TcPO₂ measurements in patients with peripheral ischaemia of the lower limb could accurately reflect the clinical state of the skin.

The grading of degrees of skin ischaemia in patients recruited in this study was based on assessment of clinical findings, which may be subjective. This may potentially lead to bias due to heterogeneous inter‐rater reliability, leading to weakening of internal and external validity. However, this study provided evidence for the construct validity, which is the primary form of validity, to demonstrate whether TcPO₂ measurement could really measure the degree of skin ischaemia.

TcPO₂ and prediction of chronic wound complications

TcPO₂ has been proposed as a method to determine which patients were at risk for non‐healing 5. Pecoraro et al. 26 conducted a prospective cohort study to identify independent factors associated with complications of chronic full thickness ulcer healing of lower extremity in 46 diabetic patients with lower extremity ulcers. The patients were managed with wound cleansing and dressing, until reaching one of the following six specified definitive outcomes; reepithelialisation, chronic healing failure without evidence for progress after 6 months, documented interim occurrence of osteomyelitis, surgical revascularisation, amputation of the involved extremity or death before complete ulcer healing. The authors identified that a higher peri‐wound TcPO₂ was associated with higher rate of early healing within the first 4 weeks (56·3 versus 26·9 mmHg, P = 0·003) and also a higher rate of reepithelialisation (55·3 versus 40·85 mmHg, P = 0·038). Moreover, an average peri‐wound TcPO₂ <20 mmHg was associated with a 39‐fold increased risk of early healing failure [95% confidence interval (CI) 2·8–1211, P < 0·001]. The authors concluded that peri‐wound cutaneous perfusion was the critical physiological determinant of diabetic ulcer healing.

However, the study included only diabetic patients; had small sample size, giving rise to a wide range of CI and standard error (SE), and hence reducing its reliability. Moreover, wound cleansing and dressing techniques were not standardised among all the subjects being treated, compromising the validity of the study. These findings may not be appropriate to be extrapolated to non‐diabetic patients with similar problems. However, the study does identify that TcPO₂, as a marker of peri‐wound cutaneous perfusion, is a critical physiological determinant of diabetic ulcer healing.

Fife et al. 27 conducted a retrospective cohort study to determine the reliability of TcPO₂ in predicting outcomes for patients with diabetic ulcerations at lower extremities and undergoing hyperbaric oxygen therapy (HBOT). The authors designed a categorical scale for grading, which was determined by an independent reviewer using images. A total of 629 diabetic patients were recruited, with a follow‐up period of 3 weeks. Fife et al. observed that the TcPO₂ value, measured in the chamber during HBOT, was the best to predict the success or failure of treatment using the cut‐off value of 200 mmHg, with a reliability of 74%. However, it was also found that the TcPO₂ value measured at sea‐level in air had no such predictive value.

This study recruited a large sample size with wound assessment by a third‐party assessor to reduce any potential bias and hence to promote its validity. However, it is a retrospective study, with no blinding to result assessment, which may limit its internal and external validity.

TcPO₂ and prediction of lower limb amputation complications

Carter and Tate 28 conducted a prospective cohort study on patients referred to a vascular laboratory with ischaemic wounds and peripheral arterial disease. The authors determined the relationship between TcPO₂ measured at the foot approximately 1 cm above the toes in the first inter‐metatarsal space and subsequent major amputation due to failed healing. A total of 75 limbs in 66 patients were included with a median follow‐up period of 4·2 years. A standard wound cleaning and dressing technique was followed. Potential confounders such as demographic variables and clinical characteristics related to arterial disease were adjusted with multivariable analyses. The authors found that low oxygen tension was associated with increased risk of amputation, that is, a TcPO₂ ≤10 mmHg and also that a TcPO₂ ≤20 mmHg was associated with increased risk of amputation, with relative risks of 2·16 and 2·55, respectively (P < 0·05).

This is a prospective study that minimised recall errors and limited confounding variables and bias. Moreover, the authors adjusted for the most clinically important confounders affecting wound healing such as age, smoking status and diabetes, and TcPO₂ measurement and wound assessment methods were consistent for all patients within the study. Such measures minimised the risk of bias and improved the internal validity.

Recently, Andrews et al. 29 conducted a retrospective and observational study on 307 patients undergoing partial foot amputation and non‐invasive arterial studies including TcPO₂ measurement during a 15‐year period. The authors found that TcPO₂ values were significantly predictive of wound healing within 3 months, with the area under the receiver operating characteristic curve (AUC) value being 0·75 (95% CI 0·70–0·81). Using a cut‐off value of 38 mmHg, the sensitivity and specificity were both estimated to be 71%.

It is a retrospective study recruiting only those undergoing amputations. Bias may exist as the decisions regarding amputations made by different surgeons may not be standardised. Moreover, a heterogeneous group of patients with different indications for amputations was recruited, which may affect the external validity of the findings, and make it difficult to compare with other similar studies.

Conclusion

This review has highlighted some of the continuing controversies in relation to the validity and reliability of TcPO₂ measurement. However, it is increasingly being employed in a variety of clinical practices because it is non‐invasive with encouraging results reported in some studies 21, 23, 24. From the evidence reviewed, it appears that TcPO₂ does correlate well with increasing skin ischaemia 25. Moreover, relevant TcPO₂ measurements have been shown to have predictive value in ulcer healing of diabetic patients 26, success rate of HBOT 27 and amputation rate of patients suffering from peripheral vascular disease or with ischaemic ulcers 28, 29. It is well established that all wounds need good blood supply to heal 30, therefore TcPO₂ measurement may have a potential role in clinical management of other wound conditions such as wound infections or venous ulcers.

The quality of the studies reviewed here varied, and not all areas of wound healing related to TcPO₂ have been studied to date, for example, research into TcPO₂ measurement in patients with venous ulcers, wound infection or surgical wounds (other than amputations) may be worth exploring in the future.