| | An evaluation of the reliability and validity of capillary refill time testAbstract BackgroundWhile capillary refill time test (CRTT) has been commonly used as a quick and convenient clinical vascular assessment of the lower limb, the validity of this test has not been established. ObjectivesThere were three aims to this study: the first was to evaluate the reliability of CRTT, the second to investigate the premise that CRTT represents skin perfusion, and thirdly to determine whether CRTT can detect lower limb vascular disease. MethodIntra- and inter-tester reliability was evaluated for five experienced physicians who used CRTT on 10 participants. Criterion validity for measurement of skin perfusion was investigated by comparison with laser Doppler flowmetry on 49 participants, and for this same group the usefulness of detecting those persons with peripheral arterial disease and a state of impaired healing was evaluated. ConclusionThese results question the usefulness of capillary refill time test for assessment of vascular status in the lower limb. 1. Introduction  Capillary refill time test (CRTT), also known as subcutaneous venous plexus filling time, subpapillary plexus filling time, capillary fill time and capillary return test, is one of the most commonly used vascular assessments of the lower limb. As a measure of cutaneous perfusion, the importance of CRTT is based on the premise that skin vascularity is a valid indicator of peripheral vascular status. Buerger first described assessment of vascular disease by blanching the skin and observing delayed skin colour return in 1924 [1], and since that time CRTT has become widely used and accepted by health professionals. It has been suggested that clinical assessment of peripheral macrovascular disease and cutaneous microvascular disease should include the use of CRTT [2], [3], [4], [5], [6], [7], [8]. The widespread popularity of CRTT may be because it is accessible, easy to perform and requires no expense. The acceptance of CRTT has occurred despite a notable deficiency of research to determine its usefulness and application in a clinical setting. To date, there is only one study that has investigated the validity of toe CRTT. Boyko et al. [9] investigated the ability of CRTT to determine macrovascular pathology in persons with diabetes mellitus with an ankle-brachial pressure index (ABPI) of less than 0.5. They found CRTT, using a 5 s cut-off score, was sensitive to identifying large vessel pathology in only 28.3% of participants, although the specificity of the test was higher (85.3%). As these results are not generalisable beyond this specific population and ABPIs are known to be poor determinants of microcirculatory status and healing potential [10], [11], [12], [13], CRTT's clinical value for these conditions remains unknown. Sosa et al. [8] found the capillary refill index (CRI) to be useful for persons requiring re-vascularisation surgery, but this test differs from toe CRTT by also involving dorsal foot skin testing and therefore cannot be appropriately compared. As CRTT is thought to be a measure of skin perfusion which can detect vascular disease, there are three aims to this research: (1)to evaluate the intra-rater (re-test) and inter-rater reliability of CRTT; (2)to evaluate whether CRTT is an accurate measure of skin perfusion by comparing against laser Doppler flowmetry (LDF); (3)to evaluate whether CRTT can detect both large and small vessel vascular pathology as represented by peripheral arterial disease (PAD) and microvascular dysfunction associated with impaired healing status, respectively. 2. Methods This research received ethical approval from University of Western Sydney Ethics Review Committee (2000/12) and South Western Sydney Area Health Service Research Ethics Committee (2000/25), and informed consent was obtained from all participants. 2.1. Phase 1: reliability study To explore the intra-tester (re-test) and inter-tester reliability of CRTT, five podiatric physicians, experienced and familiar with the application and interpretation of CRTT were recruited as testers. Ten patients from the University of Western Sydney Uniclinic were recruited as participants (mean age=49.2 years, S.D.=21.25 years). For Trial 1, each tester performed CRTT on both halluces of all 10 participants, the order of which was according to a random allocation model and differed for each clinician; this was repeated an hour later, and again, the order of testing was determined in a random order (Trial 2). Data were assessed for normality and groups were compared using non-parametric testing. No significant carryover test effect (z=−1.02, P=0.057) was found between Trial 1 (median=3.73s) and Trial 2 (median=3.89s). For intra-tester reliability, intraclass correlation coefficients (ICC), standard error of the measurement (S.E.M. measurement) and 95% confidence intervals (CI) were calculated for each tester. An average ICC was derived from z-score transformation [14]. For inter-tester reliability, the trials were analysed separately to determine paired comparison and group ICC, from which S.E.M. (measurement) and 95% CI were derived. For all analyses of both reliability and validity studies, the level of significance was set at α=0.05. 2.2. Phase 2: criterion validity study To evaluate whether CRTT is an accurate measure of skin perfusion, a mixed pathology cohort of 49 participants (32males/17 females; age range 20–95 years) were recruited from University of Western Sydney Uniclinic and Liverpool Hospital Podiatry Outpatients Clinic. The diagnostic criteria for participant characteristics (Table 1) were pragmatic and based on diagnosis by an experienced physician using recent clinical and laboratory results. Persons excluded from participating included those who were under the age of 18 years, unable to provide informed consent, unable to lie supine for 20min and persons with missing halluces. | | |  | Diagnostic group | Number | Diagnostic criteria | |
|---|
| Diabetes mellitus (all type 2) | 22 (44.9%) | Confirmatory diagnosis by physician and requiring management plan | | | Hypertension | 14 (28.6%) | Confirmatory diagnosis by physician and requiring management plan | | | Peripheral arterial disease (PAD) | 15 (30.6%) | Symptoms, signs and assessment findings meet moderate to severe diagnostic criteria from physician or vascular surgeon | | | Peripheral neuropathy | 18 (36.7%) | Symptoms, signs and assessment findings meet moderate to severe diagnostic criteria from physician or endocrinologist | | | Impaired healing | 20 (40.8%) | History of, or current, ulceration or gangrene of the feet or history of non-traumatic amputation within the feet | | | | |
CRTT was benchmarked against resting flux laser Doppler flowmetry, a considered clinical gold standard, in order to investigate its validity for measurement of cutaneous perfusion. Infrared LDF was chosen for its qualities as a non-invasive, direct, quantitative measure of the microcirculation, which measures to a depth inclusive of capillaries, arterioles, venules, metarterioles and arteriovenous anastomoses (AVA) [11], [15]. LDF is also characterised by excellent resolution, allowing it to capture normal spatial and temporal variability, which results from morphological non-uniformity and irregular contractibility of microcirculatory vessels [16], [17], [18], [19]. To improve resting flux reproducibility, a mean value was derived from multiple recordings as recommended [20], [21]. This study used a single channel infrared 780nm diode laser Doppler flowmeter (Model BLF21 Transonic Systems Inc., NY) (Fig. 1). The signal is processed in accordance with algorithms developed to represent flow as mLmin−1100−1g, such that a numerical value, Tissue Perfusion Units (TPU), represents volume and velocity [22]. To standardise pressure application, the laser probe was attached to plantar hallux skin with double-sided adhesive discs. The fibre optic cable was taped to adjacent surfaces to prevent fibre-induced motion artefact. After 30s signal stabilisation period, the lowest value consistently noted over three oscillatory cycles or over 2min was recorded. For the validity study, CRTT was performed with an instrument instead of a tester's thumb, in order to provide a consistent pressure when performing CRTT. A calibrated standardised pressure device was designed (Fig. 2) that consisted of a spring inside a 5ml syringe. As long as the spring maintained its length and the plunger was depressed to the same calibration point, constant force and pressure, similar to that exerted by a thumb, can be assumed for all applications. The capillary refill time was determined to be the time taken for return of colour to the blanched toe-testing site as measured by a stopwatch and conducted in an isolated, quiet room, with an ambient temperature maintained between 23 and 24°C. Participants were rested in a supine position for 10min prior to, and for the duration of, testing. Three measurements of both CRTT and LDF were performed on each of the right and left plantar halluces. In order to investigate whether there was a difference over time with the three repeated CRTT and LDF measures, repeated measures ANOVA was used to explore the data. As no differences were noted, mean values of the three trials were used for analyses. All variables were examined for normality. While LDF values were suitable for log10 transformation, CRTT data could not be transformed into a normal distribution and therefore required non-parametric analysis. A difference was detected between right and left sides in the impaired healing group. Therefore right and left data were considered to be unrelated and were combined for all further analyses [23]. Correlation statistics were used to compare CRTT and LDF, and transformation of LDF also made this data suitable for parametric analysis and linear regression modelling. 2.3. Phase 3: predictive validity study To evaluate whether CRTT could accurately predict the presence of disease, Mann–Whitney U-testing and independent t-testing were used to determine statistical significance for CRTT and LDF, respectively, for each of the diagnostic subgroups of the mixed cohort. In addition, for PAD and impaired healing, sensitivity, specificity and predictive values were also calculated. For these further analyses, cut-off scores for CRTT were required, however with little research in this field, CRTT cut-off scores cited in the literature are subjective opinions of those authors. Cut-off scores between 1 and 5s are frequently recommended [2], [5], [6] and therefore, two scores of 3 and 5s [6], [9] were chosen for investigation in this study. PAD, a common macrovascular disease of the lower limb, is frequently cited in the literature as the target condition for CRTT and is also likely to be the condition for which CRTT was originally intended. However, CRTT is also recommended for assessment of microvascular status. Microcirculatory disease or dysfunction, which commonly results from diabetes mellitus, peripheral neuropathy and PAD, is more complex, multi-systemic, and difficult to define with regard to disease presentation [24], [25], [26], [27], [28], [29]. For these reasons, a state of impaired healing in the foot, as determined by history or presence of ulceration or gangrene, was regarded to be a clear indicator of cutaneous microcirculatory dysfunction, irrespective of aetiology or complexity of pathogenesis. 3. Results 3.2. Phase 3: predictive validity study Statistical differences between diagnostic groups were examined for CRTT and log10LDF. For CRTT, no significant difference was found between groups with and without PAD (z=−1.86, P=0.062), impaired healing (z=−0.21, P=0.830), diabetes mellitus (z=−0.23, P=0.818), neuropathy (z=−0.91, P=0.364) and hypertension (z=−0.78, P=0.062). No difference was found according to gender (z=−1.00, P=0.317) and no correlation was found between age and CRTT values (rs=0.014, P=0.891). LDF measurement detected significant differences according to gender (t96=2.23, P=0.028), and groups with and without diabetes mellitus (t96=2.8, P=0.006) and hypertension (t96=2.2, P=0.031). Of note, no difference was found by LDF values for detection of PAD (t96=0.27, P=0.787), impaired healing (t96=1.95, P=0.054) and neuropathy (t96=1.78, P=0.079). For correlation between LDFlog10 and age, r=0.29 (P=0.003) and for prediction by linear regression r2=0.09. Results for sensitivity, specificity and predictive values for PAD and impaired healing are presented in Table 4. | | | | | PAD | Impaired healing | |
|---|
| | >3s cut-off score | >5s cut-off score | >3s cut-off score | >5s cut-off score | |
|---|
| Sensitivity (%) | 63 | 10 | 50 | 5 | | | Specificity (%) | 59 | 84 | 53 | 20 | | | +ve PV (%) | 40 | 21 | 42 | 14 | | | −ve PV (%) | 78 | 68 | 61 | 56 | | | | |
4. Discussion The usefulness of any clinical measurement tool is dependent upon the extent to which the test result can be relied upon to be accurate, free of error and truly reflective of that which it is intended to measure. The results of this study suggest CRTT has not only questionable reliability, but is also unable to detect large or small vessel vascular disease in the lower limb. For our small group of experienced podiatric physicians, intra-tester (re-test) reliability was varied. Even though the overall ICC of 0.72 was ‘reasonable to good’, the lowest individual ICC was 0.1. More importantly, the overall standard error of the measure was ±1.94s. As CRTT cut-off values recommended in the literature usually range from only 1 to 5s, this re-test error of nearly 2s is proportionally large and therefore clinically significant. Not unexpectedly, ICC for agreement between testers was much lower than re-test values. In terms of criterion validity, the results suggest CRTT provides a fair, but not good, assessment of skin perfusion. While the degree of linear association between CRTT and LDF was moderate (r=−0.61), CRTT was only 38% predictive of the LDF scores in this study. When evaluating disease predictive value, there was no significant difference between the CRTT results of those with and without PAD in this sample. Further, the sensitivity using the 5s cut-off score was only 10%, compared with 63% for the 3s cut-off score. It can be concluded that in this sample, longer refill times were not a result of PAD. The specificity was 59% for 3s cut-off score and 84% for 5s cut-off score. This compared similarly to Boyko et al.'s [9] study, which found a 5s cut-off had 85% specificity but only 28% sensitivity, in persons with diabetes mellitus. These results may also be temperature dependent. Because the room was maintained at 23–24°C, varying degrees of vasospasm and prolonged refill times are less likely to be induced, therefore overestimating CRTT ability to predict a true negative. CRTT was also unable to adequately predict participants with impaired healing (n=20). Sensitivity and specificity for CRTT detection of impaired healing was 50% and 53%, respectively, for a 3s cut-off score and, as found for prediction of PAD (n=15), a significant drop in sensitivity was found for the 5s cut-off (5%). It remains unclear whether shorter or longer cut-off points would have different predictive values for PAD and impaired healing. However, while CRTT was unable to discriminate impaired healing or PAD, it is important to note that the LDF was also unable to predict these groups. These findings are consistent with criticism directed towards the appropriateness of resting microvascular measurements, even using sophisticated instrumentation, for detection of vascular pathology. As most vascular related conditions involve complex and superimposing processes, a range of microcirculatory results might be expected. In particular, the effects of peripheral neuropathy and microvascular dysfunction in diabetes mellitus, are more likely to result in overall increased resting microcirculatory flow [27], [29], [30], [31], [32]. Although some studies have found clinical resting measures to be of some use [11], [33] the majority of evidence suggests microcirculatory provocation tests, which determine reserve flow potential, are required for diagnosis of these conditions and prediction of their outcomes [31], [34], [35], [36], [37], [38]. The results of our study need to be viewed in light of its limitations. Although we used LDF as a “gold standard” of vascular perfusion, it is acknowledged that invasive and more complex procedures, such as isotope clearance, may be more accurate. Our sample size for each of the vascular pathologies was small and results should be considered as a platform for further research, incorporating larger numbers and strict disease classification. Finally, our lack of a comparison arm of participants with no vascular pathology is acknowledged and should be included in any further investigations of CRTT. 5. Conclusion Clinical evaluation of cutaneous vascular perfusion is complex. In this study CRTT reliability was found to be poor and both intra- and inter-tester comparisons produced error that is clinically significant. CRTT was found to be a fair measure of cutaneous perfusion but could not determine the presence or absence of PAD or microvascular disease, the latter as represented by a state of impaired healing, in persons of this sample group. The usefulness of resting measures as an indicator of PAD and impaired healing was also questioned, which is consistent with more recent vascular research findings but further challenges CRTT theory. As such, the use of capillary refill time test for assessment of microvascular and macrovascular disease states of the lower limb should be questioned. Acknowledgements The authors wish to acknowledge the assistance of Dr. Hylton Menz for his assistance in the initial preparation of this manuscript, Dr. Stephen Brodie for development of instrumentation and Rosalind Kidd and staff of Liverpool Hospital Podiatry Outpatients Clinic and University of Western Sydney Uniclinic for recruitment of participants and provision of facilities. References [1]. [1]Buerger L. The circulatory disorders of the extremities. Philadelphia: W.B. Saunders; 1924;. [2]. [2]Campbell J, Bradley A, Milns D, White D, Turner W, Luxton D. Do ‘low-risk’ older people need podiatry care? Preliminary results of a follow up study of discharged patients. Br J Podiatry. 2000;3(2):39–45. [3]. [3]Australian Diabetes Association. Australian podiatric guidelines for diabetes. 2nd ed. 1997. [4]. [4]Diacogiogis D. Clinical indicators for intervention in the rehabilitation of inpatients. Australas J Podiatr Med. 2003;37(4):95–99. [5]. [5]Nelson J. The vascular history and patient examination. In: Robbins J editors. Peripheral vascular disease in the lower extremity. Philadelphia: W.B. Saunders Company; 1992;p. 1–18. [6]. [6]McLeod-Roberts J. Vascular assessment. In: Merriman L, Tollafield D editor. Assessment of the lower limb. New York: Churchill Livingston; 1995;p. 75–102. [7]. [7]Boegel W. Expert diabetic foot care for the primary care physician. Clin Diab. 1999;17(1):37–45. [8]. [8]Sosa T, Loncaric L, Boljesic S. Capillary refill test—a possible auxilliary diagnostic method in obliterative arteriosclerosis. Cor Vasa. 1991;33(5):404–413. MEDLINE [9]. [9]Boyko E, Ahroni J, Davignon D, Stense V, Prigeon R, Smith D. Diagnostic utility of the history and physical examination for peripheral vascular disease among patients with diabetes mellitus. J Clin Epidemiol. 1997;50(6):659–668. Abstract |
Full-Text PDF (1509 KB)
|
CrossRef
[10]. [10]Karanfilian R, Lynch T, Zirul V, Padberg F, Jamil Z, Hobson R. The value of laser Doppler velocimetry and transcutaneous oxygen tension determination in predicting healing of ischaemic forefoot ulcerations and amputations in diabetic and nondiabetic patients. J Vasc Surg. 1986;4(5):511–516. Abstract | Full Text |
Full-Text PDF (562 KB)
|
CrossRef
[11]. [11]Kvernebo K, Slagsvold C, Stranden E, Kroese A, Larsen S. Laser Doppler flowmetry in evaluation of lower limb resting skin circulation. A study in healthy controls and atherosclerotic patients. Scand J Clin Lab Invest. 1988;48:621–626. MEDLINE |
CrossRef
[12]. [12]Ramsey D, Manke D, Sumner D. Toe blood pressure: a valuable adjunct to ankle pressure measurement for assessing peripheral arterial disease. J Cardiovasc Surg. 1983;24:43–48. [13]. [13]Ubbink D, Spincemaille G, Reneman R, Jacobs M. Prediction of imminent amputation in patients with non-reconstructible leg ischaemia by means of microcirculatory investigations. J Vasc Surg. 1999;30:114–121. Abstract | Full Text |
Full-Text PDF (93 KB)
|
CrossRef
[14]. [14]Silver N, Dunlap W. Averaging correlation coefficients: should Fisher's Z transformation be used. J Appl Psychol. 1987;72(1):146–148.
CrossRef
[15]. [15]Hales J, Roberts R, Westerman R, Stephens F, Fawcett A. Evidence for skin microvascular compartmentalization by laser-Doppler and photoplethysmographic techniques. Int J Microcirc: Clin Exp. 1993;12:99–104. MEDLINE [16]. [16]Fagrell B, Nilsson G. Advantages and limitations of one-point laser Doppler perfusion monitoring in clinical practice. Vasc Med Rev. 1995;6:97–101. [17]. [17]Salerud E, Tenland T, Nilsson G, Oberg P. Rhythmical variations in human skin blood flow. Int J Microcirc: Clin Exp. 1983;2:91–102. MEDLINE [18]. [18]Tenland T, Salerud E, Nilsson G, Oberg P. Spatial and temporal variations in human skin blood flow. Int J Microcirc: Clin Exp. 1983;2:81–90. MEDLINE [19]. [19]Wardell K, Braverman I, Silverman D, Nilsson G. Spatial heterogeneity in normal skin perfusion recorded with laser Doppler imaging and flowmetry. Microvasc Res. 1994;48:26–38. MEDLINE |
CrossRef
[20]. [20]Braverman I, Keh A, Goldminz D. Correlation of laser Doppler wave patterns with underlying microvascular anatomy. J Invest Dermatol. 1990;95(3):283–286. [21]. [21]Van den Brande P, von Kemp K, De Coninck A, Debing E. Laser Doppler flux characteristics at the skin of the dorsum of the foot in young and in elderly healthy human subjects. Microvasc Res. 1997;53:156–162. MEDLINE |
CrossRef
[22]. [22]Transonic SI. Transonic laser Doppler monitors BLF 21 series operator's manual. New York: Transonic Systems Incorporated; 1997. [23]. [23]Menz H. Two feet, or one person? Problems associated with statistical analysis of paired data in foot and ankle medicine. Foot. 2004;14:2–5. [24]. [24]Flynn D, Tooke J. Aetiology of diabetic foot ulceration: role for the microcirculation?. Diabet Med. 1992;8:320–329. [25]. [25]Khan F, Elhadd T, Greene S, Belch J. Impaired skin microvascular function in children, adolescents, and young adults with type 1 diabetes. Diabetes Care. 2000;23(2):215–219. MEDLINE |
CrossRef
[26]. [26]Murphie P. Macrovascular disease aetiology and diabetic foot ulceration. J Wound Care. 2001;10(4):103–107. MEDLINE [27]. [27]Netten P, Wollersheim H, Thien T, Lutterman J. Skin microcirculation of the foot in diabetic neuropathy. Clin Sci. 1996;91:559–565. [28]. [28]Tooke J, Sandeman D, Shore A. Microvascular hemodynamics in hypertension and diabetes. J Cardiovasc Pharmacol. 1991;18(2):51–53. MEDLINE [29]. [29]Tooke J. Microvasculature in diabetes. Cardiovasc Res. 1996;32:764–771. MEDLINE |
CrossRef
[30]. [30]Flynn M, Edmonds M, Tooke J, Watkins P. Direct measurement of capillary flow in the diabetic neuropathic foot. Diabetologia. 1988;31:652–656.
CrossRef
[31]. [31]Jorneskog G, Brismar K, Fagrell B. Skin capillary circulation is more impaired than non-diabetic patients with peripheral vascular disease. Diabet Med. 1995;12:36–41. MEDLINE |
CrossRef
[32]. [32]Oomen P, Kant G, Dullart R, Reitsma W, Smit A. Acute hyperglycemia and hyperinsulinemia enhance vasodilation in type 1 diabetes mellitus without increasing capillary permeability and inducing endothelial dysfunction. Microvasc Res 2002;63:1–9. [33]. [33]Winsor T, Haumschild D, Winsor D, Wang Y, Luong T. Clinical application of laser Doppler flowmetry for measurement of cutaneous circulation in health and disease. Angiol-J Vasc Dis. 1987;38(10):727–736. [34]. [34]Diamantopoulos E, Stavreas N, Roussis D, Charitos D, Vasdekis S, Raptis S. Simultaneous laser Doppler and transcutanous oxygen tension measurements in claudicant patients. Int Angiol. 1994;14(1):53–59. MEDLINE [35]. [35]Fairs S, Ham R, Conway B, Roberts V. Limb perfusion in the lower limb amputee—a comparative study using a laser Doppler flowmeter and a transcutaneous oxygen electrode. Prosthet Orthot Int. 1987;11:80–84. MEDLINE [36]. [36]Gebuhr P, Jorgenson J, Vollmer-Larsen B, Nielsen S, Alsbjorn B. Estimation of amputation level with a laser Doppler flowmeter. J Bone Joint Surg. 1989;71-B3:514–517. [37]. [37]Leonardo G, Arpaia M, Del Guercio R. A new method for quantitative assessment of arterial insufficiency of the limbs: cutaneous postischemic hyperemia test by laser Doppler. Angiol-J Vasc Dis. 1987;38(5):378–385. [38]. [38]Karnafel W, Juskowa J, Maniewski R, Liebert A, Jasik M, Zbiec A. Microcirculation in the diabetic foot as measured by a multichannel laser Doppler instrument. Med Sci Monit. 2002;8(7):MT137–MT144. MEDLINE a School of Biomedical and Health Sciences, B24 Campbelltown Campus, University of Western Sydney, Locked Bag 1797, Penrith South DC, NSW 1797, Australia b Academic Unit of Musculoskeletal Disease, University of Leeds, Chapel Allerton Hospital, Chapeltown Rd, Leeds LS7 4SA, United Kingdom  Corresponding author. Tel.: +61 2 4620 3759; fax: +61 2 4620 3792.
PII: S0958-2592(06)00094-0 doi:10.1016/j.foot.2006.08.006 © 2006 Elsevier Ltd. All rights reserved. | |
|
FULL TEXT00094-0/journalimage?img=PIIS0958259206000940.gr1.lrg.gif&fig=fig1&kwhquery=&issn=0958-2592&ishighres=false&allhighres=false&free=yes','journalimage');)
00094-0/journalimage?img=PIIS0958259206000940.gr2.lrg.gif&fig=fig2&kwhquery=&issn=0958-2592&ishighres=false&allhighres=false&free=yes','journalimage');)
00094-0/journalimage?img=PIIS0958259206000940.gr3.lrg.gif&fig=fig3&kwhquery=&issn=0958-2592&ishighres=false&allhighres=false&free=yes','journalimage');)