Reliability of capillary refill time for evaluation of tissue perfusion in simulated vascular occluded limbs

Abstract

Purpose: No standardized execution or evidence demonstrates the area of the digit giving the most accurate capillary refill time (CRT). This study investigated the reliability and validity of CRT, and the relative merits of areas where the test could be performed. Methods: In all, 127 healthy volunteers were assessed for normal CRT at the fingernail, lateral paronychia, and proximal and distal pulps of the index finger. The predictive validity of the CRT for the diagnosis of compromised vascular perfusion was also investigated on 24 subjects, using an inflated tourniquet. Three raters assessed interobserver reliability. Results: The mean fingernail, lateral paronychia, proximal pulp, and distal pulp CRTs were 1.93, 1.78, 1.70, and 1.57 s, respectively. The tourniquet and non-tourniquet results demonstrated significant mean differences; however, the fingernail showed a subtle difference (1.22 s) compared with the proximal pulp (4.46 s). The CRT interobserver reliability was fair at the fingernail (intraclass correlation coefficient [ICC] = 0.51), but very poor in occluded limbs (ICC = 0.13). At the lateral paronychia and finger pulp, the interobserver reliability was reasonable (ICC = 0.75–0.81 [non-tourniquet] vs 0.62–0.68 [tourniquet]). In a receiver-operating characteristic curve analysis, the proximal pulp demonstrated better discrimination (area under the curve = 0.93, 95% CI 0.89–0.97, p < 0.0001); the best cutoff point was calculated to be 3 s at the proximal pulp. Conclusions: CRT use at appropriate areas is reliable. The most dependable site is the finger pulp, and the proposed cutoff is 3 s.