Docs are all told to check it, but here’s why that might not be necessary
On a busy shift the triage nurse comes to find you. She has just put an infant in a room and reports that he has prolonged capillary refill. She walks off before you can ask about the other vital signs. How worried should you be about this baby? Is capillary refill time reliable? Does it depend on where you check it? What happens if the baby is cold or has a fever, does that affect it?
You’re about to go down the hall and check on the baby when another nurse tells you that she put a potentially sick child with a high fever in a room. Earlier, you had been called by the child’s pediatrician who was worried this child has something “serious.” The nurse reports that the child is alert, but has a fever and tachycardia. The blood pressure and capillary refill time are normal. Does that mean he is okay? After all, the tachycardia could be a physiologic response to fever, couldn’t it?
Every PALS course you have ever taken has included capillary refill time (CRT) in the assessment protocol. PALS considers capillary refill to be prolonged if it is greater than two seconds. Your hospital’s sepsis screening initiative includes CRT as a component of their algorithm. CRT is included in your ATLS training. So how well does it work?
So the first question, is it reliable? If you and the triage nurse both independently evaluated the infant referred to above, would you both reach the same conclusion? The answer is probably not. A recent systematic review of the validity and reliability of CRT in children found that reliability is poor. In order to improve reliability, you would need to standardize the site where you checked the CRT, how long you pressed before letting go and in measuring the CRT and you would need to use a stopwatch or timer.
Where you check CRT matters. That same study found that CRT in healthy children and older infants was two seconds or less in the fingertips but up to four seconds in the feet. In a prospective observational study of 1,193 febrile children in an ED setting, peripheral CRT measured at the fingertip was compared to central CRT at the sternum. Agreement between measures was only “fair” (0.35), according to the authors.
Can temperature affect CRT? Cold temperatures do. For instance, in one study looking at healthy, adult volunteers, CRT was measured in the fingertip at room temperature, after immersion in cold water and after immersion in warm water. Warm water was chosen to potentially mimic the effect of a fever. The researchers found considerable variability in CRT, but the only significant effect was a prolongation of CRT with cold temperature. You could argue that warm water doesn’t really approximate the physiologic effects of a fever. But in the systemic review mentioned earlier, there was no clear evidence that fever prolongs CRT, although lower skin temperatures were associated with longer CRT.
Does age matter? Fleming and colleagues also reported on neonates. In the first week of life normal infants may have CRT of five to seven seconds in some sites.
So far, it would appear that CRT has poor reliability, that it can vary depending on where it is checked and that cooler temperatures can prolong it. It is hard to interpret in the first week of life as it is prolonged even in healthy neonates. Fever may not affect it so you are not going to worry about that for the moment. But the bigger question is: if it is so unreliable, why check it at all?
Does CRT add anything in the assessment of the severity of illness or dehydration in children? We need something to go by. Heart rate can be elevated with pain, anxiety, crying or as a response to fever so it is not always predictive. It is well known that blood pressure is maintained in children who are in the early stages of shock. So some additional measure would be useful. How does CRT perform?
Leonard and Beattie asked that very question in an ED setting. They looked at 4,878 children presenting due to illness and measured their CRT at triage. Then they followed their course to see what happened to them. They found that CRT was not associated with meningococcal disease, significant bacterial illness or white count. Having a prolonged CRT was associated with having a higher triage category, getting a fluid bolus and a longer hospital stay.
In the study CRT performed best if greater than three seconds. The authors’ conclusion was that it had limited value as a predictor of serious illness. In a study looking at intubated ICU patients, Tibby and colleagues found CRT weakly correlated with lactate levels. They concluded that a normal CRT had little predictive value. Fleming and colleagues in their systemic review concluded that CRT had high specificity, but low sensitivity.
CRT can be helpful as a “red flag” for serious illness or dehydration. It appears that a prolonged CRT does increase the likelihood of a negative outcome. CRT is not very sensitive however. Do not assume that a normal CRT means that all is well with your pediatric patient. Look at all the vital signs, the general appearance and level of alertness. And above all, stay vigilant.
 Fleming S, et al. Arch Dis Child 2015;100:239-249.
 De Vos-Kerkohf, et al. Arch Dis Child 2017;102:17-21.
 Shinozaki K, et al. J Clin Monitoring and Computing 2018; https://doi.org/10.1007/s10877-018-0159-7.
 Leonard P, et al. European J Emerg Med. 2004;11(3):158-163.
 Tibby S, et al. Arch Dis Child 1999;80:163-166.