Article Text
Abstract
Objectives To determine the agreement of fingertip and sternum capillary refill time (CRT) in children.
Design Prospective, method-comparison study.
Setting Single children's emergency department, UK
Participants 92 children aged 0–12 years, with clinical observations within normal ranges for their age, no relevant medical history and presenting to hospital with a minor illness or injury.
Main outcome measures Agreement between fingertip and sternum CRT measurements.
Results Fingertip CRT ranged from 0.05 to 2.78 s with a mean of 1.08±0.44 and sternum CRT ranged from 0.85 to 2.38 s with a mean of 1.5±0.33. There was a significant difference between fingertip and sternum CRT (t=−9.2, df=91, p=<0.001) and a weak association between the two measurements (r=0.18, p=0.9). A Bland Altman comparison showed the mean difference between fingertip and sternum CRT was −0.49±0.51 with an upper and lower limit of agreement ranging from −1.5 (95% CI −1.69 to −1.32) to 0.53 (95% CI 0.34 to 0.71).
Conclusions Measurements of CRT taken at the fingertip and sternum are not comparable. Fingertip CRT was faster than sternum CRT. Normal CRT is 2–3 s. The current study questions the usefulness of CRT in the assessment of circulation in children.
- Circulatory
- Data Collection
- Paediatric Practice
- Accident & Emergency
Statistics from Altmetric.com
What is already known about this topic
-
That capillary refill time (CRT) is affected by the temperature of the limb, ambient temperature and lighting.
-
Fingertip pulp and sternum sites are recommended sites to gain CRT.
What this study adds
-
Fingertip CRT is faster than sternum CRT.
-
There is no relationship between sternum and fingertip CRT and therefore measurements gained at each site cannot be used comparatively.
Introduction
Capillary refill time (CRT) is an assessment tool routinely used to evaluate circulatory status and the diagnosis of shock and dehydration, with a prolonged refill time suggesting poor peripheral perfusion.1 This has been extrapolated to assume poor perfusion to other tissue and vital organs.2–4 Historically, CRT was initially used in the 1940s to estimate the degree of shock in a battlefield survivor5 however there was no published reference range until Champion et al6 proposed CRT as one of five elements of a trauma score. An upper limit of 2 s was arbitrarily chosen during the development of the scoring system.7 The trauma score was approved by the American College of Surgeons8 and became widely accepted in practice as a means of assessment in trauma and as a general method of assessment, including those conducted on children. Champion et al9 later advised that CRT be removed in the revised trauma score because of difficulties in measuring it in the dark.
A number of other external factors have been shown to affect CRT recording, for example, lighting,9–11 temperature of limb and the environment.10 ,12–14 Other potential variables, such as variation in the procedure for conducting CRT have largely not been considered in previous paediatric studies; however, in neonates pressing for an increased length of time14 ,15 or with increased pressure16 resulted in a prolonged CRT.
The current guidelines provided by the Advanced Life support Group,17 the Resuscitation Council (RC)1 ,18 ,19 and the National Institute for Health and Clinical Excellence20 ,21 state either the fingertip or sternum site for CRT. This is reflected in practice whereby CRT sites are used interchangeably and comparatively, with peripheral circulation viewed to be represented by fingertip CRT and central circulation by sternum CRT.20 ,21 There is no evidence to substantiate the interchangeable or comparative use of CRT sites. All studies in children have used solely fingertip pulp or nail bed; other than in neonates there is no research on sternum CRT. As there is no research on sternum CRT, there is also no description of the interchangeable or comparative use of sternum CRT and fingertip CRT. The aim of this study was to determine the agreement between CRT measurement done on the fingertip and the sternum in children.
Method
This study was conducted in a single children's emergency department in the UK on ‘well’ children. To substantiate ‘wellness’ their observations had to be within normal limits for their age. A history was taken to establish previous, ongoing or recent health problems, and the diagnosis was obtained to ensure that the complaint they presented with was in agreement. Participants had to be over 1 month and up to 12 years of age. Children were excluded from the study if they had taken any medications in the last 6 h; were crying or distressed; had a recent history of diarrhoea, vomiting or poor fluid intake; or had a pre-existing cardiac or metabolic condition.
Due to the lack of literature examining CRT sites in children the sample size for this study was based on existing literature on neonates.14 A sample size of 168 children ensured 95% power to detect at the 5% significance level differences in CRT to the order of 0.5 s.
The study was approved by an NHS Research Ethics Committee. Parents gave informed written consent. Children who were assessed as able to understand the study had to demonstrate understanding of the procedure prior to giving written assent. As potential participants could not have substantial time between receiving information about the study and giving consent,22 (due to the short duration of attendance in the emergency department), families were given an age-appropriate leaflet explaining the procedure with contact details and a participant number; this number was linked to their hospital number so that if, at a later date, they wished to withdraw their child from the study they could do so. None of the participants later contacted the hospital to withdraw from the study.
Data collection
Data were collected by one observer because significant inter-rater variability has been shown in other studies.13 ,14 ,23–25 The observer was a senior children's emergency nurse experienced in the usage of the test. Participants were approached after registering at the emergency department reception and if they presented with a minor illness or injury they were approached to participate. To control for known confounding variables the room temperature of the waiting area and the triage room were monitored and maintained at a constant temperature. Thirty minutes was allowed before CRT measurements were taken to allow the participant to acclimatise to the room temperature. The examination always took place in the same room, in bright artificial light. Data were collected at triage. A record was made of the child's gender, ethnicity, skin colour, reason for attendance and diagnosis.
The reliability of the CRT measurement required standardisation of the technique. The researcher's index finger was used to press on the sternum with enough pressure to cause the skin to blanche. This was done to the manubrium part of the sternum, level with the first intercostal space; pressure was applied for 5 s, as recommended by the RC1 ,18 ,19 and the Advanced Life support Group.17 The time taken for the skin to return to its pretest colour after the finger was removed was recorded in milliseconds by a digital stopwatch. Directly after this, the measurement was then repeated at the fingertip pulp of the child's right index finger, between the observer's thumb and index finger, the hand was held at heart level.1 ,18
Statistical analysis
A difference of 0.5 s between sternum and fingertip CRT was taken to be clinically significant. This was chosen because 0.5 is the usual mathematical amount that is used to round up to the next whole digit. Data were summarised using mean and SD. Differences between fingertip and sternum CRT measurements were analysed using a two-tailed t test; the relationship between measurements was assessed through Pearson's correlation coefficient and the level of agreement was determined using the Bland Altman method of comparison.26
Results
The study was conducted in March 2011. Data collection needed to be by a single researcher to ensure measurements were standardised, which limited the number of children who could be approached to participate. During the data collection period 1524 children attended the emergency department of whom a convenience sample of 103 children fulfilling the inclusion criteria were approached to be included in the study. Of the 103 prospective participants, three declined to take part. Sternal CRT in darker skinned children was difficult to assess and as a result eight children were excluded from the analysis. These were Somali (n=2, 2%), afro-Caribbean (n=3, 3%), Indian (n=1, 1%), mixed race (n=1, 1%) and Sri Lankan (n=1, 1%). Data for 92 children were therefore included in the analysis.
There were a similar number of male (n=44, 48%) and female (n=48, 52%) children that participated. Their mean age was 6.3 years±3.7 (range from 7 weeks to 12 years). Children were from a range of ethnic groups: white children (n=37, 40%), Asian children (n=33, 36%), mixed race children (n=19, 21%) and black children (n=3, 3%).
There were various reasons for attendance and in most of the cases the diagnosis was the same, for the purpose of analysis some reasons for attendance were grouped (table 1). The majority of participants did not have any previous medical history (n=76, 83%). Other medical histories stated were asthma (n=10, 11%), prematurity at birth (n=2, 2%), eczema (n=1, 1%), bladder problem (n=1, 1%), tuberculosis (n=1, 1%) and epilepsy (n=1, 1%).
All the clinical observations were within the expected range17 (table 2). Fingertip CRT was an average 0.42 s quicker than sternum CRT. There was a significant difference between fingertip and sternum CRTs (t=−9.2, df=91, p=<0.001). Similarly, there was a weak relationship between fingertip and sternum CRT (r=0.18, p>0.05; figure 1). The Bland Altman comparison showed the mean difference between fingertip and sternum CRT was −0.49±0.51 (figure 2) with an upper and lower limit of agreement ranging from −1.5 (95% CI −1.69 to −1.32) to 0.53 (95% CI 0.34 to 0.71) indicating sternum CRT ranged between 1.5 s slower to 0.53 s quicker than fingertip CRT.
Discussion
This study aimed to determine whether there was agreement between CRT taken at the fingertip and sternum in a non-critically ill population of children. We found that there were significant differences in measurements, no-relationship between measurements at the two sites and the limits of agreement were so variable that it suggests that CRT measurements taken at fingertip and sternum were not interchangeable. However, what we are unable to conclude is if one site is preferable to the other. Fingertip CRT demonstrated greater variability than sternum CRT as it had the quickest and slowest CRT. Moreover, our results support guidance of a normal CRT of 2–3 s,17 rather than the RC's advice of less than 2 s.1 ,18
Our results support those shown in other studies in children23 and neonates13–15 ,27 ,28 that there is wide variability in CRT measurements taken at different sites. For example, CRT measurement taken on a heel was significantly slower than the fingertip14 ,15 ,23 ,27 CRT is a commonly conducted procedure for assessing circulation yet there has been little research exploring confounding factors. The results of eight children needed to be excluded from the analysis on the basis of skin colour, the darker the skin was the more difficult it was to establish sternum CRT. This could lead to the conclusion that CRT should be carried out on only the fingertip, the problem with this being that it can be difficult to obtain a CRT on the tiny fingertip of a baby. This is reflected in studies conducted on neonates, where forehead and chest sites were found to be preferable.13–15 ,27 ,29
Another aspect of CRT measurement in which there is a dearth of evidence is the influence of environmental factors and inter-rater variability. This study aimed to control for these factors by standardising where the procedure was conducted, allowing time for body temperature to acclimatise to the environmental temperature and being conducted by a single researcher. Even controlling for these confounding factors there were wide variations in results, suggesting the procedure conducted in ‘real life’ in a variety of settings by multiple health professionals would produce even greater variability.
There are a number of limitations to the current study. First, due to time constraints we were unable to achieve our planned sample of 168 children. However, the wide variability suggests 100 children were a sufficient number to adequately power the study in order to make conclusions. Second, the study was single centred and there was a predominance of participants who were either white or fair skinned, the results can therefore not be generalised to other ethnic groups. Similarly, this study only included a non-critically ill group of children so the degree of variability between fingertip and sternum measurements needs to be established across a wide range of physiological conditions.
Conclusion
We have shown that fingertip and sternum CRT measurements in children are not comparable and the variability is such that assumptions about results from one site cannot be made for the other. Fingertip is faster than sternum CRT however this is not consistent so a standardised correction cannot be applied. We have suggested a way in which CRT can be standardised; however the influence of so many variables cast doubt over the usefulness of this method of assessment and it may be time for the RC and others to re-evaluate their recommendations.
Recommendations for practice
The current results suggest more precise guidance for performing CRT on children is required. The lack of guidance may be due to the limited evidence available to underpin this method of assessment. In the absence of evidence guidance needs to account for the influence of various extraneous factors that may influence CRT, which need be considered in practice when the CRT test is carried out. This should be in a similar way in which any test is calibrated before use. Based on the current results a five-stage recommendation is being proposed for conducting CRT in children (box 1).
Five point recommendation for carrying out the capillary refill time test
-
Consider any factors that are likely to alter capillary refill time (CRT) (eg, consider not doing CRT if coldness of poor lighting has the potential to effect the result).
-
Choose either the sternum or the fingertip pulp site (consider skin colour and size of child's finger).
-
Using your index finger apply enough pressure to cause the skin to blanche (turn pale).
-
Apply pressure for 5 s.
-
Remove pressure and immediately count aloud how long it takes for skin to return from blanched back to pretest colour.
CRT Should be 2–3 s
-
One person should carry out CRT and it should not be repeated or sites compared.
-
When recording CRT in the patients notes the letter F or S should be written next to the number to indicate whether the CRT was taken at the fingertip pulp or sternum, respectively.
-
Fingertip CRT requires raising the hand to heart level.
References
Footnotes
-
Contributors All authors helped to plan the research, and took part in the writing of the final article. This research was carried out as part of MSc Paediatric Advanced nurse practitioner degree from London South Bank University. At the time of research the Author JC was employed by Ealing Hospital NHS Trust.
-
Competing interests None.
-
Ethics approval The London Central REC 2 Committee granted ethical approval, as well as the hospital research and development centre at Ealing Hospital, and London South Bank University Faculty Ethics Committee.
-
Provenance and peer review Not commissioned; externally peer reviewed.