Saloni Malik PGY-1

The Article: Mackenzie DC, et al. Carotid Flow Time Test Performance for the Detection of Dehydration in Children With Diarrhea. J Ultrasound Med 2017. doi:10.1002/jum.14478

The Objective:

Accurate classification of dehydration severity can improve patient care by allowing providers to appropriately select the subset of patients that be treated effectively with oral rehydration alone. However, clinical assessments of volume status in children are often inaccurate. In adults, point of care ultrasound corrected carotid artery flow time measurements have shown utility in characterizing the extent of dehydration. This study aimed to evaluate whether this modality could also predict dehydration severity in a population of children in Dhaka, Bangladesh presenting with acute diarrhea. Corrected carotid artery flow time (CFT) was compared with classification of dehydration based on weight change after rehydration.

The Study:

The parent study of this study was a prospective cohort study of pediatric patients presenting to the International Center of Diarrheal Disease Research in Bangladesh with acute diarrhea. These patients were originally enrolled under a study validating the DHAKA score for dehydration, and a subset of them were included in the ultrasound portion of the study as well. Children were randomly selected for screening by selecting colored marbles from a pouch.

The study design of this paper is best described as a cross-sectional study of those children who were included in the ultrasound portion, as the paper describes different data collected in these children rather than comparing any features which children who did not undergo ultrasound or who had other meaningfully different characteristics.

Inclusion Criteria:

1) Patients younger than 60 months

2) Diarrhea defined as 3 or more loose stools in less than 24 hours

Exclusion Criteria:

1) Children older than 60 months

2) Less than 3 loose stools in 24 hours

3) Diarrhea for more than 14 days

4) A diagnosis other than gastroenteritis

5) Prior enrollment in the study

Data Collection:

Research nurses obtained informed consent, demographic data, historical data, and a clinical assessment of dehydration severity using the DHAKA score and the Integrated Management of Childhood Illness algorithm. Mid-upper-arm measurements were taken to characterize nutritional status. Patients were treated according to rehydration protocols and were undressed and weighed every 8 hours to assess response to rehydration. Percentage of dehydration was calculated as (stable weight - admission weight)/stable weight * 100. Severe dehydration was defined as >9%, some dehydration as 3-9%, and no dehydration as <3%.

Corrected Carotid Artery Flow time:

Using the uSmart 3200T US System, the linear probe was used to visualize the right common carotid artery. Pulsed wave Doppler tracing of flow through the artery was obtained, and cardiac cycle time was determined by measuring the distance between heartbeats (points A to C). Carotid flow time was measured between the upstroke of the flow tracing and the dicrotic notch (points A to B). Corrected carotid artery flow time was calculated by dividing flow time by square root of the cardiac cycle time.

Analysis to determine performance of corrected carotid flow time:

- Area Under the Receiver Operator Characteristic Curve

- Linear regression modeling to describe the relationship between corrected carotid flow time and percentage of dehydration

Results:

After subtracting ineligible participants, participants with uninterpretable images, and those lost to follow up, 251 pediatric patients were included in the study out of a possible 496 initially enrolled in the DHAKA score study. The median age was 21 months. Based on the pre-post weight assessment, 97 patients (39%) had no dehydration, 107 (43%) had some dehydration, and 19% had severe dehydration. The mean corrected carotid artery flow time in this study population was 326 milliseconds.

The AUROCs for corrected carotid artery flow time to predict severe dehydration was 0.51 (95% CI 0.42-0.61) for the research nurses, 0.52 (95% CI 0.40-0.65) for the expert reviewers. A subset of studies identified as good quality had AUROC of 0.53 (95% CI 0.40-0.65). Similarly, linear regression modeling of the association between corrected carotid artery flow time and percentage of dehydration was weak; research nurses r = 0.03 (95% CI -0.05, 0.12, p = .44), and expert reviewer r = 0.1 (95% CI, -0.05, 0.25, p = 0.18).

Discussion:

Although ultrasound has shown utility in predicting severity of dehydration in adults, in this study, corrected carotid artery flow time was not strongly predictive of degree of dehydration as defined by percentage change in body weight.

The authors note that the results may have been stronger with dynamic measurements of volume status instead of a single static measurement, and that using the passive leg raise (PLR) maneuver may also have improved the ability to predict dehydration.

In addition, 326 milliseconds as the mean corrected carotid artery flow time in this study population is comparatively high when compared to the populations in adult studies, indicating either that children differ from adults in their corrected carotid artery flow time, or that this study population was skewed towards higher values, limiting the ability to identify differences.

It may be the case that the failure of this modality to accurately predict severity of dehydration may be due to differences in physiology between children and adults; as children respond to dehydration through higher heart rate rather than stroke volume, which may make interpreting the waveforms more challenging.

The poor performance of this modality have also have been affected by technical challenges in obtaining adequate doppler tracing in young children, due to factors such as patient cooperation, shorter vessel length, and smaller necks. The poor correlation between nurses and expert reviewers may be a reflection of the operator-dependent limitation of this modality.

The Takeaway:

A single static measurement of carotid flow time does not predict volume status or responsiveness.
Utilize change in CFT before and after PLR.
Future studies needed to assess the accuracy of CFT.