Development of an appropriate consumable for point of care bilirubin measurement in low resource settings.

Introduction: Severe neonatal jaundice is a condition where high concentrations of bilirubin in blood can cause irreversible brain damage in newborns, leading to lifelong disability and death. Incidence of neonatal jaundice is much higher in low- and middle-income countries (LMICs) with 667.8 cases per 10,000 live births in Africa compared to 4.4 and 3.2 cases per 10,000 births in the Americas and Europe [1]. The lack of appropriate diagnostic tools for measuring blood bilirubin levels in LMICs is a major contributor to this disparity. Though multiple point-of-care (POC) bilirubinometers have been developed to fill this gap, these devices have yet to be widely adopted in LMICs because their consumables are expensive and difficult to use. The existing consumable for the POC bilirubinometer BiliDx consists of two parts: a lateral flow strip, and a plastic cassette. Eliminating the cassette has potential to greatly reduce consumable cost, however, this change requires designing an alternative method of containing blood while the strip is inserted into the reader. Additionally, users desire a consumable that requires lower blood volumes to simplify collection and want to collect blood without the need for a transfer pipette, necessitating a consumable that can manage any excess blood that is collected. In response to these needs, Rice360 developed a consumable that does not require a disposable plastic cassette and significantly reduces the required blood volume for BiliDx in order to facilitate the adoption of low-cost jaundice diagnostics in LMICs.

Materials and

Methods: We developed a lateral flow strip which is assembled as depicted in Figure 1-A. To use the strip, the user pricks the heel of the newborn, visually fills the collection pad with blood directly from the heel prick, waits for plasma to separate from blood, removes the release paper, and folds the strip along perforation – soaking up any excess blood with the cellulose pad. The strip is then placed into a reusable holder, and inserted into BiliDx to perform a reading (Figure 1-C).

To assess accuracy, blood was applied to prototyped strips as well as the strips supplied with BiliDx and measured using the BiliDx reader. Adult blood was collected from two participants under Rice IRB approval and spiked with bilirubin ranging from 0mg/dL to 25mg/dL and adjusted to 55% hematocrit to simulate hematocrit levels of newborn blood. To assess minimum blood volume required to saturate strips we adjusted adult blood from one participant to 50%, 60% and 65% hematocrit, which represent the 50th, 90th, and 95th percentile of newborn blood in Malawi [2]. Blood volumes ranging from 30 uL to 50 uL were applied to strips, BiliDx measurements were taken every 30 seconds until results reached steady state.
To support use without a transfer pipette to meter blood volume, strips must contain any excess blood. Adult blood from one participant was applied to strips and strips were folded and compressed until blood was visually forced out of them. The compressive force was measured and recorded.

Results, Conclusions, and Discussions: The algorithm BiliDx uses to convert optical densities to bilirubin concentration is trained on data collected using BiliDx cassettes. Because of this, we compared precision of an intermediate optical density variable which is independent of strip materials with the understanding that a new consumable would require the algorithm to be retrained. A quadratic fit of this intermediate variable to calculated bilirubin concentration produced an r-squared value of 0.990 for prototyped strips compared to 0.986 for the consumable that accompanies BiliDx (Figure 1-D). Adjustments made to the strip design appear to have minimal impact on accuracy.

Red optical density, which does not vary with hemoglobin or bilirubin concentration, is used to assess the wetness of the strip. When this value stops changing, the strip is saturated. Blood volume testing showed strips reached fully saturated steady state when supplied 35 uL or more of blood (Figure 1-E). When compared to the commercial consumable which requires 75 uL of blood, this constitutes a 53% reduction in required blood volume.

Blood containment testing showed on average 81 N were required to expel blood from the strip. Li et al determined the average maximum pinch force of adult males is 68.5 ± 23.6 N [3]. This implies the strip can withstand the range of forces it would be subjected to by a user without leaking blood. This data supports the conclusion that this design can be used without a disposable plastic cassette housing.

This design has significant potential to reduce consumable costs for point of care bilirubinometers by eliminating the plastic cassette while maintaining accuracy. Additionally, because these strips are one dimensional, they can be manufactured using existing automated lateral flow processes which can also contribute to cost reduction. Finally, the lower blood volume and ability to collect blood directly has potential to increase usability. Future work includes usability testing comparing our prototyped strip to the current consumable for BiliDx as well as other industry standard cassettes. These reductions in cost and increase in the clinical usability have potential to increase the adoption of point of care bilirubinometers in LMICs for easier identification of jaundice.

Acknowledgements (Optional): 1. Slusher TM, Zamora TG, Appiah D, et al. Burden of severe neonatal jaundice: a systematic review and meta-analysis. BMJ Pediatrics Open. 2017;1(1):e000105. doi:https://doi.org/10.1136/bmjpo-2017-000105

2. Shapiro A, Mtenthaonga P, Rowland Mjumira, et al. Design and field evaluation of a lateral flow cassette device for point-of-care bilirubin measurement. PLOS global public health. 2023;3(8): e0002262-e0002262. doi:https://doi.org/10.1371/journal.pgph.0002262

3. Li L, Li Y, Wang H, Chen W, Liu X. Effect of Force Level and Gender on Pinch Force Perception in Healthy Adults. i-Perception. 2020;11(3):204166952092704. doi:https://doi.org/10.1177/2041669520927043