Factors Associated With Missing Biological Samples in the Dog Aging Project

Author

Sydney Holland

Where was it published?

Veterinary Medicine and Science
Volume 10, Issue 6
11 November 2024

 What is this paper about?

In scientific research that follows the same group of dogs over time – like the Dog Aging Project – each sample helps to tell an important part of each dog’s story. When samples are missing or unusable, it creates gaps in that story that make it harder for researchers to draw accurate conclusions from their research. To help prevent these gaps, it is important to study research processes to understand what might cause certain samples to be missing.

Members of our Precision Cohort participate in an annual collection of routine samples like bloodwork (similar to an annual check-up), urine, feces, and hair collected by their primary care veterinarian. Our goal was to understand why certain samples may be unable to be collected at those appointments, leading them to be considered “missing”. In this paper, we assessed 90 sample kits received from our Precision Cohort members to gather data on any samples they were missing and study factors that might be related to sample missingness. We assessed the participant’s geographic region, the dog’s level of cooperation during sample collection, and the ease of blood draw (as scored by the primary care veterinarian).

What do these results mean for me and my dog?

Out of the samples returned in a Precision Cohort sample kit, our blood samples were found to be the most frequently missing. However, when compared to our studied factors like a participant’s geographic region, a dog’s cooperation during sample collection, or the ease of blood collection (cooperation and ease of blood collection as scored by the primary care veterinarian), no studied factors showed a relationship with the missingness of any of our different types of samples.

This means that blood samples from these kits are more likely to be missing than other samples like hair, urine, or feces. However, the increased likeliness that blood samples will be missing is equal across all dogs and is not higher in certain groups than others.

What causes our blood sample to be “missing” can be caused by obvious things like a blood tube not being entirely full, but it can also be caused by other factors like dehydration or an individual dog’s differences in their blood composition. While our Precision Cohort members are asked to fast before their sample collection, we still encourage access to water before their appointment to ensure their dog is well hydrated.

Where can I learn more?

Holland, S.N., Tinkle, A.K., Prescott, J.N., Blattman, B.L., McClelland, R.L., Nam, Y., , Creevy, K.E. and Fajt, V.R. (2024), Factors Associated With Missing Biological Samples in the Dog Aging Project. Vet Med Sci, 10: e70113. https://doi.org/10.1002/vms3.70113

Abstract

Background: The Dog Aging Project (DAP) is a longitudinal study examining aging in US companion dogs, where missing data pose a challenge to result validity and statistical power. Early recognition and procedural adjustments are vital to address missing samples.

Objectives: This research assesses missing biospecimen samples within the Precision Cohort of the DAP, aiming to identify contributing factors and propose mitigation strategies.

Methods: Ninety sample kits collected between August 2021 and October 2021 were evaluated for the missingness of individual kit components. A primary assessment reviewed the missingness of each sample type with hypothesized associated factors. A secondary assessment addressed the missingness of plasma and serum aliquots made after check-in.

Results: EDTA samples for flow cytometry, peripheral blood mononuclear cells (PBMCs), DNA and biobanking were missing at rates of 18.9%, 13.3%, 6.7% and 5.6%, respectively. Faecal samples were missing at 5.6%. Urine and plasma were each missing at 2.2%. Serum, hair and dried blood spot cards were each missing at 1.1%. No statistically significant difference in missingness was seen across regions, ease of blood collection, or cooperativeness of dog during sample collection. Higher time-to-receipt was associated with a higher missingness proportion. Increasing missingness proportions were seen across plasma and serum aliquots that were created in sequential order.

Conclusions: The study suggests reevaluating sample collection instructions to address frequently missing sample types and prevent further loss. Investigation into collection procedures for these types is recommended to identify potential causes of sample yield loss.