Feline leukemia virus point-of-care lateral flow tests have low positive predictive value in apparently healthy shelter cats

TL;DR

A study found that a common test for feline leukemia virus is not reliable in healthy cats, but more accurate in sick ones.

Contribution

This study evaluates the effectiveness of lateral flow tests for FeLV in shelter cats and provides prevalence estimates using Bayesian modeling.

Findings

The true prevalence of FeLV in healthy cats was estimated at 0.8%.

Positive predictive value of lateral flow tests in healthy cats was only 40.1%.

Sick cats had a much higher positive predictive value of 99.0%.

Abstract

This study aimed to determine the true prevalence of feline leukemia virus (FeLV) infections in apparently healthy and sick shelter cats in Mississippi and estimate the predictive value of a lateral flow test results at the estimated true prevalences to guide testing recommendations. Blood samples (n = 383) were collected from a convenience sample of healthy and sick cats from February 2021 to July 2023. Blood serum samples from both apparently healthy and sick cats were tested for FeLV using lateral flow and insulated isothermal PCR (iiPCR) testing methods. Bayesian latent class modeling was used to estimate test performance and predictive value in both populations. The estimated true prevalence of FeLV in apparently healthy cats was 0.8% (95% CI 0.1%, 2.2%) and 5.3% (95% CI 1.3%, 11.5%) in sick cats. At these prevalences, the estimated positive predictive value of FeLV in healthy cats was 40.1% (95% CI 4.1%, 85.2%) and 99.0% (95% CI 99.4%, 100%) in sick cats. Negative predictive value of FeLV in healthy cats was 79.8% (95% CI 45.2%, 96.8%) and 99.2% (95% CI 97.1%, 100%) in sick cats. The predictive value of a positive test is low for healthy cats. Serial testing may not meaningfully improve the positive predictive value. Positive and negative predictive values were better for sick cats and may have diagnostic utility. Current testing methods may not be adequate for controlling the disease in population.

Figures