An age-stratified semiquantitative urine ACR triage strategy for quantitative confirmation: a retrospective...

Early detection of albuminuria is important for identifying individuals at risk of chronic kidney disease (CKD) progression. Semiquantitative urine albumin-to-creatinine ratio (ACR) strips may serve as an initial laboratory screening or triage tool. However, screening strategies generally apply a single fixed threshold, regardless of patient age. Because CKD prevalence increases with age, a single cut-off does not account for differences in pre-test probability across age groups, potentially resulting in different trade-offs between quantitative confirmation and false-negative classification. We evaluated an age-adjusted threshold approach to improve diagnostic safety while reducing the number of quantitative ACR confirmations required.

Semiquantitative ACR strip results (UC-3500, Sysmex) and quantitative ACR values (Alinity c, Abbott) were collected and analyzed. Diagnostic performance indicators for ACR strip testing were evaluated across different decision thresholds in the overall population and among age groups (18-44, 45-64 and ≥65 years), using quantitative ACR measurement as the reference standard.

Selected thresholds were 30, 30, and 15 mg/g creatinine for the 18-44, 45-64, and ≥65-year age groups with false omission rates (FOR) of 1.36 %, 3.97 % and 4.47 %, respectively. Overall, the age-stratified strategy yielded a sensitivity of 85.60 %, a NPV of 96.44 % and a FOR of 3.56 %, while avoiding 65.16 % of quantitative ACR measurements.

In this retrospective simulation, an age-stratified reflex-testing strategy provided a balance between reducing quantitative ACR measurements and limiting the residual probability of ACR ≥30 mg/g creatinine following a negative semiquantitative result. The strategy used more sensitive thresholds in older patients, in whom albuminuria prevalence was higher, while allowing a greater reduction in quantitative testing in younger groups. These findings were reproducible in a temporal internal validation cohort but require external and local validation before implementation in other laboratories.