Back
Community screening of high-risk populations for albuminuria with point of care (POC) devices provides an innovative way to reach underserved groups with identification of early disease to improve outcomes. We evaluated the test performance of POC devices which assess semi-quantitative (Siemens Clinitek, Healthy.io Minuteful kidney test) and quantitative (Allegro) urine albumin : creatinine ratio (ACR).
245 unselected urine samples from renal and diabetes outpatient clinics were tested on all three devices and compared with clinically validated measurement by Roche Cobas 8000 – c702 module. All samples were tested within 4 hours of sampling by the same user. Semi quantitative thresholds included: normal (<3.39mg/mmol); abnormal (3.39mg/mmol – 33.9mg/mmol); high abnormal (>33.9mg/mmol). Quantitative Allegro samples with readings that were not possible to differentiate between abnormal and high abnormal due to low urine creatinine were excluded. The uACR results from the Roche Cobas 8000 – c702 module standard and Allegro devices were converted to semi-quantitative thresholds to allow sensitivity, sensitivity, positive and negative predictive values and area under ROC curves to be calculated. The weighted Kappa statistic was calculated for each POC device to compare agreement with the Roche ACR assay.
245 results were suitable for analysis and Roche ACR assay ranged from 0.3mg/mmol to 936.2mg/mmol with a mean of 44.83mg/mmol; when grouped this equated to 90 (37%) normal, 91 (37%) abnormal and 64 (26%) high abnormal results.
Performance of the POC uACR devices paralleled their accessibility and ability to be mobilised. The Allegro device had excellent test performance at all thresholds but requires calibration as well as storage and use in a stable location. The Siemens Clinitek device is smaller allowing it to be moved easily to community settings with high negative predictive value and specificity, and good sensitivity. The Minuteful kidney device provides a home-based service using smartphones and also had high sensitivity and negative predictive value but a high false positive rate. These findings allow healthcare providers to understand the limitations of test performance when employing POC testing to screen for CKD.