Predicted Albumin-Creatinine Ratio Performs Well for CKD Screening, Staging, and Prognosis

By Megan Brooks

July 14, 2020

(Reuters Health) - Researchers with the Chronic Kidney Disease Prognosis Consortium have developed equations to predict urine albumin-creatinine ratio (ACR) from urine protein-creatinine ratio (PCR) and dipstick protein.

In testing, predicted ACR performed well in CKD screening and staging, as well as prognostically with the two-year, four-variable kidney failure risk equation (KFRE).

Urine ACR is the preferred way to measure albuminuria. However, if ACR is not available, predicted ACR from PCR or urine dipstick protein may be helpful for CKD screening, staging, and prognosis, the investigators said in Annals of Internal Medicine.

Measurement of urine ACR is a "gold standard for quantifying albuminuria, but sometimes total protein assays or semiquantitative urine dipstick may be measured instead due to lower costs or other considerations," Dr. Morgan Grams, from Johns Hopkins University in Baltimore, told Reuters Health by email.

"How to relate these different measures of albuminuria in clinical practice and in research settings is uncertain. There have been some rough estimates provided in guidelines and single studies that examined the associations between alternate measures of urinary protein and urinary albumin to creatinine ratio. We decided to pursue a meta-analysis to develop the most robust conversion factor as possible," said Dr. Grams.

To develop equations to predict ACR, the researchers collated data from 12 research and 21 clinical cohorts of 919,383 adults (mean age, 61, 50% women) with same-day measures of ACR and PCR or dipstick protein.

They compared the accuracy of predicted ACR at the screening threshold for CKD (ACR ≥ 30 mg/g) and CKD staging categories for albuminuria (stage A2: ACR of 30 to 299 mg/g, stage A3: ACR ≥300 mg/g).

For screening and classification into stages A2 and A3, the PCR-predicted ACR equations had moderate sensitivity (91%, 75%, and 87%, respectively) and specificity (87%, 89%, and 98%, respectively).

The urine dipstick categories of trace to + and ++ had high specificity (>88%) but lower sensitivity (<78%) for identifying CKD stage A2 and A3, respectively.

For individual risk prediction, the estimated two-year, four-variable KFRE using predicted ACR from PCR was "very similar" to the one using observed ACR, the authors report, with a c-statistic of 0.883 in the crude and adjusted models compared with 0.879 using observed ACR.

The researchers have developed a calculator ( that allows for easy conversion.

"The process is very quick and simple for use in the clinic: plug in the protein-creatinine ratio or dipstick value as well as whether the patient is male or female. There is the option to include the patient's diabetes and hypertension statuses if known, but it is not required," Dr. Grams said.

"It is important to note that the conversion is an estimate, and that the most precise value of albuminuria would be obtained from directly measuring the urine albumin to creatinine ratio. The web calculator emphasizes this uncertainty by providing the prediction interval -- the range of values within which 95% of the corresponding urine albumin to creatinine ratios would fall," she noted.

The authors of an editorial say the implications of accurate and high-performing equations to estimate ACR from urine protein measures are "immense."

Dr. Tyrone Harrison and Dr. Brenda Hemmelgarn from University of Calgary in Canada note the reagent costs to measure protein are "considerably less than those for measuring albumin; therefore, it is highly likely that in some health care systems, only PCR or urine dipstick measures are available. In this setting, use of conversion equations from the inexpensive and widely available PCR or dipstick gives patients and clinicians in economically disparate locations access to such tools as the KFRE."

"Although ACR measurement is not ubiquitous in all health care settings and is not available in all research settings, ACR-estimating equations, such as those reported in this study, may enhance both research and clinical care," the editorialists say.

SOURCE: and Annals of Internal Medicine, online July 13, 2020.