Abstract: TH-PO392
Conversion of Urine Protein-Creatinine to Albumin-Creatinine Ratio for Use in CKD Risk Equations
Session Information
- CKD: Risk Scores and Translational Epidemiology
November 07, 2019 | Location: Exhibit Hall, Walter E. Washington Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: CKD (Non-Dialysis)
- 2101 CKD (Non-Dialysis): Epidemiology, Risk Factors, and Prevention
Authors
- Sumida, Keiichi, CKD Prognosis Consortium, Baltimore, Maryland, United States
- Nadkarni, Girish N., CKD Prognosis Consortium, Baltimore, Maryland, United States
- Polkinghorne, Kevan, CKD Prognosis Consortium, Baltimore, Maryland, United States
- L Heerspink, Hiddo Jan, CKD Prognosis Consortium, Baltimore, Maryland, United States
Background
Urine albumin-creatinine ratio (ACR) is a core component of CKD staging and used in equations to predict adverse outcomes. However, many cohorts and health systems preferentially measure urine protein-creatinine ratio (PCR) rather than ACR. These assays measure different protein components and the degree to which levels of PCR may be convertible to ACR is untested.
Methods
Cohorts in the CKD Prognosis Consortium with outpatient measures of PCR and ACR performed <90 days apart were included in the development of an equation for conversion of PCR to ACR using linear regression. Analyses were performed using all available data within individual cohorts, accounting for multiple records per person. Data were then meta-analyzed using random effects and linear splines to model log-PCR with a knot at 500 mg/g.
Results
There were 11 cohorts (2 general population, 3 high cardiovascular risk, 6 CKD cohorts) with 34,708 participants included from North America, Europe, and Japan. Average age was 58 years (SD, 16), 50% were female, and 7.4% were black. Median ACR was 181 mg/g (IQR, 51-304), and median PCR was 373 mg/g (152-654). There was no relationship between ACR and PCR at PCR <50 mg/g; thus, these values were excluded in the development of the equation. Above PCR 50 mg/g, there was a log-linear relationship with a slightly shallower slope at PCR >500 mg/g (p<0.001; footnote of figure; median RMSE 0.83). Relationships between PCR and ACR were similar across cohorts (Figure), as well as demographics, and hypertension, cardiovascular disease, and diabetes status.
Conclusion
Guidelines recommend measurement of ACR. However, when ACR is not available, we developed an equation to convert PCR levels >50mg/g to ACR for use in risk equations. Lower levels of PCR were not amenable to harmonization.
Funding
- NIDDK Support