Urine Albumin-Creatinine Ratio Variability in People With Type 2 Diabetes: Clinical and Research Implications

Am J Kidney Dis. 2024 Jul;84(1):8-17.e1. doi: 10.1053/j.ajkd.2023.12.018. Epub 2024 Mar 15.


Rationale & objective: Evidence has demonstrated that albuminuria is a key diagnostic and prognostic marker of diabetic chronic kidney disease, but the impact of its day-to-day variability has not been adequately considered. This study quantified within-individual variability of albuminuria in people with type 2 diabetes to inform clinical albuminuria monitoring.

Study design: Descriptive cross-sectional analysis.

Setting & participants: People with type 2 diabetes (n=826, 67.1 [IQR, 60.3-72.4] years, 64.9% male) participating in the Progression of Diabetic Complications (PREDICT) cohort study.

Exposure: Four spot urine collections for measurement of urinary albumin-creatinine ratio (UACR) within 4 weeks.

Outcome: Variability of UACR.

Analytical approach: We characterized within-individual variability (coefficient of variation [CV], 95% limits of random variation, intraclass correlation coefficient), developed a calculator displaying probabilities that any observed difference between a pair of UACR values truly exceeded a 30% difference, and estimated the ranges of diagnostic uncertainty to inform a need for additional UACR collections to exclude or confirm albuminuria. Multiple linear regression examined factors influencing UACR variability.

Results: We observed high within-individual variability (CV 48.8%; 95% limits of random variation showed a repeated UACR to be as high/low as 3.78/0.26 times the first). If a single-collection UACR increased from 2 to 5mg/mmol, the probability that UACR actually increased by at least 30% was only 50%, rising to 97% when 2 collections were obtained at each time point. The ranges of diagnostic uncertainty were 2.0-4.0mg/mmol after an initial UACR test, narrowing to 2.4-3.2 and 2.7-2.9mg/mmol for the mean of 2 and 3 collections, respectively. Some factors correlated with higher (female sex; moderately increased albuminuria) or lower (reduced estimated glomerular filtration rate and sodium-glucose cotransporter 2 inhibitor/angiotensin-converting enzyme inhibitor/angiotensin receptor blocker treatment) within-individual UACR variability.

Limitations: Reliance on the mean of 4 UACR collections as the reference standard for albuminuria.

Conclusions: UACR demonstrates a high degree of within-individual variability among individuals with type 2 diabetes. Multiple urine collections for UACR may improve capacity to monitor changes over time in clinical and research settings but may not be necessary for the diagnosis of albuminuria.

Plain-language summary: Albuminuria (albumin in urine) is a diagnostic and prognostic marker of diabetic chronic kidney disease. However, albuminuria can vary within an individual from day to day. We compared 4 random spot urinary albumin-creatinine ratio (UACR) samples from 826 participants. We found that a second UACR collection may be as small as a fourth or as large as almost 4 times the first sample's UACR level. This high degree of variability presents a challenge to our ability to interpret changes in albuminuria. Multiple collections have been suggested as a solution. We have constructed tools that may aid clinicians in deciding how many urine collections are required to monitor and diagnose albuminuria. Multiple urine collections may be required for individual monitoring but not necessarily for diagnosis.

Keywords: Albuminuria; UACR; reliability; reproducibility; type 2 diabetes; urine albumin-creatinine ratio; within-individual variability.

MeSH terms

  • Aged
  • Albuminuria* / diagnosis
  • Albuminuria* / urine
  • Cohort Studies
  • Creatinine* / urine
  • Cross-Sectional Studies
  • Diabetes Mellitus, Type 2* / complications
  • Diabetes Mellitus, Type 2* / diagnosis
  • Diabetes Mellitus, Type 2* / urine
  • Diabetic Nephropathies* / diagnosis
  • Diabetic Nephropathies* / urine
  • Female
  • Humans
  • Male
  • Middle Aged


  • Creatinine