# Urine Matrix Metalloproteinase-7 (MMP-7) Versus Urine Albumin-to-Creatinine Ratio (ACR) as Predictors of Renal Dysfunction: A Decision Curve Analysis

## TL;DR

This study compares urine MMP-7 and urine ACR as biomarkers for kidney dysfunction and finds that MMP-7 performs better in predicting renal impairment.

## Contribution

The study introduces urine MMP-7 as a more effective predictor of renal dysfunction compared to the widely used urine ACR.

## Key findings

- Urine MMP-7 showed stronger correlation with serum creatinine than urine ACR.
- Urine MMP-7 had higher sensitivity, specificity, and diagnostic accuracy compared to urine ACR.
- Decision curve analysis confirmed higher net benefits of urine MMP-7 over urine ACR for predicting renal impairment.

## Abstract

Background and objectives: The level of matrix metalloproteinase-7 (MMP-7) in diabetic urine samples escalates owing to reduced renal function. Renal biopsy is rarely recommended due to its invasive nature. Nowadays, urine albumin-to-creatinine ratio (ACR) is widely used to assess renal impairment. We mapped this study to compare urine MMP-7 and urine ACR as indicators of renal impairment.

Methods: This cross-sectional study was conducted at Kalinga Institute of Medical Sciences (KIMS), Bhubaneswar, India, from February 2020 to January 2023. Adult patients with either type 2 diabetes mellitus (T2DM), kidney disease, or hypertension were scrutinized. Their serum creatinine, urine albumin, urine creatinine, urine ACR, and urine MMP-7 levels were evaluated. We correlated serum creatinine values with urine ACR and urine MMP-7 levels. For predictive modeling, we developed two models: ACR_model and MMP7_model. The ACR_model and MMP7_model weighed renal activity by analyzing participants with urine ACR > 30 mg/g and normalized urine MMP-7 > 10 µg/L, respectively. Each model's predictive accuracy was computed using the area under the receiver-operating characteristic (ROC) curve (AUC). For predictive modeling, we deployed the bootstrap method and decision curve analysis. We used R software (version 4.4.2) for data analysis.

Results: A total of 287 (87.5%) of the 328 patients we scrutinized were deemed eligible for the study. We found statistically significant correlation coefficients of serum creatinine with urine MMP-7 than with urine ACR. It suggested a stronger association of serum creatinine with urine MMP-7. The study revealed that urine ACR had lower values of the following parameters than urine MMP-7: sensitivity (81.3% versus 86.7%), specificity (64.4% versus 68.2%), and diagnostic accuracy (78.2% versus 86.6%). The decision curve analysis unveiled that urine MMP-7 demonstrated higher net benefits juxtaposed with urine ACR, regardless of the threshold probability.

Conclusion: This study analyzed the role of urine ACR and urine MMP-7 as biomarkers of renal failure. We discovered stronger correlations between serum creatinine and urine MMP-7 contrasted with urine ACR. Urine MMP-7 offered better specificity, sensitivity, and diagnostic accuracy than urine ACR. The decision curve analysis also revealed that urine MMP-7 outperformed urine ACR in forecasting renal impairment.