March 3, 2026Open Access

To explore the performance of ultrasound elastography in staging diabetic kidney disease: a systematic review and meta-analysis

Q: What is the clinical evidence from this study?

Study design: Meta-Analysis. Population: Diabetic kidney disease (n=2814). Intervention: Ultrasound elastography (USE) vs. Healthy participants and type 2 diabetic patients without nephropathy. Primary outcome: Diagnostic accuracy for differentiating five stages of DKD from two non-DKD groups (AUC 0.88, 95% CI 0.85-0.90).

Key Result

Ultrasound elastography differentiated diabetic kidney disease from non-diabetic groups with a pooled sensitivity of 79%, specificity of 83%, and an area under the curve of 0.88.

Key Points

Ultrasound elastography identifies five stages of diabetic kidney disease with an area under curve of 0.88, indicating high diagnostic accuracy.
Sensitivity and specificity for differentiating stages 3 to 5 of diabetic kidney disease were 80% and 79%, respectively, suggesting its reliability.
Analysis involved 18 studies and 2814 participants, providing a comprehensive evaluation of various USE parameters for kidney disease staging and assessment methods used included QUADAS-2 for risk of bias evaluation and GRADE for certainty of evidence assessments.

Study Design

Type

Meta-Analysis (n=2,814)

Structured PICO

Does quantitative ultrasound elastography accurately differentiate stages of diabetic kidney disease?

Population

2814 participants from 18 studies, including patients with five stages of diabetic kidney disease (DKD), healthy participants, and type 2 diabetic patients without nephropathy

Intervention

Quantitative ultrasound elastography (USE) evaluating cortical stiffness

Comparator

Healthy participants, type 2 diabetic patients without nephropathy, and inter-stage comparisons

Outcome

Diagnostic accuracy (sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio, and area under curve) for differentiating stages of DKDsurrogate

Ultrasound elastography shows promise as a noninvasive and quantitative diagnostic tool with good accuracy for staging diabetic kidney disease.

Main Result

Effect estimate: AUC 0.88 (95% CI 0.85-0.90)

Limitations

Exclusion of patients with prevalent diabetes-related comorbidities (e.g., hypertension, cirrhosis) introducing potential selection bias.
Inability to perform meta-analysis comparing macroalbuminuria to other patients due to the absence of reported 2x2 diagnostic data.
Low number of primary studies assessing diabetic individuals without DKD.
Did not evaluate other elastography parameters such as Emax and Emin.
Low inter-observer agreement due to reliance on operator expertise.
Measurements can be affected by ROI depth, geometry of acquisition, subcutaneous fat thickness, and BMI.

Abstract

To assess the diagnostic performance of different quantitative ultrasound elastography (USE) parameters in differentiating five stages of diabetic kidney disease (DKD) from each other as well as from healthy participants and type 2 diabetic patients without nephropathy. The protocol was pre-registered at https://osf.io/dfu8b/ . Using PubMed, Web of Science, Embase, and Cochrane Library, we found studies up to October 10, 2023. Risk of bias was assessed using Quality Assessment of Diagnostic Accuracy Studies-2 (QUADAS-2). The USE parameter of cortical stiffness was compared in an inter-stage manner based on mean differences, standardized mean differences, and ratios of means. In addition, cortical stiffness was evaluated for its diagnostic accuracy in staging DKD. Sensitivity and publication bias analyses were also conducted. Certainty of evidence was assessed using Grading of Recommendations, Assessment, Development, and Evaluation (GRADE). Eighteen studies were included with 2814 participants. The mean stiffness value consistently increased with the progression of the disease. Differentiation of five stages of DKD from two non-DKD groups, respectively, showed sensitivity, specificity, positive likelihood ratio, negative likelihood ratio, diagnostic odds ratio, and area under curve of 79%, 83%, 4.70, 0.25, 18.68, and 0.88. These values for differentiation of stages 3 to 5 of DKD from the other four milder groups were 80%, 79%, 3.85, 0.25, 15.53, and 0.87, respectively. USE shows promise as a noninvasive, simple, and quantitative diagnostic tool for staging DKD, especially in its early stages. With standardized protocols, its application could benefit patient management by enhancing the accuracy of DKD staging.

Bookmark

View Full Paper