Biomarker-Based Stratification for Risk of Long-Term Outcome after Acute Kidney Injury

Can Biomarkers Predict Long-Term Risk After Acute Kidney Injury?

 

Acute kidney injury (AKI) is incredibly common, affecting up to 1 in 5 hospitalised patients, and its consequences don’t end at discharge. Many patients go on to develop chronic kidney disease (CKD), require dialysis, or even die within years of the initial episode.

Yet one key challenge remains – How do we identify which patients are truly at risk after AKI? A new study by Noble et al. (2026) brings us a step closer to answering that question.

 

 

Moving Beyond Creatinine

 

Traditionally, clinicians rely on serum creatinine and eGFR to assess kidney function. But these markers alone are blunt tools—they often fail to capture the biological processes driving long-term damage. Instead, this study focuses on a panel of four biomarkers:

 

• sTNFR1 and sTNFR2 (markers of inflammation and injury)
• Cystatin C (a sensitive marker of kidney function)
• eGFR (standard clinical measure)

 

Together, they offer a more nuanced snapshot of kidney health during recovery.

 

 

The Study

 

This was a prospective cohort study of patients with AKI, followed for one year. Biomarkers were measured at:

 

• 30 days
• 60 days
• 90 days post-AKI

 

The researchers then asked a critical question – Can these biomarkers predict major adverse kidney events (MAKE) or disease progression at one year?

 

 

Key Findings – Strong Predictive Performance

 

At 90 days post-AKI, the 4-biomarker panel achieved:

 

• AUC = 0.79 for predicting adverse outcomes

 

This level of performance is considered good discrimination, a strong result in clinical prediction modelling.

 

 

Excellent “rule-out” capability

 

What really stands out is the model’s ability to identify low-risk patients:

 

• Sensitivity: 100%
• Negative predictive value (NPV): 100%

 

In simple terms: if the test says you’re low risk, it’s highly likely to be correct. This opens the door to reducing unnecessary follow-up for patients unlikely to deteriorate.

 

Interestingly, the model performed almost as well at earlier timepoints. This suggests clinicians may not need to wait 3 months to assess long-term risk.

 

• Day 30: AUC 0.75
• Day 60: AUC 0.83

 

 

Biomarker levels tell a story

 

• Biomarker levels are highest during AKI
• They fall by Day 30 and then stabilise
• Patients with worse outcomes consistently show higher levels

 

This reinforces the biological relevance of the markers, not just statistical associations. It has also be noted in the study that the addition of H-FABP and Midkine may increase predictive performance.

 

 

Why this matters?

 

Healthcare systems face a major challenge:

 

• Millions survive AKI each year
• Not all need intensive follow-up
• Resources are limited

 

This study suggests a solution that allows the use of biomarkers to personalise care into the below categories, in other words, a shift toward precision nephrology.

 

• Low-risk patients → minimal follow-up
• High-risk patients → closer monitoring and early intervention