Early Nephrology Consultation Didn’t Improve AKI Outcomes

Key Takeaways

• At a single U.S. center, hospitalized adults without AKI who were flagged as high risk by ESTOP-AKI had no significant between-group differences in serum creatinine rise or major downstream kidney outcomes with early nephrology consultation.
• During the 7-day study period, 38.9% of flagged patients developed AKI and 16.1% developed stage 2 or 3 AKI.
• Early nephrology consultation changed process measures far more than clinical outcomes, increasing specialist activity without a corresponding improvement in reported kidney endpoints.
• Recommendation follow-through for medication-related, fluid or diuretic, and vasopressor-related advice was lower with early consultation than with usual care, at 41% vs 68%.
• Investigators described ESTOP-AKI as a feasible tool for real-time EMR monitoring that identified hospitalized patients before creatinine-based injury was evident.

Hospitalized adults can be on a path toward acute kidney injury (AKI) before serum creatinine rises enough to make the problem clinically obvious, leaving a narrow window in which risk stratification might prompt earlier specialist input. ESTOP-AKI was used as a dynamic machine-learning score to flag inpatients at heightened risk for more severe kidney injury before overt AKI appeared, raising the practical question of whether earlier nephrology involvement could alter that trajectory. At the University of Chicago, investigators tested that question in a randomized trial.

Investigators at the University of Chicago reported a randomized AKI consultation trial in JAMA Network Open conducted from 2019 to 2024 in 180 hospitalized adults without AKI who were considered at high risk for stage 2 AKI with an ESTOP-AKI score of at least 0.01.

Patients were assigned to early nephrology consultation or usual care. Early consultation included an in-person assessment and clinical recommendations, whereas nephrology became involved in usual care only when the primary team requested it. The primary endpoint was peak change in serum creatinine from enrollment over 7 days, with secondary outcomes including AKI development, inpatient kidney replacement therapy, hospital length of stay, and inpatient and 90-day mortality.

AKI developed in 38.9% of patients during the 7-day study period, and 16.1% developed stage 2 or 3 AKI. Investigators reported no significant between-group differences in peak serum creatinine change, AKI development, inpatient kidney replacement therapy, readmission, or 90-day mortality.

Process measures shifted far more than clinical outcomes. The early-consultation arm generated 121 nephrology consultations with 270 recommendations. Usual care generated 19 consultations with 36 recommendations. Follow-through on medication-related, fluid or diuretic, and vasopressor-related recommendations was 41% in the early-consultation group versus 68% in usual care.

Because the trial was conducted at a single center, the findings are best framed as evidence from one inpatient system rather than as a broadly generalizable practice signal. In investigator remarks on recommendation uptake and score feasibility, Koyner attributed lower follow-through to the timing of consultation, saying recommendations arrived before creatinine-based AKI was clinically evident and may have seemed less urgent to primary teams. Investigators also described ESTOP-AKI as feasible for real-time electronic medical record (EMR) monitoring, even though the consultation strategy linked to that prediction did not improve outcomes in this trial.

The trial separated two related questions in AKI prevention: whether a high-risk inpatient population can be identified before overt injury, and whether an intervention tied to that prediction changes short-term kidney outcomes. In this study, early nephrology involvement expanded consultation activity without improving the reported clinical endpoints. Investigators said further research is needed on adherence to early recommendations and on related risk-stratification strategies.

Clinician Questions

Which hospitalized adults does the ESTOP-AKI early nephrology consultation trial apply to?

The randomized trial involved hospitalized adults at the University of Chicago who did not have AKI at enrollment but were considered at high risk for stage 2 AKI because their ESTOP-AKI score was greater than 0.01. The findings therefore apply to inpatient risk identification before overt AKI develops, not to patients who already had AKI at enrollment or to outpatient use.

How was AKI risk identified before creatinine-based injury was evident in the ESTOP-AKI trial?

Investigators used ESTOP-AKI as a dynamic machine-learning score that updated as laboratory values and vital signs changed, allowing AKI risk to be flagged before creatinine-based injury was clinically apparent. According to Koyner, the model had been validated in patient encounters across Chicagoland institutions and was feasible for real-time electronic medical record monitoring.

What questions did investigators say remain after early nephrology consultation failed to improve AKI outcomes?

Investigators said future research should examine whether greater adherence to early recommendations changes AKI outcomes, which specific guideline-based recommendations may be useful, and whether combining the risk score with novel AKI biomarkers can better identify high-risk hospitalized patients.