乌鸦传媒

SAN DIEGO -- Early, personalized recommendations did not improve outcomes for hospitalized patients with acute kidney injury (AKI), the KAT-AKI randomized trial showed.

The composite outcome of progression to a higher AKI stage, dialysis, or mortality occurred in 19.8% of patients who received recommendations delivered via electronic health records by a team of physicians and pharmacists within 1 hour of AKI detection compared with 18.4% of patients who received usual care (P=0.28), reported Abinet Aklilu, MD, MPH, of Yale University School of Medicine in New Haven, Connecticut, at the American Society of Nephrology Kidney Week meeting.

The proportion of patients progressing to any of the individual outcomes of the composite also did not differ between the intervention and usual care groups:

• AKI progression: 13.5% vs 13.0%
• Dialysis receipt: 1.6% vs 1.5%
• Mortality: 9.6% vs 9.2%

This held true for subgroups stratified by age, sex, baseline comorbidities, location, hospital site, baseline creatinine strata, and recommendation categories, Aklilu and colleagues noted in , where the findings were simultaneously published.

Affecting one in five hospitalized patients, AKI is associated with worse outcomes including death, chronic kidney disease (CKD), and higher healthcare costs, Aklilu said. Clinical decision support tools, like the one used in the trial, have been incorporated as the "standard of AKI care in some countries," she added.

However, "there is no standard of care for AKI notification in the U.S. or in any of the hospitals that participated in this study," the researchers wrote.

For the trial, the researchers included 4,003 patients who were hospitalized within the Yale New Haven Health System (80%) or Johns Hopkins Health System (20%) from October 2021 to February 2024. AKI cases were identified via an electronic alert algorithm that compared each creatinine value measured during hospitalization to the lowest creatinine value in the preceding 7 days and to creatinine values in the past 48 hours using KDIGO criteria. The team of 20 physicians and 12 pharmacists then reviewed patients' electronic health records to confirm AKI.

Median age was 72, 47% were women, 66.5% were white, and 23% were Black. Of these patients, 47.1% had diabetes, 41.6% had CKD, 81.1% had hypertension, 44.6% had heart failure, and 6.7% had cirrhosis.

After AKI confirmation, the physician and pharmacist provided a set of recommendations focused on evaluation and management of AKI, all individualized for each patient. Of note, these recommendations were only provided if they weren't already ordered or performed by the patient care team. For example, they wouldn't recommend a kidney ultrasound if one was already done.

The team then entered their recommendations in a form, which addressed five major areas of AKI care -- general AKI evaluation, volume, potassium, acid-base management, and medication modifications -- and the need for nephrology consultation. "The physician recommendation form contained 34 binary options for recommendations that included 11 potential diagnostic tests, 10 interventions for volume management, 7 for potassium management, 5 for acidosis, and a single category for a nephrology consult recommendation," the researchers wrote. "Additional recommendations via free text were allowed."

The pharmacist focused on two categories: medications to be discontinued or substituted, as well as medications that required dose adjustment or monitoring. Physician and pharmacist notes were automatically generated in the patient's electronic health record within the hour. Median time from AKI diagnosis to randomization was 56 minutes, and patients received a median of three recommendations.

As for the half of participants randomized to usual care, they also had recommendations made for them by the KAT-AKI team prior to randomization, but these were never delivered to the care team, and were stored and used for the analysis.

No adverse events occurred, but there was one report of an inappropriate intravenous fluid recommendation for a patient in the intervention group whom the team believed had volume overload. However, the patient didn't receive the recommended intervention.

There were no significant differences between the intervention and usual care groups for rates of discharge to hospice (4.4% vs 5.3%) or nephrology consults within 14 days of AKI detection (16.1% vs 14.2%).

Looking at secondary outcomes, significantly more recommendations were implemented in the intervention group compared with the usual care group within 24 hours (33.8% vs 24.3%, P<0.001). This was largely driven by a higher rate of general (30.2% vs 16.5%) and medication-related (43% vs 27.1%) recommendations. While this didn't improve patient outcomes, Aklilu said this does show that the "rapid, personalized recommendation delivery for AKI ... was feasible and ... improved the rate of AKI best-practice implementation."

One limitation to the study was the dependence on a clinician checking the patient's chart in a timely manner to implement the recommendations.

Comment