Even when patients who were recovering from serious acute kidney injury (AKI) had low creatinine clearance, discontinuation of dialysis could be feasible and safe, new research showed.

Furthermore, creatinine clearance appeared to be a more accurate predictor of recovery than urine output or estimated glomerular filtration rate (eGFR).

“Traditional notions of creatinine clearance or urine output [cutoffs] may overestimate subsequent dialysis need in patients with dialysis-requiring AKI who no longer require vasopressor support or mechanical ventilation,” first author Ian McCoy, MD, assistant professor in the Division of Nephrology at the University of California, San Francisco, told Medscape Medical News.

Many patients did not require dialysis during the week after having creatinine clearance < 10 mL/min, according to the study, which was published in the Journal of the American Society of Nephrology.

Dialysis-requiring AKI, the most severe form of AKI, is associated with an increased risk for morbidity and mortality and is commonly treated with a standard regimen of thrice-weekly hemodialysis.

However, results from the recent Liberation from Acute Dialysis (LIBERATE-D) trial showed that a more conservative strategy of providing dialysis only as needed in response to key clinical indications resulted in better outcomes. These included a shorter time to and higher rates of kidney function recovery, with no significant increase in adverse events.

Potential explanations for the improvement include that dialysis may mask true kidney recovery and can cause adverse effects of its own.

Urine collection metrics, including creatinine clearance and urine output, are among the clinical indicators used to guide dialysis decisions. However, specific thresholds associated with safe dialysis discontinuation have not been well established.

Creatinine Proved the Superior Predictor

To investigate, McCoy and colleagues conducted an ancillary study of the LIBERATE-D trial, evaluating 147 timed urine collections from 85 patients with dialysis-requiring AKI in the intervention arm who received dialysis only as needed in response to the clinical indicators.

Although most patients had experienced critical illness earlier in their hospitalization, 64% were no longer in the ICU during their first timed urine collection, indicating that their critical illness was resolving.

The timed urine collections showed a median creatinine clearance of 7.1 mL/min and a median urine output of 930 mL/d.

Dialysis was performed within 7 days after urine collection in 49% of cases.

Creatinine clearance exhibited significantly greater discrimination power for predicting which patients remained dialysis free for 7 days after urine collection than urine output (area under the curve [AUC], 0.84 vs 0.78, respectively).

Among those with creatinine clearance > 20 mL/min, 96% remained dialysis free for 7 days after urine collection. The rate remained at 81% among patients with a creatinine clearance of 10-20 mL/min, which is indicative of severe impairment.

Even among patients with a creatinine clearance of 5 mL/min to < 10 mL/min, suggestive of stage 5 chronic kidney disease, 44% did not require dialysis during the next 7 days.

The odds of requiring dialysis decreased by 18% for every 1-mL/min increase in creatinine clearance.

For urine output, the odds of requiring dialysis within 7 days decreased by 9% for every additional 100 mL of urine. Among patients with a urine output > 1000 mL/d, 73% remained dialysis free for a week compared with 28% of those with a urine output of 200-499 mL/d.

Among 75 patients who did not receive dialysis within 7 days after urine collection, only two required dialysis again during hospitalization, at a median of 14 days later. In both cases, the creatinine clearance was > 8 mL/min and urine output was > 1000 mL/d, suggesting recurrent AKI rather than failure to recover.

Urea clearance and the average of urea and creatinine clearances had predictive performance for dialysis within 7 days similar to that of creatinine clearance alone, with AUC values of 0.82 and 0.84, respectively.

By contrast, eGFR exhibited significantly lower predictive performance than creatinine clearance or urine output, with an AUC of 0.67. Measuring urea clearance in addition to creatinine clearance did not improve prognostic performance.

Results Represent an ‘Actively Improving’ Population

McCoy said the low creatinine clearance levels associated with safe discontinuation were “a surprise,” particularly when compared with other recommendations.

He emphasized that the differences likely reflected the clinical status of the study population. Most patients were no longer in the ICU and were “actively improving,” without the need for vasopressor support or mechanical ventilation, unlike patients in many previous studies of dialysis discontinuation.

“This is important because many dialysis-requiring AKI patients do not survive ICU admission or have recurrent AKI during ICU admission, and therefore, studies from the ICU population will have more pessimistic estimates of the odds of achieving dialysis independence,” he said.

The findings therefore should not be generalized to patients who remain critically ill.

Dialysis Discontinuation May Help ‘Unmask’ Kidney Recovery

The potential challenges of switching to an as-needed dialysis approach include more complicated planning for case managers and pharmacists. Clinicians may also reasonably be concerned about complications from discontinuing dialysis too early, McCoy noted.

However, the LIBERATE-D trial showed that a more conservative approach could give patients a greater chance of achieving dialysis independence.

In fact, conservative treatment is the only treatment ever shown “to foster faster recovery” in patients with dialysis-requiring AKI, said the trial’s first author Chi-yuan Hsu, MD, also from the Division of Nephrology at the University of California, San Francisco, while presenting the LIBERATE-D trial at Kidney Week 2025: American Society of Nephrology Annual Meeting, as previously reported by Medscape Medical News.

Results Confirm Suspicions Regarding Stable Patients

Commenting on the new analysis, Lui G. Forni, MD, PhD, professor of intensive care at the University of Surrey’s School of Medicine in Surrey, England, said the finding that creatinine clearance outperformed urine output and eGFR confirms suspicions that thresholds in current guidelines do not apply to a relatively stable population.

Forni said the stronger performance of creatinine clearance made sense.

“Urine output can stay deceptively preserved at very low GFR because of impaired tubular reabsorption. Therefore, you may guess that creatinine clearance outperforms it. The authors have shown it does,” he told Medscape Medical News.

Importantly, “the cohort in the study was captured later in the clinical course than almost every prior study on this question,” Forni said.

One caveat was the limited duration of follow-up.

“Although this is a pragmatic approach, it does not answer whether the patients have achieved durable kidney recovery. For that, a longer-term [endpoint] would be needed,” Forni said.

Reluctance Is Reasonable

Although discontinuing dialysis at such low levels of creatinine clearance may give clinicians pause, Forni emphasized the importance of daily monitoring.

“Note that at creatinine clearance levels of 5 to < 10 mL/min [in the study], dialysis was needed again within the week 56% of the time, but daily reassessment and an indication-triggered approach should be adopted,” he said.

Ultimately, dialysis should still be given when emergent criteria are reached, “just not scheduled prophylactically while the patient is stable and asymptomatic,” Forni said.

McCoy and Forni reported no disclosures.