Creatinine and urine-output acute kidney injury criteria at a 24-hour landmark and subsequent severe kidney progression in critically ill adults with cirrhosis: a two-database cohort study

In the intensive care units of hospitals around the world, doctors constantly monitor the kidneys of patients who are critically ill. For people with cirrhosis, a severe scarring of the liver, this monitoring is especially urgent because their kidneys are often the next organ to fail. The standard way to check kidney health involves looking at two things: a blood test that measures a waste product called creatinine, and a count of how much urine the patient produces. In a healthy body, creatinine levels rise when the kidneys struggle to filter blood, and urine output drops when the kidneys cannot process fluids. However, in patients with cirrhosis, the blood test can be misleading. Because these patients often have very little muscle mass, their bodies produce less creatinine to begin with, so a rising level might not look as high as it should, hiding the true severity of the damage. Doctors have long suspected that watching urine output might provide a clearer picture, but it has been difficult to know exactly how much weight to give that second clue when the first one seems quiet.

A new study set out to solve this puzzle by looking at a specific window of time for patients who were already in the intensive care unit. The researchers focused on adults with cirrhosis who had survived their first day in the hospital without needing immediate dialysis or having reached a severe stage of kidney failure. They asked a simple but vital question: if a patient shows signs of trouble on just one of the two tests, or on both, does that predict what will happen in the next two days? By treating the first twenty-four hours as a starting line, the team could see how the combination of these two signals—the blood test and the urine count—forecasted the future health of the kidneys.

The researchers analyzed data from two large collections of medical records from hospitals in the United States. One collection came from a single academic medical center, while the other gathered information from over a hundred different hospitals. They looked at nearly 1,000 patients from the first center and about 670 from the second. For every patient, the team checked their status at the twenty-four-hour mark. They counted how many of the warning signs were present: zero, one, or two. A zero meant the patient looked fine on both counts. A one meant they had a problem with either the blood test or the urine output, but not both. A two meant they had warning signs on both fronts. The goal was to see if this simple count could predict whether the patient would develop severe kidney failure or need dialysis in the following forty-eight hours.

The results showed a clear and steady pattern. In both groups of hospitals, the more warning signs a patient had at the twenty-four-hour mark, the more likely they were to get worse. Among the patients who had no warning signs at all, about 19 percent in the single-center group and 10 percent in the multi-center group went on to develop severe kidney problems. For those with one warning sign, the risk jumped to roughly 33 percent and 20 percent, respectively. For the small group of patients who showed problems on both the blood test and the urine count, the risk was highest, reaching nearly 47 percent in the single-center group and 50 percent in the multi-center group. The study found that for every additional warning sign a patient had, their risk of severe progression increased significantly, regardless of which hospital database was being examined.

The researchers were careful to explain that this connection does not prove that the warning signs cause the failure, nor does it suggest a new rule for how doctors should treat patients. The study was observational, meaning it watched what happened naturally without changing any treatments. The team also noted that the strength of the link depended on how the urine output was measured. In some cases, using the patient's actual weight rather than a standard ideal weight changed the classification of who was at risk, which altered the numbers slightly but did not change the overall trend. The study also highlighted that the two databases were not perfectly identical; one had more detailed historical blood test records than the other, which is why the researchers analyzed them separately rather than mixing the numbers together.

Ultimately, the work suggests that doctors should not rely on just one of these two signals when assessing a cirrhosis patient in the intensive care unit. If a patient has a normal blood test but is producing very little urine, or vice versa, they are at a higher risk of rapid decline than someone with no signs on either test. The most dangerous group is the one showing trouble on both fronts. While the study does not offer a magic formula for saving lives, it provides a clearer map of the short-term risks. It indicates that paying attention to the combination of these two routine checks can help medical teams identify patients who are likely to deteriorate quickly, giving them a better chance to intervene before the kidneys fail completely. The findings serve as a reminder that in the complex landscape of critical illness, looking at multiple pieces of evidence together often reveals a truth that a single test might miss.