Nephrotoxic Drug Exposure and Acute Kidney Injury Risk in Hospitalised Patients: A MIMIC-IV Cohort Study
This MIMIC-IV cohort study demonstrates that exposure to nephrotoxic drugs, particularly in polypharmacy combinations, significantly increases the risk of acute kidney injury, hospital length of stay, and mortality, while establishing a polypharmacy score as a practical tool for real-time risk stratification and prevention.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Inside the walls of a modern hospital, a silent complication often strikes patients who are already fighting for their health. It is called acute kidney injury, a sudden failure of the kidneys to filter waste from the blood. This condition is not rare; it affects a significant portion of people admitted to hospitals, and when it happens, it makes recovery harder, longer, and more dangerous. While many factors can trigger this failure, one major cause sits squarely in the medicine cabinets of the hospital itself. Certain drugs, prescribed to treat infections, pain, or heart conditions, can accidentally damage the kidneys. This is known as nephrotoxicity. The challenge for doctors has long been knowing exactly when a patient is walking a dangerous line, especially when they are taking several of these medications at once.
For decades, medical textbooks have listed which individual drugs carry this risk. Doctors know that some antibiotics, pain relievers, and diuretics can harm the kidneys if used carelessly. However, a critical piece of the puzzle remained missing. No one had looked at the big picture to see how these drugs interact when given together to thousands of patients. It was unclear how the risk builds up when a patient takes multiple nephrotoxic drugs, or which specific combinations create the most immediate danger. Without this large-scale map, hospitals could not easily predict which patients were at the highest risk before the damage occurred, leaving them to react only after the kidneys had already started to fail.
To fill this gap, researchers turned to a massive digital archive of hospital records known as MIMIC-IV. This database contains the complete medical histories of over 300,000 hospital admissions from a major medical center in Boston, spanning more than a decade. The researchers focused on adults who stayed in the hospital for at least two days and had their kidney function monitored through blood tests. They carefully removed patients who already had severe, long-term kidney disease before they arrived, ensuring that any kidney injury found during the stay was truly new. By filtering the data this way, they created a clean group of 280,468 admissions to study.
The team then built a detailed profile for every patient in this group, tracking exactly which potentially harmful drugs they received. They did not just look at whether a patient took a drug; they calculated how long the patient was on it, the total amount received, and most importantly, how many different types of harmful drugs were being used at the same time. They created a simple score to represent this mix, counting how many distinct categories of risky drugs a patient was exposed to. They also looked for specific pairs of drugs that are known to be dangerous when combined, such as certain antibiotics given alongside diuretics.
The results revealed a stark and clear pattern. Among the patients studied, nearly 17 percent developed acute kidney injury during their stay. The data showed that patients who were exposed to at least one nephrotoxic drug were more than three times as likely to develop this injury compared to those who were not exposed. But the most striking finding was the relationship between the number of drugs and the risk of injury. The risk did not just rise slightly; it climbed in a steady, predictable ladder. Patients who took no nephrotoxic drugs had a very low chance of injury. However, for every additional type of harmful drug added to their regimen, the risk jumped significantly. When a patient was exposed to six different categories of nephrotoxic drugs at the same time, the chance of developing kidney injury soared to nearly 96 percent.
This dose-response relationship held true even when looking at specific drug combinations. The researchers found that certain pairings were particularly deadly. For instance, the combination of an aminoglycoside antibiotic and a loop diuretic resulted in a kidney injury rate of nearly 48 percent. Another common and dangerous pairing, involving vancomycin and piperacillin-tazobactam, led to injury in over 42 percent of cases. These numbers were far higher than the risk associated with taking any single one of these drugs alone. The study also noted that some drugs, like non-steroidal anti-inflammatory pain relievers, appeared to have a lower risk in the data, but the researchers explained this was likely because these drugs were often given to patients who were already less sick to begin with, rather than because the drugs were safe.
The consequences of this kidney injury were severe and measurable. Patients who developed the condition stayed in the hospital for an average of 13.7 days, which is more than double the six-day average for those who did not. The impact on survival was even more profound. The death rate for patients with kidney injury was nearly 11 percent, compared to just over 1 percent for those without it. This means that for every patient who suffered this injury, the hospital stay was extended by roughly seven and a half days, and the likelihood of dying in the hospital increased by nearly ten times.
The study concludes that the number of different nephrotoxic drugs a patient is taking is a powerful, real-time indicator of danger. Because this information is available the moment a doctor writes a prescription, it offers a practical tool for hospitals to identify high-risk patients immediately. By monitoring the count of these drugs, medical teams can intervene before the kidneys are damaged, potentially preventing thousands of extra hospital days and saving lives. The research provides a solid foundation for building computer systems that can alert doctors to these risks automatically, turning a complex medical problem into a manageable one through simple observation and timely action.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.