📄 medicine

Effect of Proton Pump Inhibitors on the Development of Acute Kidney Injury in Critically Ill Patients: A Clinical Trial Emulation

This retrospective cohort study utilizing the MIMIC-IV database found that critically ill patients receiving proton pump inhibitor (PPI) prophylaxis have a significantly higher risk of developing acute kidney injury compared to those not exposed, with an adjusted odds ratio of 1.78.

Original authors: Henry Oliveros, Jessica Barrera, Eduardo Tuta-Quintero, Henry Robayo-Amortegui, Luis Felipe Reyes

Published 2026-07-09

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Henry Oliveros, Jessica Barrera, Eduardo Tuta-Quintero, Henry Robayo-Amortegui, Luis Felipe Reyes

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

Imagine the Intensive Care Unit (ICU) as a high-stakes, chaotic construction site. The patients are the fragile structures being repaired, and the medical team is the crew trying to keep everything standing. In this environment, there's a constant fear of "leaks" (bleeding) from the stomach, so the crew often hands out a specific type of sealant called Proton Pump Inhibitors (PPIs) to everyone, just in case.

This study is like a detective investigation that asked a simple question: Is this "preventative sealant" accidentally damaging the building's plumbing (the kidneys)?

Here is the story of what the researchers found, broken down into everyday concepts:

The Setup: A Massive Digital Archive

The researchers didn't go into a hospital to watch patients; instead, they went into a giant digital library called MIMIC-IV. This library holds the medical records of over 53,000 people who had been in the ICU at a specific hospital (Beth Israel Deaconess Medical Center) between 2008 and 2019.

They filtered this massive pile of records down to 8,161 patients. They threw out anyone who:

Was in the ICU for less than 4 days (too short a time to see the damage).
Already had broken plumbing (kidney failure) before they arrived.
Were taking other similar drugs that would confuse the results.

The Experiment: Emulating a "What If" Scenario

Since you can't ethically force some patients to take a drug and others not to in a real-life emergency, the researchers used a clever computer trick called "Clinical Trial Emulation."

Think of it like a video game simulation. They took the real-world data and said, "Okay, let's pretend we ran a perfect experiment." They used a mathematical tool called Inverse Probability Weighting (IPW).

The Analogy: Imagine you have two groups of people: Group A (who got the PPIs) and Group B (who didn't). In real life, Group A was probably sicker to begin with. The IPW tool acts like a digital scale, adding "weight" to the healthier people in Group B and "lightening" the burden on the sicker people in Group A, so the two groups look as identical as possible on paper. This way, if one group gets sick, it's likely because of the drug, not because they were already weaker.

The Findings: The "Leak" in the Plumbing

The investigation revealed a clear connection between the sealant and the plumbing damage.

The Risk: Patients who received PPIs were 1.27 times more likely to develop Acute Kidney Injury (AKI) compared to those who didn't, even after the researchers balanced the groups and accounted for other illnesses.
- Simple Math: If 100 people without the drug got kidney trouble, about 127 people with the drug got it. It's not a guarantee, but the risk goes up noticeably.
The Real Culprits (The "Big Bad" Villains): The study found that while PPIs were a problem, they weren't the only problem. The biggest threats to the kidneys were:
- Vancomycin (an antibiotic): This was the strongest villain, making kidney injury 2.4 times more likely.
- Aminoglycosides (another antibiotic): Made it 1.55 times more likely.
- Contrast Media (the dye used for X-rays): Made it 1.44 times more likely.
- Sepsis (a severe body-wide infection): Made it 1.47 times more likely.

The "Negative Control" Check

To make sure their detective work was sound, the researchers used a "negative control." This is like checking if a smoke detector is working by seeing if it goes off when there is no fire.

They checked if PPIs caused Ventilator-Associated Pneumonia (a lung infection).
The Result: After adjusting for other factors, PPIs did not significantly increase the risk of pneumonia in this specific analysis. This suggests that the link between PPIs and kidney injury is real and specific, not just a sign that the drug makes everything worse.

The Bottom Line

The study concludes that while PPIs are often given to ICU patients to prevent stomach bleeding, they act like a double-edged sword. They might stop a leak in the stomach, but they seem to increase the chance of a leak in the kidneys.

The authors suggest that doctors should be very careful about handing out these "sealants" to critically ill patients, especially when those patients are already being hit by other kidney stressors like strong antibiotics, infection, or contrast dye. It's a reminder that in the ICU, every medicine is a trade-off, and sometimes the cure for one problem can create a new one.

Technical Summary: Effect of Proton Pump Inhibitors on the Development of Acute Kidney Injury in Critically Ill Patients

Problem Statement
Acute Kidney Injury (AKI) represents a significant clinical and economic burden in Intensive Care Units (ICUs), with incidence rates ranging from 30% to 70% depending on classification criteria. While various factors contribute to AKI, including antibiotics, contrast agents, and nonsteroidal anti-inflammatory drugs, the role of Proton Pump Inhibitors (PPIs) remains controversial. PPIs are frequently administered to critically ill patients for gastrointestinal bleeding prophylaxis, particularly those with risk factors such as mechanical ventilation, sepsis, and shock. However, existing literature presents contradictory findings regarding whether PPI exposure increases the risk of AKI. Some studies suggest no increased risk, while others indicate a correlation, particularly in specific surgical populations. This study aims to resolve this uncertainty by determining the association between PPI exposure and the development of AKI in critically ill patients during the first seven days of ICU admission.

Methodology
The study employed a retrospective cohort design using data from the MIMIC-IV database (version 2.2), encompassing patients admitted to the Beth Israel Deaconess Medical Center (BIDMC) between 2008 and 2019. The researchers utilized a "clinical trial emulation" framework, adhering to guidelines established by Hernán to minimize bias inherent in observational studies.

Study Population: From an initial pool of 53,569 ICU patients, the final cohort consisted of 8,161 adults (≥18 years) who had no history of prior renal insufficiency, no prior PPI or H2 blocker use upon hospitalization, and an ICU stay of at least 96 hours.
Exposure Definition: Patients were classified as "exposed" if they received PPIs during their ICU stay, with treatment initiation defined as time zero. Those who did not receive PPIs were "unexposed." Patients with prior PPI use or concomitant H2 blocker use were excluded to avoid prevalence and immortal time biases.
Outcome: The primary outcome was the development of AKI (Stages I, II, or III) within seven days of exposure, defined by changes in serum creatinine relative to baseline.
Statistical Analysis: To address confounding, the study employed multivariate logistic regression and Inverse Probability Weighting (IPW) based on propensity scores. Covariates included demographic factors, comorbidities (Charlson index), severity scores (SAPS II, SOFA), and exposure to other nephrotoxic agents (e.g., vancomycin, aminoglycosides, contrast media).
Residual Confounding Check: A negative control outcome, ventilator-associated pneumonia (VAP), was analyzed to evaluate the robustness of the causal inference methods.

Key Results
The final analysis included 8,161 patients, of whom 3,592 (44%) received PPI prophylaxis and 4,569 (56%) did not.

AKI Incidence: AKI developed in 855 patients (10.48%). The incidence was significantly higher in the PPI-exposed group (13.5%) compared to the unexposed group (8.1%) ( $p < 0.001$ ).
Risk Association:
- Crude Odds Ratio (OR): The unadjusted OR for AKI in PPI-exposed patients was 1.78 (95% CI: 1.54–2.05).
- Adjusted OR: After adjusting for demographic variables, comorbidities, and severity scores via multivariate logistic regression, the OR was 1.27 (95% CI: 1.09–1.48).
- IPW OR: Using inverse probability weighting to balance baseline characteristics, the OR remained 1.27 (95% CI: 1.08–1.48).
Other Risk Factors: The strongest associations with AKI were observed with vancomycin (OR 2.40), aminoglycosides (OR 1.55), and contrast media (OR 1.44). Sepsis and metastatic tumors also showed significant associations.
Negative Control: The association between PPIs and VAP was not statistically significant after adjustment (Adjusted OR 1.01; 95% CI: 0.83–1.22), suggesting that the observed association with AKI is not merely a result of unmeasured confounding factors affecting general ICU outcomes.
Clinical Impact: Patients who developed AKI had a significantly higher ICU mortality rate (39.6%) compared to those without AKI (15.1%) and a longer average length of stay (11.9 days vs. 8.0 days).

Key Contributions

Clinical Trial Emulation: The study rigorously applied causal inference methods to emulate a randomized clinical trial, specifically defining "time zero" at ICU admission and excluding patients with prior PPI use to mitigate selection and immortal time biases.
Specificity of Exposure: Unlike previous studies that grouped PPIs and H2 blockers together, this research isolated PPI exposure, providing a clearer picture of their specific renal risks.
Robust Adjustment: The use of both multivariate regression and IPW, combined with a negative control outcome (VAP), strengthens the validity of the causal link between PPIs and AKI in a critically ill population.
Identification of Confounding Trends: The study highlighted a discrepancy in clinical practice, noting that 53.4% of patients with valid indications for prophylaxis did not receive PPIs, while 37.4% of those who received them lacked a clear indication.

Significance and Claims
The paper concludes that there is a clear association between PPI exposure and the development of acute kidney injury in critically ill patients, even after rigorous adjustment for confounding variables. The authors assert that while PPIs are a contributing factor, they act in conjunction with other significant risk factors such as the administration of nephrotoxic antibiotics (vancomycin, aminoglycosides), contrast media, and the presence of sepsis and liver disease.

The significance of these findings lies in the call for a more precise and personalized approach to PPI administration in the ICU. The authors suggest that prophylaxis should be carefully considered, particularly when patients are simultaneously exposed to other factors that contribute to renal injury. The study does not claim PPIs are the sole cause of AKI but emphasizes that their use adds to the cumulative risk profile of critically ill patients, warranting a re-evaluation of current prophylaxis protocols.