This cohort study examines the association of ibuprofen with acute kidney injury (AKI) risk in hospitalized adults and how the finding varies among patients with AKI risk factors.
Key Points
Question
Is use of ibuprofen associated with acute kidney injury (AKI) in hospitalized patients and do findings vary across subgroups?
Findings
In this cohort study of 78 656 hospitalized patients, ibuprofen was associated with higher rates of AKI, severe AKI, and kidney replacement therapy compared with oxycodone. Ibuprofen-associated AKI risk was greater among patients with critical illness and patients who were recovering from surgery.
Meaning
This study found that the excess risk of ibuprofen-associated AKI is concentrated in those already at high risk of AKI, suggesting that ibuprofen may be a reasonable alternative to opioids in patients with low AKI risk.
Abstract
Importance
Although nonsteroidal anti-inflammatory drugs (NSAIDs) are the most effective nonopioid analgesics for acute pain, they may be underused in the inpatient setting due to uncertainty regarding the risk of acute kidney injury (AKI). Whether inpatient exposure to NSAIDs, such as ibuprofen, substantially increases AKI risk and how ibuprofen-associated AKI varies across patient subgroups remain unknown.
Objective
To determine whether ibuprofen use is associated with AKI in hospitalized patients and whether the finding varies across AKI risk factor subgroups.
Design, Setting, and Participants
A retrospective active-comparator cohort study of hospitalized patients (aged ≥18 years) without end-stage kidney disease or contraindications to NSAID treatment admitted from January 1, 2004, to June 30, 2017, in a large academic health system including 3 hospitals in Philadelphia, Pennsylvania, with follow-up until 48 hours after treatment end, hospital discharge, or 14 days elapsed, whichever occurred first. Analyses were conducted from March 2023 through August 2026.
Exposures
At least 24 hours of treatment with ibuprofen or oxycodone for acute pain.
Main Outcomes and Measures
The primary outcome was AKI, defined by changes in serum creatinine and incident kidney replacement therapy (KRT). Balancing weights were used to adjust for baseline covariates and weighted Poisson regression used to estimate incidence rate ratios (IRRs) and 95% CIs.
Results
This study included 78 656 patients (median age, 59 [IQR, 47-69] years; 38 092 females [48%]) treated with ibuprofen (n = 6313) or oxycodone (n = 72 343). There were 4799 AKI events (79% stage 1) during 270 472 days of follow-up (17.7 events per 1000 days). After adjustment for baseline differences, patients treated with ibuprofen had a higher rate of any AKI (IRR, 1.22; 95% CI, 1.09-1.38), severe AKI (IRR, 1.36; 95% CI, 1.08-1.72), and KRT (IRR, 3.29; 95% CI, 1.35-8.03). IRRs were greater in postoperative (IRR, 1.70 [95% CI, 1.28-2.27] vs 1.12 [95% CI, 0.98-1.28]) and critically ill (IRR, 1.65 [95% CI, 1.29-2.11] vs 1.16 [95% CI, 1.02-1.33]) patients. In patients at low risk (those non–critically ill, nonpostoperative, aged <60 years without kidney dysfunction; ibuprofen, n = 2966; oxycodone, n = 20 440), ibuprofen was not associated with higher AKI risk (IRR, 0.94; 95% CI, 0.75-1.18).
Conclusions and Relevance
This study in an inpatient cohort found that ibuprofen was associated with higher rates of AKI and KRT, but this excess risk was concentrated in patients at high risk. Patients at low risk may be able to receive ibuprofen safely yet were often prescribed oxycodone.
Introduction
Treatment of acute pain is one of the most common interventions during hospitalization, occurring in tens of millions of patients annually.1,2,3 Although opioids are a mainstay of treatment, they have an extensive adverse-effect profile2,4 and inpatient exposure is linked to chronic dependence.5,6 This has created substantial interest in opioid minimization strategies, including the use of nonopioid analgesics as a part of multimodal regimens.4,7,8
Nonsteroidal anti-inflammatory drugs (NSAIDs) are the most effective nonopioid analgesics available for acute pain.4,7,9 However, NSAIDs may be underused in the inpatient setting due to concerns for adverse effects.8,10,11 One of the most important NSAID-associated adverse effects is acute kidney injury (AKI). NSAIDs cause nephrotoxic effects, primarily through vasoconstriction of afferent arterioles in the glomerulus, which can reduce kidney blood flow, leading to ischemic AKI in vulnerable patients.12 Most evidence regarding NSAID-AKI risk comes from outpatient settings, where patients often use NSAIDs for extended durations. In contrast, hospitalized patients typically receive NSAIDs for at most a few days.13 Meta-analyses of available clinical trials among inpatients have concluded that available evidence is generally low quality and insufficient to guide clinical decision-making.9,14 Available observational studies have been underpowered to examine key patient-centered outcomes such as severe AKI or need for kidney replacement therapy (KRT).13,15,16,17,18 The kidney-related safety of inpatient NSAID analgesia thus remains unclear.
Clinicians often avoid NSAIDs among inpatients with established AKI risk factors, such as increased age, kidney dysfunction, obesity, heart failure, and diabetes.19,20,21 However, to our knowledge, no studies have rigorously evaluated whether these factors alter NSAID-induced AKI risk. Characterizing relative contributions of preexisting AKI risk factors, NSAID-specific nephrotoxic effects, and their potential interactions is essential for developing individualized analgesic regimens that balance the risks and benefits of NSAIDs and opioids. Given the large population of inpatients treated for pain each year,1 and the high prevalence of AKI risk factors in this population, the potential impact on kidney disease is substantial. We conducted a large multihospital cohort study to examine whether ibuprofen, a commonly used NSAID for analgesia in the inpatient setting, is associated with AKI risk and to examine whether findings vary among patients with AKI risk factors.
Methods
Study Design
This cohort study compared AKI rates in hospitalized patients treated with ibuprofen with those treated with oxycodone, an active comparator.22 We focused on ibuprofen because it is the most frequently used NSAID at the study hospitals, is primarily administered orally, and is an alternative to opioids for moderate pain. We chose oxycodone as the active comparator since it is not known to be nephrotoxic, is administered orally, and is one of the most commonly used opioids in the inpatient setting.2 The University of Pennsylvania Institutional Review Board approved the study with a waiver of informed consent because the retrospective design was determined to involve no more than minimal risk. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.
Identification of the Study Cohort
We included patients aged 18 years and older who had 24 hours or more of exposure to either ibuprofen or oxycodone between January 1, 2004, and June 30, 2017, who were admitted to any of 3 University of Pennsylvania Health System hospitals. Analyses were conducted from March 2023 through August 2026. The index date was the date of first dose of study drug within an inpatient encounter. Baseline comorbid conditions were assessed during the 2 years prior to the index date. Baseline laboratory values and medications were assessed within 72 hours before the index date.
We excluded patients with conditions that might preclude NSAID treatment, including baseline platelet count less than 100 × 103/μL (to convert to ×109/L, multiply by 1), end-stage kidney disease, KRT, baseline serum creatinine (SCr) greater than 2 mg/dL (to convert SCr to micromoles per liter multiply by 88.4), or incident AKI within 2 weeks prior to, and remaining unresolved by, the index date. AKI prior to the index date was defined using Kidney Disease: Improving Global Outcomes (KDIGO) creatinine and dialysis criteria from the date of hospital admission to the index date.23 Baseline SCr for AKI prior to the index date was the average of prior outpatient or hospital discharge values from 1 year to 1 week before the date of current hospitalization admission.24 If these data were missing, baseline SCr was defined as the lowest value during the first 7 days of hospitalization up to the index date. AKI prior to the index date was considered resolved if SCr returned to within 25% of baseline.25 We excluded encounters with receipt of an intravenous (IV) NSAID or extended-release opioid during the baseline period. We excluded those exposed to IV NSAIDs because such exposure could cause subclinical toxic effects that manifest during follow-up. We excluded those with extended-release opioid exposure because these agents are often used for chronic pain, whereas our focus was on the treatment of acute pain. We also excluded patients with receipt of both study drugs on the index date, history of solid organ transplant, admission for parturition, lack of a baseline SCr value, and lack of at least 1 postbaseline SCr value. Additionally, we excluded patients with extremes of body mass index (BMI [calculated as weight in kilograms divided by height in meters squared]) (<14 or >70) values because these observations were rare and may represent distinct clinical scenarios for which we did not have enough data to evaluate. In patients with multiple eligible encounters, only the first encounter was included.
Follow-Up
Treatment courses were defined using medication administration date-time stamps in the electronic health record (EHR). Follow-up began with the first dose of study drug (the index date) and continued until a lapse in therapy of more than 48 hours, occurrence of the primary outcome, hospital discharge, death, or 14 days after the index date, whichever occurred first. Because treatment was classified at initiation and patients were not censored for crossover during follow-up, our analysis corresponds to an intention-to-treat type of estimate of the association between ibuprofen and AKI in clinical conditions.
Covariate Definitions
Potential confounders were selected a priori based on clinical knowledge and prior evidence regarding AKI risk factors.26 Comorbidities were identified using diagnosis codes, as described previously.26 Resolved prior AKI events were AKI events prior to the index date that had resolved by the index date. Baseline estimated glomerular filtration rate (eGFR) was calculated using the CKD Epidemiology Collaboration equation.27 The weight and height parameters most proximal and prior to study entry were used to calculate BMI. Sex and race were extracted from EHR records, which were populated by patient self-report. Race was categorized as Black, White, or other (including American Indian, Asian or Other Pacific Islander, or unknown). Race was included as a proxy to capture potential confounding from social and structural factors associated with kidney disease outcomes.28
Outcome Definition
The primary outcome was AKI of any stage. AKI was defined by changes in SCr and initiation of KRT according to KDIGO criteria.23 The reference value for AKI phenotyping was the SCr value most proximal and prior to the index date. We staged AKI over 7 days after AKI onset. Secondary outcomes were the incidence of severe (stage 2 or higher) AKI and need for KRT.
Subgroup Analyses
We evaluated heterogeneity by age, obesity, heart failure, type 2 diabetes, baseline eGFR, critical illness (intensive care unit admission at the index date), and postoperative status. These factors were selected because of their potential to modify vulnerability to ibuprofen-induced AKI. Kidney function is a major risk factor for adverse drug reactions29 and modifies the risk of adverse effects from other medications, including warfarin and amiodarone.30 Age, obesity, heart failure, and diabetes may alter susceptibility to ibuprofen-induced AKI via effects on renal blood flow, the renin-angiotensin-aldosterone system, and systemic inflammation.19,31,32,33 Critically ill and postoperative patients have a high baseline risk of AKI due to hypotension, sepsis, volume depletion, and other factors.23
Statistical Analysis
We used balancing weights to balance baseline characteristics between treatment groups, targeting an average treatment effect among the treated population (those who received ibuprofen).34,35,36 R version 4.2.1 (R Foundation for Statistical Computing) and the BalanceR package were used to calculate weights.37 Absolute standardized mean differences (SMDs) greater than 0.10 were considered imbalanced.38 We estimated AKI incidence rate ratios (IRRs) using weighted Poisson regression with the log of follow-up time as an offset. For each risk factor, we re-estimated weights within subgroups to ensure balance of patient characteristics within subgroups, which is necessary for unconfounded interaction estimates.39,40,41 We targeted balance within quintiles of covariates, age, baseline eGFR, and BMI.30 We evaluated interaction using treatment-by-risk factor interaction terms in weighted Poisson models. Age, baseline eGFR, and BMI were modeled continuously using restricted cubic splines. Interactions were assessed using model-based P values and, for spline-modeled variables, joint Wald test of treatment-spline interaction terms (statistical significance set at 2-sided α = .05). Standardized incidence rates and measures of interaction were estimated with Stata version 18.5 (StataCorp LLC).40
We additionally conducted the following post hoc analyses to examine the robustness of our findings: dose–response analysis, adjustment for nonoxycodone opioid use in the baseline period, restriction to patients exposed to only oral analgesics, evaluation of SCr monitoring during follow-up, extension of the follow-up window after last exposure to the study drug to 5 days, and subgroup analysis by concomitant nephrotoxin exposure. Complete methods for these are described in the eMethods in Supplement 1.
Results
Population Characteristics
This study cohort included 78 656 patients (median age, 59 [IQR, 47-69] years; 38 092 females [48%]; 24 507 Black [31%], 46 217 White [59%], 7932 other [10%]) (Table 1), of whom 6313 received ibuprofen and 72 343 received oxycodone (Figure 1). The median duration of follow-up was 3.0 days (IQR, 1.8-4.2 days) (ibuprofen group: 2.8 [IQR, 1.7-3.8] days; oxycodone group: 3.1 [IQR, 1.9-4.3] days). Before weighting, the prevalence of AKI risk factors was lower among patients treated with ibuprofen compared with those treated with oxycodone: for example, they were younger (median age, 52 [IQR, 40-64] years vs 59 [IQR, 47-69] years), more likely to be not postoperative (81.7% vs 65.3%), less likely to be critically ill (11.4% vs 21.0% admitted to the ICU at index date), and had higher median baseline eGFR (97.3 mL/minute per 1.73 m2 [IQR, 80.5-113.2 mL/minute per 1.73 m2] vs 90.1 mL/minute per 1.73 m2 [IQR, 71.2-105.5 mL/minute per 1.73 m2]) (Table 1). Balancing weights reduced all absolute SMDs to less than 0.001 overall and in subgroups (eFigures 1-8 in Supplement 1).
Table 1. Baseline Characteristics of the Cohort in Unweighted and Weighted Populations.
| Characteristic | Unweighted | Weighted | ||||
|---|---|---|---|---|---|---|
| Oxycodone | Ibuprofen | SMDa | Oxycodone | Ibuprofen | SMDa | |
| No. of patients | 72 343b | 6313b | 6313c | 6313c | ||
| Demographic characteristics | ||||||
| Age, median (IQR), y | 59 (47-69) | 52 (40-64) | 0.331 | 52 (40-64) | 52 (40-64) | <0.001 |
| Sex (reference = female), No. (%) | ||||||
| Female | 34 353 (47.5) | 3739 (59.2) | 0.237 | 3739 (59.2) | 3739 (59.2) | <0.001 |
| Male | 37 990 (52.5) | 2574 (40.8) | 2574 (40.8) | 2574 (40.8) | ||
| Race,d No. (%) | ||||||
| Black | 22 294 (30.8) | 2213 (35.1) | 0.09 | 2213 (35.1) | 2213 (35.1) | <0.001 |
| White | 42 731 (59.1) | 3486 (55.2) | 0.078 | 3486 (55.2) | 3486 (55.2) | <0.001 |
| Other | 7318 (10.1) | 614 (9.7) | 0.013 | 614 (9.7) | 614 (9.7) | <0.001 |
| BMI, median (IQR) | 28.3 (24.3-33.4) | 27.4 (23.4-33) | 0.079 | 27.5 (23.4-33) | 27.4 (23.4-33) | <0.001 |
| Hospital admission characteristics, No. (%) | ||||||
| Admission type (surgery)e | 48 561 (67.1) | 2629 (41.6) | 0.529 | 2629 (41.6) | 2629 (41.6) | <0.001 |
| Hospital center | ||||||
| HUP | 37 462 (51.8) | 4011 (63.5) | 0.24 | 4011 (63.5) | 4011 (63.5) | <0.001 |
| Presbyterian | 17 148 (23.7) | 998 (15.8) | 0.199 | 998 (15.8) | 998 (15.8) | <0.001 |
| Pennsylvania | 17 733 (24.5) | 1304 (20.7) | 0.092 | 1304 (20.7) | 1304 (20.7) | <0.001 |
| Location of initial presentationf | ||||||
| ED | 21 764 (30.1) | 3022 (47.9) | 0.371 | 3022 (47.9) | 3022 (47.9) | <0.001 |
| ICU | 4188 (5.8) | 275 (4.4) | 0.065 | 275 (4.4) | 275 (4.4) | <0.001 |
| OR | 22 221 (30.7) | 1242 (19.7) | 0.256 | 1242 (19.7) | 1242 (19.7) | <0.001 |
| Floor | 19 431 (26.9) | 1533 (24.3) | 0.059 | 1533 (24.3) | 1533 (24.3) | <0.001 |
| Other | 4739 (6.6) | 241 (3.8) | 0.124 | 241 (3.8) | 241 (3.8) | <0.001 |
| Prior length of stay, median (IQR), d | 1 (0-2) | 1 (1-2) | 0.112 | 1 (0-2) | 1 (1-2) | <0.001 |
| ICU patient at index date, No. (%) | 15 215 (21.0) | 717 (11.4) | 0.265 | 717 (11.4) | 717 (11.4) | <0.001 |
| Perioperative recency,g No. (%) | ||||||
| Not postoperative | 47 271 (65.3) | 5156 (81.7) | 0.377 | 5156 (81.7) | 5156 (81.7) | <0.001 |
| POD 0 | 6534 (9.0) | 311 (4.9) | 0.162 | 311 (4.9) | 311 (4.9) | <0.001 |
| POD 1 | 11 646 (16.1) | 500 (7.9) | 0.254 | 500 (7.9) | 500 (7.9) | <0.001 |
| POD 2 | 5191 (7.2) | 231 (3.7) | 0.156 | 231 (3.7) | 231 (3.7) | <0.001 |
| POD ≥3 | 1701 (2.4) | 115 (1.8) | 0.037 | 115 (1.8) | 115 (1.8) | <0.001 |
| Invasive mechanical ventilation at baseline, No. (%) | 5666 (7.8) | 239 (3.8) | 0.174 | 239 (3.8) | 239 (3.8) | <0.001 |
| Any prior ventilator use during hospitalization | 12 955 (17.9) | 514 (8.1) | 0.293 | 514 (8.1) | 514 (8.1) | <0.001 |
| Comorbidities, No. (%) | ||||||
| Heart failure | 11 570 (16.0) | 711 (11.3) | 0.138 | 711 (11.3) | 711 (11.3) | <0.001 |
| Myocardial infarction | 6542 (9.0) | 392 (6.2) | 0.107 | 392 (6.2) | 392 (6.2) | <0.001 |
| Arrhythmia | 11 605 (16.0) | 1036 (16.4) | 0.01 | 1036 (16.4) | 1036 (16.4) | <0.001 |
| Atrial fibrillation | 9843 (13.6) | 693 (11.0) | 0.08 | 693 (11.0) | 693 (11.0) | <0.001 |
| Valvular disease | 9167 (12.7) | 450 (7.1) | 0.186 | 450 (7.1) | 450 (7.1) | <0.001 |
| Stroke | 5284 (7.3) | 447 (7.1) | 0.009 | 447 (7.1) | 447 (7.1) | <0.001 |
| Peripheral vascular disease | 8039 (11.1) | 313 (5.0) | 0.228 | 313 (5.0) | 313 (5.0) | <0.001 |
| Pulmonary circulation disorder | 4860 (6.7) | 334 (5.3) | 0.06 | 334 (5.3) | 334 (5.3) | <0.001 |
| Chronic pulmonary disease | 16 082 (22.2) | 1530 (24.2) | 0.048 | 1530 (24.2) | 1530 (24.2) | <0.001 |
| Liver disease | 3183 (4.4) | 284 (4.5) | 0.005 | 284 (4.5) | 284 (4.5) | <0.001 |
| Diabetes, noncomplicated | 12 679 (17.5) | 821 (13.0) | 0.126 | 821 (13.0) | 821 (13.0) | <0.001 |
| Diabetes, complicated | 3586 (5.0) | 203 (3.2) | 0.088 | 203 (3.2) | 203 (3.2) | <0.001 |
| Chronic kidney disease | 6265 (8.7) | 228 (3.6) | 0.212 | 228 (3.6) | 228 (3.6) | <0.001 |
| Weight loss | 4465 (6.2) | 495 (7.8) | 0.065 | 495 (7.8) | 495 (7.8) | <0.001 |
| Fluid and electrolyte disorder | 16 060 (22.2) | 1377 (21.8) | 0.009 | 1377 (21.8) | 1377 (21.8) | <0.001 |
| Cancer, noncomplicated | 10 113 (14.0) | 834 (13.2) | 0.022 | 834 (13.2) | 834 (13.2) | <0.001 |
| Cancer, metastatic | 4415 (6.1) | 458 (7.3) | 0.046 | 458 (7.3) | 458 (7.3) | <0.001 |
| HIV | 1206 (1.7) | 174 (2.8) | 0.074 | 174 (2.8) | 174 (2.8) | <0.001 |
| Kidney function | ||||||
| eGFR, median (IQR), mL/min per 1.73 m2 | 90.1 (71.2-105.5) | 97.3 (80.5-113.2) | 0.322 | 97.3 (80.5-113.2) | 97.3 (80.5-113.2) | <0.001 |
| Resolved prior AKI, No. (%) | 5375 (7.4) | 518 (8.2) | 0.029 | 518 (8.2) | 518 (8.2) | <0.001 |
| Laboratory values | ||||||
| WBC, median (IQR), ×103 cells/μL | 9.6 (7.3-12.6) | 8.9 (6.7-11.8) | 0.147 | 8.9 (6.8-11.9) | 8.9 (6.7-11.8) | <0.001 |
| Hemoglobin, median (IQR), g/dL | 11.3 (9.9-12.7) | 11.8 (10.3-13.1) | 0.192 | 11.7 (10.2-13.2) | 11.8 (10.3-13.1) | <0.001 |
| Platelets, median (IQR), ×103/μL | 216 (169-277) | 232 (184-297) | 0.177 | 235 (185-299) | 232 (184-297) | <0.001 |
| Chloride, median (IQR), mEq/L | 104 (101-107) | 104 (101-107) | 0.041 | 104 (101-107) | 104 (101-107) | <0.001 |
| Potassium, median (IQR), mEq/L | 4.1 (3.8-4.4) | 4 (3.7-4.3) | 0.157 | 4 (3.7-4.3) | 4 (3.7-4.3) | <0.001 |
| Medications, No. (%) | ||||||
| RAS inhibitorh | 11 580 (16.0) | 924 (14.6) | 0.038 | 924 (14.6) | 924 (14.6) | <0.001 |
| Metoprolol | 15 563 (21.5) | 1237 (19.6) | 0.047 | 1237 (19.6) | 1237 (19.6) | <0.001 |
| Combined alpha and beta blocker | 4857 (6.7) | 253 (4.0) | 0.12 | 253 (4.0) | 253 (4.0) | <0.001 |
| Hydrochlorothiazide | 3588 (5.0) | 342 (5.4) | 0.021 | 342 (5.4) | 342 (5.4) | <0.001 |
| Hydralazine | 2915 (4.0) | 118 (1.9) | 0.128 | 118 (1.9) | 118 (1.9) | <0.001 |
| Loop diuretic | 10 158 (14.0) | 741 (11.7) | 0.069 | 741 (11.7) | 741 (11.7) | <0.001 |
| Other antihypertensivei | 3258 (4.5) | 292 (4.6) | 0.006 | 292 (4.6) | 292 (4.6) | <0.001 |
| Acid suppressant, H2RA | 19 705 (27.2) | 1207 (19.1) | 0.193 | 1207 (19.1) | 1207 (19.1) | <0.001 |
| Acid suppressant, PPI | 15 155 (20.9) | 1538 (24.4) | 0.082 | 1538 (24.4) | 1538 (24.4) | <0.001 |
| Broad-spectrum antibioticj | 8041 (11.1) | 1022 (16.2) | 0.148 | 1022 (16.2) | 1022 (16.2) | <0.001 |
| Narrow-spectrum antibiotick | 31 272 (43.2) | 1554 (24.6) | 0.401 | 1554 (24.6) | 1554 (24.6) | <0.001 |
| Vancomycin | 14 944 (20.7) | 817 (12.9) | 0.207 | 817 (12.9) | 817 (12.9) | <0.001 |
| Bactrim | 1720 (2.4) | 141 (2.2) | 0.01 | 141 (2.2) | 141 (2.2) | <0.001 |
| Other nephrotoxic antibioticl | 2349 (3.2) | 123 (1.9) | 0.082 | 123 (1.9) | 123 (1.9) | <0.001 |
| Other nephrotoxinm | 1517 (2.1) | 152 (2.4) | 0.021 | 152 (2.4) | 152 (2.4) | <0.001 |
| Vasopressor | 4252 (5.9) | 182 (2.9) | 0.147 | 182 (2.9) | 182 (2.9) | <0.001 |
Abbreviations: AKI, acute kidney injury; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); ED, emergency department; eGFR, estimated glomerular filtration rate; H2RA, histamine 2 receptor antagonist; HUP, Hospital of the University of Pennsylvania; ICU, intensive care unit; OR, operating room; POD, postoperative day; PPI, proton pump inhibitor; RAS, renin-angiotensin-aldosterone system; SMD, standardized mean difference; WBC, white blood cell count.
SI unit conversion factors: To convert WBC ×103 cells/μL to ×109 cells/L, multiply by 1; to convert hemoglobin g/dL to g/L, multiply by 10; to convert platelets ×103/μL to ×109/L multiply by 1; to convert chloride and potassium mEq/L to mmol/L, multiply by 1.
SMDs for continuous variables were calculated comparing differences in means and pooled SDs relative to the oxycodone group.
Actual sample size.
Weighted sample size after balance weights applied.
Race was categorized by patient self-report as Black, White, or other (including American India, Asian or Other Pacific Islander, or unknown).
Admission type was derived from the administratively assigned diagnostic-related group and categorized as surgery or nonsurgery.
Initial presentation represents the floor to which the patient was first admitted.
Perioperative recency was derived from time since patient was admitted to the OR and categorized as 0, 1, 2, or ≥3 days (POD 0, 1, 2, or ≥3, respectively).
Lisinopril, enalapril, ramipril, benazepril, quinapril, captopril, losartan, irbesartan, and valsartan.
Propranolol, clonidine, doxazosin, terazosin.
Carbapenems, cefepime, piperacillin-tazobactam, fluoroquinolones, aztreonam.
First- and second-generation cephalosporins, macrolides, amoxicillin, penicillin, tetracycline, nitrofurantoin, ampicillin-sulbactam.
Aminoglycosides (amikacin, gentamicin, tobramycin), colistin.
Carboplatin, cisplatin, ifosfamide, cyclosporine, tacrolimus, methotrexate, amphotericin, acyclovir.
Figure 1 . Flow Chart of Selection of Patients Into the Study.

SI unit conversion factors: To convert platelets to x109/L, multiply by 1; SCr to micromoles per liter multiply by 88.4. AKI indicates acute kidney injury; BMI, body mass index; ESKD, end-stage kidney disease; IV, intravenous; KRT, kidney replacement therapy; NSAID, nonsteroidal anti-inflammatory drug; SCr, serum creatinine.
Analysis in the Full Cohort
There were 4799 AKI events (320 [5.1%] in the ibuprofen group and 4479 [6.2%] in the oxycodone group) during 270 472 days of patient follow-up (overall rate, 17.7 events per 1000 days). Most AKI events were stage 1 (79%), with 15% stage 2 and 6% stage 3 events. In the weighted cohort, ibuprofen was associated with a significantly higher risk of any AKI (IRR, 1.22; 95% CI, 1.09-1.38), severe AKI (RR, 1.36; 95% CI, 1.08-1.72), and incident KRT (IRR, 3.29; 95% CI, 1.35-8.03) (Table 2).
Table 2. Outcomes According to Study Group.
| Outcome | Ibuprofen (n = 6313) | Oxycodone (n = 72 343) | IRR (95% CI)a |
|---|---|---|---|
| Any AKI, No. (rate per 1000 d) | |||
| Crude | 320 (16.4) | 4479 (17.8) | 0.92 (0.82-1.03) |
| Weighted | NA | NA | 1.22 (1.09-1.38) |
| Stage 2/3 AKI, No. (rate per 1000 d) | |||
| Crude | 84 (4.3) | 914 (3.6) | 1.18 (0.95-1.48) |
| Weighted | NA | NA | 1.36 (1.08-1.72) |
| Incident dialysis, No. (rate per 1000 d) | |||
| Crude | 7 (0.36) | 57 (0.23) | 1.58 (0.72-3.46) |
| Weighted | NA | NA | 3.29 (1.35-8.03) |
Abbreviations AKI, acute kidney injury; IRR, incidence rate ratio; NA, not applicable.
IRRs are from weighted Poisson models with a person-time offset.
Subgroup Analyses
The IRRs of ibuprofen with AKI within subgroups defined by AKI risk factors are summarized in Figure 2 and Figure 3. The IRRs were greater in postoperative (IRR, 1.70 [95% CI, 1.28-2.27] vs 1.12 [95% CI, 0.98-1.28]) and critically ill (IRR, 1.65 [95% CI, 1.29-2.11] vs 1.16 [95% CI, 1.02-1.33]) patients. The IRRs were lower in patients with diabetes (IRR, 0.92 [95% CI, 0.71-1.20] vs 1.30 [95% CI, 1.14-1.49]). Interaction terms between ibuprofen use and eGFR, age, and BMI as continuous variables were nonsignificant. In a post hoc subgroup analysis, the IRR was 1.48 (95% CI, 1.18-1.86) among patients with concomitant nephrotoxin exposure and 1.15 (95% CI, 1.00-1.32) among patients without concomitant nephrotoxin exposure, although the P value for interaction was nonsignificant (eFigures 9 and 10 in Supplement 1).
Figure 2. Dot Plot Showing the Heterogeneity in Ibuprofen-Associated AKI by Patient Subgroups.

The IRRs for ibuprofen compared with oxycodone are shown. Blue squares represent point estimates and the horizontal lines indicate 95% CIs. The overall IRRs for the full cohort is shown in the top row. Also shown are the P values for the test of an interaction, which indicate whether treatment effects differ significantly across subgroup strata. AKI indicates acute kidney injury; IRR, incidence rate ratio.
Figure 3. Line Graph Showing Heterogeneity in Ibuprofen-Associated AKI by Continuous AKI Risk Factors.

The estimated incidence rate of AKI between the oxycodone (blue) and ibuprofen (orange) treatment groups was estimated across the spectrum of (A) eGFR, (B) age, and (C) BMI. A joint Wald test of the treatment-spline interaction terms was used to assess statistical evidence of variation in the association. AKI indicates acute kidney injury; BMI, body mass index (calculated as weight in kilograms divided by height in meters squared); eGFR, estimated glomerular filtration rate.
Analysis in Patients at Low Risk of AKI
In a post hoc analysis, we restricted the cohort to patients who were not critically ill, not postoperative, were aged less than 60 years, and had eGFR values greater than 60 mL/minute per 1.73 m2 (ibuprofen, n = 2966; oxycodone, n = 20 440). We re-estimated weights with restriction covariates excluded (eFigure 11 in Supplement 1). In this subgroup, the risk of AKI was substantially lower than in the overall cohort (10.7 events per 1000 days), and ibuprofen treatment was not associated with higher risk (IRR, 0.94; 95% CI, 0.75-1.18). Despite being at low risk of AKI, 87% of these patients were treated with oxycodone.
Additional Post Hoc Analyses
We examined the dose–response relationship between ibuprofen exposure and AKI by classifying patients based on the ibuprofen dose over the first 24 hours of treatment: none (oxycodone), low dose (≤600 mg), medium dose (601-1200 mg), or high dose (>1200 mg). We observed a monotonically increasing IRRs between ibuprofen exposure and AKI incidence (eTables 1-3 in Supplement 1). Additional post hoc analyses yielded results consistent with the main analyses (eFigures 12-19 and eTables 4-9 in Supplement 1).
Discussion
In this cohort of hospitalized patients treated for pain, those treated with ibuprofen were, on average, healthier, younger, and had fewer comorbidities than patients treated with oxycodone. These differences in patient characteristics reflect current clinical practice, where NSAIDs are preferentially used in patients whom clinicians deem to be at low risk of NSAID-induced AKI. However, our results suggest that, despite such judicious prescribing, brief exposure to ibuprofen was associated with meaningful nephrotoxic effects, including increased risk of severe AKI and KRT. In contrast, the results of our subgroup analyses suggest that there may be a subset of patients who can safely be treated with NSAIDs, such as younger patients with preserved kidney function who are not critically ill and are not postoperative. In the subset of the study population defined by these characteristics, ibuprofen use was not associated with excess AKI risk. Yet this subset was infrequently treated with ibuprofen. Thus our analysis suggests that outcomes might be improved with enhanced risk stratification and individualized analgesic prescribing that effectively identifies patients at low risk of NSAID-associated AKI. To our knowledge, this is the largest inpatient study of ibuprofen-associated AKI, which allowed us to study events that are important to patients, including severe AKI and need for KRT.
Our findings differ from those of other cohort studies, which have generally concluded that short-term treatment with NSAIDs is not associated with meaningful increases in AKI risk.13,15,16,17,18 However, none of these studies used an active-comparator design, which substantially increases the risk of bias from residual confounding and immortal time.22 Additionally, to our knowledge, no previous studies examined associations with severe AKI or KRT. The available evidence from randomized clinical trials is similarly limited. A Cochrane meta-analysis identified 26 trials that examined NSAID treatment in the perioperative setting.14 The authors noted that most trials had a high risk of bias, and only 2 studies specifically reported AKI outcomes, leading them to conclude that the available data were insufficient to confirm the safety of NSAID treatment.14 A second meta-analysis of trials that focused on critically ill postoperative patients had similar conclusions: although NSAID use significantly reduced opioid requirements, inconsistent reporting of adverse events precluded analysis of safety end points.9 Our results add substantially to this evidence, representing, to our knowledge, the largest examination of inpatient NSAID exposure enabling evaluation of severe AKI and KRT outcomes.
Our subgroup analyses found increased IRRs for AKI with ibuprofen use in critically ill patients and those recovering from surgery—high-risk groups in whom factors that may predispose individuals to NSAID nephrotoxic effects are common, including rapidly changing volume status, hypotension, sepsis, and profound inflammation in response to surgery.24 Given the variability in underlying AKI risk among patients in these vulnerable subgroups, additional studies to investigate sources that modify ibuprofen-associated AKI risk within them are warranted, which might enable the strategic use of these drugs as part of multimodal analgesia.7
The findings also varied as a function of age and baseline eGFR, with risk increasing sharply as age increased beyond 60 years and eGFR decreased below 60 mL/minute per 1.73 m2. These results suggest potential thresholds that might be useful for clinical decision-making. However, the estimates were imprecise, and the interaction tests were nonsignificant, likely due to small numbers of patients treated with ibuprofen at the extremes of age and eGFR. None of the other factors investigated were associated with differential nephrotoxic effects, apart from diabetes, which was associated with reduced ibuprofen-associated AKI risk. The mechanism underlying this observation is unclear. Patients with diabetes often have altered kidney blood flow, characterized by hyperfiltration mediated by afferent arteriolar vasodilation.32,33 We speculate that, in this setting, NSAID-induced vasoconstriction might reduce excessive intraglomerular pressure and hyperfiltration-mediated injury, analogous to the mechanism of kidney protection afforded by SGLT2 inhibitors.42 However, these results may also be spurious due to residual confounding. Additional studies of NSAID effects in diabetes and on kidney blood flow are needed.
Limitations
Our study has several limitations. Despite the active comparator design, which provides rigorous control of confounding by indication, and the comprehensive set of covariates used to formulate balancing weights, the nonrandomized design remains susceptible to residual confounding. We did not have information regarding some AKI risk factors, such as hypotension, volume depletion, and presence of sepsis. However, given the pattern of confounding observed in the cohort, in which patients treated with ibuprofen appeared to be systematically at lower risk compared with patients treated with oxycodone, any residual confounding would likely bias effect estimates toward lower AKI risk with ibuprofen. Thus our estimates may underestimate the risk associated with ibuprofen. Additionally, other active comparators, such as acetaminophen, could provide further insight into the association and should be included in future work.
We examined differential ibuprofen-associated AKI risk across subgroups defined by key AKI risk factors. However, subgroup analyses consider characteristics separately, potentially oversimplifying complex drug–disease interactions. Future studies should evaluate heterogeneity by multiple characteristics simultaneously to better understand variability in the ibuprofen–AKI relationship.43,44 While our analysis was focused on kidney outcomes, other NSAID-mediated adverse effects (eg, bleeding) also inform prescribing decisions. Research evaluating the broader comparative safety of these classes is needed to optimize inpatient analgesia. In addition, although our study had a large sample size, we were still underpowered to examine interactions in which ibuprofen treatment was uncommon, such as extremes of age and eGFR values.45 We also excluded patients who received both ibuprofen and oxycodone within 24 hours of follow-up, which may limit the generalizability of the results to patients treated with both drugs. However, this was uncommon (<1% of all screened courses). Finally, our cohort was restricted to patients admitted to the University of Pennsylvania Health System from 2004 to 2017, which may not fully reflect contemporary prescribing practices. However, recent evidence suggests that the patterns of NSAID prescribing have changed little in the intervening years: relatively few patients are treated with NSAIDs and opioids remain the dominant treatment for acute pain in hospitalized patients.10,13,15,16,17,18 Further, the pathophysiologic mechanisms underlying NSAID nephrotoxic effects and potential interactions between NSAID exposure and patient characteristics are unlikely to change over time.
Conclusions
In this cohort study of hospitalized patients, ibuprofen was associated with higher rates of AKI and KRT, but this excess risk was concentrated in patients at high risk. Patients at low risk may be able to receive ibuprofen safely yet are often prescribed oxycodone.
eMethods
eFigure 1. Balance of Baseline Covariates in Study Cohort
eFigure 2. Balance of Baseline Covariates in eGFR Quintile
eFigure 3. Balance of Baseline Covariates in Age Quintile
eFigure 4. Balance of Baseline Covariates in BMI Quintile
eFigure 5. Balance of Baseline Covariates in Critical Illness Subgroups
eFigure 6. Balance of Baseline Covariates in Postoperative Status Subgroups.
eFigure 7. Balance of Baseline Covariates in Patients With and Without Heart Failure.
eFigure 8. Balance of Baseline Covariates in Patients with and without Diabetes.
eFigure 9. Balance of Baseline Covariates by Concomitant Nephrotoxin Exposure
eFigure 10. AKI Incidence with Ibuprofen by Concomitant Nephrotoxin Exposure
eFigure 11. Balance of Baseline Covariates in the Low-Risk Subset.
eTable 1. Balance of Baseline Covariates by Dose Exposure Category Before Weighting
eTable 2. Balance of Baseline Covariates by Dose Exposure Category After Weighting
eTable 3. Adjusted AKI Incidence Rate Ratios by Dose Category
eTable 4. Baseline Opioid Exposure in Full Study Cohort
eTable 5. Baseline Opioid Exposure in Critically Ill and Postoperative Subgroups
eFigure 12. Balance of Baseline Covariates in Full Study Cohort, Including Baseline Opioid Exposure
eFigure 13. Balance of Baseline Covariates by Critical Illness Status, Including Baseline Opioid Exposure
eFigure 14. Balance of Baseline Covariates by Postoperative Status, Including Baseline Opioid Exposure
eFigure 15. AKI Incidence with Ibuprofen Adjusting for Opioid Exposure During Baseline Period
eTable 6. Identification of Oral-Analgesics-Only Cohort
eFigure 16. Balance of Baseline Covariates in Oral-Analgesics-Only Cohort
eFigure 17. Balance of Baseline Covariates in Oral-Analgesics-Only Cohort by Critical Illness Status
eFigure 18. Balance of Baseline Covariates in Oral-Analgesics-Only Cohort by Postoperative Status
eFigure 19. AKI Incidence with Ibuprofen in the Oral-Analgesics-Only Cohort
eTable 7. Serum Creatinine Monitoring Over Follow-Up
eTable 8. Follow-up Duration and AKI Events Under Extended Follow-up
eTable 9. Adjusted Incidence Rate Ratio of Ibuprofen Associated AKI Under Extended Follow-up
eReference
Data Sharing Statement
References
- 1.American Hospital Association. Fast facts on U.S. hospitals. 2026. Accessed June 17, 2026. https://www.aha.org/statistics/fast-facts-us-hospitals#:~:text=Among%20other%20data%20points%2C%20the%20survey%20includes,find%20in%20AHA%20Hospital%20Statistics%2C%202026%20edition
- 2.Herzig SJ, Rothberg MB, Cheung M, Ngo LH, Marcantonio ER. Opioid utilization and opioid-related adverse events in nonsurgical patients in US hospitals. J Hosp Med. 2014;9(2):73-81. doi: 10.1002/jhm.2102 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Wu CL, Hung YL, Wang YR, et al. Pain prevalence in hospitalized patients at a tertiary academic medical center: exploring severe persistent pain. PLoS One. 2020;15(12):e0243574. doi: 10.1371/journal.pone.0243574 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Qaseem A, McLean RM, O’Gurek D, et al. ; Clinical Guidelines Committee of the American College of Physicians; Commission on Health of the Public and Science of the American Academy of Family Physicians . Nonpharmacologic and pharmacologic management of acute pain from non-low back, musculoskeletal injuries in adults: a clinical guideline from the American College of Physicians and American Academy of Family Physicians. Ann Intern Med. 2020;173(9):739-748. doi: 10.7326/M19-3602 [DOI] [PubMed] [Google Scholar]
- 5.Calcaterra SL, Yamashita TE, Min SJ, Keniston A, Frank JW, Binswanger IA. Opioid prescribing at hospital discharge contributes to chronic opioid use. J Gen Intern Med. 2016;31(5):478-485. doi: 10.1007/s11606-015-3539-4 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Donohue JM, Kennedy JN, Seymour CW, et al. Patterns of opioid administration among opioid-naive inpatients and associations with postdischarge opioid use: a cohort study. Ann Intern Med. 2019;171(2):81-90. doi: 10.7326/M18-2864 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Maund E, McDaid C, Rice S, Wright K, Jenkins B, Woolacott N. Paracetamol and selective and non-selective non-steroidal anti-inflammatory drugs for the reduction in morphine-related side-effects after major surgery: a systematic review. Br J Anaesth. 2011;106(3):292-297. doi: 10.1093/bja/aeq406 [DOI] [PubMed] [Google Scholar]
- 8.Du Toit LM, Bagshaw SM, Basmaji J, et al. Stakeholder perspectives on the reduction of opioid exposure in ICU: a modified Delphi survey. Intensive Crit Care Nurs. 2025;91:104089. doi: 10.1016/j.iccn.2025.104089 [DOI] [PubMed] [Google Scholar]
- 9.Ma CH, Tworek KB, Kung JY, et al. Systemic nonsteroidal anti-inflammatories for analgesia in postoperative critical care patients: a systematic review and meta-analysis of randomized control trials. Crit Care Explor. 2023;5(7):e0938. doi: 10.1097/CCE.0000000000000938 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Bykov K, Patorno E, Franklin JM, Vine SM, Bateman BT. Postoperative inpatient utilization of opioid and opioid-sparing analgesics in the United States hospitals, 2007-2017. Pharmacoepidemiol Drug Saf. 2021;30(3):390-394. doi: 10.1002/pds.5187 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.2023 American Geriatrics Society Beers Criteria® Update Expert Panel . American Geriatrics Society 2023 updated AGS Beers Criteria® for potentially inappropriate medication use in older adults. J Am Geriatr Soc. 2023;71(7):2052-2081. doi: 10.1111/jgs.18372 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Whelton A. Nephrotoxicity of nonsteroidal anti-inflammatory drugs: physiologic foundations and clinical implications. Am J Med. 1999;106(5B)(suppl 2):13S-24S. doi: 10.1016/S0002-9343(99)00113-8 [DOI] [PubMed] [Google Scholar]
- 13.Jeon N, Park H, Segal R, Brumback B, Winterstein AG. Non-steroidal anti-inflammatory drug-associated acute kidney injury: does short-term NSAID use pose a risk in hospitalized patients? Eur J Clin Pharmacol. 2021;77(9):1409-1417. doi: 10.1007/s00228-021-03121-0 [DOI] [PubMed] [Google Scholar]
- 14.Bell S, Rennie T, Marwick CA, Davey P. Effects of peri-operative nonsteroidal anti-inflammatory drugs on post-operative kidney function for adults with normal kidney function. Cochrane Database Syst Rev. 2018;11(11):CD011274. doi: 10.1002/14651858.CD011274.pub2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Rieck KM, Schroeder DR, Mao MA. Inpatient NSAID exposure and development of acute kidney injury in medical inpatients: a retrospective cohort study. Am J Med Open. 2025;15:100119. doi: 10.1016/j.ajmo.2025.100119 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.STARSurg Collaborative . Perioperative nonsteroidal anti-inflammatory drugs (NSAID) administration and acute kidney injury (AKI) in major gastrointestinal surgery: a prospective, multicenter, propensity matched cohort study. Ann Surg. 2022;275(5):904-910. doi: 10.1097/SLA.0000000000004314 [DOI] [PubMed] [Google Scholar]
- 17.Oh AR, Lee JW, Chung C. Perioperative ketorolac use and postoperative acute kidney injury in elderly patients undergoing noncardiac surgery: a retrospective cohort study. Sci Rep. 2026;16(1):2232. doi: 10.1038/s41598-025-25792-w [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Hatton GE, Bell C, Wei S, Wade CE, Kao LS, Harvin JA. Do early non-steroidal anti-inflammatory drugs for analgesia worsen acute kidney injury in critically ill trauma patients? An inverse probability of treatment weighted analysis. J Trauma Acute Care Surg. 2020;89(4):673-678. doi: 10.1097/TA.0000000000002875 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Perazella MA, Rosner MH. Drug-induced acute kidney injury. Clin J Am Soc Nephrol. 2022;17(8):1220-1233. doi: 10.2215/CJN.11290821 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Takeuchi T, Rahman AKMF, Ghazi L, et al. Epidemiological risk factors for acute kidney injury outcomes in hospitalized adult patients: a multicenter cohort study. Clin Kidney J. 2025;18(2):sfae426. doi: 10.1093/ckj/sfae426 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Girman CJ, Kou TD, Brodovicz K, et al. Risk of acute renal failure in patients with type 2 diabetes mellitus. Diabet Med. 2012;29(5):614-621. doi: 10.1111/j.1464-5491.2011.03498.x [DOI] [PubMed] [Google Scholar]
- 22.Lund JL, Richardson DB, Stürmer T. The active comparator, new user study design in pharmacoepidemiology: historical foundations and contemporary application. Curr Epidemiol Rep. 2015;2(4):221-228. doi: 10.1007/s40471-015-0053-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Khwaja A. KDIGO clinical practice guidelines for acute kidney injury. Nephron Clin Pract. 2012;120(4):c179-c184. doi: 10.1159/000339789 [DOI] [PubMed] [Google Scholar]
- 24.Siew ED, Ikizler TA, Matheny ME, et al. Estimating baseline kidney function in hospitalized patients with impaired kidney function. Clin J Am Soc Nephrol. 2012;7(5):712-719. doi: 10.2215/CJN.10821011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Pannu N, James M, Hemmelgarn B, Klarenbach S; Alberta Kidney Disease Network . Association between AKI, recovery of renal function, and long-term outcomes after hospital discharge. Clin J Am Soc Nephrol. 2013;8(2):194-202. doi: 10.2215/CJN.06480612 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Miano TA, Shashaty MGS, Yang W, Brown JR, Zuppa A, Hennessy S. Effect of renin-angiotensin system inhibitors on the comparative nephrotoxicity of NSAIDs and opioids during hospitalization. Kidney360. 2020;1(7):604-613. doi: 10.34067/KID.0001432020 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Levey AS, Stevens LA, Schmid CH, et al. ; CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) . A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150(9):604-612. doi: 10.7326/0003-4819-150-9-200905050-00006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Mohottige D, Boulware LE, Ford CL, Jones C, Norris KC. Use of race in kidney research and medicine: concepts, principles, and practice. Clin J Am Soc Nephrol. 2022;17(2):314-322. doi: 10.2215/CJN.04890421 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Laville SM, Gras-Champel V, Hamroun A, et al. ; CKD-REIN Study Group . Kidney function decline and serious adverse drug reactions in patients with CKD. Am J Kidney Dis. 2024;83(5):601-614.e1. doi: 10.1053/j.ajkd.2023.09.012 [DOI] [PubMed] [Google Scholar]
- 30.Miano TA, Yang W, Shashaty MGS, Zuppa A, Brown JR, Hennessy S. The magnitude of the warfarin-amiodarone drug-drug interaction varies with renal function: a propensity-matched cohort study. Clin Pharmacol Ther. 2020;107(6):1446-1456. doi: 10.1002/cpt.1819 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Tsuboi N, Okabayashi Y, Shimizu A, Yokoo T. The renal pathology of obesity. Kidney Int Rep. 2017;2(2):251-260. doi: 10.1016/j.ekir.2017.01.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32.Chagnac A, Zingerman B, Rozen-Zvi B, Herman-Edelstein M. Consequences of glomerular hyperfiltration: the role of physical forces in the pathogenesis of chronic kidney disease in diabetes and obesity. Nephron. 2019;143(1):38-42. doi: 10.1159/000499486 [DOI] [PubMed] [Google Scholar]
- 33.Martinez Leon V, Hilburg R, Susztak K. Mechanisms of diabetic kidney disease and established and emerging treatments. Nat Rev Endocrinol. 2026;22(1):21-35. doi: 10.1038/s41574-025-01171-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Thelen HE, Miano TA, Keele L. A primer on using balancing weights to estimate inverse probability weights. Am J Epidemiol. 2025;194(10):2761-2764. doi: 10.1093/aje/kwaf060 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Ben-Michael E, Keele L. Using balancing weights to target the treatment effect on the treated when overlap is poor. Epidemiology. 2023;34(5):637-644. doi: 10.1097/EDE.0000000000001644 [DOI] [PubMed] [Google Scholar]
- 36.Chattopadhyay A, Hase CH, Zubizarreta JR. Balancing vs modeling approaches to weighting in practice. Stat Med. 2020;39(24):3227-3254. doi: 10.1002/sim.8659 [DOI] [PubMed] [Google Scholar]
- 37.Ben-Michael E. Nalancer. Published online September 12, 2023. Accessed October 16, 2023. https://github.com/ebenmichael/balancer
- 38.Austin PC. Using the standardized difference to compare the prevalence of a binary variable between two groups in observational research. Commun Stat Simul Comput. 2009;38(6):1228-1234. doi: 10.1080/03610910902859574 [DOI] [Google Scholar]
- 39.Ben-Michael E, Feller A, Rothstein J. Varying impacts of letters of recommendation on college admissions: approximate balancing weights for subgroup effects in observational studies. arXiv. Preprint. Posted online February 22, 2021. doi: 10.48550/arXiv.2008.04394 [DOI]
- 40.Wang SV, Jin Y, Fireman B, et al. relative performance of propensity score matching strategies for subgroup analyses. Am J Epidemiol. 2018;187(8):1799-1807. doi: 10.1093/aje/kwy049 [DOI] [PubMed] [Google Scholar]
- 41.VanderWeele TJ. On the distinction between interaction and effect modification. Epidemiology. 2009;20(6):863-871. doi: 10.1097/EDE.0b013e3181ba333c [DOI] [PubMed] [Google Scholar]
- 42.Upadhyay A. SGLT2 inhibitors and kidney protection: mechanisms beyond tubuloglomerular feedback. Kidney360. 2024;5(5):771-782. doi: 10.34067/KID.0000000000000425 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Thelen H, Hennessy S. Characterizing treatment effect heterogeneity using real-world data. Clin Pharmacol Ther. 2025;117(5):1209-1216. doi: 10.1002/cpt.3627 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Kent DM, Paulus JK, van Klaveren D, et al. The Predictive Approaches to Treatment effect Heterogeneity (PATH) statement. Ann Intern Med. 2020;172(1):35-45. doi: 10.7326/M18-3667 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Brookes ST, Whitely E, Egger M, Smith GD, Mulheran PA, Peters TJ. Subgroup analyses in randomized trials: risks of subgroup-specific analyses; power and sample size for the interaction test. J Clin Epidemiol. 2004;57(3):229-236. doi: 10.1016/j.jclinepi.2003.08.009 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
eMethods
eFigure 1. Balance of Baseline Covariates in Study Cohort
eFigure 2. Balance of Baseline Covariates in eGFR Quintile
eFigure 3. Balance of Baseline Covariates in Age Quintile
eFigure 4. Balance of Baseline Covariates in BMI Quintile
eFigure 5. Balance of Baseline Covariates in Critical Illness Subgroups
eFigure 6. Balance of Baseline Covariates in Postoperative Status Subgroups.
eFigure 7. Balance of Baseline Covariates in Patients With and Without Heart Failure.
eFigure 8. Balance of Baseline Covariates in Patients with and without Diabetes.
eFigure 9. Balance of Baseline Covariates by Concomitant Nephrotoxin Exposure
eFigure 10. AKI Incidence with Ibuprofen by Concomitant Nephrotoxin Exposure
eFigure 11. Balance of Baseline Covariates in the Low-Risk Subset.
eTable 1. Balance of Baseline Covariates by Dose Exposure Category Before Weighting
eTable 2. Balance of Baseline Covariates by Dose Exposure Category After Weighting
eTable 3. Adjusted AKI Incidence Rate Ratios by Dose Category
eTable 4. Baseline Opioid Exposure in Full Study Cohort
eTable 5. Baseline Opioid Exposure in Critically Ill and Postoperative Subgroups
eFigure 12. Balance of Baseline Covariates in Full Study Cohort, Including Baseline Opioid Exposure
eFigure 13. Balance of Baseline Covariates by Critical Illness Status, Including Baseline Opioid Exposure
eFigure 14. Balance of Baseline Covariates by Postoperative Status, Including Baseline Opioid Exposure
eFigure 15. AKI Incidence with Ibuprofen Adjusting for Opioid Exposure During Baseline Period
eTable 6. Identification of Oral-Analgesics-Only Cohort
eFigure 16. Balance of Baseline Covariates in Oral-Analgesics-Only Cohort
eFigure 17. Balance of Baseline Covariates in Oral-Analgesics-Only Cohort by Critical Illness Status
eFigure 18. Balance of Baseline Covariates in Oral-Analgesics-Only Cohort by Postoperative Status
eFigure 19. AKI Incidence with Ibuprofen in the Oral-Analgesics-Only Cohort
eTable 7. Serum Creatinine Monitoring Over Follow-Up
eTable 8. Follow-up Duration and AKI Events Under Extended Follow-up
eTable 9. Adjusted Incidence Rate Ratio of Ibuprofen Associated AKI Under Extended Follow-up
eReference
Data Sharing Statement
