Abstract BACKGROUND AND AIMS Acute kidney injury (AKI) is a common problem in patients admitted to the intensive care unit (ICU). Nephrotoxic drugs play a causal role in 14–28% of the ICU AKI cases and are one of the few preventable AKI causes. The Dutch multicenter project ‘Towards a leaRning mEdication Safety system in a national network of Intensive Care Units—timely detection of adverse drug Events’ (RESCUE) aims to improve medication safety in ICU patients with respect to drug-induced AKI. This requires knowledge on which drugs may cause AKI and the strength of their effects. Obtaining such knowledge is challenging due to confounding. Especially in the ICU, where patients often have multiple comorbidities, AKI risk factors may also be indications or contraindications for drugs. We aimed to estimate the independent associations between nephrotoxic drug groups and AKI in ICU patients by adjusting for drug group-specific confounders. METHOD We used a detailed electronic health record database of ICU patients admitted to 13 different ICUs between 2010 and 2019 in the Netherlands. The database contains admission characteristics, (chronic) comorbidities, physiology measurements, Acute Physiology and Chronic Health Evaluation IV admission diagnoses and predicted mortality probabilities, serum creatinine (SCr), drug administrations, arterial blood pressure and renal replacement therapy (RRT) treatments. We included admissions with at least two SCr values and excluded admissions with one or more missing AKI risk factor values. Using the Kidney Disease: Improving Global Outcomes Clinical Practice Guideline for Acute Kidney Injury, we diagnosed and staged AKI according to the SCr and RRT criteria. We recorded exposure to eight drug groups widely regarded as acutely nephrotoxic [1] until AKI onset, discharge or death. Multilevel logistic regression models were applied to estimate the independent association between each drug group and AKI. The estimates were adjusted for drug group-specific confounders: we registered which AKI risk factors were indications or contraindications for each drug group. RESULTS We included 98 391 ICU admissions, of which 14 074 (14%) were diagnosed with AKI, ranging between 6 and 21% across the ICUs. Of all admissions with AKI, 8962 (64%), 1283 (9%) and 3829 (27%) developed AKI stage 1, stage 2 or stage 3, respectively (Figure 1). The percentage of admissions exposed to nephrotoxic drug groups was 7% for aminoglycosides, 6% for nonsteroidal anti-inflammatory drugs (NSAIDs), 6% for penicillin combinations, 5% for glycopeptide antibiotics, 2% for sulfonamides, 2% for antivirals for herpes infections, 1% for immunosuppressants and 1% for antimycotic antibiotics. All drug groups were associated with a higher odds of AKI except NSAIDs and penicillin combinations (Table 1). While penicillin combinations were not associated with AKI, NSAIDs were associated with a lower odds of AKI. The drug groups with the highest odds of AKI were immunosuppressants [odds ratio (OR) 2.26, 95% confidence interval (95% CI) 1.97–2.60; P < 0.001), sulfonamides (OR 1.84, 95% CI 1.65–2.04; P < 0.001) and aminoglycosides (OR 1.70, 95% CI 1.59–1.81; P < 0.001). The NSAID group was the only group that showed a higher OR after adjustment. CONCLUSION We estimated the independent associations between nephrotoxic drug groups and AKI by adjusting for drug-group specific confounders. Our results provide new estimates of the strengths of the associations. Surprisingly, NSAIDs were associated with a lower odds of AKI. This association might be biased due to residual confounding by contraindication. Future studies may investigate strategies to avoid drug-induced AKI.
Read full abstract