Drugs to avoid in renal failure: medications to avoid in kidney disease, the nephrotoxic drugs list, and what needs a renal dose adjustment

True contraindications are a much shorter list than the list of drugs that merely need adjusting. The ones stated as contraindications in US labeling include metformin below an eGFR of 30, dofetilide below a creatinine clearance of 20, nitrofurantoin below a creatinine clearance of 60 in the FDA label, and several diabetes and gout agents at product specific cutoffs. Meperidine is avoided rather than formally contraindicated, but the accumulation of normeperidine and the resulting seizure risk make the practical answer the same, and it appears on the AGS Beers Criteria medication list for the same reason.

Everything else on the list above is a dose question, not a prohibition. That distinction matters clinically because reflexively stopping renally cleared drugs in a patient with CKD causes its own harm: undertreated pain, uncontrolled atrial fibrillation, untreated infection. The right move is almost always to adjust the dose or switch within the class, not to leave the indication untreated. Anticoagulation is the clearest example: the four direct oral anticoagulants differ enough in renal clearance that poor kidney function usually changes which agent you pick rather than whether the patient gets one. Confirm each cutoff in the current labeling, because they differ by product and sometimes by indication within the same product.

This is the question that trips up more dosing decisions than any single drug. Most older US drug labels state their renal cutoffs as creatinine clearance estimated by the Cockcroft-Gault equation, in mL/min. What your laboratory reports is an eGFR from the 2021 race free CKD-EPI creatinine equation, in mL/min/1.73 m2. Those are different numbers computed from different inputs, and they diverge most in exactly the patients where the decision is hardest: the very old, the very small, the obese.

The direction of travel is clear. The 2024 FDA guidance for industry recommends eGFR rather than Cockcroft-Gault creatinine clearance for evaluating pharmacokinetics in impaired kidney function, and a National Kidney Foundation workgroup consensus published in the American Journal of Health-System Pharmacy recommends moving to the race free 2021 CKD-EPI equation, individualized for body surface area, for medication related decisions in adults. In practice, until every label is rewritten, the safe habit is to know which equation a given threshold came from, to de-index eGFR for body surface area when the patient is far from average size, and to treat a value close to a cutoff as a prompt to look harder rather than a verdict.

The most common error in CKD prescribing is not giving something nephrotoxic. It is stopping something protective because the creatinine went up. ACE inhibitors and ARBs cause an expected creatinine rise of up to roughly 30 percent when started, reflecting reduced intraglomerular pressure rather than injury, and that hemodynamic change is the mechanism by which they slow CKD progression. Stopping them for a small, stable rise gives up the benefit and rarely helps.

SGLT2 inhibitors behave the same way. They produce an initial eGFR dip followed by slower long term decline, they have CKD indications with initiation thresholds down around an eGFR of 20 to 25 depending on the product, and treatment is generally continued as function declines rather than stopped at an arbitrary number. The nonsteroidal mineralocorticoid receptor antagonist finerenone sits alongside them, with its own eGFR and potassium based criteria. Statins likewise need no routine renal reduction for atorvastatin or pravastatin, though rosuvastatin, simvastatin and lovastatin carry product specific renal guidance. The clinically useful framing is that some drugs damage the kidney, some accumulate in it, and some protect it, and the third group must not be swept up with the first two.

The two situations call for different reflexes. In acute kidney injury the kidney function is moving, so any estimating equation is already out of date by the time you read it, and the priority is removing whatever is driving the injury: NSAIDs, contrast exposure, aminoglycosides, and the renin angiotensin blockade that is protective in stable CKD but harmful during acute hypoperfusion. Holding an ACE inhibitor during an acute illness with volume depletion is a different decision from stopping it permanently, and the restart plan should be documented at the time it is held.

In chronic kidney disease the function is comparatively stable, so the work is systematic rather than urgent: dose adjustment by band, agent substitution within a class, an annual look at whether each renally cleared drug still fits the current eGFR, and vigilance about the additive combinations. The recurring failure in stable chronic kidney disease is not a dramatic mistake, it is a gabapentin dose that was correct three years ago at an eGFR of 55 and is now sedating at an eGFR of 28, with nobody having rechecked it. A scheduled renal dose adjustment review catches that; an interaction check on its own never will.

Several common antibiotics need no renal adjustment at all, which makes them the practical first choices when kidney function is poor or unknown: azithromycin, ceftriaxone, doxycycline, clindamycin, moxifloxacin, linezolid and metronidazole are the ones cited most often. Most beta lactams, the other fluoroquinolones, vancomycin, trimethoprim/sulfamethoxazole, fluconazole and acyclovir do need adjustment, and the beta lactams in particular carry a neurotoxicity risk at unadjusted doses, with cefepime the best documented example.

Nitrofurantoin is the outlier worth naming twice, because it is the reflex choice for uncomplicated cystitis and the one antibiotic that becomes both less effective and more toxic as filtration falls. Our full breakdown of which antibiotics are safe in renal failure covers agent by agent handling, the adjustment thresholds, and what to reach for in a urinary tract infection when eGFR is low, and the antibiotic dosing calculator runs the same check against the regimen in front of you.

Renal dosing fails in practice for a mundane reason: it requires three lookups at once. You need the current kidney function, the threshold in the labeling for this specific product and indication, and whether anything else in the regimen is adding to the load or blocking the same clearance pathway. Doing that by hand for a ten drug list, at the end of a clinic session, is where the misses happen.

Prescriber.io takes the regimen and returns interaction flags, contraindications, renal and hepatic dose considerations and guideline based alternatives in one card, with the source of each flag cited so it can be checked rather than trusted. It is decision-support for licensed US clinicians and it does not prescribe. Every flag is a prompt for you to consider. You verify against the current labeling and your own institutional references, apply clinical judgment, and sign. If you are still comparing options, our guide to clinical decision support software weighs them on the point-of-care job specifically.