Antibiotic Dosing in Renal Failure: The Thinking, Not a Table
Here is a scene you will live through many times. It is 2 am, you are admitting a septic 78 year old with a creatinine of 2.4, and the order set is asking what dose of cefepime you would like. You do not know. So you search, and you find two tables that disagree with each other. This post is not going to hand you a third. Renal dosing numbers are institution specific, reference specific, and they change. Your hospital pharmacy already owns the current version. What I can give you is the thinking, because the thinking does not change, and once you have it, the tables stop being frightening and start being tools.
Why kidney function changes everything
Most of the antibiotics you reach for on an inpatient service leave the body through the kidneys. Vancomycin, the aminoglycosides, cefepime, piperacillin-tazobactam, meropenem: renally cleared, all of them. When the kidneys fail, the drug does not stop coming in. It stops going out. Serum levels climb, and with them climbs whatever toxicity the drug carries: more kidney injury, ototoxicity, seizures.
The core logic fits in two sentences. A loading dose fills the tank, and the size of the tank, the volume of distribution, has almost nothing to do with kidney function. A maintenance dose replaces what the kidneys remove, so maintenance dosing has everything to do with kidney function. This is why the classic teaching says give the full loading dose even in renal failure, then adjust everything after it. The mistake I see every month is the reverse: a terrified intern gives a septic dialysis patient a timid first dose of vancomycin because "the kidneys are bad." The kidneys are bad. The tank is the same size. That patient spent twelve hours underdosed during the exact window when adequate antibiotics matter most. Load fully, then slow down.
Estimate the function honestly
Every renal dosing decision starts with an estimate of kidney function, and the estimate is where people quietly go wrong. The equations, whatever flavor your hospital uses, all assume the creatinine is at steady state. In acute kidney injury it is not. A creatinine of 1.8 that was 0.9 this morning does not mean "mild impairment." It means the kidneys may have essentially stopped and the creatinine simply has not caught up yet. Creatinine lags injury by a day or more. If the number is moving, assume the function is worse than any equation says, dose for the trend and not the snapshot, and recheck soon.
Then remember where creatinine comes from: muscle. The 92 year old woman who weighs 48 kg with a creatinine of 1.0 does not have normal kidneys. She has no muscle. Her real clearance may be half of what the number suggests. The amputee has the same problem for the same reason. Meanwhile the 28 year old bodybuilder with a creatinine of 1.4 may have perfectly fine kidneys and a great deal of meat generating creatinine. Run the estimate, then look at the patient the estimate is describing, and ask whether the two match. When they do not, believe the patient.
Drugs that need levels, drugs that need intervals
Sort the renally cleared antibiotics into two mental bins.
The first bin is drugs we monitor with levels because the toxic range sits close to the therapeutic range. Vancomycin is the everyday example: nephrotoxic, renally cleared, and dosed against measured exposure, which is why most hospitals hand it to a pharmacy dosing service, and why you should let them have it. The aminoglycosides, gentamicin, tobramycin, and amikacin, are the sharper version of the same idea: superb gram negative killers that take out kidneys and eighth cranial nerves when they accumulate. If your patient is on either class, know when the last level was drawn, what it showed, and who is following it.
The second bin is drugs we adjust by dose and interval rather than levels, which is most of the beta-lactams. Cefepime, piperacillin-tazobactam, meropenem: as clearance falls, the drug is given less often, in smaller amounts, or both, and the specific scheme lives in your institution's references, not in your memory. Do not memorize numbers. Memorize the reflex: renally cleared beta-lactam plus reduced clearance equals an adjustment I need to look up or ask about, every single time, including every time the creatinine changes mid-course.
One drug deserves its own sentence. Nitrofurantoin needs to be concentrated in the urine to work, and failing kidneys cannot concentrate it there. The standard teaching is to avoid it at low GFR: the bladder does not get treated and the patient keeps the systemic exposure. An outpatient cystitis reflex prescription in a CKD patient is a classic miss. Check the clearance before you sign it, and check a current reference like the Sanford guide for what to use instead.
Dialysis is its own pharmacology
The moment a patient is on renal replacement, stop thinking about GFR and start asking logistics questions, because the machine is now the kidney and the machine runs on a schedule.
Ask what modality it is. Intermittent hemodialysis three times a week is a completely different clearance situation from continuous renal replacement therapy in the ICU, which removes drug around the clock and often needs doses closer to normal, not smaller. Ask when the next session is, because for dialyzable drugs the classic move is dosing after dialysis so the session does not immediately wash out what you gave. Ask whether the drug is dialyzed at all, because some are barely touched, and whether the patient still makes urine, because residual function changes the math. You are not expected to know these answers from memory. You are expected to know the questions, and to ask them out loud to nephrology and pharmacy instead of guessing. If you want the deeper physiology behind those questions, the nephrology board review covers it properly.
Cefepime neurotoxicity: the encephalopathy you will meet
Learn this one now, because you will see it and the diagnosis is embarrassingly easy to miss. Cefepime accumulates in renal failure, crosses into the brain, and produces an encephalopathy: confusion, decreased responsiveness, myoclonic jerks, sometimes nonconvulsive status epilepticus. The setup is always the same. An elderly patient with impaired or worsening renal function is a few days into cefepime, and now the family says grandma is not herself. The team works up "altered mental status," gets the head CT, checks the ammonia, blames the ICU, and the antibiotic sits there on the MAR the entire time.
So build the reflex: new confusion on cefepime equals check the renal function and the dose before you do anything clever. Was the dose ever adjusted? Did the creatinine drift up mid-course while everyone watched the fever curve instead? The treatment is stopping or adjusting the drug, and in the right patient an EEG earns its keep. You will look like a wizard for asking one question about a medication list.
The phone call that makes you look smart
The strongest renal dosing move you own is a pharmacist call that sounds like this: "I am starting drug X for indication Y. Weight is 70 kg, creatinine was 1.1 two days ago and is 2.3 today, so this is an unstable AKI, and dialysis has not started. What dose and interval do you want, and what monitoring?" That call takes ninety seconds. Notice what it contains: the drug, the indication, the weight, the creatinine trend rather than a single value, and the dialysis status. Those five items are the entire input a pharmacist needs, and offering them up front marks you as someone worth helping.
The alternative is the resident who copies a dose from an old note, never rechecks it, and gets found out at the end of the week when the creatinine has doubled and the vancomycin trough is in the stratosphere. Renal dosing is not a one time decision. It is a daily one, made every morning when you look at the creatinine, which is also the right moment to ask whether the antibiotic should be narrowed or stopped entirely. That discipline is its own topic, and I wrote about it in antibiotic duration and de-escalation. While you are watching the morning labs, watch the potassium and magnesium too, because sick kidneys scramble those as well, and electrolyte replacement in renal failure has its own traps.
None of this requires memorizing a table. It requires knowing that loading and maintenance are different problems, that creatinine lies in predictable ways, which drugs need levels, what to ask when dialysis enters the picture, one famous toxicity, and one phone number. That is the thinking. The numbers belong to your pharmacist and your current references, and the residents who understand that division of labor are the ones who never hurt anyone with an antibiotic.
If this way of thinking through problems is useful to you, it is exactly how the board review question banks on this site are written: reasoning first, trivia second, with rationales that teach the principle behind the answer. And if you want all of it in one place, the practice exams, the clinical guides, and the community of residents and attendings working through the same cases, take a look at membership. It exists so that the thinking in posts like this one becomes a habit before your patients need it to be.
Where HistoryandPhysical.net fits in
Reasoning about renal dosing is the skill. Doing the sums at 2 a.m. is the failure point.
- Free calculators for the inputs. The calculator set covers the values this decision actually turns on, all free and all showing their working.
- Guides for the surrounding decisions. Antibiotic coverage, duration and de escalation and electrolyte replacement.
- Nephrology and ID question banks. A 120 question nephrology board review bank and a 120 question infectious disease bank, each with its own sixty question timed exam and per-topic score report.
- Evidence you can check. The Evidence tool cites PubMed, Europe PMC and published guidelines rather than asserting.
Membership is $10 a month, $79 a year, or $59 a year at the founding rate, with a 30 day refund window on any payment.
This article is education, not medical advice, and it deliberately contains no doses. Antibiotic dosing in renal impairment must follow your institution's protocols, current references, and your pharmacist's guidance for the individual patient in front of you.
Keep going with Alo Academy
Writing a good note is one part of the job. Passing the exams that get you there is another. Alo Academy covers those with question banks kept current with the exams, explanations that show the reasoning rather than the answer, and AI analysis that finds the specific mistakes costing you points, plus a dashboard that turns it into what to study next.