Digging through the canine literature, the most important thing is not to lump normal clinical dosing together with studies specifically designed to manufacture kidney failure. When you separate those, the picture changes quite a bit.
| Drug / study | Dogs | Dose / schedule | Duration | What happened | How I’d interpret it |
|---|---|---|---|---|---|
| Amikacin efficacy trials | 80 infected dogs | 10 mg/kg BID = 20 mg/kg/day | 8–21 days | 0/80 had recognized nephrotoxicity or other toxicity | Probably the most clinically relevant amikacin safety number available. Older data, but actual infected patients rather than an AKI model. (Drugs.com) |
| Amikacin toxicology | dogs, exact N not stated in label summary | 30 mg/kg/day | 90 days | Minimal renal alterations, considered reversible; no urinary casts reported | Strikingly high cumulative exposure compared with a 5–10 day clinical course. (Drugs.com) |
| Amikacin toxicology | dogs | 45 mg/kg/day | 14 days | Minimal-to-mild renal histopathologic changes | About 2–3× a common clinical daily dose before microscopic lesions become evident. (Drugs.com) |
| Amikacin chronic study | 18 Beagles | 30, 60, or 90 mg/kg/day IM | 100 days | Dose-related tubular degeneration/necrosis, especially 60–90 mg/kg/day; high-dose animals had substantial toxicity | This is genuine amikacin nephrotoxicity—but at enormous and prolonged exposures. (HRES PDF) |
| Amikacin extreme-dose study | dogs | 100, 200, 400 mg/kg/day IV | repeated dosing | 100 mg/kg considered no-effect level in that study; at 200 mg/kg, renal changes in 5/10 dogs; 400 mg/kg, all affected and 3 died | Useful toxicology, but essentially irrelevant to routine therapeutic dosing because the doses are gigantic. (PubMed) |
| Gentamicin once-daily safety | 6 healthy dogs | 6 mg/kg IM q24h | 5 days | No clinically important renal failure. 1/6 slight creatinine/BUN rise; 1/6 ↓USG; 2/6 a few granular casts | Demonstrates that subtle tubular effects can occur while overt nephrotoxicity remains uncommon. (PubMed) |
| Gentamicin experimental toxicity | 6 dogs | 10 mg/kg q8h = 30 mg/kg/day | 10 days | Urinary GGT increased by day 5; creatinine >2 mg/dL around day 9; glycosuria 3/6, casts 1/6 | Classic paper often underlying the scary reputation—but this is deliberately nephrotoxic dosing. (PubMed) |
| Gentamicin diet/toxicity model | 18 Beagles | 10 mg/kg IM q8h = 30 mg/kg/day | 8 days | Progressive renal tubular injury/enzymuria | Again, 30 mg/kg/day divided TID was being used as a nephrotoxicity model. (PubMed) |
| Gentamicin AKI model | 5 dogs | 8–10 mg/kg SC q8h = 24–30 mg/kg/day | until AKI developed | 5/5 intentionally developed AKI; urine NGAL rose before creatinine | Extremely useful for understanding biomarkers, terrible for estimating the risk from ordinary clinical dosing. (PubMed) |
| Gentamicin + misoprostol toxicity study | 12 dogs | 10 mg/kg IV q8h = 30 mg/kg/day | 8 days | Significant loss of renal function and histologic injury | Investigators explicitly describe the dose as a nephrotoxic dose. (PubMed) |
| Gentamicin dosing in compromised kidneys | 26 subtotally nephrectomized dogs | only ~3.75 mg/kg/day, divided BID/TID or once-daily infusion | 12 days | No major serum chemistry/histopathology difference between groups, but gentamicin clearance deteriorated most with TID dosing; once-daily infusion preserved clearance | Excellent evidence that schedule matters, even when total daily dose is identical. (PubMed) |
What jumps out at me
For amikacin, the number I keep coming back to is:
0/80 dogs with detected clinical nephrotoxicity at 10 mg/kg BID for 8–21 days. (Drugs.com)
That doesn’t prove the true risk is zero; 80 dogs isn’t enough to detect rare events, and those older studies probably weren’t measuring SDMA, NGAL, microalbuminuria, etc. But it makes it awfully difficult to characterize ordinary amikacin therapy as something that routinely wrecks kidneys.
For gentamicin, the nasty experimental results almost invariably start looking like:
8–10 mg/kg q8h = 24–30 mg/kg/day for 8–10+ days.
And yes—at 30 mg/kg/day, you can reliably hurt a dog’s kidneys. In one study, creatinine didn’t significantly rise until day 9, despite earlier tubular injury. (PubMed)
That’s important because it explains both sides of the argument. The nephrotoxicity isn’t imaginary. But the frequently cited studies are often essentially:
“We wanted an experimental model of acute kidney injury, so we gave dogs huge amounts of gentamicin three times a day until their kidneys were injured.”
Then the clinical takeaway gradually morphs into “gentamicin is terribly nephrotoxic.”
Those are not equivalent statements.
And once-daily dosing really matters
Modern recommendations deliberately move away from divided aminoglycoside administration. Current ISCAID guidance lists approximately:
Amikacin: 15 mg/kg q24h
Gentamicin: 9–14 mg/kg q24h
for appropriate canine infections, with renal monitoring. (Wiley Online Library)
The pharmacologic rationale is strong: aminoglycoside bacterial killing is concentration dependent, there’s a significant post-antibiotic effect, and renal tubular uptake is saturable. High peak + very low trough is generally preferable to keeping the kidney bathing in aminoglycoside around the clock.
The Frazier study is especially neat. Dogs with experimentally reduced renal reserve received the same total daily gentamicin dose, ~3.75 mg/kg/day, but given BID, TID, or as once-daily infusions. Gentamicin clearance deteriorated dramatically in the TID group, whereas it essentially stayed unchanged in the once-daily infusion groups. (PubMed)
That’s pretty compelling evidence that “how much aminoglycoside per day” isn’t the whole toxicity equation.
So how dangerous is a short therapeutic course?
Here’s where I’m going to be very careful with the statistics:
There is not a good, large, modern canine prospective study from which anyone can legitimately say, “amikacin causes renal failure in X% of dogs.”
That number simply doesn’t exist.
The available canine evidence supports something more nuanced:
Clinically important nephrotoxicity appears uncommon during a short course in a well-hydrated dog with normal renal function, particularly using once-daily administration. Subclinical proximal tubular injury may occur before creatinine changes, and the risk rises with duration, renal impairment, dehydration/hypotension, persistent trough levels and concurrent nephrotoxins. (PubMed)
And the amikacin data arguably look friendlier than the gentamicin data.
One more fascinating bit: a 2026 pharmacokinetic modeling study estimated essentially 0% AUC-defined nephrotoxicity in virtual healthy dogs across gentamicin doses up through 10 mg/kg q24h. Using a more conservative trough criterion, predicted risk increased as the dose went up—about 1% at 6 mg/kg, 5% at 8 mg/kg and 15% at 10 mg/kg q24h—but those are modeled probabilities based on a surrogate exposure threshold, not observed rates of renal failure in dogs. Kidney impairment dramatically increased predicted risk. (PubMed Central (PMC))
So I would not quote that as “15% of dogs get nephrotoxicity.” That’s not what the paper showed.
My bottom-line read of the literature
If somebody told me:
“Aminoglycosides are nephrotoxic, so I use them cautiously and monitor renal function,”
I’d agree completely.
If somebody told me:
“You shouldn’t give a normal, hydrated dog amikacin for five or seven days because there’s a substantial likelihood it’ll go into renal failure,”
I’d ask them to show me the canine study demonstrating that, because I can’t find it.
The literature I’ve found actually points much more toward:
short therapeutic course + normal kidneys + hydration + once-daily dosing = relatively low clinical risk
versus
large total daily dose + q8h exposure + prolonged treatment + impaired renal clearance/dehydration = very real nephrotoxicity.
That 0/80 amikacin result over as long as 21 days is considerably more reassuring than I expected when I started digging. (Drugs.com)
