The classic multiple-daily-dose method — peak-and-trough serum sampling for gentamicin, still the standard in endocarditis synergy dosing, pregnancy, and unpredictable renal function
Extended-interval once-daily aminoglycoside dosing (using a single nomogram-derived level) became the default in the 1990s because it is simpler, exploits concentration-dependent killing, and reduces nephrotoxicity in most patients. But traditional multiple-daily dosing with explicit peak-and-trough monitoring has never been retired — in several clinical situations its more granular, two-parameter pharmacokinetic picture is exactly what is needed.
For enterococcal or streptococcal infective endocarditis, gentamicin is not used as a standalone bactericidal agent — it is added at low dose to a cell-wall-active agent (penicillin, ampicillin, or vancomycin) to achieve synergistic bactericidal killing. The aminoglycoside's job is to reach the bacterial ribosome, which it can only do efficiently once the cell-wall agent has damaged the envelope.
Because the goal is a modest, steady synergy concentration rather than a maximal concentration-dependent kill, synergy dosing uses low doses (roughly 1 mg/kg) given every 8 hours, targeting a peak of only 3–5 mg/L and a trough under 1 mg/L. A once-daily high-dose strategy would produce unnecessary peaks without added synergistic benefit and would leave long trough-free stretches that do not match the twice/thrice-daily co-administration schedule of the primary agent.
Synergy dosing intentionally targets a lower, tightly-bounded peak than conventional gram-negative treatment — a nuance a single-level once-daily nomogram is not built to express.
Pregnancy substantially changes aminoglycoside pharmacokinetics: plasma volume expands by roughly 30–50%, glomerular filtration rate rises, and body composition shifts throughout gestation. These changes increase the volume of distribution (Vd) and drug clearance, making a single population-based once-daily nomogram unreliable — the same weight-based dose can land at a markedly different peak and trough than it would outside pregnancy.
Traditional multiple-daily dosing with individualized peak/trough sampling lets the clinician directly observe the patient's actual achieved concentrations and clearance rather than relying on a nomogram derived from non-pregnant populations, and adjust dose and interval empirically.
Once-daily nomograms assume a reasonably stable, estimable creatinine clearance. In patients with significant renal impairment, rapidly changing renal function (acute kidney injury), dialysis, or extremes of body habitus, that assumption breaks down — a single trough-timed level cannot reliably characterize both the peak and the elimination half-life.
In these patients, traditional dosing with directly measured peak and trough levels (and, when needed, a third mid-interval level for formal pharmacokinetic calculation of Vd and elimination rate constant) gives a much more individualized and defensible picture than extrapolating from a nomogram built for patients with predictable, stable renal clearance.
Aminoglycosides like gentamicin kill bacteria in a concentration-dependent fashion: the higher the peak concentration relative to the pathogen's minimum inhibitory concentration (MIC), the faster and more complete the bactericidal effect, and the longer the post-antibiotic effect that follows. The peak level is drawn specifically to confirm that the administered dose actually achieved a concentration high enough to kill.
The peak sample must be timed carefully: drawn too early (during infusion or immediately after), it captures a distribution-phase artifact rather than the true post-distribution serum concentration; drawn too late, it underestimates the achieved Cmax because elimination has already begun.
Standard practice draws the peak 30–60 minutes after the end of a 30–60 minute IV infusion, allowing the drug to distribute out of the central compartment into tissue before sampling. For intramuscular dosing, the peak is typically drawn around 60 minutes post-injection to account for the additional absorption phase.
Unlike a single once-daily nomogram target, traditional dosing explicitly tailors the peak goal to clinical context:
• Uncomplicated urinary tract or lower-severity infection: peak 4–6 mg/L is often adequate • Conventional moderate systemic gram-negative infection: peak 5–10 mg/L • Severe, life-threatening infection (sepsis, ventilator-associated pneumonia, bacteremia): peak toward 8–10 mg/L to maximize the Cmax:MIC ratio against less-susceptible organisms • Endocarditis synergy dosing: peak deliberately kept low, 3–5 mg/L (see Stage 1)
A peak below the target range for the clinical scenario suggests the dose is too low to reliably achieve bactericidal killing and prompts a dose increase; a peak above target increases toxicity risk without added benefit and prompts a dose decrease.
Concentration-dependent killing means a single very high peak clears more bacteria, faster, than the same total daily dose spread flatter — the pharmacologic rationale peak monitoring exists to verify.
While the peak confirms the dose was high enough to kill, the trough confirms the interval was long enough for the drug to clear before the next dose. Gentamicin accumulates preferentially in the renal cortex and inner-ear structures; a persistently elevated trough is the earliest pharmacokinetic warning sign that drug is accumulating faster than it is being eliminated.
Gentamicin is cleared almost entirely by glomerular filtration. Its terminal half-life in tissue compartments (renal cortex, inner ear perilymph) is far longer than its plasma half-life, because the drug binds avidly to megalin receptors on proximal tubule cells and is slowly released back into circulation. A trough that fails to fall below target before the next dose is evidence that plasma clearance has slowed — often because early, sub-clinical proximal tubule injury is already reducing filtration, well before serum creatinine rises detectably.
Because of this, the trough functions less as a simple safety cutoff and more as an early functional readout of the patient's real-time aminoglycoside clearance capacity.
A rising trough on serial monitoring can precede a measurable creatinine increase by 1–2 days, which is why trough trends — not just a single value — are tracked throughout a treatment course.
Conventional gram-negative multiple-daily dosing targets a trough below roughly 2 mg/L. Endocarditis synergy dosing, given its lower peak target and its longer intended treatment duration (often 2–6 weeks alongside a cell-wall agent), targets an even tighter trough below about 1 mg/L, since cumulative exposure over a prolonged course is itself a major toxicity determinant independent of any single level.
An elevated trough at any target threshold does not necessarily call for a dose change — because the peak may still be appropriate — it calls first for extending the dosing interval, giving the kidneys more time to clear the drug between doses while preserving the bactericidal peak.
The defining feature of traditional peak/trough monitoring is that it yields two independent pieces of information mapped to two independent knobs: the peak result tells the clinician whether to change the dose amount, and the trough result tells the clinician whether to change the dosing interval. Once-daily nomogram approaches collapse this to a single trough-timed level and a single lookup — traditional dosing keeps the two axes separate.
A practical way to think about the adjustment grid: peak status (below / within / above target) sets the dose-axis move, and trough status (acceptable / elevated) sets the interval-axis move — and the two are applied independently and simultaneously:
• Peak low, trough acceptable → raise the dose; interval unchanged • Peak within target, trough acceptable → regimen is appropriate; continue unchanged • Peak within target, trough elevated → extend the interval; keep the same dose amount • Peak high, trough acceptable → lower the dose; interval unchanged • Peak high, trough elevated → lower the dose and extend the interval — both efficacy and safety targets have been missed • Peak low, trough elevated → clearance is reduced enough that simply raising the dose would worsen accumulation; extend the interval first, then reassess whether the dose also needs adjustment on repeat levels
When levels are meaningfully off target, many clinicians move beyond simple directional adjustment to a full individualized pharmacokinetic recalculation (the Sawchuk-Zaske method or equivalent): using the measured peak and trough (and their exact sampling times relative to the dose), the patient's elimination rate constant (kel) and volume of distribution (Vd) are back-calculated, and a new dose and interval are computed algebraically to hit the desired peak and trough exactly.
This is the pharmacokinetic power that distinguishes traditional dosing from a once-daily nomogram: rather than looking up a population-based recommendation from a single level, the clinician derives the individual patient's own kinetic parameters and designs a regimen specifically for them.
Because it takes 3–5 half-lives to reach a new steady state, levels are typically rechecked around the third or fourth dose after any adjustment — not the very next dose — to judge whether the new regimen is actually hitting target.
Hitting every peak and trough target does not guarantee a patient will be spared aminoglycoside toxicity. Ototoxicity and nephrotoxicity have individual susceptibility components — including cumulative dose, treatment duration, concurrent nephrotoxins, and even mitochondrial genetic variants — that a pharmacokinetic level alone cannot capture. Peak/trough monitoring is therefore always paired with structured clinical and laboratory surveillance.
Aminoglycoside nephrotoxicity is typically non-oliguric acute tubular injury from proximal tubule drug accumulation, usually emerging after 5–7 days of therapy and often reversible if caught early and the drug is stopped or adjusted. Surveillance includes serum creatinine measured every 2–3 days during treatment (more frequently in unstable renal function), tracking the trend rather than a single value, alongside urine output monitoring and avoidance of concurrent nephrotoxins (contrast, NSAIDs, other nephrotoxic antibiotics) whenever possible.
A rising creatinine or a persistently elevated trough should prompt reassessment of the entire regimen — interval extension, dose reduction, or in some cases discontinuation in favor of an alternative agent.
Ototoxicity has two distinct clinical presentations that must be monitored separately:
• Cochlear (auditory) toxicity: progressive, typically starting with high-frequency hearing loss that may be missed on casual conversation testing, then tinnitus, then loss extending into speech frequencies. Baseline and serial audiometry is recommended for patients on courses beyond a few days, especially with prior aminoglycoside exposure or renal impairment. • Vestibular toxicity: dizziness, disequilibrium, ataxia, oscillopsia (the visual world appears to bounce with head movement) — often subtler and easier to miss than hearing loss, since patients may attribute unsteadiness to illness rather than the drug. Directed bedside questioning about imbalance and gait observation should accompany every clinical review.
Both forms of ototoxicity can be substantially or fully irreversible even when levels have been within target range throughout — reflecting direct hair-cell toxicity that is not fully predicted by serum drug concentration.
A well-described contributor to unpredictable ototoxicity is the mitochondrial 12S rRNA variant m.1555A>G, which confers profound aminoglycoside-induced hearing loss risk even at single, appropriately-dosed exposures — a reminder that PK targets manage average risk, not individual susceptibility.
In practice, a traditional peak/trough monitoring plan bundles together: (1) serial peak and trough levels to verify the regimen is pharmacokinetically on target, (2) serial serum creatinine to catch early nephrotoxicity, (3) baseline plus periodic audiometry for courses expected to exceed roughly a week, and (4) routine bedside questioning about tinnitus, hearing changes, and balance at every clinical encounter.
No single one of these substitutes for the others — a perfect trough does not rule out ototoxicity, and normal hearing at one check does not guarantee it will remain normal a week later on a prolonged course. The combination is what allows clinicians to use a genuinely effective but narrow-therapeutic-index drug with reasonable safety margins.