Once-Daily Aminoglycoside Dosing Against a Rapidly Improving Creatinine
A single patient, hours after an obstructing stone was relieved. The disagreement isn't whether the antibiotic will work — the organism is fully susceptible — it's whether a dosing interval built for a stable kidney can be trusted on one that's actively changing.
Dennis R., a 45-year-old man, has taught woodshop at the same high school for seventeen years and by his own description has ‘never missed a first day’ until this week, when a dull right flank ache he'd been ignoring for two days turned into rigors and a fever of 103.4°F on a Tuesday morning. Imaging in the emergency department found a 7mm obstructing stone at the right ureterovesical junction with moderate hydronephrosis, and urology placed a stent emergently once his blood pressure began trending down despite two liters of fluid. Blood and urine cultures both grew Escherichia coli, and by the time he reached the ICU he was on low-dose norepinephrine with a lactate of 3.2 — septic shock from an obstructed, infected collecting system, now source-controlled but not yet clinically turned around.
The antibiotic question the team is working through isn't which drug covers the organism — the isolate is susceptible to nearly everything tested — but how to dose gentamicin, added empirically at admission for broader gram-negative coverage before susceptibilities returned and now being reconsidered now that they have. His creatinine has moved from 1.4 at presentation, reflecting acute obstruction, down to 1.0 six hours after stent placement, and the team expects it to keep improving as the obstruction resolves — a genuinely moving renal-function target, not a stable one. Once-daily extended-interval dosing depends on concentration-dependent bactericidal killing and a real post-antibiotic effect, and large trials and meta-analyses going back three decades — Barza et al.'s 1996 BMJ meta-analysis of twenty-one randomized trials and 3,091 patients chief among them — found no efficacy penalty against traditional multiple-daily dosing and a lower risk of nephrotoxicity, a risk ratio of 0.74 whose confidence interval ran right up to 1.00. What none of those trials specifically isolated is a patient whose creatinine is actively falling by nearly a third within the first day of therapy, which is exactly the situation extended-interval nomograms like Hartford's were built to accommodate through renal-function-based interval adjustment, not exactly the situation those nomograms were validated against in their original form. His extended-interval dose was calculated using the Cockcroft-Gault estimate from his admission creatinine of 1.4, a number the team already expects six hours' worth of stent-relieved obstruction to have outdated by the time the next dose is due.
In the ICU, six hours after the stent
Continue gentamicin once-daily, extended interval, now that the organism's confirmed susceptible and we're deciding how to actually dose it rather than just cover broadly. Concentration-dependent killing means the peak-to-MIC ratio, not the time above MIC, drives efficacy for aminoglycosides, and Barza and colleagues' 1996 BMJ meta-analysis of twenty-one randomized trials found once-daily dosing at least as effective as multiple-daily regimens — the difference in treatment failures wasn't significant — with less nephrotoxicity, and it specifically noted the efficacy advantage grew with the proportion of Pseudomonas isolates in a trial. His post-antibiotic effect at an adequate peak concentration covers the interval between doses even as the drug level itself falls.
The efficacy data doesn't worry me — it's the extended-interval nomograms I'd be cautious about applying to him today. The Hartford nomogram Nicolau and colleagues validated in 1995 assumes a reasonably stable creatinine clearance to select the dosing interval; his creatinine dropped by nearly a third in six hours and is still moving. Pick an interval off a moving number and you risk under-dosing tomorrow if it keeps improving, or accumulating drug if it plateaus higher than expected — the nomogram wasn't built to be re-derived hourly, and I'd rather dose conventionally with tighter peak/trough monitoring until his renal function actually settles.
I'd keep once-daily dosing rather than switch him to conventional multiple-daily — the efficacy and nephrotoxicity case for extended-interval is real and I don't think an improving kidney argues against using it, only against setting the interval once and forgetting it.
Draw a level earlier than we'd normally schedule one — at eight hours instead of waiting for a standard trough — specifically because his creatinine is still moving, and recheck renal function daily rather than at the usual interval. That answers the nephrologist's concern about a stale interval without giving up the actual advantage of the dosing strategy itself.
Agreed same day: once-daily dosing continued, with an earlier-than-standard confirmatory level and daily renal function rechecks rather than a strategy change.
Not agreed: at what point a still-improving creatinine should be treated as ‘settled enough’ to return to the standard trough-timing schedule — the nephrologist wants two consecutive stable creatinine values before stepping down monitoring frequency; the infectious disease physician sees that as more caution than the pharmacokinetics actually require once the first confirmatory level comes back on target.