Vancomycin Redosing After a Change in CRRT Settings
A single patient, on continuous renal replacement therapy for an unrelated reason. The disagreement is whether to redose vancomycin off a published nomogram, wait for a confirmatory level, or do both.
Harlan W., a 49-year-old man, has owned and run the same auto-repair shop for over two decades, the kind of place where he still does most of the diagnostic work himself rather than handing it to an employee, and his wife has been coming by the ICU every evening with photos of the shop just to keep him updated. He was admitted twelve days ago with severe necrotizing pancreatitis, and while imaging so far shows the necrosis remaining sterile enough to avoid surgical debridement, he developed a secondary MRSA bacteremia eight days in, presumed to originate from a central line, and acute kidney injury severe enough to require continuous renal replacement therapy shortly after. He has been on vancomycin since the bacteremia was identified, dosed initially according to standard CRRT nomograms for his original CVVHDF settings.
Two days ago, the ICU team increased his CRRT effluent rate to improve clearance of inflammatory mediators associated with his pancreatitis — a decision made for reasons unrelated to his antibiotic dosing, but one that changes it anyway, since a higher effluent rate increases the convective clearance of vancomycin along with everything else being filtered. A trough level drawn yesterday, at what should have been steady state under the original dosing, came back at 8.2 mcg/mL — below the 15 to 20 mcg/mL range targeted for MRSA bacteremia, and consistent with a drug being cleared faster than the original nomogram assumed now that the effluent rate has changed. Trotman and colleagues' 2005 review in Clinical Infectious Diseases on antimicrobial dosing during continuous renal replacement therapy is explicit that dosing needs re-derivation whenever CRRT settings change, not just at initiation — a principle the team agrees with in the abstract and is now applying to an active, ongoing bacteremia rather than a hypothetical one. That the prescription was adjusted under the ICU team's own protocol rather than through a formal nephrology-pharmacy consult is why nobody with an antibiotic-dosing brief was in the room when the setting changed — and why the vancomycin question surfaced as a low trough the next morning rather than as an anticipated consequence the same afternoon.
In the ICU, after the low trough result
Increase the vancomycin dose now, using published CRRT dosing tables re-derived for his new effluent rate. Trotman and colleagues' 2005 review is direct on this point: CRRT should be treated as a dynamic clearance modality that needs antibiotic re-dosing whenever the settings themselves change, not just at CRRT initiation. His trough is already confirmed low, at a rate change we made two days ago for an unrelated reason — waiting further only extends the time he spends underdosed for an active MRSA bacteremia.
I'd be cautious about adjusting purely off a nomogram here. Those tables are derived from population averages across CRRT modalities and settings; they don't reliably predict any one patient's actual clearance, and his own trough already shows the original nomogram missed by a wide margin once his effluent rate changed. Adjusting again by table alone risks either still under-dosing if his real clearance is even higher than the nomogram assumes, or overshooting into toxicity if it isn't. I'd rather increase modestly and confirm with a level at the new steady state before committing to a full nomogram-predicted dose.
I don't think this needs to be either-or. He has an active, culture-confirmed MRSA bacteremia today — waiting on a confirmatory level before making any change costs him real time at a subtherapeutic dose, which the nephrologist's own caution doesn't actually require.
Increase the dose now, using the nomogram as our best current estimate the way the pharmacologist proposes. But draw a level at the new steady state, the way the nephrologist wants, before making any further adjustment beyond this one — that gets him off a known-low dose today without pretending the nomogram is the last word.
Agreed same day: dose increased per re-derived nomogram, with a confirmatory level drawn at the new steady state rather than after another blind adjustment.
Not agreed: whether future CRRT setting changes on this unit should automatically trigger an antibiotic dose review, or whether that should remain something the treating team has to remember to flag each time — the nephrologist favors a standing protocol; the pharmacologist and intensivist see this case as a reason to flag it going forward rather than evidence a formal rule is needed yet.