Page 1798 · Dosing vancomycin during renal replacement therapy — modality-specific clearance, post-dialysis supplementation, and level-guided individualization
Vancomycin is a hydrophilic glycopeptide with a molecular weight (~1,450 Da) and moderate protein binding (~30–55%) that make it dialyzable to a meaningful degree — but exactly how much is removed depends enormously on which renal replacement therapy (RRT) modality is running. Intermittent hemodialysis (IHD), continuous renal replacement therapy (CRRT, in its CVVH/CVVHD/CVVHDF forms), and peritoneal dialysis (PD) differ in membrane surface area, blood/dialysate flow rates, treatment duration, and diffusive versus convective clearance mechanisms — producing dramatically different vancomycin removal profiles that each require their own dosing logic.
Older, low-flux dialysis membranes removed little vancomycin, and generations of clinicians learned to treat it as essentially "non-dialyzable." Modern high-flux and high-efficiency membranes, widely used today, behave very differently — they can remove a clinically significant fraction of circulating drug during a single session.
The three modalities differ along several axes that all affect drug removal:
• Contact time: an intermittent hemodialysis session runs for a few hours a few times per week; CRRT runs continuously, 24 hours a day; peritoneal dialysis exchanges occur continuously but far more slowly. • Clearance mechanism: hemodialysis relies mainly on diffusion across a membrane driven by a concentration gradient; hemofiltration relies on convection (solvent drag) which can behave differently for a moderately protein-bound, mid-size molecule; peritoneal dialysis relies on slow diffusion and convection across the peritoneal membrane itself. • Membrane characteristics: flux (pore size), surface area, and material all influence how efficiently vancomycin crosses. • Flow rates: blood flow rate and dialysate/effluent flow rate set the pace of removal.
Because of these differences, a dosing plan that is correct for one modality can be badly wrong for another — either under-dosing (risking treatment failure and resistance) or over-dosing (risking nephrotoxicity, though residual renal function is usually minimal in these patients, or other toxicity).
Although exact numbers vary by patient and prescription, the qualitative pattern of vancomycin removal differs predictably by modality:
• Intermittent hemodialysis: removal is concentrated into a short window — while the patient is on the machine — and is often substantial for a several-hour high-efficiency session. Serum levels fall in a step-like fashion during the session, then rebound somewhat afterward as drug redistributes from tissue compartments. This session-driven, "sawtooth" pattern favors dosing that gives a base dose plus a supplemental dose after dialysis. • Continuous RRT: removal proceeds around the clock at a comparatively steady rate, more analogous (though not identical) to a low level of preserved native kidney function. This steadier clearance pattern generally supports a more standard maintenance-interval dosing approach rather than session-triggered supplementation. • Peritoneal dialysis: removal is typically the slowest and least efficient of the three for a molecule of vancomycin's size and protein binding, so systemic dosing generally needs comparatively little adjustment for removal via the peritoneal route — though intraperitoneal dosing has its own specific considerations for treating peritonitis.
The simulator above lets you toggle between these three modalities and see how each one's characteristic removal pattern reshapes the recommended dosing logic.
The single most important practical takeaway from comparing modalities: identify exactly which RRT modality — and which specific prescription within that modality — a patient is receiving before applying any dosing rule. "The patient is on dialysis" is not enough information.
Intermittent hemodialysis (IHD) concentrates vancomycin removal into discrete, several-hour sessions rather than spreading it evenly across the day. With modern high-flux membranes, a substantial fraction of circulating vancomycin can be removed during a single session — meaning that if no adjustment is made, drug levels can fall below the therapeutic target between sessions. The practical response is straightforward in concept: give a dose, and after the fraction removed by dialysis is accounted for, supplement what the session cleared.
Picture vancomycin serum concentration over the days between hemodialysis sessions: it rises after a dose, drifts down slowly as it distributes and is minimally cleared by whatever residual mechanisms remain, then falls more steeply during each hemodialysis session as the membrane actively removes drug, and finally shows a modest rebound afterward as vancomycin re-equilibrates from peripheral tissue compartments back into blood. Repeated across sessions, this produces a characteristic sawtooth pattern rather than the smoother decline seen with continuous clearance.
This pattern has direct dosing implications:
• Because a large, predictable drop occurs during each session, replacing that loss with a supplemental dose after the session helps keep trough levels from falling too low before the next dose. • Because levels rebound for a period after dialysis ends, a level drawn immediately at the end of a session can underestimate the "true" post-distribution concentration — timing of level draws relative to the session matters. • Because sessions are intermittent, the interval between doses is often built around the dialysis schedule itself (e.g., dosing after each session) rather than a fixed clock-hour interval.
The supplemental, post-hemodialysis dose is the direct practical answer to session-driven removal: rather than trying to guess in advance exactly how much a future session will remove and front-load the regimen, many approaches instead give a maintenance dose, let hemodialysis do its work, and then supplement afterward once the removal for that session has occurred.
This approach has intuitive appeal:
• It responds to what actually happened during the session rather than a prediction of what might happen. • It aligns naturally with when the patient is already at the dialysis unit and often already having blood drawn, making it a convenient time to also check a level. • It avoids "double dosing" risk that could occur if a scheduled maintenance dose were given right before a session, only to have much of it removed anyway.
The simulator's "time since last dialysis session" slider illustrates this: shortly after a session ends, a supplemental dose is typically due; as more hours pass without a session, that specific post-session dosing decision is no longer the active consideration until the next session approaches.
Continuous renal replacement therapy (CRRT) — delivered as continuous venovenous hemofiltration (CVVH), hemodialysis (CVVHD), or hemodiafiltration (CVVHDF) — runs 24 hours a day rather than in discrete sessions. Because clearance is applied essentially continuously rather than in bursts, the resulting vancomycin concentration profile is comparatively smooth, more reminiscent in shape (though not in magnitude) of clearance provided by a degree of preserved native kidney function than of the sawtooth pattern seen with intermittent hemodialysis.
Because CRRT clearance is ongoing rather than concentrated into a few hours a few times a week, there is no single "session" whose removal needs to be replaced afterward. Instead, the ongoing clearance behaves more like a component of total body clearance that is simply present around the clock — conceptually closer to residual native renal clearance than to the abrupt, large removal of an intermittent hemodialysis session.
This generally supports:
• A standard maintenance-style dosing interval (a loading dose followed by doses at a fixed interval, or a continuous infusion) rather than a regimen built around discrete post-treatment supplementation. • Relative stability of the dosing plan as long as the CRRT prescription itself does not change — because the clearance being provided is roughly constant hour to hour. • Careful attention to what happens if CRRT is interrupted (for filter clotting, line issues, or transport off the unit) — because clearance can drop suddenly if therapy pauses, even though the regimen was designed around continuous clearance being present.
It is important not to over-simplify "steady" as "negligible." Continuous therapies, especially at higher effluent flow rates or with certain filter types, can provide clearance that is not trivial compared to normal renal function — meaning that CRRT patients are not simply anuric patients who can be dosed as if RRT weren't running. The steadiness of the clearance changes the shape of the dosing problem (no abrupt session-driven drops to chase) but does not mean the clearance can be ignored.
Because the achieved clearance depends heavily on the specific CRRT prescription — filter, effluent rate, and modality variant — the next stage looks specifically at how those settings change the picture, and why "the patient is on CRRT" still leaves an important question unanswered: which CRRT, running how?
Not all CRRT prescriptions are equivalent. The membrane/filter used, whether clearance is achieved by convection (hemofiltration), diffusion (hemodialysis), or both (hemodiafiltration), and — critically — the prescribed effluent flow rate (the combined rate of ultrafiltrate and/or dialysate flow) together determine how much vancomycin clearance a given CRRT circuit is actually providing at any moment. Dosing should be built around the specific settings in use, not a generic assumption that "CRRT" means one fixed level of clearance.
CRRT can be delivered in several configurations that clear drugs by different physical mechanisms:
• CVVH (continuous venovenous hemofiltration): clearance is convective — solute is dragged along with fluid across the membrane (solvent drag). For a molecule like vancomycin with meaningful protein binding, only the unbound fraction moves with the filtrate. • CVVHD (continuous venovenous hemodialysis): clearance is diffusive — a countercurrent dialysate flow creates a concentration gradient across the membrane, similar in principle to intermittent hemodialysis but continuous and lower-flow. • CVVHDF (continuous venovenous hemodiafiltration): combines both convective and diffusive clearance, generally providing higher combined clearance than either alone at comparable flow settings.
The filter/membrane itself — its surface area, pore structure, and flux characteristics — also sets an upper bound on how efficiently vancomycin can cross, independent of flow rates.
Effluent flow rate — the total rate at which fluid (ultrafiltrate plus, where applicable, dialysate) leaves the circuit — is one of the strongest levers on achieved drug clearance in CRRT. Higher effluent rates generally translate into higher vancomycin clearance, all else equal, because more plasma water (carrying unbound drug) is being processed and removed per unit time.
Practical consequences for dosing:
• Two patients both described simply as "on CRRT" can have meaningfully different vancomycin clearance if their effluent rates differ. • If the ICU team increases or decreases the effluent flow rate — a routine adjustment made for reasons unrelated to vancomycin, such as fluid balance or metabolic control — the vancomycin clearance being provided changes too, and the dosing plan (and the interpretation of levels drawn under the old settings) should be reconsidered. • Filter changes (clotting and replacement, or a change to a different membrane) can also shift clearance and are a reasonable trigger to re-evaluate dosing and re-check a level.
In short: the CRRT prescription is a moving target that vancomycin dosing needs to track, not a fixed background condition.
A useful mental model: CRRT clearance of vancomycin scales roughly with how much plasma water is being processed by the circuit. Effluent flow rate is the number that most directly captures that — which is why it deserves attention alongside modality type when reasoning about dosing.
Across every modality discussed so far, one theme recurs: drug removal on RRT is highly variable — between patients, between modalities, and even between two patients nominally on "the same" modality but with different prescriptions, residual native kidney function, volume status, or protein binding. Standardized nomograms provide a reasonable starting point, but therapeutic drug monitoring (TDM) — measuring actual vancomycin concentrations and adjusting the regimen accordingly — is what actually individualizes dosing safely and effectively in this population.
In patients with stable, normal renal function, population pharmacokinetic estimates and simple weight-based dosing can get reasonably close to target exposure without frequent level checks. RRT patients are a different situation entirely:
• Their non-renal clearance may itself be altered by critical illness, other organ dysfunction, or the underlying condition that led to RRT. • The RRT-related clearance component can change hour to hour (CRRT settings, filter changes) or session to session (IHD schedule, session length, blood flow rate, whether a session was shortened or skipped). • Volume of distribution is frequently altered by critical illness — capillary leak, aggressive fluid resuscitation, and third-spacing are common and change how a given dose translates into peak and trough concentrations. • Residual native kidney function, when present, adds an additional and sometimes fluctuating clearance pathway on top of the RRT itself.
Given this combination of factors, a dose that is calculated from a population-average nomogram can be substantially wrong for a specific patient on a specific day — in either direction.
Level-guided (TDM-based) dosing does not replace an initial dosing estimate — it builds on one. The general pattern is:
1. Start with a reasonable initial regimen based on the specific modality and prescription in use (as discussed in the previous stages), informed by patient weight and clinical severity. 2. Draw a level at a clinically meaningful time point — for intermittent hemodialysis, this often means levels timed around the session (e.g., before the next session, and sometimes after, accounting for post-dialysis rebound); for continuous therapies, a level can generally be drawn once steady state is reasonably expected. 3. Compare the measured concentration to the therapeutic target for the indication being treated. 4. Adjust the dose, the interval, or both based on the result — rather than assuming the initial estimate remains correct indefinitely. 5. Re-check levels after any meaningful change: a new hemodialysis schedule, a CRRT setting change, a filter change, or a change in the patient's clinical trajectory (e.g., recovering native renal function, worsening hemodynamics).
This loop — estimate, measure, adjust, re-measure — is what actually keeps a patient in the therapeutic range despite the many moving parts that RRT introduces, and is the reason TDM is emphasized as essential rather than optional in this population.
The core message across every stage of this simulator: modality-specific starting doses are necessary but not sufficient. Because removal on RRT is so variable, measured drug levels — not the nomogram alone — should ultimately guide the dose a given patient receives.