Безперервна інфузія ванкоміцину до рівноважного стану — steady-state pharmacokinetics vs. intermittent peak/trough dosing
Vancomycin has traditionally been dosed intermittently: a fixed dose infused over 1–2 hours every 8–12 hours, producing a sawtooth concentration-time profile with a peak shortly after infusion ends and a trough just before the next dose. Continuous infusion (CI) replaces this cycle with a constant-rate infusion pump running around the clock, aiming to hold the plasma concentration at one stable value once steady state is reached — trading the peak/trough swings of intermittent dosing for a flat, predictable exposure.
With intermittent infusion, vancomycin is delivered as a bolus-like dose over a short window (commonly 60–120 minutes), then the drug distributes and is eliminated until the next dose. Because vancomycin follows roughly first-order, two-compartment kinetics, the concentration curve within each interval looks like a rapid rise during infusion followed by an exponential decline:
C(t) = C_peak · e^(−k·t)
This produces a repeating pattern: a peak shortly after the infusion ends, a trough just before the next dose, and everything in between. The peak-to-trough ratio can be substantial — a typical q12h regimen may show peaks near 30–40 mg/L and troughs near 10–15 mg/L, even though the average concentration sits in a therapeutic zone. Clinicians must time blood draws precisely (usually just before the next dose, at true trough) to interpret the level correctly; drawing too early or too late systematically biases the result.
Continuous infusion removes the on/off cycle entirely. The drug is delivered at a constant rate (mg/hour) via infusion pump, so instead of alternating between infusion and elimination-only phases, absorption (infusion) and elimination happen simultaneously and continuously. Mathematically, the concentration approaches a steady-state value:
Css = Rate / CL
where CL is the patient's vancomycin clearance (largely determined by renal function). Because the input rate never stops, the curve does not have a decay phase — it simply approaches Css asymptotically and then stays there (barring changes in rate or clearance). The practical consequence is that a single concentration measured any time after steady state is reached is directly interpretable, without needing peak/trough timing precision.
A hallmark of first-order pharmacokinetics is that any constant-rate input approaches its steady-state concentration exponentially, governed entirely by the drug's elimination half-life — not by the infusion rate itself. For vancomycin, with a typical half-life of roughly 4–6 hours in patients with normal renal function, reaching steady state through the infusion alone would take the better part of a day. A loading dose front-loads the body with drug so that a therapeutic concentration is achieved almost immediately, while the continuous infusion then simply maintains it.
When a constant infusion begins in a drug-naive patient, plasma concentration rises according to:
C(t) = Css · (1 − e^(−k·t))
where k = 0.693 / t½. This curve is entirely predictable in shape: after one half-life the concentration is at 50% of Css, after two half-lives 75%, after roughly 3.3 half-lives about 90%, and after 4–5 half-lives it is considered to have essentially reached steady state (>94–97%). Critically, this timeline depends only on the half-life — increasing the infusion rate changes what Css will ultimately be, but not how quickly (in relative terms) the curve approaches it. This is why simply "running the pump faster" cannot shortcut the wait for steady state; it only changes the plateau it eventually reaches.
Because waiting 24–30 hours for therapeutic concentrations is clinically unacceptable in serious infections (bacteremia, sepsis, endocarditis, CNS infections), a loading dose is administered at the start of therapy — typically 20–25 mg/kg based on actual body weight, infused over 1.5–2 hours to reduce infusion-related reactions. This bolus immediately raises the plasma concentration close to the intended steady-state target. The continuous infusion is started at (or shortly after) the loading dose and simply has to maintain that level rather than build it up from zero — collapsing the effective "time to therapeutic concentration" from roughly a day down to 1–2 hours.
Because the loading dose distributes into peripheral tissue compartments over the first few hours, a confirmatory level is usually not drawn until the infusion itself has run long enough to approach steady state (often re-checked at 24 hours), even though the loading dose already achieved an early therapeutic concentration.
Because a continuous infusion produces a flat concentration-time curve once steady state is reached, monitoring is dramatically simplified compared to intermittent dosing. A single random blood level — drawn at any convenient time after steady state has been confirmed — is directly equal to Css, and from Css the total daily drug exposure (AUC24) can be estimated with a simple multiplication, without needing to time the draw around a dose.
Under intermittent dosing, estimating total drug exposure (AUC) requires either a validated peak-and-trough pair fed into a two-compartment model, or a Bayesian software estimate, precisely because the concentration is constantly changing across the dosing interval. Under continuous infusion, the concentration curve is effectively flat at steady state, so:
AUC24 ≈ Css × 24 hours
This relationship is illustrative and assumes the level truly reflects steady state (not still rising) and that the infusion rate has not changed recently. It removes the sampling-time sensitivity that makes intermittent-dosing troughs notoriously easy to mistime in busy clinical settings — a level drawn an hour early or late during continuous infusion barely changes the result, whereas the same timing error on an intermittent trough can meaningfully bias dosing decisions.
A typical continuous-infusion monitoring workflow: give the loading dose, start the infusion at the calculated maintenance rate, then draw a random level roughly 24 hours later (allowing the curve to approach steady state after any loading-dose distribution effects settle). If the level falls within the target range, monitoring can often shift to periodic checks (e.g., every 24–48 hours, or with changes in renal function). If the level is outside the target range, the infusion rate is adjusted and a new level is checked after another approach-to-steady-state interval — again roughly 4–5 half-lives after the rate change, though in practice many protocols re-check sooner given clinical urgency.
Because steady-state concentration is directly proportional to infusion rate (Css = Rate / CL, with clearance assumed roughly constant over the short term), titrating the infusion rate based on a measured level is a straightforward proportional adjustment — unlike intermittent dosing, where changing peak and trough targets can require recalculating both dose size and interval.
If a measured steady-state level (Css_measured) differs from the target, and clearance has not meaningfully changed since the level was drawn, then a new rate can be estimated by simple proportion:
New Rate = Current Rate × (Target Css / Measured Css)
For example, if a patient is on 50 mg/hour and the measured level comes back at 30 mg/L against a target midpoint of 22.5 mg/L, the suggested new rate is 50 × (22.5/30) ≈ 37.5 mg/hour — a reduction. Conversely, a measured level of 15 mg/L on the same 50 mg/hour rate suggests increasing toward 50 × (22.5/15) = 75 mg/hour. This is only a starting estimate: clinicians round to practical pump settings, consider the trend of renal function, and re-check a level after allowing the system to re-equilibrate before making further changes.
This simple scaling works because, at steady state, input rate exactly equals elimination rate (Rate = CL × Css), so Css and Rate move together linearly as long as clearance stays fixed over the adjustment interval. This is a meaningfully simpler relationship than intermittent dosing, where both the size of each dose and the interval between doses jointly determine peak, trough, and AUC — meaning a change intended to fix a high trough can inadvertently also lower the peak below target, requiring a more complex two-parameter adjustment. With continuous infusion, one parameter (rate) maps directly onto one outcome (Css), which is part of why titration is considered comparatively straightforward once a patient is stable at steady state.
This proportional approach is illustrative rather than a substitute for full pharmacokinetic dosing software — it assumes stable renal function and clearance between the measured level and the adjustment, which may not hold in patients with rapidly changing kidney function.
Continuous infusion is not the default strategy for every patient, but specific clinical situations make its steady-state delivery particularly attractive: patients in whom intermittent dosing struggles to hit AUC-based targets without excessive peaks, and patients receiving renal replacement therapy, where drug clearance is heavily influenced by the dialysis or filtration circuit rather than native kidney function alone.
Current vancomycin dosing guidance emphasizes AUC-guided targets (commonly AUC24/MIC ≥ 400, balanced against nephrotoxicity risk at higher exposures) rather than trough-only targets. In some patients — particularly those with augmented renal clearance, where glomerular filtration runs well above normal — intermittent dosing intervals may need to be shortened or doses increased substantially to maintain adequate trough/AUC, sometimes pushing peak concentrations uncomfortably high or requiring impractically frequent dosing. In these patients, continuous infusion can deliver the same total daily dose while avoiding the exaggerated peaks that come from compressing a large dose into a short infusion window, making the AUC target easier to hit predictably.
In patients receiving continuous renal replacement therapy (CRRT — continuous veno-venous hemofiltration/hemodialysis), vancomycin clearance becomes dominated by the filter/dialysis circuit rather than native renal function, and that clearance is itself continuous and relatively stable (governed by effluent flow rate, filter type, and convection/diffusion settings). A continuous vancomycin infusion pairs naturally with continuous clearance: because the removal process runs steadily around the clock, matching it with a steady input avoids the repeated large boluses that would otherwise need individual adjustment for whatever the circuit happened to remove since the last dose. Intermittent hemodialysis (running only a few hours, a few times a week) instead requires dosing strategies timed around each dialysis session — a different but related challenge outside the scope of continuous infusion.
The combination of augmented renal clearance (where intermittent dosing struggles to sustain AUC) and continuous renal replacement therapy (where clearance itself is continuous) are the two scenarios most consistently cited as favoring continuous infusion — both reflect situations where clearance behaves in a way that a matched, steady input handles more gracefully than a series of discrete doses.