Why Similar Titration Patterns Can Have Different Causes
During titration, a dose may be increased because the measured response is below target, yet the response may change little—or may later decline. That pattern alone does not reveal why. Drug concentration may still be rising as doses accumulate. Clearance may be changing, altering exposure at a given dose. Alternatively, the pharmacodynamic response may diminish at comparable exposure, consistent with tolerance. These explanations imply different expectations for what will happen after the next dose adjustment, so distinguishing them matters for interpreting titration data.
The key is to examine dose history, exposure, and response together. A response measured at one visit is rarely enough. Concentration sampling must be interpreted in relation to dose timing, and response must be compared at similar exposures while allowing for measurement noise, response delay, and changes in the underlying condition.
Start With the Expected Accumulation Pattern
A simple reference model helps establish what accumulation should look like. Assume instantaneous intravenous bolus doses \(D\) are administered every \(\tau\) hours to a one-compartment system with volume \(V\) and first-order elimination rate constant \(k\). Assume \(V\) and \(k\) remain constant, doses are taken as scheduled, and concentrations from separate doses add by linear superposition. After a single dose, concentration declines exponentially with half-life \(\ln(2)/k\).
Immediately after the \(n\)th dose, the contribution from the current dose is \(D/V\); contributions from prior doses have decayed for one or more dosing intervals. Thus:
The sum is a finite geometric series. If dosing continues long enough under these assumptions, the post-dose concentration approaches \( (D/V)/(1-e^{-k\tau}) \). Relative to the post-dose concentration after a single dose, the steady-state accumulation factor is \(1/(1-e^{-k\tau})\). This is a model-based expectation, not a rule that every patient must follow exactly.
Worked Example: Recognizing Accumulation at a Constant Dose
Suppose a hypothetical patient receives 100 mg intravenously every 24 hours. Under a one-compartment model, the half-life is 24 hours and \(D/V=10\) mg/L. Assume constant clearance, complete adherence, and linear elimination. What post-dose concentrations does the model predict after the first three doses, and what is the steady-state accumulation factor?
The elimination constant is \(k=\ln(2)/24=0.02888\ \mathrm{h}^{-1}\). Over one 24-hour interval, \(e^{-k\tau}=0.5\). The first post-dose concentration is \(10\) mg/L. After the second dose:
After the third dose, the first dose has undergone two intervals of elimination, so:
The steady-state post-dose concentration is \(10/(1-0.5)=20\) mg/L, giving an accumulation factor of \(2.000\) relative to the first-dose post-dose concentration. Under these assumptions, rising concentrations at an unchanged dose are expected; they do not by themselves indicate tolerance or a change in clearance.
In practice, a trough may be more useful than an immediate post-dose concentration, but its value also depends on sampling time. For this example, the trough just before the second dose is \(10(0.5)=5\) mg/L; the trough just before the third is \(15(0.5)=7.5\) mg/L. Comparing samples taken at different points in the dosing interval can therefore create an apparent trend that is only a timing difference.
Compare Response at Comparable Exposure
Accumulation is a pharmacokinetic change in concentration over time. Tolerance is a pharmacodynamic change in the response conditional on exposure. A useful conceptual model separates these components:
Here \(E(t)\) is a baseline-adjusted response, \(C(t)\) is the relevant exposure measure at time \(t\), \(E_{\max}\) is the maximal response in the absence of adaptation, and \(EC_{50}\) is the concentration associated with half that maximal response. The factor \(q(t)\) represents time-varying pharmacodynamic sensitivity; a decline in \(q(t)\) can represent tolerance. The residual \(\epsilon(t)\) represents unexplained measurement and biological variation. This model is a simplifying description: the relevant exposure may be a concentration at a particular time, an average, or a delayed exposure metric, depending on the drug and response.
If \(q(t)\) is constant and the concentration-response relation is stable, higher exposure should produce the response predicted by that relation. If exposure is comparable but response is repeatedly lower later, tolerance is one possible explanation. It is not the only one: disease progression, a change in concomitant treatment, altered response measurement, or a delay between exposure and effect can produce a similar pattern.
Worked Example: Lower Response at the Same Concentration
In a hypothetical titration, the same concentration metric is 10 mg/L at an early visit and a later visit. Assume, for illustration, that \(E_{\max}=100\) response units and \(EC_{50}=10\) mg/L. The measured baseline-adjusted responses are 40 units early and 24 units later. Estimate the sensitivity factor \(q(t)\) at each visit and interpret the comparison.
At 10 mg/L, the concentration-response component before applying \(q(t)\) is:
Ignoring residual error for this simple calculation, \(q_{\mathrm{early}}=40/50=0.8000\), while \(q_{\mathrm{late}}=24/50=0.4800\). The later response is 16 units lower, a decrease of \(16/40=0.4000\), or 40.00%, relative to the early response at the same concentration. This pattern is consistent with reduced pharmacodynamic sensitivity. It does not prove tolerance: the interpretation depends on the exposure metric being comparable and on other determinants of response remaining sufficiently stable.
Check Whether Clearance Could Explain the Exposure Pattern
For linear pharmacokinetics after an intravenous dose, exposure over a complete concentration-time profile is related to dose and clearance by \(AUC=D/CL\), where \(AUC\) is the area under the concentration-time curve and \(CL\) is clearance. This relationship assumes the dose is known, the profile captures the relevant interval, and the linear model is appropriate. For an oral dose, bioavailability also affects AUC, so a change in absorption or bioavailability can resemble a change in clearance unless it is addressed.
With repeated dosing or changing doses, compare exposure with the dose history rather than interpreting concentration alone. At a fixed dose and stable bioavailability, a rising AUC across comparable intervals can be consistent with falling clearance; a falling AUC can be consistent with rising clearance. An accumulation pattern under stable clearance should instead be assessed against the expected approach to steady state. If dose, interval, adherence, or sampling time changes, the comparison requires a model that accounts for those changes.
Worked Example: Detecting a Possible Change in Clearance
A hypothetical patient receives the same 100 mg intravenous dose at two separated assessments. Complete profiles give AUC values of 20 and 12.5 mg·h/L, respectively. Assume linear pharmacokinetics and that dose and profile collection are comparable. Calculate apparent clearance at each assessment.
At the first assessment:
At the second assessment:
The later apparent clearance is \(8/5=1.600\), or 60.00% higher, and AUC is lower despite the same dose. If response also declined, reduced exposure from increased clearance could explain the pattern without invoking tolerance. The calculation identifies an exposure change consistent with altered clearance under the stated assumptions; it does not establish its biological cause, and alternative explanations such as changed bioavailability would need consideration for non-intravenous dosing.
A Practical Interpretation Sequence
List dose amounts, dosing times, holds, restarts, and adherence information. Mark the exact times of concentration and response measurements. Do not treat a sample after a dose change as though it came from a stable regimen.
At unchanged dosing, compare measured concentrations with the expected accumulation trajectory under a plausible PK model. Check sampling time and whether the observation period is long enough to approach steady state.
Use dose-normalized exposure or a fitted PK model when profiles and assumptions permit. A changing dose-normalized exposure pattern suggests that clearance, bioavailability, or another PK process may have changed.
Evaluate whether response is lower at comparable exposure, allowing for expected response delay and measurement variability. A reproducible decline conditional on exposure supports tolerance as a candidate explanation, not as a unique diagnosis.
Diagnostics, Alternatives, and Limitations
The strongest evidence comes from multiple observations that distinguish the competing explanations. A concentration increase after each dose at an unchanged regimen, with a trajectory consistent with the PK model, supports accumulation. A sustained shift in dose-normalized exposure at a fixed regimen suggests a PK change, although clearance cannot always be separated from absorption or adherence using sparse concentrations. A lower response at similar relevant exposure, replicated over time and not explained by delayed response or other changes, supports a time-varying PD effect.
These patterns can overlap. Accumulation and tolerance may occur together: concentration can rise while response sensitivity falls. A dose escalation can also mask a decline in clearance by maintaining exposure, or mask tolerance by raising exposure enough to preserve response. In such cases, interpreting dose and response alone is especially unreliable.
| Candidate explanation | Pattern that supports it | Important check |
|---|---|---|
| Accumulation | Concentration rises at an unchanged dose in a pattern compatible with residual drug from prior doses. | Verify dose timing, sampling time, and whether the expected accumulation period has elapsed. |
| Time-varying clearance | Exposure at a comparable dose changes over time; the direction of change is consistent with altered elimination. | Consider adherence, bioavailability, dose history, and whether AUC or the PK model is adequately characterized. |
| Tolerance | Response falls at comparable relevant exposure, or greater exposure is needed for a similar response. | Consider response delay, disease change, concomitant treatment, and measurement reliability. |
A fitted joint PK-PD model can formally represent concentration and response over time, including dose history and a time-varying sensitivity term. It can be more informative than examining concentrations or responses separately, but a complicated model does not create information absent from the data. If concentration sampling is sparse, dose and clearance effects may be difficult to separate. If exposure and time move together—for example, concentrations always rise as response sensitivity is hypothesized to fall—the data may not distinguish accumulation from tolerance without observations at overlapping exposure levels.
The simpler alternative is descriptive comparison: inspect concentration-time and response-time plots, stratified by dose and sampling time, and compare periods with similar exposure. This is transparent and useful for identifying patterns, but it does not fully adjust for changing dose history or uncertainty. Neither approach justifies a causal claim from a simple temporal association. The interpretation should state what the observations support and what remains unresolved.
Common Mistake
The central limitation is identifiability: different mechanisms can produce similar observed dose-response histories. Better-timed concentration samples, stable-dose observation periods when clinically appropriate, and repeated response measurements at overlapping exposures can help separate them. Clinical decisions must also account for safety; observation or sampling should not delay a necessary dose reduction or other intervention.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- Under constant clearance and a fixed dosing interval, why can concentration rise after repeated doses even when each dose is unchanged?
- A response is lower at a later visit, but concentration was not measured and the dose was increased shortly before that visit. Can this establish tolerance? Why or why not?
- At the same intravenous dose, AUC falls from 30 to 20 mg·h/L. By what factor does apparent clearance change, assuming linear PK?
- What observation most directly supports tolerance rather than simple accumulation?
- Name two reasons why falling response at similar measured concentration might not be tolerance.
Solutions. (1) Residual drug from earlier doses remains when the next dose is administered, so successive dose contributions add until elimination and input balance at steady state. (2) No. The dose history and unmeasured exposure prevent assessment of response conditional on exposure; delayed response, ongoing accumulation, and changing clearance remain possible. (3) Since \(CL=D/AUC\), \(CL_2/CL_1=30/20=1.500\): apparent clearance is 1.5 times its earlier value, a 50.00% increase. (4) A reproducibly lower response at comparable relevant exposure, after accounting for response delay and other changes, supports tolerance. (5) Examples include disease progression, a change in concomitant treatment, response-measurement error, or an exposure metric that fails to capture delayed effect-site exposure.