Interpreting Paired Peak-and-Trough Samples

Why Pair a Peak With a Trough?

A trough records concentration near the end of a dosing interval, but it does not show how high concentration rose earlier in that interval or how sharply it fell. Two dosing profiles can therefore have the same trough and different peaks. A paired peak-and-trough assessment adds information about the observed within-interval fluctuation and can flag a profile that deserves closer review.

The pair is interpretable only in context. The concentrations should come from the same patient and dosing interval, and their collection times should be considered relative to dose administration and, where relevant, infusion duration. A sample labelled “peak” is not automatically the true maximum concentration; it is a measurement taken at a specified postdose time. Likewise, a sample is a trough measurement only if its timing is sufficiently close to the next dose for the intended interpretation.

Definition: For paired peak and trough concentrations \(C_{\text{peak}}\) and \(C_{\text{trough}}\), the peak-to-trough ratio is \(C_{\text{peak}}/C_{\text{trough}}\), the peak-to-trough difference is \(C_{\text{peak}}-C_{\text{trough}}\), and the relative decline from peak to trough is \((C_{\text{peak}}-C_{\text{trough}})/C_{\text{peak}}\). These describe the sampled concentrations; they represent the true within-interval maximum and minimum only when the sampling times capture those points adequately.

The ratio is dimensionless, while the difference retains the concentration units. The relative decline is often expressed as a percentage. Reporting more than one measure can help: a ratio emphasizes proportional fluctuation, whereas an absolute difference shows the concentration change in clinically familiar units. None of these quantities alone establishes whether a regimen is appropriate; interpretation still depends on the drug-specific target, dosing history, clinical context, and sample timing.

Relating the Pair to Elimination

A simple model can help determine whether a pair is consistent with a plausible decline between samples. Suppose drug input has ended, distribution is complete, and concentration declines by first-order elimination in a one-compartment model. Let \(k\) be the elimination rate constant, and let \(\Delta t\) be the elapsed time from the first sample to the second. Under these assumptions, concentration follows an exponential decline.

\frac{dC}{dt}=-kC

Separating variables and integrating from the first sample, with concentration \(C_1\), to the second sample, with concentration \(C_2\), gives:

\int_{C_1}^{C_2}\frac{1}{C}\,dC = -k\int_{t_1}^{t_2}dt \quad\Longrightarrow\quad \ln\left(\frac{C_2}{C_1}\right)=-k(t_2-t_1)

Thus, if the first sample is the post-input peak sample and the second is the later trough sample, the model-based apparent elimination rate constant and half-life are:

k_{\text{app}}=\frac{\ln(C_{\text{peak}}/C_{\text{trough}})}{\Delta t}, \qquad t_{1/2,\text{app}}=\frac{\ln 2}{k_{\text{app}}}

This derivation depends on a real decline between samples, accurate concentration measurements and times, and a suitable monoexponential elimination phase. If drug is still being absorbed or infused, distribution is ongoing, the kinetics are not first order, or either sample time is wrong, the estimated value may not represent elimination. In contrast, the descriptive ratio and difference can still be calculated, but they must be described as summaries of the sampled values rather than as model-based pharmacokinetic estimates.

Worked Example: Predicting a Trough From a Six-Hour Half-Life

Consider a hypothetical patient with a measured post-input peak of 24 mg/L and a trough sample collected 13 hours later. Assume, for this calculation, that both samples lie on a monoexponential elimination phase and that the relevant half-life is 6 hours. What trough does the model predict, and how much decline does that represent?

A half-life of 6 hours means concentration is multiplied by \(2^{-\Delta t/6}\) over an elapsed time of \(\Delta t\). The predicted trough is therefore:

C_{\text{trough,pred}} = 24\times 2^{-13/6} = \frac{24}{2^{13/6}}

Since \(2^{13/6}\approx 4.48985\), the predicted concentration is \(24/4.48985\approx 5.345\) mg/L, or about 5.35 mg/L. The peak-to-trough ratio is approximately 4.49. The relative decline is \((24-5.345)/24\approx 0.7773\), or 77.73%; the absolute decline is approximately 18.66 mg/L.

These are predictions under the stated model, not a claim that every patient with a six-hour half-life will have this observed pair. If the measured trough differed, the discrepancy would prompt a check of dose and sample times, the assumed half-life, and the suitability of the model before drawing conclusions.

Using the Pair to Check Timing

Documented sample times can be compared with the decline implied by a model or an independently supported half-life. This is a consistency check, not a way to prove that the timestamps are correct. A mismatch can arise from inaccurate times, an incorrect assumption about the elimination phase, patient-specific kinetics that differ from the assumed value, or other features of the concentration profile.

Worked Example: Checking a Pair Against Documented Times

In a hypothetical infusion case, the post-input sample is 18 mg/L and is collected 1 hour after the infusion ends. The trough sample is 6 mg/L and is collected 12 hours after the infusion ends. Thus, the elapsed time between samples is 11 hours. Suppose a six-hour half-life is being used as a provisional expectation, and assume both samples are in a monoexponential elimination phase.

The predicted peak-to-trough ratio across 11 hours is \(2^{11/6}\approx 3.56359\). The corresponding predicted trough is:

C_{\text{trough,pred}} = \frac{18}{2^{11/6}} \approx \frac{18}{3.56359} \approx 5.051\ \text{mg/L}

The observed ratio is \(18/6=3\), and the observed trough is 6 mg/L, about 0.949 mg/L above the prediction. If the monoexponential assumptions hold, the pair implies:

k_{\text{app}}=\frac{\ln(18/6)}{11} =\frac{\ln 3}{11} \approx 0.09987\ \text{h}^{-1}, \qquad t_{1/2,\text{app}} =\frac{\ln 2}{0.09987} \approx 6.941\ \text{h}

The apparent half-life is longer than the provisional six-hour value. That difference does not establish a timing error: it indicates that the pair and the provisional expectation do not match exactly under the assumed model. Review the recorded administration and collection times, and consider whether the samples actually represent post-input elimination and a predose trough. Other explanations remain possible, so the pair alone cannot identify which explanation is correct.

What the Trough Alone Can Hide

When a clinical decision depends on both a low concentration near the end of the interval and a high concentration after dosing, a trough alone cannot characterize both parts of the profile. A paired sample makes the observed fluctuation explicit. It can show, for example, that a trough that appears acceptable coexists with a much higher sampled peak. Whether that peak is concerning must be judged against the relevant drug-specific target and clinical context; the arithmetic does not create a target.

Worked Example: Same Trough, Different Fluctuation

Two hypothetical profiles have trough concentrations of 4 mg/L, with each peak sampled 8 hours before its paired trough. Profile A has a peak of 16 mg/L; Profile B has a peak of 8 mg/L. Compare their sampled fluctuation and, conditionally, their apparent half-lives.

For Profile A, the peak-to-trough ratio is \(16/4=4\), the difference is \(16-4=12\) mg/L, and the relative decline is \(12/16=0.75\), or 75%. For Profile B, the ratio is \(8/4=2\), the difference is \(8-4=4\) mg/L, and the relative decline is \(4/8=0.50\), or 50%.

If both pairs are in monoexponential elimination phases, Profile A implies \(k_{\text{app}}=\ln(4)/8\approx 0.1733\ \text{h}^{-1}\) and \(t_{1/2,\text{app}}=\ln(2)/0.1733\approx 4.000\) hours. Profile B implies \(k_{\text{app}}=\ln(2)/8\approx 0.08664\ \text{h}^{-1}\) and \(t_{1/2,\text{app}}=\ln(2)/0.08664\approx 8.000\) hours.

The common trough does not make these profiles equivalent: their sampled peaks and fluctuation differ. The half-lives are conditional model-based summaries, not evidence that the profiles differ only because of elimination. Differences in dose, input, sampling time, or model fit could also matter.

Assumptions, Diagnostics, and Alternatives

Before treating a pair as a peak-to-trough profile, establish that both measurements belong to the same patient and dosing interval and check the actual collection times against the dose or infusion record. Confirm that the “peak” sample was collected at a time appropriate for the drug and route, and that the trough was collected sufficiently close to the next dose. For oral drugs, absorption may still be occurring near a sampled high concentration; for some drugs, distribution may produce an early concentration decline that is not the terminal elimination phase. The sampled maximum and minimum may therefore differ from the true within-interval maximum and minimum.

When estimating \(k_{\text{app}}\), verify \(C_{\text{peak}}>C_{\text{trough}}>0\) and \(\Delta t>0\). A nonpositive concentration, reversed ordering, or nonpositive time interval makes the stated calculation invalid. An unexpectedly large or small ratio should lead to review of the sample and dose timestamps and the model assumptions, not automatic rejection of one value. If \(C_{\text{peak}}/C_{\text{trough}}\) is close to 1, the logarithm is close to zero, so a half-life estimate becomes very large and highly sensitive to small concentration differences. A very large ratio can likewise magnify the consequences of incorrect timing or a sample taken during distribution.

Paired peak-and-trough interpretation complements, rather than replaces, other monitoring approaches. A trough-only approach directly addresses a trough target but cannot describe the sampled peak. A paired assessment describes two points and their fluctuation, but those two points do not by themselves provide the full concentration-time curve or establish total exposure. An AUC-based approach addresses exposure over an interval using its own sampling and estimation requirements. Which approach is appropriate depends on the drug-specific monitoring objective; this comparison does not make the measures interchangeable.

Approach	What it directly summarizes	Important limitation
Trough alone	Concentration near the end of an interval	Does not show the earlier sampled peak or within-interval fluctuation
Paired peak and trough	Two concentrations, their ratio or difference, and—under assumptions—an apparent decline rate	Two samples may miss the true maximum, minimum, or important profile features
AUC-based monitoring	Exposure integrated over a dosing interval, when suitably estimated	Requires appropriate sampling or estimation; it is not obtained merely by taking a peak and trough ratio

Common Mistake and Key Takeaway

Wrong reasoning: “The trough is in range, so the whole concentration profile must be acceptable.” A trough describes concentration near one part of the interval; it cannot establish what happened earlier. Correction: If peak exposure or fluctuation is relevant to the monitoring question, interpret the paired peak and trough with their actual times and appropriate drug-specific targets. Conversely, do not infer a true maximum or elimination half-life from two samples unless the sampling and model assumptions support that interpretation.

Key takeaway: A paired peak-and-trough result adds information about observed within-interval fluctuation and may expose a profile that a trough alone obscures. The concentration ratio can be translated into an apparent elimination rate only when timing and monoexponential post-input decline assumptions are reasonable.

Check Your Understanding

Try each problem, then compare your reasoning with the solutions below. All concentrations and data are hypothetical.

A peak is 20 mg/L and its paired trough is 5 mg/L. Calculate the ratio, difference, and relative decline.
A peak of 12 mg/L and trough of 3 mg/L are separated by 6 hours. Under a monoexponential elimination assumption, calculate \(k_{\text{app}}\) and \(t_{1/2,\text{app}}\).
A peak sample is collected 2 hours after an infusion ends and a trough sample 14 hours after the infusion ends. What elapsed time belongs in the paired-sample calculation?
Two hypothetical profiles share a trough of 2 mg/L, but their peaks are 6 mg/L and 10 mg/L. What conclusion can be made from the pair, and what cannot be concluded without more context?

Solutions. (1) The ratio is \(20/5=4\), the difference is \(20-5=15\) mg/L, and the relative decline is \(15/20=0.75\), or 75%. (2) The elapsed time is 6 hours, so \(k_{\text{app}}=\ln(12/3)/6=\ln(4)/6\approx 0.2310\ \text{h}^{-1}\). The apparent half-life is \(\ln(2)/0.2310\approx 3.000\) hours. This estimate depends on the assumed monoexponential elimination phase and accurate times. (3) The elapsed time between the samples is \(14-2=12\) hours; it is not 14 hours, because the first sample was already collected 2 hours after infusion end. (4) The sampled ratios are \(6/2=3\) and \(10/2=5\), so the profiles have the same trough but different sampled peaks and ratios. Without information about sampling times, targets, dosing, and clinical context, one cannot determine whether either profile is appropriate or attribute the difference to a particular pharmacokinetic cause.