Why Recovery and Assay Controls Matter
A plasma binding assay estimates how much drug is unbound under specified conditions. In equilibrium dialysis, for example, drug can move across a membrane from plasma into buffer, while plasma proteins remain in the plasma chamber. At equilibrium, a concentration ratio can estimate the fraction unbound—but only if the measured concentrations represent the same drug, the relevant free concentrations have equilibrated, and the assay has not lost a consequential amount of drug to the apparatus or another process.
The central diagnostic distinction is between redistribution and loss. Binding to plasma proteins redistributes drug between free and bound forms; it does not, by itself, remove drug from the assay. By contrast, adsorption to a membrane, degradation, precipitation, or unmeasured sampling can reduce the amount recovered from the compartments being analyzed. A low recovery therefore signals that the binding result needs investigation, not that the drug must be highly bound.
Mass Balance in a Two-Chamber Assay
Consider an equilibrium dialysis setup in which drug is initially added to the plasma chamber. Let \(V_p\) and \(V_b\) be the plasma- and buffer-chamber volumes, \(C_{p,0}\) the initial total drug concentration in plasma, and \(C_p\) and \(C_b\) the measured concentrations at the end of the assay. If the initial buffer contains no drug, then \(A_0=C_{p,0}V_p\). Ignoring sampled amounts for the moment, the liquid-phase recovery is
The products of concentration and volume are amounts. For example, \(1\ \mu\mathrm{M}\times1\ \mathrm{mL}=1\ \mathrm{nmol}\). If the assay uses unequal chamber volumes, both terms in the numerator are necessary; adding the two concentrations alone is not a mass balance. If samples are withdrawn during the assay, their drug amounts must also be added to the numerator unless the protocol replaces them and accounts for that replacement.
Under the standard interpretation of an equilibrium dialysis fraction unbound result, the free drug concentration is equal in the two chambers at equilibrium. This requires conditions that allow the relevant unbound drug to cross the membrane and avoid a persistent pH gradient. With matched conditions, the estimated fraction unbound is
Here \(C_p\) is the total concentration measured in plasma, whereas \(C_b\) represents the unbound concentration at equilibrium. The ratio is not justified merely because two concentrations were measured: equilibration, compatible pH conditions, and an appropriate assay mass balance are part of its interpretation.
Worked Example: Calculating Recovery and an Apparent Fraction Unbound
In a hypothetical dialysis assay, \(0.200\ \mathrm{mL}\) of plasma is initially spiked to \(100\ \mu\mathrm{M}\). The buffer chamber contains \(0.200\ \mathrm{mL}\). At the endpoint, the measured concentrations are \(80\ \mu\mathrm{M}\) in plasma and \(8.0\ \mu\mathrm{M}\) in buffer. Assume the reported volumes are the final chamber volumes and no samples were removed.
The initial amount is
The measured liquid-phase amount is
Thus,
The endpoint concentration ratio is \(C_b/C_p=8.0/80=0.100\). This is an apparent \(f_{u,p}\) of 0.100. The 12.0% of initial drug not accounted for in the liquid chambers cannot be classified as protein-bound: it is absent from the measured liquid-phase mass balance. Whether the ratio is usable depends on assay-specific acceptance criteria and controls that assess the missing amount, pH, and equilibration.
Diagnosing Low Recovery and Membrane Adsorption
Low recovery means that the measured compartments do not account for the expected initial amount within the laboratory’s predefined criteria. It does not identify the cause. Potential explanations include adsorption to the dialysis membrane, chamber walls, seals, or other surfaces; chemical or enzymatic degradation; precipitation; errors in the initial concentration or final volumes; and unrecorded sample removal. A low recovery can also arise from an analytical problem, such as matrix-dependent quantification or a concentration below the validated quantification range.
A useful control is to expose the drug to the assay apparatus under conditions that minimize or exclude plasma binding, then measure the liquid phase and, where feasible, extract or rinse the membrane and other surfaces. Drug recovered in a membrane rinse supports adsorption as a contributor to the missing liquid-phase amount. A buffer-only control can also reveal losses to the apparatus, although its conditions may not reproduce every aspect of a plasma-containing assay. The control should be interpreted as evidence about a possible mechanism, not as proof that the binding-assay result is corrected by a simple arithmetic adjustment.
Worked Example: Using a Membrane Rinse to Locate Missing Drug
In a hypothetical assay, \(0.200\ \mathrm{mL}\) of plasma is spiked to \(50.0\ \mu\mathrm{M}\), giving \(10.0\ \mathrm{nmol}\) initially. At the endpoint, \(0.200\ \mathrm{mL}\) of plasma contains \(32.0\ \mu\mathrm{M}\), and \(0.200\ \mathrm{mL}\) of buffer contains \(2.0\ \mu\mathrm{M}\).
The liquid-phase amount is
The liquid-phase recovery is \(6.80/10.0=68.0\%\). A validated membrane rinse then contains \(2.60\ \mathrm{nmol}\). Including that amount gives
The rinse accounts for a substantial part of the liquid-phase deficit, supporting membrane-associated drug as a likely source of loss. The endpoint ratio is \(2.0/32.0=0.0625\), but the rinse does not prove that this ratio is unbiased: adsorption may have changed the conditions during equilibration, and the rinse procedure itself must be validated. A repeat with a lower-binding apparatus or a suitable membrane control is more informative than treating the 94.0% recovery as an automatic correction to \(f_{u,p}\).
pH Shifts and Inadequate Equilibration
For an ionizable compound, pH can affect the distribution of ionized and un-ionized species. If the plasma and buffer chambers finish at different pH values, the measured total unbound concentrations need not be equal even when the species able to cross the membrane have equilibrated. Consequently, \(C_b/C_p\) may not estimate the intended fraction unbound under a common pH condition. Measuring only the starting pH is insufficient: check the pH in both chambers at the endpoint, using a procedure that does not materially disturb the assay. A pH difference is a warning to investigate and repeat under controlled conditions, not a general invitation to apply a universal numerical correction.
Equilibration is a separate requirement. An endpoint taken too early can make the buffer concentration appear low relative to the plasma concentration and therefore make the ratio appear smaller than its equilibrium value. A time-course control addresses this directly: measure both chambers at multiple times while keeping temperature, volumes, and other assay conditions consistent. Evidence for a practical plateau is that subsequent measurements are stable within the method’s analytical variability. A single endpoint cannot establish that equilibrium was reached.
Worked Example: Recognizing a pH-Related Interpretation Problem
A hypothetical assay reports \(C_p=50.0\ \mu\mathrm{M}\) and \(C_b=20.0\ \mu\mathrm{M}\), with endpoint pH 7.4 in plasma and 6.5 in buffer. The numerical ratio is
That arithmetic is correct, but interpreting 0.400 as the fraction unbound at a common pH is not justified: the chambers have a substantial pH difference, and the distribution of ionizable species can differ between them. In a separate hypothetical repeat with pH controlled at 7.4 in both chambers, suppose \(C_p=50.0\ \mu\mathrm{M}\) and \(C_b=5.0\ \mu\mathrm{M}\). The matched-pH ratio is \(5.0/50.0=0.100\). The repeat supports a different estimate under the controlled condition; it does not establish a universal correction factor of four for the first assay.
Worked Example: Checking for Inadequate Equilibration
In a hypothetical time-course experiment, chamber volumes are equal and unchanged, pH is stable and matched, and recovery is between 95% and 97% at each time. The measured plasma and buffer concentrations are:
| Time | \(C_p\) (µM) | \(C_b\) (µM) | \(C_b/C_p\) |
|---|---|---|---|
| 2 h | 90.0 | 4.0 | 0.0444 |
| 4 h | 82.0 | 7.5 | 0.0915 |
| 8 h | 80.0 | 8.0 | 0.100 |
| 12 h | 79.9 | 8.0 | 0.100 |
At 2 h, the ratio is \(4.0/90.0=0.0444\); at 4 h, it is \(7.5/82.0=0.0915\); and at 8 h, it is \(8.0/80.0=0.100\). The 12 h ratio is \(8.0/79.9=0.1001\), which rounds to 0.100. The last two measurements are effectively stable at the displayed precision, whereas earlier measurements differ markedly. Under the stated stable pH and recovery conditions, this pattern supports inadequate equilibration at the shorter times and a practical plateau by 8 h. The result should still be judged against the method’s analytical variability and prespecified criteria.
Controls, Alternatives, and Limitations
A practical diagnostic sequence is to verify the initial amount and final volumes, calculate recovery from the measured compartments, check endpoint pH in both chambers, and examine a time course. If recovery is low, use appropriate controls to investigate adsorption, degradation, precipitation, and analytical recovery before interpreting the concentration ratio. These checks answer different questions: mass balance assesses accounted-for drug, pH checks whether the comparison is on compatible conditions, and a time course assesses whether concentrations have stabilized.
Equilibrium dialysis is not the only approach. Ultrafiltration can separate a filtrate enriched in unbound drug from a protein-containing retentate and may be faster, but filter adsorption, concentration polarization, pressure, and the separation procedure itself can affect recovery or the measured free concentration. Dialysis avoids pressure-driven filtration but may take longer and remains vulnerable to membrane or apparatus adsorption and pH or time-course problems. Neither method is automatically correct; the choice depends on the compound and the assay’s validated performance. Results from different methods should not be treated as interchangeable without considering their conditions and controls.
There is no single recovery percentage that guarantees validity for every compound, concentration, matrix, and analytical method. Prespecified acceptance criteria should be based on method validation and intended use. Even apparently acceptable total recovery does not prove that equilibrium was reached, that pH was suitable, or that the assay measured the correct concentrations. Conversely, low recovery is a reason to investigate and qualify or repeat the result, not a direct estimate of \(f_{u,p}\). This tutorial concerns diagnosing those assay artifacts; interpretation of binding that varies with drug concentration requires a separate analysis.
Common Pitfall
- Wrong reasoning: “The assay recovered only 70%, so the drug is 30% bound,” or “I can divide the measured fraction unbound by 0.70 to correct it.”
- Correction: Recovery describes the fraction of initial drug accounted for in specified compartments. Protein binding is a distribution between free and bound drug, not the unexplained remainder in a mass balance. A correction is not justified without a validated model for the loss and evidence that the loss has not distorted the assay ratio.
- Wrong reasoning: “The buffer concentration was measured, so equilibrium must have been reached.”
- Correction: A measurable concentration is not evidence of a plateau. Use time-course measurements and confirm that pH and recovery are also acceptable.
Check Your Understanding
Answer each question, then compare your reasoning with the solutions.
- A \(0.100\ \mathrm{mL}\) plasma sample initially contains \(40.0\ \mu\mathrm{M}\) drug. At the endpoint, plasma is \(30.0\ \mu\mathrm{M}\) and buffer is \(5.0\ \mu\mathrm{M}\); both chambers are \(0.100\ \mathrm{mL}\). What is liquid-phase recovery?
- Why does drug found in a membrane rinse support adsorption but not automatically validate a corrected fraction unbound?
- What is the main interpretive concern when endpoint pH differs between chambers for an ionizable drug?
- In a time course, the buffer concentration continues rising and the plasma concentration continues falling. What should be concluded about the endpoint ratio?
Solutions. 1. The initial amount is \(40.0\times0.100=4.00\ \mathrm{nmol}\). The measured liquid amount is \((30.0+5.0)\times0.100=3.50\ \mathrm{nmol}\), so recovery is \(3.50/4.00=87.5\%\). 2. The rinse identifies drug associated with the membrane, but adsorption may have altered conditions during the assay; a validated control or repeat is needed to establish whether the ratio is reliable. 3. A pH difference can change the distribution of ionized species, so the measured concentrations may not represent equal free concentrations under a common pH condition. 4. The measurements do not yet show a plateau, so the endpoint ratio should not be treated as an established equilibrium estimate; continue the time course under controlled conditions.