Identifiability of Filtration, Secretion, and Reabsorption

Why a Measured Renal Clearance May Not Reveal Its Pathways

A renal clearance summarizes the net result of several processes. Filtration and secretion move drug toward the tubular lumen, while reabsorption moves drug back toward blood. If only plasma exposure and total urinary recovery are measured, different combinations of these processes can produce the same net urinary amount. The challenge is not merely to estimate a total; it is to determine whether the measurements distinguish the underlying pathways.

The model below is a deliberately simplified, operational description. It helps expose the identifiability problem; it is not a claim that all transport along the nephron is instantaneous or linear. Let \(C\) denote unbound plasma concentration in mg/L, and let \(R_{\mathrm{urine}}\) denote the rate of unchanged drug excretion in mg/h. The model assumes constant glomerular filtration rate (GFR), a single saturable secretory pathway, and linear reabsorption over the concentration range considered.

Here \(V_{\max}\) is the maximum secretory rate in mg/h, \(K_m\) is the unbound concentration in mg/L giving half of \(V_{\max}\), and \(CL_{\mathrm{reab}}\) is an apparent reabsorptive clearance in L/h. The linear reabsorption term is a convenient approximation: it represents reabsorptive flux as proportional to plasma concentration, rather than explicitly modeling tubular-lumen concentration, urine flow, or transit. The net renal clearance referenced to unbound concentration is \(CL_{\mathrm{R}}=R_{\mathrm{urine}}/C\), for \(C>0\).

This static relationship assumes that the measured concentration is representative of the condition being summarized, that urinary recovery is complete, and that unchanged drug in urine is correctly quantified. It also assumes that GFR, unbound fraction, and reabsorptive behavior do not change across the comparison. If plasma concentration varies over time, cumulative urinary recovery and plasma exposure instead provide an exposure-averaged quantity; they do not automatically identify the instantaneous terms in this equation.

One Net Observation, Many Pathway Combinations

At one concentration, a measured clearance gives one equation involving secretion and reabsorption, in addition to filtration. Even if GFR is known, that single equation generally cannot determine \(V_{\max}\), \(K_m\), and \(CL_{\mathrm{reab}}\) separately. The ambiguity is structural, not a defect that can be removed simply by collecting the same kind of measurement more precisely.

The same issue arises when the observation is cumulative urinary recovery \(A_{e}\) and plasma area under the concentration-time curve (AUC). Their ratio, \(A_{e}/\mathrm{AUC}\), can summarize net renal clearance under appropriate conditions, but it remains a net quantity. A single total does not say how much drug was filtered, secreted, or reabsorbed along the way.

How Additional Concentrations Can Help

Dose levels can be informative when they produce meaningfully different unbound concentrations and the model’s other terms are constrained. Saturable secretion contributes a concentration-dependent clearance, \(V_{\max}/(K_m+C)\); the filtration and linear reabsorption terms are constant clearances in this simplified model. A concentration series can therefore reveal curvature that a single concentration cannot show.

A dose series is useful only if it spans concentrations that distinguish the candidate mechanisms. When all concentrations are much smaller than \(K_m\), secretion is approximately linear: \(V_{\max}/(K_m+C)\approx V_{\max}/K_m\). The data may then identify a ratio of parameters rather than \(V_{\max}\) and \(K_m\) separately. When all concentrations are much larger than \(K_m\), the secretory rate approaches \(V_{\max}\), and the data may provide little information about \(K_m\). A broad, well-measured range is more informative, provided dose-dependent changes in other processes are not mistaken for saturation.

Time Courses and Perturbations Add Different Information

Timed urine collections preserve information lost when all urine is pooled into one total. They can show whether urinary excretion is delayed relative to plasma exposure or whether excretion changes across intervals. That timing information can test a dynamic renal model, but it is not automatically a unique signature of secretion or reabsorption. Urine flow, tubular transit, collection timing, and changing plasma concentrations can also alter the observed profile.

A selective inhibitor can provide a different kind of information by perturbing one pathway. The interpretation requires evidence that the inhibitor reaches the relevant site and affects the intended transporter, while leaving GFR, unbound exposure, and other renal pathways sufficiently unchanged. Comparing conditions at matched concentration, or explicitly accounting for concentration differences, is important because saturation itself changes clearance with concentration.

Checking the Model and Choosing Measurements

A fitted pathway model should be checked against more than its ability to reproduce a single clearance. Useful diagnostics include whether observed clearance changes systematically with concentration, whether predicted urinary amounts agree with complete collections, and whether predictions match the timing of interval samples. Large, structured residuals or implausible parameter estimates can signal model misspecification, poor measurements, or an uninformative design. A negative apparent reabsorptive clearance or a negative net excretion prediction, for example, should prompt scrutiny of assumptions and data rather than be accepted as a biological conclusion.

The main alternatives answer different questions. A clearance-ratio comparison summarizes net renal handling relative to a reference such as filtration, but does not by itself partition secretion and reabsorption. A dose-ranging study can reveal concentration dependence, but attribution depends on controlling or modeling other dose-related changes. Timed urine data add temporal constraints, but require a dynamic account of transit and urine flow. An inhibitor study adds a pathway perturbation, but depends on selectivity and matched exposure. Combining these designs can constrain a model more strongly than relying on any one summary.

The practical limitation is that every added parameter requires information. More samples of the same kind may improve precision, but cannot resolve a structural ambiguity if the model predicts the same observations for different parameter values. Conversely, more complex dynamic models can create additional parameters that are difficult to estimate. Model complexity should therefore be matched to the measurements and the scientific question.

A common mistake is to subtract a filtration estimate from total renal clearance and call the remainder “secretion.” That reasoning ignores reabsorption: in the model, \(CL_{\mathrm{R}}-\mathrm{GFR}=V_{\max}/(K_m+C)-CL_{\mathrm{reab}}\), a net of secretion and reabsorption, not secretion alone. The correction is to report the net difference as such unless reabsorption is negligible or independently constrained by suitable data and assumptions.