Intermittent Hemodialysis and Supplemental Doses

Why Hemodialysis Changes a Dosing Schedule

For a drug substantially removed by intermittent hemodialysis, the concentration at the end of a session may be lower than expected from the patient's usual elimination alone. The size of that difference depends not only on how readily the drug crosses the dialyzer, but also on treatment duration, dialysis clearance, distribution volume, and the concentration entering the session. A dosing plan therefore needs to consider both the ordinary maintenance regimen and any drug removed during dialysis.

The calculation is a model-based estimate, not a standalone prescription. It is most useful when the drug's pharmacokinetic behavior is reasonably described by a one-compartment model over the interval of interest, and when dialysis clearance and treatment timing are known or estimated. The earlier material on renal function covariates and nonlinear renal elimination concerns changes in the patient's intrinsic elimination; here, dialysis is treated as an additional, time-limited elimination pathway.

A Simple Model for a Dialysis Session

Definition: Let \(C(t)\) be the drug concentration in a well-mixed apparent distribution volume \(V\). Let \(CL_0\) denote non-dialysis systemic clearance during the interval, \(CL_D\) the effective drug clearance attributable to the dialyzer, and \(T\) the session duration. If these clearances and \(V\) are constant during dialysis, elimination is first order, and there is no drug input during the session, then the elimination rate constants are \(k_0=CL_0/V\) without dialysis and \(k_{\mathrm{on}}=(CL_0+CL_D)/V\) during dialysis. Clearances are in volume per time, \(V\) is in volume, and rate constants are in inverse time.

The model assumes a single, well-mixed compartment, constant clearances and volume, and no important concentration rebound during the calculation interval. It also assumes the stated \(CL_D\) represents effective drug removal under the relevant dialysis settings. Real dialysis clearance can depend on blood and dialysate flow, membrane, treatment interruptions, protein binding, and the patient's access and hemodynamics. A measured dialyzer extraction or a validated drug-specific estimate is preferable to assuming that every drug is removed at the same rate.

With \(C_{\mathrm{start}}\) as the concentration at the beginning of dialysis, the predicted end-session concentration is obtained by integrating the first-order model. A no-dialysis prediction at the same time provides a useful counterfactual for quantifying the additional concentration deficit associated with the session.

\frac{dC(t)}{dt}=-\frac{CL}{V}C(t) \qquad\Longrightarrow\qquad C_{\mathrm{HD}}(T)=C_{\mathrm{start}} \exp\left[-\frac{CL_0+CL_D}{V}T\right], \quad C_{\mathrm{no\,HD}}(T)=C_{\mathrm{start}} \exp\left[-\frac{CL_0}{V}T\right].

The fraction of the starting concentration remaining at the end of dialysis is \(C_{\mathrm{HD}}(T)/C_{\mathrm{start}}\). The concentration deficit relative to the no-dialysis prediction is \(C_{\mathrm{no\,HD}}(T)-C_{\mathrm{HD}}(T)\). These quantities are not the same as the mass physically removed by the dialyzer. Under this model, the dialyzer-attributable mass removed is the time integral of \(CL_D C(t)\):

A_D=\int_0^T CL_D C(t)\,dt =V C_{\mathrm{start}}\frac{CL_D}{CL_0+CL_D} \left(1-\exp\left[-\frac{CL_0+CL_D}{V}T\right]\right).

This expression follows by substituting the predicted concentration during the session into the integral and integrating the exponential. It attributes removal through the dialyzer only; ordinary systemic elimination continues at the same time. Because systemic elimination occurs from a concentration that is itself falling during dialysis, the dialyzer mass removed is not generally equal to \(V\) times the concentration deficit against the no-dialysis trajectory.

Worked Example: Estimating Dialyzer Removal and the Concentration Deficit

Suppose a hypothetical patient has \(C_{\mathrm{start}}=8.00\) mg/L, \(V=30.0\) L, \(CL_0=2.00\) L/h, and \(CL_D=6.00\) L/h during a 4.00-hour session. Assume the one-compartment, constant-clearance model applies.

The elimination rates are \(k_0=2.00/30.0=0.06667\ \mathrm{h}^{-1}\) and \(k_{\mathrm{on}}=(2.00+6.00)/30.0=0.2667\ \mathrm{h}^{-1}\). Thus:

C_{\mathrm{HD}}(4) =8.00\exp[-(8.00/30.0)(4.00)] =8.00e^{-1.0667} \approx 2.753\ \mathrm{mg/L}.

For the same four hours without dialysis, the predicted concentration is:

C_{\mathrm{no\,HD}}(4) =8.00\exp[-(2.00/30.0)(4.00)] =8.00e^{-0.2667} \approx 6.127\ \mathrm{mg/L}.

The end-session concentration deficit relative to that prediction is \(6.127-2.753=3.374\) mg/L. The model-based dialyzer mass removal is instead:

A_D =(30.0\ \mathrm{L})(8.00\ \mathrm{mg/L}) \frac{6.00}{8.00}(1-e^{-1.0667}) \approx 118.1\ \mathrm{mg}.

The distinction matters: multiplying the 3.374 mg/L deficit by 30.0 L gives about 101.2 mg, not 118.1 mg. The former is a concentration difference against a counterfactual; the latter is the integrated amount cleared through the dialyzer under the model.

Choosing a Postdialysis Supplement

A supplement can be estimated by asking how much drug would raise the modeled postdialysis concentration to a specified target. For a rapidly available dose with bioavailability \(F\), and assuming it distributes into the same apparent volume \(V\), an administered amount \(D_{\mathrm{sup}}\) produces an immediate modeled concentration increment \(F D_{\mathrm{sup}}/V\). If the target is the concentration predicted at the same time without dialysis, the supplement is:

D_{\mathrm{sup}} =\frac{V}{F} \left[C_{\mathrm{target}}-C_{\mathrm{HD}}(T)\right], \qquad C_{\mathrm{target}}=C_{\mathrm{no\,HD}}(T).

This calculation assumes the supplement is given after dialysis, its absorption and distribution can be represented by \(F\) and \(V\), and the target is appropriate for the drug and patient. If the routine dose is due at that time, the schedule must state whether that dose is being replaced, supplemented, or administered separately. A supplement should not be added blindly on top of a routine dose that already provides the intended replacement.

Worked Example: Calculating a Supplement From Two Model Predictions

In a hypothetical concentration-based calculation, a pharmacokinetic model predicts a no-dialysis concentration of \(4e^{-0.03}\) mg/L at the comparison time and an end-session concentration of \(4e^{-0.01}e^{-0.4508}\) mg/L under the specified dialysis scenario. These are treated as predictions at the same comparison time. Suppose \(V=40.0\) L and the supplement has \(F=1\).

First evaluate the predictions without rounding the intermediate values:

C_{\mathrm{no\,HD}}=4e^{-0.03}=3.8817821\ \mathrm{mg/L}, \qquad C_{\mathrm{HD}}=4e^{-0.01}e^{-0.4508} =4e^{-0.4608} =2.5231153\ \mathrm{mg/L}.

The modeled deficit is \(3.8817821-2.5231153=1.3586668\) mg/L, or approximately \(1.359\) mg/L. The supplement estimate is:

D_{\mathrm{sup}} =\frac{40.0\ \mathrm{L}}{1}(1.3586668\ \mathrm{mg/L}) =54.3467\ \mathrm{mg} \approx 54.35\ \mathrm{mg}.

The units reduce to mg because L multiplied by mg/L is mg. This result is a model-based amount for restoring the specified concentration, not a claim that this dose is clinically suitable or available as a practical dosage form.

Timing Routine Doses and Supplements

The timing of the regular dose relative to dialysis can change both the concentration entering treatment and the need for a supplement. A dose administered before dialysis may itself be partly removed; a dose administered after dialysis is not exposed to that session. The relevant calculation is therefore based on the concentration at session start and a clearly specified comparison target at the same post-session time. When a scheduled dose falls near dialysis, explicitly model or document that dose rather than treating all drug loss as an additional amount to replace.

Worked Example: Timing a Supplement After a Session

Assume a hypothetical drug follows the same one-compartment model. At dialysis start, \(C_{\mathrm{start}}=10.0\) mg/L; \(V=25.0\) L; \(CL_0/V=0.0400\ \mathrm{h}^{-1}\); \(CL_D/V=0.250\ \mathrm{h}^{-1}\); and treatment lasts 3.00 h. The proposed supplement has \(F=0.800\). Suppose the desired post-session target is the concentration that would have remained at that time without dialysis.

The no-dialysis concentration is:

C_{\mathrm{target}} =10.0e^{-(0.0400)(3.00)} =10.0e^{-0.120} =8.869\ \mathrm{mg/L}.

During dialysis, the total rate constant is \(0.0400+0.250=0.290\ \mathrm{h}^{-1}\), so:

C_{\mathrm{HD}}(3) =10.0e^{-(0.290)(3.00)} =10.0e^{-0.870} =4.190\ \mathrm{mg/L}.

The deficit is \(8.869-4.190=4.679\) mg/L. Therefore, the estimated supplement is:

D_{\mathrm{sup}} =\frac{25.0\ \mathrm{L}}{0.800}(4.679\ \mathrm{mg/L}) \approx 146.2\ \mathrm{mg}.

Under these assumptions, the proposed postdialysis amount would restore the no-dialysis target in the model. If a routine dose is also due, its contribution must be included in the dosing plan; otherwise the total administered amount could exceed the amount intended to replace the modeled deficit.

Assumptions, Diagnostics, and Alternatives

A simple exponential estimate is most defensible when the drug behaves approximately as a single compartment over the session, clearance is reasonably stable, and the treatment duration and dialysis settings are known. Check that the concentration used is tied to the correct time, that units are consistent, and that a target concentration is clinically justified. For a reported fraction removed \(f\), check its definition: a fraction of drug removed across the session is not interchangeable with a dialyzer extraction ratio measured at one instant.

Several patterns indicate that the model may be inadequate. A measured concentration that falls less than predicted may reflect lower effective dialyzer clearance, shorter or interrupted treatment, or redistribution from tissues into plasma. A concentration that rises after dialysis without a dose can indicate postdialysis rebound as drug redistributes from peripheral compartments. In either case, an immediate end-session concentration may not represent the later concentration relevant to dosing. Repeated levels at defined times, drug-specific dialysis data, or a multicompartment model may be needed.

Approach	What it estimates	Main limitation
Clearance-based model	Concentration trajectory from \(CL_0\), \(CL_D\), \(V\), and session duration	Depends on accurate parameters and model assumptions
Measured pre/post concentrations	Observed concentration change during a particular session	Sampling time, rebound, and non-dialysis elimination affect interpretation
Published or protocol-based supplement	A practical dose recommendation for a specified drug and dialysis context	May not fit the patient's treatment settings, target, or dose timing

Measured pre- and postdialysis concentrations can help assess a model, but a single pair of levels does not by itself isolate dialyzer removal: ordinary elimination also occurs, and the post-session sample may precede redistribution. A published dosing protocol may be more practical when supported for the drug and treatment, while a clearance-based model makes assumptions and timing explicit. Neither removes the need to individualize the target and monitor response.

Common Pitfalls and Limitations

Equating a concentration deficit with mass removed by the dialyzer. The deficit compares two concentration trajectories; dialyzer mass removal is an integral over the session. They need not match.
Using a fraction removed without its definition. A percentage may refer to a measured concentration change, a mass balance, or a model prediction. Confirm the time points and reference before applying it.
Ignoring dose timing. A pre-session dose may be dialyzed, whereas a post-session dose is not exposed to that session. State whether the routine dose is included in the replacement plan.
Assuming a postdialysis level is stable. Redistribution can produce rebound, so the immediate concentration may not equal the concentration at a later clinically relevant time.
Interpreting the calculated amount as a universal recommendation. The estimated supplement depends on the selected target, bioavailability, distribution volume, and clearance assumptions; clinical dosing and monitoring remain necessary.

Key takeaway: Predict the concentration during dialysis from the total elimination rate, compare it with a no-dialysis prediction at the same time if a replacement target is intended, and calculate a supplement only after specifying the target, bioavailability, and treatment-dose timing.

Check Your Understanding

Use the assumptions stated in each question. Solutions are provided below the questions.

A hypothetical drug has \(CL_0=1.5\) L/h, \(CL_D=4.5\) L/h, \(V=30\) L, and a 2-hour session. What fraction of its starting concentration remains at the end of dialysis under the one-compartment model?
Why is \(V[C_{\mathrm{no\,HD}}(T)-C_{\mathrm{HD}}(T)]\) not generally the same as the amount removed through the dialyzer?
A model gives a postdialysis deficit of 2.0 mg/L. If \(V=20\) L and \(F=0.80\), what supplement restores that modeled deficit?
What can a rise in concentration after dialysis, before another dose, suggest, and why could it make an immediate post-session supplement calculation unreliable?

Solutions. (1) The total rate constant is \((1.5+4.5)/30=0.200\ \mathrm{h}^{-1}\). The fraction remaining is \(e^{-(0.200)(2)}=e^{-0.400}=0.6703\), or about 67.0%. (2) The expression using \(V\) is a counterfactual concentration gap converted to an amount-equivalent; it is not the integral of dialyzer clearance times the changing concentration. The dialyzer-attributable amount is \(\int_0^T CL_D C(t)\,dt\). (3) \(D_{\mathrm{sup}}=(20/0.80)(2.0)=50\) mg. (4) It may indicate postdialysis redistribution, or rebound, from tissues to plasma. The immediate level may not represent the later concentration, so the target time and rebound need to be considered rather than treating the immediate level as a stable post-session value.